diff options
| author | Maxime Ripard <maxime.ripard@bootlin.com> | 2019-02-11 10:35:35 +0100 |
|---|---|---|
| committer | Maxime Ripard <maxime.ripard@bootlin.com> | 2019-02-11 10:35:35 +0100 |
| commit | d588100baa28dae6a5c32d02bfe744d0792ed2ad (patch) | |
| tree | 3be3984cea35a07b4259f8ebe55ce0e52e416d1c | |
| parent | 7bd0a3271e239d34bae5fa51e4f44ed9ee861a59 (diff) | |
| parent | 16065fcdd19ddb9e093192914ac863884f308766 (diff) | |
Merge drm/drm-next into drm-misc-next
We need to backmerge drm-next to fix the komeda build failure.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
509 files changed, 23792 insertions, 27656 deletions
diff --git a/Documentation/devicetree/bindings/display/arm,komeda.txt b/Documentation/devicetree/bindings/display/arm,komeda.txt new file mode 100644 index 000000000000..02b226532ebd --- /dev/null +++ b/Documentation/devicetree/bindings/display/arm,komeda.txt @@ -0,0 +1,73 @@ +Device Tree bindings for Arm Komeda display driver + +Required properties: +- compatible: Should be "arm,mali-d71" +- reg: Physical base address and length of the registers in the system +- interrupts: the interrupt line number of the device in the system +- clocks: A list of phandle + clock-specifier pairs, one for each entry + in 'clock-names' +- clock-names: A list of clock names. It should contain: + - "mclk": for the main processor clock + - "pclk": for the APB interface clock +- #address-cells: Must be 1 +- #size-cells: Must be 0 + +Required properties for sub-node: pipeline@nq +Each device contains one or two pipeline sub-nodes (at least one), each +pipeline node should provide properties: +- reg: Zero-indexed identifier for the pipeline +- clocks: A list of phandle + clock-specifier pairs, one for each entry + in 'clock-names' +- clock-names: should contain: + - "pxclk": pixel clock + - "aclk": AXI interface clock + +- port: each pipeline connect to an encoder input port. The connection is + modeled using the OF graph bindings specified in + Documentation/devicetree/bindings/graph.txt + +Optional properties: + - memory-region: phandle to a node describing memory (see + Documentation/devicetree/bindings/reserved-memory/reserved-memory.txt) + to be used for the framebuffer; if not present, the framebuffer may + be located anywhere in memory. + +Example: +/ { + ... + + dp0: display@c00000 { + #address-cells = <1>; + #size-cells = <0>; + compatible = "arm,mali-d71"; + reg = <0xc00000 0x20000>; + interrupts = <0 168 4>; + clocks = <&dpu_mclk>, <&dpu_aclk>; + clock-names = "mclk", "pclk"; + + dp0_pipe0: pipeline@0 { + clocks = <&fpgaosc2>, <&dpu_aclk>; + clock-names = "pxclk", "aclk"; + reg = <0>; + + port { + dp0_pipe0_out: endpoint { + remote-endpoint = <&db_dvi0_in>; + }; + }; + }; + + dp0_pipe1: pipeline@1 { + clocks = <&fpgaosc2>, <&dpu_aclk>; + clock-names = "pxclk", "aclk"; + reg = <1>; + + port { + dp0_pipe1_out: endpoint { + remote-endpoint = <&db_dvi1_in>; + }; + }; + }; + }; + ... +}; diff --git a/Documentation/devicetree/bindings/display/bridge/renesas,lvds.txt b/Documentation/devicetree/bindings/display/bridge/renesas,lvds.txt index 27a054e1bb5f..900a884ad9f5 100644 --- a/Documentation/devicetree/bindings/display/bridge/renesas,lvds.txt +++ b/Documentation/devicetree/bindings/display/bridge/renesas,lvds.txt @@ -8,6 +8,7 @@ Required properties: - compatible : Shall contain one of - "renesas,r8a7743-lvds" for R8A7743 (RZ/G1M) compatible LVDS encoders + - "renesas,r8a7744-lvds" for R8A7744 (RZ/G1N) compatible LVDS encoders - "renesas,r8a774c0-lvds" for R8A774C0 (RZ/G2E) compatible LVDS encoders - "renesas,r8a7790-lvds" for R8A7790 (R-Car H2) compatible LVDS encoders - "renesas,r8a7791-lvds" for R8A7791 (R-Car M2-W) compatible LVDS encoders diff --git a/Documentation/devicetree/bindings/display/tegra/nvidia,tegra20-host1x.txt b/Documentation/devicetree/bindings/display/tegra/nvidia,tegra20-host1x.txt index 593be44a53c9..9999255ac5b6 100644 --- a/Documentation/devicetree/bindings/display/tegra/nvidia,tegra20-host1x.txt +++ b/Documentation/devicetree/bindings/display/tegra/nvidia,tegra20-host1x.txt @@ -238,6 +238,9 @@ of the following host1x client modules: - nvidia,hpd-gpio: specifies a GPIO used for hotplug detection - nvidia,edid: supplies a binary EDID blob - nvidia,panel: phandle of a display panel + - nvidia,xbar-cfg: 5 cells containing the crossbar configuration. Each lane + of the SOR, identified by the cell's index, is mapped via the crossbar to + the pad specified by the cell's value. Optional properties when driving an eDP output: - nvidia,dpaux: phandle to a DispayPort AUX interface diff --git a/Documentation/devicetree/bindings/gpu/samsung-rotator.txt b/Documentation/devicetree/bindings/gpu/samsung-rotator.txt index 82cd1ed0be93..3aca2578da0b 100644 --- a/Documentation/devicetree/bindings/gpu/samsung-rotator.txt +++ b/Documentation/devicetree/bindings/gpu/samsung-rotator.txt @@ -2,9 +2,10 @@ Required properties: - compatible : value should be one of the following: - (a) "samsung,exynos4210-rotator" for Rotator IP in Exynos4210 - (b) "samsung,exynos4212-rotator" for Rotator IP in Exynos4212/4412 - (c) "samsung,exynos5250-rotator" for Rotator IP in Exynos5250 + * "samsung,s5pv210-rotator" for Rotator IP in S5PV210 + * "samsung,exynos4210-rotator" for Rotator IP in Exynos4210 + * "samsung,exynos4212-rotator" for Rotator IP in Exynos4212/4412 + * "samsung,exynos5250-rotator" for Rotator IP in Exynos5250 - reg : Physical base address of the IP registers and length of memory mapped region. diff --git a/Documentation/gpu/afbc.rst b/Documentation/gpu/afbc.rst new file mode 100644 index 000000000000..4d38dc49d105 --- /dev/null +++ b/Documentation/gpu/afbc.rst @@ -0,0 +1,235 @@ +.. SPDX-License-Identifier: GPL-2.0+ + +=================================== + Arm Framebuffer Compression (AFBC) +=================================== + +AFBC is a proprietary lossless image compression protocol and format. +It provides fine-grained random access and minimizes the amount of +data transferred between IP blocks. + +AFBC can be enabled on drivers which support it via use of the AFBC +format modifiers defined in drm_fourcc.h. See DRM_FORMAT_MOD_ARM_AFBC(*). + +All users of the AFBC modifiers must follow the usage guidelines laid +out in this document, to ensure compatibility across different AFBC +producers and consumers. + +Components and Ordering +======================= + +AFBC streams can contain several components - where a component +corresponds to a color channel (i.e. R, G, B, X, A, Y, Cb, Cr). +The assignment of input/output color channels must be consistent +between the encoder and the decoder for correct operation, otherwise +the consumer will interpret the decoded data incorrectly. + +Furthermore, when the lossless colorspace transform is used +(AFBC_FORMAT_MOD_YTR, which should be enabled for RGB buffers for +maximum compression efficiency), the component order must be: + + * Component 0: R + * Component 1: G + * Component 2: B + +The component ordering is communicated via the fourcc code in the +fourcc:modifier pair. In general, component '0' is considered to +reside in the least-significant bits of the corresponding linear +format. For example, COMP(bits): + + * DRM_FORMAT_ABGR8888 + + * Component 0: R(8) + * Component 1: G(8) + * Component 2: B(8) + * Component 3: A(8) + + * DRM_FORMAT_BGR888 + + * Component 0: R(8) + * Component 1: G(8) + * Component 2: B(8) + + * DRM_FORMAT_YUYV + + * Component 0: Y(8) + * Component 1: Cb(8, 2x1 subsampled) + * Component 2: Cr(8, 2x1 subsampled) + +In AFBC, 'X' components are not treated any differently from any other +component. Therefore, an AFBC buffer with fourcc DRM_FORMAT_XBGR8888 +encodes with 4 components, like so: + + * DRM_FORMAT_XBGR8888 + + * Component 0: R(8) + * Component 1: G(8) + * Component 2: B(8) + * Component 3: X(8) + +Please note, however, that the inclusion of a "wasted" 'X' channel is +bad for compression efficiency, and so it's recommended to avoid +formats containing 'X' bits. If a fourth component is +required/expected by the encoder/decoder, then it is recommended to +instead use an equivalent format with alpha, setting all alpha bits to +'1'. If there is no requirement for a fourth component, then a format +which doesn't include alpha can be used, e.g. DRM_FORMAT_BGR888. + +Number of Planes +================ + +Formats which are typically multi-planar in linear layouts (e.g. YUV +420), can be encoded into one, or multiple, AFBC planes. As with +component order, the encoder and decoder must agree about the number +of planes in order to correctly decode the buffer. The fourcc code is +used to determine the number of encoded planes in an AFBC buffer, +matching the number of planes for the linear (unmodified) format. +Within each plane, the component ordering also follows the fourcc +code: + +For example: + + * DRM_FORMAT_YUYV: nplanes = 1 + + * Plane 0: + + * Component 0: Y(8) + * Component 1: Cb(8, 2x1 subsampled) + * Component 2: Cr(8, 2x1 subsampled) + + * DRM_FORMAT_NV12: nplanes = 2 + + * Plane 0: + + * Component 0: Y(8) + + * Plane 1: + + * Component 0: Cb(8, 2x1 subsampled) + * Component 1: Cr(8, 2x1 subsampled) + +Cross-device interoperability +============================= + +For maximum compatibility across devices, the table below defines +canonical formats for use between AFBC-enabled devices. Formats which +are listed here must be used exactly as specified when using the AFBC +modifiers. Formats which are not listed should be avoided. + +.. flat-table:: AFBC formats + + * - Fourcc code + - Description + - Planes/Components + + * - DRM_FORMAT_ABGR2101010 + - 10-bit per component RGB, with 2-bit alpha + - Plane 0: 4 components + * Component 0: R(10) + * Component 1: G(10) + * Component 2: B(10) + * Component 3: A(2) + + * - DRM_FORMAT_ABGR8888 + - 8-bit per component RGB, with 8-bit alpha + - Plane 0: 4 components + * Component 0: R(8) + * Component 1: G(8) + * Component 2: B(8) + * Component 3: A(8) + + * - DRM_FORMAT_BGR888 + - 8-bit per component RGB + - Plane 0: 3 components + * Component 0: R(8) + * Component 1: G(8) + * Component 2: B(8) + + * - DRM_FORMAT_BGR565 + - 5/6-bit per component RGB + - Plane 0: 3 components + * Component 0: R(5) + * Component 1: G(6) + * Component 2: B(5) + + * - DRM_FORMAT_ABGR1555 + - 5-bit per component RGB, with 1-bit alpha + - Plane 0: 4 components + * Component 0: R(5) + * Component 1: G(5) + * Component 2: B(5) + * Component 3: A(1) + + * - DRM_FORMAT_VUY888 + - 8-bit per component YCbCr 444, single plane + - Plane 0: 3 components + * Component 0: Y(8) + * Component 1: Cb(8) + * Component 2: Cr(8) + + * - DRM_FORMAT_VUY101010 + - 10-bit per component YCbCr 444, single plane + - Plane 0: 3 components + * Component 0: Y(10) + * Component 1: Cb(10) + * Component 2: Cr(10) + + * - DRM_FORMAT_YUYV + - 8-bit per component YCbCr 422, single plane + - Plane 0: 3 components + * Component 0: Y(8) + * Component 1: Cb(8, 2x1 subsampled) + * Component 2: Cr(8, 2x1 subsampled) + + * - DRM_FORMAT_NV16 + - 8-bit per component YCbCr 422, two plane + - Plane 0: 1 component + * Component 0: Y(8) + Plane 1: 2 components + * Component 0: Cb(8, 2x1 subsampled) + * Component 1: Cr(8, 2x1 subsampled) + + * - DRM_FORMAT_Y210 + - 10-bit per component YCbCr 422, single plane + - Plane 0: 3 components + * Component 0: Y(10) + * Component 1: Cb(10, 2x1 subsampled) + * Component 2: Cr(10, 2x1 subsampled) + + * - DRM_FORMAT_P210 + - 10-bit per component YCbCr 422, two plane + - Plane 0: 1 component + * Component 0: Y(10) + Plane 1: 2 components + * Component 0: Cb(10, 2x1 subsampled) + * Component 1: Cr(10, 2x1 subsampled) + + * - DRM_FORMAT_YUV420_8BIT + - 8-bit per component YCbCr 420, single plane + - Plane 0: 3 components + * Component 0: Y(8) + * Component 1: Cb(8, 2x2 subsampled) + * Component 2: Cr(8, 2x2 subsampled) + + * - DRM_FORMAT_YUV420_10BIT + - 10-bit per component YCbCr 420, single plane + - Plane 0: 3 components + * Component 0: Y(10) + * Component 1: Cb(10, 2x2 subsampled) + * Component 2: Cr(10, 2x2 subsampled) + + * - DRM_FORMAT_NV12 + - 8-bit per component YCbCr 420, two plane + - Plane 0: 1 component + * Component 0: Y(8) + Plane 1: 2 components + * Component 0: Cb(8, 2x2 subsampled) + * Component 1: Cr(8, 2x2 subsampled) + + * - DRM_FORMAT_P010 + - 10-bit per component YCbCr 420, two plane + - Plane 0: 1 component + * Component 0: Y(10) + Plane 1: 2 components + * Component 0: Cb(10, 2x2 subsampled) + * Component 1: Cr(10, 2x2 subsampled) diff --git a/Documentation/gpu/drivers.rst b/Documentation/gpu/drivers.rst index 7c1672118a73..044a7025477c 100644 --- a/Documentation/gpu/drivers.rst +++ b/Documentation/gpu/drivers.rst @@ -17,6 +17,8 @@ GPU Driver Documentation vkms bridge/dw-hdmi xen-front + afbc + komeda-kms .. only:: subproject and html diff --git a/Documentation/gpu/komeda-kms.rst b/Documentation/gpu/komeda-kms.rst new file mode 100644 index 000000000000..b08da1cffecc --- /dev/null +++ b/Documentation/gpu/komeda-kms.rst @@ -0,0 +1,488 @@ +.. SPDX-License-Identifier: GPL-2.0 + +============================== + drm/komeda Arm display driver +============================== + +The drm/komeda driver supports the Arm display processor D71 and later products, +this document gives a brief overview of driver design: how it works and why +design it like that. + +Overview of D71 like display IPs +================================ + +From D71, Arm display IP begins to adopt a flexible and modularized +architecture. A display pipeline is made up of multiple individual and +functional pipeline stages called components, and every component has some +specific capabilities that can give the flowed pipeline pixel data a +particular processing. + +Typical D71 components: + +Layer +----- +Layer is the first pipeline stage, which prepares the pixel data for the next +stage. It fetches the pixel from memory, decodes it if it's AFBC, rotates the +source image, unpacks or converts YUV pixels to the device internal RGB pixels, +then adjusts the color_space of pixels if needed. + +Scaler +------ +As its name suggests, scaler takes responsibility for scaling, and D71 also +supports image enhancements by scaler. +The usage of scaler is very flexible and can be connected to layer output +for layer scaling, or connected to compositor and scale the whole display +frame and then feed the output data into wb_layer which will then write it +into memory. + +Compositor (compiz) +------------------- +Compositor blends multiple layers or pixel data flows into one single display +frame. its output frame can be fed into post image processor for showing it on +the monitor or fed into wb_layer and written to memory at the same time. +user can also insert a scaler between compositor and wb_layer to down scale +the display frame first and and then write to memory. + +Writeback Layer (wb_layer) +-------------------------- +Writeback layer does the opposite things of Layer, which connects to compiz +and writes the composition result to memory. + +Post image processor (improc) +----------------------------- +Post image processor adjusts frame data like gamma and color space to fit the +requirements of the monitor. + +Timing controller (timing_ctrlr) +-------------------------------- +Final stage of display pipeline, Timing controller is not for the pixel +handling, but only for controlling the display timing. + +Merger +------ +D71 scaler mostly only has the half horizontal input/output capabilities +compared with Layer, like if Layer supports 4K input size, the scaler only can +support 2K input/output in the same time. To achieve the ful frame scaling, D71 +introduces Layer Split, which splits the whole image to two half parts and feeds +them to two Layers A and B, and does the scaling independently. After scaling +the result need to be fed to merger to merge two part images together, and then +output merged result to compiz. + +Splitter +-------- +Similar to Layer Split, but Splitter is used for writeback, which splits the +compiz result to two parts and then feed them to two scalers. + +Possible D71 Pipeline usage +=========================== + +Benefitting from the modularized architecture, D71 pipelines can be easily +adjusted to fit different usages. And D71 has two pipelines, which support two +types of working mode: + +- Dual display mode + Two pipelines work independently and separately to drive two display outputs. + +- Single display mode + Two pipelines work together to drive only one display output. + + On this mode, pipeline_B doesn't work indenpendently, but outputs its + composition result into pipeline_A, and its pixel timing also derived from + pipeline_A.timing_ctrlr. The pipeline_B works just like a "slave" of + pipeline_A(master) + +Single pipeline data flow +------------------------- + +.. kernel-render:: DOT + :alt: Single pipeline digraph + :caption: Single pipeline data flow + + digraph single_ppl { + rankdir=LR; + + subgraph { + "Memory"; + "Monitor"; + } + + subgraph cluster_pipeline { + style=dashed + node [shape=box] + { + node [bgcolor=grey style=dashed] + "Scaler-0"; + "Scaler-1"; + "Scaler-0/1" + } + + node [bgcolor=grey style=filled] + "Layer-0" -> "Scaler-0" + "Layer-1" -> "Scaler-0" + "Layer-2" -> "Scaler-1" + "Layer-3" -> "Scaler-1" + + "Layer-0" -> "Compiz" + "Layer-1" -> "Compiz" + "Layer-2" -> "Compiz" + "Layer-3" -> "Compiz" + "Scaler-0" -> "Compiz" + "Scaler-1" -> "Compiz" + + "Compiz" -> "Scaler-0/1" -> "Wb_layer" + "Compiz" -> "Improc" -> "Timing Controller" + } + + "Wb_layer" -> "Memory" + "Timing Controller" -> "Monitor" + } + +Dual pipeline with Slave enabled +-------------------------------- + +.. kernel-render:: DOT + :alt: Slave pipeline digraph + :caption: Slave pipeline enabled data flow + + digraph slave_ppl { + rankdir=LR; + + subgraph { + "Memory"; + "Monitor"; + } + node [shape=box] + subgraph cluster_pipeline_slave { + style=dashed + label="Slave Pipeline_B" + node [shape=box] + { + node [bgcolor=grey style=dashed] + "Slave.Scaler-0"; + "Slave.Scaler-1"; + } + + node [bgcolor=grey style=filled] + "Slave.Layer-0" -> "Slave.Scaler-0" + "Slave.Layer-1" -> "Slave.Scaler-0" + "Slave.Layer-2" -> "Slave.Scaler-1" + "Slave.Layer-3" -> "Slave.Scaler-1" + + "Slave.Layer-0" -> "Slave.Compiz" + "Slave.Layer-1" -> "Slave.Compiz" + "Slave.Layer-2" -> "Slave.Compiz" + "Slave.Layer-3" -> "Slave.Compiz" + "Slave.Scaler-0" -> "Slave.Compiz" + "Slave.Scaler-1" -> "Slave.Compiz" + } + + subgraph cluster_pipeline_master { + style=dashed + label="Master Pipeline_A" + node [shape=box] + { + node [bgcolor=grey style=dashed] + "Scaler-0"; + "Scaler-1"; + "Scaler-0/1" + } + + node [bgcolor=grey style=filled] + "Layer-0" -> "Scaler-0" + "Layer-1" -> "Scaler-0" + "Layer-2" -> "Scaler-1" + "Layer-3" -> "Scaler-1" + + "Slave.Compiz" -> "Compiz" + "Layer-0" -> "Compiz" + "Layer-1" -> "Compiz" + "Layer-2" -> "Compiz" + "Layer-3" -> "Compiz" + "Scaler-0" -> "Compiz" + "Scaler-1" -> "Compiz" + + "Compiz" -> "Scaler-0/1" -> "Wb_layer" + "Compiz" -> "Improc" -> "Timing Controller" + } + + "Wb_layer" -> "Memory" + "Timing Controller" -> "Monitor" + } + +Sub-pipelines for input and output +---------------------------------- + +A complete display pipeline can be easily divided into three sub-pipelines +according to the in/out usage. + +Layer(input) pipeline +~~~~~~~~~~~~~~~~~~~~~ + +.. kernel-render:: DOT + :alt: Layer data digraph + :caption: Layer (input) data flow + + digraph layer_data_flow { + rankdir=LR; + node [shape=box] + + { + node [bgcolor=grey style=dashed] + "Scaler-n"; + } + + "Layer-n" -> "Scaler-n" -> "Compiz" + } + +.. kernel-render:: DOT + :alt: Layer Split digraph + :caption: Layer Split pipeline + + digraph layer_data_flow { + rankdir=LR; + node [shape=box] + + "Layer-0/1" -> "Scaler-0" -> "Merger" + "Layer-2/3" -> "Scaler-1" -> "Merger" + "Merger" -> "Compiz" + } + +Writeback(output) pipeline +~~~~~~~~~~~~~~~~~~~~~~~~~~ +.. kernel-render:: DOT + :alt: writeback digraph + :caption: Writeback(output) data flow + + digraph writeback_data_flow { + rankdir=LR; + node [shape=box] + + { + node [bgcolor=grey style=dashed] + "Scaler-n"; + } + + "Compiz" -> "Scaler-n" -> "Wb_layer" + } + +.. kernel-render:: DOT + :alt: split writeback digraph + :caption: Writeback(output) Split data flow + + digraph writeback_data_flow { + rankdir=LR; + node [shape=box] + + "Compiz" -> "Splitter" + "Splitter" -> "Scaler-0" -> "Merger" + "Splitter" -> "Scaler-1" -> "Merger" + "Merger" -> "Wb_layer" + } + +Display output pipeline +~~~~~~~~~~~~~~~~~~~~~~~ +.. kernel-render:: DOT + :alt: display digraph + :caption: display output data flow + + digraph single_ppl { + rankdir=LR; + node [shape=box] + + "Compiz" -> "Improc" -> "Timing Controller" + } + +In the following section we'll see these three sub-pipelines will be handled +by KMS-plane/wb_conn/crtc respectively. + +Komeda Resource abstraction +=========================== + +struct komeda_pipeline/component +-------------------------------- + +To fully utilize and easily access/configure the HW, the driver side also uses +a similar architecture: Pipeline/Component to describe the HW features and +capabilities, and a specific component includes two parts: + +- Data flow controlling. +- Specific component capabilities and features. + +So the driver defines a common header struct komeda_component to describe the +data flow control and all specific components are a subclass of this base +structure. + +.. kernel-doc:: drivers/gpu/drm/arm/display/komeda/komeda_pipeline.h + :internal: + +Resource discovery and initialization +===================================== + +Pipeline and component are used to describe how to handle the pixel data. We +still need a @struct komeda_dev to describe the whole view of the device, and +the control-abilites of device. + +We have &komeda_dev, &komeda_pipeline, &komeda_component. Now fill devices with +pipelines. Since komeda is not for D71 only but also intended for later products, +of course we’d better share as much as possible between different products. To +achieve this, split the komeda device into two layers: CORE and CHIP. + +- CORE: for common features and capabilities handling. +- CHIP: for register programing and HW specific feature (limitation) handling. + +CORE can access CHIP by three chip function structures: + +- struct komeda_dev_funcs +- struct komeda_pipeline_funcs +- struct komeda_component_funcs + +.. kernel-doc:: drivers/gpu/drm/arm/display/komeda/komeda_dev.h + :internal: + +Format handling +=============== + +.. kernel-doc:: drivers/gpu/drm/arm/display/komeda/komeda_format_caps.h + :internal: +.. kernel-doc:: drivers/gpu/drm/arm/display/komeda/komeda_framebuffer.h + :internal: + +Attach komeda_dev to DRM-KMS +============================ + +Komeda abstracts resources by pipeline/component, but DRM-KMS uses +crtc/plane/connector. One KMS-obj cannot represent only one single component, +since the requirements of a single KMS object cannot simply be achieved by a +single component, usually that needs multiple components to fit the requirement. +Like set mode, gamma, ctm for KMS all target on CRTC-obj, but komeda needs +compiz, improc and timing_ctrlr to work together to fit these requirements. +And a KMS-Plane may require multiple komeda resources: layer/scaler/compiz. + +So, one KMS-Obj represents a sub-pipeline of komeda resources. + +- Plane: `Layer(input) pipeline`_ +- Wb_connector: `Writeback(output) pipeline`_ +- Crtc: `Display output pipeline`_ + +So, for komeda, we treat KMS crtc/plane/connector as users of pipeline and +component, and at any one time a pipeline/component only can be used by one +user. And pipeline/component will be treated as private object of DRM-KMS; the +state will be managed by drm_atomic_state as well. + +How to map plane to Layer(input) pipeline +----------------------------------------- + +Komeda has multiple Layer input pipelines, see: +- `Single pipeline data flow`_ +- `Dual pipeline with Slave enabled`_ + +The easiest way is binding a plane to a fixed Layer pipeline, but consider the +komeda capabilities: + +- Layer Split, See `Layer(input) pipeline`_ + + Layer_Split is quite complicated feature, which splits a big image into two + parts and handles it by two layers and two scalers individually. But it + imports an edge problem or effect in the middle of the image after the split. + To avoid such a problem, it needs a complicated Split calculation and some + special configurations to the layer and scaler. We'd better hide such HW + related complexity to user mode. + +- Slave pipeline, See `Dual pipeline with Slave enabled`_ + + Since the compiz component doesn't output alpha value, the slave pipeline + only can be used for bottom layers composition. The komeda driver wants to + hide this limitation to the user. The way to do this is to pick a suitable + Layer according to plane_state->zpos. + +So for komeda, the KMS-plane doesn't represent a fixed komeda layer pipeline, +but multiple Layers with same capabilities. Komeda will select one or more +Layers to fit the requirement of one KMS-plane. + +Make component/pipeline to be drm_private_obj +--------------------------------------------- + +Add :c:type:`drm_private_obj` to :c:type:`komeda_component`, :c:type:`komeda_pipeline` + +.. code-block:: c + + struct komeda_component { + struct drm_private_obj obj; + ... + } + + struct komeda_pipeline { + struct drm_private_obj obj; + ... + } + +Tracking component_state/pipeline_state by drm_atomic_state +----------------------------------------------------------- + +Add :c:type:`drm_private_state` and user to :c:type:`komeda_component_state`, +:c:type:`komeda_pipeline_state` + +.. code-block:: c + + struct komeda_component_state { + struct drm_private_state obj; + void *binding_user; + ... + } + + struct komeda_pipeline_state { + struct drm_private_state obj; + struct drm_crtc *crtc; + ... + } + +komeda component validation +--------------------------- + +Komeda has multiple types of components, but the process of validation are +similar, usually including the following steps: + +.. code-block:: c + + int komeda_xxxx_validate(struct komeda_component_xxx xxx_comp, + struct komeda_component_output *input_dflow, + struct drm_plane/crtc/connector *user, + struct drm_plane/crtc/connector_state, *user_state) + { + setup 1: check if component is needed, like the scaler is optional depending + on the user_state; if unneeded, just return, and the caller will + put the data flow into next stage. + Setup 2: check user_state with component features and capabilities to see + if requirements can be met; if not, return fail. + Setup 3: get component_state from drm_atomic_state, and try set to set + user to component; fail if component has been assigned to another + user already. + Setup 3: configure the component_state, like set its input component, + convert user_state to component specific state. + Setup 4: adjust the input_dflow and prepare it for the next stage. + } + +komeda_kms Abstraction +---------------------- + +.. kernel-doc:: drivers/gpu/drm/arm/display/komeda/komeda_kms.h + :internal: + +komde_kms Functions +------------------- +.. kernel-doc:: drivers/gpu/drm/arm/display/komeda/komeda_crtc.c + :internal: +.. kernel-doc:: drivers/gpu/drm/arm/display/komeda/komeda_plane.c + :internal: + +Build komeda to be a Linux module driver +======================================== + +Now we have two level devices: + +- komeda_dev: describes the real display hardware. +- komeda_kms_dev: attachs or connects komeda_dev to DRM-KMS. + +All komeda operations are supplied or operated by komeda_dev or komeda_kms_dev, +the module driver is only a simple wrapper to pass the Linux command +(probe/remove/pm) into komeda_dev or komeda_kms_dev. diff --git a/MAINTAINERS b/MAINTAINERS index df5668bc1c88..e7e81fadff65 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1133,13 +1133,26 @@ S: Supported F: drivers/gpu/drm/arm/hdlcd_* F: Documentation/devicetree/bindings/display/arm,hdlcd.txt +ARM KOMEDA DRM-KMS DRIVER +M: James (Qian) Wang <james.qian.wang@arm.com> +M: Liviu Dudau <liviu.dudau@arm.com> +L: Mali DP Maintainers <malidp@foss.arm.com> +S: Supported +T: git git://linux-arm.org/linux-ld.git for-upstream/mali-dp +F: drivers/gpu/drm/arm/display/include/ +F: drivers/gpu/drm/arm/display/komeda/ +F: Documentation/devicetree/bindings/display/arm/arm,komeda.txt +F: Documentation/gpu/komeda-kms.rst + ARM MALI-DP DRM DRIVER M: Liviu Dudau <liviu.dudau@arm.com> M: Brian Starkey <brian.starkey@arm.com> -M: Mali DP Maintainers <malidp@foss.arm.com> +L: Mali DP Maintainers <malidp@foss.arm.com> S: Supported +T: git git://linux-arm.org/linux-ld.git for-upstream/mali-dp F: drivers/gpu/drm/arm/ F: Documentation/devicetree/bindings/display/arm,malidp.txt +F: Documentation/gpu/afbc.rst ARM MFM AND FLOPPY DRIVERS M: Ian Molton <spyro@f2s.com> diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile index ce8d1d384319..1ac55c65eac0 100644 --- a/drivers/gpu/drm/Makefile +++ b/drivers/gpu/drm/Makefile @@ -51,7 +51,7 @@ obj-$(CONFIG_DRM_DEBUG_SELFTEST) += selftests/ obj-$(CONFIG_DRM) += drm.o obj-$(CONFIG_DRM_MIPI_DSI) += drm_mipi_dsi.o obj-$(CONFIG_DRM_PANEL_ORIENTATION_QUIRKS) += drm_panel_orientation_quirks.o -obj-$(CONFIG_DRM_ARM) += arm/ +obj-y += arm/ obj-$(CONFIG_DRM_TTM) += ttm/ obj-$(CONFIG_DRM_SCHED) += scheduler/ obj-$(CONFIG_DRM_TDFX) += tdfx/ @@ -81,7 +81,7 @@ obj-$(CONFIG_DRM_UDL) += udl/ obj-$(CONFIG_DRM_AST) += ast/ obj-$(CONFIG_DRM_ARMADA) += armada/ obj-$(CONFIG_DRM_ATMEL_HLCDC) += atmel-hlcdc/ -obj-$(CONFIG_DRM_RCAR_DU) += rcar-du/ +obj-y += rcar-du/ obj-$(CONFIG_DRM_SHMOBILE) +=shmobile/ obj-y += omapdrm/ obj-$(CONFIG_DRM_SUN4I) += sun4i/ diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile index f76bcb9c45e4..466da5954a68 100644 --- a/drivers/gpu/drm/amd/amdgpu/Makefile +++ b/drivers/gpu/drm/amd/amdgpu/Makefile @@ -57,7 +57,7 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.o \ # add asic specific block amdgpu-$(CONFIG_DRM_AMDGPU_CIK)+= cik.o cik_ih.o kv_smc.o kv_dpm.o \ - ci_smc.o ci_dpm.o dce_v8_0.o gfx_v7_0.o cik_sdma.o uvd_v4_2.o vce_v2_0.o + dce_v8_0.o gfx_v7_0.o cik_sdma.o uvd_v4_2.o vce_v2_0.o amdgpu-$(CONFIG_DRM_AMDGPU_SI)+= si.o gmc_v6_0.o gfx_v6_0.o si_ih.o si_dma.o dce_v6_0.o si_dpm.o si_smc.o diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index bcef6ea4bcf9..9efa681d0878 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -542,6 +542,11 @@ struct amdgpu_asic_funcs { bool (*need_full_reset)(struct amdgpu_device *adev); /* initialize doorbell layout for specific asic*/ void (*init_doorbell_index)(struct amdgpu_device *adev); + /* PCIe bandwidth usage */ + void (*get_pcie_usage)(struct amdgpu_device *adev, uint64_t *count0, + uint64_t *count1); + /* do we need to reset the asic at init time (e.g., kexec) */ + bool (*need_reset_on_init)(struct amdgpu_device *adev); }; /* @@ -634,7 +639,7 @@ struct amdgpu_nbio_funcs { void (*hdp_flush)(struct amdgpu_device *adev, struct amdgpu_ring *ring); u32 (*get_memsize)(struct amdgpu_device *adev); void (*sdma_doorbell_range)(struct amdgpu_device *adev, int instance, - bool use_doorbell, int doorbell_index); + bool use_doorbell, int doorbell_index, int doorbell_size); void (*enable_doorbell_aperture)(struct amdgpu_device *adev, bool enable); void (*enable_doorbell_selfring_aperture)(struct amdgpu_device *adev, @@ -1042,6 +1047,8 @@ int emu_soc_asic_init(struct amdgpu_device *adev); #define amdgpu_asic_invalidate_hdp(adev, r) (adev)->asic_funcs->invalidate_hdp((adev), (r)) #define amdgpu_asic_need_full_reset(adev) (adev)->asic_funcs->need_full_reset((adev)) #define amdgpu_asic_init_doorbell_index(adev) (adev)->asic_funcs->init_doorbell_index((adev)) +#define amdgpu_asic_get_pcie_usage(adev, cnt0, cnt1) ((adev)->asic_funcs->get_pcie_usage((adev), (cnt0), (cnt1))) +#define amdgpu_asic_need_reset_on_init(adev) (adev)->asic_funcs->need_reset_on_init((adev)) /* Common functions */ bool amdgpu_device_should_recover_gpu(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 2dfaf158ef07..e957e42c539a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -28,8 +28,6 @@ #include <linux/module.h> #include <linux/dma-buf.h> -const struct kgd2kfd_calls *kgd2kfd; - static const unsigned int compute_vmid_bitmap = 0xFF00; /* Total memory size in system memory and all GPU VRAM. Used to @@ -47,12 +45,9 @@ int amdgpu_amdkfd_init(void) amdgpu_amdkfd_total_mem_size *= si.mem_unit; #ifdef CONFIG_HSA_AMD - ret = kgd2kfd_init(KFD_INTERFACE_VERSION, &kgd2kfd); - if (ret) - kgd2kfd = NULL; + ret = kgd2kfd_init(); amdgpu_amdkfd_gpuvm_init_mem_limits(); #else - kgd2kfd = NULL; ret = -ENOENT; #endif @@ -61,17 +56,13 @@ int amdgpu_amdkfd_init(void) void amdgpu_amdkfd_fini(void) { - if (kgd2kfd) - kgd2kfd->exit(); + kgd2kfd_exit(); } void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev) { const struct kfd2kgd_calls *kfd2kgd; - if (!kgd2kfd) - return; - switch (adev->asic_type) { #ifdef CONFIG_DRM_AMDGPU_CIK case CHIP_KAVERI: @@ -98,8 +89,8 @@ void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev) return; } - adev->kfd.dev = kgd2kfd->probe((struct kgd_dev *)adev, - adev->pdev, kfd2kgd); + adev->kfd.dev = kgd2kfd_probe((struct kgd_dev *)adev, + adev->pdev, kfd2kgd); if (adev->kfd.dev) amdgpu_amdkfd_total_mem_size += adev->gmc.real_vram_size; @@ -182,7 +173,7 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) &gpu_resources.doorbell_start_offset); if (adev->asic_type < CHIP_VEGA10) { - kgd2kfd->device_init(adev->kfd.dev, &gpu_resources); + kgd2kfd_device_init(adev->kfd.dev, &gpu_resources); return; } @@ -197,13 +188,13 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) * can use each doorbell assignment twice. */ gpu_resources.sdma_doorbell[0][i] = - adev->doorbell_index.sdma_engine0 + (i >> 1); + adev->doorbell_index.sdma_engine[0] + (i >> 1); gpu_resources.sdma_doorbell[0][i+1] = - adev->doorbell_index.sdma_engine0 + 0x200 + (i >> 1); + adev->doorbell_index.sdma_engine[0] + 0x200 + (i >> 1); gpu_resources.sdma_doorbell[1][i] = - adev->doorbell_index.sdma_engine1 + (i >> 1); + adev->doorbell_index.sdma_engine[1] + (i >> 1); gpu_resources.sdma_doorbell[1][i+1] = - adev->doorbell_index.sdma_engine1 + 0x200 + (i >> 1); + adev->doorbell_index.sdma_engine[1] + 0x200 + (i >> 1); } /* Doorbells 0x0e0-0ff and 0x2e0-2ff are reserved for * SDMA, IH and VCN. So don't use them for the CP. @@ -211,14 +202,14 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev) gpu_resources.reserved_doorbell_mask = 0x1e0; gpu_resources.reserved_doorbell_val = 0x0e0; - kgd2kfd->device_init(adev->kfd.dev, &gpu_resources); + kgd2kfd_device_init(adev->kfd.dev, &gpu_resources); } } void amdgpu_amdkfd_device_fini(struct amdgpu_device *adev) { if (adev->kfd.dev) { - kgd2kfd->device_exit(adev->kfd.dev); + kgd2kfd_device_exit(adev->kfd.dev); adev->kfd.dev = NULL; } } @@ -227,13 +218,13 @@ void amdgpu_amdkfd_interrupt(struct amdgpu_device *adev, const void *ih_ring_entry) { if (adev->kfd.dev) - kgd2kfd->interrupt(adev->kfd.dev, ih_ring_entry); + kgd2kfd_interrupt(adev->kfd.dev, ih_ring_entry); } void amdgpu_amdkfd_suspend(struct amdgpu_device *adev) { if (adev->kfd.dev) - kgd2kfd->suspend(adev->kfd.dev); + kgd2kfd_suspend(adev->kfd.dev); } int amdgpu_amdkfd_resume(struct amdgpu_device *adev) @@ -241,7 +232,7 @@ int amdgpu_amdkfd_resume(struct amdgpu_device *adev) int r = 0; if (adev->kfd.dev) - r = kgd2kfd->resume(adev->kfd.dev); + r = kgd2kfd_resume(adev->kfd.dev); return r; } @@ -251,7 +242,7 @@ int amdgpu_amdkfd_pre_reset(struct amdgpu_device *adev) int r = 0; if (adev->kfd.dev) - r = kgd2kfd->pre_reset(adev->kfd.dev); + r = kgd2kfd_pre_reset(adev->kfd.dev); return r; } @@ -261,7 +252,7 @@ int amdgpu_amdkfd_post_reset(struct amdgpu_device *adev) int r = 0; if (adev->kfd.dev) - r = kgd2kfd->post_reset(adev->kfd.dev); + r = kgd2kfd_post_reset(adev->kfd.dev); return r; } @@ -619,4 +610,47 @@ struct kfd2kgd_calls *amdgpu_amdkfd_gfx_9_0_get_functions(void) { return NULL; } + +struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, struct pci_dev *pdev, + const struct kfd2kgd_calls *f2g) +{ + return NULL; +} + +bool kgd2kfd_device_init(struct kfd_dev *kfd, + const struct kgd2kfd_shared_resources *gpu_resources) +{ + return false; +} + +void kgd2kfd_device_exit(struct kfd_dev *kfd) +{ +} + +void kgd2kfd_exit(void) +{ +} + +void kgd2kfd_suspend(struct kfd_dev *kfd) +{ +} + +int kgd2kfd_resume(struct kfd_dev *kfd) +{ + return 0; +} + +int kgd2kfd_pre_reset(struct kfd_dev *kfd) +{ + return 0; +} + +int kgd2kfd_post_reset(struct kfd_dev *kfd) +{ + return 0; +} + +void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry) +{ +} #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index 70429f7aa9a8..0b31a1859023 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h @@ -33,7 +33,6 @@ #include "amdgpu_sync.h" #include "amdgpu_vm.h" -extern const struct kgd2kfd_calls *kgd2kfd; extern uint64_t amdgpu_amdkfd_total_mem_size; struct amdgpu_device; @@ -214,4 +213,22 @@ int amdgpu_amdkfd_gpuvm_import_dmabuf(struct kgd_dev *kgd, void amdgpu_amdkfd_gpuvm_init_mem_limits(void); void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo); +/* KGD2KFD callbacks */ +int kgd2kfd_init(void); +void kgd2kfd_exit(void); +struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, struct pci_dev *pdev, + const struct kfd2kgd_calls *f2g); +bool kgd2kfd_device_init(struct kfd_dev *kfd, + const struct kgd2kfd_shared_resources *gpu_resources); +void kgd2kfd_device_exit(struct kfd_dev *kfd); +void kgd2kfd_suspend(struct kfd_dev *kfd); +int kgd2kfd_resume(struct kfd_dev *kfd); +int kgd2kfd_pre_reset(struct kfd_dev *kfd); +int kgd2kfd_post_reset(struct kfd_dev *kfd); +void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry); +int kgd2kfd_quiesce_mm(struct mm_struct *mm); +int kgd2kfd_resume_mm(struct mm_struct *mm); +int kgd2kfd_schedule_evict_and_restore_process(struct mm_struct *mm, + struct dma_fence *fence); + #endif /* AMDGPU_AMDKFD_H_INCLUDED */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c index 574c1181ae9a..3107b9575929 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c @@ -122,7 +122,7 @@ static bool amdkfd_fence_enable_signaling(struct dma_fence *f) if (dma_fence_is_signaled(f)) return true; - if (!kgd2kfd->schedule_evict_and_restore_process(fence->mm, f)) + if (!kgd2kfd_schedule_evict_and_restore_process(fence->mm, f)) return true; return false; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index be1ab43473c6..d7b10d79f1de 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -1790,7 +1790,7 @@ int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, evicted_bos = atomic_inc_return(&process_info->evicted_bos); if (evicted_bos == 1) { /* First eviction, stop the queues */ - r = kgd2kfd->quiesce_mm(mm); + r = kgd2kfd_quiesce_mm(mm); if (r) pr_err("Failed to quiesce KFD\n"); schedule_delayed_work(&process_info->restore_userptr_work, @@ -2082,7 +2082,7 @@ static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work) evicted_bos) goto unlock_out; evicted_bos = 0; - if (kgd2kfd->resume_mm(mm)) { + if (kgd2kfd_resume_mm(mm)) { pr_err("%s: Failed to resume KFD\n", __func__); /* No recovery from this failure. Probably the CP is * hanging. No point trying again. diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c index a028661d9e20..92b11de19581 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c @@ -576,6 +576,7 @@ static const struct amdgpu_px_quirk amdgpu_px_quirk_list[] = { { 0x1002, 0x6900, 0x1028, 0x0812, AMDGPU_PX_QUIRK_FORCE_ATPX }, { 0x1002, 0x6900, 0x1028, 0x0813, AMDGPU_PX_QUIRK_FORCE_ATPX }, { 0x1002, 0x6900, 0x1025, 0x125A, AMDGPU_PX_QUIRK_FORCE_ATPX }, + { 0x1002, 0x6900, 0x17AA, 0x3806, AMDGPU_PX_QUIRK_FORCE_ATPX }, { 0, 0, 0, 0, 0 }, }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 1c49b8266d69..52a5e4fdc95b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -214,6 +214,7 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs case AMDGPU_CHUNK_ID_DEPENDENCIES: case AMDGPU_CHUNK_ID_SYNCOBJ_IN: case AMDGPU_CHUNK_ID_SYNCOBJ_OUT: + case AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES: break; default: @@ -1090,6 +1091,15 @@ static int amdgpu_cs_process_fence_dep(struct amdgpu_cs_parser *p, fence = amdgpu_ctx_get_fence(ctx, entity, deps[i].handle); + + if (chunk->chunk_id == AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES) { + struct drm_sched_fence *s_fence = to_drm_sched_fence(fence); + struct dma_fence *old = fence; + + fence = dma_fence_get(&s_fence->scheduled); + dma_fence_put(old); + } + if (IS_ERR(fence)) { r = PTR_ERR(fence); amdgpu_ctx_put(ctx); @@ -1177,7 +1187,8 @@ static int amdgpu_cs_dependencies(struct amdgpu_device *adev, chunk = &p->chunks[i]; - if (chunk->chunk_id == AMDGPU_CHUNK_ID_DEPENDENCIES) { + if (chunk->chunk_id == AMDGPU_CHUNK_ID_DEPENDENCIES || + chunk->chunk_id == AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES) { r = amdgpu_cs_process_fence_dep(p, chunk); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index f5816737c648..4f8fb4ecde34 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -1645,7 +1645,7 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) if (r) { DRM_ERROR("sw_init of IP block <%s> failed %d\n", adev->ip_blocks[i].version->funcs->name, r); - return r; + goto init_failed; } adev->ip_blocks[i].status.sw = true; @@ -1654,17 +1654,17 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) r = amdgpu_device_vram_scratch_init(adev); if (r) { DRM_ERROR("amdgpu_vram_scratch_init failed %d\n", r); - return r; + goto init_failed; } r = adev->ip_blocks[i].version->funcs->hw_init((void *)adev); if (r) { DRM_ERROR("hw_init %d failed %d\n", i, r); - return r; + goto init_failed; } r = amdgpu_device_wb_init(adev); if (r) { DRM_ERROR("amdgpu_device_wb_init failed %d\n", r); - return r; + goto init_failed; } adev->ip_blocks[i].status.hw = true; @@ -1675,7 +1675,7 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) AMDGPU_CSA_SIZE); if (r) { DRM_ERROR("allocate CSA failed %d\n", r); - return r; + goto init_failed; } } } @@ -1683,28 +1683,32 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) r = amdgpu_ucode_create_bo(adev); /* create ucode bo when sw_init complete*/ if (r) - return r; + goto init_failed; r = amdgpu_device_ip_hw_init_phase1(adev); if (r) - return r; + goto init_failed; r = amdgpu_device_fw_loading(adev); if (r) - return r; + goto init_failed; r = amdgpu_device_ip_hw_init_phase2(adev); if (r) - return r; + goto init_failed; if (adev->gmc.xgmi.num_physical_nodes > 1) amdgpu_xgmi_add_device(adev); amdgpu_amdkfd_device_init(adev); - if (amdgpu_sriov_vf(adev)) +init_failed: + if (amdgpu_sriov_vf(adev)) { + if (!r) + amdgpu_virt_init_data_exchange(adev); amdgpu_virt_release_full_gpu(adev, true); + } - return 0; + return r; } /** @@ -2131,7 +2135,7 @@ static int amdgpu_device_ip_reinit_early_sriov(struct amdgpu_device *adev) continue; r = block->version->funcs->hw_init(adev); - DRM_INFO("RE-INIT: %s %s\n", block->version->funcs->name, r?"failed":"succeeded"); + DRM_INFO("RE-INIT-early: %s %s\n", block->version->funcs->name, r?"failed":"succeeded"); if (r) return r; } @@ -2165,7 +2169,7 @@ static int amdgpu_device_ip_reinit_late_sriov(struct amdgpu_device *adev) continue; r = block->version->funcs->hw_init(adev); - DRM_INFO("RE-INIT: %s %s\n", block->version->funcs->name, r?"failed":"succeeded"); + DRM_INFO("RE-INIT-late: %s %s\n", block->version->funcs->name, r?"failed":"succeeded"); if (r) return r; } @@ -2546,6 +2550,17 @@ int amdgpu_device_init(struct amdgpu_device *adev, /* detect if we are with an SRIOV vbios */ amdgpu_device_detect_sriov_bios(adev); + /* check if we need to reset the asic + * E.g., driver was not cleanly unloaded previously, etc. + */ + if (!amdgpu_sriov_vf(adev) && amdgpu_asic_need_reset_on_init(adev)) { + r = amdgpu_asic_reset(adev); + if (r) { + dev_err(adev->dev, "asic reset on init failed\n"); + goto failed; + } + } + /* Post card if necessary */ if (amdgpu_device_need_post(adev)) { if (!adev->bios) { @@ -2610,6 +2625,8 @@ fence_driver_init: } dev_err(adev->dev, "amdgpu_device_ip_init failed\n"); amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_AMDGPU_INIT_FAIL, 0, 0); + if (amdgpu_virt_request_full_gpu(adev, false)) + amdgpu_virt_release_full_gpu(adev, false); goto failed; } @@ -2632,9 +2649,6 @@ fence_driver_init: goto failed; } - if (amdgpu_sriov_vf(adev)) - amdgpu_virt_init_data_exchange(adev); - amdgpu_fbdev_init(adev); r = amdgpu_pm_sysfs_init(adev); @@ -2798,7 +2812,7 @@ int amdgpu_device_suspend(struct drm_device *dev, bool suspend, bool fbcon) struct drm_framebuffer *fb = crtc->primary->fb; struct amdgpu_bo *robj; - if (amdgpu_crtc->cursor_bo) { + if (amdgpu_crtc->cursor_bo && !adev->enable_virtual_display) { struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo); r = amdgpu_bo_reserve(aobj, true); if (r == 0) { @@ -2906,7 +2920,7 @@ int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon) list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); - if (amdgpu_crtc->cursor_bo) { + if (amdgpu_crtc->cursor_bo && !adev->enable_virtual_display) { struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo); r = amdgpu_bo_reserve(aobj, true); if (r == 0) { @@ -3226,6 +3240,7 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev, r = amdgpu_ib_ring_tests(adev); error: + amdgpu_virt_init_data_exchange(adev); amdgpu_virt_release_full_gpu(adev, true); if (!r && adev->virt.gim_feature & AMDGIM_FEATURE_GIM_FLR_VRAMLOST) { atomic_inc(&adev->vram_lost_counter); @@ -3298,17 +3313,15 @@ static int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, if (!ring || !ring->sched.thread) continue; - kthread_park(ring->sched.thread); - - if (job && job->base.sched != &ring->sched) - continue; - - drm_sched_hw_job_reset(&ring->sched, job ? &job->base : NULL); + drm_sched_stop(&ring->sched); /* after all hw jobs are reset, hw fence is meaningless, so force_completion */ amdgpu_fence_driver_force_completion(ring); } + if(job) + drm_sched_increase_karma(&job->base); + if (!amdgpu_sriov_vf(adev)) { @@ -3454,14 +3467,10 @@ static void amdgpu_device_post_asic_reset(struct amdgpu_device *adev, if (!ring || !ring->sched.thread) continue; - /* only need recovery sched of the given job's ring - * or all rings (in the case @job is NULL) - * after above amdgpu_reset accomplished - */ - if ((!job || job->base.sched == &ring->sched) && !adev->asic_reset_res) - drm_sched_job_recovery(&ring->sched); + if (!adev->asic_reset_res) + drm_sched_resubmit_jobs(&ring->sched); - kthread_unpark(ring->sched.thread); + drm_sched_start(&ring->sched, !adev->asic_reset_res); } if (!amdgpu_device_has_dc_support(adev)) { @@ -3521,9 +3530,9 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, * by different nodes. No point also since the one node already executing * reset will also reset all the other nodes in the hive. */ - hive = amdgpu_get_xgmi_hive(adev); + hive = amdgpu_get_xgmi_hive(adev, 0); if (hive && adev->gmc.xgmi.num_physical_nodes > 1 && - !mutex_trylock(&hive->hive_lock)) + !mutex_trylock(&hive->reset_lock)) return 0; /* Start with adev pre asic reset first for soft reset check.*/ @@ -3602,13 +3611,45 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ } if (hive && adev->gmc.xgmi.num_physical_nodes > 1) - mutex_unlock(&hive->hive_lock); + mutex_unlock(&hive->reset_lock); if (r) dev_info(adev->dev, "GPU reset end with ret = %d\n", r); return r; } +static void amdgpu_device_get_min_pci_speed_width(struct amdgpu_device *adev, + enum pci_bus_speed *speed, + enum pcie_link_width *width) +{ + struct pci_dev *pdev = adev->pdev; + enum pci_bus_speed cur_speed; + enum pcie_link_width cur_width; + + *speed = PCI_SPEED_UNKNOWN; + *width = PCIE_LNK_WIDTH_UNKNOWN; + + while (pdev) { + cur_speed = pcie_get_speed_cap(pdev); + cur_width = pcie_get_width_cap(pdev); + + if (cur_speed != PCI_SPEED_UNKNOWN) { + if (*speed == PCI_SPEED_UNKNOWN) + *speed = cur_speed; + else if (cur_speed < *speed) + *speed = cur_speed; + } + + if (cur_width != PCIE_LNK_WIDTH_UNKNOWN) { + if (*width == PCIE_LNK_WIDTH_UNKNOWN) + *width = cur_width; + else if (cur_width < *width) + *width = cur_width; + } + pdev = pci_upstream_bridge(pdev); + } +} + /** * amdgpu_device_get_pcie_info - fence pcie info about the PCIE slot * @@ -3621,8 +3662,8 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev) { struct pci_dev *pdev; - enum pci_bus_speed speed_cap; - enum pcie_link_width link_width; + enum pci_bus_speed speed_cap, platform_speed_cap; + enum pcie_link_width platform_link_width; if (amdgpu_pcie_gen_cap) adev->pm.pcie_gen_mask = amdgpu_pcie_gen_cap; @@ -3639,6 +3680,12 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev) return; } + if (adev->pm.pcie_gen_mask && adev->pm.pcie_mlw_mask) + return; + + amdgpu_device_get_min_pci_speed_width(adev, &platform_speed_cap, + &platform_link_width); + if (adev->pm.pcie_gen_mask == 0) { /* asic caps */ pdev = adev->pdev; @@ -3664,22 +3711,20 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev) adev->pm.pcie_gen_mask |= CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1; } /* platform caps */ - pdev = adev->ddev->pdev->bus->self; - speed_cap = pcie_get_speed_cap(pdev); - if (speed_cap == PCI_SPEED_UNKNOWN) { + if (platform_speed_cap == PCI_SPEED_UNKNOWN) { adev->pm.pcie_gen_mask |= (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 | CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2); } else { - if (speed_cap == PCIE_SPEED_16_0GT) + if (platform_speed_cap == PCIE_SPEED_16_0GT) adev->pm.pcie_gen_mask |= (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 | CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2 | CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3 | CAIL_PCIE_LINK_SPEED_SUPPORT_GEN4); - else if (speed_cap == PCIE_SPEED_8_0GT) + else if (platform_speed_cap == PCIE_SPEED_8_0GT) adev->pm.pcie_gen_mask |= (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 | CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2 | CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3); - else if (speed_cap == PCIE_SPEED_5_0GT) + else if (platform_speed_cap == PCIE_SPEED_5_0GT) adev->pm.pcie_gen_mask |= (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 | CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2); else @@ -3688,12 +3733,10 @@ static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev) } } if (adev->pm.pcie_mlw_mask == 0) { - pdev = adev->ddev->pdev->bus->self; - link_width = pcie_get_width_cap(pdev); - if (link_width == PCIE_LNK_WIDTH_UNKNOWN) { + if (platform_link_width == PCIE_LNK_WIDTH_UNKNOWN) { adev->pm.pcie_mlw_mask |= AMDGPU_DEFAULT_PCIE_MLW_MASK; } else { - switch (link_width) { + switch (platform_link_width) { case PCIE_LNK_X32: adev->pm.pcie_mlw_mask = (CAIL_PCIE_LINK_WIDTH_SUPPORT_X32 | CAIL_PCIE_LINK_WIDTH_SUPPORT_X16 | diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c index 15ce7e681d67..b083b219b1a9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c @@ -188,10 +188,12 @@ int amdgpu_display_crtc_page_flip_target(struct drm_crtc *crtc, goto cleanup; } - r = amdgpu_bo_pin(new_abo, amdgpu_display_supported_domains(adev)); - if (unlikely(r != 0)) { - DRM_ERROR("failed to pin new abo buffer before flip\n"); - goto unreserve; + if (!adev->enable_virtual_display) { + r = amdgpu_bo_pin(new_abo, amdgpu_display_supported_domains(adev)); + if (unlikely(r != 0)) { + DRM_ERROR("failed to pin new abo buffer before flip\n"); + goto unreserve; + } } r = amdgpu_ttm_alloc_gart(&new_abo->tbo); @@ -211,7 +213,8 @@ int amdgpu_display_crtc_page_flip_target(struct drm_crtc *crtc, amdgpu_bo_get_tiling_flags(new_abo, &tiling_flags); amdgpu_bo_unreserve(new_abo); - work->base = amdgpu_bo_gpu_offset(new_abo); + if (!adev->enable_virtual_display) + work->base = amdgpu_bo_gpu_offset(new_abo); work->target_vblank = target - (uint32_t)drm_crtc_vblank_count(crtc) + amdgpu_get_vblank_counter_kms(dev, work->crtc_id); @@ -242,9 +245,10 @@ pflip_cleanup: goto cleanup; } unpin: - if (unlikely(amdgpu_bo_unpin(new_abo) != 0)) { - DRM_ERROR("failed to unpin new abo in error path\n"); - } + if (!adev->enable_virtual_display) + if (unlikely(amdgpu_bo_unpin(new_abo) != 0)) + DRM_ERROR("failed to unpin new abo in error path\n"); + unreserve: amdgpu_bo_unreserve(new_abo); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h index be620b29f4aa..1cfec06f81d4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_doorbell.h @@ -51,14 +51,7 @@ struct amdgpu_doorbell_index { uint32_t userqueue_start; uint32_t userqueue_end; uint32_t gfx_ring0; - uint32_t sdma_engine0; - uint32_t sdma_engine1; - uint32_t sdma_engine2; - uint32_t sdma_engine3; - uint32_t sdma_engine4; - uint32_t sdma_engine5; - uint32_t sdma_engine6; - uint32_t sdma_engine7; + uint32_t sdma_engine[8]; uint32_t ih; union { struct { @@ -79,6 +72,8 @@ struct amdgpu_doorbell_index { } uvd_vce; }; uint32_t max_assignment; + /* Per engine SDMA doorbell size in dword */ + uint32_t sdma_doorbell_range; }; typedef enum _AMDGPU_DOORBELL_ASSIGNMENT diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h index f972cd156795..2f61e9edb1c1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h @@ -364,6 +364,14 @@ enum amdgpu_pcie_gen { ((adev)->powerplay.pp_funcs->enable_mgpu_fan_boost(\ (adev)->powerplay.pp_handle)) +#define amdgpu_dpm_get_ppfeature_status(adev, buf) \ + ((adev)->powerplay.pp_funcs->get_ppfeature_status(\ + (adev)->powerplay.pp_handle, (buf))) + +#define amdgpu_dpm_set_ppfeature_status(adev, ppfeatures) \ + ((adev)->powerplay.pp_funcs->set_ppfeature_status(\ + (adev)->powerplay.pp_handle, (ppfeatures))) + struct amdgpu_dpm { struct amdgpu_ps *ps; /* number of valid power states */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index a1bb3773087b..7f3aa7b7e1d8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -71,9 +71,11 @@ * - 3.25.0 - Add support for sensor query info (stable pstate sclk/mclk). * - 3.26.0 - GFX9: Process AMDGPU_IB_FLAG_TC_WB_NOT_INVALIDATE. * - 3.27.0 - Add new chunk to to AMDGPU_CS to enable BO_LIST creation. + * - 3.28.0 - Add AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES + * - 3.29.0 - Add AMDGPU_IB_FLAG_RESET_GDS_MAX_WAVE_ID */ #define KMS_DRIVER_MAJOR 3 -#define KMS_DRIVER_MINOR 27 +#define KMS_DRIVER_MINOR 29 #define KMS_DRIVER_PATCHLEVEL 0 int amdgpu_vram_limit = 0; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gds.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gds.h index ecbcefe49a98..f89f5734d985 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gds.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gds.h @@ -37,6 +37,8 @@ struct amdgpu_gds { struct amdgpu_gds_asic_info mem; struct amdgpu_gds_asic_info gws; struct amdgpu_gds_asic_info oa; + uint32_t gds_compute_max_wave_id; + /* At present, GDS, GWS and OA resources for gfx (graphics) * is always pre-allocated and available for graphics operation. * Such resource is shared between all gfx clients. diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index f4f00217546e..d21dd2f369da 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -54,10 +54,6 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size, memset(&bp, 0, sizeof(bp)); *obj = NULL; - /* At least align on page size */ - if (alignment < PAGE_SIZE) { - alignment = PAGE_SIZE; - } bp.size = size; bp.byte_align = alignment; @@ -244,9 +240,6 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data, return -EINVAL; } flags |= AMDGPU_GEM_CREATE_NO_CPU_ACCESS; - /* GDS allocations must be DW aligned */ - if (args->in.domains & AMDGPU_GEM_DOMAIN_GDS) - size = ALIGN(size, 4); } if (flags & AMDGPU_GEM_CREATE_VM_ALWAYS_VALID) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c index c48207b377bc..0b8ef2d27d6b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c @@ -202,12 +202,12 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, amdgpu_asic_flush_hdp(adev, ring); } + if (need_ctx_switch) + status |= AMDGPU_HAVE_CTX_SWITCH; + skip_preamble = ring->current_ctx == fence_ctx; if (job && ring->funcs->emit_cntxcntl) { - if (need_ctx_switch) - status |= AMDGPU_HAVE_CTX_SWITCH; status |= job->preamble_status; - amdgpu_ring_emit_cntxcntl(ring, status); } @@ -221,8 +221,8 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, !amdgpu_sriov_vf(adev)) /* for SRIOV preemption, Preamble CE ib must be inserted anyway */ continue; - amdgpu_ring_emit_ib(ring, job, ib, need_ctx_switch); - need_ctx_switch = false; + amdgpu_ring_emit_ib(ring, job, ib, status); + status &= ~AMDGPU_HAVE_CTX_SWITCH; } if (ring->funcs->emit_tmz) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c index 8af67f649660..d0a5db777b6d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c @@ -52,6 +52,8 @@ int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, ih->use_bus_addr = use_bus_addr; if (use_bus_addr) { + dma_addr_t dma_addr; + if (ih->ring) return 0; @@ -59,21 +61,26 @@ int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, * add them to the end of the ring allocation. */ ih->ring = dma_alloc_coherent(adev->dev, ih->ring_size + 8, - &ih->rb_dma_addr, GFP_KERNEL); + &dma_addr, GFP_KERNEL); if (ih->ring == NULL) return -ENOMEM; memset((void *)ih->ring, 0, ih->ring_size + 8); - ih->wptr_offs = (ih->ring_size / 4) + 0; - ih->rptr_offs = (ih->ring_size / 4) + 1; + ih->gpu_addr = dma_addr; + ih->wptr_addr = dma_addr + ih->ring_size; + ih->wptr_cpu = &ih->ring[ih->ring_size / 4]; + ih->rptr_addr = dma_addr + ih->ring_size + 4; + ih->rptr_cpu = &ih->ring[(ih->ring_size / 4) + 1]; } else { - r = amdgpu_device_wb_get(adev, &ih->wptr_offs); + unsigned wptr_offs, rptr_offs; + + r = amdgpu_device_wb_get(adev, &wptr_offs); if (r) return r; - r = amdgpu_device_wb_get(adev, &ih->rptr_offs); + r = amdgpu_device_wb_get(adev, &rptr_offs); if (r) { - amdgpu_device_wb_free(adev, ih->wptr_offs); + amdgpu_device_wb_free(adev, wptr_offs); return r; } @@ -82,10 +89,15 @@ int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, &ih->ring_obj, &ih->gpu_addr, (void **)&ih->ring); if (r) { - amdgpu_device_wb_free(adev, ih->rptr_offs); - amdgpu_device_wb_free(adev, ih->wptr_offs); + amdgpu_device_wb_free(adev, rptr_offs); + amdgpu_device_wb_free(adev, wptr_offs); return r; } + + ih->wptr_addr = adev->wb.gpu_addr + wptr_offs * 4; + ih->wptr_cpu = &adev->wb.wb[wptr_offs]; + ih->rptr_addr = adev->wb.gpu_addr + rptr_offs * 4; + ih->rptr_cpu = &adev->wb.wb[rptr_offs]; } return 0; } @@ -109,13 +121,13 @@ void amdgpu_ih_ring_fini(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih) * add them to the end of the ring allocation. */ dma_free_coherent(adev->dev, ih->ring_size + 8, - (void *)ih->ring, ih->rb_dma_addr); + (void *)ih->ring, ih->gpu_addr); ih->ring = NULL; } else { amdgpu_bo_free_kernel(&ih->ring_obj, &ih->gpu_addr, (void **)&ih->ring); - amdgpu_device_wb_free(adev, ih->wptr_offs); - amdgpu_device_wb_free(adev, ih->rptr_offs); + amdgpu_device_wb_free(adev, (ih->wptr_addr - ih->gpu_addr) / 4); + amdgpu_device_wb_free(adev, (ih->rptr_addr - ih->gpu_addr) / 4); } } @@ -137,7 +149,7 @@ int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, if (!ih->enabled || adev->shutdown) return IRQ_NONE; - wptr = amdgpu_ih_get_wptr(adev); + wptr = amdgpu_ih_get_wptr(adev, ih); restart_ih: /* is somebody else already processing irqs? */ @@ -154,11 +166,11 @@ restart_ih: ih->rptr &= ih->ptr_mask; } - amdgpu_ih_set_rptr(adev); + amdgpu_ih_set_rptr(adev, ih); atomic_set(&ih->lock, 0); /* make sure wptr hasn't changed while processing */ - wptr = amdgpu_ih_get_wptr(adev); + wptr = amdgpu_ih_get_wptr(adev, ih); if (wptr != ih->rptr) goto restart_ih; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h index f877bb78d10a..1ccb1831382a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h @@ -31,34 +31,40 @@ struct amdgpu_iv_entry; * R6xx+ IH ring */ struct amdgpu_ih_ring { - struct amdgpu_bo *ring_obj; - volatile uint32_t *ring; - unsigned rptr; unsigned ring_size; - uint64_t gpu_addr; uint32_t ptr_mask; - atomic_t lock; - bool enabled; - unsigned wptr_offs; - unsigned rptr_offs; u32 doorbell_index; bool use_doorbell; bool use_bus_addr; - dma_addr_t rb_dma_addr; /* only used when use_bus_addr = true */ + + struct amdgpu_bo *ring_obj; + volatile uint32_t *ring; + uint64_t gpu_addr; + + uint64_t wptr_addr; + volatile uint32_t *wptr_cpu; + + uint64_t rptr_addr; + volatile uint32_t *rptr_cpu; + + bool enabled; + unsigned rptr; + atomic_t lock; }; /* provided by the ih block */ struct amdgpu_ih_funcs { /* ring read/write ptr handling, called from interrupt context */ - u32 (*get_wptr)(struct amdgpu_device *adev); - void (*decode_iv)(struct amdgpu_device *adev, + u32 (*get_wptr)(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih); + void (*decode_iv)(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, struct amdgpu_iv_entry *entry); - void (*set_rptr)(struct amdgpu_device *adev); + void (*set_rptr)(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih); }; -#define amdgpu_ih_get_wptr(adev) (adev)->irq.ih_funcs->get_wptr((adev)) -#define amdgpu_ih_decode_iv(adev, iv) (adev)->irq.ih_funcs->decode_iv((adev), (iv)) -#define amdgpu_ih_set_rptr(adev) (adev)->irq.ih_funcs->set_rptr((adev)) +#define amdgpu_ih_get_wptr(adev, ih) (adev)->irq.ih_funcs->get_wptr((adev), (ih)) +#define amdgpu_ih_decode_iv(adev, iv) \ + (adev)->irq.ih_funcs->decode_iv((adev), (ih), (iv)) +#define amdgpu_ih_set_rptr(adev, ih) (adev)->irq.ih_funcs->set_rptr((adev), (ih)) int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih, unsigned ring_size, bool use_bus_addr); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c index b7968f426862..8bfb3dab46f7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c @@ -148,6 +148,8 @@ static void amdgpu_irq_callback(struct amdgpu_device *adev, entry.iv_entry = (const uint32_t *)&ih->ring[ring_index]; amdgpu_ih_decode_iv(adev, &entry); + trace_amdgpu_iv(ih - &adev->irq.ih, &entry); + amdgpu_irq_dispatch(adev, &entry); } @@ -175,6 +177,36 @@ irqreturn_t amdgpu_irq_handler(int irq, void *arg) } /** + * amdgpu_irq_handle_ih1 - kick of processing for IH1 + * + * @work: work structure in struct amdgpu_irq + * + * Kick of processing IH ring 1. + */ +static void amdgpu_irq_handle_ih1(struct work_struct *work) +{ + struct amdgpu_device *adev = container_of(work, struct amdgpu_device, + irq.ih1_work); + + amdgpu_ih_process(adev, &adev->irq.ih1, amdgpu_irq_callback); +} + +/** + * amdgpu_irq_handle_ih2 - kick of processing for IH2 + * + * @work: work structure in struct amdgpu_irq + * + * Kick of processing IH ring 2. + */ +static void amdgpu_irq_handle_ih2(struct work_struct *work) +{ + struct amdgpu_device *adev = container_of(work, struct amdgpu_device, + irq.ih2_work); + + amdgpu_ih_process(adev, &adev->irq.ih2, amdgpu_irq_callback); +} + +/** * amdgpu_msi_ok - check whether MSI functionality is enabled * * @adev: amdgpu device pointer (unused) @@ -238,6 +270,9 @@ int amdgpu_irq_init(struct amdgpu_device *adev) amdgpu_hotplug_work_func); } + INIT_WORK(&adev->irq.ih1_work, amdgpu_irq_handle_ih1); + INIT_WORK(&adev->irq.ih2_work, amdgpu_irq_handle_ih2); + adev->irq.installed = true; r = drm_irq_install(adev->ddev, adev->ddev->pdev->irq); if (r) { @@ -367,8 +402,6 @@ void amdgpu_irq_dispatch(struct amdgpu_device *adev, bool handled = false; int r; - trace_amdgpu_iv(entry); - if (client_id >= AMDGPU_IRQ_CLIENTID_MAX) { DRM_DEBUG("Invalid client_id in IV: %d\n", client_id); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h index f6ce171cb8aa..c27decfda494 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h @@ -87,9 +87,11 @@ struct amdgpu_irq { /* status, etc. */ bool msi_enabled; /* msi enabled */ - /* interrupt ring */ - struct amdgpu_ih_ring ih; - const struct amdgpu_ih_funcs *ih_funcs; + /* interrupt rings */ + struct amdgpu_ih_ring ih, ih1, ih2; + const struct amdgpu_ih_funcs *ih_funcs; + struct work_struct ih1_work, ih2_work; + struct amdgpu_irq_src self_irq; /* gen irq stuff */ struct irq_domain *domain; /* GPU irq controller domain */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 728e15e5d68a..fd9c4beeaaa4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -426,12 +426,20 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev, size_t acc_size; int r; - page_align = roundup(bp->byte_align, PAGE_SIZE) >> PAGE_SHIFT; - if (bp->domain & (AMDGPU_GEM_DOMAIN_GDS | AMDGPU_GEM_DOMAIN_GWS | - AMDGPU_GEM_DOMAIN_OA)) + /* Note that GDS/GWS/OA allocates 1 page per byte/resource. */ + if (bp->domain & (AMDGPU_GEM_DOMAIN_GWS | AMDGPU_GEM_DOMAIN_OA)) { + /* GWS and OA don't need any alignment. */ + page_align = bp->byte_align; size <<= PAGE_SHIFT; - else + } else if (bp->domain & AMDGPU_GEM_DOMAIN_GDS) { + /* Both size and alignment must be a multiple of 4. */ + page_align = ALIGN(bp->byte_align, 4); + size = ALIGN(size, 4) << PAGE_SHIFT; + } else { + /* Memory should be aligned at least to a page size. */ + page_align = ALIGN(bp->byte_align, PAGE_SIZE) >> PAGE_SHIFT; size = ALIGN(size, PAGE_SIZE); + } if (!amdgpu_bo_validate_size(adev, size, bp->domain)) return -ENOMEM; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c index 1f61ed95727c..a7adb7b6bd98 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c @@ -626,11 +626,71 @@ static ssize_t amdgpu_get_pp_od_clk_voltage(struct device *dev, } /** - * DOC: pp_dpm_sclk pp_dpm_mclk pp_dpm_pcie + * DOC: ppfeatures + * + * The amdgpu driver provides a sysfs API for adjusting what powerplay + * features to be enabled. The file ppfeatures is used for this. And + * this is only available for Vega10 and later dGPUs. + * + * Reading back the file will show you the followings: + * - Current ppfeature masks + * - List of the all supported powerplay features with their naming, + * bitmasks and enablement status('Y'/'N' means "enabled"/"disabled"). + * + * To manually enable or disable a specific feature, just set or clear + * the corresponding bit from original ppfeature masks and input the + * new ppfeature masks. + */ +static ssize_t amdgpu_set_ppfeature_status(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + uint64_t featuremask; + int ret; + + ret = kstrtou64(buf, 0, &featuremask); + if (ret) + return -EINVAL; + + pr_debug("featuremask = 0x%llx\n", featuremask); + + if (adev->powerplay.pp_funcs->set_ppfeature_status) { + ret = amdgpu_dpm_set_ppfeature_status(adev, featuremask); + if (ret) + return -EINVAL; + } + + return count; +} + +static ssize_t amdgpu_get_ppfeature_status(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + + if (adev->powerplay.pp_funcs->get_ppfeature_status) + return amdgpu_dpm_get_ppfeature_status(adev, buf); + + return snprintf(buf, PAGE_SIZE, "\n"); +} + +/** + * DOC: pp_dpm_sclk pp_dpm_mclk pp_dpm_socclk pp_dpm_fclk pp_dpm_dcefclk + * pp_dpm_pcie * * The amdgpu driver provides a sysfs API for adjusting what power levels * are enabled for a given power state. The files pp_dpm_sclk, pp_dpm_mclk, - * and pp_dpm_pcie are used for this. + * pp_dpm_socclk, pp_dpm_fclk, pp_dpm_dcefclk and pp_dpm_pcie are used for + * this. + * + * pp_dpm_socclk and pp_dpm_dcefclk interfaces are only available for + * Vega10 and later ASICs. + * pp_dpm_fclk interface is only available for Vega20 and later ASICs. * * Reading back the files will show you the available power levels within * the power state and the clock information for those levels. @@ -640,6 +700,8 @@ static ssize_t amdgpu_get_pp_od_clk_voltage(struct device *dev, * Secondly,Enter a new value for each level by inputing a string that * contains " echo xx xx xx > pp_dpm_sclk/mclk/pcie" * E.g., echo 4 5 6 to > pp_dpm_sclk will enable sclk levels 4, 5, and 6. + * + * NOTE: change to the dcefclk max dpm level is not supported now */ static ssize_t amdgpu_get_pp_dpm_sclk(struct device *dev, @@ -750,6 +812,114 @@ static ssize_t amdgpu_set_pp_dpm_mclk(struct device *dev, return count; } +static ssize_t amdgpu_get_pp_dpm_socclk(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + + if (adev->powerplay.pp_funcs->print_clock_levels) + return amdgpu_dpm_print_clock_levels(adev, PP_SOCCLK, buf); + else + return snprintf(buf, PAGE_SIZE, "\n"); +} + +static ssize_t amdgpu_set_pp_dpm_socclk(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + int ret; + uint32_t mask = 0; + + ret = amdgpu_read_mask(buf, count, &mask); + if (ret) + return ret; + + if (adev->powerplay.pp_funcs->force_clock_level) + ret = amdgpu_dpm_force_clock_level(adev, PP_SOCCLK, mask); + + if (ret) + return -EINVAL; + + return count; +} + +static ssize_t amdgpu_get_pp_dpm_fclk(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + + if (adev->powerplay.pp_funcs->print_clock_levels) + return amdgpu_dpm_print_clock_levels(adev, PP_FCLK, buf); + else + return snprintf(buf, PAGE_SIZE, "\n"); +} + +static ssize_t amdgpu_set_pp_dpm_fclk(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + int ret; + uint32_t mask = 0; + + ret = amdgpu_read_mask(buf, count, &mask); + if (ret) + return ret; + + if (adev->powerplay.pp_funcs->force_clock_level) + ret = amdgpu_dpm_force_clock_level(adev, PP_FCLK, mask); + + if (ret) + return -EINVAL; + + return count; +} + +static ssize_t amdgpu_get_pp_dpm_dcefclk(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + + if (adev->powerplay.pp_funcs->print_clock_levels) + return amdgpu_dpm_print_clock_levels(adev, PP_DCEFCLK, buf); + else + return snprintf(buf, PAGE_SIZE, "\n"); +} + +static ssize_t amdgpu_set_pp_dpm_dcefclk(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + int ret; + uint32_t mask = 0; + + ret = amdgpu_read_mask(buf, count, &mask); + if (ret) + return ret; + + if (adev->powerplay.pp_funcs->force_clock_level) + ret = amdgpu_dpm_force_clock_level(adev, PP_DCEFCLK, mask); + + if (ret) + return -EINVAL; + + return count; +} + static ssize_t amdgpu_get_pp_dpm_pcie(struct device *dev, struct device_attribute *attr, char *buf) @@ -990,6 +1160,31 @@ static ssize_t amdgpu_get_busy_percent(struct device *dev, return snprintf(buf, PAGE_SIZE, "%d\n", value); } +/** + * DOC: pcie_bw + * + * The amdgpu driver provides a sysfs API for estimating how much data + * has been received and sent by the GPU in the last second through PCIe. + * The file pcie_bw is used for this. + * The Perf counters count the number of received and sent messages and return + * those values, as well as the maximum payload size of a PCIe packet (mps). + * Note that it is not possible to easily and quickly obtain the size of each + * packet transmitted, so we output the max payload size (mps) to allow for + * quick estimation of the PCIe bandwidth usage + */ +static ssize_t amdgpu_get_pcie_bw(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + uint64_t count0, count1; + + amdgpu_asic_get_pcie_usage(adev, &count0, &count1); + return snprintf(buf, PAGE_SIZE, "%llu %llu %i\n", + count0, count1, pcie_get_mps(adev->pdev)); +} + static DEVICE_ATTR(power_dpm_state, S_IRUGO | S_IWUSR, amdgpu_get_dpm_state, amdgpu_set_dpm_state); static DEVICE_ATTR(power_dpm_force_performance_level, S_IRUGO | S_IWUSR, amdgpu_get_dpm_forced_performance_level, @@ -1008,6 +1203,15 @@ static DEVICE_ATTR(pp_dpm_sclk, S_IRUGO | S_IWUSR, static DEVICE_ATTR(pp_dpm_mclk, S_IRUGO | S_IWUSR, amdgpu_get_pp_dpm_mclk, amdgpu_set_pp_dpm_mclk); +static DEVICE_ATTR(pp_dpm_socclk, S_IRUGO | S_IWUSR, + amdgpu_get_pp_dpm_socclk, + amdgpu_set_pp_dpm_socclk); +static DEVICE_ATTR(pp_dpm_fclk, S_IRUGO | S_IWUSR, + amdgpu_get_pp_dpm_fclk, + amdgpu_set_pp_dpm_fclk); +static DEVICE_ATTR(pp_dpm_dcefclk, S_IRUGO | S_IWUSR, + amdgpu_get_pp_dpm_dcefclk, + amdgpu_set_pp_dpm_dcefclk); static DEVICE_ATTR(pp_dpm_pcie, S_IRUGO | S_IWUSR, amdgpu_get_pp_dpm_pcie, amdgpu_set_pp_dpm_pcie); @@ -1025,6 +1229,10 @@ static DEVICE_ATTR(pp_od_clk_voltage, S_IRUGO | S_IWUSR, amdgpu_set_pp_od_clk_voltage); static DEVICE_ATTR(gpu_busy_percent, S_IRUGO, amdgpu_get_busy_percent, NULL); +static DEVICE_ATTR(pcie_bw, S_IRUGO, amdgpu_get_pcie_bw, NULL); +static DEVICE_ATTR(ppfeatures, S_IRUGO | S_IWUSR, + amdgpu_get_ppfeature_status, + amdgpu_set_ppfeature_status); static ssize_t amdgpu_hwmon_show_temp(struct device *dev, struct device_attribute *attr, @@ -1516,6 +1724,75 @@ static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev, return count; } +static ssize_t amdgpu_hwmon_show_sclk(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct amdgpu_device *adev = dev_get_drvdata(dev); + struct drm_device *ddev = adev->ddev; + uint32_t sclk; + int r, size = sizeof(sclk); + + /* Can't get voltage when the card is off */ + if ((adev->flags & AMD_IS_PX) && + (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) + return -EINVAL; + + /* sanity check PP is enabled */ + if (!(adev->powerplay.pp_funcs && + adev->powerplay.pp_funcs->read_sensor)) + return -EINVAL; + + /* get the sclk */ + r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GFX_SCLK, + (void *)&sclk, &size); + if (r) + return r; + + return snprintf(buf, PAGE_SIZE, "%d\n", sclk * 10 * 1000); +} + +static ssize_t amdgpu_hwmon_show_sclk_label(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + return snprintf(buf, PAGE_SIZE, "sclk\n"); +} + +static ssize_t amdgpu_hwmon_show_mclk(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct amdgpu_device *adev = dev_get_drvdata(dev); + struct drm_device *ddev = adev->ddev; + uint32_t mclk; + int r, size = sizeof(mclk); + + /* Can't get voltage when the card is off */ + if ((adev->flags & AMD_IS_PX) && + (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) + return -EINVAL; + + /* sanity check PP is enabled */ + if (!(adev->powerplay.pp_funcs && + adev->powerplay.pp_funcs->read_sensor)) + return -EINVAL; + + /* get the sclk */ + r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GFX_MCLK, + (void *)&mclk, &size); + if (r) + return r; + + return snprintf(buf, PAGE_SIZE, "%d\n", mclk * 10 * 1000); +} + +static ssize_t amdgpu_hwmon_show_mclk_label(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + return snprintf(buf, PAGE_SIZE, "mclk\n"); +} /** * DOC: hwmon @@ -1532,6 +1809,10 @@ static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev, * * - GPU fan * + * - GPU gfx/compute engine clock + * + * - GPU memory clock (dGPU only) + * * hwmon interfaces for GPU temperature: * * - temp1_input: the on die GPU temperature in millidegrees Celsius @@ -1576,6 +1857,12 @@ static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev, * * - fan[1-*]_enable: Enable or disable the sensors.1: Enable 0: Disable * + * hwmon interfaces for GPU clocks: + * + * - freq1_input: the gfx/compute clock in hertz + * + * - freq2_input: the memory clock in hertz + * * You can use hwmon tools like sensors to view this information on your system. * */ @@ -1600,6 +1887,10 @@ static SENSOR_DEVICE_ATTR(power1_average, S_IRUGO, amdgpu_hwmon_show_power_avg, static SENSOR_DEVICE_ATTR(power1_cap_max, S_IRUGO, amdgpu_hwmon_show_power_cap_max, NULL, 0); static SENSOR_DEVICE_ATTR(power1_cap_min, S_IRUGO, amdgpu_hwmon_show_power_cap_min, NULL, 0); static SENSOR_DEVICE_ATTR(power1_cap, S_IRUGO | S_IWUSR, amdgpu_hwmon_show_power_cap, amdgpu_hwmon_set_power_cap, 0); +static SENSOR_DEVICE_ATTR(freq1_input, S_IRUGO, amdgpu_hwmon_show_sclk, NULL, 0); +static SENSOR_DEVICE_ATTR(freq1_label, S_IRUGO, amdgpu_hwmon_show_sclk_label, NULL, 0); +static SENSOR_DEVICE_ATTR(freq2_input, S_IRUGO, amdgpu_hwmon_show_mclk, NULL, 0); +static SENSOR_DEVICE_ATTR(freq2_label, S_IRUGO, amdgpu_hwmon_show_mclk_label, NULL, 0); static struct attribute *hwmon_attributes[] = { &sensor_dev_attr_temp1_input.dev_attr.attr, @@ -1622,6 +1913,10 @@ static struct attribute *hwmon_attributes[] = { &sensor_dev_attr_power1_cap_max.dev_attr.attr, &sensor_dev_attr_power1_cap_min.dev_attr.attr, &sensor_dev_attr_power1_cap.dev_attr.attr, + &sensor_dev_attr_freq1_input.dev_attr.attr, + &sensor_dev_attr_freq1_label.dev_attr.attr, + &sensor_dev_attr_freq2_input.dev_attr.attr, + &sensor_dev_attr_freq2_label.dev_attr.attr, NULL }; @@ -1686,7 +1981,8 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj, effective_mode &= ~S_IWUSR; if ((adev->flags & AMD_IS_APU) && - (attr == &sensor_dev_attr_power1_cap_max.dev_attr.attr || + (attr == &sensor_dev_attr_power1_average.dev_attr.attr || + attr == &sensor_dev_attr_power1_cap_max.dev_attr.attr || attr == &sensor_dev_attr_power1_cap_min.dev_attr.attr|| attr == &sensor_dev_attr_power1_cap.dev_attr.attr)) return 0; @@ -1712,6 +2008,12 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj, attr == &sensor_dev_attr_in1_label.dev_attr.attr)) return 0; + /* no mclk on APUs */ + if ((adev->flags & AMD_IS_APU) && + (attr == &sensor_dev_attr_freq2_input.dev_attr.attr || + attr == &sensor_dev_attr_freq2_label.dev_attr.attr)) + return 0; + return effective_mode; } @@ -2008,6 +2310,7 @@ void amdgpu_pm_print_power_states(struct amdgpu_device *adev) int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) { + struct pp_hwmgr *hwmgr = adev->powerplay.pp_handle; int ret; if (adev->pm.sysfs_initialized) @@ -2069,6 +2372,25 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) DRM_ERROR("failed to create device file pp_dpm_mclk\n"); return ret; } + if (adev->asic_type >= CHIP_VEGA10) { + ret = device_create_file(adev->dev, &dev_attr_pp_dpm_socclk); + if (ret) { + DRM_ERROR("failed to create device file pp_dpm_socclk\n"); + return ret; + } + ret = device_create_file(adev->dev, &dev_attr_pp_dpm_dcefclk); + if (ret) { + DRM_ERROR("failed to create device file pp_dpm_dcefclk\n"); + return ret; + } + } + if (adev->asic_type >= CHIP_VEGA20) { + ret = device_create_file(adev->dev, &dev_attr_pp_dpm_fclk); + if (ret) { + DRM_ERROR("failed to create device file pp_dpm_fclk\n"); + return ret; + } + } ret = device_create_file(adev->dev, &dev_attr_pp_dpm_pcie); if (ret) { DRM_ERROR("failed to create device file pp_dpm_pcie\n"); @@ -2091,12 +2413,14 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) "pp_power_profile_mode\n"); return ret; } - ret = device_create_file(adev->dev, - &dev_attr_pp_od_clk_voltage); - if (ret) { - DRM_ERROR("failed to create device file " - "pp_od_clk_voltage\n"); - return ret; + if (hwmgr->od_enabled) { + ret = device_create_file(adev->dev, + &dev_attr_pp_od_clk_voltage); + if (ret) { + DRM_ERROR("failed to create device file " + "pp_od_clk_voltage\n"); + return ret; + } } ret = device_create_file(adev->dev, &dev_attr_gpu_busy_percent); @@ -2105,12 +2429,31 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) "gpu_busy_level\n"); return ret; } + /* PCIe Perf counters won't work on APU nodes */ + if (!(adev->flags & AMD_IS_APU)) { + ret = device_create_file(adev->dev, &dev_attr_pcie_bw); + if (ret) { + DRM_ERROR("failed to create device file pcie_bw\n"); + return ret; + } + } ret = amdgpu_debugfs_pm_init(adev); if (ret) { DRM_ERROR("Failed to register debugfs file for dpm!\n"); return ret; } + if ((adev->asic_type >= CHIP_VEGA10) && + !(adev->flags & AMD_IS_APU)) { + ret = device_create_file(adev->dev, + &dev_attr_ppfeatures); + if (ret) { + DRM_ERROR("failed to create device file " + "ppfeatures\n"); + return ret; + } + } + adev->pm.sysfs_initialized = true; return 0; @@ -2118,6 +2461,8 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) void amdgpu_pm_sysfs_fini(struct amdgpu_device *adev) { + struct pp_hwmgr *hwmgr = adev->powerplay.pp_handle; + if (adev->pm.dpm_enabled == 0) return; @@ -2133,14 +2478,26 @@ void amdgpu_pm_sysfs_fini(struct amdgpu_device *adev) device_remove_file(adev->dev, &dev_attr_pp_dpm_sclk); device_remove_file(adev->dev, &dev_attr_pp_dpm_mclk); + if (adev->asic_type >= CHIP_VEGA10) { + device_remove_file(adev->dev, &dev_attr_pp_dpm_socclk); + device_remove_file(adev->dev, &dev_attr_pp_dpm_dcefclk); + } device_remove_file(adev->dev, &dev_attr_pp_dpm_pcie); + if (adev->asic_type >= CHIP_VEGA20) + device_remove_file(adev->dev, &dev_attr_pp_dpm_fclk); device_remove_file(adev->dev, &dev_attr_pp_sclk_od); device_remove_file(adev->dev, &dev_attr_pp_mclk_od); device_remove_file(adev->dev, &dev_attr_pp_power_profile_mode); - device_remove_file(adev->dev, - &dev_attr_pp_od_clk_voltage); + if (hwmgr->od_enabled) + device_remove_file(adev->dev, + &dev_attr_pp_od_clk_voltage); device_remove_file(adev->dev, &dev_attr_gpu_busy_percent); + if (!(adev->flags & AMD_IS_APU)) + device_remove_file(adev->dev, &dev_attr_pcie_bw); + if ((adev->asic_type >= CHIP_VEGA10) && + !(adev->flags & AMD_IS_APU)) + device_remove_file(adev->dev, &dev_attr_ppfeatures); } void amdgpu_pm_compute_clocks(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index 8fab0d637ee5..d87f165e3a23 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -67,9 +67,6 @@ static int psp_sw_init(void *handle) psp->adev = adev; - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) - return 0; - ret = psp_init_microcode(psp); if (ret) { DRM_ERROR("Failed to load psp firmware!\n"); @@ -83,9 +80,6 @@ static int psp_sw_fini(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) - return 0; - release_firmware(adev->psp.sos_fw); adev->psp.sos_fw = NULL; release_firmware(adev->psp.asd_fw); @@ -140,13 +134,24 @@ psp_cmd_submit_buf(struct psp_context *psp, while (*((unsigned int *)psp->fence_buf) != index) msleep(1); - /* the status field must be 0 after FW is loaded */ - if (ucode && psp->cmd_buf_mem->resp.status) { - DRM_ERROR("failed loading with status (%d) and ucode id (%d)\n", - psp->cmd_buf_mem->resp.status, ucode->ucode_id); - return -EINVAL; + /* In some cases, psp response status is not 0 even there is no + * problem while the command is submitted. Some version of PSP FW + * doesn't write 0 to that field. + * So here we would like to only print a warning instead of an error + * during psp initialization to avoid breaking hw_init and it doesn't + * return -EINVAL. + */ + if (psp->cmd_buf_mem->resp.status) { + if (ucode) + DRM_WARN("failed to load ucode id (%d) ", + ucode->ucode_id); + DRM_WARN("psp command failed and response status is (%d)\n", + psp->cmd_buf_mem->resp.status); } + /* get xGMI session id from response buffer */ + cmd->resp.session_id = psp->cmd_buf_mem->resp.session_id; + if (ucode) { ucode->tmr_mc_addr_lo = psp->cmd_buf_mem->resp.fw_addr_lo; ucode->tmr_mc_addr_hi = psp->cmd_buf_mem->resp.fw_addr_hi; @@ -495,6 +500,98 @@ static int psp_hw_start(struct psp_context *psp) return 0; } +static int psp_get_fw_type(struct amdgpu_firmware_info *ucode, + enum psp_gfx_fw_type *type) +{ + switch (ucode->ucode_id) { + case AMDGPU_UCODE_ID_SDMA0: + *type = GFX_FW_TYPE_SDMA0; + break; + case AMDGPU_UCODE_ID_SDMA1: + *type = GFX_FW_TYPE_SDMA1; + break; + case AMDGPU_UCODE_ID_CP_CE: + *type = GFX_FW_TYPE_CP_CE; + break; + case AMDGPU_UCODE_ID_CP_PFP: + *type = GFX_FW_TYPE_CP_PFP; + break; + case AMDGPU_UCODE_ID_CP_ME: + *type = GFX_FW_TYPE_CP_ME; + break; + case AMDGPU_UCODE_ID_CP_MEC1: + *type = GFX_FW_TYPE_CP_MEC; + break; + case AMDGPU_UCODE_ID_CP_MEC1_JT: + *type = GFX_FW_TYPE_CP_MEC_ME1; + break; + case AMDGPU_UCODE_ID_CP_MEC2: + *type = GFX_FW_TYPE_CP_MEC; + break; + case AMDGPU_UCODE_ID_CP_MEC2_JT: + *type = GFX_FW_TYPE_CP_MEC_ME2; + break; + case AMDGPU_UCODE_ID_RLC_G: + *type = GFX_FW_TYPE_RLC_G; + break; + case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL: + *type = GFX_FW_TYPE_RLC_RESTORE_LIST_SRM_CNTL; + break; + case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM: + *type = GFX_FW_TYPE_RLC_RESTORE_LIST_GPM_MEM; + break; + case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM: + *type = GFX_FW_TYPE_RLC_RESTORE_LIST_SRM_MEM; + break; + case AMDGPU_UCODE_ID_SMC: + *type = GFX_FW_TYPE_SMU; + break; + case AMDGPU_UCODE_ID_UVD: + *type = GFX_FW_TYPE_UVD; + break; + case AMDGPU_UCODE_ID_UVD1: + *type = GFX_FW_TYPE_UVD1; + break; + case AMDGPU_UCODE_ID_VCE: + *type = GFX_FW_TYPE_VCE; + break; + case AMDGPU_UCODE_ID_VCN: + *type = GFX_FW_TYPE_VCN; + break; + case AMDGPU_UCODE_ID_DMCU_ERAM: + *type = GFX_FW_TYPE_DMCU_ERAM; + break; + case AMDGPU_UCODE_ID_DMCU_INTV: + *type = GFX_FW_TYPE_DMCU_ISR; + break; + case AMDGPU_UCODE_ID_MAXIMUM: + default: + return -EINVAL; + } + + return 0; +} + +static int psp_prep_load_ip_fw_cmd_buf(struct amdgpu_firmware_info *ucode, + struct psp_gfx_cmd_resp *cmd) +{ + int ret; + uint64_t fw_mem_mc_addr = ucode->mc_addr; + + memset(cmd, 0, sizeof(struct psp_gfx_cmd_resp)); + + cmd->cmd_id = GFX_CMD_ID_LOAD_IP_FW; + cmd->cmd.cmd_load_ip_fw.fw_phy_addr_lo = lower_32_bits(fw_mem_mc_addr); + cmd->cmd.cmd_load_ip_fw.fw_phy_addr_hi = upper_32_bits(fw_mem_mc_addr); + cmd->cmd.cmd_load_ip_fw.fw_size = ucode->ucode_size; + + ret = psp_get_fw_type(ucode, &cmd->cmd.cmd_load_ip_fw.fw_type); + if (ret) + DRM_ERROR("Unknown firmware type\n"); + + return ret; +} + static int psp_np_fw_load(struct psp_context *psp) { int i, ret; @@ -516,7 +613,7 @@ static int psp_np_fw_load(struct psp_context *psp) /*skip ucode loading in SRIOV VF */ continue; - ret = psp_prep_cmd_buf(ucode, psp->cmd); + ret = psp_prep_load_ip_fw_cmd_buf(ucode, psp->cmd); if (ret) return ret; @@ -541,7 +638,7 @@ static int psp_load_fw(struct amdgpu_device *adev) struct psp_context *psp = &adev->psp; if (amdgpu_sriov_vf(adev) && adev->in_gpu_reset) { - psp_ring_destroy(psp, PSP_RING_TYPE__KM); + psp_ring_stop(psp, PSP_RING_TYPE__KM); /* should not destroy ring, only stop */ goto skip_memalloc; } @@ -618,10 +715,6 @@ static int psp_hw_init(void *handle) int ret; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) - return 0; - mutex_lock(&adev->firmware.mutex); /* * This sequence is just used on hw_init only once, no need on @@ -651,9 +744,6 @@ static int psp_hw_fini(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; struct psp_context *psp = &adev->psp; - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) - return 0; - if (adev->gmc.xgmi.num_physical_nodes > 1 && psp->xgmi_context.initialized == 1) psp_xgmi_terminate(psp); @@ -682,9 +772,6 @@ static int psp_suspend(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; struct psp_context *psp = &adev->psp; - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) - return 0; - if (adev->gmc.xgmi.num_physical_nodes > 1 && psp->xgmi_context.initialized == 1) { ret = psp_xgmi_terminate(psp); @@ -709,9 +796,6 @@ static int psp_resume(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; struct psp_context *psp = &adev->psp; - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) - return 0; - DRM_INFO("PSP is resuming...\n"); mutex_lock(&adev->firmware.mutex); @@ -747,11 +831,6 @@ static bool psp_check_fw_loading_status(struct amdgpu_device *adev, { struct amdgpu_firmware_info *ucode = NULL; - if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP) { - DRM_INFO("firmware is not loaded by PSP\n"); - return true; - } - if (!adev->firmware.fw_size) return false; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h index 3ee573b4016e..2ef98cc755d6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h @@ -65,8 +65,6 @@ struct psp_funcs int (*init_microcode)(struct psp_context *psp); int (*bootloader_load_sysdrv)(struct psp_context *psp); int (*bootloader_load_sos)(struct psp_context *psp); - int (*prep_cmd_buf)(struct amdgpu_firmware_info *ucode, - struct psp_gfx_cmd_resp *cmd); int (*ring_init)(struct psp_context *psp, enum psp_ring_type ring_type); int (*ring_create)(struct psp_context *psp, enum psp_ring_type ring_type); @@ -176,7 +174,6 @@ struct psp_xgmi_topology_info { struct psp_xgmi_node_info nodes[AMDGPU_XGMI_MAX_CONNECTED_NODES]; }; -#define psp_prep_cmd_buf(ucode, type) (psp)->funcs->prep_cmd_buf((ucode), (type)) #define psp_ring_init(psp, type) (psp)->funcs->ring_init((psp), (type)) #define psp_ring_create(psp, type) (psp)->funcs->ring_create((psp), (type)) #define psp_ring_stop(psp, type) (psp)->funcs->ring_stop((psp), (type)) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index d87e828a084b..d7fae2676269 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -131,7 +131,7 @@ struct amdgpu_ring_funcs { void (*emit_ib)(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch); + uint32_t flags); void (*emit_fence)(struct amdgpu_ring *ring, uint64_t addr, uint64_t seq, unsigned flags); void (*emit_pipeline_sync)(struct amdgpu_ring *ring); @@ -229,7 +229,7 @@ struct amdgpu_ring { #define amdgpu_ring_get_rptr(r) (r)->funcs->get_rptr((r)) #define amdgpu_ring_get_wptr(r) (r)->funcs->get_wptr((r)) #define amdgpu_ring_set_wptr(r) (r)->funcs->set_wptr((r)) -#define amdgpu_ring_emit_ib(r, job, ib, c) ((r)->funcs->emit_ib((r), (job), (ib), (c))) +#define amdgpu_ring_emit_ib(r, job, ib, flags) ((r)->funcs->emit_ib((r), (job), (ib), (flags))) #define amdgpu_ring_emit_pipeline_sync(r) (r)->funcs->emit_pipeline_sync((r)) #define amdgpu_ring_emit_vm_flush(r, vmid, addr) (r)->funcs->emit_vm_flush((r), (vmid), (addr)) #define amdgpu_ring_emit_fence(r, addr, seq, flags) (r)->funcs->emit_fence((r), (addr), (seq), (flags)) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h index 626abca770a0..d3ca2424b5fe 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h @@ -76,9 +76,10 @@ TRACE_EVENT(amdgpu_mm_wreg, ); TRACE_EVENT(amdgpu_iv, - TP_PROTO(struct amdgpu_iv_entry *iv), - TP_ARGS(iv), + TP_PROTO(unsigned ih, struct amdgpu_iv_entry *iv), + TP_ARGS(ih, iv), TP_STRUCT__entry( + __field(unsigned, ih) __field(unsigned, client_id) __field(unsigned, src_id) __field(unsigned, ring_id) @@ -90,6 +91,7 @@ TRACE_EVENT(amdgpu_iv, __array(unsigned, src_data, 4) ), TP_fast_assign( + __entry->ih = ih; __entry->client_id = iv->client_id; __entry->src_id = iv->src_id; __entry->ring_id = iv->ring_id; @@ -103,8 +105,9 @@ TRACE_EVENT(amdgpu_iv, __entry->src_data[2] = iv->src_data[2]; __entry->src_data[3] = iv->src_data[3]; ), - TP_printk("client_id:%u src_id:%u ring:%u vmid:%u timestamp: %llu pasid:%u src_data: %08x %08x %08x %08x", - __entry->client_id, __entry->src_id, + TP_printk("ih:%u client_id:%u src_id:%u ring:%u vmid:%u " + "timestamp: %llu pasid:%u src_data: %08x %08x %08x %08x", + __entry->ih, __entry->client_id, __entry->src_id, __entry->ring_id, __entry->vmid, __entry->timestamp, __entry->pasid, __entry->src_data[0], __entry->src_data[1], diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index c91ec3101d00..73e71e61dc99 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -1546,7 +1546,8 @@ static struct ttm_bo_driver amdgpu_bo_driver = { .io_mem_reserve = &amdgpu_ttm_io_mem_reserve, .io_mem_free = &amdgpu_ttm_io_mem_free, .io_mem_pfn = amdgpu_ttm_io_mem_pfn, - .access_memory = &amdgpu_ttm_access_memory + .access_memory = &amdgpu_ttm_access_memory, + .del_from_lru_notify = &amdgpu_vm_del_from_lru_notify }; /* @@ -1755,7 +1756,7 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) } r = amdgpu_bo_create_kernel(adev, adev->gds.mem.gfx_partition_size, - PAGE_SIZE, AMDGPU_GEM_DOMAIN_GDS, + 4, AMDGPU_GEM_DOMAIN_GDS, &adev->gds.gds_gfx_bo, NULL, NULL); if (r) return r; @@ -1768,7 +1769,7 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) } r = amdgpu_bo_create_kernel(adev, adev->gds.gws.gfx_partition_size, - PAGE_SIZE, AMDGPU_GEM_DOMAIN_GWS, + 1, AMDGPU_GEM_DOMAIN_GWS, &adev->gds.gws_gfx_bo, NULL, NULL); if (r) return r; @@ -1781,7 +1782,7 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) } r = amdgpu_bo_create_kernel(adev, adev->gds.oa.gfx_partition_size, - PAGE_SIZE, AMDGPU_GEM_DOMAIN_OA, + 1, AMDGPU_GEM_DOMAIN_OA, &adev->gds.oa_gfx_bo, NULL, NULL); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c index 98a1b2ce2b9d..c021b114c8a4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c @@ -1035,7 +1035,7 @@ out: void amdgpu_vce_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { amdgpu_ring_write(ring, VCE_CMD_IB); amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr)); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h index 50293652af14..30ea54dd9117 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h @@ -66,7 +66,7 @@ void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp); int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx); int amdgpu_vce_ring_parse_cs_vm(struct amdgpu_cs_parser *p, uint32_t ib_idx); void amdgpu_vce_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, - struct amdgpu_ib *ib, bool ctx_switch); + struct amdgpu_ib *ib, uint32_t flags); void amdgpu_vce_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq, unsigned flags); int amdgpu_vce_ring_test_ring(struct amdgpu_ring *ring); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index e73d152659a2..75481cf3348f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -107,14 +107,6 @@ struct amdgpu_pte_update_params { * DMA addresses to use for mapping, used during VM update by CPU */ dma_addr_t *pages_addr; - - /** - * @kptr: - * - * Kernel pointer of PD/PT BO that needs to be updated, - * used during VM update by CPU - */ - void *kptr; }; /** @@ -623,6 +615,28 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, list_add(&entry->tv.head, validated); } +void amdgpu_vm_del_from_lru_notify(struct ttm_buffer_object *bo) +{ + struct amdgpu_bo *abo; + struct amdgpu_vm_bo_base *bo_base; + + if (!amdgpu_bo_is_amdgpu_bo(bo)) + return; + + if (bo->mem.placement & TTM_PL_FLAG_NO_EVICT) + return; + + abo = ttm_to_amdgpu_bo(bo); + if (!abo->parent) + return; + for (bo_base = abo->vm_bo; bo_base; bo_base = bo_base->next) { + struct amdgpu_vm *vm = bo_base->vm; + + if (abo->tbo.resv == vm->root.base.bo->tbo.resv) + vm->bulk_moveable = false; + } + +} /** * amdgpu_vm_move_to_lru_tail - move all BOs to the end of LRU * @@ -799,9 +813,16 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, addr += ats_entries * 8; } - if (entries) + if (entries) { + uint64_t value = 0; + + /* Workaround for fault priority problem on GMC9 */ + if (level == AMDGPU_VM_PTB && adev->asic_type >= CHIP_VEGA10) + value = AMDGPU_PTE_EXECUTABLE; + amdgpu_vm_set_pte_pde(adev, &job->ibs[0], addr, 0, - entries, 0, 0); + entries, 0, value); + } amdgpu_ring_pad_ib(ring, &job->ibs[0]); @@ -847,9 +868,6 @@ static void amdgpu_vm_bo_param(struct amdgpu_device *adev, struct amdgpu_vm *vm, bp->size = amdgpu_vm_bo_size(adev, level); bp->byte_align = AMDGPU_GPU_PAGE_SIZE; bp->domain = AMDGPU_GEM_DOMAIN_VRAM; - if (bp->size <= PAGE_SIZE && adev->asic_type >= CHIP_VEGA10 && - adev->flags & AMD_IS_APU) - bp->domain |= AMDGPU_GEM_DOMAIN_GTT; bp->domain = amdgpu_bo_get_preferred_pin_domain(adev, bp->domain); bp->flags = AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS | AMDGPU_GEM_CREATE_CPU_GTT_USWC; @@ -1506,20 +1524,27 @@ error: } /** - * amdgpu_vm_update_huge - figure out parameters for PTE updates + * amdgpu_vm_update_flags - figure out flags for PTE updates * * Make sure to set the right flags for the PTEs at the desired level. */ -static void amdgpu_vm_update_huge(struct amdgpu_pte_update_params *params, - struct amdgpu_bo *bo, unsigned level, - uint64_t pe, uint64_t addr, - unsigned count, uint32_t incr, - uint64_t flags) +static void amdgpu_vm_update_flags(struct amdgpu_pte_update_params *params, + struct amdgpu_bo *bo, unsigned level, + uint64_t pe, uint64_t addr, + unsigned count, uint32_t incr, + uint64_t flags) { if (level != AMDGPU_VM_PTB) { flags |= AMDGPU_PDE_PTE; amdgpu_gmc_get_vm_pde(params->adev, level, &addr, &flags); + + } else if (params->adev->asic_type >= CHIP_VEGA10 && + !(flags & AMDGPU_PTE_VALID) && + !(flags & AMDGPU_PTE_PRT)) { + + /* Workaround for fault priority problem on GMC9 */ + flags |= AMDGPU_PTE_EXECUTABLE; } amdgpu_vm_update_func(params, bo, pe, addr, count, incr, flags); @@ -1676,9 +1701,9 @@ static int amdgpu_vm_update_ptes(struct amdgpu_pte_update_params *params, uint64_t upd_end = min(entry_end, frag_end); unsigned nptes = (upd_end - frag_start) >> shift; - amdgpu_vm_update_huge(params, pt, cursor.level, - pe_start, dst, nptes, incr, - flags | AMDGPU_PTE_FRAG(frag)); + amdgpu_vm_update_flags(params, pt, cursor.level, + pe_start, dst, nptes, incr, + flags | AMDGPU_PTE_FRAG(frag)); pe_start += nptes * 8; dst += (uint64_t)nptes * AMDGPU_GPU_PAGE_SIZE << shift; @@ -1756,13 +1781,20 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, if (pages_addr) params.src = ~0; - /* Wait for PT BOs to be free. PTs share the same resv. object + /* Wait for PT BOs to be idle. PTs share the same resv. object * as the root PD BO */ r = amdgpu_vm_wait_pd(adev, vm, owner); if (unlikely(r)) return r; + /* Wait for any BO move to be completed */ + if (exclusive) { + r = dma_fence_wait(exclusive, true); + if (unlikely(r)) + return r; + } + params.func = amdgpu_vm_cpu_set_ptes; params.pages_addr = pages_addr; return amdgpu_vm_update_ptes(¶ms, start, last + 1, @@ -1776,13 +1808,12 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, /* * reserve space for two commands every (1 << BLOCK_SIZE) * entries or 2k dwords (whatever is smaller) - * - * The second command is for the shadow pagetables. */ + ncmds = ((nptes >> min(adev->vm_manager.block_size, 11u)) + 1); + + /* The second command is for the shadow pagetables. */ if (vm->root.base.bo->shadow) - ncmds = ((nptes >> min(adev->vm_manager.block_size, 11u)) + 1) * 2; - else - ncmds = ((nptes >> min(adev->vm_manager.block_size, 11u)) + 1); + ncmds *= 2; /* padding, etc. */ ndw = 64; @@ -1801,10 +1832,11 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, ndw += ncmds * 10; /* extra commands for begin/end fragments */ + ncmds = 2 * adev->vm_manager.fragment_size; if (vm->root.base.bo->shadow) - ndw += 2 * 10 * adev->vm_manager.fragment_size * 2; - else - ndw += 2 * 10 * adev->vm_manager.fragment_size; + ncmds *= 2; + + ndw += 10 * ncmds; params.func = amdgpu_vm_do_set_ptes; } @@ -3006,7 +3038,7 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, } DRM_DEBUG_DRIVER("VM update mode is %s\n", vm->use_cpu_for_update ? "CPU" : "SDMA"); - WARN_ONCE((vm->use_cpu_for_update & !amdgpu_gmc_vram_full_visible(&adev->gmc)), + WARN_ONCE((vm->use_cpu_for_update && !amdgpu_gmc_vram_full_visible(&adev->gmc)), "CPU update of VM recommended only for large BAR system\n"); vm->last_update = NULL; @@ -3136,7 +3168,7 @@ int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm, uns vm->pte_support_ats = pte_support_ats; DRM_DEBUG_DRIVER("VM update mode is %s\n", vm->use_cpu_for_update ? "CPU" : "SDMA"); - WARN_ONCE((vm->use_cpu_for_update & !amdgpu_gmc_vram_full_visible(&adev->gmc)), + WARN_ONCE((vm->use_cpu_for_update && !amdgpu_gmc_vram_full_visible(&adev->gmc)), "CPU update of VM recommended only for large BAR system\n"); if (vm->pasid) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h index e8dcfd59fc93..81ff8177f092 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h @@ -363,4 +363,6 @@ int amdgpu_vm_add_fault(struct amdgpu_retryfault_hashtable *fault_hash, u64 key) void amdgpu_vm_clear_fault(struct amdgpu_retryfault_hashtable *fault_hash, u64 key); +void amdgpu_vm_del_from_lru_notify(struct ttm_buffer_object *bo); + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c index 8a8bc60cb6b4..407dd16cc35c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c @@ -40,26 +40,40 @@ void *amdgpu_xgmi_hive_try_lock(struct amdgpu_hive_info *hive) return &hive->device_list; } -struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev) +struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev, int lock) { int i; struct amdgpu_hive_info *tmp; if (!adev->gmc.xgmi.hive_id) return NULL; + + mutex_lock(&xgmi_mutex); + for (i = 0 ; i < hive_count; ++i) { tmp = &xgmi_hives[i]; - if (tmp->hive_id == adev->gmc.xgmi.hive_id) + if (tmp->hive_id == adev->gmc.xgmi.hive_id) { + if (lock) + mutex_lock(&tmp->hive_lock); + mutex_unlock(&xgmi_mutex); return tmp; + } } - if (i >= AMDGPU_MAX_XGMI_HIVE) + if (i >= AMDGPU_MAX_XGMI_HIVE) { + mutex_unlock(&xgmi_mutex); return NULL; + } /* initialize new hive if not exist */ tmp = &xgmi_hives[hive_count++]; tmp->hive_id = adev->gmc.xgmi.hive_id; INIT_LIST_HEAD(&tmp->device_list); mutex_init(&tmp->hive_lock); + mutex_init(&tmp->reset_lock); + if (lock) + mutex_lock(&tmp->hive_lock); + + mutex_unlock(&xgmi_mutex); return tmp; } @@ -77,10 +91,6 @@ int amdgpu_xgmi_update_topology(struct amdgpu_hive_info *hive, struct amdgpu_dev "XGMI: Set topology failure on device %llx, hive %llx, ret %d", adev->gmc.xgmi.node_id, adev->gmc.xgmi.hive_id, ret); - else - dev_info(adev->dev, "XGMI: Set topology for node %d, hive 0x%llx.\n", - adev->gmc.xgmi.physical_node_id, - adev->gmc.xgmi.hive_id); return ret; } @@ -111,10 +121,14 @@ int amdgpu_xgmi_add_device(struct amdgpu_device *adev) return ret; } - mutex_lock(&xgmi_mutex); - hive = amdgpu_get_xgmi_hive(adev); - if (!hive) + hive = amdgpu_get_xgmi_hive(adev, 1); + if (!hive) { + ret = -EINVAL; + dev_err(adev->dev, + "XGMI: node 0x%llx, can not match hive 0x%llx in the hive list.\n", + adev->gmc.xgmi.node_id, adev->gmc.xgmi.hive_id); goto exit; + } hive_topology = &hive->topology_info; @@ -142,8 +156,11 @@ int amdgpu_xgmi_add_device(struct amdgpu_device *adev) break; } + dev_info(adev->dev, "XGMI: Add node %d, hive 0x%llx.\n", + adev->gmc.xgmi.physical_node_id, adev->gmc.xgmi.hive_id); + + mutex_unlock(&hive->hive_lock); exit: - mutex_unlock(&xgmi_mutex); return ret; } @@ -154,15 +171,14 @@ void amdgpu_xgmi_remove_device(struct amdgpu_device *adev) if (!adev->gmc.xgmi.supported) return; - mutex_lock(&xgmi_mutex); - - hive = amdgpu_get_xgmi_hive(adev); + hive = amdgpu_get_xgmi_hive(adev, 1); if (!hive) - goto exit; + return; - if (!(hive->number_devices--)) + if (!(hive->number_devices--)) { mutex_destroy(&hive->hive_lock); - -exit: - mutex_unlock(&xgmi_mutex); + mutex_destroy(&hive->reset_lock); + } else { + mutex_unlock(&hive->hive_lock); + } } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h index 6151eb9c8ad3..14bc60664159 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.h @@ -29,10 +29,11 @@ struct amdgpu_hive_info { struct list_head device_list; struct psp_xgmi_topology_info topology_info; int number_devices; - struct mutex hive_lock; + struct mutex hive_lock, + reset_lock; }; -struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev); +struct amdgpu_hive_info *amdgpu_get_xgmi_hive(struct amdgpu_device *adev, int lock); int amdgpu_xgmi_update_topology(struct amdgpu_hive_info *hive, struct amdgpu_device *adev); int amdgpu_xgmi_add_device(struct amdgpu_device *adev); void amdgpu_xgmi_remove_device(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c deleted file mode 100644 index 86e14c754dd4..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c +++ /dev/null @@ -1,6844 +0,0 @@ -/* - * Copyright 2013 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - */ - -#include <linux/firmware.h> -#include <drm/drmP.h> -#include "amdgpu.h" -#include "amdgpu_pm.h" -#include "amdgpu_ucode.h" -#include "cikd.h" -#include "amdgpu_dpm.h" -#include "ci_dpm.h" -#include "gfx_v7_0.h" -#include "atom.h" -#include "amd_pcie.h" -#include <linux/seq_file.h> - -#include "smu/smu_7_0_1_d.h" -#include "smu/smu_7_0_1_sh_mask.h" - -#include "dce/dce_8_0_d.h" -#include "dce/dce_8_0_sh_mask.h" - -#include "bif/bif_4_1_d.h" -#include "bif/bif_4_1_sh_mask.h" - -#include "gca/gfx_7_2_d.h" -#include "gca/gfx_7_2_sh_mask.h" - -#include "gmc/gmc_7_1_d.h" -#include "gmc/gmc_7_1_sh_mask.h" - -MODULE_FIRMWARE("amdgpu/bonaire_smc.bin"); -MODULE_FIRMWARE("amdgpu/bonaire_k_smc.bin"); -MODULE_FIRMWARE("amdgpu/hawaii_smc.bin"); -MODULE_FIRMWARE("amdgpu/hawaii_k_smc.bin"); - -#define MC_CG_ARB_FREQ_F0 0x0a -#define MC_CG_ARB_FREQ_F1 0x0b -#define MC_CG_ARB_FREQ_F2 0x0c -#define MC_CG_ARB_FREQ_F3 0x0d - -#define SMC_RAM_END 0x40000 - -#define VOLTAGE_SCALE 4 -#define VOLTAGE_VID_OFFSET_SCALE1 625 -#define VOLTAGE_VID_OFFSET_SCALE2 100 - -static const struct amd_pm_funcs ci_dpm_funcs; - -static const struct ci_pt_defaults defaults_hawaii_xt = -{ - 1, 0xF, 0xFD, 0x19, 5, 0x14, 0, 0xB0000, - { 0x2E, 0x00, 0x00, 0x88, 0x00, 0x00, 0x72, 0x60, 0x51, 0xA7, 0x79, 0x6B, 0x90, 0xBD, 0x79 }, - { 0x217, 0x217, 0x217, 0x242, 0x242, 0x242, 0x269, 0x269, 0x269, 0x2A1, 0x2A1, 0x2A1, 0x2C9, 0x2C9, 0x2C9 } -}; - -static const struct ci_pt_defaults defaults_hawaii_pro = -{ - 1, 0xF, 0xFD, 0x19, 5, 0x14, 0, 0x65062, - { 0x2E, 0x00, 0x00, 0x88, 0x00, 0x00, 0x72, 0x60, 0x51, 0xA7, 0x79, 0x6B, 0x90, 0xBD, 0x79 }, - { 0x217, 0x217, 0x217, 0x242, 0x242, 0x242, 0x269, 0x269, 0x269, 0x2A1, 0x2A1, 0x2A1, 0x2C9, 0x2C9, 0x2C9 } -}; - -static const struct ci_pt_defaults defaults_bonaire_xt = -{ - 1, 0xF, 0xFD, 0x19, 5, 45, 0, 0xB0000, - { 0x79, 0x253, 0x25D, 0xAE, 0x72, 0x80, 0x83, 0x86, 0x6F, 0xC8, 0xC9, 0xC9, 0x2F, 0x4D, 0x61 }, - { 0x17C, 0x172, 0x180, 0x1BC, 0x1B3, 0x1BD, 0x206, 0x200, 0x203, 0x25D, 0x25A, 0x255, 0x2C3, 0x2C5, 0x2B4 } -}; - -#if 0 -static const struct ci_pt_defaults defaults_bonaire_pro = -{ - 1, 0xF, 0xFD, 0x19, 5, 45, 0, 0x65062, - { 0x8C, 0x23F, 0x244, 0xA6, 0x83, 0x85, 0x86, 0x86, 0x83, 0xDB, 0xDB, 0xDA, 0x67, 0x60, 0x5F }, - { 0x187, 0x193, 0x193, 0x1C7, 0x1D1, 0x1D1, 0x210, 0x219, 0x219, 0x266, 0x26C, 0x26C, 0x2C9, 0x2CB, 0x2CB } -}; -#endif - -static const struct ci_pt_defaults defaults_saturn_xt = -{ - 1, 0xF, 0xFD, 0x19, 5, 55, 0, 0x70000, - { 0x8C, 0x247, 0x249, 0xA6, 0x80, 0x81, 0x8B, 0x89, 0x86, 0xC9, 0xCA, 0xC9, 0x4D, 0x4D, 0x4D }, - { 0x187, 0x187, 0x187, 0x1C7, 0x1C7, 0x1C7, 0x210, 0x210, 0x210, 0x266, 0x266, 0x266, 0x2C9, 0x2C9, 0x2C9 } -}; - -#if 0 -static const struct ci_pt_defaults defaults_saturn_pro = -{ - 1, 0xF, 0xFD, 0x19, 5, 55, 0, 0x30000, - { 0x96, 0x21D, 0x23B, 0xA1, 0x85, 0x87, 0x83, 0x84, 0x81, 0xE6, 0xE6, 0xE6, 0x71, 0x6A, 0x6A }, - { 0x193, 0x19E, 0x19E, 0x1D2, 0x1DC, 0x1DC, 0x21A, 0x223, 0x223, 0x26E, 0x27E, 0x274, 0x2CF, 0x2D2, 0x2D2 } -}; -#endif - -static const struct ci_pt_config_reg didt_config_ci[] = -{ - { 0x10, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x10, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x10, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x10, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x11, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x11, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x11, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x11, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x12, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x12, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x12, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x12, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x2, 0x00003fff, 0, 0x4, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x2, 0x03ff0000, 16, 0x80, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x2, 0x78000000, 27, 0x3, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x1, 0x0000ffff, 0, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x1, 0xffff0000, 16, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x0, 0x00000001, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x30, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x30, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x30, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x30, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x31, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x31, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x31, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x31, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x32, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x32, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x32, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x32, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x22, 0x00003fff, 0, 0x4, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x22, 0x03ff0000, 16, 0x80, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x22, 0x78000000, 27, 0x3, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x21, 0x0000ffff, 0, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x21, 0xffff0000, 16, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x20, 0x00000001, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x50, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x50, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x50, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x50, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x51, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x51, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x51, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x51, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x52, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x52, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x52, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x52, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x42, 0x00003fff, 0, 0x4, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x42, 0x03ff0000, 16, 0x80, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x42, 0x78000000, 27, 0x3, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x41, 0x0000ffff, 0, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x41, 0xffff0000, 16, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x40, 0x00000001, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x70, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x70, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x70, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x70, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x71, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x71, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x71, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x71, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x72, 0x000000ff, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x72, 0x0000ff00, 8, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x72, 0x00ff0000, 16, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x72, 0xff000000, 24, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x62, 0x00003fff, 0, 0x4, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x62, 0x03ff0000, 16, 0x80, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x62, 0x78000000, 27, 0x3, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x61, 0x0000ffff, 0, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x61, 0xffff0000, 16, 0x3FFF, CISLANDS_CONFIGREG_DIDT_IND }, - { 0x60, 0x00000001, 0, 0x0, CISLANDS_CONFIGREG_DIDT_IND }, - { 0xFFFFFFFF } -}; - -static u8 ci_get_memory_module_index(struct amdgpu_device *adev) -{ - return (u8) ((RREG32(mmBIOS_SCRATCH_4) >> 16) & 0xff); -} - -#define MC_CG_ARB_FREQ_F0 0x0a -#define MC_CG_ARB_FREQ_F1 0x0b -#define MC_CG_ARB_FREQ_F2 0x0c -#define MC_CG_ARB_FREQ_F3 0x0d - -static int ci_copy_and_switch_arb_sets(struct amdgpu_device *adev, - u32 arb_freq_src, u32 arb_freq_dest) -{ - u32 mc_arb_dram_timing; - u32 mc_arb_dram_timing2; - u32 burst_time; - u32 mc_cg_config; - - switch (arb_freq_src) { - case MC_CG_ARB_FREQ_F0: - mc_arb_dram_timing = RREG32(mmMC_ARB_DRAM_TIMING); - mc_arb_dram_timing2 = RREG32(mmMC_ARB_DRAM_TIMING2); - burst_time = (RREG32(mmMC_ARB_BURST_TIME) & MC_ARB_BURST_TIME__STATE0_MASK) >> - MC_ARB_BURST_TIME__STATE0__SHIFT; - break; - case MC_CG_ARB_FREQ_F1: - mc_arb_dram_timing = RREG32(mmMC_ARB_DRAM_TIMING_1); - mc_arb_dram_timing2 = RREG32(mmMC_ARB_DRAM_TIMING2_1); - burst_time = (RREG32(mmMC_ARB_BURST_TIME) & MC_ARB_BURST_TIME__STATE1_MASK) >> - MC_ARB_BURST_TIME__STATE1__SHIFT; - break; - default: - return -EINVAL; - } - - switch (arb_freq_dest) { - case MC_CG_ARB_FREQ_F0: - WREG32(mmMC_ARB_DRAM_TIMING, mc_arb_dram_timing); - WREG32(mmMC_ARB_DRAM_TIMING2, mc_arb_dram_timing2); - WREG32_P(mmMC_ARB_BURST_TIME, (burst_time << MC_ARB_BURST_TIME__STATE0__SHIFT), - ~MC_ARB_BURST_TIME__STATE0_MASK); - break; - case MC_CG_ARB_FREQ_F1: - WREG32(mmMC_ARB_DRAM_TIMING_1, mc_arb_dram_timing); - WREG32(mmMC_ARB_DRAM_TIMING2_1, mc_arb_dram_timing2); - WREG32_P(mmMC_ARB_BURST_TIME, (burst_time << MC_ARB_BURST_TIME__STATE1__SHIFT), - ~MC_ARB_BURST_TIME__STATE1_MASK); - break; - default: - return -EINVAL; - } - - mc_cg_config = RREG32(mmMC_CG_CONFIG) | 0x0000000F; - WREG32(mmMC_CG_CONFIG, mc_cg_config); - WREG32_P(mmMC_ARB_CG, (arb_freq_dest) << MC_ARB_CG__CG_ARB_REQ__SHIFT, - ~MC_ARB_CG__CG_ARB_REQ_MASK); - - return 0; -} - -static u8 ci_get_ddr3_mclk_frequency_ratio(u32 memory_clock) -{ - u8 mc_para_index; - - if (memory_clock < 10000) - mc_para_index = 0; - else if (memory_clock >= 80000) - mc_para_index = 0x0f; - else - mc_para_index = (u8)((memory_clock - 10000) / 5000 + 1); - return mc_para_index; -} - -static u8 ci_get_mclk_frequency_ratio(u32 memory_clock, bool strobe_mode) -{ - u8 mc_para_index; - - if (strobe_mode) { - if (memory_clock < 12500) - mc_para_index = 0x00; - else if (memory_clock > 47500) - mc_para_index = 0x0f; - else - mc_para_index = (u8)((memory_clock - 10000) / 2500); - } else { - if (memory_clock < 65000) - mc_para_index = 0x00; - else if (memory_clock > 135000) - mc_para_index = 0x0f; - else - mc_para_index = (u8)((memory_clock - 60000) / 5000); - } - return mc_para_index; -} - -static void ci_trim_voltage_table_to_fit_state_table(struct amdgpu_device *adev, - u32 max_voltage_steps, - struct atom_voltage_table *voltage_table) -{ - unsigned int i, diff; - - if (voltage_table->count <= max_voltage_steps) - return; - - diff = voltage_table->count - max_voltage_steps; - - for (i = 0; i < max_voltage_steps; i++) - voltage_table->entries[i] = voltage_table->entries[i + diff]; - - voltage_table->count = max_voltage_steps; -} - -static int ci_get_std_voltage_value_sidd(struct amdgpu_device *adev, - struct atom_voltage_table_entry *voltage_table, - u16 *std_voltage_hi_sidd, u16 *std_voltage_lo_sidd); -static int ci_set_power_limit(struct amdgpu_device *adev, u32 n); -static int ci_set_overdrive_target_tdp(struct amdgpu_device *adev, - u32 target_tdp); -static int ci_update_uvd_dpm(struct amdgpu_device *adev, bool gate); -static void ci_dpm_set_irq_funcs(struct amdgpu_device *adev); - -static PPSMC_Result amdgpu_ci_send_msg_to_smc_with_parameter(struct amdgpu_device *adev, - PPSMC_Msg msg, u32 parameter); -static void ci_thermal_start_smc_fan_control(struct amdgpu_device *adev); -static void ci_fan_ctrl_set_default_mode(struct amdgpu_device *adev); - -static struct ci_power_info *ci_get_pi(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = adev->pm.dpm.priv; - - return pi; -} - -static struct ci_ps *ci_get_ps(struct amdgpu_ps *rps) -{ - struct ci_ps *ps = rps->ps_priv; - - return ps; -} - -static void ci_initialize_powertune_defaults(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - switch (adev->pdev->device) { - case 0x6649: - case 0x6650: - case 0x6651: - case 0x6658: - case 0x665C: - case 0x665D: - default: - pi->powertune_defaults = &defaults_bonaire_xt; - break; - case 0x6640: - case 0x6641: - case 0x6646: - case 0x6647: - pi->powertune_defaults = &defaults_saturn_xt; - break; - case 0x67B8: - case 0x67B0: - pi->powertune_defaults = &defaults_hawaii_xt; - break; - case 0x67BA: - case 0x67B1: - pi->powertune_defaults = &defaults_hawaii_pro; - break; - case 0x67A0: - case 0x67A1: - case 0x67A2: - case 0x67A8: - case 0x67A9: - case 0x67AA: - case 0x67B9: - case 0x67BE: - pi->powertune_defaults = &defaults_bonaire_xt; - break; - } - - pi->dte_tj_offset = 0; - - pi->caps_power_containment = true; - pi->caps_cac = false; - pi->caps_sq_ramping = false; - pi->caps_db_ramping = false; - pi->caps_td_ramping = false; - pi->caps_tcp_ramping = false; - - if (pi->caps_power_containment) { - pi->caps_cac = true; - if (adev->asic_type == CHIP_HAWAII) - pi->enable_bapm_feature = false; - else - pi->enable_bapm_feature = true; - pi->enable_tdc_limit_feature = true; - pi->enable_pkg_pwr_tracking_feature = true; - } -} - -static u8 ci_convert_to_vid(u16 vddc) -{ - return (6200 - (vddc * VOLTAGE_SCALE)) / 25; -} - -static int ci_populate_bapm_vddc_vid_sidd(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u8 *hi_vid = pi->smc_powertune_table.BapmVddCVidHiSidd; - u8 *lo_vid = pi->smc_powertune_table.BapmVddCVidLoSidd; - u8 *hi2_vid = pi->smc_powertune_table.BapmVddCVidHiSidd2; - u32 i; - - if (adev->pm.dpm.dyn_state.cac_leakage_table.entries == NULL) - return -EINVAL; - if (adev->pm.dpm.dyn_state.cac_leakage_table.count > 8) - return -EINVAL; - if (adev->pm.dpm.dyn_state.cac_leakage_table.count != - adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count) - return -EINVAL; - - for (i = 0; i < adev->pm.dpm.dyn_state.cac_leakage_table.count; i++) { - if (adev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_EVV) { - lo_vid[i] = ci_convert_to_vid(adev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc1); - hi_vid[i] = ci_convert_to_vid(adev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc2); - hi2_vid[i] = ci_convert_to_vid(adev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc3); - } else { - lo_vid[i] = ci_convert_to_vid(adev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc); - hi_vid[i] = ci_convert_to_vid((u16)adev->pm.dpm.dyn_state.cac_leakage_table.entries[i].leakage); - } - } - return 0; -} - -static int ci_populate_vddc_vid(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u8 *vid = pi->smc_powertune_table.VddCVid; - u32 i; - - if (pi->vddc_voltage_table.count > 8) - return -EINVAL; - - for (i = 0; i < pi->vddc_voltage_table.count; i++) - vid[i] = ci_convert_to_vid(pi->vddc_voltage_table.entries[i].value); - - return 0; -} - -static int ci_populate_svi_load_line(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - const struct ci_pt_defaults *pt_defaults = pi->powertune_defaults; - - pi->smc_powertune_table.SviLoadLineEn = pt_defaults->svi_load_line_en; - pi->smc_powertune_table.SviLoadLineVddC = pt_defaults->svi_load_line_vddc; - pi->smc_powertune_table.SviLoadLineTrimVddC = 3; - pi->smc_powertune_table.SviLoadLineOffsetVddC = 0; - - return 0; -} - -static int ci_populate_tdc_limit(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - const struct ci_pt_defaults *pt_defaults = pi->powertune_defaults; - u16 tdc_limit; - - tdc_limit = adev->pm.dpm.dyn_state.cac_tdp_table->tdc * 256; - pi->smc_powertune_table.TDC_VDDC_PkgLimit = cpu_to_be16(tdc_limit); - pi->smc_powertune_table.TDC_VDDC_ThrottleReleaseLimitPerc = - pt_defaults->tdc_vddc_throttle_release_limit_perc; - pi->smc_powertune_table.TDC_MAWt = pt_defaults->tdc_mawt; - - return 0; -} - -static int ci_populate_dw8(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - const struct ci_pt_defaults *pt_defaults = pi->powertune_defaults; - int ret; - - ret = amdgpu_ci_read_smc_sram_dword(adev, - SMU7_FIRMWARE_HEADER_LOCATION + - offsetof(SMU7_Firmware_Header, PmFuseTable) + - offsetof(SMU7_Discrete_PmFuses, TdcWaterfallCtl), - (u32 *)&pi->smc_powertune_table.TdcWaterfallCtl, - pi->sram_end); - if (ret) - return -EINVAL; - else - pi->smc_powertune_table.TdcWaterfallCtl = pt_defaults->tdc_waterfall_ctl; - - return 0; -} - -static int ci_populate_fuzzy_fan(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - if ((adev->pm.dpm.fan.fan_output_sensitivity & (1 << 15)) || - (adev->pm.dpm.fan.fan_output_sensitivity == 0)) - adev->pm.dpm.fan.fan_output_sensitivity = - adev->pm.dpm.fan.default_fan_output_sensitivity; - - pi->smc_powertune_table.FuzzyFan_PwmSetDelta = - cpu_to_be16(adev->pm.dpm.fan.fan_output_sensitivity); - - return 0; -} - -static int ci_min_max_v_gnbl_pm_lid_from_bapm_vddc(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u8 *hi_vid = pi->smc_powertune_table.BapmVddCVidHiSidd; - u8 *lo_vid = pi->smc_powertune_table.BapmVddCVidLoSidd; - int i, min, max; - - min = max = hi_vid[0]; - for (i = 0; i < 8; i++) { - if (0 != hi_vid[i]) { - if (min > hi_vid[i]) - min = hi_vid[i]; - if (max < hi_vid[i]) - max = hi_vid[i]; - } - - if (0 != lo_vid[i]) { - if (min > lo_vid[i]) - min = lo_vid[i]; - if (max < lo_vid[i]) - max = lo_vid[i]; - } - } - - if ((min == 0) || (max == 0)) - return -EINVAL; - pi->smc_powertune_table.GnbLPMLMaxVid = (u8)max; - pi->smc_powertune_table.GnbLPMLMinVid = (u8)min; - - return 0; -} - -static int ci_populate_bapm_vddc_base_leakage_sidd(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u16 hi_sidd = pi->smc_powertune_table.BapmVddCBaseLeakageHiSidd; - u16 lo_sidd = pi->smc_powertune_table.BapmVddCBaseLeakageLoSidd; - struct amdgpu_cac_tdp_table *cac_tdp_table = - adev->pm.dpm.dyn_state.cac_tdp_table; - - hi_sidd = cac_tdp_table->high_cac_leakage / 100 * 256; - lo_sidd = cac_tdp_table->low_cac_leakage / 100 * 256; - - pi->smc_powertune_table.BapmVddCBaseLeakageHiSidd = cpu_to_be16(hi_sidd); - pi->smc_powertune_table.BapmVddCBaseLeakageLoSidd = cpu_to_be16(lo_sidd); - - return 0; -} - -static int ci_populate_bapm_parameters_in_dpm_table(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - const struct ci_pt_defaults *pt_defaults = pi->powertune_defaults; - SMU7_Discrete_DpmTable *dpm_table = &pi->smc_state_table; - struct amdgpu_cac_tdp_table *cac_tdp_table = - adev->pm.dpm.dyn_state.cac_tdp_table; - struct amdgpu_ppm_table *ppm = adev->pm.dpm.dyn_state.ppm_table; - int i, j, k; - const u16 *def1; - const u16 *def2; - - dpm_table->DefaultTdp = cac_tdp_table->tdp * 256; - dpm_table->TargetTdp = cac_tdp_table->configurable_tdp * 256; - - dpm_table->DTETjOffset = (u8)pi->dte_tj_offset; - dpm_table->GpuTjMax = - (u8)(pi->thermal_temp_setting.temperature_high / 1000); - dpm_table->GpuTjHyst = 8; - - dpm_table->DTEAmbientTempBase = pt_defaults->dte_ambient_temp_base; - - if (ppm) { - dpm_table->PPM_PkgPwrLimit = cpu_to_be16((u16)ppm->dgpu_tdp * 256 / 1000); - dpm_table->PPM_TemperatureLimit = cpu_to_be16((u16)ppm->tj_max * 256); - } else { - dpm_table->PPM_PkgPwrLimit = cpu_to_be16(0); - dpm_table->PPM_TemperatureLimit = cpu_to_be16(0); - } - - dpm_table->BAPM_TEMP_GRADIENT = cpu_to_be32(pt_defaults->bapm_temp_gradient); - def1 = pt_defaults->bapmti_r; - def2 = pt_defaults->bapmti_rc; - - for (i = 0; i < SMU7_DTE_ITERATIONS; i++) { - for (j = 0; j < SMU7_DTE_SOURCES; j++) { - for (k = 0; k < SMU7_DTE_SINKS; k++) { - dpm_table->BAPMTI_R[i][j][k] = cpu_to_be16(*def1); - dpm_table->BAPMTI_RC[i][j][k] = cpu_to_be16(*def2); - def1++; - def2++; - } - } - } - - return 0; -} - -static int ci_populate_pm_base(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 pm_fuse_table_offset; - int ret; - - if (pi->caps_power_containment) { - ret = amdgpu_ci_read_smc_sram_dword(adev, - SMU7_FIRMWARE_HEADER_LOCATION + - offsetof(SMU7_Firmware_Header, PmFuseTable), - &pm_fuse_table_offset, pi->sram_end); - if (ret) - return ret; - ret = ci_populate_bapm_vddc_vid_sidd(adev); - if (ret) - return ret; - ret = ci_populate_vddc_vid(adev); - if (ret) - return ret; - ret = ci_populate_svi_load_line(adev); - if (ret) - return ret; - ret = ci_populate_tdc_limit(adev); - if (ret) - return ret; - ret = ci_populate_dw8(adev); - if (ret) - return ret; - ret = ci_populate_fuzzy_fan(adev); - if (ret) - return ret; - ret = ci_min_max_v_gnbl_pm_lid_from_bapm_vddc(adev); - if (ret) - return ret; - ret = ci_populate_bapm_vddc_base_leakage_sidd(adev); - if (ret) - return ret; - ret = amdgpu_ci_copy_bytes_to_smc(adev, pm_fuse_table_offset, - (u8 *)&pi->smc_powertune_table, - sizeof(SMU7_Discrete_PmFuses), pi->sram_end); - if (ret) - return ret; - } - - return 0; -} - -static void ci_do_enable_didt(struct amdgpu_device *adev, const bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 data; - - if (pi->caps_sq_ramping) { - data = RREG32_DIDT(ixDIDT_SQ_CTRL0); - if (enable) - data |= DIDT_SQ_CTRL0__DIDT_CTRL_EN_MASK; - else - data &= ~DIDT_SQ_CTRL0__DIDT_CTRL_EN_MASK; - WREG32_DIDT(ixDIDT_SQ_CTRL0, data); - } - - if (pi->caps_db_ramping) { - data = RREG32_DIDT(ixDIDT_DB_CTRL0); - if (enable) - data |= DIDT_DB_CTRL0__DIDT_CTRL_EN_MASK; - else - data &= ~DIDT_DB_CTRL0__DIDT_CTRL_EN_MASK; - WREG32_DIDT(ixDIDT_DB_CTRL0, data); - } - - if (pi->caps_td_ramping) { - data = RREG32_DIDT(ixDIDT_TD_CTRL0); - if (enable) - data |= DIDT_TD_CTRL0__DIDT_CTRL_EN_MASK; - else - data &= ~DIDT_TD_CTRL0__DIDT_CTRL_EN_MASK; - WREG32_DIDT(ixDIDT_TD_CTRL0, data); - } - - if (pi->caps_tcp_ramping) { - data = RREG32_DIDT(ixDIDT_TCP_CTRL0); - if (enable) - data |= DIDT_TCP_CTRL0__DIDT_CTRL_EN_MASK; - else - data &= ~DIDT_TCP_CTRL0__DIDT_CTRL_EN_MASK; - WREG32_DIDT(ixDIDT_TCP_CTRL0, data); - } -} - -static int ci_program_pt_config_registers(struct amdgpu_device *adev, - const struct ci_pt_config_reg *cac_config_regs) -{ - const struct ci_pt_config_reg *config_regs = cac_config_regs; - u32 data; - u32 cache = 0; - - if (config_regs == NULL) - return -EINVAL; - - while (config_regs->offset != 0xFFFFFFFF) { - if (config_regs->type == CISLANDS_CONFIGREG_CACHE) { - cache |= ((config_regs->value << config_regs->shift) & config_regs->mask); - } else { - switch (config_regs->type) { - case CISLANDS_CONFIGREG_SMC_IND: - data = RREG32_SMC(config_regs->offset); - break; - case CISLANDS_CONFIGREG_DIDT_IND: - data = RREG32_DIDT(config_regs->offset); - break; - default: - data = RREG32(config_regs->offset); - break; - } - - data &= ~config_regs->mask; - data |= ((config_regs->value << config_regs->shift) & config_regs->mask); - data |= cache; - - switch (config_regs->type) { - case CISLANDS_CONFIGREG_SMC_IND: - WREG32_SMC(config_regs->offset, data); - break; - case CISLANDS_CONFIGREG_DIDT_IND: - WREG32_DIDT(config_regs->offset, data); - break; - default: - WREG32(config_regs->offset, data); - break; - } - cache = 0; - } - config_regs++; - } - return 0; -} - -static int ci_enable_didt(struct amdgpu_device *adev, bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - int ret; - - if (pi->caps_sq_ramping || pi->caps_db_ramping || - pi->caps_td_ramping || pi->caps_tcp_ramping) { - amdgpu_gfx_rlc_enter_safe_mode(adev); - - if (enable) { - ret = ci_program_pt_config_registers(adev, didt_config_ci); - if (ret) { - amdgpu_gfx_rlc_exit_safe_mode(adev); - return ret; - } - } - - ci_do_enable_didt(adev, enable); - - amdgpu_gfx_rlc_exit_safe_mode(adev); - } - - return 0; -} - -static int ci_enable_power_containment(struct amdgpu_device *adev, bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - PPSMC_Result smc_result; - int ret = 0; - - if (enable) { - pi->power_containment_features = 0; - if (pi->caps_power_containment) { - if (pi->enable_bapm_feature) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_EnableDTE); - if (smc_result != PPSMC_Result_OK) - ret = -EINVAL; - else - pi->power_containment_features |= POWERCONTAINMENT_FEATURE_BAPM; - } - - if (pi->enable_tdc_limit_feature) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_TDCLimitEnable); - if (smc_result != PPSMC_Result_OK) - ret = -EINVAL; - else - pi->power_containment_features |= POWERCONTAINMENT_FEATURE_TDCLimit; - } - - if (pi->enable_pkg_pwr_tracking_feature) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_PkgPwrLimitEnable); - if (smc_result != PPSMC_Result_OK) { - ret = -EINVAL; - } else { - struct amdgpu_cac_tdp_table *cac_tdp_table = - adev->pm.dpm.dyn_state.cac_tdp_table; - u32 default_pwr_limit = - (u32)(cac_tdp_table->maximum_power_delivery_limit * 256); - - pi->power_containment_features |= POWERCONTAINMENT_FEATURE_PkgPwrLimit; - - ci_set_power_limit(adev, default_pwr_limit); - } - } - } - } else { - if (pi->caps_power_containment && pi->power_containment_features) { - if (pi->power_containment_features & POWERCONTAINMENT_FEATURE_TDCLimit) - amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_TDCLimitDisable); - - if (pi->power_containment_features & POWERCONTAINMENT_FEATURE_BAPM) - amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_DisableDTE); - - if (pi->power_containment_features & POWERCONTAINMENT_FEATURE_PkgPwrLimit) - amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_PkgPwrLimitDisable); - pi->power_containment_features = 0; - } - } - - return ret; -} - -static int ci_enable_smc_cac(struct amdgpu_device *adev, bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - PPSMC_Result smc_result; - int ret = 0; - - if (pi->caps_cac) { - if (enable) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_EnableCac); - if (smc_result != PPSMC_Result_OK) { - ret = -EINVAL; - pi->cac_enabled = false; - } else { - pi->cac_enabled = true; - } - } else if (pi->cac_enabled) { - amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_DisableCac); - pi->cac_enabled = false; - } - } - - return ret; -} - -static int ci_enable_thermal_based_sclk_dpm(struct amdgpu_device *adev, - bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - PPSMC_Result smc_result = PPSMC_Result_OK; - - if (pi->thermal_sclk_dpm_enabled) { - if (enable) - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_ENABLE_THERMAL_DPM); - else - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_DISABLE_THERMAL_DPM); - } - - if (smc_result == PPSMC_Result_OK) - return 0; - else - return -EINVAL; -} - -static int ci_power_control_set_level(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct amdgpu_cac_tdp_table *cac_tdp_table = - adev->pm.dpm.dyn_state.cac_tdp_table; - s32 adjust_percent; - s32 target_tdp; - int ret = 0; - bool adjust_polarity = false; /* ??? */ - - if (pi->caps_power_containment) { - adjust_percent = adjust_polarity ? - adev->pm.dpm.tdp_adjustment : (-1 * adev->pm.dpm.tdp_adjustment); - target_tdp = ((100 + adjust_percent) * - (s32)cac_tdp_table->configurable_tdp) / 100; - - ret = ci_set_overdrive_target_tdp(adev, (u32)target_tdp); - } - - return ret; -} - -static void ci_dpm_powergate_uvd(void *handle, bool gate) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - - pi->uvd_power_gated = gate; - - if (gate) { - /* stop the UVD block */ - amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD, - AMD_PG_STATE_GATE); - ci_update_uvd_dpm(adev, gate); - } else { - amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD, - AMD_PG_STATE_UNGATE); - ci_update_uvd_dpm(adev, gate); - } -} - -static bool ci_dpm_vblank_too_short(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - u32 vblank_time = amdgpu_dpm_get_vblank_time(adev); - u32 switch_limit = adev->gmc.vram_type == AMDGPU_VRAM_TYPE_GDDR5 ? 450 : 300; - - /* disable mclk switching if the refresh is >120Hz, even if the - * blanking period would allow it - */ - if (amdgpu_dpm_get_vrefresh(adev) > 120) - return true; - - if (vblank_time < switch_limit) - return true; - else - return false; - -} - -static void ci_apply_state_adjust_rules(struct amdgpu_device *adev, - struct amdgpu_ps *rps) -{ - struct ci_ps *ps = ci_get_ps(rps); - struct ci_power_info *pi = ci_get_pi(adev); - struct amdgpu_clock_and_voltage_limits *max_limits; - bool disable_mclk_switching; - u32 sclk, mclk; - int i; - - if (rps->vce_active) { - rps->evclk = adev->pm.dpm.vce_states[adev->pm.dpm.vce_level].evclk; - rps->ecclk = adev->pm.dpm.vce_states[adev->pm.dpm.vce_level].ecclk; - } else { - rps->evclk = 0; - rps->ecclk = 0; - } - - if ((adev->pm.dpm.new_active_crtc_count > 1) || - ci_dpm_vblank_too_short(adev)) - disable_mclk_switching = true; - else - disable_mclk_switching = false; - - if ((rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) == ATOM_PPLIB_CLASSIFICATION_UI_BATTERY) - pi->battery_state = true; - else - pi->battery_state = false; - - if (adev->pm.ac_power) - max_limits = &adev->pm.dpm.dyn_state.max_clock_voltage_on_ac; - else - max_limits = &adev->pm.dpm.dyn_state.max_clock_voltage_on_dc; - - if (adev->pm.ac_power == false) { - for (i = 0; i < ps->performance_level_count; i++) { - if (ps->performance_levels[i].mclk > max_limits->mclk) - ps->performance_levels[i].mclk = max_limits->mclk; - if (ps->performance_levels[i].sclk > max_limits->sclk) - ps->performance_levels[i].sclk = max_limits->sclk; - } - } - - /* XXX validate the min clocks required for display */ - - if (disable_mclk_switching) { - mclk = ps->performance_levels[ps->performance_level_count - 1].mclk; - sclk = ps->performance_levels[0].sclk; - } else { - mclk = ps->performance_levels[0].mclk; - sclk = ps->performance_levels[0].sclk; - } - - if (adev->pm.pm_display_cfg.min_core_set_clock > sclk) - sclk = adev->pm.pm_display_cfg.min_core_set_clock; - - if (adev->pm.pm_display_cfg.min_mem_set_clock > mclk) - mclk = adev->pm.pm_display_cfg.min_mem_set_clock; - - if (rps->vce_active) { - if (sclk < adev->pm.dpm.vce_states[adev->pm.dpm.vce_level].sclk) - sclk = adev->pm.dpm.vce_states[adev->pm.dpm.vce_level].sclk; - if (mclk < adev->pm.dpm.vce_states[adev->pm.dpm.vce_level].mclk) - mclk = adev->pm.dpm.vce_states[adev->pm.dpm.vce_level].mclk; - } - - ps->performance_levels[0].sclk = sclk; - ps->performance_levels[0].mclk = mclk; - - if (ps->performance_levels[1].sclk < ps->performance_levels[0].sclk) - ps->performance_levels[1].sclk = ps->performance_levels[0].sclk; - - if (disable_mclk_switching) { - if (ps->performance_levels[0].mclk < ps->performance_levels[1].mclk) - ps->performance_levels[0].mclk = ps->performance_levels[1].mclk; - } else { - if (ps->performance_levels[1].mclk < ps->performance_levels[0].mclk) - ps->performance_levels[1].mclk = ps->performance_levels[0].mclk; - } -} - -static int ci_thermal_set_temperature_range(struct amdgpu_device *adev, - int min_temp, int max_temp) -{ - int low_temp = 0 * 1000; - int high_temp = 255 * 1000; - u32 tmp; - - if (low_temp < min_temp) - low_temp = min_temp; - if (high_temp > max_temp) - high_temp = max_temp; - if (high_temp < low_temp) { - DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp); - return -EINVAL; - } - - tmp = RREG32_SMC(ixCG_THERMAL_INT); - tmp &= ~(CG_THERMAL_INT__DIG_THERM_INTH_MASK | CG_THERMAL_INT__DIG_THERM_INTL_MASK); - tmp |= ((high_temp / 1000) << CG_THERMAL_INT__DIG_THERM_INTH__SHIFT) | - ((low_temp / 1000)) << CG_THERMAL_INT__DIG_THERM_INTL__SHIFT; - WREG32_SMC(ixCG_THERMAL_INT, tmp); - -#if 0 - /* XXX: need to figure out how to handle this properly */ - tmp = RREG32_SMC(ixCG_THERMAL_CTRL); - tmp &= DIG_THERM_DPM_MASK; - tmp |= DIG_THERM_DPM(high_temp / 1000); - WREG32_SMC(ixCG_THERMAL_CTRL, tmp); -#endif - - adev->pm.dpm.thermal.min_temp = low_temp; - adev->pm.dpm.thermal.max_temp = high_temp; - return 0; -} - -static int ci_thermal_enable_alert(struct amdgpu_device *adev, - bool enable) -{ - u32 thermal_int = RREG32_SMC(ixCG_THERMAL_INT); - PPSMC_Result result; - - if (enable) { - thermal_int &= ~(CG_THERMAL_INT_CTRL__THERM_INTH_MASK_MASK | - CG_THERMAL_INT_CTRL__THERM_INTL_MASK_MASK); - WREG32_SMC(ixCG_THERMAL_INT, thermal_int); - result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_Thermal_Cntl_Enable); - if (result != PPSMC_Result_OK) { - DRM_DEBUG_KMS("Could not enable thermal interrupts.\n"); - return -EINVAL; - } - } else { - thermal_int |= CG_THERMAL_INT_CTRL__THERM_INTH_MASK_MASK | - CG_THERMAL_INT_CTRL__THERM_INTL_MASK_MASK; - WREG32_SMC(ixCG_THERMAL_INT, thermal_int); - result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_Thermal_Cntl_Disable); - if (result != PPSMC_Result_OK) { - DRM_DEBUG_KMS("Could not disable thermal interrupts.\n"); - return -EINVAL; - } - } - - return 0; -} - -static void ci_fan_ctrl_set_static_mode(struct amdgpu_device *adev, u32 mode) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 tmp; - - if (pi->fan_ctrl_is_in_default_mode) { - tmp = (RREG32_SMC(ixCG_FDO_CTRL2) & CG_FDO_CTRL2__FDO_PWM_MODE_MASK) - >> CG_FDO_CTRL2__FDO_PWM_MODE__SHIFT; - pi->fan_ctrl_default_mode = tmp; - tmp = (RREG32_SMC(ixCG_FDO_CTRL2) & CG_FDO_CTRL2__TMIN_MASK) - >> CG_FDO_CTRL2__TMIN__SHIFT; - pi->t_min = tmp; - pi->fan_ctrl_is_in_default_mode = false; - } - - tmp = RREG32_SMC(ixCG_FDO_CTRL2) & ~CG_FDO_CTRL2__TMIN_MASK; - tmp |= 0 << CG_FDO_CTRL2__TMIN__SHIFT; - WREG32_SMC(ixCG_FDO_CTRL2, tmp); - - tmp = RREG32_SMC(ixCG_FDO_CTRL2) & ~CG_FDO_CTRL2__FDO_PWM_MODE_MASK; - tmp |= mode << CG_FDO_CTRL2__FDO_PWM_MODE__SHIFT; - WREG32_SMC(ixCG_FDO_CTRL2, tmp); -} - -static int ci_thermal_setup_fan_table(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - SMU7_Discrete_FanTable fan_table = { FDO_MODE_HARDWARE }; - u32 duty100; - u32 t_diff1, t_diff2, pwm_diff1, pwm_diff2; - u16 fdo_min, slope1, slope2; - u32 reference_clock, tmp; - int ret; - u64 tmp64; - - if (!pi->fan_table_start) { - adev->pm.dpm.fan.ucode_fan_control = false; - return 0; - } - - duty100 = (RREG32_SMC(ixCG_FDO_CTRL1) & CG_FDO_CTRL1__FMAX_DUTY100_MASK) - >> CG_FDO_CTRL1__FMAX_DUTY100__SHIFT; - - if (duty100 == 0) { - adev->pm.dpm.fan.ucode_fan_control = false; - return 0; - } - - tmp64 = (u64)adev->pm.dpm.fan.pwm_min * duty100; - do_div(tmp64, 10000); - fdo_min = (u16)tmp64; - - t_diff1 = adev->pm.dpm.fan.t_med - adev->pm.dpm.fan.t_min; - t_diff2 = adev->pm.dpm.fan.t_high - adev->pm.dpm.fan.t_med; - - pwm_diff1 = adev->pm.dpm.fan.pwm_med - adev->pm.dpm.fan.pwm_min; - pwm_diff2 = adev->pm.dpm.fan.pwm_high - adev->pm.dpm.fan.pwm_med; - - slope1 = (u16)((50 + ((16 * duty100 * pwm_diff1) / t_diff1)) / 100); - slope2 = (u16)((50 + ((16 * duty100 * pwm_diff2) / t_diff2)) / 100); - - fan_table.TempMin = cpu_to_be16((50 + adev->pm.dpm.fan.t_min) / 100); - fan_table.TempMed = cpu_to_be16((50 + adev->pm.dpm.fan.t_med) / 100); - fan_table.TempMax = cpu_to_be16((50 + adev->pm.dpm.fan.t_max) / 100); - - fan_table.Slope1 = cpu_to_be16(slope1); - fan_table.Slope2 = cpu_to_be16(slope2); - - fan_table.FdoMin = cpu_to_be16(fdo_min); - - fan_table.HystDown = cpu_to_be16(adev->pm.dpm.fan.t_hyst); - - fan_table.HystUp = cpu_to_be16(1); - - fan_table.HystSlope = cpu_to_be16(1); - - fan_table.TempRespLim = cpu_to_be16(5); - - reference_clock = amdgpu_asic_get_xclk(adev); - - fan_table.RefreshPeriod = cpu_to_be32((adev->pm.dpm.fan.cycle_delay * - reference_clock) / 1600); - - fan_table.FdoMax = cpu_to_be16((u16)duty100); - - tmp = (RREG32_SMC(ixCG_MULT_THERMAL_CTRL) & CG_MULT_THERMAL_CTRL__TEMP_SEL_MASK) - >> CG_MULT_THERMAL_CTRL__TEMP_SEL__SHIFT; - fan_table.TempSrc = (uint8_t)tmp; - - ret = amdgpu_ci_copy_bytes_to_smc(adev, - pi->fan_table_start, - (u8 *)(&fan_table), - sizeof(fan_table), - pi->sram_end); - - if (ret) { - DRM_ERROR("Failed to load fan table to the SMC."); - adev->pm.dpm.fan.ucode_fan_control = false; - } - - return 0; -} - -static int ci_fan_ctrl_start_smc_fan_control(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - PPSMC_Result ret; - - if (pi->caps_od_fuzzy_fan_control_support) { - ret = amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_StartFanControl, - FAN_CONTROL_FUZZY); - if (ret != PPSMC_Result_OK) - return -EINVAL; - ret = amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_SetFanPwmMax, - adev->pm.dpm.fan.default_max_fan_pwm); - if (ret != PPSMC_Result_OK) - return -EINVAL; - } else { - ret = amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_StartFanControl, - FAN_CONTROL_TABLE); - if (ret != PPSMC_Result_OK) - return -EINVAL; - } - - pi->fan_is_controlled_by_smc = true; - return 0; -} - - -static int ci_fan_ctrl_stop_smc_fan_control(struct amdgpu_device *adev) -{ - PPSMC_Result ret; - struct ci_power_info *pi = ci_get_pi(adev); - - ret = amdgpu_ci_send_msg_to_smc(adev, PPSMC_StopFanControl); - if (ret == PPSMC_Result_OK) { - pi->fan_is_controlled_by_smc = false; - return 0; - } else { - return -EINVAL; - } -} - -static int ci_dpm_get_fan_speed_percent(void *handle, - u32 *speed) -{ - u32 duty, duty100; - u64 tmp64; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (adev->pm.no_fan) - return -ENOENT; - - duty100 = (RREG32_SMC(ixCG_FDO_CTRL1) & CG_FDO_CTRL1__FMAX_DUTY100_MASK) - >> CG_FDO_CTRL1__FMAX_DUTY100__SHIFT; - duty = (RREG32_SMC(ixCG_THERMAL_STATUS) & CG_THERMAL_STATUS__FDO_PWM_DUTY_MASK) - >> CG_THERMAL_STATUS__FDO_PWM_DUTY__SHIFT; - - if (duty100 == 0) - return -EINVAL; - - tmp64 = (u64)duty * 100; - do_div(tmp64, duty100); - *speed = (u32)tmp64; - - if (*speed > 100) - *speed = 100; - - return 0; -} - -static int ci_dpm_set_fan_speed_percent(void *handle, - u32 speed) -{ - u32 tmp; - u32 duty, duty100; - u64 tmp64; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - - if (adev->pm.no_fan) - return -ENOENT; - - if (pi->fan_is_controlled_by_smc) - return -EINVAL; - - if (speed > 100) - return -EINVAL; - - duty100 = (RREG32_SMC(ixCG_FDO_CTRL1) & CG_FDO_CTRL1__FMAX_DUTY100_MASK) - >> CG_FDO_CTRL1__FMAX_DUTY100__SHIFT; - - if (duty100 == 0) - return -EINVAL; - - tmp64 = (u64)speed * duty100; - do_div(tmp64, 100); - duty = (u32)tmp64; - - tmp = RREG32_SMC(ixCG_FDO_CTRL0) & ~CG_FDO_CTRL0__FDO_STATIC_DUTY_MASK; - tmp |= duty << CG_FDO_CTRL0__FDO_STATIC_DUTY__SHIFT; - WREG32_SMC(ixCG_FDO_CTRL0, tmp); - - return 0; -} - -static void ci_dpm_set_fan_control_mode(void *handle, u32 mode) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - switch (mode) { - case AMD_FAN_CTRL_NONE: - if (adev->pm.dpm.fan.ucode_fan_control) - ci_fan_ctrl_stop_smc_fan_control(adev); - ci_dpm_set_fan_speed_percent(adev, 100); - break; - case AMD_FAN_CTRL_MANUAL: - if (adev->pm.dpm.fan.ucode_fan_control) - ci_fan_ctrl_stop_smc_fan_control(adev); - break; - case AMD_FAN_CTRL_AUTO: - if (adev->pm.dpm.fan.ucode_fan_control) - ci_thermal_start_smc_fan_control(adev); - break; - default: - break; - } -} - -static u32 ci_dpm_get_fan_control_mode(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - - if (pi->fan_is_controlled_by_smc) - return AMD_FAN_CTRL_AUTO; - else - return AMD_FAN_CTRL_MANUAL; -} - -#if 0 -static int ci_fan_ctrl_get_fan_speed_rpm(struct amdgpu_device *adev, - u32 *speed) -{ - u32 tach_period; - u32 xclk = amdgpu_asic_get_xclk(adev); - - if (adev->pm.no_fan) - return -ENOENT; - - if (adev->pm.fan_pulses_per_revolution == 0) - return -ENOENT; - - tach_period = (RREG32_SMC(ixCG_TACH_STATUS) & CG_TACH_STATUS__TACH_PERIOD_MASK) - >> CG_TACH_STATUS__TACH_PERIOD__SHIFT; - if (tach_period == 0) - return -ENOENT; - - *speed = 60 * xclk * 10000 / tach_period; - - return 0; -} - -static int ci_fan_ctrl_set_fan_speed_rpm(struct amdgpu_device *adev, - u32 speed) -{ - u32 tach_period, tmp; - u32 xclk = amdgpu_asic_get_xclk(adev); - - if (adev->pm.no_fan) - return -ENOENT; - - if (adev->pm.fan_pulses_per_revolution == 0) - return -ENOENT; - - if ((speed < adev->pm.fan_min_rpm) || - (speed > adev->pm.fan_max_rpm)) - return -EINVAL; - - if (adev->pm.dpm.fan.ucode_fan_control) - ci_fan_ctrl_stop_smc_fan_control(adev); - - tach_period = 60 * xclk * 10000 / (8 * speed); - tmp = RREG32_SMC(ixCG_TACH_CTRL) & ~CG_TACH_CTRL__TARGET_PERIOD_MASK; - tmp |= tach_period << CG_TACH_CTRL__TARGET_PERIOD__SHIFT; - WREG32_SMC(CG_TACH_CTRL, tmp); - - ci_fan_ctrl_set_static_mode(adev, FDO_PWM_MODE_STATIC_RPM); - - return 0; -} -#endif - -static void ci_fan_ctrl_set_default_mode(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 tmp; - - if (!pi->fan_ctrl_is_in_default_mode) { - tmp = RREG32_SMC(ixCG_FDO_CTRL2) & ~CG_FDO_CTRL2__FDO_PWM_MODE_MASK; - tmp |= pi->fan_ctrl_default_mode << CG_FDO_CTRL2__FDO_PWM_MODE__SHIFT; - WREG32_SMC(ixCG_FDO_CTRL2, tmp); - - tmp = RREG32_SMC(ixCG_FDO_CTRL2) & ~CG_FDO_CTRL2__TMIN_MASK; - tmp |= pi->t_min << CG_FDO_CTRL2__TMIN__SHIFT; - WREG32_SMC(ixCG_FDO_CTRL2, tmp); - pi->fan_ctrl_is_in_default_mode = true; - } -} - -static void ci_thermal_start_smc_fan_control(struct amdgpu_device *adev) -{ - if (adev->pm.dpm.fan.ucode_fan_control) { - ci_fan_ctrl_start_smc_fan_control(adev); - ci_fan_ctrl_set_static_mode(adev, FDO_PWM_MODE_STATIC); - } -} - -static void ci_thermal_initialize(struct amdgpu_device *adev) -{ - u32 tmp; - - if (adev->pm.fan_pulses_per_revolution) { - tmp = RREG32_SMC(ixCG_TACH_CTRL) & ~CG_TACH_CTRL__EDGE_PER_REV_MASK; - tmp |= (adev->pm.fan_pulses_per_revolution - 1) - << CG_TACH_CTRL__EDGE_PER_REV__SHIFT; - WREG32_SMC(ixCG_TACH_CTRL, tmp); - } - - tmp = RREG32_SMC(ixCG_FDO_CTRL2) & ~CG_FDO_CTRL2__TACH_PWM_RESP_RATE_MASK; - tmp |= 0x28 << CG_FDO_CTRL2__TACH_PWM_RESP_RATE__SHIFT; - WREG32_SMC(ixCG_FDO_CTRL2, tmp); -} - -static int ci_thermal_start_thermal_controller(struct amdgpu_device *adev) -{ - int ret; - - ci_thermal_initialize(adev); - ret = ci_thermal_set_temperature_range(adev, CISLANDS_TEMP_RANGE_MIN, CISLANDS_TEMP_RANGE_MAX); - if (ret) - return ret; - ret = ci_thermal_enable_alert(adev, true); - if (ret) - return ret; - if (adev->pm.dpm.fan.ucode_fan_control) { - ret = ci_thermal_setup_fan_table(adev); - if (ret) - return ret; - ci_thermal_start_smc_fan_control(adev); - } - - return 0; -} - -static void ci_thermal_stop_thermal_controller(struct amdgpu_device *adev) -{ - if (!adev->pm.no_fan) - ci_fan_ctrl_set_default_mode(adev); -} - -static int ci_read_smc_soft_register(struct amdgpu_device *adev, - u16 reg_offset, u32 *value) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - return amdgpu_ci_read_smc_sram_dword(adev, - pi->soft_regs_start + reg_offset, - value, pi->sram_end); -} - -static int ci_write_smc_soft_register(struct amdgpu_device *adev, - u16 reg_offset, u32 value) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - return amdgpu_ci_write_smc_sram_dword(adev, - pi->soft_regs_start + reg_offset, - value, pi->sram_end); -} - -static void ci_init_fps_limits(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - SMU7_Discrete_DpmTable *table = &pi->smc_state_table; - - if (pi->caps_fps) { - u16 tmp; - - tmp = 45; - table->FpsHighT = cpu_to_be16(tmp); - - tmp = 30; - table->FpsLowT = cpu_to_be16(tmp); - } -} - -static int ci_update_sclk_t(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - int ret = 0; - u32 low_sclk_interrupt_t = 0; - - if (pi->caps_sclk_throttle_low_notification) { - low_sclk_interrupt_t = cpu_to_be32(pi->low_sclk_interrupt_t); - - ret = amdgpu_ci_copy_bytes_to_smc(adev, - pi->dpm_table_start + - offsetof(SMU7_Discrete_DpmTable, LowSclkInterruptT), - (u8 *)&low_sclk_interrupt_t, - sizeof(u32), pi->sram_end); - - } - - return ret; -} - -static void ci_get_leakage_voltages(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u16 leakage_id, virtual_voltage_id; - u16 vddc, vddci; - int i; - - pi->vddc_leakage.count = 0; - pi->vddci_leakage.count = 0; - - if (adev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_EVV) { - for (i = 0; i < CISLANDS_MAX_LEAKAGE_COUNT; i++) { - virtual_voltage_id = ATOM_VIRTUAL_VOLTAGE_ID0 + i; - if (amdgpu_atombios_get_voltage_evv(adev, virtual_voltage_id, &vddc) != 0) - continue; - if (vddc != 0 && vddc != virtual_voltage_id) { - pi->vddc_leakage.actual_voltage[pi->vddc_leakage.count] = vddc; - pi->vddc_leakage.leakage_id[pi->vddc_leakage.count] = virtual_voltage_id; - pi->vddc_leakage.count++; - } - } - } else if (amdgpu_atombios_get_leakage_id_from_vbios(adev, &leakage_id) == 0) { - for (i = 0; i < CISLANDS_MAX_LEAKAGE_COUNT; i++) { - virtual_voltage_id = ATOM_VIRTUAL_VOLTAGE_ID0 + i; - if (amdgpu_atombios_get_leakage_vddc_based_on_leakage_params(adev, &vddc, &vddci, - virtual_voltage_id, - leakage_id) == 0) { - if (vddc != 0 && vddc != virtual_voltage_id) { - pi->vddc_leakage.actual_voltage[pi->vddc_leakage.count] = vddc; - pi->vddc_leakage.leakage_id[pi->vddc_leakage.count] = virtual_voltage_id; - pi->vddc_leakage.count++; - } - if (vddci != 0 && vddci != virtual_voltage_id) { - pi->vddci_leakage.actual_voltage[pi->vddci_leakage.count] = vddci; - pi->vddci_leakage.leakage_id[pi->vddci_leakage.count] = virtual_voltage_id; - pi->vddci_leakage.count++; - } - } - } - } -} - -static void ci_set_dpm_event_sources(struct amdgpu_device *adev, u32 sources) -{ - struct ci_power_info *pi = ci_get_pi(adev); - bool want_thermal_protection; - enum amdgpu_dpm_event_src dpm_event_src; - u32 tmp; - - switch (sources) { - case 0: - default: - want_thermal_protection = false; - break; - case (1 << AMDGPU_DPM_AUTO_THROTTLE_SRC_THERMAL): - want_thermal_protection = true; - dpm_event_src = AMDGPU_DPM_EVENT_SRC_DIGITAL; - break; - case (1 << AMDGPU_DPM_AUTO_THROTTLE_SRC_EXTERNAL): - want_thermal_protection = true; - dpm_event_src = AMDGPU_DPM_EVENT_SRC_EXTERNAL; - break; - case ((1 << AMDGPU_DPM_AUTO_THROTTLE_SRC_EXTERNAL) | - (1 << AMDGPU_DPM_AUTO_THROTTLE_SRC_THERMAL)): - want_thermal_protection = true; - dpm_event_src = AMDGPU_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL; - break; - } - - if (want_thermal_protection) { -#if 0 - /* XXX: need to figure out how to handle this properly */ - tmp = RREG32_SMC(ixCG_THERMAL_CTRL); - tmp &= DPM_EVENT_SRC_MASK; - tmp |= DPM_EVENT_SRC(dpm_event_src); - WREG32_SMC(ixCG_THERMAL_CTRL, tmp); -#endif - - tmp = RREG32_SMC(ixGENERAL_PWRMGT); - if (pi->thermal_protection) - tmp &= ~GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK; - else - tmp |= GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK; - WREG32_SMC(ixGENERAL_PWRMGT, tmp); - } else { - tmp = RREG32_SMC(ixGENERAL_PWRMGT); - tmp |= GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK; - WREG32_SMC(ixGENERAL_PWRMGT, tmp); - } -} - -static void ci_enable_auto_throttle_source(struct amdgpu_device *adev, - enum amdgpu_dpm_auto_throttle_src source, - bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - if (enable) { - if (!(pi->active_auto_throttle_sources & (1 << source))) { - pi->active_auto_throttle_sources |= 1 << source; - ci_set_dpm_event_sources(adev, pi->active_auto_throttle_sources); - } - } else { - if (pi->active_auto_throttle_sources & (1 << source)) { - pi->active_auto_throttle_sources &= ~(1 << source); - ci_set_dpm_event_sources(adev, pi->active_auto_throttle_sources); - } - } -} - -static void ci_enable_vr_hot_gpio_interrupt(struct amdgpu_device *adev) -{ - if (adev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REGULATOR_HOT) - amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_EnableVRHotGPIOInterrupt); -} - -static int ci_unfreeze_sclk_mclk_dpm(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - PPSMC_Result smc_result; - - if (!pi->need_update_smu7_dpm_table) - return 0; - - if ((!pi->sclk_dpm_key_disabled) && - (pi->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK | DPMTABLE_UPDATE_SCLK))) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_SCLKDPM_UnfreezeLevel); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - if ((!pi->mclk_dpm_key_disabled) && - (pi->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_MCLKDPM_UnfreezeLevel); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - pi->need_update_smu7_dpm_table = 0; - return 0; -} - -static int ci_enable_sclk_mclk_dpm(struct amdgpu_device *adev, bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - PPSMC_Result smc_result; - - if (enable) { - if (!pi->sclk_dpm_key_disabled) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_DPM_Enable); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - if (!pi->mclk_dpm_key_disabled) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_MCLKDPM_Enable); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - - WREG32_P(mmMC_SEQ_CNTL_3, MC_SEQ_CNTL_3__CAC_EN_MASK, - ~MC_SEQ_CNTL_3__CAC_EN_MASK); - - WREG32_SMC(ixLCAC_MC0_CNTL, 0x05); - WREG32_SMC(ixLCAC_MC1_CNTL, 0x05); - WREG32_SMC(ixLCAC_CPL_CNTL, 0x100005); - - udelay(10); - - WREG32_SMC(ixLCAC_MC0_CNTL, 0x400005); - WREG32_SMC(ixLCAC_MC1_CNTL, 0x400005); - WREG32_SMC(ixLCAC_CPL_CNTL, 0x500005); - } - } else { - if (!pi->sclk_dpm_key_disabled) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_DPM_Disable); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - if (!pi->mclk_dpm_key_disabled) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_MCLKDPM_Disable); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - } - - return 0; -} - -static int ci_start_dpm(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - PPSMC_Result smc_result; - int ret; - u32 tmp; - - tmp = RREG32_SMC(ixGENERAL_PWRMGT); - tmp |= GENERAL_PWRMGT__GLOBAL_PWRMGT_EN_MASK; - WREG32_SMC(ixGENERAL_PWRMGT, tmp); - - tmp = RREG32_SMC(ixSCLK_PWRMGT_CNTL); - tmp |= SCLK_PWRMGT_CNTL__DYNAMIC_PM_EN_MASK; - WREG32_SMC(ixSCLK_PWRMGT_CNTL, tmp); - - ci_write_smc_soft_register(adev, offsetof(SMU7_SoftRegisters, VoltageChangeTimeout), 0x1000); - - WREG32_P(mmBIF_LNCNT_RESET, 0, ~BIF_LNCNT_RESET__RESET_LNCNT_EN_MASK); - - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_Voltage_Cntl_Enable); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - - ret = ci_enable_sclk_mclk_dpm(adev, true); - if (ret) - return ret; - - if (!pi->pcie_dpm_key_disabled) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_PCIeDPM_Enable); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - return 0; -} - -static int ci_freeze_sclk_mclk_dpm(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - PPSMC_Result smc_result; - - if (!pi->need_update_smu7_dpm_table) - return 0; - - if ((!pi->sclk_dpm_key_disabled) && - (pi->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK | DPMTABLE_UPDATE_SCLK))) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_SCLKDPM_FreezeLevel); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - if ((!pi->mclk_dpm_key_disabled) && - (pi->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_MCLKDPM_FreezeLevel); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - return 0; -} - -static int ci_stop_dpm(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - PPSMC_Result smc_result; - int ret; - u32 tmp; - - tmp = RREG32_SMC(ixGENERAL_PWRMGT); - tmp &= ~GENERAL_PWRMGT__GLOBAL_PWRMGT_EN_MASK; - WREG32_SMC(ixGENERAL_PWRMGT, tmp); - - tmp = RREG32_SMC(ixSCLK_PWRMGT_CNTL); - tmp &= ~SCLK_PWRMGT_CNTL__DYNAMIC_PM_EN_MASK; - WREG32_SMC(ixSCLK_PWRMGT_CNTL, tmp); - - if (!pi->pcie_dpm_key_disabled) { - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_PCIeDPM_Disable); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - ret = ci_enable_sclk_mclk_dpm(adev, false); - if (ret) - return ret; - - smc_result = amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_Voltage_Cntl_Disable); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - - return 0; -} - -static void ci_enable_sclk_control(struct amdgpu_device *adev, bool enable) -{ - u32 tmp = RREG32_SMC(ixSCLK_PWRMGT_CNTL); - - if (enable) - tmp &= ~SCLK_PWRMGT_CNTL__SCLK_PWRMGT_OFF_MASK; - else - tmp |= SCLK_PWRMGT_CNTL__SCLK_PWRMGT_OFF_MASK; - WREG32_SMC(ixSCLK_PWRMGT_CNTL, tmp); -} - -#if 0 -static int ci_notify_hw_of_power_source(struct amdgpu_device *adev, - bool ac_power) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct amdgpu_cac_tdp_table *cac_tdp_table = - adev->pm.dpm.dyn_state.cac_tdp_table; - u32 power_limit; - - if (ac_power) - power_limit = (u32)(cac_tdp_table->maximum_power_delivery_limit * 256); - else - power_limit = (u32)(cac_tdp_table->battery_power_limit * 256); - - ci_set_power_limit(adev, power_limit); - - if (pi->caps_automatic_dc_transition) { - if (ac_power) - amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_RunningOnAC); - else - amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_Remove_DC_Clamp); - } - - return 0; -} -#endif - -static PPSMC_Result amdgpu_ci_send_msg_to_smc_with_parameter(struct amdgpu_device *adev, - PPSMC_Msg msg, u32 parameter) -{ - WREG32(mmSMC_MSG_ARG_0, parameter); - return amdgpu_ci_send_msg_to_smc(adev, msg); -} - -static PPSMC_Result amdgpu_ci_send_msg_to_smc_return_parameter(struct amdgpu_device *adev, - PPSMC_Msg msg, u32 *parameter) -{ - PPSMC_Result smc_result; - - smc_result = amdgpu_ci_send_msg_to_smc(adev, msg); - - if ((smc_result == PPSMC_Result_OK) && parameter) - *parameter = RREG32(mmSMC_MSG_ARG_0); - - return smc_result; -} - -static int ci_dpm_force_state_sclk(struct amdgpu_device *adev, u32 n) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - if (!pi->sclk_dpm_key_disabled) { - PPSMC_Result smc_result = - amdgpu_ci_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_SCLKDPM_SetEnabledMask, 1 << n); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - return 0; -} - -static int ci_dpm_force_state_mclk(struct amdgpu_device *adev, u32 n) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - if (!pi->mclk_dpm_key_disabled) { - PPSMC_Result smc_result = - amdgpu_ci_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_MCLKDPM_SetEnabledMask, 1 << n); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - return 0; -} - -static int ci_dpm_force_state_pcie(struct amdgpu_device *adev, u32 n) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - if (!pi->pcie_dpm_key_disabled) { - PPSMC_Result smc_result = - amdgpu_ci_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_PCIeDPM_ForceLevel, n); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - return 0; -} - -static int ci_set_power_limit(struct amdgpu_device *adev, u32 n) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - if (pi->power_containment_features & POWERCONTAINMENT_FEATURE_PkgPwrLimit) { - PPSMC_Result smc_result = - amdgpu_ci_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_PkgPwrSetLimit, n); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - - return 0; -} - -static int ci_set_overdrive_target_tdp(struct amdgpu_device *adev, - u32 target_tdp) -{ - PPSMC_Result smc_result = - amdgpu_ci_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_OverDriveSetTargetTdp, target_tdp); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - return 0; -} - -#if 0 -static int ci_set_boot_state(struct amdgpu_device *adev) -{ - return ci_enable_sclk_mclk_dpm(adev, false); -} -#endif - -static u32 ci_get_average_sclk_freq(struct amdgpu_device *adev) -{ - u32 sclk_freq; - PPSMC_Result smc_result = - amdgpu_ci_send_msg_to_smc_return_parameter(adev, - PPSMC_MSG_API_GetSclkFrequency, - &sclk_freq); - if (smc_result != PPSMC_Result_OK) - sclk_freq = 0; - - return sclk_freq; -} - -static u32 ci_get_average_mclk_freq(struct amdgpu_device *adev) -{ - u32 mclk_freq; - PPSMC_Result smc_result = - amdgpu_ci_send_msg_to_smc_return_parameter(adev, - PPSMC_MSG_API_GetMclkFrequency, - &mclk_freq); - if (smc_result != PPSMC_Result_OK) - mclk_freq = 0; - - return mclk_freq; -} - -static void ci_dpm_start_smc(struct amdgpu_device *adev) -{ - int i; - - amdgpu_ci_program_jump_on_start(adev); - amdgpu_ci_start_smc_clock(adev); - amdgpu_ci_start_smc(adev); - for (i = 0; i < adev->usec_timeout; i++) { - if (RREG32_SMC(ixFIRMWARE_FLAGS) & FIRMWARE_FLAGS__INTERRUPTS_ENABLED_MASK) - break; - } -} - -static void ci_dpm_stop_smc(struct amdgpu_device *adev) -{ - amdgpu_ci_reset_smc(adev); - amdgpu_ci_stop_smc_clock(adev); -} - -static int ci_process_firmware_header(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 tmp; - int ret; - - ret = amdgpu_ci_read_smc_sram_dword(adev, - SMU7_FIRMWARE_HEADER_LOCATION + - offsetof(SMU7_Firmware_Header, DpmTable), - &tmp, pi->sram_end); - if (ret) - return ret; - - pi->dpm_table_start = tmp; - - ret = amdgpu_ci_read_smc_sram_dword(adev, - SMU7_FIRMWARE_HEADER_LOCATION + - offsetof(SMU7_Firmware_Header, SoftRegisters), - &tmp, pi->sram_end); - if (ret) - return ret; - - pi->soft_regs_start = tmp; - - ret = amdgpu_ci_read_smc_sram_dword(adev, - SMU7_FIRMWARE_HEADER_LOCATION + - offsetof(SMU7_Firmware_Header, mcRegisterTable), - &tmp, pi->sram_end); - if (ret) - return ret; - - pi->mc_reg_table_start = tmp; - - ret = amdgpu_ci_read_smc_sram_dword(adev, - SMU7_FIRMWARE_HEADER_LOCATION + - offsetof(SMU7_Firmware_Header, FanTable), - &tmp, pi->sram_end); - if (ret) - return ret; - - pi->fan_table_start = tmp; - - ret = amdgpu_ci_read_smc_sram_dword(adev, - SMU7_FIRMWARE_HEADER_LOCATION + - offsetof(SMU7_Firmware_Header, mcArbDramTimingTable), - &tmp, pi->sram_end); - if (ret) - return ret; - - pi->arb_table_start = tmp; - - return 0; -} - -static void ci_read_clock_registers(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - pi->clock_registers.cg_spll_func_cntl = - RREG32_SMC(ixCG_SPLL_FUNC_CNTL); - pi->clock_registers.cg_spll_func_cntl_2 = - RREG32_SMC(ixCG_SPLL_FUNC_CNTL_2); - pi->clock_registers.cg_spll_func_cntl_3 = - RREG32_SMC(ixCG_SPLL_FUNC_CNTL_3); - pi->clock_registers.cg_spll_func_cntl_4 = - RREG32_SMC(ixCG_SPLL_FUNC_CNTL_4); - pi->clock_registers.cg_spll_spread_spectrum = - RREG32_SMC(ixCG_SPLL_SPREAD_SPECTRUM); - pi->clock_registers.cg_spll_spread_spectrum_2 = - RREG32_SMC(ixCG_SPLL_SPREAD_SPECTRUM_2); - pi->clock_registers.dll_cntl = RREG32(mmDLL_CNTL); - pi->clock_registers.mclk_pwrmgt_cntl = RREG32(mmMCLK_PWRMGT_CNTL); - pi->clock_registers.mpll_ad_func_cntl = RREG32(mmMPLL_AD_FUNC_CNTL); - pi->clock_registers.mpll_dq_func_cntl = RREG32(mmMPLL_DQ_FUNC_CNTL); - pi->clock_registers.mpll_func_cntl = RREG32(mmMPLL_FUNC_CNTL); - pi->clock_registers.mpll_func_cntl_1 = RREG32(mmMPLL_FUNC_CNTL_1); - pi->clock_registers.mpll_func_cntl_2 = RREG32(mmMPLL_FUNC_CNTL_2); - pi->clock_registers.mpll_ss1 = RREG32(mmMPLL_SS1); - pi->clock_registers.mpll_ss2 = RREG32(mmMPLL_SS2); -} - -static void ci_init_sclk_t(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - pi->low_sclk_interrupt_t = 0; -} - -static void ci_enable_thermal_protection(struct amdgpu_device *adev, - bool enable) -{ - u32 tmp = RREG32_SMC(ixGENERAL_PWRMGT); - - if (enable) - tmp &= ~GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK; - else - tmp |= GENERAL_PWRMGT__THERMAL_PROTECTION_DIS_MASK; - WREG32_SMC(ixGENERAL_PWRMGT, tmp); -} - -static void ci_enable_acpi_power_management(struct amdgpu_device *adev) -{ - u32 tmp = RREG32_SMC(ixGENERAL_PWRMGT); - - tmp |= GENERAL_PWRMGT__STATIC_PM_EN_MASK; - - WREG32_SMC(ixGENERAL_PWRMGT, tmp); -} - -#if 0 -static int ci_enter_ulp_state(struct amdgpu_device *adev) -{ - - WREG32(mmSMC_MESSAGE_0, PPSMC_MSG_SwitchToMinimumPower); - - udelay(25000); - - return 0; -} - -static int ci_exit_ulp_state(struct amdgpu_device *adev) -{ - int i; - - WREG32(mmSMC_MESSAGE_0, PPSMC_MSG_ResumeFromMinimumPower); - - udelay(7000); - - for (i = 0; i < adev->usec_timeout; i++) { - if (RREG32(mmSMC_RESP_0) == 1) - break; - udelay(1000); - } - - return 0; -} -#endif - -static int ci_notify_smc_display_change(struct amdgpu_device *adev, - bool has_display) -{ - PPSMC_Msg msg = has_display ? PPSMC_MSG_HasDisplay : PPSMC_MSG_NoDisplay; - - return (amdgpu_ci_send_msg_to_smc(adev, msg) == PPSMC_Result_OK) ? 0 : -EINVAL; -} - -static int ci_enable_ds_master_switch(struct amdgpu_device *adev, - bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - if (enable) { - if (pi->caps_sclk_ds) { - if (amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_MASTER_DeepSleep_ON) != PPSMC_Result_OK) - return -EINVAL; - } else { - if (amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_MASTER_DeepSleep_OFF) != PPSMC_Result_OK) - return -EINVAL; - } - } else { - if (pi->caps_sclk_ds) { - if (amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_MASTER_DeepSleep_OFF) != PPSMC_Result_OK) - return -EINVAL; - } - } - - return 0; -} - -static void ci_program_display_gap(struct amdgpu_device *adev) -{ - u32 tmp = RREG32_SMC(ixCG_DISPLAY_GAP_CNTL); - u32 pre_vbi_time_in_us; - u32 frame_time_in_us; - u32 ref_clock = adev->clock.spll.reference_freq; - u32 refresh_rate = amdgpu_dpm_get_vrefresh(adev); - u32 vblank_time = amdgpu_dpm_get_vblank_time(adev); - - tmp &= ~CG_DISPLAY_GAP_CNTL__DISP_GAP_MASK; - if (adev->pm.dpm.new_active_crtc_count > 0) - tmp |= (AMDGPU_PM_DISPLAY_GAP_VBLANK_OR_WM << CG_DISPLAY_GAP_CNTL__DISP_GAP__SHIFT); - else - tmp |= (AMDGPU_PM_DISPLAY_GAP_IGNORE << CG_DISPLAY_GAP_CNTL__DISP_GAP__SHIFT); - WREG32_SMC(ixCG_DISPLAY_GAP_CNTL, tmp); - - if (refresh_rate == 0) - refresh_rate = 60; - if (vblank_time == 0xffffffff) - vblank_time = 500; - frame_time_in_us = 1000000 / refresh_rate; - pre_vbi_time_in_us = - frame_time_in_us - 200 - vblank_time; - tmp = pre_vbi_time_in_us * (ref_clock / 100); - - WREG32_SMC(ixCG_DISPLAY_GAP_CNTL2, tmp); - ci_write_smc_soft_register(adev, offsetof(SMU7_SoftRegisters, PreVBlankGap), 0x64); - ci_write_smc_soft_register(adev, offsetof(SMU7_SoftRegisters, VBlankTimeout), (frame_time_in_us - pre_vbi_time_in_us)); - - - ci_notify_smc_display_change(adev, (adev->pm.dpm.new_active_crtc_count == 1)); - -} - -static void ci_enable_spread_spectrum(struct amdgpu_device *adev, bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 tmp; - - if (enable) { - if (pi->caps_sclk_ss_support) { - tmp = RREG32_SMC(ixGENERAL_PWRMGT); - tmp |= GENERAL_PWRMGT__DYN_SPREAD_SPECTRUM_EN_MASK; - WREG32_SMC(ixGENERAL_PWRMGT, tmp); - } - } else { - tmp = RREG32_SMC(ixCG_SPLL_SPREAD_SPECTRUM); - tmp &= ~CG_SPLL_SPREAD_SPECTRUM__SSEN_MASK; - WREG32_SMC(ixCG_SPLL_SPREAD_SPECTRUM, tmp); - - tmp = RREG32_SMC(ixGENERAL_PWRMGT); - tmp &= ~GENERAL_PWRMGT__DYN_SPREAD_SPECTRUM_EN_MASK; - WREG32_SMC(ixGENERAL_PWRMGT, tmp); - } -} - -static void ci_program_sstp(struct amdgpu_device *adev) -{ - WREG32_SMC(ixCG_STATIC_SCREEN_PARAMETER, - ((CISLANDS_SSTU_DFLT << CG_STATIC_SCREEN_PARAMETER__STATIC_SCREEN_THRESHOLD_UNIT__SHIFT) | - (CISLANDS_SST_DFLT << CG_STATIC_SCREEN_PARAMETER__STATIC_SCREEN_THRESHOLD__SHIFT))); -} - -static void ci_enable_display_gap(struct amdgpu_device *adev) -{ - u32 tmp = RREG32_SMC(ixCG_DISPLAY_GAP_CNTL); - - tmp &= ~(CG_DISPLAY_GAP_CNTL__DISP_GAP_MASK | - CG_DISPLAY_GAP_CNTL__DISP_GAP_MCHG_MASK); - tmp |= ((AMDGPU_PM_DISPLAY_GAP_IGNORE << CG_DISPLAY_GAP_CNTL__DISP_GAP__SHIFT) | - (AMDGPU_PM_DISPLAY_GAP_VBLANK << CG_DISPLAY_GAP_CNTL__DISP_GAP_MCHG__SHIFT)); - - WREG32_SMC(ixCG_DISPLAY_GAP_CNTL, tmp); -} - -static void ci_program_vc(struct amdgpu_device *adev) -{ - u32 tmp; - - tmp = RREG32_SMC(ixSCLK_PWRMGT_CNTL); - tmp &= ~(SCLK_PWRMGT_CNTL__RESET_SCLK_CNT_MASK | SCLK_PWRMGT_CNTL__RESET_BUSY_CNT_MASK); - WREG32_SMC(ixSCLK_PWRMGT_CNTL, tmp); - - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_0, CISLANDS_VRC_DFLT0); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_1, CISLANDS_VRC_DFLT1); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_2, CISLANDS_VRC_DFLT2); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_3, CISLANDS_VRC_DFLT3); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_4, CISLANDS_VRC_DFLT4); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_5, CISLANDS_VRC_DFLT5); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_6, CISLANDS_VRC_DFLT6); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_7, CISLANDS_VRC_DFLT7); -} - -static void ci_clear_vc(struct amdgpu_device *adev) -{ - u32 tmp; - - tmp = RREG32_SMC(ixSCLK_PWRMGT_CNTL); - tmp |= (SCLK_PWRMGT_CNTL__RESET_SCLK_CNT_MASK | SCLK_PWRMGT_CNTL__RESET_BUSY_CNT_MASK); - WREG32_SMC(ixSCLK_PWRMGT_CNTL, tmp); - - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_0, 0); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_1, 0); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_2, 0); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_3, 0); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_4, 0); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_5, 0); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_6, 0); - WREG32_SMC(ixCG_FREQ_TRAN_VOTING_7, 0); -} - -static int ci_upload_firmware(struct amdgpu_device *adev) -{ - int i, ret; - - if (amdgpu_ci_is_smc_running(adev)) { - DRM_INFO("smc is running, no need to load smc firmware\n"); - return 0; - } - - for (i = 0; i < adev->usec_timeout; i++) { - if (RREG32_SMC(ixRCU_UC_EVENTS) & RCU_UC_EVENTS__boot_seq_done_MASK) - break; - } - WREG32_SMC(ixSMC_SYSCON_MISC_CNTL, 1); - - amdgpu_ci_stop_smc_clock(adev); - amdgpu_ci_reset_smc(adev); - - ret = amdgpu_ci_load_smc_ucode(adev, SMC_RAM_END); - - return ret; - -} - -static int ci_get_svi2_voltage_table(struct amdgpu_device *adev, - struct amdgpu_clock_voltage_dependency_table *voltage_dependency_table, - struct atom_voltage_table *voltage_table) -{ - u32 i; - - if (voltage_dependency_table == NULL) - return -EINVAL; - - voltage_table->mask_low = 0; - voltage_table->phase_delay = 0; - - voltage_table->count = voltage_dependency_table->count; - for (i = 0; i < voltage_table->count; i++) { - voltage_table->entries[i].value = voltage_dependency_table->entries[i].v; - voltage_table->entries[i].smio_low = 0; - } - - return 0; -} - -static int ci_construct_voltage_tables(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - int ret; - - if (pi->voltage_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) { - ret = amdgpu_atombios_get_voltage_table(adev, VOLTAGE_TYPE_VDDC, - VOLTAGE_OBJ_GPIO_LUT, - &pi->vddc_voltage_table); - if (ret) - return ret; - } else if (pi->voltage_control == CISLANDS_VOLTAGE_CONTROL_BY_SVID2) { - ret = ci_get_svi2_voltage_table(adev, - &adev->pm.dpm.dyn_state.vddc_dependency_on_mclk, - &pi->vddc_voltage_table); - if (ret) - return ret; - } - - if (pi->vddc_voltage_table.count > SMU7_MAX_LEVELS_VDDC) - ci_trim_voltage_table_to_fit_state_table(adev, SMU7_MAX_LEVELS_VDDC, - &pi->vddc_voltage_table); - - if (pi->vddci_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) { - ret = amdgpu_atombios_get_voltage_table(adev, VOLTAGE_TYPE_VDDCI, - VOLTAGE_OBJ_GPIO_LUT, - &pi->vddci_voltage_table); - if (ret) - return ret; - } else if (pi->vddci_control == CISLANDS_VOLTAGE_CONTROL_BY_SVID2) { - ret = ci_get_svi2_voltage_table(adev, - &adev->pm.dpm.dyn_state.vddci_dependency_on_mclk, - &pi->vddci_voltage_table); - if (ret) - return ret; - } - - if (pi->vddci_voltage_table.count > SMU7_MAX_LEVELS_VDDCI) - ci_trim_voltage_table_to_fit_state_table(adev, SMU7_MAX_LEVELS_VDDCI, - &pi->vddci_voltage_table); - - if (pi->mvdd_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) { - ret = amdgpu_atombios_get_voltage_table(adev, VOLTAGE_TYPE_MVDDC, - VOLTAGE_OBJ_GPIO_LUT, - &pi->mvdd_voltage_table); - if (ret) - return ret; - } else if (pi->mvdd_control == CISLANDS_VOLTAGE_CONTROL_BY_SVID2) { - ret = ci_get_svi2_voltage_table(adev, - &adev->pm.dpm.dyn_state.mvdd_dependency_on_mclk, - &pi->mvdd_voltage_table); - if (ret) - return ret; - } - - if (pi->mvdd_voltage_table.count > SMU7_MAX_LEVELS_MVDD) - ci_trim_voltage_table_to_fit_state_table(adev, SMU7_MAX_LEVELS_MVDD, - &pi->mvdd_voltage_table); - - return 0; -} - -static void ci_populate_smc_voltage_table(struct amdgpu_device *adev, - struct atom_voltage_table_entry *voltage_table, - SMU7_Discrete_VoltageLevel *smc_voltage_table) -{ - int ret; - - ret = ci_get_std_voltage_value_sidd(adev, voltage_table, - &smc_voltage_table->StdVoltageHiSidd, - &smc_voltage_table->StdVoltageLoSidd); - - if (ret) { - smc_voltage_table->StdVoltageHiSidd = voltage_table->value * VOLTAGE_SCALE; - smc_voltage_table->StdVoltageLoSidd = voltage_table->value * VOLTAGE_SCALE; - } - - smc_voltage_table->Voltage = cpu_to_be16(voltage_table->value * VOLTAGE_SCALE); - smc_voltage_table->StdVoltageHiSidd = - cpu_to_be16(smc_voltage_table->StdVoltageHiSidd); - smc_voltage_table->StdVoltageLoSidd = - cpu_to_be16(smc_voltage_table->StdVoltageLoSidd); -} - -static int ci_populate_smc_vddc_table(struct amdgpu_device *adev, - SMU7_Discrete_DpmTable *table) -{ - struct ci_power_info *pi = ci_get_pi(adev); - unsigned int count; - - table->VddcLevelCount = pi->vddc_voltage_table.count; - for (count = 0; count < table->VddcLevelCount; count++) { - ci_populate_smc_voltage_table(adev, - &pi->vddc_voltage_table.entries[count], - &table->VddcLevel[count]); - - if (pi->voltage_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) - table->VddcLevel[count].Smio |= - pi->vddc_voltage_table.entries[count].smio_low; - else - table->VddcLevel[count].Smio = 0; - } - table->VddcLevelCount = cpu_to_be32(table->VddcLevelCount); - - return 0; -} - -static int ci_populate_smc_vddci_table(struct amdgpu_device *adev, - SMU7_Discrete_DpmTable *table) -{ - unsigned int count; - struct ci_power_info *pi = ci_get_pi(adev); - - table->VddciLevelCount = pi->vddci_voltage_table.count; - for (count = 0; count < table->VddciLevelCount; count++) { - ci_populate_smc_voltage_table(adev, - &pi->vddci_voltage_table.entries[count], - &table->VddciLevel[count]); - - if (pi->vddci_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) - table->VddciLevel[count].Smio |= - pi->vddci_voltage_table.entries[count].smio_low; - else - table->VddciLevel[count].Smio = 0; - } - table->VddciLevelCount = cpu_to_be32(table->VddciLevelCount); - - return 0; -} - -static int ci_populate_smc_mvdd_table(struct amdgpu_device *adev, - SMU7_Discrete_DpmTable *table) -{ - struct ci_power_info *pi = ci_get_pi(adev); - unsigned int count; - - table->MvddLevelCount = pi->mvdd_voltage_table.count; - for (count = 0; count < table->MvddLevelCount; count++) { - ci_populate_smc_voltage_table(adev, - &pi->mvdd_voltage_table.entries[count], - &table->MvddLevel[count]); - - if (pi->mvdd_control == CISLANDS_VOLTAGE_CONTROL_BY_GPIO) - table->MvddLevel[count].Smio |= - pi->mvdd_voltage_table.entries[count].smio_low; - else - table->MvddLevel[count].Smio = 0; - } - table->MvddLevelCount = cpu_to_be32(table->MvddLevelCount); - - return 0; -} - -static int ci_populate_smc_voltage_tables(struct amdgpu_device *adev, - SMU7_Discrete_DpmTable *table) -{ - int ret; - - ret = ci_populate_smc_vddc_table(adev, table); - if (ret) - return ret; - - ret = ci_populate_smc_vddci_table(adev, table); - if (ret) - return ret; - - ret = ci_populate_smc_mvdd_table(adev, table); - if (ret) - return ret; - - return 0; -} - -static int ci_populate_mvdd_value(struct amdgpu_device *adev, u32 mclk, - SMU7_Discrete_VoltageLevel *voltage) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 i = 0; - - if (pi->mvdd_control != CISLANDS_VOLTAGE_CONTROL_NONE) { - for (i = 0; i < adev->pm.dpm.dyn_state.mvdd_dependency_on_mclk.count; i++) { - if (mclk <= adev->pm.dpm.dyn_state.mvdd_dependency_on_mclk.entries[i].clk) { - voltage->Voltage = pi->mvdd_voltage_table.entries[i].value; - break; - } - } - - if (i >= adev->pm.dpm.dyn_state.mvdd_dependency_on_mclk.count) - return -EINVAL; - } - - return -EINVAL; -} - -static int ci_get_std_voltage_value_sidd(struct amdgpu_device *adev, - struct atom_voltage_table_entry *voltage_table, - u16 *std_voltage_hi_sidd, u16 *std_voltage_lo_sidd) -{ - u16 v_index, idx; - bool voltage_found = false; - *std_voltage_hi_sidd = voltage_table->value * VOLTAGE_SCALE; - *std_voltage_lo_sidd = voltage_table->value * VOLTAGE_SCALE; - - if (adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries == NULL) - return -EINVAL; - - if (adev->pm.dpm.dyn_state.cac_leakage_table.entries) { - for (v_index = 0; (u32)v_index < adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count; v_index++) { - if (voltage_table->value == - adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[v_index].v) { - voltage_found = true; - if ((u32)v_index < adev->pm.dpm.dyn_state.cac_leakage_table.count) - idx = v_index; - else - idx = adev->pm.dpm.dyn_state.cac_leakage_table.count - 1; - *std_voltage_lo_sidd = - adev->pm.dpm.dyn_state.cac_leakage_table.entries[idx].vddc * VOLTAGE_SCALE; - *std_voltage_hi_sidd = - adev->pm.dpm.dyn_state.cac_leakage_table.entries[idx].leakage * VOLTAGE_SCALE; - break; - } - } - - if (!voltage_found) { - for (v_index = 0; (u32)v_index < adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count; v_index++) { - if (voltage_table->value <= - adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[v_index].v) { - voltage_found = true; - if ((u32)v_index < adev->pm.dpm.dyn_state.cac_leakage_table.count) - idx = v_index; - else - idx = adev->pm.dpm.dyn_state.cac_leakage_table.count - 1; - *std_voltage_lo_sidd = - adev->pm.dpm.dyn_state.cac_leakage_table.entries[idx].vddc * VOLTAGE_SCALE; - *std_voltage_hi_sidd = - adev->pm.dpm.dyn_state.cac_leakage_table.entries[idx].leakage * VOLTAGE_SCALE; - break; - } - } - } - } - - return 0; -} - -static void ci_populate_phase_value_based_on_sclk(struct amdgpu_device *adev, - const struct amdgpu_phase_shedding_limits_table *limits, - u32 sclk, - u32 *phase_shedding) -{ - unsigned int i; - - *phase_shedding = 1; - - for (i = 0; i < limits->count; i++) { - if (sclk < limits->entries[i].sclk) { - *phase_shedding = i; - break; - } - } -} - -static void ci_populate_phase_value_based_on_mclk(struct amdgpu_device *adev, - const struct amdgpu_phase_shedding_limits_table *limits, - u32 mclk, - u32 *phase_shedding) -{ - unsigned int i; - - *phase_shedding = 1; - - for (i = 0; i < limits->count; i++) { - if (mclk < limits->entries[i].mclk) { - *phase_shedding = i; - break; - } - } -} - -static int ci_init_arb_table_index(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 tmp; - int ret; - - ret = amdgpu_ci_read_smc_sram_dword(adev, pi->arb_table_start, - &tmp, pi->sram_end); - if (ret) - return ret; - - tmp &= 0x00FFFFFF; - tmp |= MC_CG_ARB_FREQ_F1 << 24; - - return amdgpu_ci_write_smc_sram_dword(adev, pi->arb_table_start, - tmp, pi->sram_end); -} - -static int ci_get_dependency_volt_by_clk(struct amdgpu_device *adev, - struct amdgpu_clock_voltage_dependency_table *allowed_clock_voltage_table, - u32 clock, u32 *voltage) -{ - u32 i = 0; - - if (allowed_clock_voltage_table->count == 0) - return -EINVAL; - - for (i = 0; i < allowed_clock_voltage_table->count; i++) { - if (allowed_clock_voltage_table->entries[i].clk >= clock) { - *voltage = allowed_clock_voltage_table->entries[i].v; - return 0; - } - } - - *voltage = allowed_clock_voltage_table->entries[i-1].v; - - return 0; -} - -static u8 ci_get_sleep_divider_id_from_clock(u32 sclk, u32 min_sclk_in_sr) -{ - u32 i; - u32 tmp; - u32 min = max(min_sclk_in_sr, (u32)CISLAND_MINIMUM_ENGINE_CLOCK); - - if (sclk < min) - return 0; - - for (i = CISLAND_MAX_DEEPSLEEP_DIVIDER_ID; ; i--) { - tmp = sclk >> i; - if (tmp >= min || i == 0) - break; - } - - return (u8)i; -} - -static int ci_initial_switch_from_arb_f0_to_f1(struct amdgpu_device *adev) -{ - return ci_copy_and_switch_arb_sets(adev, MC_CG_ARB_FREQ_F0, MC_CG_ARB_FREQ_F1); -} - -static int ci_reset_to_default(struct amdgpu_device *adev) -{ - return (amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_ResetToDefaults) == PPSMC_Result_OK) ? - 0 : -EINVAL; -} - -static int ci_force_switch_to_arb_f0(struct amdgpu_device *adev) -{ - u32 tmp; - - tmp = (RREG32_SMC(ixSMC_SCRATCH9) & 0x0000ff00) >> 8; - - if (tmp == MC_CG_ARB_FREQ_F0) - return 0; - - return ci_copy_and_switch_arb_sets(adev, tmp, MC_CG_ARB_FREQ_F0); -} - -static void ci_register_patching_mc_arb(struct amdgpu_device *adev, - const u32 engine_clock, - const u32 memory_clock, - u32 *dram_timimg2) -{ - bool patch; - u32 tmp, tmp2; - - tmp = RREG32(mmMC_SEQ_MISC0); - patch = ((tmp & 0x0000f00) == 0x300) ? true : false; - - if (patch && - ((adev->pdev->device == 0x67B0) || - (adev->pdev->device == 0x67B1))) { - if ((memory_clock > 100000) && (memory_clock <= 125000)) { - tmp2 = (((0x31 * engine_clock) / 125000) - 1) & 0xff; - *dram_timimg2 &= ~0x00ff0000; - *dram_timimg2 |= tmp2 << 16; - } else if ((memory_clock > 125000) && (memory_clock <= 137500)) { - tmp2 = (((0x36 * engine_clock) / 137500) - 1) & 0xff; - *dram_timimg2 &= ~0x00ff0000; - *dram_timimg2 |= tmp2 << 16; - } - } -} - -static int ci_populate_memory_timing_parameters(struct amdgpu_device *adev, - u32 sclk, - u32 mclk, - SMU7_Discrete_MCArbDramTimingTableEntry *arb_regs) -{ - u32 dram_timing; - u32 dram_timing2; - u32 burst_time; - - amdgpu_atombios_set_engine_dram_timings(adev, sclk, mclk); - - dram_timing = RREG32(mmMC_ARB_DRAM_TIMING); - dram_timing2 = RREG32(mmMC_ARB_DRAM_TIMING2); - burst_time = RREG32(mmMC_ARB_BURST_TIME) & MC_ARB_BURST_TIME__STATE0_MASK; - - ci_register_patching_mc_arb(adev, sclk, mclk, &dram_timing2); - - arb_regs->McArbDramTiming = cpu_to_be32(dram_timing); - arb_regs->McArbDramTiming2 = cpu_to_be32(dram_timing2); - arb_regs->McArbBurstTime = (u8)burst_time; - - return 0; -} - -static int ci_do_program_memory_timing_parameters(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - SMU7_Discrete_MCArbDramTimingTable arb_regs; - u32 i, j; - int ret = 0; - - memset(&arb_regs, 0, sizeof(SMU7_Discrete_MCArbDramTimingTable)); - - for (i = 0; i < pi->dpm_table.sclk_table.count; i++) { - for (j = 0; j < pi->dpm_table.mclk_table.count; j++) { - ret = ci_populate_memory_timing_parameters(adev, - pi->dpm_table.sclk_table.dpm_levels[i].value, - pi->dpm_table.mclk_table.dpm_levels[j].value, - &arb_regs.entries[i][j]); - if (ret) - break; - } - } - - if (ret == 0) - ret = amdgpu_ci_copy_bytes_to_smc(adev, - pi->arb_table_start, - (u8 *)&arb_regs, - sizeof(SMU7_Discrete_MCArbDramTimingTable), - pi->sram_end); - - return ret; -} - -static int ci_program_memory_timing_parameters(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - if (pi->need_update_smu7_dpm_table == 0) - return 0; - - return ci_do_program_memory_timing_parameters(adev); -} - -static void ci_populate_smc_initial_state(struct amdgpu_device *adev, - struct amdgpu_ps *amdgpu_boot_state) -{ - struct ci_ps *boot_state = ci_get_ps(amdgpu_boot_state); - struct ci_power_info *pi = ci_get_pi(adev); - u32 level = 0; - - for (level = 0; level < adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count; level++) { - if (adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[level].clk >= - boot_state->performance_levels[0].sclk) { - pi->smc_state_table.GraphicsBootLevel = level; - break; - } - } - - for (level = 0; level < adev->pm.dpm.dyn_state.vddc_dependency_on_mclk.count; level++) { - if (adev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries[level].clk >= - boot_state->performance_levels[0].mclk) { - pi->smc_state_table.MemoryBootLevel = level; - break; - } - } -} - -static u32 ci_get_dpm_level_enable_mask_value(struct ci_single_dpm_table *dpm_table) -{ - u32 i; - u32 mask_value = 0; - - for (i = dpm_table->count; i > 0; i--) { - mask_value = mask_value << 1; - if (dpm_table->dpm_levels[i-1].enabled) - mask_value |= 0x1; - else - mask_value &= 0xFFFFFFFE; - } - - return mask_value; -} - -static void ci_populate_smc_link_level(struct amdgpu_device *adev, - SMU7_Discrete_DpmTable *table) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_dpm_table *dpm_table = &pi->dpm_table; - u32 i; - - for (i = 0; i < dpm_table->pcie_speed_table.count; i++) { - table->LinkLevel[i].PcieGenSpeed = - (u8)dpm_table->pcie_speed_table.dpm_levels[i].value; - table->LinkLevel[i].PcieLaneCount = - amdgpu_encode_pci_lane_width(dpm_table->pcie_speed_table.dpm_levels[i].param1); - table->LinkLevel[i].EnabledForActivity = 1; - table->LinkLevel[i].DownT = cpu_to_be32(5); - table->LinkLevel[i].UpT = cpu_to_be32(30); - } - - pi->smc_state_table.LinkLevelCount = (u8)dpm_table->pcie_speed_table.count; - pi->dpm_level_enable_mask.pcie_dpm_enable_mask = - ci_get_dpm_level_enable_mask_value(&dpm_table->pcie_speed_table); -} - -static int ci_populate_smc_uvd_level(struct amdgpu_device *adev, - SMU7_Discrete_DpmTable *table) -{ - u32 count; - struct atom_clock_dividers dividers; - int ret = -EINVAL; - - table->UvdLevelCount = - adev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.count; - - for (count = 0; count < table->UvdLevelCount; count++) { - table->UvdLevel[count].VclkFrequency = - adev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries[count].vclk; - table->UvdLevel[count].DclkFrequency = - adev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries[count].dclk; - table->UvdLevel[count].MinVddc = - adev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries[count].v * VOLTAGE_SCALE; - table->UvdLevel[count].MinVddcPhases = 1; - - ret = amdgpu_atombios_get_clock_dividers(adev, - COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, - table->UvdLevel[count].VclkFrequency, false, ÷rs); - if (ret) - return ret; - - table->UvdLevel[count].VclkDivider = (u8)dividers.post_divider; - - ret = amdgpu_atombios_get_clock_dividers(adev, - COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, - table->UvdLevel[count].DclkFrequency, false, ÷rs); - if (ret) - return ret; - - table->UvdLevel[count].DclkDivider = (u8)dividers.post_divider; - - table->UvdLevel[count].VclkFrequency = cpu_to_be32(table->UvdLevel[count].VclkFrequency); - table->UvdLevel[count].DclkFrequency = cpu_to_be32(table->UvdLevel[count].DclkFrequency); - table->UvdLevel[count].MinVddc = cpu_to_be16(table->UvdLevel[count].MinVddc); - } - - return ret; -} - -static int ci_populate_smc_vce_level(struct amdgpu_device *adev, - SMU7_Discrete_DpmTable *table) -{ - u32 count; - struct atom_clock_dividers dividers; - int ret = -EINVAL; - - table->VceLevelCount = - adev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.count; - - for (count = 0; count < table->VceLevelCount; count++) { - table->VceLevel[count].Frequency = - adev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries[count].evclk; - table->VceLevel[count].MinVoltage = - (u16)adev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries[count].v * VOLTAGE_SCALE; - table->VceLevel[count].MinPhases = 1; - - ret = amdgpu_atombios_get_clock_dividers(adev, - COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, - table->VceLevel[count].Frequency, false, ÷rs); - if (ret) - return ret; - - table->VceLevel[count].Divider = (u8)dividers.post_divider; - - table->VceLevel[count].Frequency = cpu_to_be32(table->VceLevel[count].Frequency); - table->VceLevel[count].MinVoltage = cpu_to_be16(table->VceLevel[count].MinVoltage); - } - - return ret; - -} - -static int ci_populate_smc_acp_level(struct amdgpu_device *adev, - SMU7_Discrete_DpmTable *table) -{ - u32 count; - struct atom_clock_dividers dividers; - int ret = -EINVAL; - - table->AcpLevelCount = (u8) - (adev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.count); - - for (count = 0; count < table->AcpLevelCount; count++) { - table->AcpLevel[count].Frequency = - adev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.entries[count].clk; - table->AcpLevel[count].MinVoltage = - adev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.entries[count].v; - table->AcpLevel[count].MinPhases = 1; - - ret = amdgpu_atombios_get_clock_dividers(adev, - COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, - table->AcpLevel[count].Frequency, false, ÷rs); - if (ret) - return ret; - - table->AcpLevel[count].Divider = (u8)dividers.post_divider; - - table->AcpLevel[count].Frequency = cpu_to_be32(table->AcpLevel[count].Frequency); - table->AcpLevel[count].MinVoltage = cpu_to_be16(table->AcpLevel[count].MinVoltage); - } - - return ret; -} - -static int ci_populate_smc_samu_level(struct amdgpu_device *adev, - SMU7_Discrete_DpmTable *table) -{ - u32 count; - struct atom_clock_dividers dividers; - int ret = -EINVAL; - - table->SamuLevelCount = - adev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.count; - - for (count = 0; count < table->SamuLevelCount; count++) { - table->SamuLevel[count].Frequency = - adev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.entries[count].clk; - table->SamuLevel[count].MinVoltage = - adev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.entries[count].v * VOLTAGE_SCALE; - table->SamuLevel[count].MinPhases = 1; - - ret = amdgpu_atombios_get_clock_dividers(adev, - COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, - table->SamuLevel[count].Frequency, false, ÷rs); - if (ret) - return ret; - - table->SamuLevel[count].Divider = (u8)dividers.post_divider; - - table->SamuLevel[count].Frequency = cpu_to_be32(table->SamuLevel[count].Frequency); - table->SamuLevel[count].MinVoltage = cpu_to_be16(table->SamuLevel[count].MinVoltage); - } - - return ret; -} - -static int ci_calculate_mclk_params(struct amdgpu_device *adev, - u32 memory_clock, - SMU7_Discrete_MemoryLevel *mclk, - bool strobe_mode, - bool dll_state_on) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 dll_cntl = pi->clock_registers.dll_cntl; - u32 mclk_pwrmgt_cntl = pi->clock_registers.mclk_pwrmgt_cntl; - u32 mpll_ad_func_cntl = pi->clock_registers.mpll_ad_func_cntl; - u32 mpll_dq_func_cntl = pi->clock_registers.mpll_dq_func_cntl; - u32 mpll_func_cntl = pi->clock_registers.mpll_func_cntl; - u32 mpll_func_cntl_1 = pi->clock_registers.mpll_func_cntl_1; - u32 mpll_func_cntl_2 = pi->clock_registers.mpll_func_cntl_2; - u32 mpll_ss1 = pi->clock_registers.mpll_ss1; - u32 mpll_ss2 = pi->clock_registers.mpll_ss2; - struct atom_mpll_param mpll_param; - int ret; - - ret = amdgpu_atombios_get_memory_pll_dividers(adev, memory_clock, strobe_mode, &mpll_param); - if (ret) - return ret; - - mpll_func_cntl &= ~MPLL_FUNC_CNTL__BWCTRL_MASK; - mpll_func_cntl |= (mpll_param.bwcntl << MPLL_FUNC_CNTL__BWCTRL__SHIFT); - - mpll_func_cntl_1 &= ~(MPLL_FUNC_CNTL_1__CLKF_MASK | MPLL_FUNC_CNTL_1__CLKFRAC_MASK | - MPLL_FUNC_CNTL_1__VCO_MODE_MASK); - mpll_func_cntl_1 |= (mpll_param.clkf) << MPLL_FUNC_CNTL_1__CLKF__SHIFT | - (mpll_param.clkfrac << MPLL_FUNC_CNTL_1__CLKFRAC__SHIFT) | - (mpll_param.vco_mode << MPLL_FUNC_CNTL_1__VCO_MODE__SHIFT); - - mpll_ad_func_cntl &= ~MPLL_AD_FUNC_CNTL__YCLK_POST_DIV_MASK; - mpll_ad_func_cntl |= (mpll_param.post_div << MPLL_AD_FUNC_CNTL__YCLK_POST_DIV__SHIFT); - - if (adev->gmc.vram_type == AMDGPU_VRAM_TYPE_GDDR5) { - mpll_dq_func_cntl &= ~(MPLL_DQ_FUNC_CNTL__YCLK_SEL_MASK | - MPLL_AD_FUNC_CNTL__YCLK_POST_DIV_MASK); - mpll_dq_func_cntl |= (mpll_param.yclk_sel << MPLL_DQ_FUNC_CNTL__YCLK_SEL__SHIFT) | - (mpll_param.post_div << MPLL_AD_FUNC_CNTL__YCLK_POST_DIV__SHIFT); - } - - if (pi->caps_mclk_ss_support) { - struct amdgpu_atom_ss ss; - u32 freq_nom; - u32 tmp; - u32 reference_clock = adev->clock.mpll.reference_freq; - - if (mpll_param.qdr == 1) - freq_nom = memory_clock * 4 * (1 << mpll_param.post_div); - else - freq_nom = memory_clock * 2 * (1 << mpll_param.post_div); - - tmp = (freq_nom / reference_clock); - tmp = tmp * tmp; - if (amdgpu_atombios_get_asic_ss_info(adev, &ss, - ASIC_INTERNAL_MEMORY_SS, freq_nom)) { - u32 clks = reference_clock * 5 / ss.rate; - u32 clkv = (u32)((((131 * ss.percentage * ss.rate) / 100) * tmp) / freq_nom); - - mpll_ss1 &= ~MPLL_SS1__CLKV_MASK; - mpll_ss1 |= (clkv << MPLL_SS1__CLKV__SHIFT); - - mpll_ss2 &= ~MPLL_SS2__CLKS_MASK; - mpll_ss2 |= (clks << MPLL_SS2__CLKS__SHIFT); - } - } - - mclk_pwrmgt_cntl &= ~MCLK_PWRMGT_CNTL__DLL_SPEED_MASK; - mclk_pwrmgt_cntl |= (mpll_param.dll_speed << MCLK_PWRMGT_CNTL__DLL_SPEED__SHIFT); - - if (dll_state_on) - mclk_pwrmgt_cntl |= MCLK_PWRMGT_CNTL__MRDCK0_PDNB_MASK | - MCLK_PWRMGT_CNTL__MRDCK1_PDNB_MASK; - else - mclk_pwrmgt_cntl &= ~(MCLK_PWRMGT_CNTL__MRDCK0_PDNB_MASK | - MCLK_PWRMGT_CNTL__MRDCK1_PDNB_MASK); - - mclk->MclkFrequency = memory_clock; - mclk->MpllFuncCntl = mpll_func_cntl; - mclk->MpllFuncCntl_1 = mpll_func_cntl_1; - mclk->MpllFuncCntl_2 = mpll_func_cntl_2; - mclk->MpllAdFuncCntl = mpll_ad_func_cntl; - mclk->MpllDqFuncCntl = mpll_dq_func_cntl; - mclk->MclkPwrmgtCntl = mclk_pwrmgt_cntl; - mclk->DllCntl = dll_cntl; - mclk->MpllSs1 = mpll_ss1; - mclk->MpllSs2 = mpll_ss2; - - return 0; -} - -static int ci_populate_single_memory_level(struct amdgpu_device *adev, - u32 memory_clock, - SMU7_Discrete_MemoryLevel *memory_level) -{ - struct ci_power_info *pi = ci_get_pi(adev); - int ret; - bool dll_state_on; - - if (adev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries) { - ret = ci_get_dependency_volt_by_clk(adev, - &adev->pm.dpm.dyn_state.vddc_dependency_on_mclk, - memory_clock, &memory_level->MinVddc); - if (ret) - return ret; - } - - if (adev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries) { - ret = ci_get_dependency_volt_by_clk(adev, - &adev->pm.dpm.dyn_state.vddci_dependency_on_mclk, - memory_clock, &memory_level->MinVddci); - if (ret) - return ret; - } - - if (adev->pm.dpm.dyn_state.mvdd_dependency_on_mclk.entries) { - ret = ci_get_dependency_volt_by_clk(adev, - &adev->pm.dpm.dyn_state.mvdd_dependency_on_mclk, - memory_clock, &memory_level->MinMvdd); - if (ret) - return ret; - } - - memory_level->MinVddcPhases = 1; - - if (pi->vddc_phase_shed_control) - ci_populate_phase_value_based_on_mclk(adev, - &adev->pm.dpm.dyn_state.phase_shedding_limits_table, - memory_clock, - &memory_level->MinVddcPhases); - - memory_level->EnabledForActivity = 1; - memory_level->EnabledForThrottle = 1; - memory_level->UpH = 0; - memory_level->DownH = 100; - memory_level->VoltageDownH = 0; - memory_level->ActivityLevel = (u16)pi->mclk_activity_target; - - memory_level->StutterEnable = false; - memory_level->StrobeEnable = false; - memory_level->EdcReadEnable = false; - memory_level->EdcWriteEnable = false; - memory_level->RttEnable = false; - - memory_level->DisplayWatermark = PPSMC_DISPLAY_WATERMARK_LOW; - - if (pi->mclk_stutter_mode_threshold && - (memory_clock <= pi->mclk_stutter_mode_threshold) && - (!pi->uvd_enabled) && - (RREG32(mmDPG_PIPE_STUTTER_CONTROL) & DPG_PIPE_STUTTER_CONTROL__STUTTER_ENABLE_MASK) && - (adev->pm.dpm.new_active_crtc_count <= 2)) - memory_level->StutterEnable = true; - - if (pi->mclk_strobe_mode_threshold && - (memory_clock <= pi->mclk_strobe_mode_threshold)) - memory_level->StrobeEnable = 1; - - if (adev->gmc.vram_type == AMDGPU_VRAM_TYPE_GDDR5) { - memory_level->StrobeRatio = - ci_get_mclk_frequency_ratio(memory_clock, memory_level->StrobeEnable); - if (pi->mclk_edc_enable_threshold && - (memory_clock > pi->mclk_edc_enable_threshold)) - memory_level->EdcReadEnable = true; - - if (pi->mclk_edc_wr_enable_threshold && - (memory_clock > pi->mclk_edc_wr_enable_threshold)) - memory_level->EdcWriteEnable = true; - - if (memory_level->StrobeEnable) { - if (ci_get_mclk_frequency_ratio(memory_clock, true) >= - ((RREG32(mmMC_SEQ_MISC7) >> 16) & 0xf)) - dll_state_on = ((RREG32(mmMC_SEQ_MISC5) >> 1) & 0x1) ? true : false; - else - dll_state_on = ((RREG32(mmMC_SEQ_MISC6) >> 1) & 0x1) ? true : false; - } else { - dll_state_on = pi->dll_default_on; - } - } else { - memory_level->StrobeRatio = ci_get_ddr3_mclk_frequency_ratio(memory_clock); - dll_state_on = ((RREG32(mmMC_SEQ_MISC5) >> 1) & 0x1) ? true : false; - } - - ret = ci_calculate_mclk_params(adev, memory_clock, memory_level, memory_level->StrobeEnable, dll_state_on); - if (ret) - return ret; - - memory_level->MinVddc = cpu_to_be32(memory_level->MinVddc * VOLTAGE_SCALE); - memory_level->MinVddcPhases = cpu_to_be32(memory_level->MinVddcPhases); - memory_level->MinVddci = cpu_to_be32(memory_level->MinVddci * VOLTAGE_SCALE); - memory_level->MinMvdd = cpu_to_be32(memory_level->MinMvdd * VOLTAGE_SCALE); - - memory_level->MclkFrequency = cpu_to_be32(memory_level->MclkFrequency); - memory_level->ActivityLevel = cpu_to_be16(memory_level->ActivityLevel); - memory_level->MpllFuncCntl = cpu_to_be32(memory_level->MpllFuncCntl); - memory_level->MpllFuncCntl_1 = cpu_to_be32(memory_level->MpllFuncCntl_1); - memory_level->MpllFuncCntl_2 = cpu_to_be32(memory_level->MpllFuncCntl_2); - memory_level->MpllAdFuncCntl = cpu_to_be32(memory_level->MpllAdFuncCntl); - memory_level->MpllDqFuncCntl = cpu_to_be32(memory_level->MpllDqFuncCntl); - memory_level->MclkPwrmgtCntl = cpu_to_be32(memory_level->MclkPwrmgtCntl); - memory_level->DllCntl = cpu_to_be32(memory_level->DllCntl); - memory_level->MpllSs1 = cpu_to_be32(memory_level->MpllSs1); - memory_level->MpllSs2 = cpu_to_be32(memory_level->MpllSs2); - - return 0; -} - -static int ci_populate_smc_acpi_level(struct amdgpu_device *adev, - SMU7_Discrete_DpmTable *table) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct atom_clock_dividers dividers; - SMU7_Discrete_VoltageLevel voltage_level; - u32 spll_func_cntl = pi->clock_registers.cg_spll_func_cntl; - u32 spll_func_cntl_2 = pi->clock_registers.cg_spll_func_cntl_2; - u32 dll_cntl = pi->clock_registers.dll_cntl; - u32 mclk_pwrmgt_cntl = pi->clock_registers.mclk_pwrmgt_cntl; - int ret; - - table->ACPILevel.Flags &= ~PPSMC_SWSTATE_FLAG_DC; - - if (pi->acpi_vddc) - table->ACPILevel.MinVddc = cpu_to_be32(pi->acpi_vddc * VOLTAGE_SCALE); - else - table->ACPILevel.MinVddc = cpu_to_be32(pi->min_vddc_in_pp_table * VOLTAGE_SCALE); - - table->ACPILevel.MinVddcPhases = pi->vddc_phase_shed_control ? 0 : 1; - - table->ACPILevel.SclkFrequency = adev->clock.spll.reference_freq; - - ret = amdgpu_atombios_get_clock_dividers(adev, - COMPUTE_GPUCLK_INPUT_FLAG_SCLK, - table->ACPILevel.SclkFrequency, false, ÷rs); - if (ret) - return ret; - - table->ACPILevel.SclkDid = (u8)dividers.post_divider; - table->ACPILevel.DisplayWatermark = PPSMC_DISPLAY_WATERMARK_LOW; - table->ACPILevel.DeepSleepDivId = 0; - - spll_func_cntl &= ~CG_SPLL_FUNC_CNTL__SPLL_PWRON_MASK; - spll_func_cntl |= CG_SPLL_FUNC_CNTL__SPLL_RESET_MASK; - - spll_func_cntl_2 &= ~CG_SPLL_FUNC_CNTL_2__SCLK_MUX_SEL_MASK; - spll_func_cntl_2 |= (4 << CG_SPLL_FUNC_CNTL_2__SCLK_MUX_SEL__SHIFT); - - table->ACPILevel.CgSpllFuncCntl = spll_func_cntl; - table->ACPILevel.CgSpllFuncCntl2 = spll_func_cntl_2; - table->ACPILevel.CgSpllFuncCntl3 = pi->clock_registers.cg_spll_func_cntl_3; - table->ACPILevel.CgSpllFuncCntl4 = pi->clock_registers.cg_spll_func_cntl_4; - table->ACPILevel.SpllSpreadSpectrum = pi->clock_registers.cg_spll_spread_spectrum; - table->ACPILevel.SpllSpreadSpectrum2 = pi->clock_registers.cg_spll_spread_spectrum_2; - table->ACPILevel.CcPwrDynRm = 0; - table->ACPILevel.CcPwrDynRm1 = 0; - - table->ACPILevel.Flags = cpu_to_be32(table->ACPILevel.Flags); - table->ACPILevel.MinVddcPhases = cpu_to_be32(table->ACPILevel.MinVddcPhases); - table->ACPILevel.SclkFrequency = cpu_to_be32(table->ACPILevel.SclkFrequency); - table->ACPILevel.CgSpllFuncCntl = cpu_to_be32(table->ACPILevel.CgSpllFuncCntl); - table->ACPILevel.CgSpllFuncCntl2 = cpu_to_be32(table->ACPILevel.CgSpllFuncCntl2); - table->ACPILevel.CgSpllFuncCntl3 = cpu_to_be32(table->ACPILevel.CgSpllFuncCntl3); - table->ACPILevel.CgSpllFuncCntl4 = cpu_to_be32(table->ACPILevel.CgSpllFuncCntl4); - table->ACPILevel.SpllSpreadSpectrum = cpu_to_be32(table->ACPILevel.SpllSpreadSpectrum); - table->ACPILevel.SpllSpreadSpectrum2 = cpu_to_be32(table->ACPILevel.SpllSpreadSpectrum2); - table->ACPILevel.CcPwrDynRm = cpu_to_be32(table->ACPILevel.CcPwrDynRm); - table->ACPILevel.CcPwrDynRm1 = cpu_to_be32(table->ACPILevel.CcPwrDynRm1); - - table->MemoryACPILevel.MinVddc = table->ACPILevel.MinVddc; - table->MemoryACPILevel.MinVddcPhases = table->ACPILevel.MinVddcPhases; - - if (pi->vddci_control != CISLANDS_VOLTAGE_CONTROL_NONE) { - if (pi->acpi_vddci) - table->MemoryACPILevel.MinVddci = - cpu_to_be32(pi->acpi_vddci * VOLTAGE_SCALE); - else - table->MemoryACPILevel.MinVddci = - cpu_to_be32(pi->min_vddci_in_pp_table * VOLTAGE_SCALE); - } - - if (ci_populate_mvdd_value(adev, 0, &voltage_level)) - table->MemoryACPILevel.MinMvdd = 0; - else - table->MemoryACPILevel.MinMvdd = - cpu_to_be32(voltage_level.Voltage * VOLTAGE_SCALE); - - mclk_pwrmgt_cntl |= MCLK_PWRMGT_CNTL__MRDCK0_RESET_MASK | - MCLK_PWRMGT_CNTL__MRDCK1_RESET_MASK; - mclk_pwrmgt_cntl &= ~(MCLK_PWRMGT_CNTL__MRDCK0_PDNB_MASK | - MCLK_PWRMGT_CNTL__MRDCK1_PDNB_MASK); - - dll_cntl &= ~(DLL_CNTL__MRDCK0_BYPASS_MASK | DLL_CNTL__MRDCK1_BYPASS_MASK); - - table->MemoryACPILevel.DllCntl = cpu_to_be32(dll_cntl); - table->MemoryACPILevel.MclkPwrmgtCntl = cpu_to_be32(mclk_pwrmgt_cntl); - table->MemoryACPILevel.MpllAdFuncCntl = - cpu_to_be32(pi->clock_registers.mpll_ad_func_cntl); - table->MemoryACPILevel.MpllDqFuncCntl = - cpu_to_be32(pi->clock_registers.mpll_dq_func_cntl); - table->MemoryACPILevel.MpllFuncCntl = - cpu_to_be32(pi->clock_registers.mpll_func_cntl); - table->MemoryACPILevel.MpllFuncCntl_1 = - cpu_to_be32(pi->clock_registers.mpll_func_cntl_1); - table->MemoryACPILevel.MpllFuncCntl_2 = - cpu_to_be32(pi->clock_registers.mpll_func_cntl_2); - table->MemoryACPILevel.MpllSs1 = cpu_to_be32(pi->clock_registers.mpll_ss1); - table->MemoryACPILevel.MpllSs2 = cpu_to_be32(pi->clock_registers.mpll_ss2); - - table->MemoryACPILevel.EnabledForThrottle = 0; - table->MemoryACPILevel.EnabledForActivity = 0; - table->MemoryACPILevel.UpH = 0; - table->MemoryACPILevel.DownH = 100; - table->MemoryACPILevel.VoltageDownH = 0; - table->MemoryACPILevel.ActivityLevel = - cpu_to_be16((u16)pi->mclk_activity_target); - - table->MemoryACPILevel.StutterEnable = false; - table->MemoryACPILevel.StrobeEnable = false; - table->MemoryACPILevel.EdcReadEnable = false; - table->MemoryACPILevel.EdcWriteEnable = false; - table->MemoryACPILevel.RttEnable = false; - - return 0; -} - - -static int ci_enable_ulv(struct amdgpu_device *adev, bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_ulv_parm *ulv = &pi->ulv; - - if (ulv->supported) { - if (enable) - return (amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_EnableULV) == PPSMC_Result_OK) ? - 0 : -EINVAL; - else - return (amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_DisableULV) == PPSMC_Result_OK) ? - 0 : -EINVAL; - } - - return 0; -} - -static int ci_populate_ulv_level(struct amdgpu_device *adev, - SMU7_Discrete_Ulv *state) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u16 ulv_voltage = adev->pm.dpm.backbias_response_time; - - state->CcPwrDynRm = 0; - state->CcPwrDynRm1 = 0; - - if (ulv_voltage == 0) { - pi->ulv.supported = false; - return 0; - } - - if (pi->voltage_control != CISLANDS_VOLTAGE_CONTROL_BY_SVID2) { - if (ulv_voltage > adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[0].v) - state->VddcOffset = 0; - else - state->VddcOffset = - adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[0].v - ulv_voltage; - } else { - if (ulv_voltage > adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[0].v) - state->VddcOffsetVid = 0; - else - state->VddcOffsetVid = (u8) - ((adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[0].v - ulv_voltage) * - VOLTAGE_VID_OFFSET_SCALE2 / VOLTAGE_VID_OFFSET_SCALE1); - } - state->VddcPhase = pi->vddc_phase_shed_control ? 0 : 1; - - state->CcPwrDynRm = cpu_to_be32(state->CcPwrDynRm); - state->CcPwrDynRm1 = cpu_to_be32(state->CcPwrDynRm1); - state->VddcOffset = cpu_to_be16(state->VddcOffset); - - return 0; -} - -static int ci_calculate_sclk_params(struct amdgpu_device *adev, - u32 engine_clock, - SMU7_Discrete_GraphicsLevel *sclk) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct atom_clock_dividers dividers; - u32 spll_func_cntl_3 = pi->clock_registers.cg_spll_func_cntl_3; - u32 spll_func_cntl_4 = pi->clock_registers.cg_spll_func_cntl_4; - u32 cg_spll_spread_spectrum = pi->clock_registers.cg_spll_spread_spectrum; - u32 cg_spll_spread_spectrum_2 = pi->clock_registers.cg_spll_spread_spectrum_2; - u32 reference_clock = adev->clock.spll.reference_freq; - u32 reference_divider; - u32 fbdiv; - int ret; - - ret = amdgpu_atombios_get_clock_dividers(adev, - COMPUTE_GPUCLK_INPUT_FLAG_SCLK, - engine_clock, false, ÷rs); - if (ret) - return ret; - - reference_divider = 1 + dividers.ref_div; - fbdiv = dividers.fb_div & 0x3FFFFFF; - - spll_func_cntl_3 &= ~CG_SPLL_FUNC_CNTL_3__SPLL_FB_DIV_MASK; - spll_func_cntl_3 |= (fbdiv << CG_SPLL_FUNC_CNTL_3__SPLL_FB_DIV__SHIFT); - spll_func_cntl_3 |= CG_SPLL_FUNC_CNTL_3__SPLL_DITHEN_MASK; - - if (pi->caps_sclk_ss_support) { - struct amdgpu_atom_ss ss; - u32 vco_freq = engine_clock * dividers.post_div; - - if (amdgpu_atombios_get_asic_ss_info(adev, &ss, - ASIC_INTERNAL_ENGINE_SS, vco_freq)) { - u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate); - u32 clk_v = 4 * ss.percentage * fbdiv / (clk_s * 10000); - - cg_spll_spread_spectrum &= ~(CG_SPLL_SPREAD_SPECTRUM__CLKS_MASK | CG_SPLL_SPREAD_SPECTRUM__SSEN_MASK); - cg_spll_spread_spectrum |= (clk_s << CG_SPLL_SPREAD_SPECTRUM__CLKS__SHIFT); - cg_spll_spread_spectrum |= (1 << CG_SPLL_SPREAD_SPECTRUM__SSEN__SHIFT); - - cg_spll_spread_spectrum_2 &= ~CG_SPLL_SPREAD_SPECTRUM_2__CLKV_MASK; - cg_spll_spread_spectrum_2 |= (clk_v << CG_SPLL_SPREAD_SPECTRUM_2__CLKV__SHIFT); - } - } - - sclk->SclkFrequency = engine_clock; - sclk->CgSpllFuncCntl3 = spll_func_cntl_3; - sclk->CgSpllFuncCntl4 = spll_func_cntl_4; - sclk->SpllSpreadSpectrum = cg_spll_spread_spectrum; - sclk->SpllSpreadSpectrum2 = cg_spll_spread_spectrum_2; - sclk->SclkDid = (u8)dividers.post_divider; - - return 0; -} - -static int ci_populate_single_graphic_level(struct amdgpu_device *adev, - u32 engine_clock, - u16 sclk_activity_level_t, - SMU7_Discrete_GraphicsLevel *graphic_level) -{ - struct ci_power_info *pi = ci_get_pi(adev); - int ret; - - ret = ci_calculate_sclk_params(adev, engine_clock, graphic_level); - if (ret) - return ret; - - ret = ci_get_dependency_volt_by_clk(adev, - &adev->pm.dpm.dyn_state.vddc_dependency_on_sclk, - engine_clock, &graphic_level->MinVddc); - if (ret) - return ret; - - graphic_level->SclkFrequency = engine_clock; - - graphic_level->Flags = 0; - graphic_level->MinVddcPhases = 1; - - if (pi->vddc_phase_shed_control) - ci_populate_phase_value_based_on_sclk(adev, - &adev->pm.dpm.dyn_state.phase_shedding_limits_table, - engine_clock, - &graphic_level->MinVddcPhases); - - graphic_level->ActivityLevel = sclk_activity_level_t; - - graphic_level->CcPwrDynRm = 0; - graphic_level->CcPwrDynRm1 = 0; - graphic_level->EnabledForThrottle = 1; - graphic_level->UpH = 0; - graphic_level->DownH = 0; - graphic_level->VoltageDownH = 0; - graphic_level->PowerThrottle = 0; - - if (pi->caps_sclk_ds) - graphic_level->DeepSleepDivId = ci_get_sleep_divider_id_from_clock(engine_clock, - CISLAND_MINIMUM_ENGINE_CLOCK); - - graphic_level->DisplayWatermark = PPSMC_DISPLAY_WATERMARK_LOW; - - graphic_level->Flags = cpu_to_be32(graphic_level->Flags); - graphic_level->MinVddc = cpu_to_be32(graphic_level->MinVddc * VOLTAGE_SCALE); - graphic_level->MinVddcPhases = cpu_to_be32(graphic_level->MinVddcPhases); - graphic_level->SclkFrequency = cpu_to_be32(graphic_level->SclkFrequency); - graphic_level->ActivityLevel = cpu_to_be16(graphic_level->ActivityLevel); - graphic_level->CgSpllFuncCntl3 = cpu_to_be32(graphic_level->CgSpllFuncCntl3); - graphic_level->CgSpllFuncCntl4 = cpu_to_be32(graphic_level->CgSpllFuncCntl4); - graphic_level->SpllSpreadSpectrum = cpu_to_be32(graphic_level->SpllSpreadSpectrum); - graphic_level->SpllSpreadSpectrum2 = cpu_to_be32(graphic_level->SpllSpreadSpectrum2); - graphic_level->CcPwrDynRm = cpu_to_be32(graphic_level->CcPwrDynRm); - graphic_level->CcPwrDynRm1 = cpu_to_be32(graphic_level->CcPwrDynRm1); - - return 0; -} - -static int ci_populate_all_graphic_levels(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_dpm_table *dpm_table = &pi->dpm_table; - u32 level_array_address = pi->dpm_table_start + - offsetof(SMU7_Discrete_DpmTable, GraphicsLevel); - u32 level_array_size = sizeof(SMU7_Discrete_GraphicsLevel) * - SMU7_MAX_LEVELS_GRAPHICS; - SMU7_Discrete_GraphicsLevel *levels = pi->smc_state_table.GraphicsLevel; - u32 i, ret; - - memset(levels, 0, level_array_size); - - for (i = 0; i < dpm_table->sclk_table.count; i++) { - ret = ci_populate_single_graphic_level(adev, - dpm_table->sclk_table.dpm_levels[i].value, - (u16)pi->activity_target[i], - &pi->smc_state_table.GraphicsLevel[i]); - if (ret) - return ret; - if (i > 1) - pi->smc_state_table.GraphicsLevel[i].DeepSleepDivId = 0; - if (i == (dpm_table->sclk_table.count - 1)) - pi->smc_state_table.GraphicsLevel[i].DisplayWatermark = - PPSMC_DISPLAY_WATERMARK_HIGH; - } - pi->smc_state_table.GraphicsLevel[0].EnabledForActivity = 1; - - pi->smc_state_table.GraphicsDpmLevelCount = (u8)dpm_table->sclk_table.count; - pi->dpm_level_enable_mask.sclk_dpm_enable_mask = - ci_get_dpm_level_enable_mask_value(&dpm_table->sclk_table); - - ret = amdgpu_ci_copy_bytes_to_smc(adev, level_array_address, - (u8 *)levels, level_array_size, - pi->sram_end); - if (ret) - return ret; - - return 0; -} - -static int ci_populate_ulv_state(struct amdgpu_device *adev, - SMU7_Discrete_Ulv *ulv_level) -{ - return ci_populate_ulv_level(adev, ulv_level); -} - -static int ci_populate_all_memory_levels(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_dpm_table *dpm_table = &pi->dpm_table; - u32 level_array_address = pi->dpm_table_start + - offsetof(SMU7_Discrete_DpmTable, MemoryLevel); - u32 level_array_size = sizeof(SMU7_Discrete_MemoryLevel) * - SMU7_MAX_LEVELS_MEMORY; - SMU7_Discrete_MemoryLevel *levels = pi->smc_state_table.MemoryLevel; - u32 i, ret; - - memset(levels, 0, level_array_size); - - for (i = 0; i < dpm_table->mclk_table.count; i++) { - if (dpm_table->mclk_table.dpm_levels[i].value == 0) - return -EINVAL; - ret = ci_populate_single_memory_level(adev, - dpm_table->mclk_table.dpm_levels[i].value, - &pi->smc_state_table.MemoryLevel[i]); - if (ret) - return ret; - } - - if ((dpm_table->mclk_table.count >= 2) && - ((adev->pdev->device == 0x67B0) || (adev->pdev->device == 0x67B1))) { - pi->smc_state_table.MemoryLevel[1].MinVddc = - pi->smc_state_table.MemoryLevel[0].MinVddc; - pi->smc_state_table.MemoryLevel[1].MinVddcPhases = - pi->smc_state_table.MemoryLevel[0].MinVddcPhases; - } - - pi->smc_state_table.MemoryLevel[0].ActivityLevel = cpu_to_be16(0x1F); - - pi->smc_state_table.MemoryDpmLevelCount = (u8)dpm_table->mclk_table.count; - pi->dpm_level_enable_mask.mclk_dpm_enable_mask = - ci_get_dpm_level_enable_mask_value(&dpm_table->mclk_table); - - pi->smc_state_table.MemoryLevel[dpm_table->mclk_table.count - 1].DisplayWatermark = - PPSMC_DISPLAY_WATERMARK_HIGH; - - ret = amdgpu_ci_copy_bytes_to_smc(adev, level_array_address, - (u8 *)levels, level_array_size, - pi->sram_end); - if (ret) - return ret; - - return 0; -} - -static void ci_reset_single_dpm_table(struct amdgpu_device *adev, - struct ci_single_dpm_table* dpm_table, - u32 count) -{ - u32 i; - - dpm_table->count = count; - for (i = 0; i < MAX_REGULAR_DPM_NUMBER; i++) - dpm_table->dpm_levels[i].enabled = false; -} - -static void ci_setup_pcie_table_entry(struct ci_single_dpm_table* dpm_table, - u32 index, u32 pcie_gen, u32 pcie_lanes) -{ - dpm_table->dpm_levels[index].value = pcie_gen; - dpm_table->dpm_levels[index].param1 = pcie_lanes; - dpm_table->dpm_levels[index].enabled = true; -} - -static int ci_setup_default_pcie_tables(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - if (!pi->use_pcie_performance_levels && !pi->use_pcie_powersaving_levels) - return -EINVAL; - - if (pi->use_pcie_performance_levels && !pi->use_pcie_powersaving_levels) { - pi->pcie_gen_powersaving = pi->pcie_gen_performance; - pi->pcie_lane_powersaving = pi->pcie_lane_performance; - } else if (!pi->use_pcie_performance_levels && pi->use_pcie_powersaving_levels) { - pi->pcie_gen_performance = pi->pcie_gen_powersaving; - pi->pcie_lane_performance = pi->pcie_lane_powersaving; - } - - ci_reset_single_dpm_table(adev, - &pi->dpm_table.pcie_speed_table, - SMU7_MAX_LEVELS_LINK); - - if (adev->asic_type == CHIP_BONAIRE) - ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 0, - pi->pcie_gen_powersaving.min, - pi->pcie_lane_powersaving.max); - else - ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 0, - pi->pcie_gen_powersaving.min, - pi->pcie_lane_powersaving.min); - ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 1, - pi->pcie_gen_performance.min, - pi->pcie_lane_performance.min); - ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 2, - pi->pcie_gen_powersaving.min, - pi->pcie_lane_powersaving.max); - ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 3, - pi->pcie_gen_performance.min, - pi->pcie_lane_performance.max); - ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 4, - pi->pcie_gen_powersaving.max, - pi->pcie_lane_powersaving.max); - ci_setup_pcie_table_entry(&pi->dpm_table.pcie_speed_table, 5, - pi->pcie_gen_performance.max, - pi->pcie_lane_performance.max); - - pi->dpm_table.pcie_speed_table.count = 6; - - return 0; -} - -static int ci_setup_default_dpm_tables(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct amdgpu_clock_voltage_dependency_table *allowed_sclk_vddc_table = - &adev->pm.dpm.dyn_state.vddc_dependency_on_sclk; - struct amdgpu_clock_voltage_dependency_table *allowed_mclk_table = - &adev->pm.dpm.dyn_state.vddc_dependency_on_mclk; - struct amdgpu_cac_leakage_table *std_voltage_table = - &adev->pm.dpm.dyn_state.cac_leakage_table; - u32 i; - - if (allowed_sclk_vddc_table == NULL) - return -EINVAL; - if (allowed_sclk_vddc_table->count < 1) - return -EINVAL; - if (allowed_mclk_table == NULL) - return -EINVAL; - if (allowed_mclk_table->count < 1) - return -EINVAL; - - memset(&pi->dpm_table, 0, sizeof(struct ci_dpm_table)); - - ci_reset_single_dpm_table(adev, - &pi->dpm_table.sclk_table, - SMU7_MAX_LEVELS_GRAPHICS); - ci_reset_single_dpm_table(adev, - &pi->dpm_table.mclk_table, - SMU7_MAX_LEVELS_MEMORY); - ci_reset_single_dpm_table(adev, - &pi->dpm_table.vddc_table, - SMU7_MAX_LEVELS_VDDC); - ci_reset_single_dpm_table(adev, - &pi->dpm_table.vddci_table, - SMU7_MAX_LEVELS_VDDCI); - ci_reset_single_dpm_table(adev, - &pi->dpm_table.mvdd_table, - SMU7_MAX_LEVELS_MVDD); - - pi->dpm_table.sclk_table.count = 0; - for (i = 0; i < allowed_sclk_vddc_table->count; i++) { - if ((i == 0) || - (pi->dpm_table.sclk_table.dpm_levels[pi->dpm_table.sclk_table.count-1].value != - allowed_sclk_vddc_table->entries[i].clk)) { - pi->dpm_table.sclk_table.dpm_levels[pi->dpm_table.sclk_table.count].value = - allowed_sclk_vddc_table->entries[i].clk; - pi->dpm_table.sclk_table.dpm_levels[pi->dpm_table.sclk_table.count].enabled = - (i == 0) ? true : false; - pi->dpm_table.sclk_table.count++; - } - } - - pi->dpm_table.mclk_table.count = 0; - for (i = 0; i < allowed_mclk_table->count; i++) { - if ((i == 0) || - (pi->dpm_table.mclk_table.dpm_levels[pi->dpm_table.mclk_table.count-1].value != - allowed_mclk_table->entries[i].clk)) { - pi->dpm_table.mclk_table.dpm_levels[pi->dpm_table.mclk_table.count].value = - allowed_mclk_table->entries[i].clk; - pi->dpm_table.mclk_table.dpm_levels[pi->dpm_table.mclk_table.count].enabled = - (i == 0) ? true : false; - pi->dpm_table.mclk_table.count++; - } - } - - for (i = 0; i < allowed_sclk_vddc_table->count; i++) { - pi->dpm_table.vddc_table.dpm_levels[i].value = - allowed_sclk_vddc_table->entries[i].v; - pi->dpm_table.vddc_table.dpm_levels[i].param1 = - std_voltage_table->entries[i].leakage; - pi->dpm_table.vddc_table.dpm_levels[i].enabled = true; - } - pi->dpm_table.vddc_table.count = allowed_sclk_vddc_table->count; - - allowed_mclk_table = &adev->pm.dpm.dyn_state.vddci_dependency_on_mclk; - if (allowed_mclk_table) { - for (i = 0; i < allowed_mclk_table->count; i++) { - pi->dpm_table.vddci_table.dpm_levels[i].value = - allowed_mclk_table->entries[i].v; - pi->dpm_table.vddci_table.dpm_levels[i].enabled = true; - } - pi->dpm_table.vddci_table.count = allowed_mclk_table->count; - } - - allowed_mclk_table = &adev->pm.dpm.dyn_state.mvdd_dependency_on_mclk; - if (allowed_mclk_table) { - for (i = 0; i < allowed_mclk_table->count; i++) { - pi->dpm_table.mvdd_table.dpm_levels[i].value = - allowed_mclk_table->entries[i].v; - pi->dpm_table.mvdd_table.dpm_levels[i].enabled = true; - } - pi->dpm_table.mvdd_table.count = allowed_mclk_table->count; - } - - ci_setup_default_pcie_tables(adev); - - /* save a copy of the default DPM table */ - memcpy(&(pi->golden_dpm_table), &(pi->dpm_table), - sizeof(struct ci_dpm_table)); - - return 0; -} - -static int ci_find_boot_level(struct ci_single_dpm_table *table, - u32 value, u32 *boot_level) -{ - u32 i; - int ret = -EINVAL; - - for(i = 0; i < table->count; i++) { - if (value == table->dpm_levels[i].value) { - *boot_level = i; - ret = 0; - } - } - - return ret; -} - -static int ci_init_smc_table(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_ulv_parm *ulv = &pi->ulv; - struct amdgpu_ps *amdgpu_boot_state = adev->pm.dpm.boot_ps; - SMU7_Discrete_DpmTable *table = &pi->smc_state_table; - int ret; - - ret = ci_setup_default_dpm_tables(adev); - if (ret) - return ret; - - if (pi->voltage_control != CISLANDS_VOLTAGE_CONTROL_NONE) - ci_populate_smc_voltage_tables(adev, table); - - ci_init_fps_limits(adev); - - if (adev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC) - table->SystemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC; - - if (adev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC) - table->SystemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC; - - if (adev->gmc.vram_type == AMDGPU_VRAM_TYPE_GDDR5) - table->SystemFlags |= PPSMC_SYSTEMFLAG_GDDR5; - - if (ulv->supported) { - ret = ci_populate_ulv_state(adev, &pi->smc_state_table.Ulv); - if (ret) - return ret; - WREG32_SMC(ixCG_ULV_PARAMETER, ulv->cg_ulv_parameter); - } - - ret = ci_populate_all_graphic_levels(adev); - if (ret) - return ret; - - ret = ci_populate_all_memory_levels(adev); - if (ret) - return ret; - - ci_populate_smc_link_level(adev, table); - - ret = ci_populate_smc_acpi_level(adev, table); - if (ret) - return ret; - - ret = ci_populate_smc_vce_level(adev, table); - if (ret) - return ret; - - ret = ci_populate_smc_acp_level(adev, table); - if (ret) - return ret; - - ret = ci_populate_smc_samu_level(adev, table); - if (ret) - return ret; - - ret = ci_do_program_memory_timing_parameters(adev); - if (ret) - return ret; - - ret = ci_populate_smc_uvd_level(adev, table); - if (ret) - return ret; - - table->UvdBootLevel = 0; - table->VceBootLevel = 0; - table->AcpBootLevel = 0; - table->SamuBootLevel = 0; - table->GraphicsBootLevel = 0; - table->MemoryBootLevel = 0; - - ret = ci_find_boot_level(&pi->dpm_table.sclk_table, - pi->vbios_boot_state.sclk_bootup_value, - (u32 *)&pi->smc_state_table.GraphicsBootLevel); - - ret = ci_find_boot_level(&pi->dpm_table.mclk_table, - pi->vbios_boot_state.mclk_bootup_value, - (u32 *)&pi->smc_state_table.MemoryBootLevel); - - table->BootVddc = pi->vbios_boot_state.vddc_bootup_value; - table->BootVddci = pi->vbios_boot_state.vddci_bootup_value; - table->BootMVdd = pi->vbios_boot_state.mvdd_bootup_value; - - ci_populate_smc_initial_state(adev, amdgpu_boot_state); - - ret = ci_populate_bapm_parameters_in_dpm_table(adev); - if (ret) - return ret; - - table->UVDInterval = 1; - table->VCEInterval = 1; - table->ACPInterval = 1; - table->SAMUInterval = 1; - table->GraphicsVoltageChangeEnable = 1; - table->GraphicsThermThrottleEnable = 1; - table->GraphicsInterval = 1; - table->VoltageInterval = 1; - table->ThermalInterval = 1; - table->TemperatureLimitHigh = (u16)((pi->thermal_temp_setting.temperature_high * - CISLANDS_Q88_FORMAT_CONVERSION_UNIT) / 1000); - table->TemperatureLimitLow = (u16)((pi->thermal_temp_setting.temperature_low * - CISLANDS_Q88_FORMAT_CONVERSION_UNIT) / 1000); - table->MemoryVoltageChangeEnable = 1; - table->MemoryInterval = 1; - table->VoltageResponseTime = 0; - table->VddcVddciDelta = 4000; - table->PhaseResponseTime = 0; - table->MemoryThermThrottleEnable = 1; - table->PCIeBootLinkLevel = pi->dpm_table.pcie_speed_table.count - 1; - table->PCIeGenInterval = 1; - if (pi->voltage_control == CISLANDS_VOLTAGE_CONTROL_BY_SVID2) - table->SVI2Enable = 1; - else - table->SVI2Enable = 0; - - table->ThermGpio = 17; - table->SclkStepSize = 0x4000; - - table->SystemFlags = cpu_to_be32(table->SystemFlags); - table->SmioMaskVddcVid = cpu_to_be32(table->SmioMaskVddcVid); - table->SmioMaskVddcPhase = cpu_to_be32(table->SmioMaskVddcPhase); - table->SmioMaskVddciVid = cpu_to_be32(table->SmioMaskVddciVid); - table->SmioMaskMvddVid = cpu_to_be32(table->SmioMaskMvddVid); - table->SclkStepSize = cpu_to_be32(table->SclkStepSize); - table->TemperatureLimitHigh = cpu_to_be16(table->TemperatureLimitHigh); - table->TemperatureLimitLow = cpu_to_be16(table->TemperatureLimitLow); - table->VddcVddciDelta = cpu_to_be16(table->VddcVddciDelta); - table->VoltageResponseTime = cpu_to_be16(table->VoltageResponseTime); - table->PhaseResponseTime = cpu_to_be16(table->PhaseResponseTime); - table->BootVddc = cpu_to_be16(table->BootVddc * VOLTAGE_SCALE); - table->BootVddci = cpu_to_be16(table->BootVddci * VOLTAGE_SCALE); - table->BootMVdd = cpu_to_be16(table->BootMVdd * VOLTAGE_SCALE); - - ret = amdgpu_ci_copy_bytes_to_smc(adev, - pi->dpm_table_start + - offsetof(SMU7_Discrete_DpmTable, SystemFlags), - (u8 *)&table->SystemFlags, - sizeof(SMU7_Discrete_DpmTable) - 3 * sizeof(SMU7_PIDController), - pi->sram_end); - if (ret) - return ret; - - return 0; -} - -static void ci_trim_single_dpm_states(struct amdgpu_device *adev, - struct ci_single_dpm_table *dpm_table, - u32 low_limit, u32 high_limit) -{ - u32 i; - - for (i = 0; i < dpm_table->count; i++) { - if ((dpm_table->dpm_levels[i].value < low_limit) || - (dpm_table->dpm_levels[i].value > high_limit)) - dpm_table->dpm_levels[i].enabled = false; - else - dpm_table->dpm_levels[i].enabled = true; - } -} - -static void ci_trim_pcie_dpm_states(struct amdgpu_device *adev, - u32 speed_low, u32 lanes_low, - u32 speed_high, u32 lanes_high) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_single_dpm_table *pcie_table = &pi->dpm_table.pcie_speed_table; - u32 i, j; - - for (i = 0; i < pcie_table->count; i++) { - if ((pcie_table->dpm_levels[i].value < speed_low) || - (pcie_table->dpm_levels[i].param1 < lanes_low) || - (pcie_table->dpm_levels[i].value > speed_high) || - (pcie_table->dpm_levels[i].param1 > lanes_high)) - pcie_table->dpm_levels[i].enabled = false; - else - pcie_table->dpm_levels[i].enabled = true; - } - - for (i = 0; i < pcie_table->count; i++) { - if (pcie_table->dpm_levels[i].enabled) { - for (j = i + 1; j < pcie_table->count; j++) { - if (pcie_table->dpm_levels[j].enabled) { - if ((pcie_table->dpm_levels[i].value == pcie_table->dpm_levels[j].value) && - (pcie_table->dpm_levels[i].param1 == pcie_table->dpm_levels[j].param1)) - pcie_table->dpm_levels[j].enabled = false; - } - } - } - } -} - -static int ci_trim_dpm_states(struct amdgpu_device *adev, - struct amdgpu_ps *amdgpu_state) -{ - struct ci_ps *state = ci_get_ps(amdgpu_state); - struct ci_power_info *pi = ci_get_pi(adev); - u32 high_limit_count; - - if (state->performance_level_count < 1) - return -EINVAL; - - if (state->performance_level_count == 1) - high_limit_count = 0; - else - high_limit_count = 1; - - ci_trim_single_dpm_states(adev, - &pi->dpm_table.sclk_table, - state->performance_levels[0].sclk, - state->performance_levels[high_limit_count].sclk); - - ci_trim_single_dpm_states(adev, - &pi->dpm_table.mclk_table, - state->performance_levels[0].mclk, - state->performance_levels[high_limit_count].mclk); - - ci_trim_pcie_dpm_states(adev, - state->performance_levels[0].pcie_gen, - state->performance_levels[0].pcie_lane, - state->performance_levels[high_limit_count].pcie_gen, - state->performance_levels[high_limit_count].pcie_lane); - - return 0; -} - -static int ci_apply_disp_minimum_voltage_request(struct amdgpu_device *adev) -{ - struct amdgpu_clock_voltage_dependency_table *disp_voltage_table = - &adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk; - struct amdgpu_clock_voltage_dependency_table *vddc_table = - &adev->pm.dpm.dyn_state.vddc_dependency_on_sclk; - u32 requested_voltage = 0; - u32 i; - - if (disp_voltage_table == NULL) - return -EINVAL; - if (!disp_voltage_table->count) - return -EINVAL; - - for (i = 0; i < disp_voltage_table->count; i++) { - if (adev->clock.current_dispclk == disp_voltage_table->entries[i].clk) - requested_voltage = disp_voltage_table->entries[i].v; - } - - for (i = 0; i < vddc_table->count; i++) { - if (requested_voltage <= vddc_table->entries[i].v) { - requested_voltage = vddc_table->entries[i].v; - return (amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_VddC_Request, - requested_voltage * VOLTAGE_SCALE) == PPSMC_Result_OK) ? - 0 : -EINVAL; - } - } - - return -EINVAL; -} - -static int ci_upload_dpm_level_enable_mask(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - PPSMC_Result result; - - ci_apply_disp_minimum_voltage_request(adev); - - if (!pi->sclk_dpm_key_disabled) { - if (pi->dpm_level_enable_mask.sclk_dpm_enable_mask) { - result = amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_SCLKDPM_SetEnabledMask, - pi->dpm_level_enable_mask.sclk_dpm_enable_mask); - if (result != PPSMC_Result_OK) - return -EINVAL; - } - } - - if (!pi->mclk_dpm_key_disabled) { - if (pi->dpm_level_enable_mask.mclk_dpm_enable_mask) { - result = amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_MCLKDPM_SetEnabledMask, - pi->dpm_level_enable_mask.mclk_dpm_enable_mask); - if (result != PPSMC_Result_OK) - return -EINVAL; - } - } - -#if 0 - if (!pi->pcie_dpm_key_disabled) { - if (pi->dpm_level_enable_mask.pcie_dpm_enable_mask) { - result = amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_PCIeDPM_SetEnabledMask, - pi->dpm_level_enable_mask.pcie_dpm_enable_mask); - if (result != PPSMC_Result_OK) - return -EINVAL; - } - } -#endif - - return 0; -} - -static void ci_find_dpm_states_clocks_in_dpm_table(struct amdgpu_device *adev, - struct amdgpu_ps *amdgpu_state) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_ps *state = ci_get_ps(amdgpu_state); - struct ci_single_dpm_table *sclk_table = &pi->dpm_table.sclk_table; - u32 sclk = state->performance_levels[state->performance_level_count-1].sclk; - struct ci_single_dpm_table *mclk_table = &pi->dpm_table.mclk_table; - u32 mclk = state->performance_levels[state->performance_level_count-1].mclk; - u32 i; - - pi->need_update_smu7_dpm_table = 0; - - for (i = 0; i < sclk_table->count; i++) { - if (sclk == sclk_table->dpm_levels[i].value) - break; - } - - if (i >= sclk_table->count) { - pi->need_update_smu7_dpm_table |= DPMTABLE_OD_UPDATE_SCLK; - } else { - /* XXX check display min clock requirements */ - if (CISLAND_MINIMUM_ENGINE_CLOCK != CISLAND_MINIMUM_ENGINE_CLOCK) - pi->need_update_smu7_dpm_table |= DPMTABLE_UPDATE_SCLK; - } - - for (i = 0; i < mclk_table->count; i++) { - if (mclk == mclk_table->dpm_levels[i].value) - break; - } - - if (i >= mclk_table->count) - pi->need_update_smu7_dpm_table |= DPMTABLE_OD_UPDATE_MCLK; - - if (adev->pm.dpm.current_active_crtc_count != - adev->pm.dpm.new_active_crtc_count) - pi->need_update_smu7_dpm_table |= DPMTABLE_UPDATE_MCLK; -} - -static int ci_populate_and_upload_sclk_mclk_dpm_levels(struct amdgpu_device *adev, - struct amdgpu_ps *amdgpu_state) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_ps *state = ci_get_ps(amdgpu_state); - u32 sclk = state->performance_levels[state->performance_level_count-1].sclk; - u32 mclk = state->performance_levels[state->performance_level_count-1].mclk; - struct ci_dpm_table *dpm_table = &pi->dpm_table; - int ret; - - if (!pi->need_update_smu7_dpm_table) - return 0; - - if (pi->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_SCLK) - dpm_table->sclk_table.dpm_levels[dpm_table->sclk_table.count-1].value = sclk; - - if (pi->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK) - dpm_table->mclk_table.dpm_levels[dpm_table->mclk_table.count-1].value = mclk; - - if (pi->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK | DPMTABLE_UPDATE_SCLK)) { - ret = ci_populate_all_graphic_levels(adev); - if (ret) - return ret; - } - - if (pi->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_MCLK | DPMTABLE_UPDATE_MCLK)) { - ret = ci_populate_all_memory_levels(adev); - if (ret) - return ret; - } - - return 0; -} - -static int ci_enable_uvd_dpm(struct amdgpu_device *adev, bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - const struct amdgpu_clock_and_voltage_limits *max_limits; - int i; - - if (adev->pm.ac_power) - max_limits = &adev->pm.dpm.dyn_state.max_clock_voltage_on_ac; - else - max_limits = &adev->pm.dpm.dyn_state.max_clock_voltage_on_dc; - - if (enable) { - pi->dpm_level_enable_mask.uvd_dpm_enable_mask = 0; - - for (i = adev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.count - 1; i >= 0; i--) { - if (adev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.entries[i].v <= max_limits->vddc) { - pi->dpm_level_enable_mask.uvd_dpm_enable_mask |= 1 << i; - - if (!pi->caps_uvd_dpm) - break; - } - } - - amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_UVDDPM_SetEnabledMask, - pi->dpm_level_enable_mask.uvd_dpm_enable_mask); - - if (pi->last_mclk_dpm_enable_mask & 0x1) { - pi->uvd_enabled = true; - pi->dpm_level_enable_mask.mclk_dpm_enable_mask &= 0xFFFFFFFE; - amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_MCLKDPM_SetEnabledMask, - pi->dpm_level_enable_mask.mclk_dpm_enable_mask); - } - } else { - if (pi->uvd_enabled) { - pi->uvd_enabled = false; - pi->dpm_level_enable_mask.mclk_dpm_enable_mask |= 1; - amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_MCLKDPM_SetEnabledMask, - pi->dpm_level_enable_mask.mclk_dpm_enable_mask); - } - } - - return (amdgpu_ci_send_msg_to_smc(adev, enable ? - PPSMC_MSG_UVDDPM_Enable : PPSMC_MSG_UVDDPM_Disable) == PPSMC_Result_OK) ? - 0 : -EINVAL; -} - -static int ci_enable_vce_dpm(struct amdgpu_device *adev, bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - const struct amdgpu_clock_and_voltage_limits *max_limits; - int i; - - if (adev->pm.ac_power) - max_limits = &adev->pm.dpm.dyn_state.max_clock_voltage_on_ac; - else - max_limits = &adev->pm.dpm.dyn_state.max_clock_voltage_on_dc; - - if (enable) { - pi->dpm_level_enable_mask.vce_dpm_enable_mask = 0; - for (i = adev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.count - 1; i >= 0; i--) { - if (adev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table.entries[i].v <= max_limits->vddc) { - pi->dpm_level_enable_mask.vce_dpm_enable_mask |= 1 << i; - - if (!pi->caps_vce_dpm) - break; - } - } - - amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_VCEDPM_SetEnabledMask, - pi->dpm_level_enable_mask.vce_dpm_enable_mask); - } - - return (amdgpu_ci_send_msg_to_smc(adev, enable ? - PPSMC_MSG_VCEDPM_Enable : PPSMC_MSG_VCEDPM_Disable) == PPSMC_Result_OK) ? - 0 : -EINVAL; -} - -#if 0 -static int ci_enable_samu_dpm(struct amdgpu_device *adev, bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - const struct amdgpu_clock_and_voltage_limits *max_limits; - int i; - - if (adev->pm.ac_power) - max_limits = &adev->pm.dpm.dyn_state.max_clock_voltage_on_ac; - else - max_limits = &adev->pm.dpm.dyn_state.max_clock_voltage_on_dc; - - if (enable) { - pi->dpm_level_enable_mask.samu_dpm_enable_mask = 0; - for (i = adev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.count - 1; i >= 0; i--) { - if (adev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table.entries[i].v <= max_limits->vddc) { - pi->dpm_level_enable_mask.samu_dpm_enable_mask |= 1 << i; - - if (!pi->caps_samu_dpm) - break; - } - } - - amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_SAMUDPM_SetEnabledMask, - pi->dpm_level_enable_mask.samu_dpm_enable_mask); - } - return (amdgpu_ci_send_msg_to_smc(adev, enable ? - PPSMC_MSG_SAMUDPM_Enable : PPSMC_MSG_SAMUDPM_Disable) == PPSMC_Result_OK) ? - 0 : -EINVAL; -} - -static int ci_enable_acp_dpm(struct amdgpu_device *adev, bool enable) -{ - struct ci_power_info *pi = ci_get_pi(adev); - const struct amdgpu_clock_and_voltage_limits *max_limits; - int i; - - if (adev->pm.ac_power) - max_limits = &adev->pm.dpm.dyn_state.max_clock_voltage_on_ac; - else - max_limits = &adev->pm.dpm.dyn_state.max_clock_voltage_on_dc; - - if (enable) { - pi->dpm_level_enable_mask.acp_dpm_enable_mask = 0; - for (i = adev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.count - 1; i >= 0; i--) { - if (adev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table.entries[i].v <= max_limits->vddc) { - pi->dpm_level_enable_mask.acp_dpm_enable_mask |= 1 << i; - - if (!pi->caps_acp_dpm) - break; - } - } - - amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_ACPDPM_SetEnabledMask, - pi->dpm_level_enable_mask.acp_dpm_enable_mask); - } - - return (amdgpu_ci_send_msg_to_smc(adev, enable ? - PPSMC_MSG_ACPDPM_Enable : PPSMC_MSG_ACPDPM_Disable) == PPSMC_Result_OK) ? - 0 : -EINVAL; -} -#endif - -static int ci_update_uvd_dpm(struct amdgpu_device *adev, bool gate) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 tmp; - int ret = 0; - - if (!gate) { - /* turn the clocks on when decoding */ - if (pi->caps_uvd_dpm || - (adev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.count <= 0)) - pi->smc_state_table.UvdBootLevel = 0; - else - pi->smc_state_table.UvdBootLevel = - adev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table.count - 1; - - tmp = RREG32_SMC(ixDPM_TABLE_475); - tmp &= ~DPM_TABLE_475__UvdBootLevel_MASK; - tmp |= (pi->smc_state_table.UvdBootLevel << DPM_TABLE_475__UvdBootLevel__SHIFT); - WREG32_SMC(ixDPM_TABLE_475, tmp); - ret = ci_enable_uvd_dpm(adev, true); - } else { - ret = ci_enable_uvd_dpm(adev, false); - if (ret) - return ret; - } - - return ret; -} - -static u8 ci_get_vce_boot_level(struct amdgpu_device *adev) -{ - u8 i; - u32 min_evclk = 30000; /* ??? */ - struct amdgpu_vce_clock_voltage_dependency_table *table = - &adev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table; - - for (i = 0; i < table->count; i++) { - if (table->entries[i].evclk >= min_evclk) - return i; - } - - return table->count - 1; -} - -static int ci_update_vce_dpm(struct amdgpu_device *adev, - struct amdgpu_ps *amdgpu_new_state, - struct amdgpu_ps *amdgpu_current_state) -{ - struct ci_power_info *pi = ci_get_pi(adev); - int ret = 0; - u32 tmp; - - if (amdgpu_current_state->evclk != amdgpu_new_state->evclk) { - if (amdgpu_new_state->evclk) { - pi->smc_state_table.VceBootLevel = ci_get_vce_boot_level(adev); - tmp = RREG32_SMC(ixDPM_TABLE_475); - tmp &= ~DPM_TABLE_475__VceBootLevel_MASK; - tmp |= (pi->smc_state_table.VceBootLevel << DPM_TABLE_475__VceBootLevel__SHIFT); - WREG32_SMC(ixDPM_TABLE_475, tmp); - - ret = ci_enable_vce_dpm(adev, true); - } else { - ret = ci_enable_vce_dpm(adev, false); - if (ret) - return ret; - } - } - return ret; -} - -#if 0 -static int ci_update_samu_dpm(struct amdgpu_device *adev, bool gate) -{ - return ci_enable_samu_dpm(adev, gate); -} - -static int ci_update_acp_dpm(struct amdgpu_device *adev, bool gate) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 tmp; - - if (!gate) { - pi->smc_state_table.AcpBootLevel = 0; - - tmp = RREG32_SMC(ixDPM_TABLE_475); - tmp &= ~AcpBootLevel_MASK; - tmp |= AcpBootLevel(pi->smc_state_table.AcpBootLevel); - WREG32_SMC(ixDPM_TABLE_475, tmp); - } - - return ci_enable_acp_dpm(adev, !gate); -} -#endif - -static int ci_generate_dpm_level_enable_mask(struct amdgpu_device *adev, - struct amdgpu_ps *amdgpu_state) -{ - struct ci_power_info *pi = ci_get_pi(adev); - int ret; - - ret = ci_trim_dpm_states(adev, amdgpu_state); - if (ret) - return ret; - - pi->dpm_level_enable_mask.sclk_dpm_enable_mask = - ci_get_dpm_level_enable_mask_value(&pi->dpm_table.sclk_table); - pi->dpm_level_enable_mask.mclk_dpm_enable_mask = - ci_get_dpm_level_enable_mask_value(&pi->dpm_table.mclk_table); - pi->last_mclk_dpm_enable_mask = - pi->dpm_level_enable_mask.mclk_dpm_enable_mask; - if (pi->uvd_enabled) { - if (pi->dpm_level_enable_mask.mclk_dpm_enable_mask & 1) - pi->dpm_level_enable_mask.mclk_dpm_enable_mask &= 0xFFFFFFFE; - } - pi->dpm_level_enable_mask.pcie_dpm_enable_mask = - ci_get_dpm_level_enable_mask_value(&pi->dpm_table.pcie_speed_table); - - return 0; -} - -static u32 ci_get_lowest_enabled_level(struct amdgpu_device *adev, - u32 level_mask) -{ - u32 level = 0; - - while ((level_mask & (1 << level)) == 0) - level++; - - return level; -} - - -static int ci_dpm_force_performance_level(void *handle, - enum amd_dpm_forced_level level) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - u32 tmp, levels, i; - int ret; - - if (level == AMD_DPM_FORCED_LEVEL_HIGH) { - if ((!pi->pcie_dpm_key_disabled) && - pi->dpm_level_enable_mask.pcie_dpm_enable_mask) { - levels = 0; - tmp = pi->dpm_level_enable_mask.pcie_dpm_enable_mask; - while (tmp >>= 1) - levels++; - if (levels) { - ret = ci_dpm_force_state_pcie(adev, level); - if (ret) - return ret; - for (i = 0; i < adev->usec_timeout; i++) { - tmp = (RREG32_SMC(ixTARGET_AND_CURRENT_PROFILE_INDEX_1) & - TARGET_AND_CURRENT_PROFILE_INDEX_1__CURR_PCIE_INDEX_MASK) >> - TARGET_AND_CURRENT_PROFILE_INDEX_1__CURR_PCIE_INDEX__SHIFT; - if (tmp == levels) - break; - udelay(1); - } - } - } - if ((!pi->sclk_dpm_key_disabled) && - pi->dpm_level_enable_mask.sclk_dpm_enable_mask) { - levels = 0; - tmp = pi->dpm_level_enable_mask.sclk_dpm_enable_mask; - while (tmp >>= 1) - levels++; - if (levels) { - ret = ci_dpm_force_state_sclk(adev, levels); - if (ret) - return ret; - for (i = 0; i < adev->usec_timeout; i++) { - tmp = (RREG32_SMC(ixTARGET_AND_CURRENT_PROFILE_INDEX) & - TARGET_AND_CURRENT_PROFILE_INDEX__CURR_SCLK_INDEX_MASK) >> - TARGET_AND_CURRENT_PROFILE_INDEX__CURR_SCLK_INDEX__SHIFT; - if (tmp == levels) - break; - udelay(1); - } - } - } - if ((!pi->mclk_dpm_key_disabled) && - pi->dpm_level_enable_mask.mclk_dpm_enable_mask) { - levels = 0; - tmp = pi->dpm_level_enable_mask.mclk_dpm_enable_mask; - while (tmp >>= 1) - levels++; - if (levels) { - ret = ci_dpm_force_state_mclk(adev, levels); - if (ret) - return ret; - for (i = 0; i < adev->usec_timeout; i++) { - tmp = (RREG32_SMC(ixTARGET_AND_CURRENT_PROFILE_INDEX) & - TARGET_AND_CURRENT_PROFILE_INDEX__CURR_MCLK_INDEX_MASK) >> - TARGET_AND_CURRENT_PROFILE_INDEX__CURR_MCLK_INDEX__SHIFT; - if (tmp == levels) - break; - udelay(1); - } - } - } - } else if (level == AMD_DPM_FORCED_LEVEL_LOW) { - if ((!pi->sclk_dpm_key_disabled) && - pi->dpm_level_enable_mask.sclk_dpm_enable_mask) { - levels = ci_get_lowest_enabled_level(adev, - pi->dpm_level_enable_mask.sclk_dpm_enable_mask); - ret = ci_dpm_force_state_sclk(adev, levels); - if (ret) - return ret; - for (i = 0; i < adev->usec_timeout; i++) { - tmp = (RREG32_SMC(ixTARGET_AND_CURRENT_PROFILE_INDEX) & - TARGET_AND_CURRENT_PROFILE_INDEX__CURR_SCLK_INDEX_MASK) >> - TARGET_AND_CURRENT_PROFILE_INDEX__CURR_SCLK_INDEX__SHIFT; - if (tmp == levels) - break; - udelay(1); - } - } - if ((!pi->mclk_dpm_key_disabled) && - pi->dpm_level_enable_mask.mclk_dpm_enable_mask) { - levels = ci_get_lowest_enabled_level(adev, - pi->dpm_level_enable_mask.mclk_dpm_enable_mask); - ret = ci_dpm_force_state_mclk(adev, levels); - if (ret) - return ret; - for (i = 0; i < adev->usec_timeout; i++) { - tmp = (RREG32_SMC(ixTARGET_AND_CURRENT_PROFILE_INDEX) & - TARGET_AND_CURRENT_PROFILE_INDEX__CURR_MCLK_INDEX_MASK) >> - TARGET_AND_CURRENT_PROFILE_INDEX__CURR_MCLK_INDEX__SHIFT; - if (tmp == levels) - break; - udelay(1); - } - } - if ((!pi->pcie_dpm_key_disabled) && - pi->dpm_level_enable_mask.pcie_dpm_enable_mask) { - levels = ci_get_lowest_enabled_level(adev, - pi->dpm_level_enable_mask.pcie_dpm_enable_mask); - ret = ci_dpm_force_state_pcie(adev, levels); - if (ret) - return ret; - for (i = 0; i < adev->usec_timeout; i++) { - tmp = (RREG32_SMC(ixTARGET_AND_CURRENT_PROFILE_INDEX_1) & - TARGET_AND_CURRENT_PROFILE_INDEX_1__CURR_PCIE_INDEX_MASK) >> - TARGET_AND_CURRENT_PROFILE_INDEX_1__CURR_PCIE_INDEX__SHIFT; - if (tmp == levels) - break; - udelay(1); - } - } - } else if (level == AMD_DPM_FORCED_LEVEL_AUTO) { - if (!pi->pcie_dpm_key_disabled) { - PPSMC_Result smc_result; - - smc_result = amdgpu_ci_send_msg_to_smc(adev, - PPSMC_MSG_PCIeDPM_UnForceLevel); - if (smc_result != PPSMC_Result_OK) - return -EINVAL; - } - ret = ci_upload_dpm_level_enable_mask(adev); - if (ret) - return ret; - } - - adev->pm.dpm.forced_level = level; - - return 0; -} - -static int ci_set_mc_special_registers(struct amdgpu_device *adev, - struct ci_mc_reg_table *table) -{ - u8 i, j, k; - u32 temp_reg; - - for (i = 0, j = table->last; i < table->last; i++) { - if (j >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; - switch(table->mc_reg_address[i].s1) { - case mmMC_SEQ_MISC1: - temp_reg = RREG32(mmMC_PMG_CMD_EMRS); - table->mc_reg_address[j].s1 = mmMC_PMG_CMD_EMRS; - table->mc_reg_address[j].s0 = mmMC_SEQ_PMG_CMD_EMRS_LP; - for (k = 0; k < table->num_entries; k++) { - table->mc_reg_table_entry[k].mc_data[j] = - ((temp_reg & 0xffff0000)) | ((table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16); - } - j++; - - if (j >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; - temp_reg = RREG32(mmMC_PMG_CMD_MRS); - table->mc_reg_address[j].s1 = mmMC_PMG_CMD_MRS; - table->mc_reg_address[j].s0 = mmMC_SEQ_PMG_CMD_MRS_LP; - for (k = 0; k < table->num_entries; k++) { - table->mc_reg_table_entry[k].mc_data[j] = - (temp_reg & 0xffff0000) | (table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff); - if (adev->gmc.vram_type != AMDGPU_VRAM_TYPE_GDDR5) - table->mc_reg_table_entry[k].mc_data[j] |= 0x100; - } - j++; - - if (adev->gmc.vram_type != AMDGPU_VRAM_TYPE_GDDR5) { - if (j >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; - table->mc_reg_address[j].s1 = mmMC_PMG_AUTO_CMD; - table->mc_reg_address[j].s0 = mmMC_PMG_AUTO_CMD; - for (k = 0; k < table->num_entries; k++) { - table->mc_reg_table_entry[k].mc_data[j] = - (table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16; - } - j++; - } - break; - case mmMC_SEQ_RESERVE_M: - temp_reg = RREG32(mmMC_PMG_CMD_MRS1); - table->mc_reg_address[j].s1 = mmMC_PMG_CMD_MRS1; - table->mc_reg_address[j].s0 = mmMC_SEQ_PMG_CMD_MRS1_LP; - for (k = 0; k < table->num_entries; k++) { - table->mc_reg_table_entry[k].mc_data[j] = - (temp_reg & 0xffff0000) | (table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff); - } - j++; - break; - default: - break; - } - - } - - table->last = j; - - return 0; -} - -static bool ci_check_s0_mc_reg_index(u16 in_reg, u16 *out_reg) -{ - bool result = true; - - switch(in_reg) { - case mmMC_SEQ_RAS_TIMING: - *out_reg = mmMC_SEQ_RAS_TIMING_LP; - break; - case mmMC_SEQ_DLL_STBY: - *out_reg = mmMC_SEQ_DLL_STBY_LP; - break; - case mmMC_SEQ_G5PDX_CMD0: - *out_reg = mmMC_SEQ_G5PDX_CMD0_LP; - break; - case mmMC_SEQ_G5PDX_CMD1: - *out_reg = mmMC_SEQ_G5PDX_CMD1_LP; - break; - case mmMC_SEQ_G5PDX_CTRL: - *out_reg = mmMC_SEQ_G5PDX_CTRL_LP; - break; - case mmMC_SEQ_CAS_TIMING: - *out_reg = mmMC_SEQ_CAS_TIMING_LP; - break; - case mmMC_SEQ_MISC_TIMING: - *out_reg = mmMC_SEQ_MISC_TIMING_LP; - break; - case mmMC_SEQ_MISC_TIMING2: - *out_reg = mmMC_SEQ_MISC_TIMING2_LP; - break; - case mmMC_SEQ_PMG_DVS_CMD: - *out_reg = mmMC_SEQ_PMG_DVS_CMD_LP; - break; - case mmMC_SEQ_PMG_DVS_CTL: - *out_reg = mmMC_SEQ_PMG_DVS_CTL_LP; - break; - case mmMC_SEQ_RD_CTL_D0: - *out_reg = mmMC_SEQ_RD_CTL_D0_LP; - break; - case mmMC_SEQ_RD_CTL_D1: - *out_reg = mmMC_SEQ_RD_CTL_D1_LP; - break; - case mmMC_SEQ_WR_CTL_D0: - *out_reg = mmMC_SEQ_WR_CTL_D0_LP; - break; - case mmMC_SEQ_WR_CTL_D1: - *out_reg = mmMC_SEQ_WR_CTL_D1_LP; - break; - case mmMC_PMG_CMD_EMRS: - *out_reg = mmMC_SEQ_PMG_CMD_EMRS_LP; - break; - case mmMC_PMG_CMD_MRS: - *out_reg = mmMC_SEQ_PMG_CMD_MRS_LP; - break; - case mmMC_PMG_CMD_MRS1: - *out_reg = mmMC_SEQ_PMG_CMD_MRS1_LP; - break; - case mmMC_SEQ_PMG_TIMING: - *out_reg = mmMC_SEQ_PMG_TIMING_LP; - break; - case mmMC_PMG_CMD_MRS2: - *out_reg = mmMC_SEQ_PMG_CMD_MRS2_LP; - break; - case mmMC_SEQ_WR_CTL_2: - *out_reg = mmMC_SEQ_WR_CTL_2_LP; - break; - default: - result = false; - break; - } - - return result; -} - -static void ci_set_valid_flag(struct ci_mc_reg_table *table) -{ - u8 i, j; - - for (i = 0; i < table->last; i++) { - for (j = 1; j < table->num_entries; j++) { - if (table->mc_reg_table_entry[j-1].mc_data[i] != - table->mc_reg_table_entry[j].mc_data[i]) { - table->valid_flag |= 1 << i; - break; - } - } - } -} - -static void ci_set_s0_mc_reg_index(struct ci_mc_reg_table *table) -{ - u32 i; - u16 address; - - for (i = 0; i < table->last; i++) { - table->mc_reg_address[i].s0 = - ci_check_s0_mc_reg_index(table->mc_reg_address[i].s1, &address) ? - address : table->mc_reg_address[i].s1; - } -} - -static int ci_copy_vbios_mc_reg_table(const struct atom_mc_reg_table *table, - struct ci_mc_reg_table *ci_table) -{ - u8 i, j; - - if (table->last > SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; - if (table->num_entries > MAX_AC_TIMING_ENTRIES) - return -EINVAL; - - for (i = 0; i < table->last; i++) - ci_table->mc_reg_address[i].s1 = table->mc_reg_address[i].s1; - - ci_table->last = table->last; - - for (i = 0; i < table->num_entries; i++) { - ci_table->mc_reg_table_entry[i].mclk_max = - table->mc_reg_table_entry[i].mclk_max; - for (j = 0; j < table->last; j++) - ci_table->mc_reg_table_entry[i].mc_data[j] = - table->mc_reg_table_entry[i].mc_data[j]; - } - ci_table->num_entries = table->num_entries; - - return 0; -} - -static int ci_register_patching_mc_seq(struct amdgpu_device *adev, - struct ci_mc_reg_table *table) -{ - u8 i, k; - u32 tmp; - bool patch; - - tmp = RREG32(mmMC_SEQ_MISC0); - patch = ((tmp & 0x0000f00) == 0x300) ? true : false; - - if (patch && - ((adev->pdev->device == 0x67B0) || - (adev->pdev->device == 0x67B1))) { - for (i = 0; i < table->last; i++) { - if (table->last >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; - switch (table->mc_reg_address[i].s1) { - case mmMC_SEQ_MISC1: - for (k = 0; k < table->num_entries; k++) { - if ((table->mc_reg_table_entry[k].mclk_max == 125000) || - (table->mc_reg_table_entry[k].mclk_max == 137500)) - table->mc_reg_table_entry[k].mc_data[i] = - (table->mc_reg_table_entry[k].mc_data[i] & 0xFFFFFFF8) | - 0x00000007; - } - break; - case mmMC_SEQ_WR_CTL_D0: - for (k = 0; k < table->num_entries; k++) { - if ((table->mc_reg_table_entry[k].mclk_max == 125000) || - (table->mc_reg_table_entry[k].mclk_max == 137500)) - table->mc_reg_table_entry[k].mc_data[i] = - (table->mc_reg_table_entry[k].mc_data[i] & 0xFFFF0F00) | - 0x0000D0DD; - } - break; - case mmMC_SEQ_WR_CTL_D1: - for (k = 0; k < table->num_entries; k++) { - if ((table->mc_reg_table_entry[k].mclk_max == 125000) || - (table->mc_reg_table_entry[k].mclk_max == 137500)) - table->mc_reg_table_entry[k].mc_data[i] = - (table->mc_reg_table_entry[k].mc_data[i] & 0xFFFF0F00) | - 0x0000D0DD; - } - break; - case mmMC_SEQ_WR_CTL_2: - for (k = 0; k < table->num_entries; k++) { - if ((table->mc_reg_table_entry[k].mclk_max == 125000) || - (table->mc_reg_table_entry[k].mclk_max == 137500)) - table->mc_reg_table_entry[k].mc_data[i] = 0; - } - break; - case mmMC_SEQ_CAS_TIMING: - for (k = 0; k < table->num_entries; k++) { - if (table->mc_reg_table_entry[k].mclk_max == 125000) - table->mc_reg_table_entry[k].mc_data[i] = - (table->mc_reg_table_entry[k].mc_data[i] & 0xFFE0FE0F) | - 0x000C0140; - else if (table->mc_reg_table_entry[k].mclk_max == 137500) - table->mc_reg_table_entry[k].mc_data[i] = - (table->mc_reg_table_entry[k].mc_data[i] & 0xFFE0FE0F) | - 0x000C0150; - } - break; - case mmMC_SEQ_MISC_TIMING: - for (k = 0; k < table->num_entries; k++) { - if (table->mc_reg_table_entry[k].mclk_max == 125000) - table->mc_reg_table_entry[k].mc_data[i] = - (table->mc_reg_table_entry[k].mc_data[i] & 0xFFFFFFE0) | - 0x00000030; - else if (table->mc_reg_table_entry[k].mclk_max == 137500) - table->mc_reg_table_entry[k].mc_data[i] = - (table->mc_reg_table_entry[k].mc_data[i] & 0xFFFFFFE0) | - 0x00000035; - } - break; - default: - break; - } - } - - WREG32(mmMC_SEQ_IO_DEBUG_INDEX, 3); - tmp = RREG32(mmMC_SEQ_IO_DEBUG_DATA); - tmp = (tmp & 0xFFF8FFFF) | (1 << 16); - WREG32(mmMC_SEQ_IO_DEBUG_INDEX, 3); - WREG32(mmMC_SEQ_IO_DEBUG_DATA, tmp); - } - - return 0; -} - -static int ci_initialize_mc_reg_table(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct atom_mc_reg_table *table; - struct ci_mc_reg_table *ci_table = &pi->mc_reg_table; - u8 module_index = ci_get_memory_module_index(adev); - int ret; - - table = kzalloc(sizeof(struct atom_mc_reg_table), GFP_KERNEL); - if (!table) - return -ENOMEM; - - WREG32(mmMC_SEQ_RAS_TIMING_LP, RREG32(mmMC_SEQ_RAS_TIMING)); - WREG32(mmMC_SEQ_CAS_TIMING_LP, RREG32(mmMC_SEQ_CAS_TIMING)); - WREG32(mmMC_SEQ_DLL_STBY_LP, RREG32(mmMC_SEQ_DLL_STBY)); - WREG32(mmMC_SEQ_G5PDX_CMD0_LP, RREG32(mmMC_SEQ_G5PDX_CMD0)); - WREG32(mmMC_SEQ_G5PDX_CMD1_LP, RREG32(mmMC_SEQ_G5PDX_CMD1)); - WREG32(mmMC_SEQ_G5PDX_CTRL_LP, RREG32(mmMC_SEQ_G5PDX_CTRL)); - WREG32(mmMC_SEQ_PMG_DVS_CMD_LP, RREG32(mmMC_SEQ_PMG_DVS_CMD)); - WREG32(mmMC_SEQ_PMG_DVS_CTL_LP, RREG32(mmMC_SEQ_PMG_DVS_CTL)); - WREG32(mmMC_SEQ_MISC_TIMING_LP, RREG32(mmMC_SEQ_MISC_TIMING)); - WREG32(mmMC_SEQ_MISC_TIMING2_LP, RREG32(mmMC_SEQ_MISC_TIMING2)); - WREG32(mmMC_SEQ_PMG_CMD_EMRS_LP, RREG32(mmMC_PMG_CMD_EMRS)); - WREG32(mmMC_SEQ_PMG_CMD_MRS_LP, RREG32(mmMC_PMG_CMD_MRS)); - WREG32(mmMC_SEQ_PMG_CMD_MRS1_LP, RREG32(mmMC_PMG_CMD_MRS1)); - WREG32(mmMC_SEQ_WR_CTL_D0_LP, RREG32(mmMC_SEQ_WR_CTL_D0)); - WREG32(mmMC_SEQ_WR_CTL_D1_LP, RREG32(mmMC_SEQ_WR_CTL_D1)); - WREG32(mmMC_SEQ_RD_CTL_D0_LP, RREG32(mmMC_SEQ_RD_CTL_D0)); - WREG32(mmMC_SEQ_RD_CTL_D1_LP, RREG32(mmMC_SEQ_RD_CTL_D1)); - WREG32(mmMC_SEQ_PMG_TIMING_LP, RREG32(mmMC_SEQ_PMG_TIMING)); - WREG32(mmMC_SEQ_PMG_CMD_MRS2_LP, RREG32(mmMC_PMG_CMD_MRS2)); - WREG32(mmMC_SEQ_WR_CTL_2_LP, RREG32(mmMC_SEQ_WR_CTL_2)); - - ret = amdgpu_atombios_init_mc_reg_table(adev, module_index, table); - if (ret) - goto init_mc_done; - - ret = ci_copy_vbios_mc_reg_table(table, ci_table); - if (ret) - goto init_mc_done; - - ci_set_s0_mc_reg_index(ci_table); - - ret = ci_register_patching_mc_seq(adev, ci_table); - if (ret) - goto init_mc_done; - - ret = ci_set_mc_special_registers(adev, ci_table); - if (ret) - goto init_mc_done; - - ci_set_valid_flag(ci_table); - -init_mc_done: - kfree(table); - - return ret; -} - -static int ci_populate_mc_reg_addresses(struct amdgpu_device *adev, - SMU7_Discrete_MCRegisters *mc_reg_table) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 i, j; - - for (i = 0, j = 0; j < pi->mc_reg_table.last; j++) { - if (pi->mc_reg_table.valid_flag & (1 << j)) { - if (i >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) - return -EINVAL; - mc_reg_table->address[i].s0 = cpu_to_be16(pi->mc_reg_table.mc_reg_address[j].s0); - mc_reg_table->address[i].s1 = cpu_to_be16(pi->mc_reg_table.mc_reg_address[j].s1); - i++; - } - } - - mc_reg_table->last = (u8)i; - - return 0; -} - -static void ci_convert_mc_registers(const struct ci_mc_reg_entry *entry, - SMU7_Discrete_MCRegisterSet *data, - u32 num_entries, u32 valid_flag) -{ - u32 i, j; - - for (i = 0, j = 0; j < num_entries; j++) { - if (valid_flag & (1 << j)) { - data->value[i] = cpu_to_be32(entry->mc_data[j]); - i++; - } - } -} - -static void ci_convert_mc_reg_table_entry_to_smc(struct amdgpu_device *adev, - const u32 memory_clock, - SMU7_Discrete_MCRegisterSet *mc_reg_table_data) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 i = 0; - - for(i = 0; i < pi->mc_reg_table.num_entries; i++) { - if (memory_clock <= pi->mc_reg_table.mc_reg_table_entry[i].mclk_max) - break; - } - - if ((i == pi->mc_reg_table.num_entries) && (i > 0)) - --i; - - ci_convert_mc_registers(&pi->mc_reg_table.mc_reg_table_entry[i], - mc_reg_table_data, pi->mc_reg_table.last, - pi->mc_reg_table.valid_flag); -} - -static void ci_convert_mc_reg_table_to_smc(struct amdgpu_device *adev, - SMU7_Discrete_MCRegisters *mc_reg_table) -{ - struct ci_power_info *pi = ci_get_pi(adev); - u32 i; - - for (i = 0; i < pi->dpm_table.mclk_table.count; i++) - ci_convert_mc_reg_table_entry_to_smc(adev, - pi->dpm_table.mclk_table.dpm_levels[i].value, - &mc_reg_table->data[i]); -} - -static int ci_populate_initial_mc_reg_table(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - int ret; - - memset(&pi->smc_mc_reg_table, 0, sizeof(SMU7_Discrete_MCRegisters)); - - ret = ci_populate_mc_reg_addresses(adev, &pi->smc_mc_reg_table); - if (ret) - return ret; - ci_convert_mc_reg_table_to_smc(adev, &pi->smc_mc_reg_table); - - return amdgpu_ci_copy_bytes_to_smc(adev, - pi->mc_reg_table_start, - (u8 *)&pi->smc_mc_reg_table, - sizeof(SMU7_Discrete_MCRegisters), - pi->sram_end); -} - -static int ci_update_and_upload_mc_reg_table(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - - if (!(pi->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) - return 0; - - memset(&pi->smc_mc_reg_table, 0, sizeof(SMU7_Discrete_MCRegisters)); - - ci_convert_mc_reg_table_to_smc(adev, &pi->smc_mc_reg_table); - - return amdgpu_ci_copy_bytes_to_smc(adev, - pi->mc_reg_table_start + - offsetof(SMU7_Discrete_MCRegisters, data[0]), - (u8 *)&pi->smc_mc_reg_table.data[0], - sizeof(SMU7_Discrete_MCRegisterSet) * - pi->dpm_table.mclk_table.count, - pi->sram_end); -} - -static void ci_enable_voltage_control(struct amdgpu_device *adev) -{ - u32 tmp = RREG32_SMC(ixGENERAL_PWRMGT); - - tmp |= GENERAL_PWRMGT__VOLT_PWRMGT_EN_MASK; - WREG32_SMC(ixGENERAL_PWRMGT, tmp); -} - -static enum amdgpu_pcie_gen ci_get_maximum_link_speed(struct amdgpu_device *adev, - struct amdgpu_ps *amdgpu_state) -{ - struct ci_ps *state = ci_get_ps(amdgpu_state); - int i; - u16 pcie_speed, max_speed = 0; - - for (i = 0; i < state->performance_level_count; i++) { - pcie_speed = state->performance_levels[i].pcie_gen; - if (max_speed < pcie_speed) - max_speed = pcie_speed; - } - - return max_speed; -} - -static u16 ci_get_current_pcie_speed(struct amdgpu_device *adev) -{ - u32 speed_cntl = 0; - - speed_cntl = RREG32_PCIE(ixPCIE_LC_SPEED_CNTL) & - PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE_MASK; - speed_cntl >>= PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE__SHIFT; - - return (u16)speed_cntl; -} - -static int ci_get_current_pcie_lane_number(struct amdgpu_device *adev) -{ - u32 link_width = 0; - - link_width = RREG32_PCIE(ixPCIE_LC_LINK_WIDTH_CNTL) & - PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD_MASK; - link_width >>= PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD__SHIFT; - - switch (link_width) { - case 1: - return 1; - case 2: - return 2; - case 3: - return 4; - case 4: - return 8; - case 0: - case 6: - default: - return 16; - } -} - -static void ci_request_link_speed_change_before_state_change(struct amdgpu_device *adev, - struct amdgpu_ps *amdgpu_new_state, - struct amdgpu_ps *amdgpu_current_state) -{ - struct ci_power_info *pi = ci_get_pi(adev); - enum amdgpu_pcie_gen target_link_speed = - ci_get_maximum_link_speed(adev, amdgpu_new_state); - enum amdgpu_pcie_gen current_link_speed; - - if (pi->force_pcie_gen == AMDGPU_PCIE_GEN_INVALID) - current_link_speed = ci_get_maximum_link_speed(adev, amdgpu_current_state); - else - current_link_speed = pi->force_pcie_gen; - - pi->force_pcie_gen = AMDGPU_PCIE_GEN_INVALID; - pi->pspp_notify_required = false; - if (target_link_speed > current_link_speed) { - switch (target_link_speed) { -#ifdef CONFIG_ACPI - case AMDGPU_PCIE_GEN3: - if (amdgpu_acpi_pcie_performance_request(adev, PCIE_PERF_REQ_PECI_GEN3, false) == 0) - break; - pi->force_pcie_gen = AMDGPU_PCIE_GEN2; - if (current_link_speed == AMDGPU_PCIE_GEN2) - break; - case AMDGPU_PCIE_GEN2: - if (amdgpu_acpi_pcie_performance_request(adev, PCIE_PERF_REQ_PECI_GEN2, false) == 0) - break; -#endif - default: - pi->force_pcie_gen = ci_get_current_pcie_speed(adev); - break; - } - } else { - if (target_link_speed < current_link_speed) - pi->pspp_notify_required = true; - } -} - -static void ci_notify_link_speed_change_after_state_change(struct amdgpu_device *adev, - struct amdgpu_ps *amdgpu_new_state, - struct amdgpu_ps *amdgpu_current_state) -{ - struct ci_power_info *pi = ci_get_pi(adev); - enum amdgpu_pcie_gen target_link_speed = - ci_get_maximum_link_speed(adev, amdgpu_new_state); - u8 request; - - if (pi->pspp_notify_required) { - if (target_link_speed == AMDGPU_PCIE_GEN3) - request = PCIE_PERF_REQ_PECI_GEN3; - else if (target_link_speed == AMDGPU_PCIE_GEN2) - request = PCIE_PERF_REQ_PECI_GEN2; - else - request = PCIE_PERF_REQ_PECI_GEN1; - - if ((request == PCIE_PERF_REQ_PECI_GEN1) && - (ci_get_current_pcie_speed(adev) > 0)) - return; - -#ifdef CONFIG_ACPI - amdgpu_acpi_pcie_performance_request(adev, request, false); -#endif - } -} - -static int ci_set_private_data_variables_based_on_pptable(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct amdgpu_clock_voltage_dependency_table *allowed_sclk_vddc_table = - &adev->pm.dpm.dyn_state.vddc_dependency_on_sclk; - struct amdgpu_clock_voltage_dependency_table *allowed_mclk_vddc_table = - &adev->pm.dpm.dyn_state.vddc_dependency_on_mclk; - struct amdgpu_clock_voltage_dependency_table *allowed_mclk_vddci_table = - &adev->pm.dpm.dyn_state.vddci_dependency_on_mclk; - - if (allowed_sclk_vddc_table == NULL) - return -EINVAL; - if (allowed_sclk_vddc_table->count < 1) - return -EINVAL; - if (allowed_mclk_vddc_table == NULL) - return -EINVAL; - if (allowed_mclk_vddc_table->count < 1) - return -EINVAL; - if (allowed_mclk_vddci_table == NULL) - return -EINVAL; - if (allowed_mclk_vddci_table->count < 1) - return -EINVAL; - - pi->min_vddc_in_pp_table = allowed_sclk_vddc_table->entries[0].v; - pi->max_vddc_in_pp_table = - allowed_sclk_vddc_table->entries[allowed_sclk_vddc_table->count - 1].v; - - pi->min_vddci_in_pp_table = allowed_mclk_vddci_table->entries[0].v; - pi->max_vddci_in_pp_table = - allowed_mclk_vddci_table->entries[allowed_mclk_vddci_table->count - 1].v; - - adev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk = - allowed_sclk_vddc_table->entries[allowed_sclk_vddc_table->count - 1].clk; - adev->pm.dpm.dyn_state.max_clock_voltage_on_ac.mclk = - allowed_mclk_vddc_table->entries[allowed_sclk_vddc_table->count - 1].clk; - adev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddc = - allowed_sclk_vddc_table->entries[allowed_sclk_vddc_table->count - 1].v; - adev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddci = - allowed_mclk_vddci_table->entries[allowed_mclk_vddci_table->count - 1].v; - - return 0; -} - -static void ci_patch_with_vddc_leakage(struct amdgpu_device *adev, u16 *vddc) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_leakage_voltage *leakage_table = &pi->vddc_leakage; - u32 leakage_index; - - for (leakage_index = 0; leakage_index < leakage_table->count; leakage_index++) { - if (leakage_table->leakage_id[leakage_index] == *vddc) { - *vddc = leakage_table->actual_voltage[leakage_index]; - break; - } - } -} - -static void ci_patch_with_vddci_leakage(struct amdgpu_device *adev, u16 *vddci) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_leakage_voltage *leakage_table = &pi->vddci_leakage; - u32 leakage_index; - - for (leakage_index = 0; leakage_index < leakage_table->count; leakage_index++) { - if (leakage_table->leakage_id[leakage_index] == *vddci) { - *vddci = leakage_table->actual_voltage[leakage_index]; - break; - } - } -} - -static void ci_patch_clock_voltage_dependency_table_with_vddc_leakage(struct amdgpu_device *adev, - struct amdgpu_clock_voltage_dependency_table *table) -{ - u32 i; - - if (table) { - for (i = 0; i < table->count; i++) - ci_patch_with_vddc_leakage(adev, &table->entries[i].v); - } -} - -static void ci_patch_clock_voltage_dependency_table_with_vddci_leakage(struct amdgpu_device *adev, - struct amdgpu_clock_voltage_dependency_table *table) -{ - u32 i; - - if (table) { - for (i = 0; i < table->count; i++) - ci_patch_with_vddci_leakage(adev, &table->entries[i].v); - } -} - -static void ci_patch_vce_clock_voltage_dependency_table_with_vddc_leakage(struct amdgpu_device *adev, - struct amdgpu_vce_clock_voltage_dependency_table *table) -{ - u32 i; - - if (table) { - for (i = 0; i < table->count; i++) - ci_patch_with_vddc_leakage(adev, &table->entries[i].v); - } -} - -static void ci_patch_uvd_clock_voltage_dependency_table_with_vddc_leakage(struct amdgpu_device *adev, - struct amdgpu_uvd_clock_voltage_dependency_table *table) -{ - u32 i; - - if (table) { - for (i = 0; i < table->count; i++) - ci_patch_with_vddc_leakage(adev, &table->entries[i].v); - } -} - -static void ci_patch_vddc_phase_shed_limit_table_with_vddc_leakage(struct amdgpu_device *adev, - struct amdgpu_phase_shedding_limits_table *table) -{ - u32 i; - - if (table) { - for (i = 0; i < table->count; i++) - ci_patch_with_vddc_leakage(adev, &table->entries[i].voltage); - } -} - -static void ci_patch_clock_voltage_limits_with_vddc_leakage(struct amdgpu_device *adev, - struct amdgpu_clock_and_voltage_limits *table) -{ - if (table) { - ci_patch_with_vddc_leakage(adev, (u16 *)&table->vddc); - ci_patch_with_vddci_leakage(adev, (u16 *)&table->vddci); - } -} - -static void ci_patch_cac_leakage_table_with_vddc_leakage(struct amdgpu_device *adev, - struct amdgpu_cac_leakage_table *table) -{ - u32 i; - - if (table) { - for (i = 0; i < table->count; i++) - ci_patch_with_vddc_leakage(adev, &table->entries[i].vddc); - } -} - -static void ci_patch_dependency_tables_with_leakage(struct amdgpu_device *adev) -{ - - ci_patch_clock_voltage_dependency_table_with_vddc_leakage(adev, - &adev->pm.dpm.dyn_state.vddc_dependency_on_sclk); - ci_patch_clock_voltage_dependency_table_with_vddc_leakage(adev, - &adev->pm.dpm.dyn_state.vddc_dependency_on_mclk); - ci_patch_clock_voltage_dependency_table_with_vddc_leakage(adev, - &adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk); - ci_patch_clock_voltage_dependency_table_with_vddci_leakage(adev, - &adev->pm.dpm.dyn_state.vddci_dependency_on_mclk); - ci_patch_vce_clock_voltage_dependency_table_with_vddc_leakage(adev, - &adev->pm.dpm.dyn_state.vce_clock_voltage_dependency_table); - ci_patch_uvd_clock_voltage_dependency_table_with_vddc_leakage(adev, - &adev->pm.dpm.dyn_state.uvd_clock_voltage_dependency_table); - ci_patch_clock_voltage_dependency_table_with_vddc_leakage(adev, - &adev->pm.dpm.dyn_state.samu_clock_voltage_dependency_table); - ci_patch_clock_voltage_dependency_table_with_vddc_leakage(adev, - &adev->pm.dpm.dyn_state.acp_clock_voltage_dependency_table); - ci_patch_vddc_phase_shed_limit_table_with_vddc_leakage(adev, - &adev->pm.dpm.dyn_state.phase_shedding_limits_table); - ci_patch_clock_voltage_limits_with_vddc_leakage(adev, - &adev->pm.dpm.dyn_state.max_clock_voltage_on_ac); - ci_patch_clock_voltage_limits_with_vddc_leakage(adev, - &adev->pm.dpm.dyn_state.max_clock_voltage_on_dc); - ci_patch_cac_leakage_table_with_vddc_leakage(adev, - &adev->pm.dpm.dyn_state.cac_leakage_table); - -} - -static void ci_update_current_ps(struct amdgpu_device *adev, - struct amdgpu_ps *rps) -{ - struct ci_ps *new_ps = ci_get_ps(rps); - struct ci_power_info *pi = ci_get_pi(adev); - - pi->current_rps = *rps; - pi->current_ps = *new_ps; - pi->current_rps.ps_priv = &pi->current_ps; - adev->pm.dpm.current_ps = &pi->current_rps; -} - -static void ci_update_requested_ps(struct amdgpu_device *adev, - struct amdgpu_ps *rps) -{ - struct ci_ps *new_ps = ci_get_ps(rps); - struct ci_power_info *pi = ci_get_pi(adev); - - pi->requested_rps = *rps; - pi->requested_ps = *new_ps; - pi->requested_rps.ps_priv = &pi->requested_ps; - adev->pm.dpm.requested_ps = &pi->requested_rps; -} - -static int ci_dpm_pre_set_power_state(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - struct amdgpu_ps requested_ps = *adev->pm.dpm.requested_ps; - struct amdgpu_ps *new_ps = &requested_ps; - - ci_update_requested_ps(adev, new_ps); - - ci_apply_state_adjust_rules(adev, &pi->requested_rps); - - return 0; -} - -static void ci_dpm_post_set_power_state(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - struct amdgpu_ps *new_ps = &pi->requested_rps; - - ci_update_current_ps(adev, new_ps); -} - - -static void ci_dpm_setup_asic(struct amdgpu_device *adev) -{ - ci_read_clock_registers(adev); - ci_enable_acpi_power_management(adev); - ci_init_sclk_t(adev); -} - -static int ci_dpm_enable(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct amdgpu_ps *boot_ps = adev->pm.dpm.boot_ps; - int ret; - - if (pi->voltage_control != CISLANDS_VOLTAGE_CONTROL_NONE) { - ci_enable_voltage_control(adev); - ret = ci_construct_voltage_tables(adev); - if (ret) { - DRM_ERROR("ci_construct_voltage_tables failed\n"); - return ret; - } - } - if (pi->caps_dynamic_ac_timing) { - ret = ci_initialize_mc_reg_table(adev); - if (ret) - pi->caps_dynamic_ac_timing = false; - } - if (pi->dynamic_ss) - ci_enable_spread_spectrum(adev, true); - if (pi->thermal_protection) - ci_enable_thermal_protection(adev, true); - ci_program_sstp(adev); - ci_enable_display_gap(adev); - ci_program_vc(adev); - ret = ci_upload_firmware(adev); - if (ret) { - DRM_ERROR("ci_upload_firmware failed\n"); - return ret; - } - ret = ci_process_firmware_header(adev); - if (ret) { - DRM_ERROR("ci_process_firmware_header failed\n"); - return ret; - } - ret = ci_initial_switch_from_arb_f0_to_f1(adev); - if (ret) { - DRM_ERROR("ci_initial_switch_from_arb_f0_to_f1 failed\n"); - return ret; - } - ret = ci_init_smc_table(adev); - if (ret) { - DRM_ERROR("ci_init_smc_table failed\n"); - return ret; - } - ret = ci_init_arb_table_index(adev); - if (ret) { - DRM_ERROR("ci_init_arb_table_index failed\n"); - return ret; - } - if (pi->caps_dynamic_ac_timing) { - ret = ci_populate_initial_mc_reg_table(adev); - if (ret) { - DRM_ERROR("ci_populate_initial_mc_reg_table failed\n"); - return ret; - } - } - ret = ci_populate_pm_base(adev); - if (ret) { - DRM_ERROR("ci_populate_pm_base failed\n"); - return ret; - } - ci_dpm_start_smc(adev); - ci_enable_vr_hot_gpio_interrupt(adev); - ret = ci_notify_smc_display_change(adev, false); - if (ret) { - DRM_ERROR("ci_notify_smc_display_change failed\n"); - return ret; - } - ci_enable_sclk_control(adev, true); - ret = ci_enable_ulv(adev, true); - if (ret) { - DRM_ERROR("ci_enable_ulv failed\n"); - return ret; - } - ret = ci_enable_ds_master_switch(adev, true); - if (ret) { - DRM_ERROR("ci_enable_ds_master_switch failed\n"); - return ret; - } - ret = ci_start_dpm(adev); - if (ret) { - DRM_ERROR("ci_start_dpm failed\n"); - return ret; - } - ret = ci_enable_didt(adev, true); - if (ret) { - DRM_ERROR("ci_enable_didt failed\n"); - return ret; - } - ret = ci_enable_smc_cac(adev, true); - if (ret) { - DRM_ERROR("ci_enable_smc_cac failed\n"); - return ret; - } - ret = ci_enable_power_containment(adev, true); - if (ret) { - DRM_ERROR("ci_enable_power_containment failed\n"); - return ret; - } - - ret = ci_power_control_set_level(adev); - if (ret) { - DRM_ERROR("ci_power_control_set_level failed\n"); - return ret; - } - - ci_enable_auto_throttle_source(adev, AMDGPU_DPM_AUTO_THROTTLE_SRC_THERMAL, true); - - ret = ci_enable_thermal_based_sclk_dpm(adev, true); - if (ret) { - DRM_ERROR("ci_enable_thermal_based_sclk_dpm failed\n"); - return ret; - } - - ci_thermal_start_thermal_controller(adev); - - ci_update_current_ps(adev, boot_ps); - - return 0; -} - -static void ci_dpm_disable(struct amdgpu_device *adev) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct amdgpu_ps *boot_ps = adev->pm.dpm.boot_ps; - - amdgpu_irq_put(adev, &adev->pm.dpm.thermal.irq, - AMDGPU_THERMAL_IRQ_LOW_TO_HIGH); - amdgpu_irq_put(adev, &adev->pm.dpm.thermal.irq, - AMDGPU_THERMAL_IRQ_HIGH_TO_LOW); - - ci_dpm_powergate_uvd(adev, true); - - if (!amdgpu_ci_is_smc_running(adev)) - return; - - ci_thermal_stop_thermal_controller(adev); - - if (pi->thermal_protection) - ci_enable_thermal_protection(adev, false); - ci_enable_power_containment(adev, false); - ci_enable_smc_cac(adev, false); - ci_enable_didt(adev, false); - ci_enable_spread_spectrum(adev, false); - ci_enable_auto_throttle_source(adev, AMDGPU_DPM_AUTO_THROTTLE_SRC_THERMAL, false); - ci_stop_dpm(adev); - ci_enable_ds_master_switch(adev, false); - ci_enable_ulv(adev, false); - ci_clear_vc(adev); - ci_reset_to_default(adev); - ci_dpm_stop_smc(adev); - ci_force_switch_to_arb_f0(adev); - ci_enable_thermal_based_sclk_dpm(adev, false); - - ci_update_current_ps(adev, boot_ps); -} - -static int ci_dpm_set_power_state(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - struct amdgpu_ps *new_ps = &pi->requested_rps; - struct amdgpu_ps *old_ps = &pi->current_rps; - int ret; - - ci_find_dpm_states_clocks_in_dpm_table(adev, new_ps); - if (pi->pcie_performance_request) - ci_request_link_speed_change_before_state_change(adev, new_ps, old_ps); - ret = ci_freeze_sclk_mclk_dpm(adev); - if (ret) { - DRM_ERROR("ci_freeze_sclk_mclk_dpm failed\n"); - return ret; - } - ret = ci_populate_and_upload_sclk_mclk_dpm_levels(adev, new_ps); - if (ret) { - DRM_ERROR("ci_populate_and_upload_sclk_mclk_dpm_levels failed\n"); - return ret; - } - ret = ci_generate_dpm_level_enable_mask(adev, new_ps); - if (ret) { - DRM_ERROR("ci_generate_dpm_level_enable_mask failed\n"); - return ret; - } - - ret = ci_update_vce_dpm(adev, new_ps, old_ps); - if (ret) { - DRM_ERROR("ci_update_vce_dpm failed\n"); - return ret; - } - - ret = ci_update_sclk_t(adev); - if (ret) { - DRM_ERROR("ci_update_sclk_t failed\n"); - return ret; - } - if (pi->caps_dynamic_ac_timing) { - ret = ci_update_and_upload_mc_reg_table(adev); - if (ret) { - DRM_ERROR("ci_update_and_upload_mc_reg_table failed\n"); - return ret; - } - } - ret = ci_program_memory_timing_parameters(adev); - if (ret) { - DRM_ERROR("ci_program_memory_timing_parameters failed\n"); - return ret; - } - ret = ci_unfreeze_sclk_mclk_dpm(adev); - if (ret) { - DRM_ERROR("ci_unfreeze_sclk_mclk_dpm failed\n"); - return ret; - } - ret = ci_upload_dpm_level_enable_mask(adev); - if (ret) { - DRM_ERROR("ci_upload_dpm_level_enable_mask failed\n"); - return ret; - } - if (pi->pcie_performance_request) - ci_notify_link_speed_change_after_state_change(adev, new_ps, old_ps); - - return 0; -} - -#if 0 -static void ci_dpm_reset_asic(struct amdgpu_device *adev) -{ - ci_set_boot_state(adev); -} -#endif - -static void ci_dpm_display_configuration_changed(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - ci_program_display_gap(adev); -} - -union power_info { - struct _ATOM_POWERPLAY_INFO info; - struct _ATOM_POWERPLAY_INFO_V2 info_2; - struct _ATOM_POWERPLAY_INFO_V3 info_3; - struct _ATOM_PPLIB_POWERPLAYTABLE pplib; - struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2; - struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3; -}; - -union pplib_clock_info { - struct _ATOM_PPLIB_R600_CLOCK_INFO r600; - struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780; - struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen; - struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo; - struct _ATOM_PPLIB_SI_CLOCK_INFO si; - struct _ATOM_PPLIB_CI_CLOCK_INFO ci; -}; - -union pplib_power_state { - struct _ATOM_PPLIB_STATE v1; - struct _ATOM_PPLIB_STATE_V2 v2; -}; - -static void ci_parse_pplib_non_clock_info(struct amdgpu_device *adev, - struct amdgpu_ps *rps, - struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info, - u8 table_rev) -{ - rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings); - rps->class = le16_to_cpu(non_clock_info->usClassification); - rps->class2 = le16_to_cpu(non_clock_info->usClassification2); - - if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) { - rps->vclk = le32_to_cpu(non_clock_info->ulVCLK); - rps->dclk = le32_to_cpu(non_clock_info->ulDCLK); - } else { - rps->vclk = 0; - rps->dclk = 0; - } - - if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) - adev->pm.dpm.boot_ps = rps; - if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE) - adev->pm.dpm.uvd_ps = rps; -} - -static void ci_parse_pplib_clock_info(struct amdgpu_device *adev, - struct amdgpu_ps *rps, int index, - union pplib_clock_info *clock_info) -{ - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_ps *ps = ci_get_ps(rps); - struct ci_pl *pl = &ps->performance_levels[index]; - - ps->performance_level_count = index + 1; - - pl->sclk = le16_to_cpu(clock_info->ci.usEngineClockLow); - pl->sclk |= clock_info->ci.ucEngineClockHigh << 16; - pl->mclk = le16_to_cpu(clock_info->ci.usMemoryClockLow); - pl->mclk |= clock_info->ci.ucMemoryClockHigh << 16; - - pl->pcie_gen = amdgpu_get_pcie_gen_support(adev, - pi->sys_pcie_mask, - pi->vbios_boot_state.pcie_gen_bootup_value, - clock_info->ci.ucPCIEGen); - pl->pcie_lane = amdgpu_get_pcie_lane_support(adev, - pi->vbios_boot_state.pcie_lane_bootup_value, - le16_to_cpu(clock_info->ci.usPCIELane)); - - if (rps->class & ATOM_PPLIB_CLASSIFICATION_ACPI) { - pi->acpi_pcie_gen = pl->pcie_gen; - } - - if (rps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV) { - pi->ulv.supported = true; - pi->ulv.pl = *pl; - pi->ulv.cg_ulv_parameter = CISLANDS_CGULVPARAMETER_DFLT; - } - - /* patch up boot state */ - if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) { - pl->mclk = pi->vbios_boot_state.mclk_bootup_value; - pl->sclk = pi->vbios_boot_state.sclk_bootup_value; - pl->pcie_gen = pi->vbios_boot_state.pcie_gen_bootup_value; - pl->pcie_lane = pi->vbios_boot_state.pcie_lane_bootup_value; - } - - switch (rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) { - case ATOM_PPLIB_CLASSIFICATION_UI_BATTERY: - pi->use_pcie_powersaving_levels = true; - if (pi->pcie_gen_powersaving.max < pl->pcie_gen) - pi->pcie_gen_powersaving.max = pl->pcie_gen; - if (pi->pcie_gen_powersaving.min > pl->pcie_gen) - pi->pcie_gen_powersaving.min = pl->pcie_gen; - if (pi->pcie_lane_powersaving.max < pl->pcie_lane) - pi->pcie_lane_powersaving.max = pl->pcie_lane; - if (pi->pcie_lane_powersaving.min > pl->pcie_lane) - pi->pcie_lane_powersaving.min = pl->pcie_lane; - break; - case ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE: - pi->use_pcie_performance_levels = true; - if (pi->pcie_gen_performance.max < pl->pcie_gen) - pi->pcie_gen_performance.max = pl->pcie_gen; - if (pi->pcie_gen_performance.min > pl->pcie_gen) - pi->pcie_gen_performance.min = pl->pcie_gen; - if (pi->pcie_lane_performance.max < pl->pcie_lane) - pi->pcie_lane_performance.max = pl->pcie_lane; - if (pi->pcie_lane_performance.min > pl->pcie_lane) - pi->pcie_lane_performance.min = pl->pcie_lane; - break; - default: - break; - } -} - -static int ci_parse_power_table(struct amdgpu_device *adev) -{ - struct amdgpu_mode_info *mode_info = &adev->mode_info; - struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info; - union pplib_power_state *power_state; - int i, j, k, non_clock_array_index, clock_array_index; - union pplib_clock_info *clock_info; - struct _StateArray *state_array; - struct _ClockInfoArray *clock_info_array; - struct _NonClockInfoArray *non_clock_info_array; - union power_info *power_info; - int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo); - u16 data_offset; - u8 frev, crev; - u8 *power_state_offset; - struct ci_ps *ps; - - if (!amdgpu_atom_parse_data_header(mode_info->atom_context, index, NULL, - &frev, &crev, &data_offset)) - return -EINVAL; - power_info = (union power_info *)(mode_info->atom_context->bios + data_offset); - - amdgpu_add_thermal_controller(adev); - - state_array = (struct _StateArray *) - (mode_info->atom_context->bios + data_offset + - le16_to_cpu(power_info->pplib.usStateArrayOffset)); - clock_info_array = (struct _ClockInfoArray *) - (mode_info->atom_context->bios + data_offset + - le16_to_cpu(power_info->pplib.usClockInfoArrayOffset)); - non_clock_info_array = (struct _NonClockInfoArray *) - (mode_info->atom_context->bios + data_offset + - le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset)); - - adev->pm.dpm.ps = kcalloc(state_array->ucNumEntries, - sizeof(struct amdgpu_ps), - GFP_KERNEL); - if (!adev->pm.dpm.ps) - return -ENOMEM; - power_state_offset = (u8 *)state_array->states; - for (i = 0; i < state_array->ucNumEntries; i++) { - u8 *idx; - power_state = (union pplib_power_state *)power_state_offset; - non_clock_array_index = power_state->v2.nonClockInfoIndex; - non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *) - &non_clock_info_array->nonClockInfo[non_clock_array_index]; - ps = kzalloc(sizeof(struct ci_ps), GFP_KERNEL); - if (ps == NULL) { - kfree(adev->pm.dpm.ps); - return -ENOMEM; - } - adev->pm.dpm.ps[i].ps_priv = ps; - ci_parse_pplib_non_clock_info(adev, &adev->pm.dpm.ps[i], - non_clock_info, - non_clock_info_array->ucEntrySize); - k = 0; - idx = (u8 *)&power_state->v2.clockInfoIndex[0]; - for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) { - clock_array_index = idx[j]; - if (clock_array_index >= clock_info_array->ucNumEntries) - continue; - if (k >= CISLANDS_MAX_HARDWARE_POWERLEVELS) - break; - clock_info = (union pplib_clock_info *) - ((u8 *)&clock_info_array->clockInfo[0] + - (clock_array_index * clock_info_array->ucEntrySize)); - ci_parse_pplib_clock_info(adev, - &adev->pm.dpm.ps[i], k, - clock_info); - k++; - } - power_state_offset += 2 + power_state->v2.ucNumDPMLevels; - } - adev->pm.dpm.num_ps = state_array->ucNumEntries; - - /* fill in the vce power states */ - for (i = 0; i < adev->pm.dpm.num_of_vce_states; i++) { - u32 sclk, mclk; - clock_array_index = adev->pm.dpm.vce_states[i].clk_idx; - clock_info = (union pplib_clock_info *) - &clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize]; - sclk = le16_to_cpu(clock_info->ci.usEngineClockLow); - sclk |= clock_info->ci.ucEngineClockHigh << 16; - mclk = le16_to_cpu(clock_info->ci.usMemoryClockLow); - mclk |= clock_info->ci.ucMemoryClockHigh << 16; - adev->pm.dpm.vce_states[i].sclk = sclk; - adev->pm.dpm.vce_states[i].mclk = mclk; - } - - return 0; -} - -static int ci_get_vbios_boot_values(struct amdgpu_device *adev, - struct ci_vbios_boot_state *boot_state) -{ - struct amdgpu_mode_info *mode_info = &adev->mode_info; - int index = GetIndexIntoMasterTable(DATA, FirmwareInfo); - ATOM_FIRMWARE_INFO_V2_2 *firmware_info; - u8 frev, crev; - u16 data_offset; - - if (amdgpu_atom_parse_data_header(mode_info->atom_context, index, NULL, - &frev, &crev, &data_offset)) { - firmware_info = - (ATOM_FIRMWARE_INFO_V2_2 *)(mode_info->atom_context->bios + - data_offset); - boot_state->mvdd_bootup_value = le16_to_cpu(firmware_info->usBootUpMVDDCVoltage); - boot_state->vddc_bootup_value = le16_to_cpu(firmware_info->usBootUpVDDCVoltage); - boot_state->vddci_bootup_value = le16_to_cpu(firmware_info->usBootUpVDDCIVoltage); - boot_state->pcie_gen_bootup_value = ci_get_current_pcie_speed(adev); - boot_state->pcie_lane_bootup_value = ci_get_current_pcie_lane_number(adev); - boot_state->sclk_bootup_value = le32_to_cpu(firmware_info->ulDefaultEngineClock); - boot_state->mclk_bootup_value = le32_to_cpu(firmware_info->ulDefaultMemoryClock); - - return 0; - } - return -EINVAL; -} - -static void ci_dpm_fini(struct amdgpu_device *adev) -{ - int i; - - for (i = 0; i < adev->pm.dpm.num_ps; i++) { - kfree(adev->pm.dpm.ps[i].ps_priv); - } - kfree(adev->pm.dpm.ps); - kfree(adev->pm.dpm.priv); - kfree(adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries); - amdgpu_free_extended_power_table(adev); -} - -/** - * ci_dpm_init_microcode - load ucode images from disk - * - * @adev: amdgpu_device pointer - * - * Use the firmware interface to load the ucode images into - * the driver (not loaded into hw). - * Returns 0 on success, error on failure. - */ -static int ci_dpm_init_microcode(struct amdgpu_device *adev) -{ - const char *chip_name; - char fw_name[30]; - int err; - - DRM_DEBUG("\n"); - - switch (adev->asic_type) { - case CHIP_BONAIRE: - if ((adev->pdev->revision == 0x80) || - (adev->pdev->revision == 0x81) || - (adev->pdev->device == 0x665f)) - chip_name = "bonaire_k"; - else - chip_name = "bonaire"; - break; - case CHIP_HAWAII: - if (adev->pdev->revision == 0x80) - chip_name = "hawaii_k"; - else - chip_name = "hawaii"; - break; - case CHIP_KAVERI: - case CHIP_KABINI: - case CHIP_MULLINS: - default: BUG(); - } - - snprintf(fw_name, sizeof(fw_name), "amdgpu/%s_smc.bin", chip_name); - err = request_firmware(&adev->pm.fw, fw_name, adev->dev); - if (err) - goto out; - err = amdgpu_ucode_validate(adev->pm.fw); - -out: - if (err) { - pr_err("cik_smc: Failed to load firmware \"%s\"\n", fw_name); - release_firmware(adev->pm.fw); - adev->pm.fw = NULL; - } - return err; -} - -static int ci_dpm_init(struct amdgpu_device *adev) -{ - int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info); - SMU7_Discrete_DpmTable *dpm_table; - struct amdgpu_gpio_rec gpio; - u16 data_offset, size; - u8 frev, crev; - struct ci_power_info *pi; - int ret; - - pi = kzalloc(sizeof(struct ci_power_info), GFP_KERNEL); - if (pi == NULL) - return -ENOMEM; - adev->pm.dpm.priv = pi; - - pi->sys_pcie_mask = - adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_MASK; - - pi->force_pcie_gen = AMDGPU_PCIE_GEN_INVALID; - - pi->pcie_gen_performance.max = AMDGPU_PCIE_GEN1; - pi->pcie_gen_performance.min = AMDGPU_PCIE_GEN3; - pi->pcie_gen_powersaving.max = AMDGPU_PCIE_GEN1; - pi->pcie_gen_powersaving.min = AMDGPU_PCIE_GEN3; - - pi->pcie_lane_performance.max = 0; - pi->pcie_lane_performance.min = 16; - pi->pcie_lane_powersaving.max = 0; - pi->pcie_lane_powersaving.min = 16; - - ret = ci_get_vbios_boot_values(adev, &pi->vbios_boot_state); - if (ret) { - ci_dpm_fini(adev); - return ret; - } - - ret = amdgpu_get_platform_caps(adev); - if (ret) { - ci_dpm_fini(adev); - return ret; - } - - ret = amdgpu_parse_extended_power_table(adev); - if (ret) { - ci_dpm_fini(adev); - return ret; - } - - ret = ci_parse_power_table(adev); - if (ret) { - ci_dpm_fini(adev); - return ret; - } - - pi->dll_default_on = false; - pi->sram_end = SMC_RAM_END; - - pi->activity_target[0] = CISLAND_TARGETACTIVITY_DFLT; - pi->activity_target[1] = CISLAND_TARGETACTIVITY_DFLT; - pi->activity_target[2] = CISLAND_TARGETACTIVITY_DFLT; - pi->activity_target[3] = CISLAND_TARGETACTIVITY_DFLT; - pi->activity_target[4] = CISLAND_TARGETACTIVITY_DFLT; - pi->activity_target[5] = CISLAND_TARGETACTIVITY_DFLT; - pi->activity_target[6] = CISLAND_TARGETACTIVITY_DFLT; - pi->activity_target[7] = CISLAND_TARGETACTIVITY_DFLT; - - pi->mclk_activity_target = CISLAND_MCLK_TARGETACTIVITY_DFLT; - - pi->sclk_dpm_key_disabled = 0; - pi->mclk_dpm_key_disabled = 0; - pi->pcie_dpm_key_disabled = 0; - pi->thermal_sclk_dpm_enabled = 0; - - if (adev->powerplay.pp_feature & PP_SCLK_DEEP_SLEEP_MASK) - pi->caps_sclk_ds = true; - else - pi->caps_sclk_ds = false; - - pi->mclk_strobe_mode_threshold = 40000; - pi->mclk_stutter_mode_threshold = 40000; - pi->mclk_edc_enable_threshold = 40000; - pi->mclk_edc_wr_enable_threshold = 40000; - - ci_initialize_powertune_defaults(adev); - - pi->caps_fps = false; - - pi->caps_sclk_throttle_low_notification = false; - - pi->caps_uvd_dpm = true; - pi->caps_vce_dpm = true; - - ci_get_leakage_voltages(adev); - ci_patch_dependency_tables_with_leakage(adev); - ci_set_private_data_variables_based_on_pptable(adev); - - adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries = - kcalloc(4, - sizeof(struct amdgpu_clock_voltage_dependency_entry), - GFP_KERNEL); - if (!adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries) { - ci_dpm_fini(adev); - return -ENOMEM; - } - adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.count = 4; - adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].clk = 0; - adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].v = 0; - adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].clk = 36000; - adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].v = 720; - adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].clk = 54000; - adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].v = 810; - adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].clk = 72000; - adev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].v = 900; - - adev->pm.dpm.dyn_state.mclk_sclk_ratio = 4; - adev->pm.dpm.dyn_state.sclk_mclk_delta = 15000; - adev->pm.dpm.dyn_state.vddc_vddci_delta = 200; - - adev->pm.dpm.dyn_state.valid_sclk_values.count = 0; - adev->pm.dpm.dyn_state.valid_sclk_values.values = NULL; - adev->pm.dpm.dyn_state.valid_mclk_values.count = 0; - adev->pm.dpm.dyn_state.valid_mclk_values.values = NULL; - - if (adev->asic_type == CHIP_HAWAII) { - pi->thermal_temp_setting.temperature_low = 94500; - pi->thermal_temp_setting.temperature_high = 95000; - pi->thermal_temp_setting.temperature_shutdown = 104000; - } else { - pi->thermal_temp_setting.temperature_low = 99500; - pi->thermal_temp_setting.temperature_high = 100000; - pi->thermal_temp_setting.temperature_shutdown = 104000; - } - - pi->uvd_enabled = false; - - dpm_table = &pi->smc_state_table; - - gpio = amdgpu_atombios_lookup_gpio(adev, VDDC_VRHOT_GPIO_PINID); - if (gpio.valid) { - dpm_table->VRHotGpio = gpio.shift; - adev->pm.dpm.platform_caps |= ATOM_PP_PLATFORM_CAP_REGULATOR_HOT; - } else { - dpm_table->VRHotGpio = CISLANDS_UNUSED_GPIO_PIN; - adev->pm.dpm.platform_caps &= ~ATOM_PP_PLATFORM_CAP_REGULATOR_HOT; - } - - gpio = amdgpu_atombios_lookup_gpio(adev, PP_AC_DC_SWITCH_GPIO_PINID); - if (gpio.valid) { - dpm_table->AcDcGpio = gpio.shift; - adev->pm.dpm.platform_caps |= ATOM_PP_PLATFORM_CAP_HARDWAREDC; - } else { - dpm_table->AcDcGpio = CISLANDS_UNUSED_GPIO_PIN; - adev->pm.dpm.platform_caps &= ~ATOM_PP_PLATFORM_CAP_HARDWAREDC; - } - - gpio = amdgpu_atombios_lookup_gpio(adev, VDDC_PCC_GPIO_PINID); - if (gpio.valid) { - u32 tmp = RREG32_SMC(ixCNB_PWRMGT_CNTL); - - switch (gpio.shift) { - case 0: - tmp &= ~CNB_PWRMGT_CNTL__GNB_SLOW_MODE_MASK; - tmp |= 1 << CNB_PWRMGT_CNTL__GNB_SLOW_MODE__SHIFT; - break; - case 1: - tmp &= ~CNB_PWRMGT_CNTL__GNB_SLOW_MODE_MASK; - tmp |= 2 << CNB_PWRMGT_CNTL__GNB_SLOW_MODE__SHIFT; - break; - case 2: - tmp |= CNB_PWRMGT_CNTL__GNB_SLOW_MASK; - break; - case 3: - tmp |= CNB_PWRMGT_CNTL__FORCE_NB_PS1_MASK; - break; - case 4: - tmp |= CNB_PWRMGT_CNTL__DPM_ENABLED_MASK; - break; - default: - DRM_INFO("Invalid PCC GPIO: %u!\n", gpio.shift); - break; - } - WREG32_SMC(ixCNB_PWRMGT_CNTL, tmp); - } - - pi->voltage_control = CISLANDS_VOLTAGE_CONTROL_NONE; - pi->vddci_control = CISLANDS_VOLTAGE_CONTROL_NONE; - pi->mvdd_control = CISLANDS_VOLTAGE_CONTROL_NONE; - if (amdgpu_atombios_is_voltage_gpio(adev, VOLTAGE_TYPE_VDDC, VOLTAGE_OBJ_GPIO_LUT)) - pi->voltage_control = CISLANDS_VOLTAGE_CONTROL_BY_GPIO; - else if (amdgpu_atombios_is_voltage_gpio(adev, VOLTAGE_TYPE_VDDC, VOLTAGE_OBJ_SVID2)) - pi->voltage_control = CISLANDS_VOLTAGE_CONTROL_BY_SVID2; - - if (adev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_VDDCI_CONTROL) { - if (amdgpu_atombios_is_voltage_gpio(adev, VOLTAGE_TYPE_VDDCI, VOLTAGE_OBJ_GPIO_LUT)) - pi->vddci_control = CISLANDS_VOLTAGE_CONTROL_BY_GPIO; - else if (amdgpu_atombios_is_voltage_gpio(adev, VOLTAGE_TYPE_VDDCI, VOLTAGE_OBJ_SVID2)) - pi->vddci_control = CISLANDS_VOLTAGE_CONTROL_BY_SVID2; - else - adev->pm.dpm.platform_caps &= ~ATOM_PP_PLATFORM_CAP_VDDCI_CONTROL; - } - - if (adev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_MVDDCONTROL) { - if (amdgpu_atombios_is_voltage_gpio(adev, VOLTAGE_TYPE_MVDDC, VOLTAGE_OBJ_GPIO_LUT)) - pi->mvdd_control = CISLANDS_VOLTAGE_CONTROL_BY_GPIO; - else if (amdgpu_atombios_is_voltage_gpio(adev, VOLTAGE_TYPE_MVDDC, VOLTAGE_OBJ_SVID2)) - pi->mvdd_control = CISLANDS_VOLTAGE_CONTROL_BY_SVID2; - else - adev->pm.dpm.platform_caps &= ~ATOM_PP_PLATFORM_CAP_MVDDCONTROL; - } - - pi->vddc_phase_shed_control = true; - -#if defined(CONFIG_ACPI) - pi->pcie_performance_request = - amdgpu_acpi_is_pcie_performance_request_supported(adev); -#else - pi->pcie_performance_request = false; -#endif - - if (amdgpu_atom_parse_data_header(adev->mode_info.atom_context, index, &size, - &frev, &crev, &data_offset)) { - pi->caps_sclk_ss_support = true; - pi->caps_mclk_ss_support = true; - pi->dynamic_ss = true; - } else { - pi->caps_sclk_ss_support = false; - pi->caps_mclk_ss_support = false; - pi->dynamic_ss = true; - } - - if (adev->pm.int_thermal_type != THERMAL_TYPE_NONE) - pi->thermal_protection = true; - else - pi->thermal_protection = false; - - pi->caps_dynamic_ac_timing = true; - - pi->uvd_power_gated = true; - - /* make sure dc limits are valid */ - if ((adev->pm.dpm.dyn_state.max_clock_voltage_on_dc.sclk == 0) || - (adev->pm.dpm.dyn_state.max_clock_voltage_on_dc.mclk == 0)) - adev->pm.dpm.dyn_state.max_clock_voltage_on_dc = - adev->pm.dpm.dyn_state.max_clock_voltage_on_ac; - - pi->fan_ctrl_is_in_default_mode = true; - - return 0; -} - -static void -ci_dpm_debugfs_print_current_performance_level(void *handle, - struct seq_file *m) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - struct amdgpu_ps *rps = &pi->current_rps; - u32 sclk = ci_get_average_sclk_freq(adev); - u32 mclk = ci_get_average_mclk_freq(adev); - u32 activity_percent = 50; - int ret; - - ret = ci_read_smc_soft_register(adev, offsetof(SMU7_SoftRegisters, AverageGraphicsA), - &activity_percent); - - if (ret == 0) { - activity_percent += 0x80; - activity_percent >>= 8; - activity_percent = activity_percent > 100 ? 100 : activity_percent; - } - - seq_printf(m, "uvd %sabled\n", pi->uvd_power_gated ? "dis" : "en"); - seq_printf(m, "vce %sabled\n", rps->vce_active ? "en" : "dis"); - seq_printf(m, "power level avg sclk: %u mclk: %u\n", - sclk, mclk); - seq_printf(m, "GPU load: %u %%\n", activity_percent); -} - -static void ci_dpm_print_power_state(void *handle, void *current_ps) -{ - struct amdgpu_ps *rps = (struct amdgpu_ps *)current_ps; - struct ci_ps *ps = ci_get_ps(rps); - struct ci_pl *pl; - int i; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - amdgpu_dpm_print_class_info(rps->class, rps->class2); - amdgpu_dpm_print_cap_info(rps->caps); - printk("\tuvd vclk: %d dclk: %d\n", rps->vclk, rps->dclk); - for (i = 0; i < ps->performance_level_count; i++) { - pl = &ps->performance_levels[i]; - printk("\t\tpower level %d sclk: %u mclk: %u pcie gen: %u pcie lanes: %u\n", - i, pl->sclk, pl->mclk, pl->pcie_gen + 1, pl->pcie_lane); - } - amdgpu_dpm_print_ps_status(adev, rps); -} - -static inline bool ci_are_power_levels_equal(const struct ci_pl *ci_cpl1, - const struct ci_pl *ci_cpl2) -{ - return ((ci_cpl1->mclk == ci_cpl2->mclk) && - (ci_cpl1->sclk == ci_cpl2->sclk) && - (ci_cpl1->pcie_gen == ci_cpl2->pcie_gen) && - (ci_cpl1->pcie_lane == ci_cpl2->pcie_lane)); -} - -static int ci_check_state_equal(void *handle, - void *current_ps, - void *request_ps, - bool *equal) -{ - struct ci_ps *ci_cps; - struct ci_ps *ci_rps; - int i; - struct amdgpu_ps *cps = (struct amdgpu_ps *)current_ps; - struct amdgpu_ps *rps = (struct amdgpu_ps *)request_ps; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (adev == NULL || cps == NULL || rps == NULL || equal == NULL) - return -EINVAL; - - ci_cps = ci_get_ps((struct amdgpu_ps *)cps); - ci_rps = ci_get_ps((struct amdgpu_ps *)rps); - - if (ci_cps == NULL) { - *equal = false; - return 0; - } - - if (ci_cps->performance_level_count != ci_rps->performance_level_count) { - - *equal = false; - return 0; - } - - for (i = 0; i < ci_cps->performance_level_count; i++) { - if (!ci_are_power_levels_equal(&(ci_cps->performance_levels[i]), - &(ci_rps->performance_levels[i]))) { - *equal = false; - return 0; - } - } - - /* If all performance levels are the same try to use the UVD clocks to break the tie.*/ - *equal = ((cps->vclk == rps->vclk) && (cps->dclk == rps->dclk)); - *equal &= ((cps->evclk == rps->evclk) && (cps->ecclk == rps->ecclk)); - - return 0; -} - -static u32 ci_dpm_get_sclk(void *handle, bool low) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_ps *requested_state = ci_get_ps(&pi->requested_rps); - - if (low) - return requested_state->performance_levels[0].sclk; - else - return requested_state->performance_levels[requested_state->performance_level_count - 1].sclk; -} - -static u32 ci_dpm_get_mclk(void *handle, bool low) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_ps *requested_state = ci_get_ps(&pi->requested_rps); - - if (low) - return requested_state->performance_levels[0].mclk; - else - return requested_state->performance_levels[requested_state->performance_level_count - 1].mclk; -} - -/* get temperature in millidegrees */ -static int ci_dpm_get_temp(void *handle) -{ - u32 temp; - int actual_temp = 0; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - temp = (RREG32_SMC(ixCG_MULT_THERMAL_STATUS) & CG_MULT_THERMAL_STATUS__CTF_TEMP_MASK) >> - CG_MULT_THERMAL_STATUS__CTF_TEMP__SHIFT; - - if (temp & 0x200) - actual_temp = 255; - else - actual_temp = temp & 0x1ff; - - actual_temp = actual_temp * 1000; - - return actual_temp; -} - -static int ci_set_temperature_range(struct amdgpu_device *adev) -{ - int ret; - - ret = ci_thermal_enable_alert(adev, false); - if (ret) - return ret; - ret = ci_thermal_set_temperature_range(adev, CISLANDS_TEMP_RANGE_MIN, - CISLANDS_TEMP_RANGE_MAX); - if (ret) - return ret; - ret = ci_thermal_enable_alert(adev, true); - if (ret) - return ret; - return ret; -} - -static int ci_dpm_early_init(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - adev->powerplay.pp_funcs = &ci_dpm_funcs; - adev->powerplay.pp_handle = adev; - ci_dpm_set_irq_funcs(adev); - - return 0; -} - -static int ci_dpm_late_init(void *handle) -{ - int ret; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (!adev->pm.dpm_enabled) - return 0; - - /* init the sysfs and debugfs files late */ - ret = amdgpu_pm_sysfs_init(adev); - if (ret) - return ret; - - ret = ci_set_temperature_range(adev); - if (ret) - return ret; - - return 0; -} - -static int ci_dpm_sw_init(void *handle) -{ - int ret; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 230, - &adev->pm.dpm.thermal.irq); - if (ret) - return ret; - - ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 231, - &adev->pm.dpm.thermal.irq); - if (ret) - return ret; - - /* default to balanced state */ - adev->pm.dpm.state = POWER_STATE_TYPE_BALANCED; - adev->pm.dpm.user_state = POWER_STATE_TYPE_BALANCED; - adev->pm.dpm.forced_level = AMD_DPM_FORCED_LEVEL_AUTO; - adev->pm.default_sclk = adev->clock.default_sclk; - adev->pm.default_mclk = adev->clock.default_mclk; - adev->pm.current_sclk = adev->clock.default_sclk; - adev->pm.current_mclk = adev->clock.default_mclk; - adev->pm.int_thermal_type = THERMAL_TYPE_NONE; - - ret = ci_dpm_init_microcode(adev); - if (ret) - return ret; - - if (amdgpu_dpm == 0) - return 0; - - INIT_WORK(&adev->pm.dpm.thermal.work, amdgpu_dpm_thermal_work_handler); - mutex_lock(&adev->pm.mutex); - ret = ci_dpm_init(adev); - if (ret) - goto dpm_failed; - adev->pm.dpm.current_ps = adev->pm.dpm.requested_ps = adev->pm.dpm.boot_ps; - if (amdgpu_dpm == 1) - amdgpu_pm_print_power_states(adev); - mutex_unlock(&adev->pm.mutex); - DRM_INFO("amdgpu: dpm initialized\n"); - - return 0; - -dpm_failed: - ci_dpm_fini(adev); - mutex_unlock(&adev->pm.mutex); - DRM_ERROR("amdgpu: dpm initialization failed\n"); - return ret; -} - -static int ci_dpm_sw_fini(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - flush_work(&adev->pm.dpm.thermal.work); - - mutex_lock(&adev->pm.mutex); - ci_dpm_fini(adev); - mutex_unlock(&adev->pm.mutex); - - release_firmware(adev->pm.fw); - adev->pm.fw = NULL; - - return 0; -} - -static int ci_dpm_hw_init(void *handle) -{ - int ret; - - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (!amdgpu_dpm) { - ret = ci_upload_firmware(adev); - if (ret) { - DRM_ERROR("ci_upload_firmware failed\n"); - return ret; - } - ci_dpm_start_smc(adev); - return 0; - } - - mutex_lock(&adev->pm.mutex); - ci_dpm_setup_asic(adev); - ret = ci_dpm_enable(adev); - if (ret) - adev->pm.dpm_enabled = false; - else - adev->pm.dpm_enabled = true; - mutex_unlock(&adev->pm.mutex); - - return ret; -} - -static int ci_dpm_hw_fini(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (adev->pm.dpm_enabled) { - mutex_lock(&adev->pm.mutex); - ci_dpm_disable(adev); - mutex_unlock(&adev->pm.mutex); - } else { - ci_dpm_stop_smc(adev); - } - - return 0; -} - -static int ci_dpm_suspend(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (adev->pm.dpm_enabled) { - mutex_lock(&adev->pm.mutex); - amdgpu_irq_put(adev, &adev->pm.dpm.thermal.irq, - AMDGPU_THERMAL_IRQ_LOW_TO_HIGH); - amdgpu_irq_put(adev, &adev->pm.dpm.thermal.irq, - AMDGPU_THERMAL_IRQ_HIGH_TO_LOW); - adev->pm.dpm.last_user_state = adev->pm.dpm.user_state; - adev->pm.dpm.last_state = adev->pm.dpm.state; - adev->pm.dpm.user_state = POWER_STATE_TYPE_INTERNAL_BOOT; - adev->pm.dpm.state = POWER_STATE_TYPE_INTERNAL_BOOT; - mutex_unlock(&adev->pm.mutex); - amdgpu_pm_compute_clocks(adev); - - } - - return 0; -} - -static int ci_dpm_resume(void *handle) -{ - int ret; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - if (adev->pm.dpm_enabled) { - /* asic init will reset to the boot state */ - mutex_lock(&adev->pm.mutex); - ci_dpm_setup_asic(adev); - ret = ci_dpm_enable(adev); - if (ret) - adev->pm.dpm_enabled = false; - else - adev->pm.dpm_enabled = true; - adev->pm.dpm.user_state = adev->pm.dpm.last_user_state; - adev->pm.dpm.state = adev->pm.dpm.last_state; - mutex_unlock(&adev->pm.mutex); - if (adev->pm.dpm_enabled) - amdgpu_pm_compute_clocks(adev); - } - return 0; -} - -static bool ci_dpm_is_idle(void *handle) -{ - /* XXX */ - return true; -} - -static int ci_dpm_wait_for_idle(void *handle) -{ - /* XXX */ - return 0; -} - -static int ci_dpm_soft_reset(void *handle) -{ - return 0; -} - -static int ci_dpm_set_interrupt_state(struct amdgpu_device *adev, - struct amdgpu_irq_src *source, - unsigned type, - enum amdgpu_interrupt_state state) -{ - u32 cg_thermal_int; - - switch (type) { - case AMDGPU_THERMAL_IRQ_LOW_TO_HIGH: - switch (state) { - case AMDGPU_IRQ_STATE_DISABLE: - cg_thermal_int = RREG32_SMC(ixCG_THERMAL_INT); - cg_thermal_int |= CG_THERMAL_INT_CTRL__THERM_INTH_MASK_MASK; - WREG32_SMC(ixCG_THERMAL_INT, cg_thermal_int); - break; - case AMDGPU_IRQ_STATE_ENABLE: - cg_thermal_int = RREG32_SMC(ixCG_THERMAL_INT); - cg_thermal_int &= ~CG_THERMAL_INT_CTRL__THERM_INTH_MASK_MASK; - WREG32_SMC(ixCG_THERMAL_INT, cg_thermal_int); - break; - default: - break; - } - break; - - case AMDGPU_THERMAL_IRQ_HIGH_TO_LOW: - switch (state) { - case AMDGPU_IRQ_STATE_DISABLE: - cg_thermal_int = RREG32_SMC(ixCG_THERMAL_INT); - cg_thermal_int |= CG_THERMAL_INT_CTRL__THERM_INTL_MASK_MASK; - WREG32_SMC(ixCG_THERMAL_INT, cg_thermal_int); - break; - case AMDGPU_IRQ_STATE_ENABLE: - cg_thermal_int = RREG32_SMC(ixCG_THERMAL_INT); - cg_thermal_int &= ~CG_THERMAL_INT_CTRL__THERM_INTL_MASK_MASK; - WREG32_SMC(ixCG_THERMAL_INT, cg_thermal_int); - break; - default: - break; - } - break; - - default: - break; - } - return 0; -} - -static int ci_dpm_process_interrupt(struct amdgpu_device *adev, - struct amdgpu_irq_src *source, - struct amdgpu_iv_entry *entry) -{ - bool queue_thermal = false; - - if (entry == NULL) - return -EINVAL; - - switch (entry->src_id) { - case 230: /* thermal low to high */ - DRM_DEBUG("IH: thermal low to high\n"); - adev->pm.dpm.thermal.high_to_low = false; - queue_thermal = true; - break; - case 231: /* thermal high to low */ - DRM_DEBUG("IH: thermal high to low\n"); - adev->pm.dpm.thermal.high_to_low = true; - queue_thermal = true; - break; - default: - break; - } - - if (queue_thermal) - schedule_work(&adev->pm.dpm.thermal.work); - - return 0; -} - -static int ci_dpm_set_clockgating_state(void *handle, - enum amd_clockgating_state state) -{ - return 0; -} - -static int ci_dpm_set_powergating_state(void *handle, - enum amd_powergating_state state) -{ - return 0; -} - -static int ci_dpm_print_clock_levels(void *handle, - enum pp_clock_type type, char *buf) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_single_dpm_table *sclk_table = &pi->dpm_table.sclk_table; - struct ci_single_dpm_table *mclk_table = &pi->dpm_table.mclk_table; - struct ci_single_dpm_table *pcie_table = &pi->dpm_table.pcie_speed_table; - - int i, now, size = 0; - uint32_t clock, pcie_speed; - - switch (type) { - case PP_SCLK: - amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_API_GetSclkFrequency); - clock = RREG32(mmSMC_MSG_ARG_0); - - for (i = 0; i < sclk_table->count; i++) { - if (clock > sclk_table->dpm_levels[i].value) - continue; - break; - } - now = i; - - for (i = 0; i < sclk_table->count; i++) - size += sprintf(buf + size, "%d: %uMhz %s\n", - i, sclk_table->dpm_levels[i].value / 100, - (i == now) ? "*" : ""); - break; - case PP_MCLK: - amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_API_GetMclkFrequency); - clock = RREG32(mmSMC_MSG_ARG_0); - - for (i = 0; i < mclk_table->count; i++) { - if (clock > mclk_table->dpm_levels[i].value) - continue; - break; - } - now = i; - - for (i = 0; i < mclk_table->count; i++) - size += sprintf(buf + size, "%d: %uMhz %s\n", - i, mclk_table->dpm_levels[i].value / 100, - (i == now) ? "*" : ""); - break; - case PP_PCIE: - pcie_speed = ci_get_current_pcie_speed(adev); - for (i = 0; i < pcie_table->count; i++) { - if (pcie_speed != pcie_table->dpm_levels[i].value) - continue; - break; - } - now = i; - - for (i = 0; i < pcie_table->count; i++) - size += sprintf(buf + size, "%d: %s %s\n", i, - (pcie_table->dpm_levels[i].value == 0) ? "2.5GT/s, x1" : - (pcie_table->dpm_levels[i].value == 1) ? "5.0GT/s, x16" : - (pcie_table->dpm_levels[i].value == 2) ? "8.0GT/s, x16" : "", - (i == now) ? "*" : ""); - break; - default: - break; - } - - return size; -} - -static int ci_dpm_force_clock_level(void *handle, - enum pp_clock_type type, uint32_t mask) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - - if (adev->pm.dpm.forced_level != AMD_DPM_FORCED_LEVEL_MANUAL) - return -EINVAL; - - if (mask == 0) - return -EINVAL; - - switch (type) { - case PP_SCLK: - if (!pi->sclk_dpm_key_disabled) - amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_SCLKDPM_SetEnabledMask, - pi->dpm_level_enable_mask.sclk_dpm_enable_mask & mask); - break; - - case PP_MCLK: - if (!pi->mclk_dpm_key_disabled) - amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_MCLKDPM_SetEnabledMask, - pi->dpm_level_enable_mask.mclk_dpm_enable_mask & mask); - break; - - case PP_PCIE: - { - uint32_t tmp = mask & pi->dpm_level_enable_mask.pcie_dpm_enable_mask; - - if (!pi->pcie_dpm_key_disabled) { - if (fls(tmp) != ffs(tmp)) - amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_PCIeDPM_UnForceLevel); - else - amdgpu_ci_send_msg_to_smc_with_parameter(adev, - PPSMC_MSG_PCIeDPM_ForceLevel, - fls(tmp) - 1); - } - break; - } - default: - break; - } - - return 0; -} - -static int ci_dpm_get_sclk_od(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_single_dpm_table *sclk_table = &(pi->dpm_table.sclk_table); - struct ci_single_dpm_table *golden_sclk_table = - &(pi->golden_dpm_table.sclk_table); - int value; - - value = (sclk_table->dpm_levels[sclk_table->count - 1].value - - golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value) * - 100 / - golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value; - - return value; -} - -static int ci_dpm_set_sclk_od(void *handle, uint32_t value) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_ps *ps = ci_get_ps(adev->pm.dpm.requested_ps); - struct ci_single_dpm_table *golden_sclk_table = - &(pi->golden_dpm_table.sclk_table); - - if (value > 20) - value = 20; - - ps->performance_levels[ps->performance_level_count - 1].sclk = - golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value * - value / 100 + - golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value; - - return 0; -} - -static int ci_dpm_get_mclk_od(void *handle) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_single_dpm_table *mclk_table = &(pi->dpm_table.mclk_table); - struct ci_single_dpm_table *golden_mclk_table = - &(pi->golden_dpm_table.mclk_table); - int value; - - value = (mclk_table->dpm_levels[mclk_table->count - 1].value - - golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value) * - 100 / - golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value; - - return value; -} - -static int ci_dpm_set_mclk_od(void *handle, uint32_t value) -{ - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - struct ci_power_info *pi = ci_get_pi(adev); - struct ci_ps *ps = ci_get_ps(adev->pm.dpm.requested_ps); - struct ci_single_dpm_table *golden_mclk_table = - &(pi->golden_dpm_table.mclk_table); - - if (value > 20) - value = 20; - - ps->performance_levels[ps->performance_level_count - 1].mclk = - golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value * - value / 100 + - golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value; - - return 0; -} - -static int ci_dpm_read_sensor(void *handle, int idx, - void *value, int *size) -{ - u32 activity_percent = 50; - int ret; - struct amdgpu_device *adev = (struct amdgpu_device *)handle; - - /* size must be at least 4 bytes for all sensors */ - if (*size < 4) - return -EINVAL; - - switch (idx) { - case AMDGPU_PP_SENSOR_GFX_SCLK: - *((uint32_t *)value) = ci_get_average_sclk_freq(adev); - *size = 4; - return 0; - case AMDGPU_PP_SENSOR_GFX_MCLK: - *((uint32_t *)value) = ci_get_average_mclk_freq(adev); - *size = 4; - return 0; - case AMDGPU_PP_SENSOR_GPU_TEMP: - *((uint32_t *)value) = ci_dpm_get_temp(adev); - *size = 4; - return 0; - case AMDGPU_PP_SENSOR_GPU_LOAD: - ret = ci_read_smc_soft_register(adev, - offsetof(SMU7_SoftRegisters, - AverageGraphicsA), - &activity_percent); - if (ret == 0) { - activity_percent += 0x80; - activity_percent >>= 8; - activity_percent = - activity_percent > 100 ? 100 : activity_percent; - } - *((uint32_t *)value) = activity_percent; - *size = 4; - return 0; - default: - return -EINVAL; - } -} - -static int ci_set_powergating_by_smu(void *handle, - uint32_t block_type, bool gate) -{ - switch (block_type) { - case AMD_IP_BLOCK_TYPE_UVD: - ci_dpm_powergate_uvd(handle, gate); - break; - default: - break; - } - return 0; -} - -static const struct amd_ip_funcs ci_dpm_ip_funcs = { - .name = "ci_dpm", - .early_init = ci_dpm_early_init, - .late_init = ci_dpm_late_init, - .sw_init = ci_dpm_sw_init, - .sw_fini = ci_dpm_sw_fini, - .hw_init = ci_dpm_hw_init, - .hw_fini = ci_dpm_hw_fini, - .suspend = ci_dpm_suspend, - .resume = ci_dpm_resume, - .is_idle = ci_dpm_is_idle, - .wait_for_idle = ci_dpm_wait_for_idle, - .soft_reset = ci_dpm_soft_reset, - .set_clockgating_state = ci_dpm_set_clockgating_state, - .set_powergating_state = ci_dpm_set_powergating_state, -}; - -const struct amdgpu_ip_block_version ci_smu_ip_block = -{ - .type = AMD_IP_BLOCK_TYPE_SMC, - .major = 7, - .minor = 0, - .rev = 0, - .funcs = &ci_dpm_ip_funcs, -}; - -static const struct amd_pm_funcs ci_dpm_funcs = { - .pre_set_power_state = &ci_dpm_pre_set_power_state, - .set_power_state = &ci_dpm_set_power_state, - .post_set_power_state = &ci_dpm_post_set_power_state, - .display_configuration_changed = &ci_dpm_display_configuration_changed, - .get_sclk = &ci_dpm_get_sclk, - .get_mclk = &ci_dpm_get_mclk, - .print_power_state = &ci_dpm_print_power_state, - .debugfs_print_current_performance_level = &ci_dpm_debugfs_print_current_performance_level, - .force_performance_level = &ci_dpm_force_performance_level, - .vblank_too_short = &ci_dpm_vblank_too_short, - .set_powergating_by_smu = &ci_set_powergating_by_smu, - .set_fan_control_mode = &ci_dpm_set_fan_control_mode, - .get_fan_control_mode = &ci_dpm_get_fan_control_mode, - .set_fan_speed_percent = &ci_dpm_set_fan_speed_percent, - .get_fan_speed_percent = &ci_dpm_get_fan_speed_percent, - .print_clock_levels = ci_dpm_print_clock_levels, - .force_clock_level = ci_dpm_force_clock_level, - .get_sclk_od = ci_dpm_get_sclk_od, - .set_sclk_od = ci_dpm_set_sclk_od, - .get_mclk_od = ci_dpm_get_mclk_od, - .set_mclk_od = ci_dpm_set_mclk_od, - .check_state_equal = ci_check_state_equal, - .get_vce_clock_state = amdgpu_get_vce_clock_state, - .read_sensor = ci_dpm_read_sensor, -}; - -static const struct amdgpu_irq_src_funcs ci_dpm_irq_funcs = { - .set = ci_dpm_set_interrupt_state, - .process = ci_dpm_process_interrupt, -}; - -static void ci_dpm_set_irq_funcs(struct amdgpu_device *adev) -{ - adev->pm.dpm.thermal.irq.num_types = AMDGPU_THERMAL_IRQ_LAST; - adev->pm.dpm.thermal.irq.funcs = &ci_dpm_irq_funcs; -} diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.h b/drivers/gpu/drm/amd/amdgpu/ci_dpm.h deleted file mode 100644 index 91be2996ae7c..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.h +++ /dev/null @@ -1,349 +0,0 @@ -/* - * Copyright 2013 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - */ -#ifndef __CI_DPM_H__ -#define __CI_DPM_H__ - -#include "amdgpu_atombios.h" -#include "ppsmc.h" - -#define SMU__NUM_SCLK_DPM_STATE 8 -#define SMU__NUM_MCLK_DPM_LEVELS 6 -#define SMU__NUM_LCLK_DPM_LEVELS 8 -#define SMU__NUM_PCIE_DPM_LEVELS 8 -#include "smu7_discrete.h" - -#define CISLANDS_MAX_HARDWARE_POWERLEVELS 2 - -#define CISLANDS_UNUSED_GPIO_PIN 0x7F - -struct ci_pl { - u32 mclk; - u32 sclk; - enum amdgpu_pcie_gen pcie_gen; - u16 pcie_lane; -}; - -struct ci_ps { - u16 performance_level_count; - bool dc_compatible; - u32 sclk_t; - struct ci_pl performance_levels[CISLANDS_MAX_HARDWARE_POWERLEVELS]; -}; - -struct ci_dpm_level { - bool enabled; - u32 value; - u32 param1; -}; - -#define CISLAND_MAX_DEEPSLEEP_DIVIDER_ID 5 -#define MAX_REGULAR_DPM_NUMBER 8 -#define CISLAND_MINIMUM_ENGINE_CLOCK 800 - -struct ci_single_dpm_table { - u32 count; - struct ci_dpm_level dpm_levels[MAX_REGULAR_DPM_NUMBER]; -}; - -struct ci_dpm_table { - struct ci_single_dpm_table sclk_table; - struct ci_single_dpm_table mclk_table; - struct ci_single_dpm_table pcie_speed_table; - struct ci_single_dpm_table vddc_table; - struct ci_single_dpm_table vddci_table; - struct ci_single_dpm_table mvdd_table; -}; - -struct ci_mc_reg_entry { - u32 mclk_max; - u32 mc_data[SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE]; -}; - -struct ci_mc_reg_table { - u8 last; - u8 num_entries; - u16 valid_flag; - struct ci_mc_reg_entry mc_reg_table_entry[MAX_AC_TIMING_ENTRIES]; - SMU7_Discrete_MCRegisterAddress mc_reg_address[SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE]; -}; - -struct ci_ulv_parm -{ - bool supported; - u32 cg_ulv_parameter; - u32 volt_change_delay; - struct ci_pl pl; -}; - -#define CISLANDS_MAX_LEAKAGE_COUNT 8 - -struct ci_leakage_voltage { - u16 count; - u16 leakage_id[CISLANDS_MAX_LEAKAGE_COUNT]; - u16 actual_voltage[CISLANDS_MAX_LEAKAGE_COUNT]; -}; - -struct ci_dpm_level_enable_mask { - u32 uvd_dpm_enable_mask; - u32 vce_dpm_enable_mask; - u32 acp_dpm_enable_mask; - u32 samu_dpm_enable_mask; - u32 sclk_dpm_enable_mask; - u32 mclk_dpm_enable_mask; - u32 pcie_dpm_enable_mask; -}; - -struct ci_vbios_boot_state -{ - u16 mvdd_bootup_value; - u16 vddc_bootup_value; - u16 vddci_bootup_value; - u32 sclk_bootup_value; - u32 mclk_bootup_value; - u16 pcie_gen_bootup_value; - u16 pcie_lane_bootup_value; -}; - -struct ci_clock_registers { - u32 cg_spll_func_cntl; - u32 cg_spll_func_cntl_2; - u32 cg_spll_func_cntl_3; - u32 cg_spll_func_cntl_4; - u32 cg_spll_spread_spectrum; - u32 cg_spll_spread_spectrum_2; - u32 dll_cntl; - u32 mclk_pwrmgt_cntl; - u32 mpll_ad_func_cntl; - u32 mpll_dq_func_cntl; - u32 mpll_func_cntl; - u32 mpll_func_cntl_1; - u32 mpll_func_cntl_2; - u32 mpll_ss1; - u32 mpll_ss2; -}; - -struct ci_thermal_temperature_setting { - s32 temperature_low; - s32 temperature_high; - s32 temperature_shutdown; -}; - -struct ci_pcie_perf_range { - u16 max; - u16 min; -}; - -enum ci_pt_config_reg_type { - CISLANDS_CONFIGREG_MMR = 0, - CISLANDS_CONFIGREG_SMC_IND, - CISLANDS_CONFIGREG_DIDT_IND, - CISLANDS_CONFIGREG_CACHE, - CISLANDS_CONFIGREG_MAX -}; - -#define POWERCONTAINMENT_FEATURE_BAPM 0x00000001 -#define POWERCONTAINMENT_FEATURE_TDCLimit 0x00000002 -#define POWERCONTAINMENT_FEATURE_PkgPwrLimit 0x00000004 - -struct ci_pt_config_reg { - u32 offset; - u32 mask; - u32 shift; - u32 value; - enum ci_pt_config_reg_type type; -}; - -struct ci_pt_defaults { - u8 svi_load_line_en; - u8 svi_load_line_vddc; - u8 tdc_vddc_throttle_release_limit_perc; - u8 tdc_mawt; - u8 tdc_waterfall_ctl; - u8 dte_ambient_temp_base; - u32 display_cac; - u32 bapm_temp_gradient; - u16 bapmti_r[SMU7_DTE_ITERATIONS * SMU7_DTE_SOURCES * SMU7_DTE_SINKS]; - u16 bapmti_rc[SMU7_DTE_ITERATIONS * SMU7_DTE_SOURCES * SMU7_DTE_SINKS]; -}; - -#define DPMTABLE_OD_UPDATE_SCLK 0x00000001 -#define DPMTABLE_OD_UPDATE_MCLK 0x00000002 -#define DPMTABLE_UPDATE_SCLK 0x00000004 -#define DPMTABLE_UPDATE_MCLK 0x00000008 - -struct ci_power_info { - struct ci_dpm_table dpm_table; - struct ci_dpm_table golden_dpm_table; - u32 voltage_control; - u32 mvdd_control; - u32 vddci_control; - u32 active_auto_throttle_sources; - struct ci_clock_registers clock_registers; - u16 acpi_vddc; - u16 acpi_vddci; - enum amdgpu_pcie_gen force_pcie_gen; - enum amdgpu_pcie_gen acpi_pcie_gen; - struct ci_leakage_voltage vddc_leakage; - struct ci_leakage_voltage vddci_leakage; - u16 max_vddc_in_pp_table; - u16 min_vddc_in_pp_table; - u16 max_vddci_in_pp_table; - u16 min_vddci_in_pp_table; - u32 mclk_strobe_mode_threshold; - u32 mclk_stutter_mode_threshold; - u32 mclk_edc_enable_threshold; - u32 mclk_edc_wr_enable_threshold; - struct ci_vbios_boot_state vbios_boot_state; - /* smc offsets */ - u32 sram_end; - u32 dpm_table_start; - u32 soft_regs_start; - u32 mc_reg_table_start; - u32 fan_table_start; - u32 arb_table_start; - /* smc tables */ - SMU7_Discrete_DpmTable smc_state_table; - SMU7_Discrete_MCRegisters smc_mc_reg_table; - SMU7_Discrete_PmFuses smc_powertune_table; - /* other stuff */ - struct ci_mc_reg_table mc_reg_table; - struct atom_voltage_table vddc_voltage_table; - struct atom_voltage_table vddci_voltage_table; - struct atom_voltage_table mvdd_voltage_table; - struct ci_ulv_parm ulv; - u32 power_containment_features; - const struct ci_pt_defaults *powertune_defaults; - u32 dte_tj_offset; - bool vddc_phase_shed_control; - struct ci_thermal_temperature_setting thermal_temp_setting; - struct ci_dpm_level_enable_mask dpm_level_enable_mask; - u32 need_update_smu7_dpm_table; - u32 sclk_dpm_key_disabled; - u32 mclk_dpm_key_disabled; - u32 pcie_dpm_key_disabled; - u32 thermal_sclk_dpm_enabled; - struct ci_pcie_perf_range pcie_gen_performance; - struct ci_pcie_perf_range pcie_lane_performance; - struct ci_pcie_perf_range pcie_gen_powersaving; - struct ci_pcie_perf_range pcie_lane_powersaving; - u32 activity_target[SMU7_MAX_LEVELS_GRAPHICS]; - u32 mclk_activity_target; - u32 low_sclk_interrupt_t; - u32 last_mclk_dpm_enable_mask; - u32 sys_pcie_mask; - /* caps */ - bool caps_power_containment; - bool caps_cac; - bool caps_sq_ramping; - bool caps_db_ramping; - bool caps_td_ramping; - bool caps_tcp_ramping; - bool caps_fps; - bool caps_sclk_ds; - bool caps_sclk_ss_support; - bool caps_mclk_ss_support; - bool caps_uvd_dpm; - bool caps_vce_dpm; - bool caps_samu_dpm; - bool caps_acp_dpm; - bool caps_automatic_dc_transition; - bool caps_sclk_throttle_low_notification; - bool caps_dynamic_ac_timing; - bool caps_od_fuzzy_fan_control_support; - /* flags */ - bool thermal_protection; - bool pcie_performance_request; - bool dynamic_ss; - bool dll_default_on; - bool cac_enabled; - bool uvd_enabled; - bool battery_state; - bool pspp_notify_required; - bool enable_bapm_feature; - bool enable_tdc_limit_feature; - bool enable_pkg_pwr_tracking_feature; - bool use_pcie_performance_levels; - bool use_pcie_powersaving_levels; - bool uvd_power_gated; - /* driver states */ - struct amdgpu_ps current_rps; - struct ci_ps current_ps; - struct amdgpu_ps requested_rps; - struct ci_ps requested_ps; - /* fan control */ - bool fan_ctrl_is_in_default_mode; - bool fan_is_controlled_by_smc; - u32 t_min; - u32 fan_ctrl_default_mode; -}; - -#define CISLANDS_VOLTAGE_CONTROL_NONE 0x0 -#define CISLANDS_VOLTAGE_CONTROL_BY_GPIO 0x1 -#define CISLANDS_VOLTAGE_CONTROL_BY_SVID2 0x2 - -#define CISLANDS_Q88_FORMAT_CONVERSION_UNIT 256 - -#define CISLANDS_VRC_DFLT0 0x3FFFC000 -#define CISLANDS_VRC_DFLT1 0x000400 -#define CISLANDS_VRC_DFLT2 0xC00080 -#define CISLANDS_VRC_DFLT3 0xC00200 -#define CISLANDS_VRC_DFLT4 0xC01680 -#define CISLANDS_VRC_DFLT5 0xC00033 -#define CISLANDS_VRC_DFLT6 0xC00033 -#define CISLANDS_VRC_DFLT7 0x3FFFC000 - -#define CISLANDS_CGULVPARAMETER_DFLT 0x00040035 -#define CISLAND_TARGETACTIVITY_DFLT 30 -#define CISLAND_MCLK_TARGETACTIVITY_DFLT 10 - -#define PCIE_PERF_REQ_REMOVE_REGISTRY 0 -#define PCIE_PERF_REQ_FORCE_LOWPOWER 1 -#define PCIE_PERF_REQ_PECI_GEN1 2 -#define PCIE_PERF_REQ_PECI_GEN2 3 -#define PCIE_PERF_REQ_PECI_GEN3 4 - -#define CISLANDS_SSTU_DFLT 0 -#define CISLANDS_SST_DFLT 0x00C8 - -/* XXX are these ok? */ -#define CISLANDS_TEMP_RANGE_MIN (90 * 1000) -#define CISLANDS_TEMP_RANGE_MAX (120 * 1000) - -int amdgpu_ci_copy_bytes_to_smc(struct amdgpu_device *adev, - u32 smc_start_address, - const u8 *src, u32 byte_count, u32 limit); -void amdgpu_ci_start_smc(struct amdgpu_device *adev); -void amdgpu_ci_reset_smc(struct amdgpu_device *adev); -int amdgpu_ci_program_jump_on_start(struct amdgpu_device *adev); -void amdgpu_ci_stop_smc_clock(struct amdgpu_device *adev); -void amdgpu_ci_start_smc_clock(struct amdgpu_device *adev); -bool amdgpu_ci_is_smc_running(struct amdgpu_device *adev); -PPSMC_Result amdgpu_ci_send_msg_to_smc(struct amdgpu_device *adev, PPSMC_Msg msg); -PPSMC_Result amdgpu_ci_wait_for_smc_inactive(struct amdgpu_device *adev); -int amdgpu_ci_load_smc_ucode(struct amdgpu_device *adev, u32 limit); -int amdgpu_ci_read_smc_sram_dword(struct amdgpu_device *adev, - u32 smc_address, u32 *value, u32 limit); -int amdgpu_ci_write_smc_sram_dword(struct amdgpu_device *adev, - u32 smc_address, u32 value, u32 limit); - -#endif diff --git a/drivers/gpu/drm/amd/amdgpu/ci_smc.c b/drivers/gpu/drm/amd/amdgpu/ci_smc.c deleted file mode 100644 index b8ba51e045b5..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/ci_smc.c +++ /dev/null @@ -1,279 +0,0 @@ -/* - * Copyright 2011 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: Alex Deucher - */ - -#include <linux/firmware.h> -#include <drm/drmP.h> -#include "amdgpu.h" -#include "cikd.h" -#include "ppsmc.h" -#include "amdgpu_ucode.h" -#include "ci_dpm.h" - -#include "smu/smu_7_0_1_d.h" -#include "smu/smu_7_0_1_sh_mask.h" - -static int ci_set_smc_sram_address(struct amdgpu_device *adev, - u32 smc_address, u32 limit) -{ - if (smc_address & 3) - return -EINVAL; - if ((smc_address + 3) > limit) - return -EINVAL; - - WREG32(mmSMC_IND_INDEX_0, smc_address); - WREG32_P(mmSMC_IND_ACCESS_CNTL, 0, ~SMC_IND_ACCESS_CNTL__AUTO_INCREMENT_IND_0_MASK); - - return 0; -} - -int amdgpu_ci_copy_bytes_to_smc(struct amdgpu_device *adev, - u32 smc_start_address, - const u8 *src, u32 byte_count, u32 limit) -{ - unsigned long flags; - u32 data, original_data; - u32 addr; - u32 extra_shift; - int ret = 0; - - if (smc_start_address & 3) - return -EINVAL; - if ((smc_start_address + byte_count) > limit) - return -EINVAL; - - addr = smc_start_address; - - spin_lock_irqsave(&adev->smc_idx_lock, flags); - while (byte_count >= 4) { - /* SMC address space is BE */ - data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3]; - - ret = ci_set_smc_sram_address(adev, addr, limit); - if (ret) - goto done; - - WREG32(mmSMC_IND_DATA_0, data); - - src += 4; - byte_count -= 4; - addr += 4; - } - - /* RMW for the final bytes */ - if (byte_count > 0) { - data = 0; - - ret = ci_set_smc_sram_address(adev, addr, limit); - if (ret) - goto done; - - original_data = RREG32(mmSMC_IND_DATA_0); - - extra_shift = 8 * (4 - byte_count); - - while (byte_count > 0) { - data = (data << 8) + *src++; - byte_count--; - } - - data <<= extra_shift; - - data |= (original_data & ~((~0UL) << extra_shift)); - - ret = ci_set_smc_sram_address(adev, addr, limit); - if (ret) - goto done; - - WREG32(mmSMC_IND_DATA_0, data); - } - -done: - spin_unlock_irqrestore(&adev->smc_idx_lock, flags); - - return ret; -} - -void amdgpu_ci_start_smc(struct amdgpu_device *adev) -{ - u32 tmp = RREG32_SMC(ixSMC_SYSCON_RESET_CNTL); - - tmp &= ~SMC_SYSCON_RESET_CNTL__rst_reg_MASK; - WREG32_SMC(ixSMC_SYSCON_RESET_CNTL, tmp); -} - -void amdgpu_ci_reset_smc(struct amdgpu_device *adev) -{ - u32 tmp = RREG32_SMC(ixSMC_SYSCON_RESET_CNTL); - - tmp |= SMC_SYSCON_RESET_CNTL__rst_reg_MASK; - WREG32_SMC(ixSMC_SYSCON_RESET_CNTL, tmp); -} - -int amdgpu_ci_program_jump_on_start(struct amdgpu_device *adev) -{ - static u8 data[] = { 0xE0, 0x00, 0x80, 0x40 }; - - return amdgpu_ci_copy_bytes_to_smc(adev, 0x0, data, 4, sizeof(data)+1); -} - -void amdgpu_ci_stop_smc_clock(struct amdgpu_device *adev) -{ - u32 tmp = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0); - - tmp |= SMC_SYSCON_CLOCK_CNTL_0__ck_disable_MASK; - - WREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0, tmp); -} - -void amdgpu_ci_start_smc_clock(struct amdgpu_device *adev) -{ - u32 tmp = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0); - - tmp &= ~SMC_SYSCON_CLOCK_CNTL_0__ck_disable_MASK; - - WREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0, tmp); -} - -bool amdgpu_ci_is_smc_running(struct amdgpu_device *adev) -{ - u32 clk = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0); - u32 pc_c = RREG32_SMC(ixSMC_PC_C); - - if (!(clk & SMC_SYSCON_CLOCK_CNTL_0__ck_disable_MASK) && (0x20100 <= pc_c)) - return true; - - return false; -} - -PPSMC_Result amdgpu_ci_send_msg_to_smc(struct amdgpu_device *adev, PPSMC_Msg msg) -{ - u32 tmp; - int i; - - if (!amdgpu_ci_is_smc_running(adev)) - return PPSMC_Result_Failed; - - WREG32(mmSMC_MESSAGE_0, msg); - - for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(mmSMC_RESP_0); - if (tmp != 0) - break; - udelay(1); - } - tmp = RREG32(mmSMC_RESP_0); - - return (PPSMC_Result)tmp; -} - -PPSMC_Result amdgpu_ci_wait_for_smc_inactive(struct amdgpu_device *adev) -{ - u32 tmp; - int i; - - if (!amdgpu_ci_is_smc_running(adev)) - return PPSMC_Result_OK; - - for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0); - if ((tmp & SMC_SYSCON_CLOCK_CNTL_0__cken_MASK) == 0) - break; - udelay(1); - } - - return PPSMC_Result_OK; -} - -int amdgpu_ci_load_smc_ucode(struct amdgpu_device *adev, u32 limit) -{ - const struct smc_firmware_header_v1_0 *hdr; - unsigned long flags; - u32 ucode_start_address; - u32 ucode_size; - const u8 *src; - u32 data; - - if (!adev->pm.fw) - return -EINVAL; - - hdr = (const struct smc_firmware_header_v1_0 *)adev->pm.fw->data; - amdgpu_ucode_print_smc_hdr(&hdr->header); - - adev->pm.fw_version = le32_to_cpu(hdr->header.ucode_version); - ucode_start_address = le32_to_cpu(hdr->ucode_start_addr); - ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes); - src = (const u8 *) - (adev->pm.fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes)); - - if (ucode_size & 3) - return -EINVAL; - - spin_lock_irqsave(&adev->smc_idx_lock, flags); - WREG32(mmSMC_IND_INDEX_0, ucode_start_address); - WREG32_P(mmSMC_IND_ACCESS_CNTL, SMC_IND_ACCESS_CNTL__AUTO_INCREMENT_IND_0_MASK, - ~SMC_IND_ACCESS_CNTL__AUTO_INCREMENT_IND_0_MASK); - while (ucode_size >= 4) { - /* SMC address space is BE */ - data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3]; - - WREG32(mmSMC_IND_DATA_0, data); - - src += 4; - ucode_size -= 4; - } - WREG32_P(mmSMC_IND_ACCESS_CNTL, 0, ~SMC_IND_ACCESS_CNTL__AUTO_INCREMENT_IND_0_MASK); - spin_unlock_irqrestore(&adev->smc_idx_lock, flags); - - return 0; -} - -int amdgpu_ci_read_smc_sram_dword(struct amdgpu_device *adev, - u32 smc_address, u32 *value, u32 limit) -{ - unsigned long flags; - int ret; - - spin_lock_irqsave(&adev->smc_idx_lock, flags); - ret = ci_set_smc_sram_address(adev, smc_address, limit); - if (ret == 0) - *value = RREG32(mmSMC_IND_DATA_0); - spin_unlock_irqrestore(&adev->smc_idx_lock, flags); - - return ret; -} - -int amdgpu_ci_write_smc_sram_dword(struct amdgpu_device *adev, - u32 smc_address, u32 value, u32 limit) -{ - unsigned long flags; - int ret; - - spin_lock_irqsave(&adev->smc_idx_lock, flags); - ret = ci_set_smc_sram_address(adev, smc_address, limit); - if (ret == 0) - WREG32(mmSMC_IND_DATA_0, value); - spin_unlock_irqrestore(&adev->smc_idx_lock, flags); - - return ret; -} diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c index 71c50d8900e3..07c1f239e9c3 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik.c +++ b/drivers/gpu/drm/amd/amdgpu/cik.c @@ -1741,6 +1741,69 @@ static bool cik_need_full_reset(struct amdgpu_device *adev) return true; } +static void cik_get_pcie_usage(struct amdgpu_device *adev, uint64_t *count0, + uint64_t *count1) +{ + uint32_t perfctr = 0; + uint64_t cnt0_of, cnt1_of; + int tmp; + + /* This reports 0 on APUs, so return to avoid writing/reading registers + * that may or may not be different from their GPU counterparts + */ + if (adev->flags & AMD_IS_APU) + return; + + /* Set the 2 events that we wish to watch, defined above */ + /* Reg 40 is # received msgs, Reg 104 is # of posted requests sent */ + perfctr = REG_SET_FIELD(perfctr, PCIE_PERF_CNTL_TXCLK, EVENT0_SEL, 40); + perfctr = REG_SET_FIELD(perfctr, PCIE_PERF_CNTL_TXCLK, EVENT1_SEL, 104); + + /* Write to enable desired perf counters */ + WREG32_PCIE(ixPCIE_PERF_CNTL_TXCLK, perfctr); + /* Zero out and enable the perf counters + * Write 0x5: + * Bit 0 = Start all counters(1) + * Bit 2 = Global counter reset enable(1) + */ + WREG32_PCIE(ixPCIE_PERF_COUNT_CNTL, 0x00000005); + + msleep(1000); + + /* Load the shadow and disable the perf counters + * Write 0x2: + * Bit 0 = Stop counters(0) + * Bit 1 = Load the shadow counters(1) + */ + WREG32_PCIE(ixPCIE_PERF_COUNT_CNTL, 0x00000002); + + /* Read register values to get any >32bit overflow */ + tmp = RREG32_PCIE(ixPCIE_PERF_CNTL_TXCLK); + cnt0_of = REG_GET_FIELD(tmp, PCIE_PERF_CNTL_TXCLK, COUNTER0_UPPER); + cnt1_of = REG_GET_FIELD(tmp, PCIE_PERF_CNTL_TXCLK, COUNTER1_UPPER); + + /* Get the values and add the overflow */ + *count0 = RREG32_PCIE(ixPCIE_PERF_COUNT0_TXCLK) | (cnt0_of << 32); + *count1 = RREG32_PCIE(ixPCIE_PERF_COUNT1_TXCLK) | (cnt1_of << 32); +} + +static bool cik_need_reset_on_init(struct amdgpu_device *adev) +{ + u32 clock_cntl, pc; + + if (adev->flags & AMD_IS_APU) + return false; + + /* check if the SMC is already running */ + clock_cntl = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0); + pc = RREG32_SMC(ixSMC_PC_C); + if ((0 == REG_GET_FIELD(clock_cntl, SMC_SYSCON_CLOCK_CNTL_0, ck_disable)) && + (0x20100 <= pc)) + return true; + + return false; +} + static const struct amdgpu_asic_funcs cik_asic_funcs = { .read_disabled_bios = &cik_read_disabled_bios, @@ -1756,6 +1819,8 @@ static const struct amdgpu_asic_funcs cik_asic_funcs = .invalidate_hdp = &cik_invalidate_hdp, .need_full_reset = &cik_need_full_reset, .init_doorbell_index = &legacy_doorbell_index_init, + .get_pcie_usage = &cik_get_pcie_usage, + .need_reset_on_init = &cik_need_reset_on_init, }; static int cik_common_early_init(void *handle) @@ -2005,10 +2070,7 @@ int cik_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &cik_ih_ip_block); amdgpu_device_ip_block_add(adev, &gfx_v7_2_ip_block); amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block); - if (amdgpu_dpm == -1) - amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); - else - amdgpu_device_ip_block_add(adev, &ci_smu_ip_block); + amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); if (adev->enable_virtual_display) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); #if defined(CONFIG_DRM_AMD_DC) @@ -2026,10 +2088,7 @@ int cik_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &cik_ih_ip_block); amdgpu_device_ip_block_add(adev, &gfx_v7_3_ip_block); amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block); - if (amdgpu_dpm == -1) - amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); - else - amdgpu_device_ip_block_add(adev, &ci_smu_ip_block); + amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); if (adev->enable_virtual_display) amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block); #if defined(CONFIG_DRM_AMD_DC) diff --git a/drivers/gpu/drm/amd/amdgpu/cik_dpm.h b/drivers/gpu/drm/amd/amdgpu/cik_dpm.h index 2a086610f74d..2fcc4b60153c 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_dpm.h +++ b/drivers/gpu/drm/amd/amdgpu/cik_dpm.h @@ -24,7 +24,6 @@ #ifndef __CIK_DPM_H__ #define __CIK_DPM_H__ -extern const struct amdgpu_ip_block_version ci_smu_ip_block; extern const struct amdgpu_ip_block_version kv_smu_ip_block; #endif diff --git a/drivers/gpu/drm/amd/amdgpu/cik_ih.c b/drivers/gpu/drm/amd/amdgpu/cik_ih.c index 8a8b4967a101..721c757156e8 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_ih.c @@ -103,9 +103,9 @@ static void cik_ih_disable_interrupts(struct amdgpu_device *adev) */ static int cik_ih_irq_init(struct amdgpu_device *adev) { + struct amdgpu_ih_ring *ih = &adev->irq.ih; int rb_bufsz; u32 interrupt_cntl, ih_cntl, ih_rb_cntl; - u64 wptr_off; /* disable irqs */ cik_ih_disable_interrupts(adev); @@ -131,9 +131,8 @@ static int cik_ih_irq_init(struct amdgpu_device *adev) ih_rb_cntl |= IH_RB_CNTL__WPTR_WRITEBACK_ENABLE_MASK; /* set the writeback address whether it's enabled or not */ - wptr_off = adev->wb.gpu_addr + (adev->irq.ih.wptr_offs * 4); - WREG32(mmIH_RB_WPTR_ADDR_LO, lower_32_bits(wptr_off)); - WREG32(mmIH_RB_WPTR_ADDR_HI, upper_32_bits(wptr_off) & 0xFF); + WREG32(mmIH_RB_WPTR_ADDR_LO, lower_32_bits(ih->wptr_addr)); + WREG32(mmIH_RB_WPTR_ADDR_HI, upper_32_bits(ih->wptr_addr) & 0xFF); WREG32(mmIH_RB_CNTL, ih_rb_cntl); @@ -183,11 +182,12 @@ static void cik_ih_irq_disable(struct amdgpu_device *adev) * Used by cik_irq_process(). * Returns the value of the wptr. */ -static u32 cik_ih_get_wptr(struct amdgpu_device *adev) +static u32 cik_ih_get_wptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { u32 wptr, tmp; - wptr = le32_to_cpu(adev->wb.wb[adev->irq.ih.wptr_offs]); + wptr = le32_to_cpu(*ih->wptr_cpu); if (wptr & IH_RB_WPTR__RB_OVERFLOW_MASK) { wptr &= ~IH_RB_WPTR__RB_OVERFLOW_MASK; @@ -196,13 +196,13 @@ static u32 cik_ih_get_wptr(struct amdgpu_device *adev) * this should allow us to catchup. */ dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n", - wptr, adev->irq.ih.rptr, (wptr + 16) & adev->irq.ih.ptr_mask); - adev->irq.ih.rptr = (wptr + 16) & adev->irq.ih.ptr_mask; + wptr, ih->rptr, (wptr + 16) & ih->ptr_mask); + ih->rptr = (wptr + 16) & ih->ptr_mask; tmp = RREG32(mmIH_RB_CNTL); tmp |= IH_RB_CNTL__WPTR_OVERFLOW_CLEAR_MASK; WREG32(mmIH_RB_CNTL, tmp); } - return (wptr & adev->irq.ih.ptr_mask); + return (wptr & ih->ptr_mask); } /* CIK IV Ring @@ -237,16 +237,17 @@ static u32 cik_ih_get_wptr(struct amdgpu_device *adev) * position and also advance the position. */ static void cik_ih_decode_iv(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih, struct amdgpu_iv_entry *entry) { /* wptr/rptr are in bytes! */ - u32 ring_index = adev->irq.ih.rptr >> 2; + u32 ring_index = ih->rptr >> 2; uint32_t dw[4]; - dw[0] = le32_to_cpu(adev->irq.ih.ring[ring_index + 0]); - dw[1] = le32_to_cpu(adev->irq.ih.ring[ring_index + 1]); - dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]); - dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); + dw[0] = le32_to_cpu(ih->ring[ring_index + 0]); + dw[1] = le32_to_cpu(ih->ring[ring_index + 1]); + dw[2] = le32_to_cpu(ih->ring[ring_index + 2]); + dw[3] = le32_to_cpu(ih->ring[ring_index + 3]); entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY; entry->src_id = dw[0] & 0xff; @@ -256,7 +257,7 @@ static void cik_ih_decode_iv(struct amdgpu_device *adev, entry->pasid = (dw[2] >> 16) & 0xffff; /* wptr/rptr are in bytes! */ - adev->irq.ih.rptr += 16; + ih->rptr += 16; } /** @@ -266,9 +267,10 @@ static void cik_ih_decode_iv(struct amdgpu_device *adev, * * Set the IH ring buffer rptr. */ -static void cik_ih_set_rptr(struct amdgpu_device *adev) +static void cik_ih_set_rptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { - WREG32(mmIH_RB_RPTR, adev->irq.ih.rptr); + WREG32(mmIH_RB_RPTR, ih->rptr); } static int cik_ih_early_init(void *handle) diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c index 45795191de1f..189599b694e8 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c @@ -220,7 +220,7 @@ static void cik_sdma_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count) static void cik_sdma_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 extra_bits = vmid & 0xf; diff --git a/drivers/gpu/drm/amd/amdgpu/cz_ih.c b/drivers/gpu/drm/amd/amdgpu/cz_ih.c index 9d3ea298e116..61024b9c7a4b 100644 --- a/drivers/gpu/drm/amd/amdgpu/cz_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/cz_ih.c @@ -103,9 +103,9 @@ static void cz_ih_disable_interrupts(struct amdgpu_device *adev) */ static int cz_ih_irq_init(struct amdgpu_device *adev) { - int rb_bufsz; + struct amdgpu_ih_ring *ih = &adev->irq.ih; u32 interrupt_cntl, ih_cntl, ih_rb_cntl; - u64 wptr_off; + int rb_bufsz; /* disable irqs */ cz_ih_disable_interrupts(adev); @@ -133,9 +133,8 @@ static int cz_ih_irq_init(struct amdgpu_device *adev) ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, WPTR_WRITEBACK_ENABLE, 1); /* set the writeback address whether it's enabled or not */ - wptr_off = adev->wb.gpu_addr + (adev->irq.ih.wptr_offs * 4); - WREG32(mmIH_RB_WPTR_ADDR_LO, lower_32_bits(wptr_off)); - WREG32(mmIH_RB_WPTR_ADDR_HI, upper_32_bits(wptr_off) & 0xFF); + WREG32(mmIH_RB_WPTR_ADDR_LO, lower_32_bits(ih->wptr_addr)); + WREG32(mmIH_RB_WPTR_ADDR_HI, upper_32_bits(ih->wptr_addr) & 0xFF); WREG32(mmIH_RB_CNTL, ih_rb_cntl); @@ -185,11 +184,12 @@ static void cz_ih_irq_disable(struct amdgpu_device *adev) * Used by cz_irq_process(VI). * Returns the value of the wptr. */ -static u32 cz_ih_get_wptr(struct amdgpu_device *adev) +static u32 cz_ih_get_wptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { u32 wptr, tmp; - wptr = le32_to_cpu(adev->wb.wb[adev->irq.ih.wptr_offs]); + wptr = le32_to_cpu(*ih->wptr_cpu); if (REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) { wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0); @@ -198,13 +198,13 @@ static u32 cz_ih_get_wptr(struct amdgpu_device *adev) * this should allow us to catchup. */ dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n", - wptr, adev->irq.ih.rptr, (wptr + 16) & adev->irq.ih.ptr_mask); - adev->irq.ih.rptr = (wptr + 16) & adev->irq.ih.ptr_mask; + wptr, ih->rptr, (wptr + 16) & ih->ptr_mask); + ih->rptr = (wptr + 16) & ih->ptr_mask; tmp = RREG32(mmIH_RB_CNTL); tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); WREG32(mmIH_RB_CNTL, tmp); } - return (wptr & adev->irq.ih.ptr_mask); + return (wptr & ih->ptr_mask); } /** @@ -216,16 +216,17 @@ static u32 cz_ih_get_wptr(struct amdgpu_device *adev) * position and also advance the position. */ static void cz_ih_decode_iv(struct amdgpu_device *adev, - struct amdgpu_iv_entry *entry) + struct amdgpu_ih_ring *ih, + struct amdgpu_iv_entry *entry) { /* wptr/rptr are in bytes! */ - u32 ring_index = adev->irq.ih.rptr >> 2; + u32 ring_index = ih->rptr >> 2; uint32_t dw[4]; - dw[0] = le32_to_cpu(adev->irq.ih.ring[ring_index + 0]); - dw[1] = le32_to_cpu(adev->irq.ih.ring[ring_index + 1]); - dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]); - dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); + dw[0] = le32_to_cpu(ih->ring[ring_index + 0]); + dw[1] = le32_to_cpu(ih->ring[ring_index + 1]); + dw[2] = le32_to_cpu(ih->ring[ring_index + 2]); + dw[3] = le32_to_cpu(ih->ring[ring_index + 3]); entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY; entry->src_id = dw[0] & 0xff; @@ -235,7 +236,7 @@ static void cz_ih_decode_iv(struct amdgpu_device *adev, entry->pasid = (dw[2] >> 16) & 0xffff; /* wptr/rptr are in bytes! */ - adev->irq.ih.rptr += 16; + ih->rptr += 16; } /** @@ -245,9 +246,10 @@ static void cz_ih_decode_iv(struct amdgpu_device *adev, * * Set the IH ring buffer rptr. */ -static void cz_ih_set_rptr(struct amdgpu_device *adev) +static void cz_ih_set_rptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { - WREG32(mmIH_RB_RPTR, adev->irq.ih.rptr); + WREG32(mmIH_RB_RPTR, ih->rptr); } static int cz_ih_early_init(void *handle) diff --git a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c index fdace004544d..e4cc1d48eaab 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_virtual.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_virtual.c @@ -167,19 +167,6 @@ static void dce_virtual_crtc_disable(struct drm_crtc *crtc) struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); dce_virtual_crtc_dpms(crtc, DRM_MODE_DPMS_OFF); - if (crtc->primary->fb) { - int r; - struct amdgpu_bo *abo; - - abo = gem_to_amdgpu_bo(crtc->primary->fb->obj[0]); - r = amdgpu_bo_reserve(abo, true); - if (unlikely(r)) - DRM_ERROR("failed to reserve abo before unpin\n"); - else { - amdgpu_bo_unpin(abo); - amdgpu_bo_unreserve(abo); - } - } amdgpu_crtc->pll_id = ATOM_PPLL_INVALID; amdgpu_crtc->encoder = NULL; @@ -692,7 +679,9 @@ static int dce_virtual_pageflip(struct amdgpu_device *adev, spin_unlock_irqrestore(&adev->ddev->event_lock, flags); drm_crtc_vblank_put(&amdgpu_crtc->base); - schedule_work(&works->unpin_work); + amdgpu_bo_unref(&works->old_abo); + kfree(works->shared); + kfree(works); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c index 1dc3013ea1d5..305276c7e4bf 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c @@ -1842,13 +1842,13 @@ static void gfx_v6_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, static void gfx_v6_0_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 header, control = 0; /* insert SWITCH_BUFFER packet before first IB in the ring frame */ - if (ctx_switch) { + if (flags & AMDGPU_HAVE_CTX_SWITCH) { amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0)); amdgpu_ring_write(ring, 0); } diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index 3a9fb6018c16..a59e0fdf5a97 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -2228,13 +2228,13 @@ static void gfx_v7_0_ring_emit_fence_compute(struct amdgpu_ring *ring, static void gfx_v7_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 header, control = 0; /* insert SWITCH_BUFFER packet before first IB in the ring frame */ - if (ctx_switch) { + if (flags & AMDGPU_HAVE_CTX_SWITCH) { amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0)); amdgpu_ring_write(ring, 0); } @@ -2259,11 +2259,27 @@ static void gfx_v7_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, static void gfx_v7_0_ring_emit_ib_compute(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24); + /* Currently, there is a high possibility to get wave ID mismatch + * between ME and GDS, leading to a hw deadlock, because ME generates + * different wave IDs than the GDS expects. This situation happens + * randomly when at least 5 compute pipes use GDS ordered append. + * The wave IDs generated by ME are also wrong after suspend/resume. + * Those are probably bugs somewhere else in the kernel driver. + * + * Writing GDS_COMPUTE_MAX_WAVE_ID resets wave ID counters in ME and + * GDS to 0 for this ring (me/pipe). + */ + if (ib->flags & AMDGPU_IB_FLAG_RESET_GDS_MAX_WAVE_ID) { + amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + amdgpu_ring_write(ring, mmGDS_COMPUTE_MAX_WAVE_ID - PACKET3_SET_CONFIG_REG_START); + amdgpu_ring_write(ring, ring->adev->gds.gds_compute_max_wave_id); + } + amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); amdgpu_ring_write(ring, #ifdef __BIG_ENDIAN @@ -5000,7 +5016,7 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = { 7 + /* gfx_v7_0_ring_emit_pipeline_sync */ CIK_FLUSH_GPU_TLB_NUM_WREG * 5 + 7 + /* gfx_v7_0_ring_emit_vm_flush */ 7 + 7 + 7, /* gfx_v7_0_ring_emit_fence_compute x3 for user fence, vm fence */ - .emit_ib_size = 4, /* gfx_v7_0_ring_emit_ib_compute */ + .emit_ib_size = 7, /* gfx_v7_0_ring_emit_ib_compute */ .emit_ib = gfx_v7_0_ring_emit_ib_compute, .emit_fence = gfx_v7_0_ring_emit_fence_compute, .emit_pipeline_sync = gfx_v7_0_ring_emit_pipeline_sync, @@ -5057,6 +5073,7 @@ static void gfx_v7_0_set_gds_init(struct amdgpu_device *adev) adev->gds.mem.total_size = RREG32(mmGDS_VMID0_SIZE); adev->gds.gws.total_size = 64; adev->gds.oa.total_size = 16; + adev->gds.gds_compute_max_wave_id = RREG32(mmGDS_COMPUTE_MAX_WAVE_ID); if (adev->gds.mem.total_size == 64 * 1024) { adev->gds.mem.gfx_partition_size = 4096; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index 381f593b0cda..b8e50a34bdb3 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -4233,7 +4233,6 @@ static int gfx_v8_0_cp_gfx_resume(struct amdgpu_device *adev) u32 tmp; u32 rb_bufsz; u64 rb_addr, rptr_addr, wptr_gpu_addr; - int r; /* Set the write pointer delay */ WREG32(mmCP_RB_WPTR_DELAY, 0); @@ -4278,9 +4277,8 @@ static int gfx_v8_0_cp_gfx_resume(struct amdgpu_device *adev) amdgpu_ring_clear_ring(ring); gfx_v8_0_cp_gfx_start(adev); ring->sched.ready = true; - r = amdgpu_ring_test_helper(ring); - return r; + return 0; } static void gfx_v8_0_cp_compute_enable(struct amdgpu_device *adev, bool enable) @@ -4369,10 +4367,9 @@ static int gfx_v8_0_kiq_kcq_enable(struct amdgpu_device *adev) amdgpu_ring_write(kiq_ring, upper_32_bits(wptr_addr)); } - r = amdgpu_ring_test_helper(kiq_ring); - if (r) - DRM_ERROR("KCQ enable failed\n"); - return r; + amdgpu_ring_commit(kiq_ring); + + return 0; } static int gfx_v8_0_deactivate_hqd(struct amdgpu_device *adev, u32 req) @@ -4709,16 +4706,32 @@ static int gfx_v8_0_kcq_resume(struct amdgpu_device *adev) if (r) goto done; - /* Test KCQs - reversing the order of rings seems to fix ring test failure - * after GPU reset - */ - for (i = adev->gfx.num_compute_rings - 1; i >= 0; i--) { +done: + return r; +} + +static int gfx_v8_0_cp_test_all_rings(struct amdgpu_device *adev) +{ + int r, i; + struct amdgpu_ring *ring; + + /* collect all the ring_tests here, gfx, kiq, compute */ + ring = &adev->gfx.gfx_ring[0]; + r = amdgpu_ring_test_helper(ring); + if (r) + return r; + + ring = &adev->gfx.kiq.ring; + r = amdgpu_ring_test_helper(ring); + if (r) + return r; + + for (i = 0; i < adev->gfx.num_compute_rings; i++) { ring = &adev->gfx.compute_ring[i]; - r = amdgpu_ring_test_helper(ring); + amdgpu_ring_test_helper(ring); } -done: - return r; + return 0; } static int gfx_v8_0_cp_resume(struct amdgpu_device *adev) @@ -4739,6 +4752,11 @@ static int gfx_v8_0_cp_resume(struct amdgpu_device *adev) r = gfx_v8_0_kcq_resume(adev); if (r) return r; + + r = gfx_v8_0_cp_test_all_rings(adev); + if (r) + return r; + gfx_v8_0_enable_gui_idle_interrupt(adev, true); return 0; @@ -5086,6 +5104,8 @@ static int gfx_v8_0_post_soft_reset(void *handle) REG_GET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_GFX)) gfx_v8_0_cp_gfx_resume(adev); + gfx_v8_0_cp_test_all_rings(adev); + adev->gfx.rlc.funcs->start(adev); return 0; @@ -6027,7 +6047,7 @@ static void gfx_v8_0_ring_emit_vgt_flush(struct amdgpu_ring *ring) static void gfx_v8_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 header, control = 0; @@ -6059,11 +6079,27 @@ static void gfx_v8_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, static void gfx_v8_0_ring_emit_ib_compute(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24); + /* Currently, there is a high possibility to get wave ID mismatch + * between ME and GDS, leading to a hw deadlock, because ME generates + * different wave IDs than the GDS expects. This situation happens + * randomly when at least 5 compute pipes use GDS ordered append. + * The wave IDs generated by ME are also wrong after suspend/resume. + * Those are probably bugs somewhere else in the kernel driver. + * + * Writing GDS_COMPUTE_MAX_WAVE_ID resets wave ID counters in ME and + * GDS to 0 for this ring (me/pipe). + */ + if (ib->flags & AMDGPU_IB_FLAG_RESET_GDS_MAX_WAVE_ID) { + amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + amdgpu_ring_write(ring, mmGDS_COMPUTE_MAX_WAVE_ID - PACKET3_SET_CONFIG_REG_START); + amdgpu_ring_write(ring, ring->adev->gds.gds_compute_max_wave_id); + } + amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); amdgpu_ring_write(ring, #ifdef __BIG_ENDIAN @@ -6870,7 +6906,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = { 7 + /* gfx_v8_0_ring_emit_pipeline_sync */ VI_FLUSH_GPU_TLB_NUM_WREG * 5 + 7 + /* gfx_v8_0_ring_emit_vm_flush */ 7 + 7 + 7, /* gfx_v8_0_ring_emit_fence_compute x3 for user fence, vm fence */ - .emit_ib_size = 4, /* gfx_v8_0_ring_emit_ib_compute */ + .emit_ib_size = 7, /* gfx_v8_0_ring_emit_ib_compute */ .emit_ib = gfx_v8_0_ring_emit_ib_compute, .emit_fence = gfx_v8_0_ring_emit_fence_compute, .emit_pipeline_sync = gfx_v8_0_ring_emit_pipeline_sync, @@ -6900,7 +6936,7 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_kiq = { 7 + /* gfx_v8_0_ring_emit_pipeline_sync */ 17 + /* gfx_v8_0_ring_emit_vm_flush */ 7 + 7 + 7, /* gfx_v8_0_ring_emit_fence_kiq x3 for user fence, vm fence */ - .emit_ib_size = 4, /* gfx_v8_0_ring_emit_ib_compute */ + .emit_ib_size = 7, /* gfx_v8_0_ring_emit_ib_compute */ .emit_fence = gfx_v8_0_ring_emit_fence_kiq, .test_ring = gfx_v8_0_ring_test_ring, .insert_nop = amdgpu_ring_insert_nop, @@ -6976,6 +7012,7 @@ static void gfx_v8_0_set_gds_init(struct amdgpu_device *adev) adev->gds.mem.total_size = RREG32(mmGDS_VMID0_SIZE); adev->gds.gws.total_size = 64; adev->gds.oa.total_size = 16; + adev->gds.gds_compute_max_wave_id = RREG32(mmGDS_COMPUTE_MAX_WAVE_ID); if (adev->gds.mem.total_size == 64 * 1024) { adev->gds.mem.gfx_partition_size = 4096; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index 7556716038d3..5533f6e4f4a4 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -113,7 +113,10 @@ static const struct soc15_reg_golden golden_settings_gc_9_0[] = SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_HI, 0xffffffff, 0x4a2c0e68), SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_LO, 0xffffffff, 0xb5d3f197), SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_CACHE_INVALIDATION, 0x3fff3af3, 0x19200000), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_GS_MAX_WAVE_ID, 0x00000fff, 0x000003ff) + SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_GS_MAX_WAVE_ID, 0x00000fff, 0x000003ff), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800), + SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000) }; static const struct soc15_reg_golden golden_settings_gc_9_0_vg10[] = @@ -135,10 +138,7 @@ static const struct soc15_reg_golden golden_settings_gc_9_0_vg10[] = SOC15_REG_GOLDEN_VALUE(GC, 0, mmRMI_UTCL1_CNTL2, 0x00030000, 0x00020000), SOC15_REG_GOLDEN_VALUE(GC, 0, mmSPI_CONFIG_CNTL_1, 0x0000000f, 0x01000107), SOC15_REG_GOLDEN_VALUE(GC, 0, mmTD_CNTL, 0x00001800, 0x00000800), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800), - SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000) + SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080) }; static const struct soc15_reg_golden golden_settings_gc_9_0_vg20[] = @@ -3587,6 +3587,8 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev { uint32_t data, def; + amdgpu_gfx_rlc_enter_safe_mode(adev); + /* It is disabled by HW by default */ if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) { /* 1 - RLC_CGTT_MGCG_OVERRIDE */ @@ -3651,6 +3653,8 @@ static void gfx_v9_0_update_medium_grain_clock_gating(struct amdgpu_device *adev WREG32_SOC15(GC, 0, mmCP_MEM_SLP_CNTL, data); } } + + amdgpu_gfx_rlc_exit_safe_mode(adev); } static void gfx_v9_0_update_3d_clock_gating(struct amdgpu_device *adev, @@ -3968,7 +3972,7 @@ static void gfx_v9_0_ring_emit_hdp_flush(struct amdgpu_ring *ring) static void gfx_v9_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 header, control = 0; @@ -4001,11 +4005,27 @@ static void gfx_v9_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, static void gfx_v9_0_ring_emit_ib_compute(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vmid << 24); + /* Currently, there is a high possibility to get wave ID mismatch + * between ME and GDS, leading to a hw deadlock, because ME generates + * different wave IDs than the GDS expects. This situation happens + * randomly when at least 5 compute pipes use GDS ordered append. + * The wave IDs generated by ME are also wrong after suspend/resume. + * Those are probably bugs somewhere else in the kernel driver. + * + * Writing GDS_COMPUTE_MAX_WAVE_ID resets wave ID counters in ME and + * GDS to 0 for this ring (me/pipe). + */ + if (ib->flags & AMDGPU_IB_FLAG_RESET_GDS_MAX_WAVE_ID) { + amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1)); + amdgpu_ring_write(ring, mmGDS_COMPUTE_MAX_WAVE_ID); + amdgpu_ring_write(ring, ring->adev->gds.gds_compute_max_wave_id); + } + amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); BUG_ON(ib->gpu_addr & 0x3); /* Dword align */ amdgpu_ring_write(ring, @@ -4725,7 +4745,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_compute = { SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 + 2 + /* gfx_v9_0_ring_emit_vm_flush */ 8 + 8 + 8, /* gfx_v9_0_ring_emit_fence x3 for user fence, vm fence */ - .emit_ib_size = 4, /* gfx_v9_0_ring_emit_ib_compute */ + .emit_ib_size = 7, /* gfx_v9_0_ring_emit_ib_compute */ .emit_ib = gfx_v9_0_ring_emit_ib_compute, .emit_fence = gfx_v9_0_ring_emit_fence, .emit_pipeline_sync = gfx_v9_0_ring_emit_pipeline_sync, @@ -4760,7 +4780,7 @@ static const struct amdgpu_ring_funcs gfx_v9_0_ring_funcs_kiq = { SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 7 + 2 + /* gfx_v9_0_ring_emit_vm_flush */ 8 + 8 + 8, /* gfx_v9_0_ring_emit_fence_kiq x3 for user fence, vm fence */ - .emit_ib_size = 4, /* gfx_v9_0_ring_emit_ib_compute */ + .emit_ib_size = 7, /* gfx_v9_0_ring_emit_ib_compute */ .emit_fence = gfx_v9_0_ring_emit_fence_kiq, .test_ring = gfx_v9_0_ring_test_ring, .insert_nop = amdgpu_ring_insert_nop, @@ -4842,6 +4862,26 @@ static void gfx_v9_0_set_gds_init(struct amdgpu_device *adev) break; } + switch (adev->asic_type) { + case CHIP_VEGA10: + case CHIP_VEGA20: + adev->gds.gds_compute_max_wave_id = 0x7ff; + break; + case CHIP_VEGA12: + adev->gds.gds_compute_max_wave_id = 0x27f; + break; + case CHIP_RAVEN: + if (adev->rev_id >= 0x8) + adev->gds.gds_compute_max_wave_id = 0x77; /* raven2 */ + else + adev->gds.gds_compute_max_wave_id = 0x15f; /* raven1 */ + break; + default: + /* this really depends on the chip */ + adev->gds.gds_compute_max_wave_id = 0x7ff; + break; + } + adev->gds.gws.total_size = 64; adev->gds.oa.total_size = 16; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index 1ad7e6b8ed1d..34440672f938 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -1471,8 +1471,9 @@ static int gmc_v8_0_process_interrupt(struct amdgpu_device *adev, gmc_v8_0_set_fault_enable_default(adev, false); if (printk_ratelimit()) { - struct amdgpu_task_info task_info = { 0 }; + struct amdgpu_task_info task_info; + memset(&task_info, 0, sizeof(struct amdgpu_task_info)); amdgpu_vm_get_task_info(adev, entry->pasid, &task_info); dev_err(adev->dev, "GPU fault detected: %d 0x%08x for process %s pid %d thread %s pid %d\n", diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index bacdaef77b6c..600259b4e291 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -305,6 +305,7 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, struct amdgpu_iv_entry *entry) { struct amdgpu_vmhub *hub = &adev->vmhub[entry->vmid_src]; + bool retry_fault = !!(entry->src_data[1] & 0x80); uint32_t status = 0; u64 addr; @@ -320,13 +321,16 @@ static int gmc_v9_0_process_interrupt(struct amdgpu_device *adev, } if (printk_ratelimit()) { - struct amdgpu_task_info task_info = { 0 }; + struct amdgpu_task_info task_info; + memset(&task_info, 0, sizeof(struct amdgpu_task_info)); amdgpu_vm_get_task_info(adev, entry->pasid, &task_info); dev_err(adev->dev, - "[%s] VMC page fault (src_id:%u ring:%u vmid:%u pasid:%u, for process %s pid %d thread %s pid %d)\n", + "[%s] %s page fault (src_id:%u ring:%u vmid:%u " + "pasid:%u, for process %s pid %d thread %s pid %d)\n", entry->vmid_src ? "mmhub" : "gfxhub", + retry_fault ? "retry" : "no-retry", entry->src_id, entry->ring_id, entry->vmid, entry->pasid, task_info.process_name, task_info.tgid, task_info.task_name, task_info.pid); @@ -961,7 +965,11 @@ static int gmc_v9_0_sw_init(void *handle) * vm size is 256TB (48bit), maximum size of Vega10, * block size 512 (9bit) */ - amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48); + /* sriov restrict max_pfn below AMDGPU_GMC_HOLE */ + if (amdgpu_sriov_vf(adev)) + amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 47); + else + amdgpu_vm_adjust_size(adev, 256 * 1024, 9, 3, 48); break; default: break; diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c index a3984d10b604..b1626e1d2f5d 100644 --- a/drivers/gpu/drm/amd/amdgpu/iceland_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/iceland_ih.c @@ -103,9 +103,9 @@ static void iceland_ih_disable_interrupts(struct amdgpu_device *adev) */ static int iceland_ih_irq_init(struct amdgpu_device *adev) { + struct amdgpu_ih_ring *ih = &adev->irq.ih; int rb_bufsz; u32 interrupt_cntl, ih_cntl, ih_rb_cntl; - u64 wptr_off; /* disable irqs */ iceland_ih_disable_interrupts(adev); @@ -133,9 +133,8 @@ static int iceland_ih_irq_init(struct amdgpu_device *adev) ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, WPTR_WRITEBACK_ENABLE, 1); /* set the writeback address whether it's enabled or not */ - wptr_off = adev->wb.gpu_addr + (adev->irq.ih.wptr_offs * 4); - WREG32(mmIH_RB_WPTR_ADDR_LO, lower_32_bits(wptr_off)); - WREG32(mmIH_RB_WPTR_ADDR_HI, upper_32_bits(wptr_off) & 0xFF); + WREG32(mmIH_RB_WPTR_ADDR_LO, lower_32_bits(ih->wptr_addr)); + WREG32(mmIH_RB_WPTR_ADDR_HI, upper_32_bits(ih->wptr_addr) & 0xFF); WREG32(mmIH_RB_CNTL, ih_rb_cntl); @@ -185,11 +184,12 @@ static void iceland_ih_irq_disable(struct amdgpu_device *adev) * Used by cz_irq_process(VI). * Returns the value of the wptr. */ -static u32 iceland_ih_get_wptr(struct amdgpu_device *adev) +static u32 iceland_ih_get_wptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { u32 wptr, tmp; - wptr = le32_to_cpu(adev->wb.wb[adev->irq.ih.wptr_offs]); + wptr = le32_to_cpu(*ih->wptr_cpu); if (REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) { wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0); @@ -198,13 +198,13 @@ static u32 iceland_ih_get_wptr(struct amdgpu_device *adev) * this should allow us to catchup. */ dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n", - wptr, adev->irq.ih.rptr, (wptr + 16) & adev->irq.ih.ptr_mask); - adev->irq.ih.rptr = (wptr + 16) & adev->irq.ih.ptr_mask; + wptr, ih->rptr, (wptr + 16) & ih->ptr_mask); + ih->rptr = (wptr + 16) & ih->ptr_mask; tmp = RREG32(mmIH_RB_CNTL); tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); WREG32(mmIH_RB_CNTL, tmp); } - return (wptr & adev->irq.ih.ptr_mask); + return (wptr & ih->ptr_mask); } /** @@ -216,16 +216,17 @@ static u32 iceland_ih_get_wptr(struct amdgpu_device *adev) * position and also advance the position. */ static void iceland_ih_decode_iv(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih, struct amdgpu_iv_entry *entry) { /* wptr/rptr are in bytes! */ - u32 ring_index = adev->irq.ih.rptr >> 2; + u32 ring_index = ih->rptr >> 2; uint32_t dw[4]; - dw[0] = le32_to_cpu(adev->irq.ih.ring[ring_index + 0]); - dw[1] = le32_to_cpu(adev->irq.ih.ring[ring_index + 1]); - dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]); - dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); + dw[0] = le32_to_cpu(ih->ring[ring_index + 0]); + dw[1] = le32_to_cpu(ih->ring[ring_index + 1]); + dw[2] = le32_to_cpu(ih->ring[ring_index + 2]); + dw[3] = le32_to_cpu(ih->ring[ring_index + 3]); entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY; entry->src_id = dw[0] & 0xff; @@ -235,7 +236,7 @@ static void iceland_ih_decode_iv(struct amdgpu_device *adev, entry->pasid = (dw[2] >> 16) & 0xffff; /* wptr/rptr are in bytes! */ - adev->irq.ih.rptr += 16; + ih->rptr += 16; } /** @@ -245,9 +246,10 @@ static void iceland_ih_decode_iv(struct amdgpu_device *adev, * * Set the IH ring buffer rptr. */ -static void iceland_ih_set_rptr(struct amdgpu_device *adev) +static void iceland_ih_set_rptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { - WREG32(mmIH_RB_RPTR, adev->irq.ih.rptr); + WREG32(mmIH_RB_RPTR, ih->rptr); } static int iceland_ih_early_init(void *handle) diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c index 8cbb4655896a..b11a1c17a7f2 100644 --- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c +++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c @@ -174,7 +174,7 @@ static int xgpu_ai_send_access_requests(struct amdgpu_device *adev, return r; } /* Retrieve checksum from mailbox2 */ - if (req == IDH_REQ_GPU_INIT_ACCESS) { + if (req == IDH_REQ_GPU_INIT_ACCESS || req == IDH_REQ_GPU_RESET_ACCESS) { adev->virt.fw_reserve.checksum_key = RREG32_NO_KIQ(SOC15_REG_OFFSET(NBIO, 0, mmBIF_BX_PF0_MAILBOX_MSGBUF_RCV_DW2)); diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c index accdedd63c98..cc967dbfd631 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c @@ -27,13 +27,9 @@ #include "nbio/nbio_6_1_default.h" #include "nbio/nbio_6_1_offset.h" #include "nbio/nbio_6_1_sh_mask.h" +#include "nbio/nbio_6_1_smn.h" #include "vega10_enum.h" -#define smnCPM_CONTROL 0x11180460 -#define smnPCIE_CNTL2 0x11180070 -#define smnPCIE_CONFIG_CNTL 0x11180044 -#define smnPCIE_CI_CNTL 0x11180080 - static u32 nbio_v6_1_get_rev_id(struct amdgpu_device *adev) { u32 tmp = RREG32_SOC15(NBIO, 0, mmRCC_DEV0_EPF0_STRAP0); @@ -72,7 +68,7 @@ static u32 nbio_v6_1_get_memsize(struct amdgpu_device *adev) } static void nbio_v6_1_sdma_doorbell_range(struct amdgpu_device *adev, int instance, - bool use_doorbell, int doorbell_index) + bool use_doorbell, int doorbell_index, int doorbell_size) { u32 reg = instance == 0 ? SOC15_REG_OFFSET(NBIO, 0, mmBIF_SDMA0_DOORBELL_RANGE) : SOC15_REG_OFFSET(NBIO, 0, mmBIF_SDMA1_DOORBELL_RANGE); @@ -81,7 +77,7 @@ static void nbio_v6_1_sdma_doorbell_range(struct amdgpu_device *adev, int instan if (use_doorbell) { doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, OFFSET, doorbell_index); - doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, SIZE, 2); + doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, SIZE, doorbell_size); } else doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, SIZE, 0); diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c index df34dc79d444..1cdb98ad2db3 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_0.c @@ -27,13 +27,11 @@ #include "nbio/nbio_7_0_default.h" #include "nbio/nbio_7_0_offset.h" #include "nbio/nbio_7_0_sh_mask.h" +#include "nbio/nbio_7_0_smn.h" #include "vega10_enum.h" #define smnNBIF_MGCG_CTRL_LCLK 0x1013a05c -#define smnCPM_CONTROL 0x11180460 -#define smnPCIE_CNTL2 0x11180070 - static u32 nbio_v7_0_get_rev_id(struct amdgpu_device *adev) { u32 tmp = RREG32_SOC15(NBIO, 0, mmRCC_DEV0_EPF0_STRAP0); @@ -69,7 +67,7 @@ static u32 nbio_v7_0_get_memsize(struct amdgpu_device *adev) } static void nbio_v7_0_sdma_doorbell_range(struct amdgpu_device *adev, int instance, - bool use_doorbell, int doorbell_index) + bool use_doorbell, int doorbell_index, int doorbell_size) { u32 reg = instance == 0 ? SOC15_REG_OFFSET(NBIO, 0, mmBIF_SDMA0_DOORBELL_RANGE) : SOC15_REG_OFFSET(NBIO, 0, mmBIF_SDMA1_DOORBELL_RANGE); @@ -78,7 +76,7 @@ static void nbio_v7_0_sdma_doorbell_range(struct amdgpu_device *adev, int instan if (use_doorbell) { doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, OFFSET, doorbell_index); - doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, SIZE, 2); + doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, SIZE, doorbell_size); } else doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, SIZE, 0); diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c index 4cd31a276dcd..e347b407bd03 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_4.c @@ -26,13 +26,10 @@ #include "nbio/nbio_7_4_offset.h" #include "nbio/nbio_7_4_sh_mask.h" +#include "nbio/nbio_7_4_0_smn.h" #define smnNBIF_MGCG_CTRL_LCLK 0x1013a21c -#define smnCPM_CONTROL 0x11180460 -#define smnPCIE_CNTL2 0x11180070 -#define smnPCIE_CI_CNTL 0x11180080 - static u32 nbio_v7_4_get_rev_id(struct amdgpu_device *adev) { u32 tmp = RREG32_SOC15(NBIO, 0, mmRCC_DEV0_EPF0_STRAP0); @@ -68,7 +65,7 @@ static u32 nbio_v7_4_get_memsize(struct amdgpu_device *adev) } static void nbio_v7_4_sdma_doorbell_range(struct amdgpu_device *adev, int instance, - bool use_doorbell, int doorbell_index) + bool use_doorbell, int doorbell_index, int doorbell_size) { u32 reg = instance == 0 ? SOC15_REG_OFFSET(NBIO, 0, mmBIF_SDMA0_DOORBELL_RANGE) : SOC15_REG_OFFSET(NBIO, 0, mmBIF_SDMA1_DOORBELL_RANGE); @@ -77,7 +74,7 @@ static void nbio_v7_4_sdma_doorbell_range(struct amdgpu_device *adev, int instan if (use_doorbell) { doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, OFFSET, doorbell_index); - doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, SIZE, 2); + doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, SIZE, doorbell_size); } else doorbell_range = REG_SET_FIELD(doorbell_range, BIF_SDMA0_DOORBELL_RANGE, SIZE, 0); diff --git a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h index 0de00fbe9233..f3a7d207af07 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h +++ b/drivers/gpu/drm/amd/amdgpu/psp_gfx_if.h @@ -191,7 +191,7 @@ enum psp_gfx_fw_type GFX_FW_TYPE_MMSCH = 19, GFX_FW_TYPE_RLC_RESTORE_LIST_GPM_MEM = 20, GFX_FW_TYPE_RLC_RESTORE_LIST_SRM_MEM = 21, - GFX_FW_TYPE_RLC_RESTORE_LIST_CNTL = 22, + GFX_FW_TYPE_RLC_RESTORE_LIST_SRM_CNTL = 22, GFX_FW_TYPE_UVD1 = 23, GFX_FW_TYPE_MAX = 24 }; diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c index d78b4306a36f..77c2bc344dfc 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v10_0.c @@ -38,75 +38,6 @@ MODULE_FIRMWARE("amdgpu/raven_asd.bin"); MODULE_FIRMWARE("amdgpu/picasso_asd.bin"); MODULE_FIRMWARE("amdgpu/raven2_asd.bin"); -static int -psp_v10_0_get_fw_type(struct amdgpu_firmware_info *ucode, enum psp_gfx_fw_type *type) -{ - switch(ucode->ucode_id) { - case AMDGPU_UCODE_ID_SDMA0: - *type = GFX_FW_TYPE_SDMA0; - break; - case AMDGPU_UCODE_ID_SDMA1: - *type = GFX_FW_TYPE_SDMA1; - break; - case AMDGPU_UCODE_ID_CP_CE: - *type = GFX_FW_TYPE_CP_CE; - break; - case AMDGPU_UCODE_ID_CP_PFP: - *type = GFX_FW_TYPE_CP_PFP; - break; - case AMDGPU_UCODE_ID_CP_ME: - *type = GFX_FW_TYPE_CP_ME; - break; - case AMDGPU_UCODE_ID_CP_MEC1: - *type = GFX_FW_TYPE_CP_MEC; - break; - case AMDGPU_UCODE_ID_CP_MEC1_JT: - *type = GFX_FW_TYPE_CP_MEC_ME1; - break; - case AMDGPU_UCODE_ID_CP_MEC2: - *type = GFX_FW_TYPE_CP_MEC; - break; - case AMDGPU_UCODE_ID_CP_MEC2_JT: - *type = GFX_FW_TYPE_CP_MEC_ME2; - break; - case AMDGPU_UCODE_ID_RLC_G: - *type = GFX_FW_TYPE_RLC_G; - break; - case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL: - *type = GFX_FW_TYPE_RLC_RESTORE_LIST_CNTL; - break; - case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM: - *type = GFX_FW_TYPE_RLC_RESTORE_LIST_GPM_MEM; - break; - case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM: - *type = GFX_FW_TYPE_RLC_RESTORE_LIST_SRM_MEM; - break; - case AMDGPU_UCODE_ID_SMC: - *type = GFX_FW_TYPE_SMU; - break; - case AMDGPU_UCODE_ID_UVD: - *type = GFX_FW_TYPE_UVD; - break; - case AMDGPU_UCODE_ID_VCE: - *type = GFX_FW_TYPE_VCE; - break; - case AMDGPU_UCODE_ID_VCN: - *type = GFX_FW_TYPE_VCN; - break; - case AMDGPU_UCODE_ID_DMCU_ERAM: - *type = GFX_FW_TYPE_DMCU_ERAM; - break; - case AMDGPU_UCODE_ID_DMCU_INTV: - *type = GFX_FW_TYPE_DMCU_ISR; - break; - case AMDGPU_UCODE_ID_MAXIMUM: - default: - return -EINVAL; - } - - return 0; -} - static int psp_v10_0_init_microcode(struct psp_context *psp) { struct amdgpu_device *adev = psp->adev; @@ -158,26 +89,6 @@ out: return err; } -static int psp_v10_0_prep_cmd_buf(struct amdgpu_firmware_info *ucode, - struct psp_gfx_cmd_resp *cmd) -{ - int ret; - uint64_t fw_mem_mc_addr = ucode->mc_addr; - - memset(cmd, 0, sizeof(struct psp_gfx_cmd_resp)); - - cmd->cmd_id = GFX_CMD_ID_LOAD_IP_FW; - cmd->cmd.cmd_load_ip_fw.fw_phy_addr_lo = lower_32_bits(fw_mem_mc_addr); - cmd->cmd.cmd_load_ip_fw.fw_phy_addr_hi = upper_32_bits(fw_mem_mc_addr); - cmd->cmd.cmd_load_ip_fw.fw_size = ucode->ucode_size; - - ret = psp_v10_0_get_fw_type(ucode, &cmd->cmd.cmd_load_ip_fw.fw_type); - if (ret) - DRM_ERROR("Unknown firmware type\n"); - - return ret; -} - static int psp_v10_0_ring_init(struct psp_context *psp, enum psp_ring_type ring_type) { @@ -454,7 +365,6 @@ static int psp_v10_0_mode1_reset(struct psp_context *psp) static const struct psp_funcs psp_v10_0_funcs = { .init_microcode = psp_v10_0_init_microcode, - .prep_cmd_buf = psp_v10_0_prep_cmd_buf, .ring_init = psp_v10_0_ring_init, .ring_create = psp_v10_0_ring_create, .ring_stop = psp_v10_0_ring_stop, diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c index 0c6e7f9b143f..f71384be1f97 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v11_0.c @@ -40,60 +40,6 @@ MODULE_FIRMWARE("amdgpu/vega20_ta.bin"); /* address block */ #define smnMP1_FIRMWARE_FLAGS 0x3010024 -static int -psp_v11_0_get_fw_type(struct amdgpu_firmware_info *ucode, enum psp_gfx_fw_type *type) -{ - switch (ucode->ucode_id) { - case AMDGPU_UCODE_ID_SDMA0: - *type = GFX_FW_TYPE_SDMA0; - break; - case AMDGPU_UCODE_ID_SDMA1: - *type = GFX_FW_TYPE_SDMA1; - break; - case AMDGPU_UCODE_ID_CP_CE: - *type = GFX_FW_TYPE_CP_CE; - break; - case AMDGPU_UCODE_ID_CP_PFP: - *type = GFX_FW_TYPE_CP_PFP; - break; - case AMDGPU_UCODE_ID_CP_ME: - *type = GFX_FW_TYPE_CP_ME; - break; - case AMDGPU_UCODE_ID_CP_MEC1: - *type = GFX_FW_TYPE_CP_MEC; - break; - case AMDGPU_UCODE_ID_CP_MEC1_JT: - *type = GFX_FW_TYPE_CP_MEC_ME1; - break; - case AMDGPU_UCODE_ID_CP_MEC2: - *type = GFX_FW_TYPE_CP_MEC; - break; - case AMDGPU_UCODE_ID_CP_MEC2_JT: - *type = GFX_FW_TYPE_CP_MEC_ME2; - break; - case AMDGPU_UCODE_ID_RLC_G: - *type = GFX_FW_TYPE_RLC_G; - break; - case AMDGPU_UCODE_ID_SMC: - *type = GFX_FW_TYPE_SMU; - break; - case AMDGPU_UCODE_ID_UVD: - *type = GFX_FW_TYPE_UVD; - break; - case AMDGPU_UCODE_ID_VCE: - *type = GFX_FW_TYPE_VCE; - break; - case AMDGPU_UCODE_ID_UVD1: - *type = GFX_FW_TYPE_UVD1; - break; - case AMDGPU_UCODE_ID_MAXIMUM: - default: - return -EINVAL; - } - - return 0; -} - static int psp_v11_0_init_microcode(struct psp_context *psp) { struct amdgpu_device *adev = psp->adev; @@ -267,26 +213,6 @@ static int psp_v11_0_bootloader_load_sos(struct psp_context *psp) return ret; } -static int psp_v11_0_prep_cmd_buf(struct amdgpu_firmware_info *ucode, - struct psp_gfx_cmd_resp *cmd) -{ - int ret; - uint64_t fw_mem_mc_addr = ucode->mc_addr; - - memset(cmd, 0, sizeof(struct psp_gfx_cmd_resp)); - - cmd->cmd_id = GFX_CMD_ID_LOAD_IP_FW; - cmd->cmd.cmd_load_ip_fw.fw_phy_addr_lo = lower_32_bits(fw_mem_mc_addr); - cmd->cmd.cmd_load_ip_fw.fw_phy_addr_hi = upper_32_bits(fw_mem_mc_addr); - cmd->cmd.cmd_load_ip_fw.fw_size = ucode->ucode_size; - - ret = psp_v11_0_get_fw_type(ucode, &cmd->cmd.cmd_load_ip_fw.fw_type); - if (ret) - DRM_ERROR("Unknown firmware type\n"); - - return ret; -} - static int psp_v11_0_ring_init(struct psp_context *psp, enum psp_ring_type ring_type) { @@ -753,7 +679,6 @@ static const struct psp_funcs psp_v11_0_funcs = { .init_microcode = psp_v11_0_init_microcode, .bootloader_load_sysdrv = psp_v11_0_bootloader_load_sysdrv, .bootloader_load_sos = psp_v11_0_bootloader_load_sos, - .prep_cmd_buf = psp_v11_0_prep_cmd_buf, .ring_init = psp_v11_0_ring_init, .ring_create = psp_v11_0_ring_create, .ring_stop = psp_v11_0_ring_stop, diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c index 79694ff16969..c63de945c021 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v3_1.c @@ -47,57 +47,6 @@ MODULE_FIRMWARE("amdgpu/vega12_asd.bin"); static uint32_t sos_old_versions[] = {1517616, 1510592, 1448594, 1446554}; -static int -psp_v3_1_get_fw_type(struct amdgpu_firmware_info *ucode, enum psp_gfx_fw_type *type) -{ - switch(ucode->ucode_id) { - case AMDGPU_UCODE_ID_SDMA0: - *type = GFX_FW_TYPE_SDMA0; - break; - case AMDGPU_UCODE_ID_SDMA1: - *type = GFX_FW_TYPE_SDMA1; - break; - case AMDGPU_UCODE_ID_CP_CE: - *type = GFX_FW_TYPE_CP_CE; - break; - case AMDGPU_UCODE_ID_CP_PFP: - *type = GFX_FW_TYPE_CP_PFP; - break; - case AMDGPU_UCODE_ID_CP_ME: - *type = GFX_FW_TYPE_CP_ME; - break; - case AMDGPU_UCODE_ID_CP_MEC1: - *type = GFX_FW_TYPE_CP_MEC; - break; - case AMDGPU_UCODE_ID_CP_MEC1_JT: - *type = GFX_FW_TYPE_CP_MEC_ME1; - break; - case AMDGPU_UCODE_ID_CP_MEC2: - *type = GFX_FW_TYPE_CP_MEC; - break; - case AMDGPU_UCODE_ID_CP_MEC2_JT: - *type = GFX_FW_TYPE_CP_MEC_ME2; - break; - case AMDGPU_UCODE_ID_RLC_G: - *type = GFX_FW_TYPE_RLC_G; - break; - case AMDGPU_UCODE_ID_SMC: - *type = GFX_FW_TYPE_SMU; - break; - case AMDGPU_UCODE_ID_UVD: - *type = GFX_FW_TYPE_UVD; - break; - case AMDGPU_UCODE_ID_VCE: - *type = GFX_FW_TYPE_VCE; - break; - case AMDGPU_UCODE_ID_MAXIMUM: - default: - return -EINVAL; - } - - return 0; -} - static int psp_v3_1_init_microcode(struct psp_context *psp) { struct amdgpu_device *adev = psp->adev; @@ -277,26 +226,6 @@ static int psp_v3_1_bootloader_load_sos(struct psp_context *psp) return ret; } -static int psp_v3_1_prep_cmd_buf(struct amdgpu_firmware_info *ucode, - struct psp_gfx_cmd_resp *cmd) -{ - int ret; - uint64_t fw_mem_mc_addr = ucode->mc_addr; - - memset(cmd, 0, sizeof(struct psp_gfx_cmd_resp)); - - cmd->cmd_id = GFX_CMD_ID_LOAD_IP_FW; - cmd->cmd.cmd_load_ip_fw.fw_phy_addr_lo = lower_32_bits(fw_mem_mc_addr); - cmd->cmd.cmd_load_ip_fw.fw_phy_addr_hi = upper_32_bits(fw_mem_mc_addr); - cmd->cmd.cmd_load_ip_fw.fw_size = ucode->ucode_size; - - ret = psp_v3_1_get_fw_type(ucode, &cmd->cmd.cmd_load_ip_fw.fw_type); - if (ret) - DRM_ERROR("Unknown firmware type\n"); - - return ret; -} - static int psp_v3_1_ring_init(struct psp_context *psp, enum psp_ring_type ring_type) { @@ -615,7 +544,6 @@ static const struct psp_funcs psp_v3_1_funcs = { .init_microcode = psp_v3_1_init_microcode, .bootloader_load_sysdrv = psp_v3_1_bootloader_load_sysdrv, .bootloader_load_sos = psp_v3_1_bootloader_load_sos, - .prep_cmd_buf = psp_v3_1_prep_cmd_buf, .ring_init = psp_v3_1_ring_init, .ring_create = psp_v3_1_ring_create, .ring_stop = psp_v3_1_ring_stop, diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c index 9f3cb2aec7c2..cca3552b36ed 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c @@ -247,7 +247,7 @@ static void sdma_v2_4_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count) static void sdma_v2_4_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c index 1bccc5fe2d9d..0ce8331baeb2 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c @@ -421,7 +421,7 @@ static void sdma_v3_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count) static void sdma_v3_0_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); @@ -1145,8 +1145,7 @@ static int sdma_v3_0_sw_init(void *handle) ring->ring_obj = NULL; if (!amdgpu_sriov_vf(adev)) { ring->use_doorbell = true; - ring->doorbell_index = (i == 0) ? - adev->doorbell_index.sdma_engine0 : adev->doorbell_index.sdma_engine1; + ring->doorbell_index = adev->doorbell_index.sdma_engine[i]; } else { ring->use_pollmem = true; } diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c index fd0bfe140ee0..127b85983e8f 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c @@ -78,7 +78,6 @@ static const struct soc15_reg_golden golden_settings_sdma_4[] = { SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_RLC1_RB_WPTR_POLL_CNTL, 0x0000fff0, 0x00403000), SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_UTCL1_PAGE, 0x000003ff, 0x000003c0), SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_UTCL1_WATERMK, 0xfc000000, 0x00000000), - SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CHICKEN_BITS, 0xfe931f07, 0x02831f07), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CLK_CTRL, 0xffffffff, 0x3f000100), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GFX_IB_CNTL, 0x800f0100, 0x00000100), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GFX_RB_WPTR_POLL_CNTL, 0x0000fff0, 0x00403000), @@ -96,6 +95,7 @@ static const struct soc15_reg_golden golden_settings_sdma_4[] = { static const struct soc15_reg_golden golden_settings_sdma_vg10[] = { SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG, 0x0018773f, 0x00104002), SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104002), + SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CHICKEN_BITS, 0xfe931f07, 0x02831d07), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG, 0x0018773f, 0x00104002), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104002) }; @@ -103,6 +103,7 @@ static const struct soc15_reg_golden golden_settings_sdma_vg10[] = { static const struct soc15_reg_golden golden_settings_sdma_vg12[] = { SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG, 0x0018773f, 0x00104001), SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104001), + SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_CHICKEN_BITS, 0xfe931f07, 0x02831d07), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG, 0x0018773f, 0x00104001), SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_GB_ADDR_CONFIG_READ, 0x0018773f, 0x00104001) }; @@ -499,7 +500,7 @@ static void sdma_v4_0_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count) static void sdma_v4_0_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); @@ -833,8 +834,6 @@ static void sdma_v4_0_gfx_resume(struct amdgpu_device *adev, unsigned int i) OFFSET, ring->doorbell_index); WREG32_SDMA(i, mmSDMA0_GFX_DOORBELL, doorbell); WREG32_SDMA(i, mmSDMA0_GFX_DOORBELL_OFFSET, doorbell_offset); - adev->nbio_funcs->sdma_doorbell_range(adev, i, ring->use_doorbell, - ring->doorbell_index); sdma_v4_0_ring_set_wptr(ring); @@ -1521,9 +1520,7 @@ static int sdma_v4_0_sw_init(void *handle) ring->use_doorbell?"true":"false"); /* doorbell size is 2 dwords, get DWORD offset */ - ring->doorbell_index = (i == 0) ? - (adev->doorbell_index.sdma_engine0 << 1) - : (adev->doorbell_index.sdma_engine1 << 1); + ring->doorbell_index = adev->doorbell_index.sdma_engine[i] << 1; sprintf(ring->name, "sdma%d", i); r = amdgpu_ring_init(adev, ring, 1024, @@ -1542,9 +1539,7 @@ static int sdma_v4_0_sw_init(void *handle) /* paging queue use same doorbell index/routing as gfx queue * with 0x400 (4096 dwords) offset on second doorbell page */ - ring->doorbell_index = (i == 0) ? - (adev->doorbell_index.sdma_engine0 << 1) - : (adev->doorbell_index.sdma_engine1 << 1); + ring->doorbell_index = adev->doorbell_index.sdma_engine[i] << 1; ring->doorbell_index += 0x400; sprintf(ring->name, "page%d", i); diff --git a/drivers/gpu/drm/amd/amdgpu/si.c b/drivers/gpu/drm/amd/amdgpu/si.c index f8408f88cd37..79c1a9bbcc21 100644 --- a/drivers/gpu/drm/amd/amdgpu/si.c +++ b/drivers/gpu/drm/amd/amdgpu/si.c @@ -47,6 +47,7 @@ #include "dce/dce_6_0_d.h" #include "uvd/uvd_4_0_d.h" #include "bif/bif_3_0_d.h" +#include "bif/bif_3_0_sh_mask.h" static const u32 tahiti_golden_registers[] = { @@ -1258,6 +1259,11 @@ static bool si_need_full_reset(struct amdgpu_device *adev) return true; } +static bool si_need_reset_on_init(struct amdgpu_device *adev) +{ + return false; +} + static int si_get_pcie_lanes(struct amdgpu_device *adev) { u32 link_width_cntl; @@ -1323,6 +1329,52 @@ static void si_set_pcie_lanes(struct amdgpu_device *adev, int lanes) WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl); } +static void si_get_pcie_usage(struct amdgpu_device *adev, uint64_t *count0, + uint64_t *count1) +{ + uint32_t perfctr = 0; + uint64_t cnt0_of, cnt1_of; + int tmp; + + /* This reports 0 on APUs, so return to avoid writing/reading registers + * that may or may not be different from their GPU counterparts + */ + if (adev->flags & AMD_IS_APU) + return; + + /* Set the 2 events that we wish to watch, defined above */ + /* Reg 40 is # received msgs, Reg 104 is # of posted requests sent */ + perfctr = REG_SET_FIELD(perfctr, PCIE_PERF_CNTL_TXCLK, EVENT0_SEL, 40); + perfctr = REG_SET_FIELD(perfctr, PCIE_PERF_CNTL_TXCLK, EVENT1_SEL, 104); + + /* Write to enable desired perf counters */ + WREG32_PCIE(ixPCIE_PERF_CNTL_TXCLK, perfctr); + /* Zero out and enable the perf counters + * Write 0x5: + * Bit 0 = Start all counters(1) + * Bit 2 = Global counter reset enable(1) + */ + WREG32_PCIE(ixPCIE_PERF_COUNT_CNTL, 0x00000005); + + msleep(1000); + + /* Load the shadow and disable the perf counters + * Write 0x2: + * Bit 0 = Stop counters(0) + * Bit 1 = Load the shadow counters(1) + */ + WREG32_PCIE(ixPCIE_PERF_COUNT_CNTL, 0x00000002); + + /* Read register values to get any >32bit overflow */ + tmp = RREG32_PCIE(ixPCIE_PERF_CNTL_TXCLK); + cnt0_of = REG_GET_FIELD(tmp, PCIE_PERF_CNTL_TXCLK, COUNTER0_UPPER); + cnt1_of = REG_GET_FIELD(tmp, PCIE_PERF_CNTL_TXCLK, COUNTER1_UPPER); + + /* Get the values and add the overflow */ + *count0 = RREG32_PCIE(ixPCIE_PERF_COUNT0_TXCLK) | (cnt0_of << 32); + *count1 = RREG32_PCIE(ixPCIE_PERF_COUNT1_TXCLK) | (cnt1_of << 32); +} + static const struct amdgpu_asic_funcs si_asic_funcs = { .read_disabled_bios = &si_read_disabled_bios, @@ -1339,6 +1391,8 @@ static const struct amdgpu_asic_funcs si_asic_funcs = .flush_hdp = &si_flush_hdp, .invalidate_hdp = &si_invalidate_hdp, .need_full_reset = &si_need_full_reset, + .get_pcie_usage = &si_get_pcie_usage, + .need_reset_on_init = &si_need_reset_on_init, }; static uint32_t si_get_rev_id(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/si_dma.c b/drivers/gpu/drm/amd/amdgpu/si_dma.c index b6e473134e19..f15f196684ba 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_dma.c +++ b/drivers/gpu/drm/amd/amdgpu/si_dma.c @@ -63,7 +63,7 @@ static void si_dma_ring_set_wptr(struct amdgpu_ring *ring) static void si_dma_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); /* The indirect buffer packet must end on an 8 DW boundary in the DMA ring. diff --git a/drivers/gpu/drm/amd/amdgpu/si_ih.c b/drivers/gpu/drm/amd/amdgpu/si_ih.c index 2938fb9f17cc..8c50c9cab455 100644 --- a/drivers/gpu/drm/amd/amdgpu/si_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/si_ih.c @@ -57,9 +57,9 @@ static void si_ih_disable_interrupts(struct amdgpu_device *adev) static int si_ih_irq_init(struct amdgpu_device *adev) { + struct amdgpu_ih_ring *ih = &adev->irq.ih; int rb_bufsz; u32 interrupt_cntl, ih_cntl, ih_rb_cntl; - u64 wptr_off; si_ih_disable_interrupts(adev); WREG32(INTERRUPT_CNTL2, adev->irq.ih.gpu_addr >> 8); @@ -76,9 +76,8 @@ static int si_ih_irq_init(struct amdgpu_device *adev) (rb_bufsz << 1) | IH_WPTR_WRITEBACK_ENABLE; - wptr_off = adev->wb.gpu_addr + (adev->irq.ih.wptr_offs * 4); - WREG32(IH_RB_WPTR_ADDR_LO, lower_32_bits(wptr_off)); - WREG32(IH_RB_WPTR_ADDR_HI, upper_32_bits(wptr_off) & 0xFF); + WREG32(IH_RB_WPTR_ADDR_LO, lower_32_bits(ih->wptr_addr)); + WREG32(IH_RB_WPTR_ADDR_HI, upper_32_bits(ih->wptr_addr) & 0xFF); WREG32(IH_RB_CNTL, ih_rb_cntl); WREG32(IH_RB_RPTR, 0); WREG32(IH_RB_WPTR, 0); @@ -100,34 +99,36 @@ static void si_ih_irq_disable(struct amdgpu_device *adev) mdelay(1); } -static u32 si_ih_get_wptr(struct amdgpu_device *adev) +static u32 si_ih_get_wptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { u32 wptr, tmp; - wptr = le32_to_cpu(adev->wb.wb[adev->irq.ih.wptr_offs]); + wptr = le32_to_cpu(*ih->wptr_cpu); if (wptr & IH_RB_WPTR__RB_OVERFLOW_MASK) { wptr &= ~IH_RB_WPTR__RB_OVERFLOW_MASK; dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n", - wptr, adev->irq.ih.rptr, (wptr + 16) & adev->irq.ih.ptr_mask); - adev->irq.ih.rptr = (wptr + 16) & adev->irq.ih.ptr_mask; + wptr, ih->rptr, (wptr + 16) & ih->ptr_mask); + ih->rptr = (wptr + 16) & ih->ptr_mask; tmp = RREG32(IH_RB_CNTL); tmp |= IH_RB_CNTL__WPTR_OVERFLOW_CLEAR_MASK; WREG32(IH_RB_CNTL, tmp); } - return (wptr & adev->irq.ih.ptr_mask); + return (wptr & ih->ptr_mask); } static void si_ih_decode_iv(struct amdgpu_device *adev, - struct amdgpu_iv_entry *entry) + struct amdgpu_ih_ring *ih, + struct amdgpu_iv_entry *entry) { - u32 ring_index = adev->irq.ih.rptr >> 2; + u32 ring_index = ih->rptr >> 2; uint32_t dw[4]; - dw[0] = le32_to_cpu(adev->irq.ih.ring[ring_index + 0]); - dw[1] = le32_to_cpu(adev->irq.ih.ring[ring_index + 1]); - dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]); - dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); + dw[0] = le32_to_cpu(ih->ring[ring_index + 0]); + dw[1] = le32_to_cpu(ih->ring[ring_index + 1]); + dw[2] = le32_to_cpu(ih->ring[ring_index + 2]); + dw[3] = le32_to_cpu(ih->ring[ring_index + 3]); entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY; entry->src_id = dw[0] & 0xff; @@ -135,12 +136,13 @@ static void si_ih_decode_iv(struct amdgpu_device *adev, entry->ring_id = dw[2] & 0xff; entry->vmid = (dw[2] >> 8) & 0xff; - adev->irq.ih.rptr += 16; + ih->rptr += 16; } -static void si_ih_set_rptr(struct amdgpu_device *adev) +static void si_ih_set_rptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { - WREG32(IH_RB_RPTR, adev->irq.ih.rptr); + WREG32(IH_RB_RPTR, ih->rptr); } static int si_ih_early_init(void *handle) diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index 8849b74078d6..62d272b4be19 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.c +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c @@ -43,6 +43,10 @@ #include "hdp/hdp_4_0_sh_mask.h" #include "smuio/smuio_9_0_offset.h" #include "smuio/smuio_9_0_sh_mask.h" +#include "nbio/nbio_7_0_default.h" +#include "nbio/nbio_7_0_sh_mask.h" +#include "nbio/nbio_7_0_smn.h" +#include "mp/mp_9_0_offset.h" #include "soc15.h" #include "soc15_common.h" @@ -385,14 +389,13 @@ void soc15_program_register_sequence(struct amdgpu_device *adev, } - -static int soc15_asic_reset(struct amdgpu_device *adev) +static int soc15_asic_mode1_reset(struct amdgpu_device *adev) { u32 i; amdgpu_atombios_scratch_regs_engine_hung(adev, true); - dev_info(adev->dev, "GPU reset\n"); + dev_info(adev->dev, "GPU mode1 reset\n"); /* disable BM */ pci_clear_master(adev->pdev); @@ -417,6 +420,63 @@ static int soc15_asic_reset(struct amdgpu_device *adev) return 0; } +static int soc15_asic_get_baco_capability(struct amdgpu_device *adev, bool *cap) +{ + void *pp_handle = adev->powerplay.pp_handle; + const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs; + + if (!pp_funcs || !pp_funcs->get_asic_baco_capability) { + *cap = false; + return -ENOENT; + } + + return pp_funcs->get_asic_baco_capability(pp_handle, cap); +} + +static int soc15_asic_baco_reset(struct amdgpu_device *adev) +{ + void *pp_handle = adev->powerplay.pp_handle; + const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs; + + if (!pp_funcs ||!pp_funcs->get_asic_baco_state ||!pp_funcs->set_asic_baco_state) + return -ENOENT; + + /* enter BACO state */ + if (pp_funcs->set_asic_baco_state(pp_handle, 1)) + return -EIO; + + /* exit BACO state */ + if (pp_funcs->set_asic_baco_state(pp_handle, 0)) + return -EIO; + + dev_info(adev->dev, "GPU BACO reset\n"); + + return 0; +} + +static int soc15_asic_reset(struct amdgpu_device *adev) +{ + int ret; + bool baco_reset; + + switch (adev->asic_type) { + case CHIP_VEGA10: + case CHIP_VEGA20: + soc15_asic_get_baco_capability(adev, &baco_reset); + break; + default: + baco_reset = false; + break; + } + + if (baco_reset) + ret = soc15_asic_baco_reset(adev); + else + ret = soc15_asic_mode1_reset(adev); + + return ret; +} + /*static int soc15_set_uvd_clock(struct amdgpu_device *adev, u32 clock, u32 cntl_reg, u32 status_reg) { @@ -535,10 +595,12 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &vega10_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block); amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block); - if (adev->asic_type == CHIP_VEGA20) - amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); - else - amdgpu_device_ip_block_add(adev, &psp_v3_1_ip_block); + if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) { + if (adev->asic_type == CHIP_VEGA20) + amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block); + else + amdgpu_device_ip_block_add(adev, &psp_v3_1_ip_block); + } amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); if (!amdgpu_sriov_vf(adev)) @@ -560,7 +622,8 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev) amdgpu_device_ip_block_add(adev, &vega10_common_ip_block); amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block); amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block); - amdgpu_device_ip_block_add(adev, &psp_v10_0_ip_block); + if (likely(adev->firmware.load_type == AMDGPU_FW_LOAD_PSP)) + amdgpu_device_ip_block_add(adev, &psp_v10_0_ip_block); amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block); amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block); amdgpu_device_ip_block_add(adev, &pp_smu_ip_block); @@ -601,6 +664,68 @@ static bool soc15_need_full_reset(struct amdgpu_device *adev) /* change this when we implement soft reset */ return true; } +static void soc15_get_pcie_usage(struct amdgpu_device *adev, uint64_t *count0, + uint64_t *count1) +{ + uint32_t perfctr = 0; + uint64_t cnt0_of, cnt1_of; + int tmp; + + /* This reports 0 on APUs, so return to avoid writing/reading registers + * that may or may not be different from their GPU counterparts + */ + if (adev->flags & AMD_IS_APU) + return; + + /* Set the 2 events that we wish to watch, defined above */ + /* Reg 40 is # received msgs, Reg 104 is # of posted requests sent */ + perfctr = REG_SET_FIELD(perfctr, PCIE_PERF_CNTL_TXCLK, EVENT0_SEL, 40); + perfctr = REG_SET_FIELD(perfctr, PCIE_PERF_CNTL_TXCLK, EVENT1_SEL, 104); + + /* Write to enable desired perf counters */ + WREG32_PCIE(smnPCIE_PERF_CNTL_TXCLK, perfctr); + /* Zero out and enable the perf counters + * Write 0x5: + * Bit 0 = Start all counters(1) + * Bit 2 = Global counter reset enable(1) + */ + WREG32_PCIE(smnPCIE_PERF_COUNT_CNTL, 0x00000005); + + msleep(1000); + + /* Load the shadow and disable the perf counters + * Write 0x2: + * Bit 0 = Stop counters(0) + * Bit 1 = Load the shadow counters(1) + */ + WREG32_PCIE(smnPCIE_PERF_COUNT_CNTL, 0x00000002); + + /* Read register values to get any >32bit overflow */ + tmp = RREG32_PCIE(smnPCIE_PERF_CNTL_TXCLK); + cnt0_of = REG_GET_FIELD(tmp, PCIE_PERF_CNTL_TXCLK, COUNTER0_UPPER); + cnt1_of = REG_GET_FIELD(tmp, PCIE_PERF_CNTL_TXCLK, COUNTER1_UPPER); + + /* Get the values and add the overflow */ + *count0 = RREG32_PCIE(smnPCIE_PERF_COUNT0_TXCLK) | (cnt0_of << 32); + *count1 = RREG32_PCIE(smnPCIE_PERF_COUNT1_TXCLK) | (cnt1_of << 32); +} + +static bool soc15_need_reset_on_init(struct amdgpu_device *adev) +{ + u32 sol_reg; + + if (adev->flags & AMD_IS_APU) + return false; + + /* Check sOS sign of life register to confirm sys driver and sOS + * are already been loaded. + */ + sol_reg = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81); + if (sol_reg) + return true; + + return false; +} static const struct amdgpu_asic_funcs soc15_asic_funcs = { @@ -617,6 +742,8 @@ static const struct amdgpu_asic_funcs soc15_asic_funcs = .invalidate_hdp = &soc15_invalidate_hdp, .need_full_reset = &soc15_need_full_reset, .init_doorbell_index = &vega10_doorbell_index_init, + .get_pcie_usage = &soc15_get_pcie_usage, + .need_reset_on_init = &soc15_need_reset_on_init, }; static const struct amdgpu_asic_funcs vega20_asic_funcs = @@ -634,6 +761,8 @@ static const struct amdgpu_asic_funcs vega20_asic_funcs = .invalidate_hdp = &soc15_invalidate_hdp, .need_full_reset = &soc15_need_full_reset, .init_doorbell_index = &vega20_doorbell_index_init, + .get_pcie_usage = &soc15_get_pcie_usage, + .need_reset_on_init = &soc15_need_reset_on_init, }; static int soc15_common_early_init(void *handle) @@ -840,6 +969,22 @@ static int soc15_common_sw_fini(void *handle) return 0; } +static void soc15_doorbell_range_init(struct amdgpu_device *adev) +{ + int i; + struct amdgpu_ring *ring; + + for (i = 0; i < adev->sdma.num_instances; i++) { + ring = &adev->sdma.instance[i].ring; + adev->nbio_funcs->sdma_doorbell_range(adev, i, + ring->use_doorbell, ring->doorbell_index, + adev->doorbell_index.sdma_doorbell_range); + } + + adev->nbio_funcs->ih_doorbell_range(adev, adev->irq.ih.use_doorbell, + adev->irq.ih.doorbell_index); +} + static int soc15_common_hw_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; @@ -852,6 +997,12 @@ static int soc15_common_hw_init(void *handle) adev->nbio_funcs->init_registers(adev); /* enable the doorbell aperture */ soc15_enable_doorbell_aperture(adev, true); + /* HW doorbell routing policy: doorbell writing not + * in SDMA/IH/MM/ACV range will be routed to CP. So + * we need to init SDMA/IH/MM/ACV doorbell range prior + * to CP ip block init and ring test. + */ + soc15_doorbell_range_init(adev); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c index 15da06ddeb75..a20b711a6756 100644 --- a/drivers/gpu/drm/amd/amdgpu/tonga_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/tonga_ih.c @@ -99,9 +99,9 @@ static void tonga_ih_disable_interrupts(struct amdgpu_device *adev) */ static int tonga_ih_irq_init(struct amdgpu_device *adev) { - int rb_bufsz; u32 interrupt_cntl, ih_rb_cntl, ih_doorbell_rtpr; - u64 wptr_off; + struct amdgpu_ih_ring *ih = &adev->irq.ih; + int rb_bufsz; /* disable irqs */ tonga_ih_disable_interrupts(adev); @@ -118,10 +118,7 @@ static int tonga_ih_irq_init(struct amdgpu_device *adev) WREG32(mmINTERRUPT_CNTL, interrupt_cntl); /* Ring Buffer base. [39:8] of 40-bit address of the beginning of the ring buffer*/ - if (adev->irq.ih.use_bus_addr) - WREG32(mmIH_RB_BASE, adev->irq.ih.rb_dma_addr >> 8); - else - WREG32(mmIH_RB_BASE, adev->irq.ih.gpu_addr >> 8); + WREG32(mmIH_RB_BASE, ih->gpu_addr >> 8); rb_bufsz = order_base_2(adev->irq.ih.ring_size / 4); ih_rb_cntl = REG_SET_FIELD(0, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); @@ -136,12 +133,8 @@ static int tonga_ih_irq_init(struct amdgpu_device *adev) WREG32(mmIH_RB_CNTL, ih_rb_cntl); /* set the writeback address whether it's enabled or not */ - if (adev->irq.ih.use_bus_addr) - wptr_off = adev->irq.ih.rb_dma_addr + (adev->irq.ih.wptr_offs * 4); - else - wptr_off = adev->wb.gpu_addr + (adev->irq.ih.wptr_offs * 4); - WREG32(mmIH_RB_WPTR_ADDR_LO, lower_32_bits(wptr_off)); - WREG32(mmIH_RB_WPTR_ADDR_HI, upper_32_bits(wptr_off) & 0xFF); + WREG32(mmIH_RB_WPTR_ADDR_LO, lower_32_bits(ih->wptr_addr)); + WREG32(mmIH_RB_WPTR_ADDR_HI, upper_32_bits(ih->wptr_addr) & 0xFF); /* set rptr, wptr to 0 */ WREG32(mmIH_RB_RPTR, 0); @@ -193,14 +186,12 @@ static void tonga_ih_irq_disable(struct amdgpu_device *adev) * Used by cz_irq_process(VI). * Returns the value of the wptr. */ -static u32 tonga_ih_get_wptr(struct amdgpu_device *adev) +static u32 tonga_ih_get_wptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { u32 wptr, tmp; - if (adev->irq.ih.use_bus_addr) - wptr = le32_to_cpu(adev->irq.ih.ring[adev->irq.ih.wptr_offs]); - else - wptr = le32_to_cpu(adev->wb.wb[adev->irq.ih.wptr_offs]); + wptr = le32_to_cpu(*ih->wptr_cpu); if (REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) { wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0); @@ -209,13 +200,13 @@ static u32 tonga_ih_get_wptr(struct amdgpu_device *adev) * this should allow us to catchup. */ dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n", - wptr, adev->irq.ih.rptr, (wptr + 16) & adev->irq.ih.ptr_mask); - adev->irq.ih.rptr = (wptr + 16) & adev->irq.ih.ptr_mask; + wptr, ih->rptr, (wptr + 16) & ih->ptr_mask); + ih->rptr = (wptr + 16) & ih->ptr_mask; tmp = RREG32(mmIH_RB_CNTL); tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); WREG32(mmIH_RB_CNTL, tmp); } - return (wptr & adev->irq.ih.ptr_mask); + return (wptr & ih->ptr_mask); } /** @@ -227,16 +218,17 @@ static u32 tonga_ih_get_wptr(struct amdgpu_device *adev) * position and also advance the position. */ static void tonga_ih_decode_iv(struct amdgpu_device *adev, - struct amdgpu_iv_entry *entry) + struct amdgpu_ih_ring *ih, + struct amdgpu_iv_entry *entry) { /* wptr/rptr are in bytes! */ - u32 ring_index = adev->irq.ih.rptr >> 2; + u32 ring_index = ih->rptr >> 2; uint32_t dw[4]; - dw[0] = le32_to_cpu(adev->irq.ih.ring[ring_index + 0]); - dw[1] = le32_to_cpu(adev->irq.ih.ring[ring_index + 1]); - dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]); - dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); + dw[0] = le32_to_cpu(ih->ring[ring_index + 0]); + dw[1] = le32_to_cpu(ih->ring[ring_index + 1]); + dw[2] = le32_to_cpu(ih->ring[ring_index + 2]); + dw[3] = le32_to_cpu(ih->ring[ring_index + 3]); entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY; entry->src_id = dw[0] & 0xff; @@ -246,7 +238,7 @@ static void tonga_ih_decode_iv(struct amdgpu_device *adev, entry->pasid = (dw[2] >> 16) & 0xffff; /* wptr/rptr are in bytes! */ - adev->irq.ih.rptr += 16; + ih->rptr += 16; } /** @@ -256,17 +248,15 @@ static void tonga_ih_decode_iv(struct amdgpu_device *adev, * * Set the IH ring buffer rptr. */ -static void tonga_ih_set_rptr(struct amdgpu_device *adev) +static void tonga_ih_set_rptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { - if (adev->irq.ih.use_doorbell) { + if (ih->use_doorbell) { /* XXX check if swapping is necessary on BE */ - if (adev->irq.ih.use_bus_addr) - adev->irq.ih.ring[adev->irq.ih.rptr_offs] = adev->irq.ih.rptr; - else - adev->wb.wb[adev->irq.ih.rptr_offs] = adev->irq.ih.rptr; - WDOORBELL32(adev->irq.ih.doorbell_index, adev->irq.ih.rptr); + *ih->rptr_cpu = ih->rptr; + WDOORBELL32(ih->doorbell_index, ih->rptr); } else { - WREG32(mmIH_RB_RPTR, adev->irq.ih.rptr); + WREG32(mmIH_RB_RPTR, ih->rptr); } } diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c index d69c8f6daaf8..c4fb58667fd4 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c @@ -511,7 +511,7 @@ static int uvd_v4_2_ring_test_ring(struct amdgpu_ring *ring) static void uvd_v4_2_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { amdgpu_ring_write(ring, PACKET0(mmUVD_RBC_IB_BASE, 0)); amdgpu_ring_write(ring, ib->gpu_addr); diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c index ee8cd06ddc38..52bd8a654734 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c @@ -526,7 +526,7 @@ static int uvd_v5_0_ring_test_ring(struct amdgpu_ring *ring) static void uvd_v5_0_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { amdgpu_ring_write(ring, PACKET0(mmUVD_LMI_RBC_IB_64BIT_BAR_LOW, 0)); amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr)); diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c index d4f4a66f8324..c9edddf9f88a 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c @@ -977,7 +977,7 @@ static int uvd_v6_0_ring_test_ring(struct amdgpu_ring *ring) static void uvd_v6_0_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); @@ -1003,7 +1003,7 @@ static void uvd_v6_0_ring_emit_ib(struct amdgpu_ring *ring, static void uvd_v6_0_enc_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c index aef924026a28..dc461df48da0 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c @@ -1272,7 +1272,7 @@ static int uvd_v7_0_ring_patch_cs_in_place(struct amdgpu_cs_parser *p, static void uvd_v7_0_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { struct amdgpu_device *adev = ring->adev; unsigned vmid = AMDGPU_JOB_GET_VMID(job); @@ -1303,7 +1303,7 @@ static void uvd_v7_0_ring_emit_ib(struct amdgpu_ring *ring, static void uvd_v7_0_enc_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c index 2668effadd27..6ec65cf11112 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c @@ -834,7 +834,7 @@ out: static void vce_v3_0_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c index 9fb34b7d8e03..aadc3e66ebd7 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v4_0.c @@ -947,7 +947,7 @@ static int vce_v4_0_set_powergating_state(void *handle, #endif static void vce_v4_0_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, - struct amdgpu_ib *ib, bool ctx_switch) + struct amdgpu_ib *ib, uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c index 89bb2fef90eb..3dbc51f9d3b9 100644 --- a/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c @@ -1371,7 +1371,7 @@ static void vcn_v1_0_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 static void vcn_v1_0_dec_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { struct amdgpu_device *adev = ring->adev; unsigned vmid = AMDGPU_JOB_GET_VMID(job); @@ -1531,7 +1531,7 @@ static void vcn_v1_0_enc_ring_insert_end(struct amdgpu_ring *ring) static void vcn_v1_0_enc_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { unsigned vmid = AMDGPU_JOB_GET_VMID(job); @@ -1736,7 +1736,7 @@ static void vcn_v1_0_jpeg_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u6 static void vcn_v1_0_jpeg_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_job *job, struct amdgpu_ib *ib, - bool ctx_switch) + uint32_t flags) { struct amdgpu_device *adev = ring->adev; unsigned vmid = AMDGPU_JOB_GET_VMID(job); diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c index 2c250b01a903..6d1f804277f8 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega10_ih.c +++ b/drivers/gpu/drm/amd/amdgpu/vega10_ih.c @@ -50,6 +50,22 @@ static void vega10_ih_enable_interrupts(struct amdgpu_device *adev) ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, ENABLE_INTR, 1); WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl); adev->irq.ih.enabled = true; + + if (adev->irq.ih1.ring_size) { + ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1); + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING1, + RB_ENABLE, 1); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1, ih_rb_cntl); + adev->irq.ih1.enabled = true; + } + + if (adev->irq.ih2.ring_size) { + ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2); + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING2, + RB_ENABLE, 1); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2, ih_rb_cntl); + adev->irq.ih2.enabled = true; + } } /** @@ -71,6 +87,53 @@ static void vega10_ih_disable_interrupts(struct amdgpu_device *adev) WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR, 0); adev->irq.ih.enabled = false; adev->irq.ih.rptr = 0; + + if (adev->irq.ih1.ring_size) { + ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1); + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING1, + RB_ENABLE, 0); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1, ih_rb_cntl); + /* set rptr, wptr to 0 */ + WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING1, 0); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING1, 0); + adev->irq.ih1.enabled = false; + adev->irq.ih1.rptr = 0; + } + + if (adev->irq.ih2.ring_size) { + ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2); + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL_RING2, + RB_ENABLE, 0); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2, ih_rb_cntl); + /* set rptr, wptr to 0 */ + WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING2, 0); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING2, 0); + adev->irq.ih2.enabled = false; + adev->irq.ih2.rptr = 0; + } +} + +static uint32_t vega10_ih_rb_cntl(struct amdgpu_ih_ring *ih, uint32_t ih_rb_cntl) +{ + int rb_bufsz = order_base_2(ih->ring_size / 4); + + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, + MC_SPACE, ih->use_bus_addr ? 1 : 4); + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, + WPTR_OVERFLOW_CLEAR, 1); + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, + WPTR_OVERFLOW_ENABLE, 1); + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_SIZE, rb_bufsz); + /* Ring Buffer write pointer writeback. If enabled, IH_RB_WPTR register + * value is written to memory + */ + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, + WPTR_WRITEBACK_ENABLE, 1); + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_SNOOP, 1); + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_RO, 0); + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_VMID, 0); + + return ih_rb_cntl; } /** @@ -86,50 +149,32 @@ static void vega10_ih_disable_interrupts(struct amdgpu_device *adev) */ static int vega10_ih_irq_init(struct amdgpu_device *adev) { + struct amdgpu_ih_ring *ih; int ret = 0; - int rb_bufsz; u32 ih_rb_cntl, ih_doorbell_rtpr; u32 tmp; - u64 wptr_off; /* disable irqs */ vega10_ih_disable_interrupts(adev); adev->nbio_funcs->ih_control(adev); - ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL); + ih = &adev->irq.ih; /* Ring Buffer base. [39:8] of 40-bit address of the beginning of the ring buffer*/ - if (adev->irq.ih.use_bus_addr) { - WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE, adev->irq.ih.rb_dma_addr >> 8); - WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI, ((u64)adev->irq.ih.rb_dma_addr >> 40) & 0xff); - ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_SPACE, 1); - } else { - WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE, adev->irq.ih.gpu_addr >> 8); - WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI, (adev->irq.ih.gpu_addr >> 40) & 0xff); - ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_SPACE, 4); - } - rb_bufsz = order_base_2(adev->irq.ih.ring_size / 4); - ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); - ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, WPTR_OVERFLOW_ENABLE, 1); - ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RB_SIZE, rb_bufsz); - /* Ring Buffer write pointer writeback. If enabled, IH_RB_WPTR register value is written to memory */ - ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, WPTR_WRITEBACK_ENABLE, 1); - ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_SNOOP, 1); - ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_RO, 0); - ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, MC_VMID, 0); - - if (adev->irq.msi_enabled) - ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RPTR_REARM, 1); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE, ih->gpu_addr >> 8); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI, (ih->gpu_addr >> 40) & 0xff); + ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL); + ih_rb_cntl = vega10_ih_rb_cntl(ih, ih_rb_cntl); + ih_rb_cntl = REG_SET_FIELD(ih_rb_cntl, IH_RB_CNTL, RPTR_REARM, + !!adev->irq.msi_enabled); WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL, ih_rb_cntl); /* set the writeback address whether it's enabled or not */ - if (adev->irq.ih.use_bus_addr) - wptr_off = adev->irq.ih.rb_dma_addr + (adev->irq.ih.wptr_offs * 4); - else - wptr_off = adev->wb.gpu_addr + (adev->irq.ih.wptr_offs * 4); - WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_ADDR_LO, lower_32_bits(wptr_off)); - WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_ADDR_HI, upper_32_bits(wptr_off) & 0xFFFF); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_ADDR_LO, + lower_32_bits(ih->wptr_addr)); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_ADDR_HI, + upper_32_bits(ih->wptr_addr) & 0xFFFF); /* set rptr, wptr to 0 */ WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, 0); @@ -137,17 +182,48 @@ static int vega10_ih_irq_init(struct amdgpu_device *adev) ih_doorbell_rtpr = RREG32_SOC15(OSSSYS, 0, mmIH_DOORBELL_RPTR); if (adev->irq.ih.use_doorbell) { - ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr, IH_DOORBELL_RPTR, - OFFSET, adev->irq.ih.doorbell_index); - ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr, IH_DOORBELL_RPTR, + ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr, + IH_DOORBELL_RPTR, OFFSET, + adev->irq.ih.doorbell_index); + ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr, + IH_DOORBELL_RPTR, ENABLE, 1); } else { - ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr, IH_DOORBELL_RPTR, + ih_doorbell_rtpr = REG_SET_FIELD(ih_doorbell_rtpr, + IH_DOORBELL_RPTR, ENABLE, 0); } WREG32_SOC15(OSSSYS, 0, mmIH_DOORBELL_RPTR, ih_doorbell_rtpr); - adev->nbio_funcs->ih_doorbell_range(adev, adev->irq.ih.use_doorbell, - adev->irq.ih.doorbell_index); + + ih = &adev->irq.ih1; + if (ih->ring_size) { + WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_RING1, ih->gpu_addr >> 8); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI_RING1, + (ih->gpu_addr >> 40) & 0xff); + + ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1); + ih_rb_cntl = vega10_ih_rb_cntl(ih, ih_rb_cntl); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1, ih_rb_cntl); + + /* set rptr, wptr to 0 */ + WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING1, 0); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING1, 0); + } + + ih = &adev->irq.ih2; + if (ih->ring_size) { + WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_RING2, ih->gpu_addr >> 8); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_BASE_HI_RING2, + (ih->gpu_addr >> 40) & 0xff); + + ih_rb_cntl = RREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING1); + ih_rb_cntl = vega10_ih_rb_cntl(ih, ih_rb_cntl); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_CNTL_RING2, ih_rb_cntl); + + /* set rptr, wptr to 0 */ + WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING2, 0); + WREG32_SOC15(OSSSYS, 0, mmIH_RB_WPTR_RING2, 0); + } tmp = RREG32_SOC15(OSSSYS, 0, mmIH_STORM_CLIENT_LIST_CNTL); tmp = REG_SET_FIELD(tmp, IH_STORM_CLIENT_LIST_CNTL, @@ -191,32 +267,58 @@ static void vega10_ih_irq_disable(struct amdgpu_device *adev) * ring buffer overflow and deal with it. * Returns the value of the wptr. */ -static u32 vega10_ih_get_wptr(struct amdgpu_device *adev) +static u32 vega10_ih_get_wptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { - u32 wptr, tmp; + u32 wptr, reg, tmp; - if (adev->irq.ih.use_bus_addr) - wptr = le32_to_cpu(adev->irq.ih.ring[adev->irq.ih.wptr_offs]); + wptr = le32_to_cpu(*ih->wptr_cpu); + + if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) + goto out; + + /* Double check that the overflow wasn't already cleared. */ + + if (ih == &adev->irq.ih) + reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR); + else if (ih == &adev->irq.ih1) + reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING1); + else if (ih == &adev->irq.ih2) + reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_WPTR_RING2); else - wptr = le32_to_cpu(adev->wb.wb[adev->irq.ih.wptr_offs]); - - if (REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) { - wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0); - - /* When a ring buffer overflow happen start parsing interrupt - * from the last not overwritten vector (wptr + 32). Hopefully - * this should allow us to catchup. - */ - tmp = (wptr + 32) & adev->irq.ih.ptr_mask; - dev_warn(adev->dev, "IH ring buffer overflow (0x%08X, 0x%08X, 0x%08X)\n", - wptr, adev->irq.ih.rptr, tmp); - adev->irq.ih.rptr = tmp; - - tmp = RREG32_NO_KIQ(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL)); - tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); - WREG32_NO_KIQ(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL), tmp); - } - return (wptr & adev->irq.ih.ptr_mask); + BUG(); + + wptr = RREG32_NO_KIQ(reg); + if (!REG_GET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW)) + goto out; + + wptr = REG_SET_FIELD(wptr, IH_RB_WPTR, RB_OVERFLOW, 0); + + /* When a ring buffer overflow happen start parsing interrupt + * from the last not overwritten vector (wptr + 32). Hopefully + * this should allow us to catchup. + */ + tmp = (wptr + 32) & ih->ptr_mask; + dev_warn(adev->dev, "IH ring buffer overflow " + "(0x%08X, 0x%08X, 0x%08X)\n", + wptr, ih->rptr, tmp); + ih->rptr = tmp; + + if (ih == &adev->irq.ih) + reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL); + else if (ih == &adev->irq.ih1) + reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING1); + else if (ih == &adev->irq.ih2) + reg = SOC15_REG_OFFSET(OSSSYS, 0, mmIH_RB_CNTL_RING2); + else + BUG(); + + tmp = RREG32_NO_KIQ(reg); + tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1); + WREG32_NO_KIQ(reg, tmp); + +out: + return (wptr & ih->ptr_mask); } /** @@ -228,20 +330,21 @@ static u32 vega10_ih_get_wptr(struct amdgpu_device *adev) * position and also advance the position. */ static void vega10_ih_decode_iv(struct amdgpu_device *adev, - struct amdgpu_iv_entry *entry) + struct amdgpu_ih_ring *ih, + struct amdgpu_iv_entry *entry) { /* wptr/rptr are in bytes! */ - u32 ring_index = adev->irq.ih.rptr >> 2; + u32 ring_index = ih->rptr >> 2; uint32_t dw[8]; - dw[0] = le32_to_cpu(adev->irq.ih.ring[ring_index + 0]); - dw[1] = le32_to_cpu(adev->irq.ih.ring[ring_index + 1]); - dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]); - dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]); - dw[4] = le32_to_cpu(adev->irq.ih.ring[ring_index + 4]); - dw[5] = le32_to_cpu(adev->irq.ih.ring[ring_index + 5]); - dw[6] = le32_to_cpu(adev->irq.ih.ring[ring_index + 6]); - dw[7] = le32_to_cpu(adev->irq.ih.ring[ring_index + 7]); + dw[0] = le32_to_cpu(ih->ring[ring_index + 0]); + dw[1] = le32_to_cpu(ih->ring[ring_index + 1]); + dw[2] = le32_to_cpu(ih->ring[ring_index + 2]); + dw[3] = le32_to_cpu(ih->ring[ring_index + 3]); + dw[4] = le32_to_cpu(ih->ring[ring_index + 4]); + dw[5] = le32_to_cpu(ih->ring[ring_index + 5]); + dw[6] = le32_to_cpu(ih->ring[ring_index + 6]); + dw[7] = le32_to_cpu(ih->ring[ring_index + 7]); entry->client_id = dw[0] & 0xff; entry->src_id = (dw[0] >> 8) & 0xff; @@ -257,9 +360,8 @@ static void vega10_ih_decode_iv(struct amdgpu_device *adev, entry->src_data[2] = dw[6]; entry->src_data[3] = dw[7]; - /* wptr/rptr are in bytes! */ - adev->irq.ih.rptr += 32; + ih->rptr += 32; } /** @@ -269,37 +371,95 @@ static void vega10_ih_decode_iv(struct amdgpu_device *adev, * * Set the IH ring buffer rptr. */ -static void vega10_ih_set_rptr(struct amdgpu_device *adev) +static void vega10_ih_set_rptr(struct amdgpu_device *adev, + struct amdgpu_ih_ring *ih) { - if (adev->irq.ih.use_doorbell) { + if (ih->use_doorbell) { /* XXX check if swapping is necessary on BE */ - if (adev->irq.ih.use_bus_addr) - adev->irq.ih.ring[adev->irq.ih.rptr_offs] = adev->irq.ih.rptr; - else - adev->wb.wb[adev->irq.ih.rptr_offs] = adev->irq.ih.rptr; - WDOORBELL32(adev->irq.ih.doorbell_index, adev->irq.ih.rptr); - } else { - WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, adev->irq.ih.rptr); + *ih->rptr_cpu = ih->rptr; + WDOORBELL32(ih->doorbell_index, ih->rptr); + } else if (ih == &adev->irq.ih) { + WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR, ih->rptr); + } else if (ih == &adev->irq.ih1) { + WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING1, ih->rptr); + } else if (ih == &adev->irq.ih2) { + WREG32_SOC15(OSSSYS, 0, mmIH_RB_RPTR_RING2, ih->rptr); } } +/** + * vega10_ih_self_irq - dispatch work for ring 1 and 2 + * + * @adev: amdgpu_device pointer + * @source: irq source + * @entry: IV with WPTR update + * + * Update the WPTR from the IV and schedule work to handle the entries. + */ +static int vega10_ih_self_irq(struct amdgpu_device *adev, + struct amdgpu_irq_src *source, + struct amdgpu_iv_entry *entry) +{ + uint32_t wptr = cpu_to_le32(entry->src_data[0]); + + switch (entry->ring_id) { + case 1: + *adev->irq.ih1.wptr_cpu = wptr; + schedule_work(&adev->irq.ih1_work); + break; + case 2: + *adev->irq.ih2.wptr_cpu = wptr; + schedule_work(&adev->irq.ih2_work); + break; + default: break; + } + return 0; +} + +static const struct amdgpu_irq_src_funcs vega10_ih_self_irq_funcs = { + .process = vega10_ih_self_irq, +}; + +static void vega10_ih_set_self_irq_funcs(struct amdgpu_device *adev) +{ + adev->irq.self_irq.num_types = 0; + adev->irq.self_irq.funcs = &vega10_ih_self_irq_funcs; +} + static int vega10_ih_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; vega10_ih_set_interrupt_funcs(adev); + vega10_ih_set_self_irq_funcs(adev); return 0; } static int vega10_ih_sw_init(void *handle) { - int r; struct amdgpu_device *adev = (struct amdgpu_device *)handle; + int r; + + r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_IH, 0, + &adev->irq.self_irq); + if (r) + return r; r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 256 * 1024, true); if (r) return r; + if (adev->asic_type == CHIP_VEGA10) { + r = amdgpu_ih_ring_init(adev, &adev->irq.ih1, PAGE_SIZE, true); + if (r) + return r; + + r = amdgpu_ih_ring_init(adev, &adev->irq.ih2, PAGE_SIZE, true); + if (r) + return r; + } + + /* TODO add doorbell for IH1 & IH2 as well */ adev->irq.ih.use_doorbell = true; adev->irq.ih.doorbell_index = adev->doorbell_index.ih << 1; @@ -313,6 +473,8 @@ static int vega10_ih_sw_fini(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; amdgpu_irq_fini(adev); + amdgpu_ih_ring_fini(adev, &adev->irq.ih2); + amdgpu_ih_ring_fini(adev, &adev->irq.ih1); amdgpu_ih_ring_fini(adev, &adev->irq.ih); return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/vega10_reg_init.c b/drivers/gpu/drm/amd/amdgpu/vega10_reg_init.c index 422674bb3cdf..4b5d60ea3e78 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega10_reg_init.c +++ b/drivers/gpu/drm/amd/amdgpu/vega10_reg_init.c @@ -70,8 +70,8 @@ void vega10_doorbell_index_init(struct amdgpu_device *adev) adev->doorbell_index.userqueue_start = AMDGPU_DOORBELL64_USERQUEUE_START; adev->doorbell_index.userqueue_end = AMDGPU_DOORBELL64_USERQUEUE_END; adev->doorbell_index.gfx_ring0 = AMDGPU_DOORBELL64_GFX_RING0; - adev->doorbell_index.sdma_engine0 = AMDGPU_DOORBELL64_sDMA_ENGINE0; - adev->doorbell_index.sdma_engine1 = AMDGPU_DOORBELL64_sDMA_ENGINE1; + adev->doorbell_index.sdma_engine[0] = AMDGPU_DOORBELL64_sDMA_ENGINE0; + adev->doorbell_index.sdma_engine[1] = AMDGPU_DOORBELL64_sDMA_ENGINE1; adev->doorbell_index.ih = AMDGPU_DOORBELL64_IH; adev->doorbell_index.uvd_vce.uvd_ring0_1 = AMDGPU_DOORBELL64_UVD_RING0_1; adev->doorbell_index.uvd_vce.uvd_ring2_3 = AMDGPU_DOORBELL64_UVD_RING2_3; @@ -83,5 +83,6 @@ void vega10_doorbell_index_init(struct amdgpu_device *adev) adev->doorbell_index.uvd_vce.vce_ring6_7 = AMDGPU_DOORBELL64_VCE_RING6_7; /* In unit of dword doorbell */ adev->doorbell_index.max_assignment = AMDGPU_DOORBELL64_MAX_ASSIGNMENT << 1; + adev->doorbell_index.sdma_doorbell_range = 4; } diff --git a/drivers/gpu/drm/amd/amdgpu/vega20_reg_init.c b/drivers/gpu/drm/amd/amdgpu/vega20_reg_init.c index edce413fda9a..53716c593b2b 100644 --- a/drivers/gpu/drm/amd/amdgpu/vega20_reg_init.c +++ b/drivers/gpu/drm/amd/amdgpu/vega20_reg_init.c @@ -68,14 +68,14 @@ void vega20_doorbell_index_init(struct amdgpu_device *adev) adev->doorbell_index.userqueue_start = AMDGPU_VEGA20_DOORBELL_USERQUEUE_START; adev->doorbell_index.userqueue_end = AMDGPU_VEGA20_DOORBELL_USERQUEUE_END; adev->doorbell_index.gfx_ring0 = AMDGPU_VEGA20_DOORBELL_GFX_RING0; - adev->doorbell_index.sdma_engine0 = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE0; - adev->doorbell_index.sdma_engine1 = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE1; - adev->doorbell_index.sdma_engine2 = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE2; - adev->doorbell_index.sdma_engine3 = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE3; - adev->doorbell_index.sdma_engine4 = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE4; - adev->doorbell_index.sdma_engine5 = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE5; - adev->doorbell_index.sdma_engine6 = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE6; - adev->doorbell_index.sdma_engine7 = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE7; + adev->doorbell_index.sdma_engine[0] = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE0; + adev->doorbell_index.sdma_engine[1] = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE1; + adev->doorbell_index.sdma_engine[2] = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE2; + adev->doorbell_index.sdma_engine[3] = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE3; + adev->doorbell_index.sdma_engine[4] = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE4; + adev->doorbell_index.sdma_engine[5] = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE5; + adev->doorbell_index.sdma_engine[6] = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE6; + adev->doorbell_index.sdma_engine[7] = AMDGPU_VEGA20_DOORBELL_sDMA_ENGINE7; adev->doorbell_index.ih = AMDGPU_VEGA20_DOORBELL_IH; adev->doorbell_index.uvd_vce.uvd_ring0_1 = AMDGPU_VEGA20_DOORBELL64_UVD_RING0_1; adev->doorbell_index.uvd_vce.uvd_ring2_3 = AMDGPU_VEGA20_DOORBELL64_UVD_RING2_3; @@ -86,5 +86,6 @@ void vega20_doorbell_index_init(struct amdgpu_device *adev) adev->doorbell_index.uvd_vce.vce_ring4_5 = AMDGPU_VEGA20_DOORBELL64_VCE_RING4_5; adev->doorbell_index.uvd_vce.vce_ring6_7 = AMDGPU_VEGA20_DOORBELL64_VCE_RING6_7; adev->doorbell_index.max_assignment = AMDGPU_VEGA20_DOORBELL_MAX_ASSIGNMENT << 1; + adev->doorbell_index.sdma_doorbell_range = 20; } diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c index 77e367459101..5e5b42a0744a 100644 --- a/drivers/gpu/drm/amd/amdgpu/vi.c +++ b/drivers/gpu/drm/amd/amdgpu/vi.c @@ -941,6 +941,69 @@ static bool vi_need_full_reset(struct amdgpu_device *adev) } } +static void vi_get_pcie_usage(struct amdgpu_device *adev, uint64_t *count0, + uint64_t *count1) +{ + uint32_t perfctr = 0; + uint64_t cnt0_of, cnt1_of; + int tmp; + + /* This reports 0 on APUs, so return to avoid writing/reading registers + * that may or may not be different from their GPU counterparts + */ + if (adev->flags & AMD_IS_APU) + return; + + /* Set the 2 events that we wish to watch, defined above */ + /* Reg 40 is # received msgs, Reg 104 is # of posted requests sent */ + perfctr = REG_SET_FIELD(perfctr, PCIE_PERF_CNTL_TXCLK, EVENT0_SEL, 40); + perfctr = REG_SET_FIELD(perfctr, PCIE_PERF_CNTL_TXCLK, EVENT1_SEL, 104); + + /* Write to enable desired perf counters */ + WREG32_PCIE(ixPCIE_PERF_CNTL_TXCLK, perfctr); + /* Zero out and enable the perf counters + * Write 0x5: + * Bit 0 = Start all counters(1) + * Bit 2 = Global counter reset enable(1) + */ + WREG32_PCIE(ixPCIE_PERF_COUNT_CNTL, 0x00000005); + + msleep(1000); + + /* Load the shadow and disable the perf counters + * Write 0x2: + * Bit 0 = Stop counters(0) + * Bit 1 = Load the shadow counters(1) + */ + WREG32_PCIE(ixPCIE_PERF_COUNT_CNTL, 0x00000002); + + /* Read register values to get any >32bit overflow */ + tmp = RREG32_PCIE(ixPCIE_PERF_CNTL_TXCLK); + cnt0_of = REG_GET_FIELD(tmp, PCIE_PERF_CNTL_TXCLK, COUNTER0_UPPER); + cnt1_of = REG_GET_FIELD(tmp, PCIE_PERF_CNTL_TXCLK, COUNTER1_UPPER); + + /* Get the values and add the overflow */ + *count0 = RREG32_PCIE(ixPCIE_PERF_COUNT0_TXCLK) | (cnt0_of << 32); + *count1 = RREG32_PCIE(ixPCIE_PERF_COUNT1_TXCLK) | (cnt1_of << 32); +} + +static bool vi_need_reset_on_init(struct amdgpu_device *adev) +{ + u32 clock_cntl, pc; + + if (adev->flags & AMD_IS_APU) + return false; + + /* check if the SMC is already running */ + clock_cntl = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0); + pc = RREG32_SMC(ixSMC_PC_C); + if ((0 == REG_GET_FIELD(clock_cntl, SMC_SYSCON_CLOCK_CNTL_0, ck_disable)) && + (0x20100 <= pc)) + return true; + + return false; +} + static const struct amdgpu_asic_funcs vi_asic_funcs = { .read_disabled_bios = &vi_read_disabled_bios, @@ -956,6 +1019,8 @@ static const struct amdgpu_asic_funcs vi_asic_funcs = .invalidate_hdp = &vi_invalidate_hdp, .need_full_reset = &vi_need_full_reset, .init_doorbell_index = &legacy_doorbell_index_init, + .get_pcie_usage = &vi_get_pcie_usage, + .need_reset_on_init = &vi_need_reset_on_init, }; #define CZ_REV_BRISTOL(rev) \ @@ -1726,8 +1791,8 @@ void legacy_doorbell_index_init(struct amdgpu_device *adev) adev->doorbell_index.mec_ring6 = AMDGPU_DOORBELL_MEC_RING6; adev->doorbell_index.mec_ring7 = AMDGPU_DOORBELL_MEC_RING7; adev->doorbell_index.gfx_ring0 = AMDGPU_DOORBELL_GFX_RING0; - adev->doorbell_index.sdma_engine0 = AMDGPU_DOORBELL_sDMA_ENGINE0; - adev->doorbell_index.sdma_engine1 = AMDGPU_DOORBELL_sDMA_ENGINE1; + adev->doorbell_index.sdma_engine[0] = AMDGPU_DOORBELL_sDMA_ENGINE0; + adev->doorbell_index.sdma_engine[1] = AMDGPU_DOORBELL_sDMA_ENGINE1; adev->doorbell_index.ih = AMDGPU_DOORBELL_IH; adev->doorbell_index.max_assignment = AMDGPU_DOORBELL_MAX_ASSIGNMENT; } diff --git a/drivers/gpu/drm/amd/amdkfd/Kconfig b/drivers/gpu/drm/amd/amdkfd/Kconfig index fbf0ee5201c3..c3613604a4f8 100644 --- a/drivers/gpu/drm/amd/amdkfd/Kconfig +++ b/drivers/gpu/drm/amd/amdkfd/Kconfig @@ -4,8 +4,8 @@ config HSA_AMD bool "HSA kernel driver for AMD GPU devices" - depends on DRM_AMDGPU && X86_64 - imply AMD_IOMMU_V2 + depends on DRM_AMDGPU && (X86_64 || ARM64) + imply AMD_IOMMU_V2 if X86_64 select MMU_NOTIFIER help Enable this if you want to use HSA features on AMD GPU devices. diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c index b7bc7d7d048f..5d85ff341385 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c @@ -863,6 +863,7 @@ static int kfd_fill_mem_info_for_cpu(int numa_node_id, int *avail_size, return 0; } +#if CONFIG_X86_64 static int kfd_fill_iolink_info_for_cpu(int numa_node_id, int *avail_size, uint32_t *num_entries, struct crat_subtype_iolink *sub_type_hdr) @@ -905,6 +906,7 @@ static int kfd_fill_iolink_info_for_cpu(int numa_node_id, int *avail_size, return 0; } +#endif /* kfd_create_vcrat_image_cpu - Create Virtual CRAT for CPU * @@ -920,7 +922,9 @@ static int kfd_create_vcrat_image_cpu(void *pcrat_image, size_t *size) struct crat_subtype_generic *sub_type_hdr; int avail_size = *size; int numa_node_id; +#ifdef CONFIG_X86_64 uint32_t entries = 0; +#endif int ret = 0; if (!pcrat_image || avail_size < VCRAT_SIZE_FOR_CPU) @@ -982,6 +986,7 @@ static int kfd_create_vcrat_image_cpu(void *pcrat_image, size_t *size) sub_type_hdr->length); /* Fill in Subtype: IO Link */ +#ifdef CONFIG_X86_64 ret = kfd_fill_iolink_info_for_cpu(numa_node_id, &avail_size, &entries, (struct crat_subtype_iolink *)sub_type_hdr); @@ -992,6 +997,9 @@ static int kfd_create_vcrat_image_cpu(void *pcrat_image, size_t *size) sub_type_hdr = (typeof(sub_type_hdr))((char *)sub_type_hdr + sub_type_hdr->length * entries); +#else + pr_info("IO link not available for non x86 platforms\n"); +#endif crat_table->num_domains++; } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_module.c b/drivers/gpu/drm/amd/amdkfd/kfd_module.c index 8018163414ff..932007eb9168 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_module.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_module.c @@ -23,22 +23,7 @@ #include <linux/sched.h> #include <linux/device.h> #include "kfd_priv.h" - -static const struct kgd2kfd_calls kgd2kfd = { - .exit = kgd2kfd_exit, - .probe = kgd2kfd_probe, - .device_init = kgd2kfd_device_init, - .device_exit = kgd2kfd_device_exit, - .interrupt = kgd2kfd_interrupt, - .suspend = kgd2kfd_suspend, - .resume = kgd2kfd_resume, - .quiesce_mm = kgd2kfd_quiesce_mm, - .resume_mm = kgd2kfd_resume_mm, - .schedule_evict_and_restore_process = - kgd2kfd_schedule_evict_and_restore_process, - .pre_reset = kgd2kfd_pre_reset, - .post_reset = kgd2kfd_post_reset, -}; +#include "amdgpu_amdkfd.h" static int kfd_init(void) { @@ -91,20 +76,10 @@ static void kfd_exit(void) kfd_chardev_exit(); } -int kgd2kfd_init(unsigned int interface_version, - const struct kgd2kfd_calls **g2f) +int kgd2kfd_init() { - int err; - - err = kfd_init(); - if (err) - return err; - - *g2f = &kgd2kfd; - - return 0; + return kfd_init(); } -EXPORT_SYMBOL(kgd2kfd_init); void kgd2kfd_exit(void) { diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h index 0689d4ccbbc0..12b66330fc6d 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h @@ -266,14 +266,6 @@ struct kfd_dev { bool pci_atomic_requested; }; -/* KGD2KFD callbacks */ -void kgd2kfd_exit(void); -struct kfd_dev *kgd2kfd_probe(struct kgd_dev *kgd, - struct pci_dev *pdev, const struct kfd2kgd_calls *f2g); -bool kgd2kfd_device_init(struct kfd_dev *kfd, - const struct kgd2kfd_shared_resources *gpu_resources); -void kgd2kfd_device_exit(struct kfd_dev *kfd); - enum kfd_mempool { KFD_MEMPOOL_SYSTEM_CACHEABLE = 1, KFD_MEMPOOL_SYSTEM_WRITECOMBINE = 2, @@ -541,11 +533,6 @@ struct qcm_process_device { /* Approx. time before evicting the process again */ #define PROCESS_ACTIVE_TIME_MS 10 -int kgd2kfd_quiesce_mm(struct mm_struct *mm); -int kgd2kfd_resume_mm(struct mm_struct *mm); -int kgd2kfd_schedule_evict_and_restore_process(struct mm_struct *mm, - struct dma_fence *fence); - /* 8 byte handle containing GPU ID in the most significant 4 bytes and * idr_handle in the least significant 4 bytes */ @@ -800,20 +787,11 @@ int kfd_numa_node_to_apic_id(int numa_node_id); /* Interrupts */ int kfd_interrupt_init(struct kfd_dev *dev); void kfd_interrupt_exit(struct kfd_dev *dev); -void kgd2kfd_interrupt(struct kfd_dev *kfd, const void *ih_ring_entry); bool enqueue_ih_ring_entry(struct kfd_dev *kfd, const void *ih_ring_entry); bool interrupt_is_wanted(struct kfd_dev *dev, const uint32_t *ih_ring_entry, uint32_t *patched_ihre, bool *flag); -/* Power Management */ -void kgd2kfd_suspend(struct kfd_dev *kfd); -int kgd2kfd_resume(struct kfd_dev *kfd); - -/* GPU reset */ -int kgd2kfd_pre_reset(struct kfd_dev *kfd); -int kgd2kfd_post_reset(struct kfd_dev *kfd); - /* amdkfd Apertures */ int kfd_init_apertures(struct kfd_process *process); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c index 5f5b2acedbac..09da91644f9f 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c @@ -1093,8 +1093,6 @@ static uint32_t kfd_generate_gpu_id(struct kfd_dev *gpu) * the GPU device is not already present in the topology device * list then return NULL. This means a new topology device has to * be created for this GPU. - * TODO: Rather than assiging @gpu to first topology device withtout - * gpu attached, it will better to have more stringent check. */ static struct kfd_topology_device *kfd_assign_gpu(struct kfd_dev *gpu) { @@ -1102,12 +1100,20 @@ static struct kfd_topology_device *kfd_assign_gpu(struct kfd_dev *gpu) struct kfd_topology_device *out_dev = NULL; down_write(&topology_lock); - list_for_each_entry(dev, &topology_device_list, list) + list_for_each_entry(dev, &topology_device_list, list) { + /* Discrete GPUs need their own topology device list + * entries. Don't assign them to CPU/APU nodes. + */ + if (!gpu->device_info->needs_iommu_device && + dev->node_props.cpu_cores_count) + continue; + if (!dev->gpu && (dev->node_props.simd_count > 0)) { dev->gpu = gpu; out_dev = dev; break; } + } up_write(&topology_lock); return out_dev; } @@ -1392,7 +1398,6 @@ int kfd_topology_enum_kfd_devices(uint8_t idx, struct kfd_dev **kdev) static int kfd_cpumask_to_apic_id(const struct cpumask *cpumask) { - const struct cpuinfo_x86 *cpuinfo; int first_cpu_of_numa_node; if (!cpumask || cpumask == cpu_none_mask) @@ -1400,9 +1405,11 @@ static int kfd_cpumask_to_apic_id(const struct cpumask *cpumask) first_cpu_of_numa_node = cpumask_first(cpumask); if (first_cpu_of_numa_node >= nr_cpu_ids) return -1; - cpuinfo = &cpu_data(first_cpu_of_numa_node); - - return cpuinfo->apicid; +#ifdef CONFIG_X86_64 + return cpu_data(first_cpu_of_numa_node).apicid; +#else + return first_cpu_of_numa_node; +#endif } /* kfd_numa_node_to_apic_id - Returns the APIC ID of the first logical processor diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 8a626d16e8e3..ad31d7b9912f 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -1760,7 +1760,7 @@ static int amdgpu_dm_backlight_update_status(struct backlight_device *bd) + caps.min_input_signal * 0x101; if (dc_link_set_backlight_level(dm->backlight_link, - brightness, 0, 0)) + brightness, 0)) return 0; else return 1; @@ -2284,6 +2284,71 @@ static int get_fb_info(const struct amdgpu_framebuffer *amdgpu_fb, return r; } +static inline uint64_t get_dcc_address(uint64_t address, uint64_t tiling_flags) +{ + uint32_t offset = AMDGPU_TILING_GET(tiling_flags, DCC_OFFSET_256B); + + return offset ? (address + offset * 256) : 0; +} + +static bool fill_plane_dcc_attributes(struct amdgpu_device *adev, + const struct amdgpu_framebuffer *afb, + struct dc_plane_state *plane_state, + uint64_t info) +{ + struct dc *dc = adev->dm.dc; + struct dc_dcc_surface_param input; + struct dc_surface_dcc_cap output; + uint32_t offset = AMDGPU_TILING_GET(info, DCC_OFFSET_256B); + uint32_t i64b = AMDGPU_TILING_GET(info, DCC_INDEPENDENT_64B) != 0; + uint64_t dcc_address; + + memset(&input, 0, sizeof(input)); + memset(&output, 0, sizeof(output)); + + if (!offset) + return false; + + if (!dc->cap_funcs.get_dcc_compression_cap) + return false; + + input.format = plane_state->format; + input.surface_size.width = + plane_state->plane_size.grph.surface_size.width; + input.surface_size.height = + plane_state->plane_size.grph.surface_size.height; + input.swizzle_mode = plane_state->tiling_info.gfx9.swizzle; + + if (plane_state->rotation == ROTATION_ANGLE_0 || + plane_state->rotation == ROTATION_ANGLE_180) + input.scan = SCAN_DIRECTION_HORIZONTAL; + else if (plane_state->rotation == ROTATION_ANGLE_90 || + plane_state->rotation == ROTATION_ANGLE_270) + input.scan = SCAN_DIRECTION_VERTICAL; + + if (!dc->cap_funcs.get_dcc_compression_cap(dc, &input, &output)) + return false; + + if (!output.capable) + return false; + + if (i64b == 0 && output.grph.rgb.independent_64b_blks != 0) + return false; + + plane_state->dcc.enable = 1; + plane_state->dcc.grph.meta_pitch = + AMDGPU_TILING_GET(info, DCC_PITCH_MAX) + 1; + plane_state->dcc.grph.independent_64b_blks = i64b; + + dcc_address = get_dcc_address(afb->address, info); + plane_state->address.grph.meta_addr.low_part = + lower_32_bits(dcc_address); + plane_state->address.grph.meta_addr.high_part = + upper_32_bits(dcc_address); + + return true; +} + static int fill_plane_attributes_from_fb(struct amdgpu_device *adev, struct dc_plane_state *plane_state, const struct amdgpu_framebuffer *amdgpu_fb) @@ -2336,6 +2401,10 @@ static int fill_plane_attributes_from_fb(struct amdgpu_device *adev, return -EINVAL; } + memset(&plane_state->address, 0, sizeof(plane_state->address)); + memset(&plane_state->tiling_info, 0, sizeof(plane_state->tiling_info)); + memset(&plane_state->dcc, 0, sizeof(plane_state->dcc)); + if (plane_state->format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN) { plane_state->address.type = PLN_ADDR_TYPE_GRAPHICS; plane_state->plane_size.grph.surface_size.x = 0; @@ -2367,8 +2436,6 @@ static int fill_plane_attributes_from_fb(struct amdgpu_device *adev, plane_state->color_space = COLOR_SPACE_YCBCR709; } - memset(&plane_state->tiling_info, 0, sizeof(plane_state->tiling_info)); - /* Fill GFX8 params */ if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE) == DC_ARRAY_2D_TILED_THIN1) { unsigned int bankw, bankh, mtaspect, tile_split, num_banks; @@ -2417,6 +2484,9 @@ static int fill_plane_attributes_from_fb(struct amdgpu_device *adev, plane_state->tiling_info.gfx9.swizzle = AMDGPU_TILING_GET(tiling_flags, SWIZZLE_MODE); plane_state->tiling_info.gfx9.shaderEnable = 1; + + fill_plane_dcc_attributes(adev, amdgpu_fb, plane_state, + tiling_flags); } plane_state->visible = true; @@ -2580,7 +2650,7 @@ get_output_color_space(const struct dc_crtc_timing *dc_crtc_timing) * according to HDMI spec, we use YCbCr709 and YCbCr601 * respectively */ - if (dc_crtc_timing->pix_clk_khz > 27030) { + if (dc_crtc_timing->pix_clk_100hz > 270300) { if (dc_crtc_timing->flags.Y_ONLY) color_space = COLOR_SPACE_YCBCR709_LIMITED; @@ -2623,7 +2693,7 @@ static void adjust_colour_depth_from_display_info(struct dc_crtc_timing *timing_ if (timing_out->display_color_depth <= COLOR_DEPTH_888) return; do { - normalized_clk = timing_out->pix_clk_khz; + normalized_clk = timing_out->pix_clk_100hz / 10; /* YCbCr 4:2:0 requires additional adjustment of 1/2 */ if (timing_out->pixel_encoding == PIXEL_ENCODING_YCBCR420) normalized_clk /= 2; @@ -2666,10 +2736,10 @@ fill_stream_properties_from_drm_display_mode(struct dc_stream_state *stream, timing_out->v_border_bottom = 0; /* TODO: un-hardcode */ if (drm_mode_is_420_only(info, mode_in) - && stream->sink->sink_signal == SIGNAL_TYPE_HDMI_TYPE_A) + && stream->signal == SIGNAL_TYPE_HDMI_TYPE_A) timing_out->pixel_encoding = PIXEL_ENCODING_YCBCR420; else if ((connector->display_info.color_formats & DRM_COLOR_FORMAT_YCRCB444) - && stream->sink->sink_signal == SIGNAL_TYPE_HDMI_TYPE_A) + && stream->signal == SIGNAL_TYPE_HDMI_TYPE_A) timing_out->pixel_encoding = PIXEL_ENCODING_YCBCR444; else timing_out->pixel_encoding = PIXEL_ENCODING_RGB; @@ -2704,14 +2774,14 @@ fill_stream_properties_from_drm_display_mode(struct dc_stream_state *stream, mode_in->crtc_vsync_start - mode_in->crtc_vdisplay; timing_out->v_sync_width = mode_in->crtc_vsync_end - mode_in->crtc_vsync_start; - timing_out->pix_clk_khz = mode_in->crtc_clock; + timing_out->pix_clk_100hz = mode_in->crtc_clock * 10; timing_out->aspect_ratio = get_aspect_ratio(mode_in); stream->output_color_space = get_output_color_space(timing_out); stream->out_transfer_func->type = TF_TYPE_PREDEFINED; stream->out_transfer_func->tf = TRANSFER_FUNCTION_SRGB; - if (stream->sink->sink_signal == SIGNAL_TYPE_HDMI_TYPE_A) + if (stream->signal == SIGNAL_TYPE_HDMI_TYPE_A) adjust_colour_depth_from_display_info(timing_out, info); } @@ -2832,7 +2902,7 @@ static void set_master_stream(struct dc_stream_state *stream_set[], if (stream_set[j] && stream_set[j]->triggered_crtc_reset.enabled) { int refresh_rate = 0; - refresh_rate = (stream_set[j]->timing.pix_clk_khz*1000)/ + refresh_rate = (stream_set[j]->timing.pix_clk_100hz*100)/ (stream_set[j]->timing.h_total*stream_set[j]->timing.v_total); if (refresh_rate > highest_rfr) { highest_rfr = refresh_rate; @@ -2889,11 +2959,9 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector, drm_connector = &aconnector->base; if (!aconnector->dc_sink) { - if (!aconnector->mst_port) { - sink = create_fake_sink(aconnector); - if (!sink) - return stream; - } + sink = create_fake_sink(aconnector); + if (!sink) + return stream; } else { sink = aconnector->dc_sink; } @@ -2905,6 +2973,8 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector, goto finish; } + stream->dm_stream_context = aconnector; + list_for_each_entry(preferred_mode, &aconnector->base.modes, head) { /* Search for preferred mode */ if (preferred_mode->type & DRM_MODE_TYPE_PREFERRED) { @@ -2956,10 +3026,7 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector, drm_connector, sink); - update_stream_signal(stream); - - if (dm_state && dm_state->freesync_capable) - stream->ignore_msa_timing_param = true; + update_stream_signal(stream, sink); finish: if (sink && sink->sink_signal == SIGNAL_TYPE_VIRTUAL && aconnector->base.force != DRM_FORCE_ON) @@ -3532,6 +3599,7 @@ static int dm_plane_helper_prepare_fb(struct drm_plane *plane, struct amdgpu_bo *rbo; uint64_t chroma_addr = 0; struct dm_plane_state *dm_plane_state_new, *dm_plane_state_old; + uint64_t tiling_flags, dcc_address; unsigned int awidth; uint32_t domain; int r; @@ -3572,6 +3640,9 @@ static int dm_plane_helper_prepare_fb(struct drm_plane *plane, DRM_ERROR("%p bind failed\n", rbo); return r; } + + amdgpu_bo_get_tiling_flags(rbo, &tiling_flags); + amdgpu_bo_unreserve(rbo); afb->address = amdgpu_bo_gpu_offset(rbo); @@ -3585,6 +3656,13 @@ static int dm_plane_helper_prepare_fb(struct drm_plane *plane, if (plane_state->format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN) { plane_state->address.grph.addr.low_part = lower_32_bits(afb->address); plane_state->address.grph.addr.high_part = upper_32_bits(afb->address); + + dcc_address = + get_dcc_address(afb->address, tiling_flags); + plane_state->address.grph.meta_addr.low_part = + lower_32_bits(dcc_address); + plane_state->address.grph.meta_addr.high_part = + upper_32_bits(dcc_address); } else { awidth = ALIGN(new_state->fb->width, 64); plane_state->address.type = PLN_ADDR_TYPE_VIDEO_PROGRESSIVE; @@ -4456,20 +4534,6 @@ static void prepare_flip_isr(struct amdgpu_crtc *acrtc) acrtc->crtc_id); } -struct dc_stream_status *dc_state_get_stream_status( - struct dc_state *state, - struct dc_stream_state *stream) -{ - uint8_t i; - - for (i = 0; i < state->stream_count; i++) { - if (stream == state->streams[i]) - return &state->stream_status[i]; - } - - return NULL; -} - static void update_freesync_state_on_stream( struct amdgpu_display_manager *dm, struct dm_crtc_state *new_crtc_state, @@ -4523,12 +4587,12 @@ static void update_freesync_state_on_stream( TRANSFER_FUNC_UNKNOWN, &vrr_infopacket); - new_crtc_state->freesync_timing_changed = + new_crtc_state->freesync_timing_changed |= (memcmp(&new_crtc_state->vrr_params.adjust, &vrr_params.adjust, sizeof(vrr_params.adjust)) != 0); - new_crtc_state->freesync_vrr_info_changed = + new_crtc_state->freesync_vrr_info_changed |= (memcmp(&new_crtc_state->vrr_infopacket, &vrr_infopacket, sizeof(vrr_infopacket)) != 0); @@ -4544,254 +4608,6 @@ static void update_freesync_state_on_stream( new_crtc_state->base.crtc->base.id, (int)new_crtc_state->base.vrr_enabled, (int)vrr_params.state); - - if (new_crtc_state->freesync_timing_changed) - DRM_DEBUG_KMS("VRR timing update: crtc=%u min=%u max=%u\n", - new_crtc_state->base.crtc->base.id, - vrr_params.adjust.v_total_min, - vrr_params.adjust.v_total_max); -} - -/* - * Executes flip - * - * Waits on all BO's fences and for proper vblank count - */ -static void amdgpu_dm_do_flip(struct drm_crtc *crtc, - struct drm_framebuffer *fb, - uint32_t target, - struct dc_state *state) -{ - unsigned long flags; - uint64_t timestamp_ns; - uint32_t target_vblank; - int r, vpos, hpos; - struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc); - struct amdgpu_framebuffer *afb = to_amdgpu_framebuffer(fb); - struct amdgpu_bo *abo = gem_to_amdgpu_bo(fb->obj[0]); - struct amdgpu_device *adev = crtc->dev->dev_private; - bool async_flip = (crtc->state->pageflip_flags & DRM_MODE_PAGE_FLIP_ASYNC) != 0; - struct dc_flip_addrs addr = { {0} }; - /* TODO eliminate or rename surface_update */ - struct dc_surface_update surface_updates[1] = { {0} }; - struct dc_stream_update stream_update = {0}; - struct dm_crtc_state *acrtc_state = to_dm_crtc_state(crtc->state); - struct dc_stream_status *stream_status; - struct dc_plane_state *surface; - - - /* Prepare wait for target vblank early - before the fence-waits */ - target_vblank = target - (uint32_t)drm_crtc_vblank_count(crtc) + - amdgpu_get_vblank_counter_kms(crtc->dev, acrtc->crtc_id); - - /* - * TODO This might fail and hence better not used, wait - * explicitly on fences instead - * and in general should be called for - * blocking commit to as per framework helpers - */ - r = amdgpu_bo_reserve(abo, true); - if (unlikely(r != 0)) { - DRM_ERROR("failed to reserve buffer before flip\n"); - WARN_ON(1); - } - - /* Wait for all fences on this FB */ - WARN_ON(reservation_object_wait_timeout_rcu(abo->tbo.resv, true, false, - MAX_SCHEDULE_TIMEOUT) < 0); - - amdgpu_bo_unreserve(abo); - - /* - * Wait until we're out of the vertical blank period before the one - * targeted by the flip - */ - while ((acrtc->enabled && - (amdgpu_display_get_crtc_scanoutpos(adev->ddev, acrtc->crtc_id, - 0, &vpos, &hpos, NULL, - NULL, &crtc->hwmode) - & (DRM_SCANOUTPOS_VALID | DRM_SCANOUTPOS_IN_VBLANK)) == - (DRM_SCANOUTPOS_VALID | DRM_SCANOUTPOS_IN_VBLANK) && - (int)(target_vblank - - amdgpu_get_vblank_counter_kms(adev->ddev, acrtc->crtc_id)) > 0)) { - usleep_range(1000, 1100); - } - - /* Flip */ - spin_lock_irqsave(&crtc->dev->event_lock, flags); - - WARN_ON(acrtc->pflip_status != AMDGPU_FLIP_NONE); - WARN_ON(!acrtc_state->stream); - - addr.address.grph.addr.low_part = lower_32_bits(afb->address); - addr.address.grph.addr.high_part = upper_32_bits(afb->address); - addr.flip_immediate = async_flip; - - timestamp_ns = ktime_get_ns(); - addr.flip_timestamp_in_us = div_u64(timestamp_ns, 1000); - - - if (acrtc->base.state->event) - prepare_flip_isr(acrtc); - - spin_unlock_irqrestore(&crtc->dev->event_lock, flags); - - stream_status = dc_stream_get_status(acrtc_state->stream); - if (!stream_status) { - DRM_ERROR("No stream status for CRTC: id=%d\n", - acrtc->crtc_id); - return; - } - - surface = stream_status->plane_states[0]; - surface_updates->surface = surface; - - if (!surface) { - DRM_ERROR("No surface for CRTC: id=%d\n", - acrtc->crtc_id); - return; - } - surface_updates->flip_addr = &addr; - - if (acrtc_state->stream) { - update_freesync_state_on_stream( - &adev->dm, - acrtc_state, - acrtc_state->stream, - surface, - addr.flip_timestamp_in_us); - - if (acrtc_state->freesync_timing_changed) - stream_update.adjust = - &acrtc_state->stream->adjust; - - if (acrtc_state->freesync_vrr_info_changed) - stream_update.vrr_infopacket = - &acrtc_state->stream->vrr_infopacket; - } - - /* Update surface timing information. */ - surface->time.time_elapsed_in_us[surface->time.index] = - addr.flip_timestamp_in_us - surface->time.prev_update_time_in_us; - surface->time.prev_update_time_in_us = addr.flip_timestamp_in_us; - surface->time.index++; - if (surface->time.index >= DC_PLANE_UPDATE_TIMES_MAX) - surface->time.index = 0; - - mutex_lock(&adev->dm.dc_lock); - - dc_commit_updates_for_stream(adev->dm.dc, - surface_updates, - 1, - acrtc_state->stream, - &stream_update, - &surface_updates->surface, - state); - mutex_unlock(&adev->dm.dc_lock); - - DRM_DEBUG_DRIVER("%s Flipping to hi: 0x%x, low: 0x%x \n", - __func__, - addr.address.grph.addr.high_part, - addr.address.grph.addr.low_part); -} - -/* - * TODO this whole function needs to go - * - * dc_surface_update is needlessly complex. See if we can just replace this - * with a dc_plane_state and follow the atomic model a bit more closely here. - */ -static bool commit_planes_to_stream( - struct amdgpu_display_manager *dm, - struct dc *dc, - struct dc_plane_state **plane_states, - uint8_t new_plane_count, - struct dm_crtc_state *dm_new_crtc_state, - struct dm_crtc_state *dm_old_crtc_state, - struct dc_state *state) -{ - /* no need to dynamically allocate this. it's pretty small */ - struct dc_surface_update updates[MAX_SURFACES]; - struct dc_flip_addrs *flip_addr; - struct dc_plane_info *plane_info; - struct dc_scaling_info *scaling_info; - int i; - struct dc_stream_state *dc_stream = dm_new_crtc_state->stream; - struct dc_stream_update *stream_update = - kzalloc(sizeof(struct dc_stream_update), GFP_KERNEL); - unsigned int abm_level; - - if (!stream_update) { - BREAK_TO_DEBUGGER(); - return false; - } - - flip_addr = kcalloc(MAX_SURFACES, sizeof(struct dc_flip_addrs), - GFP_KERNEL); - plane_info = kcalloc(MAX_SURFACES, sizeof(struct dc_plane_info), - GFP_KERNEL); - scaling_info = kcalloc(MAX_SURFACES, sizeof(struct dc_scaling_info), - GFP_KERNEL); - - if (!flip_addr || !plane_info || !scaling_info) { - kfree(flip_addr); - kfree(plane_info); - kfree(scaling_info); - kfree(stream_update); - return false; - } - - memset(updates, 0, sizeof(updates)); - - stream_update->src = dc_stream->src; - stream_update->dst = dc_stream->dst; - stream_update->out_transfer_func = dc_stream->out_transfer_func; - - if (dm_new_crtc_state->abm_level != dm_old_crtc_state->abm_level) { - abm_level = dm_new_crtc_state->abm_level; - stream_update->abm_level = &abm_level; - } - - for (i = 0; i < new_plane_count; i++) { - updates[i].surface = plane_states[i]; - updates[i].gamma = - (struct dc_gamma *)plane_states[i]->gamma_correction; - updates[i].in_transfer_func = plane_states[i]->in_transfer_func; - flip_addr[i].address = plane_states[i]->address; - flip_addr[i].flip_immediate = plane_states[i]->flip_immediate; - plane_info[i].color_space = plane_states[i]->color_space; - plane_info[i].format = plane_states[i]->format; - plane_info[i].plane_size = plane_states[i]->plane_size; - plane_info[i].rotation = plane_states[i]->rotation; - plane_info[i].horizontal_mirror = plane_states[i]->horizontal_mirror; - plane_info[i].stereo_format = plane_states[i]->stereo_format; - plane_info[i].tiling_info = plane_states[i]->tiling_info; - plane_info[i].visible = plane_states[i]->visible; - plane_info[i].per_pixel_alpha = plane_states[i]->per_pixel_alpha; - plane_info[i].dcc = plane_states[i]->dcc; - scaling_info[i].scaling_quality = plane_states[i]->scaling_quality; - scaling_info[i].src_rect = plane_states[i]->src_rect; - scaling_info[i].dst_rect = plane_states[i]->dst_rect; - scaling_info[i].clip_rect = plane_states[i]->clip_rect; - - updates[i].flip_addr = &flip_addr[i]; - updates[i].plane_info = &plane_info[i]; - updates[i].scaling_info = &scaling_info[i]; - } - - mutex_lock(&dm->dc_lock); - dc_commit_updates_for_stream( - dc, - updates, - new_plane_count, - dc_stream, stream_update, plane_states, state); - mutex_unlock(&dm->dc_lock); - - kfree(flip_addr); - kfree(plane_info); - kfree(scaling_info); - kfree(stream_update); - return true; } static void amdgpu_dm_commit_planes(struct drm_atomic_state *state, @@ -4801,26 +4617,51 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state, struct drm_crtc *pcrtc, bool *wait_for_vblank) { - uint32_t i; + uint32_t i, r; + uint64_t timestamp_ns; struct drm_plane *plane; struct drm_plane_state *old_plane_state, *new_plane_state; - struct dc_stream_state *dc_stream_attach; - struct dc_plane_state *plane_states_constructed[MAX_SURFACES]; struct amdgpu_crtc *acrtc_attach = to_amdgpu_crtc(pcrtc); struct drm_crtc_state *new_pcrtc_state = drm_atomic_get_new_crtc_state(state, pcrtc); struct dm_crtc_state *acrtc_state = to_dm_crtc_state(new_pcrtc_state); struct dm_crtc_state *dm_old_crtc_state = to_dm_crtc_state(drm_atomic_get_old_crtc_state(state, pcrtc)); - int planes_count = 0; + int flip_count = 0, planes_count = 0, vpos, hpos; unsigned long flags; + struct amdgpu_bo *abo; + uint64_t tiling_flags, dcc_address; + uint32_t target, target_vblank; + + struct { + struct dc_surface_update surface_updates[MAX_SURFACES]; + struct dc_flip_addrs flip_addrs[MAX_SURFACES]; + struct dc_stream_update stream_update; + } *flip; + + struct { + struct dc_surface_update surface_updates[MAX_SURFACES]; + struct dc_plane_info plane_infos[MAX_SURFACES]; + struct dc_scaling_info scaling_infos[MAX_SURFACES]; + struct dc_stream_update stream_update; + } *full; + + flip = kzalloc(sizeof(*flip), GFP_KERNEL); + full = kzalloc(sizeof(*full), GFP_KERNEL); + + if (!flip || !full) { + dm_error("Failed to allocate update bundles\n"); + goto cleanup; + } /* update planes when needed */ for_each_oldnew_plane_in_state(state, plane, old_plane_state, new_plane_state, i) { struct drm_crtc *crtc = new_plane_state->crtc; struct drm_crtc_state *new_crtc_state; struct drm_framebuffer *fb = new_plane_state->fb; + struct amdgpu_framebuffer *afb = to_amdgpu_framebuffer(fb); bool pflip_needed; + struct dc_plane_state *dc_plane; struct dm_plane_state *dm_new_plane_state = to_dm_plane_state(new_plane_state); if (plane->type == DRM_PLANE_TYPE_CURSOR) { @@ -4835,73 +4676,193 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state, if (!new_crtc_state->active) continue; - pflip_needed = !state->allow_modeset; - - spin_lock_irqsave(&crtc->dev->event_lock, flags); - if (acrtc_attach->pflip_status != AMDGPU_FLIP_NONE) { - DRM_ERROR("%s: acrtc %d, already busy\n", - __func__, - acrtc_attach->crtc_id); - /* In commit tail framework this cannot happen */ - WARN_ON(1); - } - spin_unlock_irqrestore(&crtc->dev->event_lock, flags); - - if (!pflip_needed || plane->type == DRM_PLANE_TYPE_OVERLAY) { - WARN_ON(!dm_new_plane_state->dc_state); - - plane_states_constructed[planes_count] = dm_new_plane_state->dc_state; + pflip_needed = old_plane_state->fb && + old_plane_state->fb != new_plane_state->fb; - dc_stream_attach = acrtc_state->stream; - planes_count++; + dc_plane = dm_new_plane_state->dc_state; - } else if (new_crtc_state->planes_changed) { - /* Assume even ONE crtc with immediate flip means + if (pflip_needed) { + /* + * Assume even ONE crtc with immediate flip means * entire can't wait for VBLANK * TODO Check if it's correct */ - *wait_for_vblank = - new_pcrtc_state->pageflip_flags & DRM_MODE_PAGE_FLIP_ASYNC ? - false : true; - - /* TODO: Needs rework for multiplane flip */ - if (plane->type == DRM_PLANE_TYPE_PRIMARY) - drm_crtc_vblank_get(crtc); - - amdgpu_dm_do_flip( - crtc, - fb, - (uint32_t)drm_crtc_vblank_count(crtc) + *wait_for_vblank, - dc_state); + if (new_pcrtc_state->pageflip_flags & DRM_MODE_PAGE_FLIP_ASYNC) + *wait_for_vblank = false; + + /* + * TODO This might fail and hence better not used, wait + * explicitly on fences instead + * and in general should be called for + * blocking commit to as per framework helpers + */ + abo = gem_to_amdgpu_bo(fb->obj[0]); + r = amdgpu_bo_reserve(abo, true); + if (unlikely(r != 0)) { + DRM_ERROR("failed to reserve buffer before flip\n"); + WARN_ON(1); + } + + /* Wait for all fences on this FB */ + WARN_ON(reservation_object_wait_timeout_rcu(abo->tbo.resv, true, false, + MAX_SCHEDULE_TIMEOUT) < 0); + + amdgpu_bo_get_tiling_flags(abo, &tiling_flags); + + amdgpu_bo_unreserve(abo); + + flip->flip_addrs[flip_count].address.grph.addr.low_part = lower_32_bits(afb->address); + flip->flip_addrs[flip_count].address.grph.addr.high_part = upper_32_bits(afb->address); + + dcc_address = get_dcc_address(afb->address, tiling_flags); + flip->flip_addrs[flip_count].address.grph.meta_addr.low_part = lower_32_bits(dcc_address); + flip->flip_addrs[flip_count].address.grph.meta_addr.high_part = upper_32_bits(dcc_address); + + flip->flip_addrs[flip_count].flip_immediate = + (crtc->state->pageflip_flags & DRM_MODE_PAGE_FLIP_ASYNC) != 0; + + timestamp_ns = ktime_get_ns(); + flip->flip_addrs[flip_count].flip_timestamp_in_us = div_u64(timestamp_ns, 1000); + flip->surface_updates[flip_count].flip_addr = &flip->flip_addrs[flip_count]; + flip->surface_updates[flip_count].surface = dc_plane; + + if (!flip->surface_updates[flip_count].surface) { + DRM_ERROR("No surface for CRTC: id=%d\n", + acrtc_attach->crtc_id); + continue; + } + + if (plane == pcrtc->primary) + update_freesync_state_on_stream( + dm, + acrtc_state, + acrtc_state->stream, + dc_plane, + flip->flip_addrs[flip_count].flip_timestamp_in_us); + + DRM_DEBUG_DRIVER("%s Flipping to hi: 0x%x, low: 0x%x\n", + __func__, + flip->flip_addrs[flip_count].address.grph.addr.high_part, + flip->flip_addrs[flip_count].address.grph.addr.low_part); + + flip_count += 1; } + full->surface_updates[planes_count].surface = dc_plane; + if (new_pcrtc_state->color_mgmt_changed) { + full->surface_updates[planes_count].gamma = dc_plane->gamma_correction; + full->surface_updates[planes_count].in_transfer_func = dc_plane->in_transfer_func; + } + + + full->scaling_infos[planes_count].scaling_quality = dc_plane->scaling_quality; + full->scaling_infos[planes_count].src_rect = dc_plane->src_rect; + full->scaling_infos[planes_count].dst_rect = dc_plane->dst_rect; + full->scaling_infos[planes_count].clip_rect = dc_plane->clip_rect; + full->surface_updates[planes_count].scaling_info = &full->scaling_infos[planes_count]; + + + full->plane_infos[planes_count].color_space = dc_plane->color_space; + full->plane_infos[planes_count].format = dc_plane->format; + full->plane_infos[planes_count].plane_size = dc_plane->plane_size; + full->plane_infos[planes_count].rotation = dc_plane->rotation; + full->plane_infos[planes_count].horizontal_mirror = dc_plane->horizontal_mirror; + full->plane_infos[planes_count].stereo_format = dc_plane->stereo_format; + full->plane_infos[planes_count].tiling_info = dc_plane->tiling_info; + full->plane_infos[planes_count].visible = dc_plane->visible; + full->plane_infos[planes_count].per_pixel_alpha = dc_plane->per_pixel_alpha; + full->plane_infos[planes_count].dcc = dc_plane->dcc; + full->surface_updates[planes_count].plane_info = &full->plane_infos[planes_count]; + + planes_count += 1; + } - if (planes_count) { - unsigned long flags; + /* + * TODO: For proper atomic behaviour, we should be calling into DC once with + * all the changes. However, DC refuses to do pageflips and non-pageflip + * changes in the same call. Change DC to respect atomic behaviour, + * hopefully eliminating dc_*_update structs in their entirety. + */ + if (flip_count) { + target = (uint32_t)drm_crtc_vblank_count(pcrtc) + *wait_for_vblank; + /* Prepare wait for target vblank early - before the fence-waits */ + target_vblank = target - (uint32_t)drm_crtc_vblank_count(pcrtc) + + amdgpu_get_vblank_counter_kms(pcrtc->dev, acrtc_attach->crtc_id); - if (new_pcrtc_state->event) { + /* + * Wait until we're out of the vertical blank period before the one + * targeted by the flip + */ + while ((acrtc_attach->enabled && + (amdgpu_display_get_crtc_scanoutpos(dm->ddev, acrtc_attach->crtc_id, + 0, &vpos, &hpos, NULL, + NULL, &pcrtc->hwmode) + & (DRM_SCANOUTPOS_VALID | DRM_SCANOUTPOS_IN_VBLANK)) == + (DRM_SCANOUTPOS_VALID | DRM_SCANOUTPOS_IN_VBLANK) && + (int)(target_vblank - + amdgpu_get_vblank_counter_kms(dm->ddev, acrtc_attach->crtc_id)) > 0)) { + usleep_range(1000, 1100); + } + if (acrtc_attach->base.state->event) { drm_crtc_vblank_get(pcrtc); spin_lock_irqsave(&pcrtc->dev->event_lock, flags); + + WARN_ON(acrtc_attach->pflip_status != AMDGPU_FLIP_NONE); prepare_flip_isr(acrtc_attach); + spin_unlock_irqrestore(&pcrtc->dev->event_lock, flags); } - dc_stream_attach->abm_level = acrtc_state->abm_level; + if (acrtc_state->stream) { - if (false == commit_planes_to_stream(dm, - dm->dc, - plane_states_constructed, - planes_count, - acrtc_state, - dm_old_crtc_state, - dc_state)) - dm_error("%s: Failed to attach plane!\n", __func__); - } else { - /*TODO BUG Here should go disable planes on CRTC. */ + if (acrtc_state->freesync_timing_changed) + flip->stream_update.adjust = + &acrtc_state->stream->adjust; + + if (acrtc_state->freesync_vrr_info_changed) + flip->stream_update.vrr_infopacket = + &acrtc_state->stream->vrr_infopacket; + } + + mutex_lock(&dm->dc_lock); + dc_commit_updates_for_stream(dm->dc, + flip->surface_updates, + flip_count, + acrtc_state->stream, + &flip->stream_update, + dc_state); + mutex_unlock(&dm->dc_lock); } + + if (planes_count) { + if (new_pcrtc_state->mode_changed) { + full->stream_update.src = acrtc_state->stream->src; + full->stream_update.dst = acrtc_state->stream->dst; + } + + if (new_pcrtc_state->color_mgmt_changed) + full->stream_update.out_transfer_func = acrtc_state->stream->out_transfer_func; + + acrtc_state->stream->abm_level = acrtc_state->abm_level; + if (acrtc_state->abm_level != dm_old_crtc_state->abm_level) + full->stream_update.abm_level = &acrtc_state->abm_level; + + mutex_lock(&dm->dc_lock); + dc_commit_updates_for_stream(dm->dc, + full->surface_updates, + planes_count, + acrtc_state->stream, + &full->stream_update, + dc_state); + mutex_unlock(&dm->dc_lock); + } + +cleanup: + kfree(flip); + kfree(full); } /* @@ -4915,7 +4876,8 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state, static void amdgpu_dm_crtc_copy_transient_flags(struct drm_crtc_state *crtc_state, struct dc_stream_state *stream_state) { - stream_state->mode_changed = crtc_state->mode_changed; + stream_state->mode_changed = + crtc_state->mode_changed || crtc_state->active_changed; } static int amdgpu_dm_atomic_commit(struct drm_device *dev, @@ -4936,10 +4898,25 @@ static int amdgpu_dm_atomic_commit(struct drm_device *dev, */ for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) { struct dm_crtc_state *dm_old_crtc_state = to_dm_crtc_state(old_crtc_state); + struct dm_crtc_state *dm_new_crtc_state = to_dm_crtc_state(new_crtc_state); struct amdgpu_crtc *acrtc = to_amdgpu_crtc(crtc); - if (drm_atomic_crtc_needs_modeset(new_crtc_state) && dm_old_crtc_state->stream) + if (drm_atomic_crtc_needs_modeset(new_crtc_state) + && dm_old_crtc_state->stream) { + /* + * If the stream is removed and CRC capture was + * enabled on the CRTC the extra vblank reference + * needs to be dropped since CRC capture will be + * disabled. + */ + if (!dm_new_crtc_state->stream + && dm_new_crtc_state->crc_enabled) { + drm_crtc_vblank_put(crtc); + dm_new_crtc_state->crc_enabled = false; + } + manage_dm_interrupts(adev, acrtc, false); + } } /* * Add check here for SoC's that support hardware cursor plane, to @@ -5077,8 +5054,8 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) dc_stream_get_status(dm_new_crtc_state->stream); if (!status) - status = dc_state_get_stream_status(dc_state, - dm_new_crtc_state->stream); + status = dc_stream_get_status_from_state(dc_state, + dm_new_crtc_state->stream); if (!status) DC_ERR("got no status for stream %p on acrtc%p\n", dm_new_crtc_state->stream, acrtc); @@ -5087,13 +5064,18 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) } } - /* Handle scaling, underscan, and abm changes*/ + /* Handle connector state changes */ for_each_oldnew_connector_in_state(state, connector, old_con_state, new_con_state, i) { struct dm_connector_state *dm_new_con_state = to_dm_connector_state(new_con_state); struct dm_connector_state *dm_old_con_state = to_dm_connector_state(old_con_state); struct amdgpu_crtc *acrtc = to_amdgpu_crtc(dm_new_con_state->base.crtc); + struct dc_surface_update dummy_updates[MAX_SURFACES]; + struct dc_stream_update stream_update; struct dc_stream_status *status = NULL; + memset(&dummy_updates, 0, sizeof(dummy_updates)); + memset(&stream_update, 0, sizeof(stream_update)); + if (acrtc) { new_crtc_state = drm_atomic_get_new_crtc_state(state, &acrtc->base); old_crtc_state = drm_atomic_get_old_crtc_state(state, &acrtc->base); @@ -5103,37 +5085,48 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) if (!acrtc || drm_atomic_crtc_needs_modeset(new_crtc_state)) continue; - dm_new_crtc_state = to_dm_crtc_state(new_crtc_state); dm_old_crtc_state = to_dm_crtc_state(old_crtc_state); - /* Skip anything that is not scaling or underscan changes */ if (!is_scaling_state_different(dm_new_con_state, dm_old_con_state) && (dm_new_crtc_state->abm_level == dm_old_crtc_state->abm_level)) continue; - update_stream_scaling_settings(&dm_new_con_state->base.crtc->mode, - dm_new_con_state, (struct dc_stream_state *)dm_new_crtc_state->stream); + if (is_scaling_state_different(dm_new_con_state, dm_old_con_state)) { + update_stream_scaling_settings(&dm_new_con_state->base.crtc->mode, + dm_new_con_state, (struct dc_stream_state *)dm_new_crtc_state->stream); - if (!dm_new_crtc_state->stream) - continue; + stream_update.src = dm_new_crtc_state->stream->src; + stream_update.dst = dm_new_crtc_state->stream->dst; + } + + if (dm_new_crtc_state->abm_level != dm_old_crtc_state->abm_level) { + dm_new_crtc_state->stream->abm_level = dm_new_crtc_state->abm_level; + + stream_update.abm_level = &dm_new_crtc_state->abm_level; + } status = dc_stream_get_status(dm_new_crtc_state->stream); WARN_ON(!status); WARN_ON(!status->plane_count); - dm_new_crtc_state->stream->abm_level = dm_new_crtc_state->abm_level; + /* + * TODO: DC refuses to perform stream updates without a dc_surface_update. + * Here we create an empty update on each plane. + * To fix this, DC should permit updating only stream properties. + */ + for (j = 0; j < status->plane_count; j++) + dummy_updates[j].surface = status->plane_states[0]; - /*TODO How it works with MPO ?*/ - if (!commit_planes_to_stream( - dm, - dm->dc, - status->plane_states, - status->plane_count, - dm_new_crtc_state, - to_dm_crtc_state(old_crtc_state), - dc_state)) - dm_error("%s: Failed to update stream scaling!\n", __func__); + + mutex_lock(&dm->dc_lock); + dc_commit_updates_for_stream(dm->dc, + dummy_updates, + status->plane_count, + dm_new_crtc_state->stream, + &stream_update, + dc_state); + mutex_unlock(&dm->dc_lock); } for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, @@ -5158,6 +5151,12 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) continue; manage_dm_interrupts(adev, acrtc, true); + +#ifdef CONFIG_DEBUG_FS + /* The stream has changed so CRC capture needs to re-enabled. */ + if (dm_new_crtc_state->crc_enabled) + amdgpu_dm_crtc_set_crc_source(crtc, "auto"); +#endif } /* update planes when needed per crtc*/ @@ -5184,18 +5183,12 @@ static void amdgpu_dm_atomic_commit_tail(struct drm_atomic_state *state) } spin_unlock_irqrestore(&adev->ddev->event_lock, flags); + /* Signal HW programming completion */ + drm_atomic_helper_commit_hw_done(state); if (wait_for_vblank) drm_atomic_helper_wait_for_flip_done(dev, state); - /* - * FIXME: - * Delay hw_done() until flip_done() is signaled. This is to block - * another commit from freeing the CRTC state while we're still - * waiting on flip_done. - */ - drm_atomic_helper_commit_hw_done(state); - drm_atomic_helper_cleanup_planes(dev, state); /* @@ -5359,10 +5352,13 @@ static void get_freesync_config_for_crtc( struct mod_freesync_config config = {0}; struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(new_con_state->base.connector); + struct drm_display_mode *mode = &new_crtc_state->base.mode; - new_crtc_state->vrr_supported = new_con_state->freesync_capable; + new_crtc_state->vrr_supported = new_con_state->freesync_capable && + aconnector->min_vfreq <= drm_mode_vrefresh(mode); - if (new_con_state->freesync_capable) { + if (new_crtc_state->vrr_supported) { + new_crtc_state->stream->ignore_msa_timing_param = true; config.state = new_crtc_state->base.vrr_enabled ? VRR_STATE_ACTIVE_VARIABLE : VRR_STATE_INACTIVE; @@ -5388,15 +5384,15 @@ static void reset_freesync_config_for_crtc( sizeof(new_crtc_state->vrr_infopacket)); } -static int dm_update_crtcs_state(struct amdgpu_display_manager *dm, - struct drm_atomic_state *state, - bool enable, - bool *lock_and_validation_needed) +static int dm_update_crtc_state(struct amdgpu_display_manager *dm, + struct drm_atomic_state *state, + struct drm_crtc *crtc, + struct drm_crtc_state *old_crtc_state, + struct drm_crtc_state *new_crtc_state, + bool enable, + bool *lock_and_validation_needed) { struct dm_atomic_state *dm_state = NULL; - struct drm_crtc *crtc; - struct drm_crtc_state *old_crtc_state, *new_crtc_state; - int i; struct dm_crtc_state *dm_old_crtc_state, *dm_new_crtc_state; struct dc_stream_state *new_stream; int ret = 0; @@ -5405,200 +5401,203 @@ static int dm_update_crtcs_state(struct amdgpu_display_manager *dm, * TODO Move this code into dm_crtc_atomic_check once we get rid of dc_validation_set * update changed items */ - for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) { - struct amdgpu_crtc *acrtc = NULL; - struct amdgpu_dm_connector *aconnector = NULL; - struct drm_connector_state *drm_new_conn_state = NULL, *drm_old_conn_state = NULL; - struct dm_connector_state *dm_new_conn_state = NULL, *dm_old_conn_state = NULL; - struct drm_plane_state *new_plane_state = NULL; + struct amdgpu_crtc *acrtc = NULL; + struct amdgpu_dm_connector *aconnector = NULL; + struct drm_connector_state *drm_new_conn_state = NULL, *drm_old_conn_state = NULL; + struct dm_connector_state *dm_new_conn_state = NULL, *dm_old_conn_state = NULL; + struct drm_plane_state *new_plane_state = NULL; - new_stream = NULL; + new_stream = NULL; - dm_old_crtc_state = to_dm_crtc_state(old_crtc_state); - dm_new_crtc_state = to_dm_crtc_state(new_crtc_state); - acrtc = to_amdgpu_crtc(crtc); + dm_old_crtc_state = to_dm_crtc_state(old_crtc_state); + dm_new_crtc_state = to_dm_crtc_state(new_crtc_state); + acrtc = to_amdgpu_crtc(crtc); - new_plane_state = drm_atomic_get_new_plane_state(state, new_crtc_state->crtc->primary); + new_plane_state = drm_atomic_get_new_plane_state(state, new_crtc_state->crtc->primary); - if (new_crtc_state->enable && new_plane_state && !new_plane_state->fb) { - ret = -EINVAL; - goto fail; - } + if (new_crtc_state->enable && new_plane_state && !new_plane_state->fb) { + ret = -EINVAL; + goto fail; + } - aconnector = amdgpu_dm_find_first_crtc_matching_connector(state, crtc); + aconnector = amdgpu_dm_find_first_crtc_matching_connector(state, crtc); - /* TODO This hack should go away */ - if (aconnector && enable) { - /* Make sure fake sink is created in plug-in scenario */ - drm_new_conn_state = drm_atomic_get_new_connector_state(state, - &aconnector->base); - drm_old_conn_state = drm_atomic_get_old_connector_state(state, - &aconnector->base); + /* TODO This hack should go away */ + if (aconnector && enable) { + /* Make sure fake sink is created in plug-in scenario */ + drm_new_conn_state = drm_atomic_get_new_connector_state(state, + &aconnector->base); + drm_old_conn_state = drm_atomic_get_old_connector_state(state, + &aconnector->base); - if (IS_ERR(drm_new_conn_state)) { - ret = PTR_ERR_OR_ZERO(drm_new_conn_state); - break; - } - - dm_new_conn_state = to_dm_connector_state(drm_new_conn_state); - dm_old_conn_state = to_dm_connector_state(drm_old_conn_state); + if (IS_ERR(drm_new_conn_state)) { + ret = PTR_ERR_OR_ZERO(drm_new_conn_state); + goto fail; + } - new_stream = create_stream_for_sink(aconnector, - &new_crtc_state->mode, - dm_new_conn_state, - dm_old_crtc_state->stream); + dm_new_conn_state = to_dm_connector_state(drm_new_conn_state); + dm_old_conn_state = to_dm_connector_state(drm_old_conn_state); - /* - * we can have no stream on ACTION_SET if a display - * was disconnected during S3, in this case it is not an - * error, the OS will be updated after detection, and - * will do the right thing on next atomic commit - */ + if (!drm_atomic_crtc_needs_modeset(new_crtc_state)) + goto skip_modeset; - if (!new_stream) { - DRM_DEBUG_DRIVER("%s: Failed to create new stream for crtc %d\n", - __func__, acrtc->base.base.id); - break; - } + new_stream = create_stream_for_sink(aconnector, + &new_crtc_state->mode, + dm_new_conn_state, + dm_old_crtc_state->stream); - dm_new_crtc_state->abm_level = dm_new_conn_state->abm_level; + /* + * we can have no stream on ACTION_SET if a display + * was disconnected during S3, in this case it is not an + * error, the OS will be updated after detection, and + * will do the right thing on next atomic commit + */ - if (dc_is_stream_unchanged(new_stream, dm_old_crtc_state->stream) && - dc_is_stream_scaling_unchanged(new_stream, dm_old_crtc_state->stream)) { - new_crtc_state->mode_changed = false; - DRM_DEBUG_DRIVER("Mode change not required, setting mode_changed to %d", - new_crtc_state->mode_changed); - } + if (!new_stream) { + DRM_DEBUG_DRIVER("%s: Failed to create new stream for crtc %d\n", + __func__, acrtc->base.base.id); + ret = -ENOMEM; + goto fail; } - if (!drm_atomic_crtc_needs_modeset(new_crtc_state)) - goto next_crtc; + dm_new_crtc_state->abm_level = dm_new_conn_state->abm_level; - DRM_DEBUG_DRIVER( - "amdgpu_crtc id:%d crtc_state_flags: enable:%d, active:%d, " - "planes_changed:%d, mode_changed:%d,active_changed:%d," - "connectors_changed:%d\n", - acrtc->crtc_id, - new_crtc_state->enable, - new_crtc_state->active, - new_crtc_state->planes_changed, - new_crtc_state->mode_changed, - new_crtc_state->active_changed, - new_crtc_state->connectors_changed); + if (dc_is_stream_unchanged(new_stream, dm_old_crtc_state->stream) && + dc_is_stream_scaling_unchanged(new_stream, dm_old_crtc_state->stream)) { + new_crtc_state->mode_changed = false; + DRM_DEBUG_DRIVER("Mode change not required, setting mode_changed to %d", + new_crtc_state->mode_changed); + } + } - /* Remove stream for any changed/disabled CRTC */ - if (!enable) { + /* mode_changed flag may get updated above, need to check again */ + if (!drm_atomic_crtc_needs_modeset(new_crtc_state)) + goto skip_modeset; - if (!dm_old_crtc_state->stream) - goto next_crtc; + DRM_DEBUG_DRIVER( + "amdgpu_crtc id:%d crtc_state_flags: enable:%d, active:%d, " + "planes_changed:%d, mode_changed:%d,active_changed:%d," + "connectors_changed:%d\n", + acrtc->crtc_id, + new_crtc_state->enable, + new_crtc_state->active, + new_crtc_state->planes_changed, + new_crtc_state->mode_changed, + new_crtc_state->active_changed, + new_crtc_state->connectors_changed); - ret = dm_atomic_get_state(state, &dm_state); - if (ret) - goto fail; + /* Remove stream for any changed/disabled CRTC */ + if (!enable) { - DRM_DEBUG_DRIVER("Disabling DRM crtc: %d\n", - crtc->base.id); + if (!dm_old_crtc_state->stream) + goto skip_modeset; - /* i.e. reset mode */ - if (dc_remove_stream_from_ctx( - dm->dc, - dm_state->context, - dm_old_crtc_state->stream) != DC_OK) { - ret = -EINVAL; - goto fail; - } + ret = dm_atomic_get_state(state, &dm_state); + if (ret) + goto fail; - dc_stream_release(dm_old_crtc_state->stream); - dm_new_crtc_state->stream = NULL; + DRM_DEBUG_DRIVER("Disabling DRM crtc: %d\n", + crtc->base.id); + + /* i.e. reset mode */ + if (dc_remove_stream_from_ctx( + dm->dc, + dm_state->context, + dm_old_crtc_state->stream) != DC_OK) { + ret = -EINVAL; + goto fail; + } - reset_freesync_config_for_crtc(dm_new_crtc_state); + dc_stream_release(dm_old_crtc_state->stream); + dm_new_crtc_state->stream = NULL; - *lock_and_validation_needed = true; + reset_freesync_config_for_crtc(dm_new_crtc_state); - } else {/* Add stream for any updated/enabled CRTC */ - /* - * Quick fix to prevent NULL pointer on new_stream when - * added MST connectors not found in existing crtc_state in the chained mode - * TODO: need to dig out the root cause of that - */ - if (!aconnector || (!aconnector->dc_sink && aconnector->mst_port)) - goto next_crtc; + *lock_and_validation_needed = true; - if (modereset_required(new_crtc_state)) - goto next_crtc; + } else {/* Add stream for any updated/enabled CRTC */ + /* + * Quick fix to prevent NULL pointer on new_stream when + * added MST connectors not found in existing crtc_state in the chained mode + * TODO: need to dig out the root cause of that + */ + if (!aconnector || (!aconnector->dc_sink && aconnector->mst_port)) + goto skip_modeset; - if (modeset_required(new_crtc_state, new_stream, - dm_old_crtc_state->stream)) { + if (modereset_required(new_crtc_state)) + goto skip_modeset; - WARN_ON(dm_new_crtc_state->stream); + if (modeset_required(new_crtc_state, new_stream, + dm_old_crtc_state->stream)) { - ret = dm_atomic_get_state(state, &dm_state); - if (ret) - goto fail; + WARN_ON(dm_new_crtc_state->stream); - dm_new_crtc_state->stream = new_stream; + ret = dm_atomic_get_state(state, &dm_state); + if (ret) + goto fail; - dc_stream_retain(new_stream); + dm_new_crtc_state->stream = new_stream; - DRM_DEBUG_DRIVER("Enabling DRM crtc: %d\n", - crtc->base.id); + dc_stream_retain(new_stream); - if (dc_add_stream_to_ctx( - dm->dc, - dm_state->context, - dm_new_crtc_state->stream) != DC_OK) { - ret = -EINVAL; - goto fail; - } + DRM_DEBUG_DRIVER("Enabling DRM crtc: %d\n", + crtc->base.id); - *lock_and_validation_needed = true; + if (dc_add_stream_to_ctx( + dm->dc, + dm_state->context, + dm_new_crtc_state->stream) != DC_OK) { + ret = -EINVAL; + goto fail; } - } -next_crtc: - /* Release extra reference */ - if (new_stream) - dc_stream_release(new_stream); + *lock_and_validation_needed = true; + } + } - /* - * We want to do dc stream updates that do not require a - * full modeset below. - */ - if (!(enable && aconnector && new_crtc_state->enable && - new_crtc_state->active)) - continue; - /* - * Given above conditions, the dc state cannot be NULL because: - * 1. We're in the process of enabling CRTCs (just been added - * to the dc context, or already is on the context) - * 2. Has a valid connector attached, and - * 3. Is currently active and enabled. - * => The dc stream state currently exists. - */ - BUG_ON(dm_new_crtc_state->stream == NULL); +skip_modeset: + /* Release extra reference */ + if (new_stream) + dc_stream_release(new_stream); - /* Scaling or underscan settings */ - if (is_scaling_state_different(dm_old_conn_state, dm_new_conn_state)) - update_stream_scaling_settings( - &new_crtc_state->mode, dm_new_conn_state, dm_new_crtc_state->stream); + /* + * We want to do dc stream updates that do not require a + * full modeset below. + */ + if (!(enable && aconnector && new_crtc_state->enable && + new_crtc_state->active)) + return 0; + /* + * Given above conditions, the dc state cannot be NULL because: + * 1. We're in the process of enabling CRTCs (just been added + * to the dc context, or already is on the context) + * 2. Has a valid connector attached, and + * 3. Is currently active and enabled. + * => The dc stream state currently exists. + */ + BUG_ON(dm_new_crtc_state->stream == NULL); - /* - * Color management settings. We also update color properties - * when a modeset is needed, to ensure it gets reprogrammed. - */ - if (dm_new_crtc_state->base.color_mgmt_changed || - drm_atomic_crtc_needs_modeset(new_crtc_state)) { - ret = amdgpu_dm_set_regamma_lut(dm_new_crtc_state); - if (ret) - goto fail; - amdgpu_dm_set_ctm(dm_new_crtc_state); - } + /* Scaling or underscan settings */ + if (is_scaling_state_different(dm_old_conn_state, dm_new_conn_state)) + update_stream_scaling_settings( + &new_crtc_state->mode, dm_new_conn_state, dm_new_crtc_state->stream); - /* Update Freesync settings. */ - get_freesync_config_for_crtc(dm_new_crtc_state, - dm_new_conn_state); + /* + * Color management settings. We also update color properties + * when a modeset is needed, to ensure it gets reprogrammed. + */ + if (dm_new_crtc_state->base.color_mgmt_changed || + drm_atomic_crtc_needs_modeset(new_crtc_state)) { + ret = amdgpu_dm_set_regamma_lut(dm_new_crtc_state); + if (ret) + goto fail; + amdgpu_dm_set_ctm(dm_new_crtc_state); } + /* Update Freesync settings. */ + get_freesync_config_for_crtc(dm_new_crtc_state, + dm_new_conn_state); + return ret; fail: @@ -5607,145 +5606,141 @@ fail: return ret; } -static int dm_update_planes_state(struct dc *dc, - struct drm_atomic_state *state, - bool enable, - bool *lock_and_validation_needed) +static int dm_update_plane_state(struct dc *dc, + struct drm_atomic_state *state, + struct drm_plane *plane, + struct drm_plane_state *old_plane_state, + struct drm_plane_state *new_plane_state, + bool enable, + bool *lock_and_validation_needed) { struct dm_atomic_state *dm_state = NULL; struct drm_crtc *new_plane_crtc, *old_plane_crtc; struct drm_crtc_state *old_crtc_state, *new_crtc_state; - struct drm_plane *plane; - struct drm_plane_state *old_plane_state, *new_plane_state; struct dm_crtc_state *dm_new_crtc_state, *dm_old_crtc_state; struct dm_plane_state *dm_new_plane_state, *dm_old_plane_state; - int i ; /* TODO return page_flip_needed() function */ bool pflip_needed = !state->allow_modeset; int ret = 0; - /* Add new planes, in reverse order as DC expectation */ - for_each_oldnew_plane_in_state_reverse(state, plane, old_plane_state, new_plane_state, i) { - new_plane_crtc = new_plane_state->crtc; - old_plane_crtc = old_plane_state->crtc; - dm_new_plane_state = to_dm_plane_state(new_plane_state); - dm_old_plane_state = to_dm_plane_state(old_plane_state); + new_plane_crtc = new_plane_state->crtc; + old_plane_crtc = old_plane_state->crtc; + dm_new_plane_state = to_dm_plane_state(new_plane_state); + dm_old_plane_state = to_dm_plane_state(old_plane_state); - /*TODO Implement atomic check for cursor plane */ - if (plane->type == DRM_PLANE_TYPE_CURSOR) - continue; + /*TODO Implement atomic check for cursor plane */ + if (plane->type == DRM_PLANE_TYPE_CURSOR) + return 0; - /* Remove any changed/removed planes */ - if (!enable) { - if (pflip_needed && - plane->type != DRM_PLANE_TYPE_OVERLAY) - continue; + /* Remove any changed/removed planes */ + if (!enable) { + if (pflip_needed && + plane->type != DRM_PLANE_TYPE_OVERLAY) + return 0; - if (!old_plane_crtc) - continue; + if (!old_plane_crtc) + return 0; - old_crtc_state = drm_atomic_get_old_crtc_state( - state, old_plane_crtc); - dm_old_crtc_state = to_dm_crtc_state(old_crtc_state); + old_crtc_state = drm_atomic_get_old_crtc_state( + state, old_plane_crtc); + dm_old_crtc_state = to_dm_crtc_state(old_crtc_state); - if (!dm_old_crtc_state->stream) - continue; + if (!dm_old_crtc_state->stream) + return 0; - DRM_DEBUG_ATOMIC("Disabling DRM plane: %d on DRM crtc %d\n", - plane->base.id, old_plane_crtc->base.id); + DRM_DEBUG_ATOMIC("Disabling DRM plane: %d on DRM crtc %d\n", + plane->base.id, old_plane_crtc->base.id); - ret = dm_atomic_get_state(state, &dm_state); - if (ret) - return ret; + ret = dm_atomic_get_state(state, &dm_state); + if (ret) + return ret; - if (!dc_remove_plane_from_context( - dc, - dm_old_crtc_state->stream, - dm_old_plane_state->dc_state, - dm_state->context)) { + if (!dc_remove_plane_from_context( + dc, + dm_old_crtc_state->stream, + dm_old_plane_state->dc_state, + dm_state->context)) { - ret = EINVAL; - return ret; - } + ret = EINVAL; + return ret; + } - dc_plane_state_release(dm_old_plane_state->dc_state); - dm_new_plane_state->dc_state = NULL; + dc_plane_state_release(dm_old_plane_state->dc_state); + dm_new_plane_state->dc_state = NULL; - *lock_and_validation_needed = true; + *lock_and_validation_needed = true; - } else { /* Add new planes */ - struct dc_plane_state *dc_new_plane_state; + } else { /* Add new planes */ + struct dc_plane_state *dc_new_plane_state; - if (drm_atomic_plane_disabling(plane->state, new_plane_state)) - continue; + if (drm_atomic_plane_disabling(plane->state, new_plane_state)) + return 0; - if (!new_plane_crtc) - continue; + if (!new_plane_crtc) + return 0; - new_crtc_state = drm_atomic_get_new_crtc_state(state, new_plane_crtc); - dm_new_crtc_state = to_dm_crtc_state(new_crtc_state); + new_crtc_state = drm_atomic_get_new_crtc_state(state, new_plane_crtc); + dm_new_crtc_state = to_dm_crtc_state(new_crtc_state); - if (!dm_new_crtc_state->stream) - continue; + if (!dm_new_crtc_state->stream) + return 0; - if (pflip_needed && - plane->type != DRM_PLANE_TYPE_OVERLAY) - continue; + if (pflip_needed && plane->type != DRM_PLANE_TYPE_OVERLAY) + return 0; - WARN_ON(dm_new_plane_state->dc_state); + WARN_ON(dm_new_plane_state->dc_state); - dc_new_plane_state = dc_create_plane_state(dc); - if (!dc_new_plane_state) - return -ENOMEM; + dc_new_plane_state = dc_create_plane_state(dc); + if (!dc_new_plane_state) + return -ENOMEM; - DRM_DEBUG_DRIVER("Enabling DRM plane: %d on DRM crtc %d\n", - plane->base.id, new_plane_crtc->base.id); + DRM_DEBUG_DRIVER("Enabling DRM plane: %d on DRM crtc %d\n", + plane->base.id, new_plane_crtc->base.id); - ret = fill_plane_attributes( - new_plane_crtc->dev->dev_private, - dc_new_plane_state, - new_plane_state, - new_crtc_state); - if (ret) { - dc_plane_state_release(dc_new_plane_state); - return ret; - } + ret = fill_plane_attributes( + new_plane_crtc->dev->dev_private, + dc_new_plane_state, + new_plane_state, + new_crtc_state); + if (ret) { + dc_plane_state_release(dc_new_plane_state); + return ret; + } - ret = dm_atomic_get_state(state, &dm_state); - if (ret) { - dc_plane_state_release(dc_new_plane_state); - return ret; - } + ret = dm_atomic_get_state(state, &dm_state); + if (ret) { + dc_plane_state_release(dc_new_plane_state); + return ret; + } - /* - * Any atomic check errors that occur after this will - * not need a release. The plane state will be attached - * to the stream, and therefore part of the atomic - * state. It'll be released when the atomic state is - * cleaned. - */ - if (!dc_add_plane_to_context( - dc, - dm_new_crtc_state->stream, - dc_new_plane_state, - dm_state->context)) { - - dc_plane_state_release(dc_new_plane_state); - return -EINVAL; - } + /* + * Any atomic check errors that occur after this will + * not need a release. The plane state will be attached + * to the stream, and therefore part of the atomic + * state. It'll be released when the atomic state is + * cleaned. + */ + if (!dc_add_plane_to_context( + dc, + dm_new_crtc_state->stream, + dc_new_plane_state, + dm_state->context)) { - dm_new_plane_state->dc_state = dc_new_plane_state; + dc_plane_state_release(dc_new_plane_state); + return -EINVAL; + } - /* Tell DC to do a full surface update every time there - * is a plane change. Inefficient, but works for now. - */ - dm_new_plane_state->dc_state->update_flags.bits.full_update = 1; + dm_new_plane_state->dc_state = dc_new_plane_state; - *lock_and_validation_needed = true; - } + /* Tell DC to do a full surface update every time there + * is a plane change. Inefficient, but works for now. + */ + dm_new_plane_state->dc_state->update_flags.bits.full_update = 1; + + *lock_and_validation_needed = true; } @@ -5769,11 +5764,13 @@ dm_determine_update_type_for_commit(struct dc *dc, struct dm_crtc_state *new_dm_crtc_state, *old_dm_crtc_state; struct dc_stream_status *status = NULL; - struct dc_surface_update *updates = kzalloc(MAX_SURFACES * sizeof(struct dc_surface_update), GFP_KERNEL); - struct dc_plane_state *surface = kzalloc(MAX_SURFACES * sizeof(struct dc_plane_state), GFP_KERNEL); - struct dc_stream_update stream_update; + struct dc_surface_update *updates; + struct dc_plane_state *surface; enum surface_update_type update_type = UPDATE_TYPE_FAST; + updates = kcalloc(MAX_SURFACES, sizeof(*updates), GFP_KERNEL); + surface = kcalloc(MAX_SURFACES, sizeof(*surface), GFP_KERNEL); + if (!updates || !surface) { DRM_ERROR("Plane or surface update failed to allocate"); /* Set type to FULL to avoid crashing in DC*/ @@ -5782,79 +5779,85 @@ dm_determine_update_type_for_commit(struct dc *dc, } for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) { + struct dc_stream_update stream_update = { 0 }; + new_dm_crtc_state = to_dm_crtc_state(new_crtc_state); old_dm_crtc_state = to_dm_crtc_state(old_crtc_state); num_plane = 0; - if (new_dm_crtc_state->stream) { - - for_each_oldnew_plane_in_state(state, plane, old_plane_state, new_plane_state, j) { - new_plane_crtc = new_plane_state->crtc; - old_plane_crtc = old_plane_state->crtc; - new_dm_plane_state = to_dm_plane_state(new_plane_state); - old_dm_plane_state = to_dm_plane_state(old_plane_state); - - if (plane->type == DRM_PLANE_TYPE_CURSOR) - continue; - - if (!state->allow_modeset) - continue; - - if (crtc == new_plane_crtc) { - updates[num_plane].surface = &surface[num_plane]; - - if (new_crtc_state->mode_changed) { - updates[num_plane].surface->src_rect = - new_dm_plane_state->dc_state->src_rect; - updates[num_plane].surface->dst_rect = - new_dm_plane_state->dc_state->dst_rect; - updates[num_plane].surface->rotation = - new_dm_plane_state->dc_state->rotation; - updates[num_plane].surface->in_transfer_func = - new_dm_plane_state->dc_state->in_transfer_func; - stream_update.dst = new_dm_crtc_state->stream->dst; - stream_update.src = new_dm_crtc_state->stream->src; - } - - if (new_crtc_state->color_mgmt_changed) { - updates[num_plane].gamma = - new_dm_plane_state->dc_state->gamma_correction; - updates[num_plane].in_transfer_func = - new_dm_plane_state->dc_state->in_transfer_func; - stream_update.gamut_remap = - &new_dm_crtc_state->stream->gamut_remap_matrix; - stream_update.out_transfer_func = - new_dm_crtc_state->stream->out_transfer_func; - } - - num_plane++; - } + if (!new_dm_crtc_state->stream) { + if (!new_dm_crtc_state->stream && old_dm_crtc_state->stream) { + update_type = UPDATE_TYPE_FULL; + goto cleanup; } - if (num_plane > 0) { - ret = dm_atomic_get_state(state, &dm_state); - if (ret) - goto cleanup; + continue; + } - old_dm_state = dm_atomic_get_old_state(state); - if (!old_dm_state) { - ret = -EINVAL; - goto cleanup; - } + for_each_oldnew_plane_in_state(state, plane, old_plane_state, new_plane_state, j) { + new_plane_crtc = new_plane_state->crtc; + old_plane_crtc = old_plane_state->crtc; + new_dm_plane_state = to_dm_plane_state(new_plane_state); + old_dm_plane_state = to_dm_plane_state(old_plane_state); - status = dc_state_get_stream_status(old_dm_state->context, - new_dm_crtc_state->stream); + if (plane->type == DRM_PLANE_TYPE_CURSOR) + continue; - update_type = dc_check_update_surfaces_for_stream(dc, updates, num_plane, - &stream_update, status); + if (!state->allow_modeset) + continue; - if (update_type > UPDATE_TYPE_MED) { - update_type = UPDATE_TYPE_FULL; - goto cleanup; - } + if (crtc != new_plane_crtc) + continue; + + updates[num_plane].surface = &surface[num_plane]; + + if (new_crtc_state->mode_changed) { + updates[num_plane].surface->src_rect = + new_dm_plane_state->dc_state->src_rect; + updates[num_plane].surface->dst_rect = + new_dm_plane_state->dc_state->dst_rect; + updates[num_plane].surface->rotation = + new_dm_plane_state->dc_state->rotation; + updates[num_plane].surface->in_transfer_func = + new_dm_plane_state->dc_state->in_transfer_func; + stream_update.dst = new_dm_crtc_state->stream->dst; + stream_update.src = new_dm_crtc_state->stream->src; } - } else if (!new_dm_crtc_state->stream && old_dm_crtc_state->stream) { + if (new_crtc_state->color_mgmt_changed) { + updates[num_plane].gamma = + new_dm_plane_state->dc_state->gamma_correction; + updates[num_plane].in_transfer_func = + new_dm_plane_state->dc_state->in_transfer_func; + stream_update.gamut_remap = + &new_dm_crtc_state->stream->gamut_remap_matrix; + stream_update.out_transfer_func = + new_dm_crtc_state->stream->out_transfer_func; + } + + num_plane++; + } + + if (num_plane == 0) + continue; + + ret = dm_atomic_get_state(state, &dm_state); + if (ret) + goto cleanup; + + old_dm_state = dm_atomic_get_old_state(state); + if (!old_dm_state) { + ret = -EINVAL; + goto cleanup; + } + + status = dc_stream_get_status_from_state(old_dm_state->context, + new_dm_crtc_state->stream); + + update_type = dc_check_update_surfaces_for_stream(dc, updates, num_plane, + &stream_update, status); + + if (update_type > UPDATE_TYPE_MED) { update_type = UPDATE_TYPE_FULL; goto cleanup; } @@ -5903,6 +5906,8 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev, struct drm_connector_state *old_con_state, *new_con_state; struct drm_crtc *crtc; struct drm_crtc_state *old_crtc_state, *new_crtc_state; + struct drm_plane *plane; + struct drm_plane_state *old_plane_state, *new_plane_state; enum surface_update_type update_type = UPDATE_TYPE_FAST; enum surface_update_type overall_update_type = UPDATE_TYPE_FAST; @@ -5921,7 +5926,7 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev, for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) { if (!drm_atomic_crtc_needs_modeset(new_crtc_state) && !new_crtc_state->color_mgmt_changed && - !new_crtc_state->vrr_enabled) + old_crtc_state->vrr_enabled == new_crtc_state->vrr_enabled) continue; if (!new_crtc_state->enable) @@ -5937,27 +5942,47 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev, } /* Remove exiting planes if they are modified */ - ret = dm_update_planes_state(dc, state, false, &lock_and_validation_needed); - if (ret) { - goto fail; + for_each_oldnew_plane_in_state_reverse(state, plane, old_plane_state, new_plane_state, i) { + ret = dm_update_plane_state(dc, state, plane, + old_plane_state, + new_plane_state, + false, + &lock_and_validation_needed); + if (ret) + goto fail; } /* Disable all crtcs which require disable */ - ret = dm_update_crtcs_state(&adev->dm, state, false, &lock_and_validation_needed); - if (ret) { - goto fail; + for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) { + ret = dm_update_crtc_state(&adev->dm, state, crtc, + old_crtc_state, + new_crtc_state, + false, + &lock_and_validation_needed); + if (ret) + goto fail; } /* Enable all crtcs which require enable */ - ret = dm_update_crtcs_state(&adev->dm, state, true, &lock_and_validation_needed); - if (ret) { - goto fail; + for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) { + ret = dm_update_crtc_state(&adev->dm, state, crtc, + old_crtc_state, + new_crtc_state, + true, + &lock_and_validation_needed); + if (ret) + goto fail; } /* Add new/modified planes */ - ret = dm_update_planes_state(dc, state, true, &lock_and_validation_needed); - if (ret) { - goto fail; + for_each_oldnew_plane_in_state_reverse(state, plane, old_plane_state, new_plane_state, i) { + ret = dm_update_plane_state(dc, state, plane, + old_plane_state, + new_plane_state, + true, + &lock_and_validation_needed); + if (ret) + goto fail; } /* Run this here since we want to validate the streams we created */ diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c index f088ac585978..a10e3a50d9ef 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crc.c @@ -64,8 +64,10 @@ amdgpu_dm_crtc_verify_crc_source(struct drm_crtc *crtc, const char *src_name, int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name) { + struct amdgpu_device *adev = crtc->dev->dev_private; struct dm_crtc_state *crtc_state = to_dm_crtc_state(crtc->state); struct dc_stream_state *stream_state = crtc_state->stream; + bool enable; enum amdgpu_dm_pipe_crc_source source = dm_parse_crc_source(src_name); @@ -80,29 +82,33 @@ int amdgpu_dm_crtc_set_crc_source(struct drm_crtc *crtc, const char *src_name) return -EINVAL; } - /* When enabling CRC, we should also disable dithering. */ - if (source == AMDGPU_DM_PIPE_CRC_SOURCE_AUTO) { - if (dc_stream_configure_crc(stream_state->ctx->dc, - stream_state, - true, true)) { - crtc_state->crc_enabled = true; - dc_stream_set_dither_option(stream_state, - DITHER_OPTION_TRUN8); - } - else - return -EINVAL; - } else { - if (dc_stream_configure_crc(stream_state->ctx->dc, - stream_state, - false, false)) { - crtc_state->crc_enabled = false; - dc_stream_set_dither_option(stream_state, - DITHER_OPTION_DEFAULT); - } - else - return -EINVAL; + enable = (source == AMDGPU_DM_PIPE_CRC_SOURCE_AUTO); + + mutex_lock(&adev->dm.dc_lock); + if (!dc_stream_configure_crc(stream_state->ctx->dc, stream_state, + enable, enable)) { + mutex_unlock(&adev->dm.dc_lock); + return -EINVAL; } + /* When enabling CRC, we should also disable dithering. */ + dc_stream_set_dither_option(stream_state, + enable ? DITHER_OPTION_TRUN8 + : DITHER_OPTION_DEFAULT); + + mutex_unlock(&adev->dm.dc_lock); + + /* + * Reading the CRC requires the vblank interrupt handler to be + * enabled. Keep a reference until CRC capture stops. + */ + if (!crtc_state->crc_enabled && enable) + drm_crtc_vblank_get(crtc); + else if (crtc_state->crc_enabled && !enable) + drm_crtc_vblank_put(crtc); + + crtc_state->crc_enabled = enable; + /* Reset crc_skipped on dm state */ crtc_state->crc_skip_count = 0; return 0; diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c index 9a7ac58eb18e..cca3e16cda4f 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c @@ -783,6 +783,45 @@ static ssize_t dtn_log_write( return size; } +/* + * Backlight at this moment. Read only. + * As written to display, taking ABM and backlight lut into account. + * Ranges from 0x0 to 0x10000 (= 100% PWM) + */ +static int current_backlight_read(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *)m->private; + struct drm_device *dev = node->minor->dev; + struct amdgpu_device *adev = dev->dev_private; + struct dc *dc = adev->dm.dc; + unsigned int backlight = dc_get_current_backlight_pwm(dc); + + seq_printf(m, "0x%x\n", backlight); + return 0; +} + +/* + * Backlight value that is being approached. Read only. + * As written to display, taking ABM and backlight lut into account. + * Ranges from 0x0 to 0x10000 (= 100% PWM) + */ +static int target_backlight_read(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *)m->private; + struct drm_device *dev = node->minor->dev; + struct amdgpu_device *adev = dev->dev_private; + struct dc *dc = adev->dm.dc; + unsigned int backlight = dc_get_target_backlight_pwm(dc); + + seq_printf(m, "0x%x\n", backlight); + return 0; +} + +static const struct drm_info_list amdgpu_dm_debugfs_list[] = { + {"amdgpu_current_backlight_pwm", ¤t_backlight_read}, + {"amdgpu_target_backlight_pwm", &target_backlight_read}, +}; + int dtn_debugfs_init(struct amdgpu_device *adev) { static const struct file_operations dtn_log_fops = { @@ -793,9 +832,15 @@ int dtn_debugfs_init(struct amdgpu_device *adev) }; struct drm_minor *minor = adev->ddev->primary; - struct dentry *root = minor->debugfs_root; + struct dentry *ent, *root = minor->debugfs_root; + int ret; + + ret = amdgpu_debugfs_add_files(adev, amdgpu_dm_debugfs_list, + ARRAY_SIZE(amdgpu_dm_debugfs_list)); + if (ret) + return ret; - struct dentry *ent = debugfs_create_file( + ent = debugfs_create_file( "amdgpu_dm_dtn_log", 0644, root, diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c index 78173311f718..b39766bd2840 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c @@ -192,7 +192,7 @@ bool dm_helpers_dp_mst_write_payload_allocation_table( int bpp = 0; int pbn = 0; - aconnector = stream->sink->priv; + aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context; if (!aconnector || !aconnector->mst_port) return false; @@ -205,7 +205,7 @@ bool dm_helpers_dp_mst_write_payload_allocation_table( mst_port = aconnector->port; if (enable) { - clock = stream->timing.pix_clk_khz; + clock = stream->timing.pix_clk_100hz / 10; switch (stream->timing.display_color_depth) { @@ -263,6 +263,13 @@ bool dm_helpers_dp_mst_write_payload_allocation_table( return true; } +/* + * poll pending down reply before clear payload allocation table + */ +void dm_helpers_dp_mst_poll_pending_down_reply( + struct dc_context *ctx, + const struct dc_link *link) +{} /* * Clear payload allocation table before enable MST DP link. @@ -284,7 +291,7 @@ bool dm_helpers_dp_mst_poll_for_allocation_change_trigger( struct drm_dp_mst_topology_mgr *mst_mgr; int ret; - aconnector = stream->sink->priv; + aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context; if (!aconnector || !aconnector->mst_port) return false; @@ -312,7 +319,7 @@ bool dm_helpers_dp_mst_send_payload_allocation( struct drm_dp_mst_port *mst_port; int ret; - aconnector = stream->sink->priv; + aconnector = (struct amdgpu_dm_connector *)stream->dm_stream_context; if (!aconnector || !aconnector->mst_port) return false; diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c index 24632727e127..f51d52eb52e6 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c @@ -35,6 +35,8 @@ #include "dc_link_ddc.h" +#include "i2caux_interface.h" + /* #define TRACE_DPCD */ #ifdef TRACE_DPCD @@ -81,80 +83,24 @@ static ssize_t dm_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) { ssize_t result = 0; - enum i2caux_transaction_action action; - enum aux_transaction_type type; + struct aux_payload payload; if (WARN_ON(msg->size > 16)) return -E2BIG; - switch (msg->request & ~DP_AUX_I2C_MOT) { - case DP_AUX_NATIVE_READ: - type = AUX_TRANSACTION_TYPE_DP; - action = I2CAUX_TRANSACTION_ACTION_DP_READ; - - result = dc_link_aux_transfer(TO_DM_AUX(aux)->ddc_service, - msg->address, - &msg->reply, - msg->buffer, - msg->size, - type, - action); - break; - case DP_AUX_NATIVE_WRITE: - type = AUX_TRANSACTION_TYPE_DP; - action = I2CAUX_TRANSACTION_ACTION_DP_WRITE; - - dc_link_aux_transfer(TO_DM_AUX(aux)->ddc_service, - msg->address, - &msg->reply, - msg->buffer, - msg->size, - type, - action); - result = msg->size; - break; - case DP_AUX_I2C_READ: - type = AUX_TRANSACTION_TYPE_I2C; - if (msg->request & DP_AUX_I2C_MOT) - action = I2CAUX_TRANSACTION_ACTION_I2C_READ_MOT; - else - action = I2CAUX_TRANSACTION_ACTION_I2C_READ; - - result = dc_link_aux_transfer(TO_DM_AUX(aux)->ddc_service, - msg->address, - &msg->reply, - msg->buffer, - msg->size, - type, - action); - break; - case DP_AUX_I2C_WRITE: - type = AUX_TRANSACTION_TYPE_I2C; - if (msg->request & DP_AUX_I2C_MOT) - action = I2CAUX_TRANSACTION_ACTION_I2C_WRITE_MOT; - else - action = I2CAUX_TRANSACTION_ACTION_I2C_WRITE; - - dc_link_aux_transfer(TO_DM_AUX(aux)->ddc_service, - msg->address, - &msg->reply, - msg->buffer, - msg->size, - type, - action); - result = msg->size; - break; - default: - return -EINVAL; - } + payload.address = msg->address; + payload.data = msg->buffer; + payload.length = msg->size; + payload.reply = &msg->reply; + payload.i2c_over_aux = (msg->request & DP_AUX_NATIVE_WRITE) == 0; + payload.write = (msg->request & DP_AUX_I2C_READ) == 0; + payload.mot = (msg->request & DP_AUX_I2C_MOT) != 0; + payload.defer_delay = 0; -#ifdef TRACE_DPCD - log_dpcd(msg->request, - msg->address, - msg->buffer, - msg->size, - r == DDC_RESULT_SUCESSFULL); -#endif + result = dc_link_aux_transfer(TO_DM_AUX(aux)->ddc_service, &payload); + + if (payload.write) + result = msg->size; if (result < 0) /* DC doesn't know about kernel error codes */ result = -EIO; @@ -228,6 +174,11 @@ static int dm_dp_mst_get_modes(struct drm_connector *connector) aconnector->edid = edid; } + if (aconnector->dc_sink && aconnector->dc_sink->sink_signal == SIGNAL_TYPE_VIRTUAL) { + dc_sink_release(aconnector->dc_sink); + aconnector->dc_sink = NULL; + } + if (!aconnector->dc_sink) { struct dc_sink *dc_sink; struct dc_sink_init_data init_params = { diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c index 7ef99037167a..a114954d6a5b 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_pp_smu.c @@ -559,6 +559,58 @@ void pp_rv_set_pme_wa_enable(struct pp_smu *pp) pp_funcs->notify_smu_enable_pwe(pp_handle); } +void pp_rv_set_active_display_count(struct pp_smu *pp, int count) +{ + const struct dc_context *ctx = pp->dm; + struct amdgpu_device *adev = ctx->driver_context; + void *pp_handle = adev->powerplay.pp_handle; + const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs; + + if (!pp_funcs || !pp_funcs->set_active_display_count) + return; + + pp_funcs->set_active_display_count(pp_handle, count); +} + +void pp_rv_set_min_deep_sleep_dcfclk(struct pp_smu *pp, int clock) +{ + const struct dc_context *ctx = pp->dm; + struct amdgpu_device *adev = ctx->driver_context; + void *pp_handle = adev->powerplay.pp_handle; + const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs; + + if (!pp_funcs || !pp_funcs->set_min_deep_sleep_dcefclk) + return; + + pp_funcs->set_min_deep_sleep_dcefclk(pp_handle, clock); +} + +void pp_rv_set_hard_min_dcefclk_by_freq(struct pp_smu *pp, int clock) +{ + const struct dc_context *ctx = pp->dm; + struct amdgpu_device *adev = ctx->driver_context; + void *pp_handle = adev->powerplay.pp_handle; + const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs; + + if (!pp_funcs || !pp_funcs->set_hard_min_dcefclk_by_freq) + return; + + pp_funcs->set_hard_min_dcefclk_by_freq(pp_handle, clock); +} + +void pp_rv_set_hard_min_fclk_by_freq(struct pp_smu *pp, int mhz) +{ + const struct dc_context *ctx = pp->dm; + struct amdgpu_device *adev = ctx->driver_context; + void *pp_handle = adev->powerplay.pp_handle; + const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs; + + if (!pp_funcs || !pp_funcs->set_hard_min_fclk_by_freq) + return; + + pp_funcs->set_hard_min_fclk_by_freq(pp_handle, mhz); +} + void dm_pp_get_funcs_rv( struct dc_context *ctx, struct pp_smu_funcs_rv *funcs) @@ -567,4 +619,9 @@ void dm_pp_get_funcs_rv( funcs->set_display_requirement = pp_rv_set_display_requirement; funcs->set_wm_ranges = pp_rv_set_wm_ranges; funcs->set_pme_wa_enable = pp_rv_set_pme_wa_enable; + funcs->set_display_count = pp_rv_set_active_display_count; + funcs->set_min_deep_sleep_dcfclk = pp_rv_set_min_deep_sleep_dcfclk; + funcs->set_hard_min_dcfclk_by_freq = pp_rv_set_hard_min_dcefclk_by_freq; + funcs->set_hard_min_fclk_by_freq = pp_rv_set_hard_min_fclk_by_freq; } + diff --git a/drivers/gpu/drm/amd/display/dc/Makefile b/drivers/gpu/drm/amd/display/dc/Makefile index aed538a4d1ba..b8ddb4acccdb 100644 --- a/drivers/gpu/drm/amd/display/dc/Makefile +++ b/drivers/gpu/drm/amd/display/dc/Makefile @@ -23,7 +23,7 @@ # Makefile for Display Core (dc) component. # -DC_LIBS = basics bios calcs dce gpio i2caux irq virtual +DC_LIBS = basics bios calcs dce gpio irq virtual ifdef CONFIG_DRM_AMD_DC_DCN1_0 DC_LIBS += dcn10 dml @@ -41,7 +41,8 @@ AMD_DC = $(addsuffix /Makefile, $(addprefix $(FULL_AMD_DISPLAY_PATH)/dc/,$(DC_LI include $(AMD_DC) DISPLAY_CORE = dc.o dc_link.o dc_resource.o dc_hw_sequencer.o dc_sink.o \ -dc_surface.o dc_link_hwss.o dc_link_dp.o dc_link_ddc.o dc_debug.o dc_stream.o +dc_surface.o dc_link_hwss.o dc_link_dp.o dc_link_ddc.o dc_debug.o dc_stream.o \ +dc_vm_helper.o AMD_DISPLAY_CORE = $(addprefix $(AMDDALPATH)/dc/core/,$(DISPLAY_CORE)) diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c index c2ab026aee91..a4c97d32e751 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c +++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser.c @@ -835,18 +835,6 @@ static enum bp_result bios_parser_enable_crtc( return bp->cmd_tbl.enable_crtc(bp, id, enable); } -static enum bp_result bios_parser_crtc_source_select( - struct dc_bios *dcb, - struct bp_crtc_source_select *bp_params) -{ - struct bios_parser *bp = BP_FROM_DCB(dcb); - - if (!bp->cmd_tbl.select_crtc_source) - return BP_RESULT_FAILURE; - - return bp->cmd_tbl.select_crtc_source(bp, bp_params); -} - static enum bp_result bios_parser_enable_disp_power_gating( struct dc_bios *dcb, enum controller_id controller_id, @@ -2842,8 +2830,6 @@ static const struct dc_vbios_funcs vbios_funcs = { .program_crtc_timing = bios_parser_program_crtc_timing, /* still use. should probably retire and program directly */ - .crtc_source_select = bios_parser_crtc_source_select, /* still use. should probably retire and program directly */ - .program_display_engine_pll = bios_parser_program_display_engine_pll, .enable_disp_power_gating = bios_parser_enable_disp_power_gating, diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c index c513ab6f3843..a1c56f29cfeb 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c +++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c @@ -1083,18 +1083,6 @@ static enum bp_result bios_parser_enable_crtc( return bp->cmd_tbl.enable_crtc(bp, id, enable); } -static enum bp_result bios_parser_crtc_source_select( - struct dc_bios *dcb, - struct bp_crtc_source_select *bp_params) -{ - struct bios_parser *bp = BP_FROM_DCB(dcb); - - if (!bp->cmd_tbl.select_crtc_source) - return BP_RESULT_FAILURE; - - return bp->cmd_tbl.select_crtc_source(bp, bp_params); -} - static enum bp_result bios_parser_enable_disp_power_gating( struct dc_bios *dcb, enum controller_id controller_id, @@ -1899,8 +1887,6 @@ static const struct dc_vbios_funcs vbios_funcs = { .is_accelerated_mode = bios_parser_is_accelerated_mode, - .is_active_display = bios_is_active_display, - .set_scratch_critical_state = bios_parser_set_scratch_critical_state, @@ -1917,8 +1903,6 @@ static const struct dc_vbios_funcs vbios_funcs = { .program_crtc_timing = bios_parser_program_crtc_timing, - .crtc_source_select = bios_parser_crtc_source_select, - .enable_disp_power_gating = bios_parser_enable_disp_power_gating, .bios_parser_destroy = firmware_parser_destroy, diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.c index fdda8aa8e303..fce46ab54c54 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.c +++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.c @@ -83,101 +83,7 @@ uint32_t bios_get_vga_enabled_displays( { uint32_t active_disp = 1; - if (bios->regs->BIOS_SCRATCH_3) /*follow up with other asic, todo*/ - active_disp = REG_READ(BIOS_SCRATCH_3) & 0XFFFF; + active_disp = REG_READ(BIOS_SCRATCH_3) & 0XFFFF; return active_disp; } -bool bios_is_active_display( - struct dc_bios *bios, - enum signal_type signal, - const struct connector_device_tag_info *device_tag) -{ - uint32_t active = 0; - uint32_t connected = 0; - uint32_t bios_scratch_0 = 0; - uint32_t bios_scratch_3 = 0; - - switch (signal) { - case SIGNAL_TYPE_DVI_SINGLE_LINK: - case SIGNAL_TYPE_DVI_DUAL_LINK: - case SIGNAL_TYPE_HDMI_TYPE_A: - case SIGNAL_TYPE_DISPLAY_PORT: - case SIGNAL_TYPE_DISPLAY_PORT_MST: - { - if (device_tag->dev_id.device_type == DEVICE_TYPE_DFP) { - switch (device_tag->dev_id.enum_id) { - case 1: - { - active = ATOM_S3_DFP1_ACTIVE; - connected = 0x0008; //ATOM_DISPLAY_DFP1_CONNECT - } - break; - - case 2: - { - active = ATOM_S3_DFP2_ACTIVE; - connected = 0x0080; //ATOM_DISPLAY_DFP2_CONNECT - } - break; - - case 3: - { - active = ATOM_S3_DFP3_ACTIVE; - connected = 0x0200; //ATOM_DISPLAY_DFP3_CONNECT - } - break; - - case 4: - { - active = ATOM_S3_DFP4_ACTIVE; - connected = 0x0400; //ATOM_DISPLAY_DFP4_CONNECT - } - break; - - case 5: - { - active = ATOM_S3_DFP5_ACTIVE; - connected = 0x0800; //ATOM_DISPLAY_DFP5_CONNECT - } - break; - - case 6: - { - active = ATOM_S3_DFP6_ACTIVE; - connected = 0x0040; //ATOM_DISPLAY_DFP6_CONNECT - } - break; - - default: - break; - } - } - } - break; - - case SIGNAL_TYPE_LVDS: - case SIGNAL_TYPE_EDP: - { - active = ATOM_S3_LCD1_ACTIVE; - connected = 0x0002; //ATOM_DISPLAY_LCD1_CONNECT - } - break; - - default: - break; - } - - - if (bios->regs->BIOS_SCRATCH_0) /*follow up with other asic, todo*/ - bios_scratch_0 = REG_READ(BIOS_SCRATCH_0); - if (bios->regs->BIOS_SCRATCH_3) /*follow up with other asic, todo*/ - bios_scratch_3 = REG_READ(BIOS_SCRATCH_3); - - bios_scratch_3 &= ATOM_S3_DEVICE_ACTIVE_MASK; - if ((active & bios_scratch_3) && (connected & bios_scratch_0)) - return true; - - return false; -} - diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.h b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.h index f33cac2147e3..75a29e68fb27 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.h +++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser_helper.h @@ -35,10 +35,6 @@ bool bios_is_accelerated_mode(struct dc_bios *bios); void bios_set_scratch_acc_mode_change(struct dc_bios *bios); void bios_set_scratch_critical_state(struct dc_bios *bios, bool state); uint32_t bios_get_vga_enabled_displays(struct dc_bios *bios); -bool bios_is_active_display( - struct dc_bios *bios, - enum signal_type signal, - const struct connector_device_tag_info *device_tag); #define GET_IMAGE(type, offset) ((type *) bios_get_image(&bp->base, offset, sizeof(type))) diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table.c b/drivers/gpu/drm/amd/display/dc/bios/command_table.c index 2bd7cd97e00d..5815983caaf8 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/command_table.c +++ b/drivers/gpu/drm/amd/display/dc/bios/command_table.c @@ -55,7 +55,6 @@ static void init_adjust_display_pll(struct bios_parser *bp); static void init_dac_encoder_control(struct bios_parser *bp); static void init_dac_output_control(struct bios_parser *bp); static void init_set_crtc_timing(struct bios_parser *bp); -static void init_select_crtc_source(struct bios_parser *bp); static void init_enable_crtc(struct bios_parser *bp); static void init_enable_crtc_mem_req(struct bios_parser *bp); static void init_external_encoder_control(struct bios_parser *bp); @@ -73,7 +72,6 @@ void dal_bios_parser_init_cmd_tbl(struct bios_parser *bp) init_dac_encoder_control(bp); init_dac_output_control(bp); init_set_crtc_timing(bp); - init_select_crtc_source(bp); init_enable_crtc(bp); init_enable_crtc_mem_req(bp); init_program_clock(bp); @@ -964,9 +962,9 @@ static enum bp_result set_pixel_clock_v3( allocation.sPCLKInput.ucPostDiv = (uint8_t)bp_params->pixel_clock_post_divider; - /* We need to convert from KHz units into 10KHz units */ + /* We need to convert from 100Hz units into 10KHz units */ allocation.sPCLKInput.usPixelClock = - cpu_to_le16((uint16_t)(bp_params->target_pixel_clock / 10)); + cpu_to_le16((uint16_t)(bp_params->target_pixel_clock_100hz / 100)); params = (PIXEL_CLOCK_PARAMETERS_V3 *)&allocation.sPCLKInput; params->ucTransmitterId = @@ -1042,9 +1040,9 @@ static enum bp_result set_pixel_clock_v5( (uint8_t)bp->cmd_helper->encoder_mode_bp_to_atom( bp_params->signal_type, false); - /* We need to convert from KHz units into 10KHz units */ + /* We need to convert from 100Hz units into 10KHz units */ clk.sPCLKInput.usPixelClock = - cpu_to_le16((uint16_t)(bp_params->target_pixel_clock / 10)); + cpu_to_le16((uint16_t)(bp_params->target_pixel_clock_100hz / 100)); if (bp_params->flags.FORCE_PROGRAMMING_OF_PLL) clk.sPCLKInput.ucMiscInfo |= @@ -1118,9 +1116,9 @@ static enum bp_result set_pixel_clock_v6( (uint8_t) bp->cmd_helper->encoder_mode_bp_to_atom( bp_params->signal_type, false); - /* We need to convert from KHz units into 10KHz units */ + /* We need to convert from 100 Hz units into 10KHz units */ clk.sPCLKInput.ulCrtcPclkFreq.ulPixelClock = - cpu_to_le32(bp_params->target_pixel_clock / 10); + cpu_to_le32(bp_params->target_pixel_clock_100hz / 100); if (bp_params->flags.FORCE_PROGRAMMING_OF_PLL) { clk.sPCLKInput.ucMiscInfo |= @@ -1182,8 +1180,7 @@ static enum bp_result set_pixel_clock_v7( clk.ucTransmitterID = bp->cmd_helper->encoder_id_to_atom(dal_graphics_object_id_get_encoder_id(bp_params->encoder_object_id)); clk.ucEncoderMode = (uint8_t) bp->cmd_helper->encoder_mode_bp_to_atom(bp_params->signal_type, false); - /* We need to convert from KHz units into 10KHz units */ - clk.ulPixelClock = cpu_to_le32(bp_params->target_pixel_clock * 10); + clk.ulPixelClock = cpu_to_le32(bp_params->target_pixel_clock_100hz); clk.ucDeepColorRatio = (uint8_t) bp->cmd_helper->transmitter_color_depth_to_atom(bp_params->color_depth); @@ -1899,120 +1896,6 @@ static enum bp_result set_crtc_using_dtd_timing_v3( /******************************************************************************* ******************************************************************************** ** - ** SELECT CRTC SOURCE - ** - ******************************************************************************** - *******************************************************************************/ - -static enum bp_result select_crtc_source_v2( - struct bios_parser *bp, - struct bp_crtc_source_select *bp_params); -static enum bp_result select_crtc_source_v3( - struct bios_parser *bp, - struct bp_crtc_source_select *bp_params); - -static void init_select_crtc_source(struct bios_parser *bp) -{ - switch (BIOS_CMD_TABLE_PARA_REVISION(SelectCRTC_Source)) { - case 2: - bp->cmd_tbl.select_crtc_source = select_crtc_source_v2; - break; - case 3: - bp->cmd_tbl.select_crtc_source = select_crtc_source_v3; - break; - default: - dm_output_to_console("Don't select_crtc_source enable_crtc for v%d\n", - BIOS_CMD_TABLE_PARA_REVISION(SelectCRTC_Source)); - bp->cmd_tbl.select_crtc_source = NULL; - break; - } -} - -static enum bp_result select_crtc_source_v2( - struct bios_parser *bp, - struct bp_crtc_source_select *bp_params) -{ - enum bp_result result = BP_RESULT_FAILURE; - SELECT_CRTC_SOURCE_PARAMETERS_V2 params; - uint8_t atom_controller_id; - uint32_t atom_engine_id; - enum signal_type s = bp_params->signal; - - memset(¶ms, 0, sizeof(params)); - - /* set controller id */ - if (bp->cmd_helper->controller_id_to_atom( - bp_params->controller_id, &atom_controller_id)) - params.ucCRTC = atom_controller_id; - else - return BP_RESULT_FAILURE; - - /* set encoder id */ - if (bp->cmd_helper->engine_bp_to_atom( - bp_params->engine_id, &atom_engine_id)) - params.ucEncoderID = (uint8_t)atom_engine_id; - else - return BP_RESULT_FAILURE; - - if (SIGNAL_TYPE_EDP == s || - (SIGNAL_TYPE_DISPLAY_PORT == s && - SIGNAL_TYPE_LVDS == bp_params->sink_signal)) - s = SIGNAL_TYPE_LVDS; - - params.ucEncodeMode = - (uint8_t)bp->cmd_helper->encoder_mode_bp_to_atom( - s, bp_params->enable_dp_audio); - - if (EXEC_BIOS_CMD_TABLE(SelectCRTC_Source, params)) - result = BP_RESULT_OK; - - return result; -} - -static enum bp_result select_crtc_source_v3( - struct bios_parser *bp, - struct bp_crtc_source_select *bp_params) -{ - bool result = BP_RESULT_FAILURE; - SELECT_CRTC_SOURCE_PARAMETERS_V3 params; - uint8_t atom_controller_id; - uint32_t atom_engine_id; - enum signal_type s = bp_params->signal; - - memset(¶ms, 0, sizeof(params)); - - if (bp->cmd_helper->controller_id_to_atom(bp_params->controller_id, - &atom_controller_id)) - params.ucCRTC = atom_controller_id; - else - return result; - - if (bp->cmd_helper->engine_bp_to_atom(bp_params->engine_id, - &atom_engine_id)) - params.ucEncoderID = (uint8_t)atom_engine_id; - else - return result; - - if (SIGNAL_TYPE_EDP == s || - (SIGNAL_TYPE_DISPLAY_PORT == s && - SIGNAL_TYPE_LVDS == bp_params->sink_signal)) - s = SIGNAL_TYPE_LVDS; - - params.ucEncodeMode = - bp->cmd_helper->encoder_mode_bp_to_atom( - s, bp_params->enable_dp_audio); - /* Needed for VBIOS Random Spatial Dithering feature */ - params.ucDstBpc = (uint8_t)(bp_params->display_output_bit_depth); - - if (EXEC_BIOS_CMD_TABLE(SelectCRTC_Source, params)) - result = BP_RESULT_OK; - - return result; -} - -/******************************************************************************* - ******************************************************************************** - ** ** ENABLE CRTC ** ******************************************************************************** @@ -2164,7 +2047,7 @@ static enum bp_result program_clock_v5( /* We need to convert from KHz units into 10KHz units */ params.sPCLKInput.ucPpll = (uint8_t) atom_pll_id; params.sPCLKInput.usPixelClock = - cpu_to_le16((uint16_t) (bp_params->target_pixel_clock / 10)); + cpu_to_le16((uint16_t) (bp_params->target_pixel_clock_100hz / 100)); params.sPCLKInput.ucCRTC = (uint8_t) ATOM_CRTC_INVALID; if (bp_params->flags.SET_EXTERNAL_REF_DIV_SRC) @@ -2196,7 +2079,7 @@ static enum bp_result program_clock_v6( /* We need to convert from KHz units into 10KHz units */ params.sPCLKInput.ucPpll = (uint8_t)atom_pll_id; params.sPCLKInput.ulDispEngClkFreq = - cpu_to_le32(bp_params->target_pixel_clock / 10); + cpu_to_le32(bp_params->target_pixel_clock_100hz / 100); if (bp_params->flags.SET_EXTERNAL_REF_DIV_SRC) params.sPCLKInput.ucMiscInfo |= PIXEL_CLOCK_MISC_REF_DIV_SRC; diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table.h b/drivers/gpu/drm/amd/display/dc/bios/command_table.h index 94f3d43a7471..ad533775e724 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/command_table.h +++ b/drivers/gpu/drm/amd/display/dc/bios/command_table.h @@ -71,9 +71,6 @@ struct cmd_tbl { enum bp_result (*set_crtc_timing)( struct bios_parser *bp, struct bp_hw_crtc_timing_parameters *bp_params); - enum bp_result (*select_crtc_source)( - struct bios_parser *bp, - struct bp_crtc_source_select *bp_params); enum bp_result (*enable_crtc)( struct bios_parser *bp, enum controller_id controller_id, diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table2.c b/drivers/gpu/drm/amd/display/dc/bios/command_table2.c index 2b5dc499a35e..bb2e8105e6ab 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/command_table2.c +++ b/drivers/gpu/drm/amd/display/dc/bios/command_table2.c @@ -301,17 +301,17 @@ static enum bp_result set_pixel_clock_v7( cmd_helper->encoder_mode_bp_to_atom( bp_params->signal_type, false); - /* We need to convert from KHz units into 10KHz units */ - clk.pixclk_100hz = cpu_to_le32(bp_params->target_pixel_clock * - 10); + clk.pixclk_100hz = cpu_to_le32(bp_params->target_pixel_clock_100hz); clk.deep_color_ratio = (uint8_t) bp->cmd_helper-> transmitter_color_depth_to_atom( bp_params->color_depth); - DC_LOG_BIOS("%s:program display clock = %d"\ - "colorDepth = %d\n", __func__,\ - bp_params->target_pixel_clock, bp_params->color_depth); + + DC_LOG_BIOS("%s:program display clock = %d, tg = %d, pll = %d, "\ + "colorDepth = %d\n", __func__, + bp_params->target_pixel_clock_100hz, (int)controller_id, + pll_id, bp_params->color_depth); if (bp_params->flags.FORCE_PROGRAMMING_OF_PLL) clk.miscinfo |= PIXEL_CLOCK_V7_MISC_FORCE_PROG_PPLL; @@ -463,75 +463,6 @@ static enum bp_result set_crtc_using_dtd_timing_v3( /****************************************************************************** ****************************************************************************** ** - ** SELECT CRTC SOURCE - ** - ****************************************************************************** - *****************************************************************************/ - - -static enum bp_result select_crtc_source_v3( - struct bios_parser *bp, - struct bp_crtc_source_select *bp_params); - -static void init_select_crtc_source(struct bios_parser *bp) -{ - switch (BIOS_CMD_TABLE_PARA_REVISION(selectcrtc_source)) { - case 3: - bp->cmd_tbl.select_crtc_source = select_crtc_source_v3; - break; - default: - dm_output_to_console("Don't select_crtc_source enable_crtc for v%d\n", - BIOS_CMD_TABLE_PARA_REVISION(selectcrtc_source)); - bp->cmd_tbl.select_crtc_source = NULL; - break; - } -} - - -static enum bp_result select_crtc_source_v3( - struct bios_parser *bp, - struct bp_crtc_source_select *bp_params) -{ - bool result = BP_RESULT_FAILURE; - struct select_crtc_source_parameters_v2_3 params; - uint8_t atom_controller_id; - uint32_t atom_engine_id; - enum signal_type s = bp_params->signal; - - memset(¶ms, 0, sizeof(params)); - - if (bp->cmd_helper->controller_id_to_atom(bp_params->controller_id, - &atom_controller_id)) - params.crtc_id = atom_controller_id; - else - return result; - - if (bp->cmd_helper->engine_bp_to_atom(bp_params->engine_id, - &atom_engine_id)) - params.encoder_id = (uint8_t)atom_engine_id; - else - return result; - - if (s == SIGNAL_TYPE_EDP || - (s == SIGNAL_TYPE_DISPLAY_PORT && bp_params->sink_signal == - SIGNAL_TYPE_LVDS)) - s = SIGNAL_TYPE_LVDS; - - params.encode_mode = - bp->cmd_helper->encoder_mode_bp_to_atom( - s, bp_params->enable_dp_audio); - /* Needed for VBIOS Random Spatial Dithering feature */ - params.dst_bpc = (uint8_t)(bp_params->display_output_bit_depth); - - if (EXEC_BIOS_CMD_TABLE(selectcrtc_source, params)) - result = BP_RESULT_OK; - - return result; -} - -/****************************************************************************** - ****************************************************************************** - ** ** ENABLE CRTC ** ****************************************************************************** @@ -808,7 +739,6 @@ void dal_firmware_parser_init_cmd_tbl(struct bios_parser *bp) init_set_crtc_timing(bp); - init_select_crtc_source(bp); init_enable_crtc(bp); init_external_encoder_control(bp); diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table2.h b/drivers/gpu/drm/amd/display/dc/bios/command_table2.h index ec1c0c9f3f1d..7a2af24dfe60 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/command_table2.h +++ b/drivers/gpu/drm/amd/display/dc/bios/command_table2.h @@ -71,9 +71,6 @@ struct cmd_tbl { enum bp_result (*set_crtc_timing)( struct bios_parser *bp, struct bp_hw_crtc_timing_parameters *bp_params); - enum bp_result (*select_crtc_source)( - struct bios_parser *bp, - struct bp_crtc_source_select *bp_params); enum bp_result (*enable_crtc)( struct bios_parser *bp, enum controller_id controller_id, diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dce_calcs.c b/drivers/gpu/drm/amd/display/dc/calcs/dce_calcs.c index 9ebe30ba4dab..f3aa7b53d2aa 100644 --- a/drivers/gpu/drm/amd/display/dc/calcs/dce_calcs.c +++ b/drivers/gpu/drm/amd/display/dc/calcs/dce_calcs.c @@ -2792,7 +2792,7 @@ static void populate_initial_data( data->lpt_en[num_displays + 4] = false; data->h_total[num_displays + 4] = bw_int_to_fixed(pipe[i].stream->timing.h_total); data->v_total[num_displays + 4] = bw_int_to_fixed(pipe[i].stream->timing.v_total); - data->pixel_rate[num_displays + 4] = bw_frc_to_fixed(pipe[i].stream->timing.pix_clk_khz, 1000); + data->pixel_rate[num_displays + 4] = bw_frc_to_fixed(pipe[i].stream->timing.pix_clk_100hz, 10000); data->src_width[num_displays + 4] = bw_int_to_fixed(pipe[i].plane_res.scl_data.viewport.width); data->pitch_in_pixels[num_displays + 4] = data->src_width[num_displays + 4]; data->src_height[num_displays + 4] = bw_int_to_fixed(pipe[i].plane_res.scl_data.viewport.height); @@ -2881,7 +2881,7 @@ static void populate_initial_data( /* Pipes without underlay after */ for (i = 0; i < pipe_count; i++) { - unsigned int pixel_clock_khz; + unsigned int pixel_clock_100hz; if (!pipe[i].stream || pipe[i].bottom_pipe) continue; @@ -2890,10 +2890,10 @@ static void populate_initial_data( data->lpt_en[num_displays + 4] = false; data->h_total[num_displays + 4] = bw_int_to_fixed(pipe[i].stream->timing.h_total); data->v_total[num_displays + 4] = bw_int_to_fixed(pipe[i].stream->timing.v_total); - pixel_clock_khz = pipe[i].stream->timing.pix_clk_khz; + pixel_clock_100hz = pipe[i].stream->timing.pix_clk_100hz; if (pipe[i].stream->timing.timing_3d_format == TIMING_3D_FORMAT_HW_FRAME_PACKING) - pixel_clock_khz *= 2; - data->pixel_rate[num_displays + 4] = bw_frc_to_fixed(pixel_clock_khz, 1000); + pixel_clock_100hz *= 2; + data->pixel_rate[num_displays + 4] = bw_frc_to_fixed(pixel_clock_100hz, 10000); if (pipe[i].plane_state) { data->src_width[num_displays + 4] = bw_int_to_fixed(pipe[i].plane_res.scl_data.viewport.width); data->pitch_in_pixels[num_displays + 4] = data->src_width[num_displays + 4]; diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c index d0fc54f8fb1c..1ef0074302c5 100644 --- a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c +++ b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_auto.c @@ -63,7 +63,7 @@ void scaler_settings_calculation(struct dcn_bw_internal_vars *v) if (v->interlace_output[k] == 1.0) { v->v_ratio[k] = 2.0 * v->v_ratio[k]; } - if ((v->underscan_output[k] == 1.0)) { + if (v->underscan_output[k] == 1.0) { v->h_ratio[k] = v->h_ratio[k] * v->under_scan_factor; v->v_ratio[k] = v->v_ratio[k] * v->under_scan_factor; } @@ -797,9 +797,40 @@ void mode_support_and_system_configuration(struct dcn_bw_internal_vars *v) else { v->maximum_vstartup = v->v_sync_plus_back_porch[k] - 1.0; } - v->line_times_for_prefetch[k] = v->maximum_vstartup - v->urgent_latency / (v->htotal[k] / v->pixel_clock[k]) - (v->time_calc + v->time_setup) / (v->htotal[k] / v->pixel_clock[k]) - (v->dst_y_after_scaler + v->dst_x_after_scaler / v->htotal[k]); - v->line_times_for_prefetch[k] =dcn_bw_floor2(4.0 * (v->line_times_for_prefetch[k] + 0.125), 1.0) / 4; - v->prefetch_bw[k] = (v->meta_pte_bytes_per_frame[k] + 2.0 * v->meta_row_bytes[k] + 2.0 * v->dpte_bytes_per_row[k] + v->prefetch_lines_y[k] * v->swath_width_yper_state[i][j][k] *dcn_bw_ceil2(v->byte_per_pixel_in_dety[k], 1.0) + v->prefetch_lines_c[k] * v->swath_width_yper_state[i][j][k] / 2.0 *dcn_bw_ceil2(v->byte_per_pixel_in_detc[k], 2.0)) / (v->line_times_for_prefetch[k] * v->htotal[k] / v->pixel_clock[k]); + + do { + v->line_times_for_prefetch[k] = v->maximum_vstartup - v->urgent_latency / (v->htotal[k] / v->pixel_clock[k]) - (v->time_calc + v->time_setup) / (v->htotal[k] / v->pixel_clock[k]) - (v->dst_y_after_scaler + v->dst_x_after_scaler / v->htotal[k]); + v->line_times_for_prefetch[k] =dcn_bw_floor2(4.0 * (v->line_times_for_prefetch[k] + 0.125), 1.0) / 4; + v->prefetch_bw[k] = (v->meta_pte_bytes_per_frame[k] + 2.0 * v->meta_row_bytes[k] + 2.0 * v->dpte_bytes_per_row[k] + v->prefetch_lines_y[k] * v->swath_width_yper_state[i][j][k] *dcn_bw_ceil2(v->byte_per_pixel_in_dety[k], 1.0) + v->prefetch_lines_c[k] * v->swath_width_yper_state[i][j][k] / 2.0 *dcn_bw_ceil2(v->byte_per_pixel_in_detc[k], 2.0)) / (v->line_times_for_prefetch[k] * v->htotal[k] / v->pixel_clock[k]); + + if (v->pte_enable == dcn_bw_yes && v->dcc_enable[k] == dcn_bw_yes) { + v->time_for_meta_pte_without_immediate_flip = dcn_bw_max3( + v->meta_pte_bytes_frame[k] / v->prefetch_bandwidth[k], + v->extra_latency, + v->htotal[k] / v->pixel_clock[k] / 4.0); + } else { + v->time_for_meta_pte_without_immediate_flip = v->htotal[k] / v->pixel_clock[k] / 4.0; + } + + if (v->pte_enable == dcn_bw_yes || v->dcc_enable[k] == dcn_bw_yes) { + v->time_for_meta_and_dpte_row_without_immediate_flip = dcn_bw_max3(( + v->meta_row_bytes[k] + v->dpte_bytes_per_row[k]) / v->prefetch_bandwidth[k], + v->htotal[k] / v->pixel_clock[k] - v->time_for_meta_pte_without_immediate_flip, + v->extra_latency); + } else { + v->time_for_meta_and_dpte_row_without_immediate_flip = dcn_bw_max2( + v->htotal[k] / v->pixel_clock[k] - v->time_for_meta_pte_without_immediate_flip, + v->extra_latency - v->time_for_meta_pte_with_immediate_flip); + } + + v->lines_for_meta_pte_without_immediate_flip[k] =dcn_bw_floor2(4.0 * (v->time_for_meta_pte_without_immediate_flip / (v->htotal[k] / v->pixel_clock[k]) + 0.125), 1.0) / 4; + v->lines_for_meta_and_dpte_row_without_immediate_flip[k] =dcn_bw_floor2(4.0 * (v->time_for_meta_and_dpte_row_without_immediate_flip / (v->htotal[k] / v->pixel_clock[k]) + 0.125), 1.0) / 4; + v->maximum_vstartup = v->maximum_vstartup - 1; + + if (v->lines_for_meta_pte_without_immediate_flip[k] < 8.0 && v->lines_for_meta_and_dpte_row_without_immediate_flip[k] < 16.0) + break; + + } while(1); } v->bw_available_for_immediate_flip = v->return_bw_per_state[i]; for (k = 0; k <= v->number_of_active_planes - 1; k++) { @@ -814,24 +845,18 @@ void mode_support_and_system_configuration(struct dcn_bw_internal_vars *v) for (k = 0; k <= v->number_of_active_planes - 1; k++) { if (v->pte_enable == dcn_bw_yes && v->dcc_enable[k] == dcn_bw_yes) { v->time_for_meta_pte_with_immediate_flip =dcn_bw_max5(v->meta_pte_bytes_per_frame[k] / v->prefetch_bw[k], v->meta_pte_bytes_per_frame[k] * v->total_immediate_flip_bytes[k] / (v->bw_available_for_immediate_flip * (v->meta_pte_bytes_per_frame[k] + v->meta_row_bytes[k] + v->dpte_bytes_per_row[k])), v->extra_latency, v->urgent_latency, v->htotal[k] / v->pixel_clock[k] / 4.0); - v->time_for_meta_pte_without_immediate_flip =dcn_bw_max3(v->meta_pte_bytes_per_frame[k] / v->prefetch_bw[k], v->extra_latency, v->htotal[k] / v->pixel_clock[k] / 4.0); } else { v->time_for_meta_pte_with_immediate_flip = v->htotal[k] / v->pixel_clock[k] / 4.0; - v->time_for_meta_pte_without_immediate_flip = v->htotal[k] / v->pixel_clock[k] / 4.0; } if (v->pte_enable == dcn_bw_yes || v->dcc_enable[k] == dcn_bw_yes) { v->time_for_meta_and_dpte_row_with_immediate_flip =dcn_bw_max5((v->meta_row_bytes[k] + v->dpte_bytes_per_row[k]) / v->prefetch_bw[k], (v->meta_row_bytes[k] + v->dpte_bytes_per_row[k]) * v->total_immediate_flip_bytes[k] / (v->bw_available_for_immediate_flip * (v->meta_pte_bytes_per_frame[k] + v->meta_row_bytes[k] + v->dpte_bytes_per_row[k])), v->htotal[k] / v->pixel_clock[k] - v->time_for_meta_pte_with_immediate_flip, v->extra_latency, 2.0 * v->urgent_latency); - v->time_for_meta_and_dpte_row_without_immediate_flip =dcn_bw_max3((v->meta_row_bytes[k] + v->dpte_bytes_per_row[k]) / v->prefetch_bw[k], v->htotal[k] / v->pixel_clock[k] - v->time_for_meta_pte_without_immediate_flip, v->extra_latency); } else { v->time_for_meta_and_dpte_row_with_immediate_flip =dcn_bw_max2(v->htotal[k] / v->pixel_clock[k] - v->time_for_meta_pte_with_immediate_flip, v->extra_latency - v->time_for_meta_pte_with_immediate_flip); - v->time_for_meta_and_dpte_row_without_immediate_flip =dcn_bw_max2(v->htotal[k] / v->pixel_clock[k] - v->time_for_meta_pte_without_immediate_flip, v->extra_latency - v->time_for_meta_pte_without_immediate_flip); } v->lines_for_meta_pte_with_immediate_flip[k] =dcn_bw_floor2(4.0 * (v->time_for_meta_pte_with_immediate_flip / (v->htotal[k] / v->pixel_clock[k]) + 0.125), 1.0) / 4; - v->lines_for_meta_pte_without_immediate_flip[k] =dcn_bw_floor2(4.0 * (v->time_for_meta_pte_without_immediate_flip / (v->htotal[k] / v->pixel_clock[k]) + 0.125), 1.0) / 4; v->lines_for_meta_and_dpte_row_with_immediate_flip[k] =dcn_bw_floor2(4.0 * (v->time_for_meta_and_dpte_row_with_immediate_flip / (v->htotal[k] / v->pixel_clock[k]) + 0.125), 1.0) / 4; - v->lines_for_meta_and_dpte_row_without_immediate_flip[k] =dcn_bw_floor2(4.0 * (v->time_for_meta_and_dpte_row_without_immediate_flip / (v->htotal[k] / v->pixel_clock[k]) + 0.125), 1.0) / 4; v->line_times_to_request_prefetch_pixel_data_with_immediate_flip = v->line_times_for_prefetch[k] - v->lines_for_meta_pte_with_immediate_flip[k] - v->lines_for_meta_and_dpte_row_with_immediate_flip[k]; v->line_times_to_request_prefetch_pixel_data_without_immediate_flip = v->line_times_for_prefetch[k] - v->lines_for_meta_pte_without_immediate_flip[k] - v->lines_for_meta_and_dpte_row_without_immediate_flip[k]; if (v->line_times_to_request_prefetch_pixel_data_with_immediate_flip > 0.0) { diff --git a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c index 43e4a2be0fa6..12d1842079ae 100644 --- a/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +++ b/drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c @@ -290,41 +290,34 @@ static void pipe_ctx_to_e2e_pipe_params ( switch (pipe->plane_state->tiling_info.gfx9.swizzle) { /* for 4/8/16 high tiles */ case DC_SW_LINEAR: - input->src.is_display_sw = 1; input->src.macro_tile_size = dm_4k_tile; break; case DC_SW_4KB_S: case DC_SW_4KB_S_X: - input->src.is_display_sw = 0; input->src.macro_tile_size = dm_4k_tile; break; case DC_SW_64KB_S: case DC_SW_64KB_S_X: case DC_SW_64KB_S_T: - input->src.is_display_sw = 0; input->src.macro_tile_size = dm_64k_tile; break; case DC_SW_VAR_S: case DC_SW_VAR_S_X: - input->src.is_display_sw = 0; input->src.macro_tile_size = dm_256k_tile; break; /* For 64bpp 2 high tiles */ case DC_SW_4KB_D: case DC_SW_4KB_D_X: - input->src.is_display_sw = 1; input->src.macro_tile_size = dm_4k_tile; break; case DC_SW_64KB_D: case DC_SW_64KB_D_X: case DC_SW_64KB_D_T: - input->src.is_display_sw = 1; input->src.macro_tile_size = dm_64k_tile; break; case DC_SW_VAR_D: case DC_SW_VAR_D_X: - input->src.is_display_sw = 1; input->src.macro_tile_size = dm_256k_tile; break; @@ -423,7 +416,7 @@ static void pipe_ctx_to_e2e_pipe_params ( - pipe->stream->timing.v_addressable - pipe->stream->timing.v_border_bottom - pipe->stream->timing.v_border_top; - input->dest.pixel_rate_mhz = pipe->stream->timing.pix_clk_khz/1000.0; + input->dest.pixel_rate_mhz = pipe->stream->timing.pix_clk_100hz/10000.0; input->dest.vstartup_start = pipe->pipe_dlg_param.vstartup_start; input->dest.vupdate_offset = pipe->pipe_dlg_param.vupdate_offset; input->dest.vupdate_offset = pipe->pipe_dlg_param.vupdate_offset; @@ -670,9 +663,9 @@ static void hack_disable_optional_pipe_split(struct dcn_bw_internal_vars *v) } static void hack_force_pipe_split(struct dcn_bw_internal_vars *v, - unsigned int pixel_rate_khz) + unsigned int pixel_rate_100hz) { - float pixel_rate_mhz = pixel_rate_khz / 1000; + float pixel_rate_mhz = pixel_rate_100hz / 10000; /* * force enabling pipe split by lower dpp clock for DPM0 to just @@ -695,7 +688,7 @@ static void hack_bounding_box(struct dcn_bw_internal_vars *v, if (context->stream_count == 1 && dbg->force_single_disp_pipe_split) - hack_force_pipe_split(v, context->streams[0]->timing.pix_clk_khz); + hack_force_pipe_split(v, context->streams[0]->timing.pix_clk_100hz); } bool dcn_validate_bandwidth( @@ -852,7 +845,7 @@ bool dcn_validate_bandwidth( v->v_sync_plus_back_porch[input_idx] = pipe->stream->timing.v_total - v->vactive[input_idx] - pipe->stream->timing.v_front_porch; - v->pixel_clock[input_idx] = pipe->stream->timing.pix_clk_khz/1000.0; + v->pixel_clock[input_idx] = pipe->stream->timing.pix_clk_100hz/10000.0; if (pipe->stream->timing.timing_3d_format == TIMING_3D_FORMAT_HW_FRAME_PACKING) v->pixel_clock[input_idx] *= 2; if (!pipe->plane_state) { @@ -961,7 +954,7 @@ bool dcn_validate_bandwidth( v->dcc_rate[input_idx] = 1; /*TODO: Worst case? does this change?*/ v->output_format[input_idx] = pipe->stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR420 ? dcn_bw_420 : dcn_bw_444; - v->output[input_idx] = pipe->stream->sink->sink_signal == + v->output[input_idx] = pipe->stream->signal == SIGNAL_TYPE_HDMI_TYPE_A ? dcn_bw_hdmi : dcn_bw_dp; v->output_deep_color[input_idx] = dcn_bw_encoder_8bpc; if (v->output[input_idx] == dcn_bw_hdmi) { diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c index 5fd52094d459..52f838442e21 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c @@ -384,7 +384,7 @@ void dc_stream_set_dither_option(struct dc_stream_state *stream, enum dc_dither_option option) { struct bit_depth_reduction_params params; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; struct pipe_ctx *pipes = NULL; int i; @@ -451,7 +451,7 @@ bool dc_stream_program_csc_matrix(struct dc *dc, struct dc_stream_state *stream) pipes, stream->output_color_space, stream->csc_color_matrix.matrix, - pipes->plane_res.hubp->opp_id); + pipes->plane_res.hubp ? pipes->plane_res.hubp->opp_id : 0); ret = true; } } @@ -526,9 +526,8 @@ void dc_link_set_preferred_link_settings(struct dc *dc, for (i = 0; i < MAX_PIPES; i++) { pipe = &dc->current_state->res_ctx.pipe_ctx[i]; - if (pipe->stream && pipe->stream->sink - && pipe->stream->sink->link) { - if (pipe->stream->sink->link == link) + if (pipe->stream && pipe->stream->link) { + if (pipe->stream->link == link) break; } } @@ -586,9 +585,6 @@ static void destruct(struct dc *dc) if (dc->ctx->gpio_service) dal_gpio_service_destroy(&dc->ctx->gpio_service); - if (dc->ctx->i2caux) - dal_i2caux_destroy(&dc->ctx->i2caux); - if (dc->ctx->created_bios) dal_bios_parser_destroy(&dc->ctx->dc_bios); @@ -625,7 +621,6 @@ static bool construct(struct dc *dc, #endif enum dce_version dc_version = DCE_VERSION_UNKNOWN; - dc_dceip = kzalloc(sizeof(*dc_dceip), GFP_KERNEL); if (!dc_dceip) { dm_error("%s: failed to create dceip\n", __func__); @@ -670,6 +665,7 @@ static bool construct(struct dc *dc, dc_ctx->dc = dc; dc_ctx->asic_id = init_params->asic_id; dc_ctx->dc_sink_id_count = 0; + dc_ctx->dc_stream_id_count = 0; dc->ctx = dc_ctx; dc->current_state = dc_create_state(); @@ -709,14 +705,6 @@ static bool construct(struct dc *dc, dc_ctx->created_bios = true; } - /* Create I2C AUX */ - dc_ctx->i2caux = dal_i2caux_create(dc_ctx); - - if (!dc_ctx->i2caux) { - ASSERT_CRITICAL(false); - goto fail; - } - dc_ctx->perf_trace = dc_perf_trace_create(); if (!dc_ctx->perf_trace) { ASSERT_CRITICAL(false); @@ -840,6 +828,11 @@ alloc_fail: return NULL; } +void dc_init_callbacks(struct dc *dc, + const struct dc_callback_init *init_params) +{ +} + void dc_destroy(struct dc **dc) { destruct(*dc); @@ -875,8 +868,9 @@ static void program_timing_sync( struct dc *dc, struct dc_state *ctx) { - int i, j; + int i, j, k; int group_index = 0; + int num_group = 0; int pipe_count = dc->res_pool->pipe_count; struct pipe_ctx *unsynced_pipes[MAX_PIPES] = { NULL }; @@ -913,11 +907,11 @@ static void program_timing_sync( } } - /* set first unblanked pipe as master */ + /* set first pipe with plane as master */ for (j = 0; j < group_size; j++) { struct pipe_ctx *temp; - if (pipe_set[j]->stream_res.tg->funcs->is_blanked && !pipe_set[j]->stream_res.tg->funcs->is_blanked(pipe_set[j]->stream_res.tg)) { + if (pipe_set[j]->plane_state) { if (j == 0) break; @@ -928,9 +922,21 @@ static void program_timing_sync( } } - /* remove any other unblanked pipes as they have already been synced */ + + for (k = 0; k < group_size; k++) { + struct dc_stream_status *status = dc_stream_get_status_from_state(ctx, pipe_set[k]->stream); + + status->timing_sync_info.group_id = num_group; + status->timing_sync_info.group_size = group_size; + if (k == 0) + status->timing_sync_info.master = true; + else + status->timing_sync_info.master = false; + + } + /* remove any other pipes with plane as they have already been synced */ for (j = j + 1; j < group_size; j++) { - if (pipe_set[j]->stream_res.tg->funcs->is_blanked && !pipe_set[j]->stream_res.tg->funcs->is_blanked(pipe_set[j]->stream_res.tg)) { + if (pipe_set[j]->plane_state) { group_size--; pipe_set[j] = pipe_set[group_size]; j--; @@ -942,6 +948,7 @@ static void program_timing_sync( dc, group_index, group_size, pipe_set); group_index++; } + num_group++; } } @@ -962,6 +969,52 @@ static bool context_changed( return false; } +bool dc_validate_seamless_boot_timing(struct dc *dc, + const struct dc_sink *sink, + struct dc_crtc_timing *crtc_timing) +{ + struct timing_generator *tg; + struct dc_link *link = sink->link; + unsigned int inst; + + /* Check for enabled DIG to identify enabled display */ + if (!link->link_enc->funcs->is_dig_enabled(link->link_enc)) + return false; + + /* Check for which front end is used by this encoder. + * Note the inst is 1 indexed, where 0 is undefined. + * Note that DIG_FE can source from different OTG but our + * current implementation always map 1-to-1, so this code makes + * the same assumption and doesn't check OTG source. + */ + inst = link->link_enc->funcs->get_dig_frontend(link->link_enc) - 1; + + /* Instance should be within the range of the pool */ + if (inst >= dc->res_pool->pipe_count) + return false; + + tg = dc->res_pool->timing_generators[inst]; + + if (!tg->funcs->is_matching_timing) + return false; + + if (!tg->funcs->is_matching_timing(tg, crtc_timing)) + return false; + + if (dc_is_dp_signal(link->connector_signal)) { + unsigned int pix_clk_100hz; + + dc->res_pool->dp_clock_source->funcs->get_pixel_clk_frequency_100hz( + dc->res_pool->dp_clock_source, + inst, &pix_clk_100hz); + + if (crtc_timing->pix_clk_100hz != pix_clk_100hz) + return false; + } + + return true; +} + bool dc_enable_stereo( struct dc *dc, struct dc_state *context, @@ -1040,7 +1093,11 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c /* Program all planes within new context*/ for (i = 0; i < context->stream_count; i++) { - const struct dc_sink *sink = context->streams[i]->sink; + const struct dc_link *link = context->streams[i]->link; + struct dc_stream_status *status; + + if (context->streams[i]->apply_seamless_boot_optimization) + context->streams[i]->apply_seamless_boot_optimization = false; if (!context->streams[i]->mode_changed) continue; @@ -1065,12 +1122,15 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c } } - CONN_MSG_MODE(sink->link, "{%dx%d, %dx%d@%dKhz}", + status = dc_stream_get_status_from_state(context, context->streams[i]); + context->streams[i]->out.otg_offset = status->primary_otg_inst; + + CONN_MSG_MODE(link, "{%dx%d, %dx%d@%dKhz}", context->streams[i]->timing.h_addressable, context->streams[i]->timing.v_addressable, context->streams[i]->timing.h_total, context->streams[i]->timing.v_total, - context->streams[i]->timing.pix_clk_khz); + context->streams[i]->timing.pix_clk_100hz / 10); } dc_enable_stereo(dc, context, dc_streams, context->stream_count); @@ -1114,6 +1174,9 @@ bool dc_post_update_surfaces_to_stream(struct dc *dc) int i; struct dc_state *context = dc->current_state; + if (dc->optimized_required == false) + return true; + post_surface_trace(dc); for (i = 0; i < dc->res_pool->pipe_count; i++) @@ -1215,6 +1278,12 @@ static enum surface_update_type get_plane_info_update_type(const struct dc_surfa */ update_flags->bits.bpp_change = 1; + if (u->plane_info->plane_size.grph.surface_pitch != u->surface->plane_size.grph.surface_pitch + || u->plane_info->plane_size.video.luma_pitch != u->surface->plane_size.video.luma_pitch + || u->plane_info->plane_size.video.chroma_pitch != u->surface->plane_size.video.chroma_pitch) + update_flags->bits.plane_size_change = 1; + + if (memcmp(&u->plane_info->tiling_info, &u->surface->tiling_info, sizeof(union dc_tiling_info)) != 0) { update_flags->bits.swizzle_change = 1; @@ -1236,7 +1305,7 @@ static enum surface_update_type get_plane_info_update_type(const struct dc_surfa || update_flags->bits.output_tf_change) return UPDATE_TYPE_FULL; - return UPDATE_TYPE_MED; + return update_flags->raw ? UPDATE_TYPE_MED : UPDATE_TYPE_FAST; } static enum surface_update_type get_scaling_info_update_type( @@ -1436,6 +1505,101 @@ static struct dc_stream_status *stream_get_status( static const enum surface_update_type update_surface_trace_level = UPDATE_TYPE_FULL; +static void copy_surface_update_to_plane( + struct dc_plane_state *surface, + struct dc_surface_update *srf_update) +{ + if (srf_update->flip_addr) { + surface->address = srf_update->flip_addr->address; + surface->flip_immediate = + srf_update->flip_addr->flip_immediate; + surface->time.time_elapsed_in_us[surface->time.index] = + srf_update->flip_addr->flip_timestamp_in_us - + surface->time.prev_update_time_in_us; + surface->time.prev_update_time_in_us = + srf_update->flip_addr->flip_timestamp_in_us; + surface->time.index++; + if (surface->time.index >= DC_PLANE_UPDATE_TIMES_MAX) + surface->time.index = 0; + } + + if (srf_update->scaling_info) { + surface->scaling_quality = + srf_update->scaling_info->scaling_quality; + surface->dst_rect = + srf_update->scaling_info->dst_rect; + surface->src_rect = + srf_update->scaling_info->src_rect; + surface->clip_rect = + srf_update->scaling_info->clip_rect; + } + + if (srf_update->plane_info) { + surface->color_space = + srf_update->plane_info->color_space; + surface->format = + srf_update->plane_info->format; + surface->plane_size = + srf_update->plane_info->plane_size; + surface->rotation = + srf_update->plane_info->rotation; + surface->horizontal_mirror = + srf_update->plane_info->horizontal_mirror; + surface->stereo_format = + srf_update->plane_info->stereo_format; + surface->tiling_info = + srf_update->plane_info->tiling_info; + surface->visible = + srf_update->plane_info->visible; + surface->per_pixel_alpha = + srf_update->plane_info->per_pixel_alpha; + surface->global_alpha = + srf_update->plane_info->global_alpha; + surface->global_alpha_value = + srf_update->plane_info->global_alpha_value; + surface->dcc = + srf_update->plane_info->dcc; + surface->sdr_white_level = + srf_update->plane_info->sdr_white_level; + } + + if (srf_update->gamma && + (surface->gamma_correction != + srf_update->gamma)) { + memcpy(&surface->gamma_correction->entries, + &srf_update->gamma->entries, + sizeof(struct dc_gamma_entries)); + surface->gamma_correction->is_identity = + srf_update->gamma->is_identity; + surface->gamma_correction->num_entries = + srf_update->gamma->num_entries; + surface->gamma_correction->type = + srf_update->gamma->type; + } + + if (srf_update->in_transfer_func && + (surface->in_transfer_func != + srf_update->in_transfer_func)) { + surface->in_transfer_func->sdr_ref_white_level = + srf_update->in_transfer_func->sdr_ref_white_level; + surface->in_transfer_func->tf = + srf_update->in_transfer_func->tf; + surface->in_transfer_func->type = + srf_update->in_transfer_func->type; + memcpy(&surface->in_transfer_func->tf_pts, + &srf_update->in_transfer_func->tf_pts, + sizeof(struct dc_transfer_func_distributed_points)); + } + + if (srf_update->input_csc_color_matrix) + surface->input_csc_color_matrix = + *srf_update->input_csc_color_matrix; + + if (srf_update->coeff_reduction_factor) + surface->coeff_reduction_factor = + *srf_update->coeff_reduction_factor; +} + static void commit_planes_do_stream_update(struct dc *dc, struct dc_stream_state *stream, struct dc_stream_update *stream_update, @@ -1459,11 +1623,13 @@ static void commit_planes_do_stream_update(struct dc *dc, stream_update->adjust->v_total_min, stream_update->adjust->v_total_max); - if (stream_update->periodic_fn_vsync_delta && - pipe_ctx->stream_res.tg->funcs->program_vline_interrupt) + if (stream_update->periodic_vsync_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt) pipe_ctx->stream_res.tg->funcs->program_vline_interrupt( - pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, - pipe_ctx->stream->periodic_fn_vsync_delta); + pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, VLINE0, &stream->periodic_vsync_config); + + if (stream_update->enhanced_sync_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt) + pipe_ctx->stream_res.tg->funcs->program_vline_interrupt( + pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, VLINE1, &stream->enhanced_sync_config); if ((stream_update->hdr_static_metadata && !stream->use_dynamic_meta) || stream_update->vrr_infopacket || @@ -1605,7 +1771,6 @@ void dc_commit_updates_for_stream(struct dc *dc, int surface_count, struct dc_stream_state *stream, struct dc_stream_update *stream_update, - struct dc_plane_state **plane_states, struct dc_state *state) { const struct dc_stream_status *stream_status; @@ -1640,14 +1805,7 @@ void dc_commit_updates_for_stream(struct dc *dc, for (i = 0; i < surface_count; i++) { struct dc_plane_state *surface = srf_updates[i].surface; - /* TODO: On flip we don't build the state, so it still has the - * old address. Which is why we are updating the address here - */ - if (srf_updates[i].flip_addr) { - surface->address = srf_updates[i].flip_addr->address; - surface->flip_immediate = srf_updates[i].flip_addr->flip_immediate; - - } + copy_surface_update_to_plane(surface, &srf_updates[i]); if (update_type >= UPDATE_TYPE_MED) { for (j = 0; j < dc->res_pool->pipe_count; j++) { @@ -1764,6 +1922,26 @@ void dc_resume(struct dc *dc) core_link_resume(dc->links[i]); } +unsigned int dc_get_current_backlight_pwm(struct dc *dc) +{ + struct abm *abm = dc->res_pool->abm; + + if (abm) + return abm->funcs->get_current_backlight(abm); + + return 0; +} + +unsigned int dc_get_target_backlight_pwm(struct dc *dc) +{ + struct abm *abm = dc->res_pool->abm; + + if (abm) + return abm->funcs->get_target_backlight(abm); + + return 0; +} + bool dc_is_dmcu_initialized(struct dc *dc) { struct dmcu *dmcu = dc->res_pool->dmcu; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link.c b/drivers/gpu/drm/amd/display/dc/core/dc_link.c index 52deacf39841..7f5a947ad31d 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_link.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_link.c @@ -43,10 +43,6 @@ #include "dpcd_defs.h" #include "dmcu.h" -#include "dce/dce_11_0_d.h" -#include "dce/dce_11_0_enum.h" -#include "dce/dce_11_0_sh_mask.h" - #define DC_LOGGER_INIT(logger) @@ -80,6 +76,12 @@ static void destruct(struct dc_link *link) { int i; + if (link->hpd_gpio != NULL) { + dal_gpio_close(link->hpd_gpio); + dal_gpio_destroy_irq(&link->hpd_gpio); + link->hpd_gpio = NULL; + } + if (link->ddc) dal_ddc_service_destroy(&link->ddc); @@ -789,7 +791,7 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason) return false; } - sink->dongle_max_pix_clk = sink_caps.max_hdmi_pixel_clock; + sink->link->dongle_max_pix_clk = sink_caps.max_hdmi_pixel_clock; sink->converter_disable_audio = converter_disable_audio; link->local_sink = sink; @@ -935,18 +937,11 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason) bool dc_link_get_hpd_state(struct dc_link *dc_link) { - struct gpio *hpd_pin; uint32_t state; - hpd_pin = get_hpd_gpio(dc_link->ctx->dc_bios, - dc_link->link_id, dc_link->ctx->gpio_service); - if (hpd_pin == NULL) - ASSERT(false); - - dal_gpio_open(hpd_pin, GPIO_MODE_INTERRUPT); - dal_gpio_get_value(hpd_pin, &state); - dal_gpio_close(hpd_pin); - dal_gpio_destroy_irq(&hpd_pin); + dal_gpio_lock_pin(dc_link->hpd_gpio); + dal_gpio_get_value(dc_link->hpd_gpio, &state); + dal_gpio_unlock_pin(dc_link->hpd_gpio); return state; } @@ -1102,7 +1097,6 @@ static bool construct( const struct link_init_data *init_params) { uint8_t i; - struct gpio *hpd_gpio = NULL; struct ddc_service_init_data ddc_service_init_data = { { 0 } }; struct dc_context *dc_ctx = init_params->ctx; struct encoder_init_data enc_init_data = { 0 }; @@ -1132,10 +1126,12 @@ static bool construct( if (link->dc->res_pool->funcs->link_init) link->dc->res_pool->funcs->link_init(link); - hpd_gpio = get_hpd_gpio(link->ctx->dc_bios, link->link_id, link->ctx->gpio_service); - - if (hpd_gpio != NULL) - link->irq_source_hpd = dal_irq_get_source(hpd_gpio); + link->hpd_gpio = get_hpd_gpio(link->ctx->dc_bios, link->link_id, link->ctx->gpio_service); + if (link->hpd_gpio != NULL) { + dal_gpio_open(link->hpd_gpio, GPIO_MODE_INTERRUPT); + dal_gpio_unlock_pin(link->hpd_gpio); + link->irq_source_hpd = dal_irq_get_source(link->hpd_gpio); + } switch (link->link_id.id) { case CONNECTOR_ID_HDMI_TYPE_A: @@ -1153,18 +1149,18 @@ static bool construct( case CONNECTOR_ID_DISPLAY_PORT: link->connector_signal = SIGNAL_TYPE_DISPLAY_PORT; - if (hpd_gpio != NULL) + if (link->hpd_gpio != NULL) link->irq_source_hpd_rx = - dal_irq_get_rx_source(hpd_gpio); + dal_irq_get_rx_source(link->hpd_gpio); break; case CONNECTOR_ID_EDP: link->connector_signal = SIGNAL_TYPE_EDP; - if (hpd_gpio != NULL) { + if (link->hpd_gpio != NULL) { link->irq_source_hpd = DC_IRQ_SOURCE_INVALID; link->irq_source_hpd_rx = - dal_irq_get_rx_source(hpd_gpio); + dal_irq_get_rx_source(link->hpd_gpio); } break; case CONNECTOR_ID_LVDS: @@ -1175,10 +1171,7 @@ static bool construct( goto create_fail; } - if (hpd_gpio != NULL) { - dal_gpio_destroy_irq(&hpd_gpio); - hpd_gpio = NULL; - } + /* TODO: #DAL3 Implement id to str function.*/ LINK_INFO("Connector[%d] description:" @@ -1281,8 +1274,9 @@ link_enc_create_fail: ddc_create_fail: create_fail: - if (hpd_gpio != NULL) { - dal_gpio_destroy_irq(&hpd_gpio); + if (link->hpd_gpio != NULL) { + dal_gpio_destroy_irq(&link->hpd_gpio); + link->hpd_gpio = NULL; } return false; @@ -1372,7 +1366,7 @@ static void dpcd_configure_panel_mode( static void enable_stream_features(struct pipe_ctx *pipe_ctx) { struct dc_stream_state *stream = pipe_ctx->stream; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; union down_spread_ctrl old_downspread; union down_spread_ctrl new_downspread; @@ -1397,7 +1391,7 @@ static enum dc_status enable_link_dp( struct dc_stream_state *stream = pipe_ctx->stream; enum dc_status status; bool skip_video_pattern; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; struct dc_link_settings link_settings = {0}; enum dp_panel_mode panel_mode; @@ -1414,8 +1408,8 @@ static enum dc_status enable_link_dp( pipe_ctx->clock_source->id, &link_settings); - if (stream->sink->edid_caps.panel_patch.dppowerup_delay > 0) { - int delay_dp_power_up_in_ms = stream->sink->edid_caps.panel_patch.dppowerup_delay; + if (stream->sink_patches.dppowerup_delay > 0) { + int delay_dp_power_up_in_ms = stream->sink_patches.dppowerup_delay; msleep(delay_dp_power_up_in_ms); } @@ -1448,7 +1442,7 @@ static enum dc_status enable_link_edp( { enum dc_status status; struct dc_stream_state *stream = pipe_ctx->stream; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; /*in case it is not on*/ link->dc->hwss.edp_power_control(link, true); link->dc->hwss.edp_wait_for_hpd_ready(link, true); @@ -1463,7 +1457,7 @@ static enum dc_status enable_link_dp_mst( struct dc_state *state, struct pipe_ctx *pipe_ctx) { - struct dc_link *link = pipe_ctx->stream->sink->link; + struct dc_link *link = pipe_ctx->stream->link; /* sink signal type after MST branch is MST. Multiple MST sinks * share one link. Link DP PHY is enable or training only once. @@ -1471,6 +1465,11 @@ static enum dc_status enable_link_dp_mst( if (link->cur_link_settings.lane_count != LANE_COUNT_UNKNOWN) return DC_OK; + /* to make sure the pending down rep can be processed + * before clear payload table + */ + dm_helpers_dp_mst_poll_pending_down_reply(link->ctx, link); + /* clear payload table */ dm_helpers_dp_mst_clear_payload_allocation_table(link->ctx, link); @@ -1597,7 +1596,7 @@ static bool i2c_write(struct pipe_ctx *pipe_ctx, cmd.payloads = &payload; if (dm_helpers_submit_i2c(pipe_ctx->stream->ctx, - pipe_ctx->stream->sink->link, &cmd)) + pipe_ctx->stream->link, &cmd)) return true; return false; @@ -1651,7 +1650,7 @@ static void write_i2c_retimer_setting( else { i2c_success = dal_ddc_service_query_ddc_data( - pipe_ctx->stream->sink->link->ddc, + pipe_ctx->stream->link->ddc, slave_address, &offset, 1, &value, 1); if (!i2c_success) /* Write failure */ @@ -1704,7 +1703,7 @@ static void write_i2c_retimer_setting( else { i2c_success = dal_ddc_service_query_ddc_data( - pipe_ctx->stream->sink->link->ddc, + pipe_ctx->stream->link->ddc, slave_address, &offset, 1, &value, 1); if (!i2c_success) /* Write failure */ @@ -1929,7 +1928,7 @@ static void write_i2c_redriver_setting( static void enable_link_hdmi(struct pipe_ctx *pipe_ctx) { struct dc_stream_state *stream = pipe_ctx->stream; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; enum dc_color_depth display_color_depth; enum engine_id eng_id; struct ext_hdmi_settings settings = {0}; @@ -1938,12 +1937,12 @@ static void enable_link_hdmi(struct pipe_ctx *pipe_ctx) && (stream->timing.v_addressable == 480); if (stream->phy_pix_clk == 0) - stream->phy_pix_clk = stream->timing.pix_clk_khz; + stream->phy_pix_clk = stream->timing.pix_clk_100hz / 10; if (stream->phy_pix_clk > 340000) is_over_340mhz = true; if (dc_is_hdmi_signal(pipe_ctx->stream->signal)) { - unsigned short masked_chip_caps = pipe_ctx->stream->sink->link->chip_caps & + unsigned short masked_chip_caps = pipe_ctx->stream->link->chip_caps & EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK; if (masked_chip_caps == EXT_DISPLAY_PATH_CAPS__HDMI20_TISN65DP159RSBT) { /* DP159, Retimer settings */ @@ -1964,11 +1963,11 @@ static void enable_link_hdmi(struct pipe_ctx *pipe_ctx) if (dc_is_hdmi_signal(pipe_ctx->stream->signal)) dal_ddc_service_write_scdc_data( - stream->sink->link->ddc, + stream->link->ddc, stream->phy_pix_clk, stream->timing.flags.LTE_340MCSC_SCRAMBLE); - memset(&stream->sink->link->cur_link_settings, 0, + memset(&stream->link->cur_link_settings, 0, sizeof(struct dc_link_settings)); display_color_depth = stream->timing.display_color_depth; @@ -1989,12 +1988,12 @@ static void enable_link_hdmi(struct pipe_ctx *pipe_ctx) static void enable_link_lvds(struct pipe_ctx *pipe_ctx) { struct dc_stream_state *stream = pipe_ctx->stream; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; if (stream->phy_pix_clk == 0) - stream->phy_pix_clk = stream->timing.pix_clk_khz; + stream->phy_pix_clk = stream->timing.pix_clk_100hz / 10; - memset(&stream->sink->link->cur_link_settings, 0, + memset(&stream->link->cur_link_settings, 0, sizeof(struct dc_link_settings)); link->link_enc->funcs->enable_lvds_output( @@ -2067,7 +2066,7 @@ static bool dp_active_dongle_validate_timing( const struct dc_crtc_timing *timing, const struct dpcd_caps *dpcd_caps) { - unsigned int required_pix_clk = timing->pix_clk_khz; + unsigned int required_pix_clk_100hz = timing->pix_clk_100hz; const struct dc_dongle_caps *dongle_caps = &dpcd_caps->dongle_caps; switch (dpcd_caps->dongle_type) { @@ -2107,9 +2106,9 @@ static bool dp_active_dongle_validate_timing( /* Check Color Depth and Pixel Clock */ if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR420) - required_pix_clk /= 2; + required_pix_clk_100hz /= 2; else if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) - required_pix_clk = required_pix_clk * 2 / 3; + required_pix_clk_100hz = required_pix_clk_100hz * 2 / 3; switch (timing->display_color_depth) { case COLOR_DEPTH_666: @@ -2119,12 +2118,12 @@ static bool dp_active_dongle_validate_timing( case COLOR_DEPTH_101010: if (dongle_caps->dp_hdmi_max_bpc < 10) return false; - required_pix_clk = required_pix_clk * 10 / 8; + required_pix_clk_100hz = required_pix_clk_100hz * 10 / 8; break; case COLOR_DEPTH_121212: if (dongle_caps->dp_hdmi_max_bpc < 12) return false; - required_pix_clk = required_pix_clk * 12 / 8; + required_pix_clk_100hz = required_pix_clk_100hz * 12 / 8; break; case COLOR_DEPTH_141414: @@ -2134,7 +2133,7 @@ static bool dp_active_dongle_validate_timing( return false; } - if (required_pix_clk > dongle_caps->dp_hdmi_max_pixel_clk) + if (required_pix_clk_100hz > (dongle_caps->dp_hdmi_max_pixel_clk * 10)) return false; return true; @@ -2145,7 +2144,7 @@ enum dc_status dc_link_validate_mode_timing( struct dc_link *link, const struct dc_crtc_timing *timing) { - uint32_t max_pix_clk = stream->sink->dongle_max_pix_clk; + uint32_t max_pix_clk = stream->link->dongle_max_pix_clk * 10; struct dpcd_caps *dpcd_caps = &link->dpcd_caps; /* A hack to avoid failing any modes for EDID override feature on @@ -2155,7 +2154,7 @@ enum dc_status dc_link_validate_mode_timing( return DC_OK; /* Passive Dongle */ - if (0 != max_pix_clk && timing->pix_clk_khz > max_pix_clk) + if (0 != max_pix_clk && timing->pix_clk_100hz > max_pix_clk) return DC_EXCEED_DONGLE_CAP; /* Active Dongle*/ @@ -2190,8 +2189,7 @@ int dc_link_get_backlight_level(const struct dc_link *link) bool dc_link_set_backlight_level(const struct dc_link *link, uint32_t backlight_pwm_u16_16, - uint32_t frame_ramp, - const struct dc_stream_state *stream) + uint32_t frame_ramp) { struct dc *core_dc = link->ctx->dc; struct abm *abm = core_dc->res_pool->abm; @@ -2206,10 +2204,6 @@ bool dc_link_set_backlight_level(const struct dc_link *link, (abm->funcs->set_backlight_level_pwm == NULL)) return false; - if (stream) - ((struct dc_stream_state *)stream)->bl_pwm_level = - backlight_pwm_u16_16; - use_smooth_brightness = dmcu->funcs->is_dmcu_initialized(dmcu); DC_LOG_BACKLIGHT("New Backlight level: %d (0x%X)\n", @@ -2219,7 +2213,7 @@ bool dc_link_set_backlight_level(const struct dc_link *link, for (i = 0; i < MAX_PIPES; i++) { if (core_dc->current_state->res_ctx.pipe_ctx[i].stream) { if (core_dc->current_state->res_ctx. - pipe_ctx[i].stream->sink->link + pipe_ctx[i].stream->link == link) /* DMCU -1 for all controller id values, * therefore +1 here @@ -2279,7 +2273,7 @@ void core_link_resume(struct dc_link *link) static struct fixed31_32 get_pbn_per_slot(struct dc_stream_state *stream) { struct dc_link_settings *link_settings = - &stream->sink->link->cur_link_settings; + &stream->link->cur_link_settings; uint32_t link_rate_in_mbps = link_settings->link_rate * LINK_RATE_REF_FREQ_IN_MHZ; struct fixed31_32 mbps = dc_fixpt_from_int( @@ -2310,7 +2304,7 @@ static struct fixed31_32 get_pbn_from_timing(struct pipe_ctx *pipe_ctx) uint32_t denominator; bpc = get_color_depth(pipe_ctx->stream_res.pix_clk_params.color_depth); - kbps = pipe_ctx->stream_res.pix_clk_params.requested_pix_clk * bpc * 3; + kbps = pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz / 10 * bpc * 3; /* * margin 5300ppm + 300ppm ~ 0.6% as per spec, factor is 1.006 @@ -2386,7 +2380,7 @@ static void update_mst_stream_alloc_table( static enum dc_status allocate_mst_payload(struct pipe_ctx *pipe_ctx) { struct dc_stream_state *stream = pipe_ctx->stream; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; struct link_encoder *link_encoder = link->link_enc; struct stream_encoder *stream_encoder = pipe_ctx->stream_res.stream_enc; struct dp_mst_stream_allocation_table proposed_table = {0}; @@ -2466,7 +2460,7 @@ static enum dc_status allocate_mst_payload(struct pipe_ctx *pipe_ctx) static enum dc_status deallocate_mst_payload(struct pipe_ctx *pipe_ctx) { struct dc_stream_state *stream = pipe_ctx->stream; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; struct link_encoder *link_encoder = link->link_enc; struct stream_encoder *stream_encoder = pipe_ctx->stream_res.stream_enc; struct dp_mst_stream_allocation_table proposed_table = {0}; @@ -2551,8 +2545,8 @@ void core_link_enable_stream( DC_LOGGER_INIT(pipe_ctx->stream->ctx->logger); if (pipe_ctx->stream->signal != SIGNAL_TYPE_VIRTUAL) { - stream->sink->link->link_enc->funcs->setup( - stream->sink->link->link_enc, + stream->link->link_enc->funcs->setup( + stream->link->link_enc, pipe_ctx->stream->signal); pipe_ctx->stream_res.stream_enc->funcs->setup_stereo_sync( pipe_ctx->stream_res.stream_enc, @@ -2586,13 +2580,23 @@ void core_link_enable_stream( &stream->timing); if (!IS_FPGA_MAXIMUS_DC(core_dc->ctx->dce_environment)) { + bool apply_edp_fast_boot_optimization = + pipe_ctx->stream->apply_edp_fast_boot_optimization; + + pipe_ctx->stream->apply_edp_fast_boot_optimization = false; + resource_build_info_frame(pipe_ctx); core_dc->hwss.update_info_frame(pipe_ctx); + /* Do not touch link on seamless boot optimization. */ + if (pipe_ctx->stream->apply_seamless_boot_optimization) { + pipe_ctx->stream->dpms_off = false; + return; + } + /* eDP lit up by bios already, no need to enable again. */ if (pipe_ctx->stream->signal == SIGNAL_TYPE_EDP && - pipe_ctx->stream->apply_edp_fast_boot_optimization) { - pipe_ctx->stream->apply_edp_fast_boot_optimization = false; + apply_edp_fast_boot_optimization) { pipe_ctx->stream->dpms_off = false; return; } @@ -2604,7 +2608,7 @@ void core_link_enable_stream( if (status != DC_OK) { DC_LOG_WARNING("enabling link %u failed: %d\n", - pipe_ctx->stream->sink->link->link_index, + pipe_ctx->stream->link->link_index, status); /* Abort stream enable *unless* the failure was due to @@ -2619,6 +2623,8 @@ void core_link_enable_stream( } } + stream->link->link_status.link_active = true; + core_dc->hwss.enable_audio_stream(pipe_ctx); /* turn off otg test pattern if enable */ @@ -2633,15 +2639,10 @@ void core_link_enable_stream( allocate_mst_payload(pipe_ctx); core_dc->hwss.unblank_stream(pipe_ctx, - &pipe_ctx->stream->sink->link->cur_link_settings); + &pipe_ctx->stream->link->cur_link_settings); if (dc_is_dp_signal(pipe_ctx->stream->signal)) enable_stream_features(pipe_ctx); - - dc_link_set_backlight_level(pipe_ctx->stream->sink->link, - pipe_ctx->stream->bl_pwm_level, - 0, - pipe_ctx->stream); } } @@ -2657,7 +2658,9 @@ void core_link_disable_stream(struct pipe_ctx *pipe_ctx, int option) core_dc->hwss.disable_stream(pipe_ctx, option); - disable_link(pipe_ctx->stream->sink->link, pipe_ctx->stream->signal); + disable_link(pipe_ctx->stream->link, pipe_ctx->stream->signal); + + pipe_ctx->stream->link->link_status.link_active = false; } void core_link_set_avmute(struct pipe_ctx *pipe_ctx, bool enable) diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_ddc.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_ddc.c index 506a97e16956..b7ee63cd8dc7 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_link_ddc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_ddc.c @@ -33,7 +33,7 @@ #include "include/vector.h" #include "core_types.h" #include "dc_link_ddc.h" -#include "aux_engine.h" +#include "dce/dce_aux.h" #define AUX_POWER_UP_WA_DELAY 500 #define I2C_OVER_AUX_DEFER_WA_DELAY 70 @@ -42,7 +42,6 @@ #define CV_SMART_DONGLE_ADDRESS 0x20 /* DVI-HDMI dongle slave address for retrieving dongle signature*/ #define DVI_HDMI_DONGLE_ADDRESS 0x68 -static const int8_t dvi_hdmi_dongle_signature_str[] = "6140063500G"; struct dvi_hdmi_dongle_signature_data { int8_t vendor[3];/* "AMD" */ uint8_t version[2]; @@ -165,43 +164,6 @@ static void dal_ddc_i2c_payloads_destroy(struct i2c_payloads **p) } -static struct aux_payloads *dal_ddc_aux_payloads_create(struct dc_context *ctx, uint32_t count) -{ - struct aux_payloads *payloads; - - payloads = kzalloc(sizeof(struct aux_payloads), GFP_KERNEL); - - if (!payloads) - return NULL; - - if (dal_vector_construct( - &payloads->payloads, ctx, count, sizeof(struct aux_payload))) - return payloads; - - kfree(payloads); - return NULL; -} - -static struct aux_payload *dal_ddc_aux_payloads_get(struct aux_payloads *p) -{ - return (struct aux_payload *)p->payloads.container; -} - -static uint32_t dal_ddc_aux_payloads_get_count(struct aux_payloads *p) -{ - return p->payloads.count; -} - -static void dal_ddc_aux_payloads_destroy(struct aux_payloads **p) -{ - if (!p || !*p) - return; - - dal_vector_destruct(&(*p)->payloads); - kfree(*p); - *p = NULL; -} - #define DDC_MIN(a, b) (((a) < (b)) ? (a) : (b)) void dal_ddc_i2c_payloads_add( @@ -225,27 +187,6 @@ void dal_ddc_i2c_payloads_add( } -void dal_ddc_aux_payloads_add( - struct aux_payloads *payloads, - uint32_t address, - uint32_t len, - uint8_t *data, - bool write) -{ - uint32_t payload_size = DEFAULT_AUX_MAX_DATA_SIZE; - uint32_t pos; - - for (pos = 0; pos < len; pos += payload_size) { - struct aux_payload payload = { - .i2c_over_aux = true, - .write = write, - .address = address, - .length = DDC_MIN(payload_size, len - pos), - .data = data + pos }; - dal_vector_append(&payloads->payloads, &payload); - } -} - static void construct( struct ddc_service *ddc_service, struct ddc_service_init_data *init_data) @@ -574,32 +515,34 @@ bool dal_ddc_service_query_ddc_data( /*TODO: len of payload data for i2c and aux is uint8!!!!, * but we want to read 256 over i2c!!!!*/ if (dal_ddc_service_is_in_aux_transaction_mode(ddc)) { - - struct aux_payloads *payloads = - dal_ddc_aux_payloads_create(ddc->ctx, payloads_num); - - struct aux_command command = { - .payloads = dal_ddc_aux_payloads_get(payloads), - .number_of_payloads = 0, + struct aux_payload write_payload = { + .i2c_over_aux = true, + .write = true, + .mot = true, + .address = address, + .length = write_size, + .data = write_buf, + .reply = NULL, .defer_delay = get_defer_delay(ddc), - .max_defer_write_retry = 0 }; + }; - dal_ddc_aux_payloads_add( - payloads, address, write_size, write_buf, true); - - dal_ddc_aux_payloads_add( - payloads, address, read_size, read_buf, false); - - command.number_of_payloads = - dal_ddc_aux_payloads_get_count(payloads); + struct aux_payload read_payload = { + .i2c_over_aux = true, + .write = false, + .mot = false, + .address = address, + .length = read_size, + .data = read_buf, + .reply = NULL, + .defer_delay = get_defer_delay(ddc), + }; - ret = dal_i2caux_submit_aux_command( - ddc->ctx->i2caux, - ddc->ddc_pin, - &command); + ret = dc_link_aux_transfer_with_retries(ddc, &write_payload); - dal_ddc_aux_payloads_destroy(&payloads); + if (!ret) + return false; + ret = dc_link_aux_transfer_with_retries(ddc, &read_payload); } else { struct i2c_payloads *payloads = dal_ddc_i2c_payloads_create(ddc->ctx, payloads_num); @@ -631,56 +574,15 @@ bool dal_ddc_service_query_ddc_data( } int dc_link_aux_transfer(struct ddc_service *ddc, - unsigned int address, - uint8_t *reply, - void *buffer, - unsigned int size, - enum aux_transaction_type type, - enum i2caux_transaction_action action) + struct aux_payload *payload) { - struct ddc *ddc_pin = ddc->ddc_pin; - struct aux_engine *aux_engine; - enum aux_channel_operation_result operation_result; - struct aux_request_transaction_data aux_req; - struct aux_reply_transaction_data aux_rep; - uint8_t returned_bytes = 0; - int res = -1; - uint32_t status; - - memset(&aux_req, 0, sizeof(aux_req)); - memset(&aux_rep, 0, sizeof(aux_rep)); - - aux_engine = ddc->ctx->dc->res_pool->engines[ddc_pin->pin_data->en]; - aux_engine->funcs->acquire(aux_engine, ddc_pin); - - aux_req.type = type; - aux_req.action = action; - - aux_req.address = address; - aux_req.delay = 0; - aux_req.length = size; - aux_req.data = buffer; - - aux_engine->funcs->submit_channel_request(aux_engine, &aux_req); - operation_result = aux_engine->funcs->get_channel_status(aux_engine, &returned_bytes); - - switch (operation_result) { - case AUX_CHANNEL_OPERATION_SUCCEEDED: - res = aux_engine->funcs->read_channel_reply(aux_engine, size, - buffer, reply, - &status); - break; - case AUX_CHANNEL_OPERATION_FAILED_HPD_DISCON: - res = 0; - break; - case AUX_CHANNEL_OPERATION_FAILED_REASON_UNKNOWN: - case AUX_CHANNEL_OPERATION_FAILED_INVALID_REPLY: - case AUX_CHANNEL_OPERATION_FAILED_TIMEOUT: - res = -1; - break; - } - aux_engine->funcs->release_engine(aux_engine); - return res; + return dce_aux_transfer(ddc, payload); +} + +bool dc_link_aux_transfer_with_retries(struct ddc_service *ddc, + struct aux_payload *payload) +{ + return dce_aux_transfer_with_retries(ddc, payload); } /*test only function*/ diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c index 0caacb60b02f..09d301216076 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c @@ -49,6 +49,8 @@ static void wait_for_training_aux_rd_interval( { union training_aux_rd_interval training_rd_interval; + memset(&training_rd_interval, 0, sizeof(training_rd_interval)); + /* overwrite the delay if rev > 1.1*/ if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_12) { /* DP 1.2 or later - retrieve delay through @@ -117,6 +119,13 @@ static void dpcd_set_link_settings( core_link_write_dpcd(link, DP_DOWNSPREAD_CTRL, &downspread.raw, sizeof(downspread)); + if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_14 && + (link->dpcd_caps.link_rate_set >= 1 && + link->dpcd_caps.link_rate_set <= 8)) { + core_link_write_dpcd(link, DP_LINK_RATE_SET, + &link->dpcd_caps.link_rate_set, 1); + } + DC_LOG_HW_LINK_TRAINING("%s\n %x rate = %x\n %x lane = %x\n %x spread = %x\n", __func__, DP_LINK_BW_SET, @@ -1542,7 +1551,7 @@ static uint32_t bandwidth_in_kbps_from_timing( ASSERT(bits_per_channel != 0); - kbps = timing->pix_clk_khz; + kbps = timing->pix_clk_100hz / 10; kbps *= bits_per_channel; if (timing->flags.Y_ONLY != 1) { @@ -1584,7 +1593,7 @@ bool dp_validate_mode_timing( const struct dc_link_settings *link_setting; /*always DP fail safe mode*/ - if (timing->pix_clk_khz == (uint32_t) 25175 && + if ((timing->pix_clk_100hz / 10) == (uint32_t) 25175 && timing->h_addressable == (uint32_t) 640 && timing->v_addressable == (uint32_t) 480) return true; @@ -1634,7 +1643,7 @@ void decide_link_settings(struct dc_stream_state *stream, req_bw = bandwidth_in_kbps_from_timing(&stream->timing); - link = stream->sink->link; + link = stream->link; /* if preferred is specified through AMDDP, use it, if it's enough * to drive the mode @@ -1656,7 +1665,7 @@ void decide_link_settings(struct dc_stream_state *stream, } /* EDP use the link cap setting */ - if (stream->sink->sink_signal == SIGNAL_TYPE_EDP) { + if (link->connector_signal == SIGNAL_TYPE_EDP) { *link_setting = link->verified_link_cap; return; } @@ -2002,11 +2011,7 @@ static void handle_automated_test(struct dc_link *link) dp_test_send_phy_test_pattern(link); test_response.bits.ACK = 1; } - if (!test_request.raw) - /* no requests, revert all test signals - * TODO: revert all test signals - */ - test_response.bits.ACK = 1; + /* send request acknowledgment */ if (test_response.bits.ACK) core_link_write_dpcd( @@ -2493,13 +2498,72 @@ bool detect_dp_sink_caps(struct dc_link *link) /* TODO save sink caps in link->sink */ } +enum dc_link_rate linkRateInKHzToLinkRateMultiplier(uint32_t link_rate_in_khz) +{ + enum dc_link_rate link_rate; + // LinkRate is normally stored as a multiplier of 0.27 Gbps per lane. Do the translation. + switch (link_rate_in_khz) { + case 1620000: + link_rate = LINK_RATE_LOW; // Rate_1 (RBR) - 1.62 Gbps/Lane + break; + case 2160000: + link_rate = LINK_RATE_RATE_2; // Rate_2 - 2.16 Gbps/Lane + break; + case 2430000: + link_rate = LINK_RATE_RATE_3; // Rate_3 - 2.43 Gbps/Lane + break; + case 2700000: + link_rate = LINK_RATE_HIGH; // Rate_4 (HBR) - 2.70 Gbps/Lane + break; + case 3240000: + link_rate = LINK_RATE_RBR2; // Rate_5 (RBR2) - 3.24 Gbps/Lane + break; + case 4320000: + link_rate = LINK_RATE_RATE_6; // Rate_6 - 4.32 Gbps/Lane + break; + case 5400000: + link_rate = LINK_RATE_HIGH2; // Rate_7 (HBR2) - 5.40 Gbps/Lane + break; + case 8100000: + link_rate = LINK_RATE_HIGH3; // Rate_8 (HBR3) - 8.10 Gbps/Lane + break; + default: + link_rate = LINK_RATE_UNKNOWN; + break; + } + return link_rate; +} + void detect_edp_sink_caps(struct dc_link *link) { - retrieve_link_cap(link); + uint8_t supported_link_rates[16] = {0}; + uint32_t entry; + uint32_t link_rate_in_khz; + enum dc_link_rate link_rate = LINK_RATE_UNKNOWN; - if (link->reported_link_cap.link_rate == LINK_RATE_UNKNOWN) - link->reported_link_cap.link_rate = LINK_RATE_HIGH2; + retrieve_link_cap(link); + if (link->dpcd_caps.dpcd_rev.raw >= DPCD_REV_14) { + // Read DPCD 00010h - 0001Fh 16 bytes at one shot + core_link_read_dpcd(link, DP_SUPPORTED_LINK_RATES, + supported_link_rates, sizeof(supported_link_rates)); + + link->dpcd_caps.link_rate_set = 0; + for (entry = 0; entry < 16; entry += 2) { + // DPCD register reports per-lane link rate = 16-bit link rate capability + // value X 200 kHz. Need multipler to find link rate in kHz. + link_rate_in_khz = (supported_link_rates[entry+1] * 0x100 + + supported_link_rates[entry]) * 200; + + if (link_rate_in_khz != 0) { + link_rate = linkRateInKHzToLinkRateMultiplier(link_rate_in_khz); + if (link->reported_link_cap.link_rate < link_rate) { + link->reported_link_cap.link_rate = link_rate; + link->dpcd_caps.link_rate_set = entry; + } + } + } + } link->verified_link_cap = link->reported_link_cap; } @@ -2621,7 +2685,7 @@ bool dc_link_dp_set_test_pattern( memset(&training_pattern, 0, sizeof(training_pattern)); for (i = 0; i < MAX_PIPES; i++) { - if (pipes[i].stream->sink->link == link) { + if (pipes[i].stream->link == link) { pipe_ctx = &pipes[i]; break; } diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c index 0065ec7d5330..f7f7515f65f4 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c @@ -70,13 +70,12 @@ void dp_enable_link_phy( */ for (i = 0; i < MAX_PIPES; i++) { if (pipes[i].stream != NULL && - pipes[i].stream->sink != NULL && - pipes[i].stream->sink->link == link) { + pipes[i].stream->link == link) { if (pipes[i].clock_source != NULL && pipes[i].clock_source->id != CLOCK_SOURCE_ID_DP_DTO) { pipes[i].clock_source = dp_cs; - pipes[i].stream_res.pix_clk_params.requested_pix_clk = - pipes[i].stream->timing.pix_clk_khz; + pipes[i].stream_res.pix_clk_params.requested_pix_clk_100hz = + pipes[i].stream->timing.pix_clk_100hz; pipes[i].clock_source->funcs->program_pix_clk( pipes[i].clock_source, &pipes[i].stream_res.pix_clk_params, @@ -120,6 +119,10 @@ bool edp_receiver_ready_T9(struct dc_link *link) break; udelay(100); //MAx T9 } while (++tries < 50); + + if (link->local_sink->edid_caps.panel_patch.extra_delay_backlight_off > 0) + udelay(link->local_sink->edid_caps.panel_patch.extra_delay_backlight_off * 1000); + return result; } bool edp_receiver_ready_T7(struct dc_link *link) @@ -279,10 +282,8 @@ void dp_retrain_link_dp_test(struct dc_link *link, for (i = 0; i < MAX_PIPES; i++) { if (pipes[i].stream != NULL && !pipes[i].top_pipe && - pipes[i].stream->sink != NULL && - pipes[i].stream->sink->link != NULL && - pipes[i].stream_res.stream_enc != NULL && - pipes[i].stream->sink->link == link) { + pipes[i].stream->link != NULL && + pipes[i].stream_res.stream_enc != NULL) { udelay(100); pipes[i].stream_res.stream_enc->funcs->dp_blank( diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c index 76137df74a53..349ab8017776 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c @@ -355,8 +355,8 @@ bool resource_are_streams_timing_synchronizable( != stream2->timing.v_addressable) return false; - if (stream1->timing.pix_clk_khz - != stream2->timing.pix_clk_khz) + if (stream1->timing.pix_clk_100hz + != stream2->timing.pix_clk_100hz) return false; if (stream1->clamping.c_depth != stream2->clamping.c_depth) @@ -1559,7 +1559,7 @@ static struct stream_encoder *find_first_free_match_stream_enc_for_link( { int i; int j = -1; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; for (i = 0; i < pool->stream_enc_count; i++) { if (!res_ctx->is_stream_enc_acquired[i] && @@ -1748,7 +1748,7 @@ static struct dc_stream_state *find_pll_sharable_stream( if (resource_are_streams_timing_synchronizable( stream_needs_pll, stream_has_pll) && !dc_is_dp_signal(stream_has_pll->signal) - && stream_has_pll->sink->link->connector_signal + && stream_has_pll->link->connector_signal != SIGNAL_TYPE_VIRTUAL) return stream_has_pll; @@ -1759,7 +1759,7 @@ static struct dc_stream_state *find_pll_sharable_stream( static int get_norm_pix_clk(const struct dc_crtc_timing *timing) { - uint32_t pix_clk = timing->pix_clk_khz; + uint32_t pix_clk = timing->pix_clk_100hz; uint32_t normalized_pix_clk = pix_clk; if (timing->pixel_encoding == PIXEL_ENCODING_YCBCR420) @@ -1791,15 +1791,60 @@ static void calculate_phy_pix_clks(struct dc_stream_state *stream) /* update actual pixel clock on all streams */ if (dc_is_hdmi_signal(stream->signal)) stream->phy_pix_clk = get_norm_pix_clk( - &stream->timing); + &stream->timing) / 10; else stream->phy_pix_clk = - stream->timing.pix_clk_khz; + stream->timing.pix_clk_100hz / 10; if (stream->timing.timing_3d_format == TIMING_3D_FORMAT_HW_FRAME_PACKING) stream->phy_pix_clk *= 2; } +static int acquire_resource_from_hw_enabled_state( + struct resource_context *res_ctx, + const struct resource_pool *pool, + struct dc_stream_state *stream) +{ + struct dc_link *link = stream->link; + unsigned int inst; + + /* Check for enabled DIG to identify enabled display */ + if (!link->link_enc->funcs->is_dig_enabled(link->link_enc)) + return -1; + + /* Check for which front end is used by this encoder. + * Note the inst is 1 indexed, where 0 is undefined. + * Note that DIG_FE can source from different OTG but our + * current implementation always map 1-to-1, so this code makes + * the same assumption and doesn't check OTG source. + */ + inst = link->link_enc->funcs->get_dig_frontend(link->link_enc) - 1; + + /* Instance should be within the range of the pool */ + if (inst >= pool->pipe_count) + return -1; + + if (!res_ctx->pipe_ctx[inst].stream) { + struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[inst]; + + pipe_ctx->stream_res.tg = pool->timing_generators[inst]; + pipe_ctx->plane_res.mi = pool->mis[inst]; + pipe_ctx->plane_res.hubp = pool->hubps[inst]; + pipe_ctx->plane_res.ipp = pool->ipps[inst]; + pipe_ctx->plane_res.xfm = pool->transforms[inst]; + pipe_ctx->plane_res.dpp = pool->dpps[inst]; + pipe_ctx->stream_res.opp = pool->opps[inst]; + if (pool->dpps[inst]) + pipe_ctx->plane_res.mpcc_inst = pool->dpps[inst]->inst; + pipe_ctx->pipe_idx = inst; + + pipe_ctx->stream = stream; + return inst; + } + + return -1; +} + enum dc_status resource_map_pool_resources( const struct dc *dc, struct dc_state *context, @@ -1824,8 +1869,15 @@ enum dc_status resource_map_pool_resources( calculate_phy_pix_clks(stream); - /* acquire new resources */ - pipe_idx = acquire_first_free_pipe(&context->res_ctx, pool, stream); + if (stream->apply_seamless_boot_optimization) + pipe_idx = acquire_resource_from_hw_enabled_state( + &context->res_ctx, + pool, + stream); + + if (pipe_idx < 0) + /* acquire new resources */ + pipe_idx = acquire_first_free_pipe(&context->res_ctx, pool, stream); #ifdef CONFIG_DRM_AMD_DC_DCN1_0 if (pipe_idx < 0) @@ -1842,7 +1894,7 @@ enum dc_status resource_map_pool_resources( &context->res_ctx, pool, stream); if (!pipe_ctx->stream_res.stream_enc) - return DC_NO_STREAM_ENG_RESOURCE; + return DC_NO_STREAM_ENC_RESOURCE; update_stream_engine_usage( &context->res_ctx, pool, @@ -1850,7 +1902,7 @@ enum dc_status resource_map_pool_resources( true); /* TODO: Add check if ASIC support and EDID audio */ - if (!stream->sink->converter_disable_audio && + if (!stream->converter_disable_audio && dc_is_audio_capable_signal(pipe_ctx->stream->signal) && stream->audio_info.mode_count) { pipe_ctx->stream_res.audio = find_first_free_audio( @@ -2112,7 +2164,7 @@ static void set_avi_info_frame( itc = true; itc_value = 1; - support = stream->sink->edid_caps.content_support; + support = stream->content_support; if (itc) { if (!support.bits.valid_content_type) { @@ -2151,8 +2203,8 @@ static void set_avi_info_frame( /* TODO : We should handle YCC quantization */ /* but we do not have matrix calculation */ - if (stream->sink->edid_caps.qs_bit == 1 && - stream->sink->edid_caps.qy_bit == 1) { + if (stream->qs_bit == 1 && + stream->qy_bit == 1) { if (color_space == COLOR_SPACE_SRGB || color_space == COLOR_SPACE_2020_RGB_FULLRANGE) { hdmi_info.bits.Q0_Q1 = RGB_QUANTIZATION_FULL_RANGE; @@ -2596,7 +2648,7 @@ void resource_build_bit_depth_reduction_params(struct dc_stream_state *stream, enum dc_status dc_validate_stream(struct dc *dc, struct dc_stream_state *stream) { struct dc *core_dc = dc; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; struct timing_generator *tg = core_dc->res_pool->timing_generators[0]; enum dc_status res = DC_OK; diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c index 66e5c4623a49..996298c35f42 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c @@ -35,20 +35,17 @@ /******************************************************************************* * Private functions ******************************************************************************/ -void update_stream_signal(struct dc_stream_state *stream) +void update_stream_signal(struct dc_stream_state *stream, struct dc_sink *sink) { - - struct dc_sink *dc_sink = stream->sink; - - if (dc_sink->sink_signal == SIGNAL_TYPE_NONE) - stream->signal = stream->sink->link->connector_signal; + if (sink->sink_signal == SIGNAL_TYPE_NONE) + stream->signal = stream->link->connector_signal; else - stream->signal = dc_sink->sink_signal; + stream->signal = sink->sink_signal; if (dc_is_dvi_signal(stream->signal)) { if (stream->ctx->dc->caps.dual_link_dvi && - stream->timing.pix_clk_khz > TMDS_MAX_PIXEL_CLOCK && - stream->sink->sink_signal != SIGNAL_TYPE_DVI_SINGLE_LINK) + (stream->timing.pix_clk_100hz / 10) > TMDS_MAX_PIXEL_CLOCK && + sink->sink_signal != SIGNAL_TYPE_DVI_SINGLE_LINK) stream->signal = SIGNAL_TYPE_DVI_DUAL_LINK; else stream->signal = SIGNAL_TYPE_DVI_SINGLE_LINK; @@ -61,10 +58,15 @@ static void construct(struct dc_stream_state *stream, uint32_t i = 0; stream->sink = dc_sink_data; - stream->ctx = stream->sink->ctx; - dc_sink_retain(dc_sink_data); + stream->ctx = dc_sink_data->ctx; + stream->link = dc_sink_data->link; + stream->sink_patches = dc_sink_data->edid_caps.panel_patch; + stream->converter_disable_audio = dc_sink_data->converter_disable_audio; + stream->qs_bit = dc_sink_data->edid_caps.qs_bit; + stream->qy_bit = dc_sink_data->edid_caps.qy_bit; + /* Copy audio modes */ /* TODO - Remove this translation */ for (i = 0; i < (dc_sink_data->edid_caps.audio_mode_count); i++) @@ -100,11 +102,14 @@ static void construct(struct dc_stream_state *stream, /* EDID CAP translation for HDMI 2.0 */ stream->timing.flags.LTE_340MCSC_SCRAMBLE = dc_sink_data->edid_caps.lte_340mcsc_scramble; - update_stream_signal(stream); + update_stream_signal(stream, dc_sink_data); stream->out_transfer_func = dc_create_transfer_func(); stream->out_transfer_func->type = TF_TYPE_BYPASS; stream->out_transfer_func->ctx = stream->ctx; + + stream->stream_id = stream->ctx->dc_stream_id_count; + stream->ctx->dc_stream_id_count++; } static void destruct(struct dc_stream_state *stream) @@ -155,21 +160,43 @@ struct dc_stream_state *dc_create_stream_for_sink( return stream; } -struct dc_stream_status *dc_stream_get_status( +/** + * dc_stream_get_status_from_state - Get stream status from given dc state + * @state: DC state to find the stream status in + * @stream: The stream to get the stream status for + * + * The given stream is expected to exist in the given dc state. Otherwise, NULL + * will be returned. + */ +struct dc_stream_status *dc_stream_get_status_from_state( + struct dc_state *state, struct dc_stream_state *stream) { uint8_t i; - struct dc *dc = stream->ctx->dc; - for (i = 0; i < dc->current_state->stream_count; i++) { - if (stream == dc->current_state->streams[i]) - return &dc->current_state->stream_status[i]; + for (i = 0; i < state->stream_count; i++) { + if (stream == state->streams[i]) + return &state->stream_status[i]; } return NULL; } /** + * dc_stream_get_status() - Get current stream status of the given stream state + * @stream: The stream to get the stream status for. + * + * The given stream is expected to exist in dc->current_state. Otherwise, NULL + * will be returned. + */ +struct dc_stream_status *dc_stream_get_status( + struct dc_stream_state *stream) +{ + struct dc *dc = stream->ctx->dc; + return dc_stream_get_status_from_state(dc->current_state, stream); +} + +/** * dc_stream_set_cursor_attributes() - Update cursor attributes and set cursor surface address */ bool dc_stream_set_cursor_attributes( @@ -334,16 +361,12 @@ void dc_stream_log(const struct dc *dc, const struct dc_stream_state *stream) stream->output_color_space); DC_LOG_DC( "\tpix_clk_khz: %d, h_total: %d, v_total: %d, pixelencoder:%d, displaycolorDepth:%d\n", - stream->timing.pix_clk_khz, + stream->timing.pix_clk_100hz / 10, stream->timing.h_total, stream->timing.v_total, stream->timing.pixel_encoding, stream->timing.display_color_depth); DC_LOG_DC( - "\tsink name: %s, serial: %d\n", - stream->sink->edid_caps.display_name, - stream->sink->edid_caps.serial_number); - DC_LOG_DC( "\tlink: %d\n", - stream->sink->link->link_index); + stream->link->link_index); } diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_surface.c b/drivers/gpu/drm/amd/display/dc/core/dc_surface.c index c60c9b4c3075..ee6bd50f60b8 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_surface.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_surface.c @@ -40,11 +40,14 @@ static void construct(struct dc_context *ctx, struct dc_plane_state *plane_state plane_state->ctx = ctx; plane_state->gamma_correction = dc_create_gamma(); - plane_state->gamma_correction->is_identity = true; + if (plane_state->gamma_correction != NULL) + plane_state->gamma_correction->is_identity = true; plane_state->in_transfer_func = dc_create_transfer_func(); - plane_state->in_transfer_func->type = TF_TYPE_BYPASS; - plane_state->in_transfer_func->ctx = ctx; + if (plane_state->in_transfer_func != NULL) { + plane_state->in_transfer_func->type = TF_TYPE_BYPASS; + plane_state->in_transfer_func->ctx = ctx; + } } static void destruct(struct dc_plane_state *plane_state) diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_vm_helper.c b/drivers/gpu/drm/amd/display/dc/core/dc_vm_helper.c new file mode 100644 index 000000000000..6ce87b682a32 --- /dev/null +++ b/drivers/gpu/drm/amd/display/dc/core/dc_vm_helper.c @@ -0,0 +1,123 @@ +/* + * Copyright 2018 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: AMD + * + */ + +#include "vm_helper.h" + +static void mark_vmid_used(struct vm_helper *vm_helper, unsigned int pos, uint8_t hubp_idx) +{ + struct vmid_usage vmids = vm_helper->hubp_vmid_usage[hubp_idx]; + + vmids.vmid_usage[0] = vmids.vmid_usage[1]; + vmids.vmid_usage[1] = 1 << pos; +} + +static void add_ptb_to_table(struct vm_helper *vm_helper, unsigned int vmid, uint64_t ptb) +{ + vm_helper->ptb_assigned_to_vmid[vmid] = ptb; + vm_helper->num_vmids_available--; +} + +static void clear_entry_from_vmid_table(struct vm_helper *vm_helper, unsigned int vmid) +{ + vm_helper->ptb_assigned_to_vmid[vmid] = 0; + vm_helper->num_vmids_available++; +} + +static void evict_vmids(struct vm_helper *vm_helper) +{ + int i; + uint16_t ord = 0; + + for (i = 0; i < vm_helper->num_vmid; i++) + ord |= vm_helper->hubp_vmid_usage[i].vmid_usage[0] | vm_helper->hubp_vmid_usage[i].vmid_usage[1]; + + // At this point any positions with value 0 are unused vmids, evict them + for (i = 1; i < vm_helper->num_vmid; i++) { + if (ord & (1u << i)) + clear_entry_from_vmid_table(vm_helper, i); + } +} + +// Return value of -1 indicates vmid table unitialized or ptb dne in the table +static int get_existing_vmid_for_ptb(struct vm_helper *vm_helper, uint64_t ptb) +{ + int i; + + for (i = 0; i < vm_helper->num_vmid; i++) { + if (vm_helper->ptb_assigned_to_vmid[i] == ptb) + return i; + } + + return -1; +} + +// Expected to be called only when there's an available vmid +static int get_next_available_vmid(struct vm_helper *vm_helper) +{ + int i; + + for (i = 1; i < vm_helper->num_vmid; i++) { + if (vm_helper->ptb_assigned_to_vmid[i] == 0) + return i; + } + + return -1; +} + +uint8_t get_vmid_for_ptb(struct vm_helper *vm_helper, int64_t ptb, uint8_t hubp_idx) +{ + unsigned int vmid = 0; + int vmid_exists = -1; + + // Physical address gets vmid 0 + if (ptb == 0) + return 0; + + vmid_exists = get_existing_vmid_for_ptb(vm_helper, ptb); + + if (vmid_exists != -1) { + mark_vmid_used(vm_helper, vmid_exists, hubp_idx); + vmid = vmid_exists; + } else { + if (vm_helper->num_vmids_available == 0) + evict_vmids(vm_helper); + + vmid = get_next_available_vmid(vm_helper); + mark_vmid_used(vm_helper, vmid, hubp_idx); + add_ptb_to_table(vm_helper, vmid, ptb); + } + + return vmid; +} + +void init_vm_helper(struct vm_helper *vm_helper, unsigned int num_vmid, unsigned int num_hubp) +{ + vm_helper->num_vmid = num_vmid; + vm_helper->num_hubp = num_hubp; + vm_helper->num_vmids_available = num_vmid - 1; + + memset(vm_helper->hubp_vmid_usage, 0, sizeof(vm_helper->hubp_vmid_usage[0]) * MAX_HUBP); + memset(vm_helper->ptb_assigned_to_vmid, 0, sizeof(vm_helper->ptb_assigned_to_vmid[0]) * MAX_VMID); +} diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h index 4b5bbb13ce7f..1a7fd6aa77eb 100644 --- a/drivers/gpu/drm/amd/display/dc/dc.h +++ b/drivers/gpu/drm/amd/display/dc/dc.h @@ -39,7 +39,7 @@ #include "inc/hw/dmcu.h" #include "dml/display_mode_lib.h" -#define DC_VER "3.2.08" +#define DC_VER "3.2.17" #define MAX_SURFACES 3 #define MAX_STREAMS 6 @@ -255,6 +255,8 @@ struct dc_debug_options { bool scl_reset_length10; bool hdmi20_disable; bool skip_detection_link_training; + unsigned int force_odm_combine; //bit vector based on otg inst + unsigned int force_fclk_khz; }; struct dc_debug_data { @@ -263,7 +265,6 @@ struct dc_debug_data { uint32_t auxErrorCount; }; - struct dc_state; struct resource_pool; struct dce_hwseq; @@ -339,8 +340,13 @@ struct dc_init_data { uint32_t log_mask; }; -struct dc *dc_create(const struct dc_init_data *init_params); +struct dc_callback_init { + uint8_t reserved; +}; +struct dc *dc_create(const struct dc_init_data *init_params); +void dc_init_callbacks(struct dc *dc, + const struct dc_callback_init *init_params); void dc_destroy(struct dc **dc); /******************************************************************************* @@ -440,6 +446,7 @@ union surface_update_flags { uint32_t coeff_reduction_change:1; uint32_t output_tf_change:1; uint32_t pixel_format_change:1; + uint32_t plane_size_change:1; /* Full updates */ uint32_t new_plane:1; @@ -587,6 +594,10 @@ struct dc_validation_set { uint8_t plane_count; }; +bool dc_validate_seamless_boot_timing(struct dc *dc, + const struct dc_sink *sink, + struct dc_crtc_timing *crtc_timing); + enum dc_status dc_validate_plane(struct dc *dc, const struct dc_plane_state *plane_state); void get_clock_requirements_for_state(struct dc_state *state, struct AsicStateEx *info); @@ -652,6 +663,7 @@ struct dpcd_caps { int8_t branch_dev_name[6]; int8_t branch_hw_revision; int8_t branch_fw_revision[2]; + uint8_t link_rate_set; bool allow_invalid_MSA_timing_param; bool panel_mode_edp; @@ -742,6 +754,9 @@ void dc_set_power_state( struct dc *dc, enum dc_acpi_cm_power_state power_state); void dc_resume(struct dc *dc); +unsigned int dc_get_current_backlight_pwm(struct dc *dc); +unsigned int dc_get_target_backlight_pwm(struct dc *dc); + bool dc_is_dmcu_initialized(struct dc *dc); #endif /* DC_INTERFACE_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/dc_bios_types.h b/drivers/gpu/drm/amd/display/dc/dc_bios_types.h index a8b3cedf9431..78c3b300ec45 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_bios_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_bios_types.h @@ -86,10 +86,6 @@ struct dc_vbios_funcs { bool (*is_accelerated_mode)( struct dc_bios *bios); - bool (*is_active_display)( - struct dc_bios *bios, - enum signal_type signal, - const struct connector_device_tag_info *device_tag); void (*set_scratch_critical_state)( struct dc_bios *bios, bool state); @@ -125,10 +121,6 @@ struct dc_vbios_funcs { enum bp_result (*program_crtc_timing)( struct dc_bios *bios, struct bp_hw_crtc_timing_parameters *bp_params); - - enum bp_result (*crtc_source_select)( - struct dc_bios *bios, - struct bp_crtc_source_select *bp_params); enum bp_result (*program_display_engine_pll)( struct dc_bios *bios, struct bp_pixel_clock_parameters *bp_params); @@ -145,7 +137,6 @@ struct dc_vbios_funcs { }; struct bios_registers { - uint32_t BIOS_SCRATCH_0; uint32_t BIOS_SCRATCH_3; uint32_t BIOS_SCRATCH_6; }; diff --git a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h index da93ab43f2d8..d4eab33c453b 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h @@ -46,11 +46,14 @@ enum dc_lane_count { */ enum dc_link_rate { LINK_RATE_UNKNOWN = 0, - LINK_RATE_LOW = 0x06, - LINK_RATE_HIGH = 0x0A, - LINK_RATE_RBR2 = 0x0C, - LINK_RATE_HIGH2 = 0x14, - LINK_RATE_HIGH3 = 0x1E + LINK_RATE_LOW = 0x06, // Rate_1 (RBR) - 1.62 Gbps/Lane + LINK_RATE_RATE_2 = 0x08, // Rate_2 - 2.16 Gbps/Lane + LINK_RATE_RATE_3 = 0x09, // Rate_3 - 2.43 Gbps/Lane + LINK_RATE_HIGH = 0x0A, // Rate_4 (HBR) - 2.70 Gbps/Lane + LINK_RATE_RBR2 = 0x0C, // Rate_5 (RBR2)- 3.24 Gbps/Lane + LINK_RATE_RATE_6 = 0x10, // Rate_6 - 4.32 Gbps/Lane + LINK_RATE_HIGH2 = 0x14, // Rate_7 (HBR2)- 5.40 Gbps/Lane + LINK_RATE_HIGH3 = 0x1E // Rate_8 (HBR3)- 8.10 Gbps/Lane }; enum dc_link_spread { diff --git a/drivers/gpu/drm/amd/display/dc/dc_helper.c b/drivers/gpu/drm/amd/display/dc/dc_helper.c index 4842d2378bbf..597d38393379 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dc_helper.c @@ -29,31 +29,59 @@ #include "dm_services.h" #include <stdarg.h> +struct dc_reg_value_masks { + uint32_t value; + uint32_t mask; +}; + +struct dc_reg_sequence { + uint32_t addr; + struct dc_reg_value_masks value_masks; +}; + +static inline void set_reg_field_value_masks( + struct dc_reg_value_masks *field_value_mask, + uint32_t value, + uint32_t mask, + uint8_t shift) +{ + ASSERT(mask != 0); + + field_value_mask->value = (field_value_mask->value & ~mask) | (mask & (value << shift)); + field_value_mask->mask = field_value_mask->mask | mask; +} + uint32_t generic_reg_update_ex(const struct dc_context *ctx, uint32_t addr, uint32_t reg_val, int n, uint8_t shift1, uint32_t mask1, uint32_t field_value1, ...) { + struct dc_reg_value_masks field_value_mask = {0}; uint32_t shift, mask, field_value; int i = 1; va_list ap; va_start(ap, field_value1); - reg_val = set_reg_field_value_ex(reg_val, field_value1, mask1, shift1); + /* gather all bits value/mask getting updated in this register */ + set_reg_field_value_masks(&field_value_mask, + field_value1, mask1, shift1); while (i < n) { shift = va_arg(ap, uint32_t); mask = va_arg(ap, uint32_t); field_value = va_arg(ap, uint32_t); - reg_val = set_reg_field_value_ex(reg_val, field_value, mask, shift); + set_reg_field_value_masks(&field_value_mask, + field_value, mask, shift); i++; } - - dm_write_reg(ctx, addr, reg_val); va_end(ap); + + /* mmio write directly */ + reg_val = (reg_val & ~field_value_mask.mask) | field_value_mask.value; + dm_write_reg(ctx, addr, reg_val); return reg_val; } diff --git a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h index e72fce4eca65..da55d623647a 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h @@ -97,6 +97,8 @@ struct dc_plane_address { union large_integer chroma_dcc_const_color; } video_progressive; }; + + union large_integer page_table_base; }; struct dc_size { @@ -730,7 +732,7 @@ struct dc_crtc_timing { uint32_t v_front_porch; uint32_t v_sync_width; - uint32_t pix_clk_khz; + uint32_t pix_clk_100hz; uint32_t vic; uint32_t hdmi_vic; diff --git a/drivers/gpu/drm/amd/display/dc/dc_link.h b/drivers/gpu/drm/amd/display/dc/dc_link.h index 29f19d57ff7a..8fc223defed4 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_link.h +++ b/drivers/gpu/drm/amd/display/dc/dc_link.h @@ -30,6 +30,7 @@ #include "grph_object_defs.h" struct dc_link_status { + bool link_active; struct dpcd_caps *dpcd_caps; }; @@ -110,6 +111,7 @@ struct dc_link { union ddi_channel_mapping ddi_channel_mapping; struct connector_device_tag_info device_tag; struct dpcd_caps dpcd_caps; + uint32_t dongle_max_pix_clk; unsigned short chip_caps; unsigned int dpcd_sink_count; enum edp_revision edp_revision; @@ -124,6 +126,7 @@ struct dc_link { struct dc_link_status link_status; struct link_trace link_trace; + struct gpio *hpd_gpio; }; const struct dc_link_status *dc_link_get_status(const struct dc_link *dc_link); @@ -146,8 +149,7 @@ static inline struct dc_link *dc_get_link_at_index(struct dc *dc, uint32_t link_ */ bool dc_link_set_backlight_level(const struct dc_link *dc_link, uint32_t backlight_pwm_u16_16, - uint32_t frame_ramp, - const struct dc_stream_state *stream); + uint32_t frame_ramp); int dc_link_get_backlight_level(const struct dc_link *dc_link); diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h index be34d638e15d..a798694992b9 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_stream.h +++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h @@ -32,17 +32,18 @@ /******************************************************************************* * Stream Interfaces ******************************************************************************/ +struct timing_sync_info { + int group_id; + int group_size; + bool master; +}; struct dc_stream_status { int primary_otg_inst; int stream_enc_inst; int plane_count; + struct timing_sync_info timing_sync_info; struct dc_plane_state *plane_states[MAX_SURFACE_NUM]; - - /* - * link this stream passes through - */ - struct dc_link *link; }; // TODO: References to this needs to be removed.. @@ -50,8 +51,20 @@ struct freesync_context { bool dummy; }; +union vline_config { + unsigned int line_number; + unsigned long long delta_in_ns; +}; + + struct dc_stream_state { + // sink is deprecated, new code should not reference + // this pointer struct dc_sink *sink; + + struct dc_link *link; + struct dc_panel_patch sink_patches; + union display_content_support content_support; struct dc_crtc_timing timing; struct dc_crtc_timing_adjust adjust; struct dc_info_packet vrr_infopacket; @@ -80,8 +93,9 @@ struct dc_stream_state { enum view_3d_format view_format; bool ignore_msa_timing_param; - - unsigned long long periodic_fn_vsync_delta; + bool converter_disable_audio; + uint8_t qs_bit; + uint8_t qy_bit; /* TODO: custom INFO packets */ /* TODO: ABM info (DMCU) */ @@ -91,7 +105,9 @@ struct dc_stream_state { /* DMCU info */ unsigned int abm_level; - unsigned int bl_pwm_level; + + union vline_config periodic_vsync_config; + union vline_config enhanced_sync_config; /* from core_stream struct */ struct dc_context *ctx; @@ -103,7 +119,8 @@ struct dc_stream_state { int phy_pix_clk; enum signal_type signal; bool dpms_off; - bool apply_edp_fast_boot_optimization; + + void *dm_stream_context; struct dc_cursor_attributes cursor_attributes; struct dc_cursor_position cursor_position; @@ -117,6 +134,21 @@ struct dc_stream_state { /* Computed state bits */ bool mode_changed : 1; + /* Output from DC when stream state is committed or altered + * DC may only access these values during: + * dc_commit_state, dc_commit_state_no_check, dc_commit_streams + * values may not change outside of those calls + */ + struct { + // For interrupt management, some hardware instance + // offsets need to be exposed to DM + uint8_t otg_offset; + } out; + + bool apply_edp_fast_boot_optimization; + bool apply_seamless_boot_optimization; + + uint32_t stream_id; }; struct dc_stream_update { @@ -126,7 +158,9 @@ struct dc_stream_update { struct dc_info_packet *hdr_static_metadata; unsigned int *abm_level; - unsigned long long *periodic_fn_vsync_delta; + union vline_config *periodic_vsync_config; + union vline_config *enhanced_sync_config; + struct dc_crtc_timing_adjust *adjust; struct dc_info_packet *vrr_infopacket; struct dc_info_packet *vsc_infopacket; @@ -163,7 +197,6 @@ void dc_commit_updates_for_stream(struct dc *dc, int surface_count, struct dc_stream_state *stream, struct dc_stream_update *stream_update, - struct dc_plane_state **plane_states, struct dc_state *state); /* * Log the current stream state. @@ -256,11 +289,14 @@ enum surface_update_type dc_check_update_surfaces_for_stream( */ struct dc_stream_state *dc_create_stream_for_sink(struct dc_sink *dc_sink); -void update_stream_signal(struct dc_stream_state *stream); +void update_stream_signal(struct dc_stream_state *stream, struct dc_sink *sink); void dc_stream_retain(struct dc_stream_state *dc_stream); void dc_stream_release(struct dc_stream_state *dc_stream); +struct dc_stream_status *dc_stream_get_status_from_state( + struct dc_state *state, + struct dc_stream_state *stream); struct dc_stream_status *dc_stream_get_status( struct dc_stream_state *dc_stream); diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h index 0b20ae23f169..da2009a108cf 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_types.h @@ -97,8 +97,8 @@ struct dc_context { struct dc_bios *dc_bios; bool created_bios; struct gpio_service *gpio_service; - struct i2caux *i2caux; uint32_t dc_sink_id_count; + uint32_t dc_stream_id_count; uint64_t fbc_gpu_addr; }; @@ -201,6 +201,7 @@ union display_content_support { struct dc_panel_patch { unsigned int dppowerup_delay; unsigned int extra_t12_ms; + unsigned int extra_delay_backlight_off; }; struct dc_edid_caps { diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c b/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c index 2a342eae80fd..01e56f1a9f34 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c @@ -314,8 +314,8 @@ static bool dce_abm_immediate_disable(struct abm *abm) /* setDMCUParam_ABMLevel */ REG_UPDATE_2(MASTER_COMM_CMD_REG, - MASTER_COMM_CMD_REG_BYTE0, MCP_ABM_LEVEL_SET, - MASTER_COMM_CMD_REG_BYTE2, MCP_DISABLE_ABM_IMMEDIATELY); + MASTER_COMM_CMD_REG_BYTE0, MCP_ABM_PIPE_SET, + MASTER_COMM_CMD_REG_BYTE1, MCP_DISABLE_ABM_IMMEDIATELY); /* notifyDMCUMsg */ REG_UPDATE(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 1); diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c index aaeb7faac0c4..4febf4ef7240 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c @@ -24,6 +24,7 @@ */ #include "dm_services.h" +#include "core_types.h" #include "dce_aux.h" #include "dce/dce_11_0_sh_mask.h" @@ -41,17 +42,17 @@ container_of((ptr), struct aux_engine_dce110, base) #define FROM_ENGINE(ptr) \ - FROM_AUX_ENGINE(container_of((ptr), struct aux_engine, base)) + FROM_AUX_ENGINE(container_of((ptr), struct dce_aux, base)) #define FROM_AUX_ENGINE_ENGINE(ptr) \ - container_of((ptr), struct aux_engine, base) + container_of((ptr), struct dce_aux, base) enum { AUX_INVALID_REPLY_RETRY_COUNTER = 1, AUX_TIMED_OUT_RETRY_COUNTER = 2, AUX_DEFER_RETRY_COUNTER = 6 }; static void release_engine( - struct aux_engine *engine) + struct dce_aux *engine) { struct aux_engine_dce110 *aux110 = FROM_AUX_ENGINE(engine); @@ -66,7 +67,7 @@ static void release_engine( #define DMCU_CAN_ACCESS_AUX 2 static bool is_engine_available( - struct aux_engine *engine) + struct dce_aux *engine) { struct aux_engine_dce110 *aux110 = FROM_AUX_ENGINE(engine); @@ -79,7 +80,7 @@ static bool is_engine_available( return (field != DMCU_CAN_ACCESS_AUX); } static bool acquire_engine( - struct aux_engine *engine) + struct dce_aux *engine) { struct aux_engine_dce110 *aux110 = FROM_AUX_ENGINE(engine); @@ -155,7 +156,7 @@ static bool acquire_engine( (0xFF & (address)) static void submit_channel_request( - struct aux_engine *engine, + struct dce_aux *engine, struct aux_request_transaction_data *request) { struct aux_engine_dce110 *aux110 = FROM_AUX_ENGINE(engine); @@ -247,7 +248,7 @@ static void submit_channel_request( REG_UPDATE(AUX_SW_CONTROL, AUX_SW_GO, 1); } -static int read_channel_reply(struct aux_engine *engine, uint32_t size, +static int read_channel_reply(struct dce_aux *engine, uint32_t size, uint8_t *buffer, uint8_t *reply_result, uint32_t *sw_status) { @@ -273,7 +274,8 @@ static int read_channel_reply(struct aux_engine *engine, uint32_t size, REG_GET(AUX_SW_DATA, AUX_SW_DATA, &reply_result_32); reply_result_32 = reply_result_32 >> 4; - *reply_result = (uint8_t)reply_result_32; + if (reply_result != NULL) + *reply_result = (uint8_t)reply_result_32; if (reply_result_32 == 0) { /* ACK */ uint32_t i = 0; @@ -299,61 +301,8 @@ static int read_channel_reply(struct aux_engine *engine, uint32_t size, return 0; } -static void process_channel_reply( - struct aux_engine *engine, - struct aux_reply_transaction_data *reply) -{ - int bytes_replied; - uint8_t reply_result; - uint32_t sw_status; - - bytes_replied = read_channel_reply(engine, reply->length, reply->data, - &reply_result, &sw_status); - - /* in case HPD is LOW, exit AUX transaction */ - if ((sw_status & AUX_SW_STATUS__AUX_SW_HPD_DISCON_MASK)) { - reply->status = AUX_TRANSACTION_REPLY_HPD_DISCON; - return; - } - - if (bytes_replied < 0) { - /* Need to handle an error case... - * Hopefully, upper layer function won't call this function if - * the number of bytes in the reply was 0, because there was - * surely an error that was asserted that should have been - * handled for hot plug case, this could happens - */ - if (!(sw_status & AUX_SW_STATUS__AUX_SW_HPD_DISCON_MASK)) { - reply->status = AUX_TRANSACTION_REPLY_INVALID; - ASSERT_CRITICAL(false); - return; - } - } else { - - switch (reply_result) { - case 0: /* ACK */ - reply->status = AUX_TRANSACTION_REPLY_AUX_ACK; - break; - case 1: /* NACK */ - reply->status = AUX_TRANSACTION_REPLY_AUX_NACK; - break; - case 2: /* DEFER */ - reply->status = AUX_TRANSACTION_REPLY_AUX_DEFER; - break; - case 4: /* AUX ACK / I2C NACK */ - reply->status = AUX_TRANSACTION_REPLY_I2C_NACK; - break; - case 8: /* AUX ACK / I2C DEFER */ - reply->status = AUX_TRANSACTION_REPLY_I2C_DEFER; - break; - default: - reply->status = AUX_TRANSACTION_REPLY_INVALID; - } - } -} - static enum aux_channel_operation_result get_channel_status( - struct aux_engine *engine, + struct dce_aux *engine, uint8_t *returned_bytes) { struct aux_engine_dce110 *aux110 = FROM_AUX_ENGINE(engine); @@ -414,469 +363,22 @@ static enum aux_channel_operation_result get_channel_status( return AUX_CHANNEL_OPERATION_FAILED_TIMEOUT; } } -static void process_read_reply( - struct aux_engine *engine, - struct read_command_context *ctx) -{ - engine->funcs->process_channel_reply(engine, &ctx->reply); - - switch (ctx->reply.status) { - case AUX_TRANSACTION_REPLY_AUX_ACK: - ctx->defer_retry_aux = 0; - if (ctx->returned_byte > ctx->current_read_length) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_PROTOCOL_ERROR; - ctx->operation_succeeded = false; - } else if (ctx->returned_byte < ctx->current_read_length) { - ctx->current_read_length -= ctx->returned_byte; - - ctx->offset += ctx->returned_byte; - - ++ctx->invalid_reply_retry_aux_on_ack; - - if (ctx->invalid_reply_retry_aux_on_ack > - AUX_INVALID_REPLY_RETRY_COUNTER) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_PROTOCOL_ERROR; - ctx->operation_succeeded = false; - } - } else { - ctx->status = I2CAUX_TRANSACTION_STATUS_SUCCEEDED; - ctx->transaction_complete = true; - ctx->operation_succeeded = true; - } - break; - case AUX_TRANSACTION_REPLY_AUX_NACK: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_NACK; - ctx->operation_succeeded = false; - break; - case AUX_TRANSACTION_REPLY_AUX_DEFER: - ++ctx->defer_retry_aux; - - if (ctx->defer_retry_aux > AUX_DEFER_RETRY_COUNTER) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } - break; - case AUX_TRANSACTION_REPLY_I2C_DEFER: - ctx->defer_retry_aux = 0; - - ++ctx->defer_retry_i2c; - - if (ctx->defer_retry_i2c > AUX_DEFER_RETRY_COUNTER) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } - break; - case AUX_TRANSACTION_REPLY_HPD_DISCON: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_HPD_DISCON; - ctx->operation_succeeded = false; - break; - default: - ctx->status = I2CAUX_TRANSACTION_STATUS_UNKNOWN; - ctx->operation_succeeded = false; - } -} -static void process_read_request( - struct aux_engine *engine, - struct read_command_context *ctx) -{ - enum aux_channel_operation_result operation_result; - engine->funcs->submit_channel_request(engine, &ctx->request); - - operation_result = engine->funcs->get_channel_status( - engine, &ctx->returned_byte); - - switch (operation_result) { - case AUX_CHANNEL_OPERATION_SUCCEEDED: - if (ctx->returned_byte > ctx->current_read_length) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_PROTOCOL_ERROR; - ctx->operation_succeeded = false; - } else { - ctx->timed_out_retry_aux = 0; - ctx->invalid_reply_retry_aux = 0; - - ctx->reply.length = ctx->returned_byte; - ctx->reply.data = ctx->buffer; - - process_read_reply(engine, ctx); - } - break; - case AUX_CHANNEL_OPERATION_FAILED_INVALID_REPLY: - ++ctx->invalid_reply_retry_aux; - - if (ctx->invalid_reply_retry_aux > - AUX_INVALID_REPLY_RETRY_COUNTER) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_PROTOCOL_ERROR; - ctx->operation_succeeded = false; - } else - udelay(400); - break; - case AUX_CHANNEL_OPERATION_FAILED_TIMEOUT: - ++ctx->timed_out_retry_aux; - - if (ctx->timed_out_retry_aux > AUX_TIMED_OUT_RETRY_COUNTER) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } else { - /* DP 1.2a, table 2-58: - * "S3: AUX Request CMD PENDING: - * retry 3 times, with 400usec wait on each" - * The HW timeout is set to 550usec, - * so we should not wait here - */ - } - break; - case AUX_CHANNEL_OPERATION_FAILED_HPD_DISCON: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_HPD_DISCON; - ctx->operation_succeeded = false; - break; - default: - ctx->status = I2CAUX_TRANSACTION_STATUS_UNKNOWN; - ctx->operation_succeeded = false; - } -} -static bool read_command( - struct aux_engine *engine, - struct i2caux_transaction_request *request, - bool middle_of_transaction) -{ - struct read_command_context ctx; - - ctx.buffer = request->payload.data; - ctx.current_read_length = request->payload.length; - ctx.offset = 0; - ctx.timed_out_retry_aux = 0; - ctx.invalid_reply_retry_aux = 0; - ctx.defer_retry_aux = 0; - ctx.defer_retry_i2c = 0; - ctx.invalid_reply_retry_aux_on_ack = 0; - ctx.transaction_complete = false; - ctx.operation_succeeded = true; - - if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_DPCD) { - ctx.request.type = AUX_TRANSACTION_TYPE_DP; - ctx.request.action = I2CAUX_TRANSACTION_ACTION_DP_READ; - ctx.request.address = request->payload.address; - } else if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_I2C) { - ctx.request.type = AUX_TRANSACTION_TYPE_I2C; - ctx.request.action = middle_of_transaction ? - I2CAUX_TRANSACTION_ACTION_I2C_READ_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_READ; - ctx.request.address = request->payload.address >> 1; - } else { - /* in DAL2, there was no return in such case */ - BREAK_TO_DEBUGGER(); - return false; - } - - ctx.request.delay = 0; - - do { - memset(ctx.buffer + ctx.offset, 0, ctx.current_read_length); - - ctx.request.data = ctx.buffer + ctx.offset; - ctx.request.length = ctx.current_read_length; - - process_read_request(engine, &ctx); - - request->status = ctx.status; - - if (ctx.operation_succeeded && !ctx.transaction_complete) - if (ctx.request.type == AUX_TRANSACTION_TYPE_I2C) - msleep(engine->delay); - } while (ctx.operation_succeeded && !ctx.transaction_complete); - - if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_DPCD) { - DC_LOG_I2C_AUX("READ: addr:0x%x value:0x%x Result:%d", - request->payload.address, - request->payload.data[0], - ctx.operation_succeeded); - } - - return ctx.operation_succeeded; -} - -static void process_write_reply( - struct aux_engine *engine, - struct write_command_context *ctx) -{ - engine->funcs->process_channel_reply(engine, &ctx->reply); - - switch (ctx->reply.status) { - case AUX_TRANSACTION_REPLY_AUX_ACK: - ctx->operation_succeeded = true; - - if (ctx->returned_byte) { - ctx->request.action = ctx->mot ? - I2CAUX_TRANSACTION_ACTION_I2C_STATUS_REQUEST_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_STATUS_REQUEST; - - ctx->current_write_length = 0; - - ++ctx->ack_m_retry; - - if (ctx->ack_m_retry > AUX_DEFER_RETRY_COUNTER) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } else - udelay(300); - } else { - ctx->status = I2CAUX_TRANSACTION_STATUS_SUCCEEDED; - ctx->defer_retry_aux = 0; - ctx->ack_m_retry = 0; - ctx->transaction_complete = true; - } - break; - case AUX_TRANSACTION_REPLY_AUX_NACK: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_NACK; - ctx->operation_succeeded = false; - break; - case AUX_TRANSACTION_REPLY_AUX_DEFER: - ++ctx->defer_retry_aux; - - if (ctx->defer_retry_aux > ctx->max_defer_retry) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } - break; - case AUX_TRANSACTION_REPLY_I2C_DEFER: - ctx->defer_retry_aux = 0; - ctx->current_write_length = 0; - - ctx->request.action = ctx->mot ? - I2CAUX_TRANSACTION_ACTION_I2C_STATUS_REQUEST_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_STATUS_REQUEST; - - ++ctx->defer_retry_i2c; - - if (ctx->defer_retry_i2c > ctx->max_defer_retry) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } - break; - case AUX_TRANSACTION_REPLY_HPD_DISCON: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_HPD_DISCON; - ctx->operation_succeeded = false; - break; - default: - ctx->status = I2CAUX_TRANSACTION_STATUS_UNKNOWN; - ctx->operation_succeeded = false; - } -} -static void process_write_request( - struct aux_engine *engine, - struct write_command_context *ctx) -{ - enum aux_channel_operation_result operation_result; - - engine->funcs->submit_channel_request(engine, &ctx->request); - - operation_result = engine->funcs->get_channel_status( - engine, &ctx->returned_byte); - - switch (operation_result) { - case AUX_CHANNEL_OPERATION_SUCCEEDED: - ctx->timed_out_retry_aux = 0; - ctx->invalid_reply_retry_aux = 0; - - ctx->reply.length = ctx->returned_byte; - ctx->reply.data = ctx->reply_data; - - process_write_reply(engine, ctx); - break; - case AUX_CHANNEL_OPERATION_FAILED_INVALID_REPLY: - ++ctx->invalid_reply_retry_aux; - - if (ctx->invalid_reply_retry_aux > - AUX_INVALID_REPLY_RETRY_COUNTER) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_PROTOCOL_ERROR; - ctx->operation_succeeded = false; - } else - udelay(400); - break; - case AUX_CHANNEL_OPERATION_FAILED_TIMEOUT: - ++ctx->timed_out_retry_aux; - - if (ctx->timed_out_retry_aux > AUX_TIMED_OUT_RETRY_COUNTER) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } else { - /* DP 1.2a, table 2-58: - * "S3: AUX Request CMD PENDING: - * retry 3 times, with 400usec wait on each" - * The HW timeout is set to 550usec, - * so we should not wait here - */ - } - break; - case AUX_CHANNEL_OPERATION_FAILED_HPD_DISCON: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_HPD_DISCON; - ctx->operation_succeeded = false; - break; - default: - ctx->status = I2CAUX_TRANSACTION_STATUS_UNKNOWN; - ctx->operation_succeeded = false; - } -} -static bool write_command( - struct aux_engine *engine, - struct i2caux_transaction_request *request, - bool middle_of_transaction) -{ - struct write_command_context ctx; - - ctx.mot = middle_of_transaction; - ctx.buffer = request->payload.data; - ctx.current_write_length = request->payload.length; - ctx.timed_out_retry_aux = 0; - ctx.invalid_reply_retry_aux = 0; - ctx.defer_retry_aux = 0; - ctx.defer_retry_i2c = 0; - ctx.ack_m_retry = 0; - ctx.transaction_complete = false; - ctx.operation_succeeded = true; - - if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_DPCD) { - ctx.request.type = AUX_TRANSACTION_TYPE_DP; - ctx.request.action = I2CAUX_TRANSACTION_ACTION_DP_WRITE; - ctx.request.address = request->payload.address; - } else if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_I2C) { - ctx.request.type = AUX_TRANSACTION_TYPE_I2C; - ctx.request.action = middle_of_transaction ? - I2CAUX_TRANSACTION_ACTION_I2C_WRITE_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_WRITE; - ctx.request.address = request->payload.address >> 1; - } else { - /* in DAL2, there was no return in such case */ - BREAK_TO_DEBUGGER(); - return false; - } - - ctx.request.delay = 0; - - ctx.max_defer_retry = - (engine->max_defer_write_retry > AUX_DEFER_RETRY_COUNTER) ? - engine->max_defer_write_retry : AUX_DEFER_RETRY_COUNTER; - - do { - ctx.request.data = ctx.buffer; - ctx.request.length = ctx.current_write_length; - - process_write_request(engine, &ctx); - - request->status = ctx.status; - - if (ctx.operation_succeeded && !ctx.transaction_complete) - if (ctx.request.type == AUX_TRANSACTION_TYPE_I2C) - msleep(engine->delay); - } while (ctx.operation_succeeded && !ctx.transaction_complete); - - if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_DPCD) { - DC_LOG_I2C_AUX("WRITE: addr:0x%x value:0x%x Result:%d", - request->payload.address, - request->payload.data[0], - ctx.operation_succeeded); - } - - return ctx.operation_succeeded; -} -static bool end_of_transaction_command( - struct aux_engine *engine, - struct i2caux_transaction_request *request) -{ - struct i2caux_transaction_request dummy_request; - uint8_t dummy_data; - - /* [tcheng] We only need to send the stop (read with MOT = 0) - * for I2C-over-Aux, not native AUX - */ - - if (request->payload.address_space != - I2CAUX_TRANSACTION_ADDRESS_SPACE_I2C) - return false; - - dummy_request.operation = request->operation; - dummy_request.payload.address_space = request->payload.address_space; - dummy_request.payload.address = request->payload.address; - - /* - * Add a dummy byte due to some receiver quirk - * where one byte is sent along with MOT = 0. - * Ideally this should be 0. - */ - - dummy_request.payload.length = 0; - dummy_request.payload.data = &dummy_data; - - if (request->operation == I2CAUX_TRANSACTION_READ) - return read_command(engine, &dummy_request, false); - else - return write_command(engine, &dummy_request, false); - - /* according Syed, it does not need now DoDummyMOT */ -} -static bool submit_request( - struct aux_engine *engine, - struct i2caux_transaction_request *request, - bool middle_of_transaction) -{ - - bool result; - bool mot_used = true; - - switch (request->operation) { - case I2CAUX_TRANSACTION_READ: - result = read_command(engine, request, mot_used); - break; - case I2CAUX_TRANSACTION_WRITE: - result = write_command(engine, request, mot_used); - break; - default: - result = false; - } - - /* [tcheng] - * need to send stop for the last transaction to free up the AUX - * if the above command fails, this would be the last transaction - */ - - if (!middle_of_transaction || !result) - end_of_transaction_command(engine, request); - - /* mask AUX interrupt */ - - return result; -} enum i2caux_engine_type get_engine_type( - const struct aux_engine *engine) + const struct dce_aux *engine) { return I2CAUX_ENGINE_TYPE_AUX; } static bool acquire( - struct aux_engine *engine, + struct dce_aux *engine, struct ddc *ddc) { enum gpio_result result; - if (engine->funcs->is_engine_available) { - /*check whether SW could use the engine*/ - if (!engine->funcs->is_engine_available(engine)) - return false; - } + if (!is_engine_available(engine)) + return false; result = dal_ddc_open(ddc, GPIO_MODE_HARDWARE, GPIO_DDC_CONFIG_TYPE_MODE_AUX); @@ -884,7 +386,7 @@ static bool acquire( if (result != GPIO_RESULT_OK) return false; - if (!engine->funcs->acquire_engine(engine)) { + if (!acquire_engine(engine)) { dal_ddc_close(ddc); return false; } @@ -894,21 +396,7 @@ static bool acquire( return true; } -static const struct aux_engine_funcs aux_engine_funcs = { - .acquire_engine = acquire_engine, - .submit_channel_request = submit_channel_request, - .process_channel_reply = process_channel_reply, - .read_channel_reply = read_channel_reply, - .get_channel_status = get_channel_status, - .is_engine_available = is_engine_available, - .release_engine = release_engine, - .destroy_engine = dce110_engine_destroy, - .submit_request = submit_request, - .get_engine_type = get_engine_type, - .acquire = acquire, -}; - -void dce110_engine_destroy(struct aux_engine **engine) +void dce110_engine_destroy(struct dce_aux **engine) { struct aux_engine_dce110 *engine110 = FROM_AUX_ENGINE(*engine); @@ -917,7 +405,7 @@ void dce110_engine_destroy(struct aux_engine **engine) *engine = NULL; } -struct aux_engine *dce110_aux_engine_construct(struct aux_engine_dce110 *aux_engine110, +struct dce_aux *dce110_aux_engine_construct(struct aux_engine_dce110 *aux_engine110, struct dc_context *ctx, uint32_t inst, uint32_t timeout_period, @@ -927,7 +415,6 @@ struct aux_engine *dce110_aux_engine_construct(struct aux_engine_dce110 *aux_eng aux_engine110->base.ctx = ctx; aux_engine110->base.delay = 0; aux_engine110->base.max_defer_write_retry = 0; - aux_engine110->base.funcs = &aux_engine_funcs; aux_engine110->base.inst = inst; aux_engine110->timeout_period = timeout_period; aux_engine110->regs = regs; @@ -935,3 +422,101 @@ struct aux_engine *dce110_aux_engine_construct(struct aux_engine_dce110 *aux_eng return &aux_engine110->base; } +static enum i2caux_transaction_action i2caux_action_from_payload(struct aux_payload *payload) +{ + if (payload->i2c_over_aux) { + if (payload->write) { + if (payload->mot) + return I2CAUX_TRANSACTION_ACTION_I2C_WRITE_MOT; + return I2CAUX_TRANSACTION_ACTION_I2C_WRITE; + } + if (payload->mot) + return I2CAUX_TRANSACTION_ACTION_I2C_READ_MOT; + return I2CAUX_TRANSACTION_ACTION_I2C_READ; + } + if (payload->write) + return I2CAUX_TRANSACTION_ACTION_DP_WRITE; + return I2CAUX_TRANSACTION_ACTION_DP_READ; +} + +int dce_aux_transfer(struct ddc_service *ddc, + struct aux_payload *payload) +{ + struct ddc *ddc_pin = ddc->ddc_pin; + struct dce_aux *aux_engine; + enum aux_channel_operation_result operation_result; + struct aux_request_transaction_data aux_req; + struct aux_reply_transaction_data aux_rep; + uint8_t returned_bytes = 0; + int res = -1; + uint32_t status; + + memset(&aux_req, 0, sizeof(aux_req)); + memset(&aux_rep, 0, sizeof(aux_rep)); + + aux_engine = ddc->ctx->dc->res_pool->engines[ddc_pin->pin_data->en]; + acquire(aux_engine, ddc_pin); + + if (payload->i2c_over_aux) + aux_req.type = AUX_TRANSACTION_TYPE_I2C; + else + aux_req.type = AUX_TRANSACTION_TYPE_DP; + + aux_req.action = i2caux_action_from_payload(payload); + + aux_req.address = payload->address; + aux_req.delay = payload->defer_delay * 10; + aux_req.length = payload->length; + aux_req.data = payload->data; + + submit_channel_request(aux_engine, &aux_req); + operation_result = get_channel_status(aux_engine, &returned_bytes); + + switch (operation_result) { + case AUX_CHANNEL_OPERATION_SUCCEEDED: + res = read_channel_reply(aux_engine, payload->length, + payload->data, payload->reply, + &status); + break; + case AUX_CHANNEL_OPERATION_FAILED_HPD_DISCON: + res = 0; + break; + case AUX_CHANNEL_OPERATION_FAILED_REASON_UNKNOWN: + case AUX_CHANNEL_OPERATION_FAILED_INVALID_REPLY: + case AUX_CHANNEL_OPERATION_FAILED_TIMEOUT: + res = -1; + break; + } + release_engine(aux_engine); + return res; +} + +#define AUX_RETRY_MAX 7 + +bool dce_aux_transfer_with_retries(struct ddc_service *ddc, + struct aux_payload *payload) +{ + int i, ret = 0; + uint8_t reply; + bool payload_reply = true; + + if (!payload->reply) { + payload_reply = false; + payload->reply = &reply; + } + + for (i = 0; i < AUX_RETRY_MAX; i++) { + ret = dce_aux_transfer(ddc, payload); + + if (ret >= 0) { + if (*payload->reply == 0) { + if (!payload_reply) + payload->reply = NULL; + return true; + } + } + + udelay(1000); + } + return false; +} diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h index f7caab85dc80..d27f22c05e4b 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.h @@ -25,7 +25,9 @@ #ifndef __DAL_AUX_ENGINE_DCE110_H__ #define __DAL_AUX_ENGINE_DCE110_H__ -#include "aux_engine.h" + +#include "i2caux_interface.h" +#include "inc/hw/aux_engine.h" #define AUX_COMMON_REG_LIST(id)\ SRI(AUX_CONTROL, DP_AUX, id), \ @@ -75,8 +77,20 @@ enum { /* This is the timeout as defined in DP 1.2a, */ SW_AUX_TIMEOUT_PERIOD_MULTIPLIER = 4 }; + +struct dce_aux { + uint32_t inst; + struct ddc *ddc; + struct dc_context *ctx; + /* following values are expressed in milliseconds */ + uint32_t delay; + uint32_t max_defer_write_retry; + + bool acquire_reset; +}; + struct aux_engine_dce110 { - struct aux_engine base; + struct dce_aux base; const struct dce110_aux_registers *regs; struct { uint32_t aux_control; @@ -96,16 +110,22 @@ struct aux_engine_dce110_init_data { const struct dce110_aux_registers *regs; }; -struct aux_engine *dce110_aux_engine_construct( +struct dce_aux *dce110_aux_engine_construct( struct aux_engine_dce110 *aux_engine110, struct dc_context *ctx, uint32_t inst, uint32_t timeout_period, const struct dce110_aux_registers *regs); -void dce110_engine_destroy(struct aux_engine **engine); +void dce110_engine_destroy(struct dce_aux **engine); bool dce110_aux_engine_acquire( - struct aux_engine *aux_engine, + struct dce_aux *aux_engine, struct ddc *ddc); + +int dce_aux_transfer(struct ddc_service *ddc, + struct aux_payload *cmd); + +bool dce_aux_transfer_with_retries(struct ddc_service *ddc, + struct aux_payload *cmd); #endif diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c index afd287f08bc9..3c52a4fc921d 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.c @@ -194,8 +194,8 @@ static uint32_t get_max_pixel_clock_for_all_paths(struct dc_state *context) if (pipe_ctx->top_pipe) continue; - if (pipe_ctx->stream_res.pix_clk_params.requested_pix_clk > max_pix_clk) - max_pix_clk = pipe_ctx->stream_res.pix_clk_params.requested_pix_clk; + if (pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz / 10 > max_pix_clk) + max_pix_clk = pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz / 10; /* raise clock state for HBR3/2 if required. Confirmed with HW DCE/DPCS * logic for HBR3 still needs Nominal (0.8V) on VDDC rail @@ -257,7 +257,7 @@ static int dce_set_clock( clk_mgr_dce->dentist_vco_freq_khz / 64); /* Prepare to program display clock*/ - pxl_clk_params.target_pixel_clock = requested_clk_khz; + pxl_clk_params.target_pixel_clock_100hz = requested_clk_khz * 10; pxl_clk_params.pll_id = CLOCK_SOURCE_ID_DFS; if (clk_mgr_dce->dfs_bypass_active) @@ -450,6 +450,42 @@ void dce_clock_read_ss_info(struct dce_clk_mgr *clk_mgr_dce) } } +/** + * dce121_clock_patch_xgmi_ss_info() - Save XGMI spread spectrum info + * @clk_mgr: clock manager base structure + * + * Reads from VBIOS the XGMI spread spectrum info and saves it within + * the dce clock manager. This operation will overwrite the existing dprefclk + * SS values if the vBIOS query succeeds. Otherwise, it does nothing. It also + * sets the ->xgmi_enabled flag. + */ +void dce121_clock_patch_xgmi_ss_info(struct clk_mgr *clk_mgr) +{ + struct dce_clk_mgr *clk_mgr_dce = TO_DCE_CLK_MGR(clk_mgr); + enum bp_result result; + struct spread_spectrum_info info = { { 0 } }; + struct dc_bios *bp = clk_mgr_dce->base.ctx->dc_bios; + + clk_mgr_dce->xgmi_enabled = false; + + result = bp->funcs->get_spread_spectrum_info(bp, AS_SIGNAL_TYPE_XGMI, + 0, &info); + if (result == BP_RESULT_OK && info.spread_spectrum_percentage != 0) { + clk_mgr_dce->xgmi_enabled = true; + clk_mgr_dce->ss_on_dprefclk = true; + clk_mgr_dce->dprefclk_ss_divider = + info.spread_percentage_divider; + + if (info.type.CENTER_MODE == 0) { + /* Currently for DP Reference clock we + * need only SS percentage for + * downspread */ + clk_mgr_dce->dprefclk_ss_percentage = + info.spread_spectrum_percentage; + } + } +} + void dce110_fill_display_configs( const struct dc_state *context, struct dm_pp_display_configuration *pp_display_cfg) @@ -483,18 +519,18 @@ void dce110_fill_display_configs( cfg->src_height = stream->src.height; cfg->src_width = stream->src.width; cfg->ddi_channel_mapping = - stream->sink->link->ddi_channel_mapping.raw; + stream->link->ddi_channel_mapping.raw; cfg->transmitter = - stream->sink->link->link_enc->transmitter; + stream->link->link_enc->transmitter; cfg->link_settings.lane_count = - stream->sink->link->cur_link_settings.lane_count; + stream->link->cur_link_settings.lane_count; cfg->link_settings.link_rate = - stream->sink->link->cur_link_settings.link_rate; + stream->link->cur_link_settings.link_rate; cfg->link_settings.link_spread = - stream->sink->link->cur_link_settings.link_spread; + stream->link->cur_link_settings.link_spread; cfg->sym_clock = stream->phy_pix_clk; /* Round v_refresh*/ - cfg->v_refresh = stream->timing.pix_clk_khz * 1000; + cfg->v_refresh = stream->timing.pix_clk_100hz * 100; cfg->v_refresh /= stream->timing.h_total; cfg->v_refresh = (cfg->v_refresh + stream->timing.v_total / 2) / stream->timing.v_total; @@ -518,7 +554,7 @@ static uint32_t dce110_get_min_vblank_time_us(const struct dc_state *context) - stream->timing.v_addressable); vertical_blank_time = vertical_blank_in_pixels - * 1000 / stream->timing.pix_clk_khz; + * 10000 / stream->timing.pix_clk_100hz; if (min_vertical_blank_time > vertical_blank_time) min_vertical_blank_time = vertical_blank_time; @@ -612,7 +648,7 @@ static void dce11_pplib_apply_display_requirements( pp_display_cfg->crtc_index = pp_display_cfg->disp_configs[0].pipe_idx; - pp_display_cfg->line_time_in_us = timing->h_total * 1000 / timing->pix_clk_khz; + pp_display_cfg->line_time_in_us = timing->h_total * 10000 / timing->pix_clk_100hz; } if (memcmp(&dc->current_state->pp_display_cfg, pp_display_cfg, sizeof(*pp_display_cfg)) != 0) @@ -625,11 +661,11 @@ static void dce_update_clocks(struct clk_mgr *clk_mgr, { struct dce_clk_mgr *clk_mgr_dce = TO_DCE_CLK_MGR(clk_mgr); struct dm_pp_power_level_change_request level_change_req; - int unpatched_disp_clk = context->bw.dce.dispclk_khz; + int patched_disp_clk = context->bw.dce.dispclk_khz; /*TODO: W/A for dal3 linux, investigate why this works */ if (!clk_mgr_dce->dfs_bypass_active) - context->bw.dce.dispclk_khz = context->bw.dce.dispclk_khz * 115 / 100; + patched_disp_clk = patched_disp_clk * 115 / 100; level_change_req.power_level = dce_get_required_clocks_state(clk_mgr, context); /* get max clock state from PPLIB */ @@ -639,13 +675,11 @@ static void dce_update_clocks(struct clk_mgr *clk_mgr, clk_mgr_dce->cur_min_clks_state = level_change_req.power_level; } - if (should_set_clock(safe_to_lower, context->bw.dce.dispclk_khz, clk_mgr->clks.dispclk_khz)) { - context->bw.dce.dispclk_khz = dce_set_clock(clk_mgr, context->bw.dce.dispclk_khz); - clk_mgr->clks.dispclk_khz = context->bw.dce.dispclk_khz; + if (should_set_clock(safe_to_lower, patched_disp_clk, clk_mgr->clks.dispclk_khz)) { + patched_disp_clk = dce_set_clock(clk_mgr, patched_disp_clk); + clk_mgr->clks.dispclk_khz = patched_disp_clk; } dce_pplib_apply_display_requirements(clk_mgr->ctx->dc, context); - - context->bw.dce.dispclk_khz = unpatched_disp_clk; } static void dce11_update_clocks(struct clk_mgr *clk_mgr, @@ -676,11 +710,11 @@ static void dce112_update_clocks(struct clk_mgr *clk_mgr, { struct dce_clk_mgr *clk_mgr_dce = TO_DCE_CLK_MGR(clk_mgr); struct dm_pp_power_level_change_request level_change_req; - int unpatched_disp_clk = context->bw.dce.dispclk_khz; + int patched_disp_clk = context->bw.dce.dispclk_khz; /*TODO: W/A for dal3 linux, investigate why this works */ if (!clk_mgr_dce->dfs_bypass_active) - context->bw.dce.dispclk_khz = context->bw.dce.dispclk_khz * 115 / 100; + patched_disp_clk = patched_disp_clk * 115 / 100; level_change_req.power_level = dce_get_required_clocks_state(clk_mgr, context); /* get max clock state from PPLIB */ @@ -690,13 +724,11 @@ static void dce112_update_clocks(struct clk_mgr *clk_mgr, clk_mgr_dce->cur_min_clks_state = level_change_req.power_level; } - if (should_set_clock(safe_to_lower, context->bw.dce.dispclk_khz, clk_mgr->clks.dispclk_khz)) { - context->bw.dce.dispclk_khz = dce112_set_clock(clk_mgr, context->bw.dce.dispclk_khz); - clk_mgr->clks.dispclk_khz = context->bw.dce.dispclk_khz; + if (should_set_clock(safe_to_lower, patched_disp_clk, clk_mgr->clks.dispclk_khz)) { + patched_disp_clk = dce112_set_clock(clk_mgr, patched_disp_clk); + clk_mgr->clks.dispclk_khz = patched_disp_clk; } dce11_pplib_apply_display_requirements(clk_mgr->ctx->dc, context); - - context->bw.dce.dispclk_khz = unpatched_disp_clk; } static void dce12_update_clocks(struct clk_mgr *clk_mgr, @@ -706,17 +738,23 @@ static void dce12_update_clocks(struct clk_mgr *clk_mgr, struct dce_clk_mgr *clk_mgr_dce = TO_DCE_CLK_MGR(clk_mgr); struct dm_pp_clock_for_voltage_req clock_voltage_req = {0}; int max_pix_clk = get_max_pixel_clock_for_all_paths(context); - int unpatched_disp_clk = context->bw.dce.dispclk_khz; + int patched_disp_clk = context->bw.dce.dispclk_khz; /*TODO: W/A for dal3 linux, investigate why this works */ if (!clk_mgr_dce->dfs_bypass_active) - context->bw.dce.dispclk_khz = context->bw.dce.dispclk_khz * 115 / 100; + patched_disp_clk = patched_disp_clk * 115 / 100; - if (should_set_clock(safe_to_lower, context->bw.dce.dispclk_khz, clk_mgr->clks.dispclk_khz)) { + if (should_set_clock(safe_to_lower, patched_disp_clk, clk_mgr->clks.dispclk_khz)) { clock_voltage_req.clk_type = DM_PP_CLOCK_TYPE_DISPLAY_CLK; - clock_voltage_req.clocks_in_khz = context->bw.dce.dispclk_khz; - context->bw.dce.dispclk_khz = dce112_set_clock(clk_mgr, context->bw.dce.dispclk_khz); - clk_mgr->clks.dispclk_khz = context->bw.dce.dispclk_khz; + /* + * When xGMI is enabled, the display clk needs to be adjusted + * with the WAFL link's SS percentage. + */ + if (clk_mgr_dce->xgmi_enabled) + patched_disp_clk = clk_mgr_adjust_dp_ref_freq_for_ss( + clk_mgr_dce, patched_disp_clk); + clock_voltage_req.clocks_in_khz = patched_disp_clk; + clk_mgr->clks.dispclk_khz = dce112_set_clock(clk_mgr, patched_disp_clk); dm_pp_apply_clock_for_voltage_request(clk_mgr->ctx, &clock_voltage_req); } @@ -729,8 +767,6 @@ static void dce12_update_clocks(struct clk_mgr *clk_mgr, dm_pp_apply_clock_for_voltage_request(clk_mgr->ctx, &clock_voltage_req); } dce11_pplib_apply_display_requirements(clk_mgr->ctx->dc, context); - - context->bw.dce.dispclk_khz = unpatched_disp_clk; } static const struct clk_mgr_funcs dce120_funcs = { @@ -882,6 +918,27 @@ struct clk_mgr *dce120_clk_mgr_create(struct dc_context *ctx) return &clk_mgr_dce->base; } +struct clk_mgr *dce121_clk_mgr_create(struct dc_context *ctx) +{ + struct dce_clk_mgr *clk_mgr_dce = kzalloc(sizeof(*clk_mgr_dce), + GFP_KERNEL); + + if (clk_mgr_dce == NULL) { + BREAK_TO_DEBUGGER(); + return NULL; + } + + memcpy(clk_mgr_dce->max_clks_by_state, dce120_max_clks_by_state, + sizeof(dce120_max_clks_by_state)); + + dce_clk_mgr_construct(clk_mgr_dce, ctx, NULL, NULL, NULL); + + clk_mgr_dce->dprefclk_khz = 625000; + clk_mgr_dce->base.funcs = &dce120_funcs; + + return &clk_mgr_dce->base; +} + void dce_clk_mgr_destroy(struct clk_mgr **clk_mgr) { struct dce_clk_mgr *clk_mgr_dce = TO_DCE_CLK_MGR(*clk_mgr); diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.h b/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.h index 3bceb31d910d..c8f8c442142a 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.h +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clk_mgr.h @@ -94,11 +94,37 @@ struct dce_clk_mgr { * This is basically "Crystal Frequency In KHz" (XTALIN) frequency */ int dfs_bypass_disp_clk; - /* Flag for Enabled SS on DPREFCLK */ + /** + * @ss_on_dprefclk: + * + * True if spread spectrum is enabled on the DP ref clock. + */ bool ss_on_dprefclk; - /* DPREFCLK SS percentage (if down-spread enabled) */ + + /** + * @xgmi_enabled: + * + * True if xGMI is enabled. On VG20, both audio and display clocks need + * to be adjusted with the WAFL link's SS info if xGMI is enabled. + */ + bool xgmi_enabled; + + /** + * @dprefclk_ss_percentage: + * + * DPREFCLK SS percentage (if down-spread enabled). + * + * Note that if XGMI is enabled, the SS info (percentage and divider) + * from the WAFL link is used instead. This is decided during + * dce_clk_mgr initialization. + */ int dprefclk_ss_percentage; - /* DPREFCLK SS percentage Divider (100 or 1000) */ + + /** + * @dprefclk_ss_divider: + * + * DPREFCLK SS percentage Divider (100 or 1000). + */ int dprefclk_ss_divider; int dprefclk_khz; @@ -163,6 +189,9 @@ struct clk_mgr *dce112_clk_mgr_create( struct clk_mgr *dce120_clk_mgr_create(struct dc_context *ctx); +struct clk_mgr *dce121_clk_mgr_create(struct dc_context *ctx); +void dce121_clock_patch_xgmi_ss_info(struct clk_mgr *clk_mgr); + void dce_clk_mgr_destroy(struct clk_mgr **clk_mgr); int dentist_get_divider_from_did(int did); diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c index 723ce80ed89c..71d5777de961 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c @@ -108,28 +108,28 @@ static const struct spread_spectrum_data *get_ss_data_entry( } /** -* Function: calculate_fb_and_fractional_fb_divider -* -* * DESCRIPTION: Calculates feedback and fractional feedback dividers values -* -*PARAMETERS: -* targetPixelClock Desired frequency in 10 KHz -* ref_divider Reference divider (already known) -* postDivider Post Divider (already known) -* feedback_divider_param Pointer where to store -* calculated feedback divider value -* fract_feedback_divider_param Pointer where to store -* calculated fract feedback divider value -* -*RETURNS: -* It fills the locations pointed by feedback_divider_param -* and fract_feedback_divider_param -* It returns - true if feedback divider not 0 -* - false should never happen) -*/ + * Function: calculate_fb_and_fractional_fb_divider + * + * * DESCRIPTION: Calculates feedback and fractional feedback dividers values + * + *PARAMETERS: + * targetPixelClock Desired frequency in 100 Hz + * ref_divider Reference divider (already known) + * postDivider Post Divider (already known) + * feedback_divider_param Pointer where to store + * calculated feedback divider value + * fract_feedback_divider_param Pointer where to store + * calculated fract feedback divider value + * + *RETURNS: + * It fills the locations pointed by feedback_divider_param + * and fract_feedback_divider_param + * It returns - true if feedback divider not 0 + * - false should never happen) + */ static bool calculate_fb_and_fractional_fb_divider( struct calc_pll_clock_source *calc_pll_cs, - uint32_t target_pix_clk_khz, + uint32_t target_pix_clk_100hz, uint32_t ref_divider, uint32_t post_divider, uint32_t *feedback_divider_param, @@ -138,11 +138,11 @@ static bool calculate_fb_and_fractional_fb_divider( uint64_t feedback_divider; feedback_divider = - (uint64_t)target_pix_clk_khz * ref_divider * post_divider; + (uint64_t)target_pix_clk_100hz * ref_divider * post_divider; feedback_divider *= 10; /* additional factor, since we divide by 10 afterwards */ feedback_divider *= (uint64_t)(calc_pll_cs->fract_fb_divider_factor); - feedback_divider = div_u64(feedback_divider, calc_pll_cs->ref_freq_khz); + feedback_divider = div_u64(feedback_divider, calc_pll_cs->ref_freq_khz * 10ull); /*Round to the number of precision * The following code replace the old code (ullfeedbackDivider + 5)/10 @@ -195,36 +195,36 @@ static bool calc_fb_divider_checking_tolerance( { uint32_t feedback_divider; uint32_t fract_feedback_divider; - uint32_t actual_calculated_clock_khz; + uint32_t actual_calculated_clock_100hz; uint32_t abs_err; - uint64_t actual_calc_clk_khz; + uint64_t actual_calc_clk_100hz; calculate_fb_and_fractional_fb_divider( calc_pll_cs, - pll_settings->adjusted_pix_clk, + pll_settings->adjusted_pix_clk_100hz, ref_divider, post_divider, &feedback_divider, &fract_feedback_divider); /*Actual calculated value*/ - actual_calc_clk_khz = (uint64_t)feedback_divider * + actual_calc_clk_100hz = (uint64_t)feedback_divider * calc_pll_cs->fract_fb_divider_factor + fract_feedback_divider; - actual_calc_clk_khz *= calc_pll_cs->ref_freq_khz; - actual_calc_clk_khz = - div_u64(actual_calc_clk_khz, + actual_calc_clk_100hz *= calc_pll_cs->ref_freq_khz * 10; + actual_calc_clk_100hz = + div_u64(actual_calc_clk_100hz, ref_divider * post_divider * calc_pll_cs->fract_fb_divider_factor); - actual_calculated_clock_khz = (uint32_t)(actual_calc_clk_khz); + actual_calculated_clock_100hz = (uint32_t)(actual_calc_clk_100hz); - abs_err = (actual_calculated_clock_khz > - pll_settings->adjusted_pix_clk) - ? actual_calculated_clock_khz - - pll_settings->adjusted_pix_clk - : pll_settings->adjusted_pix_clk - - actual_calculated_clock_khz; + abs_err = (actual_calculated_clock_100hz > + pll_settings->adjusted_pix_clk_100hz) + ? actual_calculated_clock_100hz - + pll_settings->adjusted_pix_clk_100hz + : pll_settings->adjusted_pix_clk_100hz - + actual_calculated_clock_100hz; if (abs_err <= tolerance) { /*found good values*/ @@ -233,10 +233,10 @@ static bool calc_fb_divider_checking_tolerance( pll_settings->feedback_divider = feedback_divider; pll_settings->fract_feedback_divider = fract_feedback_divider; pll_settings->pix_clk_post_divider = post_divider; - pll_settings->calculated_pix_clk = - actual_calculated_clock_khz; + pll_settings->calculated_pix_clk_100hz = + actual_calculated_clock_100hz; pll_settings->vco_freq = - actual_calculated_clock_khz * post_divider; + actual_calculated_clock_100hz * post_divider / 10; return true; } return false; @@ -257,8 +257,8 @@ static bool calc_pll_dividers_in_range( /* This is err_tolerance / 10000 = 0.0025 - acceptable error of 0.25% * This is errorTolerance / 10000 = 0.0001 - acceptable error of 0.01%*/ - tolerance = (pll_settings->adjusted_pix_clk * err_tolerance) / - 10000; + tolerance = (pll_settings->adjusted_pix_clk_100hz * err_tolerance) / + 100000; if (tolerance < CALC_PLL_CLK_SRC_ERR_TOLERANCE) tolerance = CALC_PLL_CLK_SRC_ERR_TOLERANCE; @@ -294,7 +294,7 @@ static uint32_t calculate_pixel_clock_pll_dividers( uint32_t min_ref_divider; uint32_t max_ref_divider; - if (pll_settings->adjusted_pix_clk == 0) { + if (pll_settings->adjusted_pix_clk_100hz == 0) { DC_LOG_ERROR( "%s Bad requested pixel clock", __func__); return MAX_PLL_CALC_ERROR; @@ -306,21 +306,21 @@ static uint32_t calculate_pixel_clock_pll_dividers( max_post_divider = pll_settings->pix_clk_post_divider; } else { min_post_divider = calc_pll_cs->min_pix_clock_pll_post_divider; - if (min_post_divider * pll_settings->adjusted_pix_clk < - calc_pll_cs->min_vco_khz) { - min_post_divider = calc_pll_cs->min_vco_khz / - pll_settings->adjusted_pix_clk; + if (min_post_divider * pll_settings->adjusted_pix_clk_100hz < + calc_pll_cs->min_vco_khz * 10) { + min_post_divider = calc_pll_cs->min_vco_khz * 10 / + pll_settings->adjusted_pix_clk_100hz; if ((min_post_divider * - pll_settings->adjusted_pix_clk) < - calc_pll_cs->min_vco_khz) + pll_settings->adjusted_pix_clk_100hz) < + calc_pll_cs->min_vco_khz * 10) min_post_divider++; } max_post_divider = calc_pll_cs->max_pix_clock_pll_post_divider; - if (max_post_divider * pll_settings->adjusted_pix_clk - > calc_pll_cs->max_vco_khz) - max_post_divider = calc_pll_cs->max_vco_khz / - pll_settings->adjusted_pix_clk; + if (max_post_divider * pll_settings->adjusted_pix_clk_100hz + > calc_pll_cs->max_vco_khz * 10) + max_post_divider = calc_pll_cs->max_vco_khz * 10 / + pll_settings->adjusted_pix_clk_100hz; } /* 2) Find Reference divider ranges @@ -392,47 +392,47 @@ static bool pll_adjust_pix_clk( struct pixel_clk_params *pix_clk_params, struct pll_settings *pll_settings) { - uint32_t actual_pix_clk_khz = 0; - uint32_t requested_clk_khz = 0; + uint32_t actual_pix_clk_100hz = 0; + uint32_t requested_clk_100hz = 0; struct bp_adjust_pixel_clock_parameters bp_adjust_pixel_clock_params = { 0 }; enum bp_result bp_result; switch (pix_clk_params->signal_type) { case SIGNAL_TYPE_HDMI_TYPE_A: { - requested_clk_khz = pix_clk_params->requested_pix_clk; + requested_clk_100hz = pix_clk_params->requested_pix_clk_100hz; if (pix_clk_params->pixel_encoding != PIXEL_ENCODING_YCBCR422) { switch (pix_clk_params->color_depth) { case COLOR_DEPTH_101010: - requested_clk_khz = (requested_clk_khz * 5) >> 2; + requested_clk_100hz = (requested_clk_100hz * 5) >> 2; break; /* x1.25*/ case COLOR_DEPTH_121212: - requested_clk_khz = (requested_clk_khz * 6) >> 2; + requested_clk_100hz = (requested_clk_100hz * 6) >> 2; break; /* x1.5*/ case COLOR_DEPTH_161616: - requested_clk_khz = requested_clk_khz * 2; + requested_clk_100hz = requested_clk_100hz * 2; break; /* x2.0*/ default: break; } } - actual_pix_clk_khz = requested_clk_khz; + actual_pix_clk_100hz = requested_clk_100hz; } break; case SIGNAL_TYPE_DISPLAY_PORT: case SIGNAL_TYPE_DISPLAY_PORT_MST: case SIGNAL_TYPE_EDP: - requested_clk_khz = pix_clk_params->requested_sym_clk; - actual_pix_clk_khz = pix_clk_params->requested_pix_clk; + requested_clk_100hz = pix_clk_params->requested_sym_clk * 10; + actual_pix_clk_100hz = pix_clk_params->requested_pix_clk_100hz; break; default: - requested_clk_khz = pix_clk_params->requested_pix_clk; - actual_pix_clk_khz = pix_clk_params->requested_pix_clk; + requested_clk_100hz = pix_clk_params->requested_pix_clk_100hz; + actual_pix_clk_100hz = pix_clk_params->requested_pix_clk_100hz; break; } - bp_adjust_pixel_clock_params.pixel_clock = requested_clk_khz; + bp_adjust_pixel_clock_params.pixel_clock = requested_clk_100hz / 10; bp_adjust_pixel_clock_params. encoder_object_id = pix_clk_params->encoder_object_id; bp_adjust_pixel_clock_params.signal_type = pix_clk_params->signal_type; @@ -441,9 +441,9 @@ static bool pll_adjust_pix_clk( bp_result = clk_src->bios->funcs->adjust_pixel_clock( clk_src->bios, &bp_adjust_pixel_clock_params); if (bp_result == BP_RESULT_OK) { - pll_settings->actual_pix_clk = actual_pix_clk_khz; - pll_settings->adjusted_pix_clk = - bp_adjust_pixel_clock_params.adjusted_pixel_clock; + pll_settings->actual_pix_clk_100hz = actual_pix_clk_100hz; + pll_settings->adjusted_pix_clk_100hz = + bp_adjust_pixel_clock_params.adjusted_pixel_clock * 10; pll_settings->reference_divider = bp_adjust_pixel_clock_params.reference_divider; pll_settings->pix_clk_post_divider = @@ -490,7 +490,7 @@ static uint32_t dce110_get_pix_clk_dividers_helper ( const struct spread_spectrum_data *ss_data = get_ss_data_entry( clk_src, pix_clk_params->signal_type, - pll_settings->adjusted_pix_clk); + pll_settings->adjusted_pix_clk_100hz / 10); if (NULL != ss_data) pll_settings->ss_percentage = ss_data->percentage; @@ -502,13 +502,13 @@ static uint32_t dce110_get_pix_clk_dividers_helper ( * to continue. */ DC_LOG_ERROR( "%s: Failed to adjust pixel clock!!", __func__); - pll_settings->actual_pix_clk = - pix_clk_params->requested_pix_clk; - pll_settings->adjusted_pix_clk = - pix_clk_params->requested_pix_clk; + pll_settings->actual_pix_clk_100hz = + pix_clk_params->requested_pix_clk_100hz; + pll_settings->adjusted_pix_clk_100hz = + pix_clk_params->requested_pix_clk_100hz; if (dc_is_dp_signal(pix_clk_params->signal_type)) - pll_settings->adjusted_pix_clk = 100000; + pll_settings->adjusted_pix_clk_100hz = 1000000; } /* Calculate Dividers */ @@ -533,28 +533,28 @@ static void dce112_get_pix_clk_dividers_helper ( struct pll_settings *pll_settings, struct pixel_clk_params *pix_clk_params) { - uint32_t actualPixelClockInKHz; + uint32_t actual_pixel_clock_100hz; - actualPixelClockInKHz = pix_clk_params->requested_pix_clk; + actual_pixel_clock_100hz = pix_clk_params->requested_pix_clk_100hz; /* Calculate Dividers */ if (pix_clk_params->signal_type == SIGNAL_TYPE_HDMI_TYPE_A) { switch (pix_clk_params->color_depth) { case COLOR_DEPTH_101010: - actualPixelClockInKHz = (actualPixelClockInKHz * 5) >> 2; + actual_pixel_clock_100hz = (actual_pixel_clock_100hz * 5) >> 2; break; case COLOR_DEPTH_121212: - actualPixelClockInKHz = (actualPixelClockInKHz * 6) >> 2; + actual_pixel_clock_100hz = (actual_pixel_clock_100hz * 6) >> 2; break; case COLOR_DEPTH_161616: - actualPixelClockInKHz = actualPixelClockInKHz * 2; + actual_pixel_clock_100hz = actual_pixel_clock_100hz * 2; break; default: break; } } - pll_settings->actual_pix_clk = actualPixelClockInKHz; - pll_settings->adjusted_pix_clk = actualPixelClockInKHz; - pll_settings->calculated_pix_clk = pix_clk_params->requested_pix_clk; + pll_settings->actual_pix_clk_100hz = actual_pixel_clock_100hz; + pll_settings->adjusted_pix_clk_100hz = actual_pixel_clock_100hz; + pll_settings->calculated_pix_clk_100hz = pix_clk_params->requested_pix_clk_100hz; } static uint32_t dce110_get_pix_clk_dividers( @@ -567,7 +567,7 @@ static uint32_t dce110_get_pix_clk_dividers( DC_LOGGER_INIT(); if (pix_clk_params == NULL || pll_settings == NULL - || pix_clk_params->requested_pix_clk == 0) { + || pix_clk_params->requested_pix_clk_100hz == 0) { DC_LOG_ERROR( "%s: Invalid parameters!!\n", __func__); return pll_calc_error; @@ -577,10 +577,10 @@ static uint32_t dce110_get_pix_clk_dividers( if (cs->id == CLOCK_SOURCE_ID_DP_DTO || cs->id == CLOCK_SOURCE_ID_EXTERNAL) { - pll_settings->adjusted_pix_clk = clk_src->ext_clk_khz; - pll_settings->calculated_pix_clk = clk_src->ext_clk_khz; - pll_settings->actual_pix_clk = - pix_clk_params->requested_pix_clk; + pll_settings->adjusted_pix_clk_100hz = clk_src->ext_clk_khz * 10; + pll_settings->calculated_pix_clk_100hz = clk_src->ext_clk_khz * 10; + pll_settings->actual_pix_clk_100hz = + pix_clk_params->requested_pix_clk_100hz; return 0; } @@ -599,7 +599,7 @@ static uint32_t dce112_get_pix_clk_dividers( DC_LOGGER_INIT(); if (pix_clk_params == NULL || pll_settings == NULL - || pix_clk_params->requested_pix_clk == 0) { + || pix_clk_params->requested_pix_clk_100hz == 0) { DC_LOG_ERROR( "%s: Invalid parameters!!\n", __func__); return -1; @@ -609,10 +609,10 @@ static uint32_t dce112_get_pix_clk_dividers( if (cs->id == CLOCK_SOURCE_ID_DP_DTO || cs->id == CLOCK_SOURCE_ID_EXTERNAL) { - pll_settings->adjusted_pix_clk = clk_src->ext_clk_khz; - pll_settings->calculated_pix_clk = clk_src->ext_clk_khz; - pll_settings->actual_pix_clk = - pix_clk_params->requested_pix_clk; + pll_settings->adjusted_pix_clk_100hz = clk_src->ext_clk_khz * 10; + pll_settings->calculated_pix_clk_100hz = clk_src->ext_clk_khz * 10; + pll_settings->actual_pix_clk_100hz = + pix_clk_params->requested_pix_clk_100hz; return -1; } @@ -714,7 +714,7 @@ static bool enable_spread_spectrum( ss_data = get_ss_data_entry( clk_src, signal, - pll_settings->calculated_pix_clk); + pll_settings->calculated_pix_clk_100hz / 10); /* Pixel clock PLL has been programmed to generate desired pixel clock, * now enable SS on pixel clock */ @@ -853,7 +853,7 @@ static bool dce110_program_pix_clk( /*ATOMBIOS expects pixel rate adjusted by deep color ratio)*/ bp_pc_params.controller_id = pix_clk_params->controller_id; bp_pc_params.pll_id = clock_source->id; - bp_pc_params.target_pixel_clock = pll_settings->actual_pix_clk; + bp_pc_params.target_pixel_clock_100hz = pll_settings->actual_pix_clk_100hz; bp_pc_params.encoder_object_id = pix_clk_params->encoder_object_id; bp_pc_params.signal_type = pix_clk_params->signal_type; @@ -903,12 +903,12 @@ static bool dce112_program_pix_clk( #if defined(CONFIG_DRM_AMD_DC_DCN1_0) if (IS_FPGA_MAXIMUS_DC(clock_source->ctx->dce_environment)) { unsigned int inst = pix_clk_params->controller_id - CONTROLLER_ID_D0; - unsigned dp_dto_ref_kHz = 700000; - unsigned clock_kHz = pll_settings->actual_pix_clk; + unsigned dp_dto_ref_100hz = 7000000; + unsigned clock_100hz = pll_settings->actual_pix_clk_100hz; /* Set DTO values: phase = target clock, modulo = reference clock */ - REG_WRITE(PHASE[inst], clock_kHz); - REG_WRITE(MODULO[inst], dp_dto_ref_kHz); + REG_WRITE(PHASE[inst], clock_100hz); + REG_WRITE(MODULO[inst], dp_dto_ref_100hz); /* Enable DTO */ REG_UPDATE(PIXEL_RATE_CNTL[inst], DP_DTO0_ENABLE, 1); @@ -927,7 +927,7 @@ static bool dce112_program_pix_clk( /*ATOMBIOS expects pixel rate adjusted by deep color ratio)*/ bp_pc_params.controller_id = pix_clk_params->controller_id; bp_pc_params.pll_id = clock_source->id; - bp_pc_params.target_pixel_clock = pll_settings->actual_pix_clk; + bp_pc_params.target_pixel_clock_100hz = pll_settings->actual_pix_clk_100hz; bp_pc_params.encoder_object_id = pix_clk_params->encoder_object_id; bp_pc_params.signal_type = pix_clk_params->signal_type; @@ -977,6 +977,28 @@ static bool dce110_clock_source_power_down( return bp_result == BP_RESULT_OK; } +static bool get_pixel_clk_frequency_100hz( + struct clock_source *clock_source, + unsigned int inst, + unsigned int *pixel_clk_khz) +{ + struct dce110_clk_src *clk_src = TO_DCE110_CLK_SRC(clock_source); + unsigned int clock_hz = 0; + + if (clock_source->id == CLOCK_SOURCE_ID_DP_DTO) { + clock_hz = REG_READ(PHASE[inst]); + + /* NOTE: There is agreement with VBIOS here that MODULO is + * programmed equal to DPREFCLK, in which case PHASE will be + * equivalent to pixel clock. + */ + *pixel_clk_khz = clock_hz / 100; + return true; + } + + return false; +} + /*****************************************/ /* Constructor */ /*****************************************/ @@ -984,12 +1006,14 @@ static bool dce110_clock_source_power_down( static const struct clock_source_funcs dce112_clk_src_funcs = { .cs_power_down = dce110_clock_source_power_down, .program_pix_clk = dce112_program_pix_clk, - .get_pix_clk_dividers = dce112_get_pix_clk_dividers + .get_pix_clk_dividers = dce112_get_pix_clk_dividers, + .get_pixel_clk_frequency_100hz = get_pixel_clk_frequency_100hz }; static const struct clock_source_funcs dce110_clk_src_funcs = { .cs_power_down = dce110_clock_source_power_down, .program_pix_clk = dce110_program_pix_clk, - .get_pix_clk_dividers = dce110_get_pix_clk_dividers + .get_pix_clk_dividers = dce110_get_pix_clk_dividers, + .get_pixel_clk_frequency_100hz = get_pixel_clk_frequency_100hz }; diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c b/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c index dea40b322191..c2926cf19dee 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_dmcu.c @@ -51,7 +51,6 @@ #define PSR_SET_WAITLOOP 0x31 #define MCP_INIT_DMCU 0x88 #define MCP_INIT_IRAM 0x89 -#define MCP_DMCU_VERSION 0x90 #define MASTER_COMM_CNTL_REG__MASTER_COMM_INTERRUPT_MASK 0x00000001L static bool dce_dmcu_init(struct dmcu *dmcu) @@ -317,38 +316,11 @@ static void dce_get_psr_wait_loop( } #if defined(CONFIG_DRM_AMD_DC_DCN1_0) -static void dcn10_get_dmcu_state(struct dmcu *dmcu) -{ - struct dce_dmcu *dmcu_dce = TO_DCE_DMCU(dmcu); - uint32_t dmcu_state_offset = 0xf6; - - /* Enable write access to IRAM */ - REG_UPDATE_2(DMCU_RAM_ACCESS_CTRL, - IRAM_HOST_ACCESS_EN, 1, - IRAM_RD_ADDR_AUTO_INC, 1); - - REG_WAIT(DMU_MEM_PWR_CNTL, DMCU_IRAM_MEM_PWR_STATE, 0, 2, 10); - - /* Write address to IRAM_RD_ADDR in DMCU_IRAM_RD_CTRL */ - REG_WRITE(DMCU_IRAM_RD_CTRL, dmcu_state_offset); - - /* Read data from IRAM_RD_DATA in DMCU_IRAM_RD_DATA*/ - dmcu->dmcu_state = REG_READ(DMCU_IRAM_RD_DATA); - - /* Disable write access to IRAM to allow dynamic sleep state */ - REG_UPDATE_2(DMCU_RAM_ACCESS_CTRL, - IRAM_HOST_ACCESS_EN, 0, - IRAM_RD_ADDR_AUTO_INC, 0); -} - static void dcn10_get_dmcu_version(struct dmcu *dmcu) { struct dce_dmcu *dmcu_dce = TO_DCE_DMCU(dmcu); uint32_t dmcu_version_offset = 0xf1; - /* Clear scratch */ - REG_WRITE(DC_DMCU_SCRATCH, 0); - /* Enable write access to IRAM */ REG_UPDATE_2(DMCU_RAM_ACCESS_CTRL, IRAM_HOST_ACCESS_EN, 1, @@ -359,85 +331,74 @@ static void dcn10_get_dmcu_version(struct dmcu *dmcu) /* Write address to IRAM_RD_ADDR and read from DATA register */ REG_WRITE(DMCU_IRAM_RD_CTRL, dmcu_version_offset); dmcu->dmcu_version.interface_version = REG_READ(DMCU_IRAM_RD_DATA); - dmcu->dmcu_version.year = ((REG_READ(DMCU_IRAM_RD_DATA) << 8) | + dmcu->dmcu_version.abm_version = REG_READ(DMCU_IRAM_RD_DATA); + dmcu->dmcu_version.psr_version = REG_READ(DMCU_IRAM_RD_DATA); + dmcu->dmcu_version.build_version = ((REG_READ(DMCU_IRAM_RD_DATA) << 8) | REG_READ(DMCU_IRAM_RD_DATA)); - dmcu->dmcu_version.month = REG_READ(DMCU_IRAM_RD_DATA); - dmcu->dmcu_version.date = REG_READ(DMCU_IRAM_RD_DATA); /* Disable write access to IRAM to allow dynamic sleep state */ REG_UPDATE_2(DMCU_RAM_ACCESS_CTRL, IRAM_HOST_ACCESS_EN, 0, IRAM_RD_ADDR_AUTO_INC, 0); - - /* Send MCP command message to DMCU to get version reply from FW. - * We expect this version should match the one in IRAM, otherwise - * something is wrong with DMCU and we should fail and disable UC. - */ - REG_WAIT(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 0, 100, 800); - - /* Set command to get DMCU version from microcontroller */ - REG_UPDATE(MASTER_COMM_CMD_REG, MASTER_COMM_CMD_REG_BYTE0, - MCP_DMCU_VERSION); - - /* Notify microcontroller of new command */ - REG_UPDATE(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 1); - - /* Ensure command has been executed before continuing */ - REG_WAIT(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 0, 100, 800); - - /* Somehow version does not match, so fail and return version 0 */ - if (dmcu->dmcu_version.interface_version != REG_READ(DC_DMCU_SCRATCH)) - dmcu->dmcu_version.interface_version = 0; } static bool dcn10_dmcu_init(struct dmcu *dmcu) { struct dce_dmcu *dmcu_dce = TO_DCE_DMCU(dmcu); + bool status = false; - /* DMCU FW should populate the scratch register if running */ - if (REG_READ(DC_DMCU_SCRATCH) == 0) - return false; - - /* Check state is uninitialized */ - dcn10_get_dmcu_state(dmcu); - - /* If microcontroller is already initialized, do nothing */ - if (dmcu->dmcu_state == DMCU_RUNNING) - return true; - - /* Retrieve and cache the DMCU firmware version. */ - dcn10_get_dmcu_version(dmcu); - - /* Check interface version to confirm firmware is loaded and running */ - if (dmcu->dmcu_version.interface_version == 0) - return false; + /* Definition of DC_DMCU_SCRATCH + * 0 : firmare not loaded + * 1 : PSP load DMCU FW but not initialized + * 2 : Firmware already initialized + */ + dmcu->dmcu_state = REG_READ(DC_DMCU_SCRATCH); - /* Wait until microcontroller is ready to process interrupt */ - REG_WAIT(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 0, 100, 800); + switch (dmcu->dmcu_state) { + case DMCU_UNLOADED: + status = false; + break; + case DMCU_LOADED_UNINITIALIZED: + /* Wait until microcontroller is ready to process interrupt */ + REG_WAIT(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 0, 100, 800); - /* Set initialized ramping boundary value */ - REG_WRITE(MASTER_COMM_DATA_REG1, 0xFFFF); + /* Set initialized ramping boundary value */ + REG_WRITE(MASTER_COMM_DATA_REG1, 0xFFFF); - /* Set command to initialize microcontroller */ - REG_UPDATE(MASTER_COMM_CMD_REG, MASTER_COMM_CMD_REG_BYTE0, + /* Set command to initialize microcontroller */ + REG_UPDATE(MASTER_COMM_CMD_REG, MASTER_COMM_CMD_REG_BYTE0, MCP_INIT_DMCU); - /* Notify microcontroller of new command */ - REG_UPDATE(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 1); + /* Notify microcontroller of new command */ + REG_UPDATE(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 1); - /* Ensure command has been executed before continuing */ - REG_WAIT(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 0, 100, 800); + /* Ensure command has been executed before continuing */ + REG_WAIT(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 0, 100, 800); - // Check state is initialized - dcn10_get_dmcu_state(dmcu); + // Check state is initialized + dmcu->dmcu_state = REG_READ(DC_DMCU_SCRATCH); - // If microcontroller is not in running state, fail - if (dmcu->dmcu_state != DMCU_RUNNING) - return false; + // If microcontroller is not in running state, fail + if (dmcu->dmcu_state == DMCU_RUNNING) { + /* Retrieve and cache the DMCU firmware version. */ + dcn10_get_dmcu_version(dmcu); + status = true; + } else + status = false; - return true; + break; + case DMCU_RUNNING: + status = true; + break; + default: + status = false; + break; + } + + return status; } + static bool dcn10_dmcu_load_iram(struct dmcu *dmcu, unsigned int start_offset, const char *src, diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h b/drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h index c83a7f05f14c..956bdf14503f 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h @@ -133,6 +133,10 @@ SR(DCHUB_AGP_TOP), \ BL_REG_LIST() +#define HWSEQ_VG20_REG_LIST() \ + HWSEQ_DCE120_REG_LIST(),\ + MMHUB_SR(MC_VM_XGMI_LFB_CNTL) + #define HWSEQ_DCE112_REG_LIST() \ HWSEQ_DCE10_REG_LIST(), \ HWSEQ_PIXEL_RATE_REG_LIST(CRTC), \ @@ -298,6 +302,7 @@ struct dce_hwseq_registers { uint32_t MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB; uint32_t MC_VM_SYSTEM_APERTURE_LOW_ADDR; uint32_t MC_VM_SYSTEM_APERTURE_HIGH_ADDR; + uint32_t MC_VM_XGMI_LFB_CNTL; uint32_t AZALIA_AUDIO_DTO; uint32_t AZALIA_CONTROLLER_CLOCK_GATING; }; @@ -382,6 +387,11 @@ struct dce_hwseq_registers { HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh), \ HWS_SF(, LVTMA_PWRSEQ_STATE, LVTMA_PWRSEQ_TARGET_STATE_R, mask_sh) +#define HWSEQ_VG20_MASK_SH_LIST(mask_sh)\ + HWSEQ_DCE12_MASK_SH_LIST(mask_sh),\ + HWS_SF(, MC_VM_XGMI_LFB_CNTL, PF_LFB_REGION, mask_sh),\ + HWS_SF(, MC_VM_XGMI_LFB_CNTL, PF_MAX_REGION, mask_sh) + #define HWSEQ_DCN_MASK_SH_LIST(mask_sh)\ HWSEQ_PIXEL_RATE_MASK_SH_LIST(mask_sh, OTG0_),\ HWS_SF1(OTG0_, PHYPLL_PIXEL_RATE_CNTL, PHYPLL_PIXEL_RATE_SOURCE, mask_sh), \ @@ -470,6 +480,8 @@ struct dce_hwseq_registers { type PHYSICAL_PAGE_NUMBER_MSB;\ type PHYSICAL_PAGE_NUMBER_LSB;\ type LOGICAL_ADDR; \ + type PF_LFB_REGION;\ + type PF_MAX_REGION;\ type ENABLE_L1_TLB;\ type SYSTEM_ACCESS_MODE;\ type LVTMA_BLON;\ diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c index 3e18ea84b1f9..314c04a915d2 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c @@ -599,12 +599,12 @@ bool dce110_link_encoder_validate_dvi_output( if ((connector_signal == SIGNAL_TYPE_DVI_SINGLE_LINK || connector_signal == SIGNAL_TYPE_HDMI_TYPE_A) && signal != SIGNAL_TYPE_HDMI_TYPE_A && - crtc_timing->pix_clk_khz > TMDS_MAX_PIXEL_CLOCK) + crtc_timing->pix_clk_100hz > (TMDS_MAX_PIXEL_CLOCK * 10)) return false; - if (crtc_timing->pix_clk_khz < TMDS_MIN_PIXEL_CLOCK) + if (crtc_timing->pix_clk_100hz < (TMDS_MIN_PIXEL_CLOCK * 10)) return false; - if (crtc_timing->pix_clk_khz > max_pixel_clock) + if (crtc_timing->pix_clk_100hz > (max_pixel_clock * 10)) return false; /* DVI supports 6/8bpp single-link and 10/16bpp dual-link */ @@ -788,7 +788,7 @@ bool dce110_link_encoder_validate_output_with_stream( case SIGNAL_TYPE_DVI_DUAL_LINK: is_valid = dce110_link_encoder_validate_dvi_output( enc110, - stream->sink->link->connector_signal, + stream->link->connector_signal, stream->signal, &stream->timing); break; diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c index cce0d18f91da..1fa2d4fd7a35 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c @@ -288,9 +288,18 @@ static void dce110_stream_encoder_dp_set_stream_attribute( #endif struct dce110_stream_encoder *enc110 = DCE110STRENC_FROM_STRENC(enc); - + struct dc_crtc_timing hw_crtc_timing = *crtc_timing; + if (hw_crtc_timing.flags.INTERLACE) { + /*the input timing is in VESA spec format with Interlace flag =1*/ + hw_crtc_timing.v_total /= 2; + hw_crtc_timing.v_border_top /= 2; + hw_crtc_timing.v_addressable /= 2; + hw_crtc_timing.v_border_bottom /= 2; + hw_crtc_timing.v_front_porch /= 2; + hw_crtc_timing.v_sync_width /= 2; + } /* set pixel encoding */ - switch (crtc_timing->pixel_encoding) { + switch (hw_crtc_timing.pixel_encoding) { case PIXEL_ENCODING_YCBCR422: REG_UPDATE(DP_PIXEL_FORMAT, DP_PIXEL_ENCODING, DP_PIXEL_ENCODING_TYPE_YCBCR422); @@ -299,8 +308,8 @@ static void dce110_stream_encoder_dp_set_stream_attribute( REG_UPDATE(DP_PIXEL_FORMAT, DP_PIXEL_ENCODING, DP_PIXEL_ENCODING_TYPE_YCBCR444); - if (crtc_timing->flags.Y_ONLY) - if (crtc_timing->display_color_depth != COLOR_DEPTH_666) + if (hw_crtc_timing.flags.Y_ONLY) + if (hw_crtc_timing.display_color_depth != COLOR_DEPTH_666) /* HW testing only, no use case yet. * Color depth of Y-only could be * 8, 10, 12, 16 bits */ @@ -335,7 +344,7 @@ static void dce110_stream_encoder_dp_set_stream_attribute( /* set color depth */ - switch (crtc_timing->display_color_depth) { + switch (hw_crtc_timing.display_color_depth) { case COLOR_DEPTH_666: REG_UPDATE(DP_PIXEL_FORMAT, DP_COMPONENT_DEPTH, 0); @@ -363,7 +372,7 @@ static void dce110_stream_encoder_dp_set_stream_attribute( #if defined(CONFIG_DRM_AMD_DC_DCN1_0) - switch (crtc_timing->display_color_depth) { + switch (hw_crtc_timing.display_color_depth) { case COLOR_DEPTH_666: colorimetry_bpc = 0; break; @@ -401,9 +410,9 @@ static void dce110_stream_encoder_dp_set_stream_attribute( misc0 = misc0 | 0x8; /* bit3=1, bit4=0 */ misc1 = misc1 & ~0x80; /* bit7 = 0*/ dynamic_range_ycbcr = 0; /*bt601*/ - if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) + if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR422) misc0 = misc0 | 0x2; /* bit2=0, bit1=1 */ - else if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR444) + else if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR444) misc0 = misc0 | 0x4; /* bit2=1, bit1=0 */ break; case COLOR_SPACE_YCBCR709: @@ -411,9 +420,9 @@ static void dce110_stream_encoder_dp_set_stream_attribute( misc0 = misc0 | 0x18; /* bit3=1, bit4=1 */ misc1 = misc1 & ~0x80; /* bit7 = 0*/ dynamic_range_ycbcr = 1; /*bt709*/ - if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) + if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR422) misc0 = misc0 | 0x2; /* bit2=0, bit1=1 */ - else if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR444) + else if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR444) misc0 = misc0 | 0x4; /* bit2=1, bit1=0 */ break; case COLOR_SPACE_2020_RGB_LIMITEDRANGE: @@ -453,27 +462,27 @@ static void dce110_stream_encoder_dp_set_stream_attribute( */ if (REG(DP_MSA_TIMING_PARAM1)) REG_SET_2(DP_MSA_TIMING_PARAM1, 0, - DP_MSA_HTOTAL, crtc_timing->h_total, - DP_MSA_VTOTAL, crtc_timing->v_total); + DP_MSA_HTOTAL, hw_crtc_timing.h_total, + DP_MSA_VTOTAL, hw_crtc_timing.v_total); #endif /* calcuate from vesa timing parameters * h_active_start related to leading edge of sync */ - h_blank = crtc_timing->h_total - crtc_timing->h_border_left - - crtc_timing->h_addressable - crtc_timing->h_border_right; + h_blank = hw_crtc_timing.h_total - hw_crtc_timing.h_border_left - + hw_crtc_timing.h_addressable - hw_crtc_timing.h_border_right; - h_back_porch = h_blank - crtc_timing->h_front_porch - - crtc_timing->h_sync_width; + h_back_porch = h_blank - hw_crtc_timing.h_front_porch - + hw_crtc_timing.h_sync_width; /* start at begining of left border */ - h_active_start = crtc_timing->h_sync_width + h_back_porch; + h_active_start = hw_crtc_timing.h_sync_width + h_back_porch; - v_active_start = crtc_timing->v_total - crtc_timing->v_border_top - - crtc_timing->v_addressable - crtc_timing->v_border_bottom - - crtc_timing->v_front_porch; + v_active_start = hw_crtc_timing.v_total - hw_crtc_timing.v_border_top - + hw_crtc_timing.v_addressable - hw_crtc_timing.v_border_bottom - + hw_crtc_timing.v_front_porch; #if defined(CONFIG_DRM_AMD_DC_DCN1_0) @@ -486,21 +495,21 @@ static void dce110_stream_encoder_dp_set_stream_attribute( if (REG(DP_MSA_TIMING_PARAM3)) REG_SET_4(DP_MSA_TIMING_PARAM3, 0, DP_MSA_HSYNCWIDTH, - crtc_timing->h_sync_width, + hw_crtc_timing.h_sync_width, DP_MSA_HSYNCPOLARITY, - !crtc_timing->flags.HSYNC_POSITIVE_POLARITY, + !hw_crtc_timing.flags.HSYNC_POSITIVE_POLARITY, DP_MSA_VSYNCWIDTH, - crtc_timing->v_sync_width, + hw_crtc_timing.v_sync_width, DP_MSA_VSYNCPOLARITY, - !crtc_timing->flags.VSYNC_POSITIVE_POLARITY); + !hw_crtc_timing.flags.VSYNC_POSITIVE_POLARITY); /* HWDITH include border or overscan */ if (REG(DP_MSA_TIMING_PARAM4)) REG_SET_2(DP_MSA_TIMING_PARAM4, 0, - DP_MSA_HWIDTH, crtc_timing->h_border_left + - crtc_timing->h_addressable + crtc_timing->h_border_right, - DP_MSA_VHEIGHT, crtc_timing->v_border_top + - crtc_timing->v_addressable + crtc_timing->v_border_bottom); + DP_MSA_HWIDTH, hw_crtc_timing.h_border_left + + hw_crtc_timing.h_addressable + hw_crtc_timing.h_border_right, + DP_MSA_VHEIGHT, hw_crtc_timing.v_border_top + + hw_crtc_timing.v_addressable + hw_crtc_timing.v_border_bottom); #endif } #endif @@ -662,7 +671,7 @@ static void dce110_stream_encoder_dvi_set_stream_attribute( cntl.signal = is_dual_link ? SIGNAL_TYPE_DVI_DUAL_LINK : SIGNAL_TYPE_DVI_SINGLE_LINK; cntl.enable_dp_audio = false; - cntl.pixel_clock = crtc_timing->pix_clk_khz; + cntl.pixel_clock = crtc_timing->pix_clk_100hz / 10; cntl.lanes_number = (is_dual_link) ? LANE_COUNT_EIGHT : LANE_COUNT_FOUR; if (enc110->base.bp->funcs->encoder_control( @@ -686,7 +695,7 @@ static void dce110_stream_encoder_lvds_set_stream_attribute( cntl.engine_id = enc110->base.id; cntl.signal = SIGNAL_TYPE_LVDS; cntl.enable_dp_audio = false; - cntl.pixel_clock = crtc_timing->pix_clk_khz; + cntl.pixel_clock = crtc_timing->pix_clk_100hz / 10; cntl.lanes_number = LANE_COUNT_FOUR; if (enc110->base.bp->funcs->encoder_control( @@ -1575,6 +1584,14 @@ static void setup_stereo_sync( REG_UPDATE(DIG_FE_CNTL, DIG_STEREOSYNC_GATE_EN, !enable); } +static void dig_connect_to_otg( + struct stream_encoder *enc, + int tg_inst) +{ + struct dce110_stream_encoder *enc110 = DCE110STRENC_FROM_STRENC(enc); + + REG_UPDATE(DIG_FE_CNTL, DIG_SOURCE_SELECT, tg_inst); +} static const struct stream_encoder_funcs dce110_str_enc_funcs = { .dp_set_stream_attribute = @@ -1609,7 +1626,7 @@ static const struct stream_encoder_funcs dce110_str_enc_funcs = { .hdmi_audio_disable = dce110_se_hdmi_audio_disable, .setup_stereo_sync = setup_stereo_sync, .set_avmute = dce110_stream_encoder_set_avmute, - + .dig_connect_to_otg = dig_connect_to_otg, }; void dce110_stream_encoder_construct( diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.h b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.h index 6c28229c76eb..f9cdf2b5242c 100644 --- a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.h +++ b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.h @@ -199,7 +199,8 @@ SE_SF(DP_SEC_CNTL, DP_SEC_ATP_ENABLE, mask_sh),\ SE_SF(DP_SEC_CNTL, DP_SEC_AIP_ENABLE, mask_sh),\ SE_SF(DP_SEC_CNTL, DP_SEC_ACM_ENABLE, mask_sh),\ - SE_SF(AFMT_AUDIO_PACKET_CONTROL, AFMT_AUDIO_SAMPLE_SEND, mask_sh) + SE_SF(AFMT_AUDIO_PACKET_CONTROL, AFMT_AUDIO_SAMPLE_SEND, mask_sh),\ + SE_SF(DIG_FE_CNTL, DIG_SOURCE_SELECT, mask_sh) #define SE_COMMON_MASK_SH_LIST_DCE_COMMON(mask_sh)\ SE_COMMON_MASK_SH_LIST_DCE_COMMON_BASE(mask_sh) @@ -284,7 +285,8 @@ SE_SF(DIG0_DIG_FE_CNTL, TMDS_PIXEL_ENCODING, mask_sh),\ SE_SF(DIG0_DIG_FE_CNTL, TMDS_COLOR_FORMAT, mask_sh),\ SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_SELECT, mask_sh),\ - SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_GATE_EN, mask_sh) + SE_SF(DIG0_DIG_FE_CNTL, DIG_STEREOSYNC_GATE_EN, mask_sh),\ + SE_SF(DIG0_DIG_FE_CNTL, DIG_SOURCE_SELECT, mask_sh) #define SE_COMMON_MASK_SH_LIST_SOC(mask_sh)\ SE_COMMON_MASK_SH_LIST_SOC_BASE(mask_sh) @@ -494,6 +496,7 @@ struct dce_stream_encoder_shift { uint8_t HDMI_DB_DISABLE; uint8_t DP_VID_N_MUL; uint8_t DP_VID_M_DOUBLE_VALUE_EN; + uint8_t DIG_SOURCE_SELECT; }; struct dce_stream_encoder_mask { @@ -624,6 +627,7 @@ struct dce_stream_encoder_mask { uint32_t HDMI_DB_DISABLE; uint32_t DP_VID_N_MUL; uint32_t DP_VID_M_DOUBLE_VALUE_EN; + uint32_t DIG_SOURCE_SELECT; }; struct dce110_stream_enc_registers { diff --git a/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c b/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c index 6ae51a5dfc04..23044e6723e8 100644 --- a/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c @@ -76,6 +76,7 @@ #ifndef mmBIOS_SCRATCH_2 #define mmBIOS_SCRATCH_2 0x05CB + #define mmBIOS_SCRATCH_3 0x05CC #define mmBIOS_SCRATCH_6 0x05CF #endif @@ -365,6 +366,7 @@ static const struct dce_abm_mask abm_mask = { #define DCFE_MEM_PWR_CTRL_REG_BASE 0x1b03 static const struct bios_registers bios_regs = { + .BIOS_SCRATCH_3 = mmBIOS_SCRATCH_3, .BIOS_SCRATCH_6 = mmBIOS_SCRATCH_6 }; @@ -587,7 +589,7 @@ struct output_pixel_processor *dce100_opp_create( return &opp->base; } -struct aux_engine *dce100_aux_engine_create( +struct dce_aux *dce100_aux_engine_create( struct dc_context *ctx, uint32_t inst) { diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c index 52d50e24a995..7b23239d33fe 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c @@ -62,8 +62,6 @@ static const struct dce110_compressor_reg_offsets reg_offsets[] = { } }; -static const uint32_t dce11_one_lpt_channel_max_resolution = 2560 * 1600; - static uint32_t align_to_chunks_number_per_line(uint32_t pixels) { return 256 * ((pixels + 255) / 256); diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c index 4bf24758217f..e1b285ea01ac 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c @@ -614,55 +614,6 @@ dce110_set_output_transfer_func(struct pipe_ctx *pipe_ctx, return true; } -static enum dc_status bios_parser_crtc_source_select( - struct pipe_ctx *pipe_ctx) -{ - struct dc_bios *dcb; - /* call VBIOS table to set CRTC source for the HW - * encoder block - * note: video bios clears all FMT setting here. */ - struct bp_crtc_source_select crtc_source_select = {0}; - const struct dc_sink *sink = pipe_ctx->stream->sink; - - crtc_source_select.engine_id = pipe_ctx->stream_res.stream_enc->id; - crtc_source_select.controller_id = pipe_ctx->stream_res.tg->inst + 1; - /*TODO: Need to un-hardcode color depth, dp_audio and account for - * the case where signal and sink signal is different (translator - * encoder)*/ - crtc_source_select.signal = pipe_ctx->stream->signal; - crtc_source_select.enable_dp_audio = false; - crtc_source_select.sink_signal = pipe_ctx->stream->signal; - - switch (pipe_ctx->stream->timing.display_color_depth) { - case COLOR_DEPTH_666: - crtc_source_select.display_output_bit_depth = PANEL_6BIT_COLOR; - break; - case COLOR_DEPTH_888: - crtc_source_select.display_output_bit_depth = PANEL_8BIT_COLOR; - break; - case COLOR_DEPTH_101010: - crtc_source_select.display_output_bit_depth = PANEL_10BIT_COLOR; - break; - case COLOR_DEPTH_121212: - crtc_source_select.display_output_bit_depth = PANEL_12BIT_COLOR; - break; - default: - BREAK_TO_DEBUGGER(); - crtc_source_select.display_output_bit_depth = PANEL_8BIT_COLOR; - break; - } - - dcb = sink->ctx->dc_bios; - - if (BP_RESULT_OK != dcb->funcs->crtc_source_select( - dcb, - &crtc_source_select)) { - return DC_ERROR_UNEXPECTED; - } - - return DC_OK; -} - void dce110_update_info_frame(struct pipe_ctx *pipe_ctx) { bool is_hdmi; @@ -692,10 +643,10 @@ void dce110_update_info_frame(struct pipe_ctx *pipe_ctx) void dce110_enable_stream(struct pipe_ctx *pipe_ctx) { enum dc_lane_count lane_count = - pipe_ctx->stream->sink->link->cur_link_settings.lane_count; + pipe_ctx->stream->link->cur_link_settings.lane_count; struct dc_crtc_timing *timing = &pipe_ctx->stream->timing; - struct dc_link *link = pipe_ctx->stream->sink->link; + struct dc_link *link = pipe_ctx->stream->link; uint32_t active_total_with_borders; @@ -1000,7 +951,7 @@ void dce110_enable_audio_stream(struct pipe_ctx *pipe_ctx) pipe_ctx->stream_res.audio->funcs->az_enable(pipe_ctx->stream_res.audio); - if (num_audio == 1 && pp_smu != NULL && pp_smu->set_pme_wa_enable != NULL) + if (num_audio >= 1 && pp_smu != NULL && pp_smu->set_pme_wa_enable != NULL) /*this is the first audio. apply the PME w/a in order to wake AZ from D3*/ pp_smu->set_pme_wa_enable(&pp_smu->pp_smu); /* un-mute audio */ @@ -1017,6 +968,8 @@ void dce110_disable_audio_stream(struct pipe_ctx *pipe_ctx, int option) pipe_ctx->stream_res.stream_enc->funcs->audio_mute_control( pipe_ctx->stream_res.stream_enc, true); if (pipe_ctx->stream_res.audio) { + struct pp_smu_funcs_rv *pp_smu = dc->res_pool->pp_smu; + if (option != KEEP_ACQUIRED_RESOURCE || !dc->debug.az_endpoint_mute_only) { /*only disalbe az_endpoint if power down or free*/ @@ -1036,6 +989,9 @@ void dce110_disable_audio_stream(struct pipe_ctx *pipe_ctx, int option) update_audio_usage(&dc->current_state->res_ctx, dc->res_pool, pipe_ctx->stream_res.audio, false); pipe_ctx->stream_res.audio = NULL; } + if (pp_smu != NULL && pp_smu->set_pme_wa_enable != NULL) + /*this is the first audio. apply the PME w/a in order to wake AZ from D3*/ + pp_smu->set_pme_wa_enable(&pp_smu->pp_smu); /* TODO: notify audio driver for if audio modes list changed * add audio mode list change flag */ @@ -1048,7 +1004,7 @@ void dce110_disable_audio_stream(struct pipe_ctx *pipe_ctx, int option) void dce110_disable_stream(struct pipe_ctx *pipe_ctx, int option) { struct dc_stream_state *stream = pipe_ctx->stream; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; struct dc *dc = pipe_ctx->stream->ctx->dc; if (dc_is_hdmi_signal(pipe_ctx->stream->signal)) @@ -1073,11 +1029,10 @@ void dce110_unblank_stream(struct pipe_ctx *pipe_ctx, { struct encoder_unblank_param params = { { 0 } }; struct dc_stream_state *stream = pipe_ctx->stream; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; /* only 3 items below are used by unblank */ - params.pixel_clk_khz = - pipe_ctx->stream->timing.pix_clk_khz; + params.pixel_clk_khz = pipe_ctx->stream->timing.pix_clk_100hz / 10; params.link_settings.link_rate = link_settings->link_rate; if (dc_is_dp_signal(pipe_ctx->stream->signal)) @@ -1087,10 +1042,11 @@ void dce110_unblank_stream(struct pipe_ctx *pipe_ctx, link->dc->hwss.edp_backlight_control(link, true); } } + void dce110_blank_stream(struct pipe_ctx *pipe_ctx) { struct dc_stream_state *stream = pipe_ctx->stream; - struct dc_link *link = stream->sink->link; + struct dc_link *link = stream->link; if (link->local_sink && link->local_sink->sink_signal == SIGNAL_TYPE_EDP) { link->dc->hwss.edp_backlight_control(link, false); @@ -1163,27 +1119,27 @@ static void build_audio_output( stream->timing.flags.INTERLACE; audio_output->crtc_info.refresh_rate = - (stream->timing.pix_clk_khz*1000)/ + (stream->timing.pix_clk_100hz*10000)/ (stream->timing.h_total*stream->timing.v_total); audio_output->crtc_info.color_depth = stream->timing.display_color_depth; audio_output->crtc_info.requested_pixel_clock = - pipe_ctx->stream_res.pix_clk_params.requested_pix_clk; + pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz / 10; audio_output->crtc_info.calculated_pixel_clock = - pipe_ctx->stream_res.pix_clk_params.requested_pix_clk; + pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz / 10; /*for HDMI, audio ACR is with deep color ratio factor*/ if (dc_is_hdmi_signal(pipe_ctx->stream->signal) && audio_output->crtc_info.requested_pixel_clock == - stream->timing.pix_clk_khz) { + (stream->timing.pix_clk_100hz / 10)) { if (pipe_ctx->stream_res.pix_clk_params.pixel_encoding == PIXEL_ENCODING_YCBCR420) { audio_output->crtc_info.requested_pixel_clock = audio_output->crtc_info.requested_pixel_clock/2; audio_output->crtc_info.calculated_pixel_clock = - pipe_ctx->stream_res.pix_clk_params.requested_pix_clk/2; + pipe_ctx->stream_res.pix_clk_params.requested_pix_clk_100hz/20; } } @@ -1294,8 +1250,6 @@ static enum dc_status dce110_enable_stream_timing( struct pipe_ctx *pipe_ctx_old = &dc->current_state->res_ctx. pipe_ctx[pipe_ctx->pipe_idx]; struct tg_color black_color = {0}; - struct drr_params params = {0}; - unsigned int event_triggers = 0; if (!pipe_ctx_old->stream) { @@ -1324,20 +1278,6 @@ static enum dc_status dce110_enable_stream_timing( pipe_ctx->stream_res.tg, &stream->timing, true); - - params.vertical_total_min = stream->adjust.v_total_min; - params.vertical_total_max = stream->adjust.v_total_max; - if (pipe_ctx->stream_res.tg->funcs->set_drr) - pipe_ctx->stream_res.tg->funcs->set_drr( - pipe_ctx->stream_res.tg, ¶ms); - - // DRR should set trigger event to monitor surface update event - if (stream->adjust.v_total_min != 0 && - stream->adjust.v_total_max != 0) - event_triggers = 0x80; - if (pipe_ctx->stream_res.tg->funcs->set_static_screen_control) - pipe_ctx->stream_res.tg->funcs->set_static_screen_control( - pipe_ctx->stream_res.tg, event_triggers); } if (!pipe_ctx_old->stream) { @@ -1357,6 +1297,8 @@ static enum dc_status apply_single_controller_ctx_to_hw( struct dc *dc) { struct dc_stream_state *stream = pipe_ctx->stream; + struct drr_params params = {0}; + unsigned int event_triggers = 0; if (pipe_ctx->stream_res.audio != NULL) { struct audio_output audio_output; @@ -1383,14 +1325,32 @@ static enum dc_status apply_single_controller_ctx_to_hw( } /* */ - dc->hwss.enable_stream_timing(pipe_ctx, context, dc); + /* Do not touch stream timing on seamless boot optimization. */ + if (!pipe_ctx->stream->apply_seamless_boot_optimization) + dc->hwss.enable_stream_timing(pipe_ctx, context, dc); + + if (pipe_ctx->stream_res.tg->funcs->program_vupdate_interrupt) + pipe_ctx->stream_res.tg->funcs->program_vupdate_interrupt( + pipe_ctx->stream_res.tg, + &stream->timing); + + params.vertical_total_min = stream->adjust.v_total_min; + params.vertical_total_max = stream->adjust.v_total_max; + if (pipe_ctx->stream_res.tg->funcs->set_drr) + pipe_ctx->stream_res.tg->funcs->set_drr( + pipe_ctx->stream_res.tg, ¶ms); + + // DRR should set trigger event to monitor surface update event + if (stream->adjust.v_total_min != 0 && stream->adjust.v_total_max != 0) + event_triggers = 0x80; + if (pipe_ctx->stream_res.tg->funcs->set_static_screen_control) + pipe_ctx->stream_res.tg->funcs->set_static_screen_control( + pipe_ctx->stream_res.tg, event_triggers); - /* TODO: move to stream encoder */ if (pipe_ctx->stream->signal != SIGNAL_TYPE_VIRTUAL) - if (DC_OK != bios_parser_crtc_source_select(pipe_ctx)) { - BREAK_TO_DEBUGGER(); - return DC_ERROR_UNEXPECTED; - } + pipe_ctx->stream_res.stream_enc->funcs->dig_connect_to_otg( + pipe_ctx->stream_res.stream_enc, + pipe_ctx->stream_res.tg->inst); pipe_ctx->stream_res.opp->funcs->opp_set_dyn_expansion( pipe_ctx->stream_res.opp, @@ -1408,7 +1368,7 @@ static enum dc_status apply_single_controller_ctx_to_hw( pipe_ctx->plane_res.scl_data.lb_params.alpha_en = pipe_ctx->bottom_pipe != 0; - pipe_ctx->stream->sink->link->psr_enabled = false; + pipe_ctx->stream->link->psr_enabled = false; return DC_OK; } @@ -1518,7 +1478,7 @@ static struct dc_link *get_link_for_edp(struct dc *dc) return NULL; } -static struct dc_link *get_link_for_edp_not_in_use( +static struct dc_link *get_link_for_edp_to_turn_off( struct dc *dc, struct dc_state *context) { @@ -1527,8 +1487,12 @@ static struct dc_link *get_link_for_edp_not_in_use( /* check if eDP panel is suppose to be set mode, if yes, no need to disable */ for (i = 0; i < context->stream_count; i++) { - if (context->streams[i]->signal == SIGNAL_TYPE_EDP) - return NULL; + if (context->streams[i]->signal == SIGNAL_TYPE_EDP) { + if (context->streams[i]->dpms_off == true) + return context->streams[i]->sink->link; + else + return NULL; + } } /* check if there is an eDP panel not in use */ @@ -1555,7 +1519,6 @@ void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context) int i; struct dc_link *edp_link_to_turnoff = NULL; struct dc_link *edp_link = get_link_for_edp(dc); - struct dc_bios *bios = dc->ctx->dc_bios; bool can_edp_fast_boot_optimize = false; bool apply_edp_fast_boot_optimization = false; @@ -1571,7 +1534,7 @@ void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context) } if (can_edp_fast_boot_optimize) - edp_link_to_turnoff = get_link_for_edp_not_in_use(dc, context); + edp_link_to_turnoff = get_link_for_edp_to_turn_off(dc, context); /* if OS doesn't light up eDP and eDP link is available, we want to disable * If resume from S4/S5, should optimization. @@ -1582,20 +1545,6 @@ void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context) if (context->streams[i]->signal == SIGNAL_TYPE_EDP) { context->streams[i]->apply_edp_fast_boot_optimization = true; apply_edp_fast_boot_optimization = true; - - /* When after S4 and S5, vbios may post edp and previous dpms_off - * doesn't make sense. - * Update dpms_off state to align hw and sw state via check - * vBios scratch register. - */ - if (bios->funcs->is_active_display) { - const struct connector_device_tag_info *device_tag = &(edp_link->device_tag); - - if (bios->funcs->is_active_display(bios, - context->streams[i]->signal, - device_tag)) - context->streams[i]->dpms_off = false; - } } } } @@ -1624,8 +1573,8 @@ static uint32_t compute_pstate_blackout_duration( pstate_blackout_duration_ns = 1000 * blackout_duration.value >> 24; total_dest_line_time_ns = 1000000UL * - stream->timing.h_total / - stream->timing.pix_clk_khz + + (stream->timing.h_total * 10) / + stream->timing.pix_clk_100hz + pstate_blackout_duration_ns; return total_dest_line_time_ns; @@ -1813,18 +1762,15 @@ static bool should_enable_fbc(struct dc *dc, if (i == dc->res_pool->pipe_count) return false; - if (!pipe_ctx->stream->sink) - return false; - - if (!pipe_ctx->stream->sink->link) + if (!pipe_ctx->stream->link) return false; /* Only supports eDP */ - if (pipe_ctx->stream->sink->link->connector_signal != SIGNAL_TYPE_EDP) + if (pipe_ctx->stream->link->connector_signal != SIGNAL_TYPE_EDP) return false; /* PSR should not be enabled */ - if (pipe_ctx->stream->sink->link->psr_enabled) + if (pipe_ctx->stream->link->psr_enabled) return false; /* Nothing to compress */ @@ -2329,6 +2275,11 @@ static void dce110_enable_per_frame_crtc_position_reset( } +static void init_pipes(struct dc *dc, struct dc_state *context) +{ + // Do nothing +} + static void init_hw(struct dc *dc) { int i; @@ -2573,7 +2524,7 @@ static void dce110_apply_ctx_for_surface( pipe_ctx->plane_res.mi, pipe_ctx->stream->timing.h_total, pipe_ctx->stream->timing.v_total, - pipe_ctx->stream->timing.pix_clk_khz, + pipe_ctx->stream->timing.pix_clk_100hz / 10, context->stream_count); dce110_program_front_end_for_pipe(dc, pipe_ctx); @@ -2595,7 +2546,7 @@ static void dce110_apply_ctx_for_surface( } if (dc->fbc_compressor) - enable_fbc(dc, dc->current_state); + enable_fbc(dc, context); } static void dce110_power_down_fe(struct dc *dc, struct pipe_ctx *pipe_ctx) @@ -2622,13 +2573,35 @@ static void dce110_wait_for_mpcc_disconnect( /* do nothing*/ } +static void program_output_csc(struct dc *dc, + struct pipe_ctx *pipe_ctx, + enum dc_color_space colorspace, + uint16_t *matrix, + int opp_id) +{ + int i; + struct out_csc_color_matrix tbl_entry; + + if (pipe_ctx->stream->csc_color_matrix.enable_adjustment == true) { + enum dc_color_space color_space = pipe_ctx->stream->output_color_space; + + for (i = 0; i < 12; i++) + tbl_entry.regval[i] = pipe_ctx->stream->csc_color_matrix.matrix[i]; + + tbl_entry.color_space = color_space; + + pipe_ctx->plane_res.xfm->funcs->opp_set_csc_adjustment( + pipe_ctx->plane_res.xfm, &tbl_entry); + } +} + void dce110_set_cursor_position(struct pipe_ctx *pipe_ctx) { struct dc_cursor_position pos_cpy = pipe_ctx->stream->cursor_position; struct input_pixel_processor *ipp = pipe_ctx->plane_res.ipp; struct mem_input *mi = pipe_ctx->plane_res.mi; struct dc_cursor_mi_param param = { - .pixel_clk_khz = pipe_ctx->stream->timing.pix_clk_khz, + .pixel_clk_khz = pipe_ctx->stream->timing.pix_clk_100hz / 10, .ref_clk_khz = pipe_ctx->stream->ctx->dc->res_pool->ref_clock_inKhz, .viewport = pipe_ctx->plane_res.scl_data.viewport, .h_scale_ratio = pipe_ctx->plane_res.scl_data.ratios.horz, @@ -2672,7 +2645,9 @@ void dce110_set_cursor_attribute(struct pipe_ctx *pipe_ctx) static const struct hw_sequencer_funcs dce110_funcs = { .program_gamut_remap = program_gamut_remap, + .program_output_csc = program_output_csc, .init_hw = init_hw, + .init_pipes = init_pipes, .apply_ctx_to_hw = dce110_apply_ctx_to_hw, .apply_ctx_for_surface = dce110_apply_ctx_for_surface, .update_plane_addr = update_plane_addr, diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c index e33d11785b1f..7549adaa1542 100644 --- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c @@ -84,6 +84,7 @@ #ifndef mmBIOS_SCRATCH_2 #define mmBIOS_SCRATCH_2 0x05CB + #define mmBIOS_SCRATCH_3 0x05CC #define mmBIOS_SCRATCH_6 0x05CF #endif @@ -369,6 +370,7 @@ static const struct dce110_clk_src_mask cs_mask = { }; static const struct bios_registers bios_regs = { + .BIOS_SCRATCH_3 = mmBIOS_SCRATCH_3, .BIOS_SCRATCH_6 = mmBIOS_SCRATCH_6 }; @@ -606,7 +608,7 @@ static struct output_pixel_processor *dce110_opp_create( return &opp->base; } -struct aux_engine *dce110_aux_engine_create( +struct dce_aux *dce110_aux_engine_create( struct dc_context *ctx, uint32_t inst) { @@ -779,8 +781,8 @@ static void get_pixel_clock_parameters( * the pixel clock normalization for hdmi up to here instead of doing it * in pll_adjust_pix_clk */ - pixel_clk_params->requested_pix_clk = stream->timing.pix_clk_khz; - pixel_clk_params->encoder_object_id = stream->sink->link->link_enc->id; + pixel_clk_params->requested_pix_clk_100hz = stream->timing.pix_clk_100hz; + pixel_clk_params->encoder_object_id = stream->link->link_enc->id; pixel_clk_params->signal_type = pipe_ctx->stream->signal; pixel_clk_params->controller_id = pipe_ctx->stream_res.tg->inst + 1; /* TODO: un-hardcode*/ @@ -797,10 +799,10 @@ static void get_pixel_clock_parameters( pixel_clk_params->color_depth = COLOR_DEPTH_888; } if (stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR420) { - pixel_clk_params->requested_pix_clk = pixel_clk_params->requested_pix_clk / 2; + pixel_clk_params->requested_pix_clk_100hz = pixel_clk_params->requested_pix_clk_100hz / 2; } if (stream->timing.timing_3d_format == TIMING_3D_FORMAT_HW_FRAME_PACKING) - pixel_clk_params->requested_pix_clk *= 2; + pixel_clk_params->requested_pix_clk_100hz *= 2; } @@ -874,7 +876,7 @@ static bool dce110_validate_bandwidth( __func__, context->streams[0]->timing.h_addressable, context->streams[0]->timing.v_addressable, - context->streams[0]->timing.pix_clk_khz); + context->streams[0]->timing.pix_clk_100hz / 10); if (memcmp(&dc->current_state->bw.dce, &context->bw.dce, sizeof(context->bw.dce))) { @@ -1055,7 +1057,7 @@ static struct pipe_ctx *dce110_acquire_underlay( pipe_ctx->plane_res.mi->funcs->allocate_mem_input(pipe_ctx->plane_res.mi, stream->timing.h_total, stream->timing.v_total, - stream->timing.pix_clk_khz, + stream->timing.pix_clk_100hz / 10, context->stream_count); color_space_to_black_color(dc, diff --git a/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c b/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c index 969d4e72dc94..ea3065d63372 100644 --- a/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce112/dce112_resource.c @@ -76,6 +76,7 @@ #ifndef mmBIOS_SCRATCH_2 #define mmBIOS_SCRATCH_2 0x05CB + #define mmBIOS_SCRATCH_3 0x05CC #define mmBIOS_SCRATCH_6 0x05CF #endif @@ -376,6 +377,7 @@ static const struct dce110_clk_src_mask cs_mask = { }; static const struct bios_registers bios_regs = { + .BIOS_SCRATCH_3 = mmBIOS_SCRATCH_3, .BIOS_SCRATCH_6 = mmBIOS_SCRATCH_6 }; @@ -607,7 +609,7 @@ struct output_pixel_processor *dce112_opp_create( return &opp->base; } -struct aux_engine *dce112_aux_engine_create( +struct dce_aux *dce112_aux_engine_create( struct dc_context *ctx, uint32_t inst) { @@ -763,7 +765,7 @@ static struct clock_source *find_matching_pll( const struct resource_pool *pool, const struct dc_stream_state *const stream) { - switch (stream->sink->link->link_enc->transmitter) { + switch (stream->link->link_enc->transmitter) { case TRANSMITTER_UNIPHY_A: return pool->clock_sources[DCE112_CLK_SRC_PLL0]; case TRANSMITTER_UNIPHY_B: diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c index eb0f5f9a973b..1ca30928025e 100644 --- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.c @@ -244,6 +244,21 @@ static void dce120_update_dchub( dh_data->dchub_info_valid = false; } +/** + * dce121_xgmi_enabled() - Check if xGMI is enabled + * @hws: DCE hardware sequencer object + * + * Return true if xGMI is enabled. False otherwise. + */ +bool dce121_xgmi_enabled(struct dce_hwseq *hws) +{ + uint32_t pf_max_region; + + REG_GET(MC_VM_XGMI_LFB_CNTL, PF_MAX_REGION, &pf_max_region); + /* PF_MAX_REGION == 0 means xgmi is disabled */ + return !!pf_max_region; +} + void dce120_hw_sequencer_construct(struct dc *dc) { /* All registers used by dce11.2 match those in dce11 in offset and diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.h index 77a6b86d7606..c51afbd0b012 100644 --- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.h +++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_hw_sequencer.h @@ -30,6 +30,7 @@ struct dc; +bool dce121_xgmi_enabled(struct dce_hwseq *hws); void dce120_hw_sequencer_construct(struct dc *dc); #endif /* __DC_HWSS_DCE112_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c b/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c index f12696674eb0..312a0aebf91f 100644 --- a/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce120/dce120_resource.c @@ -62,6 +62,8 @@ #include "soc15_hw_ip.h" #include "vega10_ip_offset.h" #include "nbio/nbio_6_1_offset.h" +#include "mmhub/mmhub_9_4_0_offset.h" +#include "mmhub/mmhub_9_4_0_sh_mask.h" #include "reg_helper.h" #include "dce100/dce100_resource.h" @@ -139,6 +141,17 @@ static const struct dce110_timing_generator_offsets dce120_tg_offsets[] = { .reg_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \ mm ## block ## id ## _ ## reg_name +/* MMHUB */ +#define MMHUB_BASE_INNER(seg) \ + MMHUB_BASE__INST0_SEG ## seg + +#define MMHUB_BASE(seg) \ + MMHUB_BASE_INNER(seg) + +#define MMHUB_SR(reg_name)\ + .reg_name = MMHUB_BASE(mm ## reg_name ## _BASE_IDX) + \ + mm ## reg_name + /* macros to expend register list macro defined in HW object header file * end *********************/ @@ -378,7 +391,7 @@ struct output_pixel_processor *dce120_opp_create( ctx, inst, &opp_regs[inst], &opp_shift, &opp_mask); return &opp->base; } -struct aux_engine *dce120_aux_engine_create( +struct dce_aux *dce120_aux_engine_create( struct dc_context *ctx, uint32_t inst) { @@ -429,6 +442,7 @@ struct dce_i2c_hw *dce120_i2c_hw_create( return dce_i2c_hw; } static const struct bios_registers bios_regs = { + .BIOS_SCRATCH_3 = mmBIOS_SCRATCH_3 + NBIO_BASE(mmBIOS_SCRATCH_3_BASE_IDX), .BIOS_SCRATCH_6 = mmBIOS_SCRATCH_6 + NBIO_BASE(mmBIOS_SCRATCH_6_BASE_IDX) }; @@ -681,6 +695,19 @@ static const struct dce_hwseq_mask hwseq_mask = { HWSEQ_DCE12_MASK_SH_LIST(_MASK) }; +/* HWSEQ regs for VG20 */ +static const struct dce_hwseq_registers dce121_hwseq_reg = { + HWSEQ_VG20_REG_LIST() +}; + +static const struct dce_hwseq_shift dce121_hwseq_shift = { + HWSEQ_VG20_MASK_SH_LIST(__SHIFT) +}; + +static const struct dce_hwseq_mask dce121_hwseq_mask = { + HWSEQ_VG20_MASK_SH_LIST(_MASK) +}; + static struct dce_hwseq *dce120_hwseq_create( struct dc_context *ctx) { @@ -695,6 +722,20 @@ static struct dce_hwseq *dce120_hwseq_create( return hws; } +static struct dce_hwseq *dce121_hwseq_create( + struct dc_context *ctx) +{ + struct dce_hwseq *hws = kzalloc(sizeof(struct dce_hwseq), GFP_KERNEL); + + if (hws) { + hws->ctx = ctx; + hws->regs = &dce121_hwseq_reg; + hws->shifts = &dce121_hwseq_shift; + hws->masks = &dce121_hwseq_mask; + } + return hws; +} + static const struct resource_create_funcs res_create_funcs = { .read_dce_straps = read_dce_straps, .create_audio = create_audio, @@ -702,6 +743,14 @@ static const struct resource_create_funcs res_create_funcs = { .create_hwseq = dce120_hwseq_create, }; +static const struct resource_create_funcs dce121_res_create_funcs = { + .read_dce_straps = read_dce_straps, + .create_audio = create_audio, + .create_stream_encoder = dce120_stream_encoder_create, + .create_hwseq = dce121_hwseq_create, +}; + + #define mi_inst_regs(id) { MI_DCE12_REG_LIST(id) } static const struct dce_mem_input_registers mi_regs[] = { mi_inst_regs(0), @@ -911,7 +960,8 @@ static bool construct( int j; struct dc_context *ctx = dc->ctx; struct irq_service_init_data irq_init_data; - bool harvest_enabled = ASICREV_IS_VEGA20_P(ctx->asic_id.hw_internal_rev); + static const struct resource_create_funcs *res_funcs; + bool is_vg20 = ASICREV_IS_VEGA20_P(ctx->asic_id.hw_internal_rev); uint32_t pipe_fuses; ctx->dc_bios->regs = &bios_regs; @@ -975,7 +1025,11 @@ static bool construct( } } - pool->base.clk_mgr = dce120_clk_mgr_create(ctx); + if (is_vg20) + pool->base.clk_mgr = dce121_clk_mgr_create(ctx); + else + pool->base.clk_mgr = dce120_clk_mgr_create(ctx); + if (pool->base.clk_mgr == NULL) { dm_error("DC: failed to create display clock!\n"); BREAK_TO_DEBUGGER(); @@ -1008,14 +1062,14 @@ static bool construct( if (!pool->base.irqs) goto irqs_create_fail; - /* retrieve valid pipe fuses */ - if (harvest_enabled) + /* VG20: Pipe harvesting enabled, retrieve valid pipe fuses */ + if (is_vg20) pipe_fuses = read_pipe_fuses(ctx); /* index to valid pipe resource */ j = 0; for (i = 0; i < pool->base.pipe_count; i++) { - if (harvest_enabled) { + if (is_vg20) { if ((pipe_fuses & (1 << i)) != 0) { dm_error("DC: skip invalid pipe %d!\n", i); continue; @@ -1093,10 +1147,24 @@ static bool construct( pool->base.pipe_count = j; pool->base.timing_generator_count = j; - if (!resource_construct(num_virtual_links, dc, &pool->base, - &res_create_funcs)) + if (is_vg20) + res_funcs = &dce121_res_create_funcs; + else + res_funcs = &res_create_funcs; + + if (!resource_construct(num_virtual_links, dc, &pool->base, res_funcs)) goto res_create_fail; + /* + * This is a bit of a hack. The xGMI enabled info is used to determine + * if audio and display clocks need to be adjusted with the WAFL link's + * SS info. This is a responsiblity of the clk_mgr. But since MMHUB is + * under hwseq, and the relevant register is in MMHUB, we have to do it + * here. + */ + if (is_vg20 && dce121_xgmi_enabled(dc->hwseq)) + dce121_clock_patch_xgmi_ss_info(pool->base.clk_mgr); + /* Create hardware sequencer */ if (!dce120_hw_sequencer_create(dc)) goto controller_create_fail; diff --git a/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c b/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c index cdd1d6b7b9f2..2eca81b5cf2f 100644 --- a/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dce80/dce80_resource.c @@ -77,6 +77,7 @@ #ifndef mmBIOS_SCRATCH_2 #define mmBIOS_SCRATCH_2 0x05CB + #define mmBIOS_SCRATCH_3 0x05CC #define mmBIOS_SCRATCH_6 0x05CF #endif @@ -358,6 +359,7 @@ static const struct dce110_clk_src_mask cs_mask = { }; static const struct bios_registers bios_regs = { + .BIOS_SCRATCH_3 = mmBIOS_SCRATCH_3, .BIOS_SCRATCH_6 = mmBIOS_SCRATCH_6 }; @@ -467,7 +469,7 @@ static struct output_pixel_processor *dce80_opp_create( return &opp->base; } -struct aux_engine *dce80_aux_engine_create( +struct dce_aux *dce80_aux_engine_create( struct dc_context *ctx, uint32_t inst) { diff --git a/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c b/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c index 3ba4712a35ab..8b5ce557ee71 100644 --- a/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c +++ b/drivers/gpu/drm/amd/display/dc/dce80/dce80_timing_generator.c @@ -84,17 +84,17 @@ static const struct dce110_timing_generator_offsets reg_offsets[] = { #define DCP_REG(reg) (reg + tg110->offsets.dcp) #define DMIF_REG(reg) (reg + tg110->offsets.dmif) -static void program_pix_dur(struct timing_generator *tg, uint32_t pix_clk_khz) +static void program_pix_dur(struct timing_generator *tg, uint32_t pix_clk_100hz) { uint64_t pix_dur; uint32_t addr = mmDMIF_PG0_DPG_PIPE_ARBITRATION_CONTROL1 + DCE110TG_FROM_TG(tg)->offsets.dmif; uint32_t value = dm_read_reg(tg->ctx, addr); - if (pix_clk_khz == 0) + if (pix_clk_100hz == 0) return; - pix_dur = 1000000000 / pix_clk_khz; + pix_dur = div_u64(10000000000ull, pix_clk_100hz); set_reg_field_value( value, @@ -110,7 +110,7 @@ static void program_timing(struct timing_generator *tg, bool use_vbios) { if (!use_vbios) - program_pix_dur(tg, timing->pix_clk_khz); + program_pix_dur(tg, timing->pix_clk_100hz); dce110_tg_program_timing(tg, timing, use_vbios); } diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.c index 54abedbf1b43..afe8c42211cd 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_clk_mgr.c @@ -161,69 +161,17 @@ static int get_active_display_cnt( return display_count; } -static void notify_deep_sleep_dcfclk_to_smu( - struct pp_smu_funcs_rv *pp_smu, int min_dcef_deep_sleep_clk_khz) -{ - int min_dcef_deep_sleep_clk_mhz; //minimum required DCEF Deep Sleep clock in mhz - /* - * if function pointer not set up, this message is - * sent as part of pplib_apply_display_requirements. - * So just return. - */ - if (!pp_smu || !pp_smu->set_min_deep_sleep_dcfclk) - return; - - min_dcef_deep_sleep_clk_mhz = (min_dcef_deep_sleep_clk_khz + 999) / 1000; //Round up - pp_smu->set_min_deep_sleep_dcfclk(&pp_smu->pp_smu, min_dcef_deep_sleep_clk_mhz); -} - -static void notify_hard_min_dcfclk_to_smu( - struct pp_smu_funcs_rv *pp_smu, int min_dcf_clk_khz) -{ - int min_dcf_clk_mhz; //minimum required DCF clock in mhz - - /* - * if function pointer not set up, this message is - * sent as part of pplib_apply_display_requirements. - * So just return. - */ - if (!pp_smu || !pp_smu->set_hard_min_dcfclk_by_freq) - return; - - min_dcf_clk_mhz = min_dcf_clk_khz / 1000; - - pp_smu->set_hard_min_dcfclk_by_freq(&pp_smu->pp_smu, min_dcf_clk_mhz); -} - -static void notify_hard_min_fclk_to_smu( - struct pp_smu_funcs_rv *pp_smu, int min_f_clk_khz) -{ - int min_f_clk_mhz; //minimum required F clock in mhz - - /* - * if function pointer not set up, this message is - * sent as part of pplib_apply_display_requirements. - * So just return. - */ - if (!pp_smu || !pp_smu->set_hard_min_fclk_by_freq) - return; - - min_f_clk_mhz = min_f_clk_khz / 1000; - - pp_smu->set_hard_min_fclk_by_freq(&pp_smu->pp_smu, min_f_clk_mhz); -} - static void dcn1_update_clocks(struct clk_mgr *clk_mgr, struct dc_state *context, bool safe_to_lower) { struct dc *dc = clk_mgr->ctx->dc; + struct dc_debug_options *debug = &dc->debug; struct dc_clocks *new_clocks = &context->bw.dcn.clk; struct pp_smu_display_requirement_rv *smu_req_cur = &dc->res_pool->pp_smu_req; struct pp_smu_display_requirement_rv smu_req = *smu_req_cur; struct pp_smu_funcs_rv *pp_smu = dc->res_pool->pp_smu; - uint32_t requested_dcf_clock_in_khz = 0; bool send_request_to_increase = false; bool send_request_to_lower = false; int display_count; @@ -243,9 +191,8 @@ static void dcn1_update_clocks(struct clk_mgr *clk_mgr, */ if (pp_smu->set_display_count) pp_smu->set_display_count(&pp_smu->pp_smu, display_count); - else - smu_req.display_count = display_count; + smu_req.display_count = display_count; } if (new_clocks->dispclk_khz > clk_mgr->clks.dispclk_khz @@ -261,12 +208,13 @@ static void dcn1_update_clocks(struct clk_mgr *clk_mgr, } // F Clock + if (debug->force_fclk_khz != 0) + new_clocks->fclk_khz = debug->force_fclk_khz; + if (should_set_clock(safe_to_lower, new_clocks->fclk_khz, clk_mgr->clks.fclk_khz)) { clk_mgr->clks.fclk_khz = new_clocks->fclk_khz; smu_req.hard_min_fclk_mhz = new_clocks->fclk_khz / 1000; - notify_hard_min_fclk_to_smu(pp_smu, new_clocks->fclk_khz); - send_request_to_lower = true; } @@ -281,7 +229,7 @@ static void dcn1_update_clocks(struct clk_mgr *clk_mgr, if (should_set_clock(safe_to_lower, new_clocks->dcfclk_deep_sleep_khz, clk_mgr->clks.dcfclk_deep_sleep_khz)) { clk_mgr->clks.dcfclk_deep_sleep_khz = new_clocks->dcfclk_deep_sleep_khz; - smu_req.min_deep_sleep_dcefclk_mhz = new_clocks->dcfclk_deep_sleep_khz / 1000; + smu_req.min_deep_sleep_dcefclk_mhz = (new_clocks->dcfclk_deep_sleep_khz + 999) / 1000; send_request_to_lower = true; } @@ -291,15 +239,18 @@ static void dcn1_update_clocks(struct clk_mgr *clk_mgr, */ if (send_request_to_increase) { /*use dcfclk to request voltage*/ - requested_dcf_clock_in_khz = dcn_find_dcfclk_suits_all(dc, new_clocks); - - notify_hard_min_dcfclk_to_smu(pp_smu, requested_dcf_clock_in_khz); - - if (pp_smu->set_display_requirement) - pp_smu->set_display_requirement(&pp_smu->pp_smu, &smu_req); - - notify_deep_sleep_dcfclk_to_smu(pp_smu, clk_mgr->clks.dcfclk_deep_sleep_khz); - dcn1_pplib_apply_display_requirements(dc, context); + if (pp_smu->set_hard_min_fclk_by_freq && + pp_smu->set_hard_min_dcfclk_by_freq && + pp_smu->set_min_deep_sleep_dcfclk) { + + pp_smu->set_hard_min_fclk_by_freq(&pp_smu->pp_smu, smu_req.hard_min_fclk_mhz); + pp_smu->set_hard_min_dcfclk_by_freq(&pp_smu->pp_smu, smu_req.hard_min_dcefclk_mhz); + pp_smu->set_min_deep_sleep_dcfclk(&pp_smu->pp_smu, smu_req.min_deep_sleep_dcefclk_mhz); + } else { + if (pp_smu->set_display_requirement) + pp_smu->set_display_requirement(&pp_smu->pp_smu, &smu_req); + dcn1_pplib_apply_display_requirements(dc, context); + } } /* dcn1 dppclk is tied to dispclk */ @@ -314,18 +265,20 @@ static void dcn1_update_clocks(struct clk_mgr *clk_mgr, if (!send_request_to_increase && send_request_to_lower) { /*use dcfclk to request voltage*/ - requested_dcf_clock_in_khz = dcn_find_dcfclk_suits_all(dc, new_clocks); - - notify_hard_min_dcfclk_to_smu(pp_smu, requested_dcf_clock_in_khz); - - if (pp_smu->set_display_requirement) - pp_smu->set_display_requirement(&pp_smu->pp_smu, &smu_req); - - notify_deep_sleep_dcfclk_to_smu(pp_smu, clk_mgr->clks.dcfclk_deep_sleep_khz); - dcn1_pplib_apply_display_requirements(dc, context); + if (pp_smu->set_hard_min_fclk_by_freq && + pp_smu->set_hard_min_dcfclk_by_freq && + pp_smu->set_min_deep_sleep_dcfclk) { + + pp_smu->set_hard_min_fclk_by_freq(&pp_smu->pp_smu, smu_req.hard_min_fclk_mhz); + pp_smu->set_hard_min_dcfclk_by_freq(&pp_smu->pp_smu, smu_req.hard_min_dcefclk_mhz); + pp_smu->set_min_deep_sleep_dcfclk(&pp_smu->pp_smu, smu_req.min_deep_sleep_dcefclk_mhz); + } else { + if (pp_smu->set_display_requirement) + pp_smu->set_display_requirement(&pp_smu->pp_smu, &smu_req); + dcn1_pplib_apply_display_requirements(dc, context); + } } - *smu_req_cur = smu_req; } static const struct clk_mgr_funcs dcn1_funcs = { diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c index dcb3c5530236..cd1ebe57ed59 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c @@ -463,7 +463,7 @@ void dpp1_set_cursor_position( if (src_y_offset >= (int)param->viewport.height) cur_en = 0; /* not visible beyond bottom edge*/ - if (src_y_offset < 0) + if (src_y_offset + (int)height <= 0) cur_en = 0; /* not visible beyond top edge*/ REG_UPDATE(CURSOR0_CONTROL, diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c index 116977eb24e2..41f0f4c912e7 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c @@ -51,10 +51,6 @@ #define NUM_ELEMENTS(a) (sizeof(a) / sizeof((a)[0])) -struct dcn10_input_csc_matrix { - enum dc_color_space color_space; - uint16_t regval[12]; -}; enum dcn10_coef_filter_type_sel { SCL_COEF_LUMA_VERT_FILTER = 0, @@ -99,7 +95,7 @@ enum gamut_remap_select { GAMUT_REMAP_COMB_COEFF }; -static const struct dcn10_input_csc_matrix dcn10_input_csc_matrix[] = { +static const struct dpp_input_csc_matrix dpp_input_csc_matrix[] = { {COLOR_SPACE_SRGB, {0x2000, 0, 0, 0, 0, 0x2000, 0, 0, 0, 0, 0x2000, 0} }, {COLOR_SPACE_SRGB_LIMITED, @@ -454,7 +450,7 @@ void dpp1_program_input_csc( { struct dcn10_dpp *dpp = TO_DCN10_DPP(dpp_base); int i; - int arr_size = sizeof(dcn10_input_csc_matrix)/sizeof(struct dcn10_input_csc_matrix); + int arr_size = sizeof(dpp_input_csc_matrix)/sizeof(struct dpp_input_csc_matrix); const uint16_t *regval = NULL; uint32_t cur_select = 0; enum dcn10_input_csc_select select; @@ -467,8 +463,8 @@ void dpp1_program_input_csc( if (tbl_entry == NULL) { for (i = 0; i < arr_size; i++) - if (dcn10_input_csc_matrix[i].color_space == color_space) { - regval = dcn10_input_csc_matrix[i].regval; + if (dpp_input_csc_matrix[i].color_space == color_space) { + regval = dpp_input_csc_matrix[i].regval; break; } diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c index 4a863a5dab41..c7642e748297 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c @@ -597,11 +597,13 @@ static void dpp1_dscl_set_manual_ratio_init( SCL_V_INIT_FRAC, init_frac, SCL_V_INIT_INT, init_int); - init_frac = dc_fixpt_u0d19(data->inits.v_bot) << 5; - init_int = dc_fixpt_floor(data->inits.v_bot); - REG_SET_2(SCL_VERT_FILTER_INIT_BOT, 0, - SCL_V_INIT_FRAC_BOT, init_frac, - SCL_V_INIT_INT_BOT, init_int); + if (REG(SCL_VERT_FILTER_INIT_BOT)) { + init_frac = dc_fixpt_u0d19(data->inits.v_bot) << 5; + init_int = dc_fixpt_floor(data->inits.v_bot); + REG_SET_2(SCL_VERT_FILTER_INIT_BOT, 0, + SCL_V_INIT_FRAC_BOT, init_frac, + SCL_V_INIT_INT_BOT, init_int); + } init_frac = dc_fixpt_u0d19(data->inits.v_c) << 5; init_int = dc_fixpt_floor(data->inits.v_c); @@ -609,11 +611,13 @@ static void dpp1_dscl_set_manual_ratio_init( SCL_V_INIT_FRAC_C, init_frac, SCL_V_INIT_INT_C, init_int); - init_frac = dc_fixpt_u0d19(data->inits.v_c_bot) << 5; - init_int = dc_fixpt_floor(data->inits.v_c_bot); - REG_SET_2(SCL_VERT_FILTER_INIT_BOT_C, 0, - SCL_V_INIT_FRAC_BOT_C, init_frac, - SCL_V_INIT_INT_BOT_C, init_int); + if (REG(SCL_VERT_FILTER_INIT_BOT_C)) { + init_frac = dc_fixpt_u0d19(data->inits.v_c_bot) << 5; + init_int = dc_fixpt_floor(data->inits.v_c_bot); + REG_SET_2(SCL_VERT_FILTER_INIT_BOT_C, 0, + SCL_V_INIT_FRAC_BOT_C, init_frac, + SCL_V_INIT_INT_BOT_C, init_int); + } } @@ -688,15 +692,17 @@ void dpp1_dscl_set_scaler_manual_scale( return; /* Black offsets */ - if (ycbcr) - REG_SET_2(SCL_BLACK_OFFSET, 0, - SCL_BLACK_OFFSET_RGB_Y, BLACK_OFFSET_RGB_Y, - SCL_BLACK_OFFSET_CBCR, BLACK_OFFSET_CBCR); - else + if (REG(SCL_BLACK_OFFSET)) { + if (ycbcr) + REG_SET_2(SCL_BLACK_OFFSET, 0, + SCL_BLACK_OFFSET_RGB_Y, BLACK_OFFSET_RGB_Y, + SCL_BLACK_OFFSET_CBCR, BLACK_OFFSET_CBCR); + else - REG_SET_2(SCL_BLACK_OFFSET, 0, - SCL_BLACK_OFFSET_RGB_Y, BLACK_OFFSET_RGB_Y, - SCL_BLACK_OFFSET_CBCR, BLACK_OFFSET_RGB_Y); + REG_SET_2(SCL_BLACK_OFFSET, 0, + SCL_BLACK_OFFSET_RGB_Y, BLACK_OFFSET_RGB_Y, + SCL_BLACK_OFFSET_CBCR, BLACK_OFFSET_RGB_Y); + } /* Manually calculate scale ratio and init values */ dpp1_dscl_set_manual_ratio_init(dpp, scl_data); diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c index c7d1e678ebf5..e161ad836812 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c @@ -29,19 +29,20 @@ #include "reg_helper.h" #define CTX \ - hubbub->ctx + hubbub1->base.ctx #define DC_LOGGER \ - hubbub->ctx->logger + hubbub1->base.ctx->logger #define REG(reg)\ - hubbub->regs->reg + hubbub1->regs->reg #undef FN #define FN(reg_name, field_name) \ - hubbub->shifts->field_name, hubbub->masks->field_name + hubbub1->shifts->field_name, hubbub1->masks->field_name void hubbub1_wm_read_state(struct hubbub *hubbub, struct dcn_hubbub_wm *wm) { + struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub); struct dcn_hubbub_wm_set *s; memset(wm, 0, sizeof(struct dcn_hubbub_wm)); @@ -87,14 +88,23 @@ void hubbub1_wm_read_state(struct hubbub *hubbub, s->dram_clk_chanage = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D); } -void hubbub1_disable_allow_self_refresh(struct hubbub *hubbub) +void hubbub1_allow_self_refresh_control(struct hubbub *hubbub, bool allow) { - REG_UPDATE(DCHUBBUB_ARB_DRAM_STATE_CNTL, - DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE, 0); + struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub); + + /* + * DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE = 1 means do not allow stutter + * DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE = 0 means allow stutter + */ + + REG_UPDATE_2(DCHUBBUB_ARB_DRAM_STATE_CNTL, + DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_VALUE, 0, + DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE, !allow); } bool hububu1_is_allow_self_refresh_enabled(struct hubbub *hubbub) { + struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub); uint32_t enable = 0; REG_GET(DCHUBBUB_ARB_DRAM_STATE_CNTL, @@ -107,6 +117,8 @@ bool hububu1_is_allow_self_refresh_enabled(struct hubbub *hubbub) bool hubbub1_verify_allow_pstate_change_high( struct hubbub *hubbub) { + struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub); + /* pstate latency is ~20us so if we wait over 40us and pstate allow * still not asserted, we are probably stuck and going to hang * @@ -193,7 +205,7 @@ bool hubbub1_verify_allow_pstate_change_high( * 31: SOC pstate change request */ - REG_WRITE(DCHUBBUB_TEST_DEBUG_INDEX, hubbub->debug_test_index_pstate); + REG_WRITE(DCHUBBUB_TEST_DEBUG_INDEX, hubbub1->debug_test_index_pstate); for (i = 0; i < pstate_wait_timeout_us; i++) { debug_data = REG_READ(DCHUBBUB_TEST_DEBUG_DATA); @@ -244,6 +256,8 @@ static uint32_t convert_and_clamp( void hubbub1_wm_change_req_wa(struct hubbub *hubbub) { + struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub); + REG_UPDATE_SEQ(DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL, DCHUBBUB_ARB_WATERMARK_CHANGE_REQUEST, 0, 1); } @@ -254,7 +268,7 @@ void hubbub1_program_watermarks( unsigned int refclk_mhz, bool safe_to_lower) { - uint32_t force_en = hubbub->ctx->dc->debug.disable_stutter ? 1 : 0; + struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub); /* * Need to clamp to max of the register values (i.e. no wrap) * for dcn1, all wm registers are 21-bit wide @@ -264,8 +278,8 @@ void hubbub1_program_watermarks( /* Repeat for water mark set A, B, C and D. */ /* clock state A */ - if (safe_to_lower || watermarks->a.urgent_ns > hubbub->watermarks.a.urgent_ns) { - hubbub->watermarks.a.urgent_ns = watermarks->a.urgent_ns; + if (safe_to_lower || watermarks->a.urgent_ns > hubbub1->watermarks.a.urgent_ns) { + hubbub1->watermarks.a.urgent_ns = watermarks->a.urgent_ns; prog_wm_value = convert_and_clamp(watermarks->a.urgent_ns, refclk_mhz, 0x1fffff); REG_WRITE(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_A, prog_wm_value); @@ -275,20 +289,22 @@ void hubbub1_program_watermarks( watermarks->a.urgent_ns, prog_wm_value); } - if (safe_to_lower || watermarks->a.pte_meta_urgent_ns > hubbub->watermarks.a.pte_meta_urgent_ns) { - hubbub->watermarks.a.pte_meta_urgent_ns = watermarks->a.pte_meta_urgent_ns; - prog_wm_value = convert_and_clamp(watermarks->a.pte_meta_urgent_ns, - refclk_mhz, 0x1fffff); - REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_A, prog_wm_value); - DC_LOG_BANDWIDTH_CALCS("PTE_META_URGENCY_WATERMARK_A calculated =%d\n" - "HW register value = 0x%x\n", - watermarks->a.pte_meta_urgent_ns, prog_wm_value); + if (REG(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_A)) { + if (safe_to_lower || watermarks->a.pte_meta_urgent_ns > hubbub1->watermarks.a.pte_meta_urgent_ns) { + hubbub1->watermarks.a.pte_meta_urgent_ns = watermarks->a.pte_meta_urgent_ns; + prog_wm_value = convert_and_clamp(watermarks->a.pte_meta_urgent_ns, + refclk_mhz, 0x1fffff); + REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_A, prog_wm_value); + DC_LOG_BANDWIDTH_CALCS("PTE_META_URGENCY_WATERMARK_A calculated =%d\n" + "HW register value = 0x%x\n", + watermarks->a.pte_meta_urgent_ns, prog_wm_value); + } } if (REG(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_A)) { if (safe_to_lower || watermarks->a.cstate_pstate.cstate_enter_plus_exit_ns - > hubbub->watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns) { - hubbub->watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = + > hubbub1->watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns) { + hubbub1->watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = watermarks->a.cstate_pstate.cstate_enter_plus_exit_ns; prog_wm_value = convert_and_clamp( watermarks->a.cstate_pstate.cstate_enter_plus_exit_ns, @@ -300,8 +316,8 @@ void hubbub1_program_watermarks( } if (safe_to_lower || watermarks->a.cstate_pstate.cstate_exit_ns - > hubbub->watermarks.a.cstate_pstate.cstate_exit_ns) { - hubbub->watermarks.a.cstate_pstate.cstate_exit_ns = + > hubbub1->watermarks.a.cstate_pstate.cstate_exit_ns) { + hubbub1->watermarks.a.cstate_pstate.cstate_exit_ns = watermarks->a.cstate_pstate.cstate_exit_ns; prog_wm_value = convert_and_clamp( watermarks->a.cstate_pstate.cstate_exit_ns, @@ -314,8 +330,8 @@ void hubbub1_program_watermarks( } if (safe_to_lower || watermarks->a.cstate_pstate.pstate_change_ns - > hubbub->watermarks.a.cstate_pstate.pstate_change_ns) { - hubbub->watermarks.a.cstate_pstate.pstate_change_ns = + > hubbub1->watermarks.a.cstate_pstate.pstate_change_ns) { + hubbub1->watermarks.a.cstate_pstate.pstate_change_ns = watermarks->a.cstate_pstate.pstate_change_ns; prog_wm_value = convert_and_clamp( watermarks->a.cstate_pstate.pstate_change_ns, @@ -327,8 +343,8 @@ void hubbub1_program_watermarks( } /* clock state B */ - if (safe_to_lower || watermarks->b.urgent_ns > hubbub->watermarks.b.urgent_ns) { - hubbub->watermarks.b.urgent_ns = watermarks->b.urgent_ns; + if (safe_to_lower || watermarks->b.urgent_ns > hubbub1->watermarks.b.urgent_ns) { + hubbub1->watermarks.b.urgent_ns = watermarks->b.urgent_ns; prog_wm_value = convert_and_clamp(watermarks->b.urgent_ns, refclk_mhz, 0x1fffff); REG_WRITE(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_B, prog_wm_value); @@ -338,20 +354,22 @@ void hubbub1_program_watermarks( watermarks->b.urgent_ns, prog_wm_value); } - if (safe_to_lower || watermarks->b.pte_meta_urgent_ns > hubbub->watermarks.b.pte_meta_urgent_ns) { - hubbub->watermarks.b.pte_meta_urgent_ns = watermarks->b.pte_meta_urgent_ns; - prog_wm_value = convert_and_clamp(watermarks->b.pte_meta_urgent_ns, - refclk_mhz, 0x1fffff); - REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_B, prog_wm_value); - DC_LOG_BANDWIDTH_CALCS("PTE_META_URGENCY_WATERMARK_B calculated =%d\n" - "HW register value = 0x%x\n", - watermarks->b.pte_meta_urgent_ns, prog_wm_value); + if (REG(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_B)) { + if (safe_to_lower || watermarks->b.pte_meta_urgent_ns > hubbub1->watermarks.b.pte_meta_urgent_ns) { + hubbub1->watermarks.b.pte_meta_urgent_ns = watermarks->b.pte_meta_urgent_ns; + prog_wm_value = convert_and_clamp(watermarks->b.pte_meta_urgent_ns, + refclk_mhz, 0x1fffff); + REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_B, prog_wm_value); + DC_LOG_BANDWIDTH_CALCS("PTE_META_URGENCY_WATERMARK_B calculated =%d\n" + "HW register value = 0x%x\n", + watermarks->b.pte_meta_urgent_ns, prog_wm_value); + } } if (REG(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_B)) { if (safe_to_lower || watermarks->b.cstate_pstate.cstate_enter_plus_exit_ns - > hubbub->watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns) { - hubbub->watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns = + > hubbub1->watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns) { + hubbub1->watermarks.b.cstate_pstate.cstate_enter_plus_exit_ns = watermarks->b.cstate_pstate.cstate_enter_plus_exit_ns; prog_wm_value = convert_and_clamp( watermarks->b.cstate_pstate.cstate_enter_plus_exit_ns, @@ -363,8 +381,8 @@ void hubbub1_program_watermarks( } if (safe_to_lower || watermarks->b.cstate_pstate.cstate_exit_ns - > hubbub->watermarks.b.cstate_pstate.cstate_exit_ns) { - hubbub->watermarks.b.cstate_pstate.cstate_exit_ns = + > hubbub1->watermarks.b.cstate_pstate.cstate_exit_ns) { + hubbub1->watermarks.b.cstate_pstate.cstate_exit_ns = watermarks->b.cstate_pstate.cstate_exit_ns; prog_wm_value = convert_and_clamp( watermarks->b.cstate_pstate.cstate_exit_ns, @@ -377,8 +395,8 @@ void hubbub1_program_watermarks( } if (safe_to_lower || watermarks->b.cstate_pstate.pstate_change_ns - > hubbub->watermarks.b.cstate_pstate.pstate_change_ns) { - hubbub->watermarks.b.cstate_pstate.pstate_change_ns = + > hubbub1->watermarks.b.cstate_pstate.pstate_change_ns) { + hubbub1->watermarks.b.cstate_pstate.pstate_change_ns = watermarks->b.cstate_pstate.pstate_change_ns; prog_wm_value = convert_and_clamp( watermarks->b.cstate_pstate.pstate_change_ns, @@ -390,8 +408,8 @@ void hubbub1_program_watermarks( } /* clock state C */ - if (safe_to_lower || watermarks->c.urgent_ns > hubbub->watermarks.c.urgent_ns) { - hubbub->watermarks.c.urgent_ns = watermarks->c.urgent_ns; + if (safe_to_lower || watermarks->c.urgent_ns > hubbub1->watermarks.c.urgent_ns) { + hubbub1->watermarks.c.urgent_ns = watermarks->c.urgent_ns; prog_wm_value = convert_and_clamp(watermarks->c.urgent_ns, refclk_mhz, 0x1fffff); REG_WRITE(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_C, prog_wm_value); @@ -401,20 +419,22 @@ void hubbub1_program_watermarks( watermarks->c.urgent_ns, prog_wm_value); } - if (safe_to_lower || watermarks->c.pte_meta_urgent_ns > hubbub->watermarks.c.pte_meta_urgent_ns) { - hubbub->watermarks.c.pte_meta_urgent_ns = watermarks->c.pte_meta_urgent_ns; - prog_wm_value = convert_and_clamp(watermarks->c.pte_meta_urgent_ns, - refclk_mhz, 0x1fffff); - REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_C, prog_wm_value); - DC_LOG_BANDWIDTH_CALCS("PTE_META_URGENCY_WATERMARK_C calculated =%d\n" - "HW register value = 0x%x\n", - watermarks->c.pte_meta_urgent_ns, prog_wm_value); + if (REG(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_C)) { + if (safe_to_lower || watermarks->c.pte_meta_urgent_ns > hubbub1->watermarks.c.pte_meta_urgent_ns) { + hubbub1->watermarks.c.pte_meta_urgent_ns = watermarks->c.pte_meta_urgent_ns; + prog_wm_value = convert_and_clamp(watermarks->c.pte_meta_urgent_ns, + refclk_mhz, 0x1fffff); + REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_C, prog_wm_value); + DC_LOG_BANDWIDTH_CALCS("PTE_META_URGENCY_WATERMARK_C calculated =%d\n" + "HW register value = 0x%x\n", + watermarks->c.pte_meta_urgent_ns, prog_wm_value); + } } if (REG(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_C)) { if (safe_to_lower || watermarks->c.cstate_pstate.cstate_enter_plus_exit_ns - > hubbub->watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns) { - hubbub->watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns = + > hubbub1->watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns) { + hubbub1->watermarks.c.cstate_pstate.cstate_enter_plus_exit_ns = watermarks->c.cstate_pstate.cstate_enter_plus_exit_ns; prog_wm_value = convert_and_clamp( watermarks->c.cstate_pstate.cstate_enter_plus_exit_ns, @@ -426,8 +446,8 @@ void hubbub1_program_watermarks( } if (safe_to_lower || watermarks->c.cstate_pstate.cstate_exit_ns - > hubbub->watermarks.c.cstate_pstate.cstate_exit_ns) { - hubbub->watermarks.c.cstate_pstate.cstate_exit_ns = + > hubbub1->watermarks.c.cstate_pstate.cstate_exit_ns) { + hubbub1->watermarks.c.cstate_pstate.cstate_exit_ns = watermarks->c.cstate_pstate.cstate_exit_ns; prog_wm_value = convert_and_clamp( watermarks->c.cstate_pstate.cstate_exit_ns, @@ -440,8 +460,8 @@ void hubbub1_program_watermarks( } if (safe_to_lower || watermarks->c.cstate_pstate.pstate_change_ns - > hubbub->watermarks.c.cstate_pstate.pstate_change_ns) { - hubbub->watermarks.c.cstate_pstate.pstate_change_ns = + > hubbub1->watermarks.c.cstate_pstate.pstate_change_ns) { + hubbub1->watermarks.c.cstate_pstate.pstate_change_ns = watermarks->c.cstate_pstate.pstate_change_ns; prog_wm_value = convert_and_clamp( watermarks->c.cstate_pstate.pstate_change_ns, @@ -453,8 +473,8 @@ void hubbub1_program_watermarks( } /* clock state D */ - if (safe_to_lower || watermarks->d.urgent_ns > hubbub->watermarks.d.urgent_ns) { - hubbub->watermarks.d.urgent_ns = watermarks->d.urgent_ns; + if (safe_to_lower || watermarks->d.urgent_ns > hubbub1->watermarks.d.urgent_ns) { + hubbub1->watermarks.d.urgent_ns = watermarks->d.urgent_ns; prog_wm_value = convert_and_clamp(watermarks->d.urgent_ns, refclk_mhz, 0x1fffff); REG_WRITE(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_D, prog_wm_value); @@ -464,20 +484,22 @@ void hubbub1_program_watermarks( watermarks->d.urgent_ns, prog_wm_value); } - if (safe_to_lower || watermarks->d.pte_meta_urgent_ns > hubbub->watermarks.d.pte_meta_urgent_ns) { - hubbub->watermarks.d.pte_meta_urgent_ns = watermarks->d.pte_meta_urgent_ns; - prog_wm_value = convert_and_clamp(watermarks->d.pte_meta_urgent_ns, - refclk_mhz, 0x1fffff); - REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_D, prog_wm_value); - DC_LOG_BANDWIDTH_CALCS("PTE_META_URGENCY_WATERMARK_D calculated =%d\n" - "HW register value = 0x%x\n", - watermarks->d.pte_meta_urgent_ns, prog_wm_value); + if (REG(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_D)) { + if (safe_to_lower || watermarks->d.pte_meta_urgent_ns > hubbub1->watermarks.d.pte_meta_urgent_ns) { + hubbub1->watermarks.d.pte_meta_urgent_ns = watermarks->d.pte_meta_urgent_ns; + prog_wm_value = convert_and_clamp(watermarks->d.pte_meta_urgent_ns, + refclk_mhz, 0x1fffff); + REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_D, prog_wm_value); + DC_LOG_BANDWIDTH_CALCS("PTE_META_URGENCY_WATERMARK_D calculated =%d\n" + "HW register value = 0x%x\n", + watermarks->d.pte_meta_urgent_ns, prog_wm_value); + } } if (REG(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_D)) { if (safe_to_lower || watermarks->d.cstate_pstate.cstate_enter_plus_exit_ns - > hubbub->watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns) { - hubbub->watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns = + > hubbub1->watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns) { + hubbub1->watermarks.d.cstate_pstate.cstate_enter_plus_exit_ns = watermarks->d.cstate_pstate.cstate_enter_plus_exit_ns; prog_wm_value = convert_and_clamp( watermarks->d.cstate_pstate.cstate_enter_plus_exit_ns, @@ -489,8 +511,8 @@ void hubbub1_program_watermarks( } if (safe_to_lower || watermarks->d.cstate_pstate.cstate_exit_ns - > hubbub->watermarks.d.cstate_pstate.cstate_exit_ns) { - hubbub->watermarks.d.cstate_pstate.cstate_exit_ns = + > hubbub1->watermarks.d.cstate_pstate.cstate_exit_ns) { + hubbub1->watermarks.d.cstate_pstate.cstate_exit_ns = watermarks->d.cstate_pstate.cstate_exit_ns; prog_wm_value = convert_and_clamp( watermarks->d.cstate_pstate.cstate_exit_ns, @@ -503,8 +525,8 @@ void hubbub1_program_watermarks( } if (safe_to_lower || watermarks->d.cstate_pstate.pstate_change_ns - > hubbub->watermarks.d.cstate_pstate.pstate_change_ns) { - hubbub->watermarks.d.cstate_pstate.pstate_change_ns = + > hubbub1->watermarks.d.cstate_pstate.pstate_change_ns) { + hubbub1->watermarks.d.cstate_pstate.pstate_change_ns = watermarks->d.cstate_pstate.pstate_change_ns; prog_wm_value = convert_and_clamp( watermarks->d.cstate_pstate.pstate_change_ns, @@ -520,9 +542,7 @@ void hubbub1_program_watermarks( REG_UPDATE(DCHUBBUB_ARB_DF_REQ_OUTSTAND, DCHUBBUB_ARB_MIN_REQ_OUTSTAND, 68); - REG_UPDATE_2(DCHUBBUB_ARB_DRAM_STATE_CNTL, - DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_VALUE, 0, - DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE, force_en); + hubbub1_allow_self_refresh_control(hubbub, !hubbub->ctx->dc->debug.disable_stutter); #if 0 REG_UPDATE_2(DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL, @@ -535,6 +555,8 @@ void hubbub1_update_dchub( struct hubbub *hubbub, struct dchub_init_data *dh_data) { + struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub); + if (REG(DCHUBBUB_SDPIF_FB_TOP) == 0) { ASSERT(false); /*should not come here*/ @@ -594,6 +616,8 @@ void hubbub1_update_dchub( void hubbub1_toggle_watermark_change_req(struct hubbub *hubbub) { + struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub); + uint32_t watermark_change_req; REG_GET(DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL, @@ -610,6 +634,8 @@ void hubbub1_toggle_watermark_change_req(struct hubbub *hubbub) void hubbub1_soft_reset(struct hubbub *hubbub, bool reset) { + struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub); + uint32_t reset_en = reset ? 1 : 0; REG_UPDATE(DCHUBBUB_SOFT_RESET, @@ -752,7 +778,9 @@ static bool hubbub1_get_dcc_compression_cap(struct hubbub *hubbub, const struct dc_dcc_surface_param *input, struct dc_surface_dcc_cap *output) { - struct dc *dc = hubbub->ctx->dc; + struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub); + struct dc *dc = hubbub1->base.ctx->dc; + /* implement section 1.6.2.1 of DCN1_Programming_Guide.docx */ enum dcc_control dcc_control; unsigned int bpe; @@ -764,10 +792,10 @@ static bool hubbub1_get_dcc_compression_cap(struct hubbub *hubbub, if (dc->debug.disable_dcc == DCC_DISABLE) return false; - if (!hubbub->funcs->dcc_support_pixel_format(input->format, &bpe)) + if (!hubbub1->base.funcs->dcc_support_pixel_format(input->format, &bpe)) return false; - if (!hubbub->funcs->dcc_support_swizzle(input->swizzle_mode, bpe, + if (!hubbub1->base.funcs->dcc_support_swizzle(input->swizzle_mode, bpe, &segment_order_horz, &segment_order_vert)) return false; @@ -837,6 +865,7 @@ static const struct hubbub_funcs hubbub1_funcs = { .dcc_support_swizzle = hubbub1_dcc_support_swizzle, .dcc_support_pixel_format = hubbub1_dcc_support_pixel_format, .get_dcc_compression_cap = hubbub1_get_dcc_compression_cap, + .wm_read_state = hubbub1_wm_read_state, }; void hubbub1_construct(struct hubbub *hubbub, @@ -845,18 +874,20 @@ void hubbub1_construct(struct hubbub *hubbub, const struct dcn_hubbub_shift *hubbub_shift, const struct dcn_hubbub_mask *hubbub_mask) { - hubbub->ctx = ctx; + struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub); + + hubbub1->base.ctx = ctx; - hubbub->funcs = &hubbub1_funcs; + hubbub1->base.funcs = &hubbub1_funcs; - hubbub->regs = hubbub_regs; - hubbub->shifts = hubbub_shift; - hubbub->masks = hubbub_mask; + hubbub1->regs = hubbub_regs; + hubbub1->shifts = hubbub_shift; + hubbub1->masks = hubbub_mask; - hubbub->debug_test_index_pstate = 0x7; + hubbub1->debug_test_index_pstate = 0x7; #if defined(CONFIG_DRM_AMD_DC_DCN1_01) if (ctx->dce_version == DCN_VERSION_1_01) - hubbub->debug_test_index_pstate = 0xB; + hubbub1->debug_test_index_pstate = 0xB; #endif } diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h index d0f03d152913..9cd4a5194154 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h @@ -29,6 +29,9 @@ #include "core_types.h" #include "dchubbub.h" +#define TO_DCN10_HUBBUB(hubbub)\ + container_of(hubbub, struct dcn10_hubbub, base) + #define HUBHUB_REG_LIST_DCN()\ SR(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_A),\ SR(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_A),\ @@ -107,6 +110,12 @@ struct dcn_hubbub_registers { uint32_t DCHUBBUB_SDPIF_AGP_TOP; uint32_t DCHUBBUB_CRC_CTRL; uint32_t DCHUBBUB_SOFT_RESET; + uint32_t DCN_VM_FB_LOCATION_BASE; + uint32_t DCN_VM_FB_LOCATION_TOP; + uint32_t DCN_VM_FB_OFFSET; + uint32_t DCN_VM_AGP_BOT; + uint32_t DCN_VM_AGP_TOP; + uint32_t DCN_VM_AGP_BASE; }; /* set field name */ @@ -152,7 +161,13 @@ struct dcn_hubbub_registers { type SDPIF_FB_OFFSET;\ type SDPIF_AGP_BASE;\ type SDPIF_AGP_BOT;\ - type SDPIF_AGP_TOP + type SDPIF_AGP_TOP;\ + type FB_BASE;\ + type FB_TOP;\ + type FB_OFFSET;\ + type AGP_BOT;\ + type AGP_TOP;\ + type AGP_BASE struct dcn_hubbub_shift { @@ -165,22 +180,8 @@ struct dcn_hubbub_mask { struct dc; -struct dcn_hubbub_wm_set { - uint32_t wm_set; - uint32_t data_urgent; - uint32_t pte_meta_urgent; - uint32_t sr_enter; - uint32_t sr_exit; - uint32_t dram_clk_chanage; -}; - -struct dcn_hubbub_wm { - struct dcn_hubbub_wm_set sets[4]; -}; - -struct hubbub { - const struct hubbub_funcs *funcs; - struct dc_context *ctx; +struct dcn10_hubbub { + struct hubbub base; const struct dcn_hubbub_registers *regs; const struct dcn_hubbub_shift *shifts; const struct dcn_hubbub_mask *masks; @@ -203,7 +204,7 @@ void hubbub1_program_watermarks( unsigned int refclk_mhz, bool safe_to_lower); -void hubbub1_disable_allow_self_refresh(struct hubbub *hubbub); +void hubbub1_allow_self_refresh_control(struct hubbub *hubbub, bool allow); bool hububu1_is_allow_self_refresh_enabled(struct hubbub *hubub); diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c index 345af015d061..683829466a44 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c @@ -115,7 +115,7 @@ static void hubp1_set_hubp_blank_en(struct hubp *hubp, bool blank) REG_UPDATE(DCHUBP_CNTL, HUBP_BLANK_EN, blank_en); } -static void hubp1_vready_workaround(struct hubp *hubp, +void hubp1_vready_workaround(struct hubp *hubp, struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest) { uint32_t value = 0; @@ -317,7 +317,8 @@ void hubp1_program_pixel_format( bool hubp1_program_surface_flip_and_addr( struct hubp *hubp, const struct dc_plane_address *address, - bool flip_immediate) + bool flip_immediate, + uint8_t vmid) { struct dcn10_hubp *hubp1 = TO_DCN10_HUBP(hubp); @@ -1140,7 +1141,7 @@ void hubp1_cursor_set_position( if (src_y_offset >= (int)param->viewport.height) cur_en = 0; /* not visible beyond bottom edge*/ - if (src_y_offset < 0) //+ (int)hubp->curs_attr.height + if (src_y_offset + (int)hubp->curs_attr.height <= 0) cur_en = 0; /* not visible beyond top edge*/ if (cur_en && REG_READ(CURSOR_SURFACE_ADDRESS) == 0) @@ -1149,9 +1150,28 @@ void hubp1_cursor_set_position( REG_UPDATE(CURSOR_CONTROL, CURSOR_ENABLE, cur_en); - REG_SET_2(CURSOR_POSITION, 0, - CURSOR_X_POSITION, pos->x, + //account for cases where we see negative offset relative to overlay plane + if (src_x_offset < 0 && src_y_offset < 0) { + REG_SET_2(CURSOR_POSITION, 0, + CURSOR_X_POSITION, 0, + CURSOR_Y_POSITION, 0); + x_hotspot -= src_x_offset; + y_hotspot -= src_y_offset; + } else if (src_x_offset < 0) { + REG_SET_2(CURSOR_POSITION, 0, + CURSOR_X_POSITION, 0, CURSOR_Y_POSITION, pos->y); + x_hotspot -= src_x_offset; + } else if (src_y_offset < 0) { + REG_SET_2(CURSOR_POSITION, 0, + CURSOR_X_POSITION, pos->x, + CURSOR_Y_POSITION, 0); + y_hotspot -= src_y_offset; + } else { + REG_SET_2(CURSOR_POSITION, 0, + CURSOR_X_POSITION, pos->x, + CURSOR_Y_POSITION, pos->y); + } REG_SET_2(CURSOR_HOT_SPOT, 0, CURSOR_HOT_SPOT_X, x_hotspot, diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h index 62d4232e7796..a6d6dfe00617 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h @@ -707,11 +707,6 @@ void hubp1_dcc_control(struct hubp *hubp, bool enable, bool independent_64b_blks); -bool hubp1_program_surface_flip_and_addr( - struct hubp *hubp, - const struct dc_plane_address *address, - bool flip_immediate); - bool hubp1_is_flip_pending(struct hubp *hubp); void hubp1_cursor_set_attributes( @@ -745,5 +740,7 @@ void hubp1_clear_underflow(struct hubp *hubp); enum cursor_pitch hubp1_get_cursor_pitch(unsigned int pitch); +void hubp1_vready_workaround(struct hubp *hubp, + struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest); #endif diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c index 91e015e14355..117d9d8227f7 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c @@ -40,7 +40,6 @@ #include "ipp.h" #include "mpc.h" #include "reg_helper.h" -#include "custom_float.h" #include "dcn10_hubp.h" #include "dcn10_hubbub.h" #include "dcn10_cm_common.h" @@ -92,10 +91,11 @@ static void log_mpc_crc(struct dc *dc, void dcn10_log_hubbub_state(struct dc *dc, struct dc_log_buffer_ctx *log_ctx) { struct dc_context *dc_ctx = dc->ctx; - struct dcn_hubbub_wm wm = {0}; + struct dcn_hubbub_wm wm; int i; - hubbub1_wm_read_state(dc->res_pool->hubbub, &wm); + memset(&wm, 0, sizeof(struct dcn_hubbub_wm)); + dc->res_pool->hubbub->funcs->wm_read_state(dc->res_pool->hubbub, &wm); DTN_INFO("HUBBUB WM: data_urgent pte_meta_urgent" " sr_enter sr_exit dram_clk_change\n"); @@ -636,8 +636,6 @@ static enum dc_status dcn10_enable_stream_timing( struct dc_stream_state *stream = pipe_ctx->stream; enum dc_color_space color_space; struct tg_color black_color = {0}; - struct drr_params params = {0}; - unsigned int event_triggers = 0; /* by upper caller loop, pipe0 is parent pipe and be called first. * back end is set up by for pipe0. Other children pipe share back end @@ -705,19 +703,6 @@ static enum dc_status dcn10_enable_stream_timing( return DC_ERROR_UNEXPECTED; } - params.vertical_total_min = stream->adjust.v_total_min; - params.vertical_total_max = stream->adjust.v_total_max; - if (pipe_ctx->stream_res.tg->funcs->set_drr) - pipe_ctx->stream_res.tg->funcs->set_drr( - pipe_ctx->stream_res.tg, ¶ms); - - // DRR should set trigger event to monitor surface update event - if (stream->adjust.v_total_min != 0 && stream->adjust.v_total_max != 0) - event_triggers = 0x80; - if (pipe_ctx->stream_res.tg->funcs->set_static_screen_control) - pipe_ctx->stream_res.tg->funcs->set_static_screen_control( - pipe_ctx->stream_res.tg, event_triggers); - /* TODO program crtc source select for non-virtual signal*/ /* TODO program FMT */ /* TODO setup link_enc */ @@ -971,92 +956,34 @@ static void dcn10_disable_plane(struct dc *dc, struct pipe_ctx *pipe_ctx) pipe_ctx->pipe_idx); } -static void dcn10_init_hw(struct dc *dc) +static void dcn10_init_pipes(struct dc *dc, struct dc_state *context) { int i; - struct abm *abm = dc->res_pool->abm; - struct dmcu *dmcu = dc->res_pool->dmcu; - struct dce_hwseq *hws = dc->hwseq; - struct dc_bios *dcb = dc->ctx->dc_bios; - struct dc_state *context = dc->current_state; - - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { - REG_WRITE(REFCLK_CNTL, 0); - REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1); - REG_WRITE(DIO_MEM_PWR_CTRL, 0); - - if (!dc->debug.disable_clock_gate) { - /* enable all DCN clock gating */ - REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0); - - REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0); - - REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0); - } - - enable_power_gating_plane(dc->hwseq, true); - } else { - - if (!dcb->funcs->is_accelerated_mode(dcb)) { - bool allow_self_fresh_force_enable = - hububu1_is_allow_self_refresh_enabled(dc->res_pool->hubbub); - - bios_golden_init(dc); - - /* WA for making DF sleep when idle after resume from S0i3. - * DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE is set to 1 by - * command table, if DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE = 0 - * before calling command table and it changed to 1 after, - * it should be set back to 0. - */ - if (allow_self_fresh_force_enable == false && - hububu1_is_allow_self_refresh_enabled(dc->res_pool->hubbub)) - hubbub1_disable_allow_self_refresh(dc->res_pool->hubbub); - - disable_vga(dc->hwseq); - } - - for (i = 0; i < dc->link_count; i++) { - /* Power up AND update implementation according to the - * required signal (which may be different from the - * default signal on connector). - */ - struct dc_link *link = dc->links[i]; - - if (link->link_enc->connector.id == CONNECTOR_ID_EDP) - dc->hwss.edp_power_control(link, true); - - link->link_enc->funcs->hw_init(link->link_enc); - } - } for (i = 0; i < dc->res_pool->pipe_count; i++) { struct timing_generator *tg = dc->res_pool->timing_generators[i]; if (tg->funcs->is_tg_enabled(tg)) tg->funcs->lock(tg); - } - - /* Blank controller using driver code instead of - * command table. - */ - for (i = 0; i < dc->res_pool->pipe_count; i++) { - struct timing_generator *tg = dc->res_pool->timing_generators[i]; + /* Blank controller using driver code instead of + * command table. + */ if (tg->funcs->is_tg_enabled(tg)) { tg->funcs->set_blank(tg, true); hwss_wait_for_blank_complete(tg); } } - /* Reset all MPCC muxes */ dc->res_pool->mpc->funcs->mpc_init(dc->res_pool->mpc); - for (i = 0; i < dc->res_pool->timing_generator_count; i++) { + for (i = 0; i < dc->res_pool->pipe_count; i++) { struct timing_generator *tg = dc->res_pool->timing_generators[i]; - struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; struct hubp *hubp = dc->res_pool->hubps[i]; struct dpp *dpp = dc->res_pool->dpps[i]; + struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; + + dpp->funcs->dpp_reset(dpp); pipe_ctx->stream_res.tg = tg; pipe_ctx->pipe_idx = i; @@ -1074,18 +1001,9 @@ static void dcn10_init_hw(struct dc *dc) pipe_ctx->stream_res.opp = dc->res_pool->opps[i]; hwss1_plane_atomic_disconnect(dc, pipe_ctx); - } - - for (i = 0; i < dc->res_pool->pipe_count; i++) { - struct timing_generator *tg = dc->res_pool->timing_generators[i]; if (tg->funcs->is_tg_enabled(tg)) tg->funcs->unlock(tg); - } - - for (i = 0; i < dc->res_pool->pipe_count; i++) { - struct timing_generator *tg = dc->res_pool->timing_generators[i]; - struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; dcn10_disable_plane(dc, pipe_ctx); @@ -1094,10 +1012,73 @@ static void dcn10_init_hw(struct dc *dc) tg->funcs->tg_init(tg); } +} - /* end of FPGA. Below if real ASIC */ - if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) +static void dcn10_init_hw(struct dc *dc) +{ + int i; + struct abm *abm = dc->res_pool->abm; + struct dmcu *dmcu = dc->res_pool->dmcu; + struct dce_hwseq *hws = dc->hwseq; + struct dc_bios *dcb = dc->ctx->dc_bios; + + if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) { + REG_WRITE(REFCLK_CNTL, 0); + REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1); + REG_WRITE(DIO_MEM_PWR_CTRL, 0); + + if (!dc->debug.disable_clock_gate) { + /* enable all DCN clock gating */ + REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0); + + REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0); + + REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0); + } + + enable_power_gating_plane(dc->hwseq, true); + + /* end of FPGA. Below if real ASIC */ return; + } + + if (!dcb->funcs->is_accelerated_mode(dcb)) { + bool allow_self_fresh_force_enable = + hububu1_is_allow_self_refresh_enabled( + dc->res_pool->hubbub); + + bios_golden_init(dc); + + /* WA for making DF sleep when idle after resume from S0i3. + * DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE is set to 1 by + * command table, if DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE = 0 + * before calling command table and it changed to 1 after, + * it should be set back to 0. + */ + if (allow_self_fresh_force_enable == false && + hububu1_is_allow_self_refresh_enabled(dc->res_pool->hubbub)) + hubbub1_allow_self_refresh_control(dc->res_pool->hubbub, true); + + disable_vga(dc->hwseq); + } + + for (i = 0; i < dc->link_count; i++) { + /* Power up AND update implementation according to the + * required signal (which may be different from the + * default signal on connector). + */ + struct dc_link *link = dc->links[i]; + + if (link->link_enc->connector.id == CONNECTOR_ID_EDP) + dc->hwss.edp_power_control(link, true); + + link->link_enc->funcs->hw_init(link->link_enc); + + /* Check for enabled DIG to identify enabled display */ + if (link->link_enc->funcs->is_dig_enabled && + link->link_enc->funcs->is_dig_enabled(link->link_enc)) + link->link_status.link_active = true; + } for (i = 0; i < dc->res_pool->audio_count; i++) { struct audio *audio = dc->res_pool->audios[i]; @@ -1128,6 +1109,9 @@ static void dcn10_init_hw(struct dc *dc) enable_power_gating_plane(dc->hwseq, true); memset(&dc->res_pool->clk_mgr->clks, 0, sizeof(dc->res_pool->clk_mgr->clks)); + + if (dc->hwss.init_pipes) + dc->hwss.init_pipes(dc, dc->current_state); } static void reset_hw_ctx_wrap( @@ -1202,7 +1186,8 @@ static void dcn10_update_plane_addr(const struct dc *dc, struct pipe_ctx *pipe_c pipe_ctx->plane_res.hubp->funcs->hubp_program_surface_flip_and_addr( pipe_ctx->plane_res.hubp, &plane_state->address, - plane_state->flip_immediate); + plane_state->flip_immediate, + 0); plane_state->status.requested_address = plane_state->address; @@ -2048,7 +2033,7 @@ void update_dchubp_dpp( dc->res_pool->dccg->funcs->update_dpp_dto( dc->res_pool->dccg, dpp->inst, - pipe_ctx->plane_res.bw.calc.dppclk_khz); + pipe_ctx->plane_res.bw.dppclk_khz); else dc->res_pool->clk_mgr->clks.dppclk_khz = should_divided_by_2 ? dc->res_pool->clk_mgr->clks.dispclk_khz / 2 : @@ -2125,7 +2110,8 @@ void update_dchubp_dpp( plane_state->update_flags.bits.swizzle_change || plane_state->update_flags.bits.dcc_change || plane_state->update_flags.bits.bpp_change || - plane_state->update_flags.bits.scaling_change) { + plane_state->update_flags.bits.scaling_change || + plane_state->update_flags.bits.plane_size_change) { hubp->funcs->hubp_program_surface_config( hubp, plane_state->format, @@ -2252,13 +2238,11 @@ static void program_all_pipe_in_tree( } - if (pipe_ctx->plane_state != NULL) { + if (pipe_ctx->plane_state != NULL) dcn10_program_pipe(dc, pipe_ctx, context); - } - if (pipe_ctx->bottom_pipe != NULL && pipe_ctx->bottom_pipe != pipe_ctx) { + if (pipe_ctx->bottom_pipe != NULL && pipe_ctx->bottom_pipe != pipe_ctx) program_all_pipe_in_tree(dc, pipe_ctx->bottom_pipe, context); - } } struct pipe_ctx *find_top_pipe_for_stream( @@ -2334,9 +2318,10 @@ static void dcn10_apply_ctx_for_surface( } } - if (!pipe_ctx->plane_state && - old_pipe_ctx->plane_state && - old_pipe_ctx->stream_res.tg == tg) { + if ((!pipe_ctx->plane_state || + pipe_ctx->stream_res.tg != old_pipe_ctx->stream_res.tg) && + old_pipe_ctx->plane_state && + old_pipe_ctx->stream_res.tg == tg) { dc->hwss.plane_atomic_disconnect(dc, old_pipe_ctx); removed_pipe[i] = true; @@ -2355,29 +2340,22 @@ static void dcn10_apply_ctx_for_surface( top_pipe_to_program->plane_state->update_flags.bits.full_update) for (i = 0; i < dc->res_pool->pipe_count; i++) { struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; - + tg = pipe_ctx->stream_res.tg; /* Skip inactive pipes and ones already updated */ if (!pipe_ctx->stream || pipe_ctx->stream == stream - || !pipe_ctx->plane_state) + || !pipe_ctx->plane_state + || !tg->funcs->is_tg_enabled(tg)) continue; - pipe_ctx->stream_res.tg->funcs->lock(pipe_ctx->stream_res.tg); + tg->funcs->lock(tg); pipe_ctx->plane_res.hubp->funcs->hubp_setup_interdependent( pipe_ctx->plane_res.hubp, &pipe_ctx->dlg_regs, &pipe_ctx->ttu_regs); - } - - for (i = 0; i < dc->res_pool->pipe_count; i++) { - struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i]; - - if (!pipe_ctx->stream || pipe_ctx->stream == stream - || !pipe_ctx->plane_state) - continue; - dcn10_pipe_control_lock(dc, pipe_ctx, false); - } + tg->funcs->unlock(tg); + } if (num_planes == 0) false_optc_underflow_wa(dc, stream, tg); @@ -2390,6 +2368,22 @@ static void dcn10_apply_ctx_for_surface( hubbub1_wm_change_req_wa(dc->res_pool->hubbub); } +static void dcn10_stereo_hw_frame_pack_wa(struct dc *dc, struct dc_state *context) +{ + uint8_t i; + + for (i = 0; i < context->stream_count; i++) { + if (context->streams[i]->timing.timing_3d_format + == TIMING_3D_FORMAT_HW_FRAME_PACKING) { + /* + * Disable stutter + */ + hubbub1_allow_self_refresh_control(dc->res_pool->hubbub, false); + break; + } + } +} + static void dcn10_prepare_bandwidth( struct dc *dc, struct dc_state *context) @@ -2411,6 +2405,7 @@ static void dcn10_prepare_bandwidth( &context->bw.dcn.watermarks, dc->res_pool->ref_clock_inKhz / 1000, true); + dcn10_stereo_hw_frame_pack_wa(dc, context); if (dc->debug.pplib_wm_report_mode == WM_REPORT_OVERRIDE) dcn_bw_notify_pplib_of_wm_ranges(dc); @@ -2440,6 +2435,7 @@ static void dcn10_optimize_bandwidth( &context->bw.dcn.watermarks, dc->res_pool->ref_clock_inKhz / 1000, true); + dcn10_stereo_hw_frame_pack_wa(dc, context); if (dc->debug.pplib_wm_report_mode == WM_REPORT_OVERRIDE) dcn_bw_notify_pplib_of_wm_ranges(dc); @@ -2525,7 +2521,7 @@ static void dcn10_config_stereo_parameters( timing_3d_format == TIMING_3D_FORMAT_DP_HDMI_INBAND_FA || timing_3d_format == TIMING_3D_FORMAT_SIDEBAND_FA) { enum display_dongle_type dongle = \ - stream->sink->link->ddc->dongle_type; + stream->link->ddc->dongle_type; if (dongle == DISPLAY_DONGLE_DP_VGA_CONVERTER || dongle == DISPLAY_DONGLE_DP_DVI_CONVERTER || dongle == DISPLAY_DONGLE_DP_HDMI_CONVERTER) @@ -2656,7 +2652,7 @@ static void dcn10_set_cursor_position(struct pipe_ctx *pipe_ctx) struct hubp *hubp = pipe_ctx->plane_res.hubp; struct dpp *dpp = pipe_ctx->plane_res.dpp; struct dc_cursor_mi_param param = { - .pixel_clk_khz = pipe_ctx->stream->timing.pix_clk_khz, + .pixel_clk_khz = pipe_ctx->stream->timing.pix_clk_100hz / 10, .ref_clk_khz = pipe_ctx->stream->ctx->dc->res_pool->ref_clock_inKhz, .viewport = pipe_ctx->plane_res.scl_data.viewport, .h_scale_ratio = pipe_ctx->plane_res.scl_data.ratios.horz, @@ -2716,6 +2712,7 @@ static void dcn10_set_cursor_sdr_white_level(struct pipe_ctx *pipe_ctx) static const struct hw_sequencer_funcs dcn10_funcs = { .program_gamut_remap = program_gamut_remap, .init_hw = dcn10_init_hw, + .init_pipes = dcn10_init_pipes, .apply_ctx_to_hw = dce110_apply_ctx_to_hw, .apply_ctx_for_surface = dcn10_apply_ctx_for_surface, .update_plane_addr = dcn10_update_plane_addr, diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c index cd469014baa3..98f41d250978 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c @@ -40,7 +40,6 @@ #include "ipp.h" #include "mpc.h" #include "reg_helper.h" -#include "custom_float.h" #include "dcn10_hubp.h" #include "dcn10_hubbub.h" #include "dcn10_cm_common.h" @@ -72,7 +71,7 @@ static unsigned int snprintf_count(char *pBuf, unsigned int bufSize, char *fmt, static unsigned int dcn10_get_hubbub_state(struct dc *dc, char *pBuf, unsigned int bufSize) { struct dc_context *dc_ctx = dc->ctx; - struct dcn_hubbub_wm wm = {0}; + struct dcn_hubbub_wm wm; int i; unsigned int chars_printed = 0; @@ -81,7 +80,8 @@ static unsigned int dcn10_get_hubbub_state(struct dc *dc, char *pBuf, unsigned i const uint32_t ref_clk_mhz = dc_ctx->dc->res_pool->ref_clock_inKhz / 1000; static const unsigned int frac = 1000; - hubbub1_wm_read_state(dc->res_pool->hubbub, &wm); + memset(&wm, 0, sizeof(struct dcn_hubbub_wm)); + dc->res_pool->hubbub->funcs->wm_read_state(dc->res_pool->hubbub, &wm); chars_printed = snprintf_count(pBuf, remaining_buffer, "wm_set_index,data_urgent,pte_meta_urgent,sr_enter,sr_exit,dram_clk_chanage\n"); remaining_buffer -= chars_printed; @@ -419,20 +419,22 @@ static unsigned int dcn10_get_otg_states(struct dc *dc, char *pBuf, unsigned int unsigned int remaining_buffer = bufSize; chars_printed = snprintf_count(pBuf, remaining_buffer, "instance,v_bs,v_be,v_ss,v_se,vpol,vmax,vmin,vmax_sel,vmin_sel," - "h_bs,h_be,h_ss,h_se,hpol,htot,vtot,underflow\n"); + "h_bs,h_be,h_ss,h_se,hpol,htot,vtot,underflow,pixelclk[khz]\n"); remaining_buffer -= chars_printed; pBuf += chars_printed; for (i = 0; i < pool->timing_generator_count; i++) { struct timing_generator *tg = pool->timing_generators[i]; struct dcn_otg_state s = {0}; + int pix_clk = 0; optc1_read_otg_state(DCN10TG_FROM_TG(tg), &s); + pix_clk = dc->current_state->res_ctx.pipe_ctx[i].stream_res.pix_clk_params.requested_pix_clk_100hz / 10; //only print if OTG master is enabled if (s.otg_enabled & 1) { chars_printed = snprintf_count(pBuf, remaining_buffer, "%x,%d,%d,%d,%d,%d,%d,%d,%d,%d," - "%d,%d,%d,%d,%d,%d,%d,%d" + "%d,%d,%d,%d,%d,%d,%d,%d,%d" "\n", tg->inst, s.v_blank_start, @@ -451,7 +453,8 @@ static unsigned int dcn10_get_otg_states(struct dc *dc, char *pBuf, unsigned int s.h_sync_a_pol, s.h_total, s.v_total, - s.underflow_occurred_status); + s.underflow_occurred_status, + pix_clk); remaining_buffer -= chars_printed; pBuf += chars_printed; diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c index 477ab9222216..a9db372688ff 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c @@ -85,6 +85,7 @@ static const struct link_encoder_funcs dcn10_lnk_enc_funcs = { .enable_hpd = dcn10_link_encoder_enable_hpd, .disable_hpd = dcn10_link_encoder_disable_hpd, .is_dig_enabled = dcn10_is_dig_enabled, + .get_dig_frontend = dcn10_get_dig_frontend, .destroy = dcn10_link_encoder_destroy }; @@ -440,7 +441,7 @@ static uint8_t get_frontend_source( } } -void configure_encoder( +void enc1_configure_encoder( struct dcn10_link_encoder *enc10, const struct dc_link_settings *link_settings) { @@ -495,6 +496,15 @@ bool dcn10_is_dig_enabled(struct link_encoder *enc) return value; } +unsigned int dcn10_get_dig_frontend(struct link_encoder *enc) +{ + struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc); + uint32_t value; + + REG_GET(DIG_BE_CNTL, DIG_FE_SOURCE_SELECT, &value); + return value; +} + static void link_encoder_disable(struct dcn10_link_encoder *enc10) { /* reset training pattern */ @@ -543,12 +553,12 @@ bool dcn10_link_encoder_validate_dvi_output( if ((connector_signal == SIGNAL_TYPE_DVI_SINGLE_LINK || connector_signal == SIGNAL_TYPE_HDMI_TYPE_A) && signal != SIGNAL_TYPE_HDMI_TYPE_A && - crtc_timing->pix_clk_khz > TMDS_MAX_PIXEL_CLOCK) + crtc_timing->pix_clk_100hz > (TMDS_MAX_PIXEL_CLOCK * 10)) return false; - if (crtc_timing->pix_clk_khz < TMDS_MIN_PIXEL_CLOCK) + if (crtc_timing->pix_clk_100hz < (TMDS_MIN_PIXEL_CLOCK * 10)) return false; - if (crtc_timing->pix_clk_khz > max_pixel_clock) + if (crtc_timing->pix_clk_100hz > (max_pixel_clock * 10)) return false; /* DVI supports 6/8bpp single-link and 10/16bpp dual-link */ @@ -571,7 +581,7 @@ bool dcn10_link_encoder_validate_dvi_output( static bool dcn10_link_encoder_validate_hdmi_output( const struct dcn10_link_encoder *enc10, const struct dc_crtc_timing *crtc_timing, - int adjusted_pix_clk_khz) + int adjusted_pix_clk_100hz) { enum dc_color_depth max_deep_color = enc10->base.features.max_hdmi_deep_color; @@ -581,11 +591,11 @@ static bool dcn10_link_encoder_validate_hdmi_output( if (crtc_timing->display_color_depth < COLOR_DEPTH_888) return false; - if (adjusted_pix_clk_khz < TMDS_MIN_PIXEL_CLOCK) + if (adjusted_pix_clk_100hz < (TMDS_MIN_PIXEL_CLOCK * 10)) return false; - if ((adjusted_pix_clk_khz == 0) || - (adjusted_pix_clk_khz > enc10->base.features.max_hdmi_pixel_clock)) + if ((adjusted_pix_clk_100hz == 0) || + (adjusted_pix_clk_100hz > (enc10->base.features.max_hdmi_pixel_clock * 10))) return false; /* DCE11 HW does not support 420 */ @@ -594,7 +604,7 @@ static bool dcn10_link_encoder_validate_hdmi_output( return false; if (!enc10->base.features.flags.bits.HDMI_6GB_EN && - adjusted_pix_clk_khz >= 300000) + adjusted_pix_clk_100hz >= 3000000) return false; if (enc10->base.ctx->dc->debug.hdmi20_disable && crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR420) @@ -738,7 +748,7 @@ bool dcn10_link_encoder_validate_output_with_stream( case SIGNAL_TYPE_DVI_DUAL_LINK: is_valid = dcn10_link_encoder_validate_dvi_output( enc10, - stream->sink->link->connector_signal, + stream->link->connector_signal, stream->signal, &stream->timing); break; @@ -746,7 +756,7 @@ bool dcn10_link_encoder_validate_output_with_stream( is_valid = dcn10_link_encoder_validate_hdmi_output( enc10, &stream->timing, - stream->phy_pix_clk); + stream->phy_pix_clk * 10); break; case SIGNAL_TYPE_DISPLAY_PORT: case SIGNAL_TYPE_DISPLAY_PORT_MST: @@ -910,7 +920,7 @@ void dcn10_link_encoder_enable_dp_output( * but it's not passed to asic_control. * We need to set number of lanes manually. */ - configure_encoder(enc10, link_settings); + enc1_configure_encoder(enc10, link_settings); cntl.action = TRANSMITTER_CONTROL_ENABLE; cntl.engine_id = enc->preferred_engine; @@ -949,7 +959,7 @@ void dcn10_link_encoder_enable_dp_mst_output( * but it's not passed to asic_control. * We need to set number of lanes manually. */ - configure_encoder(enc10, link_settings); + enc1_configure_encoder(enc10, link_settings); cntl.action = TRANSMITTER_CONTROL_ENABLE; cntl.engine_id = ENGINE_ID_UNKNOWN; diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h index 49ead12b2532..b74b80a247ec 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h @@ -271,7 +271,7 @@ void dcn10_link_encoder_setup( struct link_encoder *enc, enum signal_type signal); -void configure_encoder( +void enc1_configure_encoder( struct dcn10_link_encoder *enc10, const struct dc_link_settings *link_settings); @@ -336,6 +336,8 @@ void dcn10_psr_program_secondary_packet(struct link_encoder *enc, bool dcn10_is_dig_enabled(struct link_encoder *enc); +unsigned int dcn10_get_dig_frontend(struct link_encoder *enc); + void dcn10_aux_initialize(struct dcn10_link_encoder *enc10); #endif /* __DC_LINK_ENCODER__DCN10_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c index 7c138615f17d..2f78a84f0dcb 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c @@ -97,6 +97,7 @@ static uint32_t get_start_vline(struct timing_generator *optc, const struct dc_c struct dc_crtc_timing patched_crtc_timing; int vesa_sync_start; int asic_blank_end; + int interlace_factor; int vertical_line_start; patched_crtc_timing = *dc_crtc_timing; @@ -110,13 +111,16 @@ static uint32_t get_start_vline(struct timing_generator *optc, const struct dc_c vesa_sync_start - patched_crtc_timing.h_border_left; + interlace_factor = patched_crtc_timing.flags.INTERLACE ? 2 : 1; + vesa_sync_start = patched_crtc_timing.v_addressable + patched_crtc_timing.v_border_bottom + patched_crtc_timing.v_front_porch; asic_blank_end = (patched_crtc_timing.v_total - vesa_sync_start - - patched_crtc_timing.v_border_top); + patched_crtc_timing.v_border_top) + * interlace_factor; vertical_line_start = asic_blank_end - optc->dlg_otg_param.vstartup_start + 1; if (vertical_line_start < 0) { @@ -127,40 +131,95 @@ static uint32_t get_start_vline(struct timing_generator *optc, const struct dc_c return vertical_line_start; } -void optc1_program_vline_interrupt( +static void calc_vline_position( struct timing_generator *optc, const struct dc_crtc_timing *dc_crtc_timing, - unsigned long long vsync_delta) + unsigned long long vsync_delta, + uint32_t *start_line, + uint32_t *end_line) { - - struct optc *optc1 = DCN10TG_FROM_TG(optc); - unsigned long long req_delta_tens_of_usec = div64_u64((vsync_delta + 9999), 10000); - unsigned long long pix_clk_hundreds_khz = div64_u64((dc_crtc_timing->pix_clk_khz + 99), 100); + unsigned long long pix_clk_hundreds_khz = div64_u64((dc_crtc_timing->pix_clk_100hz + 999), 1000); uint32_t req_delta_lines = (uint32_t) div64_u64( (req_delta_tens_of_usec * pix_clk_hundreds_khz + dc_crtc_timing->h_total - 1), dc_crtc_timing->h_total); uint32_t vsync_line = get_start_vline(optc, dc_crtc_timing); - uint32_t start_line = 0; - uint32_t endLine = 0; if (req_delta_lines != 0) - req_delta_lines--; + req_delta_lines--; - if (req_delta_lines > vsync_line) - start_line = dc_crtc_timing->v_total - (req_delta_lines - vsync_line) + 2; - else - start_line = vsync_line - req_delta_lines; + if (req_delta_lines > vsync_line) + *start_line = dc_crtc_timing->v_total - (req_delta_lines - vsync_line) - 1; + else + *start_line = vsync_line - req_delta_lines; + + *end_line = *start_line + 2; + + if (*end_line >= dc_crtc_timing->v_total) + *end_line = 2; +} + +void optc1_program_vline_interrupt( + struct timing_generator *optc, + const struct dc_crtc_timing *dc_crtc_timing, + enum vline_select vline, + const union vline_config *vline_config) +{ + struct optc *optc1 = DCN10TG_FROM_TG(optc); + uint32_t start_line = 0; + uint32_t end_line = 0; + + switch (vline) { + case VLINE0: + calc_vline_position(optc, dc_crtc_timing, vline_config->delta_in_ns, &start_line, &end_line); + REG_SET_2(OTG_VERTICAL_INTERRUPT0_POSITION, 0, + OTG_VERTICAL_INTERRUPT0_LINE_START, start_line, + OTG_VERTICAL_INTERRUPT0_LINE_END, end_line); + break; + case VLINE1: + REG_SET(OTG_VERTICAL_INTERRUPT1_POSITION, 0, + OTG_VERTICAL_INTERRUPT1_LINE_START, vline_config->line_number); + break; + default: + break; + } +} + +void optc1_program_vupdate_interrupt( + struct timing_generator *optc, + const struct dc_crtc_timing *dc_crtc_timing) +{ + struct optc *optc1 = DCN10TG_FROM_TG(optc); + int32_t vertical_line_start; + uint32_t asic_blank_end; + uint32_t vesa_sync_start; + struct dc_crtc_timing patched_crtc_timing; + + patched_crtc_timing = *dc_crtc_timing; + optc1_apply_front_porch_workaround(optc, &patched_crtc_timing); + + /* asic_h_blank_end = HsyncWidth + HbackPorch = + * vesa. usHorizontalTotal - vesa. usHorizontalSyncStart - + * vesa.h_left_border + */ + vesa_sync_start = patched_crtc_timing.h_addressable + + patched_crtc_timing.h_border_right + + patched_crtc_timing.h_front_porch; - endLine = start_line + 2; + asic_blank_end = patched_crtc_timing.h_total - + vesa_sync_start - + patched_crtc_timing.h_border_left; - if (endLine >= dc_crtc_timing->v_total) - endLine = 2; + /* Use OTG_VERTICAL_INTERRUPT2 replace VUPDATE interrupt, + * program the reg for interrupt postition. + */ + vertical_line_start = asic_blank_end - optc->dlg_otg_param.vstartup_start + 1; + if (vertical_line_start < 0) + vertical_line_start = 0; - REG_SET_2(OTG_VERTICAL_INTERRUPT0_POSITION, 0, - OTG_VERTICAL_INTERRUPT0_LINE_START, start_line, - OTG_VERTICAL_INTERRUPT0_LINE_END, endLine); + REG_SET(OTG_VERTICAL_INTERRUPT2_POSITION, 0, + OTG_VERTICAL_INTERRUPT2_LINE_START, vertical_line_start); } /** @@ -265,22 +324,14 @@ void optc1_program_timing( patched_crtc_timing.v_addressable + patched_crtc_timing.v_border_bottom); - REG_UPDATE_2(OTG_V_BLANK_START_END, - OTG_V_BLANK_START, asic_blank_start, - OTG_V_BLANK_END, asic_blank_end); - - /* Use OTG_VERTICAL_INTERRUPT2 replace VUPDATE interrupt, - * program the reg for interrupt postition. - */ vertical_line_start = asic_blank_end - optc->dlg_otg_param.vstartup_start + 1; v_fp2 = 0; if (vertical_line_start < 0) v_fp2 = -vertical_line_start; - if (vertical_line_start < 0) - vertical_line_start = 0; - REG_SET(OTG_VERTICAL_INTERRUPT2_POSITION, 0, - OTG_VERTICAL_INTERRUPT2_LINE_START, vertical_line_start); + REG_UPDATE_2(OTG_V_BLANK_START_END, + OTG_V_BLANK_START, asic_blank_start, + OTG_V_BLANK_END, asic_blank_end); /* v_sync polarity */ v_sync_polarity = patched_crtc_timing.flags.VSYNC_POSITIVE_POLARITY ? @@ -299,16 +350,17 @@ void optc1_program_timing( } /* Interlace */ - if (patched_crtc_timing.flags.INTERLACE == 1) { - REG_UPDATE(OTG_INTERLACE_CONTROL, - OTG_INTERLACE_ENABLE, 1); - v_init = v_init / 2; - if ((optc->dlg_otg_param.vstartup_start/2)*2 > asic_blank_end) - v_fp2 = v_fp2 / 2; - } else - REG_UPDATE(OTG_INTERLACE_CONTROL, - OTG_INTERLACE_ENABLE, 0); - + if (REG(OTG_INTERLACE_CONTROL)) { + if (patched_crtc_timing.flags.INTERLACE == 1) { + REG_UPDATE(OTG_INTERLACE_CONTROL, + OTG_INTERLACE_ENABLE, 1); + v_init = v_init / 2; + if ((optc->dlg_otg_param.vstartup_start/2)*2 > asic_blank_end) + v_fp2 = v_fp2 / 2; + } else + REG_UPDATE(OTG_INTERLACE_CONTROL, + OTG_INTERLACE_ENABLE, 0); + } /* VTG enable set to 0 first VInit */ REG_UPDATE(CONTROL, @@ -338,7 +390,7 @@ void optc1_program_timing( h_div_2 = optc1_is_two_pixels_per_containter(&patched_crtc_timing); REG_UPDATE(OTG_H_TIMING_CNTL, - OTG_H_TIMING_DIV_BY2, h_div_2); + OTG_H_TIMING_DIV_BY2, h_div_2 || optc1->comb_opp_id != 0xf); } @@ -1184,6 +1236,64 @@ bool optc1_is_stereo_left_eye(struct timing_generator *optc) return ret; } +bool optc1_is_matching_timing(struct timing_generator *tg, + const struct dc_crtc_timing *otg_timing) +{ + struct dc_crtc_timing hw_crtc_timing = {0}; + struct dcn_otg_state s = {0}; + + if (tg == NULL || otg_timing == NULL) + return false; + + optc1_read_otg_state(DCN10TG_FROM_TG(tg), &s); + + hw_crtc_timing.h_total = s.h_total + 1; + hw_crtc_timing.h_addressable = s.h_total - ((s.h_total - s.h_blank_start) + s.h_blank_end); + hw_crtc_timing.h_front_porch = s.h_total + 1 - s.h_blank_start; + hw_crtc_timing.h_sync_width = s.h_sync_a_end - s.h_sync_a_start; + + hw_crtc_timing.v_total = s.v_total + 1; + hw_crtc_timing.v_addressable = s.v_total - ((s.v_total - s.v_blank_start) + s.v_blank_end); + hw_crtc_timing.v_front_porch = s.v_total + 1 - s.v_blank_start; + hw_crtc_timing.v_sync_width = s.v_sync_a_end - s.v_sync_a_start; + + if (otg_timing->h_total != hw_crtc_timing.h_total) + return false; + + if (otg_timing->h_border_left != hw_crtc_timing.h_border_left) + return false; + + if (otg_timing->h_addressable != hw_crtc_timing.h_addressable) + return false; + + if (otg_timing->h_border_right != hw_crtc_timing.h_border_right) + return false; + + if (otg_timing->h_front_porch != hw_crtc_timing.h_front_porch) + return false; + + if (otg_timing->h_sync_width != hw_crtc_timing.h_sync_width) + return false; + + if (otg_timing->v_total != hw_crtc_timing.v_total) + return false; + + if (otg_timing->v_border_top != hw_crtc_timing.v_border_top) + return false; + + if (otg_timing->v_addressable != hw_crtc_timing.v_addressable) + return false; + + if (otg_timing->v_border_bottom != hw_crtc_timing.v_border_bottom) + return false; + + if (otg_timing->v_sync_width != hw_crtc_timing.v_sync_width) + return false; + + return true; +} + + void optc1_read_otg_state(struct optc *optc1, struct dcn_otg_state *s) { @@ -1371,6 +1481,7 @@ static const struct timing_generator_funcs dcn10_tg_funcs = { .validate_timing = optc1_validate_timing, .program_timing = optc1_program_timing, .program_vline_interrupt = optc1_program_vline_interrupt, + .program_vupdate_interrupt = optc1_program_vupdate_interrupt, .program_global_sync = optc1_program_global_sync, .enable_crtc = optc1_enable_crtc, .disable_crtc = optc1_disable_crtc, @@ -1380,6 +1491,7 @@ static const struct timing_generator_funcs dcn10_tg_funcs = { .get_frame_count = optc1_get_vblank_counter, .get_scanoutpos = optc1_get_crtc_scanoutpos, .get_otg_active_size = optc1_get_otg_active_size, + .is_matching_timing = optc1_is_matching_timing, .set_early_control = optc1_set_early_control, /* used by enable_timing_synchronization. Not need for FPGA */ .wait_for_state = optc1_wait_for_state, @@ -1419,10 +1531,13 @@ void dcn10_timing_generator_init(struct optc *optc1) optc1->min_v_blank_interlace = 5; optc1->min_h_sync_width = 8; optc1->min_v_sync_width = 1; + optc1->comb_opp_id = 0xf; } bool optc1_is_two_pixels_per_containter(const struct dc_crtc_timing *timing) { - return timing->pixel_encoding == PIXEL_ENCODING_YCBCR420; + bool two_pix = timing->pixel_encoding == PIXEL_ENCODING_YCBCR420; + + return two_pix; } diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h index 8bacf0b6e27e..24452f11c598 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h @@ -67,6 +67,8 @@ SRI(OTG_CLOCK_CONTROL, OTG, inst),\ SRI(OTG_VERTICAL_INTERRUPT0_CONTROL, OTG, inst),\ SRI(OTG_VERTICAL_INTERRUPT0_POSITION, OTG, inst),\ + SRI(OTG_VERTICAL_INTERRUPT1_CONTROL, OTG, inst),\ + SRI(OTG_VERTICAL_INTERRUPT1_POSITION, OTG, inst),\ SRI(OTG_VERTICAL_INTERRUPT2_CONTROL, OTG, inst),\ SRI(OTG_VERTICAL_INTERRUPT2_POSITION, OTG, inst),\ SRI(OPTC_INPUT_CLOCK_CONTROL, ODM, inst),\ @@ -135,6 +137,8 @@ struct dcn_optc_registers { uint32_t OTG_CLOCK_CONTROL; uint32_t OTG_VERTICAL_INTERRUPT0_CONTROL; uint32_t OTG_VERTICAL_INTERRUPT0_POSITION; + uint32_t OTG_VERTICAL_INTERRUPT1_CONTROL; + uint32_t OTG_VERTICAL_INTERRUPT1_POSITION; uint32_t OTG_VERTICAL_INTERRUPT2_CONTROL; uint32_t OTG_VERTICAL_INTERRUPT2_POSITION; uint32_t OPTC_INPUT_CLOCK_CONTROL; @@ -227,6 +231,8 @@ struct dcn_optc_registers { SF(OTG0_OTG_VERTICAL_INTERRUPT0_CONTROL, OTG_VERTICAL_INTERRUPT0_INT_ENABLE, mask_sh),\ SF(OTG0_OTG_VERTICAL_INTERRUPT0_POSITION, OTG_VERTICAL_INTERRUPT0_LINE_START, mask_sh),\ SF(OTG0_OTG_VERTICAL_INTERRUPT0_POSITION, OTG_VERTICAL_INTERRUPT0_LINE_END, mask_sh),\ + SF(OTG0_OTG_VERTICAL_INTERRUPT1_CONTROL, OTG_VERTICAL_INTERRUPT1_INT_ENABLE, mask_sh),\ + SF(OTG0_OTG_VERTICAL_INTERRUPT1_POSITION, OTG_VERTICAL_INTERRUPT1_LINE_START, mask_sh),\ SF(OTG0_OTG_VERTICAL_INTERRUPT2_CONTROL, OTG_VERTICAL_INTERRUPT2_INT_ENABLE, mask_sh),\ SF(OTG0_OTG_VERTICAL_INTERRUPT2_POSITION, OTG_VERTICAL_INTERRUPT2_LINE_START, mask_sh),\ SF(ODM0_OPTC_INPUT_CLOCK_CONTROL, OPTC_INPUT_CLK_EN, mask_sh),\ @@ -361,6 +367,8 @@ struct dcn_optc_registers { type OTG_VERTICAL_INTERRUPT0_INT_ENABLE;\ type OTG_VERTICAL_INTERRUPT0_LINE_START;\ type OTG_VERTICAL_INTERRUPT0_LINE_END;\ + type OTG_VERTICAL_INTERRUPT1_INT_ENABLE;\ + type OTG_VERTICAL_INTERRUPT1_LINE_START;\ type OTG_VERTICAL_INTERRUPT2_INT_ENABLE;\ type OTG_VERTICAL_INTERRUPT2_LINE_START;\ type OPTC_INPUT_CLK_EN;\ @@ -427,7 +435,7 @@ struct optc { const struct dcn_optc_shift *tg_shift; const struct dcn_optc_mask *tg_mask; - enum controller_id controller_id; + int comb_opp_id; uint32_t max_h_total; uint32_t max_v_total; @@ -475,9 +483,11 @@ void optc1_program_timing( const struct dc_crtc_timing *dc_crtc_timing, bool use_vbios); -void optc1_program_vline_interrupt(struct timing_generator *optc, +void optc1_program_vline_interrupt( + struct timing_generator *optc, const struct dc_crtc_timing *dc_crtc_timing, - unsigned long long vsync_delta); + enum vline_select vline, + const union vline_config *vline_config); void optc1_program_global_sync( struct timing_generator *optc); diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c index 5d4772dec0ba..09d74070a49b 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c @@ -70,7 +70,7 @@ const struct _vcs_dpi_ip_params_st dcn1_0_ip = { .rob_buffer_size_kbytes = 64, .det_buffer_size_kbytes = 164, - .dpte_buffer_size_in_pte_reqs = 42, + .dpte_buffer_size_in_pte_reqs_luma = 42, .dpp_output_buffer_pixels = 2560, .opp_output_buffer_lines = 1, .pixel_chunk_size_kbytes = 8, @@ -436,7 +436,6 @@ static const struct dcn_optc_mask tg_mask = { }; static const struct bios_registers bios_regs = { - NBIO_SR(BIOS_SCRATCH_0), NBIO_SR(BIOS_SCRATCH_3), NBIO_SR(BIOS_SCRATCH_6) }; @@ -609,7 +608,7 @@ static struct output_pixel_processor *dcn10_opp_create( return &opp->base; } -struct aux_engine *dcn10_aux_engine_create( +struct dce_aux *dcn10_aux_engine_create( struct dc_context *ctx, uint32_t inst) { @@ -678,18 +677,18 @@ static struct mpc *dcn10_mpc_create(struct dc_context *ctx) static struct hubbub *dcn10_hubbub_create(struct dc_context *ctx) { - struct hubbub *hubbub = kzalloc(sizeof(struct hubbub), + struct dcn10_hubbub *dcn10_hubbub = kzalloc(sizeof(struct dcn10_hubbub), GFP_KERNEL); - if (!hubbub) + if (!dcn10_hubbub) return NULL; - hubbub1_construct(hubbub, ctx, + hubbub1_construct(&dcn10_hubbub->base, ctx, &hubbub_reg, &hubbub_shift, &hubbub_mask); - return hubbub; + return &dcn10_hubbub->base; } static struct timing_generator *dcn10_timing_generator_create( @@ -911,7 +910,7 @@ static void destruct(struct dcn10_resource_pool *pool) for (i = 0; i < pool->base.res_cap->num_ddc; i++) { if (pool->base.engines[i] != NULL) - pool->base.engines[i]->funcs->destroy_engine(&pool->base.engines[i]); + dce110_engine_destroy(&pool->base.engines[i]); if (pool->base.hw_i2cs[i] != NULL) { kfree(pool->base.hw_i2cs[i]); pool->base.hw_i2cs[i] = NULL; @@ -974,8 +973,8 @@ static void get_pixel_clock_parameters( struct pixel_clk_params *pixel_clk_params) { const struct dc_stream_state *stream = pipe_ctx->stream; - pixel_clk_params->requested_pix_clk = stream->timing.pix_clk_khz; - pixel_clk_params->encoder_object_id = stream->sink->link->link_enc->id; + pixel_clk_params->requested_pix_clk_100hz = stream->timing.pix_clk_100hz; + pixel_clk_params->encoder_object_id = stream->link->link_enc->id; pixel_clk_params->signal_type = pipe_ctx->stream->signal; pixel_clk_params->controller_id = pipe_ctx->stream_res.tg->inst + 1; /* TODO: un-hardcode*/ @@ -991,9 +990,9 @@ static void get_pixel_clock_parameters( pixel_clk_params->color_depth = COLOR_DEPTH_888; if (stream->timing.pixel_encoding == PIXEL_ENCODING_YCBCR420) - pixel_clk_params->requested_pix_clk /= 2; + pixel_clk_params->requested_pix_clk_100hz /= 2; if (stream->timing.timing_3d_format == TIMING_3D_FORMAT_HW_FRAME_PACKING) - pixel_clk_params->requested_pix_clk *= 2; + pixel_clk_params->requested_pix_clk_100hz *= 2; } @@ -1131,6 +1130,56 @@ static enum dc_status dcn10_validate_plane(const struct dc_plane_state *plane_st return DC_OK; } +static enum dc_status dcn10_validate_global(struct dc *dc, struct dc_state *context) +{ + int i, j; + bool video_down_scaled = false; + bool video_large = false; + bool desktop_large = false; + bool dcc_disabled = false; + + for (i = 0; i < context->stream_count; i++) { + if (context->stream_status[i].plane_count == 0) + continue; + + if (context->stream_status[i].plane_count > 2) + return false; + + for (j = 0; j < context->stream_status[i].plane_count; j++) { + struct dc_plane_state *plane = + context->stream_status[i].plane_states[j]; + + + if (plane->format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN) { + + if (plane->src_rect.width > plane->dst_rect.width || + plane->src_rect.height > plane->dst_rect.height) + video_down_scaled = true; + + if (plane->src_rect.width >= 3840) + video_large = true; + + } else { + if (plane->src_rect.width >= 3840) + desktop_large = true; + if (!plane->dcc.enable) + dcc_disabled = true; + } + } + } + + /* + * Workaround: On DCN10 there is UMC issue that causes underflow when + * playing 4k video on 4k desktop with video downscaled and single channel + * memory + */ + if (video_large && desktop_large && video_down_scaled && dcc_disabled && + dc->dcn_soc->number_of_channels == 1) + return DC_FAIL_SURFACE_VALIDATE; + + return DC_OK; +} + static enum dc_status dcn10_get_default_swizzle_mode(struct dc_plane_state *plane_state) { enum dc_status result = DC_OK; @@ -1159,6 +1208,7 @@ static const struct resource_funcs dcn10_res_pool_funcs = { .validate_bandwidth = dcn_validate_bandwidth, .acquire_idle_pipe_for_layer = dcn10_acquire_idle_pipe_for_layer, .validate_plane = dcn10_validate_plane, + .validate_global = dcn10_validate_global, .add_stream_to_ctx = dcn10_add_stream_to_ctx, .get_default_swizzle_mode = dcn10_get_default_swizzle_mode }; diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.c index b8b5525a389a..b08254121251 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.c @@ -261,17 +261,29 @@ void enc1_stream_encoder_dp_set_stream_attribute( uint8_t dp_component_depth = 0; struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc); + struct dc_crtc_timing hw_crtc_timing = *crtc_timing; + + if (hw_crtc_timing.flags.INTERLACE) { + /*the input timing is in VESA spec format with Interlace flag =1*/ + hw_crtc_timing.v_total /= 2; + hw_crtc_timing.v_border_top /= 2; + hw_crtc_timing.v_addressable /= 2; + hw_crtc_timing.v_border_bottom /= 2; + hw_crtc_timing.v_front_porch /= 2; + hw_crtc_timing.v_sync_width /= 2; + } + /* set pixel encoding */ - switch (crtc_timing->pixel_encoding) { + switch (hw_crtc_timing.pixel_encoding) { case PIXEL_ENCODING_YCBCR422: dp_pixel_encoding = DP_PIXEL_ENCODING_TYPE_YCBCR422; break; case PIXEL_ENCODING_YCBCR444: dp_pixel_encoding = DP_PIXEL_ENCODING_TYPE_YCBCR444; - if (crtc_timing->flags.Y_ONLY) - if (crtc_timing->display_color_depth != COLOR_DEPTH_666) + if (hw_crtc_timing.flags.Y_ONLY) + if (hw_crtc_timing.display_color_depth != COLOR_DEPTH_666) /* HW testing only, no use case yet. * Color depth of Y-only could be * 8, 10, 12, 16 bits @@ -299,7 +311,7 @@ void enc1_stream_encoder_dp_set_stream_attribute( * Pixel Encoding/Colorimetry Format and that a Sink device shall ignore MISC1, bit 7, * and MISC0, bits 7:1 (MISC1, bit 7, and MISC0, bits 7:1, become "don't care"). */ - if ((crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR420) || + if ((hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR420) || (output_color_space == COLOR_SPACE_2020_YCBCR) || (output_color_space == COLOR_SPACE_2020_RGB_FULLRANGE) || (output_color_space == COLOR_SPACE_2020_RGB_LIMITEDRANGE)) @@ -308,7 +320,7 @@ void enc1_stream_encoder_dp_set_stream_attribute( misc1 = misc1 & ~0x40; /* set color depth */ - switch (crtc_timing->display_color_depth) { + switch (hw_crtc_timing.display_color_depth) { case COLOR_DEPTH_666: dp_component_depth = DP_COMPONENT_PIXEL_DEPTH_6BPC; break; @@ -336,7 +348,7 @@ void enc1_stream_encoder_dp_set_stream_attribute( /* set dynamic range and YCbCr range */ - switch (crtc_timing->display_color_depth) { + switch (hw_crtc_timing.display_color_depth) { case COLOR_DEPTH_666: colorimetry_bpc = 0; break; @@ -372,9 +384,9 @@ void enc1_stream_encoder_dp_set_stream_attribute( misc0 = misc0 | 0x8; /* bit3=1, bit4=0 */ misc1 = misc1 & ~0x80; /* bit7 = 0*/ dynamic_range_ycbcr = 0; /*bt601*/ - if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) + if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR422) misc0 = misc0 | 0x2; /* bit2=0, bit1=1 */ - else if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR444) + else if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR444) misc0 = misc0 | 0x4; /* bit2=1, bit1=0 */ break; case COLOR_SPACE_YCBCR709: @@ -382,9 +394,9 @@ void enc1_stream_encoder_dp_set_stream_attribute( misc0 = misc0 | 0x18; /* bit3=1, bit4=1 */ misc1 = misc1 & ~0x80; /* bit7 = 0*/ dynamic_range_ycbcr = 1; /*bt709*/ - if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR422) + if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR422) misc0 = misc0 | 0x2; /* bit2=0, bit1=1 */ - else if (crtc_timing->pixel_encoding == PIXEL_ENCODING_YCBCR444) + else if (hw_crtc_timing.pixel_encoding == PIXEL_ENCODING_YCBCR444) misc0 = misc0 | 0x4; /* bit2=1, bit1=0 */ break; case COLOR_SPACE_2020_RGB_LIMITEDRANGE: @@ -414,26 +426,26 @@ void enc1_stream_encoder_dp_set_stream_attribute( * dc_crtc_timing is vesa dmt struct. data from edid */ REG_SET_2(DP_MSA_TIMING_PARAM1, 0, - DP_MSA_HTOTAL, crtc_timing->h_total, - DP_MSA_VTOTAL, crtc_timing->v_total); + DP_MSA_HTOTAL, hw_crtc_timing.h_total, + DP_MSA_VTOTAL, hw_crtc_timing.v_total); /* calculate from vesa timing parameters * h_active_start related to leading edge of sync */ - h_blank = crtc_timing->h_total - crtc_timing->h_border_left - - crtc_timing->h_addressable - crtc_timing->h_border_right; + h_blank = hw_crtc_timing.h_total - hw_crtc_timing.h_border_left - + hw_crtc_timing.h_addressable - hw_crtc_timing.h_border_right; - h_back_porch = h_blank - crtc_timing->h_front_porch - - crtc_timing->h_sync_width; + h_back_porch = h_blank - hw_crtc_timing.h_front_porch - + hw_crtc_timing.h_sync_width; /* start at beginning of left border */ - h_active_start = crtc_timing->h_sync_width + h_back_porch; + h_active_start = hw_crtc_timing.h_sync_width + h_back_porch; - v_active_start = crtc_timing->v_total - crtc_timing->v_border_top - - crtc_timing->v_addressable - crtc_timing->v_border_bottom - - crtc_timing->v_front_porch; + v_active_start = hw_crtc_timing.v_total - hw_crtc_timing.v_border_top - + hw_crtc_timing.v_addressable - hw_crtc_timing.v_border_bottom - + hw_crtc_timing.v_front_porch; /* start at beginning of left border */ @@ -443,20 +455,20 @@ void enc1_stream_encoder_dp_set_stream_attribute( REG_SET_4(DP_MSA_TIMING_PARAM3, 0, DP_MSA_HSYNCWIDTH, - crtc_timing->h_sync_width, + hw_crtc_timing.h_sync_width, DP_MSA_HSYNCPOLARITY, - !crtc_timing->flags.HSYNC_POSITIVE_POLARITY, + !hw_crtc_timing.flags.HSYNC_POSITIVE_POLARITY, DP_MSA_VSYNCWIDTH, - crtc_timing->v_sync_width, + hw_crtc_timing.v_sync_width, DP_MSA_VSYNCPOLARITY, - !crtc_timing->flags.VSYNC_POSITIVE_POLARITY); + !hw_crtc_timing.flags.VSYNC_POSITIVE_POLARITY); /* HWDITH include border or overscan */ REG_SET_2(DP_MSA_TIMING_PARAM4, 0, - DP_MSA_HWIDTH, crtc_timing->h_border_left + - crtc_timing->h_addressable + crtc_timing->h_border_right, - DP_MSA_VHEIGHT, crtc_timing->v_border_top + - crtc_timing->v_addressable + crtc_timing->v_border_bottom); + DP_MSA_HWIDTH, hw_crtc_timing.h_border_left + + hw_crtc_timing.h_addressable + hw_crtc_timing.h_border_right, + DP_MSA_VHEIGHT, hw_crtc_timing.v_border_top + + hw_crtc_timing.v_addressable + hw_crtc_timing.v_border_bottom); } static void enc1_stream_encoder_set_stream_attribute_helper( @@ -594,7 +606,7 @@ void enc1_stream_encoder_dvi_set_stream_attribute( cntl.signal = is_dual_link ? SIGNAL_TYPE_DVI_DUAL_LINK : SIGNAL_TYPE_DVI_SINGLE_LINK; cntl.enable_dp_audio = false; - cntl.pixel_clock = crtc_timing->pix_clk_khz; + cntl.pixel_clock = crtc_timing->pix_clk_100hz / 10; cntl.lanes_number = (is_dual_link) ? LANE_COUNT_EIGHT : LANE_COUNT_FOUR; if (enc1->base.bp->funcs->encoder_control( @@ -1413,6 +1425,14 @@ void enc1_setup_stereo_sync( REG_UPDATE(DIG_FE_CNTL, DIG_STEREOSYNC_GATE_EN, !enable); } +void enc1_dig_connect_to_otg( + struct stream_encoder *enc, + int tg_inst) +{ + struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc); + + REG_UPDATE(DIG_FE_CNTL, DIG_SOURCE_SELECT, tg_inst); +} static const struct stream_encoder_funcs dcn10_str_enc_funcs = { .dp_set_stream_attribute = @@ -1445,6 +1465,7 @@ static const struct stream_encoder_funcs dcn10_str_enc_funcs = { .hdmi_audio_disable = enc1_se_hdmi_audio_disable, .setup_stereo_sync = enc1_setup_stereo_sync, .set_avmute = enc1_stream_encoder_set_avmute, + .dig_connect_to_otg = enc1_dig_connect_to_otg, }; void dcn10_stream_encoder_construct( diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.h index 67f3e4dd95c1..b7c800e10a32 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.h +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_stream_encoder.h @@ -274,7 +274,8 @@ struct dcn10_stream_enc_registers { SE_SF(DP0_DP_MSA_TIMING_PARAM4, DP_MSA_HWIDTH, mask_sh),\ SE_SF(DP0_DP_MSA_TIMING_PARAM4, DP_MSA_VHEIGHT, mask_sh),\ SE_SF(DIG0_HDMI_DB_CONTROL, HDMI_DB_DISABLE, mask_sh),\ - SE_SF(DP0_DP_VID_TIMING, DP_VID_N_MUL, mask_sh) + SE_SF(DP0_DP_VID_TIMING, DP_VID_N_MUL, mask_sh),\ + SE_SF(DIG0_DIG_FE_CNTL, DIG_SOURCE_SELECT, mask_sh) #define SE_COMMON_MASK_SH_LIST_SOC(mask_sh)\ SE_COMMON_MASK_SH_LIST_SOC_BASE(mask_sh) @@ -426,7 +427,8 @@ struct dcn10_stream_enc_registers { type DP_MSA_VHEIGHT;\ type HDMI_DB_DISABLE;\ type DP_VID_N_MUL;\ - type DP_VID_M_DOUBLE_VALUE_EN + type DP_VID_M_DOUBLE_VALUE_EN;\ + type DIG_SOURCE_SELECT struct dcn10_stream_encoder_shift { SE_REG_FIELD_LIST_DCN1_0(uint8_t); @@ -523,4 +525,8 @@ void enc1_se_hdmi_audio_setup( void enc1_se_hdmi_audio_disable( struct stream_encoder *enc); +void enc1_dig_connect_to_otg( + struct stream_encoder *enc, + int tg_inst); + #endif /* __DC_STREAM_ENCODER_DCN10_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/dm_helpers.h b/drivers/gpu/drm/amd/display/dc/dm_helpers.h index 5d4527d03045..e81b24374bcb 100644 --- a/drivers/gpu/drm/amd/display/dc/dm_helpers.h +++ b/drivers/gpu/drm/amd/display/dc/dm_helpers.h @@ -58,6 +58,13 @@ bool dm_helpers_dp_mst_write_payload_allocation_table( bool enable); /* + * poll pending down reply before clear payload allocation table + */ +void dm_helpers_dp_mst_poll_pending_down_reply( + struct dc_context *ctx, + const struct dc_link *link); + +/* * Clear payload allocation table before enable MST DP link. */ void dm_helpers_dp_mst_clear_payload_allocation_table( diff --git a/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h b/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h index 0029a39efb1c..14bed5b1fa97 100644 --- a/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h +++ b/drivers/gpu/drm/amd/display/dc/dm_pp_smu.h @@ -38,7 +38,8 @@ enum pp_smu_ver { * of interface sharing between families of ASIcs. */ PP_SMU_UNSUPPORTED, - PP_SMU_VER_RV + PP_SMU_VER_RV, + PP_SMU_VER_MAX }; struct pp_smu { diff --git a/drivers/gpu/drm/amd/display/dc/dm_services_types.h b/drivers/gpu/drm/amd/display/dc/dm_services_types.h index 1af8c777b3ac..77200711abbe 100644 --- a/drivers/gpu/drm/amd/display/dc/dm_services_types.h +++ b/drivers/gpu/drm/amd/display/dc/dm_services_types.h @@ -82,9 +82,17 @@ enum dm_pp_clock_type { #define DC_DECODE_PP_CLOCK_TYPE(clk_type) \ (clk_type) == DM_PP_CLOCK_TYPE_DISPLAY_CLK ? "Display" : \ (clk_type) == DM_PP_CLOCK_TYPE_ENGINE_CLK ? "Engine" : \ - (clk_type) == DM_PP_CLOCK_TYPE_MEMORY_CLK ? "Memory" : "Invalid" - -#define DM_PP_MAX_CLOCK_LEVELS 8 + (clk_type) == DM_PP_CLOCK_TYPE_MEMORY_CLK ? "Memory" : \ + (clk_type) == DM_PP_CLOCK_TYPE_DCFCLK ? "DCF" : \ + (clk_type) == DM_PP_CLOCK_TYPE_DCEFCLK ? "DCEF" : \ + (clk_type) == DM_PP_CLOCK_TYPE_SOCCLK ? "SoC" : \ + (clk_type) == DM_PP_CLOCK_TYPE_PIXELCLK ? "Pixel" : \ + (clk_type) == DM_PP_CLOCK_TYPE_DISPLAYPHYCLK ? "Display PHY" : \ + (clk_type) == DM_PP_CLOCK_TYPE_DPPCLK ? "DPP" : \ + (clk_type) == DM_PP_CLOCK_TYPE_FCLK ? "F" : \ + "Invalid" + +#define DM_PP_MAX_CLOCK_LEVELS 16 struct dm_pp_clock_levels { uint32_t num_levels; diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h index bea4e61b94c7..c59e582c1f40 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h @@ -121,4 +121,30 @@ enum self_refresh_affinity { dm_neither_self_refresh_nor_mclk_switch }; +enum dm_validation_status { + DML_VALIDATION_OK, + DML_FAIL_SCALE_RATIO_TAP, + DML_FAIL_SOURCE_PIXEL_FORMAT, + DML_FAIL_VIEWPORT_SIZE, + DML_FAIL_TOTAL_V_ACTIVE_BW, + DML_FAIL_DIO_SUPPORT, + DML_FAIL_NOT_ENOUGH_DSC, + DML_FAIL_DSC_CLK_REQUIRED, + DML_FAIL_URGENT_LATENCY, + DML_FAIL_REORDERING_BUFFER, + DML_FAIL_DISPCLK_DPPCLK, + DML_FAIL_TOTAL_AVAILABLE_PIPES, + DML_FAIL_NUM_OTG, + DML_FAIL_WRITEBACK_MODE, + DML_FAIL_WRITEBACK_LATENCY, + DML_FAIL_WRITEBACK_SCALE_RATIO_TAP, + DML_FAIL_CURSOR_SUPPORT, + DML_FAIL_PITCH_SUPPORT, + DML_FAIL_PTE_BUFFER_SIZE, + DML_FAIL_HOST_VM_IMMEDIATE_FLIP, + DML_FAIL_DSC_INPUT_BPC, + DML_FAIL_PREFETCH_SUPPORT, + DML_FAIL_V_RATIO_PREFETCH, +}; + #endif diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c index dddeb0d4db8f..d303b789adfe 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c @@ -62,3 +62,31 @@ void dml_init_instance(struct display_mode_lib *lib, enum dml_project project) } } +const char *dml_get_status_message(enum dm_validation_status status) +{ + switch (status) { + case DML_VALIDATION_OK: return "Validation OK"; + case DML_FAIL_SCALE_RATIO_TAP: return "Scale ratio/tap"; + case DML_FAIL_SOURCE_PIXEL_FORMAT: return "Source pixel format"; + case DML_FAIL_VIEWPORT_SIZE: return "Viewport size"; + case DML_FAIL_TOTAL_V_ACTIVE_BW: return "Total vertical active bandwidth"; + case DML_FAIL_DIO_SUPPORT: return "DIO support"; + case DML_FAIL_NOT_ENOUGH_DSC: return "Not enough DSC Units"; + case DML_FAIL_DSC_CLK_REQUIRED: return "DSC clock required"; + case DML_FAIL_URGENT_LATENCY: return "Urgent latency"; + case DML_FAIL_REORDERING_BUFFER: return "Re-ordering buffer"; + case DML_FAIL_DISPCLK_DPPCLK: return "Dispclk and Dppclk"; + case DML_FAIL_TOTAL_AVAILABLE_PIPES: return "Total available pipes"; + case DML_FAIL_NUM_OTG: return "Number of OTG"; + case DML_FAIL_WRITEBACK_MODE: return "Writeback mode"; + case DML_FAIL_WRITEBACK_LATENCY: return "Writeback latency"; + case DML_FAIL_WRITEBACK_SCALE_RATIO_TAP: return "Writeback scale ratio/tap"; + case DML_FAIL_CURSOR_SUPPORT: return "Cursor support"; + case DML_FAIL_PITCH_SUPPORT: return "Pitch support"; + case DML_FAIL_PTE_BUFFER_SIZE: return "PTE buffer size"; + case DML_FAIL_DSC_INPUT_BPC: return "DSC input bpc"; + case DML_FAIL_PREFETCH_SUPPORT: return "Prefetch support"; + case DML_FAIL_V_RATIO_PREFETCH: return "Vertical ratio prefetch"; + default: return "Unknown Status"; + } +} diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h index 635206248889..a730e0209c05 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h @@ -43,4 +43,6 @@ struct display_mode_lib { void dml_init_instance(struct display_mode_lib *lib, enum dml_project project); +const char *dml_get_status_message(enum dm_validation_status status); + #endif diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h index 5dd04520ceca..391183e3428f 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h +++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h @@ -30,22 +30,15 @@ typedef struct _vcs_dpi_soc_bounding_box_st soc_bounding_box_st; typedef struct _vcs_dpi_ip_params_st ip_params_st; typedef struct _vcs_dpi_display_pipe_source_params_st display_pipe_source_params_st; typedef struct _vcs_dpi_display_output_params_st display_output_params_st; -typedef struct _vcs_dpi_display_bandwidth_st display_bandwidth_st; typedef struct _vcs_dpi_scaler_ratio_depth_st scaler_ratio_depth_st; typedef struct _vcs_dpi_scaler_taps_st scaler_taps_st; typedef struct _vcs_dpi_display_pipe_dest_params_st display_pipe_dest_params_st; typedef struct _vcs_dpi_display_pipe_params_st display_pipe_params_st; typedef struct _vcs_dpi_display_clocks_and_cfg_st display_clocks_and_cfg_st; typedef struct _vcs_dpi_display_e2e_pipe_params_st display_e2e_pipe_params_st; -typedef struct _vcs_dpi_dchub_buffer_sizing_st dchub_buffer_sizing_st; -typedef struct _vcs_dpi_watermarks_perf_st watermarks_perf_st; -typedef struct _vcs_dpi_cstate_pstate_watermarks_st cstate_pstate_watermarks_st; -typedef struct _vcs_dpi_wm_calc_pipe_params_st wm_calc_pipe_params_st; -typedef struct _vcs_dpi_vratio_pre_st vratio_pre_st; typedef struct _vcs_dpi_display_data_rq_misc_params_st display_data_rq_misc_params_st; typedef struct _vcs_dpi_display_data_rq_sizing_params_st display_data_rq_sizing_params_st; typedef struct _vcs_dpi_display_data_rq_dlg_params_st display_data_rq_dlg_params_st; -typedef struct _vcs_dpi_display_cur_rq_dlg_params_st display_cur_rq_dlg_params_st; typedef struct _vcs_dpi_display_rq_dlg_params_st display_rq_dlg_params_st; typedef struct _vcs_dpi_display_rq_sizing_params_st display_rq_sizing_params_st; typedef struct _vcs_dpi_display_rq_misc_params_st display_rq_misc_params_st; @@ -55,8 +48,6 @@ typedef struct _vcs_dpi_display_ttu_regs_st display_ttu_regs_st; typedef struct _vcs_dpi_display_data_rq_regs_st display_data_rq_regs_st; typedef struct _vcs_dpi_display_rq_regs_st display_rq_regs_st; typedef struct _vcs_dpi_display_dlg_sys_params_st display_dlg_sys_params_st; -typedef struct _vcs_dpi_display_dlg_prefetch_param_st display_dlg_prefetch_param_st; -typedef struct _vcs_dpi_display_pipe_clock_st display_pipe_clock_st; typedef struct _vcs_dpi_display_arb_params_st display_arb_params_st; struct _vcs_dpi_voltage_scaling_st { @@ -111,8 +102,6 @@ struct _vcs_dpi_soc_bounding_box_st { double xfc_bus_transport_time_us; double xfc_xbuf_latency_tolerance_us; int use_urgent_burst_bw; - double max_hscl_ratio; - double max_vscl_ratio; unsigned int num_states; struct _vcs_dpi_voltage_scaling_st clock_limits[8]; }; @@ -129,7 +118,8 @@ struct _vcs_dpi_ip_params_st { unsigned int odm_capable; unsigned int rob_buffer_size_kbytes; unsigned int det_buffer_size_kbytes; - unsigned int dpte_buffer_size_in_pte_reqs; + unsigned int dpte_buffer_size_in_pte_reqs_luma; + unsigned int dpte_buffer_size_in_pte_reqs_chroma; unsigned int pde_proc_buffer_size_64k_reqs; unsigned int dpp_output_buffer_pixels; unsigned int opp_output_buffer_lines; @@ -192,7 +182,6 @@ struct _vcs_dpi_display_xfc_params_st { struct _vcs_dpi_display_pipe_source_params_st { int source_format; unsigned char dcc; - unsigned int dcc_override; unsigned int dcc_rate; unsigned char dcc_use_global; unsigned char vm; @@ -205,7 +194,6 @@ struct _vcs_dpi_display_pipe_source_params_st { int source_scan; int sw_mode; int macro_tile_size; - unsigned char is_display_sw; unsigned int viewport_width; unsigned int viewport_height; unsigned int viewport_y_y; @@ -252,16 +240,10 @@ struct _vcs_dpi_display_output_params_st { int output_bpc; int output_type; int output_format; - int output_standard; int dsc_slices; struct writeback_st wb; }; -struct _vcs_dpi_display_bandwidth_st { - double total_bw_consumed_gbps; - double guaranteed_urgent_return_bw_gbps; -}; - struct _vcs_dpi_scaler_ratio_depth_st { double hscl_ratio; double vscl_ratio; @@ -300,11 +282,9 @@ struct _vcs_dpi_display_pipe_dest_params_st { unsigned int vupdate_width; unsigned int vready_offset; unsigned char interlaced; - unsigned char underscan; double pixel_rate_mhz; unsigned char synchronized_vblank_all_planes; unsigned char otg_inst; - unsigned char odm_split_cnt; unsigned char odm_combine; unsigned char use_maximum_vstartup; }; @@ -331,65 +311,6 @@ struct _vcs_dpi_display_e2e_pipe_params_st { display_clocks_and_cfg_st clks_cfg; }; -struct _vcs_dpi_dchub_buffer_sizing_st { - unsigned int swath_width_y; - unsigned int swath_height_y; - unsigned int swath_height_c; - unsigned int detail_buffer_size_y; -}; - -struct _vcs_dpi_watermarks_perf_st { - double stutter_eff_in_active_region_percent; - double urgent_latency_supported_us; - double non_urgent_latency_supported_us; - double dram_clock_change_margin_us; - double dram_access_eff_percent; -}; - -struct _vcs_dpi_cstate_pstate_watermarks_st { - double cstate_exit_us; - double cstate_enter_plus_exit_us; - double pstate_change_us; -}; - -struct _vcs_dpi_wm_calc_pipe_params_st { - unsigned int num_dpp; - int voltage; - int output_type; - double dcfclk_mhz; - double socclk_mhz; - double dppclk_mhz; - double pixclk_mhz; - unsigned char interlace_en; - unsigned char pte_enable; - unsigned char dcc_enable; - double dcc_rate; - double bytes_per_pixel_c; - double bytes_per_pixel_y; - unsigned int swath_width_y; - unsigned int swath_height_y; - unsigned int swath_height_c; - unsigned int det_buffer_size_y; - double h_ratio; - double v_ratio; - unsigned int h_taps; - unsigned int h_total; - unsigned int v_total; - unsigned int v_active; - unsigned int e2e_index; - double display_pipe_line_delivery_time; - double read_bw; - unsigned int lines_in_det_y; - unsigned int lines_in_det_y_rounded_down_to_swath; - double full_det_buffering_time; - double dcfclk_deepsleep_mhz_per_plane; -}; - -struct _vcs_dpi_vratio_pre_st { - double vratio_pre_l; - double vratio_pre_c; -}; - struct _vcs_dpi_display_data_rq_misc_params_st { unsigned int full_swath_bytes; unsigned int stored_swath_bytes; @@ -423,16 +344,9 @@ struct _vcs_dpi_display_data_rq_dlg_params_st { unsigned int meta_bytes_per_row_ub; }; -struct _vcs_dpi_display_cur_rq_dlg_params_st { - unsigned char enable; - unsigned int swath_height; - unsigned int req_per_line; -}; - struct _vcs_dpi_display_rq_dlg_params_st { display_data_rq_dlg_params_st rq_l; display_data_rq_dlg_params_st rq_c; - display_cur_rq_dlg_params_st rq_cur0; }; struct _vcs_dpi_display_rq_sizing_params_st { @@ -498,6 +412,10 @@ struct _vcs_dpi_display_dlg_regs_st { unsigned int xfc_reg_remote_surface_flip_latency; unsigned int xfc_reg_prefetch_margin; unsigned int dst_y_delta_drq_limit; + unsigned int refcyc_per_vm_group_vblank; + unsigned int refcyc_per_vm_group_flip; + unsigned int refcyc_per_vm_req_vblank; + unsigned int refcyc_per_vm_req_flip; }; struct _vcs_dpi_display_ttu_regs_st { @@ -556,19 +474,6 @@ struct _vcs_dpi_display_dlg_sys_params_st { unsigned int total_flip_bytes; }; -struct _vcs_dpi_display_dlg_prefetch_param_st { - double prefetch_bw; - unsigned int flip_bytes; -}; - -struct _vcs_dpi_display_pipe_clock_st { - double dcfclk_mhz; - double dispclk_mhz; - double socclk_mhz; - double dscclk_mhz[6]; - double dppclk_mhz[6]; -}; - struct _vcs_dpi_display_arb_params_st { int max_req_outstanding; int min_req_outstanding; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c b/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c index c2037daa8e66..ad8571f5a142 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c @@ -459,7 +459,7 @@ static void dml1_rq_dlg_get_row_heights( /* dpte */ /* ------ */ log2_vmpg_bytes = dml_log2(mode_lib->soc.vmm_page_size_bytes); - dpte_buf_in_pte_reqs = mode_lib->ip.dpte_buffer_size_in_pte_reqs; + dpte_buf_in_pte_reqs = mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma; log2_vmpg_height = 0; log2_vmpg_width = 0; @@ -776,7 +776,7 @@ static void get_surf_rq_param( /* dpte */ /* ------ */ log2_vmpg_bytes = dml_log2(mode_lib->soc.vmm_page_size_bytes); - dpte_buf_in_pte_reqs = mode_lib->ip.dpte_buffer_size_in_pte_reqs; + dpte_buf_in_pte_reqs = mode_lib->ip.dpte_buffer_size_in_pte_reqs_luma; log2_vmpg_height = 0; log2_vmpg_width = 0; @@ -881,7 +881,7 @@ static void get_surf_rq_param( /* the dpte_group_bytes is reduced for the specific case of vertical * access of a tile surface that has dpte request of 8x1 ptes. */ - if (!surf_linear & (log2_dpte_req_height_ptes == 0) & surf_vert) /*reduced, in this case, will have page fault within a group */ + if (!surf_linear && (log2_dpte_req_height_ptes == 0) && surf_vert) /*reduced, in this case, will have page fault within a group */ rq_sizing_param->dpte_group_bytes = 512; else /*full size */ diff --git a/drivers/gpu/drm/amd/display/dc/gpio/gpio_base.c b/drivers/gpu/drm/amd/display/dc/gpio/gpio_base.c index 1d1efd72b291..cf76ea2d9f5a 100644 --- a/drivers/gpu/drm/amd/display/dc/gpio/gpio_base.c +++ b/drivers/gpu/drm/amd/display/dc/gpio/gpio_base.c @@ -101,6 +101,18 @@ enum gpio_mode dal_gpio_get_mode( return gpio->mode; } +enum gpio_result dal_gpio_lock_pin( + struct gpio *gpio) +{ + return dal_gpio_service_lock(gpio->service, gpio->id, gpio->en); +} + +enum gpio_result dal_gpio_unlock_pin( + struct gpio *gpio) +{ + return dal_gpio_service_unlock(gpio->service, gpio->id, gpio->en); +} + enum gpio_result dal_gpio_change_mode( struct gpio *gpio, enum gpio_mode mode) diff --git a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c index dada04296025..3c63a3c04dbb 100644 --- a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c +++ b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c @@ -192,6 +192,34 @@ static void set_pin_free( service->busyness[id][en] = false; } +enum gpio_result dal_gpio_service_lock( + struct gpio_service *service, + enum gpio_id id, + uint32_t en) +{ + if (!service->busyness[id]) { + ASSERT_CRITICAL(false); + return GPIO_RESULT_OPEN_FAILED; + } + + set_pin_busy(service, id, en); + return GPIO_RESULT_OK; +} + +enum gpio_result dal_gpio_service_unlock( + struct gpio_service *service, + enum gpio_id id, + uint32_t en) +{ + if (!service->busyness[id]) { + ASSERT_CRITICAL(false); + return GPIO_RESULT_OPEN_FAILED; + } + + set_pin_free(service, id, en); + return GPIO_RESULT_OK; +} + enum gpio_result dal_gpio_service_open( struct gpio_service *service, enum gpio_id id, diff --git a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h index 1d501a43d13b..0c678af75331 100644 --- a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h +++ b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h @@ -52,4 +52,14 @@ void dal_gpio_service_close( struct gpio_service *service, struct hw_gpio_pin **ptr); +enum gpio_result dal_gpio_service_lock( + struct gpio_service *service, + enum gpio_id id, + uint32_t en); + +enum gpio_result dal_gpio_service_unlock( + struct gpio_service *service, + enum gpio_id id, + uint32_t en); + #endif diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/Makefile b/drivers/gpu/drm/amd/display/dc/i2caux/Makefile deleted file mode 100644 index 352885cb4d07..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/Makefile +++ /dev/null @@ -1,99 +0,0 @@ -# -# Copyright 2017 Advanced Micro Devices, Inc. -# -# Permission is hereby granted, free of charge, to any person obtaining a -# copy of this software and associated documentation files (the "Software"), -# to deal in the Software without restriction, including without limitation -# the rights to use, copy, modify, merge, publish, distribute, sublicense, -# and/or sell copies of the Software, and to permit persons to whom the -# Software is furnished to do so, subject to the following conditions: -# -# The above copyright notice and this permission notice shall be included in -# all copies or substantial portions of the Software. -# -# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR -# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, -# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL -# THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR -# OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, -# ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR -# OTHER DEALINGS IN THE SOFTWARE. -# -# -# Makefile for the 'i2c' sub-component of DAL. -# It provides the control and status of HW i2c engine of the adapter. - -I2CAUX = aux_engine.o engine_base.o i2caux.o i2c_engine.o \ - i2c_generic_hw_engine.o i2c_hw_engine.o i2c_sw_engine.o - -AMD_DAL_I2CAUX = $(addprefix $(AMDDALPATH)/dc/i2caux/,$(I2CAUX)) - -AMD_DISPLAY_FILES += $(AMD_DAL_I2CAUX) - -############################################################################### -# DCE 8x family -############################################################################### -I2CAUX_DCE80 = i2caux_dce80.o i2c_hw_engine_dce80.o \ - i2c_sw_engine_dce80.o - -AMD_DAL_I2CAUX_DCE80 = $(addprefix $(AMDDALPATH)/dc/i2caux/dce80/,$(I2CAUX_DCE80)) - -AMD_DISPLAY_FILES += $(AMD_DAL_I2CAUX_DCE80) - -############################################################################### -# DCE 100 family -############################################################################### -I2CAUX_DCE100 = i2caux_dce100.o - -AMD_DAL_I2CAUX_DCE100 = $(addprefix $(AMDDALPATH)/dc/i2caux/dce100/,$(I2CAUX_DCE100)) - -AMD_DISPLAY_FILES += $(AMD_DAL_I2CAUX_DCE100) - -############################################################################### -# DCE 110 family -############################################################################### -I2CAUX_DCE110 = i2caux_dce110.o i2c_sw_engine_dce110.o i2c_hw_engine_dce110.o \ - aux_engine_dce110.o - -AMD_DAL_I2CAUX_DCE110 = $(addprefix $(AMDDALPATH)/dc/i2caux/dce110/,$(I2CAUX_DCE110)) - -AMD_DISPLAY_FILES += $(AMD_DAL_I2CAUX_DCE110) - -############################################################################### -# DCE 112 family -############################################################################### -I2CAUX_DCE112 = i2caux_dce112.o - -AMD_DAL_I2CAUX_DCE112 = $(addprefix $(AMDDALPATH)/dc/i2caux/dce112/,$(I2CAUX_DCE112)) - -AMD_DISPLAY_FILES += $(AMD_DAL_I2CAUX_DCE112) - -############################################################################### -# DCN 1.0 family -############################################################################### -ifdef CONFIG_DRM_AMD_DC_DCN1_0 -I2CAUX_DCN1 = i2caux_dcn10.o - -AMD_DAL_I2CAUX_DCN1 = $(addprefix $(AMDDALPATH)/dc/i2caux/dcn10/,$(I2CAUX_DCN1)) - -AMD_DISPLAY_FILES += $(AMD_DAL_I2CAUX_DCN1) -endif - -############################################################################### -# DCE 120 family -############################################################################### -I2CAUX_DCE120 = i2caux_dce120.o - -AMD_DAL_I2CAUX_DCE120 = $(addprefix $(AMDDALPATH)/dc/i2caux/dce120/,$(I2CAUX_DCE120)) - -AMD_DISPLAY_FILES += $(AMD_DAL_I2CAUX_DCE120) - -############################################################################### -# Diagnostics on FPGA -############################################################################### -I2CAUX_DIAG = i2caux_diag.o - -AMD_DAL_I2CAUX_DIAG = $(addprefix $(AMDDALPATH)/dc/i2caux/diagnostics/,$(I2CAUX_DIAG)) - -AMD_DISPLAY_FILES += $(AMD_DAL_I2CAUX_DIAG) - diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/aux_engine.c b/drivers/gpu/drm/amd/display/dc/i2caux/aux_engine.c deleted file mode 100644 index 8cbf38b2470d..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/aux_engine.c +++ /dev/null @@ -1,606 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" -#include "dm_event_log.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "engine.h" - -/* - * Header of this unit - */ - -#include "aux_engine.h" - -/* - * Post-requisites: headers required by this unit - */ - -#include "include/link_service_types.h" - -/* - * This unit - */ - -enum { - AUX_INVALID_REPLY_RETRY_COUNTER = 1, - AUX_TIMED_OUT_RETRY_COUNTER = 2, - AUX_DEFER_RETRY_COUNTER = 6 -}; - -#define FROM_ENGINE(ptr) \ - container_of((ptr), struct aux_engine, base) -#define DC_LOGGER \ - engine->base.ctx->logger - -enum i2caux_engine_type dal_aux_engine_get_engine_type( - const struct engine *engine) -{ - return I2CAUX_ENGINE_TYPE_AUX; -} - -bool dal_aux_engine_acquire( - struct engine *engine, - struct ddc *ddc) -{ - struct aux_engine *aux_engine = FROM_ENGINE(engine); - - enum gpio_result result; - if (aux_engine->funcs->is_engine_available) { - /*check whether SW could use the engine*/ - if (!aux_engine->funcs->is_engine_available(aux_engine)) { - return false; - } - } - - result = dal_ddc_open(ddc, GPIO_MODE_HARDWARE, - GPIO_DDC_CONFIG_TYPE_MODE_AUX); - - if (result != GPIO_RESULT_OK) - return false; - - if (!aux_engine->funcs->acquire_engine(aux_engine)) { - dal_ddc_close(ddc); - return false; - } - - engine->ddc = ddc; - - return true; -} - -struct read_command_context { - uint8_t *buffer; - uint32_t current_read_length; - uint32_t offset; - enum i2caux_transaction_status status; - - struct aux_request_transaction_data request; - struct aux_reply_transaction_data reply; - - uint8_t returned_byte; - - uint32_t timed_out_retry_aux; - uint32_t invalid_reply_retry_aux; - uint32_t defer_retry_aux; - uint32_t defer_retry_i2c; - uint32_t invalid_reply_retry_aux_on_ack; - - bool transaction_complete; - bool operation_succeeded; -}; - -static void process_read_reply( - struct aux_engine *engine, - struct read_command_context *ctx) -{ - engine->funcs->process_channel_reply(engine, &ctx->reply); - - switch (ctx->reply.status) { - case AUX_TRANSACTION_REPLY_AUX_ACK: - ctx->defer_retry_aux = 0; - if (ctx->returned_byte > ctx->current_read_length) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_PROTOCOL_ERROR; - ctx->operation_succeeded = false; - } else if (ctx->returned_byte < ctx->current_read_length) { - ctx->current_read_length -= ctx->returned_byte; - - ctx->offset += ctx->returned_byte; - - ++ctx->invalid_reply_retry_aux_on_ack; - - if (ctx->invalid_reply_retry_aux_on_ack > - AUX_INVALID_REPLY_RETRY_COUNTER) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_PROTOCOL_ERROR; - ctx->operation_succeeded = false; - } - } else { - ctx->status = I2CAUX_TRANSACTION_STATUS_SUCCEEDED; - ctx->transaction_complete = true; - ctx->operation_succeeded = true; - } - break; - case AUX_TRANSACTION_REPLY_AUX_NACK: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_NACK; - ctx->operation_succeeded = false; - break; - case AUX_TRANSACTION_REPLY_AUX_DEFER: - ++ctx->defer_retry_aux; - - if (ctx->defer_retry_aux > AUX_DEFER_RETRY_COUNTER) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } - break; - case AUX_TRANSACTION_REPLY_I2C_DEFER: - ctx->defer_retry_aux = 0; - - ++ctx->defer_retry_i2c; - - if (ctx->defer_retry_i2c > AUX_DEFER_RETRY_COUNTER) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } - break; - case AUX_TRANSACTION_REPLY_HPD_DISCON: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_HPD_DISCON; - ctx->operation_succeeded = false; - break; - default: - ctx->status = I2CAUX_TRANSACTION_STATUS_UNKNOWN; - ctx->operation_succeeded = false; - } -} - -static void process_read_request( - struct aux_engine *engine, - struct read_command_context *ctx) -{ - enum aux_channel_operation_result operation_result; - - engine->funcs->submit_channel_request(engine, &ctx->request); - - operation_result = engine->funcs->get_channel_status( - engine, &ctx->returned_byte); - - switch (operation_result) { - case AUX_CHANNEL_OPERATION_SUCCEEDED: - if (ctx->returned_byte > ctx->current_read_length) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_PROTOCOL_ERROR; - ctx->operation_succeeded = false; - } else { - ctx->timed_out_retry_aux = 0; - ctx->invalid_reply_retry_aux = 0; - - ctx->reply.length = ctx->returned_byte; - ctx->reply.data = ctx->buffer; - - process_read_reply(engine, ctx); - } - break; - case AUX_CHANNEL_OPERATION_FAILED_INVALID_REPLY: - ++ctx->invalid_reply_retry_aux; - - if (ctx->invalid_reply_retry_aux > - AUX_INVALID_REPLY_RETRY_COUNTER) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_PROTOCOL_ERROR; - ctx->operation_succeeded = false; - } else - udelay(400); - break; - case AUX_CHANNEL_OPERATION_FAILED_TIMEOUT: - ++ctx->timed_out_retry_aux; - - if (ctx->timed_out_retry_aux > AUX_TIMED_OUT_RETRY_COUNTER) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } else { - /* DP 1.2a, table 2-58: - * "S3: AUX Request CMD PENDING: - * retry 3 times, with 400usec wait on each" - * The HW timeout is set to 550usec, - * so we should not wait here */ - } - break; - case AUX_CHANNEL_OPERATION_FAILED_HPD_DISCON: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_HPD_DISCON; - ctx->operation_succeeded = false; - break; - default: - ctx->status = I2CAUX_TRANSACTION_STATUS_UNKNOWN; - ctx->operation_succeeded = false; - } -} - -static bool read_command( - struct aux_engine *engine, - struct i2caux_transaction_request *request, - bool middle_of_transaction) -{ - struct read_command_context ctx; - - ctx.buffer = request->payload.data; - ctx.current_read_length = request->payload.length; - ctx.offset = 0; - ctx.timed_out_retry_aux = 0; - ctx.invalid_reply_retry_aux = 0; - ctx.defer_retry_aux = 0; - ctx.defer_retry_i2c = 0; - ctx.invalid_reply_retry_aux_on_ack = 0; - ctx.transaction_complete = false; - ctx.operation_succeeded = true; - - if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_DPCD) { - ctx.request.type = AUX_TRANSACTION_TYPE_DP; - ctx.request.action = I2CAUX_TRANSACTION_ACTION_DP_READ; - ctx.request.address = request->payload.address; - } else if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_I2C) { - ctx.request.type = AUX_TRANSACTION_TYPE_I2C; - ctx.request.action = middle_of_transaction ? - I2CAUX_TRANSACTION_ACTION_I2C_READ_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_READ; - ctx.request.address = request->payload.address >> 1; - } else { - /* in DAL2, there was no return in such case */ - BREAK_TO_DEBUGGER(); - return false; - } - - ctx.request.delay = 0; - - do { - memset(ctx.buffer + ctx.offset, 0, ctx.current_read_length); - - ctx.request.data = ctx.buffer + ctx.offset; - ctx.request.length = ctx.current_read_length; - - process_read_request(engine, &ctx); - - request->status = ctx.status; - - if (ctx.operation_succeeded && !ctx.transaction_complete) - if (ctx.request.type == AUX_TRANSACTION_TYPE_I2C) - msleep(engine->delay); - } while (ctx.operation_succeeded && !ctx.transaction_complete); - - if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_DPCD) { - DC_LOG_I2C_AUX("READ: addr:0x%x value:0x%x Result:%d", - request->payload.address, - request->payload.data[0], - ctx.operation_succeeded); - } - - return ctx.operation_succeeded; -} - -struct write_command_context { - bool mot; - - uint8_t *buffer; - uint32_t current_write_length; - enum i2caux_transaction_status status; - - struct aux_request_transaction_data request; - struct aux_reply_transaction_data reply; - - uint8_t returned_byte; - - uint32_t timed_out_retry_aux; - uint32_t invalid_reply_retry_aux; - uint32_t defer_retry_aux; - uint32_t defer_retry_i2c; - uint32_t max_defer_retry; - uint32_t ack_m_retry; - - uint8_t reply_data[DEFAULT_AUX_MAX_DATA_SIZE]; - - bool transaction_complete; - bool operation_succeeded; -}; - -static void process_write_reply( - struct aux_engine *engine, - struct write_command_context *ctx) -{ - engine->funcs->process_channel_reply(engine, &ctx->reply); - - switch (ctx->reply.status) { - case AUX_TRANSACTION_REPLY_AUX_ACK: - ctx->operation_succeeded = true; - - if (ctx->returned_byte) { - ctx->request.action = ctx->mot ? - I2CAUX_TRANSACTION_ACTION_I2C_STATUS_REQUEST_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_STATUS_REQUEST; - - ctx->current_write_length = 0; - - ++ctx->ack_m_retry; - - if (ctx->ack_m_retry > AUX_DEFER_RETRY_COUNTER) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } else - udelay(300); - } else { - ctx->status = I2CAUX_TRANSACTION_STATUS_SUCCEEDED; - ctx->defer_retry_aux = 0; - ctx->ack_m_retry = 0; - ctx->transaction_complete = true; - } - break; - case AUX_TRANSACTION_REPLY_AUX_NACK: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_NACK; - ctx->operation_succeeded = false; - break; - case AUX_TRANSACTION_REPLY_AUX_DEFER: - ++ctx->defer_retry_aux; - - if (ctx->defer_retry_aux > ctx->max_defer_retry) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } - break; - case AUX_TRANSACTION_REPLY_I2C_DEFER: - ctx->defer_retry_aux = 0; - ctx->current_write_length = 0; - - ctx->request.action = ctx->mot ? - I2CAUX_TRANSACTION_ACTION_I2C_STATUS_REQUEST_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_STATUS_REQUEST; - - ++ctx->defer_retry_i2c; - - if (ctx->defer_retry_i2c > ctx->max_defer_retry) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } - break; - case AUX_TRANSACTION_REPLY_HPD_DISCON: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_HPD_DISCON; - ctx->operation_succeeded = false; - break; - default: - ctx->status = I2CAUX_TRANSACTION_STATUS_UNKNOWN; - ctx->operation_succeeded = false; - } -} - -static void process_write_request( - struct aux_engine *engine, - struct write_command_context *ctx) -{ - enum aux_channel_operation_result operation_result; - - engine->funcs->submit_channel_request(engine, &ctx->request); - - operation_result = engine->funcs->get_channel_status( - engine, &ctx->returned_byte); - - switch (operation_result) { - case AUX_CHANNEL_OPERATION_SUCCEEDED: - ctx->timed_out_retry_aux = 0; - ctx->invalid_reply_retry_aux = 0; - - ctx->reply.length = ctx->returned_byte; - ctx->reply.data = ctx->reply_data; - - process_write_reply(engine, ctx); - break; - case AUX_CHANNEL_OPERATION_FAILED_INVALID_REPLY: - ++ctx->invalid_reply_retry_aux; - - if (ctx->invalid_reply_retry_aux > - AUX_INVALID_REPLY_RETRY_COUNTER) { - ctx->status = - I2CAUX_TRANSACTION_STATUS_FAILED_PROTOCOL_ERROR; - ctx->operation_succeeded = false; - } else - udelay(400); - break; - case AUX_CHANNEL_OPERATION_FAILED_TIMEOUT: - ++ctx->timed_out_retry_aux; - - if (ctx->timed_out_retry_aux > AUX_TIMED_OUT_RETRY_COUNTER) { - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - ctx->operation_succeeded = false; - } else { - /* DP 1.2a, table 2-58: - * "S3: AUX Request CMD PENDING: - * retry 3 times, with 400usec wait on each" - * The HW timeout is set to 550usec, - * so we should not wait here */ - } - break; - case AUX_CHANNEL_OPERATION_FAILED_HPD_DISCON: - ctx->status = I2CAUX_TRANSACTION_STATUS_FAILED_HPD_DISCON; - ctx->operation_succeeded = false; - break; - default: - ctx->status = I2CAUX_TRANSACTION_STATUS_UNKNOWN; - ctx->operation_succeeded = false; - } -} - -static bool write_command( - struct aux_engine *engine, - struct i2caux_transaction_request *request, - bool middle_of_transaction) -{ - struct write_command_context ctx; - - ctx.mot = middle_of_transaction; - ctx.buffer = request->payload.data; - ctx.current_write_length = request->payload.length; - ctx.timed_out_retry_aux = 0; - ctx.invalid_reply_retry_aux = 0; - ctx.defer_retry_aux = 0; - ctx.defer_retry_i2c = 0; - ctx.ack_m_retry = 0; - ctx.transaction_complete = false; - ctx.operation_succeeded = true; - - if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_DPCD) { - ctx.request.type = AUX_TRANSACTION_TYPE_DP; - ctx.request.action = I2CAUX_TRANSACTION_ACTION_DP_WRITE; - ctx.request.address = request->payload.address; - } else if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_I2C) { - ctx.request.type = AUX_TRANSACTION_TYPE_I2C; - ctx.request.action = middle_of_transaction ? - I2CAUX_TRANSACTION_ACTION_I2C_WRITE_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_WRITE; - ctx.request.address = request->payload.address >> 1; - } else { - /* in DAL2, there was no return in such case */ - BREAK_TO_DEBUGGER(); - return false; - } - - ctx.request.delay = 0; - - ctx.max_defer_retry = - (engine->max_defer_write_retry > AUX_DEFER_RETRY_COUNTER) ? - engine->max_defer_write_retry : AUX_DEFER_RETRY_COUNTER; - - do { - ctx.request.data = ctx.buffer; - ctx.request.length = ctx.current_write_length; - - process_write_request(engine, &ctx); - - request->status = ctx.status; - - if (ctx.operation_succeeded && !ctx.transaction_complete) - if (ctx.request.type == AUX_TRANSACTION_TYPE_I2C) - msleep(engine->delay); - } while (ctx.operation_succeeded && !ctx.transaction_complete); - - if (request->payload.address_space == - I2CAUX_TRANSACTION_ADDRESS_SPACE_DPCD) { - DC_LOG_I2C_AUX("WRITE: addr:0x%x value:0x%x Result:%d", - request->payload.address, - request->payload.data[0], - ctx.operation_succeeded); - } - - return ctx.operation_succeeded; -} - -static bool end_of_transaction_command( - struct aux_engine *engine, - struct i2caux_transaction_request *request) -{ - struct i2caux_transaction_request dummy_request; - uint8_t dummy_data; - - /* [tcheng] We only need to send the stop (read with MOT = 0) - * for I2C-over-Aux, not native AUX */ - - if (request->payload.address_space != - I2CAUX_TRANSACTION_ADDRESS_SPACE_I2C) - return false; - - dummy_request.operation = request->operation; - dummy_request.payload.address_space = request->payload.address_space; - dummy_request.payload.address = request->payload.address; - - /* - * Add a dummy byte due to some receiver quirk - * where one byte is sent along with MOT = 0. - * Ideally this should be 0. - */ - - dummy_request.payload.length = 0; - dummy_request.payload.data = &dummy_data; - - if (request->operation == I2CAUX_TRANSACTION_READ) - return read_command(engine, &dummy_request, false); - else - return write_command(engine, &dummy_request, false); - - /* according Syed, it does not need now DoDummyMOT */ -} - -bool dal_aux_engine_submit_request( - struct engine *engine, - struct i2caux_transaction_request *request, - bool middle_of_transaction) -{ - struct aux_engine *aux_engine = FROM_ENGINE(engine); - - bool result; - bool mot_used = true; - - switch (request->operation) { - case I2CAUX_TRANSACTION_READ: - result = read_command(aux_engine, request, mot_used); - break; - case I2CAUX_TRANSACTION_WRITE: - result = write_command(aux_engine, request, mot_used); - break; - default: - result = false; - } - - /* [tcheng] - * need to send stop for the last transaction to free up the AUX - * if the above command fails, this would be the last transaction */ - - if (!middle_of_transaction || !result) - end_of_transaction_command(aux_engine, request); - - /* mask AUX interrupt */ - - return result; -} - -void dal_aux_engine_construct( - struct aux_engine *engine, - struct dc_context *ctx) -{ - dal_i2caux_construct_engine(&engine->base, ctx); - engine->delay = 0; - engine->max_defer_write_retry = 0; -} - -void dal_aux_engine_destruct( - struct aux_engine *engine) -{ - dal_i2caux_destruct_engine(&engine->base); -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/aux_engine.h b/drivers/gpu/drm/amd/display/dc/i2caux/aux_engine.h deleted file mode 100644 index c33a2898d967..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/aux_engine.h +++ /dev/null @@ -1,86 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_AUX_ENGINE_H__ -#define __DAL_AUX_ENGINE_H__ - -#include "dc_ddc_types.h" - -struct aux_engine; - -struct aux_engine_funcs { - void (*destroy)( - struct aux_engine **ptr); - bool (*acquire_engine)( - struct aux_engine *engine); - void (*configure)( - struct aux_engine *engine, - union aux_config cfg); - void (*submit_channel_request)( - struct aux_engine *engine, - struct aux_request_transaction_data *request); - void (*process_channel_reply)( - struct aux_engine *engine, - struct aux_reply_transaction_data *reply); - int (*read_channel_reply)( - struct aux_engine *engine, - uint32_t size, - uint8_t *buffer, - uint8_t *reply_result, - uint32_t *sw_status); - enum aux_channel_operation_result (*get_channel_status)( - struct aux_engine *engine, - uint8_t *returned_bytes); - bool (*is_engine_available) ( - struct aux_engine *engine); -}; - -struct aux_engine { - struct engine base; - const struct aux_engine_funcs *funcs; - /* following values are expressed in milliseconds */ - uint32_t delay; - uint32_t max_defer_write_retry; - - bool acquire_reset; -}; - -void dal_aux_engine_construct( - struct aux_engine *engine, - struct dc_context *ctx); - -void dal_aux_engine_destruct( - struct aux_engine *engine); -bool dal_aux_engine_submit_request( - struct engine *ptr, - struct i2caux_transaction_request *request, - bool middle_of_transaction); -bool dal_aux_engine_acquire( - struct engine *ptr, - struct ddc *ddc); -enum i2caux_engine_type dal_aux_engine_get_engine_type( - const struct engine *engine); - -#endif diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce100/i2caux_dce100.c b/drivers/gpu/drm/amd/display/dc/i2caux/dce100/i2caux_dce100.c deleted file mode 100644 index 8b704ab0471c..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce100/i2caux_dce100.c +++ /dev/null @@ -1,106 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -#include "include/i2caux_interface.h" -#include "../i2caux.h" -#include "../engine.h" -#include "../i2c_engine.h" -#include "../i2c_sw_engine.h" -#include "../i2c_hw_engine.h" - -#include "../dce110/aux_engine_dce110.h" -#include "../dce110/i2c_hw_engine_dce110.h" -#include "../dce110/i2caux_dce110.h" - -#include "dce/dce_10_0_d.h" -#include "dce/dce_10_0_sh_mask.h" - -/* set register offset */ -#define SR(reg_name)\ - .reg_name = mm ## reg_name - -/* set register offset with instance */ -#define SRI(reg_name, block, id)\ - .reg_name = mm ## block ## id ## _ ## reg_name - -#define aux_regs(id)\ -[id] = {\ - AUX_COMMON_REG_LIST(id), \ - .AUX_RESET_MASK = 0 \ -} - -#define hw_engine_regs(id)\ -{\ - I2C_HW_ENGINE_COMMON_REG_LIST(id) \ -} - -static const struct dce110_aux_registers dce100_aux_regs[] = { - aux_regs(0), - aux_regs(1), - aux_regs(2), - aux_regs(3), - aux_regs(4), - aux_regs(5), -}; - -static const struct dce110_i2c_hw_engine_registers dce100_hw_engine_regs[] = { - hw_engine_regs(1), - hw_engine_regs(2), - hw_engine_regs(3), - hw_engine_regs(4), - hw_engine_regs(5), - hw_engine_regs(6) -}; - -static const struct dce110_i2c_hw_engine_shift i2c_shift = { - I2C_COMMON_MASK_SH_LIST_DCE100(__SHIFT) -}; - -static const struct dce110_i2c_hw_engine_mask i2c_mask = { - I2C_COMMON_MASK_SH_LIST_DCE100(_MASK) -}; - -struct i2caux *dal_i2caux_dce100_create( - struct dc_context *ctx) -{ - struct i2caux_dce110 *i2caux_dce110 = - kzalloc(sizeof(struct i2caux_dce110), GFP_KERNEL); - - if (!i2caux_dce110) { - ASSERT_CRITICAL(false); - return NULL; - } - - dal_i2caux_dce110_construct(i2caux_dce110, - ctx, - ARRAY_SIZE(dce100_aux_regs), - dce100_aux_regs, - dce100_hw_engine_regs, - &i2c_shift, - &i2c_mask); - return &i2caux_dce110->base; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce100/i2caux_dce100.h b/drivers/gpu/drm/amd/display/dc/i2caux/dce100/i2caux_dce100.h deleted file mode 100644 index 2b508d3e0ef4..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce100/i2caux_dce100.h +++ /dev/null @@ -1,32 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_AUX_DCE100_H__ -#define __DAL_I2C_AUX_DCE100_H__ - -struct i2caux *dal_i2caux_dce100_create( - struct dc_context *ctx); - -#endif /* __DAL_I2C_AUX_DCE100_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.c b/drivers/gpu/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.c deleted file mode 100644 index 59c3ed43d609..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.c +++ /dev/null @@ -1,505 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" -#include "dm_event_log.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "../engine.h" -#include "../aux_engine.h" - -/* - * Header of this unit - */ - -#include "aux_engine_dce110.h" - -/* - * Post-requisites: headers required by this unit - */ -#include "dce/dce_11_0_sh_mask.h" - -#define CTX \ - aux110->base.base.ctx -#define REG(reg_name)\ - (aux110->regs->reg_name) -#include "reg_helper.h" - -/* - * This unit - */ - -/* - * @brief - * Cast 'struct aux_engine *' - * to 'struct aux_engine_dce110 *' - */ -#define FROM_AUX_ENGINE(ptr) \ - container_of((ptr), struct aux_engine_dce110, base) - -/* - * @brief - * Cast 'struct engine *' - * to 'struct aux_engine_dce110 *' - */ -#define FROM_ENGINE(ptr) \ - FROM_AUX_ENGINE(container_of((ptr), struct aux_engine, base)) - -static void release_engine( - struct engine *engine) -{ - struct aux_engine_dce110 *aux110 = FROM_ENGINE(engine); - - REG_UPDATE(AUX_ARB_CONTROL, AUX_SW_DONE_USING_AUX_REG, 1); -} - -static void destruct( - struct aux_engine_dce110 *engine); - -static void destroy( - struct aux_engine **aux_engine) -{ - struct aux_engine_dce110 *engine = FROM_AUX_ENGINE(*aux_engine); - - destruct(engine); - - kfree(engine); - - *aux_engine = NULL; -} - -#define SW_CAN_ACCESS_AUX 1 -#define DMCU_CAN_ACCESS_AUX 2 - -static bool is_engine_available( - struct aux_engine *engine) -{ - struct aux_engine_dce110 *aux110 = FROM_AUX_ENGINE(engine); - - uint32_t value = REG_READ(AUX_ARB_CONTROL); - uint32_t field = get_reg_field_value( - value, - AUX_ARB_CONTROL, - AUX_REG_RW_CNTL_STATUS); - - return (field != DMCU_CAN_ACCESS_AUX); -} -static bool acquire_engine( - struct aux_engine *engine) -{ - struct aux_engine_dce110 *aux110 = FROM_AUX_ENGINE(engine); - - uint32_t value = REG_READ(AUX_ARB_CONTROL); - uint32_t field = get_reg_field_value( - value, - AUX_ARB_CONTROL, - AUX_REG_RW_CNTL_STATUS); - if (field == DMCU_CAN_ACCESS_AUX) - return false; - /* enable AUX before request SW to access AUX */ - value = REG_READ(AUX_CONTROL); - field = get_reg_field_value(value, - AUX_CONTROL, - AUX_EN); - - if (field == 0) { - set_reg_field_value( - value, - 1, - AUX_CONTROL, - AUX_EN); - - if (REG(AUX_RESET_MASK)) { - /*DP_AUX block as part of the enable sequence*/ - set_reg_field_value( - value, - 1, - AUX_CONTROL, - AUX_RESET); - } - - REG_WRITE(AUX_CONTROL, value); - - if (REG(AUX_RESET_MASK)) { - /*poll HW to make sure reset it done*/ - - REG_WAIT(AUX_CONTROL, AUX_RESET_DONE, 1, - 1, 11); - - set_reg_field_value( - value, - 0, - AUX_CONTROL, - AUX_RESET); - - REG_WRITE(AUX_CONTROL, value); - - REG_WAIT(AUX_CONTROL, AUX_RESET_DONE, 0, - 1, 11); - } - } /*if (field)*/ - - /* request SW to access AUX */ - REG_UPDATE(AUX_ARB_CONTROL, AUX_SW_USE_AUX_REG_REQ, 1); - - value = REG_READ(AUX_ARB_CONTROL); - field = get_reg_field_value( - value, - AUX_ARB_CONTROL, - AUX_REG_RW_CNTL_STATUS); - - return (field == SW_CAN_ACCESS_AUX); -} - -#define COMPOSE_AUX_SW_DATA_16_20(command, address) \ - ((command) | ((0xF0000 & (address)) >> 16)) - -#define COMPOSE_AUX_SW_DATA_8_15(address) \ - ((0xFF00 & (address)) >> 8) - -#define COMPOSE_AUX_SW_DATA_0_7(address) \ - (0xFF & (address)) - -static void submit_channel_request( - struct aux_engine *engine, - struct aux_request_transaction_data *request) -{ - struct aux_engine_dce110 *aux110 = FROM_AUX_ENGINE(engine); - uint32_t value; - uint32_t length; - - bool is_write = - ((request->type == AUX_TRANSACTION_TYPE_DP) && - (request->action == I2CAUX_TRANSACTION_ACTION_DP_WRITE)) || - ((request->type == AUX_TRANSACTION_TYPE_I2C) && - ((request->action == I2CAUX_TRANSACTION_ACTION_I2C_WRITE) || - (request->action == I2CAUX_TRANSACTION_ACTION_I2C_WRITE_MOT))); - if (REG(AUXN_IMPCAL)) { - /* clear_aux_error */ - REG_UPDATE_SEQ(AUXN_IMPCAL, AUXN_CALOUT_ERROR_AK, - 1, - 0); - - REG_UPDATE_SEQ(AUXP_IMPCAL, AUXP_CALOUT_ERROR_AK, - 1, - 0); - - /* force_default_calibrate */ - REG_UPDATE_1BY1_2(AUXN_IMPCAL, - AUXN_IMPCAL_ENABLE, 1, - AUXN_IMPCAL_OVERRIDE_ENABLE, 0); - - /* bug? why AUXN update EN and OVERRIDE_EN 1 by 1 while AUX P toggles OVERRIDE? */ - - REG_UPDATE_SEQ(AUXP_IMPCAL, AUXP_IMPCAL_OVERRIDE_ENABLE, - 1, - 0); - } - /* set the delay and the number of bytes to write */ - - /* The length include - * the 4 bit header and the 20 bit address - * (that is 3 byte). - * If the requested length is non zero this means - * an addition byte specifying the length is required. */ - - length = request->length ? 4 : 3; - if (is_write) - length += request->length; - - REG_UPDATE_2(AUX_SW_CONTROL, - AUX_SW_START_DELAY, request->delay, - AUX_SW_WR_BYTES, length); - - /* program action and address and payload data (if 'is_write') */ - value = REG_UPDATE_4(AUX_SW_DATA, - AUX_SW_INDEX, 0, - AUX_SW_DATA_RW, 0, - AUX_SW_AUTOINCREMENT_DISABLE, 1, - AUX_SW_DATA, COMPOSE_AUX_SW_DATA_16_20(request->action, request->address)); - - value = REG_SET_2(AUX_SW_DATA, value, - AUX_SW_AUTOINCREMENT_DISABLE, 0, - AUX_SW_DATA, COMPOSE_AUX_SW_DATA_8_15(request->address)); - - value = REG_SET(AUX_SW_DATA, value, - AUX_SW_DATA, COMPOSE_AUX_SW_DATA_0_7(request->address)); - - if (request->length) { - value = REG_SET(AUX_SW_DATA, value, - AUX_SW_DATA, request->length - 1); - } - - if (is_write) { - /* Load the HW buffer with the Data to be sent. - * This is relevant for write operation. - * For read, the data recived data will be - * processed in process_channel_reply(). */ - uint32_t i = 0; - - while (i < request->length) { - value = REG_SET(AUX_SW_DATA, value, - AUX_SW_DATA, request->data[i]); - - ++i; - } - } - - REG_UPDATE(AUX_INTERRUPT_CONTROL, AUX_SW_DONE_ACK, 1); - REG_WAIT(AUX_SW_STATUS, AUX_SW_DONE, 0, - 10, aux110->timeout_period/10); - REG_UPDATE(AUX_SW_CONTROL, AUX_SW_GO, 1); - EVENT_LOG_AUX_REQ(engine->base.ddc->pin_data->en, EVENT_LOG_AUX_ORIGIN_NATIVE, - request->action, request->address, request->length, request->data); -} - -static int read_channel_reply(struct aux_engine *engine, uint32_t size, - uint8_t *buffer, uint8_t *reply_result, - uint32_t *sw_status) -{ - struct aux_engine_dce110 *aux110 = FROM_AUX_ENGINE(engine); - uint32_t bytes_replied; - uint32_t reply_result_32; - - *sw_status = REG_GET(AUX_SW_STATUS, AUX_SW_REPLY_BYTE_COUNT, - &bytes_replied); - - /* In case HPD is LOW, exit AUX transaction */ - if ((*sw_status & AUX_SW_STATUS__AUX_SW_HPD_DISCON_MASK)) - return -1; - - /* Need at least the status byte */ - if (!bytes_replied) - return -1; - - REG_UPDATE_1BY1_3(AUX_SW_DATA, - AUX_SW_INDEX, 0, - AUX_SW_AUTOINCREMENT_DISABLE, 1, - AUX_SW_DATA_RW, 1); - - REG_GET(AUX_SW_DATA, AUX_SW_DATA, &reply_result_32); - reply_result_32 = reply_result_32 >> 4; - *reply_result = (uint8_t)reply_result_32; - - if (reply_result_32 == 0) { /* ACK */ - uint32_t i = 0; - - /* First byte was already used to get the command status */ - --bytes_replied; - - /* Do not overflow buffer */ - if (bytes_replied > size) - return -1; - - while (i < bytes_replied) { - uint32_t aux_sw_data_val; - - REG_GET(AUX_SW_DATA, AUX_SW_DATA, &aux_sw_data_val); - buffer[i] = aux_sw_data_val; - ++i; - } - - return i; - } - - return 0; -} - -static void process_channel_reply( - struct aux_engine *engine, - struct aux_reply_transaction_data *reply) -{ - int bytes_replied; - uint8_t reply_result; - uint32_t sw_status; - - bytes_replied = read_channel_reply(engine, reply->length, reply->data, - &reply_result, &sw_status); - EVENT_LOG_AUX_REP(engine->base.ddc->pin_data->en, - EVENT_LOG_AUX_ORIGIN_NATIVE, reply_result, - bytes_replied, reply->data); - - /* in case HPD is LOW, exit AUX transaction */ - if ((sw_status & AUX_SW_STATUS__AUX_SW_HPD_DISCON_MASK)) { - reply->status = AUX_TRANSACTION_REPLY_HPD_DISCON; - return; - } - - if (bytes_replied < 0) { - /* Need to handle an error case... - * Hopefully, upper layer function won't call this function if - * the number of bytes in the reply was 0, because there was - * surely an error that was asserted that should have been - * handled for hot plug case, this could happens - */ - if (!(sw_status & AUX_SW_STATUS__AUX_SW_HPD_DISCON_MASK)) { - reply->status = AUX_TRANSACTION_REPLY_INVALID; - ASSERT_CRITICAL(false); - return; - } - } else { - - switch (reply_result) { - case 0: /* ACK */ - reply->status = AUX_TRANSACTION_REPLY_AUX_ACK; - break; - case 1: /* NACK */ - reply->status = AUX_TRANSACTION_REPLY_AUX_NACK; - break; - case 2: /* DEFER */ - reply->status = AUX_TRANSACTION_REPLY_AUX_DEFER; - break; - case 4: /* AUX ACK / I2C NACK */ - reply->status = AUX_TRANSACTION_REPLY_I2C_NACK; - break; - case 8: /* AUX ACK / I2C DEFER */ - reply->status = AUX_TRANSACTION_REPLY_I2C_DEFER; - break; - default: - reply->status = AUX_TRANSACTION_REPLY_INVALID; - } - } -} - -static enum aux_channel_operation_result get_channel_status( - struct aux_engine *engine, - uint8_t *returned_bytes) -{ - struct aux_engine_dce110 *aux110 = FROM_AUX_ENGINE(engine); - - uint32_t value; - - if (returned_bytes == NULL) { - /*caller pass NULL pointer*/ - ASSERT_CRITICAL(false); - return AUX_CHANNEL_OPERATION_FAILED_REASON_UNKNOWN; - } - *returned_bytes = 0; - - /* poll to make sure that SW_DONE is asserted */ - value = REG_WAIT(AUX_SW_STATUS, AUX_SW_DONE, 1, - 10, aux110->timeout_period/10); - - /* in case HPD is LOW, exit AUX transaction */ - if ((value & AUX_SW_STATUS__AUX_SW_HPD_DISCON_MASK)) - return AUX_CHANNEL_OPERATION_FAILED_HPD_DISCON; - - /* Note that the following bits are set in 'status.bits' - * during CTS 4.2.1.2 (FW 3.3.1): - * AUX_SW_RX_MIN_COUNT_VIOL, AUX_SW_RX_INVALID_STOP, - * AUX_SW_RX_RECV_NO_DET, AUX_SW_RX_RECV_INVALID_H. - * - * AUX_SW_RX_MIN_COUNT_VIOL is an internal, - * HW debugging bit and should be ignored. */ - if (value & AUX_SW_STATUS__AUX_SW_DONE_MASK) { - if ((value & AUX_SW_STATUS__AUX_SW_RX_TIMEOUT_STATE_MASK) || - (value & AUX_SW_STATUS__AUX_SW_RX_TIMEOUT_MASK)) - return AUX_CHANNEL_OPERATION_FAILED_TIMEOUT; - - else if ((value & AUX_SW_STATUS__AUX_SW_RX_INVALID_STOP_MASK) || - (value & AUX_SW_STATUS__AUX_SW_RX_RECV_NO_DET_MASK) || - (value & - AUX_SW_STATUS__AUX_SW_RX_RECV_INVALID_H_MASK) || - (value & AUX_SW_STATUS__AUX_SW_RX_RECV_INVALID_L_MASK)) - return AUX_CHANNEL_OPERATION_FAILED_INVALID_REPLY; - - *returned_bytes = get_reg_field_value(value, - AUX_SW_STATUS, - AUX_SW_REPLY_BYTE_COUNT); - - if (*returned_bytes == 0) - return - AUX_CHANNEL_OPERATION_FAILED_INVALID_REPLY; - else { - *returned_bytes -= 1; - return AUX_CHANNEL_OPERATION_SUCCEEDED; - } - } else { - /*time_elapsed >= aux_engine->timeout_period - * AUX_SW_STATUS__AUX_SW_HPD_DISCON = at this point - */ - ASSERT_CRITICAL(false); - return AUX_CHANNEL_OPERATION_FAILED_TIMEOUT; - } -} - -static const struct aux_engine_funcs aux_engine_funcs = { - .destroy = destroy, - .acquire_engine = acquire_engine, - .submit_channel_request = submit_channel_request, - .process_channel_reply = process_channel_reply, - .read_channel_reply = read_channel_reply, - .get_channel_status = get_channel_status, - .is_engine_available = is_engine_available, -}; - -static const struct engine_funcs engine_funcs = { - .release_engine = release_engine, - .submit_request = dal_aux_engine_submit_request, - .get_engine_type = dal_aux_engine_get_engine_type, - .acquire = dal_aux_engine_acquire, -}; - -static void construct( - struct aux_engine_dce110 *engine, - const struct aux_engine_dce110_init_data *aux_init_data) -{ - dal_aux_engine_construct(&engine->base, aux_init_data->ctx); - engine->base.base.funcs = &engine_funcs; - engine->base.funcs = &aux_engine_funcs; - - engine->timeout_period = aux_init_data->timeout_period; - engine->regs = aux_init_data->regs; -} - -static void destruct( - struct aux_engine_dce110 *engine) -{ - dal_aux_engine_destruct(&engine->base); -} - -struct aux_engine *dal_aux_engine_dce110_create( - const struct aux_engine_dce110_init_data *aux_init_data) -{ - struct aux_engine_dce110 *engine; - - if (!aux_init_data) { - ASSERT_CRITICAL(false); - return NULL; - } - - engine = kzalloc(sizeof(*engine), GFP_KERNEL); - - if (!engine) { - ASSERT_CRITICAL(false); - return NULL; - } - - construct(engine, aux_init_data); - return &engine->base; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.h b/drivers/gpu/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.h deleted file mode 100644 index 85ee82162590..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.h +++ /dev/null @@ -1,78 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_AUX_ENGINE_DCE110_H__ -#define __DAL_AUX_ENGINE_DCE110_H__ - -#include "../aux_engine.h" - -#define AUX_COMMON_REG_LIST(id)\ - SRI(AUX_CONTROL, DP_AUX, id), \ - SRI(AUX_ARB_CONTROL, DP_AUX, id), \ - SRI(AUX_SW_DATA, DP_AUX, id), \ - SRI(AUX_SW_CONTROL, DP_AUX, id), \ - SRI(AUX_INTERRUPT_CONTROL, DP_AUX, id), \ - SRI(AUX_SW_STATUS, DP_AUX, id), \ - SR(AUXN_IMPCAL), \ - SR(AUXP_IMPCAL) - -struct dce110_aux_registers { - uint32_t AUX_CONTROL; - uint32_t AUX_ARB_CONTROL; - uint32_t AUX_SW_DATA; - uint32_t AUX_SW_CONTROL; - uint32_t AUX_INTERRUPT_CONTROL; - uint32_t AUX_SW_STATUS; - uint32_t AUXN_IMPCAL; - uint32_t AUXP_IMPCAL; - - uint32_t AUX_RESET_MASK; -}; - -struct aux_engine_dce110 { - struct aux_engine base; - const struct dce110_aux_registers *regs; - struct { - uint32_t aux_control; - uint32_t aux_arb_control; - uint32_t aux_sw_data; - uint32_t aux_sw_control; - uint32_t aux_interrupt_control; - uint32_t aux_sw_status; - } addr; - uint32_t timeout_period; -}; - -struct aux_engine_dce110_init_data { - uint32_t engine_id; - uint32_t timeout_period; - struct dc_context *ctx; - const struct dce110_aux_registers *regs; -}; - -struct aux_engine *dal_aux_engine_dce110_create( - const struct aux_engine_dce110_init_data *aux_init_data); - -#endif diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2c_hw_engine_dce110.c b/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2c_hw_engine_dce110.c deleted file mode 100644 index 9cbe1a7a6bcb..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2c_hw_engine_dce110.c +++ /dev/null @@ -1,574 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" -#include "include/logger_interface.h" -/* - * Pre-requisites: headers required by header of this unit - */ - -#include "include/i2caux_interface.h" -#include "../engine.h" -#include "../i2c_engine.h" -#include "../i2c_hw_engine.h" -#include "../i2c_generic_hw_engine.h" -/* - * Header of this unit - */ - -#include "i2c_hw_engine_dce110.h" - -/* - * Post-requisites: headers required by this unit - */ -#include "reg_helper.h" - -/* - * This unit - */ -#define DC_LOGGER \ - hw_engine->base.base.base.ctx->logger - -enum dc_i2c_status { - DC_I2C_STATUS__DC_I2C_STATUS_IDLE, - DC_I2C_STATUS__DC_I2C_STATUS_USED_BY_SW, - DC_I2C_STATUS__DC_I2C_STATUS_USED_BY_HW -}; - -enum dc_i2c_arbitration { - DC_I2C_ARBITRATION__DC_I2C_SW_PRIORITY_NORMAL, - DC_I2C_ARBITRATION__DC_I2C_SW_PRIORITY_HIGH -}; - - - -/* - * @brief - * Cast pointer to 'struct i2c_hw_engine *' - * to pointer 'struct i2c_hw_engine_dce110 *' - */ -#define FROM_I2C_HW_ENGINE(ptr) \ - container_of((ptr), struct i2c_hw_engine_dce110, base) -/* - * @brief - * Cast pointer to 'struct i2c_engine *' - * to pointer to 'struct i2c_hw_engine_dce110 *' - */ -#define FROM_I2C_ENGINE(ptr) \ - FROM_I2C_HW_ENGINE(container_of((ptr), struct i2c_hw_engine, base)) - -/* - * @brief - * Cast pointer to 'struct engine *' - * to 'pointer to struct i2c_hw_engine_dce110 *' - */ -#define FROM_ENGINE(ptr) \ - FROM_I2C_ENGINE(container_of((ptr), struct i2c_engine, base)) - -#define CTX \ - hw_engine->base.base.base.ctx - -#define REG(reg_name)\ - (hw_engine->regs->reg_name) - -#undef FN -#define FN(reg_name, field_name) \ - hw_engine->i2c_shift->field_name, hw_engine->i2c_mask->field_name - -#include "reg_helper.h" - -static void disable_i2c_hw_engine( - struct i2c_hw_engine_dce110 *hw_engine) -{ - REG_UPDATE_N(SETUP, 1, FN(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_ENABLE), 0); -} - -static void release_engine( - struct engine *engine) -{ - struct i2c_hw_engine_dce110 *hw_engine = FROM_ENGINE(engine); - - struct i2c_engine *base = NULL; - bool safe_to_reset; - - base = &hw_engine->base.base; - - /* Restore original HW engine speed */ - - base->funcs->set_speed(base, hw_engine->base.original_speed); - - /* Release I2C */ - REG_UPDATE(DC_I2C_ARBITRATION, DC_I2C_SW_DONE_USING_I2C_REG, 1); - - /* Reset HW engine */ - { - uint32_t i2c_sw_status = 0; - REG_GET(DC_I2C_SW_STATUS, DC_I2C_SW_STATUS, &i2c_sw_status); - /* if used by SW, safe to reset */ - safe_to_reset = (i2c_sw_status == 1); - } - - if (safe_to_reset) - REG_UPDATE_2( - DC_I2C_CONTROL, - DC_I2C_SOFT_RESET, 1, - DC_I2C_SW_STATUS_RESET, 1); - else - REG_UPDATE(DC_I2C_CONTROL, DC_I2C_SW_STATUS_RESET, 1); - - /* HW I2c engine - clock gating feature */ - if (!hw_engine->engine_keep_power_up_count) - disable_i2c_hw_engine(hw_engine); -} - -static bool setup_engine( - struct i2c_engine *i2c_engine) -{ - struct i2c_hw_engine_dce110 *hw_engine = FROM_I2C_ENGINE(i2c_engine); - uint32_t i2c_setup_limit = I2C_SETUP_TIME_LIMIT_DCE; - uint32_t reset_length = 0; - - if (hw_engine->base.base.setup_limit != 0) - i2c_setup_limit = hw_engine->base.base.setup_limit; - - /* Program pin select */ - REG_UPDATE_6( - DC_I2C_CONTROL, - DC_I2C_GO, 0, - DC_I2C_SOFT_RESET, 0, - DC_I2C_SEND_RESET, 0, - DC_I2C_SW_STATUS_RESET, 1, - DC_I2C_TRANSACTION_COUNT, 0, - DC_I2C_DDC_SELECT, hw_engine->engine_id); - - /* Program time limit */ - if (hw_engine->base.base.send_reset_length == 0) { - /*pre-dcn*/ - REG_UPDATE_N( - SETUP, 2, - FN(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_TIME_LIMIT), i2c_setup_limit, - FN(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_ENABLE), 1); - } else { - reset_length = hw_engine->base.base.send_reset_length; - } - /* Program HW priority - * set to High - interrupt software I2C at any time - * Enable restart of SW I2C that was interrupted by HW - * disable queuing of software while I2C is in use by HW */ - REG_UPDATE_2( - DC_I2C_ARBITRATION, - DC_I2C_NO_QUEUED_SW_GO, 0, - DC_I2C_SW_PRIORITY, DC_I2C_ARBITRATION__DC_I2C_SW_PRIORITY_NORMAL); - - return true; -} - -static uint32_t get_speed( - const struct i2c_engine *i2c_engine) -{ - const struct i2c_hw_engine_dce110 *hw_engine = FROM_I2C_ENGINE(i2c_engine); - uint32_t pre_scale = 0; - - REG_GET(SPEED, DC_I2C_DDC1_PRESCALE, &pre_scale); - - /* [anaumov] it seems following is unnecessary */ - /*ASSERT(value.bits.DC_I2C_DDC1_PRESCALE);*/ - return pre_scale ? - hw_engine->reference_frequency / pre_scale : - hw_engine->base.default_speed; -} - -static void set_speed( - struct i2c_engine *i2c_engine, - uint32_t speed) -{ - struct i2c_hw_engine_dce110 *hw_engine = FROM_I2C_ENGINE(i2c_engine); - - if (speed) { - if (hw_engine->i2c_mask->DC_I2C_DDC1_START_STOP_TIMING_CNTL) - REG_UPDATE_N( - SPEED, 3, - FN(DC_I2C_DDC1_SPEED, DC_I2C_DDC1_PRESCALE), hw_engine->reference_frequency / speed, - FN(DC_I2C_DDC1_SPEED, DC_I2C_DDC1_THRESHOLD), 2, - FN(DC_I2C_DDC1_SPEED, DC_I2C_DDC1_START_STOP_TIMING_CNTL), speed > 50 ? 2:1); - else - REG_UPDATE_N( - SPEED, 2, - FN(DC_I2C_DDC1_SPEED, DC_I2C_DDC1_PRESCALE), hw_engine->reference_frequency / speed, - FN(DC_I2C_DDC1_SPEED, DC_I2C_DDC1_THRESHOLD), 2); - } -} - -static inline void reset_hw_engine(struct engine *engine) -{ - struct i2c_hw_engine_dce110 *hw_engine = FROM_ENGINE(engine); - - REG_UPDATE_2( - DC_I2C_CONTROL, - DC_I2C_SW_STATUS_RESET, 1, - DC_I2C_SW_STATUS_RESET, 1); -} - -static bool is_hw_busy(struct engine *engine) -{ - struct i2c_hw_engine_dce110 *hw_engine = FROM_ENGINE(engine); - uint32_t i2c_sw_status = 0; - - REG_GET(DC_I2C_SW_STATUS, DC_I2C_SW_STATUS, &i2c_sw_status); - if (i2c_sw_status == DC_I2C_STATUS__DC_I2C_STATUS_IDLE) - return false; - - reset_hw_engine(engine); - - REG_GET(DC_I2C_SW_STATUS, DC_I2C_SW_STATUS, &i2c_sw_status); - return i2c_sw_status != DC_I2C_STATUS__DC_I2C_STATUS_IDLE; -} - - -#define STOP_TRANS_PREDICAT \ - ((hw_engine->transaction_count == 3) || \ - (request->action == I2CAUX_TRANSACTION_ACTION_I2C_WRITE) || \ - (request->action & I2CAUX_TRANSACTION_ACTION_I2C_READ)) - -#define SET_I2C_TRANSACTION(id) \ - do { \ - REG_UPDATE_N(DC_I2C_TRANSACTION##id, 5, \ - FN(DC_I2C_TRANSACTION0, DC_I2C_STOP_ON_NACK0), 1, \ - FN(DC_I2C_TRANSACTION0, DC_I2C_START0), 1, \ - FN(DC_I2C_TRANSACTION0, DC_I2C_STOP0), STOP_TRANS_PREDICAT ? 1:0, \ - FN(DC_I2C_TRANSACTION0, DC_I2C_RW0), (0 != (request->action & I2CAUX_TRANSACTION_ACTION_I2C_READ)), \ - FN(DC_I2C_TRANSACTION0, DC_I2C_COUNT0), length); \ - if (STOP_TRANS_PREDICAT) \ - last_transaction = true; \ - } while (false) - - -static bool process_transaction( - struct i2c_hw_engine_dce110 *hw_engine, - struct i2c_request_transaction_data *request) -{ - uint32_t length = request->length; - uint8_t *buffer = request->data; - uint32_t value = 0; - - bool last_transaction = false; - - struct dc_context *ctx = NULL; - - ctx = hw_engine->base.base.base.ctx; - - - - switch (hw_engine->transaction_count) { - case 0: - SET_I2C_TRANSACTION(0); - break; - case 1: - SET_I2C_TRANSACTION(1); - break; - case 2: - SET_I2C_TRANSACTION(2); - break; - case 3: - SET_I2C_TRANSACTION(3); - break; - default: - /* TODO Warning ? */ - break; - } - - - /* Write the I2C address and I2C data - * into the hardware circular buffer, one byte per entry. - * As an example, the 7-bit I2C slave address for CRT monitor - * for reading DDC/EDID information is 0b1010001. - * For an I2C send operation, the LSB must be programmed to 0; - * for I2C receive operation, the LSB must be programmed to 1. */ - if (hw_engine->transaction_count == 0) { - value = REG_SET_4(DC_I2C_DATA, 0, - DC_I2C_DATA_RW, false, - DC_I2C_DATA, request->address, - DC_I2C_INDEX, 0, - DC_I2C_INDEX_WRITE, 1); - hw_engine->buffer_used_write = 0; - } else - value = REG_SET_2(DC_I2C_DATA, 0, - DC_I2C_DATA_RW, false, - DC_I2C_DATA, request->address); - - hw_engine->buffer_used_write++; - - if (!(request->action & I2CAUX_TRANSACTION_ACTION_I2C_READ)) { - while (length) { - REG_SET_2(DC_I2C_DATA, value, - DC_I2C_INDEX_WRITE, 0, - DC_I2C_DATA, *buffer++); - hw_engine->buffer_used_write++; - --length; - } - } - - ++hw_engine->transaction_count; - hw_engine->buffer_used_bytes += length + 1; - - return last_transaction; -} - -static void execute_transaction( - struct i2c_hw_engine_dce110 *hw_engine) -{ - REG_UPDATE_N(SETUP, 5, - FN(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_DATA_DRIVE_EN), 0, - FN(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_CLK_DRIVE_EN), 0, - FN(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_DATA_DRIVE_SEL), 0, - FN(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_INTRA_TRANSACTION_DELAY), 0, - FN(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_INTRA_BYTE_DELAY), 0); - - - REG_UPDATE_5(DC_I2C_CONTROL, - DC_I2C_SOFT_RESET, 0, - DC_I2C_SW_STATUS_RESET, 0, - DC_I2C_SEND_RESET, 0, - DC_I2C_GO, 0, - DC_I2C_TRANSACTION_COUNT, hw_engine->transaction_count - 1); - - /* start I2C transfer */ - REG_UPDATE(DC_I2C_CONTROL, DC_I2C_GO, 1); - - /* all transactions were executed and HW buffer became empty - * (even though it actually happens when status becomes DONE) */ - hw_engine->transaction_count = 0; - hw_engine->buffer_used_bytes = 0; -} - -static void submit_channel_request( - struct i2c_engine *engine, - struct i2c_request_transaction_data *request) -{ - request->status = I2C_CHANNEL_OPERATION_SUCCEEDED; - - if (!process_transaction(FROM_I2C_ENGINE(engine), request)) - return; - - if (is_hw_busy(&engine->base)) { - request->status = I2C_CHANNEL_OPERATION_ENGINE_BUSY; - return; - } - - execute_transaction(FROM_I2C_ENGINE(engine)); -} - -static void process_channel_reply( - struct i2c_engine *engine, - struct i2c_reply_transaction_data *reply) -{ - uint32_t length = reply->length; - uint8_t *buffer = reply->data; - - struct i2c_hw_engine_dce110 *hw_engine = - FROM_I2C_ENGINE(engine); - - - REG_SET_3(DC_I2C_DATA, 0, - DC_I2C_INDEX, hw_engine->buffer_used_write, - DC_I2C_DATA_RW, 1, - DC_I2C_INDEX_WRITE, 1); - - while (length) { - /* after reading the status, - * if the I2C operation executed successfully - * (i.e. DC_I2C_STATUS_DONE = 1) then the I2C controller - * should read data bytes from I2C circular data buffer */ - - uint32_t i2c_data; - - REG_GET(DC_I2C_DATA, DC_I2C_DATA, &i2c_data); - *buffer++ = i2c_data; - - --length; - } -} - -static enum i2c_channel_operation_result get_channel_status( - struct i2c_engine *i2c_engine, - uint8_t *returned_bytes) -{ - uint32_t i2c_sw_status = 0; - struct i2c_hw_engine_dce110 *hw_engine = FROM_I2C_ENGINE(i2c_engine); - uint32_t value = - REG_GET(DC_I2C_SW_STATUS, DC_I2C_SW_STATUS, &i2c_sw_status); - - if (i2c_sw_status == DC_I2C_STATUS__DC_I2C_STATUS_USED_BY_SW) - return I2C_CHANNEL_OPERATION_ENGINE_BUSY; - else if (value & hw_engine->i2c_mask->DC_I2C_SW_STOPPED_ON_NACK) - return I2C_CHANNEL_OPERATION_NO_RESPONSE; - else if (value & hw_engine->i2c_mask->DC_I2C_SW_TIMEOUT) - return I2C_CHANNEL_OPERATION_TIMEOUT; - else if (value & hw_engine->i2c_mask->DC_I2C_SW_ABORTED) - return I2C_CHANNEL_OPERATION_FAILED; - else if (value & hw_engine->i2c_mask->DC_I2C_SW_DONE) - return I2C_CHANNEL_OPERATION_SUCCEEDED; - - /* - * this is the case when HW used for communication, I2C_SW_STATUS - * could be zero - */ - return I2C_CHANNEL_OPERATION_SUCCEEDED; -} - -static uint32_t get_hw_buffer_available_size( - const struct i2c_hw_engine *engine) -{ - return I2C_HW_BUFFER_SIZE - - FROM_I2C_HW_ENGINE(engine)->buffer_used_bytes; -} - -static uint32_t get_transaction_timeout( - const struct i2c_hw_engine *engine, - uint32_t length) -{ - uint32_t speed = engine->base.funcs->get_speed(&engine->base); - - uint32_t period_timeout; - uint32_t num_of_clock_stretches; - - if (!speed) - return 0; - - period_timeout = (1000 * TRANSACTION_TIMEOUT_IN_I2C_CLOCKS) / speed; - - num_of_clock_stretches = 1 + (length << 3) + 1; - num_of_clock_stretches += - (FROM_I2C_HW_ENGINE(engine)->buffer_used_bytes << 3) + - (FROM_I2C_HW_ENGINE(engine)->transaction_count << 1); - - return period_timeout * num_of_clock_stretches; -} - -static void destroy( - struct i2c_engine **i2c_engine) -{ - struct i2c_hw_engine_dce110 *engine_dce110 = - FROM_I2C_ENGINE(*i2c_engine); - - dal_i2c_hw_engine_destruct(&engine_dce110->base); - - kfree(engine_dce110); - - *i2c_engine = NULL; -} - -static const struct i2c_engine_funcs i2c_engine_funcs = { - .destroy = destroy, - .get_speed = get_speed, - .set_speed = set_speed, - .setup_engine = setup_engine, - .submit_channel_request = submit_channel_request, - .process_channel_reply = process_channel_reply, - .get_channel_status = get_channel_status, - .acquire_engine = dal_i2c_hw_engine_acquire_engine, -}; - -static const struct engine_funcs engine_funcs = { - .release_engine = release_engine, - .get_engine_type = dal_i2c_hw_engine_get_engine_type, - .acquire = dal_i2c_engine_acquire, - .submit_request = dal_i2c_hw_engine_submit_request, -}; - -static const struct i2c_hw_engine_funcs i2c_hw_engine_funcs = { - .get_hw_buffer_available_size = get_hw_buffer_available_size, - .get_transaction_timeout = get_transaction_timeout, - .wait_on_operation_result = dal_i2c_hw_engine_wait_on_operation_result, -}; - -static void construct( - struct i2c_hw_engine_dce110 *hw_engine, - const struct i2c_hw_engine_dce110_create_arg *arg) -{ - uint32_t xtal_ref_div = 0; - - dal_i2c_hw_engine_construct(&hw_engine->base, arg->ctx); - - hw_engine->base.base.base.funcs = &engine_funcs; - hw_engine->base.base.funcs = &i2c_engine_funcs; - hw_engine->base.funcs = &i2c_hw_engine_funcs; - hw_engine->base.default_speed = arg->default_speed; - - hw_engine->regs = arg->regs; - hw_engine->i2c_shift = arg->i2c_shift; - hw_engine->i2c_mask = arg->i2c_mask; - - hw_engine->engine_id = arg->engine_id; - - hw_engine->buffer_used_bytes = 0; - hw_engine->transaction_count = 0; - hw_engine->engine_keep_power_up_count = 1; - - - REG_GET(MICROSECOND_TIME_BASE_DIV, XTAL_REF_DIV, &xtal_ref_div); - - if (xtal_ref_div == 0) { - DC_LOG_WARNING("Invalid base timer divider [%s]\n", - __func__); - xtal_ref_div = 2; - } - - /*Calculating Reference Clock by divding original frequency by - * XTAL_REF_DIV. - * At upper level, uint32_t reference_frequency = - * dal_i2caux_get_reference_clock(as) >> 1 - * which already divided by 2. So we need x2 to get original - * reference clock from ppll_info - */ - hw_engine->reference_frequency = - (arg->reference_frequency * 2) / xtal_ref_div; -} - -struct i2c_engine *dal_i2c_hw_engine_dce110_create( - const struct i2c_hw_engine_dce110_create_arg *arg) -{ - struct i2c_hw_engine_dce110 *engine_dce10; - - if (!arg) { - ASSERT_CRITICAL(false); - return NULL; - } - if (!arg->reference_frequency) { - ASSERT_CRITICAL(false); - return NULL; - } - - engine_dce10 = kzalloc(sizeof(struct i2c_hw_engine_dce110), - GFP_KERNEL); - - if (!engine_dce10) { - ASSERT_CRITICAL(false); - return NULL; - } - - construct(engine_dce10, arg); - return &engine_dce10->base.base; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2c_hw_engine_dce110.h b/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2c_hw_engine_dce110.h deleted file mode 100644 index fea2946906ed..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2c_hw_engine_dce110.h +++ /dev/null @@ -1,218 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_HW_ENGINE_DCE110_H__ -#define __DAL_I2C_HW_ENGINE_DCE110_H__ - -#define I2C_HW_ENGINE_COMMON_REG_LIST(id)\ - SRI(SETUP, DC_I2C_DDC, id),\ - SRI(SPEED, DC_I2C_DDC, id),\ - SR(DC_I2C_ARBITRATION),\ - SR(DC_I2C_CONTROL),\ - SR(DC_I2C_SW_STATUS),\ - SR(DC_I2C_TRANSACTION0),\ - SR(DC_I2C_TRANSACTION1),\ - SR(DC_I2C_TRANSACTION2),\ - SR(DC_I2C_TRANSACTION3),\ - SR(DC_I2C_DATA),\ - SR(MICROSECOND_TIME_BASE_DIV) - -#define I2C_SF(reg_name, field_name, post_fix)\ - .field_name = reg_name ## __ ## field_name ## post_fix - -#define I2C_COMMON_MASK_SH_LIST_DCE_COMMON_BASE(mask_sh)\ - I2C_SF(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_ENABLE, mask_sh),\ - I2C_SF(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_TIME_LIMIT, mask_sh),\ - I2C_SF(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_DATA_DRIVE_EN, mask_sh),\ - I2C_SF(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_CLK_DRIVE_EN, mask_sh),\ - I2C_SF(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_DATA_DRIVE_SEL, mask_sh),\ - I2C_SF(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_INTRA_TRANSACTION_DELAY, mask_sh),\ - I2C_SF(DC_I2C_DDC1_SETUP, DC_I2C_DDC1_INTRA_BYTE_DELAY, mask_sh),\ - I2C_SF(DC_I2C_ARBITRATION, DC_I2C_SW_DONE_USING_I2C_REG, mask_sh),\ - I2C_SF(DC_I2C_ARBITRATION, DC_I2C_NO_QUEUED_SW_GO, mask_sh),\ - I2C_SF(DC_I2C_ARBITRATION, DC_I2C_SW_PRIORITY, mask_sh),\ - I2C_SF(DC_I2C_CONTROL, DC_I2C_SOFT_RESET, mask_sh),\ - I2C_SF(DC_I2C_CONTROL, DC_I2C_SW_STATUS_RESET, mask_sh),\ - I2C_SF(DC_I2C_CONTROL, DC_I2C_GO, mask_sh),\ - I2C_SF(DC_I2C_CONTROL, DC_I2C_SEND_RESET, mask_sh),\ - I2C_SF(DC_I2C_CONTROL, DC_I2C_TRANSACTION_COUNT, mask_sh),\ - I2C_SF(DC_I2C_CONTROL, DC_I2C_DDC_SELECT, mask_sh),\ - I2C_SF(DC_I2C_DDC1_SPEED, DC_I2C_DDC1_PRESCALE, mask_sh),\ - I2C_SF(DC_I2C_DDC1_SPEED, DC_I2C_DDC1_THRESHOLD, mask_sh),\ - I2C_SF(DC_I2C_SW_STATUS, DC_I2C_SW_STOPPED_ON_NACK, mask_sh),\ - I2C_SF(DC_I2C_SW_STATUS, DC_I2C_SW_TIMEOUT, mask_sh),\ - I2C_SF(DC_I2C_SW_STATUS, DC_I2C_SW_ABORTED, mask_sh),\ - I2C_SF(DC_I2C_SW_STATUS, DC_I2C_SW_DONE, mask_sh),\ - I2C_SF(DC_I2C_SW_STATUS, DC_I2C_SW_STATUS, mask_sh),\ - I2C_SF(DC_I2C_TRANSACTION0, DC_I2C_STOP_ON_NACK0, mask_sh),\ - I2C_SF(DC_I2C_TRANSACTION0, DC_I2C_START0, mask_sh),\ - I2C_SF(DC_I2C_TRANSACTION0, DC_I2C_RW0, mask_sh),\ - I2C_SF(DC_I2C_TRANSACTION0, DC_I2C_STOP0, mask_sh),\ - I2C_SF(DC_I2C_TRANSACTION0, DC_I2C_COUNT0, mask_sh),\ - I2C_SF(DC_I2C_DATA, DC_I2C_DATA_RW, mask_sh),\ - I2C_SF(DC_I2C_DATA, DC_I2C_DATA, mask_sh),\ - I2C_SF(DC_I2C_DATA, DC_I2C_INDEX, mask_sh),\ - I2C_SF(DC_I2C_DATA, DC_I2C_INDEX_WRITE, mask_sh),\ - I2C_SF(MICROSECOND_TIME_BASE_DIV, XTAL_REF_DIV, mask_sh) - -#define I2C_COMMON_MASK_SH_LIST_DCE100(mask_sh)\ - I2C_COMMON_MASK_SH_LIST_DCE_COMMON_BASE(mask_sh) - -#define I2C_COMMON_MASK_SH_LIST_DCE110(mask_sh)\ - I2C_COMMON_MASK_SH_LIST_DCE_COMMON_BASE(mask_sh),\ - I2C_SF(DC_I2C_DDC1_SPEED, DC_I2C_DDC1_START_STOP_TIMING_CNTL, mask_sh) - -struct dce110_i2c_hw_engine_shift { - uint8_t DC_I2C_DDC1_ENABLE; - uint8_t DC_I2C_DDC1_TIME_LIMIT; - uint8_t DC_I2C_DDC1_DATA_DRIVE_EN; - uint8_t DC_I2C_DDC1_CLK_DRIVE_EN; - uint8_t DC_I2C_DDC1_DATA_DRIVE_SEL; - uint8_t DC_I2C_DDC1_INTRA_TRANSACTION_DELAY; - uint8_t DC_I2C_DDC1_INTRA_BYTE_DELAY; - uint8_t DC_I2C_SW_DONE_USING_I2C_REG; - uint8_t DC_I2C_NO_QUEUED_SW_GO; - uint8_t DC_I2C_SW_PRIORITY; - uint8_t DC_I2C_SOFT_RESET; - uint8_t DC_I2C_SW_STATUS_RESET; - uint8_t DC_I2C_GO; - uint8_t DC_I2C_SEND_RESET; - uint8_t DC_I2C_TRANSACTION_COUNT; - uint8_t DC_I2C_DDC_SELECT; - uint8_t DC_I2C_DDC1_PRESCALE; - uint8_t DC_I2C_DDC1_THRESHOLD; - uint8_t DC_I2C_DDC1_START_STOP_TIMING_CNTL; - uint8_t DC_I2C_SW_STOPPED_ON_NACK; - uint8_t DC_I2C_SW_TIMEOUT; - uint8_t DC_I2C_SW_ABORTED; - uint8_t DC_I2C_SW_DONE; - uint8_t DC_I2C_SW_STATUS; - uint8_t DC_I2C_STOP_ON_NACK0; - uint8_t DC_I2C_START0; - uint8_t DC_I2C_RW0; - uint8_t DC_I2C_STOP0; - uint8_t DC_I2C_COUNT0; - uint8_t DC_I2C_DATA_RW; - uint8_t DC_I2C_DATA; - uint8_t DC_I2C_INDEX; - uint8_t DC_I2C_INDEX_WRITE; - uint8_t XTAL_REF_DIV; -}; - -struct dce110_i2c_hw_engine_mask { - uint32_t DC_I2C_DDC1_ENABLE; - uint32_t DC_I2C_DDC1_TIME_LIMIT; - uint32_t DC_I2C_DDC1_DATA_DRIVE_EN; - uint32_t DC_I2C_DDC1_CLK_DRIVE_EN; - uint32_t DC_I2C_DDC1_DATA_DRIVE_SEL; - uint32_t DC_I2C_DDC1_INTRA_TRANSACTION_DELAY; - uint32_t DC_I2C_DDC1_INTRA_BYTE_DELAY; - uint32_t DC_I2C_SW_DONE_USING_I2C_REG; - uint32_t DC_I2C_NO_QUEUED_SW_GO; - uint32_t DC_I2C_SW_PRIORITY; - uint32_t DC_I2C_SOFT_RESET; - uint32_t DC_I2C_SW_STATUS_RESET; - uint32_t DC_I2C_GO; - uint32_t DC_I2C_SEND_RESET; - uint32_t DC_I2C_TRANSACTION_COUNT; - uint32_t DC_I2C_DDC_SELECT; - uint32_t DC_I2C_DDC1_PRESCALE; - uint32_t DC_I2C_DDC1_THRESHOLD; - uint32_t DC_I2C_DDC1_START_STOP_TIMING_CNTL; - uint32_t DC_I2C_SW_STOPPED_ON_NACK; - uint32_t DC_I2C_SW_TIMEOUT; - uint32_t DC_I2C_SW_ABORTED; - uint32_t DC_I2C_SW_DONE; - uint32_t DC_I2C_SW_STATUS; - uint32_t DC_I2C_STOP_ON_NACK0; - uint32_t DC_I2C_START0; - uint32_t DC_I2C_RW0; - uint32_t DC_I2C_STOP0; - uint32_t DC_I2C_COUNT0; - uint32_t DC_I2C_DATA_RW; - uint32_t DC_I2C_DATA; - uint32_t DC_I2C_INDEX; - uint32_t DC_I2C_INDEX_WRITE; - uint32_t XTAL_REF_DIV; -}; - -struct dce110_i2c_hw_engine_registers { - uint32_t SETUP; - uint32_t SPEED; - uint32_t DC_I2C_ARBITRATION; - uint32_t DC_I2C_CONTROL; - uint32_t DC_I2C_SW_STATUS; - uint32_t DC_I2C_TRANSACTION0; - uint32_t DC_I2C_TRANSACTION1; - uint32_t DC_I2C_TRANSACTION2; - uint32_t DC_I2C_TRANSACTION3; - uint32_t DC_I2C_DATA; - uint32_t MICROSECOND_TIME_BASE_DIV; -}; - -struct i2c_hw_engine_dce110 { - struct i2c_hw_engine base; - const struct dce110_i2c_hw_engine_registers *regs; - const struct dce110_i2c_hw_engine_shift *i2c_shift; - const struct dce110_i2c_hw_engine_mask *i2c_mask; - struct { - uint32_t DC_I2C_DDCX_SETUP; - uint32_t DC_I2C_DDCX_SPEED; - } addr; - uint32_t engine_id; - /* expressed in kilohertz */ - uint32_t reference_frequency; - /* number of bytes currently used in HW buffer */ - uint32_t buffer_used_bytes; - /* number of bytes used for write transaction in HW buffer - * - this will be used as the index to read from*/ - uint32_t buffer_used_write; - /* number of pending transactions (before GO) */ - uint32_t transaction_count; - uint32_t engine_keep_power_up_count; - uint32_t i2_setup_time_limit; -}; - -struct i2c_hw_engine_dce110_create_arg { - uint32_t engine_id; - uint32_t reference_frequency; - uint32_t default_speed; - struct dc_context *ctx; - const struct dce110_i2c_hw_engine_registers *regs; - const struct dce110_i2c_hw_engine_shift *i2c_shift; - const struct dce110_i2c_hw_engine_mask *i2c_mask; -}; - -struct i2c_engine *dal_i2c_hw_engine_dce110_create( - const struct i2c_hw_engine_dce110_create_arg *arg); - -enum { - I2C_SETUP_TIME_LIMIT_DCE = 255, - I2C_SETUP_TIME_LIMIT_DCN = 3, - I2C_HW_BUFFER_SIZE = 538, - I2C_SEND_RESET_LENGTH_9 = 9, - I2C_SEND_RESET_LENGTH_10 = 10, -}; -#endif diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2c_sw_engine_dce110.c b/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2c_sw_engine_dce110.c deleted file mode 100644 index 3aa7f791e523..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2c_sw_engine_dce110.c +++ /dev/null @@ -1,160 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "../engine.h" -#include "../i2c_engine.h" -#include "../i2c_sw_engine.h" - -/* - * Header of this unit - */ - -#include "i2c_sw_engine_dce110.h" - -/* - * Post-requisites: headers required by this unit - */ - -/* - * This unit - */ - -/* - * @brief - * Cast 'struct i2c_sw_engine *' - * to 'struct i2c_sw_engine_dce110 *' - */ -#define FROM_I2C_SW_ENGINE(ptr) \ - container_of((ptr), struct i2c_sw_engine_dce110, base) -/* - * @brief - * Cast 'struct i2c_engine *' - * to 'struct i2c_sw_engine_dce80 *' - */ -#define FROM_I2C_ENGINE(ptr) \ - FROM_I2C_SW_ENGINE(container_of((ptr), struct i2c_sw_engine, base)) - -/* - * @brief - * Cast 'struct engine *' - * to 'struct i2c_sw_engine_dce80 *' - */ -#define FROM_ENGINE(ptr) \ - FROM_I2C_ENGINE(container_of((ptr), struct i2c_engine, base)) - -static void release_engine( - struct engine *engine) -{ -} - -static void destruct( - struct i2c_sw_engine_dce110 *engine) -{ - dal_i2c_sw_engine_destruct(&engine->base); -} - -static void destroy( - struct i2c_engine **engine) -{ - struct i2c_sw_engine_dce110 *sw_engine = FROM_I2C_ENGINE(*engine); - - destruct(sw_engine); - - kfree(sw_engine); - - *engine = NULL; -} - -static bool acquire_engine( - struct i2c_engine *engine, - struct ddc *ddc_handle) -{ - return dal_i2caux_i2c_sw_engine_acquire_engine(engine, ddc_handle); -} - -static const struct i2c_engine_funcs i2c_engine_funcs = { - .acquire_engine = acquire_engine, - .destroy = destroy, - .get_speed = dal_i2c_sw_engine_get_speed, - .set_speed = dal_i2c_sw_engine_set_speed, - .setup_engine = dal_i2c_engine_setup_i2c_engine, - .submit_channel_request = dal_i2c_sw_engine_submit_channel_request, - .process_channel_reply = dal_i2c_engine_process_channel_reply, - .get_channel_status = dal_i2c_sw_engine_get_channel_status, -}; - -static const struct engine_funcs engine_funcs = { - .release_engine = release_engine, - .get_engine_type = dal_i2c_sw_engine_get_engine_type, - .acquire = dal_i2c_engine_acquire, - .submit_request = dal_i2c_sw_engine_submit_request, -}; - -static void construct( - struct i2c_sw_engine_dce110 *engine_dce110, - const struct i2c_sw_engine_dce110_create_arg *arg_dce110) -{ - struct i2c_sw_engine_create_arg arg_base; - - arg_base.ctx = arg_dce110->ctx; - arg_base.default_speed = arg_dce110->default_speed; - - dal_i2c_sw_engine_construct(&engine_dce110->base, &arg_base); - - /*struct engine struct engine_funcs*/ - engine_dce110->base.base.base.funcs = &engine_funcs; - /*struct i2c_engine struct i2c_engine_funcs*/ - engine_dce110->base.base.funcs = &i2c_engine_funcs; - engine_dce110->base.default_speed = arg_dce110->default_speed; - engine_dce110->engine_id = arg_dce110->engine_id; -} - -struct i2c_engine *dal_i2c_sw_engine_dce110_create( - const struct i2c_sw_engine_dce110_create_arg *arg) -{ - struct i2c_sw_engine_dce110 *engine_dce110; - - if (!arg) { - ASSERT_CRITICAL(false); - return NULL; - } - - engine_dce110 = kzalloc(sizeof(struct i2c_sw_engine_dce110), - GFP_KERNEL); - - if (!engine_dce110) { - ASSERT_CRITICAL(false); - return NULL; - } - - construct(engine_dce110, arg); - return &engine_dce110->base.base; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2caux_dce110.c b/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2caux_dce110.c deleted file mode 100644 index 1d748ac1d6d6..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2caux_dce110.c +++ /dev/null @@ -1,329 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "../i2caux.h" -#include "../engine.h" -#include "../i2c_engine.h" -#include "../i2c_sw_engine.h" -#include "../i2c_hw_engine.h" - -/* - * Header of this unit - */ -#include "i2caux_dce110.h" - -#include "i2c_sw_engine_dce110.h" -#include "i2c_hw_engine_dce110.h" -#include "aux_engine_dce110.h" -#include "../../dc.h" -#include "dc_types.h" - - -/* - * Post-requisites: headers required by this unit - */ - -/* - * This unit - */ -/*cast pointer to struct i2caux TO pointer to struct i2caux_dce110*/ -#define FROM_I2C_AUX(ptr) \ - container_of((ptr), struct i2caux_dce110, base) - -static void destruct( - struct i2caux_dce110 *i2caux_dce110) -{ - dal_i2caux_destruct(&i2caux_dce110->base); -} - -static void destroy( - struct i2caux **i2c_engine) -{ - struct i2caux_dce110 *i2caux_dce110 = FROM_I2C_AUX(*i2c_engine); - - destruct(i2caux_dce110); - - kfree(i2caux_dce110); - - *i2c_engine = NULL; -} - -static struct i2c_engine *acquire_i2c_hw_engine( - struct i2caux *i2caux, - struct ddc *ddc) -{ - struct i2caux_dce110 *i2caux_dce110 = FROM_I2C_AUX(i2caux); - - struct i2c_engine *engine = NULL; - /* generic hw engine is not used for EDID read - * It may be needed for external i2c device, like thermal chip, - * TODO will be implemented when needed. - * check dce80 bool non_generic for generic hw engine; - */ - - if (!ddc) - return NULL; - - if (ddc->hw_info.hw_supported) { - enum gpio_ddc_line line = dal_ddc_get_line(ddc); - - if (line < GPIO_DDC_LINE_COUNT) - engine = i2caux->i2c_hw_engines[line]; - } - - if (!engine) - return NULL; - - if (!i2caux_dce110->i2c_hw_buffer_in_use && - engine->base.funcs->acquire(&engine->base, ddc)) { - i2caux_dce110->i2c_hw_buffer_in_use = true; - return engine; - } - - return NULL; -} - -static void release_engine( - struct i2caux *i2caux, - struct engine *engine) -{ - struct i2caux_dce110 *i2caux_dce110 = FROM_I2C_AUX(i2caux); - - if (engine->funcs->get_engine_type(engine) == - I2CAUX_ENGINE_TYPE_I2C_DDC_HW) - i2caux_dce110->i2c_hw_buffer_in_use = false; - - dal_i2caux_release_engine(i2caux, engine); -} - -static const enum gpio_ddc_line hw_ddc_lines[] = { - GPIO_DDC_LINE_DDC1, - GPIO_DDC_LINE_DDC2, - GPIO_DDC_LINE_DDC3, - GPIO_DDC_LINE_DDC4, - GPIO_DDC_LINE_DDC5, - GPIO_DDC_LINE_DDC6, -}; - -static const enum gpio_ddc_line hw_aux_lines[] = { - GPIO_DDC_LINE_DDC1, - GPIO_DDC_LINE_DDC2, - GPIO_DDC_LINE_DDC3, - GPIO_DDC_LINE_DDC4, - GPIO_DDC_LINE_DDC5, - GPIO_DDC_LINE_DDC6, -}; - -/* function table */ -static const struct i2caux_funcs i2caux_funcs = { - .destroy = destroy, - .acquire_i2c_hw_engine = acquire_i2c_hw_engine, - .release_engine = release_engine, - .acquire_i2c_sw_engine = dal_i2caux_acquire_i2c_sw_engine, - .acquire_aux_engine = dal_i2caux_acquire_aux_engine, -}; - -#include "dce/dce_11_0_d.h" -#include "dce/dce_11_0_sh_mask.h" - -/* set register offset */ -#define SR(reg_name)\ - .reg_name = mm ## reg_name - -/* set register offset with instance */ -#define SRI(reg_name, block, id)\ - .reg_name = mm ## block ## id ## _ ## reg_name - -#define aux_regs(id)\ -[id] = {\ - AUX_COMMON_REG_LIST(id), \ - .AUX_RESET_MASK = AUX_CONTROL__AUX_RESET_MASK \ -} - -#define hw_engine_regs(id)\ -{\ - I2C_HW_ENGINE_COMMON_REG_LIST(id) \ -} - -static const struct dce110_aux_registers dce110_aux_regs[] = { - aux_regs(0), - aux_regs(1), - aux_regs(2), - aux_regs(3), - aux_regs(4), - aux_regs(5) -}; - -static const struct dce110_i2c_hw_engine_registers i2c_hw_engine_regs[] = { - hw_engine_regs(1), - hw_engine_regs(2), - hw_engine_regs(3), - hw_engine_regs(4), - hw_engine_regs(5), - hw_engine_regs(6) -}; - -static const struct dce110_i2c_hw_engine_shift i2c_shift = { - I2C_COMMON_MASK_SH_LIST_DCE110(__SHIFT) -}; - -static const struct dce110_i2c_hw_engine_mask i2c_mask = { - I2C_COMMON_MASK_SH_LIST_DCE110(_MASK) -}; - -void dal_i2caux_dce110_construct( - struct i2caux_dce110 *i2caux_dce110, - struct dc_context *ctx, - unsigned int num_i2caux_inst, - const struct dce110_aux_registers aux_regs[], - const struct dce110_i2c_hw_engine_registers i2c_hw_engine_regs[], - const struct dce110_i2c_hw_engine_shift *i2c_shift, - const struct dce110_i2c_hw_engine_mask *i2c_mask) -{ - uint32_t i = 0; - uint32_t reference_frequency = 0; - bool use_i2c_sw_engine = false; - struct i2caux *base = NULL; - /*TODO: For CZ bring up, if dal_i2caux_get_reference_clock - * does not return 48KHz, we need hard coded for 48Khz. - * Some BIOS setting incorrect cause this - * For production, we always get value from BIOS*/ - reference_frequency = - dal_i2caux_get_reference_clock(ctx->dc_bios) >> 1; - - base = &i2caux_dce110->base; - - dal_i2caux_construct(base, ctx); - - i2caux_dce110->base.funcs = &i2caux_funcs; - i2caux_dce110->i2c_hw_buffer_in_use = false; - /* Create I2C engines (DDC lines per connector) - * different I2C/AUX usage cases, DDC, Generic GPIO, AUX. - */ - do { - enum gpio_ddc_line line_id = hw_ddc_lines[i]; - - struct i2c_hw_engine_dce110_create_arg hw_arg_dce110; - - if (use_i2c_sw_engine) { - struct i2c_sw_engine_dce110_create_arg sw_arg; - - sw_arg.engine_id = i; - sw_arg.default_speed = base->default_i2c_sw_speed; - sw_arg.ctx = ctx; - base->i2c_sw_engines[line_id] = - dal_i2c_sw_engine_dce110_create(&sw_arg); - } - - hw_arg_dce110.engine_id = i; - hw_arg_dce110.reference_frequency = reference_frequency; - hw_arg_dce110.default_speed = base->default_i2c_hw_speed; - hw_arg_dce110.ctx = ctx; - hw_arg_dce110.regs = &i2c_hw_engine_regs[i]; - hw_arg_dce110.i2c_shift = i2c_shift; - hw_arg_dce110.i2c_mask = i2c_mask; - - base->i2c_hw_engines[line_id] = - dal_i2c_hw_engine_dce110_create(&hw_arg_dce110); - if (base->i2c_hw_engines[line_id] != NULL) { - switch (ctx->dce_version) { - case DCN_VERSION_1_0: - base->i2c_hw_engines[line_id]->setup_limit = - I2C_SETUP_TIME_LIMIT_DCN; - base->i2c_hw_engines[line_id]->send_reset_length = 0; - break; - default: - base->i2c_hw_engines[line_id]->setup_limit = - I2C_SETUP_TIME_LIMIT_DCE; - base->i2c_hw_engines[line_id]->send_reset_length = 0; - break; - } - } - ++i; - } while (i < num_i2caux_inst); - - /* Create AUX engines for all lines which has assisted HW AUX - * 'i' (loop counter) used as DDC/AUX engine_id */ - - i = 0; - - do { - enum gpio_ddc_line line_id = hw_aux_lines[i]; - - struct aux_engine_dce110_init_data aux_init_data; - - aux_init_data.engine_id = i; - aux_init_data.timeout_period = base->aux_timeout_period; - aux_init_data.ctx = ctx; - aux_init_data.regs = &aux_regs[i]; - - base->aux_engines[line_id] = - dal_aux_engine_dce110_create(&aux_init_data); - - ++i; - } while (i < num_i2caux_inst); - - /*TODO Generic I2C SW and HW*/ -} - -/* - * dal_i2caux_dce110_create - * - * @brief - * public interface to allocate memory for DCE11 I2CAUX - * - * @param - * struct adapter_service *as - [in] - * struct dc_context *ctx - [in] - * - * @return - * pointer to the base struct of DCE11 I2CAUX - */ -struct i2caux *dal_i2caux_dce110_create( - struct dc_context *ctx) -{ - struct i2caux_dce110 *i2caux_dce110 = - kzalloc(sizeof(struct i2caux_dce110), GFP_KERNEL); - - if (!i2caux_dce110) { - ASSERT_CRITICAL(false); - return NULL; - } - - dal_i2caux_dce110_construct(i2caux_dce110, - ctx, - ARRAY_SIZE(dce110_aux_regs), - dce110_aux_regs, - i2c_hw_engine_regs, - &i2c_shift, - &i2c_mask); - return &i2caux_dce110->base; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2caux_dce110.h b/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2caux_dce110.h deleted file mode 100644 index d3d8cc58666a..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2caux_dce110.h +++ /dev/null @@ -1,54 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_AUX_DCE110_H__ -#define __DAL_I2C_AUX_DCE110_H__ - -#include "../i2caux.h" - -struct i2caux_dce110 { - struct i2caux base; - /* indicate the I2C HW circular buffer is in use */ - bool i2c_hw_buffer_in_use; -}; - -struct dce110_aux_registers; -struct dce110_i2c_hw_engine_registers; -struct dce110_i2c_hw_engine_shift; -struct dce110_i2c_hw_engine_mask; - -struct i2caux *dal_i2caux_dce110_create( - struct dc_context *ctx); - -void dal_i2caux_dce110_construct( - struct i2caux_dce110 *i2caux_dce110, - struct dc_context *ctx, - unsigned int num_i2caux_inst, - const struct dce110_aux_registers *aux_regs, - const struct dce110_i2c_hw_engine_registers *i2c_hw_engine_regs, - const struct dce110_i2c_hw_engine_shift *i2c_shift, - const struct dce110_i2c_hw_engine_mask *i2c_mask); - -#endif /* __DAL_I2C_AUX_DCE110_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce112/i2caux_dce112.c b/drivers/gpu/drm/amd/display/dc/i2caux/dce112/i2caux_dce112.c deleted file mode 100644 index a9db04738724..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce112/i2caux_dce112.c +++ /dev/null @@ -1,129 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -#include "include/i2caux_interface.h" -#include "../i2caux.h" -#include "../engine.h" -#include "../i2c_engine.h" -#include "../i2c_sw_engine.h" -#include "../i2c_hw_engine.h" - -#include "../dce110/i2caux_dce110.h" -#include "i2caux_dce112.h" - -#include "../dce110/aux_engine_dce110.h" - -#include "../dce110/i2c_hw_engine_dce110.h" - -#include "dce/dce_11_2_d.h" -#include "dce/dce_11_2_sh_mask.h" - -/* set register offset */ -#define SR(reg_name)\ - .reg_name = mm ## reg_name - -/* set register offset with instance */ -#define SRI(reg_name, block, id)\ - .reg_name = mm ## block ## id ## _ ## reg_name - -#define aux_regs(id)\ -[id] = {\ - AUX_COMMON_REG_LIST(id), \ - .AUX_RESET_MASK = AUX_CONTROL__AUX_RESET_MASK \ -} - -#define hw_engine_regs(id)\ -{\ - I2C_HW_ENGINE_COMMON_REG_LIST(id) \ -} - -static const struct dce110_aux_registers dce112_aux_regs[] = { - aux_regs(0), - aux_regs(1), - aux_regs(2), - aux_regs(3), - aux_regs(4), - aux_regs(5), -}; - -static const struct dce110_i2c_hw_engine_registers dce112_hw_engine_regs[] = { - hw_engine_regs(1), - hw_engine_regs(2), - hw_engine_regs(3), - hw_engine_regs(4), - hw_engine_regs(5), - hw_engine_regs(6) -}; - -static const struct dce110_i2c_hw_engine_shift i2c_shift = { - I2C_COMMON_MASK_SH_LIST_DCE110(__SHIFT) -}; - -static const struct dce110_i2c_hw_engine_mask i2c_mask = { - I2C_COMMON_MASK_SH_LIST_DCE110(_MASK) -}; - -static void construct( - struct i2caux_dce110 *i2caux_dce110, - struct dc_context *ctx) -{ - dal_i2caux_dce110_construct(i2caux_dce110, - ctx, - ARRAY_SIZE(dce112_aux_regs), - dce112_aux_regs, - dce112_hw_engine_regs, - &i2c_shift, - &i2c_mask); -} - -/* - * dal_i2caux_dce110_create - * - * @brief - * public interface to allocate memory for DCE11 I2CAUX - * - * @param - * struct adapter_service *as - [in] - * struct dc_context *ctx - [in] - * - * @return - * pointer to the base struct of DCE11 I2CAUX - */ -struct i2caux *dal_i2caux_dce112_create( - struct dc_context *ctx) -{ - struct i2caux_dce110 *i2caux_dce110 = - kzalloc(sizeof(struct i2caux_dce110), GFP_KERNEL); - - if (!i2caux_dce110) { - ASSERT_CRITICAL(false); - return NULL; - } - - construct(i2caux_dce110, ctx); - return &i2caux_dce110->base; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce112/i2caux_dce112.h b/drivers/gpu/drm/amd/display/dc/i2caux/dce112/i2caux_dce112.h deleted file mode 100644 index 8d35453c25b6..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce112/i2caux_dce112.h +++ /dev/null @@ -1,32 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_AUX_DCE112_H__ -#define __DAL_I2C_AUX_DCE112_H__ - -struct i2caux *dal_i2caux_dce112_create( - struct dc_context *ctx); - -#endif /* __DAL_I2C_AUX_DCE112_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce120/i2caux_dce120.c b/drivers/gpu/drm/amd/display/dc/i2caux/dce120/i2caux_dce120.c deleted file mode 100644 index 6a4f344c1db4..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce120/i2caux_dce120.c +++ /dev/null @@ -1,120 +0,0 @@ -/* - * Copyright 2012-16 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -#include "include/i2caux_interface.h" -#include "../i2caux.h" -#include "../engine.h" -#include "../i2c_engine.h" -#include "../i2c_sw_engine.h" -#include "../i2c_hw_engine.h" - -#include "../dce110/i2c_hw_engine_dce110.h" -#include "../dce110/aux_engine_dce110.h" -#include "../dce110/i2caux_dce110.h" - -#include "dce/dce_12_0_offset.h" -#include "dce/dce_12_0_sh_mask.h" -#include "soc15_hw_ip.h" -#include "vega10_ip_offset.h" - -/* begin ********************* - * macros to expend register list macro defined in HW object header file */ - -#define BASE_INNER(seg) \ - DCE_BASE__INST0_SEG ## seg - -/* compile time expand base address. */ -#define BASE(seg) \ - BASE_INNER(seg) - -#define SR(reg_name)\ - .reg_name = BASE(mm ## reg_name ## _BASE_IDX) + \ - mm ## reg_name - -#define SRI(reg_name, block, id)\ - .reg_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \ - mm ## block ## id ## _ ## reg_name -/* macros to expend register list macro defined in HW object header file - * end *********************/ - -#define aux_regs(id)\ -[id] = {\ - AUX_COMMON_REG_LIST(id), \ - .AUX_RESET_MASK = DP_AUX0_AUX_CONTROL__AUX_RESET_MASK \ -} - -static const struct dce110_aux_registers dce120_aux_regs[] = { - aux_regs(0), - aux_regs(1), - aux_regs(2), - aux_regs(3), - aux_regs(4), - aux_regs(5), -}; - -#define hw_engine_regs(id)\ -{\ - I2C_HW_ENGINE_COMMON_REG_LIST(id) \ -} - -static const struct dce110_i2c_hw_engine_registers dce120_hw_engine_regs[] = { - hw_engine_regs(1), - hw_engine_regs(2), - hw_engine_regs(3), - hw_engine_regs(4), - hw_engine_regs(5), - hw_engine_regs(6) -}; - -static const struct dce110_i2c_hw_engine_shift i2c_shift = { - I2C_COMMON_MASK_SH_LIST_DCE110(__SHIFT) -}; - -static const struct dce110_i2c_hw_engine_mask i2c_mask = { - I2C_COMMON_MASK_SH_LIST_DCE110(_MASK) -}; - -struct i2caux *dal_i2caux_dce120_create( - struct dc_context *ctx) -{ - struct i2caux_dce110 *i2caux_dce110 = - kzalloc(sizeof(struct i2caux_dce110), GFP_KERNEL); - - if (!i2caux_dce110) { - ASSERT_CRITICAL(false); - return NULL; - } - - dal_i2caux_dce110_construct(i2caux_dce110, - ctx, - ARRAY_SIZE(dce120_aux_regs), - dce120_aux_regs, - dce120_hw_engine_regs, - &i2c_shift, - &i2c_mask); - return &i2caux_dce110->base; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2c_hw_engine_dce80.c b/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2c_hw_engine_dce80.c deleted file mode 100644 index fd0832dd2c75..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2c_hw_engine_dce80.c +++ /dev/null @@ -1,875 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "../engine.h" -#include "../i2c_engine.h" -#include "../i2c_hw_engine.h" -#include "../i2c_generic_hw_engine.h" -/* - * Header of this unit - */ - -#include "i2c_hw_engine_dce80.h" - -/* - * Post-requisites: headers required by this unit - */ - -#include "dce/dce_8_0_d.h" -#include "dce/dce_8_0_sh_mask.h" -/* - * This unit - */ - -enum dc_i2c_status { - DC_I2C_STATUS__DC_I2C_STATUS_IDLE, - DC_I2C_STATUS__DC_I2C_STATUS_USED_BY_SW, - DC_I2C_STATUS__DC_I2C_STATUS_USED_BY_HW -}; - -enum dc_i2c_arbitration { - DC_I2C_ARBITRATION__DC_I2C_SW_PRIORITY_NORMAL, - DC_I2C_ARBITRATION__DC_I2C_SW_PRIORITY_HIGH -}; - -enum { - /* No timeout in HW - * (timeout implemented in SW by querying status) */ - I2C_SETUP_TIME_LIMIT = 255, - I2C_HW_BUFFER_SIZE = 144 -}; - -/* - * @brief - * Cast 'struct i2c_hw_engine *' - * to 'struct i2c_hw_engine_dce80 *' - */ -#define FROM_I2C_HW_ENGINE(ptr) \ - container_of((ptr), struct i2c_hw_engine_dce80, base) - -/* - * @brief - * Cast pointer to 'struct i2c_engine *' - * to pointer to 'struct i2c_hw_engine_dce80 *' - */ -#define FROM_I2C_ENGINE(ptr) \ - FROM_I2C_HW_ENGINE(container_of((ptr), struct i2c_hw_engine, base)) - -/* - * @brief - * Cast pointer to 'struct engine *' - * to 'pointer to struct i2c_hw_engine_dce80 *' - */ -#define FROM_ENGINE(ptr) \ - FROM_I2C_ENGINE(container_of((ptr), struct i2c_engine, base)) - -static void disable_i2c_hw_engine( - struct i2c_hw_engine_dce80 *engine) -{ - const uint32_t addr = engine->addr.DC_I2C_DDCX_SETUP; - uint32_t value = 0; - - struct dc_context *ctx = NULL; - - ctx = engine->base.base.base.ctx; - - value = dm_read_reg(ctx, addr); - - set_reg_field_value( - value, - 0, - DC_I2C_DDC1_SETUP, - DC_I2C_DDC1_ENABLE); - - dm_write_reg(ctx, addr, value); -} - -static void release_engine( - struct engine *engine) -{ - struct i2c_hw_engine_dce80 *hw_engine = FROM_ENGINE(engine); - - struct i2c_engine *base = NULL; - bool safe_to_reset; - uint32_t value = 0; - - base = &hw_engine->base.base; - - /* Restore original HW engine speed */ - - base->funcs->set_speed(base, hw_engine->base.original_speed); - - /* Release I2C */ - { - value = dm_read_reg(engine->ctx, mmDC_I2C_ARBITRATION); - - set_reg_field_value( - value, - 1, - DC_I2C_ARBITRATION, - DC_I2C_SW_DONE_USING_I2C_REG); - - dm_write_reg(engine->ctx, mmDC_I2C_ARBITRATION, value); - } - - /* Reset HW engine */ - { - uint32_t i2c_sw_status = 0; - - value = dm_read_reg(engine->ctx, mmDC_I2C_SW_STATUS); - - i2c_sw_status = get_reg_field_value( - value, - DC_I2C_SW_STATUS, - DC_I2C_SW_STATUS); - /* if used by SW, safe to reset */ - safe_to_reset = (i2c_sw_status == 1); - } - { - value = dm_read_reg(engine->ctx, mmDC_I2C_CONTROL); - - if (safe_to_reset) - set_reg_field_value( - value, - 1, - DC_I2C_CONTROL, - DC_I2C_SOFT_RESET); - - set_reg_field_value( - value, - 1, - DC_I2C_CONTROL, - DC_I2C_SW_STATUS_RESET); - - dm_write_reg(engine->ctx, mmDC_I2C_CONTROL, value); - } - - /* HW I2c engine - clock gating feature */ - if (!hw_engine->engine_keep_power_up_count) - disable_i2c_hw_engine(hw_engine); -} - -static void destruct( - struct i2c_hw_engine_dce80 *engine) -{ - dal_i2c_hw_engine_destruct(&engine->base); -} - -static void destroy( - struct i2c_engine **i2c_engine) -{ - struct i2c_hw_engine_dce80 *engine = FROM_I2C_ENGINE(*i2c_engine); - - destruct(engine); - - kfree(engine); - - *i2c_engine = NULL; -} - -static bool setup_engine( - struct i2c_engine *i2c_engine) -{ - uint32_t value = 0; - struct i2c_hw_engine_dce80 *engine = FROM_I2C_ENGINE(i2c_engine); - - /* Program pin select */ - { - const uint32_t addr = mmDC_I2C_CONTROL; - - value = dm_read_reg(i2c_engine->base.ctx, addr); - - set_reg_field_value( - value, - 0, - DC_I2C_CONTROL, - DC_I2C_GO); - - set_reg_field_value( - value, - 0, - DC_I2C_CONTROL, - DC_I2C_SOFT_RESET); - - set_reg_field_value( - value, - 0, - DC_I2C_CONTROL, - DC_I2C_SEND_RESET); - - set_reg_field_value( - value, - 0, - DC_I2C_CONTROL, - DC_I2C_SW_STATUS_RESET); - - set_reg_field_value( - value, - 0, - DC_I2C_CONTROL, - DC_I2C_TRANSACTION_COUNT); - - set_reg_field_value( - value, - engine->engine_id, - DC_I2C_CONTROL, - DC_I2C_DDC_SELECT); - - dm_write_reg(i2c_engine->base.ctx, addr, value); - } - - /* Program time limit */ - { - const uint32_t addr = engine->addr.DC_I2C_DDCX_SETUP; - - value = dm_read_reg(i2c_engine->base.ctx, addr); - - set_reg_field_value( - value, - I2C_SETUP_TIME_LIMIT, - DC_I2C_DDC1_SETUP, - DC_I2C_DDC1_TIME_LIMIT); - - set_reg_field_value( - value, - 1, - DC_I2C_DDC1_SETUP, - DC_I2C_DDC1_ENABLE); - - dm_write_reg(i2c_engine->base.ctx, addr, value); - } - - /* Program HW priority - * set to High - interrupt software I2C at any time - * Enable restart of SW I2C that was interrupted by HW - * disable queuing of software while I2C is in use by HW */ - { - value = dm_read_reg(i2c_engine->base.ctx, - mmDC_I2C_ARBITRATION); - - set_reg_field_value( - value, - 0, - DC_I2C_ARBITRATION, - DC_I2C_NO_QUEUED_SW_GO); - - set_reg_field_value( - value, - DC_I2C_ARBITRATION__DC_I2C_SW_PRIORITY_NORMAL, - DC_I2C_ARBITRATION, - DC_I2C_SW_PRIORITY); - - dm_write_reg(i2c_engine->base.ctx, - mmDC_I2C_ARBITRATION, value); - } - - return true; -} - -static uint32_t get_speed( - const struct i2c_engine *i2c_engine) -{ - const struct i2c_hw_engine_dce80 *engine = FROM_I2C_ENGINE(i2c_engine); - - const uint32_t addr = engine->addr.DC_I2C_DDCX_SPEED; - - uint32_t pre_scale = 0; - - uint32_t value = dm_read_reg(i2c_engine->base.ctx, addr); - - pre_scale = get_reg_field_value( - value, - DC_I2C_DDC1_SPEED, - DC_I2C_DDC1_PRESCALE); - - /* [anaumov] it seems following is unnecessary */ - /*ASSERT(value.bits.DC_I2C_DDC1_PRESCALE);*/ - - return pre_scale ? - engine->reference_frequency / pre_scale : - engine->base.default_speed; -} - -static void set_speed( - struct i2c_engine *i2c_engine, - uint32_t speed) -{ - struct i2c_hw_engine_dce80 *engine = FROM_I2C_ENGINE(i2c_engine); - - if (speed) { - const uint32_t addr = engine->addr.DC_I2C_DDCX_SPEED; - - uint32_t value = dm_read_reg(i2c_engine->base.ctx, addr); - - set_reg_field_value( - value, - engine->reference_frequency / speed, - DC_I2C_DDC1_SPEED, - DC_I2C_DDC1_PRESCALE); - - set_reg_field_value( - value, - 2, - DC_I2C_DDC1_SPEED, - DC_I2C_DDC1_THRESHOLD); - - dm_write_reg(i2c_engine->base.ctx, addr, value); - } -} - -static inline void reset_hw_engine(struct engine *engine) -{ - uint32_t value = dm_read_reg(engine->ctx, mmDC_I2C_CONTROL); - - set_reg_field_value( - value, - 1, - DC_I2C_CONTROL, - DC_I2C_SOFT_RESET); - - set_reg_field_value( - value, - 1, - DC_I2C_CONTROL, - DC_I2C_SW_STATUS_RESET); - - dm_write_reg(engine->ctx, mmDC_I2C_CONTROL, value); -} - -static bool is_hw_busy(struct engine *engine) -{ - uint32_t i2c_sw_status = 0; - - uint32_t value = dm_read_reg(engine->ctx, mmDC_I2C_SW_STATUS); - - i2c_sw_status = get_reg_field_value( - value, - DC_I2C_SW_STATUS, - DC_I2C_SW_STATUS); - - if (i2c_sw_status == DC_I2C_STATUS__DC_I2C_STATUS_IDLE) - return false; - - reset_hw_engine(engine); - - value = dm_read_reg(engine->ctx, mmDC_I2C_SW_STATUS); - - i2c_sw_status = get_reg_field_value( - value, - DC_I2C_SW_STATUS, - DC_I2C_SW_STATUS); - - return i2c_sw_status != DC_I2C_STATUS__DC_I2C_STATUS_IDLE; -} - -/* - * @brief - * DC_GPIO_DDC MM register offsets - */ -static const uint32_t transaction_addr[] = { - mmDC_I2C_TRANSACTION0, - mmDC_I2C_TRANSACTION1, - mmDC_I2C_TRANSACTION2, - mmDC_I2C_TRANSACTION3 -}; - -static bool process_transaction( - struct i2c_hw_engine_dce80 *engine, - struct i2c_request_transaction_data *request) -{ - uint32_t length = request->length; - uint8_t *buffer = request->data; - - bool last_transaction = false; - uint32_t value = 0; - - struct dc_context *ctx = NULL; - - ctx = engine->base.base.base.ctx; - - { - const uint32_t addr = - transaction_addr[engine->transaction_count]; - - value = dm_read_reg(ctx, addr); - - set_reg_field_value( - value, - 1, - DC_I2C_TRANSACTION0, - DC_I2C_STOP_ON_NACK0); - - set_reg_field_value( - value, - 1, - DC_I2C_TRANSACTION0, - DC_I2C_START0); - - if ((engine->transaction_count == 3) || - (request->action == I2CAUX_TRANSACTION_ACTION_I2C_WRITE) || - (request->action & I2CAUX_TRANSACTION_ACTION_I2C_READ)) { - - set_reg_field_value( - value, - 1, - DC_I2C_TRANSACTION0, - DC_I2C_STOP0); - - last_transaction = true; - } else - set_reg_field_value( - value, - 0, - DC_I2C_TRANSACTION0, - DC_I2C_STOP0); - - set_reg_field_value( - value, - (0 != (request->action & - I2CAUX_TRANSACTION_ACTION_I2C_READ)), - DC_I2C_TRANSACTION0, - DC_I2C_RW0); - - set_reg_field_value( - value, - length, - DC_I2C_TRANSACTION0, - DC_I2C_COUNT0); - - dm_write_reg(ctx, addr, value); - } - - /* Write the I2C address and I2C data - * into the hardware circular buffer, one byte per entry. - * As an example, the 7-bit I2C slave address for CRT monitor - * for reading DDC/EDID information is 0b1010001. - * For an I2C send operation, the LSB must be programmed to 0; - * for I2C receive operation, the LSB must be programmed to 1. */ - - { - value = 0; - - set_reg_field_value( - value, - false, - DC_I2C_DATA, - DC_I2C_DATA_RW); - - set_reg_field_value( - value, - request->address, - DC_I2C_DATA, - DC_I2C_DATA); - - if (engine->transaction_count == 0) { - set_reg_field_value( - value, - 0, - DC_I2C_DATA, - DC_I2C_INDEX); - - /*enable index write*/ - set_reg_field_value( - value, - 1, - DC_I2C_DATA, - DC_I2C_INDEX_WRITE); - } - - dm_write_reg(ctx, mmDC_I2C_DATA, value); - - if (!(request->action & I2CAUX_TRANSACTION_ACTION_I2C_READ)) { - - set_reg_field_value( - value, - 0, - DC_I2C_DATA, - DC_I2C_INDEX_WRITE); - - while (length) { - - set_reg_field_value( - value, - *buffer++, - DC_I2C_DATA, - DC_I2C_DATA); - - dm_write_reg(ctx, mmDC_I2C_DATA, value); - --length; - } - } - } - - ++engine->transaction_count; - engine->buffer_used_bytes += length + 1; - - return last_transaction; -} - -static void execute_transaction( - struct i2c_hw_engine_dce80 *engine) -{ - uint32_t value = 0; - struct dc_context *ctx = NULL; - - ctx = engine->base.base.base.ctx; - - { - const uint32_t addr = engine->addr.DC_I2C_DDCX_SETUP; - - value = dm_read_reg(ctx, addr); - - set_reg_field_value( - value, - 0, - DC_I2C_DDC1_SETUP, - DC_I2C_DDC1_DATA_DRIVE_EN); - - set_reg_field_value( - value, - 0, - DC_I2C_DDC1_SETUP, - DC_I2C_DDC1_CLK_DRIVE_EN); - - set_reg_field_value( - value, - 0, - DC_I2C_DDC1_SETUP, - DC_I2C_DDC1_DATA_DRIVE_SEL); - - set_reg_field_value( - value, - 0, - DC_I2C_DDC1_SETUP, - DC_I2C_DDC1_INTRA_TRANSACTION_DELAY); - - set_reg_field_value( - value, - 0, - DC_I2C_DDC1_SETUP, - DC_I2C_DDC1_INTRA_BYTE_DELAY); - - dm_write_reg(ctx, addr, value); - } - - { - const uint32_t addr = mmDC_I2C_CONTROL; - - value = dm_read_reg(ctx, addr); - - set_reg_field_value( - value, - 0, - DC_I2C_CONTROL, - DC_I2C_SOFT_RESET); - - set_reg_field_value( - value, - 0, - DC_I2C_CONTROL, - DC_I2C_SW_STATUS_RESET); - - set_reg_field_value( - value, - 0, - DC_I2C_CONTROL, - DC_I2C_SEND_RESET); - - set_reg_field_value( - value, - 0, - DC_I2C_CONTROL, - DC_I2C_GO); - - set_reg_field_value( - value, - engine->transaction_count - 1, - DC_I2C_CONTROL, - DC_I2C_TRANSACTION_COUNT); - - dm_write_reg(ctx, addr, value); - } - - /* start I2C transfer */ - { - const uint32_t addr = mmDC_I2C_CONTROL; - - value = dm_read_reg(ctx, addr); - - set_reg_field_value( - value, - 1, - DC_I2C_CONTROL, - DC_I2C_GO); - - dm_write_reg(ctx, addr, value); - } - - /* all transactions were executed and HW buffer became empty - * (even though it actually happens when status becomes DONE) */ - engine->transaction_count = 0; - engine->buffer_used_bytes = 0; -} - -static void submit_channel_request( - struct i2c_engine *engine, - struct i2c_request_transaction_data *request) -{ - request->status = I2C_CHANNEL_OPERATION_SUCCEEDED; - - if (!process_transaction(FROM_I2C_ENGINE(engine), request)) - return; - - if (is_hw_busy(&engine->base)) { - request->status = I2C_CHANNEL_OPERATION_ENGINE_BUSY; - return; - } - - execute_transaction(FROM_I2C_ENGINE(engine)); -} - -static void process_channel_reply( - struct i2c_engine *engine, - struct i2c_reply_transaction_data *reply) -{ - uint32_t length = reply->length; - uint8_t *buffer = reply->data; - - uint32_t value = 0; - - /*set index*/ - set_reg_field_value( - value, - length - 1, - DC_I2C_DATA, - DC_I2C_INDEX); - - set_reg_field_value( - value, - 1, - DC_I2C_DATA, - DC_I2C_DATA_RW); - - set_reg_field_value( - value, - 1, - DC_I2C_DATA, - DC_I2C_INDEX_WRITE); - - dm_write_reg(engine->base.ctx, mmDC_I2C_DATA, value); - - while (length) { - /* after reading the status, - * if the I2C operation executed successfully - * (i.e. DC_I2C_STATUS_DONE = 1) then the I2C controller - * should read data bytes from I2C circular data buffer */ - - value = dm_read_reg(engine->base.ctx, mmDC_I2C_DATA); - - *buffer++ = get_reg_field_value( - value, - DC_I2C_DATA, - DC_I2C_DATA); - - --length; - } -} - -static enum i2c_channel_operation_result get_channel_status( - struct i2c_engine *engine, - uint8_t *returned_bytes) -{ - uint32_t i2c_sw_status = 0; - uint32_t value = dm_read_reg(engine->base.ctx, mmDC_I2C_SW_STATUS); - - i2c_sw_status = get_reg_field_value( - value, - DC_I2C_SW_STATUS, - DC_I2C_SW_STATUS); - - if (i2c_sw_status == DC_I2C_STATUS__DC_I2C_STATUS_USED_BY_SW) - return I2C_CHANNEL_OPERATION_ENGINE_BUSY; - else if (value & DC_I2C_SW_STATUS__DC_I2C_SW_STOPPED_ON_NACK_MASK) - return I2C_CHANNEL_OPERATION_NO_RESPONSE; - else if (value & DC_I2C_SW_STATUS__DC_I2C_SW_TIMEOUT_MASK) - return I2C_CHANNEL_OPERATION_TIMEOUT; - else if (value & DC_I2C_SW_STATUS__DC_I2C_SW_ABORTED_MASK) - return I2C_CHANNEL_OPERATION_FAILED; - else if (value & DC_I2C_SW_STATUS__DC_I2C_SW_DONE_MASK) - return I2C_CHANNEL_OPERATION_SUCCEEDED; - - /* - * this is the case when HW used for communication, I2C_SW_STATUS - * could be zero - */ - return I2C_CHANNEL_OPERATION_SUCCEEDED; -} - -static uint32_t get_hw_buffer_available_size( - const struct i2c_hw_engine *engine) -{ - return I2C_HW_BUFFER_SIZE - - FROM_I2C_HW_ENGINE(engine)->buffer_used_bytes; -} - -static uint32_t get_transaction_timeout( - const struct i2c_hw_engine *engine, - uint32_t length) -{ - uint32_t speed = engine->base.funcs->get_speed(&engine->base); - - uint32_t period_timeout; - uint32_t num_of_clock_stretches; - - if (!speed) - return 0; - - period_timeout = (1000 * TRANSACTION_TIMEOUT_IN_I2C_CLOCKS) / speed; - - num_of_clock_stretches = 1 + (length << 3) + 1; - num_of_clock_stretches += - (FROM_I2C_HW_ENGINE(engine)->buffer_used_bytes << 3) + - (FROM_I2C_HW_ENGINE(engine)->transaction_count << 1); - - return period_timeout * num_of_clock_stretches; -} - -/* - * @brief - * DC_I2C_DDC1_SETUP MM register offsets - * - * @note - * The indices of this offset array are DDC engine IDs - */ -static const int32_t ddc_setup_offset[] = { - - mmDC_I2C_DDC1_SETUP - mmDC_I2C_DDC1_SETUP, /* DDC Engine 1 */ - mmDC_I2C_DDC2_SETUP - mmDC_I2C_DDC1_SETUP, /* DDC Engine 2 */ - mmDC_I2C_DDC3_SETUP - mmDC_I2C_DDC1_SETUP, /* DDC Engine 3 */ - mmDC_I2C_DDC4_SETUP - mmDC_I2C_DDC1_SETUP, /* DDC Engine 4 */ - mmDC_I2C_DDC5_SETUP - mmDC_I2C_DDC1_SETUP, /* DDC Engine 5 */ - mmDC_I2C_DDC6_SETUP - mmDC_I2C_DDC1_SETUP, /* DDC Engine 6 */ - mmDC_I2C_DDCVGA_SETUP - mmDC_I2C_DDC1_SETUP /* DDC Engine 7 */ -}; - -/* - * @brief - * DC_I2C_DDC1_SPEED MM register offsets - * - * @note - * The indices of this offset array are DDC engine IDs - */ -static const int32_t ddc_speed_offset[] = { - mmDC_I2C_DDC1_SPEED - mmDC_I2C_DDC1_SPEED, /* DDC Engine 1 */ - mmDC_I2C_DDC2_SPEED - mmDC_I2C_DDC1_SPEED, /* DDC Engine 2 */ - mmDC_I2C_DDC3_SPEED - mmDC_I2C_DDC1_SPEED, /* DDC Engine 3 */ - mmDC_I2C_DDC4_SPEED - mmDC_I2C_DDC1_SPEED, /* DDC Engine 4 */ - mmDC_I2C_DDC5_SPEED - mmDC_I2C_DDC1_SPEED, /* DDC Engine 5 */ - mmDC_I2C_DDC6_SPEED - mmDC_I2C_DDC1_SPEED, /* DDC Engine 6 */ - mmDC_I2C_DDCVGA_SPEED - mmDC_I2C_DDC1_SPEED /* DDC Engine 7 */ -}; - -static const struct i2c_engine_funcs i2c_engine_funcs = { - .destroy = destroy, - .get_speed = get_speed, - .set_speed = set_speed, - .setup_engine = setup_engine, - .submit_channel_request = submit_channel_request, - .process_channel_reply = process_channel_reply, - .get_channel_status = get_channel_status, - .acquire_engine = dal_i2c_hw_engine_acquire_engine, -}; - -static const struct engine_funcs engine_funcs = { - .release_engine = release_engine, - .get_engine_type = dal_i2c_hw_engine_get_engine_type, - .acquire = dal_i2c_engine_acquire, - .submit_request = dal_i2c_hw_engine_submit_request, -}; - -static const struct i2c_hw_engine_funcs i2c_hw_engine_funcs = { - .get_hw_buffer_available_size = - get_hw_buffer_available_size, - .get_transaction_timeout = - get_transaction_timeout, - .wait_on_operation_result = - dal_i2c_hw_engine_wait_on_operation_result, -}; - -static void construct( - struct i2c_hw_engine_dce80 *engine, - const struct i2c_hw_engine_dce80_create_arg *arg) -{ - dal_i2c_hw_engine_construct(&engine->base, arg->ctx); - - engine->base.base.base.funcs = &engine_funcs; - engine->base.base.funcs = &i2c_engine_funcs; - engine->base.funcs = &i2c_hw_engine_funcs; - engine->base.default_speed = arg->default_speed; - engine->addr.DC_I2C_DDCX_SETUP = - mmDC_I2C_DDC1_SETUP + ddc_setup_offset[arg->engine_id]; - engine->addr.DC_I2C_DDCX_SPEED = - mmDC_I2C_DDC1_SPEED + ddc_speed_offset[arg->engine_id]; - - engine->engine_id = arg->engine_id; - engine->reference_frequency = arg->reference_frequency; - engine->buffer_used_bytes = 0; - engine->transaction_count = 0; - engine->engine_keep_power_up_count = 1; -} - -struct i2c_engine *dal_i2c_hw_engine_dce80_create( - const struct i2c_hw_engine_dce80_create_arg *arg) -{ - struct i2c_hw_engine_dce80 *engine; - - if (!arg) { - BREAK_TO_DEBUGGER(); - return NULL; - } - - if ((arg->engine_id >= sizeof(ddc_setup_offset) / sizeof(int32_t)) || - (arg->engine_id >= sizeof(ddc_speed_offset) / sizeof(int32_t)) || - !arg->reference_frequency) { - BREAK_TO_DEBUGGER(); - return NULL; - } - - engine = kzalloc(sizeof(struct i2c_hw_engine_dce80), GFP_KERNEL); - - if (!engine) { - BREAK_TO_DEBUGGER(); - return NULL; - } - - construct(engine, arg); - return &engine->base.base; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2c_hw_engine_dce80.h b/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2c_hw_engine_dce80.h deleted file mode 100644 index 5c6116fb5479..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2c_hw_engine_dce80.h +++ /dev/null @@ -1,54 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_HW_ENGINE_DCE80_H__ -#define __DAL_I2C_HW_ENGINE_DCE80_H__ - -struct i2c_hw_engine_dce80 { - struct i2c_hw_engine base; - struct { - uint32_t DC_I2C_DDCX_SETUP; - uint32_t DC_I2C_DDCX_SPEED; - } addr; - uint32_t engine_id; - /* expressed in kilohertz */ - uint32_t reference_frequency; - /* number of bytes currently used in HW buffer */ - uint32_t buffer_used_bytes; - /* number of pending transactions (before GO) */ - uint32_t transaction_count; - uint32_t engine_keep_power_up_count; -}; - -struct i2c_hw_engine_dce80_create_arg { - uint32_t engine_id; - uint32_t reference_frequency; - uint32_t default_speed; - struct dc_context *ctx; -}; - -struct i2c_engine *dal_i2c_hw_engine_dce80_create( - const struct i2c_hw_engine_dce80_create_arg *arg); -#endif diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2c_sw_engine_dce80.c b/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2c_sw_engine_dce80.c deleted file mode 100644 index 4853ee26096a..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2c_sw_engine_dce80.c +++ /dev/null @@ -1,173 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "../engine.h" -#include "../i2c_engine.h" -#include "../i2c_sw_engine.h" - -/* - * Header of this unit - */ - -#include "i2c_sw_engine_dce80.h" - -/* - * Post-requisites: headers required by this unit - */ - -#include "dce/dce_8_0_d.h" -#include "dce/dce_8_0_sh_mask.h" - -/* - * This unit - */ - -static const uint32_t ddc_hw_status_addr[] = { - mmDC_I2C_DDC1_HW_STATUS, - mmDC_I2C_DDC2_HW_STATUS, - mmDC_I2C_DDC3_HW_STATUS, - mmDC_I2C_DDC4_HW_STATUS, - mmDC_I2C_DDC5_HW_STATUS, - mmDC_I2C_DDC6_HW_STATUS, - mmDC_I2C_DDCVGA_HW_STATUS -}; - -/* - * @brief - * Cast 'struct i2c_sw_engine *' - * to 'struct i2c_sw_engine_dce80 *' - */ -#define FROM_I2C_SW_ENGINE(ptr) \ - container_of((ptr), struct i2c_sw_engine_dce80, base) - -/* - * @brief - * Cast 'struct i2c_engine *' - * to 'struct i2c_sw_engine_dce80 *' - */ -#define FROM_I2C_ENGINE(ptr) \ - FROM_I2C_SW_ENGINE(container_of((ptr), struct i2c_sw_engine, base)) - -/* - * @brief - * Cast 'struct engine *' - * to 'struct i2c_sw_engine_dce80 *' - */ -#define FROM_ENGINE(ptr) \ - FROM_I2C_ENGINE(container_of((ptr), struct i2c_engine, base)) - -static void release_engine( - struct engine *engine) -{ - -} - -static void destruct( - struct i2c_sw_engine_dce80 *engine) -{ - dal_i2c_sw_engine_destruct(&engine->base); -} - -static void destroy( - struct i2c_engine **engine) -{ - struct i2c_sw_engine_dce80 *sw_engine = FROM_I2C_ENGINE(*engine); - - destruct(sw_engine); - - kfree(sw_engine); - - *engine = NULL; -} - -static bool acquire_engine( - struct i2c_engine *engine, - struct ddc *ddc_handle) -{ - return dal_i2caux_i2c_sw_engine_acquire_engine(engine, ddc_handle); -} - -static const struct i2c_engine_funcs i2c_engine_funcs = { - .acquire_engine = acquire_engine, - .destroy = destroy, - .get_speed = dal_i2c_sw_engine_get_speed, - .set_speed = dal_i2c_sw_engine_set_speed, - .setup_engine = dal_i2c_engine_setup_i2c_engine, - .submit_channel_request = dal_i2c_sw_engine_submit_channel_request, - .process_channel_reply = dal_i2c_engine_process_channel_reply, - .get_channel_status = dal_i2c_sw_engine_get_channel_status, -}; - -static const struct engine_funcs engine_funcs = { - .release_engine = release_engine, - .get_engine_type = dal_i2c_sw_engine_get_engine_type, - .acquire = dal_i2c_engine_acquire, - .submit_request = dal_i2c_sw_engine_submit_request, -}; - -static void construct( - struct i2c_sw_engine_dce80 *engine, - const struct i2c_sw_engine_dce80_create_arg *arg) -{ - struct i2c_sw_engine_create_arg arg_base; - - arg_base.ctx = arg->ctx; - arg_base.default_speed = arg->default_speed; - - dal_i2c_sw_engine_construct(&engine->base, &arg_base); - - engine->base.base.base.funcs = &engine_funcs; - engine->base.base.funcs = &i2c_engine_funcs; - engine->base.default_speed = arg->default_speed; - engine->engine_id = arg->engine_id; -} - -struct i2c_engine *dal_i2c_sw_engine_dce80_create( - const struct i2c_sw_engine_dce80_create_arg *arg) -{ - struct i2c_sw_engine_dce80 *engine; - - if (!arg) { - BREAK_TO_DEBUGGER(); - return NULL; - } - - engine = kzalloc(sizeof(struct i2c_sw_engine_dce80), GFP_KERNEL); - - if (!engine) { - BREAK_TO_DEBUGGER(); - return NULL; - } - - construct(engine, arg); - return &engine->base.base; -} - diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2caux_dce80.c b/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2caux_dce80.c deleted file mode 100644 index ed48596dd2a5..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2caux_dce80.c +++ /dev/null @@ -1,284 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "../i2caux.h" - -/* - * Header of this unit - */ - -#include "i2caux_dce80.h" - -/* - * Post-requisites: headers required by this unit - */ - -#include "../engine.h" -#include "../i2c_engine.h" -#include "../i2c_sw_engine.h" -#include "i2c_sw_engine_dce80.h" -#include "../i2c_hw_engine.h" -#include "i2c_hw_engine_dce80.h" -#include "../i2c_generic_hw_engine.h" -#include "../aux_engine.h" - - -#include "../dce110/aux_engine_dce110.h" -#include "../dce110/i2caux_dce110.h" - -#include "dce/dce_8_0_d.h" -#include "dce/dce_8_0_sh_mask.h" - - -/* set register offset */ -#define SR(reg_name)\ - .reg_name = mm ## reg_name - -/* set register offset with instance */ -#define SRI(reg_name, block, id)\ - .reg_name = mm ## block ## id ## _ ## reg_name - -#define aux_regs(id)\ -[id] = {\ - AUX_COMMON_REG_LIST(id), \ - .AUX_RESET_MASK = 0 \ -} - -static const struct dce110_aux_registers dce80_aux_regs[] = { - aux_regs(0), - aux_regs(1), - aux_regs(2), - aux_regs(3), - aux_regs(4), - aux_regs(5) -}; - -/* - * This unit - */ - -#define FROM_I2C_AUX(ptr) \ - container_of((ptr), struct i2caux_dce80, base) - -static void destruct( - struct i2caux_dce80 *i2caux_dce80) -{ - dal_i2caux_destruct(&i2caux_dce80->base); -} - -static void destroy( - struct i2caux **i2c_engine) -{ - struct i2caux_dce80 *i2caux_dce80 = FROM_I2C_AUX(*i2c_engine); - - destruct(i2caux_dce80); - - kfree(i2caux_dce80); - - *i2c_engine = NULL; -} - -static struct i2c_engine *acquire_i2c_hw_engine( - struct i2caux *i2caux, - struct ddc *ddc) -{ - struct i2caux_dce80 *i2caux_dce80 = FROM_I2C_AUX(i2caux); - - struct i2c_engine *engine = NULL; - bool non_generic; - - if (!ddc) - return NULL; - - if (ddc->hw_info.hw_supported) { - enum gpio_ddc_line line = dal_ddc_get_line(ddc); - - if (line < GPIO_DDC_LINE_COUNT) { - non_generic = true; - engine = i2caux->i2c_hw_engines[line]; - } - } - - if (!engine) { - non_generic = false; - engine = i2caux->i2c_generic_hw_engine; - } - - if (!engine) - return NULL; - - if (non_generic) { - if (!i2caux_dce80->i2c_hw_buffer_in_use && - engine->base.funcs->acquire(&engine->base, ddc)) { - i2caux_dce80->i2c_hw_buffer_in_use = true; - return engine; - } - } else { - if (engine->base.funcs->acquire(&engine->base, ddc)) - return engine; - } - - return NULL; -} - -static void release_engine( - struct i2caux *i2caux, - struct engine *engine) -{ - if (engine->funcs->get_engine_type(engine) == - I2CAUX_ENGINE_TYPE_I2C_DDC_HW) - FROM_I2C_AUX(i2caux)->i2c_hw_buffer_in_use = false; - - dal_i2caux_release_engine(i2caux, engine); -} - -static const enum gpio_ddc_line hw_ddc_lines[] = { - GPIO_DDC_LINE_DDC1, - GPIO_DDC_LINE_DDC2, - GPIO_DDC_LINE_DDC3, - GPIO_DDC_LINE_DDC4, - GPIO_DDC_LINE_DDC5, - GPIO_DDC_LINE_DDC6, - GPIO_DDC_LINE_DDC_VGA -}; - -static const enum gpio_ddc_line hw_aux_lines[] = { - GPIO_DDC_LINE_DDC1, - GPIO_DDC_LINE_DDC2, - GPIO_DDC_LINE_DDC3, - GPIO_DDC_LINE_DDC4, - GPIO_DDC_LINE_DDC5, - GPIO_DDC_LINE_DDC6 -}; - -static const struct i2caux_funcs i2caux_funcs = { - .destroy = destroy, - .acquire_i2c_hw_engine = acquire_i2c_hw_engine, - .release_engine = release_engine, - .acquire_i2c_sw_engine = dal_i2caux_acquire_i2c_sw_engine, - .acquire_aux_engine = dal_i2caux_acquire_aux_engine, -}; - -static void construct( - struct i2caux_dce80 *i2caux_dce80, - struct dc_context *ctx) -{ - /* Entire family have I2C engine reference clock frequency - * changed from XTALIN (27) to XTALIN/2 (13.5) */ - - struct i2caux *base = &i2caux_dce80->base; - - uint32_t reference_frequency = - dal_i2caux_get_reference_clock(ctx->dc_bios) >> 1; - - /*bool use_i2c_sw_engine = dal_adapter_service_is_feature_supported(as, - FEATURE_RESTORE_USAGE_I2C_SW_ENGINE);*/ - - /* Use SWI2C for dce8 currently, sicne we have bug with hwi2c */ - bool use_i2c_sw_engine = true; - - uint32_t i; - - dal_i2caux_construct(base, ctx); - - i2caux_dce80->base.funcs = &i2caux_funcs; - i2caux_dce80->i2c_hw_buffer_in_use = false; - - /* Create I2C HW engines (HW + SW pairs) - * for all lines which has assisted HW DDC - * 'i' (loop counter) used as DDC/AUX engine_id */ - - i = 0; - - do { - enum gpio_ddc_line line_id = hw_ddc_lines[i]; - - struct i2c_hw_engine_dce80_create_arg hw_arg; - - if (use_i2c_sw_engine) { - struct i2c_sw_engine_dce80_create_arg sw_arg; - - sw_arg.engine_id = i; - sw_arg.default_speed = base->default_i2c_sw_speed; - sw_arg.ctx = ctx; - base->i2c_sw_engines[line_id] = - dal_i2c_sw_engine_dce80_create(&sw_arg); - } - - hw_arg.engine_id = i; - hw_arg.reference_frequency = reference_frequency; - hw_arg.default_speed = base->default_i2c_hw_speed; - hw_arg.ctx = ctx; - - base->i2c_hw_engines[line_id] = - dal_i2c_hw_engine_dce80_create(&hw_arg); - - ++i; - } while (i < ARRAY_SIZE(hw_ddc_lines)); - - /* Create AUX engines for all lines which has assisted HW AUX - * 'i' (loop counter) used as DDC/AUX engine_id */ - - i = 0; - - do { - enum gpio_ddc_line line_id = hw_aux_lines[i]; - - struct aux_engine_dce110_init_data arg; - - arg.engine_id = i; - arg.timeout_period = base->aux_timeout_period; - arg.ctx = ctx; - arg.regs = &dce80_aux_regs[i]; - - base->aux_engines[line_id] = - dal_aux_engine_dce110_create(&arg); - - ++i; - } while (i < ARRAY_SIZE(hw_aux_lines)); - - /* TODO Generic I2C SW and HW */ -} - -struct i2caux *dal_i2caux_dce80_create( - struct dc_context *ctx) -{ - struct i2caux_dce80 *i2caux_dce80 = - kzalloc(sizeof(struct i2caux_dce80), GFP_KERNEL); - - if (!i2caux_dce80) { - BREAK_TO_DEBUGGER(); - return NULL; - } - - construct(i2caux_dce80, ctx); - return &i2caux_dce80->base; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dcn10/i2caux_dcn10.c b/drivers/gpu/drm/amd/display/dc/i2caux/dcn10/i2caux_dcn10.c deleted file mode 100644 index a59c1f50c1e8..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dcn10/i2caux_dcn10.c +++ /dev/null @@ -1,120 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -#include "include/i2caux_interface.h" -#include "../i2caux.h" -#include "../engine.h" -#include "../i2c_engine.h" -#include "../i2c_sw_engine.h" -#include "../i2c_hw_engine.h" - -#include "../dce110/aux_engine_dce110.h" -#include "../dce110/i2c_hw_engine_dce110.h" -#include "../dce110/i2caux_dce110.h" - -#include "dcn/dcn_1_0_offset.h" -#include "dcn/dcn_1_0_sh_mask.h" -#include "soc15_hw_ip.h" -#include "vega10_ip_offset.h" - -/* begin ********************* - * macros to expend register list macro defined in HW object header file */ - -#define BASE_INNER(seg) \ - DCE_BASE__INST0_SEG ## seg - -/* compile time expand base address. */ -#define BASE(seg) \ - BASE_INNER(seg) - -#define SR(reg_name)\ - .reg_name = BASE(mm ## reg_name ## _BASE_IDX) + \ - mm ## reg_name - -#define SRI(reg_name, block, id)\ - .reg_name = BASE(mm ## block ## id ## _ ## reg_name ## _BASE_IDX) + \ - mm ## block ## id ## _ ## reg_name -/* macros to expend register list macro defined in HW object header file - * end *********************/ - -#define aux_regs(id)\ -[id] = {\ - AUX_COMMON_REG_LIST(id), \ - .AUX_RESET_MASK = DP_AUX0_AUX_CONTROL__AUX_RESET_MASK \ -} - -#define hw_engine_regs(id)\ -{\ - I2C_HW_ENGINE_COMMON_REG_LIST(id) \ -} - -static const struct dce110_aux_registers dcn10_aux_regs[] = { - aux_regs(0), - aux_regs(1), - aux_regs(2), - aux_regs(3), - aux_regs(4), - aux_regs(5), -}; - -static const struct dce110_i2c_hw_engine_registers dcn10_hw_engine_regs[] = { - hw_engine_regs(1), - hw_engine_regs(2), - hw_engine_regs(3), - hw_engine_regs(4), - hw_engine_regs(5), - hw_engine_regs(6) -}; - -static const struct dce110_i2c_hw_engine_shift i2c_shift = { - I2C_COMMON_MASK_SH_LIST_DCE110(__SHIFT) -}; - -static const struct dce110_i2c_hw_engine_mask i2c_mask = { - I2C_COMMON_MASK_SH_LIST_DCE110(_MASK) -}; - -struct i2caux *dal_i2caux_dcn10_create( - struct dc_context *ctx) -{ - struct i2caux_dce110 *i2caux_dce110 = - kzalloc(sizeof(struct i2caux_dce110), GFP_KERNEL); - - if (!i2caux_dce110) { - ASSERT_CRITICAL(false); - return NULL; - } - - dal_i2caux_dce110_construct(i2caux_dce110, - ctx, - ARRAY_SIZE(dcn10_aux_regs), - dcn10_aux_regs, - dcn10_hw_engine_regs, - &i2c_shift, - &i2c_mask); - return &i2caux_dce110->base; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dcn10/i2caux_dcn10.h b/drivers/gpu/drm/amd/display/dc/i2caux/dcn10/i2caux_dcn10.h deleted file mode 100644 index aeb4a86463d4..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dcn10/i2caux_dcn10.h +++ /dev/null @@ -1,32 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_AUX_DCN10_H__ -#define __DAL_I2C_AUX_DCN10_H__ - -struct i2caux *dal_i2caux_dcn10_create( - struct dc_context *ctx); - -#endif /* __DAL_I2C_AUX_DCN10_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/diagnostics/i2caux_diag.c b/drivers/gpu/drm/amd/display/dc/i2caux/diagnostics/i2caux_diag.c deleted file mode 100644 index e6408f644086..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/diagnostics/i2caux_diag.c +++ /dev/null @@ -1,97 +0,0 @@ -/* - * Copyright 2012-16 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "../i2caux.h" -#include "../engine.h" -#include "../i2c_engine.h" -#include "../i2c_sw_engine.h" -#include "../i2c_hw_engine.h" - -/* - * Header of this unit - */ -#include "i2caux_diag.h" - -/* - * Post-requisites: headers required by this unit - */ - -/* - * This unit - */ - -static void destruct( - struct i2caux *i2caux) -{ - dal_i2caux_destruct(i2caux); -} - -static void destroy( - struct i2caux **i2c_engine) -{ - destruct(*i2c_engine); - - kfree(*i2c_engine); - - *i2c_engine = NULL; -} - -/* function table */ -static const struct i2caux_funcs i2caux_funcs = { - .destroy = destroy, - .acquire_i2c_hw_engine = NULL, - .release_engine = NULL, - .acquire_i2c_sw_engine = NULL, - .acquire_aux_engine = NULL, -}; - -static void construct( - struct i2caux *i2caux, - struct dc_context *ctx) -{ - dal_i2caux_construct(i2caux, ctx); - i2caux->funcs = &i2caux_funcs; -} - -struct i2caux *dal_i2caux_diag_fpga_create( - struct dc_context *ctx) -{ - struct i2caux *i2caux = kzalloc(sizeof(struct i2caux), - GFP_KERNEL); - - if (!i2caux) { - ASSERT_CRITICAL(false); - return NULL; - } - - construct(i2caux, ctx); - return i2caux; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/diagnostics/i2caux_diag.h b/drivers/gpu/drm/amd/display/dc/i2caux/diagnostics/i2caux_diag.h deleted file mode 100644 index a83eeb748283..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/diagnostics/i2caux_diag.h +++ /dev/null @@ -1,32 +0,0 @@ -/* - * Copyright 2012-16 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_AUX_DIAG_FPGA_H__ -#define __DAL_I2C_AUX_DIAG_FPGA_H__ - -struct i2caux *dal_i2caux_diag_fpga_create( - struct dc_context *ctx); - -#endif /* __DAL_I2C_AUX_DIAG_FPGA_H__ */ diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/engine.h b/drivers/gpu/drm/amd/display/dc/i2caux/engine.h deleted file mode 100644 index b16fb1ff687d..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/engine.h +++ /dev/null @@ -1,111 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_ENGINE_H__ -#define __DAL_ENGINE_H__ - -#include "dc_ddc_types.h" - -enum i2caux_transaction_operation { - I2CAUX_TRANSACTION_READ, - I2CAUX_TRANSACTION_WRITE -}; - -enum i2caux_transaction_address_space { - I2CAUX_TRANSACTION_ADDRESS_SPACE_I2C = 1, - I2CAUX_TRANSACTION_ADDRESS_SPACE_DPCD -}; - -struct i2caux_transaction_payload { - enum i2caux_transaction_address_space address_space; - uint32_t address; - uint32_t length; - uint8_t *data; -}; - -enum i2caux_transaction_status { - I2CAUX_TRANSACTION_STATUS_UNKNOWN = (-1L), - I2CAUX_TRANSACTION_STATUS_SUCCEEDED, - I2CAUX_TRANSACTION_STATUS_FAILED_CHANNEL_BUSY, - I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT, - I2CAUX_TRANSACTION_STATUS_FAILED_PROTOCOL_ERROR, - I2CAUX_TRANSACTION_STATUS_FAILED_NACK, - I2CAUX_TRANSACTION_STATUS_FAILED_INCOMPLETE, - I2CAUX_TRANSACTION_STATUS_FAILED_OPERATION, - I2CAUX_TRANSACTION_STATUS_FAILED_INVALID_OPERATION, - I2CAUX_TRANSACTION_STATUS_FAILED_BUFFER_OVERFLOW, - I2CAUX_TRANSACTION_STATUS_FAILED_HPD_DISCON -}; - -struct i2caux_transaction_request { - enum i2caux_transaction_operation operation; - struct i2caux_transaction_payload payload; - enum i2caux_transaction_status status; -}; - -enum i2caux_engine_type { - I2CAUX_ENGINE_TYPE_UNKNOWN = (-1L), - I2CAUX_ENGINE_TYPE_AUX, - I2CAUX_ENGINE_TYPE_I2C_DDC_HW, - I2CAUX_ENGINE_TYPE_I2C_GENERIC_HW, - I2CAUX_ENGINE_TYPE_I2C_SW -}; - -enum i2c_default_speed { - I2CAUX_DEFAULT_I2C_HW_SPEED = 50, - I2CAUX_DEFAULT_I2C_SW_SPEED = 50 -}; - -struct engine; - -struct engine_funcs { - enum i2caux_engine_type (*get_engine_type)( - const struct engine *engine); - bool (*acquire)( - struct engine *engine, - struct ddc *ddc); - bool (*submit_request)( - struct engine *engine, - struct i2caux_transaction_request *request, - bool middle_of_transaction); - void (*release_engine)( - struct engine *engine); -}; - -struct engine { - const struct engine_funcs *funcs; - uint32_t inst; - struct ddc *ddc; - struct dc_context *ctx; -}; - -void dal_i2caux_construct_engine( - struct engine *engine, - struct dc_context *ctx); - -void dal_i2caux_destruct_engine( - struct engine *engine); - -#endif diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_engine.c b/drivers/gpu/drm/amd/display/dc/i2caux/i2c_engine.c deleted file mode 100644 index 70e20bd47ce4..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_engine.c +++ /dev/null @@ -1,118 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "engine.h" - -/* - * Header of this unit - */ - -#include "i2c_engine.h" - -/* - * Post-requisites: headers required by this unit - */ - -/* - * This unit - */ - -#define FROM_ENGINE(ptr) \ - container_of((ptr), struct i2c_engine, base) - -bool dal_i2c_engine_acquire( - struct engine *engine, - struct ddc *ddc_handle) -{ - struct i2c_engine *i2c_engine = FROM_ENGINE(engine); - - uint32_t counter = 0; - bool result; - - do { - result = i2c_engine->funcs->acquire_engine( - i2c_engine, ddc_handle); - - if (result) - break; - - /* i2c_engine is busy by VBios, lets wait and retry */ - - udelay(10); - - ++counter; - } while (counter < 2); - - if (result) { - if (!i2c_engine->funcs->setup_engine(i2c_engine)) { - engine->funcs->release_engine(engine); - result = false; - } - } - - return result; -} - -bool dal_i2c_engine_setup_i2c_engine( - struct i2c_engine *engine) -{ - /* Derivative classes do not have to override this */ - - return true; -} - -void dal_i2c_engine_submit_channel_request( - struct i2c_engine *engine, - struct i2c_request_transaction_data *request) -{ - -} - -void dal_i2c_engine_process_channel_reply( - struct i2c_engine *engine, - struct i2c_reply_transaction_data *reply) -{ - -} - -void dal_i2c_engine_construct( - struct i2c_engine *engine, - struct dc_context *ctx) -{ - dal_i2caux_construct_engine(&engine->base, ctx); - engine->timeout_delay = 0; -} - -void dal_i2c_engine_destruct( - struct i2c_engine *engine) -{ - dal_i2caux_destruct_engine(&engine->base); -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_engine.h b/drivers/gpu/drm/amd/display/dc/i2caux/i2c_engine.h deleted file mode 100644 index ded6ea34b714..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_engine.h +++ /dev/null @@ -1,115 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_ENGINE_H__ -#define __DAL_I2C_ENGINE_H__ - -enum i2c_channel_operation_result { - I2C_CHANNEL_OPERATION_SUCCEEDED, - I2C_CHANNEL_OPERATION_FAILED, - I2C_CHANNEL_OPERATION_NOT_GRANTED, - I2C_CHANNEL_OPERATION_IS_BUSY, - I2C_CHANNEL_OPERATION_NO_HANDLE_PROVIDED, - I2C_CHANNEL_OPERATION_CHANNEL_IN_USE, - I2C_CHANNEL_OPERATION_CHANNEL_CLIENT_MAX_ALLOWED, - I2C_CHANNEL_OPERATION_ENGINE_BUSY, - I2C_CHANNEL_OPERATION_TIMEOUT, - I2C_CHANNEL_OPERATION_NO_RESPONSE, - I2C_CHANNEL_OPERATION_HW_REQUEST_I2C_BUS, - I2C_CHANNEL_OPERATION_WRONG_PARAMETER, - I2C_CHANNEL_OPERATION_OUT_NB_OF_RETRIES, - I2C_CHANNEL_OPERATION_NOT_STARTED -}; - -struct i2c_request_transaction_data { - enum i2caux_transaction_action action; - enum i2c_channel_operation_result status; - uint8_t address; - uint32_t length; - uint8_t *data; -}; - -struct i2c_reply_transaction_data { - uint32_t length; - uint8_t *data; -}; - -struct i2c_engine; - -struct i2c_engine_funcs { - void (*destroy)( - struct i2c_engine **ptr); - uint32_t (*get_speed)( - const struct i2c_engine *engine); - void (*set_speed)( - struct i2c_engine *engine, - uint32_t speed); - bool (*acquire_engine)( - struct i2c_engine *engine, - struct ddc *ddc); - bool (*setup_engine)( - struct i2c_engine *engine); - void (*submit_channel_request)( - struct i2c_engine *engine, - struct i2c_request_transaction_data *request); - void (*process_channel_reply)( - struct i2c_engine *engine, - struct i2c_reply_transaction_data *reply); - enum i2c_channel_operation_result (*get_channel_status)( - struct i2c_engine *engine, - uint8_t *returned_bytes); -}; - -struct i2c_engine { - struct engine base; - const struct i2c_engine_funcs *funcs; - uint32_t timeout_delay; - uint32_t setup_limit; - uint32_t send_reset_length; -}; - -void dal_i2c_engine_construct( - struct i2c_engine *engine, - struct dc_context *ctx); - -void dal_i2c_engine_destruct( - struct i2c_engine *engine); - -bool dal_i2c_engine_setup_i2c_engine( - struct i2c_engine *engine); - -void dal_i2c_engine_submit_channel_request( - struct i2c_engine *engine, - struct i2c_request_transaction_data *request); - -void dal_i2c_engine_process_channel_reply( - struct i2c_engine *engine, - struct i2c_reply_transaction_data *reply); - -bool dal_i2c_engine_acquire( - struct engine *ptr, - struct ddc *ddc_handle); - -#endif diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_generic_hw_engine.c b/drivers/gpu/drm/amd/display/dc/i2caux/i2c_generic_hw_engine.c deleted file mode 100644 index 5a4295e0fae5..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_generic_hw_engine.c +++ /dev/null @@ -1,284 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "engine.h" -#include "i2c_engine.h" -#include "i2c_hw_engine.h" - -/* - * Header of this unit - */ - -#include "i2c_generic_hw_engine.h" - -/* - * Post-requisites: headers required by this unit - */ - -/* - * This unit - */ - -/* - * @brief - * Cast 'struct i2c_hw_engine *' - * to 'struct i2c_generic_hw_engine *' - */ -#define FROM_I2C_HW_ENGINE(ptr) \ - container_of((ptr), struct i2c_generic_hw_engine, base) - -/* - * @brief - * Cast 'struct i2c_engine *' - * to 'struct i2c_generic_hw_engine *' - */ -#define FROM_I2C_ENGINE(ptr) \ - FROM_I2C_HW_ENGINE(container_of((ptr), struct i2c_hw_engine, base)) - -/* - * @brief - * Cast 'struct engine *' - * to 'struct i2c_generic_hw_engine *' - */ -#define FROM_ENGINE(ptr) \ - FROM_I2C_ENGINE(container_of((ptr), struct i2c_engine, base)) - -enum i2caux_engine_type dal_i2c_generic_hw_engine_get_engine_type( - const struct engine *engine) -{ - return I2CAUX_ENGINE_TYPE_I2C_GENERIC_HW; -} - -/* - * @brief - * Single transaction handling. - * Since transaction may be bigger than HW buffer size, - * it divides transaction to sub-transactions - * and uses batch transaction feature of the engine. - */ -bool dal_i2c_generic_hw_engine_submit_request( - struct engine *engine, - struct i2caux_transaction_request *i2caux_request, - bool middle_of_transaction) -{ - struct i2c_generic_hw_engine *hw_engine = FROM_ENGINE(engine); - - struct i2c_hw_engine *base = &hw_engine->base; - - uint32_t max_payload_size = - base->funcs->get_hw_buffer_available_size(base); - - bool initial_stop_bit = !middle_of_transaction; - - struct i2c_generic_transaction_attributes attributes; - - enum i2c_channel_operation_result operation_result = - I2C_CHANNEL_OPERATION_FAILED; - - bool result = false; - - /* setup transaction initial properties */ - - uint8_t address = i2caux_request->payload.address; - uint8_t *current_payload = i2caux_request->payload.data; - uint32_t remaining_payload_size = i2caux_request->payload.length; - - bool first_iteration = true; - - if (i2caux_request->operation == I2CAUX_TRANSACTION_READ) - attributes.action = I2CAUX_TRANSACTION_ACTION_I2C_READ; - else if (i2caux_request->operation == I2CAUX_TRANSACTION_WRITE) - attributes.action = I2CAUX_TRANSACTION_ACTION_I2C_WRITE; - else { - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_INVALID_OPERATION; - return false; - } - - /* Do batch transaction. - * Divide read/write data into payloads which fit HW buffer size. - * 1. Single transaction: - * start_bit = 1, stop_bit depends on session state, ack_on_read = 0; - * 2. Start of batch transaction: - * start_bit = 1, stop_bit = 0, ack_on_read = 1; - * 3. Middle of batch transaction: - * start_bit = 0, stop_bit = 0, ack_on_read = 1; - * 4. End of batch transaction: - * start_bit = 0, stop_bit depends on session state, ack_on_read = 0. - * Session stop bit is set if 'middle_of_transaction' = 0. */ - - while (remaining_payload_size) { - uint32_t current_transaction_size; - uint32_t current_payload_size; - - bool last_iteration; - bool stop_bit; - - /* Calculate current transaction size and payload size. - * Transaction size = total number of bytes in transaction, - * including slave's address; - * Payload size = number of data bytes in transaction. */ - - if (first_iteration) { - /* In the first sub-transaction we send slave's address - * thus we need to reserve one byte for it */ - current_transaction_size = - (remaining_payload_size > max_payload_size - 1) ? - max_payload_size : - remaining_payload_size + 1; - - current_payload_size = current_transaction_size - 1; - } else { - /* Second and further sub-transactions will have - * entire buffer reserved for data */ - current_transaction_size = - (remaining_payload_size > max_payload_size) ? - max_payload_size : - remaining_payload_size; - - current_payload_size = current_transaction_size; - } - - last_iteration = - (remaining_payload_size == current_payload_size); - - stop_bit = last_iteration ? initial_stop_bit : false; - - /* write slave device address */ - - if (first_iteration) - hw_engine->funcs->write_address(hw_engine, address); - - /* write current portion of data, if requested */ - - if (i2caux_request->operation == I2CAUX_TRANSACTION_WRITE) - hw_engine->funcs->write_data( - hw_engine, - current_payload, - current_payload_size); - - /* execute transaction */ - - attributes.start_bit = first_iteration; - attributes.stop_bit = stop_bit; - attributes.last_read = last_iteration; - attributes.transaction_size = current_transaction_size; - - hw_engine->funcs->execute_transaction(hw_engine, &attributes); - - /* wait until transaction is processed; if it fails - quit */ - - operation_result = base->funcs->wait_on_operation_result( - base, - base->funcs->get_transaction_timeout( - base, current_transaction_size), - I2C_CHANNEL_OPERATION_ENGINE_BUSY); - - if (operation_result != I2C_CHANNEL_OPERATION_SUCCEEDED) - break; - - /* read current portion of data, if requested */ - - /* the read offset should be 1 for first sub-transaction, - * and 0 for any next one */ - - if (i2caux_request->operation == I2CAUX_TRANSACTION_READ) - hw_engine->funcs->read_data(hw_engine, current_payload, - current_payload_size, first_iteration ? 1 : 0); - - /* update loop variables */ - - first_iteration = false; - current_payload += current_payload_size; - remaining_payload_size -= current_payload_size; - } - - /* update transaction status */ - - switch (operation_result) { - case I2C_CHANNEL_OPERATION_SUCCEEDED: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_SUCCEEDED; - result = true; - break; - case I2C_CHANNEL_OPERATION_NO_RESPONSE: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_NACK; - break; - case I2C_CHANNEL_OPERATION_TIMEOUT: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - break; - case I2C_CHANNEL_OPERATION_FAILED: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_INCOMPLETE; - break; - default: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_OPERATION; - } - - return result; -} - -/* - * @brief - * Returns number of microseconds to wait until timeout to be considered - */ -uint32_t dal_i2c_generic_hw_engine_get_transaction_timeout( - const struct i2c_hw_engine *engine, - uint32_t length) -{ - const struct i2c_engine *base = &engine->base; - - uint32_t speed = base->funcs->get_speed(base); - - if (!speed) - return 0; - - /* total timeout = period_timeout * (start + data bits count + stop) */ - - return ((1000 * TRANSACTION_TIMEOUT_IN_I2C_CLOCKS) / speed) * - (1 + (length << 3) + 1); -} - -void dal_i2c_generic_hw_engine_construct( - struct i2c_generic_hw_engine *engine, - struct dc_context *ctx) -{ - dal_i2c_hw_engine_construct(&engine->base, ctx); -} - -void dal_i2c_generic_hw_engine_destruct( - struct i2c_generic_hw_engine *engine) -{ - dal_i2c_hw_engine_destruct(&engine->base); -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_generic_hw_engine.h b/drivers/gpu/drm/amd/display/dc/i2caux/i2c_generic_hw_engine.h deleted file mode 100644 index 1da0397b04a2..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_generic_hw_engine.h +++ /dev/null @@ -1,77 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_GENERIC_HW_ENGINE_H__ -#define __DAL_I2C_GENERIC_HW_ENGINE_H__ - -struct i2c_generic_transaction_attributes { - enum i2caux_transaction_action action; - uint32_t transaction_size; - bool start_bit; - bool stop_bit; - bool last_read; -}; - -struct i2c_generic_hw_engine; - -struct i2c_generic_hw_engine_funcs { - void (*write_address)( - struct i2c_generic_hw_engine *engine, - uint8_t address); - void (*write_data)( - struct i2c_generic_hw_engine *engine, - const uint8_t *buffer, - uint32_t length); - void (*read_data)( - struct i2c_generic_hw_engine *engine, - uint8_t *buffer, - uint32_t length, - uint32_t offset); - void (*execute_transaction)( - struct i2c_generic_hw_engine *engine, - struct i2c_generic_transaction_attributes *attributes); -}; - -struct i2c_generic_hw_engine { - struct i2c_hw_engine base; - const struct i2c_generic_hw_engine_funcs *funcs; -}; - -void dal_i2c_generic_hw_engine_construct( - struct i2c_generic_hw_engine *engine, - struct dc_context *ctx); - -void dal_i2c_generic_hw_engine_destruct( - struct i2c_generic_hw_engine *engine); -enum i2caux_engine_type dal_i2c_generic_hw_engine_get_engine_type( - const struct engine *engine); -bool dal_i2c_generic_hw_engine_submit_request( - struct engine *ptr, - struct i2caux_transaction_request *i2caux_request, - bool middle_of_transaction); -uint32_t dal_i2c_generic_hw_engine_get_transaction_timeout( - const struct i2c_hw_engine *engine, - uint32_t length); -#endif diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_hw_engine.c b/drivers/gpu/drm/amd/display/dc/i2caux/i2c_hw_engine.c deleted file mode 100644 index 141898533e8e..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_hw_engine.c +++ /dev/null @@ -1,251 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" -#include "dm_event_log.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "engine.h" -#include "i2c_engine.h" - -/* - * Header of this unit - */ - -#include "i2c_hw_engine.h" - -/* - * Post-requisites: headers required by this unit - */ - -/* - * This unit - */ - -/* - * @brief - * Cast 'struct i2c_engine *' - * to 'struct i2c_hw_engine *' - */ -#define FROM_I2C_ENGINE(ptr) \ - container_of((ptr), struct i2c_hw_engine, base) - -/* - * @brief - * Cast 'struct engine *' - * to 'struct i2c_hw_engine *' - */ -#define FROM_ENGINE(ptr) \ - FROM_I2C_ENGINE(container_of((ptr), struct i2c_engine, base)) - -enum i2caux_engine_type dal_i2c_hw_engine_get_engine_type( - const struct engine *engine) -{ - return I2CAUX_ENGINE_TYPE_I2C_DDC_HW; -} - -bool dal_i2c_hw_engine_submit_request( - struct engine *engine, - struct i2caux_transaction_request *i2caux_request, - bool middle_of_transaction) -{ - struct i2c_hw_engine *hw_engine = FROM_ENGINE(engine); - - struct i2c_request_transaction_data request; - - uint32_t transaction_timeout; - - enum i2c_channel_operation_result operation_result; - - bool result = false; - - /* We need following: - * transaction length will not exceed - * the number of free bytes in HW buffer (minus one for address)*/ - - if (i2caux_request->payload.length >= - hw_engine->funcs->get_hw_buffer_available_size(hw_engine)) { - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_BUFFER_OVERFLOW; - return false; - } - - if (i2caux_request->operation == I2CAUX_TRANSACTION_READ) - request.action = middle_of_transaction ? - I2CAUX_TRANSACTION_ACTION_I2C_READ_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_READ; - else if (i2caux_request->operation == I2CAUX_TRANSACTION_WRITE) - request.action = middle_of_transaction ? - I2CAUX_TRANSACTION_ACTION_I2C_WRITE_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_WRITE; - else { - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_INVALID_OPERATION; - /* [anaumov] in DAL2, there was no "return false" */ - return false; - } - - request.address = (uint8_t)i2caux_request->payload.address; - request.length = i2caux_request->payload.length; - request.data = i2caux_request->payload.data; - - /* obtain timeout value before submitting request */ - - transaction_timeout = hw_engine->funcs->get_transaction_timeout( - hw_engine, i2caux_request->payload.length + 1); - - hw_engine->base.funcs->submit_channel_request( - &hw_engine->base, &request); - /* EVENT_LOG_AUX_REQ(engine->ddc->pin_data->en, EVENT_LOG_AUX_ORIGIN_I2C, */ - /* request.action, request.address, request.length, request.data); */ - - if ((request.status == I2C_CHANNEL_OPERATION_FAILED) || - (request.status == I2C_CHANNEL_OPERATION_ENGINE_BUSY)) { - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_CHANNEL_BUSY; - return false; - } - - /* wait until transaction proceed */ - - operation_result = hw_engine->funcs->wait_on_operation_result( - hw_engine, - transaction_timeout, - I2C_CHANNEL_OPERATION_ENGINE_BUSY); - - /* update transaction status */ - - switch (operation_result) { - case I2C_CHANNEL_OPERATION_SUCCEEDED: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_SUCCEEDED; - result = true; - break; - case I2C_CHANNEL_OPERATION_NO_RESPONSE: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_NACK; - break; - case I2C_CHANNEL_OPERATION_TIMEOUT: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - break; - case I2C_CHANNEL_OPERATION_FAILED: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_INCOMPLETE; - break; - default: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_OPERATION; - } - - if (result && (i2caux_request->operation == I2CAUX_TRANSACTION_READ)) { - struct i2c_reply_transaction_data reply; - - reply.data = i2caux_request->payload.data; - reply.length = i2caux_request->payload.length; - - hw_engine->base.funcs-> - process_channel_reply(&hw_engine->base, &reply); - /* EVENT_LOG_AUX_REP(engine->ddc->pin_data->en, EVENT_LOG_AUX_ORIGIN_I2C, */ - /* AUX_TRANSACTION_REPLY_I2C_ACK, reply.length, reply.data); */ - } - - - - return result; -} - -bool dal_i2c_hw_engine_acquire_engine( - struct i2c_engine *engine, - struct ddc *ddc) -{ - enum gpio_result result; - uint32_t current_speed; - - result = dal_ddc_open(ddc, GPIO_MODE_HARDWARE, - GPIO_DDC_CONFIG_TYPE_MODE_I2C); - - if (result != GPIO_RESULT_OK) - return false; - - engine->base.ddc = ddc; - - current_speed = engine->funcs->get_speed(engine); - - if (current_speed) - FROM_I2C_ENGINE(engine)->original_speed = current_speed; - - return true; -} -/* - * @brief - * Queries in a loop for current engine status - * until retrieved status matches 'expected_result', or timeout occurs. - * Timeout given in microseconds - * and the status query frequency is also one per microsecond. - */ -enum i2c_channel_operation_result dal_i2c_hw_engine_wait_on_operation_result( - struct i2c_hw_engine *engine, - uint32_t timeout, - enum i2c_channel_operation_result expected_result) -{ - enum i2c_channel_operation_result result; - uint32_t i = 0; - - if (!timeout) - return I2C_CHANNEL_OPERATION_SUCCEEDED; - - do { - result = engine->base.funcs->get_channel_status( - &engine->base, NULL); - - if (result != expected_result) - break; - - udelay(1); - - ++i; - } while (i < timeout); - - return result; -} - -void dal_i2c_hw_engine_construct( - struct i2c_hw_engine *engine, - struct dc_context *ctx) -{ - dal_i2c_engine_construct(&engine->base, ctx); - engine->original_speed = I2CAUX_DEFAULT_I2C_HW_SPEED; - engine->default_speed = I2CAUX_DEFAULT_I2C_HW_SPEED; -} - -void dal_i2c_hw_engine_destruct( - struct i2c_hw_engine *engine) -{ - dal_i2c_engine_destruct(&engine->base); -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_hw_engine.h b/drivers/gpu/drm/amd/display/dc/i2caux/i2c_hw_engine.h deleted file mode 100644 index 8936a994804a..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_hw_engine.h +++ /dev/null @@ -1,80 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_HW_ENGINE_H__ -#define __DAL_I2C_HW_ENGINE_H__ - -enum { - TRANSACTION_TIMEOUT_IN_I2C_CLOCKS = 32 -}; - -struct i2c_hw_engine; - -struct i2c_hw_engine_funcs { - uint32_t (*get_hw_buffer_available_size)( - const struct i2c_hw_engine *engine); - enum i2c_channel_operation_result (*wait_on_operation_result)( - struct i2c_hw_engine *engine, - uint32_t timeout, - enum i2c_channel_operation_result expected_result); - uint32_t (*get_transaction_timeout)( - const struct i2c_hw_engine *engine, - uint32_t length); -}; - -struct i2c_hw_engine { - struct i2c_engine base; - const struct i2c_hw_engine_funcs *funcs; - - /* Values below are in kilohertz */ - uint32_t original_speed; - uint32_t default_speed; -}; - -void dal_i2c_hw_engine_construct( - struct i2c_hw_engine *engine, - struct dc_context *ctx); - -void dal_i2c_hw_engine_destruct( - struct i2c_hw_engine *engine); - -enum i2c_channel_operation_result dal_i2c_hw_engine_wait_on_operation_result( - struct i2c_hw_engine *engine, - uint32_t timeout, - enum i2c_channel_operation_result expected_result); - -bool dal_i2c_hw_engine_acquire_engine( - struct i2c_engine *engine, - struct ddc *ddc); - -bool dal_i2c_hw_engine_submit_request( - struct engine *ptr, - struct i2caux_transaction_request *i2caux_request, - bool middle_of_transaction); - -enum i2caux_engine_type dal_i2c_hw_engine_get_engine_type( - const struct engine *engine); - -#endif diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_sw_engine.c b/drivers/gpu/drm/amd/display/dc/i2caux/i2c_sw_engine.c deleted file mode 100644 index 8e19bb629394..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_sw_engine.c +++ /dev/null @@ -1,601 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "engine.h" -#include "i2c_engine.h" - -/* - * Header of this unit - */ - -#include "i2c_sw_engine.h" - -/* - * Post-requisites: headers required by this unit - */ - -/* - * This unit - */ - -#define SCL false -#define SDA true - -static inline bool read_bit_from_ddc( - struct ddc *ddc, - bool data_nor_clock) -{ - uint32_t value = 0; - - if (data_nor_clock) - dal_gpio_get_value(ddc->pin_data, &value); - else - dal_gpio_get_value(ddc->pin_clock, &value); - - return (value != 0); -} - -static inline void write_bit_to_ddc( - struct ddc *ddc, - bool data_nor_clock, - bool bit) -{ - uint32_t value = bit ? 1 : 0; - - if (data_nor_clock) - dal_gpio_set_value(ddc->pin_data, value); - else - dal_gpio_set_value(ddc->pin_clock, value); -} - -static bool wait_for_scl_high( - struct dc_context *ctx, - struct ddc *ddc, - uint16_t clock_delay_div_4) -{ - uint32_t scl_retry = 0; - uint32_t scl_retry_max = I2C_SW_TIMEOUT_DELAY / clock_delay_div_4; - - udelay(clock_delay_div_4); - - /* 3 milliseconds delay - * to wake up some displays from "low power" state. - */ - - do { - if (read_bit_from_ddc(ddc, SCL)) - return true; - - udelay(clock_delay_div_4); - - ++scl_retry; - } while (scl_retry <= scl_retry_max); - - return false; -} - -static bool start_sync( - struct dc_context *ctx, - struct ddc *ddc_handle, - uint16_t clock_delay_div_4) -{ - uint32_t retry = 0; - - /* The I2C communications start signal is: - * the SDA going low from high, while the SCL is high. */ - - write_bit_to_ddc(ddc_handle, SCL, true); - - udelay(clock_delay_div_4); - - do { - write_bit_to_ddc(ddc_handle, SDA, true); - - if (!read_bit_from_ddc(ddc_handle, SDA)) { - ++retry; - continue; - } - - udelay(clock_delay_div_4); - - write_bit_to_ddc(ddc_handle, SCL, true); - - if (!wait_for_scl_high(ctx, ddc_handle, clock_delay_div_4)) - break; - - write_bit_to_ddc(ddc_handle, SDA, false); - - udelay(clock_delay_div_4); - - write_bit_to_ddc(ddc_handle, SCL, false); - - udelay(clock_delay_div_4); - - return true; - } while (retry <= I2C_SW_RETRIES); - - return false; -} - -static bool stop_sync( - struct dc_context *ctx, - struct ddc *ddc_handle, - uint16_t clock_delay_div_4) -{ - uint32_t retry = 0; - - /* The I2C communications stop signal is: - * the SDA going high from low, while the SCL is high. */ - - write_bit_to_ddc(ddc_handle, SCL, false); - - udelay(clock_delay_div_4); - - write_bit_to_ddc(ddc_handle, SDA, false); - - udelay(clock_delay_div_4); - - write_bit_to_ddc(ddc_handle, SCL, true); - - if (!wait_for_scl_high(ctx, ddc_handle, clock_delay_div_4)) - return false; - - write_bit_to_ddc(ddc_handle, SDA, true); - - do { - udelay(clock_delay_div_4); - - if (read_bit_from_ddc(ddc_handle, SDA)) - return true; - - ++retry; - } while (retry <= 2); - - return false; -} - -static bool write_byte( - struct dc_context *ctx, - struct ddc *ddc_handle, - uint16_t clock_delay_div_4, - uint8_t byte) -{ - int32_t shift = 7; - bool ack; - - /* bits are transmitted serially, starting from MSB */ - - do { - udelay(clock_delay_div_4); - - write_bit_to_ddc(ddc_handle, SDA, (byte >> shift) & 1); - - udelay(clock_delay_div_4); - - write_bit_to_ddc(ddc_handle, SCL, true); - - if (!wait_for_scl_high(ctx, ddc_handle, clock_delay_div_4)) - return false; - - write_bit_to_ddc(ddc_handle, SCL, false); - - --shift; - } while (shift >= 0); - - /* The display sends ACK by preventing the SDA from going high - * after the SCL pulse we use to send our last data bit. - * If the SDA goes high after that bit, it's a NACK */ - - udelay(clock_delay_div_4); - - write_bit_to_ddc(ddc_handle, SDA, true); - - udelay(clock_delay_div_4); - - write_bit_to_ddc(ddc_handle, SCL, true); - - if (!wait_for_scl_high(ctx, ddc_handle, clock_delay_div_4)) - return false; - - /* read ACK bit */ - - ack = !read_bit_from_ddc(ddc_handle, SDA); - - udelay(clock_delay_div_4 << 1); - - write_bit_to_ddc(ddc_handle, SCL, false); - - udelay(clock_delay_div_4 << 1); - - return ack; -} - -static bool read_byte( - struct dc_context *ctx, - struct ddc *ddc_handle, - uint16_t clock_delay_div_4, - uint8_t *byte, - bool more) -{ - int32_t shift = 7; - - uint8_t data = 0; - - /* The data bits are read from MSB to LSB; - * bit is read while SCL is high */ - - do { - write_bit_to_ddc(ddc_handle, SCL, true); - - if (!wait_for_scl_high(ctx, ddc_handle, clock_delay_div_4)) - return false; - - if (read_bit_from_ddc(ddc_handle, SDA)) - data |= (1 << shift); - - write_bit_to_ddc(ddc_handle, SCL, false); - - udelay(clock_delay_div_4 << 1); - - --shift; - } while (shift >= 0); - - /* read only whole byte */ - - *byte = data; - - udelay(clock_delay_div_4); - - /* send the acknowledge bit: - * SDA low means ACK, SDA high means NACK */ - - write_bit_to_ddc(ddc_handle, SDA, !more); - - udelay(clock_delay_div_4); - - write_bit_to_ddc(ddc_handle, SCL, true); - - if (!wait_for_scl_high(ctx, ddc_handle, clock_delay_div_4)) - return false; - - write_bit_to_ddc(ddc_handle, SCL, false); - - udelay(clock_delay_div_4); - - write_bit_to_ddc(ddc_handle, SDA, true); - - udelay(clock_delay_div_4); - - return true; -} - -static bool i2c_write( - struct dc_context *ctx, - struct ddc *ddc_handle, - uint16_t clock_delay_div_4, - uint8_t address, - uint32_t length, - const uint8_t *data) -{ - uint32_t i = 0; - - if (!write_byte(ctx, ddc_handle, clock_delay_div_4, address)) - return false; - - while (i < length) { - if (!write_byte(ctx, ddc_handle, clock_delay_div_4, data[i])) - return false; - ++i; - } - - return true; -} - -static bool i2c_read( - struct dc_context *ctx, - struct ddc *ddc_handle, - uint16_t clock_delay_div_4, - uint8_t address, - uint32_t length, - uint8_t *data) -{ - uint32_t i = 0; - - if (!write_byte(ctx, ddc_handle, clock_delay_div_4, address)) - return false; - - while (i < length) { - if (!read_byte(ctx, ddc_handle, clock_delay_div_4, data + i, - i < length - 1)) - return false; - ++i; - } - - return true; -} - -/* - * @brief - * Cast 'struct i2c_engine *' - * to 'struct i2c_sw_engine *' - */ -#define FROM_I2C_ENGINE(ptr) \ - container_of((ptr), struct i2c_sw_engine, base) - -/* - * @brief - * Cast 'struct engine *' - * to 'struct i2c_sw_engine *' - */ -#define FROM_ENGINE(ptr) \ - FROM_I2C_ENGINE(container_of((ptr), struct i2c_engine, base)) - -enum i2caux_engine_type dal_i2c_sw_engine_get_engine_type( - const struct engine *engine) -{ - return I2CAUX_ENGINE_TYPE_I2C_SW; -} - -bool dal_i2c_sw_engine_submit_request( - struct engine *engine, - struct i2caux_transaction_request *i2caux_request, - bool middle_of_transaction) -{ - struct i2c_sw_engine *sw_engine = FROM_ENGINE(engine); - - struct i2c_engine *base = &sw_engine->base; - - struct i2c_request_transaction_data request; - bool operation_succeeded = false; - - if (i2caux_request->operation == I2CAUX_TRANSACTION_READ) - request.action = middle_of_transaction ? - I2CAUX_TRANSACTION_ACTION_I2C_READ_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_READ; - else if (i2caux_request->operation == I2CAUX_TRANSACTION_WRITE) - request.action = middle_of_transaction ? - I2CAUX_TRANSACTION_ACTION_I2C_WRITE_MOT : - I2CAUX_TRANSACTION_ACTION_I2C_WRITE; - else { - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_INVALID_OPERATION; - /* in DAL2, there was no "return false" */ - return false; - } - - request.address = (uint8_t)i2caux_request->payload.address; - request.length = i2caux_request->payload.length; - request.data = i2caux_request->payload.data; - - base->funcs->submit_channel_request(base, &request); - - if ((request.status == I2C_CHANNEL_OPERATION_ENGINE_BUSY) || - (request.status == I2C_CHANNEL_OPERATION_FAILED)) - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_CHANNEL_BUSY; - else { - enum i2c_channel_operation_result operation_result; - - do { - operation_result = - base->funcs->get_channel_status(base, NULL); - - switch (operation_result) { - case I2C_CHANNEL_OPERATION_SUCCEEDED: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_SUCCEEDED; - operation_succeeded = true; - break; - case I2C_CHANNEL_OPERATION_NO_RESPONSE: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_NACK; - break; - case I2C_CHANNEL_OPERATION_TIMEOUT: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_TIMEOUT; - break; - case I2C_CHANNEL_OPERATION_FAILED: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_INCOMPLETE; - break; - default: - i2caux_request->status = - I2CAUX_TRANSACTION_STATUS_FAILED_OPERATION; - break; - } - } while (operation_result == I2C_CHANNEL_OPERATION_ENGINE_BUSY); - } - - return operation_succeeded; -} - -uint32_t dal_i2c_sw_engine_get_speed( - const struct i2c_engine *engine) -{ - return FROM_I2C_ENGINE(engine)->speed; -} - -void dal_i2c_sw_engine_set_speed( - struct i2c_engine *engine, - uint32_t speed) -{ - struct i2c_sw_engine *sw_engine = FROM_I2C_ENGINE(engine); - - ASSERT(speed); - - sw_engine->speed = speed ? speed : I2CAUX_DEFAULT_I2C_SW_SPEED; - - sw_engine->clock_delay = 1000 / sw_engine->speed; - - if (sw_engine->clock_delay < 12) - sw_engine->clock_delay = 12; -} - -bool dal_i2caux_i2c_sw_engine_acquire_engine( - struct i2c_engine *engine, - struct ddc *ddc) -{ - enum gpio_result result; - - result = dal_ddc_open(ddc, GPIO_MODE_FAST_OUTPUT, - GPIO_DDC_CONFIG_TYPE_MODE_I2C); - - if (result != GPIO_RESULT_OK) - return false; - - engine->base.ddc = ddc; - - return true; -} - -void dal_i2c_sw_engine_submit_channel_request( - struct i2c_engine *engine, - struct i2c_request_transaction_data *req) -{ - struct i2c_sw_engine *sw_engine = FROM_I2C_ENGINE(engine); - - struct ddc *ddc = engine->base.ddc; - uint16_t clock_delay_div_4 = sw_engine->clock_delay >> 2; - - /* send sync (start / repeated start) */ - - bool result = start_sync(engine->base.ctx, ddc, clock_delay_div_4); - - /* process payload */ - - if (result) { - switch (req->action) { - case I2CAUX_TRANSACTION_ACTION_I2C_WRITE: - case I2CAUX_TRANSACTION_ACTION_I2C_WRITE_MOT: - result = i2c_write(engine->base.ctx, ddc, clock_delay_div_4, - req->address, req->length, req->data); - break; - case I2CAUX_TRANSACTION_ACTION_I2C_READ: - case I2CAUX_TRANSACTION_ACTION_I2C_READ_MOT: - result = i2c_read(engine->base.ctx, ddc, clock_delay_div_4, - req->address, req->length, req->data); - break; - default: - result = false; - break; - } - } - - /* send stop if not 'mot' or operation failed */ - - if (!result || - (req->action == I2CAUX_TRANSACTION_ACTION_I2C_WRITE) || - (req->action == I2CAUX_TRANSACTION_ACTION_I2C_READ)) - if (!stop_sync(engine->base.ctx, ddc, clock_delay_div_4)) - result = false; - - req->status = result ? - I2C_CHANNEL_OPERATION_SUCCEEDED : - I2C_CHANNEL_OPERATION_FAILED; -} - -enum i2c_channel_operation_result dal_i2c_sw_engine_get_channel_status( - struct i2c_engine *engine, - uint8_t *returned_bytes) -{ - /* No arbitration with VBIOS is performed since DCE 6.0 */ - return I2C_CHANNEL_OPERATION_SUCCEEDED; -} - -void dal_i2c_sw_engine_destruct( - struct i2c_sw_engine *engine) -{ - dal_i2c_engine_destruct(&engine->base); -} - -static void destroy( - struct i2c_engine **ptr) -{ - dal_i2c_sw_engine_destruct(FROM_I2C_ENGINE(*ptr)); - - kfree(*ptr); - *ptr = NULL; -} - -static const struct i2c_engine_funcs i2c_engine_funcs = { - .acquire_engine = dal_i2caux_i2c_sw_engine_acquire_engine, - .destroy = destroy, - .get_speed = dal_i2c_sw_engine_get_speed, - .set_speed = dal_i2c_sw_engine_set_speed, - .setup_engine = dal_i2c_engine_setup_i2c_engine, - .submit_channel_request = dal_i2c_sw_engine_submit_channel_request, - .process_channel_reply = dal_i2c_engine_process_channel_reply, - .get_channel_status = dal_i2c_sw_engine_get_channel_status, -}; - -static void release_engine( - struct engine *engine) -{ - -} - -static const struct engine_funcs engine_funcs = { - .release_engine = release_engine, - .get_engine_type = dal_i2c_sw_engine_get_engine_type, - .acquire = dal_i2c_engine_acquire, - .submit_request = dal_i2c_sw_engine_submit_request, -}; - -void dal_i2c_sw_engine_construct( - struct i2c_sw_engine *engine, - const struct i2c_sw_engine_create_arg *arg) -{ - dal_i2c_engine_construct(&engine->base, arg->ctx); - dal_i2c_sw_engine_set_speed(&engine->base, arg->default_speed); - engine->base.funcs = &i2c_engine_funcs; - engine->base.base.funcs = &engine_funcs; -} - -struct i2c_engine *dal_i2c_sw_engine_create( - const struct i2c_sw_engine_create_arg *arg) -{ - struct i2c_sw_engine *engine; - - if (!arg) { - BREAK_TO_DEBUGGER(); - return NULL; - } - - engine = kzalloc(sizeof(struct i2c_sw_engine), GFP_KERNEL); - - if (!engine) { - BREAK_TO_DEBUGGER(); - return NULL; - } - - dal_i2c_sw_engine_construct(engine, arg); - return &engine->base; -} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_sw_engine.h b/drivers/gpu/drm/amd/display/dc/i2caux/i2c_sw_engine.h deleted file mode 100644 index 546f15b0d3f1..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/i2c_sw_engine.h +++ /dev/null @@ -1,81 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_SW_ENGINE_H__ -#define __DAL_I2C_SW_ENGINE_H__ - -enum { - I2C_SW_RETRIES = 10, - I2C_SW_SCL_READ_RETRIES = 128, - /* following value is in microseconds */ - I2C_SW_TIMEOUT_DELAY = 3000 -}; - -struct i2c_sw_engine; - -struct i2c_sw_engine { - struct i2c_engine base; - uint32_t clock_delay; - /* Values below are in KHz */ - uint32_t speed; - uint32_t default_speed; -}; - -struct i2c_sw_engine_create_arg { - uint32_t default_speed; - struct dc_context *ctx; -}; - -void dal_i2c_sw_engine_construct( - struct i2c_sw_engine *engine, - const struct i2c_sw_engine_create_arg *arg); - -bool dal_i2caux_i2c_sw_engine_acquire_engine( - struct i2c_engine *engine, - struct ddc *ddc_handle); - -void dal_i2c_sw_engine_destruct( - struct i2c_sw_engine *engine); - -struct i2c_engine *dal_i2c_sw_engine_create( - const struct i2c_sw_engine_create_arg *arg); -enum i2caux_engine_type dal_i2c_sw_engine_get_engine_type( - const struct engine *engine); -bool dal_i2c_sw_engine_submit_request( - struct engine *ptr, - struct i2caux_transaction_request *i2caux_request, - bool middle_of_transaction); -uint32_t dal_i2c_sw_engine_get_speed( - const struct i2c_engine *engine); -void dal_i2c_sw_engine_set_speed( - struct i2c_engine *ptr, - uint32_t speed); -void dal_i2c_sw_engine_submit_channel_request( - struct i2c_engine *ptr, - struct i2c_request_transaction_data *req); -enum i2c_channel_operation_result dal_i2c_sw_engine_get_channel_status( - struct i2c_engine *engine, - uint8_t *returned_bytes); -#endif diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/i2caux.c b/drivers/gpu/drm/amd/display/dc/i2caux/i2caux.c deleted file mode 100644 index 1ad6e49102ff..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/i2caux.c +++ /dev/null @@ -1,491 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#include "dm_services.h" - -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" -#include "dc_bios_types.h" - -/* - * Header of this unit - */ - -#include "i2caux.h" - -/* - * Post-requisites: headers required by this unit - */ - -#include "engine.h" -#include "i2c_engine.h" -#include "aux_engine.h" - -/* - * This unit - */ - -#include "dce80/i2caux_dce80.h" - -#include "dce100/i2caux_dce100.h" - -#include "dce110/i2caux_dce110.h" - -#include "dce112/i2caux_dce112.h" - -#include "dce120/i2caux_dce120.h" - -#if defined(CONFIG_DRM_AMD_DC_DCN1_0) -#include "dcn10/i2caux_dcn10.h" -#endif - -#include "diagnostics/i2caux_diag.h" - -/* - * @brief - * Plain API, available publicly - */ - -struct i2caux *dal_i2caux_create( - struct dc_context *ctx) -{ - if (IS_FPGA_MAXIMUS_DC(ctx->dce_environment)) { - return dal_i2caux_diag_fpga_create(ctx); - } - - switch (ctx->dce_version) { - case DCE_VERSION_8_0: - case DCE_VERSION_8_1: - case DCE_VERSION_8_3: - return dal_i2caux_dce80_create(ctx); - case DCE_VERSION_11_2: - case DCE_VERSION_11_22: - return dal_i2caux_dce112_create(ctx); - case DCE_VERSION_11_0: - return dal_i2caux_dce110_create(ctx); - case DCE_VERSION_10_0: - return dal_i2caux_dce100_create(ctx); - case DCE_VERSION_12_0: - case DCE_VERSION_12_1: - return dal_i2caux_dce120_create(ctx); -#if defined(CONFIG_DRM_AMD_DC_DCN1_0) - case DCN_VERSION_1_0: - return dal_i2caux_dcn10_create(ctx); -#endif - -#if defined(CONFIG_DRM_AMD_DC_DCN1_01) - case DCN_VERSION_1_01: - return dal_i2caux_dcn10_create(ctx); -#endif - default: - BREAK_TO_DEBUGGER(); - return NULL; - } -} - -bool dal_i2caux_submit_i2c_command( - struct i2caux *i2caux, - struct ddc *ddc, - struct i2c_command *cmd) -{ - struct i2c_engine *engine; - uint8_t index_of_payload = 0; - bool result; - - if (!ddc) { - BREAK_TO_DEBUGGER(); - return false; - } - - if (!cmd) { - BREAK_TO_DEBUGGER(); - return false; - } - - /* - * default will be SW, however there is a feature flag in adapter - * service that determines whether SW i2c_engine will be available or - * not, if sw i2c is not available we will fallback to hw. This feature - * flag is set to not creating sw i2c engine for every dce except dce80 - * currently - */ - switch (cmd->engine) { - case I2C_COMMAND_ENGINE_DEFAULT: - case I2C_COMMAND_ENGINE_SW: - /* try to acquire SW engine first, - * acquire HW engine if SW engine not available */ - engine = i2caux->funcs->acquire_i2c_sw_engine(i2caux, ddc); - - if (!engine) - engine = i2caux->funcs->acquire_i2c_hw_engine( - i2caux, ddc); - break; - case I2C_COMMAND_ENGINE_HW: - default: - /* try to acquire HW engine first, - * acquire SW engine if HW engine not available */ - engine = i2caux->funcs->acquire_i2c_hw_engine(i2caux, ddc); - - if (!engine) - engine = i2caux->funcs->acquire_i2c_sw_engine( - i2caux, ddc); - } - - if (!engine) - return false; - - engine->funcs->set_speed(engine, cmd->speed); - - result = true; - - while (index_of_payload < cmd->number_of_payloads) { - bool mot = (index_of_payload != cmd->number_of_payloads - 1); - - struct i2c_payload *payload = cmd->payloads + index_of_payload; - - struct i2caux_transaction_request request = { 0 }; - - request.operation = payload->write ? - I2CAUX_TRANSACTION_WRITE : - I2CAUX_TRANSACTION_READ; - - request.payload.address_space = - I2CAUX_TRANSACTION_ADDRESS_SPACE_I2C; - request.payload.address = (payload->address << 1) | - !payload->write; - request.payload.length = payload->length; - request.payload.data = payload->data; - - if (!engine->base.funcs->submit_request( - &engine->base, &request, mot)) { - result = false; - break; - } - - ++index_of_payload; - } - - i2caux->funcs->release_engine(i2caux, &engine->base); - - return result; -} - -bool dal_i2caux_submit_aux_command( - struct i2caux *i2caux, - struct ddc *ddc, - struct aux_command *cmd) -{ - struct aux_engine *engine; - uint8_t index_of_payload = 0; - bool result; - bool mot; - - if (!ddc) { - BREAK_TO_DEBUGGER(); - return false; - } - - if (!cmd) { - BREAK_TO_DEBUGGER(); - return false; - } - - engine = i2caux->funcs->acquire_aux_engine(i2caux, ddc); - - if (!engine) - return false; - - engine->delay = cmd->defer_delay; - engine->max_defer_write_retry = cmd->max_defer_write_retry; - - result = true; - - while (index_of_payload < cmd->number_of_payloads) { - struct aux_payload *payload = cmd->payloads + index_of_payload; - struct i2caux_transaction_request request = { 0 }; - - if (cmd->mot == I2C_MOT_UNDEF) - mot = (index_of_payload != cmd->number_of_payloads - 1); - else - mot = (cmd->mot == I2C_MOT_TRUE); - - request.operation = payload->write ? - I2CAUX_TRANSACTION_WRITE : - I2CAUX_TRANSACTION_READ; - - if (payload->i2c_over_aux) { - request.payload.address_space = - I2CAUX_TRANSACTION_ADDRESS_SPACE_I2C; - - request.payload.address = (payload->address << 1) | - !payload->write; - } else { - request.payload.address_space = - I2CAUX_TRANSACTION_ADDRESS_SPACE_DPCD; - - request.payload.address = payload->address; - } - - request.payload.length = payload->length; - request.payload.data = payload->data; - - if (!engine->base.funcs->submit_request( - &engine->base, &request, mot)) { - result = false; - break; - } - - ++index_of_payload; - } - - i2caux->funcs->release_engine(i2caux, &engine->base); - - return result; -} - -static bool get_hw_supported_ddc_line( - struct ddc *ddc, - enum gpio_ddc_line *line) -{ - enum gpio_ddc_line line_found; - - *line = GPIO_DDC_LINE_UNKNOWN; - - if (!ddc) { - BREAK_TO_DEBUGGER(); - return false; - } - - if (!ddc->hw_info.hw_supported) - return false; - - line_found = dal_ddc_get_line(ddc); - - if (line_found >= GPIO_DDC_LINE_COUNT) - return false; - - *line = line_found; - - return true; -} - -void dal_i2caux_configure_aux( - struct i2caux *i2caux, - struct ddc *ddc, - union aux_config cfg) -{ - struct aux_engine *engine = - i2caux->funcs->acquire_aux_engine(i2caux, ddc); - - if (!engine) - return; - - engine->funcs->configure(engine, cfg); - - i2caux->funcs->release_engine(i2caux, &engine->base); -} - -void dal_i2caux_destroy( - struct i2caux **i2caux) -{ - if (!i2caux || !*i2caux) { - BREAK_TO_DEBUGGER(); - return; - } - - (*i2caux)->funcs->destroy(i2caux); - - *i2caux = NULL; -} - -/* - * @brief - * An utility function used by 'struct i2caux' and its descendants - */ - -uint32_t dal_i2caux_get_reference_clock( - struct dc_bios *bios) -{ - struct dc_firmware_info info = { { 0 } }; - - if (bios->funcs->get_firmware_info(bios, &info) != BP_RESULT_OK) - return 0; - - return info.pll_info.crystal_frequency; -} - -/* - * @brief - * i2caux - */ - -enum { - /* following are expressed in KHz */ - DEFAULT_I2C_SW_SPEED = 50, - DEFAULT_I2C_HW_SPEED = 50, - - DEFAULT_I2C_SW_SPEED_100KHZ = 100, - DEFAULT_I2C_HW_SPEED_100KHZ = 100, - - /* This is the timeout as defined in DP 1.2a, - * 2.3.4 "Detailed uPacket TX AUX CH State Description". */ - AUX_TIMEOUT_PERIOD = 400, - - /* Ideally, the SW timeout should be just above 550usec - * which is programmed in HW. - * But the SW timeout of 600usec is not reliable, - * because on some systems, delay_in_microseconds() - * returns faster than it should. - * EPR #379763: by trial-and-error on different systems, - * 700usec is the minimum reliable SW timeout for polling - * the AUX_SW_STATUS.AUX_SW_DONE bit. - * This timeout expires *only* when there is - * AUX Error or AUX Timeout conditions - not during normal operation. - * During normal operation, AUX_SW_STATUS.AUX_SW_DONE bit is set - * at most within ~240usec. That means, - * increasing this timeout will not affect normal operation, - * and we'll timeout after - * SW_AUX_TIMEOUT_PERIOD_MULTIPLIER * AUX_TIMEOUT_PERIOD = 1600usec. - * This timeout is especially important for - * resume from S3 and CTS. */ - SW_AUX_TIMEOUT_PERIOD_MULTIPLIER = 4 -}; - -struct i2c_engine *dal_i2caux_acquire_i2c_sw_engine( - struct i2caux *i2caux, - struct ddc *ddc) -{ - enum gpio_ddc_line line; - struct i2c_engine *engine = NULL; - - if (get_hw_supported_ddc_line(ddc, &line)) - engine = i2caux->i2c_sw_engines[line]; - - if (!engine) - engine = i2caux->i2c_generic_sw_engine; - - if (!engine) - return NULL; - - if (!engine->base.funcs->acquire(&engine->base, ddc)) - return NULL; - - return engine; -} - -struct aux_engine *dal_i2caux_acquire_aux_engine( - struct i2caux *i2caux, - struct ddc *ddc) -{ - enum gpio_ddc_line line; - struct aux_engine *engine; - - if (!get_hw_supported_ddc_line(ddc, &line)) - return NULL; - - engine = i2caux->aux_engines[line]; - - if (!engine) - return NULL; - - if (!engine->base.funcs->acquire(&engine->base, ddc)) - return NULL; - - return engine; -} - -void dal_i2caux_release_engine( - struct i2caux *i2caux, - struct engine *engine) -{ - engine->funcs->release_engine(engine); - - dal_ddc_close(engine->ddc); - - engine->ddc = NULL; -} - -void dal_i2caux_construct( - struct i2caux *i2caux, - struct dc_context *ctx) -{ - uint32_t i = 0; - - i2caux->ctx = ctx; - do { - i2caux->i2c_sw_engines[i] = NULL; - i2caux->i2c_hw_engines[i] = NULL; - i2caux->aux_engines[i] = NULL; - - ++i; - } while (i < GPIO_DDC_LINE_COUNT); - - i2caux->i2c_generic_sw_engine = NULL; - i2caux->i2c_generic_hw_engine = NULL; - - i2caux->aux_timeout_period = - SW_AUX_TIMEOUT_PERIOD_MULTIPLIER * AUX_TIMEOUT_PERIOD; - - if (ctx->dce_version >= DCE_VERSION_11_2) { - i2caux->default_i2c_hw_speed = DEFAULT_I2C_HW_SPEED_100KHZ; - i2caux->default_i2c_sw_speed = DEFAULT_I2C_SW_SPEED_100KHZ; - } else { - i2caux->default_i2c_hw_speed = DEFAULT_I2C_HW_SPEED; - i2caux->default_i2c_sw_speed = DEFAULT_I2C_SW_SPEED; - } -} - -void dal_i2caux_destruct( - struct i2caux *i2caux) -{ - uint32_t i = 0; - - if (i2caux->i2c_generic_hw_engine) - i2caux->i2c_generic_hw_engine->funcs->destroy( - &i2caux->i2c_generic_hw_engine); - - if (i2caux->i2c_generic_sw_engine) - i2caux->i2c_generic_sw_engine->funcs->destroy( - &i2caux->i2c_generic_sw_engine); - - do { - if (i2caux->aux_engines[i]) - i2caux->aux_engines[i]->funcs->destroy( - &i2caux->aux_engines[i]); - - if (i2caux->i2c_hw_engines[i]) - i2caux->i2c_hw_engines[i]->funcs->destroy( - &i2caux->i2c_hw_engines[i]); - - if (i2caux->i2c_sw_engines[i]) - i2caux->i2c_sw_engines[i]->funcs->destroy( - &i2caux->i2c_sw_engines[i]); - - ++i; - } while (i < GPIO_DDC_LINE_COUNT); -} - diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/i2caux.h b/drivers/gpu/drm/amd/display/dc/i2caux/i2caux.h deleted file mode 100644 index 64f51bb06915..000000000000 --- a/drivers/gpu/drm/amd/display/dc/i2caux/i2caux.h +++ /dev/null @@ -1,122 +0,0 @@ -/* - * Copyright 2012-15 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: AMD - * - */ - -#ifndef __DAL_I2C_AUX_H__ -#define __DAL_I2C_AUX_H__ - -uint32_t dal_i2caux_get_reference_clock( - struct dc_bios *bios); - -struct i2caux; - -struct engine; - -struct i2caux_funcs { - void (*destroy)(struct i2caux **ptr); - struct i2c_engine * (*acquire_i2c_sw_engine)( - struct i2caux *i2caux, - struct ddc *ddc); - struct i2c_engine * (*acquire_i2c_hw_engine)( - struct i2caux *i2caux, - struct ddc *ddc); - struct aux_engine * (*acquire_aux_engine)( - struct i2caux *i2caux, - struct ddc *ddc); - void (*release_engine)( - struct i2caux *i2caux, - struct engine *engine); -}; - -struct i2c_engine; -struct aux_engine; - -struct i2caux { - struct dc_context *ctx; - const struct i2caux_funcs *funcs; - /* On ASIC we have certain amount of lines with HW DDC engine - * (4, 6, or maybe more in the future). - * For every such line, we create separate HW DDC engine - * (since we have these engines in HW) and separate SW DDC engine - * (to allow concurrent use of few lines). - * In similar way we have AUX engines. */ - - /* I2C SW engines, per DDC line. - * Only lines with HW DDC support will be initialized */ - struct i2c_engine *i2c_sw_engines[GPIO_DDC_LINE_COUNT]; - - /* I2C HW engines, per DDC line. - * Only lines with HW DDC support will be initialized */ - struct i2c_engine *i2c_hw_engines[GPIO_DDC_LINE_COUNT]; - - /* AUX engines, per DDC line. - * Only lines with HW AUX support will be initialized */ - struct aux_engine *aux_engines[GPIO_DDC_LINE_COUNT]; - - /* For all other lines, we can use - * single instance of generic I2C HW engine - * (since in HW, there is single instance of it) - * or single instance of generic I2C SW engine. - * AUX is not supported for other lines. */ - - /* General-purpose I2C SW engine. - * Can be assigned dynamically to any line per transaction */ - struct i2c_engine *i2c_generic_sw_engine; - - /* General-purpose I2C generic HW engine. - * Can be assigned dynamically to almost any line per transaction */ - struct i2c_engine *i2c_generic_hw_engine; - - /* [anaumov] in DAL2, there is a Mutex */ - - uint32_t aux_timeout_period; - - /* expressed in KHz */ - uint32_t default_i2c_sw_speed; - uint32_t default_i2c_hw_speed; -}; - -void dal_i2caux_construct( - struct i2caux *i2caux, - struct dc_context *ctx); - -void dal_i2caux_release_engine( - struct i2caux *i2caux, - struct engine *engine); - -void dal_i2caux_destruct( - struct i2caux *i2caux); - -void dal_i2caux_destroy( - struct i2caux **ptr); - -struct i2c_engine *dal_i2caux_acquire_i2c_sw_engine( - struct i2caux *i2caux, - struct ddc *ddc); - -struct aux_engine *dal_i2caux_acquire_aux_engine( - struct i2caux *i2caux, - struct ddc *ddc); - -#endif diff --git a/drivers/gpu/drm/amd/display/dc/inc/clock_source.h b/drivers/gpu/drm/amd/display/dc/inc/clock_source.h index 47ef90495376..fe6301cb8681 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/clock_source.h +++ b/drivers/gpu/drm/amd/display/dc/inc/clock_source.h @@ -78,7 +78,7 @@ struct csdp_ref_clk_ds_params { }; struct pixel_clk_params { - uint32_t requested_pix_clk; /* in KHz */ + uint32_t requested_pix_clk_100hz; /*> Requested Pixel Clock * (based on Video Timing standard used for requested mode)*/ uint32_t requested_sym_clk; /* in KHz */ @@ -104,9 +104,9 @@ struct pixel_clk_params { * with actually calculated Clock and reference Crystal frequency */ struct pll_settings { - uint32_t actual_pix_clk; - uint32_t adjusted_pix_clk; - uint32_t calculated_pix_clk; + uint32_t actual_pix_clk_100hz; + uint32_t adjusted_pix_clk_100hz; + uint32_t calculated_pix_clk_100hz; uint32_t vco_freq; uint32_t reference_freq; uint32_t reference_divider; @@ -166,6 +166,10 @@ struct clock_source_funcs { struct clock_source *, struct pixel_clk_params *, struct pll_settings *); + bool (*get_pixel_clk_frequency_100hz)( + struct clock_source *clock_source, + unsigned int inst, + unsigned int *pixel_clk_khz); }; struct clock_source { diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_status.h b/drivers/gpu/drm/amd/display/dc/inc/core_status.h index 94fc31080fda..2e61a22ef4b2 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/core_status.h +++ b/drivers/gpu/drm/amd/display/dc/inc/core_status.h @@ -30,7 +30,7 @@ enum dc_status { DC_OK = 1, DC_NO_CONTROLLER_RESOURCE = 2, - DC_NO_STREAM_ENG_RESOURCE = 3, + DC_NO_STREAM_ENC_RESOURCE = 3, DC_NO_CLOCK_SOURCE_RESOURCE = 4, DC_FAIL_CONTROLLER_VALIDATE = 5, DC_FAIL_ENC_VALIDATE = 6, diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_types.h b/drivers/gpu/drm/amd/display/dc/inc/core_types.h index b168a5e9dd9d..986ed1728644 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/core_types.h +++ b/drivers/gpu/drm/amd/display/dc/inc/core_types.h @@ -146,7 +146,7 @@ struct resource_pool { struct mpc *mpc; struct pp_smu_funcs_rv *pp_smu; struct pp_smu_display_requirement_rv pp_smu_req; - struct aux_engine *engines[MAX_PIPES]; + struct dce_aux *engines[MAX_PIPES]; struct dce_i2c_hw *hw_i2cs[MAX_PIPES]; struct dce_i2c_sw *sw_i2cs[MAX_PIPES]; bool i2c_hw_buffer_in_use; @@ -180,13 +180,8 @@ struct resource_pool { const struct resource_caps *res_cap; }; -struct dcn_fe_clocks { - int dppclk_khz; -}; - struct dcn_fe_bandwidth { - struct dcn_fe_clocks calc; - struct dcn_fe_clocks cur; + int dppclk_khz; }; struct stream_resource { diff --git a/drivers/gpu/drm/amd/display/dc/inc/dc_link_ddc.h b/drivers/gpu/drm/amd/display/dc/inc/dc_link_ddc.h index 538b83303b86..16fd4dc6c4dd 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/dc_link_ddc.h +++ b/drivers/gpu/drm/amd/display/dc/inc/dc_link_ddc.h @@ -64,13 +64,6 @@ void dal_ddc_i2c_payloads_add( uint8_t *data, bool write); -void dal_ddc_aux_payloads_add( - struct aux_payloads *payloads, - uint32_t address, - uint32_t len, - uint8_t *data, - bool write); - struct ddc_service_init_data { struct graphics_object_id id; struct dc_context *ctx; @@ -103,12 +96,10 @@ bool dal_ddc_service_query_ddc_data( uint32_t read_size); int dc_link_aux_transfer(struct ddc_service *ddc, - unsigned int address, - uint8_t *reply, - void *buffer, - unsigned int size, - enum aux_transaction_type type, - enum i2caux_transaction_action action); + struct aux_payload *payload); + +bool dc_link_aux_transfer_with_retries(struct ddc_service *ddc, + struct aux_payload *payload); void dal_ddc_service_write_scdc_data( struct ddc_service *ddc_service, diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h b/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h index 02f757dd70d4..9d2d8e51306c 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h @@ -39,6 +39,18 @@ enum segment_order { segment_order__non_contiguous, }; +struct dcn_hubbub_wm_set { + uint32_t wm_set; + uint32_t data_urgent; + uint32_t pte_meta_urgent; + uint32_t sr_enter; + uint32_t sr_exit; + uint32_t dram_clk_chanage; +}; + +struct dcn_hubbub_wm { + struct dcn_hubbub_wm_set sets[4]; +}; struct hubbub_funcs { void (*update_dchub)( @@ -58,7 +70,14 @@ struct hubbub_funcs { bool (*dcc_support_pixel_format)( enum surface_pixel_format format, unsigned int *bytes_per_element); + + void (*wm_read_state)(struct hubbub *hubbub, + struct dcn_hubbub_wm *wm); }; +struct hubbub { + const struct hubbub_funcs *funcs; + struct dc_context *ctx; +}; #endif diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/dmcu.h b/drivers/gpu/drm/amd/display/dc/inc/hw/dmcu.h index cb85eaa9857f..cbaa43853611 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/dmcu.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/dmcu.h @@ -27,16 +27,22 @@ #include "dm_services_types.h" +/* If HW itself ever powered down it will be 0. + * fwDmcuInit will write to 1. + * Driver will only call MCP init if current state is 1, + * and the MCP command will transition this to 2. + */ enum dmcu_state { - DMCU_NOT_INITIALIZED = 0, - DMCU_RUNNING = 1 + DMCU_UNLOADED = 0, + DMCU_LOADED_UNINITIALIZED = 1, + DMCU_RUNNING = 2, }; struct dmcu_version { - unsigned int date; - unsigned int month; - unsigned int year; unsigned int interface_version; + unsigned int abm_version; + unsigned int psr_version; + unsigned int build_version; }; struct dmcu { diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h b/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h index e894e649ce5a..fb7967b39edb 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h @@ -39,6 +39,11 @@ struct dpp { }; +struct dpp_input_csc_matrix { + enum dc_color_space color_space; + uint16_t regval[12]; +}; + struct dpp_grph_csc_adjustment { struct fixed31_32 temperature_matrix[CSC_TEMPERATURE_MATRIX_SIZE]; enum graphics_gamut_adjust_type gamut_adjust_type; diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h b/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h index 04c6989aac58..1cd07e94ee63 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h @@ -78,7 +78,8 @@ struct hubp_funcs { bool (*hubp_program_surface_flip_and_addr)( struct hubp *hubp, const struct dc_plane_address *address, - bool flip_immediate); + bool flip_immediate, + uint8_t vmid); void (*hubp_program_pte_vm)( struct hubp *hubp, diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h b/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h index c20fdcaac53b..c9d3e37e9531 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h @@ -153,6 +153,7 @@ struct link_encoder_funcs { void (*enable_hpd)(struct link_encoder *enc); void (*disable_hpd)(struct link_encoder *enc); bool (*is_dig_enabled)(struct link_encoder *enc); + unsigned int (*get_dig_frontend)(struct link_encoder *enc); void (*destroy)(struct link_encoder **enc); }; diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h b/drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h index 06df02ddff6a..da89c2edb07c 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h @@ -31,7 +31,7 @@ #include "dml/display_mode_structs.h" struct dchub_init_data; -struct cstate_pstate_watermarks_st1 { +struct cstate_pstate_watermarks_st { uint32_t cstate_exit_ns; uint32_t cstate_enter_plus_exit_ns; uint32_t pstate_change_ns; @@ -40,7 +40,7 @@ struct cstate_pstate_watermarks_st1 { struct dcn_watermarks { uint32_t pte_meta_urgent_ns; uint32_t urgent_ns; - struct cstate_pstate_watermarks_st1 cstate_pstate; + struct cstate_pstate_watermarks_st cstate_pstate; }; struct dcn_watermark_set { diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/stream_encoder.h b/drivers/gpu/drm/amd/display/dc/inc/hw/stream_encoder.h index 53a9b64df11a..4051493557bc 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/stream_encoder.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/stream_encoder.h @@ -161,6 +161,10 @@ struct stream_encoder_funcs { void (*set_avmute)( struct stream_encoder *enc, bool enable); + void (*dig_connect_to_otg)( + struct stream_encoder *enc, + int tg_inst); + }; #endif /* STREAM_ENCODER_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h index af700c7dac50..03ae941895f3 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h @@ -134,15 +134,29 @@ struct dc_crtc_timing; struct drr_params; +union vline_config; + + +enum vline_select { + VLINE0, + VLINE1, + VLINE2 +}; + struct timing_generator_funcs { bool (*validate_timing)(struct timing_generator *tg, const struct dc_crtc_timing *timing); void (*program_timing)(struct timing_generator *tg, const struct dc_crtc_timing *timing, bool use_vbios); - void (*program_vline_interrupt)(struct timing_generator *optc, + void (*program_vline_interrupt)( + struct timing_generator *optc, const struct dc_crtc_timing *dc_crtc_timing, - unsigned long long vsync_delta); + enum vline_select vline, + const union vline_config *vline_config); + + void (*program_vupdate_interrupt)(struct timing_generator *optc, + const struct dc_crtc_timing *dc_crtc_timing); bool (*enable_crtc)(struct timing_generator *tg); bool (*disable_crtc)(struct timing_generator *tg); bool (*is_counter_moving)(struct timing_generator *tg); @@ -159,6 +173,8 @@ struct timing_generator_funcs { bool (*get_otg_active_size)(struct timing_generator *optc, uint32_t *otg_active_width, uint32_t *otg_active_height); + bool (*is_matching_timing)(struct timing_generator *tg, + const struct dc_crtc_timing *otg_timing); void (*set_early_control)(struct timing_generator *tg, uint32_t early_cntl); void (*wait_for_state)(struct timing_generator *tg, diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/engine_base.c b/drivers/gpu/drm/amd/display/dc/inc/hw/vmid.h index 5d155d36d353..037beb0a2a27 100644 --- a/drivers/gpu/drm/amd/display/dc/i2caux/engine_base.c +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/vmid.h @@ -1,5 +1,5 @@ /* - * Copyright 2012-15 Advanced Micro Devices, Inc. + * Copyright 2018 Advanced Micro Devices, Inc. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -23,30 +23,27 @@ * */ -#include "dm_services.h" +#ifndef DAL_DC_INC_HW_VMID_H_ +#define DAL_DC_INC_HW_VMID_H_ -/* - * Pre-requisites: headers required by header of this unit - */ -#include "include/i2caux_interface.h" - -/* - * Header of this unit - */ - -#include "engine.h" +#include "core_types.h" +#include "dchubbub.h" -void dal_i2caux_construct_engine( - struct engine *engine, - struct dc_context *ctx) -{ - engine->ddc = NULL; - engine->ctx = ctx; -} +struct dcn_vmid_registers { + uint32_t CNTL; + uint32_t PAGE_TABLE_BASE_ADDR_HI32; + uint32_t PAGE_TABLE_BASE_ADDR_LO32; + uint32_t PAGE_TABLE_START_ADDR_HI32; + uint32_t PAGE_TABLE_START_ADDR_LO32; + uint32_t PAGE_TABLE_END_ADDR_HI32; + uint32_t PAGE_TABLE_END_ADDR_LO32; +}; -void dal_i2caux_destruct_engine( - struct engine *engine) -{ - /* nothing to do */ -} +struct dcn_vmid_page_table_config { + uint64_t page_table_start_addr; + uint64_t page_table_end_addr; + enum dcn_hubbub_page_table_depth depth; + enum dcn_hubbub_page_table_block_size block_size; +}; +#endif /* DAL_DC_INC_HW_VMID_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h index d6a85f48b6d1..341b4810288c 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h @@ -70,6 +70,8 @@ struct hw_sequencer_funcs { void (*init_hw)(struct dc *dc); + void (*init_pipes)(struct dc *dc, struct dc_state *context); + enum dc_status (*apply_ctx_to_hw)( struct dc *dc, struct dc_state *context); diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2c_sw_engine_dce110.h b/drivers/gpu/drm/amd/display/dc/inc/vm_helper.h index c48c61f540a8..193407f76a80 100644 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce110/i2c_sw_engine_dce110.h +++ b/drivers/gpu/drm/amd/display/dc/inc/vm_helper.h @@ -1,5 +1,5 @@ /* - * Copyright 2012-15 Advanced Micro Devices, Inc. + * Copyright 2018 Advanced Micro Devices, Inc. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -23,21 +23,34 @@ * */ -#ifndef __DAL_I2C_SW_ENGINE_DCE110_H__ -#define __DAL_I2C_SW_ENGINE_DCE110_H__ +#ifndef DC_INC_VM_HELPER_H_ +#define DC_INC_VM_HELPER_H_ -struct i2c_sw_engine_dce110 { - struct i2c_sw_engine base; - uint32_t engine_id; +#include "dc_types.h" + +#define MAX_VMID 16 +#define MAX_HUBP 6 + +struct vmid_usage { + uint16_t vmid_usage[2]; }; -struct i2c_sw_engine_dce110_create_arg { - uint32_t engine_id; - uint32_t default_speed; - struct dc_context *ctx; +struct vm_helper { + unsigned int num_vmid; + unsigned int num_hubp; + unsigned int num_vmids_available; + uint64_t ptb_assigned_to_vmid[MAX_VMID]; + struct vmid_usage hubp_vmid_usage[MAX_HUBP]; }; -struct i2c_engine *dal_i2c_sw_engine_dce110_create( - const struct i2c_sw_engine_dce110_create_arg *arg); +uint8_t get_vmid_for_ptb( + struct vm_helper *vm_helper, + int64_t ptb, + uint8_t pipe_idx); + +void init_vm_helper( + struct vm_helper *vm_helper, + unsigned int num_vmid, + unsigned int num_hubp); -#endif +#endif /* DC_INC_VM_HELPER_H_ */ diff --git a/drivers/gpu/drm/amd/display/dc/irq_types.h b/drivers/gpu/drm/amd/display/dc/irq_types.h index 0b5f3a278c22..d0ccd81ad5b4 100644 --- a/drivers/gpu/drm/amd/display/dc/irq_types.h +++ b/drivers/gpu/drm/amd/display/dc/irq_types.h @@ -144,6 +144,14 @@ enum dc_irq_source { DC_IRQ_SOURCE_DC5_VLINE0, DC_IRQ_SOURCE_DC6_VLINE0, + DC_IRQ_SOURCE_DC1_VLINE1, + DC_IRQ_SOURCE_DC2_VLINE1, + DC_IRQ_SOURCE_DC3_VLINE1, + DC_IRQ_SOURCE_DC4_VLINE1, + DC_IRQ_SOURCE_DC5_VLINE1, + DC_IRQ_SOURCE_DC6_VLINE1, + + DAL_IRQ_SOURCES_NUMBER }; diff --git a/drivers/gpu/drm/amd/display/include/bios_parser_types.h b/drivers/gpu/drm/amd/display/include/bios_parser_types.h index 7fd78a696800..01bf01a34a08 100644 --- a/drivers/gpu/drm/amd/display/include/bios_parser_types.h +++ b/drivers/gpu/drm/amd/display/include/bios_parser_types.h @@ -211,8 +211,8 @@ struct bp_pixel_clock_parameters { /* signal_type -> Encoder Mode - needed by VBIOS Exec table */ enum signal_type signal_type; /* Adjusted Pixel Clock (after VBIOS exec table) - * that becomes Target Pixel Clock (KHz) */ - uint32_t target_pixel_clock; + * that becomes Target Pixel Clock (100 Hz units) */ + uint32_t target_pixel_clock_100hz; /* Calculated Reference divider of Display PLL */ uint32_t reference_divider; /* Calculated Feedback divider of Display PLL */ diff --git a/drivers/gpu/drm/amd/display/include/gpio_interface.h b/drivers/gpu/drm/amd/display/include/gpio_interface.h index e4fd31024b92..7de64195dc33 100644 --- a/drivers/gpu/drm/amd/display/include/gpio_interface.h +++ b/drivers/gpu/drm/amd/display/include/gpio_interface.h @@ -59,6 +59,14 @@ enum gpio_result dal_gpio_change_mode( struct gpio *gpio, enum gpio_mode mode); +/* Lock Pin */ +enum gpio_result dal_gpio_lock_pin( + struct gpio *gpio); + +/* Unlock Pin */ +enum gpio_result dal_gpio_unlock_pin( + struct gpio *gpio); + /* Get the GPIO id */ enum gpio_id dal_gpio_get_id( const struct gpio *gpio); diff --git a/drivers/gpu/drm/amd/display/include/i2caux_interface.h b/drivers/gpu/drm/amd/display/include/i2caux_interface.h index 13a3c82d118f..bb012cb1a9f5 100644 --- a/drivers/gpu/drm/amd/display/include/i2caux_interface.h +++ b/drivers/gpu/drm/amd/display/include/i2caux_interface.h @@ -40,9 +40,19 @@ struct aux_payload { /* set following flag to write data, * reset it to read data */ bool write; + bool mot; uint32_t address; uint8_t length; uint8_t *data; + /* + * used to return the reply type of the transaction + * ignored if NULL + */ + uint8_t *reply; + /* expressed in milliseconds + * zero means "use default value" + */ + uint32_t defer_delay; }; struct aux_command { @@ -66,27 +76,4 @@ union aux_config { uint32_t raw; }; -struct i2caux; - -struct i2caux *dal_i2caux_create( - struct dc_context *ctx); - -bool dal_i2caux_submit_i2c_command( - struct i2caux *i2caux, - struct ddc *ddc, - struct i2c_command *cmd); - -bool dal_i2caux_submit_aux_command( - struct i2caux *i2caux, - struct ddc *ddc, - struct aux_command *cmd); - -void dal_i2caux_configure_aux( - struct i2caux *i2caux, - struct ddc *ddc, - union aux_config cfg); - -void dal_i2caux_destroy( - struct i2caux **ptr); - #endif diff --git a/drivers/gpu/drm/amd/display/modules/color/color_gamma.c b/drivers/gpu/drm/amd/display/modules/color/color_gamma.c index 479b77c2e89e..0fbc8fbc3541 100644 --- a/drivers/gpu/drm/amd/display/modules/color/color_gamma.c +++ b/drivers/gpu/drm/amd/display/modules/color/color_gamma.c @@ -823,7 +823,7 @@ static bool build_freesync_hdr(struct pwl_float_data_ex *rgb_regamma, bool is_clipped = false; struct fixed31_32 sdr_white_level; - if (fs_params == NULL || fs_params->max_content == 0 || + if (fs_params->max_content == 0 || fs_params->max_display == 0) return false; @@ -1508,7 +1508,7 @@ static bool map_regamma_hw_to_x_user( struct hw_x_point *coords = coords_x; const struct pwl_float_data_ex *regamma = rgb_regamma; - if (mapUserRamp) { + if (ramp && mapUserRamp) { copy_rgb_regamma_to_coordinates_x(coords, hw_points_num, rgb_regamma); @@ -1545,7 +1545,7 @@ bool mod_color_calculate_regamma_params(struct dc_transfer_func *output_tf, struct pwl_float_data *rgb_user = NULL; struct pwl_float_data_ex *rgb_regamma = NULL; - struct gamma_pixel *axix_x = NULL; + struct gamma_pixel *axis_x = NULL; struct pixel_gamma_point *coeff = NULL; enum dc_transfer_func_predefined tf = TRANSFER_FUNCTION_SRGB; bool ret = false; @@ -1555,47 +1555,54 @@ bool mod_color_calculate_regamma_params(struct dc_transfer_func *output_tf, /* we can use hardcoded curve for plain SRGB TF */ if (output_tf->type == TF_TYPE_PREDEFINED && canRomBeUsed == true && - output_tf->tf == TRANSFER_FUNCTION_SRGB && - (ramp->is_identity || (!mapUserRamp && ramp->type == GAMMA_RGB_256))) - return true; + output_tf->tf == TRANSFER_FUNCTION_SRGB) { + if (ramp == NULL) + return true; + if (ramp->is_identity || (!mapUserRamp && ramp->type == GAMMA_RGB_256)) + return true; + } output_tf->type = TF_TYPE_DISTRIBUTED_POINTS; - rgb_user = kvcalloc(ramp->num_entries + _EXTRA_POINTS, + if (ramp && (mapUserRamp || ramp->type != GAMMA_RGB_256)) { + rgb_user = kvcalloc(ramp->num_entries + _EXTRA_POINTS, sizeof(*rgb_user), GFP_KERNEL); - if (!rgb_user) - goto rgb_user_alloc_fail; + if (!rgb_user) + goto rgb_user_alloc_fail; + + axis_x = kvcalloc(ramp->num_entries + 3, sizeof(*axis_x), + GFP_KERNEL); + if (!axis_x) + goto axis_x_alloc_fail; + + dividers.divider1 = dc_fixpt_from_fraction(3, 2); + dividers.divider2 = dc_fixpt_from_int(2); + dividers.divider3 = dc_fixpt_from_fraction(5, 2); + + build_evenly_distributed_points( + axis_x, + ramp->num_entries, + dividers); + + if (ramp->type == GAMMA_RGB_256 && mapUserRamp) + scale_gamma(rgb_user, ramp, dividers); + else if (ramp->type == GAMMA_RGB_FLOAT_1024) + scale_gamma_dx(rgb_user, ramp, dividers); + } + rgb_regamma = kvcalloc(MAX_HW_POINTS + _EXTRA_POINTS, sizeof(*rgb_regamma), GFP_KERNEL); if (!rgb_regamma) goto rgb_regamma_alloc_fail; - axix_x = kvcalloc(ramp->num_entries + 3, sizeof(*axix_x), - GFP_KERNEL); - if (!axix_x) - goto axix_x_alloc_fail; + coeff = kvcalloc(MAX_HW_POINTS + _EXTRA_POINTS, sizeof(*coeff), GFP_KERNEL); if (!coeff) goto coeff_alloc_fail; - dividers.divider1 = dc_fixpt_from_fraction(3, 2); - dividers.divider2 = dc_fixpt_from_int(2); - dividers.divider3 = dc_fixpt_from_fraction(5, 2); - tf = output_tf->tf; - - build_evenly_distributed_points( - axix_x, - ramp->num_entries, - dividers); - - if (ramp->type == GAMMA_RGB_256 && mapUserRamp) - scale_gamma(rgb_user, ramp, dividers); - else if (ramp->type == GAMMA_RGB_FLOAT_1024) - scale_gamma_dx(rgb_user, ramp, dividers); - if (tf == TRANSFER_FUNCTION_PQ) { tf_pts->end_exponent = 7; tf_pts->x_point_at_y1_red = 125; @@ -1623,22 +1630,22 @@ bool mod_color_calculate_regamma_params(struct dc_transfer_func *output_tf, coordinates_x, tf == TRANSFER_FUNCTION_SRGB ? true:false); } map_regamma_hw_to_x_user(ramp, coeff, rgb_user, - coordinates_x, axix_x, rgb_regamma, + coordinates_x, axis_x, rgb_regamma, MAX_HW_POINTS, tf_pts, - (mapUserRamp || ramp->type != GAMMA_RGB_256) && - ramp->type != GAMMA_CS_TFM_1D); + (mapUserRamp || (ramp && ramp->type != GAMMA_RGB_256)) && + (ramp && ramp->type != GAMMA_CS_TFM_1D)); - if (ramp->type == GAMMA_CS_TFM_1D) + if (ramp && ramp->type == GAMMA_CS_TFM_1D) apply_lut_1d(ramp, MAX_HW_POINTS, tf_pts); ret = true; kvfree(coeff); coeff_alloc_fail: - kvfree(axix_x); -axix_x_alloc_fail: kvfree(rgb_regamma); rgb_regamma_alloc_fail: + kvfree(axis_x); +axis_x_alloc_fail: kvfree(rgb_user); rgb_user_alloc_fail: return ret; @@ -1758,69 +1765,85 @@ bool mod_color_calculate_degamma_params(struct dc_transfer_func *input_tf, { struct dc_transfer_func_distributed_points *tf_pts = &input_tf->tf_pts; struct dividers dividers; - struct pwl_float_data *rgb_user = NULL; struct pwl_float_data_ex *curve = NULL; struct gamma_pixel *axis_x = NULL; struct pixel_gamma_point *coeff = NULL; enum dc_transfer_func_predefined tf = TRANSFER_FUNCTION_SRGB; + uint32_t i; bool ret = false; if (input_tf->type == TF_TYPE_BYPASS) return false; - /* we can use hardcoded curve for plain SRGB TF */ + /* we can use hardcoded curve for plain SRGB TF + * If linear, it's bypass if on user ramp + */ if (input_tf->type == TF_TYPE_PREDEFINED && - input_tf->tf == TRANSFER_FUNCTION_SRGB && - (!mapUserRamp && - (ramp->type == GAMMA_RGB_256 || ramp->num_entries == 0))) + (input_tf->tf == TRANSFER_FUNCTION_SRGB || + input_tf->tf == TRANSFER_FUNCTION_LINEAR) && + !mapUserRamp) return true; input_tf->type = TF_TYPE_DISTRIBUTED_POINTS; - rgb_user = kvcalloc(ramp->num_entries + _EXTRA_POINTS, - sizeof(*rgb_user), - GFP_KERNEL); - if (!rgb_user) - goto rgb_user_alloc_fail; + if (mapUserRamp && ramp && ramp->type == GAMMA_RGB_256) { + rgb_user = kvcalloc(ramp->num_entries + _EXTRA_POINTS, + sizeof(*rgb_user), + GFP_KERNEL); + if (!rgb_user) + goto rgb_user_alloc_fail; + + axis_x = kvcalloc(ramp->num_entries + _EXTRA_POINTS, sizeof(*axis_x), + GFP_KERNEL); + if (!axis_x) + goto axis_x_alloc_fail; + + dividers.divider1 = dc_fixpt_from_fraction(3, 2); + dividers.divider2 = dc_fixpt_from_int(2); + dividers.divider3 = dc_fixpt_from_fraction(5, 2); + + build_evenly_distributed_points( + axis_x, + ramp->num_entries, + dividers); + + scale_gamma(rgb_user, ramp, dividers); + } + curve = kvcalloc(MAX_HW_POINTS + _EXTRA_POINTS, sizeof(*curve), - GFP_KERNEL); + GFP_KERNEL); if (!curve) goto curve_alloc_fail; - axis_x = kvcalloc(ramp->num_entries + _EXTRA_POINTS, sizeof(*axis_x), - GFP_KERNEL); - if (!axis_x) - goto axis_x_alloc_fail; + coeff = kvcalloc(MAX_HW_POINTS + _EXTRA_POINTS, sizeof(*coeff), - GFP_KERNEL); + GFP_KERNEL); if (!coeff) goto coeff_alloc_fail; - dividers.divider1 = dc_fixpt_from_fraction(3, 2); - dividers.divider2 = dc_fixpt_from_int(2); - dividers.divider3 = dc_fixpt_from_fraction(5, 2); - tf = input_tf->tf; - build_evenly_distributed_points( - axis_x, - ramp->num_entries, - dividers); - - if (ramp->type == GAMMA_RGB_256 && mapUserRamp) - scale_gamma(rgb_user, ramp, dividers); - else if (ramp->type == GAMMA_RGB_FLOAT_1024) - scale_gamma_dx(rgb_user, ramp, dividers); - if (tf == TRANSFER_FUNCTION_PQ) build_de_pq(curve, MAX_HW_POINTS, coordinates_x); - else + else if (tf == TRANSFER_FUNCTION_SRGB || + tf == TRANSFER_FUNCTION_BT709) build_degamma(curve, MAX_HW_POINTS, coordinates_x, - tf == TRANSFER_FUNCTION_SRGB ? true:false); + tf == TRANSFER_FUNCTION_SRGB ? true : false); + else if (tf == TRANSFER_FUNCTION_LINEAR) { + // just copy coordinates_x into curve + i = 0; + while (i != MAX_HW_POINTS + 1) { + curve[i].r = coordinates_x[i].x; + curve[i].g = curve[i].r; + curve[i].b = curve[i].r; + i++; + } + } else + goto invalid_tf_fail; tf_pts->end_exponent = 0; tf_pts->x_point_at_y1_red = 1; @@ -1830,23 +1853,21 @@ bool mod_color_calculate_degamma_params(struct dc_transfer_func *input_tf, map_regamma_hw_to_x_user(ramp, coeff, rgb_user, coordinates_x, axis_x, curve, MAX_HW_POINTS, tf_pts, - mapUserRamp && ramp->type != GAMMA_CUSTOM); - if (ramp->type == GAMMA_CUSTOM) - apply_lut_1d(ramp, MAX_HW_POINTS, tf_pts); + mapUserRamp && ramp && ramp->type == GAMMA_RGB_256); ret = true; +invalid_tf_fail: kvfree(coeff); coeff_alloc_fail: - kvfree(axis_x); -axis_x_alloc_fail: kvfree(curve); curve_alloc_fail: + kvfree(axis_x); +axis_x_alloc_fail: kvfree(rgb_user); rgb_user_alloc_fail: return ret; - } diff --git a/drivers/gpu/drm/amd/display/modules/freesync/freesync.c b/drivers/gpu/drm/amd/display/modules/freesync/freesync.c index 1544ed3f1747..94a84bc57c7a 100644 --- a/drivers/gpu/drm/amd/display/modules/freesync/freesync.c +++ b/drivers/gpu/drm/amd/display/modules/freesync/freesync.c @@ -108,8 +108,8 @@ static unsigned int calc_duration_in_us_from_v_total( { unsigned int duration_in_us = (unsigned int)(div64_u64(((unsigned long long)(v_total) - * 1000) * stream->timing.h_total, - stream->timing.pix_clk_khz)); + * 10000) * stream->timing.h_total, + stream->timing.pix_clk_100hz)); return duration_in_us; } @@ -126,7 +126,7 @@ static unsigned int calc_v_total_from_refresh( refresh_in_uhz))); v_total = div64_u64(div64_u64(((unsigned long long)( - frame_duration_in_ns) * stream->timing.pix_clk_khz), + frame_duration_in_ns) * (stream->timing.pix_clk_100hz / 10)), stream->timing.h_total), 1000000); /* v_total cannot be less than nominal */ @@ -152,7 +152,7 @@ static unsigned int calc_v_total_from_duration( duration_in_us = vrr->max_duration_in_us; v_total = div64_u64(div64_u64(((unsigned long long)( - duration_in_us) * stream->timing.pix_clk_khz), + duration_in_us) * (stream->timing.pix_clk_100hz / 10)), stream->timing.h_total), 1000); /* v_total cannot be less than nominal */ @@ -227,7 +227,7 @@ static void update_v_total_for_static_ramp( } v_total = div64_u64(div64_u64(((unsigned long long)( - current_duration_in_us) * stream->timing.pix_clk_khz), + current_duration_in_us) * (stream->timing.pix_clk_100hz / 10)), stream->timing.h_total), 1000); in_out_vrr->adjust.v_total_min = v_total; @@ -461,6 +461,26 @@ bool mod_freesync_get_v_position(struct mod_freesync *mod_freesync, return false; } +static void build_vrr_infopacket_header_vtem(enum signal_type signal, + struct dc_info_packet *infopacket) +{ + // HEADER + + // HB0, HB1, HB2 indicates PacketType VTEMPacket + infopacket->hb0 = 0x7F; + infopacket->hb1 = 0xC0; + infopacket->hb2 = 0x00; + /* HB3 Bit Fields + * Reserved :1 = 0 + * Sync :1 = 0 + * VFR :1 = 1 + * Ds_Type :2 = 0 + * End :1 = 0 + * New :1 = 0 + */ + infopacket->hb3 = 0x20; +} + static void build_vrr_infopacket_header_v1(enum signal_type signal, struct dc_info_packet *infopacket, unsigned int *payload_size) @@ -559,6 +579,54 @@ static void build_vrr_infopacket_header_v2(enum signal_type signal, } } +static void build_vrr_vtem_infopacket_data(const struct dc_stream_state *stream, + const struct mod_vrr_params *vrr, + struct dc_info_packet *infopacket) +{ + /* dc_info_packet to VtemPacket Translation of Bit-fields, + * SB[6] + * unsigned char VRR_EN :1 + * unsigned char M_CONST :1 + * unsigned char Reserved2 :2 + * unsigned char FVA_Factor_M1 :4 + * SB[7] + * unsigned char Base_Vfront :8 + * SB[8] + * unsigned char Base_Refresh_Rate_98 :2 + * unsigned char RB :1 + * unsigned char Reserved3 :5 + * SB[9] + * unsigned char Base_RefreshRate_07 :8 + */ + unsigned int fieldRateInHz; + + if (vrr->state == VRR_STATE_ACTIVE_VARIABLE || + vrr->state == VRR_STATE_ACTIVE_FIXED){ + infopacket->sb[6] |= 0x80; //VRR_EN Bit = 1 + } else { + infopacket->sb[6] &= 0x7F; //VRR_EN Bit = 0 + } + + if (!stream->timing.vic) { + infopacket->sb[7] = stream->timing.v_front_porch; + + /* TODO: In dal2, we check mode flags for a reduced blanking timing. + * Need a way to relay that information to this function. + * if("ReducedBlanking") + * { + * infopacket->sb[8] |= 0x20; //Set 3rd bit to 1 + * } + */ + fieldRateInHz = (stream->timing.pix_clk_100hz * 100)/ + (stream->timing.h_total * stream->timing.v_total); + + infopacket->sb[8] |= ((fieldRateInHz & 0x300) >> 2); + infopacket->sb[9] |= fieldRateInHz & 0xFF; + + } + infopacket->valid = true; +} + static void build_vrr_infopacket_data(const struct mod_vrr_params *vrr, struct dc_info_packet *infopacket) { @@ -672,6 +740,19 @@ static void build_vrr_infopacket_v2(enum signal_type signal, infopacket->valid = true; } +static void build_vrr_infopacket_vtem(const struct dc_stream_state *stream, + const struct mod_vrr_params *vrr, + struct dc_info_packet *infopacket) +{ + //VTEM info packet for HdmiVrr + + //VTEM Packet is structured differently + build_vrr_infopacket_header_vtem(stream->signal, infopacket); + build_vrr_vtem_infopacket_data(stream, vrr, infopacket); + + infopacket->valid = true; +} + void mod_freesync_build_vrr_infopacket(struct mod_freesync *mod_freesync, const struct dc_stream_state *stream, const struct mod_vrr_params *vrr, @@ -679,18 +760,21 @@ void mod_freesync_build_vrr_infopacket(struct mod_freesync *mod_freesync, const enum color_transfer_func *app_tf, struct dc_info_packet *infopacket) { - /* SPD info packet for FreeSync */ - - /* Check if Freesync is supported. Return if false. If true, + /* SPD info packet for FreeSync + * VTEM info packet for HdmiVRR + * Check if Freesync is supported. Return if false. If true, * set the corresponding bit in the info packet */ - if (!vrr->supported || !vrr->send_vsif) + if (!vrr->supported || (!vrr->send_info_frame && packet_type != PACKET_TYPE_VTEM)) return; switch (packet_type) { case PACKET_TYPE_FS2: build_vrr_infopacket_v2(stream->signal, vrr, app_tf, infopacket); break; + case PACKET_TYPE_VTEM: + build_vrr_infopacket_vtem(stream, vrr, infopacket); + break; case PACKET_TYPE_VRR: case PACKET_TYPE_FS1: default: @@ -739,7 +823,7 @@ void mod_freesync_build_vrr_params(struct mod_freesync *mod_freesync, return; in_out_vrr->state = in_config->state; - in_out_vrr->send_vsif = in_config->vsif_supported; + in_out_vrr->send_info_frame = in_config->vsif_supported; if (in_config->state == VRR_STATE_UNSUPPORTED) { in_out_vrr->state = VRR_STATE_UNSUPPORTED; @@ -972,7 +1056,7 @@ unsigned long long mod_freesync_calc_nominal_field_rate( unsigned long long nominal_field_rate_in_uhz = 0; /* Calculate nominal field rate for stream */ - nominal_field_rate_in_uhz = stream->timing.pix_clk_khz; + nominal_field_rate_in_uhz = stream->timing.pix_clk_100hz / 10; nominal_field_rate_in_uhz *= 1000ULL * 1000ULL * 1000ULL; nominal_field_rate_in_uhz = div_u64(nominal_field_rate_in_uhz, stream->timing.h_total); diff --git a/drivers/gpu/drm/amd/display/modules/inc/mod_freesync.h b/drivers/gpu/drm/amd/display/modules/inc/mod_freesync.h index 949a8b62aa98..4222e403b151 100644 --- a/drivers/gpu/drm/amd/display/modules/inc/mod_freesync.h +++ b/drivers/gpu/drm/amd/display/modules/inc/mod_freesync.h @@ -104,7 +104,7 @@ struct mod_vrr_params_fixed_refresh { struct mod_vrr_params { bool supported; - bool send_vsif; + bool send_info_frame; enum mod_vrr_state state; uint32_t min_refresh_in_uhz; diff --git a/drivers/gpu/drm/amd/display/modules/inc/mod_shared.h b/drivers/gpu/drm/amd/display/modules/inc/mod_shared.h index 1bd02c0ac30c..b711e7e6c204 100644 --- a/drivers/gpu/drm/amd/display/modules/inc/mod_shared.h +++ b/drivers/gpu/drm/amd/display/modules/inc/mod_shared.h @@ -41,7 +41,8 @@ enum color_transfer_func { enum vrr_packet_type { PACKET_TYPE_VRR, PACKET_TYPE_FS1, - PACKET_TYPE_FS2 + PACKET_TYPE_FS2, + PACKET_TYPE_VTEM }; diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c index 00f63b7dd32f..3ba87b076287 100644 --- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c +++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c @@ -41,6 +41,17 @@ static const unsigned char min_reduction_table[13] = { static const unsigned char max_reduction_table[13] = { 0xf5, 0xe5, 0xd9, 0xcd, 0xb1, 0xa5, 0xa5, 0x80, 0x65, 0x4d, 0x4d, 0x4d, 0x32}; +/* ABM 2.2 Min Reduction effectively disabled (100% for all configs)*/ +static const unsigned char min_reduction_table_v_2_2[13] = { +0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff}; + +/* Possible ABM 2.2 Max Reduction configs from least aggressive to most aggressive + * 0 1 2 3 4 5 6 7 8 9 10 11 12 + * 96.1 89.8 74.9 69.4 64.7 52.2 48.6 39.6 30.2 25.1 19.6 12.5 12.5 % + */ +static const unsigned char max_reduction_table_v_2_2[13] = { +0xf5, 0xe5, 0xbf, 0xb1, 0xa5, 0x85, 0x7c, 0x65, 0x4d, 0x40, 0x32, 0x20, 0x20}; + /* Predefined ABM configuration sets. We may have different configuration sets * in order to satisfy different power/quality requirements. */ @@ -56,7 +67,15 @@ static const unsigned char abm_config[abm_defines_max_config][abm_defines_max_le #define NUM_AGGR_LEVEL 4 #define NUM_POWER_FN_SEGS 8 #define NUM_BL_CURVE_SEGS 16 +#define IRAM_SIZE 256 +#define IRAM_RESERVE_AREA_START_V2 0xF0 // reserve 0xF0~0xF6 are write by DMCU only +#define IRAM_RESERVE_AREA_END_V2 0xF6 // reserve 0xF0~0xF6 are write by DMCU only + +#define IRAM_RESERVE_AREA_START_V2_2 0xF0 // reserve 0xF0~0xFF are write by DMCU only +#define IRAM_RESERVE_AREA_END_V2_2 0xFF // reserve 0xF0~0xFF are write by DMCU only + +#pragma pack(push, 1) /* NOTE: iRAM is 256B in size */ struct iram_table_v_2 { /* flags */ @@ -85,11 +104,10 @@ struct iram_table_v_2 { /* For reading PSR State directly from IRAM */ uint8_t psr_state; /* 0xf0 */ - uint8_t dmcu_interface_version; /* 0xf1 */ - uint8_t dmcu_date_version_year_b0; /* 0xf2 */ - uint8_t dmcu_date_version_year_b1; /* 0xf3 */ - uint8_t dmcu_date_version_month; /* 0xf4 */ - uint8_t dmcu_date_version_day; /* 0xf5 */ + uint8_t dmcu_mcp_interface_version; /* 0xf1 */ + uint8_t dmcu_abm_feature_version; /* 0xf2 */ + uint8_t dmcu_psr_feature_version; /* 0xf3 */ + uint16_t dmcu_version; /* 0xf4 */ uint8_t dmcu_state; /* 0xf6 */ uint16_t blRampReduction; /* 0xf7 */ @@ -101,6 +119,52 @@ struct iram_table_v_2 { uint8_t dummy9; /* 0xff */ }; +struct iram_table_v_2_2 { + /* flags */ + uint16_t flags; /* 0x00 U16 */ + + /* parameters for ABM2.2 algorithm */ + uint8_t min_reduction[NUM_AMBI_LEVEL][NUM_AGGR_LEVEL]; /* 0x02 U0.8 */ + uint8_t max_reduction[NUM_AMBI_LEVEL][NUM_AGGR_LEVEL]; /* 0x16 U0.8 */ + uint8_t bright_pos_gain[NUM_AMBI_LEVEL][NUM_AGGR_LEVEL]; /* 0x2a U2.6 */ + uint8_t dark_pos_gain[NUM_AMBI_LEVEL][NUM_AGGR_LEVEL]; /* 0x3e U2.6 */ + uint8_t hybridFactor[NUM_AGGR_LEVEL]; /* 0x52 U0.8 */ + uint8_t contrastFactor[NUM_AGGR_LEVEL]; /* 0x56 U0.8 */ + uint8_t deviation_gain[NUM_AGGR_LEVEL]; /* 0x5a U0.8 */ + uint8_t iir_curve[NUM_AMBI_LEVEL]; /* 0x5e U0.8 */ + uint8_t pad[29]; /* 0x63 U0.8 */ + + /* parameters for crgb conversion */ + uint16_t crgb_thresh[NUM_POWER_FN_SEGS]; /* 0x80 U3.13 */ + uint16_t crgb_offset[NUM_POWER_FN_SEGS]; /* 0x90 U1.15 */ + uint16_t crgb_slope[NUM_POWER_FN_SEGS]; /* 0xa0 U4.12 */ + + /* parameters for custom curve */ + /* thresholds for brightness --> backlight */ + uint16_t backlight_thresholds[NUM_BL_CURVE_SEGS]; /* 0xb0 U16.0 */ + /* offsets for brightness --> backlight */ + uint16_t backlight_offsets[NUM_BL_CURVE_SEGS]; /* 0xd0 U16.0 */ + + /* For reading PSR State directly from IRAM */ + uint8_t psr_state; /* 0xf0 */ + uint8_t dmcu_mcp_interface_version; /* 0xf1 */ + uint8_t dmcu_abm_feature_version; /* 0xf2 */ + uint8_t dmcu_psr_feature_version; /* 0xf3 */ + uint16_t dmcu_version; /* 0xf4 */ + uint8_t dmcu_state; /* 0xf6 */ + + uint8_t dummy1; /* 0xf7 */ + uint8_t dummy2; /* 0xf8 */ + uint8_t dummy3; /* 0xf9 */ + uint8_t dummy4; /* 0xfa */ + uint8_t dummy5; /* 0xfb */ + uint8_t dummy6; /* 0xfc */ + uint8_t dummy7; /* 0xfd */ + uint8_t dummy8; /* 0xfe */ + uint8_t dummy9; /* 0xff */ +}; +#pragma pack(pop) + static uint16_t backlight_8_to_16(unsigned int backlight_8bit) { return (uint16_t)(backlight_8bit * 0x101); @@ -143,11 +207,363 @@ static void fill_backlight_transform_table(struct dmcu_iram_parameters params, } } +static void fill_backlight_transform_table_v_2_2(struct dmcu_iram_parameters params, + struct iram_table_v_2_2 *table) +{ + unsigned int i; + unsigned int num_entries = NUM_BL_CURVE_SEGS; + unsigned int query_input_8bit; + unsigned int query_output_8bit; + unsigned int lut_index; + + table->backlight_thresholds[0] = 0; + table->backlight_offsets[0] = params.backlight_lut_array[0]; + table->backlight_thresholds[num_entries-1] = 0xFFFF; + table->backlight_offsets[num_entries-1] = + params.backlight_lut_array[params.backlight_lut_array_size - 1]; + + /* Setup all brightness levels between 0% and 100% exclusive + * Fills brightness-to-backlight transform table. Backlight custom curve + * describes transform from brightness to backlight. It will be defined + * as set of thresholds and set of offsets, together, implying + * extrapolation of custom curve into 16 uniformly spanned linear + * segments. Each threshold/offset represented by 16 bit entry in + * format U4.10. + */ + for (i = 1; i+1 < num_entries; i++) { + query_input_8bit = DIV_ROUNDUP((i * 256), num_entries); + + lut_index = (params.backlight_lut_array_size - 1) * i / (num_entries - 1); + ASSERT(lut_index < params.backlight_lut_array_size); + query_output_8bit = params.backlight_lut_array[lut_index] >> 8; + + table->backlight_thresholds[i] = + backlight_8_to_16(query_input_8bit); + table->backlight_offsets[i] = + backlight_8_to_16(query_output_8bit); + } +} + +void fill_iram_v_2(struct iram_table_v_2 *ram_table, struct dmcu_iram_parameters params) +{ + unsigned int set = params.set; + + ram_table->flags = 0x0; + ram_table->deviation_gain = 0xb3; + + ram_table->blRampReduction = + cpu_to_be16(params.backlight_ramping_reduction); + ram_table->blRampStart = + cpu_to_be16(params.backlight_ramping_start); + + ram_table->min_reduction[0][0] = min_reduction_table[abm_config[set][0]]; + ram_table->min_reduction[1][0] = min_reduction_table[abm_config[set][0]]; + ram_table->min_reduction[2][0] = min_reduction_table[abm_config[set][0]]; + ram_table->min_reduction[3][0] = min_reduction_table[abm_config[set][0]]; + ram_table->min_reduction[4][0] = min_reduction_table[abm_config[set][0]]; + ram_table->max_reduction[0][0] = max_reduction_table[abm_config[set][0]]; + ram_table->max_reduction[1][0] = max_reduction_table[abm_config[set][0]]; + ram_table->max_reduction[2][0] = max_reduction_table[abm_config[set][0]]; + ram_table->max_reduction[3][0] = max_reduction_table[abm_config[set][0]]; + ram_table->max_reduction[4][0] = max_reduction_table[abm_config[set][0]]; + + ram_table->min_reduction[0][1] = min_reduction_table[abm_config[set][1]]; + ram_table->min_reduction[1][1] = min_reduction_table[abm_config[set][1]]; + ram_table->min_reduction[2][1] = min_reduction_table[abm_config[set][1]]; + ram_table->min_reduction[3][1] = min_reduction_table[abm_config[set][1]]; + ram_table->min_reduction[4][1] = min_reduction_table[abm_config[set][1]]; + ram_table->max_reduction[0][1] = max_reduction_table[abm_config[set][1]]; + ram_table->max_reduction[1][1] = max_reduction_table[abm_config[set][1]]; + ram_table->max_reduction[2][1] = max_reduction_table[abm_config[set][1]]; + ram_table->max_reduction[3][1] = max_reduction_table[abm_config[set][1]]; + ram_table->max_reduction[4][1] = max_reduction_table[abm_config[set][1]]; + + ram_table->min_reduction[0][2] = min_reduction_table[abm_config[set][2]]; + ram_table->min_reduction[1][2] = min_reduction_table[abm_config[set][2]]; + ram_table->min_reduction[2][2] = min_reduction_table[abm_config[set][2]]; + ram_table->min_reduction[3][2] = min_reduction_table[abm_config[set][2]]; + ram_table->min_reduction[4][2] = min_reduction_table[abm_config[set][2]]; + ram_table->max_reduction[0][2] = max_reduction_table[abm_config[set][2]]; + ram_table->max_reduction[1][2] = max_reduction_table[abm_config[set][2]]; + ram_table->max_reduction[2][2] = max_reduction_table[abm_config[set][2]]; + ram_table->max_reduction[3][2] = max_reduction_table[abm_config[set][2]]; + ram_table->max_reduction[4][2] = max_reduction_table[abm_config[set][2]]; + + ram_table->min_reduction[0][3] = min_reduction_table[abm_config[set][3]]; + ram_table->min_reduction[1][3] = min_reduction_table[abm_config[set][3]]; + ram_table->min_reduction[2][3] = min_reduction_table[abm_config[set][3]]; + ram_table->min_reduction[3][3] = min_reduction_table[abm_config[set][3]]; + ram_table->min_reduction[4][3] = min_reduction_table[abm_config[set][3]]; + ram_table->max_reduction[0][3] = max_reduction_table[abm_config[set][3]]; + ram_table->max_reduction[1][3] = max_reduction_table[abm_config[set][3]]; + ram_table->max_reduction[2][3] = max_reduction_table[abm_config[set][3]]; + ram_table->max_reduction[3][3] = max_reduction_table[abm_config[set][3]]; + ram_table->max_reduction[4][3] = max_reduction_table[abm_config[set][3]]; + + ram_table->bright_pos_gain[0][0] = 0x20; + ram_table->bright_pos_gain[0][1] = 0x20; + ram_table->bright_pos_gain[0][2] = 0x20; + ram_table->bright_pos_gain[0][3] = 0x20; + ram_table->bright_pos_gain[1][0] = 0x20; + ram_table->bright_pos_gain[1][1] = 0x20; + ram_table->bright_pos_gain[1][2] = 0x20; + ram_table->bright_pos_gain[1][3] = 0x20; + ram_table->bright_pos_gain[2][0] = 0x20; + ram_table->bright_pos_gain[2][1] = 0x20; + ram_table->bright_pos_gain[2][2] = 0x20; + ram_table->bright_pos_gain[2][3] = 0x20; + ram_table->bright_pos_gain[3][0] = 0x20; + ram_table->bright_pos_gain[3][1] = 0x20; + ram_table->bright_pos_gain[3][2] = 0x20; + ram_table->bright_pos_gain[3][3] = 0x20; + ram_table->bright_pos_gain[4][0] = 0x20; + ram_table->bright_pos_gain[4][1] = 0x20; + ram_table->bright_pos_gain[4][2] = 0x20; + ram_table->bright_pos_gain[4][3] = 0x20; + ram_table->bright_neg_gain[0][1] = 0x00; + ram_table->bright_neg_gain[0][2] = 0x00; + ram_table->bright_neg_gain[0][3] = 0x00; + ram_table->bright_neg_gain[1][0] = 0x00; + ram_table->bright_neg_gain[1][1] = 0x00; + ram_table->bright_neg_gain[1][2] = 0x00; + ram_table->bright_neg_gain[1][3] = 0x00; + ram_table->bright_neg_gain[2][0] = 0x00; + ram_table->bright_neg_gain[2][1] = 0x00; + ram_table->bright_neg_gain[2][2] = 0x00; + ram_table->bright_neg_gain[2][3] = 0x00; + ram_table->bright_neg_gain[3][0] = 0x00; + ram_table->bright_neg_gain[3][1] = 0x00; + ram_table->bright_neg_gain[3][2] = 0x00; + ram_table->bright_neg_gain[3][3] = 0x00; + ram_table->bright_neg_gain[4][0] = 0x00; + ram_table->bright_neg_gain[4][1] = 0x00; + ram_table->bright_neg_gain[4][2] = 0x00; + ram_table->bright_neg_gain[4][3] = 0x00; + ram_table->dark_pos_gain[0][0] = 0x00; + ram_table->dark_pos_gain[0][1] = 0x00; + ram_table->dark_pos_gain[0][2] = 0x00; + ram_table->dark_pos_gain[0][3] = 0x00; + ram_table->dark_pos_gain[1][0] = 0x00; + ram_table->dark_pos_gain[1][1] = 0x00; + ram_table->dark_pos_gain[1][2] = 0x00; + ram_table->dark_pos_gain[1][3] = 0x00; + ram_table->dark_pos_gain[2][0] = 0x00; + ram_table->dark_pos_gain[2][1] = 0x00; + ram_table->dark_pos_gain[2][2] = 0x00; + ram_table->dark_pos_gain[2][3] = 0x00; + ram_table->dark_pos_gain[3][0] = 0x00; + ram_table->dark_pos_gain[3][1] = 0x00; + ram_table->dark_pos_gain[3][2] = 0x00; + ram_table->dark_pos_gain[3][3] = 0x00; + ram_table->dark_pos_gain[4][0] = 0x00; + ram_table->dark_pos_gain[4][1] = 0x00; + ram_table->dark_pos_gain[4][2] = 0x00; + ram_table->dark_pos_gain[4][3] = 0x00; + ram_table->dark_neg_gain[0][0] = 0x00; + ram_table->dark_neg_gain[0][1] = 0x00; + ram_table->dark_neg_gain[0][2] = 0x00; + ram_table->dark_neg_gain[0][3] = 0x00; + ram_table->dark_neg_gain[1][0] = 0x00; + ram_table->dark_neg_gain[1][1] = 0x00; + ram_table->dark_neg_gain[1][2] = 0x00; + ram_table->dark_neg_gain[1][3] = 0x00; + ram_table->dark_neg_gain[2][0] = 0x00; + ram_table->dark_neg_gain[2][1] = 0x00; + ram_table->dark_neg_gain[2][2] = 0x00; + ram_table->dark_neg_gain[2][3] = 0x00; + ram_table->dark_neg_gain[3][0] = 0x00; + ram_table->dark_neg_gain[3][1] = 0x00; + ram_table->dark_neg_gain[3][2] = 0x00; + ram_table->dark_neg_gain[3][3] = 0x00; + ram_table->dark_neg_gain[4][0] = 0x00; + ram_table->dark_neg_gain[4][1] = 0x00; + ram_table->dark_neg_gain[4][2] = 0x00; + ram_table->dark_neg_gain[4][3] = 0x00; + + ram_table->iir_curve[0] = 0x65; + ram_table->iir_curve[1] = 0x65; + ram_table->iir_curve[2] = 0x65; + ram_table->iir_curve[3] = 0x65; + ram_table->iir_curve[4] = 0x65; + + //Gamma 2.4 + ram_table->crgb_thresh[0] = cpu_to_be16(0x13b6); + ram_table->crgb_thresh[1] = cpu_to_be16(0x1648); + ram_table->crgb_thresh[2] = cpu_to_be16(0x18e3); + ram_table->crgb_thresh[3] = cpu_to_be16(0x1b41); + ram_table->crgb_thresh[4] = cpu_to_be16(0x1d46); + ram_table->crgb_thresh[5] = cpu_to_be16(0x1f21); + ram_table->crgb_thresh[6] = cpu_to_be16(0x2167); + ram_table->crgb_thresh[7] = cpu_to_be16(0x2384); + ram_table->crgb_offset[0] = cpu_to_be16(0x2999); + ram_table->crgb_offset[1] = cpu_to_be16(0x3999); + ram_table->crgb_offset[2] = cpu_to_be16(0x4666); + ram_table->crgb_offset[3] = cpu_to_be16(0x5999); + ram_table->crgb_offset[4] = cpu_to_be16(0x6333); + ram_table->crgb_offset[5] = cpu_to_be16(0x7800); + ram_table->crgb_offset[6] = cpu_to_be16(0x8c00); + ram_table->crgb_offset[7] = cpu_to_be16(0xa000); + ram_table->crgb_slope[0] = cpu_to_be16(0x3147); + ram_table->crgb_slope[1] = cpu_to_be16(0x2978); + ram_table->crgb_slope[2] = cpu_to_be16(0x23a2); + ram_table->crgb_slope[3] = cpu_to_be16(0x1f55); + ram_table->crgb_slope[4] = cpu_to_be16(0x1c63); + ram_table->crgb_slope[5] = cpu_to_be16(0x1a0f); + ram_table->crgb_slope[6] = cpu_to_be16(0x178d); + ram_table->crgb_slope[7] = cpu_to_be16(0x15ab); + + fill_backlight_transform_table( + params, ram_table); +} + +void fill_iram_v_2_2(struct iram_table_v_2_2 *ram_table, struct dmcu_iram_parameters params) +{ + unsigned int set = params.set; + + ram_table->flags = 0x0; + + ram_table->deviation_gain[0] = 0xb3; + ram_table->deviation_gain[1] = 0xb3; + ram_table->deviation_gain[2] = 0xb3; + ram_table->deviation_gain[3] = 0xb3; + + ram_table->min_reduction[0][0] = min_reduction_table_v_2_2[abm_config[set][0]]; + ram_table->min_reduction[1][0] = min_reduction_table_v_2_2[abm_config[set][0]]; + ram_table->min_reduction[2][0] = min_reduction_table_v_2_2[abm_config[set][0]]; + ram_table->min_reduction[3][0] = min_reduction_table_v_2_2[abm_config[set][0]]; + ram_table->min_reduction[4][0] = min_reduction_table_v_2_2[abm_config[set][0]]; + ram_table->max_reduction[0][0] = max_reduction_table_v_2_2[abm_config[set][0]]; + ram_table->max_reduction[1][0] = max_reduction_table_v_2_2[abm_config[set][0]]; + ram_table->max_reduction[2][0] = max_reduction_table_v_2_2[abm_config[set][0]]; + ram_table->max_reduction[3][0] = max_reduction_table_v_2_2[abm_config[set][0]]; + ram_table->max_reduction[4][0] = max_reduction_table_v_2_2[abm_config[set][0]]; + + ram_table->min_reduction[0][1] = min_reduction_table_v_2_2[abm_config[set][1]]; + ram_table->min_reduction[1][1] = min_reduction_table_v_2_2[abm_config[set][1]]; + ram_table->min_reduction[2][1] = min_reduction_table_v_2_2[abm_config[set][1]]; + ram_table->min_reduction[3][1] = min_reduction_table_v_2_2[abm_config[set][1]]; + ram_table->min_reduction[4][1] = min_reduction_table_v_2_2[abm_config[set][1]]; + ram_table->max_reduction[0][1] = max_reduction_table_v_2_2[abm_config[set][1]]; + ram_table->max_reduction[1][1] = max_reduction_table_v_2_2[abm_config[set][1]]; + ram_table->max_reduction[2][1] = max_reduction_table_v_2_2[abm_config[set][1]]; + ram_table->max_reduction[3][1] = max_reduction_table_v_2_2[abm_config[set][1]]; + ram_table->max_reduction[4][1] = max_reduction_table_v_2_2[abm_config[set][1]]; + + ram_table->min_reduction[0][2] = min_reduction_table_v_2_2[abm_config[set][2]]; + ram_table->min_reduction[1][2] = min_reduction_table_v_2_2[abm_config[set][2]]; + ram_table->min_reduction[2][2] = min_reduction_table_v_2_2[abm_config[set][2]]; + ram_table->min_reduction[3][2] = min_reduction_table_v_2_2[abm_config[set][2]]; + ram_table->min_reduction[4][2] = min_reduction_table_v_2_2[abm_config[set][2]]; + ram_table->max_reduction[0][2] = max_reduction_table_v_2_2[abm_config[set][2]]; + ram_table->max_reduction[1][2] = max_reduction_table_v_2_2[abm_config[set][2]]; + ram_table->max_reduction[2][2] = max_reduction_table_v_2_2[abm_config[set][2]]; + ram_table->max_reduction[3][2] = max_reduction_table_v_2_2[abm_config[set][2]]; + ram_table->max_reduction[4][2] = max_reduction_table_v_2_2[abm_config[set][2]]; + + ram_table->min_reduction[0][3] = min_reduction_table_v_2_2[abm_config[set][3]]; + ram_table->min_reduction[1][3] = min_reduction_table_v_2_2[abm_config[set][3]]; + ram_table->min_reduction[2][3] = min_reduction_table_v_2_2[abm_config[set][3]]; + ram_table->min_reduction[3][3] = min_reduction_table_v_2_2[abm_config[set][3]]; + ram_table->min_reduction[4][3] = min_reduction_table_v_2_2[abm_config[set][3]]; + ram_table->max_reduction[0][3] = max_reduction_table_v_2_2[abm_config[set][3]]; + ram_table->max_reduction[1][3] = max_reduction_table_v_2_2[abm_config[set][3]]; + ram_table->max_reduction[2][3] = max_reduction_table_v_2_2[abm_config[set][3]]; + ram_table->max_reduction[3][3] = max_reduction_table_v_2_2[abm_config[set][3]]; + ram_table->max_reduction[4][3] = max_reduction_table_v_2_2[abm_config[set][3]]; + + ram_table->bright_pos_gain[0][0] = 0x20; + ram_table->bright_pos_gain[0][1] = 0x20; + ram_table->bright_pos_gain[0][2] = 0x20; + ram_table->bright_pos_gain[0][3] = 0x20; + ram_table->bright_pos_gain[1][0] = 0x20; + ram_table->bright_pos_gain[1][1] = 0x20; + ram_table->bright_pos_gain[1][2] = 0x20; + ram_table->bright_pos_gain[1][3] = 0x20; + ram_table->bright_pos_gain[2][0] = 0x20; + ram_table->bright_pos_gain[2][1] = 0x20; + ram_table->bright_pos_gain[2][2] = 0x20; + ram_table->bright_pos_gain[2][3] = 0x20; + ram_table->bright_pos_gain[3][0] = 0x20; + ram_table->bright_pos_gain[3][1] = 0x20; + ram_table->bright_pos_gain[3][2] = 0x20; + ram_table->bright_pos_gain[3][3] = 0x20; + ram_table->bright_pos_gain[4][0] = 0x20; + ram_table->bright_pos_gain[4][1] = 0x20; + ram_table->bright_pos_gain[4][2] = 0x20; + ram_table->bright_pos_gain[4][3] = 0x20; + + ram_table->dark_pos_gain[0][0] = 0x00; + ram_table->dark_pos_gain[0][1] = 0x00; + ram_table->dark_pos_gain[0][2] = 0x00; + ram_table->dark_pos_gain[0][3] = 0x00; + ram_table->dark_pos_gain[1][0] = 0x00; + ram_table->dark_pos_gain[1][1] = 0x00; + ram_table->dark_pos_gain[1][2] = 0x00; + ram_table->dark_pos_gain[1][3] = 0x00; + ram_table->dark_pos_gain[2][0] = 0x00; + ram_table->dark_pos_gain[2][1] = 0x00; + ram_table->dark_pos_gain[2][2] = 0x00; + ram_table->dark_pos_gain[2][3] = 0x00; + ram_table->dark_pos_gain[3][0] = 0x00; + ram_table->dark_pos_gain[3][1] = 0x00; + ram_table->dark_pos_gain[3][2] = 0x00; + ram_table->dark_pos_gain[3][3] = 0x00; + ram_table->dark_pos_gain[4][0] = 0x00; + ram_table->dark_pos_gain[4][1] = 0x00; + ram_table->dark_pos_gain[4][2] = 0x00; + ram_table->dark_pos_gain[4][3] = 0x00; + + ram_table->hybridFactor[0] = 0xff; + ram_table->hybridFactor[1] = 0xff; + ram_table->hybridFactor[2] = 0xff; + ram_table->hybridFactor[3] = 0xc0; + + ram_table->contrastFactor[0] = 0x99; + ram_table->contrastFactor[1] = 0x99; + ram_table->contrastFactor[2] = 0x99; + ram_table->contrastFactor[3] = 0x80; + + ram_table->iir_curve[0] = 0x65; + ram_table->iir_curve[1] = 0x65; + ram_table->iir_curve[2] = 0x65; + ram_table->iir_curve[3] = 0x65; + ram_table->iir_curve[4] = 0x65; + + //Gamma 2.2 + ram_table->crgb_thresh[0] = cpu_to_be16(0x127c); + ram_table->crgb_thresh[1] = cpu_to_be16(0x151b); + ram_table->crgb_thresh[2] = cpu_to_be16(0x17d5); + ram_table->crgb_thresh[3] = cpu_to_be16(0x1a56); + ram_table->crgb_thresh[4] = cpu_to_be16(0x1c83); + ram_table->crgb_thresh[5] = cpu_to_be16(0x1e72); + ram_table->crgb_thresh[6] = cpu_to_be16(0x20f0); + ram_table->crgb_thresh[7] = cpu_to_be16(0x232b); + ram_table->crgb_offset[0] = cpu_to_be16(0x2999); + ram_table->crgb_offset[1] = cpu_to_be16(0x3999); + ram_table->crgb_offset[2] = cpu_to_be16(0x4666); + ram_table->crgb_offset[3] = cpu_to_be16(0x5999); + ram_table->crgb_offset[4] = cpu_to_be16(0x6333); + ram_table->crgb_offset[5] = cpu_to_be16(0x7800); + ram_table->crgb_offset[6] = cpu_to_be16(0x8c00); + ram_table->crgb_offset[7] = cpu_to_be16(0xa000); + ram_table->crgb_slope[0] = cpu_to_be16(0x3609); + ram_table->crgb_slope[1] = cpu_to_be16(0x2dfa); + ram_table->crgb_slope[2] = cpu_to_be16(0x27ea); + ram_table->crgb_slope[3] = cpu_to_be16(0x235d); + ram_table->crgb_slope[4] = cpu_to_be16(0x2042); + ram_table->crgb_slope[5] = cpu_to_be16(0x1dc3); + ram_table->crgb_slope[6] = cpu_to_be16(0x1b1a); + ram_table->crgb_slope[7] = cpu_to_be16(0x1910); + + fill_backlight_transform_table_v_2_2( + params, ram_table); +} + bool dmcu_load_iram(struct dmcu *dmcu, struct dmcu_iram_parameters params) { - struct iram_table_v_2 ram_table; - unsigned int set = params.set; + unsigned char ram_table[IRAM_SIZE]; + bool result = false; if (dmcu == NULL) return false; @@ -157,170 +573,23 @@ bool dmcu_load_iram(struct dmcu *dmcu, memset(&ram_table, 0, sizeof(ram_table)); - ram_table.flags = 0x0; - ram_table.deviation_gain = 0xb3; + if (dmcu->dmcu_version.abm_version == 0x22) { + fill_iram_v_2_2((struct iram_table_v_2_2 *)ram_table, params); - ram_table.blRampReduction = - cpu_to_be16(params.backlight_ramping_reduction); - ram_table.blRampStart = - cpu_to_be16(params.backlight_ramping_start); + result = dmcu->funcs->load_iram( + dmcu, 0, (char *)(&ram_table), IRAM_RESERVE_AREA_START_V2_2); + } else { + fill_iram_v_2((struct iram_table_v_2 *)ram_table, params); - ram_table.min_reduction[0][0] = min_reduction_table[abm_config[set][0]]; - ram_table.min_reduction[1][0] = min_reduction_table[abm_config[set][0]]; - ram_table.min_reduction[2][0] = min_reduction_table[abm_config[set][0]]; - ram_table.min_reduction[3][0] = min_reduction_table[abm_config[set][0]]; - ram_table.min_reduction[4][0] = min_reduction_table[abm_config[set][0]]; - ram_table.max_reduction[0][0] = max_reduction_table[abm_config[set][0]]; - ram_table.max_reduction[1][0] = max_reduction_table[abm_config[set][0]]; - ram_table.max_reduction[2][0] = max_reduction_table[abm_config[set][0]]; - ram_table.max_reduction[3][0] = max_reduction_table[abm_config[set][0]]; - ram_table.max_reduction[4][0] = max_reduction_table[abm_config[set][0]]; - - ram_table.min_reduction[0][1] = min_reduction_table[abm_config[set][1]]; - ram_table.min_reduction[1][1] = min_reduction_table[abm_config[set][1]]; - ram_table.min_reduction[2][1] = min_reduction_table[abm_config[set][1]]; - ram_table.min_reduction[3][1] = min_reduction_table[abm_config[set][1]]; - ram_table.min_reduction[4][1] = min_reduction_table[abm_config[set][1]]; - ram_table.max_reduction[0][1] = max_reduction_table[abm_config[set][1]]; - ram_table.max_reduction[1][1] = max_reduction_table[abm_config[set][1]]; - ram_table.max_reduction[2][1] = max_reduction_table[abm_config[set][1]]; - ram_table.max_reduction[3][1] = max_reduction_table[abm_config[set][1]]; - ram_table.max_reduction[4][1] = max_reduction_table[abm_config[set][1]]; - - ram_table.min_reduction[0][2] = min_reduction_table[abm_config[set][2]]; - ram_table.min_reduction[1][2] = min_reduction_table[abm_config[set][2]]; - ram_table.min_reduction[2][2] = min_reduction_table[abm_config[set][2]]; - ram_table.min_reduction[3][2] = min_reduction_table[abm_config[set][2]]; - ram_table.min_reduction[4][2] = min_reduction_table[abm_config[set][2]]; - ram_table.max_reduction[0][2] = max_reduction_table[abm_config[set][2]]; - ram_table.max_reduction[1][2] = max_reduction_table[abm_config[set][2]]; - ram_table.max_reduction[2][2] = max_reduction_table[abm_config[set][2]]; - ram_table.max_reduction[3][2] = max_reduction_table[abm_config[set][2]]; - ram_table.max_reduction[4][2] = max_reduction_table[abm_config[set][2]]; - - ram_table.min_reduction[0][3] = min_reduction_table[abm_config[set][3]]; - ram_table.min_reduction[1][3] = min_reduction_table[abm_config[set][3]]; - ram_table.min_reduction[2][3] = min_reduction_table[abm_config[set][3]]; - ram_table.min_reduction[3][3] = min_reduction_table[abm_config[set][3]]; - ram_table.min_reduction[4][3] = min_reduction_table[abm_config[set][3]]; - ram_table.max_reduction[0][3] = max_reduction_table[abm_config[set][3]]; - ram_table.max_reduction[1][3] = max_reduction_table[abm_config[set][3]]; - ram_table.max_reduction[2][3] = max_reduction_table[abm_config[set][3]]; - ram_table.max_reduction[3][3] = max_reduction_table[abm_config[set][3]]; - ram_table.max_reduction[4][3] = max_reduction_table[abm_config[set][3]]; - - ram_table.bright_pos_gain[0][0] = 0x20; - ram_table.bright_pos_gain[0][1] = 0x20; - ram_table.bright_pos_gain[0][2] = 0x20; - ram_table.bright_pos_gain[0][3] = 0x20; - ram_table.bright_pos_gain[1][0] = 0x20; - ram_table.bright_pos_gain[1][1] = 0x20; - ram_table.bright_pos_gain[1][2] = 0x20; - ram_table.bright_pos_gain[1][3] = 0x20; - ram_table.bright_pos_gain[2][0] = 0x20; - ram_table.bright_pos_gain[2][1] = 0x20; - ram_table.bright_pos_gain[2][2] = 0x20; - ram_table.bright_pos_gain[2][3] = 0x20; - ram_table.bright_pos_gain[3][0] = 0x20; - ram_table.bright_pos_gain[3][1] = 0x20; - ram_table.bright_pos_gain[3][2] = 0x20; - ram_table.bright_pos_gain[3][3] = 0x20; - ram_table.bright_pos_gain[4][0] = 0x20; - ram_table.bright_pos_gain[4][1] = 0x20; - ram_table.bright_pos_gain[4][2] = 0x20; - ram_table.bright_pos_gain[4][3] = 0x20; - ram_table.bright_neg_gain[0][1] = 0x00; - ram_table.bright_neg_gain[0][2] = 0x00; - ram_table.bright_neg_gain[0][3] = 0x00; - ram_table.bright_neg_gain[1][0] = 0x00; - ram_table.bright_neg_gain[1][1] = 0x00; - ram_table.bright_neg_gain[1][2] = 0x00; - ram_table.bright_neg_gain[1][3] = 0x00; - ram_table.bright_neg_gain[2][0] = 0x00; - ram_table.bright_neg_gain[2][1] = 0x00; - ram_table.bright_neg_gain[2][2] = 0x00; - ram_table.bright_neg_gain[2][3] = 0x00; - ram_table.bright_neg_gain[3][0] = 0x00; - ram_table.bright_neg_gain[3][1] = 0x00; - ram_table.bright_neg_gain[3][2] = 0x00; - ram_table.bright_neg_gain[3][3] = 0x00; - ram_table.bright_neg_gain[4][0] = 0x00; - ram_table.bright_neg_gain[4][1] = 0x00; - ram_table.bright_neg_gain[4][2] = 0x00; - ram_table.bright_neg_gain[4][3] = 0x00; - ram_table.dark_pos_gain[0][0] = 0x00; - ram_table.dark_pos_gain[0][1] = 0x00; - ram_table.dark_pos_gain[0][2] = 0x00; - ram_table.dark_pos_gain[0][3] = 0x00; - ram_table.dark_pos_gain[1][0] = 0x00; - ram_table.dark_pos_gain[1][1] = 0x00; - ram_table.dark_pos_gain[1][2] = 0x00; - ram_table.dark_pos_gain[1][3] = 0x00; - ram_table.dark_pos_gain[2][0] = 0x00; - ram_table.dark_pos_gain[2][1] = 0x00; - ram_table.dark_pos_gain[2][2] = 0x00; - ram_table.dark_pos_gain[2][3] = 0x00; - ram_table.dark_pos_gain[3][0] = 0x00; - ram_table.dark_pos_gain[3][1] = 0x00; - ram_table.dark_pos_gain[3][2] = 0x00; - ram_table.dark_pos_gain[3][3] = 0x00; - ram_table.dark_pos_gain[4][0] = 0x00; - ram_table.dark_pos_gain[4][1] = 0x00; - ram_table.dark_pos_gain[4][2] = 0x00; - ram_table.dark_pos_gain[4][3] = 0x00; - ram_table.dark_neg_gain[0][0] = 0x00; - ram_table.dark_neg_gain[0][1] = 0x00; - ram_table.dark_neg_gain[0][2] = 0x00; - ram_table.dark_neg_gain[0][3] = 0x00; - ram_table.dark_neg_gain[1][0] = 0x00; - ram_table.dark_neg_gain[1][1] = 0x00; - ram_table.dark_neg_gain[1][2] = 0x00; - ram_table.dark_neg_gain[1][3] = 0x00; - ram_table.dark_neg_gain[2][0] = 0x00; - ram_table.dark_neg_gain[2][1] = 0x00; - ram_table.dark_neg_gain[2][2] = 0x00; - ram_table.dark_neg_gain[2][3] = 0x00; - ram_table.dark_neg_gain[3][0] = 0x00; - ram_table.dark_neg_gain[3][1] = 0x00; - ram_table.dark_neg_gain[3][2] = 0x00; - ram_table.dark_neg_gain[3][3] = 0x00; - ram_table.dark_neg_gain[4][0] = 0x00; - ram_table.dark_neg_gain[4][1] = 0x00; - ram_table.dark_neg_gain[4][2] = 0x00; - ram_table.dark_neg_gain[4][3] = 0x00; - ram_table.iir_curve[0] = 0x65; - ram_table.iir_curve[1] = 0x65; - ram_table.iir_curve[2] = 0x65; - ram_table.iir_curve[3] = 0x65; - ram_table.iir_curve[4] = 0x65; - ram_table.crgb_thresh[0] = cpu_to_be16(0x13b6); - ram_table.crgb_thresh[1] = cpu_to_be16(0x1648); - ram_table.crgb_thresh[2] = cpu_to_be16(0x18e3); - ram_table.crgb_thresh[3] = cpu_to_be16(0x1b41); - ram_table.crgb_thresh[4] = cpu_to_be16(0x1d46); - ram_table.crgb_thresh[5] = cpu_to_be16(0x1f21); - ram_table.crgb_thresh[6] = cpu_to_be16(0x2167); - ram_table.crgb_thresh[7] = cpu_to_be16(0x2384); - ram_table.crgb_offset[0] = cpu_to_be16(0x2999); - ram_table.crgb_offset[1] = cpu_to_be16(0x3999); - ram_table.crgb_offset[2] = cpu_to_be16(0x4666); - ram_table.crgb_offset[3] = cpu_to_be16(0x5999); - ram_table.crgb_offset[4] = cpu_to_be16(0x6333); - ram_table.crgb_offset[5] = cpu_to_be16(0x7800); - ram_table.crgb_offset[6] = cpu_to_be16(0x8c00); - ram_table.crgb_offset[7] = cpu_to_be16(0xa000); - ram_table.crgb_slope[0] = cpu_to_be16(0x3147); - ram_table.crgb_slope[1] = cpu_to_be16(0x2978); - ram_table.crgb_slope[2] = cpu_to_be16(0x23a2); - ram_table.crgb_slope[3] = cpu_to_be16(0x1f55); - ram_table.crgb_slope[4] = cpu_to_be16(0x1c63); - ram_table.crgb_slope[5] = cpu_to_be16(0x1a0f); - ram_table.crgb_slope[6] = cpu_to_be16(0x178d); - ram_table.crgb_slope[7] = cpu_to_be16(0x15ab); + result = dmcu->funcs->load_iram( + dmcu, 0, (char *)(&ram_table), IRAM_RESERVE_AREA_START_V2); - fill_backlight_transform_table( - params, &ram_table); + if (result) + result = dmcu->funcs->load_iram( + dmcu, IRAM_RESERVE_AREA_END_V2 + 1, + (char *)(&ram_table) + IRAM_RESERVE_AREA_END_V2 + 1, + sizeof(ram_table) - IRAM_RESERVE_AREA_END_V2 - 1); + } - return dmcu->funcs->load_iram( - dmcu, 0, (char *)(&ram_table), sizeof(ram_table)); + return result; } diff --git a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_6_1_offset.h b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_6_1_offset.h index 13d4de645190..d8e0dd192fdd 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_6_1_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_6_1_offset.h @@ -2247,6 +2247,8 @@ // addressBlock: nbio_nbif_rcc_strap_BIFDEC1[13440..14975] // base address: 0x3480 +#define mmRCC_BIF_STRAP0 0x0000 +#define mmRCC_BIF_STRAP0_BASE_IDX 2 #define mmRCC_DEV0_EPF0_STRAP0 0x000f #define mmRCC_DEV0_EPF0_STRAP0_BASE_IDX 2 diff --git a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_6_1_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_6_1_sh_mask.h index a02b67943372..29af5167cd00 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_6_1_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_6_1_sh_mask.h @@ -16838,6 +16838,10 @@ // addressBlock: nbio_nbif_rcc_strap_BIFDEC1[13440..14975] +//RCC_BIF_STRAP0 +#define RCC_BIF_STRAP0__STRAP_PX_CAPABLE__SHIFT 0x7 +#define RCC_BIF_STRAP0__STRAP_PX_CAPABLE_MASK 0x00000080L + //RCC_DEV0_EPF0_STRAP0 #define RCC_DEV0_EPF0_STRAP0__STRAP_DEVICE_ID_DEV0_F0__SHIFT 0x0 #define RCC_DEV0_EPF0_STRAP0__STRAP_MAJOR_REV_ID_DEV0_F0__SHIFT 0x10 diff --git a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_6_1_smn.h b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_6_1_smn.h new file mode 100644 index 000000000000..8c75669eb500 --- /dev/null +++ b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_6_1_smn.h @@ -0,0 +1,58 @@ +/* + * Copyright (C) 2019 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef _nbio_6_1_SMN_HEADER +#define _nbio_6_1_SMN_HEADER + + +#define smnCPM_CONTROL 0x11180460 +#define smnPCIE_CNTL2 0x11180070 +#define smnPCIE_CONFIG_CNTL 0x11180044 +#define smnPCIE_CI_CNTL 0x11180080 + + +#define smnPCIE_PERF_COUNT_CNTL 0x11180200 +#define smnPCIE_PERF_CNTL_TXCLK 0x11180204 +#define smnPCIE_PERF_COUNT0_TXCLK 0x11180208 +#define smnPCIE_PERF_COUNT1_TXCLK 0x1118020c +#define smnPCIE_PERF_CNTL_MST_R_CLK 0x11180210 +#define smnPCIE_PERF_COUNT0_MST_R_CLK 0x11180214 +#define smnPCIE_PERF_COUNT1_MST_R_CLK 0x11180218 +#define smnPCIE_PERF_CNTL_MST_C_CLK 0x1118021c +#define smnPCIE_PERF_COUNT0_MST_C_CLK 0x11180220 +#define smnPCIE_PERF_COUNT1_MST_C_CLK 0x11180224 +#define smnPCIE_PERF_CNTL_SLV_R_CLK 0x11180228 +#define smnPCIE_PERF_COUNT0_SLV_R_CLK 0x1118022c +#define smnPCIE_PERF_COUNT1_SLV_R_CLK 0x11180230 +#define smnPCIE_PERF_CNTL_SLV_S_C_CLK 0x11180234 +#define smnPCIE_PERF_COUNT0_SLV_S_C_CLK 0x11180238 +#define smnPCIE_PERF_COUNT1_SLV_S_C_CLK 0x1118023c +#define smnPCIE_PERF_CNTL_SLV_NS_C_CLK 0x11180240 +#define smnPCIE_PERF_COUNT0_SLV_NS_C_CLK 0x11180244 +#define smnPCIE_PERF_COUNT1_SLV_NS_C_CLK 0x11180248 +#define smnPCIE_PERF_CNTL_EVENT0_PORT_SEL 0x1118024c +#define smnPCIE_PERF_CNTL_EVENT1_PORT_SEL 0x11180250 +#define smnPCIE_PERF_CNTL_TXCLK2 0x11180254 +#define smnPCIE_PERF_COUNT0_TXCLK2 0x11180258 +#define smnPCIE_PERF_COUNT1_TXCLK2 0x1118025c + +#endif // _nbio_6_1_SMN_HEADER + diff --git a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_0_smn.h b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_0_smn.h new file mode 100644 index 000000000000..5563f0715896 --- /dev/null +++ b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_0_smn.h @@ -0,0 +1,54 @@ +/* + * Copyright (C) 2019 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef _nbio_7_0_SMN_HEADER +#define _nbio_7_0_SMN_HEADER + + +#define smnCPM_CONTROL 0x11180460 +#define smnPCIE_CNTL2 0x11180070 + +#define smnPCIE_PERF_COUNT_CNTL 0x11180200 +#define smnPCIE_PERF_CNTL_TXCLK 0x11180204 +#define smnPCIE_PERF_COUNT0_TXCLK 0x11180208 +#define smnPCIE_PERF_COUNT1_TXCLK 0x1118020c +#define smnPCIE_PERF_CNTL_MST_R_CLK 0x11180210 +#define smnPCIE_PERF_COUNT0_MST_R_CLK 0x11180214 +#define smnPCIE_PERF_COUNT1_MST_R_CLK 0x11180218 +#define smnPCIE_PERF_CNTL_MST_C_CLK 0x1118021c +#define smnPCIE_PERF_COUNT0_MST_C_CLK 0x11180220 +#define smnPCIE_PERF_COUNT1_MST_C_CLK 0x11180224 +#define smnPCIE_PERF_CNTL_SLV_R_CLK 0x11180228 +#define smnPCIE_PERF_COUNT0_SLV_R_CLK 0x1118022c +#define smnPCIE_PERF_COUNT1_SLV_R_CLK 0x11180230 +#define smnPCIE_PERF_CNTL_SLV_S_C_CLK 0x11180234 +#define smnPCIE_PERF_COUNT0_SLV_S_C_CLK 0x11180238 +#define smnPCIE_PERF_COUNT1_SLV_S_C_CLK 0x1118023c +#define smnPCIE_PERF_CNTL_SLV_NS_C_CLK 0x11180240 +#define smnPCIE_PERF_COUNT0_SLV_NS_C_CLK 0x11180244 +#define smnPCIE_PERF_COUNT1_SLV_NS_C_CLK 0x11180248 +#define smnPCIE_PERF_CNTL_EVENT0_PORT_SEL 0x1118024c +#define smnPCIE_PERF_CNTL_EVENT1_PORT_SEL 0x11180250 +#define smnPCIE_PERF_CNTL_TXCLK2 0x11180254 +#define smnPCIE_PERF_COUNT0_TXCLK2 0x11180258 +#define smnPCIE_PERF_COUNT1_TXCLK2 0x1118025c + +#endif // _nbio_7_0_SMN_HEADER diff --git a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_4_0_smn.h b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_4_0_smn.h new file mode 100644 index 000000000000..c1457d880c4d --- /dev/null +++ b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_4_0_smn.h @@ -0,0 +1,53 @@ +/* + * Copyright (C) 2019 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef _nbio_7_4_0_SMN_HEADER +#define _nbio_7_4_0_SMN_HEADER + + +#define smnNBIF_MGCG_CTRL_LCLK 0x1013a21c +#define smnCPM_CONTROL 0x11180460 +#define smnPCIE_CNTL2 0x11180070 +#define smnPCIE_CI_CNTL 0x11180080 + +#define smnPCIE_PERF_COUNT_CNTL 0x11180200 +#define smnPCIE_PERF_CNTL_TXCLK1 0x11180204 +#define smnPCIE_PERF_COUNT0_TXCLK1 0x11180208 +#define smnPCIE_PERF_COUNT1_TXCLK1 0x1118020c +#define smnPCIE_PERF_CNTL_TXCLK2 0x11180210 +#define smnPCIE_PERF_COUNT0_TXCLK2 0x11180214 +#define smnPCIE_PERF_COUNT1_TXCLK2 0x11180218 +#define smnPCIE_PERF_CNTL_TXCLK3 0x1118021c +#define smnPCIE_PERF_COUNT0_TXCLK3 0x11180220 +#define smnPCIE_PERF_COUNT1_TXCLK3 0x11180224 +#define smnPCIE_PERF_CNTL_TXCLK4 0x11180228 +#define smnPCIE_PERF_COUNT0_TXCLK4 0x1118022c +#define smnPCIE_PERF_COUNT1_TXCLK4 0x11180230 +#define smnPCIE_PERF_CNTL_SCLK1 0x11180234 +#define smnPCIE_PERF_COUNT0_SCLK1 0x11180238 +#define smnPCIE_PERF_COUNT1_SCLK1 0x1118023c +#define smnPCIE_PERF_CNTL_SCLK2 0x11180240 +#define smnPCIE_PERF_COUNT0_SCLK2 0x11180244 +#define smnPCIE_PERF_COUNT1_SCLK2 0x11180248 +#define smnPCIE_PERF_CNTL_EVENT_LC_PORT_SEL 0x1118024c +#define smnPCIE_PERF_CNTL_EVENT_CI_PORT_SEL 0x11180250 + +#endif // _nbio_7_4_0_SMN_HEADER diff --git a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_4_offset.h b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_4_offset.h index e932213f87f0..994e796a28d7 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_4_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_4_offset.h @@ -2567,6 +2567,8 @@ // addressBlock: nbio_nbif0_rcc_strap_BIFDEC1 // base address: 0x0 +#define mmRCC_BIF_STRAP0 0x0000 +#define mmRCC_BIF_STRAP0_BASE_IDX 2 #define mmRCC_DEV0_EPF0_STRAP0 0x0011 #define mmRCC_DEV0_EPF0_STRAP0_BASE_IDX 2 diff --git a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_4_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_4_sh_mask.h index d3704b438f2d..d467b939c971 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_4_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_4_sh_mask.h @@ -19690,6 +19690,9 @@ // addressBlock: nbio_nbif0_rcc_strap_BIFDEC1 +//RCC_BIF_STRAP0 +#define RCC_BIF_STRAP0__STRAP_PX_CAPABLE__SHIFT 0x7 +#define RCC_BIF_STRAP0__STRAP_PX_CAPABLE_MASK 0x00000080L //RCC_DEV0_EPF0_STRAP0 #define RCC_DEV0_EPF0_STRAP0__STRAP_DEVICE_ID_DEV0_F0__SHIFT 0x0 #define RCC_DEV0_EPF0_STRAP0__STRAP_MAJOR_REV_ID_DEV0_F0__SHIFT 0x10 diff --git a/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_offset.h b/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_offset.h index a9eb57a53e59..a485526f3a51 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/thm/thm_11_0_2_offset.h @@ -46,4 +46,7 @@ #define mmTHM_TCON_THERM_TRIP 0x0002 #define mmTHM_TCON_THERM_TRIP_BASE_IDX 0 +#define mmTHM_BACO_CNTL 0x0081 +#define mmTHM_BACO_CNTL_BASE_IDX 0 + #endif diff --git a/drivers/gpu/drm/amd/include/atombios.h b/drivers/gpu/drm/amd/include/atombios.h index 7931502fa54f..8ba21747b40a 100644 --- a/drivers/gpu/drm/amd/include/atombios.h +++ b/drivers/gpu/drm/amd/include/atombios.h @@ -4106,7 +4106,7 @@ typedef struct _ATOM_LCD_MODE_CONTROL_CAP typedef struct _ATOM_FAKE_EDID_PATCH_RECORD { UCHAR ucRecordType; - UCHAR ucFakeEDIDLength; // = 128 means EDID lenght is 128 bytes, otherwise the EDID length = ucFakeEDIDLength*128 + UCHAR ucFakeEDIDLength; // = 128 means EDID length is 128 bytes, otherwise the EDID length = ucFakeEDIDLength*128 UCHAR ucFakeEDIDString[1]; // This actually has ucFakeEdidLength elements. } ATOM_FAKE_EDID_PATCH_RECORD; diff --git a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h index 8154d67388cc..83d960110d23 100644 --- a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h +++ b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h @@ -34,7 +34,6 @@ struct pci_dev; -#define KFD_INTERFACE_VERSION 2 #define KGD_MAX_QUEUES 128 struct kfd_dev; @@ -330,56 +329,4 @@ struct kfd2kgd_calls { }; -/** - * struct kgd2kfd_calls - * - * @exit: Notifies amdkfd that kgd module is unloaded - * - * @probe: Notifies amdkfd about a probe done on a device in the kgd driver. - * - * @device_init: Initialize the newly probed device (if it is a device that - * amdkfd supports) - * - * @device_exit: Notifies amdkfd about a removal of a kgd device - * - * @suspend: Notifies amdkfd about a suspend action done to a kgd device - * - * @resume: Notifies amdkfd about a resume action done to a kgd device - * - * @quiesce_mm: Quiesce all user queue access to specified MM address space - * - * @resume_mm: Resume user queue access to specified MM address space - * - * @schedule_evict_and_restore_process: Schedules work queue that will prepare - * for safe eviction of KFD BOs that belong to the specified process. - * - * @pre_reset: Notifies amdkfd that amdgpu about to reset the gpu - * - * @post_reset: Notify amdkfd that amgpu successfully reseted the gpu - * - * This structure contains function callback pointers so the kgd driver - * will notify to the amdkfd about certain status changes. - * - */ -struct kgd2kfd_calls { - void (*exit)(void); - struct kfd_dev* (*probe)(struct kgd_dev *kgd, struct pci_dev *pdev, - const struct kfd2kgd_calls *f2g); - bool (*device_init)(struct kfd_dev *kfd, - const struct kgd2kfd_shared_resources *gpu_resources); - void (*device_exit)(struct kfd_dev *kfd); - void (*interrupt)(struct kfd_dev *kfd, const void *ih_ring_entry); - void (*suspend)(struct kfd_dev *kfd); - int (*resume)(struct kfd_dev *kfd); - int (*quiesce_mm)(struct mm_struct *mm); - int (*resume_mm)(struct mm_struct *mm); - int (*schedule_evict_and_restore_process)(struct mm_struct *mm, - struct dma_fence *fence); - int (*pre_reset)(struct kfd_dev *kfd); - int (*post_reset)(struct kfd_dev *kfd); -}; - -int kgd2kfd_init(unsigned interface_version, - const struct kgd2kfd_calls **g2f); - #endif /* KGD_KFD_INTERFACE_H_INCLUDED */ diff --git a/drivers/gpu/drm/amd/include/kgd_pp_interface.h b/drivers/gpu/drm/amd/include/kgd_pp_interface.h index 1479ea1dc3e7..2b579ba9b685 100644 --- a/drivers/gpu/drm/amd/include/kgd_pp_interface.h +++ b/drivers/gpu/drm/amd/include/kgd_pp_interface.h @@ -92,6 +92,9 @@ enum pp_clock_type { PP_SCLK, PP_MCLK, PP_PCIE, + PP_SOCCLK, + PP_FCLK, + PP_DCEFCLK, OD_SCLK, OD_MCLK, OD_VDDC_CURVE, @@ -127,12 +130,13 @@ enum amd_pp_task { }; enum PP_SMC_POWER_PROFILE { - PP_SMC_POWER_PROFILE_FULLSCREEN3D = 0x0, - PP_SMC_POWER_PROFILE_POWERSAVING = 0x1, - PP_SMC_POWER_PROFILE_VIDEO = 0x2, - PP_SMC_POWER_PROFILE_VR = 0x3, - PP_SMC_POWER_PROFILE_COMPUTE = 0x4, - PP_SMC_POWER_PROFILE_CUSTOM = 0x5, + PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT = 0x0, + PP_SMC_POWER_PROFILE_FULLSCREEN3D = 0x1, + PP_SMC_POWER_PROFILE_POWERSAVING = 0x2, + PP_SMC_POWER_PROFILE_VIDEO = 0x3, + PP_SMC_POWER_PROFILE_VR = 0x4, + PP_SMC_POWER_PROFILE_COMPUTE = 0x5, + PP_SMC_POWER_PROFILE_CUSTOM = 0x6, }; enum { @@ -280,6 +284,11 @@ struct amd_pm_funcs { int (*set_hard_min_dcefclk_by_freq)(void *handle, uint32_t clock); int (*set_hard_min_fclk_by_freq)(void *handle, uint32_t clock); int (*set_min_deep_sleep_dcefclk)(void *handle, uint32_t clock); + int (*get_asic_baco_capability)(void *handle, bool *cap); + int (*get_asic_baco_state)(void *handle, int *state); + int (*set_asic_baco_state)(void *handle, int state); + int (*get_ppfeature_status)(void *handle, char *buf); + int (*set_ppfeature_status)(void *handle, uint64_t ppfeature_masks); }; #endif diff --git a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c index 9bc27f468d5b..3f73f7cd18b9 100644 --- a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c +++ b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c @@ -1404,6 +1404,97 @@ static int pp_set_active_display_count(void *handle, uint32_t count) return ret; } +static int pp_get_asic_baco_capability(void *handle, bool *cap) +{ + struct pp_hwmgr *hwmgr = handle; + + if (!hwmgr) + return -EINVAL; + + if (!hwmgr->pm_en || !hwmgr->hwmgr_func->get_asic_baco_capability) + return 0; + + mutex_lock(&hwmgr->smu_lock); + hwmgr->hwmgr_func->get_asic_baco_capability(hwmgr, cap); + mutex_unlock(&hwmgr->smu_lock); + + return 0; +} + +static int pp_get_asic_baco_state(void *handle, int *state) +{ + struct pp_hwmgr *hwmgr = handle; + + if (!hwmgr) + return -EINVAL; + + if (!hwmgr->pm_en || !hwmgr->hwmgr_func->get_asic_baco_state) + return 0; + + mutex_lock(&hwmgr->smu_lock); + hwmgr->hwmgr_func->get_asic_baco_state(hwmgr, (enum BACO_STATE *)state); + mutex_unlock(&hwmgr->smu_lock); + + return 0; +} + +static int pp_set_asic_baco_state(void *handle, int state) +{ + struct pp_hwmgr *hwmgr = handle; + + if (!hwmgr) + return -EINVAL; + + if (!hwmgr->pm_en || !hwmgr->hwmgr_func->set_asic_baco_state) + return 0; + + mutex_lock(&hwmgr->smu_lock); + hwmgr->hwmgr_func->set_asic_baco_state(hwmgr, (enum BACO_STATE)state); + mutex_unlock(&hwmgr->smu_lock); + + return 0; +} + +static int pp_get_ppfeature_status(void *handle, char *buf) +{ + struct pp_hwmgr *hwmgr = handle; + int ret = 0; + + if (!hwmgr || !hwmgr->pm_en || !buf) + return -EINVAL; + + if (hwmgr->hwmgr_func->get_ppfeature_status == NULL) { + pr_info_ratelimited("%s was not implemented.\n", __func__); + return -EINVAL; + } + + mutex_lock(&hwmgr->smu_lock); + ret = hwmgr->hwmgr_func->get_ppfeature_status(hwmgr, buf); + mutex_unlock(&hwmgr->smu_lock); + + return ret; +} + +static int pp_set_ppfeature_status(void *handle, uint64_t ppfeature_masks) +{ + struct pp_hwmgr *hwmgr = handle; + int ret = 0; + + if (!hwmgr || !hwmgr->pm_en) + return -EINVAL; + + if (hwmgr->hwmgr_func->set_ppfeature_status == NULL) { + pr_info_ratelimited("%s was not implemented.\n", __func__); + return -EINVAL; + } + + mutex_lock(&hwmgr->smu_lock); + ret = hwmgr->hwmgr_func->set_ppfeature_status(hwmgr, ppfeature_masks); + mutex_unlock(&hwmgr->smu_lock); + + return ret; +} + static const struct amd_pm_funcs pp_dpm_funcs = { .load_firmware = pp_dpm_load_fw, .wait_for_fw_loading_complete = pp_dpm_fw_loading_complete, @@ -1454,4 +1545,9 @@ static const struct amd_pm_funcs pp_dpm_funcs = { .set_min_deep_sleep_dcefclk = pp_set_min_deep_sleep_dcefclk, .set_hard_min_dcefclk_by_freq = pp_set_hard_min_dcefclk_by_freq, .set_hard_min_fclk_by_freq = pp_set_hard_min_fclk_by_freq, + .get_asic_baco_capability = pp_get_asic_baco_capability, + .get_asic_baco_state = pp_get_asic_baco_state, + .set_asic_baco_state = pp_set_asic_baco_state, + .get_ppfeature_status = pp_get_ppfeature_status, + .set_ppfeature_status = pp_set_ppfeature_status, }; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/Makefile b/drivers/gpu/drm/amd/powerplay/hwmgr/Makefile index ade8973b6f4d..0b3c6d1d52e4 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/Makefile +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/Makefile @@ -35,7 +35,7 @@ HARDWARE_MGR = hwmgr.o processpptables.o \ vega12_thermal.o \ pp_overdriver.o smu_helper.o \ vega20_processpptables.o vega20_hwmgr.o vega20_powertune.o \ - vega20_thermal.o + vega20_thermal.o common_baco.o vega10_baco.o vega20_baco.o AMD_PP_HWMGR = $(addprefix $(AMD_PP_PATH)/hwmgr/,$(HARDWARE_MGR)) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/common_baco.c b/drivers/gpu/drm/amd/powerplay/hwmgr/common_baco.c new file mode 100644 index 000000000000..9c57c1f67749 --- /dev/null +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/common_baco.c @@ -0,0 +1,101 @@ +/* + * Copyright 2018 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#include "common_baco.h" + + +static bool baco_wait_register(struct pp_hwmgr *hwmgr, u32 reg, u32 mask, u32 value) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev); + u32 timeout = 5000, data; + + do { + msleep(1); + data = RREG32(reg); + timeout--; + } while (value != (data & mask) && (timeout != 0)); + + if (timeout == 0) + return false; + + return true; +} + +static bool baco_cmd_handler(struct pp_hwmgr *hwmgr, u32 command, u32 reg, u32 mask, + u32 shift, u32 value, u32 timeout) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev); + u32 data; + bool ret = true; + + switch (command) { + case CMD_WRITE: + WREG32(reg, value << shift); + break; + case CMD_READMODIFYWRITE: + data = RREG32(reg); + data = (data & (~mask)) | (value << shift); + WREG32(reg, data); + break; + case CMD_WAITFOR: + ret = baco_wait_register(hwmgr, reg, mask, value); + break; + case CMD_DELAY_MS: + if (timeout) + /* Delay in milli Seconds */ + msleep(timeout); + break; + case CMD_DELAY_US: + if (timeout) + /* Delay in micro Seconds */ + udelay(timeout); + break; + + default: + dev_warn(adev->dev, "Invalid BACO command.\n"); + ret = false; + } + + return ret; +} + +bool soc15_baco_program_registers(struct pp_hwmgr *hwmgr, + const struct soc15_baco_cmd_entry *entry, + const u32 array_size) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev); + u32 i, reg = 0; + + for (i = 0; i < array_size; i++) { + if ((entry[i].cmd == CMD_WRITE) || + (entry[i].cmd == CMD_READMODIFYWRITE) || + (entry[i].cmd == CMD_WAITFOR)) + reg = adev->reg_offset[entry[i].hwip][entry[i].inst][entry[i].seg] + + entry[i].reg_offset; + if (!baco_cmd_handler(hwmgr, entry[i].cmd, reg, entry[i].mask, + entry[i].shift, entry[i].val, entry[i].timeout)) + return false; + } + + return true; +} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2c_sw_engine_dce80.h b/drivers/gpu/drm/amd/powerplay/hwmgr/common_baco.h index 26355c088746..95296c916f4e 100644 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2c_sw_engine_dce80.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/common_baco.h @@ -1,5 +1,5 @@ /* - * Copyright 2012-15 Advanced Micro Devices, Inc. + * Copyright 2018 Advanced Micro Devices, Inc. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -19,25 +19,32 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. * - * Authors: AMD - * */ +#ifndef __COMMON_BOCO_H__ +#define __COMMON_BOCO_H__ +#include "hwmgr.h" -#ifndef __DAL_I2C_SW_ENGINE_DCE80_H__ -#define __DAL_I2C_SW_ENGINE_DCE80_H__ -struct i2c_sw_engine_dce80 { - struct i2c_sw_engine base; - uint32_t engine_id; +enum baco_cmd_type { + CMD_WRITE = 0, + CMD_READMODIFYWRITE, + CMD_WAITFOR, + CMD_DELAY_MS, + CMD_DELAY_US, }; -struct i2c_sw_engine_dce80_create_arg { - uint32_t engine_id; - uint32_t default_speed; - struct dc_context *ctx; +struct soc15_baco_cmd_entry { + enum baco_cmd_type cmd; + uint32_t hwip; + uint32_t inst; + uint32_t seg; + uint32_t reg_offset; + uint32_t mask; + uint32_t shift; + uint32_t timeout; + uint32_t val; }; - -struct i2c_engine *dal_i2c_sw_engine_dce80_create( - const struct i2c_sw_engine_dce80_create_arg *arg); - +extern bool soc15_baco_program_registers(struct pp_hwmgr *hwmgr, + const struct soc15_baco_cmd_entry *entry, + const u32 array_size); #endif diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c b/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c index 1f92a9f4c9e3..c1c51c115e57 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c @@ -154,15 +154,6 @@ int phm_powerdown_uvd(struct pp_hwmgr *hwmgr) return 0; } -int phm_enable_clock_power_gatings(struct pp_hwmgr *hwmgr) -{ - PHM_FUNC_CHECK(hwmgr); - - if (NULL != hwmgr->hwmgr_func->enable_clock_power_gating) - return hwmgr->hwmgr_func->enable_clock_power_gating(hwmgr); - - return 0; -} int phm_disable_clock_power_gatings(struct pp_hwmgr *hwmgr) { diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c index 0173d0480024..6cd6497c6fc2 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c @@ -64,17 +64,19 @@ static int ci_set_asic_special_caps(struct pp_hwmgr *hwmgr); static void hwmgr_init_workload_prority(struct pp_hwmgr *hwmgr) { - hwmgr->workload_prority[PP_SMC_POWER_PROFILE_FULLSCREEN3D] = 2; - hwmgr->workload_prority[PP_SMC_POWER_PROFILE_POWERSAVING] = 0; - hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VIDEO] = 1; - hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VR] = 3; - hwmgr->workload_prority[PP_SMC_POWER_PROFILE_COMPUTE] = 4; - - hwmgr->workload_setting[0] = PP_SMC_POWER_PROFILE_POWERSAVING; - hwmgr->workload_setting[1] = PP_SMC_POWER_PROFILE_VIDEO; - hwmgr->workload_setting[2] = PP_SMC_POWER_PROFILE_FULLSCREEN3D; - hwmgr->workload_setting[3] = PP_SMC_POWER_PROFILE_VR; - hwmgr->workload_setting[4] = PP_SMC_POWER_PROFILE_COMPUTE; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT] = 0; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_FULLSCREEN3D] = 1; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_POWERSAVING] = 2; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VIDEO] = 3; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VR] = 4; + hwmgr->workload_prority[PP_SMC_POWER_PROFILE_COMPUTE] = 5; + + hwmgr->workload_setting[0] = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT; + hwmgr->workload_setting[1] = PP_SMC_POWER_PROFILE_FULLSCREEN3D; + hwmgr->workload_setting[2] = PP_SMC_POWER_PROFILE_POWERSAVING; + hwmgr->workload_setting[3] = PP_SMC_POWER_PROFILE_VIDEO; + hwmgr->workload_setting[4] = PP_SMC_POWER_PROFILE_VR; + hwmgr->workload_setting[5] = PP_SMC_POWER_PROFILE_COMPUTE; } int hwmgr_early_init(struct pp_hwmgr *hwmgr) @@ -271,7 +273,7 @@ int hwmgr_hw_fini(struct pp_hwmgr *hwmgr) phm_stop_thermal_controller(hwmgr); psm_set_boot_states(hwmgr); - psm_adjust_power_state_dynamic(hwmgr, false, NULL); + psm_adjust_power_state_dynamic(hwmgr, true, NULL); phm_disable_dynamic_state_management(hwmgr); phm_disable_clock_power_gatings(hwmgr); @@ -293,7 +295,7 @@ int hwmgr_suspend(struct pp_hwmgr *hwmgr) ret = psm_set_boot_states(hwmgr); if (ret) return ret; - ret = psm_adjust_power_state_dynamic(hwmgr, false, NULL); + ret = psm_adjust_power_state_dynamic(hwmgr, true, NULL); if (ret) return ret; ret = phm_power_down_asic(hwmgr); @@ -323,7 +325,7 @@ int hwmgr_resume(struct pp_hwmgr *hwmgr) if (ret) return ret; - ret = psm_adjust_power_state_dynamic(hwmgr, false, NULL); + ret = psm_adjust_power_state_dynamic(hwmgr, true, NULL); return ret; } @@ -377,12 +379,12 @@ int hwmgr_handle_task(struct pp_hwmgr *hwmgr, enum amd_pp_task task_id, ret = psm_set_user_performance_state(hwmgr, requested_ui_label, &requested_ps); if (ret) return ret; - ret = psm_adjust_power_state_dynamic(hwmgr, false, requested_ps); + ret = psm_adjust_power_state_dynamic(hwmgr, true, requested_ps); break; } case AMD_PP_TASK_COMPLETE_INIT: case AMD_PP_TASK_READJUST_POWER_STATE: - ret = psm_adjust_power_state_dynamic(hwmgr, false, NULL); + ret = psm_adjust_power_state_dynamic(hwmgr, true, NULL); break; default: break; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.c b/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.c index 56437866d120..ce177d7f04cb 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.c @@ -256,16 +256,14 @@ static void power_state_management(struct pp_hwmgr *hwmgr, } } -int psm_adjust_power_state_dynamic(struct pp_hwmgr *hwmgr, bool skip, +int psm_adjust_power_state_dynamic(struct pp_hwmgr *hwmgr, bool skip_display_settings, struct pp_power_state *new_ps) { uint32_t index; long workload; - if (skip) - return 0; - - phm_display_configuration_changed(hwmgr); + if (!skip_display_settings) + phm_display_configuration_changed(hwmgr); if (hwmgr->ps) power_state_management(hwmgr, new_ps); @@ -276,9 +274,11 @@ int psm_adjust_power_state_dynamic(struct pp_hwmgr *hwmgr, bool skip, */ phm_apply_clock_adjust_rules(hwmgr); - phm_notify_smc_display_config_after_ps_adjustment(hwmgr); + if (!skip_display_settings) + phm_notify_smc_display_config_after_ps_adjustment(hwmgr); - if (!phm_force_dpm_levels(hwmgr, hwmgr->request_dpm_level)) + if ((hwmgr->request_dpm_level != hwmgr->dpm_level) && + !phm_force_dpm_levels(hwmgr, hwmgr->request_dpm_level)) hwmgr->dpm_level = hwmgr->request_dpm_level; if (hwmgr->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL) { diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.h b/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.h index fa1b6825036a..b62d55f1f289 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/pp_psm.h @@ -34,7 +34,7 @@ int psm_set_user_performance_state(struct pp_hwmgr *hwmgr, enum PP_StateUILabel label_id, struct pp_power_state **state); int psm_adjust_power_state_dynamic(struct pp_hwmgr *hwmgr, - bool skip, + bool skip_display_settings, struct pp_power_state *new_ps); #endif diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c index d91390459326..c8f5c00dd1e7 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/smu7_hwmgr.c @@ -77,8 +77,9 @@ #define PCIE_BUS_CLK 10000 #define TCLK (PCIE_BUS_CLK / 10) -static const struct profile_mode_setting smu7_profiling[6] = - {{1, 0, 100, 30, 1, 0, 100, 10}, +static const struct profile_mode_setting smu7_profiling[7] = + {{0, 0, 0, 0, 0, 0, 0, 0}, + {1, 0, 100, 30, 1, 0, 100, 10}, {1, 10, 0, 30, 0, 0, 0, 0}, {0, 0, 0, 0, 1, 10, 16, 31}, {1, 0, 11, 50, 1, 0, 100, 10}, @@ -4889,7 +4890,8 @@ static int smu7_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf) uint32_t i, size = 0; uint32_t len; - static const char *profile_name[6] = {"3D_FULL_SCREEN", + static const char *profile_name[7] = {"BOOTUP_DEFAULT", + "3D_FULL_SCREEN", "POWER_SAVING", "VIDEO", "VR", diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_baco.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_baco.c new file mode 100644 index 000000000000..f94dab27f486 --- /dev/null +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_baco.c @@ -0,0 +1,136 @@ +#include "amdgpu.h" +#include "soc15.h" +#include "soc15_hw_ip.h" +#include "vega10_ip_offset.h" +#include "soc15_common.h" +#include "vega10_inc.h" +#include "vega10_ppsmc.h" +#include "vega10_baco.h" + + + +static const struct soc15_baco_cmd_entry pre_baco_tbl[] = +{ + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(NBIF, 0, mmBIF_DOORBELL_CNTL), BIF_DOORBELL_CNTL__DOORBELL_MONITOR_EN_MASK, BIF_DOORBELL_CNTL__DOORBELL_MONITOR_EN__SHIFT, 0, 1}, + {CMD_WRITE, SOC15_REG_ENTRY(NBIF, 0, mmBIF_FB_EN), 0, 0, 0, 0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_DSTATE_BYPASS_MASK, BACO_CNTL__BACO_DSTATE_BYPASS__SHIFT, 0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_RST_INTR_MASK_MASK, BACO_CNTL__BACO_RST_INTR_MASK__SHIFT, 0, 1} +}; + +static const struct soc15_baco_cmd_entry enter_baco_tbl[] = +{ + {CMD_WAITFOR, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__SOC_DOMAIN_IDLE_MASK, THM_BACO_CNTL__SOC_DOMAIN_IDLE__SHIFT, 0xffffffff, 0x80000000}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_EN_MASK, BACO_CNTL__BACO_EN__SHIFT, 0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_BIF_LCLK_SWITCH_MASK, BACO_CNTL__BACO_BIF_LCLK_SWITCH__SHIFT, 0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_DUMMY_EN_MASK, BACO_CNTL__BACO_DUMMY_EN__SHIFT, 0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_SOC_VDCI_RESET_MASK, THM_BACO_CNTL__BACO_SOC_VDCI_RESET__SHIFT, 0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_SMNCLK_MUX_MASK, THM_BACO_CNTL__BACO_SMNCLK_MUX__SHIFT,0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_ISO_EN_MASK, THM_BACO_CNTL__BACO_ISO_EN__SHIFT, 0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_AEB_ISO_EN_MASK, THM_BACO_CNTL__BACO_AEB_ISO_EN__SHIFT,0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_ANA_ISO_EN_MASK, THM_BACO_CNTL__BACO_ANA_ISO_EN__SHIFT, 0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_SOC_REFCLK_OFF_MASK, THM_BACO_CNTL__BACO_SOC_REFCLK_OFF__SHIFT, 0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_POWER_OFF_MASK, BACO_CNTL__BACO_POWER_OFF__SHIFT, 0, 1}, + {CMD_DELAY_MS, 0, 0, 0, 0, 0, 0, 5, 0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_RESET_EN_MASK, THM_BACO_CNTL__BACO_RESET_EN__SHIFT, 0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_PWROKRAW_CNTL_MASK, THM_BACO_CNTL__BACO_PWROKRAW_CNTL__SHIFT, 0, 0}, + {CMD_WAITFOR, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_MODE_MASK, BACO_CNTL__BACO_MODE__SHIFT, 0xffffffff, 0x100} +}; + +static const struct soc15_baco_cmd_entry exit_baco_tbl[] = +{ + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_POWER_OFF_MASK, BACO_CNTL__BACO_POWER_OFF__SHIFT, 0, 0}, + {CMD_DELAY_MS, 0, 0, 0, 0, 0, 0, 10,0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_SOC_REFCLK_OFF_MASK, THM_BACO_CNTL__BACO_SOC_REFCLK_OFF__SHIFT, 0,0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_ANA_ISO_EN_MASK, THM_BACO_CNTL__BACO_ANA_ISO_EN__SHIFT, 0, 0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_AEB_ISO_EN_MASK, THM_BACO_CNTL__BACO_AEB_ISO_EN__SHIFT,0, 0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_ISO_EN_MASK, THM_BACO_CNTL__BACO_ISO_EN__SHIFT, 0, 0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_PWROKRAW_CNTL_MASK, THM_BACO_CNTL__BACO_PWROKRAW_CNTL__SHIFT, 0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_SMNCLK_MUX_MASK, THM_BACO_CNTL__BACO_SMNCLK_MUX__SHIFT, 0, 0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_SOC_VDCI_RESET_MASK, THM_BACO_CNTL__BACO_SOC_VDCI_RESET__SHIFT, 0, 0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_EXIT_MASK, THM_BACO_CNTL__BACO_EXIT__SHIFT, 0, 1}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_RESET_EN_MASK, THM_BACO_CNTL__BACO_RESET_EN__SHIFT, 0, 0}, + {CMD_WAITFOR, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_EXIT_MASK, 0, 0xffffffff, 0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(THM, 0, mmTHM_BACO_CNTL), THM_BACO_CNTL__BACO_SB_AXI_FENCE_MASK, THM_BACO_CNTL__BACO_SB_AXI_FENCE__SHIFT, 0, 0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_DUMMY_EN_MASK, BACO_CNTL__BACO_DUMMY_EN__SHIFT, 0, 0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_BIF_LCLK_SWITCH_MASK ,BACO_CNTL__BACO_BIF_LCLK_SWITCH__SHIFT, 0, 0}, + {CMD_READMODIFYWRITE, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_EN_MASK , BACO_CNTL__BACO_EN__SHIFT, 0,0}, + {CMD_WAITFOR, SOC15_REG_ENTRY(NBIF, 0, mmBACO_CNTL), BACO_CNTL__BACO_MODE_MASK, 0, 0xffffffff, 0} + }; + +static const struct soc15_baco_cmd_entry clean_baco_tbl[] = +{ + {CMD_WRITE, SOC15_REG_ENTRY(NBIF, 0, mmBIOS_SCRATCH_6), 0, 0, 0, 0}, + {CMD_WRITE, SOC15_REG_ENTRY(NBIF, 0, mmBIOS_SCRATCH_7), 0, 0, 0, 0}, +}; + +int vega10_baco_get_capability(struct pp_hwmgr *hwmgr, bool *cap) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev); + uint32_t reg, data; + + *cap = false; + if (!phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_BACO)) + return 0; + + WREG32(0x12074, 0xFFF0003B); + data = RREG32(0x12075); + + if (data == 0x1) { + reg = RREG32_SOC15(NBIF, 0, mmRCC_BIF_STRAP0); + + if (reg & RCC_BIF_STRAP0__STRAP_PX_CAPABLE_MASK) + *cap = true; + } + + return 0; +} + +int vega10_baco_get_state(struct pp_hwmgr *hwmgr, enum BACO_STATE *state) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev); + uint32_t reg; + + reg = RREG32_SOC15(NBIF, 0, mmBACO_CNTL); + + if (reg & BACO_CNTL__BACO_MODE_MASK) + /* gfx has already entered BACO state */ + *state = BACO_STATE_IN; + else + *state = BACO_STATE_OUT; + return 0; +} + +int vega10_baco_set_state(struct pp_hwmgr *hwmgr, enum BACO_STATE state) +{ + enum BACO_STATE cur_state; + + vega10_baco_get_state(hwmgr, &cur_state); + + if (cur_state == state) + /* aisc already in the target state */ + return 0; + + if (state == BACO_STATE_IN) { + if (soc15_baco_program_registers(hwmgr, pre_baco_tbl, + ARRAY_SIZE(pre_baco_tbl))) { + if (smum_send_msg_to_smc(hwmgr, PPSMC_MSG_EnterBaco)) + return -1; + + if (soc15_baco_program_registers(hwmgr, enter_baco_tbl, + ARRAY_SIZE(enter_baco_tbl))) + return 0; + } + } else if (state == BACO_STATE_OUT) { + /* HW requires at least 20ms between regulator off and on */ + msleep(20); + /* Execute Hardware BACO exit sequence */ + if (soc15_baco_program_registers(hwmgr, exit_baco_tbl, + ARRAY_SIZE(exit_baco_tbl))) { + if (soc15_baco_program_registers(hwmgr, clean_baco_tbl, + ARRAY_SIZE(clean_baco_tbl))) + return 0; + } + } + + return -1; +} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2caux_dce80.h b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_baco.h index 21908629e973..a93b1e6d1c66 100644 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce80/i2caux_dce80.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_baco.h @@ -1,5 +1,5 @@ /* - * Copyright 2012-15 Advanced Micro Devices, Inc. + * Copyright 2018 Advanced Micro Devices, Inc. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -19,20 +19,14 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. * - * Authors: AMD - * */ +#ifndef __VEGA10_BOCO_H__ +#define __VEGA10_BOCO_H__ +#include "hwmgr.h" +#include "common_baco.h" -#ifndef __DAL_I2C_AUX_DCE80_H__ -#define __DAL_I2C_AUX_DCE80_H__ - -struct i2caux_dce80 { - struct i2caux base; - /* indicate the I2C HW circular buffer is in use */ - bool i2c_hw_buffer_in_use; -}; - -struct i2caux *dal_i2caux_dce80_create( - struct dc_context *ctx); +extern int vega10_baco_get_capability(struct pp_hwmgr *hwmgr, bool *cap); +extern int vega10_baco_get_state(struct pp_hwmgr *hwmgr, enum BACO_STATE *state); +extern int vega10_baco_set_state(struct pp_hwmgr *hwmgr, enum BACO_STATE state); #endif diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c index 79c86247d0ac..5479125ff4f6 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.c @@ -48,6 +48,7 @@ #include "ppinterrupt.h" #include "pp_overdriver.h" #include "pp_thermal.h" +#include "vega10_baco.h" #include "smuio/smuio_9_0_offset.h" #include "smuio/smuio_9_0_sh_mask.h" @@ -71,6 +72,21 @@ static const uint32_t channel_number[] = {1, 2, 0, 4, 0, 8, 0, 16, 2}; #define DF_CS_AON0_DramBaseAddress0__IntLvAddrSel_MASK 0x00000700L #define DF_CS_AON0_DramBaseAddress0__DramBaseAddr_MASK 0xFFFFF000L +typedef enum { + CLK_SMNCLK = 0, + CLK_SOCCLK, + CLK_MP0CLK, + CLK_MP1CLK, + CLK_LCLK, + CLK_DCEFCLK, + CLK_VCLK, + CLK_DCLK, + CLK_ECLK, + CLK_UCLK, + CLK_GFXCLK, + CLK_COUNT, +} CLOCK_ID_e; + static const ULONG PhwVega10_Magic = (ULONG)(PHM_VIslands_Magic); struct vega10_power_state *cast_phw_vega10_power_state( @@ -804,9 +820,9 @@ static int vega10_hwmgr_backend_init(struct pp_hwmgr *hwmgr) hwmgr->backend = data; - hwmgr->workload_mask = 1 << hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VIDEO]; - hwmgr->power_profile_mode = PP_SMC_POWER_PROFILE_VIDEO; - hwmgr->default_power_profile_mode = PP_SMC_POWER_PROFILE_VIDEO; + hwmgr->workload_mask = 1 << hwmgr->workload_prority[PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT]; + hwmgr->power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT; + hwmgr->default_power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT; vega10_set_default_registry_data(hwmgr); data->disable_dpm_mask = 0xff; @@ -3485,6 +3501,17 @@ static int vega10_upload_dpm_bootup_level(struct pp_hwmgr *hwmgr) } } + if (!data->registry_data.socclk_dpm_key_disabled) { + if (data->smc_state_table.soc_boot_level != + data->dpm_table.soc_table.dpm_state.soft_min_level) { + smum_send_msg_to_smc_with_parameter(hwmgr, + PPSMC_MSG_SetSoftMinSocclkByIndex, + data->smc_state_table.soc_boot_level); + data->dpm_table.soc_table.dpm_state.soft_min_level = + data->smc_state_table.soc_boot_level; + } + } + return 0; } @@ -3516,6 +3543,17 @@ static int vega10_upload_dpm_max_level(struct pp_hwmgr *hwmgr) } } + if (!data->registry_data.socclk_dpm_key_disabled) { + if (data->smc_state_table.soc_max_level != + data->dpm_table.soc_table.dpm_state.soft_max_level) { + smum_send_msg_to_smc_with_parameter(hwmgr, + PPSMC_MSG_SetSoftMaxSocclkByIndex, + data->smc_state_table.soc_max_level); + data->dpm_table.soc_table.dpm_state.soft_max_level = + data->smc_state_table.soc_max_level; + } + } + return 0; } @@ -3541,6 +3579,10 @@ static int vega10_generate_dpm_level_enable_mask( vega10_find_lowest_dpm_level(&(data->dpm_table.mem_table)); data->smc_state_table.mem_max_level = vega10_find_highest_dpm_level(&(data->dpm_table.mem_table)); + data->smc_state_table.soc_boot_level = + vega10_find_lowest_dpm_level(&(data->dpm_table.soc_table)); + data->smc_state_table.soc_max_level = + vega10_find_highest_dpm_level(&(data->dpm_table.soc_table)); PP_ASSERT_WITH_CODE(!vega10_upload_dpm_bootup_level(hwmgr), "Attempt to upload DPM Bootup Levels Failed!", @@ -3555,6 +3597,9 @@ static int vega10_generate_dpm_level_enable_mask( for(i = data->smc_state_table.mem_boot_level; i < data->smc_state_table.mem_max_level; i++) data->dpm_table.mem_table.dpm_levels[i].enabled = true; + for (i = data->smc_state_table.soc_boot_level; i < data->smc_state_table.soc_max_level; i++) + data->dpm_table.soc_table.dpm_levels[i].enabled = true; + return 0; } @@ -4028,6 +4073,24 @@ static int vega10_force_clock_level(struct pp_hwmgr *hwmgr, break; + case PP_SOCCLK: + data->smc_state_table.soc_boot_level = mask ? (ffs(mask) - 1) : 0; + data->smc_state_table.soc_max_level = mask ? (fls(mask) - 1) : 0; + + PP_ASSERT_WITH_CODE(!vega10_upload_dpm_bootup_level(hwmgr), + "Failed to upload boot level to lowest!", + return -EINVAL); + + PP_ASSERT_WITH_CODE(!vega10_upload_dpm_max_level(hwmgr), + "Failed to upload dpm max level to highest!", + return -EINVAL); + + break; + + case PP_DCEFCLK: + pr_info("Setting DCEFCLK min/max dpm level is not supported!\n"); + break; + case PP_PCIE: default: break; @@ -4267,12 +4330,113 @@ static int vega10_set_watermarks_for_clocks_ranges(struct pp_hwmgr *hwmgr, return result; } +static int vega10_get_ppfeature_status(struct pp_hwmgr *hwmgr, char *buf) +{ + static const char *ppfeature_name[] = { + "DPM_PREFETCHER", + "GFXCLK_DPM", + "UCLK_DPM", + "SOCCLK_DPM", + "UVD_DPM", + "VCE_DPM", + "ULV", + "MP0CLK_DPM", + "LINK_DPM", + "DCEFCLK_DPM", + "AVFS", + "GFXCLK_DS", + "SOCCLK_DS", + "LCLK_DS", + "PPT", + "TDC", + "THERMAL", + "GFX_PER_CU_CG", + "RM", + "DCEFCLK_DS", + "ACDC", + "VR0HOT", + "VR1HOT", + "FW_CTF", + "LED_DISPLAY", + "FAN_CONTROL", + "FAST_PPT", + "DIDT", + "ACG", + "PCC_LIMIT"}; + static const char *output_title[] = { + "FEATURES", + "BITMASK", + "ENABLEMENT"}; + uint64_t features_enabled; + int i; + int ret = 0; + int size = 0; + + ret = vega10_get_enabled_smc_features(hwmgr, &features_enabled); + PP_ASSERT_WITH_CODE(!ret, + "[EnableAllSmuFeatures] Failed to get enabled smc features!", + return ret); + + size += sprintf(buf + size, "Current ppfeatures: 0x%016llx\n", features_enabled); + size += sprintf(buf + size, "%-19s %-22s %s\n", + output_title[0], + output_title[1], + output_title[2]); + for (i = 0; i < GNLD_FEATURES_MAX; i++) { + size += sprintf(buf + size, "%-19s 0x%016llx %6s\n", + ppfeature_name[i], + 1ULL << i, + (features_enabled & (1ULL << i)) ? "Y" : "N"); + } + + return size; +} + +static int vega10_set_ppfeature_status(struct pp_hwmgr *hwmgr, uint64_t new_ppfeature_masks) +{ + uint64_t features_enabled; + uint64_t features_to_enable; + uint64_t features_to_disable; + int ret = 0; + + if (new_ppfeature_masks >= (1ULL << GNLD_FEATURES_MAX)) + return -EINVAL; + + ret = vega10_get_enabled_smc_features(hwmgr, &features_enabled); + if (ret) + return ret; + + features_to_disable = + (features_enabled ^ new_ppfeature_masks) & features_enabled; + features_to_enable = + (features_enabled ^ new_ppfeature_masks) ^ features_to_disable; + + pr_debug("features_to_disable 0x%llx\n", features_to_disable); + pr_debug("features_to_enable 0x%llx\n", features_to_enable); + + if (features_to_disable) { + ret = vega10_enable_smc_features(hwmgr, false, features_to_disable); + if (ret) + return ret; + } + + if (features_to_enable) { + ret = vega10_enable_smc_features(hwmgr, true, features_to_enable); + if (ret) + return ret; + } + + return 0; +} + static int vega10_print_clock_levels(struct pp_hwmgr *hwmgr, enum pp_clock_type type, char *buf) { struct vega10_hwmgr *data = hwmgr->backend; struct vega10_single_dpm_table *sclk_table = &(data->dpm_table.gfx_table); struct vega10_single_dpm_table *mclk_table = &(data->dpm_table.mem_table); + struct vega10_single_dpm_table *soc_table = &(data->dpm_table.soc_table); + struct vega10_single_dpm_table *dcef_table = &(data->dpm_table.dcef_table); struct vega10_pcie_table *pcie_table = &(data->dpm_table.pcie_table); struct vega10_odn_clock_voltage_dependency_table *podn_vdd_dep = NULL; @@ -4303,6 +4467,32 @@ static int vega10_print_clock_levels(struct pp_hwmgr *hwmgr, i, mclk_table->dpm_levels[i].value / 100, (i == now) ? "*" : ""); break; + case PP_SOCCLK: + if (data->registry_data.socclk_dpm_key_disabled) + break; + + smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetCurrentSocclkIndex); + now = smum_get_argument(hwmgr); + + for (i = 0; i < soc_table->count; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, soc_table->dpm_levels[i].value / 100, + (i == now) ? "*" : ""); + break; + case PP_DCEFCLK: + if (data->registry_data.dcefclk_dpm_key_disabled) + break; + + smum_send_msg_to_smc_with_parameter(hwmgr, + PPSMC_MSG_GetClockFreqMHz, CLK_DCEFCLK); + now = smum_get_argument(hwmgr); + + for (i = 0; i < dcef_table->count; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, dcef_table->dpm_levels[i].value / 100, + (dcef_table->dpm_levels[i].value / 100 == now) ? + "*" : ""); + break; case PP_PCIE: smum_send_msg_to_smc(hwmgr, PPSMC_MSG_GetCurrentLinkIndex); now = smum_get_argument(hwmgr); @@ -4668,13 +4858,15 @@ static int vega10_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf) { struct vega10_hwmgr *data = hwmgr->backend; uint32_t i, size = 0; - static const uint8_t profile_mode_setting[5][4] = {{70, 60, 1, 3,}, + static const uint8_t profile_mode_setting[6][4] = {{70, 60, 0, 0,}, + {70, 60, 1, 3,}, {90, 60, 0, 0,}, {70, 60, 0, 0,}, {70, 90, 0, 0,}, {30, 60, 0, 6,}, }; - static const char *profile_name[6] = {"3D_FULL_SCREEN", + static const char *profile_name[7] = {"BOOTUP_DEFAULT", + "3D_FULL_SCREEN", "POWER_SAVING", "VIDEO", "VR", @@ -4978,6 +5170,12 @@ static const struct pp_hwmgr_func vega10_hwmgr_funcs = { .set_power_limit = vega10_set_power_limit, .odn_edit_dpm_table = vega10_odn_edit_dpm_table, .get_performance_level = vega10_get_performance_level, + .get_asic_baco_capability = vega10_baco_get_capability, + .get_asic_baco_state = vega10_baco_get_state, + .set_asic_baco_state = vega10_baco_set_state, + .enable_mgpu_fan_boost = vega10_enable_mgpu_fan_boost, + .get_ppfeature_status = vega10_get_ppfeature_status, + .set_ppfeature_status = vega10_set_ppfeature_status, }; int vega10_hwmgr_init(struct pp_hwmgr *hwmgr) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.h b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.h index 89870556de1b..f752b4ad0c8a 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_hwmgr.h @@ -199,6 +199,7 @@ struct vega10_smc_state_table { uint32_t vce_boot_level; uint32_t gfx_max_level; uint32_t mem_max_level; + uint32_t soc_max_level; uint8_t vr_hot_gpio; uint8_t ac_dc_gpio; uint8_t therm_out_gpio; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_pptable.h b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_pptable.h index b3e63003a789..c934e9612c1b 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_pptable.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_pptable.h @@ -282,6 +282,30 @@ typedef struct _ATOM_Vega10_Fan_Table_V2 { UCHAR ucFanMaxRPM; } ATOM_Vega10_Fan_Table_V2; +typedef struct _ATOM_Vega10_Fan_Table_V3 { + UCHAR ucRevId; + USHORT usFanOutputSensitivity; + USHORT usFanAcousticLimitRpm; + USHORT usThrottlingRPM; + USHORT usTargetTemperature; + USHORT usMinimumPWMLimit; + USHORT usTargetGfxClk; + USHORT usFanGainEdge; + USHORT usFanGainHotspot; + USHORT usFanGainLiquid; + USHORT usFanGainVrVddc; + USHORT usFanGainVrMvdd; + USHORT usFanGainPlx; + USHORT usFanGainHbm; + UCHAR ucEnableZeroRPM; + USHORT usFanStopTemperature; + USHORT usFanStartTemperature; + UCHAR ucFanParameters; + UCHAR ucFanMinRPM; + UCHAR ucFanMaxRPM; + USHORT usMGpuThrottlingRPM; +} ATOM_Vega10_Fan_Table_V3; + typedef struct _ATOM_Vega10_Thermal_Controller { UCHAR ucRevId; UCHAR ucType; /* one of ATOM_VEGA10_PP_THERMALCONTROLLER_*/ diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c index b8747a5c9204..b6767d74dc85 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_processpptables.c @@ -32,6 +32,7 @@ #include "vega10_pptable.h" #define NUM_DSPCLK_LEVELS 8 +#define VEGA10_ENGINECLOCK_HARDMAX 198000 static void set_hw_cap(struct pp_hwmgr *hwmgr, bool enable, enum phm_platform_caps cap) @@ -122,6 +123,7 @@ static int init_thermal_controller( const Vega10_PPTable_Generic_SubTable_Header *header; const ATOM_Vega10_Fan_Table *fan_table_v1; const ATOM_Vega10_Fan_Table_V2 *fan_table_v2; + const ATOM_Vega10_Fan_Table_V3 *fan_table_v3; thermal_controller = (ATOM_Vega10_Thermal_Controller *) (((unsigned long)powerplay_table) + @@ -206,7 +208,7 @@ static int init_thermal_controller( le16_to_cpu(fan_table_v1->usFanStopTemperature); hwmgr->thermal_controller.advanceFanControlParameters.usZeroRPMStartTemperature = le16_to_cpu(fan_table_v1->usFanStartTemperature); - } else if (header->ucRevId > 10) { + } else if (header->ucRevId == 0xb) { fan_table_v2 = (ATOM_Vega10_Fan_Table_V2 *)header; hwmgr->thermal_controller.fanInfo.ucTachometerPulsesPerRevolution = @@ -250,7 +252,54 @@ static int init_thermal_controller( le16_to_cpu(fan_table_v2->usFanStopTemperature); hwmgr->thermal_controller.advanceFanControlParameters.usZeroRPMStartTemperature = le16_to_cpu(fan_table_v2->usFanStartTemperature); + } else if (header->ucRevId > 0xb) { + fan_table_v3 = (ATOM_Vega10_Fan_Table_V3 *)header; + + hwmgr->thermal_controller.fanInfo.ucTachometerPulsesPerRevolution = + fan_table_v3->ucFanParameters & ATOM_VEGA10_PP_FANPARAMETERS_TACHOMETER_PULSES_PER_REVOLUTION_MASK; + hwmgr->thermal_controller.fanInfo.ulMinRPM = fan_table_v3->ucFanMinRPM * 100UL; + hwmgr->thermal_controller.fanInfo.ulMaxRPM = fan_table_v3->ucFanMaxRPM * 100UL; + phm_cap_set(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_MicrocodeFanControl); + hwmgr->thermal_controller.advanceFanControlParameters.usFanOutputSensitivity = + le16_to_cpu(fan_table_v3->usFanOutputSensitivity); + hwmgr->thermal_controller.advanceFanControlParameters.usMaxFanRPM = + fan_table_v3->ucFanMaxRPM * 100UL; + hwmgr->thermal_controller.advanceFanControlParameters.usFanRPMMaxLimit = + le16_to_cpu(fan_table_v3->usThrottlingRPM); + hwmgr->thermal_controller.advanceFanControlParameters.ulMinFanSCLKAcousticLimit = + le16_to_cpu(fan_table_v3->usFanAcousticLimitRpm); + hwmgr->thermal_controller.advanceFanControlParameters.usTMax = + le16_to_cpu(fan_table_v3->usTargetTemperature); + hwmgr->thermal_controller.advanceFanControlParameters.usPWMMin = + le16_to_cpu(fan_table_v3->usMinimumPWMLimit); + hwmgr->thermal_controller.advanceFanControlParameters.ulTargetGfxClk = + le16_to_cpu(fan_table_v3->usTargetGfxClk); + hwmgr->thermal_controller.advanceFanControlParameters.usFanGainEdge = + le16_to_cpu(fan_table_v3->usFanGainEdge); + hwmgr->thermal_controller.advanceFanControlParameters.usFanGainHotspot = + le16_to_cpu(fan_table_v3->usFanGainHotspot); + hwmgr->thermal_controller.advanceFanControlParameters.usFanGainLiquid = + le16_to_cpu(fan_table_v3->usFanGainLiquid); + hwmgr->thermal_controller.advanceFanControlParameters.usFanGainVrVddc = + le16_to_cpu(fan_table_v3->usFanGainVrVddc); + hwmgr->thermal_controller.advanceFanControlParameters.usFanGainVrMvdd = + le16_to_cpu(fan_table_v3->usFanGainVrMvdd); + hwmgr->thermal_controller.advanceFanControlParameters.usFanGainPlx = + le16_to_cpu(fan_table_v3->usFanGainPlx); + hwmgr->thermal_controller.advanceFanControlParameters.usFanGainHbm = + le16_to_cpu(fan_table_v3->usFanGainHbm); + + hwmgr->thermal_controller.advanceFanControlParameters.ucEnableZeroRPM = + fan_table_v3->ucEnableZeroRPM; + hwmgr->thermal_controller.advanceFanControlParameters.usZeroRPMStopTemperature = + le16_to_cpu(fan_table_v3->usFanStopTemperature); + hwmgr->thermal_controller.advanceFanControlParameters.usZeroRPMStartTemperature = + le16_to_cpu(fan_table_v3->usFanStartTemperature); + hwmgr->thermal_controller.advanceFanControlParameters.usMGpuThrottlingRPMLimit = + le16_to_cpu(fan_table_v3->usMGpuThrottlingRPM); } + return 0; } @@ -258,7 +307,26 @@ static int init_over_drive_limits( struct pp_hwmgr *hwmgr, const ATOM_Vega10_POWERPLAYTABLE *powerplay_table) { - hwmgr->platform_descriptor.overdriveLimit.engineClock = + const ATOM_Vega10_GFXCLK_Dependency_Table *gfxclk_dep_table = + (const ATOM_Vega10_GFXCLK_Dependency_Table *) + (((unsigned long) powerplay_table) + + le16_to_cpu(powerplay_table->usGfxclkDependencyTableOffset)); + bool is_acg_enabled = false; + ATOM_Vega10_GFXCLK_Dependency_Record_V2 *patom_record_v2; + + if (gfxclk_dep_table->ucRevId == 1) { + patom_record_v2 = + (ATOM_Vega10_GFXCLK_Dependency_Record_V2 *)gfxclk_dep_table->entries; + is_acg_enabled = + (bool)patom_record_v2[gfxclk_dep_table->ucNumEntries-1].ucACGEnable; + } + + if (powerplay_table->ulMaxODEngineClock > VEGA10_ENGINECLOCK_HARDMAX && + !is_acg_enabled) + hwmgr->platform_descriptor.overdriveLimit.engineClock = + VEGA10_ENGINECLOCK_HARDMAX; + else + hwmgr->platform_descriptor.overdriveLimit.engineClock = le32_to_cpu(powerplay_table->ulMaxODEngineClock); hwmgr->platform_descriptor.overdriveLimit.memoryClock = le32_to_cpu(powerplay_table->ulMaxODMemoryClock); diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.c index 3f807d6c95ce..ba8763daa380 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.c @@ -556,6 +556,43 @@ int vega10_thermal_setup_fan_table(struct pp_hwmgr *hwmgr) return ret; } +int vega10_enable_mgpu_fan_boost(struct pp_hwmgr *hwmgr) +{ + struct vega10_hwmgr *data = hwmgr->backend; + PPTable_t *table = &(data->smc_state_table.pp_table); + int ret; + + if (!data->smu_features[GNLD_FAN_CONTROL].supported) + return 0; + + if (!hwmgr->thermal_controller.advanceFanControlParameters. + usMGpuThrottlingRPMLimit) + return 0; + + table->FanThrottlingRpm = hwmgr->thermal_controller. + advanceFanControlParameters.usMGpuThrottlingRPMLimit; + + ret = smum_smc_table_manager(hwmgr, + (uint8_t *)(&(data->smc_state_table.pp_table)), + PPTABLE, false); + if (ret) { + pr_info("Failed to update fan control table in pptable!"); + return ret; + } + + ret = vega10_disable_fan_control_feature(hwmgr); + if (ret) { + pr_info("Attempt to disable SMC fan control feature failed!"); + return ret; + } + + ret = vega10_enable_fan_control_feature(hwmgr); + if (ret) + pr_info("Attempt to enable SMC fan control feature failed!"); + + return ret; +} + /** * Start the fan control on the SMC. * @param hwmgr the address of the powerplay hardware manager. diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.h b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.h index 21e7c4dfa2ca..4a0ede7c1f07 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega10_thermal.h @@ -73,6 +73,7 @@ extern int vega10_thermal_disable_alert(struct pp_hwmgr *hwmgr); extern int vega10_fan_ctrl_start_smc_fan_control(struct pp_hwmgr *hwmgr); extern int vega10_start_thermal_controller(struct pp_hwmgr *hwmgr, struct PP_TemperatureRange *range); +extern int vega10_enable_mgpu_fan_boost(struct pp_hwmgr *hwmgr); #endif diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c index 54364444ecd1..6c8e78611c03 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega12_hwmgr.c @@ -753,6 +753,22 @@ static int vega12_init_smc_table(struct pp_hwmgr *hwmgr) return 0; } +static int vega12_run_acg_btc(struct pp_hwmgr *hwmgr) +{ + uint32_t result; + + PP_ASSERT_WITH_CODE( + smum_send_msg_to_smc(hwmgr, PPSMC_MSG_RunAcgBtc) == 0, + "[Run_ACG_BTC] Attempt to run ACG BTC failed!", + return -EINVAL); + + result = smum_get_argument(hwmgr); + PP_ASSERT_WITH_CODE(result == 1, + "Failed to run ACG BTC!", return -EINVAL); + + return 0; +} + static int vega12_set_allowed_featuresmask(struct pp_hwmgr *hwmgr) { struct vega12_hwmgr *data = @@ -931,6 +947,11 @@ static int vega12_enable_dpm_tasks(struct pp_hwmgr *hwmgr) "Failed to initialize SMC table!", result = tmp_result); + tmp_result = vega12_run_acg_btc(hwmgr); + PP_ASSERT_WITH_CODE(!tmp_result, + "Failed to run ACG BTC!", + result = tmp_result); + result = vega12_enable_all_smu_features(hwmgr); PP_ASSERT_WITH_CODE(!result, "Failed to enable all smu features!", @@ -1072,6 +1093,16 @@ static int vega12_upload_dpm_min_level(struct pp_hwmgr *hwmgr) return ret); } + if (data->smu_features[GNLD_DPM_DCEFCLK].enabled) { + min_freq = data->dpm_table.dcef_table.dpm_state.hard_min_level; + + PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( + hwmgr, PPSMC_MSG_SetHardMinByFreq, + (PPCLK_DCEFCLK << 16) | (min_freq & 0xffff))), + "Failed to set hard min dcefclk!", + return ret); + } + return ret; } @@ -1797,7 +1828,7 @@ static int vega12_force_clock_level(struct pp_hwmgr *hwmgr, enum pp_clock_type type, uint32_t mask) { struct vega12_hwmgr *data = (struct vega12_hwmgr *)(hwmgr->backend); - uint32_t soft_min_level, soft_max_level; + uint32_t soft_min_level, soft_max_level, hard_min_level; int ret = 0; switch (type) { @@ -1842,6 +1873,56 @@ static int vega12_force_clock_level(struct pp_hwmgr *hwmgr, break; + case PP_SOCCLK: + soft_min_level = mask ? (ffs(mask) - 1) : 0; + soft_max_level = mask ? (fls(mask) - 1) : 0; + + if (soft_max_level >= data->dpm_table.soc_table.count) { + pr_err("Clock level specified %d is over max allowed %d\n", + soft_max_level, + data->dpm_table.soc_table.count - 1); + return -EINVAL; + } + + data->dpm_table.soc_table.dpm_state.soft_min_level = + data->dpm_table.soc_table.dpm_levels[soft_min_level].value; + data->dpm_table.soc_table.dpm_state.soft_max_level = + data->dpm_table.soc_table.dpm_levels[soft_max_level].value; + + ret = vega12_upload_dpm_min_level(hwmgr); + PP_ASSERT_WITH_CODE(!ret, + "Failed to upload boot level to lowest!", + return ret); + + ret = vega12_upload_dpm_max_level(hwmgr); + PP_ASSERT_WITH_CODE(!ret, + "Failed to upload dpm max level to highest!", + return ret); + + break; + + case PP_DCEFCLK: + hard_min_level = mask ? (ffs(mask) - 1) : 0; + + if (hard_min_level >= data->dpm_table.dcef_table.count) { + pr_err("Clock level specified %d is over max allowed %d\n", + hard_min_level, + data->dpm_table.dcef_table.count - 1); + return -EINVAL; + } + + data->dpm_table.dcef_table.dpm_state.hard_min_level = + data->dpm_table.dcef_table.dpm_levels[hard_min_level].value; + + ret = vega12_upload_dpm_min_level(hwmgr); + PP_ASSERT_WITH_CODE(!ret, + "Failed to upload boot level to lowest!", + return ret); + + //TODO: Setting DCEFCLK max dpm level is not supported + + break; + case PP_PCIE: break; @@ -1852,6 +1933,104 @@ static int vega12_force_clock_level(struct pp_hwmgr *hwmgr, return 0; } +static int vega12_get_ppfeature_status(struct pp_hwmgr *hwmgr, char *buf) +{ + static const char *ppfeature_name[] = { + "DPM_PREFETCHER", + "GFXCLK_DPM", + "UCLK_DPM", + "SOCCLK_DPM", + "UVD_DPM", + "VCE_DPM", + "ULV", + "MP0CLK_DPM", + "LINK_DPM", + "DCEFCLK_DPM", + "GFXCLK_DS", + "SOCCLK_DS", + "LCLK_DS", + "PPT", + "TDC", + "THERMAL", + "GFX_PER_CU_CG", + "RM", + "DCEFCLK_DS", + "ACDC", + "VR0HOT", + "VR1HOT", + "FW_CTF", + "LED_DISPLAY", + "FAN_CONTROL", + "DIDT", + "GFXOFF", + "CG", + "ACG"}; + static const char *output_title[] = { + "FEATURES", + "BITMASK", + "ENABLEMENT"}; + uint64_t features_enabled; + int i; + int ret = 0; + int size = 0; + + ret = vega12_get_enabled_smc_features(hwmgr, &features_enabled); + PP_ASSERT_WITH_CODE(!ret, + "[EnableAllSmuFeatures] Failed to get enabled smc features!", + return ret); + + size += sprintf(buf + size, "Current ppfeatures: 0x%016llx\n", features_enabled); + size += sprintf(buf + size, "%-19s %-22s %s\n", + output_title[0], + output_title[1], + output_title[2]); + for (i = 0; i < GNLD_FEATURES_MAX; i++) { + size += sprintf(buf + size, "%-19s 0x%016llx %6s\n", + ppfeature_name[i], + 1ULL << i, + (features_enabled & (1ULL << i)) ? "Y" : "N"); + } + + return size; +} + +static int vega12_set_ppfeature_status(struct pp_hwmgr *hwmgr, uint64_t new_ppfeature_masks) +{ + uint64_t features_enabled; + uint64_t features_to_enable; + uint64_t features_to_disable; + int ret = 0; + + if (new_ppfeature_masks >= (1ULL << GNLD_FEATURES_MAX)) + return -EINVAL; + + ret = vega12_get_enabled_smc_features(hwmgr, &features_enabled); + if (ret) + return ret; + + features_to_disable = + (features_enabled ^ new_ppfeature_masks) & features_enabled; + features_to_enable = + (features_enabled ^ new_ppfeature_masks) ^ features_to_disable; + + pr_debug("features_to_disable 0x%llx\n", features_to_disable); + pr_debug("features_to_enable 0x%llx\n", features_to_enable); + + if (features_to_disable) { + ret = vega12_enable_smc_features(hwmgr, false, features_to_disable); + if (ret) + return ret; + } + + if (features_to_enable) { + ret = vega12_enable_smc_features(hwmgr, true, features_to_enable); + if (ret) + return ret; + } + + return 0; +} + static int vega12_print_clock_levels(struct pp_hwmgr *hwmgr, enum pp_clock_type type, char *buf) { @@ -1891,6 +2070,42 @@ static int vega12_print_clock_levels(struct pp_hwmgr *hwmgr, (clocks.data[i].clocks_in_khz / 1000 == now / 100) ? "*" : ""); break; + case PP_SOCCLK: + PP_ASSERT_WITH_CODE( + smum_send_msg_to_smc_with_parameter(hwmgr, + PPSMC_MSG_GetDpmClockFreq, (PPCLK_SOCCLK << 16)) == 0, + "Attempt to get Current SOCCLK Frequency Failed!", + return -EINVAL); + now = smum_get_argument(hwmgr); + + PP_ASSERT_WITH_CODE( + vega12_get_socclocks(hwmgr, &clocks) == 0, + "Attempt to get soc clk levels Failed!", + return -1); + for (i = 0; i < clocks.num_levels; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, clocks.data[i].clocks_in_khz / 1000, + (clocks.data[i].clocks_in_khz / 1000 == now) ? "*" : ""); + break; + + case PP_DCEFCLK: + PP_ASSERT_WITH_CODE( + smum_send_msg_to_smc_with_parameter(hwmgr, + PPSMC_MSG_GetDpmClockFreq, (PPCLK_DCEFCLK << 16)) == 0, + "Attempt to get Current DCEFCLK Frequency Failed!", + return -EINVAL); + now = smum_get_argument(hwmgr); + + PP_ASSERT_WITH_CODE( + vega12_get_dcefclocks(hwmgr, &clocks) == 0, + "Attempt to get dcef clk levels Failed!", + return -1); + for (i = 0; i < clocks.num_levels; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, clocks.data[i].clocks_in_khz / 1000, + (clocks.data[i].clocks_in_khz / 1000 == now) ? "*" : ""); + break; + case PP_PCIE: break; @@ -2411,6 +2626,8 @@ static const struct pp_hwmgr_func vega12_hwmgr_funcs = { .start_thermal_controller = vega12_start_thermal_controller, .powergate_gfx = vega12_gfx_off_control, .get_performance_level = vega12_get_performance_level, + .get_ppfeature_status = vega12_get_ppfeature_status, + .set_ppfeature_status = vega12_set_ppfeature_status, }; int vega12_hwmgr_init(struct pp_hwmgr *hwmgr) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_baco.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_baco.c new file mode 100644 index 000000000000..0d883b358df2 --- /dev/null +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_baco.c @@ -0,0 +1,81 @@ +#include "amdgpu.h" +#include "soc15.h" +#include "soc15_hw_ip.h" +#include "soc15_common.h" +#include "vega20_inc.h" +#include "vega20_ppsmc.h" +#include "vega20_baco.h" + + + +static const struct soc15_baco_cmd_entry clean_baco_tbl[] = +{ + {CMD_WRITE, SOC15_REG_ENTRY(NBIF, 0, mmBIOS_SCRATCH_6), 0, 0, 0, 0}, + {CMD_WRITE, SOC15_REG_ENTRY(NBIF, 0, mmBIOS_SCRATCH_7), 0, 0, 0, 0}, +}; + +int vega20_baco_get_capability(struct pp_hwmgr *hwmgr, bool *cap) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev); + uint32_t reg; + + *cap = false; + if (!phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_BACO)) + return 0; + + if (((RREG32(0x17569) & 0x20000000) >> 29) == 0x1) { + reg = RREG32_SOC15(NBIF, 0, mmRCC_BIF_STRAP0); + + if (reg & RCC_BIF_STRAP0__STRAP_PX_CAPABLE_MASK) + *cap = true; + } + + return 0; +} + +int vega20_baco_get_state(struct pp_hwmgr *hwmgr, enum BACO_STATE *state) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev); + uint32_t reg; + + reg = RREG32_SOC15(NBIF, 0, mmBACO_CNTL); + + if (reg & BACO_CNTL__BACO_MODE_MASK) + /* gfx has already entered BACO state */ + *state = BACO_STATE_IN; + else + *state = BACO_STATE_OUT; + return 0; +} + +int vega20_baco_set_state(struct pp_hwmgr *hwmgr, enum BACO_STATE state) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev); + enum BACO_STATE cur_state; + uint32_t data; + + vega20_baco_get_state(hwmgr, &cur_state); + + if (cur_state == state) + /* aisc already in the target state */ + return 0; + + if (state == BACO_STATE_IN) { + data = RREG32_SOC15(THM, 0, mmTHM_BACO_CNTL); + data |= 0x80000000; + WREG32_SOC15(THM, 0, mmTHM_BACO_CNTL, data); + + + if(smum_send_msg_to_smc_with_parameter(hwmgr, PPSMC_MSG_EnterBaco, 0)) + return -1; + + } else if (state == BACO_STATE_OUT) { + if (smum_send_msg_to_smc(hwmgr, PPSMC_MSG_ExitBaco)) + return -1; + if (!soc15_baco_program_registers(hwmgr, clean_baco_tbl, + ARRAY_SIZE(clean_baco_tbl))) + return -1; + } + + return 0; +} diff --git a/drivers/gpu/drm/amd/display/dc/i2caux/dce120/i2caux_dce120.h b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_baco.h index b6ac47617c70..c51988a9ed77 100644 --- a/drivers/gpu/drm/amd/display/dc/i2caux/dce120/i2caux_dce120.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_baco.h @@ -1,5 +1,5 @@ /* - * Copyright 2012-16 Advanced Micro Devices, Inc. + * Copyright 2018 Advanced Micro Devices, Inc. * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), @@ -19,14 +19,14 @@ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR * OTHER DEALINGS IN THE SOFTWARE. * - * Authors: AMD - * */ +#ifndef __VEGA20_BOCO_H__ +#define __VEGA20_BOCO_H__ +#include "hwmgr.h" +#include "common_baco.h" -#ifndef __DAL_I2C_AUX_DCE120_H__ -#define __DAL_I2C_AUX_DCE120_H__ - -struct i2caux *dal_i2caux_dce120_create( - struct dc_context *ctx); +extern int vega20_baco_get_capability(struct pp_hwmgr *hwmgr, bool *cap); +extern int vega20_baco_get_state(struct pp_hwmgr *hwmgr, enum BACO_STATE *state); +extern int vega20_baco_set_state(struct pp_hwmgr *hwmgr, enum BACO_STATE state); -#endif /* __DAL_I2C_AUX_DCE120_H__ */ +#endif diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c index 26154f9b2178..0769b1ec562b 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.c @@ -47,6 +47,7 @@ #include "pp_overdriver.h" #include "pp_thermal.h" #include "soc15_common.h" +#include "vega20_baco.h" #include "smuio/smuio_9_0_offset.h" #include "smuio/smuio_9_0_sh_mask.h" #include "nbio/nbio_7_4_sh_mask.h" @@ -390,9 +391,9 @@ static int vega20_hwmgr_backend_init(struct pp_hwmgr *hwmgr) hwmgr->backend = data; - hwmgr->workload_mask = 1 << hwmgr->workload_prority[PP_SMC_POWER_PROFILE_VIDEO]; - hwmgr->power_profile_mode = PP_SMC_POWER_PROFILE_VIDEO; - hwmgr->default_power_profile_mode = PP_SMC_POWER_PROFILE_VIDEO; + hwmgr->workload_mask = 1 << hwmgr->workload_prority[PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT]; + hwmgr->power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT; + hwmgr->default_power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT; vega20_set_default_registry_data(hwmgr); @@ -770,6 +771,54 @@ static int vega20_init_smc_table(struct pp_hwmgr *hwmgr) return 0; } +/* + * Override PCIe link speed and link width for DPM Level 1. PPTable entries + * reflect the ASIC capabilities and not the system capabilities. For e.g. + * Vega20 board in a PCI Gen3 system. In this case, when SMU's tries to switch + * to DPM1, it fails as system doesn't support Gen4. + */ +static int vega20_override_pcie_parameters(struct pp_hwmgr *hwmgr) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)(hwmgr->adev); + uint32_t pcie_gen = 0, pcie_width = 0, smu_pcie_arg; + int ret; + + if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN4) + pcie_gen = 3; + else if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3) + pcie_gen = 2; + else if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2) + pcie_gen = 1; + else if (adev->pm.pcie_gen_mask & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1) + pcie_gen = 0; + + if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X16) + pcie_width = 6; + else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X12) + pcie_width = 5; + else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X8) + pcie_width = 4; + else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X4) + pcie_width = 3; + else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X2) + pcie_width = 2; + else if (adev->pm.pcie_mlw_mask & CAIL_PCIE_LINK_WIDTH_SUPPORT_X1) + pcie_width = 1; + + /* Bit 31:16: LCLK DPM level. 0 is DPM0, and 1 is DPM1 + * Bit 15:8: PCIE GEN, 0 to 3 corresponds to GEN1 to GEN4 + * Bit 7:0: PCIE lane width, 1 to 7 corresponds is x1 to x32 + */ + smu_pcie_arg = (1 << 16) | (pcie_gen << 8) | pcie_width; + ret = smum_send_msg_to_smc_with_parameter(hwmgr, + PPSMC_MSG_OverridePcieParameters, smu_pcie_arg); + PP_ASSERT_WITH_CODE(!ret, + "[OverridePcieParameters] Attempt to override pcie params failed!", + return ret); + + return 0; +} + static int vega20_set_allowed_featuresmask(struct pp_hwmgr *hwmgr) { struct vega20_hwmgr *data = @@ -803,6 +852,11 @@ static int vega20_set_allowed_featuresmask(struct pp_hwmgr *hwmgr) return 0; } +static int vega20_run_btc(struct pp_hwmgr *hwmgr) +{ + return smum_send_msg_to_smc(hwmgr, PPSMC_MSG_RunBtc); +} + static int vega20_run_btc_afll(struct pp_hwmgr *hwmgr) { return smum_send_msg_to_smc(hwmgr, PPSMC_MSG_RunAfllBtc); @@ -980,6 +1034,9 @@ static int vega20_od8_set_feature_capabilities( pp_table->FanZeroRpmEnable) od_settings->overdrive8_capabilities |= OD8_FAN_ZERO_RPM_CONTROL; + if (!od_settings->overdrive8_capabilities) + hwmgr->od_enabled = false; + return 0; } @@ -1561,6 +1618,11 @@ static int vega20_enable_dpm_tasks(struct pp_hwmgr *hwmgr) "[EnableDPMTasks] Failed to initialize SMC table!", return result); + result = vega20_run_btc(hwmgr); + PP_ASSERT_WITH_CODE(!result, + "[EnableDPMTasks] Failed to run btc!", + return result); + result = vega20_run_btc_afll(hwmgr); PP_ASSERT_WITH_CODE(!result, "[EnableDPMTasks] Failed to run btc afll!", @@ -1571,6 +1633,11 @@ static int vega20_enable_dpm_tasks(struct pp_hwmgr *hwmgr) "[EnableDPMTasks] Failed to enable all smu features!", return result); + result = vega20_override_pcie_parameters(hwmgr); + PP_ASSERT_WITH_CODE(!result, + "[EnableDPMTasks] Failed to override pcie parameters!", + return result); + result = vega20_notify_smc_display_change(hwmgr); PP_ASSERT_WITH_CODE(!result, "[EnableDPMTasks] Failed to notify smc display change!", @@ -1689,13 +1756,6 @@ static int vega20_upload_dpm_min_level(struct pp_hwmgr *hwmgr, uint32_t feature_ (PPCLK_UCLK << 16) | (min_freq & 0xffff))), "Failed to set soft min memclk !", return ret); - - min_freq = data->dpm_table.mem_table.dpm_state.hard_min_level; - PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( - hwmgr, PPSMC_MSG_SetHardMinByFreq, - (PPCLK_UCLK << 16) | (min_freq & 0xffff))), - "Failed to set hard min memclk !", - return ret); } if (data->smu_features[GNLD_DPM_UVD].enabled && @@ -1739,6 +1799,28 @@ static int vega20_upload_dpm_min_level(struct pp_hwmgr *hwmgr, uint32_t feature_ return ret); } + if (data->smu_features[GNLD_DPM_FCLK].enabled && + (feature_mask & FEATURE_DPM_FCLK_MASK)) { + min_freq = data->dpm_table.fclk_table.dpm_state.soft_min_level; + + PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( + hwmgr, PPSMC_MSG_SetSoftMinByFreq, + (PPCLK_FCLK << 16) | (min_freq & 0xffff))), + "Failed to set soft min fclk!", + return ret); + } + + if (data->smu_features[GNLD_DPM_DCEFCLK].enabled && + (feature_mask & FEATURE_DPM_DCEFCLK_MASK)) { + min_freq = data->dpm_table.dcef_table.dpm_state.hard_min_level; + + PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( + hwmgr, PPSMC_MSG_SetHardMinByFreq, + (PPCLK_DCEFCLK << 16) | (min_freq & 0xffff))), + "Failed to set hard min dcefclk!", + return ret); + } + return ret; } @@ -1811,6 +1893,17 @@ static int vega20_upload_dpm_max_level(struct pp_hwmgr *hwmgr, uint32_t feature_ return ret); } + if (data->smu_features[GNLD_DPM_FCLK].enabled && + (feature_mask & FEATURE_DPM_FCLK_MASK)) { + max_freq = data->dpm_table.fclk_table.dpm_state.soft_max_level; + + PP_ASSERT_WITH_CODE(!(ret = smum_send_msg_to_smc_with_parameter( + hwmgr, PPSMC_MSG_SetSoftMaxByFreq, + (PPCLK_FCLK << 16) | (max_freq & 0xffff))), + "Failed to set soft max fclk!", + return ret); + } + return ret; } @@ -1918,16 +2011,36 @@ static uint32_t vega20_dpm_get_mclk(struct pp_hwmgr *hwmgr, bool low) return (mem_clk * 100); } +static int vega20_get_metrics_table(struct pp_hwmgr *hwmgr, SmuMetrics_t *metrics_table) +{ + struct vega20_hwmgr *data = + (struct vega20_hwmgr *)(hwmgr->backend); + int ret = 0; + + if (!data->metrics_time || time_after(jiffies, data->metrics_time + HZ / 2)) { + ret = smum_smc_table_manager(hwmgr, (uint8_t *)metrics_table, + TABLE_SMU_METRICS, true); + if (ret) { + pr_info("Failed to export SMU metrics table!\n"); + return ret; + } + memcpy(&data->metrics_table, metrics_table, sizeof(SmuMetrics_t)); + data->metrics_time = jiffies; + } else + memcpy(metrics_table, &data->metrics_table, sizeof(SmuMetrics_t)); + + return ret; +} + static int vega20_get_gpu_power(struct pp_hwmgr *hwmgr, uint32_t *query) { int ret = 0; SmuMetrics_t metrics_table; - ret = smum_smc_table_manager(hwmgr, (uint8_t *)&metrics_table, TABLE_SMU_METRICS, true); - PP_ASSERT_WITH_CODE(!ret, - "Failed to export SMU METRICS table!", - return ret); + ret = vega20_get_metrics_table(hwmgr, &metrics_table); + if (ret) + return ret; *query = metrics_table.CurrSocketPower << 8; @@ -1958,10 +2071,9 @@ static int vega20_get_current_activity_percent(struct pp_hwmgr *hwmgr, int ret = 0; SmuMetrics_t metrics_table; - ret = smum_smc_table_manager(hwmgr, (uint8_t *)&metrics_table, TABLE_SMU_METRICS, true); - PP_ASSERT_WITH_CODE(!ret, - "Failed to export SMU METRICS table!", - return ret); + ret = vega20_get_metrics_table(hwmgr, &metrics_table); + if (ret) + return ret; *activity_percent = metrics_table.AverageGfxActivity; @@ -1973,16 +2085,18 @@ static int vega20_read_sensor(struct pp_hwmgr *hwmgr, int idx, { struct vega20_hwmgr *data = (struct vega20_hwmgr *)(hwmgr->backend); struct amdgpu_device *adev = hwmgr->adev; + SmuMetrics_t metrics_table; uint32_t val_vid; int ret = 0; switch (idx) { case AMDGPU_PP_SENSOR_GFX_SCLK: - ret = vega20_get_current_clk_freq(hwmgr, - PPCLK_GFXCLK, - (uint32_t *)value); - if (!ret) - *size = 4; + ret = vega20_get_metrics_table(hwmgr, &metrics_table); + if (ret) + return ret; + + *((uint32_t *)value) = metrics_table.AverageGfxclkFrequency * 100; + *size = 4; break; case AMDGPU_PP_SENSOR_GFX_MCLK: ret = vega20_get_current_clk_freq(hwmgr, @@ -2140,6 +2254,12 @@ static int vega20_force_dpm_highest(struct pp_hwmgr *hwmgr) data->dpm_table.mem_table.dpm_state.soft_max_level = data->dpm_table.mem_table.dpm_levels[soft_level].value; + soft_level = vega20_find_highest_dpm_level(&(data->dpm_table.soc_table)); + + data->dpm_table.soc_table.dpm_state.soft_min_level = + data->dpm_table.soc_table.dpm_state.soft_max_level = + data->dpm_table.soc_table.dpm_levels[soft_level].value; + ret = vega20_upload_dpm_min_level(hwmgr, 0xFFFFFFFF); PP_ASSERT_WITH_CODE(!ret, "Failed to upload boot level to highest!", @@ -2172,6 +2292,12 @@ static int vega20_force_dpm_lowest(struct pp_hwmgr *hwmgr) data->dpm_table.mem_table.dpm_state.soft_max_level = data->dpm_table.mem_table.dpm_levels[soft_level].value; + soft_level = vega20_find_lowest_dpm_level(&(data->dpm_table.soc_table)); + + data->dpm_table.soc_table.dpm_state.soft_min_level = + data->dpm_table.soc_table.dpm_state.soft_max_level = + data->dpm_table.soc_table.dpm_levels[soft_level].value; + ret = vega20_upload_dpm_min_level(hwmgr, 0xFFFFFFFF); PP_ASSERT_WITH_CODE(!ret, "Failed to upload boot level to highest!", @@ -2188,8 +2314,32 @@ static int vega20_force_dpm_lowest(struct pp_hwmgr *hwmgr) static int vega20_unforce_dpm_levels(struct pp_hwmgr *hwmgr) { + struct vega20_hwmgr *data = + (struct vega20_hwmgr *)(hwmgr->backend); + uint32_t soft_min_level, soft_max_level; int ret = 0; + soft_min_level = vega20_find_lowest_dpm_level(&(data->dpm_table.gfx_table)); + soft_max_level = vega20_find_highest_dpm_level(&(data->dpm_table.gfx_table)); + data->dpm_table.gfx_table.dpm_state.soft_min_level = + data->dpm_table.gfx_table.dpm_levels[soft_min_level].value; + data->dpm_table.gfx_table.dpm_state.soft_max_level = + data->dpm_table.gfx_table.dpm_levels[soft_max_level].value; + + soft_min_level = vega20_find_lowest_dpm_level(&(data->dpm_table.mem_table)); + soft_max_level = vega20_find_highest_dpm_level(&(data->dpm_table.mem_table)); + data->dpm_table.mem_table.dpm_state.soft_min_level = + data->dpm_table.mem_table.dpm_levels[soft_min_level].value; + data->dpm_table.mem_table.dpm_state.soft_max_level = + data->dpm_table.mem_table.dpm_levels[soft_max_level].value; + + soft_min_level = vega20_find_lowest_dpm_level(&(data->dpm_table.soc_table)); + soft_max_level = vega20_find_highest_dpm_level(&(data->dpm_table.soc_table)); + data->dpm_table.soc_table.dpm_state.soft_min_level = + data->dpm_table.soc_table.dpm_levels[soft_min_level].value; + data->dpm_table.soc_table.dpm_state.soft_max_level = + data->dpm_table.soc_table.dpm_levels[soft_max_level].value; + ret = vega20_upload_dpm_min_level(hwmgr, 0xFFFFFFFF); PP_ASSERT_WITH_CODE(!ret, "Failed to upload DPM Bootup Levels!", @@ -2240,7 +2390,7 @@ static int vega20_force_clock_level(struct pp_hwmgr *hwmgr, enum pp_clock_type type, uint32_t mask) { struct vega20_hwmgr *data = (struct vega20_hwmgr *)(hwmgr->backend); - uint32_t soft_min_level, soft_max_level; + uint32_t soft_min_level, soft_max_level, hard_min_level; int ret = 0; switch (type) { @@ -2248,6 +2398,13 @@ static int vega20_force_clock_level(struct pp_hwmgr *hwmgr, soft_min_level = mask ? (ffs(mask) - 1) : 0; soft_max_level = mask ? (fls(mask) - 1) : 0; + if (soft_max_level >= data->dpm_table.gfx_table.count) { + pr_err("Clock level specified %d is over max allowed %d\n", + soft_max_level, + data->dpm_table.gfx_table.count - 1); + return -EINVAL; + } + data->dpm_table.gfx_table.dpm_state.soft_min_level = data->dpm_table.gfx_table.dpm_levels[soft_min_level].value; data->dpm_table.gfx_table.dpm_state.soft_max_level = @@ -2268,6 +2425,13 @@ static int vega20_force_clock_level(struct pp_hwmgr *hwmgr, soft_min_level = mask ? (ffs(mask) - 1) : 0; soft_max_level = mask ? (fls(mask) - 1) : 0; + if (soft_max_level >= data->dpm_table.mem_table.count) { + pr_err("Clock level specified %d is over max allowed %d\n", + soft_max_level, + data->dpm_table.mem_table.count - 1); + return -EINVAL; + } + data->dpm_table.mem_table.dpm_state.soft_min_level = data->dpm_table.mem_table.dpm_levels[soft_min_level].value; data->dpm_table.mem_table.dpm_state.soft_max_level = @@ -2285,6 +2449,84 @@ static int vega20_force_clock_level(struct pp_hwmgr *hwmgr, break; + case PP_SOCCLK: + soft_min_level = mask ? (ffs(mask) - 1) : 0; + soft_max_level = mask ? (fls(mask) - 1) : 0; + + if (soft_max_level >= data->dpm_table.soc_table.count) { + pr_err("Clock level specified %d is over max allowed %d\n", + soft_max_level, + data->dpm_table.soc_table.count - 1); + return -EINVAL; + } + + data->dpm_table.soc_table.dpm_state.soft_min_level = + data->dpm_table.soc_table.dpm_levels[soft_min_level].value; + data->dpm_table.soc_table.dpm_state.soft_max_level = + data->dpm_table.soc_table.dpm_levels[soft_max_level].value; + + ret = vega20_upload_dpm_min_level(hwmgr, FEATURE_DPM_SOCCLK_MASK); + PP_ASSERT_WITH_CODE(!ret, + "Failed to upload boot level to lowest!", + return ret); + + ret = vega20_upload_dpm_max_level(hwmgr, FEATURE_DPM_SOCCLK_MASK); + PP_ASSERT_WITH_CODE(!ret, + "Failed to upload dpm max level to highest!", + return ret); + + break; + + case PP_FCLK: + soft_min_level = mask ? (ffs(mask) - 1) : 0; + soft_max_level = mask ? (fls(mask) - 1) : 0; + + if (soft_max_level >= data->dpm_table.fclk_table.count) { + pr_err("Clock level specified %d is over max allowed %d\n", + soft_max_level, + data->dpm_table.fclk_table.count - 1); + return -EINVAL; + } + + data->dpm_table.fclk_table.dpm_state.soft_min_level = + data->dpm_table.fclk_table.dpm_levels[soft_min_level].value; + data->dpm_table.fclk_table.dpm_state.soft_max_level = + data->dpm_table.fclk_table.dpm_levels[soft_max_level].value; + + ret = vega20_upload_dpm_min_level(hwmgr, FEATURE_DPM_FCLK_MASK); + PP_ASSERT_WITH_CODE(!ret, + "Failed to upload boot level to lowest!", + return ret); + + ret = vega20_upload_dpm_max_level(hwmgr, FEATURE_DPM_FCLK_MASK); + PP_ASSERT_WITH_CODE(!ret, + "Failed to upload dpm max level to highest!", + return ret); + + break; + + case PP_DCEFCLK: + hard_min_level = mask ? (ffs(mask) - 1) : 0; + + if (hard_min_level >= data->dpm_table.dcef_table.count) { + pr_err("Clock level specified %d is over max allowed %d\n", + hard_min_level, + data->dpm_table.dcef_table.count - 1); + return -EINVAL; + } + + data->dpm_table.dcef_table.dpm_state.hard_min_level = + data->dpm_table.dcef_table.dpm_levels[hard_min_level].value; + + ret = vega20_upload_dpm_min_level(hwmgr, FEATURE_DPM_DCEFCLK_MASK); + PP_ASSERT_WITH_CODE(!ret, + "Failed to upload boot level to lowest!", + return ret); + + //TODO: Setting DCEFCLK max dpm level is not supported + + break; + case PP_PCIE: soft_min_level = mask ? (ffs(mask) - 1) : 0; soft_max_level = mask ? (fls(mask) - 1) : 0; @@ -2335,6 +2577,7 @@ static int vega20_dpm_force_dpm_level(struct pp_hwmgr *hwmgr, return ret; vega20_force_clock_level(hwmgr, PP_SCLK, 1 << sclk_mask); vega20_force_clock_level(hwmgr, PP_MCLK, 1 << mclk_mask); + vega20_force_clock_level(hwmgr, PP_SOCCLK, 1 << soc_mask); break; case AMD_DPM_FORCED_LEVEL_MANUAL: @@ -2765,6 +3008,108 @@ static int vega20_odn_edit_dpm_table(struct pp_hwmgr *hwmgr, return 0; } +static int vega20_get_ppfeature_status(struct pp_hwmgr *hwmgr, char *buf) +{ + static const char *ppfeature_name[] = { + "DPM_PREFETCHER", + "GFXCLK_DPM", + "UCLK_DPM", + "SOCCLK_DPM", + "UVD_DPM", + "VCE_DPM", + "ULV", + "MP0CLK_DPM", + "LINK_DPM", + "DCEFCLK_DPM", + "GFXCLK_DS", + "SOCCLK_DS", + "LCLK_DS", + "PPT", + "TDC", + "THERMAL", + "GFX_PER_CU_CG", + "RM", + "DCEFCLK_DS", + "ACDC", + "VR0HOT", + "VR1HOT", + "FW_CTF", + "LED_DISPLAY", + "FAN_CONTROL", + "GFX_EDC", + "GFXOFF", + "CG", + "FCLK_DPM", + "FCLK_DS", + "MP1CLK_DS", + "MP0CLK_DS", + "XGMI"}; + static const char *output_title[] = { + "FEATURES", + "BITMASK", + "ENABLEMENT"}; + uint64_t features_enabled; + int i; + int ret = 0; + int size = 0; + + ret = vega20_get_enabled_smc_features(hwmgr, &features_enabled); + PP_ASSERT_WITH_CODE(!ret, + "[EnableAllSmuFeatures] Failed to get enabled smc features!", + return ret); + + size += sprintf(buf + size, "Current ppfeatures: 0x%016llx\n", features_enabled); + size += sprintf(buf + size, "%-19s %-22s %s\n", + output_title[0], + output_title[1], + output_title[2]); + for (i = 0; i < GNLD_FEATURES_MAX; i++) { + size += sprintf(buf + size, "%-19s 0x%016llx %6s\n", + ppfeature_name[i], + 1ULL << i, + (features_enabled & (1ULL << i)) ? "Y" : "N"); + } + + return size; +} + +static int vega20_set_ppfeature_status(struct pp_hwmgr *hwmgr, uint64_t new_ppfeature_masks) +{ + uint64_t features_enabled; + uint64_t features_to_enable; + uint64_t features_to_disable; + int ret = 0; + + if (new_ppfeature_masks >= (1ULL << GNLD_FEATURES_MAX)) + return -EINVAL; + + ret = vega20_get_enabled_smc_features(hwmgr, &features_enabled); + if (ret) + return ret; + + features_to_disable = + (features_enabled ^ new_ppfeature_masks) & features_enabled; + features_to_enable = + (features_enabled ^ new_ppfeature_masks) ^ features_to_disable; + + pr_debug("features_to_disable 0x%llx\n", features_to_disable); + pr_debug("features_to_enable 0x%llx\n", features_to_enable); + + if (features_to_disable) { + ret = vega20_enable_smc_features(hwmgr, false, features_to_disable); + if (ret) + return ret; + } + + if (features_to_enable) { + ret = vega20_enable_smc_features(hwmgr, true, features_to_enable); + if (ret) + return ret; + } + + return 0; +} + static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr, enum pp_clock_type type, char *buf) { @@ -2779,6 +3124,8 @@ static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr, PPTable_t *pptable = (PPTable_t *)pptable_information->smc_pptable; struct amdgpu_device *adev = hwmgr->adev; struct pp_clock_levels_with_latency clocks; + struct vega20_single_dpm_table *fclk_dpm_table = + &(data->dpm_table.fclk_table); int i, now, size = 0; int ret = 0; uint32_t gen_speed, lane_width; @@ -2818,6 +3165,52 @@ static int vega20_print_clock_levels(struct pp_hwmgr *hwmgr, (clocks.data[i].clocks_in_khz == now * 10) ? "*" : ""); break; + case PP_SOCCLK: + ret = vega20_get_current_clk_freq(hwmgr, PPCLK_SOCCLK, &now); + PP_ASSERT_WITH_CODE(!ret, + "Attempt to get current socclk freq Failed!", + return ret); + + ret = vega20_get_socclocks(hwmgr, &clocks); + PP_ASSERT_WITH_CODE(!ret, + "Attempt to get soc clk levels Failed!", + return ret); + + for (i = 0; i < clocks.num_levels; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, clocks.data[i].clocks_in_khz / 1000, + (clocks.data[i].clocks_in_khz == now * 10) ? "*" : ""); + break; + + case PP_FCLK: + ret = vega20_get_current_clk_freq(hwmgr, PPCLK_FCLK, &now); + PP_ASSERT_WITH_CODE(!ret, + "Attempt to get current fclk freq Failed!", + return ret); + + for (i = 0; i < fclk_dpm_table->count; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, fclk_dpm_table->dpm_levels[i].value, + fclk_dpm_table->dpm_levels[i].value == (now / 100) ? "*" : ""); + break; + + case PP_DCEFCLK: + ret = vega20_get_current_clk_freq(hwmgr, PPCLK_DCEFCLK, &now); + PP_ASSERT_WITH_CODE(!ret, + "Attempt to get current dcefclk freq Failed!", + return ret); + + ret = vega20_get_dcefclocks(hwmgr, &clocks); + PP_ASSERT_WITH_CODE(!ret, + "Attempt to get dcefclk levels Failed!", + return ret); + + for (i = 0; i < clocks.num_levels; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, clocks.data[i].clocks_in_khz / 1000, + (clocks.data[i].clocks_in_khz == now * 10) ? "*" : ""); + break; + case PP_PCIE: gen_speed = (RREG32_PCIE(smnPCIE_LC_SPEED_CNTL) & PSWUSP0_PCIE_LC_SPEED_CNTL__LC_CURRENT_DATA_RATE_MASK) @@ -3261,6 +3654,9 @@ static int conv_power_profile_to_pplib_workload(int power_profile) int pplib_workload = 0; switch (power_profile) { + case PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT: + pplib_workload = WORKLOAD_DEFAULT_BIT; + break; case PP_SMC_POWER_PROFILE_FULLSCREEN3D: pplib_workload = WORKLOAD_PPLIB_FULL_SCREEN_3D_BIT; break; @@ -3290,6 +3686,7 @@ static int vega20_get_power_profile_mode(struct pp_hwmgr *hwmgr, char *buf) uint32_t i, size = 0; uint16_t workload_type = 0; static const char *profile_name[] = { + "BOOTUP_DEFAULT", "3D_FULL_SCREEN", "POWER_SAVING", "VIDEO", @@ -3557,6 +3954,8 @@ static const struct pp_hwmgr_func vega20_hwmgr_funcs = { .force_clock_level = vega20_force_clock_level, .print_clock_levels = vega20_print_clock_levels, .read_sensor = vega20_read_sensor, + .get_ppfeature_status = vega20_get_ppfeature_status, + .set_ppfeature_status = vega20_set_ppfeature_status, /* powergate related */ .powergate_uvd = vega20_power_gate_uvd, .powergate_vce = vega20_power_gate_vce, @@ -3577,6 +3976,10 @@ static const struct pp_hwmgr_func vega20_hwmgr_funcs = { /* smu memory related */ .notify_cac_buffer_info = vega20_notify_cac_buffer_info, .enable_mgpu_fan_boost = vega20_enable_mgpu_fan_boost, + /* BACO related */ + .get_asic_baco_capability = vega20_baco_get_capability, + .get_asic_baco_state = vega20_baco_get_state, + .set_asic_baco_state = vega20_baco_set_state, }; int vega20_hwmgr_init(struct pp_hwmgr *hwmgr) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.h b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.h index 25faaa5c5b10..37f5f5e657da 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_hwmgr.h @@ -520,6 +520,9 @@ struct vega20_hwmgr { /* ---- Gfxoff ---- */ bool gfxoff_allowed; uint32_t counter_gfxoff; + + unsigned long metrics_time; + SmuMetrics_t metrics_table; }; #define VEGA20_DPM2_NEAR_TDP_DEC 10 diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_inc.h b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_inc.h index 6738bad53602..613cb1989b3d 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_inc.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/vega20_inc.h @@ -31,5 +31,6 @@ #include "asic_reg/mp/mp_9_0_sh_mask.h" #include "asic_reg/nbio/nbio_7_4_offset.h" +#include "asic_reg/nbio/nbio_7_4_sh_mask.h" #endif diff --git a/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h b/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h index f4dab979a3a1..6e0be6027705 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h +++ b/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h @@ -397,7 +397,6 @@ struct phm_odn_clock_levels { }; extern int phm_disable_clock_power_gatings(struct pp_hwmgr *hwmgr); -extern int phm_enable_clock_power_gatings(struct pp_hwmgr *hwmgr); extern int phm_powerdown_uvd(struct pp_hwmgr *hwmgr); extern int phm_setup_asic(struct pp_hwmgr *hwmgr); extern int phm_enable_dynamic_state_management(struct pp_hwmgr *hwmgr); diff --git a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h index 0d298a0409f5..bac3d85e3b82 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h +++ b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h @@ -47,6 +47,11 @@ enum DISPLAY_GAP { }; typedef enum DISPLAY_GAP DISPLAY_GAP; +enum BACO_STATE { + BACO_STATE_OUT = 0, + BACO_STATE_IN, +}; + struct vi_dpm_level { bool enabled; uint32_t value; @@ -251,7 +256,6 @@ struct pp_hwmgr_func { uint32_t (*get_sclk)(struct pp_hwmgr *hwmgr, bool low); int (*power_state_set)(struct pp_hwmgr *hwmgr, const void *state); - int (*enable_clock_power_gating)(struct pp_hwmgr *hwmgr); int (*notify_smc_display_config_after_ps_adjustment)(struct pp_hwmgr *hwmgr); int (*pre_display_config_changed)(struct pp_hwmgr *hwmgr); int (*display_config_changed)(struct pp_hwmgr *hwmgr); @@ -334,6 +338,11 @@ struct pp_hwmgr_func { int (*enable_mgpu_fan_boost)(struct pp_hwmgr *hwmgr); int (*set_hard_min_dcefclk_by_freq)(struct pp_hwmgr *hwmgr, uint32_t clock); int (*set_hard_min_fclk_by_freq)(struct pp_hwmgr *hwmgr, uint32_t clock); + int (*get_asic_baco_capability)(struct pp_hwmgr *hwmgr, bool *cap); + int (*get_asic_baco_state)(struct pp_hwmgr *hwmgr, enum BACO_STATE *state); + int (*set_asic_baco_state)(struct pp_hwmgr *hwmgr, enum BACO_STATE state); + int (*get_ppfeature_status)(struct pp_hwmgr *hwmgr, char *buf); + int (*set_ppfeature_status)(struct pp_hwmgr *hwmgr, uint64_t ppfeature_masks); }; struct pp_table_func { @@ -678,6 +687,7 @@ struct pp_advance_fan_control_parameters { uint32_t ulTargetGfxClk; uint16_t usZeroRPMStartTemperature; uint16_t usZeroRPMStopTemperature; + uint16_t usMGpuThrottlingRPMLimit; }; struct pp_thermal_controller_info { @@ -705,7 +715,7 @@ enum PP_TABLE_VERSION { /** * The main hardware manager structure. */ -#define Workload_Policy_Max 5 +#define Workload_Policy_Max 6 struct pp_hwmgr { void *adev; diff --git a/drivers/gpu/drm/arm/Kconfig b/drivers/gpu/drm/arm/Kconfig index 9a18e1bd57b4..a204103b3efb 100644 --- a/drivers/gpu/drm/arm/Kconfig +++ b/drivers/gpu/drm/arm/Kconfig @@ -1,13 +1,10 @@ -config DRM_ARM - bool - help - Choose this option to select drivers for ARM's devices +# SPDX-License-Identifier: GPL-2.0 +menu "ARM devices" config DRM_HDLCD tristate "ARM HDLCD" depends on DRM && OF && (ARM || ARM64) depends on COMMON_CLK - select DRM_ARM select DRM_KMS_HELPER select DRM_KMS_CMA_HELPER help @@ -29,7 +26,6 @@ config DRM_MALI_DISPLAY tristate "ARM Mali Display Processor" depends on DRM && OF && (ARM || ARM64) depends on COMMON_CLK - select DRM_ARM select DRM_KMS_HELPER select DRM_KMS_CMA_HELPER select DRM_GEM_CMA_HELPER @@ -40,3 +36,7 @@ config DRM_MALI_DISPLAY of the hardware. If compiled as a module it will be called mali-dp. + +source "drivers/gpu/drm/arm/display/Kconfig" + +endmenu diff --git a/drivers/gpu/drm/arm/Makefile b/drivers/gpu/drm/arm/Makefile index 3bf31d1a4722..120bef801fcf 100644 --- a/drivers/gpu/drm/arm/Makefile +++ b/drivers/gpu/drm/arm/Makefile @@ -3,3 +3,4 @@ obj-$(CONFIG_DRM_HDLCD) += hdlcd.o mali-dp-y := malidp_drv.o malidp_hw.o malidp_planes.o malidp_crtc.o mali-dp-y += malidp_mw.o obj-$(CONFIG_DRM_MALI_DISPLAY) += mali-dp.o +obj-$(CONFIG_DRM_KOMEDA) += display/ diff --git a/drivers/gpu/drm/arm/display/Kbuild b/drivers/gpu/drm/arm/display/Kbuild new file mode 100644 index 000000000000..382f1ca831e4 --- /dev/null +++ b/drivers/gpu/drm/arm/display/Kbuild @@ -0,0 +1,3 @@ +# SPDX-License-Identifier: GPL-2.0 + +obj-$(CONFIG_DRM_KOMEDA) += komeda/ diff --git a/drivers/gpu/drm/arm/display/Kconfig b/drivers/gpu/drm/arm/display/Kconfig new file mode 100644 index 000000000000..cec0639e3aa1 --- /dev/null +++ b/drivers/gpu/drm/arm/display/Kconfig @@ -0,0 +1,14 @@ +# SPDX-License-Identifier: GPL-2.0 +config DRM_KOMEDA + tristate "ARM Komeda display driver" + depends on DRM && OF + depends on COMMON_CLK + select DRM_KMS_HELPER + select DRM_KMS_CMA_HELPER + select DRM_GEM_CMA_HELPER + select VIDEOMODE_HELPERS + help + Choose this option if you want to compile the ARM Komeda display + Processor driver. It supports the D71 variants of the hardware. + + If compiled as a module it will be called komeda. diff --git a/drivers/gpu/drm/arm/display/include/malidp_io.h b/drivers/gpu/drm/arm/display/include/malidp_io.h new file mode 100644 index 000000000000..4fb3caf864ce --- /dev/null +++ b/drivers/gpu/drm/arm/display/include/malidp_io.h @@ -0,0 +1,42 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#ifndef _MALIDP_IO_H_ +#define _MALIDP_IO_H_ + +#include <linux/io.h> + +static inline u32 +malidp_read32(u32 __iomem *base, u32 offset) +{ + return readl((base + (offset >> 2))); +} + +static inline void +malidp_write32(u32 __iomem *base, u32 offset, u32 v) +{ + writel(v, (base + (offset >> 2))); +} + +static inline void +malidp_write32_mask(u32 __iomem *base, u32 offset, u32 m, u32 v) +{ + u32 tmp = malidp_read32(base, offset); + + tmp &= (~m); + malidp_write32(base, offset, v | tmp); +} + +static inline void +malidp_write_group(u32 __iomem *base, u32 offset, int num, const u32 *values) +{ + int i; + + for (i = 0; i < num; i++) + malidp_write32(base, offset + i * 4, values[i]); +} + +#endif /*_MALIDP_IO_H_*/ diff --git a/drivers/gpu/drm/arm/display/include/malidp_product.h b/drivers/gpu/drm/arm/display/include/malidp_product.h new file mode 100644 index 000000000000..b35fc5db866b --- /dev/null +++ b/drivers/gpu/drm/arm/display/include/malidp_product.h @@ -0,0 +1,23 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#ifndef _MALIDP_PRODUCT_H_ +#define _MALIDP_PRODUCT_H_ + +/* Product identification */ +#define MALIDP_CORE_ID(__product, __major, __minor, __status) \ + ((((__product) & 0xFFFF) << 16) | (((__major) & 0xF) << 12) | \ + (((__minor) & 0xF) << 8) | ((__status) & 0xFF)) + +#define MALIDP_CORE_ID_PRODUCT_ID(__core_id) ((__u32)(__core_id) >> 16) +#define MALIDP_CORE_ID_MAJOR(__core_id) (((__u32)(__core_id) >> 12) & 0xF) +#define MALIDP_CORE_ID_MINOR(__core_id) (((__u32)(__core_id) >> 8) & 0xF) +#define MALIDP_CORE_ID_STATUS(__core_id) (((__u32)(__core_id)) & 0xFF) + +/* Mali-display product IDs */ +#define MALIDP_D71_PRODUCT_ID 0x0071 + +#endif /* _MALIDP_PRODUCT_H_ */ diff --git a/drivers/gpu/drm/arm/display/include/malidp_utils.h b/drivers/gpu/drm/arm/display/include/malidp_utils.h new file mode 100644 index 000000000000..63cc47cefcf8 --- /dev/null +++ b/drivers/gpu/drm/arm/display/include/malidp_utils.h @@ -0,0 +1,16 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#ifndef _MALIDP_UTILS_ +#define _MALIDP_UTILS_ + +#define has_bit(nr, mask) (BIT(nr) & (mask)) +#define has_bits(bits, mask) (((bits) & (mask)) == (bits)) + +#define dp_for_each_set_bit(bit, mask) \ + for_each_set_bit((bit), ((unsigned long *)&(mask)), sizeof(mask) * 8) + +#endif /* _MALIDP_UTILS_ */ diff --git a/drivers/gpu/drm/arm/display/komeda/Makefile b/drivers/gpu/drm/arm/display/komeda/Makefile new file mode 100644 index 000000000000..1b875e5dc0f6 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/Makefile @@ -0,0 +1,21 @@ +# SPDX-License-Identifier: GPL-2.0 + +ccflags-y := \ + -I$(src)/../include \ + -I$(src) + +komeda-y := \ + komeda_drv.o \ + komeda_dev.o \ + komeda_format_caps.o \ + komeda_pipeline.o \ + komeda_framebuffer.o \ + komeda_kms.o \ + komeda_crtc.o \ + komeda_plane.o \ + komeda_private_obj.o + +komeda-y += \ + d71/d71_dev.o + +obj-$(CONFIG_DRM_KOMEDA) += komeda.o diff --git a/drivers/gpu/drm/arm/display/komeda/d71/d71_dev.c b/drivers/gpu/drm/arm/display/komeda/d71/d71_dev.c new file mode 100644 index 000000000000..edbf9daa1545 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/d71/d71_dev.c @@ -0,0 +1,111 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#include "malidp_io.h" +#include "komeda_dev.h" + +static int d71_enum_resources(struct komeda_dev *mdev) +{ + /* TODO add enum resources */ + return -1; +} + +#define __HW_ID(__group, __format) \ + ((((__group) & 0x7) << 3) | ((__format) & 0x7)) + +#define RICH KOMEDA_FMT_RICH_LAYER +#define SIMPLE KOMEDA_FMT_SIMPLE_LAYER +#define RICH_SIMPLE (KOMEDA_FMT_RICH_LAYER | KOMEDA_FMT_SIMPLE_LAYER) +#define RICH_WB (KOMEDA_FMT_RICH_LAYER | KOMEDA_FMT_WB_LAYER) +#define RICH_SIMPLE_WB (RICH_SIMPLE | KOMEDA_FMT_WB_LAYER) + +#define Rot_0 DRM_MODE_ROTATE_0 +#define Flip_H_V (DRM_MODE_REFLECT_X | DRM_MODE_REFLECT_Y | Rot_0) +#define Rot_ALL_H_V (DRM_MODE_ROTATE_MASK | Flip_H_V) + +#define LYT_NM BIT(AFBC_FORMAT_MOD_BLOCK_SIZE_16x16) +#define LYT_WB BIT(AFBC_FORMAT_MOD_BLOCK_SIZE_32x8) +#define LYT_NM_WB (LYT_NM | LYT_WB) + +#define AFB_TH AFBC(_TILED | _SPARSE) +#define AFB_TH_SC_YTR AFBC(_TILED | _SC | _SPARSE | _YTR) +#define AFB_TH_SC_YTR_BS AFBC(_TILED | _SC | _SPARSE | _YTR | _SPLIT) + +static struct komeda_format_caps d71_format_caps_table[] = { + /* HW_ID | fourcc | tile_sz | layer_types | rots | afbc_layouts | afbc_features */ + /* ABGR_2101010*/ + {__HW_ID(0, 0), DRM_FORMAT_ARGB2101010, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + {__HW_ID(0, 1), DRM_FORMAT_ABGR2101010, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + {__HW_ID(0, 1), DRM_FORMAT_ABGR2101010, 1, RICH_SIMPLE, Rot_ALL_H_V, LYT_NM_WB, AFB_TH_SC_YTR_BS}, /* afbc */ + {__HW_ID(0, 2), DRM_FORMAT_RGBA1010102, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + {__HW_ID(0, 3), DRM_FORMAT_BGRA1010102, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + /* ABGR_8888*/ + {__HW_ID(1, 0), DRM_FORMAT_ARGB8888, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + {__HW_ID(1, 1), DRM_FORMAT_ABGR8888, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + {__HW_ID(1, 1), DRM_FORMAT_ABGR8888, 1, RICH_SIMPLE, Rot_ALL_H_V, LYT_NM_WB, AFB_TH_SC_YTR_BS}, /* afbc */ + {__HW_ID(1, 2), DRM_FORMAT_RGBA8888, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + {__HW_ID(1, 3), DRM_FORMAT_BGRA8888, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + /* XBGB_8888 */ + {__HW_ID(2, 0), DRM_FORMAT_XRGB8888, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + {__HW_ID(2, 1), DRM_FORMAT_XBGR8888, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + {__HW_ID(2, 2), DRM_FORMAT_RGBX8888, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + {__HW_ID(2, 3), DRM_FORMAT_BGRX8888, 1, RICH_SIMPLE_WB, Flip_H_V, 0, 0}, + /* BGR_888 */ /* none-afbc RGB888 doesn't support rotation and flip */ + {__HW_ID(3, 0), DRM_FORMAT_RGB888, 1, RICH_SIMPLE_WB, Rot_0, 0, 0}, + {__HW_ID(3, 1), DRM_FORMAT_BGR888, 1, RICH_SIMPLE_WB, Rot_0, 0, 0}, + {__HW_ID(3, 1), DRM_FORMAT_BGR888, 1, RICH_SIMPLE, Rot_ALL_H_V, LYT_NM_WB, AFB_TH_SC_YTR_BS}, /* afbc */ + /* BGR 16bpp */ + {__HW_ID(4, 0), DRM_FORMAT_RGBA5551, 1, RICH_SIMPLE, Flip_H_V, 0, 0}, + {__HW_ID(4, 1), DRM_FORMAT_ABGR1555, 1, RICH_SIMPLE, Flip_H_V, 0, 0}, + {__HW_ID(4, 1), DRM_FORMAT_ABGR1555, 1, RICH_SIMPLE, Rot_ALL_H_V, LYT_NM_WB, AFB_TH_SC_YTR}, /* afbc */ + {__HW_ID(4, 2), DRM_FORMAT_RGB565, 1, RICH_SIMPLE, Flip_H_V, 0, 0}, + {__HW_ID(4, 3), DRM_FORMAT_BGR565, 1, RICH_SIMPLE, Flip_H_V, 0, 0}, + {__HW_ID(4, 3), DRM_FORMAT_BGR565, 1, RICH_SIMPLE, Rot_ALL_H_V, LYT_NM_WB, AFB_TH_SC_YTR}, /* afbc */ + {__HW_ID(4, 4), DRM_FORMAT_R8, 1, SIMPLE, Rot_0, 0, 0}, + /* YUV 444/422/420 8bit */ + {__HW_ID(5, 0), 0 /*XYUV8888*/, 1, 0, 0, 0, 0}, + /* XYUV unsupported*/ + {__HW_ID(5, 1), DRM_FORMAT_YUYV, 1, RICH, Rot_ALL_H_V, LYT_NM, AFB_TH}, /* afbc */ + {__HW_ID(5, 2), DRM_FORMAT_YUYV, 1, RICH, Flip_H_V, 0, 0}, + {__HW_ID(5, 3), DRM_FORMAT_UYVY, 1, RICH, Flip_H_V, 0, 0}, + {__HW_ID(5, 4), 0, /*X0L0 */ 2, 0, 0, 0}, /* Y0L0 unsupported */ + {__HW_ID(5, 6), DRM_FORMAT_NV12, 1, RICH, Flip_H_V, 0, 0}, + {__HW_ID(5, 6), 0/*DRM_FORMAT_YUV420_8BIT*/, 1, RICH, Rot_ALL_H_V, LYT_NM, AFB_TH}, /* afbc */ + {__HW_ID(5, 7), DRM_FORMAT_YUV420, 1, RICH, Flip_H_V, 0, 0}, + /* YUV 10bit*/ + {__HW_ID(6, 0), 0,/*XVYU2101010*/ 1, 0, 0, 0, 0},/* VYV30 unsupported */ + {__HW_ID(6, 6), 0/*DRM_FORMAT_X0L2*/, 2, RICH, Flip_H_V, 0, 0}, + {__HW_ID(6, 7), 0/*DRM_FORMAT_P010*/, 1, RICH, Flip_H_V, 0, 0}, + {__HW_ID(6, 7), 0/*DRM_FORMAT_YUV420_10BIT*/, 1, RICH, Rot_ALL_H_V, LYT_NM, AFB_TH}, +}; + +static void d71_init_fmt_tbl(struct komeda_dev *mdev) +{ + struct komeda_format_caps_table *table = &mdev->fmt_tbl; + + table->format_caps = d71_format_caps_table; + table->n_formats = ARRAY_SIZE(d71_format_caps_table); +} + +static struct komeda_dev_funcs d71_chip_funcs = { + .init_format_table = d71_init_fmt_tbl, + .enum_resources = d71_enum_resources, + .cleanup = NULL, +}; + +#define GLB_ARCH_ID 0x000 +#define GLB_CORE_ID 0x004 +#define GLB_CORE_INFO 0x008 + +struct komeda_dev_funcs * +d71_identify(u32 __iomem *reg_base, struct komeda_chip_info *chip) +{ + chip->arch_id = malidp_read32(reg_base, GLB_ARCH_ID); + chip->core_id = malidp_read32(reg_base, GLB_CORE_ID); + chip->core_info = malidp_read32(reg_base, GLB_CORE_INFO); + + return &d71_chip_funcs; +} diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_crtc.c b/drivers/gpu/drm/arm/display/komeda/komeda_crtc.c new file mode 100644 index 000000000000..5bb5a55f6b31 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_crtc.c @@ -0,0 +1,106 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#include <linux/clk.h> +#include <linux/spinlock.h> +#include <drm/drm_atomic.h> +#include <drm/drm_atomic_helper.h> +#include <drm/drm_plane_helper.h> +#include <drm/drm_crtc_helper.h> +#include <linux/pm_runtime.h> +#include "komeda_dev.h" +#include "komeda_kms.h" + +struct drm_crtc_helper_funcs komeda_crtc_helper_funcs = { +}; + +static const struct drm_crtc_funcs komeda_crtc_funcs = { +}; + +int komeda_kms_setup_crtcs(struct komeda_kms_dev *kms, + struct komeda_dev *mdev) +{ + struct komeda_crtc *crtc; + struct komeda_pipeline *master; + char str[16]; + int i; + + kms->n_crtcs = 0; + + for (i = 0; i < mdev->n_pipelines; i++) { + crtc = &kms->crtcs[kms->n_crtcs]; + master = mdev->pipelines[i]; + + crtc->master = master; + crtc->slave = NULL; + + if (crtc->slave) + sprintf(str, "pipe-%d", crtc->slave->id); + else + sprintf(str, "None"); + + DRM_INFO("crtc%d: master(pipe-%d) slave(%s) output: %s.\n", + kms->n_crtcs, master->id, str, + master->of_output_dev ? + master->of_output_dev->full_name : "None"); + + kms->n_crtcs++; + } + + return 0; +} + +static struct drm_plane * +get_crtc_primary(struct komeda_kms_dev *kms, struct komeda_crtc *crtc) +{ + struct komeda_plane *kplane; + struct drm_plane *plane; + + drm_for_each_plane(plane, &kms->base) { + if (plane->type != DRM_PLANE_TYPE_PRIMARY) + continue; + + kplane = to_kplane(plane); + /* only master can be primary */ + if (kplane->layer->base.pipeline == crtc->master) + return plane; + } + + return NULL; +} + +static int komeda_crtc_add(struct komeda_kms_dev *kms, + struct komeda_crtc *kcrtc) +{ + struct drm_crtc *crtc = &kcrtc->base; + int err; + + err = drm_crtc_init_with_planes(&kms->base, crtc, + get_crtc_primary(kms, kcrtc), NULL, + &komeda_crtc_funcs, NULL); + if (err) + return err; + + drm_crtc_helper_add(crtc, &komeda_crtc_helper_funcs); + drm_crtc_vblank_reset(crtc); + + crtc->port = kcrtc->master->of_output_port; + + return 0; +} + +int komeda_kms_add_crtcs(struct komeda_kms_dev *kms, struct komeda_dev *mdev) +{ + int i, err; + + for (i = 0; i < kms->n_crtcs; i++) { + err = komeda_crtc_add(kms, &kms->crtcs[i]); + if (err) + return err; + } + + return 0; +} diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_dev.c b/drivers/gpu/drm/arm/display/komeda/komeda_dev.c new file mode 100644 index 000000000000..0fe6954fbbf4 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_dev.c @@ -0,0 +1,186 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#include <linux/platform_device.h> +#include <linux/of_device.h> +#include <linux/of_graph.h> +#include "komeda_dev.h" + +static int komeda_parse_pipe_dt(struct komeda_dev *mdev, struct device_node *np) +{ + struct komeda_pipeline *pipe; + struct clk *clk; + u32 pipe_id; + int ret = 0; + + ret = of_property_read_u32(np, "reg", &pipe_id); + if (ret != 0 || pipe_id >= mdev->n_pipelines) + return -EINVAL; + + pipe = mdev->pipelines[pipe_id]; + + clk = of_clk_get_by_name(np, "aclk"); + if (IS_ERR(clk)) { + DRM_ERROR("get aclk for pipeline %d failed!\n", pipe_id); + return PTR_ERR(clk); + } + pipe->aclk = clk; + + clk = of_clk_get_by_name(np, "pxclk"); + if (IS_ERR(clk)) { + DRM_ERROR("get pxclk for pipeline %d failed!\n", pipe_id); + return PTR_ERR(clk); + } + pipe->pxlclk = clk; + + /* enum ports */ + pipe->of_output_dev = + of_graph_get_remote_node(np, KOMEDA_OF_PORT_OUTPUT, 0); + pipe->of_output_port = + of_graph_get_port_by_id(np, KOMEDA_OF_PORT_OUTPUT); + + pipe->of_node = np; + + return 0; +} + +static int komeda_parse_dt(struct device *dev, struct komeda_dev *mdev) +{ + struct device_node *child, *np = dev->of_node; + struct clk *clk; + int ret; + + clk = devm_clk_get(dev, "mclk"); + if (IS_ERR(clk)) + return PTR_ERR(clk); + + mdev->mclk = clk; + + for_each_available_child_of_node(np, child) { + if (of_node_cmp(child->name, "pipeline") == 0) { + ret = komeda_parse_pipe_dt(mdev, child); + if (ret) { + DRM_ERROR("parse pipeline dt error!\n"); + of_node_put(child); + break; + } + } + } + + return ret; +} + +struct komeda_dev *komeda_dev_create(struct device *dev) +{ + struct platform_device *pdev = to_platform_device(dev); + const struct komeda_product_data *product; + struct komeda_dev *mdev; + struct resource *io_res; + int err = 0; + + product = of_device_get_match_data(dev); + if (!product) + return ERR_PTR(-ENODEV); + + io_res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (!io_res) { + DRM_ERROR("No registers defined.\n"); + return ERR_PTR(-ENODEV); + } + + mdev = devm_kzalloc(dev, sizeof(*mdev), GFP_KERNEL); + if (!mdev) + return ERR_PTR(-ENOMEM); + + mdev->dev = dev; + mdev->reg_base = devm_ioremap_resource(dev, io_res); + if (IS_ERR(mdev->reg_base)) { + DRM_ERROR("Map register space failed.\n"); + err = PTR_ERR(mdev->reg_base); + mdev->reg_base = NULL; + goto err_cleanup; + } + + mdev->pclk = devm_clk_get(dev, "pclk"); + if (IS_ERR(mdev->pclk)) { + DRM_ERROR("Get APB clk failed.\n"); + err = PTR_ERR(mdev->pclk); + mdev->pclk = NULL; + goto err_cleanup; + } + + /* Enable APB clock to access the registers */ + clk_prepare_enable(mdev->pclk); + + mdev->funcs = product->identify(mdev->reg_base, &mdev->chip); + if (!komeda_product_match(mdev, product->product_id)) { + DRM_ERROR("DT configured %x mismatch with real HW %x.\n", + product->product_id, + MALIDP_CORE_ID_PRODUCT_ID(mdev->chip.core_id)); + err = -ENODEV; + goto err_cleanup; + } + + DRM_INFO("Found ARM Mali-D%x version r%dp%d\n", + MALIDP_CORE_ID_PRODUCT_ID(mdev->chip.core_id), + MALIDP_CORE_ID_MAJOR(mdev->chip.core_id), + MALIDP_CORE_ID_MINOR(mdev->chip.core_id)); + + mdev->funcs->init_format_table(mdev); + + err = mdev->funcs->enum_resources(mdev); + if (err) { + DRM_ERROR("enumerate display resource failed.\n"); + goto err_cleanup; + } + + err = komeda_parse_dt(dev, mdev); + if (err) { + DRM_ERROR("parse device tree failed.\n"); + goto err_cleanup; + } + + return mdev; + +err_cleanup: + komeda_dev_destroy(mdev); + return ERR_PTR(err); +} + +void komeda_dev_destroy(struct komeda_dev *mdev) +{ + struct device *dev = mdev->dev; + struct komeda_dev_funcs *funcs = mdev->funcs; + int i; + + for (i = 0; i < mdev->n_pipelines; i++) { + komeda_pipeline_destroy(mdev, mdev->pipelines[i]); + mdev->pipelines[i] = NULL; + } + + mdev->n_pipelines = 0; + + if (funcs && funcs->cleanup) + funcs->cleanup(mdev); + + if (mdev->reg_base) { + devm_iounmap(dev, mdev->reg_base); + mdev->reg_base = NULL; + } + + if (mdev->mclk) { + devm_clk_put(dev, mdev->mclk); + mdev->mclk = NULL; + } + + if (mdev->pclk) { + clk_disable_unprepare(mdev->pclk); + devm_clk_put(dev, mdev->pclk); + mdev->pclk = NULL; + } + + devm_kfree(dev, mdev); +} diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_dev.h b/drivers/gpu/drm/arm/display/komeda/komeda_dev.h new file mode 100644 index 000000000000..0f77dead6a23 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_dev.h @@ -0,0 +1,110 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#ifndef _KOMEDA_DEV_H_ +#define _KOMEDA_DEV_H_ + +#include <linux/device.h> +#include <linux/clk.h> +#include "komeda_pipeline.h" +#include "malidp_product.h" +#include "komeda_format_caps.h" + +/* malidp device id */ +enum { + MALI_D71 = 0, +}; + +/* pipeline DT ports */ +enum { + KOMEDA_OF_PORT_OUTPUT = 0, + KOMEDA_OF_PORT_COPROC = 1, +}; + +struct komeda_chip_info { + u32 arch_id; + u32 core_id; + u32 core_info; + u32 bus_width; +}; + +struct komeda_product_data { + u32 product_id; + struct komeda_dev_funcs *(*identify)(u32 __iomem *reg, + struct komeda_chip_info *info); +}; + +struct komeda_dev; + +/** + * struct komeda_dev_funcs + * + * Supplied by chip level and returned by the chip entry function xxx_identify, + */ +struct komeda_dev_funcs { + /** + * @init_format_table: + * + * initialize &komeda_dev->format_table, this function should be called + * before the &enum_resource + */ + void (*init_format_table)(struct komeda_dev *mdev); + /** + * @enum_resources: + * + * for CHIP to report or add pipeline and component resources to CORE + */ + int (*enum_resources)(struct komeda_dev *mdev); + /** @cleanup: call to chip to cleanup komeda_dev->chip data */ + void (*cleanup)(struct komeda_dev *mdev); +}; + +/** + * struct komeda_dev + * + * Pipeline and component are used to describe how to handle the pixel data. + * komeda_device is for describing the whole view of the device, and the + * control-abilites of device. + */ +struct komeda_dev { + struct device *dev; + u32 __iomem *reg_base; + + struct komeda_chip_info chip; + /** @fmt_tbl: initialized by &komeda_dev_funcs->init_format_table */ + struct komeda_format_caps_table fmt_tbl; + /** @pclk: APB clock for register access */ + struct clk *pclk; + /** @mck: HW main engine clk */ + struct clk *mclk; + + int n_pipelines; + struct komeda_pipeline *pipelines[KOMEDA_MAX_PIPELINES]; + + /** @funcs: chip funcs to access to HW */ + struct komeda_dev_funcs *funcs; + /** + * @chip_data: + * + * chip data will be added by &komeda_dev_funcs.enum_resources() and + * destroyed by &komeda_dev_funcs.cleanup() + */ + void *chip_data; +}; + +static inline bool +komeda_product_match(struct komeda_dev *mdev, u32 target) +{ + return MALIDP_CORE_ID_PRODUCT_ID(mdev->chip.core_id) == target; +} + +struct komeda_dev_funcs * +d71_identify(u32 __iomem *reg, struct komeda_chip_info *chip); + +struct komeda_dev *komeda_dev_create(struct device *dev); +void komeda_dev_destroy(struct komeda_dev *mdev); + +#endif /*_KOMEDA_DEV_H_*/ diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_drv.c b/drivers/gpu/drm/arm/display/komeda/komeda_drv.c new file mode 100644 index 000000000000..2bdd189b041d --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_drv.c @@ -0,0 +1,144 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#include <linux/module.h> +#include <linux/kernel.h> +#include <linux/platform_device.h> +#include <linux/component.h> +#include <drm/drm_of.h> +#include "komeda_dev.h" +#include "komeda_kms.h" + +struct komeda_drv { + struct komeda_dev *mdev; + struct komeda_kms_dev *kms; +}; + +static void komeda_unbind(struct device *dev) +{ + struct komeda_drv *mdrv = dev_get_drvdata(dev); + + if (!mdrv) + return; + + komeda_kms_detach(mdrv->kms); + komeda_dev_destroy(mdrv->mdev); + + dev_set_drvdata(dev, NULL); + devm_kfree(dev, mdrv); +} + +static int komeda_bind(struct device *dev) +{ + struct komeda_drv *mdrv; + int err; + + mdrv = devm_kzalloc(dev, sizeof(*mdrv), GFP_KERNEL); + if (!mdrv) + return -ENOMEM; + + mdrv->mdev = komeda_dev_create(dev); + if (IS_ERR(mdrv->mdev)) { + err = PTR_ERR(mdrv->mdev); + goto free_mdrv; + } + + mdrv->kms = komeda_kms_attach(mdrv->mdev); + if (IS_ERR(mdrv->kms)) { + err = PTR_ERR(mdrv->kms); + goto destroy_mdev; + } + + dev_set_drvdata(dev, mdrv); + + return 0; + +destroy_mdev: + komeda_dev_destroy(mdrv->mdev); + +free_mdrv: + devm_kfree(dev, mdrv); + return err; +} + +static const struct component_master_ops komeda_master_ops = { + .bind = komeda_bind, + .unbind = komeda_unbind, +}; + +static int compare_of(struct device *dev, void *data) +{ + return dev->of_node == data; +} + +static void komeda_add_slave(struct device *master, + struct component_match **match, + struct device_node *np, int port) +{ + struct device_node *remote; + + remote = of_graph_get_remote_node(np, port, 0); + if (remote) { + drm_of_component_match_add(master, match, compare_of, remote); + of_node_put(remote); + } +} + +static int komeda_platform_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct component_match *match = NULL; + struct device_node *child; + + if (!dev->of_node) + return -ENODEV; + + for_each_available_child_of_node(dev->of_node, child) { + if (of_node_cmp(child->name, "pipeline") != 0) + continue; + + /* add connector */ + komeda_add_slave(dev, &match, child, KOMEDA_OF_PORT_OUTPUT); + } + + return component_master_add_with_match(dev, &komeda_master_ops, match); +} + +static int komeda_platform_remove(struct platform_device *pdev) +{ + component_master_del(&pdev->dev, &komeda_master_ops); + return 0; +} + +static const struct komeda_product_data komeda_products[] = { + [MALI_D71] = { + .product_id = MALIDP_D71_PRODUCT_ID, + .identify = d71_identify, + }, +}; + +const struct of_device_id komeda_of_match[] = { + { .compatible = "arm,mali-d71", .data = &komeda_products[MALI_D71], }, + {}, +}; + +MODULE_DEVICE_TABLE(of, komeda_of_match); + +static struct platform_driver komeda_platform_driver = { + .probe = komeda_platform_probe, + .remove = komeda_platform_remove, + .driver = { + .name = "komeda", + .of_match_table = komeda_of_match, + .pm = NULL, + }, +}; + +module_platform_driver(komeda_platform_driver); + +MODULE_AUTHOR("James.Qian.Wang <james.qian.wang@arm.com>"); +MODULE_DESCRIPTION("Komeda KMS driver"); +MODULE_LICENSE("GPL v2"); diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.c b/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.c new file mode 100644 index 000000000000..1e17bd6107a4 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.c @@ -0,0 +1,75 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ + +#include <linux/slab.h> +#include "komeda_format_caps.h" +#include "malidp_utils.h" + +const struct komeda_format_caps * +komeda_get_format_caps(struct komeda_format_caps_table *table, + u32 fourcc, u64 modifier) +{ + const struct komeda_format_caps *caps; + u64 afbc_features = modifier & ~(AFBC_FORMAT_MOD_BLOCK_SIZE_MASK); + u32 afbc_layout = modifier & AFBC_FORMAT_MOD_BLOCK_SIZE_MASK; + int id; + + for (id = 0; id < table->n_formats; id++) { + caps = &table->format_caps[id]; + + if (fourcc != caps->fourcc) + continue; + + if ((modifier == 0ULL) && (caps->supported_afbc_layouts == 0)) + return caps; + + if (has_bits(afbc_features, caps->supported_afbc_features) && + has_bit(afbc_layout, caps->supported_afbc_layouts)) + return caps; + } + + return NULL; +} + +u32 *komeda_get_layer_fourcc_list(struct komeda_format_caps_table *table, + u32 layer_type, u32 *n_fmts) +{ + const struct komeda_format_caps *cap; + u32 *fmts; + int i, j, n = 0; + + fmts = kcalloc(table->n_formats, sizeof(u32), GFP_KERNEL); + if (!fmts) + return NULL; + + for (i = 0; i < table->n_formats; i++) { + cap = &table->format_caps[i]; + if (!(layer_type & cap->supported_layer_types) || + (cap->fourcc == 0)) + continue; + + /* one fourcc may has two caps items in table (afbc/none-afbc), + * so check the existing list to avoid adding a duplicated one. + */ + for (j = n - 1; j >= 0; j--) + if (fmts[j] == cap->fourcc) + break; + + if (j < 0) + fmts[n++] = cap->fourcc; + } + + if (n_fmts) + *n_fmts = n; + + return fmts; +} + +void komeda_put_fourcc_list(u32 *fourcc_list) +{ + kfree(fourcc_list); +} diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.h b/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.h new file mode 100644 index 000000000000..60f39e77b098 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_format_caps.h @@ -0,0 +1,89 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ + +#ifndef _KOMEDA_FORMAT_CAPS_H_ +#define _KOMEDA_FORMAT_CAPS_H_ + +#include <linux/types.h> +#include <uapi/drm/drm_fourcc.h> +#include <drm/drm_fourcc.h> + +#define AFBC(x) DRM_FORMAT_MOD_ARM_AFBC(x) + +/* afbc layerout */ +#define AFBC_16x16(x) AFBC(AFBC_FORMAT_MOD_BLOCK_SIZE_16x16 | (x)) +#define AFBC_32x8(x) AFBC(AFBC_FORMAT_MOD_BLOCK_SIZE_32x8 | (x)) + +/* afbc features */ +#define _YTR AFBC_FORMAT_MOD_YTR +#define _SPLIT AFBC_FORMAT_MOD_SPLIT +#define _SPARSE AFBC_FORMAT_MOD_SPARSE +#define _CBR AFBC_FORMAT_MOD_CBR +#define _TILED AFBC_FORMAT_MOD_TILED +#define _SC AFBC_FORMAT_MOD_SC + +/* layer_type */ +#define KOMEDA_FMT_RICH_LAYER BIT(0) +#define KOMEDA_FMT_SIMPLE_LAYER BIT(1) +#define KOMEDA_FMT_WB_LAYER BIT(2) + +#define AFBC_TH_LAYOUT_ALIGNMENT 8 +#define AFBC_HEADER_SIZE 16 +#define AFBC_SUPERBLK_ALIGNMENT 128 +#define AFBC_SUPERBLK_PIXELS 256 +#define AFBC_BODY_START_ALIGNMENT 1024 +#define AFBC_TH_BODY_START_ALIGNMENT 4096 + +/** + * struct komeda_format_caps + * + * komeda_format_caps is for describing ARM display specific features and + * limitations for a specific format, and format_caps will be linked into + * &komeda_framebuffer like a extension of &drm_format_info. + * + * NOTE: one fourcc may has two different format_caps items for fourcc and + * fourcc+modifier + * + * @hw_id: hw format id, hw specific value. + * @fourcc: drm fourcc format. + * @tile_size: format tiled size, used by ARM format X0L0/X0L2 + * @supported_layer_types: indicate which layer supports this format + * @supported_rots: allowed rotations for this format + * @supported_afbc_layouts: supported afbc layerout + * @supported_afbc_features: supported afbc features + */ +struct komeda_format_caps { + u32 hw_id; + u32 fourcc; + u32 tile_size; + u32 supported_layer_types; + u32 supported_rots; + u32 supported_afbc_layouts; + u64 supported_afbc_features; +}; + +/** + * struct komeda_format_caps_table - format_caps mananger + * + * @n_formats: the size of format_caps list. + * @format_caps: format_caps list. + */ +struct komeda_format_caps_table { + u32 n_formats; + const struct komeda_format_caps *format_caps; +}; + +const struct komeda_format_caps * +komeda_get_format_caps(struct komeda_format_caps_table *table, + u32 fourcc, u64 modifier); + +u32 *komeda_get_layer_fourcc_list(struct komeda_format_caps_table *table, + u32 layer_type, u32 *n_fmts); + +void komeda_put_fourcc_list(u32 *fourcc_list); + +#endif diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_framebuffer.c b/drivers/gpu/drm/arm/display/komeda/komeda_framebuffer.c new file mode 100644 index 000000000000..23ee74d42239 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_framebuffer.c @@ -0,0 +1,165 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#include <drm/drm_gem.h> +#include <drm/drm_gem_framebuffer_helper.h> +#include <drm/drm_fb_cma_helper.h> +#include <drm/drm_gem_cma_helper.h> +#include "komeda_framebuffer.h" +#include "komeda_dev.h" + +static void komeda_fb_destroy(struct drm_framebuffer *fb) +{ + struct komeda_fb *kfb = to_kfb(fb); + u32 i; + + for (i = 0; i < fb->format->num_planes; i++) + drm_gem_object_put_unlocked(fb->obj[i]); + + drm_framebuffer_cleanup(fb); + kfree(kfb); +} + +static int komeda_fb_create_handle(struct drm_framebuffer *fb, + struct drm_file *file, u32 *handle) +{ + return drm_gem_handle_create(file, fb->obj[0], handle); +} + +static const struct drm_framebuffer_funcs komeda_fb_funcs = { + .destroy = komeda_fb_destroy, + .create_handle = komeda_fb_create_handle, +}; + +static int +komeda_fb_none_afbc_size_check(struct komeda_dev *mdev, struct komeda_fb *kfb, + struct drm_file *file, + const struct drm_mode_fb_cmd2 *mode_cmd) +{ + struct drm_framebuffer *fb = &kfb->base; + struct drm_gem_object *obj; + u32 min_size = 0; + u32 i; + + for (i = 0; i < fb->format->num_planes; i++) { + obj = drm_gem_object_lookup(file, mode_cmd->handles[i]); + if (!obj) { + DRM_DEBUG_KMS("Failed to lookup GEM object\n"); + fb->obj[i] = NULL; + + return -ENOENT; + } + + kfb->aligned_w = fb->width / (i ? fb->format->hsub : 1); + kfb->aligned_h = fb->height / (i ? fb->format->vsub : 1); + + if (fb->pitches[i] % mdev->chip.bus_width) { + DRM_DEBUG_KMS("Pitch[%d]: 0x%x doesn't align to 0x%x\n", + i, fb->pitches[i], mdev->chip.bus_width); + drm_gem_object_put_unlocked(obj); + fb->obj[i] = NULL; + + return -EINVAL; + } + + min_size = ((kfb->aligned_h / kfb->format_caps->tile_size - 1) + * fb->pitches[i]) + + (kfb->aligned_w * fb->format->cpp[i] + * kfb->format_caps->tile_size) + + fb->offsets[i]; + + if (obj->size < min_size) { + DRM_DEBUG_KMS("Fail to check none afbc fb size.\n"); + drm_gem_object_put_unlocked(obj); + fb->obj[i] = NULL; + + return -EINVAL; + } + + fb->obj[i] = obj; + } + + if (fb->format->num_planes == 3) { + if (fb->pitches[1] != fb->pitches[2]) { + DRM_DEBUG_KMS("The pitch[1] and [2] are not same\n"); + return -EINVAL; + } + } + + return 0; +} + +struct drm_framebuffer * +komeda_fb_create(struct drm_device *dev, struct drm_file *file, + const struct drm_mode_fb_cmd2 *mode_cmd) +{ + struct komeda_dev *mdev = dev->dev_private; + struct komeda_fb *kfb; + int ret = 0, i; + + kfb = kzalloc(sizeof(*kfb), GFP_KERNEL); + if (!kfb) + return ERR_PTR(-ENOMEM); + + kfb->format_caps = komeda_get_format_caps(&mdev->fmt_tbl, + mode_cmd->pixel_format, + mode_cmd->modifier[0]); + if (!kfb->format_caps) { + DRM_DEBUG_KMS("FMT %x is not supported.\n", + mode_cmd->pixel_format); + kfree(kfb); + return ERR_PTR(-EINVAL); + } + + drm_helper_mode_fill_fb_struct(dev, &kfb->base, mode_cmd); + + ret = komeda_fb_none_afbc_size_check(mdev, kfb, file, mode_cmd); + if (ret < 0) + goto err_cleanup; + + ret = drm_framebuffer_init(dev, &kfb->base, &komeda_fb_funcs); + if (ret < 0) { + DRM_DEBUG_KMS("failed to initialize fb\n"); + + goto err_cleanup; + } + + return &kfb->base; + +err_cleanup: + for (i = 0; i < kfb->base.format->num_planes; i++) + drm_gem_object_put_unlocked(kfb->base.obj[i]); + + kfree(kfb); + return ERR_PTR(ret); +} + +dma_addr_t +komeda_fb_get_pixel_addr(struct komeda_fb *kfb, int x, int y, int plane) +{ + struct drm_framebuffer *fb = &kfb->base; + const struct drm_gem_cma_object *obj; + u32 plane_x, plane_y, cpp, pitch, offset; + + if (plane >= fb->format->num_planes) { + DRM_DEBUG_KMS("Out of max plane num.\n"); + return -EINVAL; + } + + obj = drm_fb_cma_get_gem_obj(fb, plane); + + offset = fb->offsets[plane]; + if (!fb->modifier) { + plane_x = x / (plane ? fb->format->hsub : 1); + plane_y = y / (plane ? fb->format->vsub : 1); + cpp = fb->format->cpp[plane]; + pitch = fb->pitches[plane]; + offset += plane_x * cpp * kfb->format_caps->tile_size + + (plane_y * pitch) / kfb->format_caps->tile_size; + } + + return obj->paddr + offset; +} diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_framebuffer.h b/drivers/gpu/drm/arm/display/komeda/komeda_framebuffer.h new file mode 100644 index 000000000000..0de2e4a2afd2 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_framebuffer.h @@ -0,0 +1,34 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#ifndef _KOMEDA_FRAMEBUFFER_H_ +#define _KOMEDA_FRAMEBUFFER_H_ + +#include <drm/drm_framebuffer.h> +#include "komeda_format_caps.h" + +/** struct komeda_fb - entend drm_framebuffer with komeda attribute */ +struct komeda_fb { + /** @base: &drm_framebuffer */ + struct drm_framebuffer base; + /* @format_caps: &komeda_format_caps */ + const struct komeda_format_caps *format_caps; + /** @aligned_w: aligned frame buffer width */ + u32 aligned_w; + /** @aligned_h: aligned frame buffer height */ + u32 aligned_h; +}; + +#define to_kfb(dfb) container_of(dfb, struct komeda_fb, base) + +struct drm_framebuffer * +komeda_fb_create(struct drm_device *dev, struct drm_file *file, + const struct drm_mode_fb_cmd2 *mode_cmd); +dma_addr_t +komeda_fb_get_pixel_addr(struct komeda_fb *kfb, int x, int y, int plane); +bool komeda_fb_is_layer_supported(struct komeda_fb *kfb, u32 layer_type); + +#endif diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_kms.c b/drivers/gpu/drm/arm/display/komeda/komeda_kms.c new file mode 100644 index 000000000000..3fc096d3883e --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_kms.c @@ -0,0 +1,167 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#include <linux/component.h> +#include <drm/drm_atomic.h> +#include <drm/drm_atomic_helper.h> +#include <drm/drm_gem_framebuffer_helper.h> +#include <drm/drm_gem_cma_helper.h> +#include <drm/drm_fb_helper.h> +#include <linux/interrupt.h> +#include "komeda_dev.h" +#include "komeda_kms.h" +#include "komeda_framebuffer.h" + +DEFINE_DRM_GEM_CMA_FOPS(komeda_cma_fops); + +static int komeda_gem_cma_dumb_create(struct drm_file *file, + struct drm_device *dev, + struct drm_mode_create_dumb *args) +{ + u32 alignment = 16; /* TODO get alignment from dev */ + + args->pitch = ALIGN(DIV_ROUND_UP(args->width * args->bpp, 8), + alignment); + + return drm_gem_cma_dumb_create_internal(file, dev, args); +} + +static struct drm_driver komeda_kms_driver = { + .driver_features = DRIVER_GEM | DRIVER_MODESET | DRIVER_ATOMIC | + DRIVER_PRIME, + .lastclose = drm_fb_helper_lastclose, + .gem_free_object_unlocked = drm_gem_cma_free_object, + .gem_vm_ops = &drm_gem_cma_vm_ops, + .dumb_create = komeda_gem_cma_dumb_create, + .prime_handle_to_fd = drm_gem_prime_handle_to_fd, + .prime_fd_to_handle = drm_gem_prime_fd_to_handle, + .gem_prime_export = drm_gem_prime_export, + .gem_prime_import = drm_gem_prime_import, + .gem_prime_get_sg_table = drm_gem_cma_prime_get_sg_table, + .gem_prime_import_sg_table = drm_gem_cma_prime_import_sg_table, + .gem_prime_vmap = drm_gem_cma_prime_vmap, + .gem_prime_vunmap = drm_gem_cma_prime_vunmap, + .gem_prime_mmap = drm_gem_cma_prime_mmap, + .fops = &komeda_cma_fops, + .name = "komeda", + .desc = "Arm Komeda Display Processor driver", + .date = "20181101", + .major = 0, + .minor = 1, +}; + +static void komeda_kms_commit_tail(struct drm_atomic_state *old_state) +{ + struct drm_device *dev = old_state->dev; + + drm_atomic_helper_commit_modeset_disables(dev, old_state); + + drm_atomic_helper_commit_planes(dev, old_state, 0); + + drm_atomic_helper_commit_modeset_enables(dev, old_state); + + drm_atomic_helper_wait_for_flip_done(dev, old_state); + + drm_atomic_helper_commit_hw_done(old_state); + + drm_atomic_helper_cleanup_planes(dev, old_state); +} + +static const struct drm_mode_config_helper_funcs komeda_mode_config_helpers = { + .atomic_commit_tail = komeda_kms_commit_tail, +}; + +static const struct drm_mode_config_funcs komeda_mode_config_funcs = { + .fb_create = komeda_fb_create, + .atomic_check = drm_atomic_helper_check, + .atomic_commit = drm_atomic_helper_commit, +}; + +static void komeda_kms_mode_config_init(struct komeda_kms_dev *kms, + struct komeda_dev *mdev) +{ + struct drm_mode_config *config = &kms->base.mode_config; + + drm_mode_config_init(&kms->base); + + komeda_kms_setup_crtcs(kms, mdev); + + /* Get value from dev */ + config->min_width = 0; + config->min_height = 0; + config->max_width = 4096; + config->max_height = 4096; + config->allow_fb_modifiers = false; + + config->funcs = &komeda_mode_config_funcs; + config->helper_private = &komeda_mode_config_helpers; +} + +struct komeda_kms_dev *komeda_kms_attach(struct komeda_dev *mdev) +{ + struct komeda_kms_dev *kms = kzalloc(sizeof(*kms), GFP_KERNEL); + struct drm_device *drm; + int err; + + if (!kms) + return ERR_PTR(-ENOMEM); + + drm = &kms->base; + err = drm_dev_init(drm, &komeda_kms_driver, mdev->dev); + if (err) + goto free_kms; + + drm->dev_private = mdev; + + komeda_kms_mode_config_init(kms, mdev); + + err = komeda_kms_add_private_objs(kms, mdev); + if (err) + goto cleanup_mode_config; + + err = komeda_kms_add_planes(kms, mdev); + if (err) + goto cleanup_mode_config; + + err = drm_vblank_init(drm, kms->n_crtcs); + if (err) + goto cleanup_mode_config; + + err = komeda_kms_add_crtcs(kms, mdev); + if (err) + goto cleanup_mode_config; + + err = component_bind_all(mdev->dev, kms); + if (err) + goto cleanup_mode_config; + + drm_mode_config_reset(drm); + + err = drm_dev_register(drm, 0); + if (err) + goto cleanup_mode_config; + + return kms; + +cleanup_mode_config: + drm_mode_config_cleanup(drm); +free_kms: + kfree(kms); + return ERR_PTR(err); +} + +void komeda_kms_detach(struct komeda_kms_dev *kms) +{ + struct drm_device *drm = &kms->base; + struct komeda_dev *mdev = drm->dev_private; + + drm_dev_unregister(drm); + component_unbind_all(mdev->dev, drm); + komeda_kms_cleanup_private_objs(mdev); + drm_mode_config_cleanup(drm); + drm->dev_private = NULL; + drm_dev_put(drm); +} diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_kms.h b/drivers/gpu/drm/arm/display/komeda/komeda_kms.h new file mode 100644 index 000000000000..f13666004a42 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_kms.h @@ -0,0 +1,113 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#ifndef _KOMEDA_KMS_H_ +#define _KOMEDA_KMS_H_ + +#include <drm/drm_atomic.h> +#include <drm/drm_atomic_helper.h> +#include <drm/drm_crtc_helper.h> +#include <drm/drm_writeback.h> + +/** struct komeda_plane - komeda instance of drm_plane */ +struct komeda_plane { + /** @base: &drm_plane */ + struct drm_plane base; + /** + * @layer: + * + * represents available layer input pipelines for this plane. + * + * NOTE: + * the layer is not for a specific Layer, but indicate a group of + * Layers with same capabilities. + */ + struct komeda_layer *layer; +}; + +/** + * struct komeda_plane_state + * + * The plane_state can be split into two data flow (left/right) and handled + * by two layers &komeda_plane.layer and &komeda_plane.layer.right + */ +struct komeda_plane_state { + /** @base: &drm_plane_state */ + struct drm_plane_state base; + + /* private properties */ +}; + +/** + * struct komeda_wb_connector + */ +struct komeda_wb_connector { + /** @base: &drm_writeback_connector */ + struct drm_writeback_connector base; + + /** @wb_layer: represents associated writeback pipeline of komeda */ + struct komeda_layer *wb_layer; +}; + +/** + * struct komeda_crtc + */ +struct komeda_crtc { + /** @base: &drm_crtc */ + struct drm_crtc base; + /** @master: only master has display output */ + struct komeda_pipeline *master; + /** + * @slave: optional + * + * Doesn't have its own display output, the handled data flow will + * merge into the master. + */ + struct komeda_pipeline *slave; +}; + +/** struct komeda_crtc_state */ +struct komeda_crtc_state { + /** @base: &drm_crtc_state */ + struct drm_crtc_state base; + + /* private properties */ + + /* computed state which are used by validate/check */ + u32 affected_pipes; + u32 active_pipes; +}; + +/** struct komeda_kms_dev - for gather KMS related things */ +struct komeda_kms_dev { + /** @base: &drm_device */ + struct drm_device base; + + /** @n_crtcs: valid numbers of crtcs in &komeda_kms_dev.crtcs */ + int n_crtcs; + /** @crtcs: crtcs list */ + struct komeda_crtc crtcs[KOMEDA_MAX_PIPELINES]; +}; + +#define to_kplane(p) container_of(p, struct komeda_plane, base) +#define to_kplane_st(p) container_of(p, struct komeda_plane_state, base) +#define to_kconn(p) container_of(p, struct komeda_wb_connector, base) +#define to_kcrtc(p) container_of(p, struct komeda_crtc, base) +#define to_kcrtc_st(p) container_of(p, struct komeda_crtc_state, base) +#define to_kdev(p) container_of(p, struct komeda_kms_dev, base) + +int komeda_kms_setup_crtcs(struct komeda_kms_dev *kms, struct komeda_dev *mdev); + +int komeda_kms_add_crtcs(struct komeda_kms_dev *kms, struct komeda_dev *mdev); +int komeda_kms_add_planes(struct komeda_kms_dev *kms, struct komeda_dev *mdev); +int komeda_kms_add_private_objs(struct komeda_kms_dev *kms, + struct komeda_dev *mdev); +void komeda_kms_cleanup_private_objs(struct komeda_dev *mdev); + +struct komeda_kms_dev *komeda_kms_attach(struct komeda_dev *mdev); +void komeda_kms_detach(struct komeda_kms_dev *kms); + +#endif /*_KOMEDA_KMS_H_*/ diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_pipeline.c b/drivers/gpu/drm/arm/display/komeda/komeda_pipeline.c new file mode 100644 index 000000000000..edb1cd7795f9 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_pipeline.c @@ -0,0 +1,200 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#include "komeda_dev.h" +#include "komeda_pipeline.h" + +/** komeda_pipeline_add - Add a pipeline to &komeda_dev */ +struct komeda_pipeline * +komeda_pipeline_add(struct komeda_dev *mdev, size_t size, + struct komeda_pipeline_funcs *funcs) +{ + struct komeda_pipeline *pipe; + + if (mdev->n_pipelines + 1 > KOMEDA_MAX_PIPELINES) { + DRM_ERROR("Exceed max support %d pipelines.\n", + KOMEDA_MAX_PIPELINES); + return NULL; + } + + if (size < sizeof(*pipe)) { + DRM_ERROR("Request pipeline size too small.\n"); + return NULL; + } + + pipe = devm_kzalloc(mdev->dev, size, GFP_KERNEL); + if (!pipe) + return NULL; + + pipe->mdev = mdev; + pipe->id = mdev->n_pipelines; + pipe->funcs = funcs; + + mdev->pipelines[mdev->n_pipelines] = pipe; + mdev->n_pipelines++; + + return pipe; +} + +void komeda_pipeline_destroy(struct komeda_dev *mdev, + struct komeda_pipeline *pipe) +{ + struct komeda_component *c; + int i; + + dp_for_each_set_bit(i, pipe->avail_comps) { + c = komeda_pipeline_get_component(pipe, i); + komeda_component_destroy(mdev, c); + } + + clk_put(pipe->pxlclk); + clk_put(pipe->aclk); + + of_node_put(pipe->of_output_dev); + of_node_put(pipe->of_output_port); + of_node_put(pipe->of_node); + + devm_kfree(mdev->dev, pipe); +} + +struct komeda_component ** +komeda_pipeline_get_component_pos(struct komeda_pipeline *pipe, int id) +{ + struct komeda_dev *mdev = pipe->mdev; + struct komeda_pipeline *temp = NULL; + struct komeda_component **pos = NULL; + + switch (id) { + case KOMEDA_COMPONENT_LAYER0: + case KOMEDA_COMPONENT_LAYER1: + case KOMEDA_COMPONENT_LAYER2: + case KOMEDA_COMPONENT_LAYER3: + pos = to_cpos(pipe->layers[id - KOMEDA_COMPONENT_LAYER0]); + break; + case KOMEDA_COMPONENT_WB_LAYER: + pos = to_cpos(pipe->wb_layer); + break; + case KOMEDA_COMPONENT_COMPIZ0: + case KOMEDA_COMPONENT_COMPIZ1: + temp = mdev->pipelines[id - KOMEDA_COMPONENT_COMPIZ0]; + if (!temp) { + DRM_ERROR("compiz-%d doesn't exist.\n", id); + return NULL; + } + pos = to_cpos(temp->compiz); + break; + case KOMEDA_COMPONENT_SCALER0: + case KOMEDA_COMPONENT_SCALER1: + pos = to_cpos(pipe->scalers[id - KOMEDA_COMPONENT_SCALER0]); + break; + case KOMEDA_COMPONENT_IPS0: + case KOMEDA_COMPONENT_IPS1: + temp = mdev->pipelines[id - KOMEDA_COMPONENT_IPS0]; + if (!temp) { + DRM_ERROR("ips-%d doesn't exist.\n", id); + return NULL; + } + pos = to_cpos(temp->improc); + break; + case KOMEDA_COMPONENT_TIMING_CTRLR: + pos = to_cpos(pipe->ctrlr); + break; + default: + pos = NULL; + DRM_ERROR("Unknown pipeline resource ID: %d.\n", id); + break; + } + + return pos; +} + +struct komeda_component * +komeda_pipeline_get_component(struct komeda_pipeline *pipe, int id) +{ + struct komeda_component **pos = NULL; + struct komeda_component *c = NULL; + + pos = komeda_pipeline_get_component_pos(pipe, id); + if (pos) + c = *pos; + + return c; +} + +/** komeda_component_add - Add a component to &komeda_pipeline */ +struct komeda_component * +komeda_component_add(struct komeda_pipeline *pipe, + size_t comp_sz, u32 id, u32 hw_id, + struct komeda_component_funcs *funcs, + u8 max_active_inputs, u32 supported_inputs, + u8 max_active_outputs, u32 __iomem *reg, + const char *name_fmt, ...) +{ + struct komeda_component **pos; + struct komeda_component *c; + int idx, *num = NULL; + + if (max_active_inputs > KOMEDA_COMPONENT_N_INPUTS) { + WARN(1, "please large KOMEDA_COMPONENT_N_INPUTS to %d.\n", + max_active_inputs); + return NULL; + } + + pos = komeda_pipeline_get_component_pos(pipe, id); + if (!pos || (*pos)) + return NULL; + + if (has_bit(id, KOMEDA_PIPELINE_LAYERS)) { + idx = id - KOMEDA_COMPONENT_LAYER0; + num = &pipe->n_layers; + if (idx != pipe->n_layers) { + DRM_ERROR("please add Layer by id sequence.\n"); + return NULL; + } + } else if (has_bit(id, KOMEDA_PIPELINE_SCALERS)) { + idx = id - KOMEDA_COMPONENT_SCALER0; + num = &pipe->n_scalers; + if (idx != pipe->n_scalers) { + DRM_ERROR("please add Scaler by id sequence.\n"); + return NULL; + } + } + + c = devm_kzalloc(pipe->mdev->dev, comp_sz, GFP_KERNEL); + if (!c) + return NULL; + + c->id = id; + c->hw_id = hw_id; + c->reg = reg; + c->pipeline = pipe; + c->max_active_inputs = max_active_inputs; + c->max_active_outputs = max_active_outputs; + c->supported_inputs = supported_inputs; + c->funcs = funcs; + + if (name_fmt) { + va_list args; + + va_start(args, name_fmt); + vsnprintf(c->name, sizeof(c->name), name_fmt, args); + va_end(args); + } + + if (num) + *num = *num + 1; + + pipe->avail_comps |= BIT(c->id); + *pos = c; + + return c; +} + +void komeda_component_destroy(struct komeda_dev *mdev, + struct komeda_component *c) +{ + devm_kfree(mdev->dev, c); +} diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_pipeline.h b/drivers/gpu/drm/arm/display/komeda/komeda_pipeline.h new file mode 100644 index 000000000000..8c950bc8ae96 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_pipeline.h @@ -0,0 +1,359 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#ifndef _KOMEDA_PIPELINE_H_ +#define _KOMEDA_PIPELINE_H_ + +#include <linux/types.h> +#include <drm/drm_atomic.h> +#include <drm/drm_atomic_helper.h> +#include "malidp_utils.h" + +#define KOMEDA_MAX_PIPELINES 2 +#define KOMEDA_PIPELINE_MAX_LAYERS 4 +#define KOMEDA_PIPELINE_MAX_SCALERS 2 +#define KOMEDA_COMPONENT_N_INPUTS 5 + +/* pipeline component IDs */ +enum { + KOMEDA_COMPONENT_LAYER0 = 0, + KOMEDA_COMPONENT_LAYER1 = 1, + KOMEDA_COMPONENT_LAYER2 = 2, + KOMEDA_COMPONENT_LAYER3 = 3, + KOMEDA_COMPONENT_WB_LAYER = 7, /* write back layer */ + KOMEDA_COMPONENT_SCALER0 = 8, + KOMEDA_COMPONENT_SCALER1 = 9, + KOMEDA_COMPONENT_SPLITTER = 12, + KOMEDA_COMPONENT_MERGER = 14, + KOMEDA_COMPONENT_COMPIZ0 = 16, /* compositor */ + KOMEDA_COMPONENT_COMPIZ1 = 17, + KOMEDA_COMPONENT_IPS0 = 20, /* post image processor */ + KOMEDA_COMPONENT_IPS1 = 21, + KOMEDA_COMPONENT_TIMING_CTRLR = 22, /* timing controller */ +}; + +#define KOMEDA_PIPELINE_LAYERS (BIT(KOMEDA_COMPONENT_LAYER0) |\ + BIT(KOMEDA_COMPONENT_LAYER1) |\ + BIT(KOMEDA_COMPONENT_LAYER2) |\ + BIT(KOMEDA_COMPONENT_LAYER3)) + +#define KOMEDA_PIPELINE_SCALERS (BIT(KOMEDA_COMPONENT_SCALER0) |\ + BIT(KOMEDA_COMPONENT_SCALER1)) + +#define KOMEDA_PIPELINE_COMPIZS (BIT(KOMEDA_COMPONENT_COMPIZ0) |\ + BIT(KOMEDA_COMPONENT_COMPIZ1)) + +#define KOMEDA_PIPELINE_IMPROCS (BIT(KOMEDA_COMPONENT_IPS0) |\ + BIT(KOMEDA_COMPONENT_IPS1)) +struct komeda_component; +struct komeda_component_state; + +/** komeda_component_funcs - component control functions */ +struct komeda_component_funcs { + /** @validate: optional, + * component may has special requirements or limitations, this function + * supply HW the ability to do the further HW specific check. + */ + int (*validate)(struct komeda_component *c, + struct komeda_component_state *state); + /** @update: update is a active update */ + void (*update)(struct komeda_component *c, + struct komeda_component_state *state); + /** @disable: disable component */ + void (*disable)(struct komeda_component *c); + /** @dump_register: Optional, dump registers to seq_file */ + void (*dump_register)(struct komeda_component *c, struct seq_file *seq); +}; + +/** + * struct komeda_component + * + * struct komeda_component describe the data flow capabilities for how to link a + * component into the display pipeline. + * all specified components are subclass of this structure. + */ +struct komeda_component { + /** @obj: treat component as private obj */ + struct drm_private_obj obj; + /** @pipeline: the komeda pipeline this component belongs to */ + struct komeda_pipeline *pipeline; + /** @name: component name */ + char name[32]; + /** + * @reg: + * component register base, + * which is initialized by chip and used by chip only + */ + u32 __iomem *reg; + /** @id: component id */ + u32 id; + /** @hw_ic: component hw id, + * which is initialized by chip and used by chip only + */ + u32 hw_id; + + /** + * @max_active_inputs: + * @max_active_outpus: + * + * maximum number of inputs/outputs that can be active in the same time + * Note: + * the number isn't the bit number of @supported_inputs or + * @supported_outputs, but may be less than it, since component may not + * support enabling all @supported_inputs/outputs at the same time. + */ + u8 max_active_inputs; + u8 max_active_outputs; + /** + * @supported_inputs: + * @supported_outputs: + * + * bitmask of BIT(component->id) for the supported inputs/outputs + * describes the possibilities of how a component is linked into a + * pipeline. + */ + u32 supported_inputs; + u32 supported_outputs; + + /** + * @funcs: chip functions to access HW + */ + struct komeda_component_funcs *funcs; +}; + +/** + * struct komeda_component_output + * + * a component has multiple outputs, if want to know where the data + * comes from, only know the component is not enough, we still need to know + * its output port + */ +struct komeda_component_output { + /** @component: indicate which component the data comes from */ + struct komeda_component *component; + /** @output_port: + * the output port of the &komeda_component_output.component + */ + u8 output_port; +}; + +/** + * struct komeda_component_state + * + * component_state is the data flow configuration of the component, and it's + * the superclass of all specific component_state like @komeda_layer_state, + * @komeda_scaler_state + */ +struct komeda_component_state { + /** @obj: tracking component_state by drm_atomic_state */ + struct drm_private_state obj; + struct komeda_component *component; + /** + * @binding_user: + * currently bound user, the user can be crtc/plane/wb_conn, which is + * valid decided by @component and @inputs + * + * - Layer: its user always is plane. + * - compiz/improc/timing_ctrlr: the user is crtc. + * - wb_layer: wb_conn; + * - scaler: plane when input is layer, wb_conn if input is compiz. + */ + union { + struct drm_crtc *crtc; + struct drm_plane *plane; + struct drm_connector *wb_conn; + void *binding_user; + }; + /** + * @active_inputs: + * + * active_inputs is bitmask of @inputs index + * + * - active_inputs = changed_active_inputs + unchanged_active_inputs + * - affected_inputs = old->active_inputs + new->active_inputs; + * - disabling_inputs = affected_inputs ^ active_inputs; + * - changed_inputs = disabling_inputs + changed_active_inputs; + * + * NOTE: + * changed_inputs doesn't include all active_input but only + * @changed_active_inputs, and this bitmask can be used in chip + * level for dirty update. + */ + u16 active_inputs; + u16 changed_active_inputs; + u16 affected_inputs; + /** + * @inputs: + * + * the specific inputs[i] only valid on BIT(i) has been set in + * @active_inputs, if not the inputs[i] is undefined. + */ + struct komeda_component_output inputs[KOMEDA_COMPONENT_N_INPUTS]; +}; + +static inline u16 component_disabling_inputs(struct komeda_component_state *st) +{ + return st->affected_inputs ^ st->active_inputs; +} + +static inline u16 component_changed_inputs(struct komeda_component_state *st) +{ + return component_disabling_inputs(st) | st->changed_active_inputs; +} + +#define to_comp(__c) (((__c) == NULL) ? NULL : &((__c)->base)) +#define to_cpos(__c) ((struct komeda_component **)&(__c)) + +/* these structures are going to be filled in in uture patches */ +struct komeda_layer { + struct komeda_component base; + /* layer specific features and caps */ + int layer_type; /* RICH, SIMPLE or WB */ +}; + +struct komeda_layer_state { + struct komeda_component_state base; + /* layer specific configuration state */ +}; + +struct komeda_compiz { + struct komeda_component base; + /* compiz specific features and caps */ +}; + +struct komeda_compiz_state { + struct komeda_component_state base; + /* compiz specific configuration state */ +}; + +struct komeda_scaler { + struct komeda_component base; + /* scaler features and caps */ +}; + +struct komeda_scaler_state { + struct komeda_component_state base; +}; + +struct komeda_improc { + struct komeda_component base; +}; + +struct komeda_improc_state { + struct komeda_component_state base; +}; + +/* display timing controller */ +struct komeda_timing_ctrlr { + struct komeda_component base; +}; + +struct komeda_timing_ctrlr_state { + struct komeda_component_state base; +}; + +/** struct komeda_pipeline_funcs */ +struct komeda_pipeline_funcs { + /* dump_register: Optional, dump registers to seq_file */ + void (*dump_register)(struct komeda_pipeline *pipe, + struct seq_file *sf); +}; + +/** + * struct komeda_pipeline + * + * Represent a complete display pipeline and hold all functional components. + */ +struct komeda_pipeline { + /** @obj: link pipeline as private obj of drm_atomic_state */ + struct drm_private_obj obj; + /** @mdev: the parent komeda_dev */ + struct komeda_dev *mdev; + /** @pxlclk: pixel clock */ + struct clk *pxlclk; + /** @aclk: AXI clock */ + struct clk *aclk; + /** @id: pipeline id */ + int id; + /** @avail_comps: available components mask of pipeline */ + u32 avail_comps; + int n_layers; + struct komeda_layer *layers[KOMEDA_PIPELINE_MAX_LAYERS]; + int n_scalers; + struct komeda_scaler *scalers[KOMEDA_PIPELINE_MAX_SCALERS]; + struct komeda_compiz *compiz; + struct komeda_layer *wb_layer; + struct komeda_improc *improc; + struct komeda_timing_ctrlr *ctrlr; + struct komeda_pipeline_funcs *funcs; /* private pipeline functions */ + + /** @of_node: pipeline dt node */ + struct device_node *of_node; + /** @of_output_port: pipeline output port */ + struct device_node *of_output_port; + /** @of_output_dev: output connector device node */ + struct device_node *of_output_dev; +}; + +/** + * struct komeda_pipeline_state + * + * NOTE: + * Unlike the pipeline, pipeline_state doesn’t gather any component_state + * into it. It because all component will be managed by drm_atomic_state. + */ +struct komeda_pipeline_state { + /** @obj: tracking pipeline_state by drm_atomic_state */ + struct drm_private_state obj; + struct komeda_pipeline *pipe; + /** @crtc: currently bound crtc */ + struct drm_crtc *crtc; + /** + * @active_comps: + * + * bitmask - BIT(component->id) of active components + */ + u32 active_comps; +}; + +#define to_layer(c) container_of(c, struct komeda_layer, base) +#define to_compiz(c) container_of(c, struct komeda_compiz, base) +#define to_scaler(c) container_of(c, struct komeda_scaler, base) +#define to_improc(c) container_of(c, struct komeda_improc, base) +#define to_ctrlr(c) container_of(c, struct komeda_timing_ctrlr, base) + +#define to_layer_st(c) container_of(c, struct komeda_layer_state, base) +#define to_compiz_st(c) container_of(c, struct komeda_compiz_state, base) +#define to_scaler_st(c) container_of(c, struct komeda_scaler_state, base) +#define to_improc_st(c) container_of(c, struct komeda_improc_state, base) +#define to_ctrlr_st(c) container_of(c, struct komeda_timing_ctrlr_state, base) + +#define priv_to_comp_st(o) container_of(o, struct komeda_component_state, obj) +#define priv_to_pipe_st(o) container_of(o, struct komeda_pipeline_state, obj) + +/* pipeline APIs */ +struct komeda_pipeline * +komeda_pipeline_add(struct komeda_dev *mdev, size_t size, + struct komeda_pipeline_funcs *funcs); +void komeda_pipeline_destroy(struct komeda_dev *mdev, + struct komeda_pipeline *pipe); + +struct komeda_component * +komeda_pipeline_get_component(struct komeda_pipeline *pipe, int id); + +/* component APIs */ +struct komeda_component * +komeda_component_add(struct komeda_pipeline *pipe, + size_t comp_sz, u32 id, u32 hw_id, + struct komeda_component_funcs *funcs, + u8 max_active_inputs, u32 supported_inputs, + u8 max_active_outputs, u32 __iomem *reg, + const char *name_fmt, ...); + +void komeda_component_destroy(struct komeda_dev *mdev, + struct komeda_component *c); + +#endif /* _KOMEDA_PIPELINE_H_*/ diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_plane.c b/drivers/gpu/drm/arm/display/komeda/komeda_plane.c new file mode 100644 index 000000000000..0a4953a9a909 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_plane.c @@ -0,0 +1,109 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#include <drm/drm_atomic.h> +#include <drm/drm_atomic_helper.h> +#include <drm/drm_plane_helper.h> +#include "komeda_dev.h" +#include "komeda_kms.h" + +static const struct drm_plane_helper_funcs komeda_plane_helper_funcs = { +}; + +static void komeda_plane_destroy(struct drm_plane *plane) +{ + drm_plane_cleanup(plane); + + kfree(to_kplane(plane)); +} + +static const struct drm_plane_funcs komeda_plane_funcs = { +}; + +/* for komeda, which is pipeline can be share between crtcs */ +static u32 get_possible_crtcs(struct komeda_kms_dev *kms, + struct komeda_pipeline *pipe) +{ + struct komeda_crtc *crtc; + u32 possible_crtcs = 0; + int i; + + for (i = 0; i < kms->n_crtcs; i++) { + crtc = &kms->crtcs[i]; + + if ((pipe == crtc->master) || (pipe == crtc->slave)) + possible_crtcs |= BIT(i); + } + + return possible_crtcs; +} + +/* use Layer0 as primary */ +static u32 get_plane_type(struct komeda_kms_dev *kms, + struct komeda_component *c) +{ + bool is_primary = (c->id == KOMEDA_COMPONENT_LAYER0); + + return is_primary ? DRM_PLANE_TYPE_PRIMARY : DRM_PLANE_TYPE_OVERLAY; +} + +static int komeda_plane_add(struct komeda_kms_dev *kms, + struct komeda_layer *layer) +{ + struct komeda_dev *mdev = kms->base.dev_private; + struct komeda_component *c = &layer->base; + struct komeda_plane *kplane; + struct drm_plane *plane; + u32 *formats, n_formats = 0; + int err; + + kplane = kzalloc(sizeof(*kplane), GFP_KERNEL); + if (!kplane) + return -ENOMEM; + + plane = &kplane->base; + kplane->layer = layer; + + formats = komeda_get_layer_fourcc_list(&mdev->fmt_tbl, + layer->layer_type, &n_formats); + + err = drm_universal_plane_init(&kms->base, plane, + get_possible_crtcs(kms, c->pipeline), + &komeda_plane_funcs, + formats, n_formats, NULL, + get_plane_type(kms, c), + "%s", c->name); + + komeda_put_fourcc_list(formats); + + if (err) + goto cleanup; + + drm_plane_helper_add(plane, &komeda_plane_helper_funcs); + + return 0; +cleanup: + komeda_plane_destroy(plane); + return err; +} + +int komeda_kms_add_planes(struct komeda_kms_dev *kms, struct komeda_dev *mdev) +{ + struct komeda_pipeline *pipe; + int i, j, err; + + for (i = 0; i < mdev->n_pipelines; i++) { + pipe = mdev->pipelines[i]; + + for (j = 0; j < pipe->n_layers; j++) { + err = komeda_plane_add(kms, pipe->layers[j]); + if (err) + return err; + } + } + + return 0; +} diff --git a/drivers/gpu/drm/arm/display/komeda/komeda_private_obj.c b/drivers/gpu/drm/arm/display/komeda/komeda_private_obj.c new file mode 100644 index 000000000000..f1c9e3fefa86 --- /dev/null +++ b/drivers/gpu/drm/arm/display/komeda/komeda_private_obj.c @@ -0,0 +1,88 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * (C) COPYRIGHT 2018 ARM Limited. All rights reserved. + * Author: James.Qian.Wang <james.qian.wang@arm.com> + * + */ +#include "komeda_dev.h" +#include "komeda_kms.h" + +static struct drm_private_state * +komeda_pipeline_atomic_duplicate_state(struct drm_private_obj *obj) +{ + struct komeda_pipeline_state *st; + + st = kmemdup(obj->state, sizeof(*st), GFP_KERNEL); + if (!st) + return NULL; + + st->active_comps = 0; + + __drm_atomic_helper_private_obj_duplicate_state(obj, &st->obj); + + return &st->obj; +} + +static void +komeda_pipeline_atomic_destroy_state(struct drm_private_obj *obj, + struct drm_private_state *state) +{ + kfree(priv_to_pipe_st(state)); +} + +static const struct drm_private_state_funcs komeda_pipeline_obj_funcs = { + .atomic_duplicate_state = komeda_pipeline_atomic_duplicate_state, + .atomic_destroy_state = komeda_pipeline_atomic_destroy_state, +}; + +static int komeda_pipeline_obj_add(struct komeda_kms_dev *kms, + struct komeda_pipeline *pipe) +{ + struct komeda_pipeline_state *st; + + st = kzalloc(sizeof(*st), GFP_KERNEL); + if (!st) + return -ENOMEM; + + st->pipe = pipe; + drm_atomic_private_obj_init(&kms->base, &pipe->obj, &st->obj, + &komeda_pipeline_obj_funcs); + + return 0; +} + +int komeda_kms_add_private_objs(struct komeda_kms_dev *kms, + struct komeda_dev *mdev) +{ + struct komeda_pipeline *pipe; + int i, err; + + for (i = 0; i < mdev->n_pipelines; i++) { + pipe = mdev->pipelines[i]; + + err = komeda_pipeline_obj_add(kms, pipe); + if (err) + return err; + + /* Add component */ + } + + return 0; +} + +void komeda_kms_cleanup_private_objs(struct komeda_dev *mdev) +{ + struct komeda_pipeline *pipe; + struct komeda_component *c; + int i, id; + + for (i = 0; i < mdev->n_pipelines; i++) { + pipe = mdev->pipelines[i]; + dp_for_each_set_bit(id, pipe->avail_comps) { + c = komeda_pipeline_get_component(pipe, id); + + drm_atomic_private_obj_fini(&c->obj); + } + drm_atomic_private_obj_fini(&pipe->obj); + } +} diff --git a/drivers/gpu/drm/ast/ast_main.c b/drivers/gpu/drm/ast/ast_main.c index 373700c05a00..2854399856ba 100644 --- a/drivers/gpu/drm/ast/ast_main.c +++ b/drivers/gpu/drm/ast/ast_main.c @@ -639,13 +639,9 @@ int ast_dumb_create(struct drm_file *file, static void ast_bo_unref(struct ast_bo **bo) { - struct ttm_buffer_object *tbo; - if ((*bo) == NULL) return; - - tbo = &((*bo)->bo); - ttm_bo_unref(&tbo); + ttm_bo_put(&((*bo)->bo)); *bo = NULL; } diff --git a/drivers/gpu/drm/drm_color_mgmt.c b/drivers/gpu/drm/drm_color_mgmt.c index 07dcf47daafe..d5d34d0c79c7 100644 --- a/drivers/gpu/drm/drm_color_mgmt.c +++ b/drivers/gpu/drm/drm_color_mgmt.c @@ -462,3 +462,46 @@ int drm_plane_create_color_properties(struct drm_plane *plane, return 0; } EXPORT_SYMBOL(drm_plane_create_color_properties); + +/** + * drm_color_lut_check - check validity of lookup table + * @lut: property blob containing LUT to check + * @tests: bitmask of tests to run + * + * Helper to check whether a userspace-provided lookup table is valid and + * satisfies hardware requirements. Drivers pass a bitmask indicating which of + * the tests in &drm_color_lut_tests should be performed. + * + * Returns 0 on success, -EINVAL on failure. + */ +int drm_color_lut_check(const struct drm_property_blob *lut, u32 tests) +{ + const struct drm_color_lut *entry; + int i; + + if (!lut || !tests) + return 0; + + entry = lut->data; + for (i = 0; i < drm_color_lut_size(lut); i++) { + if (tests & DRM_COLOR_LUT_EQUAL_CHANNELS) { + if (entry[i].red != entry[i].blue || + entry[i].red != entry[i].green) { + DRM_DEBUG_KMS("All LUT entries must have equal r/g/b\n"); + return -EINVAL; + } + } + + if (i > 0 && tests & DRM_COLOR_LUT_NON_DECREASING) { + if (entry[i].red < entry[i - 1].red || + entry[i].green < entry[i - 1].green || + entry[i].blue < entry[i - 1].blue) { + DRM_DEBUG_KMS("LUT entries must never decrease.\n"); + return -EINVAL; + } + } + } + + return 0; +} +EXPORT_SYMBOL(drm_color_lut_check); diff --git a/drivers/gpu/drm/etnaviv/etnaviv_sched.c b/drivers/gpu/drm/etnaviv/etnaviv_sched.c index 49a6763693f1..67ae26602024 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_sched.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_sched.c @@ -109,16 +109,19 @@ static void etnaviv_sched_timedout_job(struct drm_sched_job *sched_job) } /* block scheduler */ - kthread_park(gpu->sched.thread); - drm_sched_hw_job_reset(&gpu->sched, sched_job); + drm_sched_stop(&gpu->sched); + + if(sched_job) + drm_sched_increase_karma(sched_job); /* get the GPU back into the init state */ etnaviv_core_dump(gpu); etnaviv_gpu_recover_hang(gpu); + drm_sched_resubmit_jobs(&gpu->sched); + /* restart scheduler after GPU is usable again */ - drm_sched_job_recovery(&gpu->sched); - kthread_unpark(gpu->sched.thread); + drm_sched_start(&gpu->sched, true); } static void etnaviv_sched_free_job(struct drm_sched_job *sched_job) diff --git a/drivers/gpu/drm/exynos/exynos_drm_rotator.c b/drivers/gpu/drm/exynos/exynos_drm_rotator.c index 8d67b2a54be3..05abfed6f7f8 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_rotator.c +++ b/drivers/gpu/drm/exynos/exynos_drm_rotator.c @@ -356,6 +356,11 @@ static int rotator_runtime_resume(struct device *dev) } #endif +static const struct drm_exynos_ipp_limit rotator_s5pv210_rbg888_limits[] = { + { IPP_SIZE_LIMIT(BUFFER, .h = { 8, SZ_16K }, .v = { 8, SZ_16K }) }, + { IPP_SIZE_LIMIT(AREA, .h.align = 2, .v.align = 2) }, +}; + static const struct drm_exynos_ipp_limit rotator_4210_rbg888_limits[] = { { IPP_SIZE_LIMIT(BUFFER, .h = { 8, SZ_16K }, .v = { 8, SZ_16K }) }, { IPP_SIZE_LIMIT(AREA, .h.align = 4, .v.align = 4) }, @@ -371,6 +376,11 @@ static const struct drm_exynos_ipp_limit rotator_5250_rbg888_limits[] = { { IPP_SIZE_LIMIT(AREA, .h.align = 2, .v.align = 2) }, }; +static const struct drm_exynos_ipp_limit rotator_s5pv210_yuv_limits[] = { + { IPP_SIZE_LIMIT(BUFFER, .h = { 32, SZ_64K }, .v = { 32, SZ_64K }) }, + { IPP_SIZE_LIMIT(AREA, .h.align = 8, .v.align = 8) }, +}; + static const struct drm_exynos_ipp_limit rotator_4210_yuv_limits[] = { { IPP_SIZE_LIMIT(BUFFER, .h = { 32, SZ_64K }, .v = { 32, SZ_64K }) }, { IPP_SIZE_LIMIT(AREA, .h.align = 8, .v.align = 8) }, @@ -381,6 +391,11 @@ static const struct drm_exynos_ipp_limit rotator_4412_yuv_limits[] = { { IPP_SIZE_LIMIT(AREA, .h.align = 8, .v.align = 8) }, }; +static const struct exynos_drm_ipp_formats rotator_s5pv210_formats[] = { + { IPP_SRCDST_FORMAT(XRGB8888, rotator_s5pv210_rbg888_limits) }, + { IPP_SRCDST_FORMAT(NV12, rotator_s5pv210_yuv_limits) }, +}; + static const struct exynos_drm_ipp_formats rotator_4210_formats[] = { { IPP_SRCDST_FORMAT(XRGB8888, rotator_4210_rbg888_limits) }, { IPP_SRCDST_FORMAT(NV12, rotator_4210_yuv_limits) }, @@ -396,6 +411,11 @@ static const struct exynos_drm_ipp_formats rotator_5250_formats[] = { { IPP_SRCDST_FORMAT(NV12, rotator_4412_yuv_limits) }, }; +static const struct rot_variant rotator_s5pv210_data = { + .formats = rotator_s5pv210_formats, + .num_formats = ARRAY_SIZE(rotator_s5pv210_formats), +}; + static const struct rot_variant rotator_4210_data = { .formats = rotator_4210_formats, .num_formats = ARRAY_SIZE(rotator_4210_formats), @@ -413,6 +433,9 @@ static const struct rot_variant rotator_5250_data = { static const struct of_device_id exynos_rotator_match[] = { { + .compatible = "samsung,s5pv210-rotator", + .data = &rotator_s5pv210_data, + }, { .compatible = "samsung,exynos4210-rotator", .data = &rotator_4210_data, }, { diff --git a/drivers/gpu/drm/exynos/exynos_drm_scaler.c b/drivers/gpu/drm/exynos/exynos_drm_scaler.c index 71270efa64f3..ed1dd1aec902 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_scaler.c +++ b/drivers/gpu/drm/exynos/exynos_drm_scaler.c @@ -1,7 +1,7 @@ /* * Copyright (C) 2017 Samsung Electronics Co.Ltd * Author: - * Andrzej Pietrasiewicz <andrzej.p@samsung.com> + * Andrzej Pietrasiewicz <andrzejtp2010@gmail.com> * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License version 2 as diff --git a/drivers/gpu/drm/exynos/regs-scaler.h b/drivers/gpu/drm/exynos/regs-scaler.h index fc7ccad75e74..512a2baced11 100644 --- a/drivers/gpu/drm/exynos/regs-scaler.h +++ b/drivers/gpu/drm/exynos/regs-scaler.h @@ -2,7 +2,7 @@ * * Copyright (c) 2017 Samsung Electronics Co., Ltd. * http://www.samsung.com/ - * Author: Andrzej Pietrasiewicz <andrzej.p@samsung.com> + * Author: Andrzej Pietrasiewicz <andrzejtp2010@gmail.com> * * Register definition file for Samsung scaler driver * diff --git a/drivers/gpu/drm/i915/Kconfig.debug b/drivers/gpu/drm/i915/Kconfig.debug index 9e36ffb5eb7c..ad4d71161dda 100644 --- a/drivers/gpu/drm/i915/Kconfig.debug +++ b/drivers/gpu/drm/i915/Kconfig.debug @@ -21,11 +21,11 @@ config DRM_I915_DEBUG select DEBUG_FS select PREEMPT_COUNT select I2C_CHARDEV + select STACKDEPOT select DRM_DP_AUX_CHARDEV select X86_MSR # used by igt/pm_rpm select DRM_VGEM # used by igt/prime_vgem (dmabuf interop checks) select DRM_DEBUG_MM if DRM=y - select STACKDEPOT if DRM=y # for DRM_DEBUG_MM select DRM_DEBUG_SELFTEST select SW_SYNC # signaling validation framework (igt/syncobj*) select DRM_I915_SW_FENCE_DEBUG_OBJECTS @@ -173,6 +173,7 @@ config DRM_I915_DEBUG_RUNTIME_PM bool "Enable extra state checking for runtime PM" depends on DRM_I915 default n + select STACKDEPOT help Choose this option to turn on extra state checking for the runtime PM functionality. This may introduce overhead during diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile index c34bee16730d..1787e1299b1b 100644 --- a/drivers/gpu/drm/i915/Makefile +++ b/drivers/gpu/drm/i915/Makefile @@ -22,6 +22,7 @@ subdir-ccflags-y += $(call cc-disable-warning, unused-but-set-variable) subdir-ccflags-y += $(call cc-disable-warning, sign-compare) subdir-ccflags-y += $(call cc-disable-warning, sometimes-uninitialized) subdir-ccflags-y += $(call cc-disable-warning, initializer-overrides) +subdir-ccflags-y += $(call cc-disable-warning, uninitialized) subdir-ccflags-$(CONFIG_DRM_I915_WERROR) += -Werror # Fine grained warnings disable @@ -40,9 +41,10 @@ i915-y := i915_drv.o \ i915_mm.o \ i915_params.o \ i915_pci.o \ + i915_reset.o \ i915_suspend.o \ - i915_syncmap.o \ i915_sw_fence.o \ + i915_syncmap.o \ i915_sysfs.o \ intel_csr.o \ intel_device_info.o \ @@ -55,7 +57,9 @@ i915-$(CONFIG_DEBUG_FS) += i915_debugfs.o intel_pipe_crc.o i915-$(CONFIG_PERF_EVENTS) += i915_pmu.o # GEM code -i915-y += i915_cmd_parser.o \ +i915-y += \ + i915_active.o \ + i915_cmd_parser.o \ i915_gem_batch_pool.o \ i915_gem_clflush.o \ i915_gem_context.o \ @@ -166,6 +170,7 @@ i915-$(CONFIG_DRM_I915_SELFTEST) += \ selftests/i915_random.o \ selftests/i915_selftest.o \ selftests/igt_flush_test.o \ + selftests/igt_live_test.o \ selftests/igt_reset.o \ selftests/igt_spinner.o @@ -198,3 +203,4 @@ endif i915-y += intel_lpe_audio.o obj-$(CONFIG_DRM_I915) += i915.o +obj-$(CONFIG_DRM_I915_GVT_KVMGT) += gvt/kvmgt.o diff --git a/drivers/gpu/drm/i915/gvt/Makefile b/drivers/gpu/drm/i915/gvt/Makefile index b016dc753db9..271fb46d4dd0 100644 --- a/drivers/gpu/drm/i915/gvt/Makefile +++ b/drivers/gpu/drm/i915/gvt/Makefile @@ -7,4 +7,3 @@ GVT_SOURCE := gvt.o aperture_gm.o handlers.o vgpu.o trace_points.o firmware.o \ ccflags-y += -I$(src) -I$(src)/$(GVT_DIR) i915-y += $(addprefix $(GVT_DIR)/, $(GVT_SOURCE)) -obj-$(CONFIG_DRM_I915_GVT_KVMGT) += $(GVT_DIR)/kvmgt.o diff --git a/drivers/gpu/drm/i915/gvt/aperture_gm.c b/drivers/gpu/drm/i915/gvt/aperture_gm.c index 359d37d5c958..1fa2f65c3cd1 100644 --- a/drivers/gpu/drm/i915/gvt/aperture_gm.c +++ b/drivers/gpu/drm/i915/gvt/aperture_gm.c @@ -180,7 +180,7 @@ static void free_vgpu_fence(struct intel_vgpu *vgpu) } mutex_unlock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put_unchecked(dev_priv); } static int alloc_vgpu_fence(struct intel_vgpu *vgpu) @@ -206,7 +206,7 @@ static int alloc_vgpu_fence(struct intel_vgpu *vgpu) _clear_vgpu_fence(vgpu); mutex_unlock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put_unchecked(dev_priv); return 0; out_free_fence: gvt_vgpu_err("Failed to alloc fences\n"); @@ -219,7 +219,7 @@ out_free_fence: vgpu->fence.regs[i] = NULL; } mutex_unlock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put_unchecked(dev_priv); return -ENOSPC; } @@ -317,7 +317,7 @@ void intel_vgpu_reset_resource(struct intel_vgpu *vgpu) intel_runtime_pm_get(dev_priv); _clear_vgpu_fence(vgpu); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put_unchecked(dev_priv); } /** diff --git a/drivers/gpu/drm/i915/gvt/cmd_parser.c b/drivers/gpu/drm/i915/gvt/cmd_parser.c index 77ae634eb11c..35b4ec3f7618 100644 --- a/drivers/gpu/drm/i915/gvt/cmd_parser.c +++ b/drivers/gpu/drm/i915/gvt/cmd_parser.c @@ -55,10 +55,10 @@ struct sub_op_bits { int low; }; struct decode_info { - char *name; + const char *name; int op_len; int nr_sub_op; - struct sub_op_bits *sub_op; + const struct sub_op_bits *sub_op; }; #define MAX_CMD_BUDGET 0x7fffffff @@ -375,7 +375,7 @@ typedef int (*parser_cmd_handler)(struct parser_exec_state *s); #define ADDR_FIX_5(x1, x2, x3, x4, x5) (ADDR_FIX_1(x1) | ADDR_FIX_4(x2, x3, x4, x5)) struct cmd_info { - char *name; + const char *name; u32 opcode; #define F_LEN_MASK (1U<<0) @@ -399,10 +399,10 @@ struct cmd_info { #define R_VECS (1 << VECS) #define R_ALL (R_RCS | R_VCS | R_BCS | R_VECS) /* rings that support this cmd: BLT/RCS/VCS/VECS */ - uint16_t rings; + u16 rings; /* devices that support this cmd: SNB/IVB/HSW/... */ - uint16_t devices; + u16 devices; /* which DWords are address that need fix up. * bit 0 means a 32-bit non address operand in command @@ -412,20 +412,20 @@ struct cmd_info { * No matter the address length, each address only takes * one bit in the bitmap. */ - uint16_t addr_bitmap; + u16 addr_bitmap; /* flag == F_LEN_CONST : command length * flag == F_LEN_VAR : length bias bits * Note: length is in DWord */ - uint8_t len; + u8 len; parser_cmd_handler handler; }; struct cmd_entry { struct hlist_node hlist; - struct cmd_info *info; + const struct cmd_info *info; }; enum { @@ -474,7 +474,7 @@ struct parser_exec_state { int saved_buf_addr_type; bool is_ctx_wa; - struct cmd_info *info; + const struct cmd_info *info; struct intel_vgpu_workload *workload; }; @@ -485,12 +485,12 @@ struct parser_exec_state { static unsigned long bypass_scan_mask = 0; /* ring ALL, type = 0 */ -static struct sub_op_bits sub_op_mi[] = { +static const struct sub_op_bits sub_op_mi[] = { {31, 29}, {28, 23}, }; -static struct decode_info decode_info_mi = { +static const struct decode_info decode_info_mi = { "MI", OP_LEN_MI, ARRAY_SIZE(sub_op_mi), @@ -498,12 +498,12 @@ static struct decode_info decode_info_mi = { }; /* ring RCS, command type 2 */ -static struct sub_op_bits sub_op_2d[] = { +static const struct sub_op_bits sub_op_2d[] = { {31, 29}, {28, 22}, }; -static struct decode_info decode_info_2d = { +static const struct decode_info decode_info_2d = { "2D", OP_LEN_2D, ARRAY_SIZE(sub_op_2d), @@ -511,14 +511,14 @@ static struct decode_info decode_info_2d = { }; /* ring RCS, command type 3 */ -static struct sub_op_bits sub_op_3d_media[] = { +static const struct sub_op_bits sub_op_3d_media[] = { {31, 29}, {28, 27}, {26, 24}, {23, 16}, }; -static struct decode_info decode_info_3d_media = { +static const struct decode_info decode_info_3d_media = { "3D_Media", OP_LEN_3D_MEDIA, ARRAY_SIZE(sub_op_3d_media), @@ -526,7 +526,7 @@ static struct decode_info decode_info_3d_media = { }; /* ring VCS, command type 3 */ -static struct sub_op_bits sub_op_mfx_vc[] = { +static const struct sub_op_bits sub_op_mfx_vc[] = { {31, 29}, {28, 27}, {26, 24}, @@ -534,7 +534,7 @@ static struct sub_op_bits sub_op_mfx_vc[] = { {20, 16}, }; -static struct decode_info decode_info_mfx_vc = { +static const struct decode_info decode_info_mfx_vc = { "MFX_VC", OP_LEN_MFX_VC, ARRAY_SIZE(sub_op_mfx_vc), @@ -542,7 +542,7 @@ static struct decode_info decode_info_mfx_vc = { }; /* ring VECS, command type 3 */ -static struct sub_op_bits sub_op_vebox[] = { +static const struct sub_op_bits sub_op_vebox[] = { {31, 29}, {28, 27}, {26, 24}, @@ -550,14 +550,14 @@ static struct sub_op_bits sub_op_vebox[] = { {20, 16}, }; -static struct decode_info decode_info_vebox = { +static const struct decode_info decode_info_vebox = { "VEBOX", OP_LEN_VEBOX, ARRAY_SIZE(sub_op_vebox), sub_op_vebox, }; -static struct decode_info *ring_decode_info[I915_NUM_ENGINES][8] = { +static const struct decode_info *ring_decode_info[I915_NUM_ENGINES][8] = { [RCS] = { &decode_info_mi, NULL, @@ -616,7 +616,7 @@ static struct decode_info *ring_decode_info[I915_NUM_ENGINES][8] = { static inline u32 get_opcode(u32 cmd, int ring_id) { - struct decode_info *d_info; + const struct decode_info *d_info; d_info = ring_decode_info[ring_id][CMD_TYPE(cmd)]; if (d_info == NULL) @@ -625,7 +625,7 @@ static inline u32 get_opcode(u32 cmd, int ring_id) return cmd >> (32 - d_info->op_len); } -static inline struct cmd_info *find_cmd_entry(struct intel_gvt *gvt, +static inline const struct cmd_info *find_cmd_entry(struct intel_gvt *gvt, unsigned int opcode, int ring_id) { struct cmd_entry *e; @@ -638,7 +638,7 @@ static inline struct cmd_info *find_cmd_entry(struct intel_gvt *gvt, return NULL; } -static inline struct cmd_info *get_cmd_info(struct intel_gvt *gvt, +static inline const struct cmd_info *get_cmd_info(struct intel_gvt *gvt, u32 cmd, int ring_id) { u32 opcode; @@ -657,7 +657,7 @@ static inline u32 sub_op_val(u32 cmd, u32 hi, u32 low) static inline void print_opcode(u32 cmd, int ring_id) { - struct decode_info *d_info; + const struct decode_info *d_info; int i; d_info = ring_decode_info[ring_id][CMD_TYPE(cmd)]; @@ -776,7 +776,7 @@ static inline int ip_gma_advance(struct parser_exec_state *s, return 0; } -static inline int get_cmd_length(struct cmd_info *info, u32 cmd) +static inline int get_cmd_length(const struct cmd_info *info, u32 cmd) { if ((info->flag & F_LEN_MASK) == F_LEN_CONST) return info->len; @@ -901,7 +901,8 @@ static int cmd_reg_handler(struct parser_exec_state *s, * It's good enough to support initializing mmio by lri command in * vgpu inhibit context on KBL. */ - if (IS_KABYLAKE(s->vgpu->gvt->dev_priv) && + if ((IS_KABYLAKE(s->vgpu->gvt->dev_priv) + || IS_COFFEELAKE(s->vgpu->gvt->dev_priv)) && intel_gvt_mmio_is_in_ctx(gvt, offset) && !strncmp(cmd, "lri", 3)) { intel_gvt_hypervisor_read_gpa(s->vgpu, @@ -1280,9 +1281,7 @@ static int gen8_check_mi_display_flip(struct parser_exec_state *s, if (!info->async_flip) return 0; - if (IS_SKYLAKE(dev_priv) - || IS_KABYLAKE(dev_priv) - || IS_BROXTON(dev_priv)) { + if (INTEL_GEN(dev_priv) >= 9) { stride = vgpu_vreg_t(s->vgpu, info->stride_reg) & GENMASK(9, 0); tile = (vgpu_vreg_t(s->vgpu, info->ctrl_reg) & GENMASK(12, 10)) >> 10; @@ -1310,9 +1309,7 @@ static int gen8_update_plane_mmio_from_mi_display_flip( set_mask_bits(&vgpu_vreg_t(vgpu, info->surf_reg), GENMASK(31, 12), info->surf_val << 12); - if (IS_SKYLAKE(dev_priv) - || IS_KABYLAKE(dev_priv) - || IS_BROXTON(dev_priv)) { + if (INTEL_GEN(dev_priv) >= 9) { set_mask_bits(&vgpu_vreg_t(vgpu, info->stride_reg), GENMASK(9, 0), info->stride_val); set_mask_bits(&vgpu_vreg_t(vgpu, info->ctrl_reg), GENMASK(12, 10), @@ -1336,9 +1333,7 @@ static int decode_mi_display_flip(struct parser_exec_state *s, if (IS_BROADWELL(dev_priv)) return gen8_decode_mi_display_flip(s, info); - if (IS_SKYLAKE(dev_priv) - || IS_KABYLAKE(dev_priv) - || IS_BROXTON(dev_priv)) + if (INTEL_GEN(dev_priv) >= 9) return skl_decode_mi_display_flip(s, info); return -ENODEV; @@ -1643,8 +1638,8 @@ static int batch_buffer_needs_scan(struct parser_exec_state *s) static int find_bb_size(struct parser_exec_state *s, unsigned long *bb_size) { unsigned long gma = 0; - struct cmd_info *info; - uint32_t cmd_len = 0; + const struct cmd_info *info; + u32 cmd_len = 0; bool bb_end = false; struct intel_vgpu *vgpu = s->vgpu; u32 cmd; @@ -1842,7 +1837,7 @@ static int cmd_handler_mi_batch_buffer_start(struct parser_exec_state *s) static int mi_noop_index; -static struct cmd_info cmd_info[] = { +static const struct cmd_info cmd_info[] = { {"MI_NOOP", OP_MI_NOOP, F_LEN_CONST, R_ALL, D_ALL, 0, 1, NULL}, {"MI_SET_PREDICATE", OP_MI_SET_PREDICATE, F_LEN_CONST, R_ALL, D_ALL, @@ -2521,7 +2516,7 @@ static void add_cmd_entry(struct intel_gvt *gvt, struct cmd_entry *e) static int cmd_parser_exec(struct parser_exec_state *s) { struct intel_vgpu *vgpu = s->vgpu; - struct cmd_info *info; + const struct cmd_info *info; u32 cmd; int ret = 0; @@ -2683,7 +2678,7 @@ static int scan_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx) I915_GTT_PAGE_SIZE))) return -EINVAL; - ring_tail = wa_ctx->indirect_ctx.size + 3 * sizeof(uint32_t); + ring_tail = wa_ctx->indirect_ctx.size + 3 * sizeof(u32); ring_size = round_up(wa_ctx->indirect_ctx.size + CACHELINE_BYTES, PAGE_SIZE); gma_head = wa_ctx->indirect_ctx.guest_gma; @@ -2850,7 +2845,7 @@ put_obj: static int combine_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx) { - uint32_t per_ctx_start[CACHELINE_DWORDS] = {0}; + u32 per_ctx_start[CACHELINE_DWORDS] = {0}; unsigned char *bb_start_sva; if (!wa_ctx->per_ctx.valid) @@ -2895,10 +2890,10 @@ int intel_gvt_scan_and_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx) return 0; } -static struct cmd_info *find_cmd_entry_any_ring(struct intel_gvt *gvt, +static const struct cmd_info *find_cmd_entry_any_ring(struct intel_gvt *gvt, unsigned int opcode, unsigned long rings) { - struct cmd_info *info = NULL; + const struct cmd_info *info = NULL; unsigned int ring; for_each_set_bit(ring, &rings, I915_NUM_ENGINES) { @@ -2913,7 +2908,7 @@ static int init_cmd_table(struct intel_gvt *gvt) { int i; struct cmd_entry *e; - struct cmd_info *info; + const struct cmd_info *info; unsigned int gen_type; gen_type = intel_gvt_get_device_type(gvt); diff --git a/drivers/gpu/drm/i915/gvt/display.c b/drivers/gpu/drm/i915/gvt/display.c index df1e14145747..035479e273be 100644 --- a/drivers/gpu/drm/i915/gvt/display.c +++ b/drivers/gpu/drm/i915/gvt/display.c @@ -198,7 +198,8 @@ static void emulate_monitor_status_change(struct intel_vgpu *vgpu) SDE_PORTC_HOTPLUG_CPT | SDE_PORTD_HOTPLUG_CPT); - if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) { + if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv) || + IS_COFFEELAKE(dev_priv)) { vgpu_vreg_t(vgpu, SDEISR) &= ~(SDE_PORTA_HOTPLUG_SPT | SDE_PORTE_HOTPLUG_SPT); vgpu_vreg_t(vgpu, SKL_FUSE_STATUS) |= @@ -273,7 +274,8 @@ static void emulate_monitor_status_change(struct intel_vgpu *vgpu) vgpu_vreg_t(vgpu, SFUSE_STRAP) |= SFUSE_STRAP_DDID_DETECTED; } - if ((IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) && + if ((IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv) || + IS_COFFEELAKE(dev_priv)) && intel_vgpu_has_monitor_on_port(vgpu, PORT_E)) { vgpu_vreg_t(vgpu, SDEISR) |= SDE_PORTE_HOTPLUG_SPT; } @@ -340,6 +342,7 @@ static int setup_virtual_dp_monitor(struct intel_vgpu *vgpu, int port_num, port->dpcd->data_valid = true; port->dpcd->data[DPCD_SINK_COUNT] = 0x1; port->type = type; + port->id = resolution; emulate_monitor_status_change(vgpu); @@ -443,6 +446,36 @@ void intel_gvt_emulate_vblank(struct intel_gvt *gvt) } /** + * intel_vgpu_emulate_hotplug - trigger hotplug event for vGPU + * @vgpu: a vGPU + * @conncted: link state + * + * This function is used to trigger hotplug interrupt for vGPU + * + */ +void intel_vgpu_emulate_hotplug(struct intel_vgpu *vgpu, bool connected) +{ + struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv; + + /* TODO: add more platforms support */ + if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) { + if (connected) { + vgpu_vreg_t(vgpu, SFUSE_STRAP) |= + SFUSE_STRAP_DDID_DETECTED; + vgpu_vreg_t(vgpu, SDEISR) |= SDE_PORTD_HOTPLUG_CPT; + } else { + vgpu_vreg_t(vgpu, SFUSE_STRAP) &= + ~SFUSE_STRAP_DDID_DETECTED; + vgpu_vreg_t(vgpu, SDEISR) &= ~SDE_PORTD_HOTPLUG_CPT; + } + vgpu_vreg_t(vgpu, SDEIIR) |= SDE_PORTD_HOTPLUG_CPT; + vgpu_vreg_t(vgpu, PCH_PORT_HOTPLUG) |= + PORTD_HOTPLUG_STATUS_MASK; + intel_vgpu_trigger_virtual_event(vgpu, DP_D_HOTPLUG); + } +} + +/** * intel_vgpu_clean_display - clean vGPU virtual display emulation * @vgpu: a vGPU * @@ -453,7 +486,8 @@ void intel_vgpu_clean_display(struct intel_vgpu *vgpu) { struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv; - if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) + if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv) || + IS_COFFEELAKE(dev_priv)) clean_virtual_dp_monitor(vgpu, PORT_D); else clean_virtual_dp_monitor(vgpu, PORT_B); @@ -476,7 +510,8 @@ int intel_vgpu_init_display(struct intel_vgpu *vgpu, u64 resolution) intel_vgpu_init_i2c_edid(vgpu); - if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) + if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv) || + IS_COFFEELAKE(dev_priv)) return setup_virtual_dp_monitor(vgpu, PORT_D, GVT_DP_D, resolution); else diff --git a/drivers/gpu/drm/i915/gvt/display.h b/drivers/gpu/drm/i915/gvt/display.h index ea7c1c525b8c..a87f33e6a23c 100644 --- a/drivers/gpu/drm/i915/gvt/display.h +++ b/drivers/gpu/drm/i915/gvt/display.h @@ -146,18 +146,19 @@ enum intel_vgpu_port_type { GVT_PORT_MAX }; +enum intel_vgpu_edid { + GVT_EDID_1024_768, + GVT_EDID_1920_1200, + GVT_EDID_NUM, +}; + struct intel_vgpu_port { /* per display EDID information */ struct intel_vgpu_edid_data *edid; /* per display DPCD information */ struct intel_vgpu_dpcd_data *dpcd; int type; -}; - -enum intel_vgpu_edid { - GVT_EDID_1024_768, - GVT_EDID_1920_1200, - GVT_EDID_NUM, + enum intel_vgpu_edid id; }; static inline char *vgpu_edid_str(enum intel_vgpu_edid id) @@ -172,6 +173,30 @@ static inline char *vgpu_edid_str(enum intel_vgpu_edid id) } } +static inline unsigned int vgpu_edid_xres(enum intel_vgpu_edid id) +{ + switch (id) { + case GVT_EDID_1024_768: + return 1024; + case GVT_EDID_1920_1200: + return 1920; + default: + return 0; + } +} + +static inline unsigned int vgpu_edid_yres(enum intel_vgpu_edid id) +{ + switch (id) { + case GVT_EDID_1024_768: + return 768; + case GVT_EDID_1920_1200: + return 1200; + default: + return 0; + } +} + void intel_gvt_emulate_vblank(struct intel_gvt *gvt); void intel_gvt_check_vblank_emulation(struct intel_gvt *gvt); diff --git a/drivers/gpu/drm/i915/gvt/dmabuf.c b/drivers/gpu/drm/i915/gvt/dmabuf.c index 51ed99a37803..3e7e2b80c857 100644 --- a/drivers/gpu/drm/i915/gvt/dmabuf.c +++ b/drivers/gpu/drm/i915/gvt/dmabuf.c @@ -29,7 +29,6 @@ */ #include <linux/dma-buf.h> -#include <drm/drmP.h> #include <linux/vfio.h> #include "i915_drv.h" @@ -164,9 +163,7 @@ static struct drm_i915_gem_object *vgpu_create_gem(struct drm_device *dev, obj->read_domains = I915_GEM_DOMAIN_GTT; obj->write_domain = 0; - if (IS_SKYLAKE(dev_priv) - || IS_KABYLAKE(dev_priv) - || IS_BROXTON(dev_priv)) { + if (INTEL_GEN(dev_priv) >= 9) { unsigned int tiling_mode = 0; unsigned int stride = 0; diff --git a/drivers/gpu/drm/i915/gvt/edid.c b/drivers/gpu/drm/i915/gvt/edid.c index 5d4bb35bb889..1fe6124918f1 100644 --- a/drivers/gpu/drm/i915/gvt/edid.c +++ b/drivers/gpu/drm/i915/gvt/edid.c @@ -77,16 +77,32 @@ static unsigned char edid_get_byte(struct intel_vgpu *vgpu) return chr; } +static inline int cnp_get_port_from_gmbus0(u32 gmbus0) +{ + int port_select = gmbus0 & _GMBUS_PIN_SEL_MASK; + int port = -EINVAL; + + if (port_select == GMBUS_PIN_1_BXT) + port = PORT_B; + else if (port_select == GMBUS_PIN_2_BXT) + port = PORT_C; + else if (port_select == GMBUS_PIN_3_BXT) + port = PORT_D; + else if (port_select == GMBUS_PIN_4_CNP) + port = PORT_E; + return port; +} + static inline int bxt_get_port_from_gmbus0(u32 gmbus0) { int port_select = gmbus0 & _GMBUS_PIN_SEL_MASK; int port = -EINVAL; - if (port_select == 1) + if (port_select == GMBUS_PIN_1_BXT) port = PORT_B; - else if (port_select == 2) + else if (port_select == GMBUS_PIN_2_BXT) port = PORT_C; - else if (port_select == 3) + else if (port_select == GMBUS_PIN_3_BXT) port = PORT_D; return port; } @@ -96,13 +112,13 @@ static inline int get_port_from_gmbus0(u32 gmbus0) int port_select = gmbus0 & _GMBUS_PIN_SEL_MASK; int port = -EINVAL; - if (port_select == 2) + if (port_select == GMBUS_PIN_VGADDC) port = PORT_E; - else if (port_select == 4) + else if (port_select == GMBUS_PIN_DPC) port = PORT_C; - else if (port_select == 5) + else if (port_select == GMBUS_PIN_DPB) port = PORT_B; - else if (port_select == 6) + else if (port_select == GMBUS_PIN_DPD) port = PORT_D; return port; } @@ -133,6 +149,8 @@ static int gmbus0_mmio_write(struct intel_vgpu *vgpu, if (IS_BROXTON(dev_priv)) port = bxt_get_port_from_gmbus0(pin_select); + else if (IS_COFFEELAKE(dev_priv)) + port = cnp_get_port_from_gmbus0(pin_select); else port = get_port_from_gmbus0(pin_select); if (WARN_ON(port < 0)) diff --git a/drivers/gpu/drm/i915/gvt/fb_decoder.c b/drivers/gpu/drm/i915/gvt/fb_decoder.c index 85e6736f0a32..65e847392aea 100644 --- a/drivers/gpu/drm/i915/gvt/fb_decoder.c +++ b/drivers/gpu/drm/i915/gvt/fb_decoder.c @@ -151,9 +151,7 @@ static u32 intel_vgpu_get_stride(struct intel_vgpu *vgpu, int pipe, u32 stride_reg = vgpu_vreg_t(vgpu, DSPSTRIDE(pipe)) & stride_mask; u32 stride = stride_reg; - if (IS_SKYLAKE(dev_priv) - || IS_KABYLAKE(dev_priv) - || IS_BROXTON(dev_priv)) { + if (INTEL_GEN(dev_priv) >= 9) { switch (tiled) { case PLANE_CTL_TILED_LINEAR: stride = stride_reg * 64; @@ -217,9 +215,7 @@ int intel_vgpu_decode_primary_plane(struct intel_vgpu *vgpu, if (!plane->enabled) return -ENODEV; - if (IS_SKYLAKE(dev_priv) - || IS_KABYLAKE(dev_priv) - || IS_BROXTON(dev_priv)) { + if (INTEL_GEN(dev_priv) >= 9) { plane->tiled = val & PLANE_CTL_TILED_MASK; fmt = skl_format_to_drm( val & PLANE_CTL_FORMAT_MASK, @@ -260,9 +256,7 @@ int intel_vgpu_decode_primary_plane(struct intel_vgpu *vgpu, } plane->stride = intel_vgpu_get_stride(vgpu, pipe, plane->tiled, - (IS_SKYLAKE(dev_priv) - || IS_KABYLAKE(dev_priv) - || IS_BROXTON(dev_priv)) ? + (INTEL_GEN(dev_priv) >= 9) ? (_PRI_PLANE_STRIDE_MASK >> 6) : _PRI_PLANE_STRIDE_MASK, plane->bpp); diff --git a/drivers/gpu/drm/i915/gvt/gvt.c b/drivers/gpu/drm/i915/gvt/gvt.c index 733a2a0d0c30..43f4242062dd 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.c +++ b/drivers/gpu/drm/i915/gvt/gvt.c @@ -185,54 +185,9 @@ static const struct intel_gvt_ops intel_gvt_ops = { .vgpu_query_plane = intel_vgpu_query_plane, .vgpu_get_dmabuf = intel_vgpu_get_dmabuf, .write_protect_handler = intel_vgpu_page_track_handler, + .emulate_hotplug = intel_vgpu_emulate_hotplug, }; -/** - * intel_gvt_init_host - Load MPT modules and detect if we're running in host - * - * This function is called at the driver loading stage. If failed to find a - * loadable MPT module or detect currently we're running in a VM, then GVT-g - * will be disabled - * - * Returns: - * Zero on success, negative error code if failed. - * - */ -int intel_gvt_init_host(void) -{ - if (intel_gvt_host.initialized) - return 0; - - /* Xen DOM U */ - if (xen_domain() && !xen_initial_domain()) - return -ENODEV; - - /* Try to load MPT modules for hypervisors */ - if (xen_initial_domain()) { - /* In Xen dom0 */ - intel_gvt_host.mpt = try_then_request_module( - symbol_get(xengt_mpt), "xengt"); - intel_gvt_host.hypervisor_type = INTEL_GVT_HYPERVISOR_XEN; - } else { -#if IS_ENABLED(CONFIG_DRM_I915_GVT_KVMGT) - /* not in Xen. Try KVMGT */ - intel_gvt_host.mpt = try_then_request_module( - symbol_get(kvmgt_mpt), "kvmgt"); - intel_gvt_host.hypervisor_type = INTEL_GVT_HYPERVISOR_KVM; -#endif - } - - /* Fail to load MPT modules - bail out */ - if (!intel_gvt_host.mpt) - return -EINVAL; - - gvt_dbg_core("Running with hypervisor %s in host mode\n", - supported_hypervisors[intel_gvt_host.hypervisor_type]); - - intel_gvt_host.initialized = true; - return 0; -} - static void init_device_info(struct intel_gvt *gvt) { struct intel_gvt_device_info *info = &gvt->device_info; @@ -316,7 +271,6 @@ void intel_gvt_clean_device(struct drm_i915_private *dev_priv) return; intel_gvt_destroy_idle_vgpu(gvt->idle_vgpu); - intel_gvt_hypervisor_host_exit(&dev_priv->drm.pdev->dev, gvt); intel_gvt_cleanup_vgpu_type_groups(gvt); intel_gvt_clean_vgpu_types(gvt); @@ -352,13 +306,6 @@ int intel_gvt_init_device(struct drm_i915_private *dev_priv) struct intel_vgpu *vgpu; int ret; - /* - * Cannot initialize GVT device without intel_gvt_host gets - * initialized first. - */ - if (WARN_ON(!intel_gvt_host.initialized)) - return -EINVAL; - if (WARN_ON(dev_priv->gvt)) return -EEXIST; @@ -420,13 +367,6 @@ int intel_gvt_init_device(struct drm_i915_private *dev_priv) goto out_clean_types; } - ret = intel_gvt_hypervisor_host_init(&dev_priv->drm.pdev->dev, gvt, - &intel_gvt_ops); - if (ret) { - gvt_err("failed to register gvt-g host device: %d\n", ret); - goto out_clean_types; - } - vgpu = intel_gvt_create_idle_vgpu(gvt); if (IS_ERR(vgpu)) { ret = PTR_ERR(vgpu); @@ -441,6 +381,8 @@ int intel_gvt_init_device(struct drm_i915_private *dev_priv) gvt_dbg_core("gvt device initialization is done\n"); dev_priv->gvt = gvt; + intel_gvt_host.dev = &dev_priv->drm.pdev->dev; + intel_gvt_host.initialized = true; return 0; out_clean_types: @@ -467,6 +409,45 @@ out_clean_idr: return ret; } -#if IS_ENABLED(CONFIG_DRM_I915_GVT_KVMGT) -MODULE_SOFTDEP("pre: kvmgt"); -#endif +int +intel_gvt_register_hypervisor(struct intel_gvt_mpt *m) +{ + int ret; + void *gvt; + + if (!intel_gvt_host.initialized) + return -ENODEV; + + if (m->type != INTEL_GVT_HYPERVISOR_KVM && + m->type != INTEL_GVT_HYPERVISOR_XEN) + return -EINVAL; + + /* Get a reference for device model module */ + if (!try_module_get(THIS_MODULE)) + return -ENODEV; + + intel_gvt_host.mpt = m; + intel_gvt_host.hypervisor_type = m->type; + gvt = (void *)kdev_to_i915(intel_gvt_host.dev)->gvt; + + ret = intel_gvt_hypervisor_host_init(intel_gvt_host.dev, gvt, + &intel_gvt_ops); + if (ret < 0) { + gvt_err("Failed to init %s hypervisor module\n", + supported_hypervisors[intel_gvt_host.hypervisor_type]); + module_put(THIS_MODULE); + return -ENODEV; + } + gvt_dbg_core("Running with hypervisor %s in host mode\n", + supported_hypervisors[intel_gvt_host.hypervisor_type]); + return 0; +} +EXPORT_SYMBOL_GPL(intel_gvt_register_hypervisor); + +void +intel_gvt_unregister_hypervisor(void) +{ + intel_gvt_hypervisor_host_exit(intel_gvt_host.dev); + module_put(THIS_MODULE); +} +EXPORT_SYMBOL_GPL(intel_gvt_unregister_hypervisor); diff --git a/drivers/gpu/drm/i915/gvt/gvt.h b/drivers/gpu/drm/i915/gvt/gvt.h index b4ab1dad0143..8bce09de4b82 100644 --- a/drivers/gpu/drm/i915/gvt/gvt.h +++ b/drivers/gpu/drm/i915/gvt/gvt.h @@ -52,12 +52,8 @@ #define GVT_MAX_VGPU 8 -enum { - INTEL_GVT_HYPERVISOR_XEN = 0, - INTEL_GVT_HYPERVISOR_KVM, -}; - struct intel_gvt_host { + struct device *dev; bool initialized; int hypervisor_type; struct intel_gvt_mpt *mpt; @@ -540,6 +536,8 @@ int intel_vgpu_emulate_cfg_read(struct intel_vgpu *vgpu, unsigned int offset, int intel_vgpu_emulate_cfg_write(struct intel_vgpu *vgpu, unsigned int offset, void *p_data, unsigned int bytes); +void intel_vgpu_emulate_hotplug(struct intel_vgpu *vgpu, bool connected); + static inline u64 intel_vgpu_get_bar_gpa(struct intel_vgpu *vgpu, int bar) { /* We are 64bit bar. */ @@ -581,6 +579,7 @@ struct intel_gvt_ops { int (*vgpu_get_dmabuf)(struct intel_vgpu *vgpu, unsigned int); int (*write_protect_handler)(struct intel_vgpu *, u64, void *, unsigned int); + void (*emulate_hotplug)(struct intel_vgpu *vgpu, bool connected); }; @@ -597,7 +596,7 @@ static inline void mmio_hw_access_pre(struct drm_i915_private *dev_priv) static inline void mmio_hw_access_post(struct drm_i915_private *dev_priv) { - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put_unchecked(dev_priv); } /** diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c index b5475c91e2ef..9c106e47e640 100644 --- a/drivers/gpu/drm/i915/gvt/handlers.c +++ b/drivers/gpu/drm/i915/gvt/handlers.c @@ -57,6 +57,8 @@ unsigned long intel_gvt_get_device_type(struct intel_gvt *gvt) return D_KBL; else if (IS_BROXTON(gvt->dev_priv)) return D_BXT; + else if (IS_COFFEELAKE(gvt->dev_priv)) + return D_CFL; return 0; } @@ -276,14 +278,12 @@ static int mul_force_wake_write(struct intel_vgpu *vgpu, unsigned int offset, void *p_data, unsigned int bytes) { u32 old, new; - uint32_t ack_reg_offset; + u32 ack_reg_offset; old = vgpu_vreg(vgpu, offset); new = CALC_MODE_MASK_REG(old, *(u32 *)p_data); - if (IS_SKYLAKE(vgpu->gvt->dev_priv) - || IS_KABYLAKE(vgpu->gvt->dev_priv) - || IS_BROXTON(vgpu->gvt->dev_priv)) { + if (INTEL_GEN(vgpu->gvt->dev_priv) >= 9) { switch (offset) { case FORCEWAKE_RENDER_GEN9_REG: ack_reg_offset = FORCEWAKE_ACK_RENDER_GEN9_REG; @@ -833,7 +833,7 @@ static int dp_aux_ch_ctl_trans_done(struct intel_vgpu *vgpu, u32 value, } static void dp_aux_ch_ctl_link_training(struct intel_vgpu_dpcd_data *dpcd, - uint8_t t) + u8 t) { if ((t & DPCD_TRAINING_PATTERN_SET_MASK) == DPCD_TRAINING_PATTERN_1) { /* training pattern 1 for CR */ @@ -889,9 +889,7 @@ static int dp_aux_ch_ctl_mmio_write(struct intel_vgpu *vgpu, write_vreg(vgpu, offset, p_data, bytes); data = vgpu_vreg(vgpu, offset); - if ((IS_SKYLAKE(vgpu->gvt->dev_priv) - || IS_KABYLAKE(vgpu->gvt->dev_priv) - || IS_BROXTON(vgpu->gvt->dev_priv)) + if ((INTEL_GEN(vgpu->gvt->dev_priv) >= 9) && offset != _REG_SKL_DP_AUX_CH_CTL(port_index)) { /* SKL DPB/C/D aux ctl register changed */ return 0; @@ -919,7 +917,7 @@ static int dp_aux_ch_ctl_mmio_write(struct intel_vgpu *vgpu, if (op == GVT_AUX_NATIVE_WRITE) { int t; - uint8_t buf[16]; + u8 buf[16]; if ((addr + len + 1) >= DPCD_SIZE) { /* @@ -1407,7 +1405,8 @@ static int mailbox_write(struct intel_vgpu *vgpu, unsigned int offset, switch (cmd) { case GEN9_PCODE_READ_MEM_LATENCY: if (IS_SKYLAKE(vgpu->gvt->dev_priv) - || IS_KABYLAKE(vgpu->gvt->dev_priv)) { + || IS_KABYLAKE(vgpu->gvt->dev_priv) + || IS_COFFEELAKE(vgpu->gvt->dev_priv)) { /** * "Read memory latency" command on gen9. * Below memory latency values are read @@ -1431,7 +1430,8 @@ static int mailbox_write(struct intel_vgpu *vgpu, unsigned int offset, break; case SKL_PCODE_CDCLK_CONTROL: if (IS_SKYLAKE(vgpu->gvt->dev_priv) - || IS_KABYLAKE(vgpu->gvt->dev_priv)) + || IS_KABYLAKE(vgpu->gvt->dev_priv) + || IS_COFFEELAKE(vgpu->gvt->dev_priv)) *data0 = SKL_CDCLK_READY_FOR_CHANGE; break; case GEN6_PCODE_READ_RC6VIDS: @@ -3041,8 +3041,8 @@ static int init_skl_mmio_info(struct intel_gvt *gvt) MMIO_DFH(GEN9_WM_CHICKEN3, D_SKL_PLUS, F_MODE_MASK | F_CMD_ACCESS, NULL, NULL); - MMIO_D(_MMIO(0x4ab8), D_KBL); - MMIO_D(_MMIO(0x2248), D_KBL | D_SKL); + MMIO_D(_MMIO(0x4ab8), D_KBL | D_CFL); + MMIO_D(_MMIO(0x2248), D_SKL_PLUS); return 0; } @@ -3302,7 +3302,8 @@ int intel_gvt_setup_mmio_info(struct intel_gvt *gvt) if (ret) goto err; } else if (IS_SKYLAKE(dev_priv) - || IS_KABYLAKE(dev_priv)) { + || IS_KABYLAKE(dev_priv) + || IS_COFFEELAKE(dev_priv)) { ret = init_broadwell_mmio_info(gvt); if (ret) goto err; diff --git a/drivers/gpu/drm/i915/gvt/hypercall.h b/drivers/gpu/drm/i915/gvt/hypercall.h index 5af11cf1b482..5e01cc8d9b16 100644 --- a/drivers/gpu/drm/i915/gvt/hypercall.h +++ b/drivers/gpu/drm/i915/gvt/hypercall.h @@ -33,13 +33,19 @@ #ifndef _GVT_HYPERCALL_H_ #define _GVT_HYPERCALL_H_ +enum hypervisor_type { + INTEL_GVT_HYPERVISOR_XEN = 0, + INTEL_GVT_HYPERVISOR_KVM, +}; + /* * Specific GVT-g MPT modules function collections. Currently GVT-g supports * both Xen and KVM by providing dedicated hypervisor-related MPT modules. */ struct intel_gvt_mpt { + enum hypervisor_type type; int (*host_init)(struct device *dev, void *gvt, const void *ops); - void (*host_exit)(struct device *dev, void *gvt); + void (*host_exit)(struct device *dev); int (*attach_vgpu)(void *vgpu, unsigned long *handle); void (*detach_vgpu)(unsigned long handle); int (*inject_msi)(unsigned long handle, u32 addr, u16 data); @@ -61,12 +67,12 @@ struct intel_gvt_mpt { int (*set_trap_area)(unsigned long handle, u64 start, u64 end, bool map); int (*set_opregion)(void *vgpu); + int (*set_edid)(void *vgpu, int port_num); int (*get_vfio_device)(void *vgpu); void (*put_vfio_device)(void *vgpu); bool (*is_valid_gfn)(unsigned long handle, unsigned long gfn); }; extern struct intel_gvt_mpt xengt_mpt; -extern struct intel_gvt_mpt kvmgt_mpt; #endif /* _GVT_HYPERCALL_H_ */ diff --git a/drivers/gpu/drm/i915/gvt/interrupt.c b/drivers/gpu/drm/i915/gvt/interrupt.c index 6b9d1354ff29..67125c5eec6e 100644 --- a/drivers/gpu/drm/i915/gvt/interrupt.c +++ b/drivers/gpu/drm/i915/gvt/interrupt.c @@ -581,9 +581,7 @@ static void gen8_init_irq( SET_BIT_INFO(irq, 4, PRIMARY_C_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_C); SET_BIT_INFO(irq, 5, SPRITE_C_FLIP_DONE, INTEL_GVT_IRQ_INFO_DE_PIPE_C); - } else if (IS_SKYLAKE(gvt->dev_priv) - || IS_KABYLAKE(gvt->dev_priv) - || IS_BROXTON(gvt->dev_priv)) { + } else if (INTEL_GEN(gvt->dev_priv) >= 9) { SET_BIT_INFO(irq, 25, AUX_CHANNEL_B, INTEL_GVT_IRQ_INFO_DE_PORT); SET_BIT_INFO(irq, 26, AUX_CHANNEL_C, INTEL_GVT_IRQ_INFO_DE_PORT); SET_BIT_INFO(irq, 27, AUX_CHANNEL_D, INTEL_GVT_IRQ_INFO_DE_PORT); diff --git a/drivers/gpu/drm/i915/gvt/kvmgt.c b/drivers/gpu/drm/i915/gvt/kvmgt.c index c1072143da1d..63eef86a2a85 100644 --- a/drivers/gpu/drm/i915/gvt/kvmgt.c +++ b/drivers/gpu/drm/i915/gvt/kvmgt.c @@ -57,6 +57,8 @@ static const struct intel_gvt_ops *intel_gvt_ops; #define VFIO_PCI_INDEX_TO_OFFSET(index) ((u64)(index) << VFIO_PCI_OFFSET_SHIFT) #define VFIO_PCI_OFFSET_MASK (((u64)(1) << VFIO_PCI_OFFSET_SHIFT) - 1) +#define EDID_BLOB_OFFSET (PAGE_SIZE/2) + #define OPREGION_SIGNATURE "IntelGraphicsMem" struct vfio_region; @@ -76,6 +78,11 @@ struct vfio_region { void *data; }; +struct vfio_edid_region { + struct vfio_region_gfx_edid vfio_edid_regs; + void *edid_blob; +}; + struct kvmgt_pgfn { gfn_t gfn; struct hlist_node hnode; @@ -427,6 +434,111 @@ static const struct intel_vgpu_regops intel_vgpu_regops_opregion = { .release = intel_vgpu_reg_release_opregion, }; +static int handle_edid_regs(struct intel_vgpu *vgpu, + struct vfio_edid_region *region, char *buf, + size_t count, u16 offset, bool is_write) +{ + struct vfio_region_gfx_edid *regs = ®ion->vfio_edid_regs; + unsigned int data; + + if (offset + count > sizeof(*regs)) + return -EINVAL; + + if (count != 4) + return -EINVAL; + + if (is_write) { + data = *((unsigned int *)buf); + switch (offset) { + case offsetof(struct vfio_region_gfx_edid, link_state): + if (data == VFIO_DEVICE_GFX_LINK_STATE_UP) { + if (!drm_edid_block_valid( + (u8 *)region->edid_blob, + 0, + true, + NULL)) { + gvt_vgpu_err("invalid EDID blob\n"); + return -EINVAL; + } + intel_gvt_ops->emulate_hotplug(vgpu, true); + } else if (data == VFIO_DEVICE_GFX_LINK_STATE_DOWN) + intel_gvt_ops->emulate_hotplug(vgpu, false); + else { + gvt_vgpu_err("invalid EDID link state %d\n", + regs->link_state); + return -EINVAL; + } + regs->link_state = data; + break; + case offsetof(struct vfio_region_gfx_edid, edid_size): + if (data > regs->edid_max_size) { + gvt_vgpu_err("EDID size is bigger than %d!\n", + regs->edid_max_size); + return -EINVAL; + } + regs->edid_size = data; + break; + default: + /* read-only regs */ + gvt_vgpu_err("write read-only EDID region at offset %d\n", + offset); + return -EPERM; + } + } else { + memcpy(buf, (char *)regs + offset, count); + } + + return count; +} + +static int handle_edid_blob(struct vfio_edid_region *region, char *buf, + size_t count, u16 offset, bool is_write) +{ + if (offset + count > region->vfio_edid_regs.edid_size) + return -EINVAL; + + if (is_write) + memcpy(region->edid_blob + offset, buf, count); + else + memcpy(buf, region->edid_blob + offset, count); + + return count; +} + +static size_t intel_vgpu_reg_rw_edid(struct intel_vgpu *vgpu, char *buf, + size_t count, loff_t *ppos, bool iswrite) +{ + int ret; + unsigned int i = VFIO_PCI_OFFSET_TO_INDEX(*ppos) - + VFIO_PCI_NUM_REGIONS; + struct vfio_edid_region *region = + (struct vfio_edid_region *)vgpu->vdev.region[i].data; + loff_t pos = *ppos & VFIO_PCI_OFFSET_MASK; + + if (pos < region->vfio_edid_regs.edid_offset) { + ret = handle_edid_regs(vgpu, region, buf, count, pos, iswrite); + } else { + pos -= EDID_BLOB_OFFSET; + ret = handle_edid_blob(region, buf, count, pos, iswrite); + } + + if (ret < 0) + gvt_vgpu_err("failed to access EDID region\n"); + + return ret; +} + +static void intel_vgpu_reg_release_edid(struct intel_vgpu *vgpu, + struct vfio_region *region) +{ + kfree(region->data); +} + +static const struct intel_vgpu_regops intel_vgpu_regops_edid = { + .rw = intel_vgpu_reg_rw_edid, + .release = intel_vgpu_reg_release_edid, +}; + static int intel_vgpu_register_reg(struct intel_vgpu *vgpu, unsigned int type, unsigned int subtype, const struct intel_vgpu_regops *ops, @@ -493,6 +605,36 @@ static int kvmgt_set_opregion(void *p_vgpu) return ret; } +static int kvmgt_set_edid(void *p_vgpu, int port_num) +{ + struct intel_vgpu *vgpu = (struct intel_vgpu *)p_vgpu; + struct intel_vgpu_port *port = intel_vgpu_port(vgpu, port_num); + struct vfio_edid_region *base; + int ret; + + base = kzalloc(sizeof(*base), GFP_KERNEL); + if (!base) + return -ENOMEM; + + /* TODO: Add multi-port and EDID extension block support */ + base->vfio_edid_regs.edid_offset = EDID_BLOB_OFFSET; + base->vfio_edid_regs.edid_max_size = EDID_SIZE; + base->vfio_edid_regs.edid_size = EDID_SIZE; + base->vfio_edid_regs.max_xres = vgpu_edid_xres(port->id); + base->vfio_edid_regs.max_yres = vgpu_edid_yres(port->id); + base->edid_blob = port->edid->edid_block; + + ret = intel_vgpu_register_reg(vgpu, + VFIO_REGION_TYPE_GFX, + VFIO_REGION_SUBTYPE_GFX_EDID, + &intel_vgpu_regops_edid, EDID_SIZE, + VFIO_REGION_INFO_FLAG_READ | + VFIO_REGION_INFO_FLAG_WRITE | + VFIO_REGION_INFO_FLAG_CAPS, base); + + return ret; +} + static void kvmgt_put_vfio_device(void *vgpu) { if (WARN_ON(!((struct intel_vgpu *)vgpu)->vdev.vfio_device)) @@ -627,6 +769,12 @@ static int intel_vgpu_open(struct mdev_device *mdev) goto undo_iommu; } + /* Take a module reference as mdev core doesn't take + * a reference for vendor driver. + */ + if (!try_module_get(THIS_MODULE)) + goto undo_group; + ret = kvmgt_guest_init(mdev); if (ret) goto undo_group; @@ -679,6 +827,9 @@ static void __intel_vgpu_release(struct intel_vgpu *vgpu) &vgpu->vdev.group_notifier); WARN(ret, "vfio_unregister_notifier for group failed: %d\n", ret); + /* dereference module reference taken at open */ + module_put(THIS_MODULE); + info = (struct kvmgt_guest_info *)vgpu->handle; kvmgt_guest_exit(info); @@ -703,7 +854,7 @@ static void intel_vgpu_release_work(struct work_struct *work) __intel_vgpu_release(vgpu); } -static uint64_t intel_vgpu_get_bar_addr(struct intel_vgpu *vgpu, int bar) +static u64 intel_vgpu_get_bar_addr(struct intel_vgpu *vgpu, int bar) { u32 start_lo, start_hi; u32 mem_type; @@ -730,10 +881,10 @@ static uint64_t intel_vgpu_get_bar_addr(struct intel_vgpu *vgpu, int bar) return ((u64)start_hi << 32) | start_lo; } -static int intel_vgpu_bar_rw(struct intel_vgpu *vgpu, int bar, uint64_t off, +static int intel_vgpu_bar_rw(struct intel_vgpu *vgpu, int bar, u64 off, void *buf, unsigned int count, bool is_write) { - uint64_t bar_start = intel_vgpu_get_bar_addr(vgpu, bar); + u64 bar_start = intel_vgpu_get_bar_addr(vgpu, bar); int ret; if (is_write) @@ -745,13 +896,13 @@ static int intel_vgpu_bar_rw(struct intel_vgpu *vgpu, int bar, uint64_t off, return ret; } -static inline bool intel_vgpu_in_aperture(struct intel_vgpu *vgpu, uint64_t off) +static inline bool intel_vgpu_in_aperture(struct intel_vgpu *vgpu, u64 off) { return off >= vgpu_aperture_offset(vgpu) && off < vgpu_aperture_offset(vgpu) + vgpu_aperture_sz(vgpu); } -static int intel_vgpu_aperture_rw(struct intel_vgpu *vgpu, uint64_t off, +static int intel_vgpu_aperture_rw(struct intel_vgpu *vgpu, u64 off, void *buf, unsigned long count, bool is_write) { void *aperture_va; @@ -783,7 +934,7 @@ static ssize_t intel_vgpu_rw(struct mdev_device *mdev, char *buf, { struct intel_vgpu *vgpu = mdev_get_drvdata(mdev); unsigned int index = VFIO_PCI_OFFSET_TO_INDEX(*ppos); - uint64_t pos = *ppos & VFIO_PCI_OFFSET_MASK; + u64 pos = *ppos & VFIO_PCI_OFFSET_MASK; int ret = -EINVAL; @@ -1029,7 +1180,7 @@ static int intel_vgpu_get_irq_count(struct intel_vgpu *vgpu, int type) static int intel_vgpu_set_intx_mask(struct intel_vgpu *vgpu, unsigned int index, unsigned int start, - unsigned int count, uint32_t flags, + unsigned int count, u32 flags, void *data) { return 0; @@ -1037,21 +1188,21 @@ static int intel_vgpu_set_intx_mask(struct intel_vgpu *vgpu, static int intel_vgpu_set_intx_unmask(struct intel_vgpu *vgpu, unsigned int index, unsigned int start, - unsigned int count, uint32_t flags, void *data) + unsigned int count, u32 flags, void *data) { return 0; } static int intel_vgpu_set_intx_trigger(struct intel_vgpu *vgpu, unsigned int index, unsigned int start, unsigned int count, - uint32_t flags, void *data) + u32 flags, void *data) { return 0; } static int intel_vgpu_set_msi_trigger(struct intel_vgpu *vgpu, unsigned int index, unsigned int start, unsigned int count, - uint32_t flags, void *data) + u32 flags, void *data) { struct eventfd_ctx *trigger; @@ -1070,12 +1221,12 @@ static int intel_vgpu_set_msi_trigger(struct intel_vgpu *vgpu, return 0; } -static int intel_vgpu_set_irqs(struct intel_vgpu *vgpu, uint32_t flags, +static int intel_vgpu_set_irqs(struct intel_vgpu *vgpu, u32 flags, unsigned int index, unsigned int start, unsigned int count, void *data) { int (*func)(struct intel_vgpu *vgpu, unsigned int index, - unsigned int start, unsigned int count, uint32_t flags, + unsigned int start, unsigned int count, u32 flags, void *data) = NULL; switch (index) { @@ -1467,7 +1618,7 @@ static int kvmgt_host_init(struct device *dev, void *gvt, const void *ops) return mdev_register_device(dev, &intel_vgpu_ops); } -static void kvmgt_host_exit(struct device *dev, void *gvt) +static void kvmgt_host_exit(struct device *dev) { mdev_unregister_device(dev); } @@ -1849,7 +2000,8 @@ static bool kvmgt_is_valid_gfn(unsigned long handle, unsigned long gfn) return ret; } -struct intel_gvt_mpt kvmgt_mpt = { +static struct intel_gvt_mpt kvmgt_mpt = { + .type = INTEL_GVT_HYPERVISOR_KVM, .host_init = kvmgt_host_init, .host_exit = kvmgt_host_exit, .attach_vgpu = kvmgt_attach_vgpu, @@ -1864,19 +2016,22 @@ struct intel_gvt_mpt kvmgt_mpt = { .dma_map_guest_page = kvmgt_dma_map_guest_page, .dma_unmap_guest_page = kvmgt_dma_unmap_guest_page, .set_opregion = kvmgt_set_opregion, + .set_edid = kvmgt_set_edid, .get_vfio_device = kvmgt_get_vfio_device, .put_vfio_device = kvmgt_put_vfio_device, .is_valid_gfn = kvmgt_is_valid_gfn, }; -EXPORT_SYMBOL_GPL(kvmgt_mpt); static int __init kvmgt_init(void) { + if (intel_gvt_register_hypervisor(&kvmgt_mpt) < 0) + return -ENODEV; return 0; } static void __exit kvmgt_exit(void) { + intel_gvt_unregister_hypervisor(); } module_init(kvmgt_init); diff --git a/drivers/gpu/drm/i915/gvt/mmio.c b/drivers/gpu/drm/i915/gvt/mmio.c index 43f65848ecd6..ed4df2f6d60b 100644 --- a/drivers/gpu/drm/i915/gvt/mmio.c +++ b/drivers/gpu/drm/i915/gvt/mmio.c @@ -57,7 +57,7 @@ int intel_vgpu_gpa_to_mmio_offset(struct intel_vgpu *vgpu, u64 gpa) (reg >= gvt->device_info.gtt_start_offset \ && reg < gvt->device_info.gtt_start_offset + gvt_ggtt_sz(gvt)) -static void failsafe_emulate_mmio_rw(struct intel_vgpu *vgpu, uint64_t pa, +static void failsafe_emulate_mmio_rw(struct intel_vgpu *vgpu, u64 pa, void *p_data, unsigned int bytes, bool read) { struct intel_gvt *gvt = NULL; @@ -99,7 +99,7 @@ static void failsafe_emulate_mmio_rw(struct intel_vgpu *vgpu, uint64_t pa, * Returns: * Zero on success, negative error code if failed */ -int intel_vgpu_emulate_mmio_read(struct intel_vgpu *vgpu, uint64_t pa, +int intel_vgpu_emulate_mmio_read(struct intel_vgpu *vgpu, u64 pa, void *p_data, unsigned int bytes) { struct intel_gvt *gvt = vgpu->gvt; @@ -171,7 +171,7 @@ out: * Returns: * Zero on success, negative error code if failed */ -int intel_vgpu_emulate_mmio_write(struct intel_vgpu *vgpu, uint64_t pa, +int intel_vgpu_emulate_mmio_write(struct intel_vgpu *vgpu, u64 pa, void *p_data, unsigned int bytes) { struct intel_gvt *gvt = vgpu->gvt; diff --git a/drivers/gpu/drm/i915/gvt/mmio.h b/drivers/gpu/drm/i915/gvt/mmio.h index 1ffc69eba30e..5874f1cb4306 100644 --- a/drivers/gpu/drm/i915/gvt/mmio.h +++ b/drivers/gpu/drm/i915/gvt/mmio.h @@ -43,15 +43,16 @@ struct intel_vgpu; #define D_SKL (1 << 1) #define D_KBL (1 << 2) #define D_BXT (1 << 3) +#define D_CFL (1 << 4) -#define D_GEN9PLUS (D_SKL | D_KBL | D_BXT) -#define D_GEN8PLUS (D_BDW | D_SKL | D_KBL | D_BXT) +#define D_GEN9PLUS (D_SKL | D_KBL | D_BXT | D_CFL) +#define D_GEN8PLUS (D_BDW | D_SKL | D_KBL | D_BXT | D_CFL) -#define D_SKL_PLUS (D_SKL | D_KBL | D_BXT) -#define D_BDW_PLUS (D_BDW | D_SKL | D_KBL | D_BXT) +#define D_SKL_PLUS (D_SKL | D_KBL | D_BXT | D_CFL) +#define D_BDW_PLUS (D_BDW | D_SKL | D_KBL | D_BXT | D_CFL) #define D_PRE_SKL (D_BDW) -#define D_ALL (D_BDW | D_SKL | D_KBL | D_BXT) +#define D_ALL (D_BDW | D_SKL | D_KBL | D_BXT | D_CFL) typedef int (*gvt_mmio_func)(struct intel_vgpu *, unsigned int, void *, unsigned int); diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.c b/drivers/gpu/drm/i915/gvt/mmio_context.c index d6e02c15ef97..7d84cfb9051a 100644 --- a/drivers/gpu/drm/i915/gvt/mmio_context.c +++ b/drivers/gpu/drm/i915/gvt/mmio_context.c @@ -353,8 +353,7 @@ static void handle_tlb_pending_event(struct intel_vgpu *vgpu, int ring_id) */ fw = intel_uncore_forcewake_for_reg(dev_priv, reg, FW_REG_READ | FW_REG_WRITE); - if (ring_id == RCS && (IS_SKYLAKE(dev_priv) || - IS_KABYLAKE(dev_priv) || IS_BROXTON(dev_priv))) + if (ring_id == RCS && (INTEL_GEN(dev_priv) >= 9)) fw |= FORCEWAKE_RENDER; intel_uncore_forcewake_get(dev_priv, fw); @@ -391,7 +390,8 @@ static void switch_mocs(struct intel_vgpu *pre, struct intel_vgpu *next, if (WARN_ON(ring_id >= ARRAY_SIZE(regs))) return; - if ((IS_KABYLAKE(dev_priv) || IS_BROXTON(dev_priv)) && ring_id == RCS) + if ((IS_KABYLAKE(dev_priv) || IS_BROXTON(dev_priv) + || IS_COFFEELAKE(dev_priv)) && ring_id == RCS) return; if (!pre && !gen9_render_mocs.initialized) @@ -457,9 +457,7 @@ static void switch_mmio(struct intel_vgpu *pre, u32 old_v, new_v; dev_priv = pre ? pre->gvt->dev_priv : next->gvt->dev_priv; - if (IS_SKYLAKE(dev_priv) - || IS_KABYLAKE(dev_priv) - || IS_BROXTON(dev_priv)) + if (INTEL_GEN(dev_priv) >= 9) switch_mocs(pre, next, ring_id); for (mmio = dev_priv->gvt->engine_mmio_list.mmio; @@ -471,8 +469,8 @@ static void switch_mmio(struct intel_vgpu *pre, * state image on kabylake, it's initialized by lri command and * save or restore with context together. */ - if ((IS_KABYLAKE(dev_priv) || IS_BROXTON(dev_priv)) - && mmio->in_context) + if ((IS_KABYLAKE(dev_priv) || IS_BROXTON(dev_priv) + || IS_COFFEELAKE(dev_priv)) && mmio->in_context) continue; // save @@ -565,9 +563,7 @@ void intel_gvt_init_engine_mmio_context(struct intel_gvt *gvt) { struct engine_mmio *mmio; - if (IS_SKYLAKE(gvt->dev_priv) || - IS_KABYLAKE(gvt->dev_priv) || - IS_BROXTON(gvt->dev_priv)) + if (INTEL_GEN(gvt->dev_priv) >= 9) gvt->engine_mmio_list.mmio = gen9_engine_mmio_list; else gvt->engine_mmio_list.mmio = gen8_engine_mmio_list; diff --git a/drivers/gpu/drm/i915/gvt/mpt.h b/drivers/gpu/drm/i915/gvt/mpt.h index 67f19992b226..5d8b8f228d8f 100644 --- a/drivers/gpu/drm/i915/gvt/mpt.h +++ b/drivers/gpu/drm/i915/gvt/mpt.h @@ -50,11 +50,10 @@ * Zero on success, negative error code if failed */ static inline int intel_gvt_hypervisor_host_init(struct device *dev, - void *gvt, const void *ops) + void *gvt, const void *ops) { - /* optional to provide */ if (!intel_gvt_host.mpt->host_init) - return 0; + return -ENODEV; return intel_gvt_host.mpt->host_init(dev, gvt, ops); } @@ -62,14 +61,13 @@ static inline int intel_gvt_hypervisor_host_init(struct device *dev, /** * intel_gvt_hypervisor_host_exit - exit GVT-g host side */ -static inline void intel_gvt_hypervisor_host_exit(struct device *dev, - void *gvt) +static inline void intel_gvt_hypervisor_host_exit(struct device *dev) { /* optional to provide */ if (!intel_gvt_host.mpt->host_exit) return; - intel_gvt_host.mpt->host_exit(dev, gvt); + intel_gvt_host.mpt->host_exit(dev); } /** @@ -316,6 +314,23 @@ static inline int intel_gvt_hypervisor_set_opregion(struct intel_vgpu *vgpu) } /** + * intel_gvt_hypervisor_set_edid - Set EDID region for guest + * @vgpu: a vGPU + * @port_num: display port number + * + * Returns: + * Zero on success, negative error code if failed. + */ +static inline int intel_gvt_hypervisor_set_edid(struct intel_vgpu *vgpu, + int port_num) +{ + if (!intel_gvt_host.mpt->set_edid) + return 0; + + return intel_gvt_host.mpt->set_edid(vgpu, port_num); +} + +/** * intel_gvt_hypervisor_get_vfio_device - increase vfio device ref count * @vgpu: a vGPU * @@ -362,4 +377,7 @@ static inline bool intel_gvt_hypervisor_is_valid_gfn( return intel_gvt_host.mpt->is_valid_gfn(vgpu->handle, gfn); } +int intel_gvt_register_hypervisor(struct intel_gvt_mpt *); +void intel_gvt_unregister_hypervisor(void); + #endif /* _GVT_MPT_H_ */ diff --git a/drivers/gpu/drm/i915/gvt/sched_policy.c b/drivers/gpu/drm/i915/gvt/sched_policy.c index c32e7d5e8629..1c763a27a412 100644 --- a/drivers/gpu/drm/i915/gvt/sched_policy.c +++ b/drivers/gpu/drm/i915/gvt/sched_policy.c @@ -94,7 +94,7 @@ static void gvt_balance_timeslice(struct gvt_sched_data *sched_data) { struct vgpu_sched_data *vgpu_data; struct list_head *pos; - static uint64_t stage_check; + static u64 stage_check; int stage = stage_check++ % GVT_TS_BALANCE_STAGE_NUM; /* The timeslice accumulation reset at stage 0, which is @@ -474,6 +474,6 @@ void intel_vgpu_stop_schedule(struct intel_vgpu *vgpu) } } spin_unlock_bh(&scheduler->mmio_context_lock); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put_unchecked(dev_priv); mutex_unlock(&vgpu->gvt->sched_lock); } diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c index 1ad8c5e1455d..b7957eefb976 100644 --- a/drivers/gpu/drm/i915/gvt/scheduler.c +++ b/drivers/gpu/drm/i915/gvt/scheduler.c @@ -299,7 +299,8 @@ static int copy_workload_to_ring_buffer(struct intel_vgpu_workload *workload) void *shadow_ring_buffer_va; u32 *cs; - if ((IS_KABYLAKE(req->i915) || IS_BROXTON(req->i915)) + if ((IS_KABYLAKE(req->i915) || IS_BROXTON(req->i915) + || IS_COFFEELAKE(req->i915)) && is_inhibit_context(req->hw_context)) intel_vgpu_restore_inhibit_context(vgpu, req); @@ -939,9 +940,7 @@ static int workload_thread(void *priv) struct intel_vgpu_workload *workload = NULL; struct intel_vgpu *vgpu = NULL; int ret; - bool need_force_wake = IS_SKYLAKE(gvt->dev_priv) - || IS_KABYLAKE(gvt->dev_priv) - || IS_BROXTON(gvt->dev_priv); + bool need_force_wake = (INTEL_GEN(gvt->dev_priv) >= 9); DEFINE_WAIT_FUNC(wait, woken_wake_function); kfree(p); @@ -997,7 +996,7 @@ complete: intel_uncore_forcewake_put(gvt->dev_priv, FORCEWAKE_ALL); - intel_runtime_pm_put(gvt->dev_priv); + intel_runtime_pm_put_unchecked(gvt->dev_priv); if (ret && (vgpu_is_vm_unhealthy(ret))) enter_failsafe_mode(vgpu, GVT_FAILSAFE_GUEST_ERR); } @@ -1451,7 +1450,7 @@ intel_vgpu_create_workload(struct intel_vgpu *vgpu, int ring_id, mutex_lock(&dev_priv->drm.struct_mutex); ret = intel_gvt_scan_and_shadow_workload(workload); mutex_unlock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put_unchecked(dev_priv); } if (ret && (vgpu_is_vm_unhealthy(ret))) { diff --git a/drivers/gpu/drm/i915/gvt/scheduler.h b/drivers/gpu/drm/i915/gvt/scheduler.h index ca5529d0e48e..1e9eec6a32fe 100644 --- a/drivers/gpu/drm/i915/gvt/scheduler.h +++ b/drivers/gpu/drm/i915/gvt/scheduler.h @@ -61,7 +61,7 @@ struct shadow_indirect_ctx { unsigned long guest_gma; unsigned long shadow_gma; void *shadow_va; - uint32_t size; + u32 size; }; #define PER_CTX_ADDR_MASK 0xfffff000 diff --git a/drivers/gpu/drm/i915/gvt/trace.h b/drivers/gpu/drm/i915/gvt/trace.h index 1fd64202d74e..6d787750d279 100644 --- a/drivers/gpu/drm/i915/gvt/trace.h +++ b/drivers/gpu/drm/i915/gvt/trace.h @@ -228,7 +228,7 @@ TRACE_EVENT(oos_sync, TRACE_EVENT(gvt_command, TP_PROTO(u8 vgpu_id, u8 ring_id, u32 ip_gma, u32 *cmd_va, u32 cmd_len, u32 buf_type, u32 buf_addr_type, - void *workload, char *cmd_name), + void *workload, const char *cmd_name), TP_ARGS(vgpu_id, ring_id, ip_gma, cmd_va, cmd_len, buf_type, buf_addr_type, workload, cmd_name), diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c index e1c860f80eb0..720e2b10adaa 100644 --- a/drivers/gpu/drm/i915/gvt/vgpu.c +++ b/drivers/gpu/drm/i915/gvt/vgpu.c @@ -428,6 +428,12 @@ static struct intel_vgpu *__intel_gvt_create_vgpu(struct intel_gvt *gvt, if (ret) goto out_clean_sched_policy; + /*TODO: add more platforms support */ + if (IS_SKYLAKE(gvt->dev_priv) || IS_KABYLAKE(gvt->dev_priv)) + ret = intel_gvt_hypervisor_set_edid(vgpu, PORT_D); + if (ret) + goto out_clean_sched_policy; + return vgpu; out_clean_sched_policy: diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c new file mode 100644 index 000000000000..215b6ff8aa73 --- /dev/null +++ b/drivers/gpu/drm/i915/i915_active.c @@ -0,0 +1,286 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2019 Intel Corporation + */ + +#include "i915_drv.h" +#include "i915_active.h" + +#define BKL(ref) (&(ref)->i915->drm.struct_mutex) + +/* + * Active refs memory management + * + * To be more economical with memory, we reap all the i915_active trees as + * they idle (when we know the active requests are inactive) and allocate the + * nodes from a local slab cache to hopefully reduce the fragmentation. + */ +static struct i915_global_active { + struct kmem_cache *slab_cache; +} global; + +struct active_node { + struct i915_active_request base; + struct i915_active *ref; + struct rb_node node; + u64 timeline; +}; + +static void +__active_park(struct i915_active *ref) +{ + struct active_node *it, *n; + + rbtree_postorder_for_each_entry_safe(it, n, &ref->tree, node) { + GEM_BUG_ON(i915_active_request_isset(&it->base)); + kmem_cache_free(global.slab_cache, it); + } + ref->tree = RB_ROOT; +} + +static void +__active_retire(struct i915_active *ref) +{ + GEM_BUG_ON(!ref->count); + if (--ref->count) + return; + + /* return the unused nodes to our slabcache */ + __active_park(ref); + + ref->retire(ref); +} + +static void +node_retire(struct i915_active_request *base, struct i915_request *rq) +{ + __active_retire(container_of(base, struct active_node, base)->ref); +} + +static void +last_retire(struct i915_active_request *base, struct i915_request *rq) +{ + __active_retire(container_of(base, struct i915_active, last)); +} + +static struct i915_active_request * +active_instance(struct i915_active *ref, u64 idx) +{ + struct active_node *node; + struct rb_node **p, *parent; + struct i915_request *old; + + /* + * We track the most recently used timeline to skip a rbtree search + * for the common case, under typical loads we never need the rbtree + * at all. We can reuse the last slot if it is empty, that is + * after the previous activity has been retired, or if it matches the + * current timeline. + * + * Note that we allow the timeline to be active simultaneously in + * the rbtree and the last cache. We do this to avoid having + * to search and replace the rbtree element for a new timeline, with + * the cost being that we must be aware that the ref may be retired + * twice for the same timeline (as the older rbtree element will be + * retired before the new request added to last). + */ + old = i915_active_request_raw(&ref->last, BKL(ref)); + if (!old || old->fence.context == idx) + goto out; + + /* Move the currently active fence into the rbtree */ + idx = old->fence.context; + + parent = NULL; + p = &ref->tree.rb_node; + while (*p) { + parent = *p; + + node = rb_entry(parent, struct active_node, node); + if (node->timeline == idx) + goto replace; + + if (node->timeline < idx) + p = &parent->rb_right; + else + p = &parent->rb_left; + } + + node = kmem_cache_alloc(global.slab_cache, GFP_KERNEL); + + /* kmalloc may retire the ref->last (thanks shrinker)! */ + if (unlikely(!i915_active_request_raw(&ref->last, BKL(ref)))) { + kmem_cache_free(global.slab_cache, node); + goto out; + } + + if (unlikely(!node)) + return ERR_PTR(-ENOMEM); + + i915_active_request_init(&node->base, NULL, node_retire); + node->ref = ref; + node->timeline = idx; + + rb_link_node(&node->node, parent, p); + rb_insert_color(&node->node, &ref->tree); + +replace: + /* + * Overwrite the previous active slot in the rbtree with last, + * leaving last zeroed. If the previous slot is still active, + * we must be careful as we now only expect to receive one retire + * callback not two, and so much undo the active counting for the + * overwritten slot. + */ + if (i915_active_request_isset(&node->base)) { + /* Retire ourselves from the old rq->active_list */ + __list_del_entry(&node->base.link); + ref->count--; + GEM_BUG_ON(!ref->count); + } + GEM_BUG_ON(list_empty(&ref->last.link)); + list_replace_init(&ref->last.link, &node->base.link); + node->base.request = fetch_and_zero(&ref->last.request); + +out: + return &ref->last; +} + +void i915_active_init(struct drm_i915_private *i915, + struct i915_active *ref, + void (*retire)(struct i915_active *ref)) +{ + ref->i915 = i915; + ref->retire = retire; + ref->tree = RB_ROOT; + i915_active_request_init(&ref->last, NULL, last_retire); + ref->count = 0; +} + +int i915_active_ref(struct i915_active *ref, + u64 timeline, + struct i915_request *rq) +{ + struct i915_active_request *active; + + active = active_instance(ref, timeline); + if (IS_ERR(active)) + return PTR_ERR(active); + + if (!i915_active_request_isset(active)) + ref->count++; + __i915_active_request_set(active, rq); + + GEM_BUG_ON(!ref->count); + return 0; +} + +bool i915_active_acquire(struct i915_active *ref) +{ + lockdep_assert_held(BKL(ref)); + return !ref->count++; +} + +void i915_active_release(struct i915_active *ref) +{ + lockdep_assert_held(BKL(ref)); + __active_retire(ref); +} + +int i915_active_wait(struct i915_active *ref) +{ + struct active_node *it, *n; + int ret = 0; + + if (i915_active_acquire(ref)) + goto out_release; + + ret = i915_active_request_retire(&ref->last, BKL(ref)); + if (ret) + goto out_release; + + rbtree_postorder_for_each_entry_safe(it, n, &ref->tree, node) { + ret = i915_active_request_retire(&it->base, BKL(ref)); + if (ret) + break; + } + +out_release: + i915_active_release(ref); + return ret; +} + +int i915_request_await_active_request(struct i915_request *rq, + struct i915_active_request *active) +{ + struct i915_request *barrier = + i915_active_request_raw(active, &rq->i915->drm.struct_mutex); + + return barrier ? i915_request_await_dma_fence(rq, &barrier->fence) : 0; +} + +int i915_request_await_active(struct i915_request *rq, struct i915_active *ref) +{ + struct active_node *it, *n; + int ret; + + ret = i915_request_await_active_request(rq, &ref->last); + if (ret) + return ret; + + rbtree_postorder_for_each_entry_safe(it, n, &ref->tree, node) { + ret = i915_request_await_active_request(rq, &it->base); + if (ret) + return ret; + } + + return 0; +} + +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM) +void i915_active_fini(struct i915_active *ref) +{ + GEM_BUG_ON(i915_active_request_isset(&ref->last)); + GEM_BUG_ON(!RB_EMPTY_ROOT(&ref->tree)); + GEM_BUG_ON(ref->count); +} +#endif + +int i915_active_request_set(struct i915_active_request *active, + struct i915_request *rq) +{ + int err; + + /* Must maintain ordering wrt previous active requests */ + err = i915_request_await_active_request(rq, active); + if (err) + return err; + + __i915_active_request_set(active, rq); + return 0; +} + +void i915_active_retire_noop(struct i915_active_request *active, + struct i915_request *request) +{ + /* Space left intentionally blank */ +} + +#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) +#include "selftests/i915_active.c" +#endif + +int __init i915_global_active_init(void) +{ + global.slab_cache = KMEM_CACHE(active_node, SLAB_HWCACHE_ALIGN); + if (!global.slab_cache) + return -ENOMEM; + + return 0; +} + +void __exit i915_global_active_exit(void) +{ + kmem_cache_destroy(global.slab_cache); +} diff --git a/drivers/gpu/drm/i915/i915_active.h b/drivers/gpu/drm/i915/i915_active.h new file mode 100644 index 000000000000..12b5c1d287d1 --- /dev/null +++ b/drivers/gpu/drm/i915/i915_active.h @@ -0,0 +1,425 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2019 Intel Corporation + */ + +#ifndef _I915_ACTIVE_H_ +#define _I915_ACTIVE_H_ + +#include <linux/lockdep.h> + +#include "i915_active_types.h" +#include "i915_request.h" + +/* + * We treat requests as fences. This is not be to confused with our + * "fence registers" but pipeline synchronisation objects ala GL_ARB_sync. + * We use the fences to synchronize access from the CPU with activity on the + * GPU, for example, we should not rewrite an object's PTE whilst the GPU + * is reading them. We also track fences at a higher level to provide + * implicit synchronisation around GEM objects, e.g. set-domain will wait + * for outstanding GPU rendering before marking the object ready for CPU + * access, or a pageflip will wait until the GPU is complete before showing + * the frame on the scanout. + * + * In order to use a fence, the object must track the fence it needs to + * serialise with. For example, GEM objects want to track both read and + * write access so that we can perform concurrent read operations between + * the CPU and GPU engines, as well as waiting for all rendering to + * complete, or waiting for the last GPU user of a "fence register". The + * object then embeds a #i915_active_request to track the most recent (in + * retirement order) request relevant for the desired mode of access. + * The #i915_active_request is updated with i915_active_request_set() to + * track the most recent fence request, typically this is done as part of + * i915_vma_move_to_active(). + * + * When the #i915_active_request completes (is retired), it will + * signal its completion to the owner through a callback as well as mark + * itself as idle (i915_active_request.request == NULL). The owner + * can then perform any action, such as delayed freeing of an active + * resource including itself. + */ + +void i915_active_retire_noop(struct i915_active_request *active, + struct i915_request *request); + +/** + * i915_active_request_init - prepares the activity tracker for use + * @active - the active tracker + * @rq - initial request to track, can be NULL + * @func - a callback when then the tracker is retired (becomes idle), + * can be NULL + * + * i915_active_request_init() prepares the embedded @active struct for use as + * an activity tracker, that is for tracking the last known active request + * associated with it. When the last request becomes idle, when it is retired + * after completion, the optional callback @func is invoked. + */ +static inline void +i915_active_request_init(struct i915_active_request *active, + struct i915_request *rq, + i915_active_retire_fn retire) +{ + RCU_INIT_POINTER(active->request, rq); + INIT_LIST_HEAD(&active->link); + active->retire = retire ?: i915_active_retire_noop; +} + +#define INIT_ACTIVE_REQUEST(name) i915_active_request_init((name), NULL, NULL) + +/** + * i915_active_request_set - updates the tracker to watch the current request + * @active - the active tracker + * @request - the request to watch + * + * __i915_active_request_set() watches the given @request for completion. Whilst + * that @request is busy, the @active reports busy. When that @request is + * retired, the @active tracker is updated to report idle. + */ +static inline void +__i915_active_request_set(struct i915_active_request *active, + struct i915_request *request) +{ + list_move(&active->link, &request->active_list); + rcu_assign_pointer(active->request, request); +} + +int __must_check +i915_active_request_set(struct i915_active_request *active, + struct i915_request *rq); + +/** + * i915_active_request_set_retire_fn - updates the retirement callback + * @active - the active tracker + * @fn - the routine called when the request is retired + * @mutex - struct_mutex used to guard retirements + * + * i915_active_request_set_retire_fn() updates the function pointer that + * is called when the final request associated with the @active tracker + * is retired. + */ +static inline void +i915_active_request_set_retire_fn(struct i915_active_request *active, + i915_active_retire_fn fn, + struct mutex *mutex) +{ + lockdep_assert_held(mutex); + active->retire = fn ?: i915_active_retire_noop; +} + +static inline struct i915_request * +__i915_active_request_peek(const struct i915_active_request *active) +{ + /* + * Inside the error capture (running with the driver in an unknown + * state), we want to bend the rules slightly (a lot). + * + * Work is in progress to make it safer, in the meantime this keeps + * the known issue from spamming the logs. + */ + return rcu_dereference_protected(active->request, 1); +} + +/** + * i915_active_request_raw - return the active request + * @active - the active tracker + * + * i915_active_request_raw() returns the current request being tracked, or NULL. + * It does not obtain a reference on the request for the caller, so the caller + * must hold struct_mutex. + */ +static inline struct i915_request * +i915_active_request_raw(const struct i915_active_request *active, + struct mutex *mutex) +{ + return rcu_dereference_protected(active->request, + lockdep_is_held(mutex)); +} + +/** + * i915_active_request_peek - report the active request being monitored + * @active - the active tracker + * + * i915_active_request_peek() returns the current request being tracked if + * still active, or NULL. It does not obtain a reference on the request + * for the caller, so the caller must hold struct_mutex. + */ +static inline struct i915_request * +i915_active_request_peek(const struct i915_active_request *active, + struct mutex *mutex) +{ + struct i915_request *request; + + request = i915_active_request_raw(active, mutex); + if (!request || i915_request_completed(request)) + return NULL; + + return request; +} + +/** + * i915_active_request_get - return a reference to the active request + * @active - the active tracker + * + * i915_active_request_get() returns a reference to the active request, or NULL + * if the active tracker is idle. The caller must hold struct_mutex. + */ +static inline struct i915_request * +i915_active_request_get(const struct i915_active_request *active, + struct mutex *mutex) +{ + return i915_request_get(i915_active_request_peek(active, mutex)); +} + +/** + * __i915_active_request_get_rcu - return a reference to the active request + * @active - the active tracker + * + * __i915_active_request_get() returns a reference to the active request, + * or NULL if the active tracker is idle. The caller must hold the RCU read + * lock, but the returned pointer is safe to use outside of RCU. + */ +static inline struct i915_request * +__i915_active_request_get_rcu(const struct i915_active_request *active) +{ + /* + * Performing a lockless retrieval of the active request is super + * tricky. SLAB_TYPESAFE_BY_RCU merely guarantees that the backing + * slab of request objects will not be freed whilst we hold the + * RCU read lock. It does not guarantee that the request itself + * will not be freed and then *reused*. Viz, + * + * Thread A Thread B + * + * rq = active.request + * retire(rq) -> free(rq); + * (rq is now first on the slab freelist) + * active.request = NULL + * + * rq = new submission on a new object + * ref(rq) + * + * To prevent the request from being reused whilst the caller + * uses it, we take a reference like normal. Whilst acquiring + * the reference we check that it is not in a destroyed state + * (refcnt == 0). That prevents the request being reallocated + * whilst the caller holds on to it. To check that the request + * was not reallocated as we acquired the reference we have to + * check that our request remains the active request across + * the lookup, in the same manner as a seqlock. The visibility + * of the pointer versus the reference counting is controlled + * by using RCU barriers (rcu_dereference and rcu_assign_pointer). + * + * In the middle of all that, we inspect whether the request is + * complete. Retiring is lazy so the request may be completed long + * before the active tracker is updated. Querying whether the + * request is complete is far cheaper (as it involves no locked + * instructions setting cachelines to exclusive) than acquiring + * the reference, so we do it first. The RCU read lock ensures the + * pointer dereference is valid, but does not ensure that the + * seqno nor HWS is the right one! However, if the request was + * reallocated, that means the active tracker's request was complete. + * If the new request is also complete, then both are and we can + * just report the active tracker is idle. If the new request is + * incomplete, then we acquire a reference on it and check that + * it remained the active request. + * + * It is then imperative that we do not zero the request on + * reallocation, so that we can chase the dangling pointers! + * See i915_request_alloc(). + */ + do { + struct i915_request *request; + + request = rcu_dereference(active->request); + if (!request || i915_request_completed(request)) + return NULL; + + /* + * An especially silly compiler could decide to recompute the + * result of i915_request_completed, more specifically + * re-emit the load for request->fence.seqno. A race would catch + * a later seqno value, which could flip the result from true to + * false. Which means part of the instructions below might not + * be executed, while later on instructions are executed. Due to + * barriers within the refcounting the inconsistency can't reach + * past the call to i915_request_get_rcu, but not executing + * that while still executing i915_request_put() creates + * havoc enough. Prevent this with a compiler barrier. + */ + barrier(); + + request = i915_request_get_rcu(request); + + /* + * What stops the following rcu_access_pointer() from occurring + * before the above i915_request_get_rcu()? If we were + * to read the value before pausing to get the reference to + * the request, we may not notice a change in the active + * tracker. + * + * The rcu_access_pointer() is a mere compiler barrier, which + * means both the CPU and compiler are free to perform the + * memory read without constraint. The compiler only has to + * ensure that any operations after the rcu_access_pointer() + * occur afterwards in program order. This means the read may + * be performed earlier by an out-of-order CPU, or adventurous + * compiler. + * + * The atomic operation at the heart of + * i915_request_get_rcu(), see dma_fence_get_rcu(), is + * atomic_inc_not_zero() which is only a full memory barrier + * when successful. That is, if i915_request_get_rcu() + * returns the request (and so with the reference counted + * incremented) then the following read for rcu_access_pointer() + * must occur after the atomic operation and so confirm + * that this request is the one currently being tracked. + * + * The corresponding write barrier is part of + * rcu_assign_pointer(). + */ + if (!request || request == rcu_access_pointer(active->request)) + return rcu_pointer_handoff(request); + + i915_request_put(request); + } while (1); +} + +/** + * i915_active_request_get_unlocked - return a reference to the active request + * @active - the active tracker + * + * i915_active_request_get_unlocked() returns a reference to the active request, + * or NULL if the active tracker is idle. The reference is obtained under RCU, + * so no locking is required by the caller. + * + * The reference should be freed with i915_request_put(). + */ +static inline struct i915_request * +i915_active_request_get_unlocked(const struct i915_active_request *active) +{ + struct i915_request *request; + + rcu_read_lock(); + request = __i915_active_request_get_rcu(active); + rcu_read_unlock(); + + return request; +} + +/** + * i915_active_request_isset - report whether the active tracker is assigned + * @active - the active tracker + * + * i915_active_request_isset() returns true if the active tracker is currently + * assigned to a request. Due to the lazy retiring, that request may be idle + * and this may report stale information. + */ +static inline bool +i915_active_request_isset(const struct i915_active_request *active) +{ + return rcu_access_pointer(active->request); +} + +/** + * i915_active_request_retire - waits until the request is retired + * @active - the active request on which to wait + * + * i915_active_request_retire() waits until the request is completed, + * and then ensures that at least the retirement handler for this + * @active tracker is called before returning. If the @active + * tracker is idle, the function returns immediately. + */ +static inline int __must_check +i915_active_request_retire(struct i915_active_request *active, + struct mutex *mutex) +{ + struct i915_request *request; + long ret; + + request = i915_active_request_raw(active, mutex); + if (!request) + return 0; + + ret = i915_request_wait(request, + I915_WAIT_INTERRUPTIBLE | I915_WAIT_LOCKED, + MAX_SCHEDULE_TIMEOUT); + if (ret < 0) + return ret; + + list_del_init(&active->link); + RCU_INIT_POINTER(active->request, NULL); + + active->retire(active, request); + + return 0; +} + +/* + * GPU activity tracking + * + * Each set of commands submitted to the GPU compromises a single request that + * signals a fence upon completion. struct i915_request combines the + * command submission, scheduling and fence signaling roles. If we want to see + * if a particular task is complete, we need to grab the fence (struct + * i915_request) for that task and check or wait for it to be signaled. More + * often though we want to track the status of a bunch of tasks, for example + * to wait for the GPU to finish accessing some memory across a variety of + * different command pipelines from different clients. We could choose to + * track every single request associated with the task, but knowing that + * each request belongs to an ordered timeline (later requests within a + * timeline must wait for earlier requests), we need only track the + * latest request in each timeline to determine the overall status of the + * task. + * + * struct i915_active provides this tracking across timelines. It builds a + * composite shared-fence, and is updated as new work is submitted to the task, + * forming a snapshot of the current status. It should be embedded into the + * different resources that need to track their associated GPU activity to + * provide a callback when that GPU activity has ceased, or otherwise to + * provide a serialisation point either for request submission or for CPU + * synchronisation. + */ + +void i915_active_init(struct drm_i915_private *i915, + struct i915_active *ref, + void (*retire)(struct i915_active *ref)); + +int i915_active_ref(struct i915_active *ref, + u64 timeline, + struct i915_request *rq); + +int i915_active_wait(struct i915_active *ref); + +int i915_request_await_active(struct i915_request *rq, + struct i915_active *ref); +int i915_request_await_active_request(struct i915_request *rq, + struct i915_active_request *active); + +bool i915_active_acquire(struct i915_active *ref); + +static inline void i915_active_cancel(struct i915_active *ref) +{ + GEM_BUG_ON(ref->count != 1); + ref->count = 0; +} + +void i915_active_release(struct i915_active *ref); + +static inline bool +i915_active_is_idle(const struct i915_active *ref) +{ + return !ref->count; +} + +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM) +void i915_active_fini(struct i915_active *ref); +#else +static inline void i915_active_fini(struct i915_active *ref) { } +#endif + +int i915_global_active_init(void); +void i915_global_active_exit(void); + +#endif /* _I915_ACTIVE_H_ */ diff --git a/drivers/gpu/drm/i915/i915_active_types.h b/drivers/gpu/drm/i915/i915_active_types.h new file mode 100644 index 000000000000..b679253b53a5 --- /dev/null +++ b/drivers/gpu/drm/i915/i915_active_types.h @@ -0,0 +1,36 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2019 Intel Corporation + */ + +#ifndef _I915_ACTIVE_TYPES_H_ +#define _I915_ACTIVE_TYPES_H_ + +#include <linux/rbtree.h> +#include <linux/rcupdate.h> + +struct drm_i915_private; +struct i915_active_request; +struct i915_request; + +typedef void (*i915_active_retire_fn)(struct i915_active_request *, + struct i915_request *); + +struct i915_active_request { + struct i915_request __rcu *request; + struct list_head link; + i915_active_retire_fn retire; +}; + +struct i915_active { + struct drm_i915_private *i915; + + struct rb_root tree; + struct i915_active_request last; + unsigned int count; + + void (*retire)(struct i915_active *ref); +}; + +#endif /* _I915_ACTIVE_TYPES_H_ */ diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 978da8838393..0bd890c04fe4 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -33,6 +33,8 @@ #include "intel_drv.h" #include "intel_guc_submission.h" +#include "i915_reset.h" + static inline struct drm_i915_private *node_to_i915(struct drm_info_node *node) { return to_i915(node->minor->dev); @@ -158,14 +160,14 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj) obj->mm.madv == I915_MADV_DONTNEED ? " purgeable" : ""); if (obj->base.name) seq_printf(m, " (name: %d)", obj->base.name); - list_for_each_entry(vma, &obj->vma_list, obj_link) { + list_for_each_entry(vma, &obj->vma.list, obj_link) { if (i915_vma_is_pinned(vma)) pin_count++; } seq_printf(m, " (pinned x %d)", pin_count); if (obj->pin_global) seq_printf(m, " (global)"); - list_for_each_entry(vma, &obj->vma_list, obj_link) { + list_for_each_entry(vma, &obj->vma.list, obj_link) { if (!drm_mm_node_allocated(&vma->node)) continue; @@ -205,7 +207,7 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj) if (vma->fence) seq_printf(m, " , fence: %d%s", vma->fence->id, - i915_gem_active_isset(&vma->last_fence) ? "*" : ""); + i915_active_request_isset(&vma->last_fence) ? "*" : ""); seq_puts(m, ")"); } if (obj->stolen) @@ -321,7 +323,7 @@ static int per_file_stats(int id, void *ptr, void *data) if (obj->base.name || obj->base.dma_buf) stats->shared += obj->base.size; - list_for_each_entry(vma, &obj->vma_list, obj_link) { + list_for_each_entry(vma, &obj->vma.list, obj_link) { if (!drm_mm_node_allocated(&vma->node)) continue; @@ -627,10 +629,12 @@ static void gen8_display_interrupt_info(struct seq_file *m) for_each_pipe(dev_priv, pipe) { enum intel_display_power_domain power_domain; + intel_wakeref_t wakeref; power_domain = POWER_DOMAIN_PIPE(pipe); - if (!intel_display_power_get_if_enabled(dev_priv, - power_domain)) { + wakeref = intel_display_power_get_if_enabled(dev_priv, + power_domain); + if (!wakeref) { seq_printf(m, "Pipe %c power disabled\n", pipe_name(pipe)); continue; @@ -645,7 +649,7 @@ static void gen8_display_interrupt_info(struct seq_file *m) pipe_name(pipe), I915_READ(GEN8_DE_PIPE_IER(pipe))); - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); } seq_printf(m, "Display Engine port interrupt mask:\t%08x\n", @@ -675,11 +679,14 @@ static int i915_interrupt_info(struct seq_file *m, void *data) struct drm_i915_private *dev_priv = node_to_i915(m->private); struct intel_engine_cs *engine; enum intel_engine_id id; + intel_wakeref_t wakeref; int i, pipe; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); if (IS_CHERRYVIEW(dev_priv)) { + intel_wakeref_t pref; + seq_printf(m, "Master Interrupt Control:\t%08x\n", I915_READ(GEN8_MASTER_IRQ)); @@ -695,8 +702,9 @@ static int i915_interrupt_info(struct seq_file *m, void *data) enum intel_display_power_domain power_domain; power_domain = POWER_DOMAIN_PIPE(pipe); - if (!intel_display_power_get_if_enabled(dev_priv, - power_domain)) { + pref = intel_display_power_get_if_enabled(dev_priv, + power_domain); + if (!pref) { seq_printf(m, "Pipe %c power disabled\n", pipe_name(pipe)); continue; @@ -706,17 +714,17 @@ static int i915_interrupt_info(struct seq_file *m, void *data) pipe_name(pipe), I915_READ(PIPESTAT(pipe))); - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, pref); } - intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); + pref = intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); seq_printf(m, "Port hotplug:\t%08x\n", I915_READ(PORT_HOTPLUG_EN)); seq_printf(m, "DPFLIPSTAT:\t%08x\n", I915_READ(VLV_DPFLIPSTAT)); seq_printf(m, "DPINVGTT:\t%08x\n", I915_READ(DPINVGTT)); - intel_display_power_put(dev_priv, POWER_DOMAIN_INIT); + intel_display_power_put(dev_priv, POWER_DOMAIN_INIT, pref); for (i = 0; i < 4; i++) { seq_printf(m, "GT Interrupt IMR %d:\t%08x\n", @@ -779,10 +787,12 @@ static int i915_interrupt_info(struct seq_file *m, void *data) I915_READ(VLV_IMR)); for_each_pipe(dev_priv, pipe) { enum intel_display_power_domain power_domain; + intel_wakeref_t pref; power_domain = POWER_DOMAIN_PIPE(pipe); - if (!intel_display_power_get_if_enabled(dev_priv, - power_domain)) { + pref = intel_display_power_get_if_enabled(dev_priv, + power_domain); + if (!pref) { seq_printf(m, "Pipe %c power disabled\n", pipe_name(pipe)); continue; @@ -791,7 +801,7 @@ static int i915_interrupt_info(struct seq_file *m, void *data) seq_printf(m, "Pipe %c stat:\t%08x\n", pipe_name(pipe), I915_READ(PIPESTAT(pipe))); - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, pref); } seq_printf(m, "Master IER:\t%08x\n", @@ -878,7 +888,7 @@ static int i915_interrupt_info(struct seq_file *m, void *data) } } - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return 0; } @@ -951,10 +961,11 @@ static int i915_gpu_info_open(struct inode *inode, struct file *file) { struct drm_i915_private *i915 = inode->i_private; struct i915_gpu_state *gpu; + intel_wakeref_t wakeref; - intel_runtime_pm_get(i915); - gpu = i915_capture_gpu_state(i915); - intel_runtime_pm_put(i915); + gpu = NULL; + with_intel_runtime_pm(i915, wakeref) + gpu = i915_capture_gpu_state(i915); if (IS_ERR(gpu)) return PTR_ERR(gpu); @@ -1013,9 +1024,10 @@ static int i915_frequency_info(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); struct intel_rps *rps = &dev_priv->gt_pm.rps; + intel_wakeref_t wakeref; int ret = 0; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); if (IS_GEN(dev_priv, 5)) { u16 rgvswctl = I915_READ16(MEMSWCTL); @@ -1227,7 +1239,7 @@ static int i915_frequency_info(struct seq_file *m, void *unused) seq_printf(m, "Max CD clock frequency: %d kHz\n", dev_priv->max_cdclk_freq); seq_printf(m, "Max pixel clock frequency: %d kHz\n", dev_priv->max_dotclk_freq); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return ret; } @@ -1266,14 +1278,13 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused) u64 acthd[I915_NUM_ENGINES]; u32 seqno[I915_NUM_ENGINES]; struct intel_instdone instdone; + intel_wakeref_t wakeref; enum intel_engine_id id; if (test_bit(I915_WEDGED, &dev_priv->gpu_error.flags)) seq_puts(m, "Wedged\n"); if (test_bit(I915_RESET_BACKOFF, &dev_priv->gpu_error.flags)) seq_puts(m, "Reset in progress: struct_mutex backoff\n"); - if (test_bit(I915_RESET_HANDOFF, &dev_priv->gpu_error.flags)) - seq_puts(m, "Reset in progress: reset handoff to waiter\n"); if (waitqueue_active(&dev_priv->gpu_error.wait_queue)) seq_puts(m, "Waiter holding struct mutex\n"); if (waitqueue_active(&dev_priv->gpu_error.reset_queue)) @@ -1284,17 +1295,15 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused) return 0; } - intel_runtime_pm_get(dev_priv); + with_intel_runtime_pm(dev_priv, wakeref) { + for_each_engine(engine, dev_priv, id) { + acthd[id] = intel_engine_get_active_head(engine); + seqno[id] = intel_engine_get_seqno(engine); + } - for_each_engine(engine, dev_priv, id) { - acthd[id] = intel_engine_get_active_head(engine); - seqno[id] = intel_engine_get_seqno(engine); + intel_engine_get_instdone(dev_priv->engine[RCS], &instdone); } - intel_engine_get_instdone(dev_priv->engine[RCS], &instdone); - - intel_runtime_pm_put(dev_priv); - if (timer_pending(&dev_priv->gpu_error.hangcheck_work.timer)) seq_printf(m, "Hangcheck active, timer fires in %dms\n", jiffies_to_msecs(dev_priv->gpu_error.hangcheck_work.timer.expires - @@ -1307,37 +1316,16 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused) seq_printf(m, "GT active? %s\n", yesno(dev_priv->gt.awake)); for_each_engine(engine, dev_priv, id) { - struct intel_breadcrumbs *b = &engine->breadcrumbs; - struct rb_node *rb; - seq_printf(m, "%s:\n", engine->name); - seq_printf(m, "\tseqno = %x [current %x, last %x]\n", + seq_printf(m, "\tseqno = %x [current %x, last %x], %dms ago\n", engine->hangcheck.seqno, seqno[id], - intel_engine_last_submit(engine)); - seq_printf(m, "\twaiters? %s, fake irq active? %s, stalled? %s, wedged? %s\n", - yesno(intel_engine_has_waiter(engine)), - yesno(test_bit(engine->id, - &dev_priv->gpu_error.missed_irq_rings)), - yesno(engine->hangcheck.stalled), - yesno(engine->hangcheck.wedged)); - - spin_lock_irq(&b->rb_lock); - for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) { - struct intel_wait *w = rb_entry(rb, typeof(*w), node); - - seq_printf(m, "\t%s [%d] waiting for %x\n", - w->tsk->comm, w->tsk->pid, w->seqno); - } - spin_unlock_irq(&b->rb_lock); + intel_engine_last_submit(engine), + jiffies_to_msecs(jiffies - + engine->hangcheck.action_timestamp)); seq_printf(m, "\tACTHD = 0x%08llx [current 0x%08llx]\n", (long long)engine->hangcheck.acthd, (long long)acthd[id]); - seq_printf(m, "\taction = %s(%d) %d ms ago\n", - hangcheck_action_to_str(engine->hangcheck.action), - engine->hangcheck.action, - jiffies_to_msecs(jiffies - - engine->hangcheck.action_timestamp)); if (engine->id == RCS) { seq_puts(m, "\tinstdone read =\n"); @@ -1569,18 +1557,17 @@ static int gen6_drpc_info(struct seq_file *m) static int i915_drpc_info(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); - int err; - - intel_runtime_pm_get(dev_priv); - - if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) - err = vlv_drpc_info(m); - else if (INTEL_GEN(dev_priv) >= 6) - err = gen6_drpc_info(m); - else - err = ironlake_drpc_info(m); - - intel_runtime_pm_put(dev_priv); + intel_wakeref_t wakeref; + int err = -ENODEV; + + with_intel_runtime_pm(dev_priv, wakeref) { + if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + err = vlv_drpc_info(m); + else if (INTEL_GEN(dev_priv) >= 6) + err = gen6_drpc_info(m); + else + err = ironlake_drpc_info(m); + } return err; } @@ -1602,11 +1589,12 @@ static int i915_fbc_status(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); struct intel_fbc *fbc = &dev_priv->fbc; + intel_wakeref_t wakeref; if (!HAS_FBC(dev_priv)) return -ENODEV; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); mutex_lock(&fbc->lock); if (intel_fbc_is_active(dev_priv)) @@ -1633,7 +1621,7 @@ static int i915_fbc_status(struct seq_file *m, void *unused) } mutex_unlock(&fbc->lock); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return 0; } @@ -1678,11 +1666,12 @@ DEFINE_SIMPLE_ATTRIBUTE(i915_fbc_false_color_fops, static int i915_ips_status(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); + intel_wakeref_t wakeref; if (!HAS_IPS(dev_priv)) return -ENODEV; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); seq_printf(m, "Enabled by kernel parameter: %s\n", yesno(i915_modparams.enable_ips)); @@ -1696,7 +1685,7 @@ static int i915_ips_status(struct seq_file *m, void *unused) seq_puts(m, "Currently: disabled\n"); } - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return 0; } @@ -1704,10 +1693,10 @@ static int i915_ips_status(struct seq_file *m, void *unused) static int i915_sr_status(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); + intel_wakeref_t wakeref; bool sr_enabled = false; - intel_runtime_pm_get(dev_priv); - intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); + wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); if (INTEL_GEN(dev_priv) >= 9) /* no global SR status; inspect per-plane WM */; @@ -1723,8 +1712,7 @@ static int i915_sr_status(struct seq_file *m, void *unused) else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) sr_enabled = I915_READ(FW_BLC_SELF_VLV) & FW_CSPWRDWNEN; - intel_display_power_put(dev_priv, POWER_DOMAIN_INIT); - intel_runtime_pm_put(dev_priv); + intel_display_power_put(dev_priv, POWER_DOMAIN_INIT, wakeref); seq_printf(m, "self-refresh: %s\n", enableddisabled(sr_enabled)); @@ -1733,31 +1721,24 @@ static int i915_sr_status(struct seq_file *m, void *unused) static int i915_emon_status(struct seq_file *m, void *unused) { - struct drm_i915_private *dev_priv = node_to_i915(m->private); - struct drm_device *dev = &dev_priv->drm; - unsigned long temp, chipset, gfx; - int ret; + struct drm_i915_private *i915 = node_to_i915(m->private); + intel_wakeref_t wakeref; - if (!IS_GEN(dev_priv, 5)) + if (!IS_GEN(i915, 5)) return -ENODEV; - intel_runtime_pm_get(dev_priv); - - ret = mutex_lock_interruptible(&dev->struct_mutex); - if (ret) - return ret; - - temp = i915_mch_val(dev_priv); - chipset = i915_chipset_val(dev_priv); - gfx = i915_gfx_val(dev_priv); - mutex_unlock(&dev->struct_mutex); + with_intel_runtime_pm(i915, wakeref) { + unsigned long temp, chipset, gfx; - seq_printf(m, "GMCH temp: %ld\n", temp); - seq_printf(m, "Chipset power: %ld\n", chipset); - seq_printf(m, "GFX power: %ld\n", gfx); - seq_printf(m, "Total power: %ld\n", chipset + gfx); + temp = i915_mch_val(i915); + chipset = i915_chipset_val(i915); + gfx = i915_gfx_val(i915); - intel_runtime_pm_put(dev_priv); + seq_printf(m, "GMCH temp: %ld\n", temp); + seq_printf(m, "Chipset power: %ld\n", chipset); + seq_printf(m, "GFX power: %ld\n", gfx); + seq_printf(m, "Total power: %ld\n", chipset + gfx); + } return 0; } @@ -1767,13 +1748,14 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused) struct drm_i915_private *dev_priv = node_to_i915(m->private); struct intel_rps *rps = &dev_priv->gt_pm.rps; unsigned int max_gpu_freq, min_gpu_freq; + intel_wakeref_t wakeref; int gpu_freq, ia_freq; int ret; if (!HAS_LLC(dev_priv)) return -ENODEV; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); ret = mutex_lock_interruptible(&dev_priv->pcu_lock); if (ret) @@ -1806,7 +1788,7 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused) mutex_unlock(&dev_priv->pcu_lock); out: - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return ret; } @@ -1979,8 +1961,9 @@ static const char *swizzle_string(unsigned swizzle) static int i915_swizzle_info(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = node_to_i915(m->private); + intel_wakeref_t wakeref; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); seq_printf(m, "bit6 swizzle for X-tiling = %s\n", swizzle_string(dev_priv->mm.bit_6_swizzle_x)); @@ -2018,23 +2001,11 @@ static int i915_swizzle_info(struct seq_file *m, void *data) if (dev_priv->quirks & QUIRK_PIN_SWIZZLED_PAGES) seq_puts(m, "L-shaped memory detected\n"); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return 0; } -static int count_irq_waiters(struct drm_i915_private *i915) -{ - struct intel_engine_cs *engine; - enum intel_engine_id id; - int count = 0; - - for_each_engine(engine, i915, id) - count += intel_engine_has_waiter(engine); - - return count; -} - static const char *rps_power_to_str(unsigned int power) { static const char * const strings[] = { @@ -2055,9 +2026,10 @@ static int i915_rps_boost_info(struct seq_file *m, void *data) struct drm_device *dev = &dev_priv->drm; struct intel_rps *rps = &dev_priv->gt_pm.rps; u32 act_freq = rps->cur_freq; + intel_wakeref_t wakeref; struct drm_file *file; - if (intel_runtime_pm_get_if_in_use(dev_priv)) { + with_intel_runtime_pm_if_in_use(dev_priv, wakeref) { if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { mutex_lock(&dev_priv->pcu_lock); act_freq = vlv_punit_read(dev_priv, @@ -2068,13 +2040,11 @@ static int i915_rps_boost_info(struct seq_file *m, void *data) act_freq = intel_get_cagf(dev_priv, I915_READ(GEN6_RPSTAT1)); } - intel_runtime_pm_put(dev_priv); } seq_printf(m, "RPS enabled? %d\n", rps->enabled); seq_printf(m, "GPU busy? %s [%d requests]\n", yesno(dev_priv->gt.awake), dev_priv->gt.active_requests); - seq_printf(m, "CPU waiting? %d\n", count_irq_waiters(dev_priv)); seq_printf(m, "Boosts outstanding? %d\n", atomic_read(&rps->num_waiters)); seq_printf(m, "Interactive? %d\n", READ_ONCE(rps->power.interactive)); @@ -2151,6 +2121,7 @@ static int i915_llc(struct seq_file *m, void *data) static int i915_huc_load_status_info(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = node_to_i915(m->private); + intel_wakeref_t wakeref; struct drm_printer p; if (!HAS_HUC(dev_priv)) @@ -2159,9 +2130,8 @@ static int i915_huc_load_status_info(struct seq_file *m, void *data) p = drm_seq_file_printer(m); intel_uc_fw_dump(&dev_priv->huc.fw, &p); - intel_runtime_pm_get(dev_priv); - seq_printf(m, "\nHuC status 0x%08x:\n", I915_READ(HUC_STATUS2)); - intel_runtime_pm_put(dev_priv); + with_intel_runtime_pm(dev_priv, wakeref) + seq_printf(m, "\nHuC status 0x%08x:\n", I915_READ(HUC_STATUS2)); return 0; } @@ -2169,8 +2139,8 @@ static int i915_huc_load_status_info(struct seq_file *m, void *data) static int i915_guc_load_status_info(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = node_to_i915(m->private); + intel_wakeref_t wakeref; struct drm_printer p; - u32 tmp, i; if (!HAS_GUC(dev_priv)) return -ENODEV; @@ -2178,22 +2148,23 @@ static int i915_guc_load_status_info(struct seq_file *m, void *data) p = drm_seq_file_printer(m); intel_uc_fw_dump(&dev_priv->guc.fw, &p); - intel_runtime_pm_get(dev_priv); - - tmp = I915_READ(GUC_STATUS); - - seq_printf(m, "\nGuC status 0x%08x:\n", tmp); - seq_printf(m, "\tBootrom status = 0x%x\n", - (tmp & GS_BOOTROM_MASK) >> GS_BOOTROM_SHIFT); - seq_printf(m, "\tuKernel status = 0x%x\n", - (tmp & GS_UKERNEL_MASK) >> GS_UKERNEL_SHIFT); - seq_printf(m, "\tMIA Core status = 0x%x\n", - (tmp & GS_MIA_MASK) >> GS_MIA_SHIFT); - seq_puts(m, "\nScratch registers:\n"); - for (i = 0; i < 16; i++) - seq_printf(m, "\t%2d: \t0x%x\n", i, I915_READ(SOFT_SCRATCH(i))); - - intel_runtime_pm_put(dev_priv); + with_intel_runtime_pm(dev_priv, wakeref) { + u32 tmp = I915_READ(GUC_STATUS); + u32 i; + + seq_printf(m, "\nGuC status 0x%08x:\n", tmp); + seq_printf(m, "\tBootrom status = 0x%x\n", + (tmp & GS_BOOTROM_MASK) >> GS_BOOTROM_SHIFT); + seq_printf(m, "\tuKernel status = 0x%x\n", + (tmp & GS_UKERNEL_MASK) >> GS_UKERNEL_SHIFT); + seq_printf(m, "\tMIA Core status = 0x%x\n", + (tmp & GS_MIA_MASK) >> GS_MIA_SHIFT); + seq_puts(m, "\nScratch registers:\n"); + for (i = 0; i < 16; i++) { + seq_printf(m, "\t%2d: \t0x%x\n", + i, I915_READ(SOFT_SCRATCH(i))); + } + } return 0; } @@ -2245,7 +2216,7 @@ static void i915_guc_client_info(struct seq_file *m, { struct intel_engine_cs *engine; enum intel_engine_id id; - uint64_t tot = 0; + u64 tot = 0; seq_printf(m, "\tPriority %d, GuC stage index: %u, PD offset 0x%x\n", client->priority, client->stage_id, client->proc_desc_offset); @@ -2500,7 +2471,8 @@ DEFINE_SHOW_ATTRIBUTE(i915_psr_sink_status); static void psr_source_status(struct drm_i915_private *dev_priv, struct seq_file *m) { - u32 val, psr_status; + u32 val, status_val; + const char *status = "unknown"; if (dev_priv->psr.psr2_enabled) { static const char * const live_status[] = { @@ -2516,14 +2488,11 @@ psr_source_status(struct drm_i915_private *dev_priv, struct seq_file *m) "BUF_ON", "TG_ON" }; - psr_status = I915_READ(EDP_PSR2_STATUS); - val = (psr_status & EDP_PSR2_STATUS_STATE_MASK) >> - EDP_PSR2_STATUS_STATE_SHIFT; - if (val < ARRAY_SIZE(live_status)) { - seq_printf(m, "Source PSR status: 0x%x [%s]\n", - psr_status, live_status[val]); - return; - } + val = I915_READ(EDP_PSR2_STATUS); + status_val = (val & EDP_PSR2_STATUS_STATE_MASK) >> + EDP_PSR2_STATUS_STATE_SHIFT; + if (status_val < ARRAY_SIZE(live_status)) + status = live_status[status_val]; } else { static const char * const live_status[] = { "IDLE", @@ -2535,74 +2504,102 @@ psr_source_status(struct drm_i915_private *dev_priv, struct seq_file *m) "SRDOFFACK", "SRDENT_ON", }; - psr_status = I915_READ(EDP_PSR_STATUS); - val = (psr_status & EDP_PSR_STATUS_STATE_MASK) >> - EDP_PSR_STATUS_STATE_SHIFT; - if (val < ARRAY_SIZE(live_status)) { - seq_printf(m, "Source PSR status: 0x%x [%s]\n", - psr_status, live_status[val]); - return; - } + val = I915_READ(EDP_PSR_STATUS); + status_val = (val & EDP_PSR_STATUS_STATE_MASK) >> + EDP_PSR_STATUS_STATE_SHIFT; + if (status_val < ARRAY_SIZE(live_status)) + status = live_status[status_val]; } - seq_printf(m, "Source PSR status: 0x%x [%s]\n", psr_status, "unknown"); + seq_printf(m, "Source PSR status: %s [0x%08x]\n", status, val); } static int i915_edp_psr_status(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = node_to_i915(m->private); - u32 psrperf = 0; - bool enabled = false; - bool sink_support; + struct i915_psr *psr = &dev_priv->psr; + intel_wakeref_t wakeref; + const char *status; + bool enabled; + u32 val; if (!HAS_PSR(dev_priv)) return -ENODEV; - sink_support = dev_priv->psr.sink_support; - seq_printf(m, "Sink_Support: %s\n", yesno(sink_support)); - if (!sink_support) - return 0; + seq_printf(m, "Sink support: %s", yesno(psr->sink_support)); + if (psr->dp) + seq_printf(m, " [0x%02x]", psr->dp->psr_dpcd[0]); + seq_puts(m, "\n"); - intel_runtime_pm_get(dev_priv); + if (!psr->sink_support) + return 0; - mutex_lock(&dev_priv->psr.lock); - seq_printf(m, "PSR mode: %s\n", - dev_priv->psr.psr2_enabled ? "PSR2" : "PSR1"); - seq_printf(m, "Enabled: %s\n", yesno(dev_priv->psr.enabled)); - seq_printf(m, "Busy frontbuffer bits: 0x%03x\n", - dev_priv->psr.busy_frontbuffer_bits); + wakeref = intel_runtime_pm_get(dev_priv); + mutex_lock(&psr->lock); - if (dev_priv->psr.psr2_enabled) - enabled = I915_READ(EDP_PSR2_CTL) & EDP_PSR2_ENABLE; + if (psr->enabled) + status = psr->psr2_enabled ? "PSR2 enabled" : "PSR1 enabled"; else - enabled = I915_READ(EDP_PSR_CTL) & EDP_PSR_ENABLE; + status = "disabled"; + seq_printf(m, "PSR mode: %s\n", status); - seq_printf(m, "Main link in standby mode: %s\n", - yesno(dev_priv->psr.link_standby)); + if (!psr->enabled) + goto unlock; - seq_printf(m, "HW Enabled & Active bit: %s\n", yesno(enabled)); + if (psr->psr2_enabled) { + val = I915_READ(EDP_PSR2_CTL); + enabled = val & EDP_PSR2_ENABLE; + } else { + val = I915_READ(EDP_PSR_CTL); + enabled = val & EDP_PSR_ENABLE; + } + seq_printf(m, "Source PSR ctl: %s [0x%08x]\n", + enableddisabled(enabled), val); + psr_source_status(dev_priv, m); + seq_printf(m, "Busy frontbuffer bits: 0x%08x\n", + psr->busy_frontbuffer_bits); /* * SKL+ Perf counter is reset to 0 everytime DC state is entered */ if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) { - psrperf = I915_READ(EDP_PSR_PERF_CNT) & - EDP_PSR_PERF_CNT_MASK; + val = I915_READ(EDP_PSR_PERF_CNT) & EDP_PSR_PERF_CNT_MASK; + seq_printf(m, "Performance counter: %u\n", val); + } - seq_printf(m, "Performance_Counter: %u\n", psrperf); + if (psr->debug & I915_PSR_DEBUG_IRQ) { + seq_printf(m, "Last attempted entry at: %lld\n", + psr->last_entry_attempt); + seq_printf(m, "Last exit at: %lld\n", psr->last_exit); } - psr_source_status(dev_priv, m); - mutex_unlock(&dev_priv->psr.lock); + if (psr->psr2_enabled) { + u32 su_frames_val[3]; + int frame; - if (READ_ONCE(dev_priv->psr.debug) & I915_PSR_DEBUG_IRQ) { - seq_printf(m, "Last attempted entry at: %lld\n", - dev_priv->psr.last_entry_attempt); - seq_printf(m, "Last exit at: %lld\n", - dev_priv->psr.last_exit); + /* + * Reading all 3 registers before hand to minimize crossing a + * frame boundary between register reads + */ + for (frame = 0; frame < PSR2_SU_STATUS_FRAMES; frame += 3) + su_frames_val[frame / 3] = I915_READ(PSR2_SU_STATUS(frame)); + + seq_puts(m, "Frame:\tPSR2 SU blocks:\n"); + + for (frame = 0; frame < PSR2_SU_STATUS_FRAMES; frame++) { + u32 su_blocks; + + su_blocks = su_frames_val[frame / 3] & + PSR2_SU_STATUS_MASK(frame); + su_blocks = su_blocks >> PSR2_SU_STATUS_SHIFT(frame); + seq_printf(m, "%d\t%d\n", frame, su_blocks); + } } - intel_runtime_pm_put(dev_priv); +unlock: + mutex_unlock(&psr->lock); + intel_runtime_pm_put(dev_priv, wakeref); + return 0; } @@ -2611,6 +2608,7 @@ i915_edp_psr_debug_set(void *data, u64 val) { struct drm_i915_private *dev_priv = data; struct drm_modeset_acquire_ctx ctx; + intel_wakeref_t wakeref; int ret; if (!CAN_PSR(dev_priv)) @@ -2618,7 +2616,7 @@ i915_edp_psr_debug_set(void *data, u64 val) DRM_DEBUG_KMS("Setting PSR debug to %llx\n", val); - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); drm_modeset_acquire_init(&ctx, DRM_MODESET_ACQUIRE_INTERRUPTIBLE); @@ -2633,7 +2631,7 @@ retry: drm_modeset_drop_locks(&ctx); drm_modeset_acquire_fini(&ctx); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return ret; } @@ -2658,24 +2656,20 @@ static int i915_energy_uJ(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = node_to_i915(m->private); unsigned long long power; + intel_wakeref_t wakeref; u32 units; if (INTEL_GEN(dev_priv) < 6) return -ENODEV; - intel_runtime_pm_get(dev_priv); - - if (rdmsrl_safe(MSR_RAPL_POWER_UNIT, &power)) { - intel_runtime_pm_put(dev_priv); + if (rdmsrl_safe(MSR_RAPL_POWER_UNIT, &power)) return -ENODEV; - } units = (power & 0x1f00) >> 8; - power = I915_READ(MCH_SECP_NRG_STTS); - power = (1000000 * power) >> units; /* convert to uJ */ - - intel_runtime_pm_put(dev_priv); + with_intel_runtime_pm(dev_priv, wakeref) + power = I915_READ(MCH_SECP_NRG_STTS); + power = (1000000 * power) >> units; /* convert to uJ */ seq_printf(m, "%llu", power); return 0; @@ -2689,6 +2683,9 @@ static int i915_runtime_pm_status(struct seq_file *m, void *unused) if (!HAS_RUNTIME_PM(dev_priv)) seq_puts(m, "Runtime power management not supported\n"); + seq_printf(m, "Runtime power status: %s\n", + enableddisabled(!dev_priv->power_domains.wakeref)); + seq_printf(m, "GPU idle: %s (epoch %u)\n", yesno(!dev_priv->gt.awake), dev_priv->gt.epoch); seq_printf(m, "IRQs disabled: %s\n", @@ -2703,6 +2700,12 @@ static int i915_runtime_pm_status(struct seq_file *m, void *unused) pci_power_name(pdev->current_state), pdev->current_state); + if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM)) { + struct drm_printer p = drm_seq_file_printer(m); + + print_intel_runtime_pm_wakeref(dev_priv, &p); + } + return 0; } @@ -2737,6 +2740,7 @@ static int i915_power_domain_info(struct seq_file *m, void *unused) static int i915_dmc_info(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); + intel_wakeref_t wakeref; struct intel_csr *csr; if (!HAS_CSR(dev_priv)) @@ -2744,7 +2748,7 @@ static int i915_dmc_info(struct seq_file *m, void *unused) csr = &dev_priv->csr; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); seq_printf(m, "fw loaded: %s\n", yesno(csr->dmc_payload != NULL)); seq_printf(m, "path: %s\n", csr->fw_path); @@ -2770,7 +2774,7 @@ out: seq_printf(m, "ssp base: 0x%08x\n", I915_READ(CSR_SSP_BASE)); seq_printf(m, "htp: 0x%08x\n", I915_READ(CSR_HTP_SKL)); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return 0; } @@ -3053,8 +3057,10 @@ static int i915_display_info(struct seq_file *m, void *unused) struct intel_crtc *crtc; struct drm_connector *connector; struct drm_connector_list_iter conn_iter; + intel_wakeref_t wakeref; + + wakeref = intel_runtime_pm_get(dev_priv); - intel_runtime_pm_get(dev_priv); seq_printf(m, "CRTC info\n"); seq_printf(m, "---------\n"); for_each_intel_crtc(dev, crtc) { @@ -3102,7 +3108,7 @@ static int i915_display_info(struct seq_file *m, void *unused) drm_connector_list_iter_end(&conn_iter); mutex_unlock(&dev->mode_config.mutex); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return 0; } @@ -3111,10 +3117,11 @@ static int i915_engine_info(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); struct intel_engine_cs *engine; + intel_wakeref_t wakeref; enum intel_engine_id id; struct drm_printer p; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); seq_printf(m, "GT awake? %s (epoch %u)\n", yesno(dev_priv->gt.awake), dev_priv->gt.epoch); @@ -3127,7 +3134,7 @@ static int i915_engine_info(struct seq_file *m, void *unused) for_each_engine(engine, dev_priv, id) intel_engine_dump(engine, &p, "%s\n", engine->name); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return 0; } @@ -3240,20 +3247,21 @@ static ssize_t i915_ipc_status_write(struct file *file, const char __user *ubuf, { struct seq_file *m = file->private_data; struct drm_i915_private *dev_priv = m->private; - int ret; + intel_wakeref_t wakeref; bool enable; + int ret; ret = kstrtobool_from_user(ubuf, len, &enable); if (ret < 0) return ret; - intel_runtime_pm_get(dev_priv); - if (!dev_priv->ipc_enabled && enable) - DRM_INFO("Enabling IPC: WM will be proper only after next commit\n"); - dev_priv->wm.distrust_bios_wm = true; - dev_priv->ipc_enabled = enable; - intel_enable_ipc(dev_priv); - intel_runtime_pm_put(dev_priv); + with_intel_runtime_pm(dev_priv, wakeref) { + if (!dev_priv->ipc_enabled && enable) + DRM_INFO("Enabling IPC: WM will be proper only after next commit\n"); + dev_priv->wm.distrust_bios_wm = true; + dev_priv->ipc_enabled = enable; + intel_enable_ipc(dev_priv); + } return len; } @@ -3621,7 +3629,7 @@ static int i915_displayport_test_type_show(struct seq_file *m, void *data) } DEFINE_SHOW_ATTRIBUTE(i915_displayport_test_type); -static void wm_latency_show(struct seq_file *m, const uint16_t wm[8]) +static void wm_latency_show(struct seq_file *m, const u16 wm[8]) { struct drm_i915_private *dev_priv = m->private; struct drm_device *dev = &dev_priv->drm; @@ -3664,7 +3672,7 @@ static void wm_latency_show(struct seq_file *m, const uint16_t wm[8]) static int pri_wm_latency_show(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = m->private; - const uint16_t *latencies; + const u16 *latencies; if (INTEL_GEN(dev_priv) >= 9) latencies = dev_priv->wm.skl_latency; @@ -3679,7 +3687,7 @@ static int pri_wm_latency_show(struct seq_file *m, void *data) static int spr_wm_latency_show(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = m->private; - const uint16_t *latencies; + const u16 *latencies; if (INTEL_GEN(dev_priv) >= 9) latencies = dev_priv->wm.skl_latency; @@ -3694,7 +3702,7 @@ static int spr_wm_latency_show(struct seq_file *m, void *data) static int cur_wm_latency_show(struct seq_file *m, void *data) { struct drm_i915_private *dev_priv = m->private; - const uint16_t *latencies; + const u16 *latencies; if (INTEL_GEN(dev_priv) >= 9) latencies = dev_priv->wm.skl_latency; @@ -3720,7 +3728,7 @@ static int spr_wm_latency_open(struct inode *inode, struct file *file) { struct drm_i915_private *dev_priv = inode->i_private; - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) return -ENODEV; return single_open(file, spr_wm_latency_show, dev_priv); @@ -3730,19 +3738,19 @@ static int cur_wm_latency_open(struct inode *inode, struct file *file) { struct drm_i915_private *dev_priv = inode->i_private; - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) return -ENODEV; return single_open(file, cur_wm_latency_show, dev_priv); } static ssize_t wm_latency_write(struct file *file, const char __user *ubuf, - size_t len, loff_t *offp, uint16_t wm[8]) + size_t len, loff_t *offp, u16 wm[8]) { struct seq_file *m = file->private_data; struct drm_i915_private *dev_priv = m->private; struct drm_device *dev = &dev_priv->drm; - uint16_t new[8] = { 0 }; + u16 new[8] = { 0 }; int num_levels; int level; int ret; @@ -3787,7 +3795,7 @@ static ssize_t pri_wm_latency_write(struct file *file, const char __user *ubuf, { struct seq_file *m = file->private_data; struct drm_i915_private *dev_priv = m->private; - uint16_t *latencies; + u16 *latencies; if (INTEL_GEN(dev_priv) >= 9) latencies = dev_priv->wm.skl_latency; @@ -3802,7 +3810,7 @@ static ssize_t spr_wm_latency_write(struct file *file, const char __user *ubuf, { struct seq_file *m = file->private_data; struct drm_i915_private *dev_priv = m->private; - uint16_t *latencies; + u16 *latencies; if (INTEL_GEN(dev_priv) >= 9) latencies = dev_priv->wm.skl_latency; @@ -3817,7 +3825,7 @@ static ssize_t cur_wm_latency_write(struct file *file, const char __user *ubuf, { struct seq_file *m = file->private_data; struct drm_i915_private *dev_priv = m->private; - uint16_t *latencies; + u16 *latencies; if (INTEL_GEN(dev_priv) >= 9) latencies = dev_priv->wm.skl_latency; @@ -3868,8 +3876,6 @@ static int i915_wedged_set(void *data, u64 val) { struct drm_i915_private *i915 = data; - struct intel_engine_cs *engine; - unsigned int tmp; /* * There is no safeguard against this debugfs entry colliding @@ -3882,18 +3888,8 @@ i915_wedged_set(void *data, u64 val) if (i915_reset_backoff(&i915->gpu_error)) return -EAGAIN; - for_each_engine_masked(engine, i915, val, tmp) { - engine->hangcheck.seqno = intel_engine_get_seqno(engine); - engine->hangcheck.stalled = true; - } - i915_handle_error(i915, val, I915_ERROR_CAPTURE, "Manually set wedged engine mask = %llx", val); - - wait_on_bit(&i915->gpu_error.flags, - I915_RESET_HANDOFF, - TASK_UNINTERRUPTIBLE); - return 0; } @@ -3901,94 +3897,6 @@ DEFINE_SIMPLE_ATTRIBUTE(i915_wedged_fops, i915_wedged_get, i915_wedged_set, "%llu\n"); -static int -fault_irq_set(struct drm_i915_private *i915, - unsigned long *irq, - unsigned long val) -{ - int err; - - err = mutex_lock_interruptible(&i915->drm.struct_mutex); - if (err) - return err; - - err = i915_gem_wait_for_idle(i915, - I915_WAIT_LOCKED | - I915_WAIT_INTERRUPTIBLE, - MAX_SCHEDULE_TIMEOUT); - if (err) - goto err_unlock; - - *irq = val; - mutex_unlock(&i915->drm.struct_mutex); - - /* Flush idle worker to disarm irq */ - drain_delayed_work(&i915->gt.idle_work); - - return 0; - -err_unlock: - mutex_unlock(&i915->drm.struct_mutex); - return err; -} - -static int -i915_ring_missed_irq_get(void *data, u64 *val) -{ - struct drm_i915_private *dev_priv = data; - - *val = dev_priv->gpu_error.missed_irq_rings; - return 0; -} - -static int -i915_ring_missed_irq_set(void *data, u64 val) -{ - struct drm_i915_private *i915 = data; - - return fault_irq_set(i915, &i915->gpu_error.missed_irq_rings, val); -} - -DEFINE_SIMPLE_ATTRIBUTE(i915_ring_missed_irq_fops, - i915_ring_missed_irq_get, i915_ring_missed_irq_set, - "0x%08llx\n"); - -static int -i915_ring_test_irq_get(void *data, u64 *val) -{ - struct drm_i915_private *dev_priv = data; - - *val = dev_priv->gpu_error.test_irq_rings; - - return 0; -} - -static int -i915_ring_test_irq_set(void *data, u64 val) -{ - struct drm_i915_private *i915 = data; - - /* GuC keeps the user interrupt permanently enabled for submission */ - if (USES_GUC_SUBMISSION(i915)) - return -ENODEV; - - /* - * From icl, we can no longer individually mask interrupt generation - * from each engine. - */ - if (INTEL_GEN(i915) >= 11) - return -ENODEV; - - val &= INTEL_INFO(i915)->ring_mask; - DRM_DEBUG_DRIVER("Masking interrupts on rings 0x%08llx\n", val); - - return fault_irq_set(i915, &i915->gpu_error.test_irq_rings, val); -} - -DEFINE_SIMPLE_ATTRIBUTE(i915_ring_test_irq_fops, - i915_ring_test_irq_get, i915_ring_test_irq_set, - "0x%08llx\n"); - #define DROP_UNBOUND BIT(0) #define DROP_BOUND BIT(1) #define DROP_RETIRE BIT(2) @@ -4019,13 +3927,15 @@ static int i915_drop_caches_set(void *data, u64 val) { struct drm_i915_private *i915 = data; + intel_wakeref_t wakeref; int ret = 0; DRM_DEBUG("Dropping caches: 0x%08llx [0x%08llx]\n", val, val & DROP_ALL); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); - if (val & DROP_RESET_ACTIVE && !intel_engines_are_idle(i915)) + if (val & DROP_RESET_ACTIVE && + wait_for(intel_engines_are_idle(i915), I915_IDLE_ENGINES_TIMEOUT)) i915_gem_set_wedged(i915); /* No need to check and wait for gpu resets, only libdrm auto-restarts @@ -4047,13 +3957,8 @@ i915_drop_caches_set(void *data, u64 val) mutex_unlock(&i915->drm.struct_mutex); } - if (val & DROP_RESET_ACTIVE && - i915_terminally_wedged(&i915->gpu_error)) { + if (val & DROP_RESET_ACTIVE && i915_terminally_wedged(&i915->gpu_error)) i915_handle_error(i915, ALL_ENGINES, 0, NULL); - wait_on_bit(&i915->gpu_error.flags, - I915_RESET_HANDOFF, - TASK_UNINTERRUPTIBLE); - } fs_reclaim_acquire(GFP_KERNEL); if (val & DROP_BOUND) @@ -4078,7 +3983,7 @@ i915_drop_caches_set(void *data, u64 val) i915_gem_drain_freed_objects(i915); out: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); return ret; } @@ -4091,16 +3996,14 @@ static int i915_cache_sharing_get(void *data, u64 *val) { struct drm_i915_private *dev_priv = data; - u32 snpcr; + intel_wakeref_t wakeref; + u32 snpcr = 0; if (!(IS_GEN_RANGE(dev_priv, 6, 7))) return -ENODEV; - intel_runtime_pm_get(dev_priv); - - snpcr = I915_READ(GEN6_MBCUNIT_SNPCR); - - intel_runtime_pm_put(dev_priv); + with_intel_runtime_pm(dev_priv, wakeref) + snpcr = I915_READ(GEN6_MBCUNIT_SNPCR); *val = (snpcr & GEN6_MBC_SNPCR_MASK) >> GEN6_MBC_SNPCR_SHIFT; @@ -4111,7 +4014,7 @@ static int i915_cache_sharing_set(void *data, u64 val) { struct drm_i915_private *dev_priv = data; - u32 snpcr; + intel_wakeref_t wakeref; if (!(IS_GEN_RANGE(dev_priv, 6, 7))) return -ENODEV; @@ -4119,16 +4022,17 @@ i915_cache_sharing_set(void *data, u64 val) if (val > 3) return -EINVAL; - intel_runtime_pm_get(dev_priv); DRM_DEBUG_DRIVER("Manually setting uncore sharing to %llu\n", val); + with_intel_runtime_pm(dev_priv, wakeref) { + u32 snpcr; + + /* Update the cache sharing policy here as well */ + snpcr = I915_READ(GEN6_MBCUNIT_SNPCR); + snpcr &= ~GEN6_MBC_SNPCR_MASK; + snpcr |= val << GEN6_MBC_SNPCR_SHIFT; + I915_WRITE(GEN6_MBCUNIT_SNPCR, snpcr); + } - /* Update the cache sharing policy here as well */ - snpcr = I915_READ(GEN6_MBCUNIT_SNPCR); - snpcr &= ~GEN6_MBC_SNPCR_MASK; - snpcr |= (val << GEN6_MBC_SNPCR_SHIFT); - I915_WRITE(GEN6_MBCUNIT_SNPCR, snpcr); - - intel_runtime_pm_put(dev_priv); return 0; } @@ -4350,6 +4254,7 @@ static int i915_sseu_status(struct seq_file *m, void *unused) { struct drm_i915_private *dev_priv = node_to_i915(m->private); struct sseu_dev_info sseu; + intel_wakeref_t wakeref; if (INTEL_GEN(dev_priv) < 8) return -ENODEV; @@ -4364,20 +4269,17 @@ static int i915_sseu_status(struct seq_file *m, void *unused) sseu.max_eus_per_subslice = RUNTIME_INFO(dev_priv)->sseu.max_eus_per_subslice; - intel_runtime_pm_get(dev_priv); - - if (IS_CHERRYVIEW(dev_priv)) { - cherryview_sseu_device_status(dev_priv, &sseu); - } else if (IS_BROADWELL(dev_priv)) { - broadwell_sseu_device_status(dev_priv, &sseu); - } else if (IS_GEN(dev_priv, 9)) { - gen9_sseu_device_status(dev_priv, &sseu); - } else if (INTEL_GEN(dev_priv) >= 10) { - gen10_sseu_device_status(dev_priv, &sseu); + with_intel_runtime_pm(dev_priv, wakeref) { + if (IS_CHERRYVIEW(dev_priv)) + cherryview_sseu_device_status(dev_priv, &sseu); + else if (IS_BROADWELL(dev_priv)) + broadwell_sseu_device_status(dev_priv, &sseu); + else if (IS_GEN(dev_priv, 9)) + gen9_sseu_device_status(dev_priv, &sseu); + else if (INTEL_GEN(dev_priv) >= 10) + gen10_sseu_device_status(dev_priv, &sseu); } - intel_runtime_pm_put(dev_priv); - i915_print_sseu_info(m, false, &sseu); return 0; @@ -4390,7 +4292,7 @@ static int i915_forcewake_open(struct inode *inode, struct file *file) if (INTEL_GEN(i915) < 6) return 0; - intel_runtime_pm_get(i915); + file->private_data = (void *)(uintptr_t)intel_runtime_pm_get(i915); intel_uncore_forcewake_user_get(i915); return 0; @@ -4404,7 +4306,8 @@ static int i915_forcewake_release(struct inode *inode, struct file *file) return 0; intel_uncore_forcewake_user_put(i915); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, + (intel_wakeref_t)(uintptr_t)file->private_data); return 0; } @@ -4757,8 +4660,6 @@ static const struct i915_debugfs_files { } i915_debugfs_files[] = { {"i915_wedged", &i915_wedged_fops}, {"i915_cache_sharing", &i915_cache_sharing_fops}, - {"i915_ring_missed_irq", &i915_ring_missed_irq_fops}, - {"i915_ring_test_irq", &i915_ring_test_irq_fops}, {"i915_gem_drop_caches", &i915_drop_caches_fops}, #if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR) {"i915_error_state", &i915_error_state_fops}, @@ -4837,7 +4738,7 @@ static int i915_dpcd_show(struct seq_file *m, void *data) struct drm_connector *connector = m->private; struct intel_dp *intel_dp = enc_to_intel_dp(&intel_attached_encoder(connector)->base); - uint8_t buf[16]; + u8 buf[16]; ssize_t err; int i; @@ -4952,9 +4853,8 @@ static int i915_dsc_fec_support_show(struct seq_file *m, void *data) crtc_state = to_intel_crtc_state(crtc->state); seq_printf(m, "DSC_Enabled: %s\n", yesno(crtc_state->dsc_params.compression_enable)); - if (intel_dp->dsc_dpcd) - seq_printf(m, "DSC_Sink_Support: %s\n", - yesno(drm_dp_sink_supports_dsc(intel_dp->dsc_dpcd))); + seq_printf(m, "DSC_Sink_Support: %s\n", + yesno(drm_dp_sink_supports_dsc(intel_dp->dsc_dpcd))); if (!intel_dp_is_edp(intel_dp)) seq_printf(m, "FEC_Sink_Support: %s\n", yesno(drm_dp_sink_supports_fec(intel_dp->fec_capable))); diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index 16fd5ae06997..6630212f2faf 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -50,6 +50,7 @@ #include "i915_drv.h" #include "i915_trace.h" #include "i915_pmu.h" +#include "i915_reset.h" #include "i915_query.h" #include "i915_vgpu.h" #include "intel_drv.h" @@ -907,6 +908,7 @@ static int i915_driver_init_early(struct drm_i915_private *dev_priv) mutex_init(&dev_priv->pps_mutex); i915_memcpy_init_early(dev_priv); + intel_runtime_pm_init_early(dev_priv); ret = i915_workqueues_init(dev_priv); if (ret < 0) @@ -1781,6 +1783,9 @@ void i915_driver_unload(struct drm_device *dev) i915_driver_unregister(dev_priv); + /* Flush any external code that still may be under the RCU lock */ + synchronize_rcu(); + if (i915_gem_suspend(dev_priv)) DRM_ERROR("failed to idle hardware; continuing to unload!\n"); @@ -1809,8 +1814,7 @@ void i915_driver_unload(struct drm_device *dev) i915_driver_cleanup_mmio(dev_priv); enable_rpm_wakeref_asserts(dev_priv); - - WARN_ON(atomic_read(&dev_priv->runtime_pm.wakeref_count)); + intel_runtime_pm_cleanup(dev_priv); } static void i915_driver_release(struct drm_device *dev) @@ -2012,6 +2016,8 @@ static int i915_drm_suspend_late(struct drm_device *dev, bool hibernation) out: enable_rpm_wakeref_asserts(dev_priv); + if (!dev_priv->uncore.user_forcewake.count) + intel_runtime_pm_cleanup(dev_priv); return ret; } @@ -2202,211 +2208,6 @@ static int i915_resume_switcheroo(struct drm_device *dev) return i915_drm_resume(dev); } -/** - * i915_reset - reset chip after a hang - * @i915: #drm_i915_private to reset - * @stalled_mask: mask of the stalled engines with the guilty requests - * @reason: user error message for why we are resetting - * - * Reset the chip. Useful if a hang is detected. Marks the device as wedged - * on failure. - * - * Caller must hold the struct_mutex. - * - * Procedure is fairly simple: - * - reset the chip using the reset reg - * - re-init context state - * - re-init hardware status page - * - re-init ring buffer - * - re-init interrupt state - * - re-init display - */ -void i915_reset(struct drm_i915_private *i915, - unsigned int stalled_mask, - const char *reason) -{ - struct i915_gpu_error *error = &i915->gpu_error; - int ret; - int i; - - GEM_TRACE("flags=%lx\n", error->flags); - - might_sleep(); - lockdep_assert_held(&i915->drm.struct_mutex); - assert_rpm_wakelock_held(i915); - GEM_BUG_ON(!test_bit(I915_RESET_BACKOFF, &error->flags)); - - if (!test_bit(I915_RESET_HANDOFF, &error->flags)) - return; - - /* Clear any previous failed attempts at recovery. Time to try again. */ - if (!i915_gem_unset_wedged(i915)) - goto wakeup; - - if (reason) - dev_notice(i915->drm.dev, "Resetting chip for %s\n", reason); - error->reset_count++; - - ret = i915_gem_reset_prepare(i915); - if (ret) { - dev_err(i915->drm.dev, "GPU recovery failed\n"); - goto taint; - } - - if (!intel_has_gpu_reset(i915)) { - if (i915_modparams.reset) - dev_err(i915->drm.dev, "GPU reset not supported\n"); - else - DRM_DEBUG_DRIVER("GPU reset disabled\n"); - goto error; - } - - for (i = 0; i < 3; i++) { - ret = intel_gpu_reset(i915, ALL_ENGINES); - if (ret == 0) - break; - - msleep(100); - } - if (ret) { - dev_err(i915->drm.dev, "Failed to reset chip\n"); - goto taint; - } - - /* Ok, now get things going again... */ - - /* - * Everything depends on having the GTT running, so we need to start - * there. - */ - ret = i915_ggtt_enable_hw(i915); - if (ret) { - DRM_ERROR("Failed to re-enable GGTT following reset (%d)\n", - ret); - goto error; - } - - i915_gem_reset(i915, stalled_mask); - intel_overlay_reset(i915); - - /* - * Next we need to restore the context, but we don't use those - * yet either... - * - * Ring buffer needs to be re-initialized in the KMS case, or if X - * was running at the time of the reset (i.e. we weren't VT - * switched away). - */ - ret = i915_gem_init_hw(i915); - if (ret) { - DRM_ERROR("Failed to initialise HW following reset (%d)\n", - ret); - goto error; - } - - i915_queue_hangcheck(i915); - -finish: - i915_gem_reset_finish(i915); -wakeup: - clear_bit(I915_RESET_HANDOFF, &error->flags); - wake_up_bit(&error->flags, I915_RESET_HANDOFF); - return; - -taint: - /* - * History tells us that if we cannot reset the GPU now, we - * never will. This then impacts everything that is run - * subsequently. On failing the reset, we mark the driver - * as wedged, preventing further execution on the GPU. - * We also want to go one step further and add a taint to the - * kernel so that any subsequent faults can be traced back to - * this failure. This is important for CI, where if the - * GPU/driver fails we would like to reboot and restart testing - * rather than continue on into oblivion. For everyone else, - * the system should still plod along, but they have been warned! - */ - add_taint(TAINT_WARN, LOCKDEP_STILL_OK); -error: - i915_gem_set_wedged(i915); - i915_retire_requests(i915); - goto finish; -} - -static inline int intel_gt_reset_engine(struct drm_i915_private *dev_priv, - struct intel_engine_cs *engine) -{ - return intel_gpu_reset(dev_priv, intel_engine_flag(engine)); -} - -/** - * i915_reset_engine - reset GPU engine to recover from a hang - * @engine: engine to reset - * @msg: reason for GPU reset; or NULL for no dev_notice() - * - * Reset a specific GPU engine. Useful if a hang is detected. - * Returns zero on successful reset or otherwise an error code. - * - * Procedure is: - * - identifies the request that caused the hang and it is dropped - * - reset engine (which will force the engine to idle) - * - re-init/configure engine - */ -int i915_reset_engine(struct intel_engine_cs *engine, const char *msg) -{ - struct i915_gpu_error *error = &engine->i915->gpu_error; - struct i915_request *active_request; - int ret; - - GEM_TRACE("%s flags=%lx\n", engine->name, error->flags); - GEM_BUG_ON(!test_bit(I915_RESET_ENGINE + engine->id, &error->flags)); - - active_request = i915_gem_reset_prepare_engine(engine); - if (IS_ERR_OR_NULL(active_request)) { - /* Either the previous reset failed, or we pardon the reset. */ - ret = PTR_ERR(active_request); - goto out; - } - - if (msg) - dev_notice(engine->i915->drm.dev, - "Resetting %s for %s\n", engine->name, msg); - error->reset_engine_count[engine->id]++; - - if (!engine->i915->guc.execbuf_client) - ret = intel_gt_reset_engine(engine->i915, engine); - else - ret = intel_guc_reset_engine(&engine->i915->guc, engine); - if (ret) { - /* If we fail here, we expect to fallback to a global reset */ - DRM_DEBUG_DRIVER("%sFailed to reset %s, ret=%d\n", - engine->i915->guc.execbuf_client ? "GuC " : "", - engine->name, ret); - goto out; - } - - /* - * The request that caused the hang is stuck on elsp, we know the - * active request and can drop it, adjust head to skip the offending - * request to resume executing remaining requests in the queue. - */ - i915_gem_reset_engine(engine, active_request, true); - - /* - * The engine and its registers (and workarounds in case of render) - * have been reset to their default values. Follow the init_ring - * process to program RING_MODE, HWSP and re-enable submission. - */ - ret = engine->init_hw(engine); - if (ret) - goto out; - -out: - intel_engine_cancel_stop_cs(engine); - i915_gem_reset_finish_engine(engine); - return ret; -} - static int i915_pm_prepare(struct device *kdev) { struct pci_dev *pdev = to_pci_dev(kdev); @@ -2744,6 +2545,10 @@ static void vlv_restore_gunit_s0ix_state(struct drm_i915_private *dev_priv) static int vlv_wait_for_pw_status(struct drm_i915_private *dev_priv, u32 mask, u32 val) { + i915_reg_t reg = VLV_GTLC_PW_STATUS; + u32 reg_value; + int ret; + /* The HW does not like us polling for PW_STATUS frequently, so * use the sleeping loop rather than risk the busy spin within * intel_wait_for_register(). @@ -2751,8 +2556,12 @@ static int vlv_wait_for_pw_status(struct drm_i915_private *dev_priv, * Transitioning between RC6 states should be at most 2ms (see * valleyview_enable_rps) so use a 3ms timeout. */ - return wait_for((I915_READ_NOTRACE(VLV_GTLC_PW_STATUS) & mask) == val, - 3); + ret = wait_for(((reg_value = I915_READ_NOTRACE(reg)) & mask) == val, 3); + + /* just trace the final value */ + trace_i915_reg_rw(false, reg, reg_value, sizeof(reg_value), true); + + return ret; } int vlv_force_gfx_clock(struct drm_i915_private *dev_priv, bool force_on) @@ -2967,7 +2776,7 @@ static int intel_runtime_suspend(struct device *kdev) } enable_rpm_wakeref_asserts(dev_priv); - WARN_ON_ONCE(atomic_read(&dev_priv->runtime_pm.wakeref_count)); + intel_runtime_pm_cleanup(dev_priv); if (intel_uncore_arm_unclaimed_mmio_detection(dev_priv)) DRM_ERROR("Unclaimed access detected prior to suspending\n"); diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 5df26ccda8a4..9adc7bb9e69c 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -45,6 +45,7 @@ #include <linux/pm_qos.h> #include <linux/reservation.h> #include <linux/shmem_fs.h> +#include <linux/stackdepot.h> #include <drm/intel-gtt.h> #include <drm/drm_legacy.h> /* for struct drm_dma_handle */ @@ -90,8 +91,8 @@ #define DRIVER_NAME "i915" #define DRIVER_DESC "Intel Graphics" -#define DRIVER_DATE "20190110" -#define DRIVER_TIMESTAMP 1547162337 +#define DRIVER_DATE "20190207" +#define DRIVER_TIMESTAMP 1549572331 /* Use I915_STATE_WARN(x) and I915_STATE_WARN_ON() (rather than WARN() and * WARN_ON()) for hw state sanity checks to check for unexpected conditions @@ -130,6 +131,8 @@ bool i915_error_injected(void); __i915_printk(i915, i915_error_injected() ? KERN_DEBUG : KERN_ERR, \ fmt, ##__VA_ARGS__) +typedef depot_stack_handle_t intel_wakeref_t; + enum hpd_pin { HPD_NONE = 0, HPD_TV = HPD_NONE, /* TV is known to be unreliable */ @@ -320,8 +323,20 @@ struct drm_i915_display_funcs { /* display clock increase/decrease */ /* pll clock increase/decrease */ - void (*load_csc_matrix)(struct intel_crtc_state *crtc_state); - void (*load_luts)(struct intel_crtc_state *crtc_state); + /* + * Program double buffered color management registers during + * vblank evasion. The registers should then latch during the + * next vblank start, alongside any other double buffered registers + * involved with the same commit. + */ + void (*color_commit)(const struct intel_crtc_state *crtc_state); + /* + * Load LUTs (and other single buffered color management + * registers). Will (hopefully) be called during the vblank + * following the latching of any double buffered registers + * involved with the same commit. + */ + void (*load_luts)(const struct intel_crtc_state *crtc_state); }; #define CSR_VERSION(major, minor) ((major) << 16 | (minor)) @@ -331,16 +346,17 @@ struct drm_i915_display_funcs { struct intel_csr { struct work_struct work; const char *fw_path; - uint32_t required_version; - uint32_t max_fw_size; /* bytes */ - uint32_t *dmc_payload; - uint32_t dmc_fw_size; /* dwords */ - uint32_t version; - uint32_t mmio_count; + u32 required_version; + u32 max_fw_size; /* bytes */ + u32 *dmc_payload; + u32 dmc_fw_size; /* dwords */ + u32 version; + u32 mmio_count; i915_reg_t mmioaddr[8]; - uint32_t mmiodata[8]; - uint32_t dc_state; - uint32_t allowed_dc_mask; + u32 mmiodata[8]; + u32 dc_state; + u32 allowed_dc_mask; + intel_wakeref_t wakeref; }; enum i915_cache_level { @@ -396,7 +412,7 @@ struct intel_fbc { struct { unsigned int mode_flags; - uint32_t hsw_bdw_pixel_rate; + u32 hsw_bdw_pixel_rate; } crtc; struct { @@ -415,7 +431,7 @@ struct intel_fbc { int y; - uint16_t pixel_blend_mode; + u16 pixel_blend_mode; } plane; struct { @@ -555,7 +571,7 @@ struct i915_suspend_saved_registers { u32 saveSWF0[16]; u32 saveSWF1[16]; u32 saveSWF3[3]; - uint64_t saveFENCE[I915_MAX_NUM_FENCES]; + u64 saveFENCE[I915_MAX_NUM_FENCES]; u32 savePCH_PORT_HOTPLUG; u16 saveGCDGMBUS; }; @@ -818,6 +834,8 @@ struct i915_power_domains { bool display_core_suspended; int power_well_count; + intel_wakeref_t wakeref; + struct mutex lock; int domain_use_count[POWER_DOMAIN_NUM]; struct i915_power_well *power_wells; @@ -900,9 +918,9 @@ struct i915_gem_mm { atomic_t bsd_engine_dispatch_index; /** Bit 6 swizzling required for X tiling */ - uint32_t bit_6_swizzle_x; + u32 bit_6_swizzle_x; /** Bit 6 swizzling required for Y tiling */ - uint32_t bit_6_swizzle_y; + u32 bit_6_swizzle_y; /* accounting, useful for userland debugging */ spinlock_t object_stat_lock; @@ -929,20 +947,20 @@ struct ddi_vbt_port_info { * populate this field. */ #define HDMI_LEVEL_SHIFT_UNKNOWN 0xff - uint8_t hdmi_level_shift; + u8 hdmi_level_shift; - uint8_t supports_dvi:1; - uint8_t supports_hdmi:1; - uint8_t supports_dp:1; - uint8_t supports_edp:1; - uint8_t supports_typec_usb:1; - uint8_t supports_tbt:1; + u8 supports_dvi:1; + u8 supports_hdmi:1; + u8 supports_dp:1; + u8 supports_edp:1; + u8 supports_typec_usb:1; + u8 supports_tbt:1; - uint8_t alternate_aux_channel; - uint8_t alternate_ddc_pin; + u8 alternate_aux_channel; + u8 alternate_ddc_pin; - uint8_t dp_boost_level; - uint8_t hdmi_boost_level; + u8 dp_boost_level; + u8 hdmi_boost_level; int dp_max_link_rate; /* 0 for not limited by VBT */ }; @@ -1033,41 +1051,41 @@ enum intel_ddb_partitioning { struct intel_wm_level { bool enable; - uint32_t pri_val; - uint32_t spr_val; - uint32_t cur_val; - uint32_t fbc_val; + u32 pri_val; + u32 spr_val; + u32 cur_val; + u32 fbc_val; }; struct ilk_wm_values { - uint32_t wm_pipe[3]; - uint32_t wm_lp[3]; - uint32_t wm_lp_spr[3]; - uint32_t wm_linetime[3]; + u32 wm_pipe[3]; + u32 wm_lp[3]; + u32 wm_lp_spr[3]; + u32 wm_linetime[3]; bool enable_fbc_wm; enum intel_ddb_partitioning partitioning; }; struct g4x_pipe_wm { - uint16_t plane[I915_MAX_PLANES]; - uint16_t fbc; + u16 plane[I915_MAX_PLANES]; + u16 fbc; }; struct g4x_sr_wm { - uint16_t plane; - uint16_t cursor; - uint16_t fbc; + u16 plane; + u16 cursor; + u16 fbc; }; struct vlv_wm_ddl_values { - uint8_t plane[I915_MAX_PLANES]; + u8 plane[I915_MAX_PLANES]; }; struct vlv_wm_values { struct g4x_pipe_wm pipe[3]; struct g4x_sr_wm sr; struct vlv_wm_ddl_values ddl[3]; - uint8_t level; + u8 level; bool cxsr; }; @@ -1081,10 +1099,10 @@ struct g4x_wm_values { }; struct skl_ddb_entry { - uint16_t start, end; /* in number of blocks, 'end' is exclusive */ + u16 start, end; /* in number of blocks, 'end' is exclusive */ }; -static inline uint16_t skl_ddb_entry_size(const struct skl_ddb_entry *entry) +static inline u16 skl_ddb_entry_size(const struct skl_ddb_entry *entry) { return entry->end - entry->start; } @@ -1108,8 +1126,9 @@ struct skl_ddb_values { }; struct skl_wm_level { - uint16_t plane_res_b; - uint8_t plane_res_l; + u16 min_ddb_alloc; + u16 plane_res_b; + u8 plane_res_l; bool plane_en; }; @@ -1118,15 +1137,15 @@ struct skl_wm_params { bool x_tiled, y_tiled; bool rc_surface; bool is_planar; - uint32_t width; - uint8_t cpp; - uint32_t plane_pixel_rate; - uint32_t y_min_scanlines; - uint32_t plane_bytes_per_line; + u32 width; + u8 cpp; + u32 plane_pixel_rate; + u32 y_min_scanlines; + u32 plane_bytes_per_line; uint_fixed_16_16_t plane_blocks_per_line; uint_fixed_16_16_t y_tile_minimum; - uint32_t linetime_us; - uint32_t dbuf_block_size; + u32 linetime_us; + u32 dbuf_block_size; }; /* @@ -1156,6 +1175,25 @@ struct i915_runtime_pm { atomic_t wakeref_count; bool suspended; bool irqs_enabled; + +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) + /* + * To aide detection of wakeref leaks and general misuse, we + * track all wakeref holders. With manual markup (i.e. returning + * a cookie to each rpm_get caller which they then supply to their + * paired rpm_put) we can remove corresponding pairs of and keep + * the array trimmed to active wakerefs. + */ + struct intel_runtime_pm_debug { + spinlock_t lock; + + depot_stack_handle_t last_acquire; + depot_stack_handle_t last_release; + + depot_stack_handle_t *owners; + unsigned long count; + } debug; +#endif }; enum intel_pipe_crc_source { @@ -1312,6 +1350,12 @@ struct i915_perf_stream { struct list_head link; /** + * @wakeref: As we keep the device awake while the perf stream is + * active, we track our runtime pm reference for later release. + */ + intel_wakeref_t wakeref; + + /** * @sample_flags: Flags representing the `DRM_I915_PERF_PROP_SAMPLE_*` * properties given when opening a stream, representing the contents * of a single sample as read() by userspace. @@ -1484,14 +1528,14 @@ struct drm_i915_private { * Base address of where the gmbus and gpio blocks are located (either * on PCH or on SoC for platforms without PCH). */ - uint32_t gpio_mmio_base; + u32 gpio_mmio_base; /* MMIO base address for MIPI regs */ - uint32_t mipi_mmio_base; + u32 mipi_mmio_base; - uint32_t psr_mmio_base; + u32 psr_mmio_base; - uint32_t pps_mmio_base; + u32 pps_mmio_base; wait_queue_head_t gmbus_wait_queue; @@ -1746,17 +1790,17 @@ struct drm_i915_private { * in 0.5us units for WM1+. */ /* primary */ - uint16_t pri_latency[5]; + u16 pri_latency[5]; /* sprite */ - uint16_t spr_latency[5]; + u16 spr_latency[5]; /* cursor */ - uint16_t cur_latency[5]; + u16 cur_latency[5]; /* * Raw watermark memory latency values * for SKL for all 8 levels * in 1us units. */ - uint16_t skl_latency[8]; + u16 skl_latency[8]; /* current hardware state */ union { @@ -1766,7 +1810,7 @@ struct drm_i915_private { struct g4x_wm_values g4x; }; - uint8_t max_level; + u8 max_level; /* * Should be held around atomic WM register writing; also @@ -1944,7 +1988,14 @@ struct drm_i915_private { void (*resume)(struct drm_i915_private *); void (*cleanup_engine)(struct intel_engine_cs *engine); - struct list_head timelines; + struct i915_gt_timelines { + struct mutex mutex; /* protects list, tainted by GPU */ + struct list_head active_list; + + /* Pack multiple timelines' seqnos into the same page */ + spinlock_t hwsp_lock; + struct list_head hwsp_free_list; + } timelines; struct list_head active_rings; struct list_head closed_vma; @@ -1957,7 +2008,7 @@ struct drm_i915_private { * In order to reduce the effect on performance, there * is a slight delay before we do so. */ - bool awake; + intel_wakeref_t awake; /** * The number of times we have woken up. @@ -2314,6 +2365,8 @@ static inline unsigned int i915_sg_segment_size(void) INTEL_INFO(dev_priv)->gt == 3) #define IS_CNL_WITH_PORT_F(dev_priv) (IS_CANNONLAKE(dev_priv) && \ (INTEL_DEVID(dev_priv) & 0x0004) == 0x0004) +#define IS_ICL_WITH_PORT_F(dev_priv) (IS_ICELAKE(dev_priv) && \ + INTEL_DEVID(dev_priv) != 0x8A51) #define IS_ALPHA_SUPPORT(intel_info) ((intel_info)->is_alpha_support) @@ -2449,7 +2502,7 @@ static inline unsigned int i915_sg_segment_size(void) #define HAS_FW_BLC(dev_priv) (INTEL_GEN(dev_priv) > 2) #define HAS_FBC(dev_priv) (INTEL_INFO(dev_priv)->display.has_fbc) -#define HAS_CUR_FBC(dev_priv) (!HAS_GMCH_DISPLAY(dev_priv) && INTEL_GEN(dev_priv) >= 7) +#define HAS_CUR_FBC(dev_priv) (!HAS_GMCH(dev_priv) && INTEL_GEN(dev_priv) >= 7) #define HAS_IPS(dev_priv) (IS_HSW_ULT(dev_priv) || IS_BROADWELL(dev_priv)) @@ -2529,7 +2582,7 @@ static inline unsigned int i915_sg_segment_size(void) #define HAS_PCH_NOP(dev_priv) (INTEL_PCH_TYPE(dev_priv) == PCH_NOP) #define HAS_PCH_SPLIT(dev_priv) (INTEL_PCH_TYPE(dev_priv) != PCH_NONE) -#define HAS_GMCH_DISPLAY(dev_priv) (INTEL_INFO(dev_priv)->display.has_gmch_display) +#define HAS_GMCH(dev_priv) (INTEL_INFO(dev_priv)->display.has_gmch) #define HAS_LSPCON(dev_priv) (INTEL_GEN(dev_priv) >= 9) @@ -2584,19 +2637,7 @@ extern const struct dev_pm_ops i915_pm_ops; extern int i915_driver_load(struct pci_dev *pdev, const struct pci_device_id *ent); extern void i915_driver_unload(struct drm_device *dev); -extern int intel_gpu_reset(struct drm_i915_private *dev_priv, u32 engine_mask); -extern bool intel_has_gpu_reset(struct drm_i915_private *dev_priv); - -extern void i915_reset(struct drm_i915_private *i915, - unsigned int stalled_mask, - const char *reason); -extern int i915_reset_engine(struct intel_engine_cs *engine, - const char *reason); - -extern bool intel_has_reset_engine(struct drm_i915_private *dev_priv); -extern int intel_reset_guc(struct drm_i915_private *dev_priv); -extern int intel_guc_reset_engine(struct intel_guc *guc, - struct intel_engine_cs *engine); + extern void intel_engine_init_hangcheck(struct intel_engine_cs *engine); extern void intel_hangcheck_init(struct drm_i915_private *dev_priv); extern unsigned long i915_chipset_val(struct drm_i915_private *dev_priv); @@ -2639,20 +2680,11 @@ static inline void i915_queue_hangcheck(struct drm_i915_private *dev_priv) &dev_priv->gpu_error.hangcheck_work, delay); } -__printf(4, 5) -void i915_handle_error(struct drm_i915_private *dev_priv, - u32 engine_mask, - unsigned long flags, - const char *fmt, ...); -#define I915_ERROR_CAPTURE BIT(0) - extern void intel_irq_init(struct drm_i915_private *dev_priv); extern void intel_irq_fini(struct drm_i915_private *dev_priv); int intel_irq_install(struct drm_i915_private *dev_priv); void intel_irq_uninstall(struct drm_i915_private *dev_priv); -void i915_clear_error_registers(struct drm_i915_private *dev_priv); - static inline bool intel_gvt_active(struct drm_i915_private *dev_priv) { return dev_priv->gvt; @@ -2676,45 +2708,45 @@ i915_disable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe, void valleyview_enable_display_irqs(struct drm_i915_private *dev_priv); void valleyview_disable_display_irqs(struct drm_i915_private *dev_priv); void i915_hotplug_interrupt_update(struct drm_i915_private *dev_priv, - uint32_t mask, - uint32_t bits); + u32 mask, + u32 bits); void ilk_update_display_irq(struct drm_i915_private *dev_priv, - uint32_t interrupt_mask, - uint32_t enabled_irq_mask); + u32 interrupt_mask, + u32 enabled_irq_mask); static inline void -ilk_enable_display_irq(struct drm_i915_private *dev_priv, uint32_t bits) +ilk_enable_display_irq(struct drm_i915_private *dev_priv, u32 bits) { ilk_update_display_irq(dev_priv, bits, bits); } static inline void -ilk_disable_display_irq(struct drm_i915_private *dev_priv, uint32_t bits) +ilk_disable_display_irq(struct drm_i915_private *dev_priv, u32 bits) { ilk_update_display_irq(dev_priv, bits, 0); } void bdw_update_pipe_irq(struct drm_i915_private *dev_priv, enum pipe pipe, - uint32_t interrupt_mask, - uint32_t enabled_irq_mask); + u32 interrupt_mask, + u32 enabled_irq_mask); static inline void bdw_enable_pipe_irq(struct drm_i915_private *dev_priv, - enum pipe pipe, uint32_t bits) + enum pipe pipe, u32 bits) { bdw_update_pipe_irq(dev_priv, pipe, bits, bits); } static inline void bdw_disable_pipe_irq(struct drm_i915_private *dev_priv, - enum pipe pipe, uint32_t bits) + enum pipe pipe, u32 bits) { bdw_update_pipe_irq(dev_priv, pipe, bits, 0); } void ibx_display_interrupt_update(struct drm_i915_private *dev_priv, - uint32_t interrupt_mask, - uint32_t enabled_irq_mask); + u32 interrupt_mask, + u32 enabled_irq_mask); static inline void -ibx_enable_display_interrupt(struct drm_i915_private *dev_priv, uint32_t bits) +ibx_enable_display_interrupt(struct drm_i915_private *dev_priv, u32 bits) { ibx_display_interrupt_update(dev_priv, bits, bits); } static inline void -ibx_disable_display_interrupt(struct drm_i915_private *dev_priv, uint32_t bits) +ibx_disable_display_interrupt(struct drm_i915_private *dev_priv, u32 bits) { ibx_display_interrupt_update(dev_priv, bits, 0); } @@ -2904,8 +2936,8 @@ enum i915_mm_subclass { /* lockdep subclass for obj->mm.lock/struct_mutex */ I915_MM_SHRINKER /* called "recursively" from direct-reclaim-esque */ }; -void __i915_gem_object_put_pages(struct drm_i915_gem_object *obj, - enum i915_mm_subclass subclass); +int __i915_gem_object_put_pages(struct drm_i915_gem_object *obj, + enum i915_mm_subclass subclass); void __i915_gem_object_invalidate(struct drm_i915_gem_object *obj); enum i915_map_type { @@ -2974,7 +3006,7 @@ int i915_gem_dumb_create(struct drm_file *file_priv, struct drm_device *dev, struct drm_mode_create_dumb *args); int i915_gem_mmap_gtt(struct drm_file *file_priv, struct drm_device *dev, - uint32_t handle, uint64_t *offset); + u32 handle, u64 *offset); int i915_gem_mmap_gtt_version(void); void i915_gem_track_fb(struct drm_i915_gem_object *old, @@ -2991,11 +3023,6 @@ static inline bool i915_reset_backoff(struct i915_gpu_error *error) return unlikely(test_bit(I915_RESET_BACKOFF, &error->flags)); } -static inline bool i915_reset_handoff(struct i915_gpu_error *error) -{ - return unlikely(test_bit(I915_RESET_HANDOFF, &error->flags)); -} - static inline bool i915_terminally_wedged(struct i915_gpu_error *error) { return unlikely(test_bit(I915_WEDGED, &error->flags)); @@ -3017,18 +3044,8 @@ static inline u32 i915_reset_engine_count(struct i915_gpu_error *error, return READ_ONCE(error->reset_engine_count[engine->id]); } -struct i915_request * -i915_gem_reset_prepare_engine(struct intel_engine_cs *engine); -int i915_gem_reset_prepare(struct drm_i915_private *dev_priv); -void i915_gem_reset(struct drm_i915_private *dev_priv, - unsigned int stalled_mask); -void i915_gem_reset_finish_engine(struct intel_engine_cs *engine); -void i915_gem_reset_finish(struct drm_i915_private *dev_priv); void i915_gem_set_wedged(struct drm_i915_private *dev_priv); bool i915_gem_unset_wedged(struct drm_i915_private *dev_priv); -void i915_gem_reset_engine(struct intel_engine_cs *engine, - struct i915_request *request, - bool stalled); void i915_gem_init_mmio(struct drm_i915_private *i915); int __must_check i915_gem_init(struct drm_i915_private *dev_priv); @@ -3125,7 +3142,7 @@ int i915_perf_remove_config_ioctl(struct drm_device *dev, void *data, struct drm_file *file); void i915_oa_init_reg_state(struct intel_engine_cs *engine, struct i915_gem_context *ctx, - uint32_t *reg_state); + u32 *reg_state); /* i915_gem_evict.c */ int __must_check i915_gem_evict_something(struct i915_address_space *vm, @@ -3300,6 +3317,20 @@ mkwrite_device_info(struct drm_i915_private *dev_priv) return (struct intel_device_info *)INTEL_INFO(dev_priv); } +static inline struct intel_sseu +intel_device_default_sseu(struct drm_i915_private *i915) +{ + const struct sseu_dev_info *sseu = &RUNTIME_INFO(i915)->sseu; + struct intel_sseu value = { + .slice_mask = sseu->slice_mask, + .subslice_mask = sseu->subslice_mask[0], + .min_eus_per_subslice = sseu->max_eus_per_subslice, + .max_eus_per_subslice = sseu->max_eus_per_subslice, + }; + + return value; +} + /* modesetting */ extern void intel_modeset_init_hw(struct drm_device *dev); extern int intel_modeset_init(struct drm_device *dev); @@ -3377,10 +3408,10 @@ bool bxt_ddi_phy_is_enabled(struct drm_i915_private *dev_priv, enum dpio_phy phy); bool bxt_ddi_phy_verify_state(struct drm_i915_private *dev_priv, enum dpio_phy phy); -uint8_t bxt_ddi_phy_calc_lane_lat_optim_mask(uint8_t lane_count); +u8 bxt_ddi_phy_calc_lane_lat_optim_mask(u8 lane_count); void bxt_ddi_phy_set_lane_optim_mask(struct intel_encoder *encoder, - uint8_t lane_lat_optim_mask); -uint8_t bxt_ddi_phy_get_lane_lat_optim_mask(struct intel_encoder *encoder); + u8 lane_lat_optim_mask); +u8 bxt_ddi_phy_get_lane_lat_optim_mask(struct intel_encoder *encoder); void chv_set_phy_signal_level(struct intel_encoder *encoder, u32 deemph_reg_value, u32 margin_reg_value, diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 04aef3f64c45..6728ea5c71d4 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -28,15 +28,6 @@ #include <drm/drm_vma_manager.h> #include <drm/drm_pci.h> #include <drm/i915_drm.h> -#include "i915_drv.h" -#include "i915_gem_clflush.h" -#include "i915_vgpu.h" -#include "i915_trace.h" -#include "intel_drv.h" -#include "intel_frontbuffer.h" -#include "intel_mocs.h" -#include "intel_workarounds.h" -#include "i915_gemfs.h" #include <linux/dma-fence-array.h> #include <linux/kthread.h> #include <linux/reservation.h> @@ -48,6 +39,18 @@ #include <linux/dma-buf.h> #include <linux/mman.h> +#include "i915_drv.h" +#include "i915_gem_clflush.h" +#include "i915_gemfs.h" +#include "i915_reset.h" +#include "i915_trace.h" +#include "i915_vgpu.h" + +#include "intel_drv.h" +#include "intel_frontbuffer.h" +#include "intel_mocs.h" +#include "intel_workarounds.h" + static void i915_gem_flush_free_objects(struct drm_i915_private *i915); static bool cpu_write_needs_clflush(struct drm_i915_gem_object *obj) @@ -140,6 +143,8 @@ int i915_mutex_lock_interruptible(struct drm_device *dev) static u32 __i915_gem_park(struct drm_i915_private *i915) { + intel_wakeref_t wakeref; + GEM_TRACE("\n"); lockdep_assert_held(&i915->drm.struct_mutex); @@ -170,14 +175,13 @@ static u32 __i915_gem_park(struct drm_i915_private *i915) i915_pmu_gt_parked(i915); i915_vma_parked(i915); - i915->gt.awake = false; + wakeref = fetch_and_zero(&i915->gt.awake); + GEM_BUG_ON(!wakeref); if (INTEL_GEN(i915) >= 6) gen6_rps_idle(i915); - intel_display_power_put(i915, POWER_DOMAIN_GT_IRQ); - - intel_runtime_pm_put(i915); + intel_display_power_put(i915, POWER_DOMAIN_GT_IRQ, wakeref); return i915->gt.epoch; } @@ -202,12 +206,11 @@ void i915_gem_unpark(struct drm_i915_private *i915) lockdep_assert_held(&i915->drm.struct_mutex); GEM_BUG_ON(!i915->gt.active_requests); + assert_rpm_wakelock_held(i915); if (i915->gt.awake) return; - intel_runtime_pm_get_noresume(i915); - /* * It seems that the DMC likes to transition between the DC states a lot * when there are no connected displays (no active power domains) during @@ -219,9 +222,9 @@ void i915_gem_unpark(struct drm_i915_private *i915) * Work around it by grabbing a GT IRQ power domain whilst there is any * GT activity, preventing any DC state transitions. */ - intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ); + i915->gt.awake = intel_display_power_get(i915, POWER_DOMAIN_GT_IRQ); + GEM_BUG_ON(!i915->gt.awake); - i915->gt.awake = true; if (unlikely(++i915->gt.epoch == 0)) /* keep 0 as invalid */ i915->gt.epoch = 1; @@ -244,21 +247,19 @@ int i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data, struct drm_file *file) { - struct drm_i915_private *dev_priv = to_i915(dev); - struct i915_ggtt *ggtt = &dev_priv->ggtt; + struct i915_ggtt *ggtt = &to_i915(dev)->ggtt; struct drm_i915_gem_get_aperture *args = data; struct i915_vma *vma; u64 pinned; + mutex_lock(&ggtt->vm.mutex); + pinned = ggtt->vm.reserved; - mutex_lock(&dev->struct_mutex); - list_for_each_entry(vma, &ggtt->vm.active_list, vm_link) + list_for_each_entry(vma, &ggtt->vm.bound_list, vm_link) if (i915_vma_is_pinned(vma)) pinned += vma->node.size; - list_for_each_entry(vma, &ggtt->vm.inactive_list, vm_link) - if (i915_vma_is_pinned(vma)) - pinned += vma->node.size; - mutex_unlock(&dev->struct_mutex); + + mutex_unlock(&ggtt->vm.mutex); args->aper_size = ggtt->vm.total; args->aper_available_size = args->aper_size - pinned; @@ -438,15 +439,19 @@ int i915_gem_object_unbind(struct drm_i915_gem_object *obj) if (ret) return ret; - while ((vma = list_first_entry_or_null(&obj->vma_list, - struct i915_vma, - obj_link))) { + spin_lock(&obj->vma.lock); + while (!ret && (vma = list_first_entry_or_null(&obj->vma.list, + struct i915_vma, + obj_link))) { list_move_tail(&vma->obj_link, &still_in_list); + spin_unlock(&obj->vma.lock); + ret = i915_vma_unbind(vma); - if (ret) - break; + + spin_lock(&obj->vma.lock); } - list_splice(&still_in_list, &obj->vma_list); + list_splice(&still_in_list, &obj->vma.list); + spin_unlock(&obj->vma.lock); return ret; } @@ -656,11 +661,6 @@ i915_gem_object_wait(struct drm_i915_gem_object *obj, struct intel_rps_client *rps_client) { might_sleep(); -#if IS_ENABLED(CONFIG_LOCKDEP) - GEM_BUG_ON(debug_locks && - !!lockdep_is_held(&obj->base.dev->struct_mutex) != - !!(flags & I915_WAIT_LOCKED)); -#endif GEM_BUG_ON(timeout < 0); timeout = i915_gem_object_wait_reservation(obj->resv, @@ -712,8 +712,8 @@ void i915_gem_object_free(struct drm_i915_gem_object *obj) static int i915_gem_create(struct drm_file *file, struct drm_i915_private *dev_priv, - uint64_t size, - uint32_t *handle_p) + u64 size, + u32 *handle_p) { struct drm_i915_gem_object *obj; int ret; @@ -784,6 +784,8 @@ fb_write_origin(struct drm_i915_gem_object *obj, unsigned int domain) void i915_gem_flush_ggtt_writes(struct drm_i915_private *dev_priv) { + intel_wakeref_t wakeref; + /* * No actual flushing is required for the GTT write domain for reads * from the GTT domain. Writes to it "immediately" go to main memory @@ -810,13 +812,13 @@ void i915_gem_flush_ggtt_writes(struct drm_i915_private *dev_priv) i915_gem_chipset_flush(dev_priv); - intel_runtime_pm_get(dev_priv); - spin_lock_irq(&dev_priv->uncore.lock); + with_intel_runtime_pm(dev_priv, wakeref) { + spin_lock_irq(&dev_priv->uncore.lock); - POSTING_READ_FW(RING_HEAD(RENDER_RING_BASE)); + POSTING_READ_FW(RING_HEAD(RENDER_RING_BASE)); - spin_unlock_irq(&dev_priv->uncore.lock); - intel_runtime_pm_put(dev_priv); + spin_unlock_irq(&dev_priv->uncore.lock); + } } static void @@ -1068,6 +1070,7 @@ i915_gem_gtt_pread(struct drm_i915_gem_object *obj, { struct drm_i915_private *i915 = to_i915(obj->base.dev); struct i915_ggtt *ggtt = &i915->ggtt; + intel_wakeref_t wakeref; struct drm_mm_node node; struct i915_vma *vma; void __user *user_data; @@ -1078,7 +1081,7 @@ i915_gem_gtt_pread(struct drm_i915_gem_object *obj, if (ret) return ret; - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE | PIN_NONFAULT | @@ -1151,7 +1154,7 @@ out_unpin: i915_vma_unpin(vma); } out_unlock: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); return ret; @@ -1252,6 +1255,7 @@ i915_gem_gtt_pwrite_fast(struct drm_i915_gem_object *obj, { struct drm_i915_private *i915 = to_i915(obj->base.dev); struct i915_ggtt *ggtt = &i915->ggtt; + intel_wakeref_t wakeref; struct drm_mm_node node; struct i915_vma *vma; u64 remain, offset; @@ -1270,13 +1274,14 @@ i915_gem_gtt_pwrite_fast(struct drm_i915_gem_object *obj, * This easily dwarfs any performance advantage from * using the cache bypass of indirect GGTT access. */ - if (!intel_runtime_pm_get_if_in_use(i915)) { + wakeref = intel_runtime_pm_get_if_in_use(i915); + if (!wakeref) { ret = -EFAULT; goto out_unlock; } } else { /* No backing pages, no fallback, we must force GGTT access */ - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); } vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, @@ -1358,7 +1363,7 @@ out_unpin: i915_vma_unpin(vma); } out_rpm: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); out_unlock: mutex_unlock(&i915->drm.struct_mutex); return ret; @@ -1531,23 +1536,21 @@ err: static void i915_gem_object_bump_inactive_ggtt(struct drm_i915_gem_object *obj) { - struct drm_i915_private *i915; + struct drm_i915_private *i915 = to_i915(obj->base.dev); struct list_head *list; struct i915_vma *vma; GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj)); + mutex_lock(&i915->ggtt.vm.mutex); for_each_ggtt_vma(vma, obj) { - if (i915_vma_is_active(vma)) - continue; - if (!drm_mm_node_allocated(&vma->node)) continue; - list_move_tail(&vma->vm_link, &vma->vm->inactive_list); + list_move_tail(&vma->vm_link, &vma->vm->bound_list); } + mutex_unlock(&i915->ggtt.vm.mutex); - i915 = to_i915(obj->base.dev); spin_lock(&i915->mm.obj_lock); list = obj->bind_count ? &i915->mm.bound_list : &i915->mm.unbound_list; list_move_tail(&obj->mm.link, list); @@ -1567,8 +1570,8 @@ i915_gem_set_domain_ioctl(struct drm_device *dev, void *data, { struct drm_i915_gem_set_domain *args = data; struct drm_i915_gem_object *obj; - uint32_t read_domains = args->read_domains; - uint32_t write_domain = args->write_domain; + u32 read_domains = args->read_domains; + u32 write_domain = args->write_domain; int err; /* Only handle setting domains to types used by the CPU. */ @@ -1678,6 +1681,16 @@ i915_gem_sw_finish_ioctl(struct drm_device *dev, void *data, return 0; } +static inline bool +__vma_matches(struct vm_area_struct *vma, struct file *filp, + unsigned long addr, unsigned long size) +{ + if (vma->vm_file != filp) + return false; + + return vma->vm_start == addr && (vma->vm_end - vma->vm_start) == size; +} + /** * i915_gem_mmap_ioctl - Maps the contents of an object, returning the address * it is mapped to. @@ -1727,6 +1740,9 @@ i915_gem_mmap_ioctl(struct drm_device *dev, void *data, addr = vm_mmap(obj->base.filp, 0, args->size, PROT_READ | PROT_WRITE, MAP_SHARED, args->offset); + if (IS_ERR_VALUE(addr)) + goto err; + if (args->flags & I915_MMAP_WC) { struct mm_struct *mm = current->mm; struct vm_area_struct *vma; @@ -1736,23 +1752,28 @@ i915_gem_mmap_ioctl(struct drm_device *dev, void *data, return -EINTR; } vma = find_vma(mm, addr); - if (vma) + if (vma && __vma_matches(vma, obj->base.filp, addr, args->size)) vma->vm_page_prot = pgprot_writecombine(vm_get_page_prot(vma->vm_flags)); else addr = -ENOMEM; up_write(&mm->mmap_sem); + if (IS_ERR_VALUE(addr)) + goto err; /* This may race, but that's ok, it only gets set */ WRITE_ONCE(obj->frontbuffer_ggtt_origin, ORIGIN_CPU); } i915_gem_object_put(obj); - if (IS_ERR((void *)addr)) - return addr; - args->addr_ptr = (uint64_t) addr; + args->addr_ptr = (u64)addr; return 0; + +err: + i915_gem_object_put(obj); + + return addr; } static unsigned int tile_row_pages(const struct drm_i915_gem_object *obj) @@ -1863,6 +1884,7 @@ vm_fault_t i915_gem_fault(struct vm_fault *vmf) struct drm_i915_private *dev_priv = to_i915(dev); struct i915_ggtt *ggtt = &dev_priv->ggtt; bool write = area->vm_flags & VM_WRITE; + intel_wakeref_t wakeref; struct i915_vma *vma; pgoff_t page_offset; int ret; @@ -1892,7 +1914,7 @@ vm_fault_t i915_gem_fault(struct vm_fault *vmf) if (ret) goto err; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); ret = i915_mutex_lock_interruptible(dev); if (ret) @@ -1970,7 +1992,7 @@ err_unpin: err_unlock: mutex_unlock(&dev->struct_mutex); err_rpm: - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); i915_gem_object_unpin_pages(obj); err: switch (ret) { @@ -2043,6 +2065,7 @@ void i915_gem_release_mmap(struct drm_i915_gem_object *obj) { struct drm_i915_private *i915 = to_i915(obj->base.dev); + intel_wakeref_t wakeref; /* Serialisation between user GTT access and our code depends upon * revoking the CPU's PTE whilst the mutex is held. The next user @@ -2053,7 +2076,7 @@ i915_gem_release_mmap(struct drm_i915_gem_object *obj) * wakeref. */ lockdep_assert_held(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); if (!obj->userfault_count) goto out; @@ -2070,7 +2093,7 @@ i915_gem_release_mmap(struct drm_i915_gem_object *obj) wmb(); out: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); } void i915_gem_runtime_suspend(struct drm_i915_private *dev_priv) @@ -2150,8 +2173,8 @@ static void i915_gem_object_free_mmap_offset(struct drm_i915_gem_object *obj) int i915_gem_mmap_gtt(struct drm_file *file, struct drm_device *dev, - uint32_t handle, - uint64_t *offset) + u32 handle, + u64 *offset) { struct drm_i915_gem_object *obj; int ret; @@ -2298,8 +2321,8 @@ __i915_gem_object_unset_pages(struct drm_i915_gem_object *obj) struct sg_table *pages; pages = fetch_and_zero(&obj->mm.pages); - if (!pages) - return NULL; + if (IS_ERR_OR_NULL(pages)) + return pages; spin_lock(&i915->mm.obj_lock); list_del(&obj->mm.link); @@ -2323,22 +2346,23 @@ __i915_gem_object_unset_pages(struct drm_i915_gem_object *obj) return pages; } -void __i915_gem_object_put_pages(struct drm_i915_gem_object *obj, - enum i915_mm_subclass subclass) +int __i915_gem_object_put_pages(struct drm_i915_gem_object *obj, + enum i915_mm_subclass subclass) { struct sg_table *pages; + int ret; if (i915_gem_object_has_pinned_pages(obj)) - return; + return -EBUSY; GEM_BUG_ON(obj->bind_count); - if (!i915_gem_object_has_pages(obj)) - return; /* May be called by shrinker from within get_pages() (on another bo) */ mutex_lock_nested(&obj->mm.lock, subclass); - if (unlikely(atomic_read(&obj->mm.pages_pin_count))) + if (unlikely(atomic_read(&obj->mm.pages_pin_count))) { + ret = -EBUSY; goto unlock; + } /* * ->put_pages might need to allocate memory for the bit17 swizzle @@ -2346,11 +2370,24 @@ void __i915_gem_object_put_pages(struct drm_i915_gem_object *obj, * lists early. */ pages = __i915_gem_object_unset_pages(obj); + + /* + * XXX Temporary hijinx to avoid updating all backends to handle + * NULL pages. In the future, when we have more asynchronous + * get_pages backends we should be better able to handle the + * cancellation of the async task in a more uniform manner. + */ + if (!pages && !i915_gem_object_needs_async_cancel(obj)) + pages = ERR_PTR(-EINVAL); + if (!IS_ERR(pages)) obj->ops->put_pages(obj, pages); + ret = 0; unlock: mutex_unlock(&obj->mm.lock); + + return ret; } bool i915_sg_trim(struct sg_table *orig_st) @@ -2854,59 +2891,12 @@ i915_gem_object_pwrite_gtt(struct drm_i915_gem_object *obj, return 0; } -static void i915_gem_client_mark_guilty(struct drm_i915_file_private *file_priv, - const struct i915_gem_context *ctx) +static bool match_ring(struct i915_request *rq) { - unsigned int score; - unsigned long prev_hang; - - if (i915_gem_context_is_banned(ctx)) - score = I915_CLIENT_SCORE_CONTEXT_BAN; - else - score = 0; - - prev_hang = xchg(&file_priv->hang_timestamp, jiffies); - if (time_before(jiffies, prev_hang + I915_CLIENT_FAST_HANG_JIFFIES)) - score += I915_CLIENT_SCORE_HANG_FAST; + struct drm_i915_private *dev_priv = rq->i915; + u32 ring = I915_READ(RING_START(rq->engine->mmio_base)); - if (score) { - atomic_add(score, &file_priv->ban_score); - - DRM_DEBUG_DRIVER("client %s: gained %u ban score, now %u\n", - ctx->name, score, - atomic_read(&file_priv->ban_score)); - } -} - -static void i915_gem_context_mark_guilty(struct i915_gem_context *ctx) -{ - unsigned int score; - bool banned, bannable; - - atomic_inc(&ctx->guilty_count); - - bannable = i915_gem_context_is_bannable(ctx); - score = atomic_add_return(CONTEXT_SCORE_GUILTY, &ctx->ban_score); - banned = score >= CONTEXT_SCORE_BAN_THRESHOLD; - - /* Cool contexts don't accumulate client ban score */ - if (!bannable) - return; - - if (banned) { - DRM_DEBUG_DRIVER("context %s: guilty %d, score %u, banned\n", - ctx->name, atomic_read(&ctx->guilty_count), - score); - i915_gem_context_set_banned(ctx); - } - - if (!IS_ERR_OR_NULL(ctx->file_priv)) - i915_gem_client_mark_guilty(ctx->file_priv, ctx); -} - -static void i915_gem_context_mark_innocent(struct i915_gem_context *ctx) -{ - atomic_inc(&ctx->active_count); + return ring == i915_ggtt_offset(rq->ring->vma); } struct i915_request * @@ -2928,9 +2918,16 @@ i915_gem_find_active_request(struct intel_engine_cs *engine) */ spin_lock_irqsave(&engine->timeline.lock, flags); list_for_each_entry(request, &engine->timeline.requests, link) { - if (__i915_request_completed(request, request->global_seqno)) + if (i915_request_completed(request)) continue; + if (!i915_request_started(request)) + break; + + /* More than one preemptible request may match! */ + if (!match_ring(request)) + break; + active = request; break; } @@ -2939,361 +2936,6 @@ i915_gem_find_active_request(struct intel_engine_cs *engine) return active; } -/* - * Ensure irq handler finishes, and not run again. - * Also return the active request so that we only search for it once. - */ -struct i915_request * -i915_gem_reset_prepare_engine(struct intel_engine_cs *engine) -{ - struct i915_request *request; - - /* - * During the reset sequence, we must prevent the engine from - * entering RC6. As the context state is undefined until we restart - * the engine, if it does enter RC6 during the reset, the state - * written to the powercontext is undefined and so we may lose - * GPU state upon resume, i.e. fail to restart after a reset. - */ - intel_uncore_forcewake_get(engine->i915, FORCEWAKE_ALL); - - request = engine->reset.prepare(engine); - if (request && request->fence.error == -EIO) - request = ERR_PTR(-EIO); /* Previous reset failed! */ - - return request; -} - -int i915_gem_reset_prepare(struct drm_i915_private *dev_priv) -{ - struct intel_engine_cs *engine; - struct i915_request *request; - enum intel_engine_id id; - int err = 0; - - for_each_engine(engine, dev_priv, id) { - request = i915_gem_reset_prepare_engine(engine); - if (IS_ERR(request)) { - err = PTR_ERR(request); - continue; - } - - engine->hangcheck.active_request = request; - } - - i915_gem_revoke_fences(dev_priv); - intel_uc_sanitize(dev_priv); - - return err; -} - -static void engine_skip_context(struct i915_request *request) -{ - struct intel_engine_cs *engine = request->engine; - struct i915_gem_context *hung_ctx = request->gem_context; - struct i915_timeline *timeline = request->timeline; - unsigned long flags; - - GEM_BUG_ON(timeline == &engine->timeline); - - spin_lock_irqsave(&engine->timeline.lock, flags); - spin_lock(&timeline->lock); - - list_for_each_entry_continue(request, &engine->timeline.requests, link) - if (request->gem_context == hung_ctx) - i915_request_skip(request, -EIO); - - list_for_each_entry(request, &timeline->requests, link) - i915_request_skip(request, -EIO); - - spin_unlock(&timeline->lock); - spin_unlock_irqrestore(&engine->timeline.lock, flags); -} - -/* Returns the request if it was guilty of the hang */ -static struct i915_request * -i915_gem_reset_request(struct intel_engine_cs *engine, - struct i915_request *request, - bool stalled) -{ - /* The guilty request will get skipped on a hung engine. - * - * Users of client default contexts do not rely on logical - * state preserved between batches so it is safe to execute - * queued requests following the hang. Non default contexts - * rely on preserved state, so skipping a batch loses the - * evolution of the state and it needs to be considered corrupted. - * Executing more queued batches on top of corrupted state is - * risky. But we take the risk by trying to advance through - * the queued requests in order to make the client behaviour - * more predictable around resets, by not throwing away random - * amount of batches it has prepared for execution. Sophisticated - * clients can use gem_reset_stats_ioctl and dma fence status - * (exported via sync_file info ioctl on explicit fences) to observe - * when it loses the context state and should rebuild accordingly. - * - * The context ban, and ultimately the client ban, mechanism are safety - * valves if client submission ends up resulting in nothing more than - * subsequent hangs. - */ - - if (i915_request_completed(request)) { - GEM_TRACE("%s pardoned global=%d (fence %llx:%lld), current %d\n", - engine->name, request->global_seqno, - request->fence.context, request->fence.seqno, - intel_engine_get_seqno(engine)); - stalled = false; - } - - if (stalled) { - i915_gem_context_mark_guilty(request->gem_context); - i915_request_skip(request, -EIO); - - /* If this context is now banned, skip all pending requests. */ - if (i915_gem_context_is_banned(request->gem_context)) - engine_skip_context(request); - } else { - /* - * Since this is not the hung engine, it may have advanced - * since the hang declaration. Double check by refinding - * the active request at the time of the reset. - */ - request = i915_gem_find_active_request(engine); - if (request) { - unsigned long flags; - - i915_gem_context_mark_innocent(request->gem_context); - dma_fence_set_error(&request->fence, -EAGAIN); - - /* Rewind the engine to replay the incomplete rq */ - spin_lock_irqsave(&engine->timeline.lock, flags); - request = list_prev_entry(request, link); - if (&request->link == &engine->timeline.requests) - request = NULL; - spin_unlock_irqrestore(&engine->timeline.lock, flags); - } - } - - return request; -} - -void i915_gem_reset_engine(struct intel_engine_cs *engine, - struct i915_request *request, - bool stalled) -{ - if (request) - request = i915_gem_reset_request(engine, request, stalled); - - /* Setup the CS to resume from the breadcrumb of the hung request */ - engine->reset.reset(engine, request); -} - -void i915_gem_reset(struct drm_i915_private *dev_priv, - unsigned int stalled_mask) -{ - struct intel_engine_cs *engine; - enum intel_engine_id id; - - lockdep_assert_held(&dev_priv->drm.struct_mutex); - - i915_retire_requests(dev_priv); - - for_each_engine(engine, dev_priv, id) { - struct intel_context *ce; - - i915_gem_reset_engine(engine, - engine->hangcheck.active_request, - stalled_mask & ENGINE_MASK(id)); - ce = fetch_and_zero(&engine->last_retired_context); - if (ce) - intel_context_unpin(ce); - - /* - * Ostensibily, we always want a context loaded for powersaving, - * so if the engine is idle after the reset, send a request - * to load our scratch kernel_context. - * - * More mysteriously, if we leave the engine idle after a reset, - * the next userspace batch may hang, with what appears to be - * an incoherent read by the CS (presumably stale TLB). An - * empty request appears sufficient to paper over the glitch. - */ - if (intel_engine_is_idle(engine)) { - struct i915_request *rq; - - rq = i915_request_alloc(engine, - dev_priv->kernel_context); - if (!IS_ERR(rq)) - i915_request_add(rq); - } - } - - i915_gem_restore_fences(dev_priv); -} - -void i915_gem_reset_finish_engine(struct intel_engine_cs *engine) -{ - engine->reset.finish(engine); - - intel_uncore_forcewake_put(engine->i915, FORCEWAKE_ALL); -} - -void i915_gem_reset_finish(struct drm_i915_private *dev_priv) -{ - struct intel_engine_cs *engine; - enum intel_engine_id id; - - lockdep_assert_held(&dev_priv->drm.struct_mutex); - - for_each_engine(engine, dev_priv, id) { - engine->hangcheck.active_request = NULL; - i915_gem_reset_finish_engine(engine); - } -} - -static void nop_submit_request(struct i915_request *request) -{ - unsigned long flags; - - GEM_TRACE("%s fence %llx:%lld -> -EIO\n", - request->engine->name, - request->fence.context, request->fence.seqno); - dma_fence_set_error(&request->fence, -EIO); - - spin_lock_irqsave(&request->engine->timeline.lock, flags); - __i915_request_submit(request); - intel_engine_write_global_seqno(request->engine, request->global_seqno); - spin_unlock_irqrestore(&request->engine->timeline.lock, flags); -} - -void i915_gem_set_wedged(struct drm_i915_private *i915) -{ - struct intel_engine_cs *engine; - enum intel_engine_id id; - - GEM_TRACE("start\n"); - - if (GEM_SHOW_DEBUG()) { - struct drm_printer p = drm_debug_printer(__func__); - - for_each_engine(engine, i915, id) - intel_engine_dump(engine, &p, "%s\n", engine->name); - } - - if (test_and_set_bit(I915_WEDGED, &i915->gpu_error.flags)) - goto out; - - /* - * First, stop submission to hw, but do not yet complete requests by - * rolling the global seqno forward (since this would complete requests - * for which we haven't set the fence error to EIO yet). - */ - for_each_engine(engine, i915, id) - i915_gem_reset_prepare_engine(engine); - - /* Even if the GPU reset fails, it should still stop the engines */ - if (INTEL_GEN(i915) >= 5) - intel_gpu_reset(i915, ALL_ENGINES); - - for_each_engine(engine, i915, id) { - engine->submit_request = nop_submit_request; - engine->schedule = NULL; - } - i915->caps.scheduler = 0; - - /* - * Make sure no request can slip through without getting completed by - * either this call here to intel_engine_write_global_seqno, or the one - * in nop_submit_request. - */ - synchronize_rcu(); - - /* Mark all executing requests as skipped */ - for_each_engine(engine, i915, id) - engine->cancel_requests(engine); - - for_each_engine(engine, i915, id) { - i915_gem_reset_finish_engine(engine); - intel_engine_wakeup(engine); - } - -out: - GEM_TRACE("end\n"); - - wake_up_all(&i915->gpu_error.reset_queue); -} - -bool i915_gem_unset_wedged(struct drm_i915_private *i915) -{ - struct i915_timeline *tl; - - lockdep_assert_held(&i915->drm.struct_mutex); - if (!test_bit(I915_WEDGED, &i915->gpu_error.flags)) - return true; - - if (!i915->gt.scratch) /* Never full initialised, recovery impossible */ - return false; - - GEM_TRACE("start\n"); - - /* - * Before unwedging, make sure that all pending operations - * are flushed and errored out - we may have requests waiting upon - * third party fences. We marked all inflight requests as EIO, and - * every execbuf since returned EIO, for consistency we want all - * the currently pending requests to also be marked as EIO, which - * is done inside our nop_submit_request - and so we must wait. - * - * No more can be submitted until we reset the wedged bit. - */ - list_for_each_entry(tl, &i915->gt.timelines, link) { - struct i915_request *rq; - - rq = i915_gem_active_peek(&tl->last_request, - &i915->drm.struct_mutex); - if (!rq) - continue; - - /* - * We can't use our normal waiter as we want to - * avoid recursively trying to handle the current - * reset. The basic dma_fence_default_wait() installs - * a callback for dma_fence_signal(), which is - * triggered by our nop handler (indirectly, the - * callback enables the signaler thread which is - * woken by the nop_submit_request() advancing the seqno - * and when the seqno passes the fence, the signaler - * then signals the fence waking us up). - */ - if (dma_fence_default_wait(&rq->fence, true, - MAX_SCHEDULE_TIMEOUT) < 0) - return false; - } - i915_retire_requests(i915); - GEM_BUG_ON(i915->gt.active_requests); - - intel_engines_sanitize(i915, false); - - /* - * Undo nop_submit_request. We prevent all new i915 requests from - * being queued (by disallowing execbuf whilst wedged) so having - * waited for all active requests above, we know the system is idle - * and do not have to worry about a thread being inside - * engine->submit_request() as we swap over. So unlike installing - * the nop_submit_request on reset, we can do this from normal - * context and do not require stop_machine(). - */ - intel_engines_reset_default_submission(i915); - i915_gem_contexts_lost(i915); - - GEM_TRACE("end\n"); - - smp_mb__before_atomic(); /* complete takeover before enabling execbuf */ - clear_bit(I915_WEDGED, &i915->gpu_error.flags); - - return true; -} - static void i915_gem_retire_work_handler(struct work_struct *work) { @@ -3396,7 +3038,7 @@ static void assert_kernel_context_is_current(struct drm_i915_private *i915) GEM_BUG_ON(i915->gt.active_requests); for_each_engine(engine, i915, id) { - GEM_BUG_ON(__i915_gem_active_peek(&engine->timeline.last_request)); + GEM_BUG_ON(__i915_active_request_peek(&engine->timeline.last_request)); GEM_BUG_ON(engine->last_retired_context != to_intel_context(i915->kernel_context, engine)); } @@ -3615,33 +3257,6 @@ i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file) return ret; } -static long wait_for_timeline(struct i915_timeline *tl, - unsigned int flags, long timeout) -{ - struct i915_request *rq; - - rq = i915_gem_active_get_unlocked(&tl->last_request); - if (!rq) - return timeout; - - /* - * "Race-to-idle". - * - * Switching to the kernel context is often used a synchronous - * step prior to idling, e.g. in suspend for flushing all - * current operations to memory before sleeping. These we - * want to complete as quickly as possible to avoid prolonged - * stalls, so allow the gpu to boost to maximum clocks. - */ - if (flags & I915_WAIT_FOR_IDLE_BOOST) - gen6_rps_boost(rq, NULL); - - timeout = i915_request_wait(rq, flags, timeout); - i915_request_put(rq); - - return timeout; -} - static int wait_for_engines(struct drm_i915_private *i915) { if (wait_for(intel_engines_are_idle(i915), I915_IDLE_ENGINES_TIMEOUT)) { @@ -3655,6 +3270,52 @@ static int wait_for_engines(struct drm_i915_private *i915) return 0; } +static long +wait_for_timelines(struct drm_i915_private *i915, + unsigned int flags, long timeout) +{ + struct i915_gt_timelines *gt = &i915->gt.timelines; + struct i915_timeline *tl; + + if (!READ_ONCE(i915->gt.active_requests)) + return timeout; + + mutex_lock(>->mutex); + list_for_each_entry(tl, >->active_list, link) { + struct i915_request *rq; + + rq = i915_active_request_get_unlocked(&tl->last_request); + if (!rq) + continue; + + mutex_unlock(>->mutex); + + /* + * "Race-to-idle". + * + * Switching to the kernel context is often used a synchronous + * step prior to idling, e.g. in suspend for flushing all + * current operations to memory before sleeping. These we + * want to complete as quickly as possible to avoid prolonged + * stalls, so allow the gpu to boost to maximum clocks. + */ + if (flags & I915_WAIT_FOR_IDLE_BOOST) + gen6_rps_boost(rq, NULL); + + timeout = i915_request_wait(rq, flags, timeout); + i915_request_put(rq); + if (timeout < 0) + return timeout; + + /* restart after reacquiring the lock */ + mutex_lock(>->mutex); + tl = list_entry(>->active_list, typeof(*tl), link); + } + mutex_unlock(>->mutex); + + return timeout; +} + int i915_gem_wait_for_idle(struct drm_i915_private *i915, unsigned int flags, long timeout) { @@ -3666,17 +3327,15 @@ int i915_gem_wait_for_idle(struct drm_i915_private *i915, if (!READ_ONCE(i915->gt.awake)) return 0; + timeout = wait_for_timelines(i915, flags, timeout); + if (timeout < 0) + return timeout; + if (flags & I915_WAIT_LOCKED) { - struct i915_timeline *tl; int err; lockdep_assert_held(&i915->drm.struct_mutex); - list_for_each_entry(tl, &i915->gt.timelines, link) { - timeout = wait_for_timeline(tl, flags, timeout); - if (timeout < 0) - return timeout; - } if (GEM_SHOW_DEBUG() && !timeout) { /* Presume that timeout was non-zero to begin with! */ dev_warn(&i915->drm.pdev->dev, @@ -3690,17 +3349,6 @@ int i915_gem_wait_for_idle(struct drm_i915_private *i915, i915_retire_requests(i915); GEM_BUG_ON(i915->gt.active_requests); - } else { - struct intel_engine_cs *engine; - enum intel_engine_id id; - - for_each_engine(engine, i915, id) { - struct i915_timeline *tl = &engine->timeline; - - timeout = wait_for_timeline(tl, flags, timeout); - if (timeout < 0) - return timeout; - } } return 0; @@ -3886,7 +3534,7 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj, * reading an invalid PTE on older architectures. */ restart: - list_for_each_entry(vma, &obj->vma_list, obj_link) { + list_for_each_entry(vma, &obj->vma.list, obj_link) { if (!drm_mm_node_allocated(&vma->node)) continue; @@ -3964,7 +3612,7 @@ restart: */ } - list_for_each_entry(vma, &obj->vma_list, obj_link) { + list_for_each_entry(vma, &obj->vma.list, obj_link) { if (!drm_mm_node_allocated(&vma->node)) continue; @@ -3974,7 +3622,7 @@ restart: } } - list_for_each_entry(vma, &obj->vma_list, obj_link) + list_for_each_entry(vma, &obj->vma.list, obj_link) vma->node.color = cache_level; i915_gem_object_set_cache_coherency(obj, cache_level); obj->cache_dirty = true; /* Always invalidate stale cachelines */ @@ -4537,7 +4185,8 @@ out: } static void -frontbuffer_retire(struct i915_gem_active *active, struct i915_request *request) +frontbuffer_retire(struct i915_active_request *active, + struct i915_request *request) { struct drm_i915_gem_object *obj = container_of(active, typeof(*obj), frontbuffer_write); @@ -4550,7 +4199,9 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj, { mutex_init(&obj->mm.lock); - INIT_LIST_HEAD(&obj->vma_list); + spin_lock_init(&obj->vma.lock); + INIT_LIST_HEAD(&obj->vma.list); + INIT_LIST_HEAD(&obj->lut_list); INIT_LIST_HEAD(&obj->batch_pool_link); @@ -4562,7 +4213,8 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj, obj->resv = &obj->__builtin_resv; obj->frontbuffer_ggtt_origin = ORIGIN_GTT; - init_request_active(&obj->frontbuffer_write, frontbuffer_retire); + i915_active_request_init(&obj->frontbuffer_write, + NULL, frontbuffer_retire); obj->mm.madv = I915_MADV_WILLNEED; INIT_RADIX_TREE(&obj->mm.get_page.radix, GFP_KERNEL | __GFP_NOWARN); @@ -4705,8 +4357,9 @@ static void __i915_gem_free_objects(struct drm_i915_private *i915, struct llist_node *freed) { struct drm_i915_gem_object *obj, *on; + intel_wakeref_t wakeref; - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); llist_for_each_entry_safe(obj, on, freed, freed) { struct i915_vma *vma, *vn; @@ -4715,14 +4368,13 @@ static void __i915_gem_free_objects(struct drm_i915_private *i915, mutex_lock(&i915->drm.struct_mutex); GEM_BUG_ON(i915_gem_object_is_active(obj)); - list_for_each_entry_safe(vma, vn, - &obj->vma_list, obj_link) { + list_for_each_entry_safe(vma, vn, &obj->vma.list, obj_link) { GEM_BUG_ON(i915_vma_is_active(vma)); vma->flags &= ~I915_VMA_PIN_MASK; i915_vma_destroy(vma); } - GEM_BUG_ON(!list_empty(&obj->vma_list)); - GEM_BUG_ON(!RB_EMPTY_ROOT(&obj->vma_tree)); + GEM_BUG_ON(!list_empty(&obj->vma.list)); + GEM_BUG_ON(!RB_EMPTY_ROOT(&obj->vma.tree)); /* This serializes freeing with the shrinker. Since the free * is delayed, first by RCU then by the workqueue, we want the @@ -4767,7 +4419,7 @@ static void __i915_gem_free_objects(struct drm_i915_private *i915, if (on) cond_resched(); } - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); } static void i915_gem_flush_free_objects(struct drm_i915_private *i915) @@ -4876,11 +4528,11 @@ void __i915_gem_object_release_unless_active(struct drm_i915_gem_object *obj) void i915_gem_sanitize(struct drm_i915_private *i915) { - GEM_TRACE("\n"); + intel_wakeref_t wakeref; - mutex_lock(&i915->drm.struct_mutex); + GEM_TRACE("\n"); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); intel_uncore_forcewake_get(i915, FORCEWAKE_ALL); /* @@ -4903,21 +4555,25 @@ void i915_gem_sanitize(struct drm_i915_private *i915) intel_engines_sanitize(i915, false); intel_uncore_forcewake_put(i915, FORCEWAKE_ALL); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); + mutex_lock(&i915->drm.struct_mutex); i915_gem_contexts_lost(i915); mutex_unlock(&i915->drm.struct_mutex); } int i915_gem_suspend(struct drm_i915_private *i915) { + intel_wakeref_t wakeref; int ret; GEM_TRACE("\n"); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); intel_suspend_gt_powersave(i915); + flush_workqueue(i915->wq); + mutex_lock(&i915->drm.struct_mutex); /* @@ -4947,11 +4603,9 @@ int i915_gem_suspend(struct drm_i915_private *i915) i915_retire_requests(i915); /* ensure we flush after wedging */ mutex_unlock(&i915->drm.struct_mutex); + i915_reset_flush(i915); - intel_uc_suspend(i915); - - cancel_delayed_work_sync(&i915->gpu_error.hangcheck_work); - cancel_delayed_work_sync(&i915->gt.retire_work); + drain_delayed_work(&i915->gt.retire_work); /* * As the idle_work is rearming if it detects a race, play safe and @@ -4959,6 +4613,8 @@ int i915_gem_suspend(struct drm_i915_private *i915) */ drain_delayed_work(&i915->gt.idle_work); + intel_uc_suspend(i915); + /* * Assert that we successfully flushed all the work and * reset the GPU back to its idle, low power state. @@ -4967,12 +4623,12 @@ int i915_gem_suspend(struct drm_i915_private *i915) if (WARN_ON(!intel_engines_are_idle(i915))) i915_gem_set_wedged(i915); /* no hope, discard everything */ - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); return 0; err_unlock: mutex_unlock(&i915->drm.struct_mutex); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); return ret; } @@ -5395,6 +5051,8 @@ int i915_gem_init(struct drm_i915_private *dev_priv) dev_priv->gt.cleanup_engine = intel_engine_cleanup; } + i915_timelines_init(dev_priv); + ret = i915_gem_init_userptr(dev_priv); if (ret) return ret; @@ -5517,8 +5175,10 @@ err_unlock: err_uc_misc: intel_uc_fini_misc(dev_priv); - if (ret != -EIO) + if (ret != -EIO) { i915_gem_cleanup_userptr(dev_priv); + i915_timelines_fini(dev_priv); + } if (ret == -EIO) { mutex_lock(&dev_priv->drm.struct_mutex); @@ -5569,6 +5229,7 @@ void i915_gem_fini(struct drm_i915_private *dev_priv) intel_uc_fini_misc(dev_priv); i915_gem_cleanup_userptr(dev_priv); + i915_timelines_fini(dev_priv); i915_gem_drain_freed_objects(dev_priv); @@ -5671,7 +5332,6 @@ int i915_gem_init_early(struct drm_i915_private *dev_priv) if (!dev_priv->priorities) goto err_dependencies; - INIT_LIST_HEAD(&dev_priv->gt.timelines); INIT_LIST_HEAD(&dev_priv->gt.active_rings); INIT_LIST_HEAD(&dev_priv->gt.closed_vma); @@ -5683,6 +5343,7 @@ int i915_gem_init_early(struct drm_i915_private *dev_priv) i915_gem_idle_work_handler); init_waitqueue_head(&dev_priv->gpu_error.wait_queue); init_waitqueue_head(&dev_priv->gpu_error.reset_queue); + mutex_init(&dev_priv->gpu_error.wedge_mutex); atomic_set(&dev_priv->mm.bsd_engine_dispatch_index, 0); @@ -5714,7 +5375,6 @@ void i915_gem_cleanup_early(struct drm_i915_private *dev_priv) GEM_BUG_ON(!llist_empty(&dev_priv->mm.free_list)); GEM_BUG_ON(atomic_read(&dev_priv->mm.free_count)); WARN_ON(dev_priv->mm.object_count); - WARN_ON(!list_empty(&dev_priv->gt.timelines)); kmem_cache_destroy(dev_priv->priorities); kmem_cache_destroy(dev_priv->dependencies); diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index 5933adbe3d99..280813a4bf82 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -89,6 +89,7 @@ #include <drm/i915_drm.h> #include "i915_drv.h" #include "i915_trace.h" +#include "intel_lrc_reg.h" #include "intel_workarounds.h" #define ALL_L3_SLICES(dev) (1 << NUM_L3_SLICES(dev)) - 1 @@ -321,6 +322,32 @@ static u32 default_desc_template(const struct drm_i915_private *i915, return desc; } +static void intel_context_retire(struct i915_active_request *active, + struct i915_request *rq) +{ + struct intel_context *ce = + container_of(active, typeof(*ce), active_tracker); + + intel_context_unpin(ce); +} + +void +intel_context_init(struct intel_context *ce, + struct i915_gem_context *ctx, + struct intel_engine_cs *engine) +{ + ce->gem_context = ctx; + + INIT_LIST_HEAD(&ce->signal_link); + INIT_LIST_HEAD(&ce->signals); + + /* Use the whole device by default */ + ce->sseu = intel_device_default_sseu(ctx->i915); + + i915_active_request_init(&ce->active_tracker, + NULL, intel_context_retire); +} + static struct i915_gem_context * __create_hw_context(struct drm_i915_private *dev_priv, struct drm_i915_file_private *file_priv) @@ -338,11 +365,8 @@ __create_hw_context(struct drm_i915_private *dev_priv, ctx->i915 = dev_priv; ctx->sched.priority = I915_USER_PRIORITY(I915_PRIORITY_NORMAL); - for (n = 0; n < ARRAY_SIZE(ctx->__engine); n++) { - struct intel_context *ce = &ctx->__engine[n]; - - ce->gem_context = ctx; - } + for (n = 0; n < ARRAY_SIZE(ctx->__engine); n++) + intel_context_init(&ctx->__engine[n], ctx, dev_priv->engine[n]); INIT_RADIX_TREE(&ctx->handles_vma, GFP_KERNEL); INIT_LIST_HEAD(&ctx->handles_list); @@ -645,8 +669,8 @@ last_request_on_engine(struct i915_timeline *timeline, GEM_BUG_ON(timeline == &engine->timeline); - rq = i915_gem_active_raw(&timeline->last_request, - &engine->i915->drm.struct_mutex); + rq = i915_active_request_raw(&timeline->last_request, + &engine->i915->drm.struct_mutex); if (rq && rq->engine == engine) { GEM_TRACE("last request for %s on engine %s: %llx:%llu\n", timeline->name, engine->name, @@ -839,6 +863,56 @@ out: return 0; } +static int get_sseu(struct i915_gem_context *ctx, + struct drm_i915_gem_context_param *args) +{ + struct drm_i915_gem_context_param_sseu user_sseu; + struct intel_engine_cs *engine; + struct intel_context *ce; + int ret; + + if (args->size == 0) + goto out; + else if (args->size < sizeof(user_sseu)) + return -EINVAL; + + if (copy_from_user(&user_sseu, u64_to_user_ptr(args->value), + sizeof(user_sseu))) + return -EFAULT; + + if (user_sseu.flags || user_sseu.rsvd) + return -EINVAL; + + engine = intel_engine_lookup_user(ctx->i915, + user_sseu.engine_class, + user_sseu.engine_instance); + if (!engine) + return -EINVAL; + + /* Only use for mutex here is to serialize get_param and set_param. */ + ret = mutex_lock_interruptible(&ctx->i915->drm.struct_mutex); + if (ret) + return ret; + + ce = to_intel_context(ctx, engine); + + user_sseu.slice_mask = ce->sseu.slice_mask; + user_sseu.subslice_mask = ce->sseu.subslice_mask; + user_sseu.min_eus_per_subslice = ce->sseu.min_eus_per_subslice; + user_sseu.max_eus_per_subslice = ce->sseu.max_eus_per_subslice; + + mutex_unlock(&ctx->i915->drm.struct_mutex); + + if (copy_to_user(u64_to_user_ptr(args->value), &user_sseu, + sizeof(user_sseu))) + return -EFAULT; + +out: + args->size = sizeof(user_sseu); + + return 0; +} + int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data, struct drm_file *file) { @@ -851,15 +925,17 @@ int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data, if (!ctx) return -ENOENT; - args->size = 0; switch (args->param) { case I915_CONTEXT_PARAM_BAN_PERIOD: ret = -EINVAL; break; case I915_CONTEXT_PARAM_NO_ZEROMAP: + args->size = 0; args->value = test_bit(UCONTEXT_NO_ZEROMAP, &ctx->user_flags); break; case I915_CONTEXT_PARAM_GTT_SIZE: + args->size = 0; + if (ctx->ppgtt) args->value = ctx->ppgtt->vm.total; else if (to_i915(dev)->mm.aliasing_ppgtt) @@ -868,14 +944,20 @@ int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data, args->value = to_i915(dev)->ggtt.vm.total; break; case I915_CONTEXT_PARAM_NO_ERROR_CAPTURE: + args->size = 0; args->value = i915_gem_context_no_error_capture(ctx); break; case I915_CONTEXT_PARAM_BANNABLE: + args->size = 0; args->value = i915_gem_context_is_bannable(ctx); break; case I915_CONTEXT_PARAM_PRIORITY: + args->size = 0; args->value = ctx->sched.priority >> I915_USER_PRIORITY_SHIFT; break; + case I915_CONTEXT_PARAM_SSEU: + ret = get_sseu(ctx, args); + break; default: ret = -EINVAL; break; @@ -885,6 +967,281 @@ int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data, return ret; } +static int gen8_emit_rpcs_config(struct i915_request *rq, + struct intel_context *ce, + struct intel_sseu sseu) +{ + u64 offset; + u32 *cs; + + cs = intel_ring_begin(rq, 4); + if (IS_ERR(cs)) + return PTR_ERR(cs); + + offset = i915_ggtt_offset(ce->state) + + LRC_STATE_PN * PAGE_SIZE + + (CTX_R_PWR_CLK_STATE + 1) * 4; + + *cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT; + *cs++ = lower_32_bits(offset); + *cs++ = upper_32_bits(offset); + *cs++ = gen8_make_rpcs(rq->i915, &sseu); + + intel_ring_advance(rq, cs); + + return 0; +} + +static int +gen8_modify_rpcs_gpu(struct intel_context *ce, + struct intel_engine_cs *engine, + struct intel_sseu sseu) +{ + struct drm_i915_private *i915 = engine->i915; + struct i915_request *rq, *prev; + intel_wakeref_t wakeref; + int ret; + + GEM_BUG_ON(!ce->pin_count); + + lockdep_assert_held(&i915->drm.struct_mutex); + + /* Submitting requests etc needs the hw awake. */ + wakeref = intel_runtime_pm_get(i915); + + rq = i915_request_alloc(engine, i915->kernel_context); + if (IS_ERR(rq)) { + ret = PTR_ERR(rq); + goto out_put; + } + + /* Queue this switch after all other activity by this context. */ + prev = i915_active_request_raw(&ce->ring->timeline->last_request, + &i915->drm.struct_mutex); + if (prev && !i915_request_completed(prev)) { + ret = i915_request_await_dma_fence(rq, &prev->fence); + if (ret < 0) + goto out_add; + } + + /* Order all following requests to be after. */ + ret = i915_timeline_set_barrier(ce->ring->timeline, rq); + if (ret) + goto out_add; + + ret = gen8_emit_rpcs_config(rq, ce, sseu); + if (ret) + goto out_add; + + /* + * Guarantee context image and the timeline remains pinned until the + * modifying request is retired by setting the ce activity tracker. + * + * But we only need to take one pin on the account of it. Or in other + * words transfer the pinned ce object to tracked active request. + */ + if (!i915_active_request_isset(&ce->active_tracker)) + __intel_context_pin(ce); + __i915_active_request_set(&ce->active_tracker, rq); + +out_add: + i915_request_add(rq); +out_put: + intel_runtime_pm_put(i915, wakeref); + + return ret; +} + +static int +__i915_gem_context_reconfigure_sseu(struct i915_gem_context *ctx, + struct intel_engine_cs *engine, + struct intel_sseu sseu) +{ + struct intel_context *ce = to_intel_context(ctx, engine); + int ret = 0; + + GEM_BUG_ON(INTEL_GEN(ctx->i915) < 8); + GEM_BUG_ON(engine->id != RCS); + + /* Nothing to do if unmodified. */ + if (!memcmp(&ce->sseu, &sseu, sizeof(sseu))) + return 0; + + /* + * If context is not idle we have to submit an ordered request to modify + * its context image via the kernel context. Pristine and idle contexts + * will be configured on pinning. + */ + if (ce->pin_count) + ret = gen8_modify_rpcs_gpu(ce, engine, sseu); + + if (!ret) + ce->sseu = sseu; + + return ret; +} + +static int +i915_gem_context_reconfigure_sseu(struct i915_gem_context *ctx, + struct intel_engine_cs *engine, + struct intel_sseu sseu) +{ + int ret; + + ret = mutex_lock_interruptible(&ctx->i915->drm.struct_mutex); + if (ret) + return ret; + + ret = __i915_gem_context_reconfigure_sseu(ctx, engine, sseu); + + mutex_unlock(&ctx->i915->drm.struct_mutex); + + return ret; +} + +static int +user_to_context_sseu(struct drm_i915_private *i915, + const struct drm_i915_gem_context_param_sseu *user, + struct intel_sseu *context) +{ + const struct sseu_dev_info *device = &RUNTIME_INFO(i915)->sseu; + + /* No zeros in any field. */ + if (!user->slice_mask || !user->subslice_mask || + !user->min_eus_per_subslice || !user->max_eus_per_subslice) + return -EINVAL; + + /* Max > min. */ + if (user->max_eus_per_subslice < user->min_eus_per_subslice) + return -EINVAL; + + /* + * Some future proofing on the types since the uAPI is wider than the + * current internal implementation. + */ + if (overflows_type(user->slice_mask, context->slice_mask) || + overflows_type(user->subslice_mask, context->subslice_mask) || + overflows_type(user->min_eus_per_subslice, + context->min_eus_per_subslice) || + overflows_type(user->max_eus_per_subslice, + context->max_eus_per_subslice)) + return -EINVAL; + + /* Check validity against hardware. */ + if (user->slice_mask & ~device->slice_mask) + return -EINVAL; + + if (user->subslice_mask & ~device->subslice_mask[0]) + return -EINVAL; + + if (user->max_eus_per_subslice > device->max_eus_per_subslice) + return -EINVAL; + + context->slice_mask = user->slice_mask; + context->subslice_mask = user->subslice_mask; + context->min_eus_per_subslice = user->min_eus_per_subslice; + context->max_eus_per_subslice = user->max_eus_per_subslice; + + /* Part specific restrictions. */ + if (IS_GEN(i915, 11)) { + unsigned int hw_s = hweight8(device->slice_mask); + unsigned int hw_ss_per_s = hweight8(device->subslice_mask[0]); + unsigned int req_s = hweight8(context->slice_mask); + unsigned int req_ss = hweight8(context->subslice_mask); + + /* + * Only full subslice enablement is possible if more than one + * slice is turned on. + */ + if (req_s > 1 && req_ss != hw_ss_per_s) + return -EINVAL; + + /* + * If more than four (SScount bitfield limit) subslices are + * requested then the number has to be even. + */ + if (req_ss > 4 && (req_ss & 1)) + return -EINVAL; + + /* + * If only one slice is enabled and subslice count is below the + * device full enablement, it must be at most half of the all + * available subslices. + */ + if (req_s == 1 && req_ss < hw_ss_per_s && + req_ss > (hw_ss_per_s / 2)) + return -EINVAL; + + /* ABI restriction - VME use case only. */ + + /* All slices or one slice only. */ + if (req_s != 1 && req_s != hw_s) + return -EINVAL; + + /* + * Half subslices or full enablement only when one slice is + * enabled. + */ + if (req_s == 1 && + (req_ss != hw_ss_per_s && req_ss != (hw_ss_per_s / 2))) + return -EINVAL; + + /* No EU configuration changes. */ + if ((user->min_eus_per_subslice != + device->max_eus_per_subslice) || + (user->max_eus_per_subslice != + device->max_eus_per_subslice)) + return -EINVAL; + } + + return 0; +} + +static int set_sseu(struct i915_gem_context *ctx, + struct drm_i915_gem_context_param *args) +{ + struct drm_i915_private *i915 = ctx->i915; + struct drm_i915_gem_context_param_sseu user_sseu; + struct intel_engine_cs *engine; + struct intel_sseu sseu; + int ret; + + if (args->size < sizeof(user_sseu)) + return -EINVAL; + + if (!IS_GEN(i915, 11)) + return -ENODEV; + + if (copy_from_user(&user_sseu, u64_to_user_ptr(args->value), + sizeof(user_sseu))) + return -EFAULT; + + if (user_sseu.flags || user_sseu.rsvd) + return -EINVAL; + + engine = intel_engine_lookup_user(i915, + user_sseu.engine_class, + user_sseu.engine_instance); + if (!engine) + return -EINVAL; + + /* Only render engine supports RPCS configuration. */ + if (engine->class != RENDER_CLASS) + return -ENODEV; + + ret = user_to_context_sseu(i915, &user_sseu, &sseu); + if (ret) + return ret; + + ret = i915_gem_context_reconfigure_sseu(ctx, engine, sseu); + if (ret) + return ret; + + args->size = sizeof(user_sseu); + + return 0; +} + int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data, struct drm_file *file) { @@ -947,7 +1304,9 @@ int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data, I915_USER_PRIORITY(priority); } break; - + case I915_CONTEXT_PARAM_SSEU: + ret = set_sseu(ctx, args); + break; default: ret = -EINVAL; break; diff --git a/drivers/gpu/drm/i915/i915_gem_context.h b/drivers/gpu/drm/i915/i915_gem_context.h index f6d870b1f73e..ca150a764c24 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.h +++ b/drivers/gpu/drm/i915/i915_gem_context.h @@ -31,6 +31,7 @@ #include "i915_gem.h" #include "i915_scheduler.h" +#include "intel_device_info.h" struct pid; @@ -53,6 +54,16 @@ struct intel_context_ops { void (*destroy)(struct intel_context *ce); }; +/* + * Powergating configuration for a particular (context,engine). + */ +struct intel_sseu { + u8 slice_mask; + u8 subslice_mask; + u8 min_eus_per_subslice; + u8 max_eus_per_subslice; +}; + /** * struct i915_gem_context - client state * @@ -164,13 +175,24 @@ struct i915_gem_context { struct intel_context { struct i915_gem_context *gem_context; struct intel_engine_cs *active; + struct list_head signal_link; + struct list_head signals; struct i915_vma *state; struct intel_ring *ring; u32 *lrc_reg_state; u64 lrc_desc; int pin_count; + /** + * active_tracker: Active tracker for the external rq activity + * on this intel_context object. + */ + struct i915_active_request active_tracker; + const struct intel_context_ops *ops; + + /** sseu: Control eu/slice partitioning */ + struct intel_sseu sseu; } __engine[I915_NUM_ENGINES]; /** ring_size: size for allocating the per-engine ring buffer */ @@ -364,4 +386,8 @@ static inline void i915_gem_context_put(struct i915_gem_context *ctx) kref_put(&ctx->ref, i915_gem_context_release); } +void intel_context_init(struct intel_context *ce, + struct i915_gem_context *ctx, + struct intel_engine_cs *engine); + #endif /* !__I915_GEM_CONTEXT_H__ */ diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c index f6855401f247..68d74c50ac39 100644 --- a/drivers/gpu/drm/i915/i915_gem_evict.c +++ b/drivers/gpu/drm/i915/i915_gem_evict.c @@ -126,31 +126,25 @@ i915_gem_evict_something(struct i915_address_space *vm, struct drm_i915_private *dev_priv = vm->i915; struct drm_mm_scan scan; struct list_head eviction_list; - struct list_head *phases[] = { - &vm->inactive_list, - &vm->active_list, - NULL, - }, **phase; struct i915_vma *vma, *next; struct drm_mm_node *node; enum drm_mm_insert_mode mode; + struct i915_vma *active; int ret; lockdep_assert_held(&vm->i915->drm.struct_mutex); trace_i915_gem_evict(vm, min_size, alignment, flags); /* - * The goal is to evict objects and amalgamate space in LRU order. - * The oldest idle objects reside on the inactive list, which is in - * retirement order. The next objects to retire are those in flight, - * on the active list, again in retirement order. + * The goal is to evict objects and amalgamate space in rough LRU order. + * Since both active and inactive objects reside on the same list, + * in a mix of creation and last scanned order, as we process the list + * we sort it into inactive/active, which keeps the active portion + * in a rough MRU order. * * The retirement sequence is thus: - * 1. Inactive objects (already retired) - * 2. Active objects (will stall on unbinding) - * - * On each list, the oldest objects lie at the HEAD with the freshest - * object on the TAIL. + * 1. Inactive objects (already retired, random order) + * 2. Active objects (will stall on unbinding, oldest scanned first) */ mode = DRM_MM_INSERT_BEST; if (flags & PIN_HIGH) @@ -169,17 +163,46 @@ i915_gem_evict_something(struct i915_address_space *vm, */ if (!(flags & PIN_NONBLOCK)) i915_retire_requests(dev_priv); - else - phases[1] = NULL; search_again: + active = NULL; INIT_LIST_HEAD(&eviction_list); - phase = phases; - do { - list_for_each_entry(vma, *phase, vm_link) - if (mark_free(&scan, vma, flags, &eviction_list)) - goto found; - } while (*++phase); + list_for_each_entry_safe(vma, next, &vm->bound_list, vm_link) { + /* + * We keep this list in a rough least-recently scanned order + * of active elements (inactive elements are cheap to reap). + * New entries are added to the end, and we move anything we + * scan to the end. The assumption is that the working set + * of applications is either steady state (and thanks to the + * userspace bo cache it almost always is) or volatile and + * frequently replaced after a frame, which are self-evicting! + * Given that assumption, the MRU order of the scan list is + * fairly static, and keeping it in least-recently scan order + * is suitable. + * + * To notice when we complete one full cycle, we record the + * first active element seen, before moving it to the tail. + */ + if (i915_vma_is_active(vma)) { + if (vma == active) { + if (flags & PIN_NONBLOCK) + break; + + active = ERR_PTR(-EAGAIN); + } + + if (active != ERR_PTR(-EAGAIN)) { + if (!active) + active = vma; + + list_move_tail(&vma->vm_link, &vm->bound_list); + continue; + } + } + + if (mark_free(&scan, vma, flags, &eviction_list)) + goto found; + } /* Nothing found, clean up and bail out! */ list_for_each_entry_safe(vma, next, &eviction_list, evict_link) { @@ -388,11 +411,6 @@ int i915_gem_evict_for_node(struct i915_address_space *vm, */ int i915_gem_evict_vm(struct i915_address_space *vm) { - struct list_head *phases[] = { - &vm->inactive_list, - &vm->active_list, - NULL - }, **phase; struct list_head eviction_list; struct i915_vma *vma, *next; int ret; @@ -412,16 +430,15 @@ int i915_gem_evict_vm(struct i915_address_space *vm) } INIT_LIST_HEAD(&eviction_list); - phase = phases; - do { - list_for_each_entry(vma, *phase, vm_link) { - if (i915_vma_is_pinned(vma)) - continue; + mutex_lock(&vm->mutex); + list_for_each_entry(vma, &vm->bound_list, vm_link) { + if (i915_vma_is_pinned(vma)) + continue; - __i915_vma_pin(vma); - list_add(&vma->evict_link, &eviction_list); - } - } while (*++phase); + __i915_vma_pin(vma); + list_add(&vma->evict_link, &eviction_list); + } + mutex_unlock(&vm->mutex); ret = 0; list_for_each_entry_safe(vma, next, &eviction_list, evict_link) { diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index e7994505d850..02adcaf6ebea 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -753,6 +753,68 @@ static int eb_select_context(struct i915_execbuffer *eb) return 0; } +static struct i915_request *__eb_wait_for_ring(struct intel_ring *ring) +{ + struct i915_request *rq; + + /* + * Completely unscientific finger-in-the-air estimates for suitable + * maximum user request size (to avoid blocking) and then backoff. + */ + if (intel_ring_update_space(ring) >= PAGE_SIZE) + return NULL; + + /* + * Find a request that after waiting upon, there will be at least half + * the ring available. The hysteresis allows us to compete for the + * shared ring and should mean that we sleep less often prior to + * claiming our resources, but not so long that the ring completely + * drains before we can submit our next request. + */ + list_for_each_entry(rq, &ring->request_list, ring_link) { + if (__intel_ring_space(rq->postfix, + ring->emit, ring->size) > ring->size / 2) + break; + } + if (&rq->ring_link == &ring->request_list) + return NULL; /* weird, we will check again later for real */ + + return i915_request_get(rq); +} + +static int eb_wait_for_ring(const struct i915_execbuffer *eb) +{ + const struct intel_context *ce; + struct i915_request *rq; + int ret = 0; + + /* + * Apply a light amount of backpressure to prevent excessive hogs + * from blocking waiting for space whilst holding struct_mutex and + * keeping all of their resources pinned. + */ + + ce = to_intel_context(eb->ctx, eb->engine); + if (!ce->ring) /* first use, assume empty! */ + return 0; + + rq = __eb_wait_for_ring(ce->ring); + if (rq) { + mutex_unlock(&eb->i915->drm.struct_mutex); + + if (i915_request_wait(rq, + I915_WAIT_INTERRUPTIBLE, + MAX_SCHEDULE_TIMEOUT) < 0) + ret = -EINTR; + + i915_request_put(rq); + + mutex_lock(&eb->i915->drm.struct_mutex); + } + + return ret; +} + static int eb_lookup_vmas(struct i915_execbuffer *eb) { struct radix_tree_root *handles_vma = &eb->ctx->handles_vma; @@ -1976,6 +2038,18 @@ static int eb_submit(struct i915_execbuffer *eb) return err; } + /* + * After we completed waiting for other engines (using HW semaphores) + * then we can signal that this request/batch is ready to run. This + * allows us to determine if the batch is still waiting on the GPU + * or actually running by checking the breadcrumb. + */ + if (eb->engine->emit_init_breadcrumb) { + err = eb->engine->emit_init_breadcrumb(eb->request); + if (err) + return err; + } + err = eb->engine->emit_bb_start(eb->request, eb->batch->node.start + eb->batch_start_offset, @@ -2202,6 +2276,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, struct i915_execbuffer eb; struct dma_fence *in_fence = NULL; struct sync_file *out_fence = NULL; + intel_wakeref_t wakeref; int out_fence_fd = -1; int err; @@ -2272,12 +2347,16 @@ i915_gem_do_execbuffer(struct drm_device *dev, * wakeref that we hold until the GPU has been idle for at least * 100ms. */ - intel_runtime_pm_get(eb.i915); + wakeref = intel_runtime_pm_get(eb.i915); err = i915_mutex_lock_interruptible(dev); if (err) goto err_rpm; + err = eb_wait_for_ring(&eb); /* may temporarily drop struct_mutex */ + if (unlikely(err)) + goto err_unlock; + err = eb_relocate(&eb); if (err) { /* @@ -2422,9 +2501,10 @@ err_batch_unpin: err_vma: if (eb.exec) eb_release_vmas(&eb); +err_unlock: mutex_unlock(&dev->struct_mutex); err_rpm: - intel_runtime_pm_put(eb.i915); + intel_runtime_pm_put(eb.i915, wakeref); i915_gem_context_put(eb.ctx); err_destroy: eb_destroy(&eb); diff --git a/drivers/gpu/drm/i915/i915_gem_fence_reg.c b/drivers/gpu/drm/i915/i915_gem_fence_reg.c index d67c07cdd0b8..e037e94792f3 100644 --- a/drivers/gpu/drm/i915/i915_gem_fence_reg.c +++ b/drivers/gpu/drm/i915/i915_gem_fence_reg.c @@ -209,6 +209,7 @@ static void fence_write(struct drm_i915_fence_reg *fence, static int fence_update(struct drm_i915_fence_reg *fence, struct i915_vma *vma) { + intel_wakeref_t wakeref; int ret; if (vma) { @@ -222,7 +223,7 @@ static int fence_update(struct drm_i915_fence_reg *fence, i915_gem_object_get_tiling(vma->obj))) return -EINVAL; - ret = i915_gem_active_retire(&vma->last_fence, + ret = i915_active_request_retire(&vma->last_fence, &vma->obj->base.dev->struct_mutex); if (ret) return ret; @@ -231,7 +232,7 @@ static int fence_update(struct drm_i915_fence_reg *fence, if (fence->vma) { struct i915_vma *old = fence->vma; - ret = i915_gem_active_retire(&old->last_fence, + ret = i915_active_request_retire(&old->last_fence, &old->obj->base.dev->struct_mutex); if (ret) return ret; @@ -256,9 +257,10 @@ static int fence_update(struct drm_i915_fence_reg *fence, * If the device is currently powered down, we will defer the write * to the runtime resume, see i915_gem_restore_fences(). */ - if (intel_runtime_pm_get_if_in_use(fence->i915)) { + wakeref = intel_runtime_pm_get_if_in_use(fence->i915); + if (wakeref) { fence_write(fence, vma); - intel_runtime_pm_put(fence->i915); + intel_runtime_pm_put(fence->i915, wakeref); } if (vma) { @@ -553,8 +555,8 @@ void i915_gem_restore_fences(struct drm_i915_private *dev_priv) void i915_gem_detect_bit_6_swizzle(struct drm_i915_private *dev_priv) { - uint32_t swizzle_x = I915_BIT_6_SWIZZLE_UNKNOWN; - uint32_t swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN; + u32 swizzle_x = I915_BIT_6_SWIZZLE_UNKNOWN; + u32 swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN; if (INTEL_GEN(dev_priv) >= 8 || IS_VALLEYVIEW(dev_priv)) { /* @@ -577,7 +579,7 @@ i915_gem_detect_bit_6_swizzle(struct drm_i915_private *dev_priv) swizzle_y = I915_BIT_6_SWIZZLE_NONE; } } else { - uint32_t dimm_c0, dimm_c1; + u32 dimm_c0, dimm_c1; dimm_c0 = I915_READ(MAD_DIMM_C0); dimm_c1 = I915_READ(MAD_DIMM_C1); dimm_c0 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK; @@ -609,7 +611,7 @@ i915_gem_detect_bit_6_swizzle(struct drm_i915_private *dev_priv) swizzle_y = I915_BIT_6_SWIZZLE_NONE; } else if (IS_MOBILE(dev_priv) || IS_I915G(dev_priv) || IS_I945G(dev_priv)) { - uint32_t dcc; + u32 dcc; /* On 9xx chipsets, channel interleave by the CPU is * determined by DCC. For single-channel, neither the CPU diff --git a/drivers/gpu/drm/i915/i915_gem_fence_reg.h b/drivers/gpu/drm/i915/i915_gem_fence_reg.h index 99a31ded4dfd..09dcaf14121b 100644 --- a/drivers/gpu/drm/i915/i915_gem_fence_reg.h +++ b/drivers/gpu/drm/i915/i915_gem_fence_reg.h @@ -50,4 +50,3 @@ struct drm_i915_fence_reg { }; #endif - diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index a8807fbed0aa..d646d37eec2f 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -37,6 +37,7 @@ #include "i915_drv.h" #include "i915_vgpu.h" +#include "i915_reset.h" #include "i915_trace.h" #include "intel_drv.h" #include "intel_frontbuffer.h" @@ -473,8 +474,7 @@ static void vm_free_page(struct i915_address_space *vm, struct page *page) spin_unlock(&vm->free_pages.lock); } -static void i915_address_space_init(struct i915_address_space *vm, - struct drm_i915_private *dev_priv) +static void i915_address_space_init(struct i915_address_space *vm, int subclass) { /* * The vm->mutex must be reclaim safe (for use in the shrinker). @@ -482,6 +482,7 @@ static void i915_address_space_init(struct i915_address_space *vm, * attempt holding the lock is immediately reported by lockdep. */ mutex_init(&vm->mutex); + lockdep_set_subclass(&vm->mutex, subclass); i915_gem_shrinker_taints_mutex(vm->i915, &vm->mutex); GEM_BUG_ON(!vm->total); @@ -490,9 +491,8 @@ static void i915_address_space_init(struct i915_address_space *vm, stash_init(&vm->free_pages); - INIT_LIST_HEAD(&vm->active_list); - INIT_LIST_HEAD(&vm->inactive_list); INIT_LIST_HEAD(&vm->unbound_list); + INIT_LIST_HEAD(&vm->bound_list); } static void i915_address_space_fini(struct i915_address_space *vm) @@ -1547,7 +1547,7 @@ static struct i915_hw_ppgtt *gen8_ppgtt_create(struct drm_i915_private *i915) /* From bdw, there is support for read-only pages in the PPGTT. */ ppgtt->vm.has_read_only = true; - i915_address_space_init(&ppgtt->vm, i915); + i915_address_space_init(&ppgtt->vm, VM_CLASS_PPGTT); /* There are only few exceptions for gen >=6. chv and bxt. * And we are not sure about the latter so play safe for now. @@ -1917,21 +1917,23 @@ static struct i915_vma *pd_vma_create(struct gen6_hw_ppgtt *ppgtt, int size) if (!vma) return ERR_PTR(-ENOMEM); - init_request_active(&vma->last_fence, NULL); + i915_active_init(i915, &vma->active, NULL); + INIT_ACTIVE_REQUEST(&vma->last_fence); vma->vm = &ggtt->vm; vma->ops = &pd_vma_ops; vma->private = ppgtt; - vma->active = RB_ROOT; - vma->size = size; vma->fence_size = size; vma->flags = I915_VMA_GGTT; vma->ggtt_view.type = I915_GGTT_VIEW_ROTATED; /* prevent fencing */ INIT_LIST_HEAD(&vma->obj_link); + + mutex_lock(&vma->vm->mutex); list_add(&vma->vm_link, &vma->vm->unbound_list); + mutex_unlock(&vma->vm->mutex); return vma; } @@ -1996,7 +1998,7 @@ static struct i915_hw_ppgtt *gen6_ppgtt_create(struct drm_i915_private *i915) ppgtt->base.vm.total = I915_PDES * GEN6_PTES * I915_GTT_PAGE_SIZE; - i915_address_space_init(&ppgtt->base.vm, i915); + i915_address_space_init(&ppgtt->base.vm, VM_CLASS_PPGTT); ppgtt->base.vm.allocate_va_range = gen6_alloc_va_range; ppgtt->base.vm.clear_range = gen6_ppgtt_clear_range; @@ -2110,8 +2112,7 @@ void i915_ppgtt_close(struct i915_address_space *vm) static void ppgtt_destroy_vma(struct i915_address_space *vm) { struct list_head *phases[] = { - &vm->active_list, - &vm->inactive_list, + &vm->bound_list, &vm->unbound_list, NULL, }, **phase; @@ -2134,8 +2135,7 @@ void i915_ppgtt_release(struct kref *kref) ppgtt_destroy_vma(&ppgtt->vm); - GEM_BUG_ON(!list_empty(&ppgtt->vm.active_list)); - GEM_BUG_ON(!list_empty(&ppgtt->vm.inactive_list)); + GEM_BUG_ON(!list_empty(&ppgtt->vm.bound_list)); GEM_BUG_ON(!list_empty(&ppgtt->vm.unbound_list)); ppgtt->vm.cleanup(&ppgtt->vm); @@ -2527,6 +2527,7 @@ static int ggtt_bind_vma(struct i915_vma *vma, { struct drm_i915_private *i915 = vma->vm->i915; struct drm_i915_gem_object *obj = vma->obj; + intel_wakeref_t wakeref; u32 pte_flags; /* Applicable to VLV (gen8+ do not support RO in the GGTT) */ @@ -2534,9 +2535,8 @@ static int ggtt_bind_vma(struct i915_vma *vma, if (i915_gem_object_is_readonly(obj)) pte_flags |= PTE_READ_ONLY; - intel_runtime_pm_get(i915); - vma->vm->insert_entries(vma->vm, vma, cache_level, pte_flags); - intel_runtime_pm_put(i915); + with_intel_runtime_pm(i915, wakeref) + vma->vm->insert_entries(vma->vm, vma, cache_level, pte_flags); vma->page_sizes.gtt = I915_GTT_PAGE_SIZE; @@ -2553,10 +2553,10 @@ static int ggtt_bind_vma(struct i915_vma *vma, static void ggtt_unbind_vma(struct i915_vma *vma) { struct drm_i915_private *i915 = vma->vm->i915; + intel_wakeref_t wakeref; - intel_runtime_pm_get(i915); - vma->vm->clear_range(vma->vm, vma->node.start, vma->size); - intel_runtime_pm_put(i915); + with_intel_runtime_pm(i915, wakeref) + vma->vm->clear_range(vma->vm, vma->node.start, vma->size); } static int aliasing_gtt_bind_vma(struct i915_vma *vma, @@ -2588,9 +2588,12 @@ static int aliasing_gtt_bind_vma(struct i915_vma *vma, } if (flags & I915_VMA_GLOBAL_BIND) { - intel_runtime_pm_get(i915); - vma->vm->insert_entries(vma->vm, vma, cache_level, pte_flags); - intel_runtime_pm_put(i915); + intel_wakeref_t wakeref; + + with_intel_runtime_pm(i915, wakeref) { + vma->vm->insert_entries(vma->vm, vma, + cache_level, pte_flags); + } } return 0; @@ -2601,9 +2604,11 @@ static void aliasing_gtt_unbind_vma(struct i915_vma *vma) struct drm_i915_private *i915 = vma->vm->i915; if (vma->flags & I915_VMA_GLOBAL_BIND) { - intel_runtime_pm_get(i915); - vma->vm->clear_range(vma->vm, vma->node.start, vma->size); - intel_runtime_pm_put(i915); + struct i915_address_space *vm = vma->vm; + intel_wakeref_t wakeref; + + with_intel_runtime_pm(i915, wakeref) + vm->clear_range(vm, vma->node.start, vma->size); } if (vma->flags & I915_VMA_LOCAL_BIND) { @@ -2795,8 +2800,7 @@ void i915_ggtt_cleanup_hw(struct drm_i915_private *dev_priv) mutex_lock(&dev_priv->drm.struct_mutex); i915_gem_fini_aliasing_ppgtt(dev_priv); - GEM_BUG_ON(!list_empty(&ggtt->vm.active_list)); - list_for_each_entry_safe(vma, vn, &ggtt->vm.inactive_list, vm_link) + list_for_each_entry_safe(vma, vn, &ggtt->vm.bound_list, vm_link) WARN_ON(i915_vma_unbind(vma)); if (drm_mm_node_allocated(&ggtt->error_capture)) @@ -3227,7 +3231,8 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt) ggtt->vm.insert_entries = gen8_ggtt_insert_entries; /* Serialize GTT updates with aperture access on BXT if VT-d is on. */ - if (intel_ggtt_update_needs_vtd_wa(dev_priv)) { + if (intel_ggtt_update_needs_vtd_wa(dev_priv) || + IS_CHERRYVIEW(dev_priv) /* fails with concurrent use/update */) { ggtt->vm.insert_entries = bxt_vtd_ggtt_insert_entries__BKL; ggtt->vm.insert_page = bxt_vtd_ggtt_insert_page__BKL; if (ggtt->vm.clear_range != nop_clear_range) @@ -3428,7 +3433,7 @@ int i915_ggtt_init_hw(struct drm_i915_private *dev_priv) * and beyond the end of the GTT if we do not provide a guard. */ mutex_lock(&dev_priv->drm.struct_mutex); - i915_address_space_init(&ggtt->vm, dev_priv); + i915_address_space_init(&ggtt->vm, VM_CLASS_GGTT); ggtt->vm.is_ggtt = true; @@ -3501,32 +3506,39 @@ void i915_gem_restore_gtt_mappings(struct drm_i915_private *dev_priv) i915_check_and_clear_faults(dev_priv); + mutex_lock(&ggtt->vm.mutex); + /* First fill our portion of the GTT with scratch pages */ ggtt->vm.clear_range(&ggtt->vm, 0, ggtt->vm.total); - ggtt->vm.closed = true; /* skip rewriting PTE on VMA unbind */ /* clflush objects bound into the GGTT and rebind them. */ - GEM_BUG_ON(!list_empty(&ggtt->vm.active_list)); - list_for_each_entry_safe(vma, vn, &ggtt->vm.inactive_list, vm_link) { + list_for_each_entry_safe(vma, vn, &ggtt->vm.bound_list, vm_link) { struct drm_i915_gem_object *obj = vma->obj; if (!(vma->flags & I915_VMA_GLOBAL_BIND)) continue; + mutex_unlock(&ggtt->vm.mutex); + if (!i915_vma_unbind(vma)) - continue; + goto lock; WARN_ON(i915_vma_bind(vma, obj ? obj->cache_level : 0, PIN_UPDATE)); if (obj) WARN_ON(i915_gem_object_set_to_gtt_domain(obj, false)); + +lock: + mutex_lock(&ggtt->vm.mutex); } ggtt->vm.closed = false; i915_ggtt_invalidate(dev_priv); + mutex_unlock(&ggtt->vm.mutex); + if (INTEL_GEN(dev_priv) >= 8) { struct intel_ppat *ppat = &dev_priv->ppat; diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h index e2360f16427a..03ade71b8d9a 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.h +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h @@ -39,6 +39,7 @@ #include <linux/pagevec.h> #include "i915_request.h" +#include "i915_reset.h" #include "i915_selftest.h" #include "i915_timeline.h" @@ -288,6 +289,8 @@ struct i915_address_space { bool closed; struct mutex mutex; /* protects vma and our lists */ +#define VM_CLASS_GGTT 0 +#define VM_CLASS_PPGTT 1 u64 scratch_pte; struct i915_page_dma scratch_page; @@ -296,32 +299,12 @@ struct i915_address_space { struct i915_page_directory_pointer *scratch_pdp; /* GEN8+ & 48b PPGTT */ /** - * List of objects currently involved in rendering. - * - * Includes buffers having the contents of their GPU caches - * flushed, not necessarily primitives. last_read_req - * represents when the rendering involved will be completed. - * - * A reference is held on the buffer while on this list. + * List of vma currently bound. */ - struct list_head active_list; + struct list_head bound_list; /** - * LRU list of objects which are not in the ringbuffer and - * are ready to unbind, but are still in the GTT. - * - * last_read_req is NULL while an object is in this list. - * - * A reference is not held on the buffer while on this list, - * as merely being GTT-bound shouldn't prevent its being - * freed, and we'll pull it off the list in the free path. - */ - struct list_head inactive_list; - - /** - * List of vma that have been unbound. - * - * A reference is not held on the buffer while on this list. + * List of vma that are not unbound. */ struct list_head unbound_list; @@ -659,19 +642,19 @@ int i915_gem_gtt_insert(struct i915_address_space *vm, /* Flags used by pin/bind&friends. */ #define PIN_NONBLOCK BIT_ULL(0) -#define PIN_MAPPABLE BIT_ULL(1) -#define PIN_ZONE_4G BIT_ULL(2) -#define PIN_NONFAULT BIT_ULL(3) -#define PIN_NOEVICT BIT_ULL(4) - -#define PIN_MBZ BIT_ULL(5) /* I915_VMA_PIN_OVERFLOW */ -#define PIN_GLOBAL BIT_ULL(6) /* I915_VMA_GLOBAL_BIND */ -#define PIN_USER BIT_ULL(7) /* I915_VMA_LOCAL_BIND */ -#define PIN_UPDATE BIT_ULL(8) - -#define PIN_HIGH BIT_ULL(9) -#define PIN_OFFSET_BIAS BIT_ULL(10) -#define PIN_OFFSET_FIXED BIT_ULL(11) +#define PIN_NONFAULT BIT_ULL(1) +#define PIN_NOEVICT BIT_ULL(2) +#define PIN_MAPPABLE BIT_ULL(3) +#define PIN_ZONE_4G BIT_ULL(4) +#define PIN_HIGH BIT_ULL(5) +#define PIN_OFFSET_BIAS BIT_ULL(6) +#define PIN_OFFSET_FIXED BIT_ULL(7) + +#define PIN_MBZ BIT_ULL(8) /* I915_VMA_PIN_OVERFLOW */ +#define PIN_GLOBAL BIT_ULL(9) /* I915_VMA_GLOBAL_BIND */ +#define PIN_USER BIT_ULL(10) /* I915_VMA_LOCAL_BIND */ +#define PIN_UPDATE BIT_ULL(11) + #define PIN_OFFSET_MASK (-I915_GTT_PAGE_SIZE) #endif diff --git a/drivers/gpu/drm/i915/i915_gem_object.h b/drivers/gpu/drm/i915/i915_gem_object.h index ff3da64470dd..fab040331cdb 100644 --- a/drivers/gpu/drm/i915/i915_gem_object.h +++ b/drivers/gpu/drm/i915/i915_gem_object.h @@ -57,6 +57,7 @@ struct drm_i915_gem_object_ops { #define I915_GEM_OBJECT_HAS_STRUCT_PAGE BIT(0) #define I915_GEM_OBJECT_IS_SHRINKABLE BIT(1) #define I915_GEM_OBJECT_IS_PROXY BIT(2) +#define I915_GEM_OBJECT_ASYNC_CANCEL BIT(3) /* Interface between the GEM object and its backing storage. * get_pages() is called once prior to the use of the associated set @@ -86,24 +87,33 @@ struct drm_i915_gem_object { const struct drm_i915_gem_object_ops *ops; - /** - * @vma_list: List of VMAs backed by this object - * - * The VMA on this list are ordered by type, all GGTT vma are placed - * at the head and all ppGTT vma are placed at the tail. The different - * types of GGTT vma are unordered between themselves, use the - * @vma_tree (which has a defined order between all VMA) to find an - * exact match. - */ - struct list_head vma_list; - /** - * @vma_tree: Ordered tree of VMAs backed by this object - * - * All VMA created for this object are placed in the @vma_tree for - * fast retrieval via a binary search in i915_vma_instance(). - * They are also added to @vma_list for easy iteration. - */ - struct rb_root vma_tree; + struct { + /** + * @vma.lock: protect the list/tree of vmas + */ + spinlock_t lock; + + /** + * @vma.list: List of VMAs backed by this object + * + * The VMA on this list are ordered by type, all GGTT vma are + * placed at the head and all ppGTT vma are placed at the tail. + * The different types of GGTT vma are unordered between + * themselves, use the @vma.tree (which has a defined order + * between all VMA) to quickly find an exact match. + */ + struct list_head list; + + /** + * @vma.tree: Ordered tree of VMAs backed by this object + * + * All VMA created for this object are placed in the @vma.tree + * for fast retrieval via a binary search in + * i915_vma_instance(). They are also added to @vma.list for + * easy iteration. + */ + struct rb_root tree; + } vma; /** * @lut_list: List of vma lookup entries in use for this object. @@ -165,7 +175,7 @@ struct drm_i915_gem_object { atomic_t frontbuffer_bits; unsigned int frontbuffer_ggtt_origin; /* write once */ - struct i915_gem_active frontbuffer_write; + struct i915_active_request frontbuffer_write; /** Current tiling stride for the object, if it's tiled. */ unsigned int tiling_and_stride; @@ -388,6 +398,12 @@ i915_gem_object_is_proxy(const struct drm_i915_gem_object *obj) } static inline bool +i915_gem_object_needs_async_cancel(const struct drm_i915_gem_object *obj) +{ + return obj->ops->flags & I915_GEM_OBJECT_ASYNC_CANCEL; +} + +static inline bool i915_gem_object_is_active(const struct drm_i915_gem_object *obj) { return obj->active_count; diff --git a/drivers/gpu/drm/i915/i915_gem_shrinker.c b/drivers/gpu/drm/i915/i915_gem_shrinker.c index 6cc2b964c955..6da795c7e62e 100644 --- a/drivers/gpu/drm/i915/i915_gem_shrinker.c +++ b/drivers/gpu/drm/i915/i915_gem_shrinker.c @@ -153,6 +153,7 @@ i915_gem_shrink(struct drm_i915_private *i915, { &i915->mm.bound_list, I915_SHRINK_BOUND }, { NULL, 0 }, }, *phase; + intel_wakeref_t wakeref = 0; unsigned long count = 0; unsigned long scanned = 0; bool unlock; @@ -182,9 +183,11 @@ i915_gem_shrink(struct drm_i915_private *i915, * device just to recover a little memory. If absolutely necessary, * we will force the wake during oom-notifier. */ - if ((flags & I915_SHRINK_BOUND) && - !intel_runtime_pm_get_if_in_use(i915)) - flags &= ~I915_SHRINK_BOUND; + if (flags & I915_SHRINK_BOUND) { + wakeref = intel_runtime_pm_get_if_in_use(i915); + if (!wakeref) + flags &= ~I915_SHRINK_BOUND; + } /* * As we may completely rewrite the (un)bound list whilst unbinding @@ -265,7 +268,7 @@ i915_gem_shrink(struct drm_i915_private *i915, } if (flags & I915_SHRINK_BOUND) - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); i915_retire_requests(i915); @@ -292,14 +295,15 @@ i915_gem_shrink(struct drm_i915_private *i915, */ unsigned long i915_gem_shrink_all(struct drm_i915_private *i915) { - unsigned long freed; - - intel_runtime_pm_get(i915); - freed = i915_gem_shrink(i915, -1UL, NULL, - I915_SHRINK_BOUND | - I915_SHRINK_UNBOUND | - I915_SHRINK_ACTIVE); - intel_runtime_pm_put(i915); + intel_wakeref_t wakeref; + unsigned long freed = 0; + + with_intel_runtime_pm(i915, wakeref) { + freed = i915_gem_shrink(i915, -1UL, NULL, + I915_SHRINK_BOUND | + I915_SHRINK_UNBOUND | + I915_SHRINK_ACTIVE); + } return freed; } @@ -370,14 +374,16 @@ i915_gem_shrinker_scan(struct shrinker *shrinker, struct shrink_control *sc) I915_SHRINK_BOUND | I915_SHRINK_UNBOUND); if (sc->nr_scanned < sc->nr_to_scan && current_is_kswapd()) { - intel_runtime_pm_get(i915); - freed += i915_gem_shrink(i915, - sc->nr_to_scan - sc->nr_scanned, - &sc->nr_scanned, - I915_SHRINK_ACTIVE | - I915_SHRINK_BOUND | - I915_SHRINK_UNBOUND); - intel_runtime_pm_put(i915); + intel_wakeref_t wakeref; + + with_intel_runtime_pm(i915, wakeref) { + freed += i915_gem_shrink(i915, + sc->nr_to_scan - sc->nr_scanned, + &sc->nr_scanned, + I915_SHRINK_ACTIVE | + I915_SHRINK_BOUND | + I915_SHRINK_UNBOUND); + } } shrinker_unlock(i915, unlock); @@ -392,12 +398,13 @@ i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr) container_of(nb, struct drm_i915_private, mm.oom_notifier); struct drm_i915_gem_object *obj; unsigned long unevictable, bound, unbound, freed_pages; + intel_wakeref_t wakeref; - intel_runtime_pm_get(i915); - freed_pages = i915_gem_shrink(i915, -1UL, NULL, - I915_SHRINK_BOUND | - I915_SHRINK_UNBOUND); - intel_runtime_pm_put(i915); + freed_pages = 0; + with_intel_runtime_pm(i915, wakeref) + freed_pages += i915_gem_shrink(i915, -1UL, NULL, + I915_SHRINK_BOUND | + I915_SHRINK_UNBOUND); /* Because we may be allocating inside our own driver, we cannot * assert that there are no objects with pinned pages that are not @@ -435,6 +442,7 @@ i915_gem_shrinker_vmap(struct notifier_block *nb, unsigned long event, void *ptr container_of(nb, struct drm_i915_private, mm.vmap_notifier); struct i915_vma *vma, *next; unsigned long freed_pages = 0; + intel_wakeref_t wakeref; bool unlock; if (!shrinker_lock(i915, 0, &unlock)) @@ -446,20 +454,27 @@ i915_gem_shrinker_vmap(struct notifier_block *nb, unsigned long event, void *ptr MAX_SCHEDULE_TIMEOUT)) goto out; - intel_runtime_pm_get(i915); - freed_pages += i915_gem_shrink(i915, -1UL, NULL, - I915_SHRINK_BOUND | - I915_SHRINK_UNBOUND | - I915_SHRINK_VMAPS); - intel_runtime_pm_put(i915); + with_intel_runtime_pm(i915, wakeref) + freed_pages += i915_gem_shrink(i915, -1UL, NULL, + I915_SHRINK_BOUND | + I915_SHRINK_UNBOUND | + I915_SHRINK_VMAPS); /* We also want to clear any cached iomaps as they wrap vmap */ + mutex_lock(&i915->ggtt.vm.mutex); list_for_each_entry_safe(vma, next, - &i915->ggtt.vm.inactive_list, vm_link) { + &i915->ggtt.vm.bound_list, vm_link) { unsigned long count = vma->node.size >> PAGE_SHIFT; - if (vma->iomap && i915_vma_unbind(vma) == 0) + + if (!vma->iomap || i915_vma_is_active(vma)) + continue; + + mutex_unlock(&i915->ggtt.vm.mutex); + if (i915_vma_unbind(vma) == 0) freed_pages += count; + mutex_lock(&i915->ggtt.vm.mutex); } + mutex_unlock(&i915->ggtt.vm.mutex); out: shrinker_unlock(i915, unlock); diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c index 9df615eea2d8..74a9661479ca 100644 --- a/drivers/gpu/drm/i915/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/i915_gem_stolen.c @@ -701,7 +701,10 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_i915_private *dev_priv vma->pages = obj->mm.pages; vma->flags |= I915_VMA_GLOBAL_BIND; __i915_vma_set_map_and_fenceable(vma); - list_move_tail(&vma->vm_link, &ggtt->vm.inactive_list); + + mutex_lock(&ggtt->vm.mutex); + list_move_tail(&vma->vm_link, &ggtt->vm.bound_list); + mutex_unlock(&ggtt->vm.mutex); spin_lock(&dev_priv->mm.obj_lock); list_move_tail(&obj->mm.link, &dev_priv->mm.bound_list); diff --git a/drivers/gpu/drm/i915/i915_gem_userptr.c b/drivers/gpu/drm/i915/i915_gem_userptr.c index 1fb6a7bb5054..1d3f9a31ad61 100644 --- a/drivers/gpu/drm/i915/i915_gem_userptr.c +++ b/drivers/gpu/drm/i915/i915_gem_userptr.c @@ -49,77 +49,67 @@ struct i915_mmu_notifier { struct hlist_node node; struct mmu_notifier mn; struct rb_root_cached objects; - struct workqueue_struct *wq; + struct i915_mm_struct *mm; }; struct i915_mmu_object { struct i915_mmu_notifier *mn; struct drm_i915_gem_object *obj; struct interval_tree_node it; - struct list_head link; - struct work_struct work; - bool attached; }; -static void cancel_userptr(struct work_struct *work) +static void add_object(struct i915_mmu_object *mo) { - struct i915_mmu_object *mo = container_of(work, typeof(*mo), work); - struct drm_i915_gem_object *obj = mo->obj; - struct work_struct *active; - - /* Cancel any active worker and force us to re-evaluate gup */ - mutex_lock(&obj->mm.lock); - active = fetch_and_zero(&obj->userptr.work); - mutex_unlock(&obj->mm.lock); - if (active) - goto out; - - i915_gem_object_wait(obj, I915_WAIT_ALL, MAX_SCHEDULE_TIMEOUT, NULL); - - mutex_lock(&obj->base.dev->struct_mutex); - - /* We are inside a kthread context and can't be interrupted */ - if (i915_gem_object_unbind(obj) == 0) - __i915_gem_object_put_pages(obj, I915_MM_NORMAL); - WARN_ONCE(i915_gem_object_has_pages(obj), - "Failed to release pages: bind_count=%d, pages_pin_count=%d, pin_global=%d\n", - obj->bind_count, - atomic_read(&obj->mm.pages_pin_count), - obj->pin_global); - - mutex_unlock(&obj->base.dev->struct_mutex); - -out: - i915_gem_object_put(obj); + GEM_BUG_ON(!RB_EMPTY_NODE(&mo->it.rb)); + interval_tree_insert(&mo->it, &mo->mn->objects); } -static void add_object(struct i915_mmu_object *mo) +static void del_object(struct i915_mmu_object *mo) { - if (mo->attached) + if (RB_EMPTY_NODE(&mo->it.rb)) return; - interval_tree_insert(&mo->it, &mo->mn->objects); - mo->attached = true; + interval_tree_remove(&mo->it, &mo->mn->objects); + RB_CLEAR_NODE(&mo->it.rb); } -static void del_object(struct i915_mmu_object *mo) +static void +__i915_gem_userptr_set_active(struct drm_i915_gem_object *obj, bool value) { - if (!mo->attached) + struct i915_mmu_object *mo = obj->userptr.mmu_object; + + /* + * During mm_invalidate_range we need to cancel any userptr that + * overlaps the range being invalidated. Doing so requires the + * struct_mutex, and that risks recursion. In order to cause + * recursion, the user must alias the userptr address space with + * a GTT mmapping (possible with a MAP_FIXED) - then when we have + * to invalidate that mmaping, mm_invalidate_range is called with + * the userptr address *and* the struct_mutex held. To prevent that + * we set a flag under the i915_mmu_notifier spinlock to indicate + * whether this object is valid. + */ + if (!mo) return; - interval_tree_remove(&mo->it, &mo->mn->objects); - mo->attached = false; + spin_lock(&mo->mn->lock); + if (value) + add_object(mo); + else + del_object(mo); + spin_unlock(&mo->mn->lock); } -static int i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn, - const struct mmu_notifier_range *range) +static int +userptr_mn_invalidate_range_start(struct mmu_notifier *_mn, + const struct mmu_notifier_range *range) { struct i915_mmu_notifier *mn = container_of(_mn, struct i915_mmu_notifier, mn); - struct i915_mmu_object *mo; struct interval_tree_node *it; - LIST_HEAD(cancelled); + struct mutex *unlock = NULL; unsigned long end; + int ret = 0; if (RB_EMPTY_ROOT(&mn->objects.rb_root)) return 0; @@ -130,11 +120,15 @@ static int i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn, spin_lock(&mn->lock); it = interval_tree_iter_first(&mn->objects, range->start, end); while (it) { + struct drm_i915_gem_object *obj; + if (!range->blockable) { - spin_unlock(&mn->lock); - return -EAGAIN; + ret = -EAGAIN; + break; } - /* The mmu_object is released late when destroying the + + /* + * The mmu_object is released late when destroying the * GEM object so it is entirely possible to gain a * reference on an object in the process of being freed * since our serialisation is via the spinlock and not @@ -143,29 +137,65 @@ static int i915_gem_userptr_mn_invalidate_range_start(struct mmu_notifier *_mn, * use-after-free we only acquire a reference on the * object if it is not in the process of being destroyed. */ - mo = container_of(it, struct i915_mmu_object, it); - if (kref_get_unless_zero(&mo->obj->base.refcount)) - queue_work(mn->wq, &mo->work); + obj = container_of(it, struct i915_mmu_object, it)->obj; + if (!kref_get_unless_zero(&obj->base.refcount)) { + it = interval_tree_iter_next(it, range->start, end); + continue; + } + spin_unlock(&mn->lock); + + if (!unlock) { + unlock = &mn->mm->i915->drm.struct_mutex; + + switch (mutex_trylock_recursive(unlock)) { + default: + case MUTEX_TRYLOCK_FAILED: + if (mutex_lock_killable_nested(unlock, I915_MM_SHRINKER)) { + i915_gem_object_put(obj); + return -EINTR; + } + /* fall through */ + case MUTEX_TRYLOCK_SUCCESS: + break; + + case MUTEX_TRYLOCK_RECURSIVE: + unlock = ERR_PTR(-EEXIST); + break; + } + } + + ret = i915_gem_object_unbind(obj); + if (ret == 0) + ret = __i915_gem_object_put_pages(obj, I915_MM_SHRINKER); + i915_gem_object_put(obj); + if (ret) + goto unlock; - list_add(&mo->link, &cancelled); - it = interval_tree_iter_next(it, range->start, end); + spin_lock(&mn->lock); + + /* + * As we do not (yet) protect the mmu from concurrent insertion + * over this range, there is no guarantee that this search will + * terminate given a pathologic workload. + */ + it = interval_tree_iter_first(&mn->objects, range->start, end); } - list_for_each_entry(mo, &cancelled, link) - del_object(mo); spin_unlock(&mn->lock); - if (!list_empty(&cancelled)) - flush_workqueue(mn->wq); +unlock: + if (!IS_ERR_OR_NULL(unlock)) + mutex_unlock(unlock); + + return ret; - return 0; } static const struct mmu_notifier_ops i915_gem_userptr_notifier = { - .invalidate_range_start = i915_gem_userptr_mn_invalidate_range_start, + .invalidate_range_start = userptr_mn_invalidate_range_start, }; static struct i915_mmu_notifier * -i915_mmu_notifier_create(struct mm_struct *mm) +i915_mmu_notifier_create(struct i915_mm_struct *mm) { struct i915_mmu_notifier *mn; @@ -176,13 +206,7 @@ i915_mmu_notifier_create(struct mm_struct *mm) spin_lock_init(&mn->lock); mn->mn.ops = &i915_gem_userptr_notifier; mn->objects = RB_ROOT_CACHED; - mn->wq = alloc_workqueue("i915-userptr-release", - WQ_UNBOUND | WQ_MEM_RECLAIM, - 0); - if (mn->wq == NULL) { - kfree(mn); - return ERR_PTR(-ENOMEM); - } + mn->mm = mm; return mn; } @@ -192,16 +216,14 @@ i915_gem_userptr_release__mmu_notifier(struct drm_i915_gem_object *obj) { struct i915_mmu_object *mo; - mo = obj->userptr.mmu_object; - if (mo == NULL) + mo = fetch_and_zero(&obj->userptr.mmu_object); + if (!mo) return; spin_lock(&mo->mn->lock); del_object(mo); spin_unlock(&mo->mn->lock); kfree(mo); - - obj->userptr.mmu_object = NULL; } static struct i915_mmu_notifier * @@ -214,7 +236,7 @@ i915_mmu_notifier_find(struct i915_mm_struct *mm) if (mn) return mn; - mn = i915_mmu_notifier_create(mm->mm); + mn = i915_mmu_notifier_create(mm); if (IS_ERR(mn)) err = PTR_ERR(mn); @@ -237,10 +259,8 @@ i915_mmu_notifier_find(struct i915_mm_struct *mm) mutex_unlock(&mm->i915->mm_lock); up_write(&mm->mm->mmap_sem); - if (mn && !IS_ERR(mn)) { - destroy_workqueue(mn->wq); + if (mn && !IS_ERR(mn)) kfree(mn); - } return err ? ERR_PTR(err) : mm->mn; } @@ -263,14 +283,14 @@ i915_gem_userptr_init__mmu_notifier(struct drm_i915_gem_object *obj, return PTR_ERR(mn); mo = kzalloc(sizeof(*mo), GFP_KERNEL); - if (mo == NULL) + if (!mo) return -ENOMEM; mo->mn = mn; mo->obj = obj; mo->it.start = obj->userptr.ptr; mo->it.last = obj->userptr.ptr + obj->base.size - 1; - INIT_WORK(&mo->work, cancel_userptr); + RB_CLEAR_NODE(&mo->it.rb); obj->userptr.mmu_object = mo; return 0; @@ -284,13 +304,17 @@ i915_mmu_notifier_free(struct i915_mmu_notifier *mn, return; mmu_notifier_unregister(&mn->mn, mm); - destroy_workqueue(mn->wq); kfree(mn); } #else static void +__i915_gem_userptr_set_active(struct drm_i915_gem_object *obj, bool value) +{ +} + +static void i915_gem_userptr_release__mmu_notifier(struct drm_i915_gem_object *obj) { } @@ -458,42 +482,6 @@ alloc_table: return st; } -static int -__i915_gem_userptr_set_active(struct drm_i915_gem_object *obj, - bool value) -{ - int ret = 0; - - /* During mm_invalidate_range we need to cancel any userptr that - * overlaps the range being invalidated. Doing so requires the - * struct_mutex, and that risks recursion. In order to cause - * recursion, the user must alias the userptr address space with - * a GTT mmapping (possible with a MAP_FIXED) - then when we have - * to invalidate that mmaping, mm_invalidate_range is called with - * the userptr address *and* the struct_mutex held. To prevent that - * we set a flag under the i915_mmu_notifier spinlock to indicate - * whether this object is valid. - */ -#if defined(CONFIG_MMU_NOTIFIER) - if (obj->userptr.mmu_object == NULL) - return 0; - - spin_lock(&obj->userptr.mmu_object->mn->lock); - /* In order to serialise get_pages with an outstanding - * cancel_userptr, we must drop the struct_mutex and try again. - */ - if (!value) - del_object(obj->userptr.mmu_object); - else if (!work_pending(&obj->userptr.mmu_object->work)) - add_object(obj->userptr.mmu_object); - else - ret = -EAGAIN; - spin_unlock(&obj->userptr.mmu_object->mn->lock); -#endif - - return ret; -} - static void __i915_gem_userptr_get_pages_worker(struct work_struct *_work) { @@ -679,8 +667,11 @@ i915_gem_userptr_put_pages(struct drm_i915_gem_object *obj, struct sgt_iter sgt_iter; struct page *page; - BUG_ON(obj->userptr.work != NULL); + /* Cancel any inflight work and force them to restart their gup */ + obj->userptr.work = NULL; __i915_gem_userptr_set_active(obj, false); + if (!pages) + return; if (obj->mm.madv != I915_MADV_WILLNEED) obj->mm.dirty = false; @@ -718,7 +709,8 @@ i915_gem_userptr_dmabuf_export(struct drm_i915_gem_object *obj) static const struct drm_i915_gem_object_ops i915_gem_userptr_ops = { .flags = I915_GEM_OBJECT_HAS_STRUCT_PAGE | - I915_GEM_OBJECT_IS_SHRINKABLE, + I915_GEM_OBJECT_IS_SHRINKABLE | + I915_GEM_OBJECT_ASYNC_CANCEL, .get_pages = i915_gem_userptr_get_pages, .put_pages = i915_gem_userptr_put_pages, .dmabuf_export = i915_gem_userptr_dmabuf_export, diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index 5eaf586c4d48..9a65341fec09 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -447,9 +447,14 @@ static void error_print_request(struct drm_i915_error_state_buf *m, if (!erq->seqno) return; - err_printf(m, "%s pid %d, ban score %d, seqno %8x:%08x, prio %d, emitted %dms, start %08x, head %08x, tail %08x\n", + err_printf(m, "%s pid %d, ban score %d, seqno %8x:%08x%s%s, prio %d, emitted %dms, start %08x, head %08x, tail %08x\n", prefix, erq->pid, erq->ban_score, - erq->context, erq->seqno, erq->sched_attr.priority, + erq->context, erq->seqno, + test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, + &erq->flags) ? "!" : "", + test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, + &erq->flags) ? "+" : "", + erq->sched_attr.priority, jiffies_to_msecs(erq->jiffies - epoch), erq->start, erq->head, erq->tail); } @@ -530,13 +535,9 @@ static void error_print_engine(struct drm_i915_error_state_buf *m, } err_printf(m, " seqno: 0x%08x\n", ee->seqno); err_printf(m, " last_seqno: 0x%08x\n", ee->last_seqno); - err_printf(m, " waiting: %s\n", yesno(ee->waiting)); err_printf(m, " ring->head: 0x%08x\n", ee->cpu_ring_head); err_printf(m, " ring->tail: 0x%08x\n", ee->cpu_ring_tail); - err_printf(m, " hangcheck stall: %s\n", yesno(ee->hangcheck_stalled)); - err_printf(m, " hangcheck action: %s\n", - hangcheck_action_to_str(ee->hangcheck_action)); - err_printf(m, " hangcheck action timestamp: %dms (%lu%s)\n", + err_printf(m, " hangcheck timestamp: %dms (%lu%s)\n", jiffies_to_msecs(ee->hangcheck_timestamp - epoch), ee->hangcheck_timestamp, ee->hangcheck_timestamp == epoch ? "; epoch" : ""); @@ -684,15 +685,15 @@ static void __err_print_to_sgl(struct drm_i915_error_state_buf *m, jiffies_to_msecs(error->capture - error->epoch)); for (i = 0; i < ARRAY_SIZE(error->engine); i++) { - if (error->engine[i].hangcheck_stalled && - error->engine[i].context.pid) { - err_printf(m, "Active process (on ring %s): %s [%d], score %d%s\n", - engine_name(m->i915, i), - error->engine[i].context.comm, - error->engine[i].context.pid, - error->engine[i].context.ban_score, - bannable(&error->engine[i].context)); - } + if (!error->engine[i].context.pid) + continue; + + err_printf(m, "Active process (on ring %s): %s [%d], score %d%s\n", + engine_name(m->i915, i), + error->engine[i].context.comm, + error->engine[i].context.pid, + error->engine[i].context.ban_score, + bannable(&error->engine[i].context)); } err_printf(m, "Reset count: %u\n", error->reset_count); err_printf(m, "Suspend count: %u\n", error->suspend_count); @@ -722,8 +723,6 @@ static void __err_print_to_sgl(struct drm_i915_error_state_buf *m, err_printf(m, "FORCEWAKE: 0x%08x\n", error->forcewake); err_printf(m, "DERRMR: 0x%08x\n", error->derrmr); err_printf(m, "CCID: 0x%08x\n", error->ccid); - err_printf(m, "Missed interrupts: 0x%08lx\n", - m->i915->gpu_error.missed_irq_rings); for (i = 0; i < error->nfence; i++) err_printf(m, " fence[%d] = %08llx\n", i, error->fence[i]); @@ -807,21 +806,6 @@ static void __err_print_to_sgl(struct drm_i915_error_state_buf *m, error->epoch); } - if (IS_ERR(ee->waiters)) { - err_printf(m, "%s --- ? waiters [unable to acquire spinlock]\n", - m->i915->engine[i]->name); - } else if (ee->num_waiters) { - err_printf(m, "%s --- %d waiters\n", - m->i915->engine[i]->name, - ee->num_waiters); - for (j = 0; j < ee->num_waiters; j++) { - err_printf(m, " seqno 0x%08x for %s [%d]\n", - ee->waiters[j].seqno, - ee->waiters[j].comm, - ee->waiters[j].pid); - } - } - print_error_obj(m, m->i915->engine[i], "ringbuffer", ee->ringbuffer); @@ -1003,8 +987,6 @@ void __i915_gpu_state_free(struct kref *error_ref) i915_error_object_free(ee->wa_ctx); kfree(ee->requests); - if (!IS_ERR_OR_NULL(ee->waiters)) - kfree(ee->waiters); } for (i = 0; i < ARRAY_SIZE(error->active_bo); i++) @@ -1080,23 +1062,23 @@ i915_error_object_create(struct drm_i915_private *i915, } /* The error capture is special as tries to run underneath the normal - * locking rules - so we use the raw version of the i915_gem_active lookup. + * locking rules - so we use the raw version of the i915_active_request lookup. */ -static inline uint32_t -__active_get_seqno(struct i915_gem_active *active) +static inline u32 +__active_get_seqno(struct i915_active_request *active) { struct i915_request *request; - request = __i915_gem_active_peek(active); + request = __i915_active_request_peek(active); return request ? request->global_seqno : 0; } static inline int -__active_get_engine_id(struct i915_gem_active *active) +__active_get_engine_id(struct i915_active_request *active) { struct i915_request *request; - request = __i915_gem_active_peek(active); + request = __i915_active_request_peek(active); return request ? request->engine->id : -1; } @@ -1124,7 +1106,9 @@ static void capture_bo(struct drm_i915_error_buffer *err, static u32 capture_error_bo(struct drm_i915_error_buffer *err, int count, struct list_head *head, - bool pinned_only) + unsigned int flags) +#define ACTIVE_ONLY BIT(0) +#define PINNED_ONLY BIT(1) { struct i915_vma *vma; int i = 0; @@ -1133,7 +1117,10 @@ static u32 capture_error_bo(struct drm_i915_error_buffer *err, if (!vma->obj) continue; - if (pinned_only && !i915_vma_is_pinned(vma)) + if (flags & ACTIVE_ONLY && !i915_vma_is_active(vma)) + continue; + + if (flags & PINNED_ONLY && !i915_vma_is_pinned(vma)) continue; capture_bo(err++, vma); @@ -1144,7 +1131,8 @@ static u32 capture_error_bo(struct drm_i915_error_buffer *err, return i; } -/* Generate a semi-unique error code. The code is not meant to have meaning, The +/* + * Generate a semi-unique error code. The code is not meant to have meaning, The * code's only purpose is to try to prevent false duplicated bug reports by * grossly estimating a GPU error state. * @@ -1153,29 +1141,23 @@ static u32 capture_error_bo(struct drm_i915_error_buffer *err, * * It's only a small step better than a random number in its current form. */ -static uint32_t i915_error_generate_code(struct drm_i915_private *dev_priv, - struct i915_gpu_state *error, - int *engine_id) +static u32 i915_error_generate_code(struct i915_gpu_state *error, + unsigned long engine_mask) { - uint32_t error_code = 0; - int i; - - /* IPEHR would be an ideal way to detect errors, as it's the gross + /* + * IPEHR would be an ideal way to detect errors, as it's the gross * measure of "the command that hung." However, has some very common * synchronization commands which almost always appear in the case * strictly a client bug. Use instdone to differentiate those some. */ - for (i = 0; i < I915_NUM_ENGINES; i++) { - if (error->engine[i].hangcheck_stalled) { - if (engine_id) - *engine_id = i; + if (engine_mask) { + struct drm_i915_error_engine *ee = + &error->engine[ffs(engine_mask)]; - return error->engine[i].ipehr ^ - error->engine[i].instdone.instdone; - } + return ee->ipehr ^ ee->instdone.instdone; } - return error_code; + return 0; } static void gem_record_fences(struct i915_gpu_state *error) @@ -1208,59 +1190,6 @@ static void gen6_record_semaphore_state(struct intel_engine_cs *engine, I915_READ(RING_SYNC_2(engine->mmio_base)); } -static void error_record_engine_waiters(struct intel_engine_cs *engine, - struct drm_i915_error_engine *ee) -{ - struct intel_breadcrumbs *b = &engine->breadcrumbs; - struct drm_i915_error_waiter *waiter; - struct rb_node *rb; - int count; - - ee->num_waiters = 0; - ee->waiters = NULL; - - if (RB_EMPTY_ROOT(&b->waiters)) - return; - - if (!spin_trylock_irq(&b->rb_lock)) { - ee->waiters = ERR_PTR(-EDEADLK); - return; - } - - count = 0; - for (rb = rb_first(&b->waiters); rb != NULL; rb = rb_next(rb)) - count++; - spin_unlock_irq(&b->rb_lock); - - waiter = NULL; - if (count) - waiter = kmalloc_array(count, - sizeof(struct drm_i915_error_waiter), - GFP_ATOMIC); - if (!waiter) - return; - - if (!spin_trylock_irq(&b->rb_lock)) { - kfree(waiter); - ee->waiters = ERR_PTR(-EDEADLK); - return; - } - - ee->waiters = waiter; - for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) { - struct intel_wait *w = rb_entry(rb, typeof(*w), node); - - strcpy(waiter->comm, w->tsk->comm); - waiter->pid = w->tsk->pid; - waiter->seqno = w->seqno; - waiter++; - - if (++ee->num_waiters == count) - break; - } - spin_unlock_irq(&b->rb_lock); -} - static void error_record_engine_registers(struct i915_gpu_state *error, struct intel_engine_cs *engine, struct drm_i915_error_engine *ee) @@ -1296,7 +1225,6 @@ static void error_record_engine_registers(struct i915_gpu_state *error, intel_engine_get_instdone(engine, &ee->instdone); - ee->waiting = intel_engine_has_waiter(engine); ee->instpm = I915_READ(RING_INSTPM(engine->mmio_base)); ee->acthd = intel_engine_get_active_head(engine); ee->seqno = intel_engine_get_seqno(engine); @@ -1338,9 +1266,8 @@ static void error_record_engine_registers(struct i915_gpu_state *error, } ee->idle = intel_engine_is_idle(engine); - ee->hangcheck_timestamp = engine->hangcheck.action_timestamp; - ee->hangcheck_action = engine->hangcheck.action; - ee->hangcheck_stalled = engine->hangcheck.stalled; + if (!ee->idle) + ee->hangcheck_timestamp = engine->hangcheck.action_timestamp; ee->reset_count = i915_reset_engine_count(&dev_priv->gpu_error, engine); @@ -1371,6 +1298,7 @@ static void record_request(struct i915_request *request, { struct i915_gem_context *ctx = request->gem_context; + erq->flags = request->fence.flags; erq->context = ctx->hw_id; erq->sched_attr = request->sched.attr; erq->ban_score = atomic_read(&ctx->ban_score); @@ -1546,7 +1474,6 @@ static void gem_record_rings(struct i915_gpu_state *error) ee->engine_id = i; error_record_engine_registers(error, engine, ee); - error_record_engine_waiters(engine, ee); error_record_engine_execlists(engine, ee); request = i915_gem_find_active_request(engine); @@ -1610,14 +1537,17 @@ static void gem_capture_vm(struct i915_gpu_state *error, int count; count = 0; - list_for_each_entry(vma, &vm->active_list, vm_link) - count++; + list_for_each_entry(vma, &vm->bound_list, vm_link) + if (i915_vma_is_active(vma)) + count++; active_bo = NULL; if (count) active_bo = kcalloc(count, sizeof(*active_bo), GFP_ATOMIC); if (active_bo) - count = capture_error_bo(active_bo, count, &vm->active_list, false); + count = capture_error_bo(active_bo, + count, &vm->bound_list, + ACTIVE_ONLY); else count = 0; @@ -1655,28 +1585,20 @@ static void capture_pinned_buffers(struct i915_gpu_state *error) struct i915_address_space *vm = &error->i915->ggtt.vm; struct drm_i915_error_buffer *bo; struct i915_vma *vma; - int count_inactive, count_active; - - count_inactive = 0; - list_for_each_entry(vma, &vm->inactive_list, vm_link) - count_inactive++; + int count; - count_active = 0; - list_for_each_entry(vma, &vm->active_list, vm_link) - count_active++; + count = 0; + list_for_each_entry(vma, &vm->bound_list, vm_link) + count++; bo = NULL; - if (count_inactive + count_active) - bo = kcalloc(count_inactive + count_active, - sizeof(*bo), GFP_ATOMIC); + if (count) + bo = kcalloc(count, sizeof(*bo), GFP_ATOMIC); if (!bo) return; - count_inactive = capture_error_bo(bo, count_inactive, - &vm->active_list, true); - count_active = capture_error_bo(bo + count_inactive, count_active, - &vm->inactive_list, true); - error->pinned_bo_count = count_inactive + count_active; + error->pinned_bo_count = + capture_error_bo(bo, count, &vm->bound_list, PINNED_ONLY); error->pinned_bo = bo; } @@ -1783,31 +1705,35 @@ static void capture_reg_state(struct i915_gpu_state *error) error->pgtbl_er = I915_READ(PGTBL_ER); } -static void i915_error_capture_msg(struct drm_i915_private *dev_priv, - struct i915_gpu_state *error, - u32 engine_mask, - const char *error_msg) +static const char * +error_msg(struct i915_gpu_state *error, unsigned long engines, const char *msg) { - u32 ecode; - int engine_id = -1, len; + int len; + int i; - ecode = i915_error_generate_code(dev_priv, error, &engine_id); + for (i = 0; i < ARRAY_SIZE(error->engine); i++) + if (!error->engine[i].context.pid) + engines &= ~BIT(i); len = scnprintf(error->error_msg, sizeof(error->error_msg), - "GPU HANG: ecode %d:%d:0x%08x", - INTEL_GEN(dev_priv), engine_id, ecode); - - if (engine_id != -1 && error->engine[engine_id].context.pid) + "GPU HANG: ecode %d:%lx:0x%08x", + INTEL_GEN(error->i915), engines, + i915_error_generate_code(error, engines)); + if (engines) { + /* Just show the first executing process, more is confusing */ + i = ffs(engines); len += scnprintf(error->error_msg + len, sizeof(error->error_msg) - len, ", in %s [%d]", - error->engine[engine_id].context.comm, - error->engine[engine_id].context.pid); + error->engine[i].context.comm, + error->engine[i].context.pid); + } + if (msg) + len += scnprintf(error->error_msg + len, + sizeof(error->error_msg) - len, + ", %s", msg); - scnprintf(error->error_msg + len, sizeof(error->error_msg) - len, - ", reason: %s, action: %s", - error_msg, - engine_mask ? "reset" : "continue"); + return error->error_msg; } static void capture_gen_state(struct i915_gpu_state *error) @@ -1847,7 +1773,7 @@ static unsigned long capture_find_epoch(const struct i915_gpu_state *error) for (i = 0; i < ARRAY_SIZE(error->engine); i++) { const struct drm_i915_error_engine *ee = &error->engine[i]; - if (ee->hangcheck_stalled && + if (ee->hangcheck_timestamp && time_before(ee->hangcheck_timestamp, epoch)) epoch = ee->hangcheck_timestamp; } @@ -1921,7 +1847,7 @@ i915_capture_gpu_state(struct drm_i915_private *i915) * i915_capture_error_state - capture an error record for later analysis * @i915: i915 device * @engine_mask: the mask of engines triggering the hang - * @error_msg: a message to insert into the error capture header + * @msg: a message to insert into the error capture header * * Should be called when an error is detected (either a hang or an error * interrupt) to capture error state from the time of the error. Fills @@ -1929,8 +1855,8 @@ i915_capture_gpu_state(struct drm_i915_private *i915) * to pick up. */ void i915_capture_error_state(struct drm_i915_private *i915, - u32 engine_mask, - const char *error_msg) + unsigned long engine_mask, + const char *msg) { static bool warned; struct i915_gpu_state *error; @@ -1946,8 +1872,7 @@ void i915_capture_error_state(struct drm_i915_private *i915, if (IS_ERR(error)) return; - i915_error_capture_msg(i915, error, engine_mask, error_msg); - DRM_INFO("%s\n", error->error_msg); + dev_info(i915->drm.dev, "%s\n", error_msg(error, engine_mask, msg)); if (!error->simulated) { spin_lock_irqsave(&i915->gpu_error.lock, flags); diff --git a/drivers/gpu/drm/i915/i915_gpu_error.h b/drivers/gpu/drm/i915/i915_gpu_error.h index 6d9f45468ac1..53b1f22dd365 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.h +++ b/drivers/gpu/drm/i915/i915_gpu_error.h @@ -82,11 +82,7 @@ struct i915_gpu_state { int engine_id; /* Software tracked state */ bool idle; - bool waiting; - int num_waiters; unsigned long hangcheck_timestamp; - bool hangcheck_stalled; - enum intel_engine_hangcheck_action hangcheck_action; struct i915_address_space *vm; int num_requests; u32 reset_count; @@ -149,6 +145,7 @@ struct i915_gpu_state { struct drm_i915_error_object *default_state; struct drm_i915_error_request { + unsigned long flags; long jiffies; pid_t pid; u32 context; @@ -161,12 +158,6 @@ struct i915_gpu_state { } *requests, execlist[EXECLIST_MAX_PORTS]; unsigned int num_ports; - struct drm_i915_error_waiter { - char comm[TASK_COMM_LEN]; - pid_t pid; - u32 seqno; - } *waiters; - struct { u32 gfx_mode; union { @@ -197,6 +188,8 @@ struct i915_gpu_state { struct scatterlist *sgl, *fit; }; +struct i915_gpu_restart; + struct i915_gpu_error { /* For hangcheck timer */ #define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */ @@ -211,8 +204,6 @@ struct i915_gpu_error { atomic_t pending_fb_pin; - unsigned long missed_irq_rings; - /** * State variable controlling the reset flow and count * @@ -247,15 +238,6 @@ struct i915_gpu_error { * i915_mutex_lock_interruptible()?). I915_RESET_BACKOFF serves a * secondary role in preventing two concurrent global reset attempts. * - * #I915_RESET_HANDOFF - To perform the actual GPU reset, we need the - * struct_mutex. We try to acquire the struct_mutex in the reset worker, - * but it may be held by some long running waiter (that we cannot - * interrupt without causing trouble). Once we are ready to do the GPU - * reset, we set the I915_RESET_HANDOFF bit and wakeup any waiters. If - * they already hold the struct_mutex and want to participate they can - * inspect the bit and do the reset directly, otherwise the worker - * waits for the struct_mutex. - * * #I915_RESET_ENGINE[num_engines] - Since the driver doesn't need to * acquire the struct_mutex to reset an engine, we need an explicit * flag to prevent two concurrent reset attempts in the same engine. @@ -269,19 +251,14 @@ struct i915_gpu_error { */ unsigned long flags; #define I915_RESET_BACKOFF 0 -#define I915_RESET_HANDOFF 1 -#define I915_RESET_MODESET 2 +#define I915_RESET_MODESET 1 +#define I915_RESET_ENGINE 2 #define I915_WEDGED (BITS_PER_LONG - 1) -#define I915_RESET_ENGINE (I915_WEDGED - I915_NUM_ENGINES) /** Number of times an engine has been reset */ u32 reset_engine_count[I915_NUM_ENGINES]; - /** Set of stalled engines with guilty requests, in the current reset */ - u32 stalled_mask; - - /** Reason for the current *global* reset */ - const char *reason; + struct mutex wedge_mutex; /* serialises wedging/unwedging */ /** * Waitqueue to signal when a hang is detected. Used to for waiters @@ -295,8 +272,7 @@ struct i915_gpu_error { */ wait_queue_head_t reset_queue; - /* For missed irq/seqno simulation. */ - unsigned long test_irq_rings; + struct i915_gpu_restart *restart; }; struct drm_i915_error_state_buf { @@ -318,7 +294,7 @@ void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...); struct i915_gpu_state *i915_capture_gpu_state(struct drm_i915_private *i915); void i915_capture_error_state(struct drm_i915_private *dev_priv, - u32 engine_mask, + unsigned long engine_mask, const char *error_msg); static inline struct i915_gpu_state * diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index c91d18707475..441d2674b272 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -225,10 +225,10 @@ static void gen9_guc_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir); /* For display hotplug interrupt */ static inline void i915_hotplug_interrupt_update_locked(struct drm_i915_private *dev_priv, - uint32_t mask, - uint32_t bits) + u32 mask, + u32 bits) { - uint32_t val; + u32 val; lockdep_assert_held(&dev_priv->irq_lock); WARN_ON(bits & ~mask); @@ -252,8 +252,8 @@ i915_hotplug_interrupt_update_locked(struct drm_i915_private *dev_priv, * version is also available. */ void i915_hotplug_interrupt_update(struct drm_i915_private *dev_priv, - uint32_t mask, - uint32_t bits) + u32 mask, + u32 bits) { spin_lock_irq(&dev_priv->irq_lock); i915_hotplug_interrupt_update_locked(dev_priv, mask, bits); @@ -302,10 +302,10 @@ static bool gen11_reset_one_iir(struct drm_i915_private * const i915, * @enabled_irq_mask: mask of interrupt bits to enable */ void ilk_update_display_irq(struct drm_i915_private *dev_priv, - uint32_t interrupt_mask, - uint32_t enabled_irq_mask) + u32 interrupt_mask, + u32 enabled_irq_mask) { - uint32_t new_val; + u32 new_val; lockdep_assert_held(&dev_priv->irq_lock); @@ -332,8 +332,8 @@ void ilk_update_display_irq(struct drm_i915_private *dev_priv, * @enabled_irq_mask: mask of interrupt bits to enable */ static void ilk_update_gt_irq(struct drm_i915_private *dev_priv, - uint32_t interrupt_mask, - uint32_t enabled_irq_mask) + u32 interrupt_mask, + u32 enabled_irq_mask) { lockdep_assert_held(&dev_priv->irq_lock); @@ -347,13 +347,13 @@ static void ilk_update_gt_irq(struct drm_i915_private *dev_priv, I915_WRITE(GTIMR, dev_priv->gt_irq_mask); } -void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask) +void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, u32 mask) { ilk_update_gt_irq(dev_priv, mask, mask); POSTING_READ_FW(GTIMR); } -void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask) +void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, u32 mask) { ilk_update_gt_irq(dev_priv, mask, 0); } @@ -392,10 +392,10 @@ static i915_reg_t gen6_pm_ier(struct drm_i915_private *dev_priv) * @enabled_irq_mask: mask of interrupt bits to enable */ static void snb_update_pm_irq(struct drm_i915_private *dev_priv, - uint32_t interrupt_mask, - uint32_t enabled_irq_mask) + u32 interrupt_mask, + u32 enabled_irq_mask) { - uint32_t new_val; + u32 new_val; WARN_ON(enabled_irq_mask & ~interrupt_mask); @@ -579,11 +579,11 @@ void gen9_disable_guc_interrupts(struct drm_i915_private *dev_priv) * @enabled_irq_mask: mask of interrupt bits to enable */ static void bdw_update_port_irq(struct drm_i915_private *dev_priv, - uint32_t interrupt_mask, - uint32_t enabled_irq_mask) + u32 interrupt_mask, + u32 enabled_irq_mask) { - uint32_t new_val; - uint32_t old_val; + u32 new_val; + u32 old_val; lockdep_assert_held(&dev_priv->irq_lock); @@ -613,10 +613,10 @@ static void bdw_update_port_irq(struct drm_i915_private *dev_priv, */ void bdw_update_pipe_irq(struct drm_i915_private *dev_priv, enum pipe pipe, - uint32_t interrupt_mask, - uint32_t enabled_irq_mask) + u32 interrupt_mask, + u32 enabled_irq_mask) { - uint32_t new_val; + u32 new_val; lockdep_assert_held(&dev_priv->irq_lock); @@ -643,10 +643,10 @@ void bdw_update_pipe_irq(struct drm_i915_private *dev_priv, * @enabled_irq_mask: mask of interrupt bits to enable */ void ibx_display_interrupt_update(struct drm_i915_private *dev_priv, - uint32_t interrupt_mask, - uint32_t enabled_irq_mask) + u32 interrupt_mask, + u32 enabled_irq_mask) { - uint32_t sdeimr = I915_READ(SDEIMR); + u32 sdeimr = I915_READ(SDEIMR); sdeimr &= ~interrupt_mask; sdeimr |= (~enabled_irq_mask & interrupt_mask); @@ -823,11 +823,26 @@ static void i915_enable_asle_pipestat(struct drm_i915_private *dev_priv) static u32 i915_get_vblank_counter(struct drm_device *dev, unsigned int pipe) { struct drm_i915_private *dev_priv = to_i915(dev); + struct drm_vblank_crtc *vblank = &dev->vblank[pipe]; + const struct drm_display_mode *mode = &vblank->hwmode; i915_reg_t high_frame, low_frame; u32 high1, high2, low, pixel, vbl_start, hsync_start, htotal; - const struct drm_display_mode *mode = &dev->vblank[pipe].hwmode; unsigned long irqflags; + /* + * On i965gm TV output the frame counter only works up to + * the point when we enable the TV encoder. After that the + * frame counter ceases to work and reads zero. We need a + * vblank wait before enabling the TV encoder and so we + * have to enable vblank interrupts while the frame counter + * is still in a working state. However the core vblank code + * does not like us returning non-zero frame counter values + * when we've told it that we don't have a working frame + * counter. Thus we must stop non-zero values leaking out. + */ + if (!vblank->max_vblank_count) + return 0; + htotal = mode->crtc_htotal; hsync_start = mode->crtc_hsync_start; vbl_start = mode->crtc_vblank_start; @@ -999,6 +1014,9 @@ static bool i915_get_crtc_scanoutpos(struct drm_device *dev, unsigned int pipe, int position; int vbl_start, vbl_end, hsync_start, htotal, vtotal; unsigned long irqflags; + bool use_scanline_counter = INTEL_GEN(dev_priv) >= 5 || + IS_G4X(dev_priv) || IS_GEN(dev_priv, 2) || + mode->private_flags & I915_MODE_FLAG_USE_SCANLINE_COUNTER; if (WARN_ON(!mode->crtc_clock)) { DRM_DEBUG_DRIVER("trying to get scanoutpos for disabled " @@ -1031,7 +1049,7 @@ static bool i915_get_crtc_scanoutpos(struct drm_device *dev, unsigned int pipe, if (stime) *stime = ktime_get(); - if (IS_GEN(dev_priv, 2) || IS_G4X(dev_priv) || INTEL_GEN(dev_priv) >= 5) { + if (use_scanline_counter) { /* No obvious pixelcount register. Only query vertical * scanout position from Display scan line register. */ @@ -1091,7 +1109,7 @@ static bool i915_get_crtc_scanoutpos(struct drm_device *dev, unsigned int pipe, else position += vtotal - vbl_end; - if (IS_GEN(dev_priv, 2) || IS_G4X(dev_priv) || INTEL_GEN(dev_priv) >= 5) { + if (use_scanline_counter) { *vpos = position; *hpos = 0; } else { @@ -1153,69 +1171,6 @@ static void ironlake_rps_change_irq_handler(struct drm_i915_private *dev_priv) return; } -static void notify_ring(struct intel_engine_cs *engine) -{ - const u32 seqno = intel_engine_get_seqno(engine); - struct i915_request *rq = NULL; - struct task_struct *tsk = NULL; - struct intel_wait *wait; - - if (unlikely(!engine->breadcrumbs.irq_armed)) - return; - - rcu_read_lock(); - - spin_lock(&engine->breadcrumbs.irq_lock); - wait = engine->breadcrumbs.irq_wait; - if (wait) { - /* - * We use a callback from the dma-fence to submit - * requests after waiting on our own requests. To - * ensure minimum delay in queuing the next request to - * hardware, signal the fence now rather than wait for - * the signaler to be woken up. We still wake up the - * waiter in order to handle the irq-seqno coherency - * issues (we may receive the interrupt before the - * seqno is written, see __i915_request_irq_complete()) - * and to handle coalescing of multiple seqno updates - * and many waiters. - */ - if (i915_seqno_passed(seqno, wait->seqno)) { - struct i915_request *waiter = wait->request; - - if (waiter && - !test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, - &waiter->fence.flags) && - intel_wait_check_request(wait, waiter)) - rq = i915_request_get(waiter); - - tsk = wait->tsk; - } - - engine->breadcrumbs.irq_count++; - } else { - if (engine->breadcrumbs.irq_armed) - __intel_engine_disarm_breadcrumbs(engine); - } - spin_unlock(&engine->breadcrumbs.irq_lock); - - if (rq) { - spin_lock(&rq->lock); - dma_fence_signal_locked(&rq->fence); - GEM_BUG_ON(!i915_request_completed(rq)); - spin_unlock(&rq->lock); - - i915_request_put(rq); - } - - if (tsk && tsk->state & TASK_NORMAL) - wake_up_process(tsk); - - rcu_read_unlock(); - - trace_intel_engine_notify(engine, wait); -} - static void vlv_c0_read(struct drm_i915_private *dev_priv, struct intel_rps_ei *ei) { @@ -1370,8 +1325,8 @@ static void ivybridge_parity_work(struct work_struct *work) container_of(work, typeof(*dev_priv), l3_parity.error_work); u32 error_status, row, bank, subbank; char *parity_event[6]; - uint32_t misccpctl; - uint8_t slice = 0; + u32 misccpctl; + u8 slice = 0; /* We must turn off DOP level clock gating to access the L3 registers. * In order to prevent a get/put style interface, acquire struct mutex @@ -1460,20 +1415,20 @@ static void ilk_gt_irq_handler(struct drm_i915_private *dev_priv, u32 gt_iir) { if (gt_iir & GT_RENDER_USER_INTERRUPT) - notify_ring(dev_priv->engine[RCS]); + intel_engine_breadcrumbs_irq(dev_priv->engine[RCS]); if (gt_iir & ILK_BSD_USER_INTERRUPT) - notify_ring(dev_priv->engine[VCS]); + intel_engine_breadcrumbs_irq(dev_priv->engine[VCS]); } static void snb_gt_irq_handler(struct drm_i915_private *dev_priv, u32 gt_iir) { if (gt_iir & GT_RENDER_USER_INTERRUPT) - notify_ring(dev_priv->engine[RCS]); + intel_engine_breadcrumbs_irq(dev_priv->engine[RCS]); if (gt_iir & GT_BSD_USER_INTERRUPT) - notify_ring(dev_priv->engine[VCS]); + intel_engine_breadcrumbs_irq(dev_priv->engine[VCS]); if (gt_iir & GT_BLT_USER_INTERRUPT) - notify_ring(dev_priv->engine[BCS]); + intel_engine_breadcrumbs_irq(dev_priv->engine[BCS]); if (gt_iir & (GT_BLT_CS_ERROR_INTERRUPT | GT_BSD_CS_ERROR_INTERRUPT | @@ -1493,7 +1448,7 @@ gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir) tasklet = true; if (iir & GT_RENDER_USER_INTERRUPT) { - notify_ring(engine); + intel_engine_breadcrumbs_irq(engine); tasklet |= USES_GUC_SUBMISSION(engine->i915); } @@ -1732,13 +1687,13 @@ static void dp_aux_irq_handler(struct drm_i915_private *dev_priv) #if defined(CONFIG_DEBUG_FS) static void display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, enum pipe pipe, - uint32_t crc0, uint32_t crc1, - uint32_t crc2, uint32_t crc3, - uint32_t crc4) + u32 crc0, u32 crc1, + u32 crc2, u32 crc3, + u32 crc4) { struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[pipe]; struct intel_crtc *crtc = intel_get_crtc_for_pipe(dev_priv, pipe); - uint32_t crcs[5]; + u32 crcs[5]; spin_lock(&pipe_crc->lock); /* @@ -1770,9 +1725,9 @@ static void display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, static inline void display_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, enum pipe pipe, - uint32_t crc0, uint32_t crc1, - uint32_t crc2, uint32_t crc3, - uint32_t crc4) {} + u32 crc0, u32 crc1, + u32 crc2, u32 crc3, + u32 crc4) {} #endif @@ -1798,7 +1753,7 @@ static void ivb_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, static void i9xx_pipe_crc_irq_handler(struct drm_i915_private *dev_priv, enum pipe pipe) { - uint32_t res1, res2; + u32 res1, res2; if (INTEL_GEN(dev_priv) >= 3) res1 = I915_READ(PIPE_CRC_RES_RES1_I915(pipe)); @@ -1839,7 +1794,7 @@ static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir) if (HAS_VEBOX(dev_priv)) { if (pm_iir & PM_VEBOX_USER_INTERRUPT) - notify_ring(dev_priv->engine[VECS]); + intel_engine_breadcrumbs_irq(dev_priv->engine[VECS]); if (pm_iir & PM_VEBOX_CS_ERROR_INTERRUPT) DRM_DEBUG("Command parser error, pm_iir 0x%08x\n", pm_iir); @@ -2932,46 +2887,6 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg) return IRQ_HANDLED; } -struct wedge_me { - struct delayed_work work; - struct drm_i915_private *i915; - const char *name; -}; - -static void wedge_me(struct work_struct *work) -{ - struct wedge_me *w = container_of(work, typeof(*w), work.work); - - dev_err(w->i915->drm.dev, - "%s timed out, cancelling all in-flight rendering.\n", - w->name); - i915_gem_set_wedged(w->i915); -} - -static void __init_wedge(struct wedge_me *w, - struct drm_i915_private *i915, - long timeout, - const char *name) -{ - w->i915 = i915; - w->name = name; - - INIT_DELAYED_WORK_ONSTACK(&w->work, wedge_me); - schedule_delayed_work(&w->work, timeout); -} - -static void __fini_wedge(struct wedge_me *w) -{ - cancel_delayed_work_sync(&w->work); - destroy_delayed_work_on_stack(&w->work); - w->i915 = NULL; -} - -#define i915_wedge_on_timeout(W, DEV, TIMEOUT) \ - for (__init_wedge((W), (DEV), (TIMEOUT), __func__); \ - (W)->i915; \ - __fini_wedge((W))) - static u32 gen11_gt_engine_identity(struct drm_i915_private * const i915, const unsigned int bank, const unsigned int bit) @@ -3182,203 +3097,6 @@ static irqreturn_t gen11_irq_handler(int irq, void *arg) return IRQ_HANDLED; } -static void i915_reset_device(struct drm_i915_private *dev_priv, - u32 engine_mask, - const char *reason) -{ - struct i915_gpu_error *error = &dev_priv->gpu_error; - struct kobject *kobj = &dev_priv->drm.primary->kdev->kobj; - char *error_event[] = { I915_ERROR_UEVENT "=1", NULL }; - char *reset_event[] = { I915_RESET_UEVENT "=1", NULL }; - char *reset_done_event[] = { I915_ERROR_UEVENT "=0", NULL }; - struct wedge_me w; - - kobject_uevent_env(kobj, KOBJ_CHANGE, error_event); - - DRM_DEBUG_DRIVER("resetting chip\n"); - kobject_uevent_env(kobj, KOBJ_CHANGE, reset_event); - - /* Use a watchdog to ensure that our reset completes */ - i915_wedge_on_timeout(&w, dev_priv, 5*HZ) { - intel_prepare_reset(dev_priv); - - error->reason = reason; - error->stalled_mask = engine_mask; - - /* Signal that locked waiters should reset the GPU */ - smp_mb__before_atomic(); - set_bit(I915_RESET_HANDOFF, &error->flags); - wake_up_all(&error->wait_queue); - - /* Wait for anyone holding the lock to wakeup, without - * blocking indefinitely on struct_mutex. - */ - do { - if (mutex_trylock(&dev_priv->drm.struct_mutex)) { - i915_reset(dev_priv, engine_mask, reason); - mutex_unlock(&dev_priv->drm.struct_mutex); - } - } while (wait_on_bit_timeout(&error->flags, - I915_RESET_HANDOFF, - TASK_UNINTERRUPTIBLE, - 1)); - - error->stalled_mask = 0; - error->reason = NULL; - - intel_finish_reset(dev_priv); - } - - if (!test_bit(I915_WEDGED, &error->flags)) - kobject_uevent_env(kobj, KOBJ_CHANGE, reset_done_event); -} - -void i915_clear_error_registers(struct drm_i915_private *dev_priv) -{ - u32 eir; - - if (!IS_GEN(dev_priv, 2)) - I915_WRITE(PGTBL_ER, I915_READ(PGTBL_ER)); - - if (INTEL_GEN(dev_priv) < 4) - I915_WRITE(IPEIR, I915_READ(IPEIR)); - else - I915_WRITE(IPEIR_I965, I915_READ(IPEIR_I965)); - - I915_WRITE(EIR, I915_READ(EIR)); - eir = I915_READ(EIR); - if (eir) { - /* - * some errors might have become stuck, - * mask them. - */ - DRM_DEBUG_DRIVER("EIR stuck: 0x%08x, masking\n", eir); - I915_WRITE(EMR, I915_READ(EMR) | eir); - I915_WRITE(IIR, I915_MASTER_ERROR_INTERRUPT); - } - - if (INTEL_GEN(dev_priv) >= 8) { - I915_WRITE(GEN8_RING_FAULT_REG, - I915_READ(GEN8_RING_FAULT_REG) & ~RING_FAULT_VALID); - POSTING_READ(GEN8_RING_FAULT_REG); - } else if (INTEL_GEN(dev_priv) >= 6) { - struct intel_engine_cs *engine; - enum intel_engine_id id; - - for_each_engine(engine, dev_priv, id) { - I915_WRITE(RING_FAULT_REG(engine), - I915_READ(RING_FAULT_REG(engine)) & - ~RING_FAULT_VALID); - } - POSTING_READ(RING_FAULT_REG(dev_priv->engine[RCS])); - } -} - -/** - * i915_handle_error - handle a gpu error - * @dev_priv: i915 device private - * @engine_mask: mask representing engines that are hung - * @flags: control flags - * @fmt: Error message format string - * - * Do some basic checking of register state at error time and - * dump it to the syslog. Also call i915_capture_error_state() to make - * sure we get a record and make it available in debugfs. Fire a uevent - * so userspace knows something bad happened (should trigger collection - * of a ring dump etc.). - */ -void i915_handle_error(struct drm_i915_private *dev_priv, - u32 engine_mask, - unsigned long flags, - const char *fmt, ...) -{ - struct intel_engine_cs *engine; - unsigned int tmp; - char error_msg[80]; - char *msg = NULL; - - if (fmt) { - va_list args; - - va_start(args, fmt); - vscnprintf(error_msg, sizeof(error_msg), fmt, args); - va_end(args); - - msg = error_msg; - } - - /* - * In most cases it's guaranteed that we get here with an RPM - * reference held, for example because there is a pending GPU - * request that won't finish until the reset is done. This - * isn't the case at least when we get here by doing a - * simulated reset via debugfs, so get an RPM reference. - */ - intel_runtime_pm_get(dev_priv); - - engine_mask &= INTEL_INFO(dev_priv)->ring_mask; - - if (flags & I915_ERROR_CAPTURE) { - i915_capture_error_state(dev_priv, engine_mask, msg); - i915_clear_error_registers(dev_priv); - } - - /* - * Try engine reset when available. We fall back to full reset if - * single reset fails. - */ - if (intel_has_reset_engine(dev_priv) && - !i915_terminally_wedged(&dev_priv->gpu_error)) { - for_each_engine_masked(engine, dev_priv, engine_mask, tmp) { - BUILD_BUG_ON(I915_RESET_MODESET >= I915_RESET_ENGINE); - if (test_and_set_bit(I915_RESET_ENGINE + engine->id, - &dev_priv->gpu_error.flags)) - continue; - - if (i915_reset_engine(engine, msg) == 0) - engine_mask &= ~intel_engine_flag(engine); - - clear_bit(I915_RESET_ENGINE + engine->id, - &dev_priv->gpu_error.flags); - wake_up_bit(&dev_priv->gpu_error.flags, - I915_RESET_ENGINE + engine->id); - } - } - - if (!engine_mask) - goto out; - - /* Full reset needs the mutex, stop any other user trying to do so. */ - if (test_and_set_bit(I915_RESET_BACKOFF, &dev_priv->gpu_error.flags)) { - wait_event(dev_priv->gpu_error.reset_queue, - !test_bit(I915_RESET_BACKOFF, - &dev_priv->gpu_error.flags)); - goto out; - } - - /* Prevent any other reset-engine attempt. */ - for_each_engine(engine, dev_priv, tmp) { - while (test_and_set_bit(I915_RESET_ENGINE + engine->id, - &dev_priv->gpu_error.flags)) - wait_on_bit(&dev_priv->gpu_error.flags, - I915_RESET_ENGINE + engine->id, - TASK_UNINTERRUPTIBLE); - } - - i915_reset_device(dev_priv, engine_mask, msg); - - for_each_engine(engine, dev_priv, tmp) { - clear_bit(I915_RESET_ENGINE + engine->id, - &dev_priv->gpu_error.flags); - } - - clear_bit(I915_RESET_BACKOFF, &dev_priv->gpu_error.flags); - wake_up_all(&dev_priv->gpu_error.reset_queue); - -out: - intel_runtime_pm_put(dev_priv); -} - /* Called from drm generic code, passed 'crtc' which * we use as a pipe index */ @@ -3411,7 +3129,7 @@ static int ironlake_enable_vblank(struct drm_device *dev, unsigned int pipe) { struct drm_i915_private *dev_priv = to_i915(dev); unsigned long irqflags; - uint32_t bit = INTEL_GEN(dev_priv) >= 7 ? + u32 bit = INTEL_GEN(dev_priv) >= 7 ? DE_PIPE_VBLANK_IVB(pipe) : DE_PIPE_VBLANK(pipe); spin_lock_irqsave(&dev_priv->irq_lock, irqflags); @@ -3473,7 +3191,7 @@ static void ironlake_disable_vblank(struct drm_device *dev, unsigned int pipe) { struct drm_i915_private *dev_priv = to_i915(dev); unsigned long irqflags; - uint32_t bit = INTEL_GEN(dev_priv) >= 7 ? + u32 bit = INTEL_GEN(dev_priv) >= 7 ? DE_PIPE_VBLANK_IVB(pipe) : DE_PIPE_VBLANK(pipe); spin_lock_irqsave(&dev_priv->irq_lock, irqflags); @@ -3691,7 +3409,7 @@ static void gen11_irq_reset(struct drm_device *dev) void gen8_irq_power_well_post_enable(struct drm_i915_private *dev_priv, u8 pipe_mask) { - uint32_t extra_ier = GEN8_PIPE_VBLANK | GEN8_PIPE_FIFO_UNDERRUN; + u32 extra_ier = GEN8_PIPE_VBLANK | GEN8_PIPE_FIFO_UNDERRUN; enum pipe pipe; spin_lock_irq(&dev_priv->irq_lock); @@ -4160,7 +3878,7 @@ static int valleyview_irq_postinstall(struct drm_device *dev) static void gen8_gt_irq_postinstall(struct drm_i915_private *dev_priv) { /* These are interrupts we'll toggle with the ring mask register */ - uint32_t gt_interrupts[] = { + u32 gt_interrupts[] = { GT_RENDER_USER_INTERRUPT << GEN8_RCS_IRQ_SHIFT | GT_CONTEXT_SWITCH_INTERRUPT << GEN8_RCS_IRQ_SHIFT | GT_RENDER_USER_INTERRUPT << GEN8_BCS_IRQ_SHIFT | @@ -4188,8 +3906,8 @@ static void gen8_gt_irq_postinstall(struct drm_i915_private *dev_priv) static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv) { - uint32_t de_pipe_masked = GEN8_PIPE_CDCLK_CRC_DONE; - uint32_t de_pipe_enables; + u32 de_pipe_masked = GEN8_PIPE_CDCLK_CRC_DONE; + u32 de_pipe_enables; u32 de_port_masked = GEN8_AUX_CHANNEL_A; u32 de_port_enables; u32 de_misc_masked = GEN8_DE_EDP_PSR; @@ -4329,6 +4047,7 @@ static int gen11_irq_postinstall(struct drm_device *dev) I915_WRITE(GEN11_DISPLAY_INT_CTL, GEN11_DISPLAY_IRQ_ENABLE); gen11_master_intr_enable(dev_priv->regs); + POSTING_READ(GEN11_GFX_MSTR_IRQ); return 0; } @@ -4499,7 +4218,7 @@ static irqreturn_t i8xx_irq_handler(int irq, void *arg) I915_WRITE16(IIR, iir); if (iir & I915_USER_INTERRUPT) - notify_ring(dev_priv->engine[RCS]); + intel_engine_breadcrumbs_irq(dev_priv->engine[RCS]); if (iir & I915_MASTER_ERROR_INTERRUPT) i8xx_error_irq_handler(dev_priv, eir, eir_stuck); @@ -4607,7 +4326,7 @@ static irqreturn_t i915_irq_handler(int irq, void *arg) I915_WRITE(IIR, iir); if (iir & I915_USER_INTERRUPT) - notify_ring(dev_priv->engine[RCS]); + intel_engine_breadcrumbs_irq(dev_priv->engine[RCS]); if (iir & I915_MASTER_ERROR_INTERRUPT) i9xx_error_irq_handler(dev_priv, eir, eir_stuck); @@ -4752,10 +4471,10 @@ static irqreturn_t i965_irq_handler(int irq, void *arg) I915_WRITE(IIR, iir); if (iir & I915_USER_INTERRUPT) - notify_ring(dev_priv->engine[RCS]); + intel_engine_breadcrumbs_irq(dev_priv->engine[RCS]); if (iir & I915_BSD_USER_INTERRUPT) - notify_ring(dev_priv->engine[VCS]); + intel_engine_breadcrumbs_irq(dev_priv->engine[VCS]); if (iir & I915_MASTER_ERROR_INTERRUPT) i9xx_error_irq_handler(dev_priv, eir, eir_stuck); @@ -4818,16 +4537,10 @@ void intel_irq_init(struct drm_i915_private *dev_priv) if (INTEL_GEN(dev_priv) >= 8) rps->pm_intrmsk_mbz |= GEN8_PMINTR_DISABLE_REDIRECT_TO_GUC; - if (IS_GEN(dev_priv, 2)) { - /* Gen2 doesn't have a hardware frame counter */ - dev->max_vblank_count = 0; - } else if (IS_G4X(dev_priv) || INTEL_GEN(dev_priv) >= 5) { - dev->max_vblank_count = 0xffffffff; /* full 32 bit counter */ + if (INTEL_GEN(dev_priv) >= 5 || IS_G4X(dev_priv)) dev->driver->get_vblank_counter = g4x_get_vblank_counter; - } else { + else if (INTEL_GEN(dev_priv) >= 3) dev->driver->get_vblank_counter = i915_get_vblank_counter; - dev->max_vblank_count = 0xffffff; /* only 24 bits of frame count */ - } /* * Opt out of the vblank disable timer on everything except gen2. diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c index 9f0539bdaa39..b5be0abbba35 100644 --- a/drivers/gpu/drm/i915/i915_params.c +++ b/drivers/gpu/drm/i915/i915_params.c @@ -97,8 +97,10 @@ i915_param_named_unsafe(disable_power_well, int, 0400, i915_param_named_unsafe(enable_ips, int, 0600, "Enable IPS (default: true)"); -i915_param_named(fastboot, bool, 0600, - "Try to skip unnecessary mode sets at boot time (default: false)"); +i915_param_named(fastboot, int, 0600, + "Try to skip unnecessary mode sets at boot time " + "(0=disabled, 1=enabled) " + "Default: -1 (use per-chip default)"); i915_param_named_unsafe(prefault_disable, bool, 0600, "Disable page prefaulting for pread/pwrite/reloc (default:false). " diff --git a/drivers/gpu/drm/i915/i915_params.h b/drivers/gpu/drm/i915/i915_params.h index 6efcf330bdab..3f14e9881a0d 100644 --- a/drivers/gpu/drm/i915/i915_params.h +++ b/drivers/gpu/drm/i915/i915_params.h @@ -63,10 +63,10 @@ struct drm_printer; param(int, edp_vswing, 0) \ param(int, reset, 2) \ param(unsigned int, inject_load_failure, 0) \ + param(int, fastboot, -1) \ /* leave bools at the end to not create holes */ \ param(bool, alpha_support, IS_ENABLED(CONFIG_DRM_I915_ALPHA_SUPPORT)) \ param(bool, enable_hangcheck, true) \ - param(bool, fastboot, false) \ param(bool, prefault_disable, false) \ param(bool, load_detect_test, false) \ param(bool, force_reset_modeset_test, false) \ diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c index 44c23ac60347..66f82f3f050f 100644 --- a/drivers/gpu/drm/i915/i915_pci.c +++ b/drivers/gpu/drm/i915/i915_pci.c @@ -28,6 +28,7 @@ #include <drm/drm_drv.h> +#include "i915_active.h" #include "i915_drv.h" #include "i915_selftest.h" @@ -69,9 +70,15 @@ #define BDW_COLORS \ .color = { .degamma_lut_size = 512, .gamma_lut_size = 512 } #define CHV_COLORS \ - .color = { .degamma_lut_size = 65, .gamma_lut_size = 257 } + .color = { .degamma_lut_size = 65, .gamma_lut_size = 257, \ + .degamma_lut_tests = DRM_COLOR_LUT_NON_DECREASING, \ + .gamma_lut_tests = DRM_COLOR_LUT_NON_DECREASING, \ + } #define GLK_COLORS \ - .color = { .degamma_lut_size = 0, .gamma_lut_size = 1024 } + .color = { .degamma_lut_size = 0, .gamma_lut_size = 1024, \ + .degamma_lut_tests = DRM_COLOR_LUT_NON_DECREASING | \ + DRM_COLOR_LUT_EQUAL_CHANNELS, \ + } /* Keep in gen based order, and chronological order within a gen */ @@ -83,7 +90,7 @@ .num_pipes = 1, \ .display.has_overlay = 1, \ .display.overlay_needs_physical = 1, \ - .display.has_gmch_display = 1, \ + .display.has_gmch = 1, \ .gpu_reset_clobbers_display = true, \ .hws_needs_physical = 1, \ .unfenced_needs_alignment = 1, \ @@ -124,7 +131,7 @@ static const struct intel_device_info intel_i865g_info = { #define GEN3_FEATURES \ GEN(3), \ .num_pipes = 2, \ - .display.has_gmch_display = 1, \ + .display.has_gmch = 1, \ .gpu_reset_clobbers_display = true, \ .ring_mask = RENDER_RING, \ .has_snoop = true, \ @@ -201,7 +208,7 @@ static const struct intel_device_info intel_pineview_info = { GEN(4), \ .num_pipes = 2, \ .display.has_hotplug = 1, \ - .display.has_gmch_display = 1, \ + .display.has_gmch = 1, \ .gpu_reset_clobbers_display = true, \ .ring_mask = RENDER_RING, \ .has_snoop = true, \ @@ -377,7 +384,7 @@ static const struct intel_device_info intel_valleyview_info = { .num_pipes = 2, .has_runtime_pm = 1, .has_rc6 = 1, - .display.has_gmch_display = 1, + .display.has_gmch = 1, .display.has_hotplug = 1, .ppgtt = INTEL_PPGTT_FULL, .has_snoop = true, @@ -469,7 +476,7 @@ static const struct intel_device_info intel_cherryview_info = { .has_runtime_pm = 1, .has_rc6 = 1, .has_logical_ring_contexts = 1, - .display.has_gmch_display = 1, + .display.has_gmch = 1, .ppgtt = INTEL_PPGTT_FULL, .has_reset_engine = 1, .has_snoop = true, @@ -707,6 +714,7 @@ static const struct pci_device_id pciidlist[] = { INTEL_AML_KBL_GT2_IDS(&intel_kabylake_gt2_info), INTEL_CFL_S_GT1_IDS(&intel_coffeelake_gt1_info), INTEL_CFL_S_GT2_IDS(&intel_coffeelake_gt2_info), + INTEL_CFL_H_GT1_IDS(&intel_coffeelake_gt1_info), INTEL_CFL_H_GT2_IDS(&intel_coffeelake_gt2_info), INTEL_CFL_U_GT2_IDS(&intel_coffeelake_gt2_info), INTEL_CFL_U_GT3_IDS(&intel_coffeelake_gt3_info), @@ -793,6 +801,8 @@ static int __init i915_init(void) bool use_kms = true; int err; + i915_global_active_init(); + err = i915_mock_selftests(); if (err) return err > 0 ? 0 : err; @@ -824,6 +834,7 @@ static void __exit i915_exit(void) return; pci_unregister_driver(&i915_pci_driver); + i915_global_active_exit(); } module_init(i915_init); diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 5b1ae5ed97b3..9ebf99f3d8d3 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -1365,7 +1365,7 @@ static void i915_oa_stream_destroy(struct i915_perf_stream *stream) free_oa_buffer(dev_priv); intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, stream->wakeref); if (stream->ctx) oa_put_render_ctx_id(stream); @@ -1677,6 +1677,11 @@ static void gen8_update_reg_state_unlocked(struct i915_gem_context *ctx, CTX_REG(reg_state, state_offset, flex_regs[i], value); } + + CTX_REG(reg_state, CTX_R_PWR_CLK_STATE, GEN8_R_PWR_CLK_STATE, + gen8_make_rpcs(dev_priv, + &to_intel_context(ctx, + dev_priv->engine[RCS])->sseu)); } /* @@ -2087,7 +2092,7 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, * In our case we are expecting that taking pm + FORCEWAKE * references will effectively disable RC6. */ - intel_runtime_pm_get(dev_priv); + stream->wakeref = intel_runtime_pm_get(dev_priv); intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); ret = alloc_oa_buffer(dev_priv); @@ -2098,21 +2103,21 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, if (ret) goto err_lock; + stream->ops = &i915_oa_stream_ops; + dev_priv->perf.oa.exclusive_stream = stream; + ret = dev_priv->perf.oa.ops.enable_metric_set(stream); if (ret) { DRM_DEBUG("Unable to enable metric set\n"); goto err_enable; } - stream->ops = &i915_oa_stream_ops; - - dev_priv->perf.oa.exclusive_stream = stream; - mutex_unlock(&dev_priv->drm.struct_mutex); return 0; err_enable: + dev_priv->perf.oa.exclusive_stream = NULL; dev_priv->perf.oa.ops.disable_metric_set(dev_priv); mutex_unlock(&dev_priv->drm.struct_mutex); @@ -2123,7 +2128,7 @@ err_oa_buf_alloc: put_oa_config(dev_priv, stream->oa_config); intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, stream->wakeref); err_config: if (stream->ctx) @@ -3021,7 +3026,7 @@ static bool chv_is_valid_mux_addr(struct drm_i915_private *dev_priv, u32 addr) (addr >= 0x182300 && addr <= 0x1823A4); } -static uint32_t mask_reg_value(u32 reg, u32 val) +static u32 mask_reg_value(u32 reg, u32 val) { /* HALF_SLICE_CHICKEN2 is programmed with a the * WaDisableSTUnitPowerOptimization workaround. Make sure the value diff --git a/drivers/gpu/drm/i915/i915_pmu.c b/drivers/gpu/drm/i915/i915_pmu.c index d6c8f8fdfda5..13d70b90dd0f 100644 --- a/drivers/gpu/drm/i915/i915_pmu.c +++ b/drivers/gpu/drm/i915/i915_pmu.c @@ -167,6 +167,7 @@ engines_sample(struct drm_i915_private *dev_priv, unsigned int period_ns) { struct intel_engine_cs *engine; enum intel_engine_id id; + intel_wakeref_t wakeref; bool fw = false; if ((dev_priv->pmu.enable & ENGINE_SAMPLE_MASK) == 0) @@ -175,7 +176,8 @@ engines_sample(struct drm_i915_private *dev_priv, unsigned int period_ns) if (!dev_priv->gt.awake) return; - if (!intel_runtime_pm_get_if_in_use(dev_priv)) + wakeref = intel_runtime_pm_get_if_in_use(dev_priv); + if (!wakeref) return; for_each_engine(engine, dev_priv, id) { @@ -210,7 +212,7 @@ engines_sample(struct drm_i915_private *dev_priv, unsigned int period_ns) if (fw) intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); } static void @@ -227,11 +229,12 @@ frequency_sample(struct drm_i915_private *dev_priv, unsigned int period_ns) u32 val; val = dev_priv->gt_pm.rps.cur_freq; - if (dev_priv->gt.awake && - intel_runtime_pm_get_if_in_use(dev_priv)) { - val = intel_get_cagf(dev_priv, - I915_READ_NOTRACE(GEN6_RPSTAT1)); - intel_runtime_pm_put(dev_priv); + if (dev_priv->gt.awake) { + intel_wakeref_t wakeref; + + with_intel_runtime_pm_if_in_use(dev_priv, wakeref) + val = intel_get_cagf(dev_priv, + I915_READ_NOTRACE(GEN6_RPSTAT1)); } add_sample_mult(&dev_priv->pmu.sample[__I915_SAMPLE_FREQ_ACT], @@ -443,12 +446,14 @@ static u64 __get_rc6(struct drm_i915_private *i915) static u64 get_rc6(struct drm_i915_private *i915) { #if IS_ENABLED(CONFIG_PM) + intel_wakeref_t wakeref; unsigned long flags; u64 val; - if (intel_runtime_pm_get_if_in_use(i915)) { + wakeref = intel_runtime_pm_get_if_in_use(i915); + if (wakeref) { val = __get_rc6(i915); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); /* * If we are coming back from being runtime suspended we must @@ -594,7 +599,8 @@ static void i915_pmu_enable(struct perf_event *event) * Update the bitmask of enabled events and increment * the event reference counter. */ - GEM_BUG_ON(bit >= I915_PMU_MASK_BITS); + BUILD_BUG_ON(ARRAY_SIZE(i915->pmu.enable_count) != I915_PMU_MASK_BITS); + GEM_BUG_ON(bit >= ARRAY_SIZE(i915->pmu.enable_count)); GEM_BUG_ON(i915->pmu.enable_count[bit] == ~0); i915->pmu.enable |= BIT_ULL(bit); i915->pmu.enable_count[bit]++; @@ -615,11 +621,16 @@ static void i915_pmu_enable(struct perf_event *event) engine = intel_engine_lookup_user(i915, engine_event_class(event), engine_event_instance(event)); - GEM_BUG_ON(!engine); - engine->pmu.enable |= BIT(sample); - GEM_BUG_ON(sample >= I915_PMU_SAMPLE_BITS); + BUILD_BUG_ON(ARRAY_SIZE(engine->pmu.enable_count) != + I915_ENGINE_SAMPLE_COUNT); + BUILD_BUG_ON(ARRAY_SIZE(engine->pmu.sample) != + I915_ENGINE_SAMPLE_COUNT); + GEM_BUG_ON(sample >= ARRAY_SIZE(engine->pmu.enable_count)); + GEM_BUG_ON(sample >= ARRAY_SIZE(engine->pmu.sample)); GEM_BUG_ON(engine->pmu.enable_count[sample] == ~0); + + engine->pmu.enable |= BIT(sample); engine->pmu.enable_count[sample]++; } @@ -649,9 +660,11 @@ static void i915_pmu_disable(struct perf_event *event) engine = intel_engine_lookup_user(i915, engine_event_class(event), engine_event_instance(event)); - GEM_BUG_ON(!engine); - GEM_BUG_ON(sample >= I915_PMU_SAMPLE_BITS); + + GEM_BUG_ON(sample >= ARRAY_SIZE(engine->pmu.enable_count)); + GEM_BUG_ON(sample >= ARRAY_SIZE(engine->pmu.sample)); GEM_BUG_ON(engine->pmu.enable_count[sample] == 0); + /* * Decrement the reference count and clear the enabled * bitmask when the last listener on an event goes away. @@ -660,7 +673,7 @@ static void i915_pmu_disable(struct perf_event *event) engine->pmu.enable &= ~BIT(sample); } - GEM_BUG_ON(bit >= I915_PMU_MASK_BITS); + GEM_BUG_ON(bit >= ARRAY_SIZE(i915->pmu.enable_count)); GEM_BUG_ON(i915->pmu.enable_count[bit] == 0); /* * Decrement the reference count and clear the enabled diff --git a/drivers/gpu/drm/i915/i915_pmu.h b/drivers/gpu/drm/i915/i915_pmu.h index 7f164ca3db12..b3728c5f13e7 100644 --- a/drivers/gpu/drm/i915/i915_pmu.h +++ b/drivers/gpu/drm/i915/i915_pmu.h @@ -31,6 +31,8 @@ enum { ((1 << I915_PMU_SAMPLE_BITS) + \ (I915_PMU_LAST + 1 - __I915_PMU_OTHER(0))) +#define I915_ENGINE_SAMPLE_COUNT (I915_SAMPLE_SEMA + 1) + struct i915_pmu_sample { u64 cur; }; diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 44958d994bfa..638a586469f9 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -117,14 +117,14 @@ */ typedef struct { - uint32_t reg; + u32 reg; } i915_reg_t; #define _MMIO(r) ((const i915_reg_t){ .reg = (r) }) #define INVALID_MMIO_REG _MMIO(0) -static inline uint32_t i915_mmio_reg_offset(i915_reg_t reg) +static inline u32 i915_mmio_reg_offset(i915_reg_t reg) { return reg.reg; } @@ -1814,7 +1814,7 @@ enum i915_power_well_id { #define _CNL_PORT_TX_C_LN0_OFFSET 0x162C40 #define _CNL_PORT_TX_D_LN0_OFFSET 0x162E40 #define _CNL_PORT_TX_F_LN0_OFFSET 0x162840 -#define _CNL_PORT_TX_DW_GRP(port, dw) (_PICK((port), \ +#define _CNL_PORT_TX_DW_GRP(dw, port) (_PICK((port), \ _CNL_PORT_TX_AE_GRP_OFFSET, \ _CNL_PORT_TX_B_GRP_OFFSET, \ _CNL_PORT_TX_B_GRP_OFFSET, \ @@ -1822,7 +1822,7 @@ enum i915_power_well_id { _CNL_PORT_TX_AE_GRP_OFFSET, \ _CNL_PORT_TX_F_GRP_OFFSET) + \ 4 * (dw)) -#define _CNL_PORT_TX_DW_LN0(port, dw) (_PICK((port), \ +#define _CNL_PORT_TX_DW_LN0(dw, port) (_PICK((port), \ _CNL_PORT_TX_AE_LN0_OFFSET, \ _CNL_PORT_TX_B_LN0_OFFSET, \ _CNL_PORT_TX_B_LN0_OFFSET, \ @@ -1858,9 +1858,9 @@ enum i915_power_well_id { #define _CNL_PORT_TX_DW4_LN0_AE 0x162450 #define _CNL_PORT_TX_DW4_LN1_AE 0x1624D0 -#define CNL_PORT_TX_DW4_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 4)) -#define CNL_PORT_TX_DW4_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 4)) -#define CNL_PORT_TX_DW4_LN(port, ln) _MMIO(_CNL_PORT_TX_DW_LN0((port), 4) + \ +#define CNL_PORT_TX_DW4_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP(4, (port))) +#define CNL_PORT_TX_DW4_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0(4, (port))) +#define CNL_PORT_TX_DW4_LN(port, ln) _MMIO(_CNL_PORT_TX_DW_LN0(4, (port)) + \ ((ln) * (_CNL_PORT_TX_DW4_LN1_AE - \ _CNL_PORT_TX_DW4_LN0_AE))) #define ICL_PORT_TX_DW4_AUX(port) _MMIO(_ICL_PORT_TX_DW_AUX(4, port)) @@ -1888,8 +1888,8 @@ enum i915_power_well_id { #define RTERM_SELECT(x) ((x) << 3) #define RTERM_SELECT_MASK (0x7 << 3) -#define CNL_PORT_TX_DW7_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP((port), 7)) -#define CNL_PORT_TX_DW7_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0((port), 7)) +#define CNL_PORT_TX_DW7_GRP(port) _MMIO(_CNL_PORT_TX_DW_GRP(7, (port))) +#define CNL_PORT_TX_DW7_LN0(port) _MMIO(_CNL_PORT_TX_DW_LN0(7, (port))) #define ICL_PORT_TX_DW7_AUX(port) _MMIO(_ICL_PORT_TX_DW_AUX(7, port)) #define ICL_PORT_TX_DW7_GRP(port) _MMIO(_ICL_PORT_TX_DW_GRP(7, port)) #define ICL_PORT_TX_DW7_LN0(port) _MMIO(_ICL_PORT_TX_DW_LN(7, 0, port)) @@ -2801,6 +2801,9 @@ enum i915_power_well_id { #define GEN6_RCS_PWR_FSM _MMIO(0x22ac) #define GEN9_RCS_FE_FSM2 _MMIO(0x22a4) +#define GEN10_CACHE_MODE_SS _MMIO(0xe420) +#define FLOAT_BLEND_OPTIMIZATION_ENABLE (1 << 4) + /* Fuse readout registers for GT */ #define HSW_PAVP_FUSE1 _MMIO(0x911C) #define HSW_F1_EU_DIS_SHIFT 16 @@ -4272,6 +4275,15 @@ enum { #define EDP_PSR2_STATUS_STATE_MASK (0xf << 28) #define EDP_PSR2_STATUS_STATE_SHIFT 28 +#define _PSR2_SU_STATUS_0 0x6F914 +#define _PSR2_SU_STATUS_1 0x6F918 +#define _PSR2_SU_STATUS_2 0x6F91C +#define _PSR2_SU_STATUS(index) _MMIO(_PICK_EVEN((index), _PSR2_SU_STATUS_0, _PSR2_SU_STATUS_1)) +#define PSR2_SU_STATUS(frame) (_PSR2_SU_STATUS((frame) / 3)) +#define PSR2_SU_STATUS_SHIFT(frame) (((frame) % 3) * 10) +#define PSR2_SU_STATUS_MASK(frame) (0x3ff << PSR2_SU_STATUS_SHIFT(frame)) +#define PSR2_SU_STATUS_FRAMES 8 + /* VGA port control */ #define ADPA _MMIO(0x61100) #define PCH_ADPA _MMIO(0xe1100) @@ -4687,7 +4699,6 @@ enum { #define EDP_FORCE_VDD (1 << 3) #define EDP_BLC_ENABLE (1 << 2) #define PANEL_POWER_RESET (1 << 1) -#define PANEL_POWER_OFF (0 << 0) #define PANEL_POWER_ON (1 << 0) #define _PP_ON_DELAYS 0x61208 @@ -4887,6 +4898,7 @@ enum { # define TV_OVERSAMPLE_NONE (2 << 18) /* Selects 8x oversampling */ # define TV_OVERSAMPLE_8X (3 << 18) +# define TV_OVERSAMPLE_MASK (3 << 18) /* Selects progressive mode rather than interlaced */ # define TV_PROGRESSIVE (1 << 17) /* Sets the colorburst to PAL mode. Required for non-M PAL modes. */ @@ -5701,6 +5713,12 @@ enum { #define PIPEMISC_DITHER_TYPE_SP (0 << 2) #define PIPEMISC(pipe) _MMIO_PIPE2(pipe, _PIPE_MISC_A) +/* Skylake+ pipe bottom (background) color */ +#define _SKL_BOTTOM_COLOR_A 0x70034 +#define SKL_BOTTOM_COLOR_GAMMA_ENABLE (1 << 31) +#define SKL_BOTTOM_COLOR_CSC_ENABLE (1 << 30) +#define SKL_BOTTOM_COLOR(pipe) _MMIO_PIPE2(pipe, _SKL_BOTTOM_COLOR_A) + #define VLV_DPFLIPSTAT _MMIO(VLV_DISPLAY_BASE + 0x70028) #define PIPEB_LINE_COMPARE_INT_EN (1 << 29) #define PIPEB_HLINE_INT_EN (1 << 28) @@ -5982,7 +6000,7 @@ enum { #define PLANE_WM_EN (1 << 31) #define PLANE_WM_LINES_SHIFT 14 #define PLANE_WM_LINES_MASK 0x1f -#define PLANE_WM_BLOCKS_MASK 0x3ff +#define PLANE_WM_BLOCKS_MASK 0x7ff /* skl+: 10 bits, icl+ 11 bits */ #define _CUR_WM_0(pipe) _PIPE(pipe, _CUR_WM_A_0, _CUR_WM_B_0) #define CUR_WM(pipe, level) _MMIO(_CUR_WM_0(pipe) + ((4) * (level))) @@ -6766,8 +6784,7 @@ enum { #define _PLANE_BUF_CFG_1_B 0x7127c #define _PLANE_BUF_CFG_2_B 0x7137c -#define SKL_DDB_ENTRY_MASK 0x3FF -#define ICL_DDB_ENTRY_MASK 0x7FF +#define DDB_ENTRY_MASK 0x7FF /* skl+: 10 bits, icl+ 11 bits */ #define DDB_ENTRY_END_SHIFT 16 #define _PLANE_BUF_CFG_1(pipe) \ _PIPE(pipe, _PLANE_BUF_CFG_1_A, _PLANE_BUF_CFG_1_B) @@ -7600,6 +7617,7 @@ enum { #define _PIPEB_CHICKEN 0x71038 #define _PIPEC_CHICKEN 0x72038 #define PER_PIXEL_ALPHA_BYPASS_EN (1 << 7) +#define PM_FILL_MAINTAIN_DBUF_FULLNESS (1 << 0) #define PIPE_CHICKEN(pipe) _MMIO_PIPE(pipe, _PIPEA_CHICKEN,\ _PIPEB_CHICKEN) @@ -9545,7 +9563,7 @@ enum skl_power_gate { #define _MG_PLL3_ENABLE 0x46038 #define _MG_PLL4_ENABLE 0x4603C /* Bits are the same as DPLL0_ENABLE */ -#define MG_PLL_ENABLE(port) _MMIO_PORT((port) - PORT_C, _MG_PLL1_ENABLE, \ +#define MG_PLL_ENABLE(tc_port) _MMIO_PORT((tc_port), _MG_PLL1_ENABLE, \ _MG_PLL2_ENABLE) #define _MG_REFCLKIN_CTL_PORT1 0x16892C @@ -9554,9 +9572,9 @@ enum skl_power_gate { #define _MG_REFCLKIN_CTL_PORT4 0x16B92C #define MG_REFCLKIN_CTL_OD_2_MUX(x) ((x) << 8) #define MG_REFCLKIN_CTL_OD_2_MUX_MASK (0x7 << 8) -#define MG_REFCLKIN_CTL(port) _MMIO_PORT((port) - PORT_C, \ - _MG_REFCLKIN_CTL_PORT1, \ - _MG_REFCLKIN_CTL_PORT2) +#define MG_REFCLKIN_CTL(tc_port) _MMIO_PORT((tc_port), \ + _MG_REFCLKIN_CTL_PORT1, \ + _MG_REFCLKIN_CTL_PORT2) #define _MG_CLKTOP2_CORECLKCTL1_PORT1 0x1688D8 #define _MG_CLKTOP2_CORECLKCTL1_PORT2 0x1698D8 @@ -9566,9 +9584,9 @@ enum skl_power_gate { #define MG_CLKTOP2_CORECLKCTL1_B_DIVRATIO_MASK (0xff << 16) #define MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO(x) ((x) << 8) #define MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO_MASK (0xff << 8) -#define MG_CLKTOP2_CORECLKCTL1(port) _MMIO_PORT((port) - PORT_C, \ - _MG_CLKTOP2_CORECLKCTL1_PORT1, \ - _MG_CLKTOP2_CORECLKCTL1_PORT2) +#define MG_CLKTOP2_CORECLKCTL1(tc_port) _MMIO_PORT((tc_port), \ + _MG_CLKTOP2_CORECLKCTL1_PORT1, \ + _MG_CLKTOP2_CORECLKCTL1_PORT2) #define _MG_CLKTOP2_HSCLKCTL_PORT1 0x1688D4 #define _MG_CLKTOP2_HSCLKCTL_PORT2 0x1698D4 @@ -9586,9 +9604,9 @@ enum skl_power_gate { #define MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO(x) ((x) << 8) #define MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_SHIFT 8 #define MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_MASK (0xf << 8) -#define MG_CLKTOP2_HSCLKCTL(port) _MMIO_PORT((port) - PORT_C, \ - _MG_CLKTOP2_HSCLKCTL_PORT1, \ - _MG_CLKTOP2_HSCLKCTL_PORT2) +#define MG_CLKTOP2_HSCLKCTL(tc_port) _MMIO_PORT((tc_port), \ + _MG_CLKTOP2_HSCLKCTL_PORT1, \ + _MG_CLKTOP2_HSCLKCTL_PORT2) #define _MG_PLL_DIV0_PORT1 0x168A00 #define _MG_PLL_DIV0_PORT2 0x169A00 @@ -9600,8 +9618,8 @@ enum skl_power_gate { #define MG_PLL_DIV0_FBDIV_FRAC(x) ((x) << 8) #define MG_PLL_DIV0_FBDIV_INT_MASK (0xff << 0) #define MG_PLL_DIV0_FBDIV_INT(x) ((x) << 0) -#define MG_PLL_DIV0(port) _MMIO_PORT((port) - PORT_C, _MG_PLL_DIV0_PORT1, \ - _MG_PLL_DIV0_PORT2) +#define MG_PLL_DIV0(tc_port) _MMIO_PORT((tc_port), _MG_PLL_DIV0_PORT1, \ + _MG_PLL_DIV0_PORT2) #define _MG_PLL_DIV1_PORT1 0x168A04 #define _MG_PLL_DIV1_PORT2 0x169A04 @@ -9615,8 +9633,8 @@ enum skl_power_gate { #define MG_PLL_DIV1_NDIVRATIO(x) ((x) << 4) #define MG_PLL_DIV1_FBPREDIV_MASK (0xf << 0) #define MG_PLL_DIV1_FBPREDIV(x) ((x) << 0) -#define MG_PLL_DIV1(port) _MMIO_PORT((port) - PORT_C, _MG_PLL_DIV1_PORT1, \ - _MG_PLL_DIV1_PORT2) +#define MG_PLL_DIV1(tc_port) _MMIO_PORT((tc_port), _MG_PLL_DIV1_PORT1, \ + _MG_PLL_DIV1_PORT2) #define _MG_PLL_LF_PORT1 0x168A08 #define _MG_PLL_LF_PORT2 0x169A08 @@ -9628,8 +9646,8 @@ enum skl_power_gate { #define MG_PLL_LF_GAINCTRL(x) ((x) << 16) #define MG_PLL_LF_INT_COEFF(x) ((x) << 8) #define MG_PLL_LF_PROP_COEFF(x) ((x) << 0) -#define MG_PLL_LF(port) _MMIO_PORT((port) - PORT_C, _MG_PLL_LF_PORT1, \ - _MG_PLL_LF_PORT2) +#define MG_PLL_LF(tc_port) _MMIO_PORT((tc_port), _MG_PLL_LF_PORT1, \ + _MG_PLL_LF_PORT2) #define _MG_PLL_FRAC_LOCK_PORT1 0x168A0C #define _MG_PLL_FRAC_LOCK_PORT2 0x169A0C @@ -9641,9 +9659,9 @@ enum skl_power_gate { #define MG_PLL_FRAC_LOCK_DCODITHEREN (1 << 10) #define MG_PLL_FRAC_LOCK_FEEDFWRDCAL_EN (1 << 8) #define MG_PLL_FRAC_LOCK_FEEDFWRDGAIN(x) ((x) << 0) -#define MG_PLL_FRAC_LOCK(port) _MMIO_PORT((port) - PORT_C, \ - _MG_PLL_FRAC_LOCK_PORT1, \ - _MG_PLL_FRAC_LOCK_PORT2) +#define MG_PLL_FRAC_LOCK(tc_port) _MMIO_PORT((tc_port), \ + _MG_PLL_FRAC_LOCK_PORT1, \ + _MG_PLL_FRAC_LOCK_PORT2) #define _MG_PLL_SSC_PORT1 0x168A10 #define _MG_PLL_SSC_PORT2 0x169A10 @@ -9655,8 +9673,8 @@ enum skl_power_gate { #define MG_PLL_SSC_STEPNUM(x) ((x) << 10) #define MG_PLL_SSC_FLLEN (1 << 9) #define MG_PLL_SSC_STEPSIZE(x) ((x) << 0) -#define MG_PLL_SSC(port) _MMIO_PORT((port) - PORT_C, _MG_PLL_SSC_PORT1, \ - _MG_PLL_SSC_PORT2) +#define MG_PLL_SSC(tc_port) _MMIO_PORT((tc_port), _MG_PLL_SSC_PORT1, \ + _MG_PLL_SSC_PORT2) #define _MG_PLL_BIAS_PORT1 0x168A14 #define _MG_PLL_BIAS_PORT2 0x169A14 @@ -9675,8 +9693,8 @@ enum skl_power_gate { #define MG_PLL_BIAS_VREF_RDAC_MASK (0x7 << 5) #define MG_PLL_BIAS_IREFTRIM(x) ((x) << 0) #define MG_PLL_BIAS_IREFTRIM_MASK (0x1f << 0) -#define MG_PLL_BIAS(port) _MMIO_PORT((port) - PORT_C, _MG_PLL_BIAS_PORT1, \ - _MG_PLL_BIAS_PORT2) +#define MG_PLL_BIAS(tc_port) _MMIO_PORT((tc_port), _MG_PLL_BIAS_PORT1, \ + _MG_PLL_BIAS_PORT2) #define _MG_PLL_TDC_COLDST_BIAS_PORT1 0x168A18 #define _MG_PLL_TDC_COLDST_BIAS_PORT2 0x169A18 @@ -9687,9 +9705,9 @@ enum skl_power_gate { #define MG_PLL_TDC_COLDST_COLDSTART (1 << 16) #define MG_PLL_TDC_TDCOVCCORR_EN (1 << 2) #define MG_PLL_TDC_TDCSEL(x) ((x) << 0) -#define MG_PLL_TDC_COLDST_BIAS(port) _MMIO_PORT((port) - PORT_C, \ - _MG_PLL_TDC_COLDST_BIAS_PORT1, \ - _MG_PLL_TDC_COLDST_BIAS_PORT2) +#define MG_PLL_TDC_COLDST_BIAS(tc_port) _MMIO_PORT((tc_port), \ + _MG_PLL_TDC_COLDST_BIAS_PORT1, \ + _MG_PLL_TDC_COLDST_BIAS_PORT2) #define _CNL_DPLL0_CFGCR0 0x6C000 #define _CNL_DPLL1_CFGCR0 0x6C080 diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index d1355154886a..c2a5c48c7541 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -29,6 +29,8 @@ #include <linux/sched/signal.h> #include "i915_drv.h" +#include "i915_active.h" +#include "i915_reset.h" static const char *i915_fence_get_driver_name(struct dma_fence *fence) { @@ -59,7 +61,7 @@ static bool i915_fence_signaled(struct dma_fence *fence) static bool i915_fence_enable_signaling(struct dma_fence *fence) { - return intel_engine_enable_signaling(to_request(fence), true); + return i915_request_enable_breadcrumb(to_request(fence)); } static signed long i915_fence_wait(struct dma_fence *fence, @@ -124,12 +126,6 @@ static void unreserve_gt(struct drm_i915_private *i915) i915_gem_park(i915); } -void i915_gem_retire_noop(struct i915_gem_active *active, - struct i915_request *request) -{ - /* Space left intentionally blank */ -} - static void advance_ring(struct i915_request *request) { struct intel_ring *ring = request->ring; @@ -181,10 +177,11 @@ static void free_capture_list(struct i915_request *request) static void __retire_engine_request(struct intel_engine_cs *engine, struct i915_request *rq) { - GEM_TRACE("%s(%s) fence %llx:%lld, global=%d, current %d\n", + GEM_TRACE("%s(%s) fence %llx:%lld, global=%d, current %d:%d\n", __func__, engine->name, rq->fence.context, rq->fence.seqno, rq->global_seqno, + hwsp_seqno(rq), intel_engine_get_seqno(engine)); GEM_BUG_ON(!i915_request_completed(rq)); @@ -197,10 +194,11 @@ static void __retire_engine_request(struct intel_engine_cs *engine, spin_unlock(&engine->timeline.lock); spin_lock(&rq->lock); - if (!test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags)) + i915_request_mark_complete(rq); + if (!i915_request_signaled(rq)) dma_fence_signal_locked(&rq->fence); if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &rq->fence.flags)) - intel_engine_cancel_signaling(rq); + i915_request_cancel_breadcrumb(rq); if (rq->waitboost) { GEM_BUG_ON(!atomic_read(&rq->i915->gt_pm.rps.num_waiters)); atomic_dec(&rq->i915->gt_pm.rps.num_waiters); @@ -241,12 +239,13 @@ static void __retire_engine_upto(struct intel_engine_cs *engine, static void i915_request_retire(struct i915_request *request) { - struct i915_gem_active *active, *next; + struct i915_active_request *active, *next; - GEM_TRACE("%s fence %llx:%lld, global=%d, current %d\n", + GEM_TRACE("%s fence %llx:%lld, global=%d, current %d:%d\n", request->engine->name, request->fence.context, request->fence.seqno, request->global_seqno, + hwsp_seqno(request), intel_engine_get_seqno(request->engine)); lockdep_assert_held(&request->i915->drm.struct_mutex); @@ -274,10 +273,10 @@ static void i915_request_retire(struct i915_request *request) * we may spend an inordinate amount of time simply handling * the retirement of requests and processing their callbacks. * Of which, this loop itself is particularly hot due to the - * cache misses when jumping around the list of i915_gem_active. - * So we try to keep this loop as streamlined as possible and - * also prefetch the next i915_gem_active to try and hide - * the likely cache miss. + * cache misses when jumping around the list of + * i915_active_request. So we try to keep this loop as + * streamlined as possible and also prefetch the next + * i915_active_request to try and hide the likely cache miss. */ prefetchw(next); @@ -306,10 +305,11 @@ void i915_request_retire_upto(struct i915_request *rq) struct intel_ring *ring = rq->ring; struct i915_request *tmp; - GEM_TRACE("%s fence %llx:%lld, global=%d, current %d\n", + GEM_TRACE("%s fence %llx:%lld, global=%d, current %d:%d\n", rq->engine->name, rq->fence.context, rq->fence.seqno, rq->global_seqno, + hwsp_seqno(rq), intel_engine_get_seqno(rq->engine)); lockdep_assert_held(&rq->i915->drm.struct_mutex); @@ -328,7 +328,7 @@ void i915_request_retire_upto(struct i915_request *rq) static u32 timeline_get_seqno(struct i915_timeline *tl) { - return ++tl->seqno; + return tl->seqno += 1 + tl->has_initial_breadcrumb; } static void move_to_timeline(struct i915_request *request, @@ -342,15 +342,23 @@ static void move_to_timeline(struct i915_request *request, spin_unlock(&request->timeline->lock); } +static u32 next_global_seqno(struct i915_timeline *tl) +{ + if (!++tl->seqno) + ++tl->seqno; + return tl->seqno; +} + void __i915_request_submit(struct i915_request *request) { struct intel_engine_cs *engine = request->engine; u32 seqno; - GEM_TRACE("%s fence %llx:%lld -> global=%d, current %d\n", + GEM_TRACE("%s fence %llx:%lld -> global=%d, current %d:%d\n", engine->name, request->fence.context, request->fence.seqno, engine->timeline.seqno + 1, + hwsp_seqno(request), intel_engine_get_seqno(engine)); GEM_BUG_ON(!irqs_disabled()); @@ -358,26 +366,27 @@ void __i915_request_submit(struct i915_request *request) GEM_BUG_ON(request->global_seqno); - seqno = timeline_get_seqno(&engine->timeline); + seqno = next_global_seqno(&engine->timeline); GEM_BUG_ON(!seqno); GEM_BUG_ON(intel_engine_signaled(engine, seqno)); /* We may be recursing from the signal callback of another i915 fence */ spin_lock_nested(&request->lock, SINGLE_DEPTH_NESTING); + GEM_BUG_ON(test_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags)); + set_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags); request->global_seqno = seqno; - if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &request->fence.flags)) - intel_engine_enable_signaling(request, false); + if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &request->fence.flags) && + !i915_request_enable_breadcrumb(request)) + intel_engine_queue_breadcrumbs(engine); spin_unlock(&request->lock); - engine->emit_breadcrumb(request, - request->ring->vaddr + request->postfix); + engine->emit_fini_breadcrumb(request, + request->ring->vaddr + request->postfix); /* Transfer from per-context onto the global per-engine timeline */ move_to_timeline(request, &engine->timeline); trace_i915_request_execute(request); - - wake_up_all(&request->execute); } void i915_request_submit(struct i915_request *request) @@ -397,10 +406,11 @@ void __i915_request_unsubmit(struct i915_request *request) { struct intel_engine_cs *engine = request->engine; - GEM_TRACE("%s fence %llx:%lld <- global=%d, current %d\n", + GEM_TRACE("%s fence %llx:%lld <- global=%d, current %d:%d\n", engine->name, request->fence.context, request->fence.seqno, request->global_seqno, + hwsp_seqno(request), intel_engine_get_seqno(engine)); GEM_BUG_ON(!irqs_disabled()); @@ -419,7 +429,9 @@ void __i915_request_unsubmit(struct i915_request *request) spin_lock_nested(&request->lock, SINGLE_DEPTH_NESTING); request->global_seqno = 0; if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &request->fence.flags)) - intel_engine_cancel_signaling(request); + i915_request_cancel_breadcrumb(request); + GEM_BUG_ON(!test_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags)); + clear_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags); spin_unlock(&request->lock); /* Transfer back from the global per-engine timeline to per-context */ @@ -509,6 +521,11 @@ out: return kmem_cache_alloc(ce->gem_context->i915->requests, GFP_KERNEL); } +static int add_timeline_barrier(struct i915_request *rq) +{ + return i915_request_await_active_request(rq, &rq->timeline->barrier); +} + /** * i915_request_alloc - allocate a request structure * @@ -565,7 +582,7 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) * We use RCU to look up requests in flight. The lookups may * race with the request being allocated from the slab freelist. * That is the request we are writing to here, may be in the process - * of being read by __i915_gem_active_get_rcu(). As such, + * of being read by __i915_active_request_get_rcu(). As such, * we have to be very careful when overwriting the contents. During * the RCU lookup, we change chase the request->engine pointer, * read the request->global_seqno and increment the reference count. @@ -608,6 +625,7 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) rq->ring = ce->ring; rq->timeline = ce->ring->timeline; GEM_BUG_ON(rq->timeline == &engine->timeline); + rq->hwsp_seqno = rq->timeline->hwsp_seqno; spin_lock_init(&rq->lock); dma_fence_init(&rq->fence, @@ -618,13 +636,11 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) /* We bump the ref for the fence chain */ i915_sw_fence_init(&i915_request_get(rq)->submit, submit_notify); - init_waitqueue_head(&rq->execute); i915_sched_node_init(&rq->sched); /* No zalloc, must clear what we need by hand */ rq->global_seqno = 0; - rq->signaling.wait.seqno = 0; rq->file_priv = NULL; rq->batch = NULL; rq->capture_list = NULL; @@ -642,7 +658,7 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) * around inside i915_request_add() there is sufficient space at * the beginning of the ring as well. */ - rq->reserved_space = 2 * engine->emit_breadcrumb_sz * sizeof(u32); + rq->reserved_space = 2 * engine->emit_fini_breadcrumb_dw * sizeof(u32); /* * Record the position of the start of the request so that @@ -652,6 +668,10 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) */ rq->head = rq->ring->emit; + ret = add_timeline_barrier(rq); + if (ret) + goto err_unwind; + ret = engine->request_alloc(rq); if (ret) goto err_unwind; @@ -893,7 +913,7 @@ void i915_request_add(struct i915_request *request) * GPU processing the request, we never over-estimate the * position of the ring's HEAD. */ - cs = intel_ring_begin(request, engine->emit_breadcrumb_sz); + cs = intel_ring_begin(request, engine->emit_fini_breadcrumb_dw); GEM_BUG_ON(IS_ERR(cs)); request->postfix = intel_ring_offset(request, cs); @@ -904,8 +924,8 @@ void i915_request_add(struct i915_request *request) * see a more recent value in the hws than we are tracking. */ - prev = i915_gem_active_raw(&timeline->last_request, - &request->i915->drm.struct_mutex); + prev = i915_active_request_raw(&timeline->last_request, + &request->i915->drm.struct_mutex); if (prev && !i915_request_completed(prev)) { i915_sw_fence_await_sw_fence(&request->submit, &prev->submit, &request->submitq); @@ -921,7 +941,7 @@ void i915_request_add(struct i915_request *request) spin_unlock_irq(&timeline->lock); GEM_BUG_ON(timeline->seqno != request->fence.seqno); - i915_gem_active_set(&timeline->last_request, request); + __i915_active_request_set(&timeline->last_request, request); list_add_tail(&request->ring_link, &ring->request_list); if (list_is_first(&request->ring_link, &ring->request_list)) { @@ -952,7 +972,7 @@ void i915_request_add(struct i915_request *request) * Allow interactive/synchronous clients to jump ahead of * the bulk clients. (FQ_CODEL) */ - if (!prev || i915_request_completed(prev)) + if (list_empty(&request->sched.signalers_list)) attr.priority |= I915_PRIORITY_NEWCLIENT; engine->schedule(request, &attr); @@ -1015,13 +1035,10 @@ static bool busywait_stop(unsigned long timeout, unsigned int cpu) return this_cpu != cpu; } -static bool __i915_spin_request(const struct i915_request *rq, - u32 seqno, int state, unsigned long timeout_us) +static bool __i915_spin_request(const struct i915_request * const rq, + int state, unsigned long timeout_us) { - struct intel_engine_cs *engine = rq->engine; - unsigned int irq, cpu; - - GEM_BUG_ON(!seqno); + unsigned int cpu; /* * Only wait for the request if we know it is likely to complete. @@ -1029,12 +1046,12 @@ static bool __i915_spin_request(const struct i915_request *rq, * We don't track the timestamps around requests, nor the average * request length, so we do not have a good indicator that this * request will complete within the timeout. What we do know is the - * order in which requests are executed by the engine and so we can - * tell if the request has started. If the request hasn't started yet, - * it is a fair assumption that it will not complete within our - * relatively short timeout. + * order in which requests are executed by the context and so we can + * tell if the request has been started. If the request is not even + * running yet, it is a fair assumption that it will not complete + * within our relatively short timeout. */ - if (!intel_engine_has_started(engine, seqno)) + if (!i915_request_is_running(rq)) return false; /* @@ -1048,20 +1065,10 @@ static bool __i915_spin_request(const struct i915_request *rq, * takes to sleep on a request, on the order of a microsecond. */ - irq = READ_ONCE(engine->breadcrumbs.irq_count); timeout_us += local_clock_us(&cpu); do { - if (intel_engine_has_completed(engine, seqno)) - return seqno == i915_request_global_seqno(rq); - - /* - * Seqno are meant to be ordered *before* the interrupt. If - * we see an interrupt without a corresponding seqno advance, - * assume we won't see one in the near future but require - * the engine->seqno_barrier() to fixup coherency. - */ - if (READ_ONCE(engine->breadcrumbs.irq_count) != irq) - break; + if (i915_request_completed(rq)) + return true; if (signal_pending_state(state, current)) break; @@ -1075,16 +1082,16 @@ static bool __i915_spin_request(const struct i915_request *rq, return false; } -static bool __i915_wait_request_check_and_reset(struct i915_request *request) -{ - struct i915_gpu_error *error = &request->i915->gpu_error; +struct request_wait { + struct dma_fence_cb cb; + struct task_struct *tsk; +}; - if (likely(!i915_reset_handoff(error))) - return false; +static void request_wait_wake(struct dma_fence *fence, struct dma_fence_cb *cb) +{ + struct request_wait *wait = container_of(cb, typeof(*wait), cb); - __set_current_state(TASK_RUNNING); - i915_reset(request->i915, error->stalled_mask, error->reason); - return true; + wake_up_process(wait->tsk); } /** @@ -1112,17 +1119,9 @@ long i915_request_wait(struct i915_request *rq, { const int state = flags & I915_WAIT_INTERRUPTIBLE ? TASK_INTERRUPTIBLE : TASK_UNINTERRUPTIBLE; - wait_queue_head_t *errq = &rq->i915->gpu_error.wait_queue; - DEFINE_WAIT_FUNC(reset, default_wake_function); - DEFINE_WAIT_FUNC(exec, default_wake_function); - struct intel_wait wait; + struct request_wait wait; might_sleep(); -#if IS_ENABLED(CONFIG_LOCKDEP) - GEM_BUG_ON(debug_locks && - !!lockdep_is_held(&rq->i915->drm.struct_mutex) != - !!(flags & I915_WAIT_LOCKED)); -#endif GEM_BUG_ON(timeout < 0); if (i915_request_completed(rq)) @@ -1133,57 +1132,23 @@ long i915_request_wait(struct i915_request *rq, trace_i915_request_wait_begin(rq, flags); - add_wait_queue(&rq->execute, &exec); - if (flags & I915_WAIT_LOCKED) - add_wait_queue(errq, &reset); + /* Optimistic short spin before touching IRQs */ + if (__i915_spin_request(rq, state, 5)) + goto out; - intel_wait_init(&wait); if (flags & I915_WAIT_PRIORITY) i915_schedule_bump_priority(rq, I915_PRIORITY_WAIT); -restart: - do { - set_current_state(state); - if (intel_wait_update_request(&wait, rq)) - break; - - if (flags & I915_WAIT_LOCKED && - __i915_wait_request_check_and_reset(rq)) - continue; - - if (signal_pending_state(state, current)) { - timeout = -ERESTARTSYS; - goto complete; - } - - if (!timeout) { - timeout = -ETIME; - goto complete; - } - - timeout = io_schedule_timeout(timeout); - } while (1); - - GEM_BUG_ON(!intel_wait_has_seqno(&wait)); - GEM_BUG_ON(!i915_sw_fence_signaled(&rq->submit)); - - /* Optimistic short spin before touching IRQs */ - if (__i915_spin_request(rq, wait.seqno, state, 5)) - goto complete; + wait.tsk = current; + if (dma_fence_add_callback(&rq->fence, &wait.cb, request_wait_wake)) + goto out; - set_current_state(state); - if (intel_engine_add_wait(rq->engine, &wait)) - /* - * In order to check that we haven't missed the interrupt - * as we enabled it, we need to kick ourselves to do a - * coherent check on the seqno before we sleep. - */ - goto wakeup; + for (;;) { + set_current_state(state); - if (flags & I915_WAIT_LOCKED) - __i915_wait_request_check_and_reset(rq); + if (i915_request_completed(rq)) + break; - for (;;) { if (signal_pending_state(state, current)) { timeout = -ERESTARTSYS; break; @@ -1195,50 +1160,13 @@ restart: } timeout = io_schedule_timeout(timeout); - - if (intel_wait_complete(&wait) && - intel_wait_check_request(&wait, rq)) - break; - - set_current_state(state); - -wakeup: - if (i915_request_completed(rq)) - break; - - /* - * If the GPU is hung, and we hold the lock, reset the GPU - * and then check for completion. On a full reset, the engine's - * HW seqno will be advanced passed us and we are complete. - * If we do a partial reset, we have to wait for the GPU to - * resume and update the breadcrumb. - * - * If we don't hold the mutex, we can just wait for the worker - * to come along and update the breadcrumb (either directly - * itself, or indirectly by recovering the GPU). - */ - if (flags & I915_WAIT_LOCKED && - __i915_wait_request_check_and_reset(rq)) - continue; - - /* Only spin if we know the GPU is processing this request */ - if (__i915_spin_request(rq, wait.seqno, state, 2)) - break; - - if (!intel_wait_check_request(&wait, rq)) { - intel_engine_remove_wait(rq->engine, &wait); - goto restart; - } } - - intel_engine_remove_wait(rq->engine, &wait); -complete: __set_current_state(TASK_RUNNING); - if (flags & I915_WAIT_LOCKED) - remove_wait_queue(errq, &reset); - remove_wait_queue(&rq->execute, &exec); - trace_i915_request_wait_end(rq); + dma_fence_remove_callback(&rq->fence, &wait.cb); + +out: + trace_i915_request_wait_end(rq); return timeout; } diff --git a/drivers/gpu/drm/i915/i915_request.h b/drivers/gpu/drm/i915/i915_request.h index d014b0605445..40f3e8dcbdd5 100644 --- a/drivers/gpu/drm/i915/i915_request.h +++ b/drivers/gpu/drm/i915/i915_request.h @@ -38,23 +38,34 @@ struct drm_i915_gem_object; struct i915_request; struct i915_timeline; -struct intel_wait { - struct rb_node node; - struct task_struct *tsk; - struct i915_request *request; - u32 seqno; -}; - -struct intel_signal_node { - struct intel_wait wait; - struct list_head link; -}; - struct i915_capture_list { struct i915_capture_list *next; struct i915_vma *vma; }; +enum { + /* + * I915_FENCE_FLAG_ACTIVE - this request is currently submitted to HW. + * + * Set by __i915_request_submit() on handing over to HW, and cleared + * by __i915_request_unsubmit() if we preempt this request. + * + * Finally cleared for consistency on retiring the request, when + * we know the HW is no longer running this request. + * + * See i915_request_is_active() + */ + I915_FENCE_FLAG_ACTIVE = DMA_FENCE_FLAG_USER_BITS, + + /* + * I915_FENCE_FLAG_SIGNAL - this request is currently on signal_list + * + * Internal bookkeeping used by the breadcrumb code to track when + * a request is on the various signal_list. + */ + I915_FENCE_FLAG_SIGNAL, +}; + /** * Request queue structure. * @@ -97,7 +108,7 @@ struct i915_request { struct intel_context *hw_context; struct intel_ring *ring; struct i915_timeline *timeline; - struct intel_signal_node signaling; + struct list_head signal_link; /* * The rcu epoch of when this request was allocated. Used to judiciously @@ -116,7 +127,6 @@ struct i915_request { */ struct i915_sw_fence submit; wait_queue_entry_t submitq; - wait_queue_head_t execute; /* * A list of everyone we wait upon, and everyone who waits upon us. @@ -130,6 +140,13 @@ struct i915_request { struct i915_sched_node sched; struct i915_dependency dep; + /* + * A convenience pointer to the current breadcrumb value stored in + * the HW status page (or our timeline's local equivalent). The full + * path would be rq->hw_context->ring->timeline->hwsp_seqno. + */ + const u32 *hwsp_seqno; + /** * GEM sequence number associated with this request on the * global execution timeline. It is zero when the request is not @@ -248,7 +265,7 @@ i915_request_put(struct i915_request *rq) * that it has passed the global seqno and the global seqno is unchanged * after the read, it is indeed complete). */ -static u32 +static inline u32 i915_request_global_seqno(const struct i915_request *request) { return READ_ONCE(request->global_seqno); @@ -270,6 +287,10 @@ void i915_request_skip(struct i915_request *request, int error); void __i915_request_unsubmit(struct i915_request *request); void i915_request_unsubmit(struct i915_request *request); +/* Note: part of the intel_breadcrumbs family */ +bool i915_request_enable_breadcrumb(struct i915_request *request); +void i915_request_cancel_breadcrumb(struct i915_request *request); + long i915_request_wait(struct i915_request *rq, unsigned int flags, long timeout) @@ -280,441 +301,106 @@ long i915_request_wait(struct i915_request *rq, #define I915_WAIT_ALL BIT(3) /* used by i915_gem_object_wait() */ #define I915_WAIT_FOR_IDLE_BOOST BIT(4) -static inline bool intel_engine_has_started(struct intel_engine_cs *engine, - u32 seqno); -static inline bool intel_engine_has_completed(struct intel_engine_cs *engine, - u32 seqno); - -/** - * Returns true if seq1 is later than seq2. - */ -static inline bool i915_seqno_passed(u32 seq1, u32 seq2) -{ - return (s32)(seq1 - seq2) >= 0; -} - -/** - * i915_request_started - check if the request has begun being executed - * @rq: the request - * - * Returns true if the request has been submitted to hardware, and the hardware - * has advanced passed the end of the previous request and so should be either - * currently processing the request (though it may be preempted and so - * not necessarily the next request to complete) or have completed the request. - */ -static inline bool i915_request_started(const struct i915_request *rq) -{ - u32 seqno; - - seqno = i915_request_global_seqno(rq); - if (!seqno) /* not yet submitted to HW */ - return false; - - return intel_engine_has_started(rq->engine, seqno); -} - -static inline bool -__i915_request_completed(const struct i915_request *rq, u32 seqno) -{ - GEM_BUG_ON(!seqno); - return intel_engine_has_completed(rq->engine, seqno) && - seqno == i915_request_global_seqno(rq); -} - -static inline bool i915_request_completed(const struct i915_request *rq) -{ - u32 seqno; - - seqno = i915_request_global_seqno(rq); - if (!seqno) - return false; - - return __i915_request_completed(rq, seqno); -} - -void i915_retire_requests(struct drm_i915_private *i915); - -/* - * We treat requests as fences. This is not be to confused with our - * "fence registers" but pipeline synchronisation objects ala GL_ARB_sync. - * We use the fences to synchronize access from the CPU with activity on the - * GPU, for example, we should not rewrite an object's PTE whilst the GPU - * is reading them. We also track fences at a higher level to provide - * implicit synchronisation around GEM objects, e.g. set-domain will wait - * for outstanding GPU rendering before marking the object ready for CPU - * access, or a pageflip will wait until the GPU is complete before showing - * the frame on the scanout. - * - * In order to use a fence, the object must track the fence it needs to - * serialise with. For example, GEM objects want to track both read and - * write access so that we can perform concurrent read operations between - * the CPU and GPU engines, as well as waiting for all rendering to - * complete, or waiting for the last GPU user of a "fence register". The - * object then embeds a #i915_gem_active to track the most recent (in - * retirement order) request relevant for the desired mode of access. - * The #i915_gem_active is updated with i915_gem_active_set() to track the - * most recent fence request, typically this is done as part of - * i915_vma_move_to_active(). - * - * When the #i915_gem_active completes (is retired), it will - * signal its completion to the owner through a callback as well as mark - * itself as idle (i915_gem_active.request == NULL). The owner - * can then perform any action, such as delayed freeing of an active - * resource including itself. - */ -struct i915_gem_active; - -typedef void (*i915_gem_retire_fn)(struct i915_gem_active *, - struct i915_request *); - -struct i915_gem_active { - struct i915_request __rcu *request; - struct list_head link; - i915_gem_retire_fn retire; -}; - -void i915_gem_retire_noop(struct i915_gem_active *, - struct i915_request *request); - -/** - * init_request_active - prepares the activity tracker for use - * @active - the active tracker - * @func - a callback when then the tracker is retired (becomes idle), - * can be NULL - * - * init_request_active() prepares the embedded @active struct for use as - * an activity tracker, that is for tracking the last known active request - * associated with it. When the last request becomes idle, when it is retired - * after completion, the optional callback @func is invoked. - */ -static inline void -init_request_active(struct i915_gem_active *active, - i915_gem_retire_fn retire) +static inline bool i915_request_signaled(const struct i915_request *rq) { - RCU_INIT_POINTER(active->request, NULL); - INIT_LIST_HEAD(&active->link); - active->retire = retire ?: i915_gem_retire_noop; + /* The request may live longer than its HWSP, so check flags first! */ + return test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags); } -/** - * i915_gem_active_set - updates the tracker to watch the current request - * @active - the active tracker - * @request - the request to watch - * - * i915_gem_active_set() watches the given @request for completion. Whilst - * that @request is busy, the @active reports busy. When that @request is - * retired, the @active tracker is updated to report idle. - */ -static inline void -i915_gem_active_set(struct i915_gem_active *active, - struct i915_request *request) +static inline bool i915_request_is_active(const struct i915_request *rq) { - list_move(&active->link, &request->active_list); - rcu_assign_pointer(active->request, request); + return test_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags); } /** - * i915_gem_active_set_retire_fn - updates the retirement callback - * @active - the active tracker - * @fn - the routine called when the request is retired - * @mutex - struct_mutex used to guard retirements - * - * i915_gem_active_set_retire_fn() updates the function pointer that - * is called when the final request associated with the @active tracker - * is retired. + * Returns true if seq1 is later than seq2. */ -static inline void -i915_gem_active_set_retire_fn(struct i915_gem_active *active, - i915_gem_retire_fn fn, - struct mutex *mutex) +static inline bool i915_seqno_passed(u32 seq1, u32 seq2) { - lockdep_assert_held(mutex); - active->retire = fn ?: i915_gem_retire_noop; + return (s32)(seq1 - seq2) >= 0; } -static inline struct i915_request * -__i915_gem_active_peek(const struct i915_gem_active *active) +static inline u32 __hwsp_seqno(const struct i915_request *rq) { - /* - * Inside the error capture (running with the driver in an unknown - * state), we want to bend the rules slightly (a lot). - * - * Work is in progress to make it safer, in the meantime this keeps - * the known issue from spamming the logs. - */ - return rcu_dereference_protected(active->request, 1); + return READ_ONCE(*rq->hwsp_seqno); } /** - * i915_gem_active_raw - return the active request - * @active - the active tracker + * hwsp_seqno - the current breadcrumb value in the HW status page + * @rq: the request, to chase the relevant HW status page * - * i915_gem_active_raw() returns the current request being tracked, or NULL. - * It does not obtain a reference on the request for the caller, so the caller - * must hold struct_mutex. - */ -static inline struct i915_request * -i915_gem_active_raw(const struct i915_gem_active *active, struct mutex *mutex) -{ - return rcu_dereference_protected(active->request, - lockdep_is_held(mutex)); -} - -/** - * i915_gem_active_peek - report the active request being monitored - * @active - the active tracker + * The emphasis in naming here is that hwsp_seqno() is not a property of the + * request, but an indication of the current HW state (associated with this + * request). Its value will change as the GPU executes more requests. * - * i915_gem_active_peek() returns the current request being tracked if - * still active, or NULL. It does not obtain a reference on the request - * for the caller, so the caller must hold struct_mutex. + * Returns the current breadcrumb value in the associated HW status page (or + * the local timeline's equivalent) for this request. The request itself + * has the associated breadcrumb value of rq->fence.seqno, when the HW + * status page has that breadcrumb or later, this request is complete. */ -static inline struct i915_request * -i915_gem_active_peek(const struct i915_gem_active *active, struct mutex *mutex) +static inline u32 hwsp_seqno(const struct i915_request *rq) { - struct i915_request *request; + u32 seqno; - request = i915_gem_active_raw(active, mutex); - if (!request || i915_request_completed(request)) - return NULL; + rcu_read_lock(); /* the HWSP may be freed at runtime */ + seqno = __hwsp_seqno(rq); + rcu_read_unlock(); - return request; + return seqno; } -/** - * i915_gem_active_get - return a reference to the active request - * @active - the active tracker - * - * i915_gem_active_get() returns a reference to the active request, or NULL - * if the active tracker is idle. The caller must hold struct_mutex. - */ -static inline struct i915_request * -i915_gem_active_get(const struct i915_gem_active *active, struct mutex *mutex) +static inline bool __i915_request_has_started(const struct i915_request *rq) { - return i915_request_get(i915_gem_active_peek(active, mutex)); + return i915_seqno_passed(hwsp_seqno(rq), rq->fence.seqno - 1); } /** - * __i915_gem_active_get_rcu - return a reference to the active request - * @active - the active tracker - * - * __i915_gem_active_get() returns a reference to the active request, or NULL - * if the active tracker is idle. The caller must hold the RCU read lock, but - * the returned pointer is safe to use outside of RCU. - */ -static inline struct i915_request * -__i915_gem_active_get_rcu(const struct i915_gem_active *active) -{ - /* - * Performing a lockless retrieval of the active request is super - * tricky. SLAB_TYPESAFE_BY_RCU merely guarantees that the backing - * slab of request objects will not be freed whilst we hold the - * RCU read lock. It does not guarantee that the request itself - * will not be freed and then *reused*. Viz, - * - * Thread A Thread B - * - * rq = active.request - * retire(rq) -> free(rq); - * (rq is now first on the slab freelist) - * active.request = NULL - * - * rq = new submission on a new object - * ref(rq) - * - * To prevent the request from being reused whilst the caller - * uses it, we take a reference like normal. Whilst acquiring - * the reference we check that it is not in a destroyed state - * (refcnt == 0). That prevents the request being reallocated - * whilst the caller holds on to it. To check that the request - * was not reallocated as we acquired the reference we have to - * check that our request remains the active request across - * the lookup, in the same manner as a seqlock. The visibility - * of the pointer versus the reference counting is controlled - * by using RCU barriers (rcu_dereference and rcu_assign_pointer). - * - * In the middle of all that, we inspect whether the request is - * complete. Retiring is lazy so the request may be completed long - * before the active tracker is updated. Querying whether the - * request is complete is far cheaper (as it involves no locked - * instructions setting cachelines to exclusive) than acquiring - * the reference, so we do it first. The RCU read lock ensures the - * pointer dereference is valid, but does not ensure that the - * seqno nor HWS is the right one! However, if the request was - * reallocated, that means the active tracker's request was complete. - * If the new request is also complete, then both are and we can - * just report the active tracker is idle. If the new request is - * incomplete, then we acquire a reference on it and check that - * it remained the active request. - * - * It is then imperative that we do not zero the request on - * reallocation, so that we can chase the dangling pointers! - * See i915_request_alloc(). - */ - do { - struct i915_request *request; - - request = rcu_dereference(active->request); - if (!request || i915_request_completed(request)) - return NULL; - - /* - * An especially silly compiler could decide to recompute the - * result of i915_request_completed, more specifically - * re-emit the load for request->fence.seqno. A race would catch - * a later seqno value, which could flip the result from true to - * false. Which means part of the instructions below might not - * be executed, while later on instructions are executed. Due to - * barriers within the refcounting the inconsistency can't reach - * past the call to i915_request_get_rcu, but not executing - * that while still executing i915_request_put() creates - * havoc enough. Prevent this with a compiler barrier. - */ - barrier(); - - request = i915_request_get_rcu(request); - - /* - * What stops the following rcu_access_pointer() from occurring - * before the above i915_request_get_rcu()? If we were - * to read the value before pausing to get the reference to - * the request, we may not notice a change in the active - * tracker. - * - * The rcu_access_pointer() is a mere compiler barrier, which - * means both the CPU and compiler are free to perform the - * memory read without constraint. The compiler only has to - * ensure that any operations after the rcu_access_pointer() - * occur afterwards in program order. This means the read may - * be performed earlier by an out-of-order CPU, or adventurous - * compiler. - * - * The atomic operation at the heart of - * i915_request_get_rcu(), see dma_fence_get_rcu(), is - * atomic_inc_not_zero() which is only a full memory barrier - * when successful. That is, if i915_request_get_rcu() - * returns the request (and so with the reference counted - * incremented) then the following read for rcu_access_pointer() - * must occur after the atomic operation and so confirm - * that this request is the one currently being tracked. - * - * The corresponding write barrier is part of - * rcu_assign_pointer(). - */ - if (!request || request == rcu_access_pointer(active->request)) - return rcu_pointer_handoff(request); - - i915_request_put(request); - } while (1); -} - -/** - * i915_gem_active_get_unlocked - return a reference to the active request - * @active - the active tracker - * - * i915_gem_active_get_unlocked() returns a reference to the active request, - * or NULL if the active tracker is idle. The reference is obtained under RCU, - * so no locking is required by the caller. + * i915_request_started - check if the request has begun being executed + * @rq: the request * - * The reference should be freed with i915_request_put(). + * Returns true if the request has been submitted to hardware, and the hardware + * has advanced passed the end of the previous request and so should be either + * currently processing the request (though it may be preempted and so + * not necessarily the next request to complete) or have completed the request. */ -static inline struct i915_request * -i915_gem_active_get_unlocked(const struct i915_gem_active *active) +static inline bool i915_request_started(const struct i915_request *rq) { - struct i915_request *request; + if (i915_request_signaled(rq)) + return true; - rcu_read_lock(); - request = __i915_gem_active_get_rcu(active); - rcu_read_unlock(); - - return request; + /* Remember: started but may have since been preempted! */ + return __i915_request_has_started(rq); } /** - * i915_gem_active_isset - report whether the active tracker is assigned - * @active - the active tracker + * i915_request_is_running - check if the request may actually be executing + * @rq: the request * - * i915_gem_active_isset() returns true if the active tracker is currently - * assigned to a request. Due to the lazy retiring, that request may be idle - * and this may report stale information. + * Returns true if the request is currently submitted to hardware, has passed + * its start point (i.e. the context is setup and not busywaiting). Note that + * it may no longer be running by the time the function returns! */ -static inline bool -i915_gem_active_isset(const struct i915_gem_active *active) +static inline bool i915_request_is_running(const struct i915_request *rq) { - return rcu_access_pointer(active->request); + if (!i915_request_is_active(rq)) + return false; + + return __i915_request_has_started(rq); } -/** - * i915_gem_active_wait - waits until the request is completed - * @active - the active request on which to wait - * @flags - how to wait - * @timeout - how long to wait at most - * @rps - userspace client to charge for a waitboost - * - * i915_gem_active_wait() waits until the request is completed before - * returning, without requiring any locks to be held. Note that it does not - * retire any requests before returning. - * - * This function relies on RCU in order to acquire the reference to the active - * request without holding any locks. See __i915_gem_active_get_rcu() for the - * glory details on how that is managed. Once the reference is acquired, we - * can then wait upon the request, and afterwards release our reference, - * free of any locking. - * - * This function wraps i915_request_wait(), see it for the full details on - * the arguments. - * - * Returns 0 if successful, or a negative error code. - */ -static inline int -i915_gem_active_wait(const struct i915_gem_active *active, unsigned int flags) +static inline bool i915_request_completed(const struct i915_request *rq) { - struct i915_request *request; - long ret = 0; - - request = i915_gem_active_get_unlocked(active); - if (request) { - ret = i915_request_wait(request, flags, MAX_SCHEDULE_TIMEOUT); - i915_request_put(request); - } + if (i915_request_signaled(rq)) + return true; - return ret < 0 ? ret : 0; + return i915_seqno_passed(hwsp_seqno(rq), rq->fence.seqno); } -/** - * i915_gem_active_retire - waits until the request is retired - * @active - the active request on which to wait - * - * i915_gem_active_retire() waits until the request is completed, - * and then ensures that at least the retirement handler for this - * @active tracker is called before returning. If the @active - * tracker is idle, the function returns immediately. - */ -static inline int __must_check -i915_gem_active_retire(struct i915_gem_active *active, - struct mutex *mutex) +static inline void i915_request_mark_complete(struct i915_request *rq) { - struct i915_request *request; - long ret; - - request = i915_gem_active_raw(active, mutex); - if (!request) - return 0; - - ret = i915_request_wait(request, - I915_WAIT_INTERRUPTIBLE | I915_WAIT_LOCKED, - MAX_SCHEDULE_TIMEOUT); - if (ret < 0) - return ret; - - list_del_init(&active->link); - RCU_INIT_POINTER(active->request, NULL); - - active->retire(active, request); - - return 0; + rq->hwsp_seqno = (u32 *)&rq->fence.seqno; /* decouple from HWSP */ } -#define for_each_active(mask, idx) \ - for (; mask ? idx = ffs(mask) - 1, 1 : 0; mask &= ~BIT(idx)) +void i915_retire_requests(struct drm_i915_private *i915); #endif /* I915_REQUEST_H */ diff --git a/drivers/gpu/drm/i915/i915_reset.c b/drivers/gpu/drm/i915/i915_reset.c new file mode 100644 index 000000000000..0e0ddf2e6815 --- /dev/null +++ b/drivers/gpu/drm/i915/i915_reset.c @@ -0,0 +1,1349 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2008-2018 Intel Corporation + */ + +#include <linux/sched/mm.h> +#include <linux/stop_machine.h> + +#include "i915_drv.h" +#include "i915_gpu_error.h" +#include "i915_reset.h" + +#include "intel_guc.h" + +#define RESET_MAX_RETRIES 3 + +/* XXX How to handle concurrent GGTT updates using tiling registers? */ +#define RESET_UNDER_STOP_MACHINE 0 + +static void engine_skip_context(struct i915_request *rq) +{ + struct intel_engine_cs *engine = rq->engine; + struct i915_gem_context *hung_ctx = rq->gem_context; + struct i915_timeline *timeline = rq->timeline; + + lockdep_assert_held(&engine->timeline.lock); + GEM_BUG_ON(timeline == &engine->timeline); + + spin_lock(&timeline->lock); + + if (i915_request_is_active(rq)) { + list_for_each_entry_continue(rq, + &engine->timeline.requests, link) + if (rq->gem_context == hung_ctx) + i915_request_skip(rq, -EIO); + } + + list_for_each_entry(rq, &timeline->requests, link) + i915_request_skip(rq, -EIO); + + spin_unlock(&timeline->lock); +} + +static void client_mark_guilty(struct drm_i915_file_private *file_priv, + const struct i915_gem_context *ctx) +{ + unsigned int score; + unsigned long prev_hang; + + if (i915_gem_context_is_banned(ctx)) + score = I915_CLIENT_SCORE_CONTEXT_BAN; + else + score = 0; + + prev_hang = xchg(&file_priv->hang_timestamp, jiffies); + if (time_before(jiffies, prev_hang + I915_CLIENT_FAST_HANG_JIFFIES)) + score += I915_CLIENT_SCORE_HANG_FAST; + + if (score) { + atomic_add(score, &file_priv->ban_score); + + DRM_DEBUG_DRIVER("client %s: gained %u ban score, now %u\n", + ctx->name, score, + atomic_read(&file_priv->ban_score)); + } +} + +static bool context_mark_guilty(struct i915_gem_context *ctx) +{ + unsigned int score; + bool banned, bannable; + + atomic_inc(&ctx->guilty_count); + + bannable = i915_gem_context_is_bannable(ctx); + score = atomic_add_return(CONTEXT_SCORE_GUILTY, &ctx->ban_score); + banned = score >= CONTEXT_SCORE_BAN_THRESHOLD; + + /* Cool contexts don't accumulate client ban score */ + if (!bannable) + return false; + + if (banned) { + DRM_DEBUG_DRIVER("context %s: guilty %d, score %u, banned\n", + ctx->name, atomic_read(&ctx->guilty_count), + score); + i915_gem_context_set_banned(ctx); + } + + if (!IS_ERR_OR_NULL(ctx->file_priv)) + client_mark_guilty(ctx->file_priv, ctx); + + return banned; +} + +static void context_mark_innocent(struct i915_gem_context *ctx) +{ + atomic_inc(&ctx->active_count); +} + +void i915_reset_request(struct i915_request *rq, bool guilty) +{ + lockdep_assert_held(&rq->engine->timeline.lock); + GEM_BUG_ON(i915_request_completed(rq)); + + if (guilty) { + i915_request_skip(rq, -EIO); + if (context_mark_guilty(rq->gem_context)) + engine_skip_context(rq); + } else { + dma_fence_set_error(&rq->fence, -EAGAIN); + context_mark_innocent(rq->gem_context); + } +} + +static void gen3_stop_engine(struct intel_engine_cs *engine) +{ + struct drm_i915_private *dev_priv = engine->i915; + const u32 base = engine->mmio_base; + + if (intel_engine_stop_cs(engine)) + DRM_DEBUG_DRIVER("%s: timed out on STOP_RING\n", engine->name); + + I915_WRITE_FW(RING_HEAD(base), I915_READ_FW(RING_TAIL(base))); + POSTING_READ_FW(RING_HEAD(base)); /* paranoia */ + + I915_WRITE_FW(RING_HEAD(base), 0); + I915_WRITE_FW(RING_TAIL(base), 0); + POSTING_READ_FW(RING_TAIL(base)); + + /* The ring must be empty before it is disabled */ + I915_WRITE_FW(RING_CTL(base), 0); + + /* Check acts as a post */ + if (I915_READ_FW(RING_HEAD(base)) != 0) + DRM_DEBUG_DRIVER("%s: ring head not parked\n", + engine->name); +} + +static void i915_stop_engines(struct drm_i915_private *i915, + unsigned int engine_mask) +{ + struct intel_engine_cs *engine; + enum intel_engine_id id; + + if (INTEL_GEN(i915) < 3) + return; + + for_each_engine_masked(engine, i915, engine_mask, id) + gen3_stop_engine(engine); +} + +static bool i915_in_reset(struct pci_dev *pdev) +{ + u8 gdrst; + + pci_read_config_byte(pdev, I915_GDRST, &gdrst); + return gdrst & GRDOM_RESET_STATUS; +} + +static int i915_do_reset(struct drm_i915_private *i915, + unsigned int engine_mask, + unsigned int retry) +{ + struct pci_dev *pdev = i915->drm.pdev; + int err; + + /* Assert reset for at least 20 usec, and wait for acknowledgement. */ + pci_write_config_byte(pdev, I915_GDRST, GRDOM_RESET_ENABLE); + udelay(50); + err = wait_for_atomic(i915_in_reset(pdev), 50); + + /* Clear the reset request. */ + pci_write_config_byte(pdev, I915_GDRST, 0); + udelay(50); + if (!err) + err = wait_for_atomic(!i915_in_reset(pdev), 50); + + return err; +} + +static bool g4x_reset_complete(struct pci_dev *pdev) +{ + u8 gdrst; + + pci_read_config_byte(pdev, I915_GDRST, &gdrst); + return (gdrst & GRDOM_RESET_ENABLE) == 0; +} + +static int g33_do_reset(struct drm_i915_private *i915, + unsigned int engine_mask, + unsigned int retry) +{ + struct pci_dev *pdev = i915->drm.pdev; + + pci_write_config_byte(pdev, I915_GDRST, GRDOM_RESET_ENABLE); + return wait_for_atomic(g4x_reset_complete(pdev), 50); +} + +static int g4x_do_reset(struct drm_i915_private *dev_priv, + unsigned int engine_mask, + unsigned int retry) +{ + struct pci_dev *pdev = dev_priv->drm.pdev; + int ret; + + /* WaVcpClkGateDisableForMediaReset:ctg,elk */ + I915_WRITE_FW(VDECCLK_GATE_D, + I915_READ(VDECCLK_GATE_D) | VCP_UNIT_CLOCK_GATE_DISABLE); + POSTING_READ_FW(VDECCLK_GATE_D); + + pci_write_config_byte(pdev, I915_GDRST, + GRDOM_MEDIA | GRDOM_RESET_ENABLE); + ret = wait_for_atomic(g4x_reset_complete(pdev), 50); + if (ret) { + DRM_DEBUG_DRIVER("Wait for media reset failed\n"); + goto out; + } + + pci_write_config_byte(pdev, I915_GDRST, + GRDOM_RENDER | GRDOM_RESET_ENABLE); + ret = wait_for_atomic(g4x_reset_complete(pdev), 50); + if (ret) { + DRM_DEBUG_DRIVER("Wait for render reset failed\n"); + goto out; + } + +out: + pci_write_config_byte(pdev, I915_GDRST, 0); + + I915_WRITE_FW(VDECCLK_GATE_D, + I915_READ(VDECCLK_GATE_D) & ~VCP_UNIT_CLOCK_GATE_DISABLE); + POSTING_READ_FW(VDECCLK_GATE_D); + + return ret; +} + +static int ironlake_do_reset(struct drm_i915_private *dev_priv, + unsigned int engine_mask, + unsigned int retry) +{ + int ret; + + I915_WRITE_FW(ILK_GDSR, ILK_GRDOM_RENDER | ILK_GRDOM_RESET_ENABLE); + ret = __intel_wait_for_register_fw(dev_priv, ILK_GDSR, + ILK_GRDOM_RESET_ENABLE, 0, + 5000, 0, + NULL); + if (ret) { + DRM_DEBUG_DRIVER("Wait for render reset failed\n"); + goto out; + } + + I915_WRITE_FW(ILK_GDSR, ILK_GRDOM_MEDIA | ILK_GRDOM_RESET_ENABLE); + ret = __intel_wait_for_register_fw(dev_priv, ILK_GDSR, + ILK_GRDOM_RESET_ENABLE, 0, + 5000, 0, + NULL); + if (ret) { + DRM_DEBUG_DRIVER("Wait for media reset failed\n"); + goto out; + } + +out: + I915_WRITE_FW(ILK_GDSR, 0); + POSTING_READ_FW(ILK_GDSR); + return ret; +} + +/* Reset the hardware domains (GENX_GRDOM_*) specified by mask */ +static int gen6_hw_domain_reset(struct drm_i915_private *dev_priv, + u32 hw_domain_mask) +{ + int err; + + /* + * GEN6_GDRST is not in the gt power well, no need to check + * for fifo space for the write or forcewake the chip for + * the read + */ + I915_WRITE_FW(GEN6_GDRST, hw_domain_mask); + + /* Wait for the device to ack the reset requests */ + err = __intel_wait_for_register_fw(dev_priv, + GEN6_GDRST, hw_domain_mask, 0, + 500, 0, + NULL); + if (err) + DRM_DEBUG_DRIVER("Wait for 0x%08x engines reset failed\n", + hw_domain_mask); + + return err; +} + +static int gen6_reset_engines(struct drm_i915_private *i915, + unsigned int engine_mask, + unsigned int retry) +{ + struct intel_engine_cs *engine; + const u32 hw_engine_mask[I915_NUM_ENGINES] = { + [RCS] = GEN6_GRDOM_RENDER, + [BCS] = GEN6_GRDOM_BLT, + [VCS] = GEN6_GRDOM_MEDIA, + [VCS2] = GEN8_GRDOM_MEDIA2, + [VECS] = GEN6_GRDOM_VECS, + }; + u32 hw_mask; + + if (engine_mask == ALL_ENGINES) { + hw_mask = GEN6_GRDOM_FULL; + } else { + unsigned int tmp; + + hw_mask = 0; + for_each_engine_masked(engine, i915, engine_mask, tmp) + hw_mask |= hw_engine_mask[engine->id]; + } + + return gen6_hw_domain_reset(i915, hw_mask); +} + +static u32 gen11_lock_sfc(struct drm_i915_private *dev_priv, + struct intel_engine_cs *engine) +{ + u8 vdbox_sfc_access = RUNTIME_INFO(dev_priv)->vdbox_sfc_access; + i915_reg_t sfc_forced_lock, sfc_forced_lock_ack; + u32 sfc_forced_lock_bit, sfc_forced_lock_ack_bit; + i915_reg_t sfc_usage; + u32 sfc_usage_bit; + u32 sfc_reset_bit; + + switch (engine->class) { + case VIDEO_DECODE_CLASS: + if ((BIT(engine->instance) & vdbox_sfc_access) == 0) + return 0; + + sfc_forced_lock = GEN11_VCS_SFC_FORCED_LOCK(engine); + sfc_forced_lock_bit = GEN11_VCS_SFC_FORCED_LOCK_BIT; + + sfc_forced_lock_ack = GEN11_VCS_SFC_LOCK_STATUS(engine); + sfc_forced_lock_ack_bit = GEN11_VCS_SFC_LOCK_ACK_BIT; + + sfc_usage = GEN11_VCS_SFC_LOCK_STATUS(engine); + sfc_usage_bit = GEN11_VCS_SFC_USAGE_BIT; + sfc_reset_bit = GEN11_VCS_SFC_RESET_BIT(engine->instance); + break; + + case VIDEO_ENHANCEMENT_CLASS: + sfc_forced_lock = GEN11_VECS_SFC_FORCED_LOCK(engine); + sfc_forced_lock_bit = GEN11_VECS_SFC_FORCED_LOCK_BIT; + + sfc_forced_lock_ack = GEN11_VECS_SFC_LOCK_ACK(engine); + sfc_forced_lock_ack_bit = GEN11_VECS_SFC_LOCK_ACK_BIT; + + sfc_usage = GEN11_VECS_SFC_USAGE(engine); + sfc_usage_bit = GEN11_VECS_SFC_USAGE_BIT; + sfc_reset_bit = GEN11_VECS_SFC_RESET_BIT(engine->instance); + break; + + default: + return 0; + } + + /* + * Tell the engine that a software reset is going to happen. The engine + * will then try to force lock the SFC (if currently locked, it will + * remain so until we tell the engine it is safe to unlock; if currently + * unlocked, it will ignore this and all new lock requests). If SFC + * ends up being locked to the engine we want to reset, we have to reset + * it as well (we will unlock it once the reset sequence is completed). + */ + I915_WRITE_FW(sfc_forced_lock, + I915_READ_FW(sfc_forced_lock) | sfc_forced_lock_bit); + + if (__intel_wait_for_register_fw(dev_priv, + sfc_forced_lock_ack, + sfc_forced_lock_ack_bit, + sfc_forced_lock_ack_bit, + 1000, 0, NULL)) { + DRM_DEBUG_DRIVER("Wait for SFC forced lock ack failed\n"); + return 0; + } + + if (I915_READ_FW(sfc_usage) & sfc_usage_bit) + return sfc_reset_bit; + + return 0; +} + +static void gen11_unlock_sfc(struct drm_i915_private *dev_priv, + struct intel_engine_cs *engine) +{ + u8 vdbox_sfc_access = RUNTIME_INFO(dev_priv)->vdbox_sfc_access; + i915_reg_t sfc_forced_lock; + u32 sfc_forced_lock_bit; + + switch (engine->class) { + case VIDEO_DECODE_CLASS: + if ((BIT(engine->instance) & vdbox_sfc_access) == 0) + return; + + sfc_forced_lock = GEN11_VCS_SFC_FORCED_LOCK(engine); + sfc_forced_lock_bit = GEN11_VCS_SFC_FORCED_LOCK_BIT; + break; + + case VIDEO_ENHANCEMENT_CLASS: + sfc_forced_lock = GEN11_VECS_SFC_FORCED_LOCK(engine); + sfc_forced_lock_bit = GEN11_VECS_SFC_FORCED_LOCK_BIT; + break; + + default: + return; + } + + I915_WRITE_FW(sfc_forced_lock, + I915_READ_FW(sfc_forced_lock) & ~sfc_forced_lock_bit); +} + +static int gen11_reset_engines(struct drm_i915_private *i915, + unsigned int engine_mask, + unsigned int retry) +{ + const u32 hw_engine_mask[I915_NUM_ENGINES] = { + [RCS] = GEN11_GRDOM_RENDER, + [BCS] = GEN11_GRDOM_BLT, + [VCS] = GEN11_GRDOM_MEDIA, + [VCS2] = GEN11_GRDOM_MEDIA2, + [VCS3] = GEN11_GRDOM_MEDIA3, + [VCS4] = GEN11_GRDOM_MEDIA4, + [VECS] = GEN11_GRDOM_VECS, + [VECS2] = GEN11_GRDOM_VECS2, + }; + struct intel_engine_cs *engine; + unsigned int tmp; + u32 hw_mask; + int ret; + + BUILD_BUG_ON(VECS2 + 1 != I915_NUM_ENGINES); + + if (engine_mask == ALL_ENGINES) { + hw_mask = GEN11_GRDOM_FULL; + } else { + hw_mask = 0; + for_each_engine_masked(engine, i915, engine_mask, tmp) { + hw_mask |= hw_engine_mask[engine->id]; + hw_mask |= gen11_lock_sfc(i915, engine); + } + } + + ret = gen6_hw_domain_reset(i915, hw_mask); + + if (engine_mask != ALL_ENGINES) + for_each_engine_masked(engine, i915, engine_mask, tmp) + gen11_unlock_sfc(i915, engine); + + return ret; +} + +static int gen8_engine_reset_prepare(struct intel_engine_cs *engine) +{ + struct drm_i915_private *dev_priv = engine->i915; + int ret; + + I915_WRITE_FW(RING_RESET_CTL(engine->mmio_base), + _MASKED_BIT_ENABLE(RESET_CTL_REQUEST_RESET)); + + ret = __intel_wait_for_register_fw(dev_priv, + RING_RESET_CTL(engine->mmio_base), + RESET_CTL_READY_TO_RESET, + RESET_CTL_READY_TO_RESET, + 700, 0, + NULL); + if (ret) + DRM_ERROR("%s: reset request timeout\n", engine->name); + + return ret; +} + +static void gen8_engine_reset_cancel(struct intel_engine_cs *engine) +{ + struct drm_i915_private *dev_priv = engine->i915; + + I915_WRITE_FW(RING_RESET_CTL(engine->mmio_base), + _MASKED_BIT_DISABLE(RESET_CTL_REQUEST_RESET)); +} + +static int gen8_reset_engines(struct drm_i915_private *i915, + unsigned int engine_mask, + unsigned int retry) +{ + struct intel_engine_cs *engine; + const bool reset_non_ready = retry >= 1; + unsigned int tmp; + int ret; + + for_each_engine_masked(engine, i915, engine_mask, tmp) { + ret = gen8_engine_reset_prepare(engine); + if (ret && !reset_non_ready) + goto skip_reset; + + /* + * If this is not the first failed attempt to prepare, + * we decide to proceed anyway. + * + * By doing so we risk context corruption and with + * some gens (kbl), possible system hang if reset + * happens during active bb execution. + * + * We rather take context corruption instead of + * failed reset with a wedged driver/gpu. And + * active bb execution case should be covered by + * i915_stop_engines we have before the reset. + */ + } + + if (INTEL_GEN(i915) >= 11) + ret = gen11_reset_engines(i915, engine_mask, retry); + else + ret = gen6_reset_engines(i915, engine_mask, retry); + +skip_reset: + for_each_engine_masked(engine, i915, engine_mask, tmp) + gen8_engine_reset_cancel(engine); + + return ret; +} + +typedef int (*reset_func)(struct drm_i915_private *, + unsigned int engine_mask, + unsigned int retry); + +static reset_func intel_get_gpu_reset(struct drm_i915_private *i915) +{ + if (!i915_modparams.reset) + return NULL; + + if (INTEL_GEN(i915) >= 8) + return gen8_reset_engines; + else if (INTEL_GEN(i915) >= 6) + return gen6_reset_engines; + else if (INTEL_GEN(i915) >= 5) + return ironlake_do_reset; + else if (IS_G4X(i915)) + return g4x_do_reset; + else if (IS_G33(i915) || IS_PINEVIEW(i915)) + return g33_do_reset; + else if (INTEL_GEN(i915) >= 3) + return i915_do_reset; + else + return NULL; +} + +int intel_gpu_reset(struct drm_i915_private *i915, unsigned int engine_mask) +{ + const int retries = engine_mask == ALL_ENGINES ? RESET_MAX_RETRIES : 1; + reset_func reset; + int ret = -ETIMEDOUT; + int retry; + + reset = intel_get_gpu_reset(i915); + if (!reset) + return -ENODEV; + + /* + * If the power well sleeps during the reset, the reset + * request may be dropped and never completes (causing -EIO). + */ + intel_uncore_forcewake_get(i915, FORCEWAKE_ALL); + for (retry = 0; ret == -ETIMEDOUT && retry < retries; retry++) { + /* + * We stop engines, otherwise we might get failed reset and a + * dead gpu (on elk). Also as modern gpu as kbl can suffer + * from system hang if batchbuffer is progressing when + * the reset is issued, regardless of READY_TO_RESET ack. + * Thus assume it is best to stop engines on all gens + * where we have a gpu reset. + * + * WaKBLVECSSemaphoreWaitPoll:kbl (on ALL_ENGINES) + * + * WaMediaResetMainRingCleanup:ctg,elk (presumably) + * + * FIXME: Wa for more modern gens needs to be validated + */ + i915_stop_engines(i915, engine_mask); + + GEM_TRACE("engine_mask=%x\n", engine_mask); + preempt_disable(); + ret = reset(i915, engine_mask, retry); + preempt_enable(); + } + intel_uncore_forcewake_put(i915, FORCEWAKE_ALL); + + return ret; +} + +bool intel_has_gpu_reset(struct drm_i915_private *i915) +{ + if (USES_GUC(i915)) + return false; + + return intel_get_gpu_reset(i915); +} + +bool intel_has_reset_engine(struct drm_i915_private *i915) +{ + return INTEL_INFO(i915)->has_reset_engine && i915_modparams.reset >= 2; +} + +int intel_reset_guc(struct drm_i915_private *i915) +{ + u32 guc_domain = + INTEL_GEN(i915) >= 11 ? GEN11_GRDOM_GUC : GEN9_GRDOM_GUC; + int ret; + + GEM_BUG_ON(!HAS_GUC(i915)); + + intel_uncore_forcewake_get(i915, FORCEWAKE_ALL); + ret = gen6_hw_domain_reset(i915, guc_domain); + intel_uncore_forcewake_put(i915, FORCEWAKE_ALL); + + return ret; +} + +/* + * Ensure irq handler finishes, and not run again. + * Also return the active request so that we only search for it once. + */ +static void reset_prepare_engine(struct intel_engine_cs *engine) +{ + /* + * During the reset sequence, we must prevent the engine from + * entering RC6. As the context state is undefined until we restart + * the engine, if it does enter RC6 during the reset, the state + * written to the powercontext is undefined and so we may lose + * GPU state upon resume, i.e. fail to restart after a reset. + */ + intel_uncore_forcewake_get(engine->i915, FORCEWAKE_ALL); + engine->reset.prepare(engine); +} + +static void reset_prepare(struct drm_i915_private *i915) +{ + struct intel_engine_cs *engine; + enum intel_engine_id id; + + for_each_engine(engine, i915, id) + reset_prepare_engine(engine); + + intel_uc_sanitize(i915); +} + +static int gt_reset(struct drm_i915_private *i915, unsigned int stalled_mask) +{ + struct intel_engine_cs *engine; + enum intel_engine_id id; + int err; + + /* + * Everything depends on having the GTT running, so we need to start + * there. + */ + err = i915_ggtt_enable_hw(i915); + if (err) + return err; + + for_each_engine(engine, i915, id) + intel_engine_reset(engine, stalled_mask & ENGINE_MASK(id)); + + i915_gem_restore_fences(i915); + + return err; +} + +static void reset_finish_engine(struct intel_engine_cs *engine) +{ + engine->reset.finish(engine); + intel_uncore_forcewake_put(engine->i915, FORCEWAKE_ALL); +} + +struct i915_gpu_restart { + struct work_struct work; + struct drm_i915_private *i915; +}; + +static void restart_work(struct work_struct *work) +{ + struct i915_gpu_restart *arg = container_of(work, typeof(*arg), work); + struct drm_i915_private *i915 = arg->i915; + struct intel_engine_cs *engine; + enum intel_engine_id id; + intel_wakeref_t wakeref; + + wakeref = intel_runtime_pm_get(i915); + mutex_lock(&i915->drm.struct_mutex); + WRITE_ONCE(i915->gpu_error.restart, NULL); + + for_each_engine(engine, i915, id) { + struct i915_request *rq; + + /* + * Ostensibily, we always want a context loaded for powersaving, + * so if the engine is idle after the reset, send a request + * to load our scratch kernel_context. + */ + if (!intel_engine_is_idle(engine)) + continue; + + rq = i915_request_alloc(engine, i915->kernel_context); + if (!IS_ERR(rq)) + i915_request_add(rq); + } + + mutex_unlock(&i915->drm.struct_mutex); + intel_runtime_pm_put(i915, wakeref); + + kfree(arg); +} + +static void reset_finish(struct drm_i915_private *i915) +{ + struct intel_engine_cs *engine; + enum intel_engine_id id; + + for_each_engine(engine, i915, id) + reset_finish_engine(engine); +} + +static void reset_restart(struct drm_i915_private *i915) +{ + struct i915_gpu_restart *arg; + + /* + * Following the reset, ensure that we always reload context for + * powersaving, and to correct engine->last_retired_context. Since + * this requires us to submit a request, queue a worker to do that + * task for us to evade any locking here. + */ + if (READ_ONCE(i915->gpu_error.restart)) + return; + + arg = kmalloc(sizeof(*arg), GFP_KERNEL); + if (arg) { + arg->i915 = i915; + INIT_WORK(&arg->work, restart_work); + + WRITE_ONCE(i915->gpu_error.restart, arg); + queue_work(i915->wq, &arg->work); + } +} + +static void nop_submit_request(struct i915_request *request) +{ + struct intel_engine_cs *engine = request->engine; + unsigned long flags; + + GEM_TRACE("%s fence %llx:%lld -> -EIO\n", + engine->name, request->fence.context, request->fence.seqno); + dma_fence_set_error(&request->fence, -EIO); + + spin_lock_irqsave(&engine->timeline.lock, flags); + __i915_request_submit(request); + i915_request_mark_complete(request); + intel_engine_write_global_seqno(engine, request->global_seqno); + spin_unlock_irqrestore(&engine->timeline.lock, flags); + + intel_engine_queue_breadcrumbs(engine); +} + +void i915_gem_set_wedged(struct drm_i915_private *i915) +{ + struct i915_gpu_error *error = &i915->gpu_error; + struct intel_engine_cs *engine; + enum intel_engine_id id; + + mutex_lock(&error->wedge_mutex); + if (test_bit(I915_WEDGED, &error->flags)) { + mutex_unlock(&error->wedge_mutex); + return; + } + + if (GEM_SHOW_DEBUG() && !intel_engines_are_idle(i915)) { + struct drm_printer p = drm_debug_printer(__func__); + + for_each_engine(engine, i915, id) + intel_engine_dump(engine, &p, "%s\n", engine->name); + } + + GEM_TRACE("start\n"); + + /* + * First, stop submission to hw, but do not yet complete requests by + * rolling the global seqno forward (since this would complete requests + * for which we haven't set the fence error to EIO yet). + */ + for_each_engine(engine, i915, id) + reset_prepare_engine(engine); + + /* Even if the GPU reset fails, it should still stop the engines */ + if (INTEL_GEN(i915) >= 5) + intel_gpu_reset(i915, ALL_ENGINES); + + for_each_engine(engine, i915, id) { + engine->submit_request = nop_submit_request; + engine->schedule = NULL; + } + i915->caps.scheduler = 0; + + /* + * Make sure no request can slip through without getting completed by + * either this call here to intel_engine_write_global_seqno, or the one + * in nop_submit_request. + */ + synchronize_rcu(); + + /* Mark all executing requests as skipped */ + for_each_engine(engine, i915, id) + engine->cancel_requests(engine); + + for_each_engine(engine, i915, id) { + reset_finish_engine(engine); + intel_engine_signal_breadcrumbs(engine); + } + + smp_mb__before_atomic(); + set_bit(I915_WEDGED, &error->flags); + + GEM_TRACE("end\n"); + mutex_unlock(&error->wedge_mutex); + + wake_up_all(&error->reset_queue); +} + +bool i915_gem_unset_wedged(struct drm_i915_private *i915) +{ + struct i915_gpu_error *error = &i915->gpu_error; + struct i915_timeline *tl; + bool ret = false; + + if (!test_bit(I915_WEDGED, &error->flags)) + return true; + + if (!i915->gt.scratch) /* Never full initialised, recovery impossible */ + return false; + + mutex_lock(&error->wedge_mutex); + + GEM_TRACE("start\n"); + + /* + * Before unwedging, make sure that all pending operations + * are flushed and errored out - we may have requests waiting upon + * third party fences. We marked all inflight requests as EIO, and + * every execbuf since returned EIO, for consistency we want all + * the currently pending requests to also be marked as EIO, which + * is done inside our nop_submit_request - and so we must wait. + * + * No more can be submitted until we reset the wedged bit. + */ + mutex_lock(&i915->gt.timelines.mutex); + list_for_each_entry(tl, &i915->gt.timelines.active_list, link) { + struct i915_request *rq; + long timeout; + + rq = i915_active_request_get_unlocked(&tl->last_request); + if (!rq) + continue; + + /* + * We can't use our normal waiter as we want to + * avoid recursively trying to handle the current + * reset. The basic dma_fence_default_wait() installs + * a callback for dma_fence_signal(), which is + * triggered by our nop handler (indirectly, the + * callback enables the signaler thread which is + * woken by the nop_submit_request() advancing the seqno + * and when the seqno passes the fence, the signaler + * then signals the fence waking us up). + */ + timeout = dma_fence_default_wait(&rq->fence, true, + MAX_SCHEDULE_TIMEOUT); + i915_request_put(rq); + if (timeout < 0) { + mutex_unlock(&i915->gt.timelines.mutex); + goto unlock; + } + } + mutex_unlock(&i915->gt.timelines.mutex); + + intel_engines_sanitize(i915, false); + + /* + * Undo nop_submit_request. We prevent all new i915 requests from + * being queued (by disallowing execbuf whilst wedged) so having + * waited for all active requests above, we know the system is idle + * and do not have to worry about a thread being inside + * engine->submit_request() as we swap over. So unlike installing + * the nop_submit_request on reset, we can do this from normal + * context and do not require stop_machine(). + */ + intel_engines_reset_default_submission(i915); + + GEM_TRACE("end\n"); + + smp_mb__before_atomic(); /* complete takeover before enabling execbuf */ + clear_bit(I915_WEDGED, &i915->gpu_error.flags); + ret = true; +unlock: + mutex_unlock(&i915->gpu_error.wedge_mutex); + + return ret; +} + +struct __i915_reset { + struct drm_i915_private *i915; + unsigned int stalled_mask; +}; + +static int __i915_reset__BKL(void *data) +{ + struct __i915_reset *arg = data; + int err; + + err = intel_gpu_reset(arg->i915, ALL_ENGINES); + if (err) + return err; + + return gt_reset(arg->i915, arg->stalled_mask); +} + +#if RESET_UNDER_STOP_MACHINE +/* + * XXX An alternative to using stop_machine would be to park only the + * processes that have a GGTT mmap. By remote parking the threads (SIGSTOP) + * we should be able to prevent their memmory accesses via the lost fence + * registers over the course of the reset without the potential recursive + * of mutexes between the pagefault handler and reset. + * + * See igt/gem_mmap_gtt/hang + */ +#define __do_reset(fn, arg) stop_machine(fn, arg, NULL) +#else +#define __do_reset(fn, arg) fn(arg) +#endif + +static int do_reset(struct drm_i915_private *i915, unsigned int stalled_mask) +{ + struct __i915_reset arg = { i915, stalled_mask }; + int err, i; + + err = __do_reset(__i915_reset__BKL, &arg); + for (i = 0; err && i < RESET_MAX_RETRIES; i++) { + msleep(100); + err = __do_reset(__i915_reset__BKL, &arg); + } + + return err; +} + +/** + * i915_reset - reset chip after a hang + * @i915: #drm_i915_private to reset + * @stalled_mask: mask of the stalled engines with the guilty requests + * @reason: user error message for why we are resetting + * + * Reset the chip. Useful if a hang is detected. Marks the device as wedged + * on failure. + * + * Caller must hold the struct_mutex. + * + * Procedure is fairly simple: + * - reset the chip using the reset reg + * - re-init context state + * - re-init hardware status page + * - re-init ring buffer + * - re-init interrupt state + * - re-init display + */ +void i915_reset(struct drm_i915_private *i915, + unsigned int stalled_mask, + const char *reason) +{ + struct i915_gpu_error *error = &i915->gpu_error; + int ret; + + GEM_TRACE("flags=%lx\n", error->flags); + + might_sleep(); + assert_rpm_wakelock_held(i915); + GEM_BUG_ON(!test_bit(I915_RESET_BACKOFF, &error->flags)); + + /* Clear any previous failed attempts at recovery. Time to try again. */ + if (!i915_gem_unset_wedged(i915)) + return; + + if (reason) + dev_notice(i915->drm.dev, "Resetting chip for %s\n", reason); + error->reset_count++; + + reset_prepare(i915); + + if (!intel_has_gpu_reset(i915)) { + if (i915_modparams.reset) + dev_err(i915->drm.dev, "GPU reset not supported\n"); + else + DRM_DEBUG_DRIVER("GPU reset disabled\n"); + goto error; + } + + if (do_reset(i915, stalled_mask)) { + dev_err(i915->drm.dev, "Failed to reset chip\n"); + goto taint; + } + + intel_overlay_reset(i915); + + /* + * Next we need to restore the context, but we don't use those + * yet either... + * + * Ring buffer needs to be re-initialized in the KMS case, or if X + * was running at the time of the reset (i.e. we weren't VT + * switched away). + */ + ret = i915_gem_init_hw(i915); + if (ret) { + DRM_ERROR("Failed to initialise HW following reset (%d)\n", + ret); + goto error; + } + + i915_queue_hangcheck(i915); + +finish: + reset_finish(i915); + if (!i915_terminally_wedged(error)) + reset_restart(i915); + return; + +taint: + /* + * History tells us that if we cannot reset the GPU now, we + * never will. This then impacts everything that is run + * subsequently. On failing the reset, we mark the driver + * as wedged, preventing further execution on the GPU. + * We also want to go one step further and add a taint to the + * kernel so that any subsequent faults can be traced back to + * this failure. This is important for CI, where if the + * GPU/driver fails we would like to reboot and restart testing + * rather than continue on into oblivion. For everyone else, + * the system should still plod along, but they have been warned! + */ + add_taint(TAINT_WARN, LOCKDEP_STILL_OK); +error: + i915_gem_set_wedged(i915); + goto finish; +} + +static inline int intel_gt_reset_engine(struct drm_i915_private *i915, + struct intel_engine_cs *engine) +{ + return intel_gpu_reset(i915, intel_engine_flag(engine)); +} + +/** + * i915_reset_engine - reset GPU engine to recover from a hang + * @engine: engine to reset + * @msg: reason for GPU reset; or NULL for no dev_notice() + * + * Reset a specific GPU engine. Useful if a hang is detected. + * Returns zero on successful reset or otherwise an error code. + * + * Procedure is: + * - identifies the request that caused the hang and it is dropped + * - reset engine (which will force the engine to idle) + * - re-init/configure engine + */ +int i915_reset_engine(struct intel_engine_cs *engine, const char *msg) +{ + struct i915_gpu_error *error = &engine->i915->gpu_error; + int ret; + + GEM_TRACE("%s flags=%lx\n", engine->name, error->flags); + GEM_BUG_ON(!test_bit(I915_RESET_ENGINE + engine->id, &error->flags)); + + reset_prepare_engine(engine); + + if (msg) + dev_notice(engine->i915->drm.dev, + "Resetting %s for %s\n", engine->name, msg); + error->reset_engine_count[engine->id]++; + + if (!engine->i915->guc.execbuf_client) + ret = intel_gt_reset_engine(engine->i915, engine); + else + ret = intel_guc_reset_engine(&engine->i915->guc, engine); + if (ret) { + /* If we fail here, we expect to fallback to a global reset */ + DRM_DEBUG_DRIVER("%sFailed to reset %s, ret=%d\n", + engine->i915->guc.execbuf_client ? "GuC " : "", + engine->name, ret); + goto out; + } + + /* + * The request that caused the hang is stuck on elsp, we know the + * active request and can drop it, adjust head to skip the offending + * request to resume executing remaining requests in the queue. + */ + intel_engine_reset(engine, true); + + /* + * The engine and its registers (and workarounds in case of render) + * have been reset to their default values. Follow the init_ring + * process to program RING_MODE, HWSP and re-enable submission. + */ + ret = engine->init_hw(engine); + if (ret) + goto out; + +out: + intel_engine_cancel_stop_cs(engine); + reset_finish_engine(engine); + return ret; +} + +static void i915_reset_device(struct drm_i915_private *i915, + u32 engine_mask, + const char *reason) +{ + struct i915_gpu_error *error = &i915->gpu_error; + struct kobject *kobj = &i915->drm.primary->kdev->kobj; + char *error_event[] = { I915_ERROR_UEVENT "=1", NULL }; + char *reset_event[] = { I915_RESET_UEVENT "=1", NULL }; + char *reset_done_event[] = { I915_ERROR_UEVENT "=0", NULL }; + struct i915_wedge_me w; + + kobject_uevent_env(kobj, KOBJ_CHANGE, error_event); + + DRM_DEBUG_DRIVER("resetting chip\n"); + kobject_uevent_env(kobj, KOBJ_CHANGE, reset_event); + + /* Use a watchdog to ensure that our reset completes */ + i915_wedge_on_timeout(&w, i915, 5 * HZ) { + intel_prepare_reset(i915); + + i915_reset(i915, engine_mask, reason); + + intel_finish_reset(i915); + } + + if (!test_bit(I915_WEDGED, &error->flags)) + kobject_uevent_env(kobj, KOBJ_CHANGE, reset_done_event); +} + +void i915_clear_error_registers(struct drm_i915_private *dev_priv) +{ + u32 eir; + + if (!IS_GEN(dev_priv, 2)) + I915_WRITE(PGTBL_ER, I915_READ(PGTBL_ER)); + + if (INTEL_GEN(dev_priv) < 4) + I915_WRITE(IPEIR, I915_READ(IPEIR)); + else + I915_WRITE(IPEIR_I965, I915_READ(IPEIR_I965)); + + I915_WRITE(EIR, I915_READ(EIR)); + eir = I915_READ(EIR); + if (eir) { + /* + * some errors might have become stuck, + * mask them. + */ + DRM_DEBUG_DRIVER("EIR stuck: 0x%08x, masking\n", eir); + I915_WRITE(EMR, I915_READ(EMR) | eir); + I915_WRITE(IIR, I915_MASTER_ERROR_INTERRUPT); + } + + if (INTEL_GEN(dev_priv) >= 8) { + I915_WRITE(GEN8_RING_FAULT_REG, + I915_READ(GEN8_RING_FAULT_REG) & ~RING_FAULT_VALID); + POSTING_READ(GEN8_RING_FAULT_REG); + } else if (INTEL_GEN(dev_priv) >= 6) { + struct intel_engine_cs *engine; + enum intel_engine_id id; + + for_each_engine(engine, dev_priv, id) { + I915_WRITE(RING_FAULT_REG(engine), + I915_READ(RING_FAULT_REG(engine)) & + ~RING_FAULT_VALID); + } + POSTING_READ(RING_FAULT_REG(dev_priv->engine[RCS])); + } +} + +/** + * i915_handle_error - handle a gpu error + * @i915: i915 device private + * @engine_mask: mask representing engines that are hung + * @flags: control flags + * @fmt: Error message format string + * + * Do some basic checking of register state at error time and + * dump it to the syslog. Also call i915_capture_error_state() to make + * sure we get a record and make it available in debugfs. Fire a uevent + * so userspace knows something bad happened (should trigger collection + * of a ring dump etc.). + */ +void i915_handle_error(struct drm_i915_private *i915, + u32 engine_mask, + unsigned long flags, + const char *fmt, ...) +{ + struct intel_engine_cs *engine; + intel_wakeref_t wakeref; + unsigned int tmp; + char error_msg[80]; + char *msg = NULL; + + if (fmt) { + va_list args; + + va_start(args, fmt); + vscnprintf(error_msg, sizeof(error_msg), fmt, args); + va_end(args); + + msg = error_msg; + } + + /* + * In most cases it's guaranteed that we get here with an RPM + * reference held, for example because there is a pending GPU + * request that won't finish until the reset is done. This + * isn't the case at least when we get here by doing a + * simulated reset via debugfs, so get an RPM reference. + */ + wakeref = intel_runtime_pm_get(i915); + + engine_mask &= INTEL_INFO(i915)->ring_mask; + + if (flags & I915_ERROR_CAPTURE) { + i915_capture_error_state(i915, engine_mask, msg); + i915_clear_error_registers(i915); + } + + /* + * Try engine reset when available. We fall back to full reset if + * single reset fails. + */ + if (intel_has_reset_engine(i915) && + !i915_terminally_wedged(&i915->gpu_error)) { + for_each_engine_masked(engine, i915, engine_mask, tmp) { + BUILD_BUG_ON(I915_RESET_MODESET >= I915_RESET_ENGINE); + if (test_and_set_bit(I915_RESET_ENGINE + engine->id, + &i915->gpu_error.flags)) + continue; + + if (i915_reset_engine(engine, msg) == 0) + engine_mask &= ~intel_engine_flag(engine); + + clear_bit(I915_RESET_ENGINE + engine->id, + &i915->gpu_error.flags); + wake_up_bit(&i915->gpu_error.flags, + I915_RESET_ENGINE + engine->id); + } + } + + if (!engine_mask) + goto out; + + /* Full reset needs the mutex, stop any other user trying to do so. */ + if (test_and_set_bit(I915_RESET_BACKOFF, &i915->gpu_error.flags)) { + wait_event(i915->gpu_error.reset_queue, + !test_bit(I915_RESET_BACKOFF, + &i915->gpu_error.flags)); + goto out; + } + + /* Prevent any other reset-engine attempt. */ + for_each_engine(engine, i915, tmp) { + while (test_and_set_bit(I915_RESET_ENGINE + engine->id, + &i915->gpu_error.flags)) + wait_on_bit(&i915->gpu_error.flags, + I915_RESET_ENGINE + engine->id, + TASK_UNINTERRUPTIBLE); + } + + i915_reset_device(i915, engine_mask, msg); + + for_each_engine(engine, i915, tmp) { + clear_bit(I915_RESET_ENGINE + engine->id, + &i915->gpu_error.flags); + } + + clear_bit(I915_RESET_BACKOFF, &i915->gpu_error.flags); + wake_up_all(&i915->gpu_error.reset_queue); + +out: + intel_runtime_pm_put(i915, wakeref); +} + +bool i915_reset_flush(struct drm_i915_private *i915) +{ + int err; + + cancel_delayed_work_sync(&i915->gpu_error.hangcheck_work); + + flush_workqueue(i915->wq); + GEM_BUG_ON(READ_ONCE(i915->gpu_error.restart)); + + mutex_lock(&i915->drm.struct_mutex); + err = i915_gem_wait_for_idle(i915, + I915_WAIT_LOCKED | + I915_WAIT_FOR_IDLE_BOOST, + MAX_SCHEDULE_TIMEOUT); + mutex_unlock(&i915->drm.struct_mutex); + + return !err; +} + +static void i915_wedge_me(struct work_struct *work) +{ + struct i915_wedge_me *w = container_of(work, typeof(*w), work.work); + + dev_err(w->i915->drm.dev, + "%s timed out, cancelling all in-flight rendering.\n", + w->name); + i915_gem_set_wedged(w->i915); +} + +void __i915_init_wedge(struct i915_wedge_me *w, + struct drm_i915_private *i915, + long timeout, + const char *name) +{ + w->i915 = i915; + w->name = name; + + INIT_DELAYED_WORK_ONSTACK(&w->work, i915_wedge_me); + schedule_delayed_work(&w->work, timeout); +} + +void __i915_fini_wedge(struct i915_wedge_me *w) +{ + cancel_delayed_work_sync(&w->work); + destroy_delayed_work_on_stack(&w->work); + w->i915 = NULL; +} diff --git a/drivers/gpu/drm/i915/i915_reset.h b/drivers/gpu/drm/i915/i915_reset.h new file mode 100644 index 000000000000..f2d347f319df --- /dev/null +++ b/drivers/gpu/drm/i915/i915_reset.h @@ -0,0 +1,59 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2008-2018 Intel Corporation + */ + +#ifndef I915_RESET_H +#define I915_RESET_H + +#include <linux/compiler.h> +#include <linux/types.h> + +struct drm_i915_private; +struct intel_engine_cs; +struct intel_guc; + +__printf(4, 5) +void i915_handle_error(struct drm_i915_private *i915, + u32 engine_mask, + unsigned long flags, + const char *fmt, ...); +#define I915_ERROR_CAPTURE BIT(0) + +void i915_clear_error_registers(struct drm_i915_private *i915); + +void i915_reset(struct drm_i915_private *i915, + unsigned int stalled_mask, + const char *reason); +int i915_reset_engine(struct intel_engine_cs *engine, + const char *reason); + +void i915_reset_request(struct i915_request *rq, bool guilty); +bool i915_reset_flush(struct drm_i915_private *i915); + +bool intel_has_gpu_reset(struct drm_i915_private *i915); +bool intel_has_reset_engine(struct drm_i915_private *i915); + +int intel_gpu_reset(struct drm_i915_private *i915, u32 engine_mask); + +int intel_reset_guc(struct drm_i915_private *i915); + +struct i915_wedge_me { + struct delayed_work work; + struct drm_i915_private *i915; + const char *name; +}; + +void __i915_init_wedge(struct i915_wedge_me *w, + struct drm_i915_private *i915, + long timeout, + const char *name); +void __i915_fini_wedge(struct i915_wedge_me *w); + +#define i915_wedge_on_timeout(W, DEV, TIMEOUT) \ + for (__i915_init_wedge((W), (DEV), (TIMEOUT), __func__); \ + (W)->i915; \ + __i915_fini_wedge((W))) + +#endif /* I915_RESET_H */ diff --git a/drivers/gpu/drm/i915/i915_scheduler.c b/drivers/gpu/drm/i915/i915_scheduler.c index 340faea6c08a..d01683167c77 100644 --- a/drivers/gpu/drm/i915/i915_scheduler.c +++ b/drivers/gpu/drm/i915/i915_scheduler.c @@ -127,8 +127,7 @@ static inline struct i915_priolist *to_priolist(struct rb_node *rb) return rb_entry(rb, struct i915_priolist, node); } -static void assert_priolists(struct intel_engine_execlists * const execlists, - long queue_priority) +static void assert_priolists(struct intel_engine_execlists * const execlists) { struct rb_node *rb; long last_prio, i; @@ -139,7 +138,7 @@ static void assert_priolists(struct intel_engine_execlists * const execlists, GEM_BUG_ON(rb_first_cached(&execlists->queue) != rb_first(&execlists->queue.rb_root)); - last_prio = (queue_priority >> I915_USER_PRIORITY_SHIFT) + 1; + last_prio = (INT_MAX >> I915_USER_PRIORITY_SHIFT) + 1; for (rb = rb_first_cached(&execlists->queue); rb; rb = rb_next(rb)) { const struct i915_priolist *p = to_priolist(rb); @@ -166,7 +165,7 @@ i915_sched_lookup_priolist(struct intel_engine_cs *engine, int prio) int idx, i; lockdep_assert_held(&engine->timeline.lock); - assert_priolists(execlists, INT_MAX); + assert_priolists(execlists); /* buckets sorted from highest [in slot 0] to lowest priority */ idx = I915_PRIORITY_COUNT - (prio & I915_PRIORITY_MASK) - 1; @@ -239,6 +238,18 @@ sched_lock_engine(struct i915_sched_node *node, struct intel_engine_cs *locked) return engine; } +static bool inflight(const struct i915_request *rq, + const struct intel_engine_cs *engine) +{ + const struct i915_request *active; + + if (!i915_request_is_active(rq)) + return false; + + active = port_request(engine->execlists.port); + return active->hw_context == rq->hw_context; +} + static void __i915_schedule(struct i915_request *rq, const struct i915_sched_attr *attr) { @@ -328,6 +339,7 @@ static void __i915_schedule(struct i915_request *rq, INIT_LIST_HEAD(&dep->dfs_link); engine = sched_lock_engine(node, engine); + lockdep_assert_held(&engine->timeline.lock); /* Recheck after acquiring the engine->timeline.lock */ if (prio <= node->attr.priority || node_signaled(node)) @@ -353,20 +365,19 @@ static void __i915_schedule(struct i915_request *rq, continue; } - if (prio <= engine->execlists.queue_priority) + if (prio <= engine->execlists.queue_priority_hint) continue; + engine->execlists.queue_priority_hint = prio; + /* * If we are already the currently executing context, don't * bother evaluating if we should preempt ourselves. */ - if (node_to_request(node)->global_seqno && - i915_seqno_passed(port_request(engine->execlists.port)->global_seqno, - node_to_request(node)->global_seqno)) + if (inflight(node_to_request(node), engine)) continue; /* Defer (tasklet) submission until after all of our updates. */ - engine->execlists.queue_priority = prio; tasklet_hi_schedule(&engine->execlists.tasklet); } diff --git a/drivers/gpu/drm/i915/i915_selftest.h b/drivers/gpu/drm/i915/i915_selftest.h index a73472dd12fd..207e21b478f2 100644 --- a/drivers/gpu/drm/i915/i915_selftest.h +++ b/drivers/gpu/drm/i915/i915_selftest.h @@ -31,6 +31,7 @@ struct i915_selftest { unsigned long timeout_jiffies; unsigned int timeout_ms; unsigned int random_seed; + char *filter; int mock; int live; }; diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c index f18afa2bac8d..d2f2a9c2fabd 100644 --- a/drivers/gpu/drm/i915/i915_suspend.c +++ b/drivers/gpu/drm/i915/i915_suspend.c @@ -86,7 +86,7 @@ int i915_save_state(struct drm_i915_private *dev_priv) } else if (IS_GEN(dev_priv, 2)) { for (i = 0; i < 7; i++) dev_priv->regfile.saveSWF1[i] = I915_READ(SWF1(i)); - } else if (HAS_GMCH_DISPLAY(dev_priv)) { + } else if (HAS_GMCH(dev_priv)) { for (i = 0; i < 16; i++) { dev_priv->regfile.saveSWF0[i] = I915_READ(SWF0(i)); dev_priv->regfile.saveSWF1[i] = I915_READ(SWF1(i)); @@ -131,7 +131,7 @@ int i915_restore_state(struct drm_i915_private *dev_priv) } else if (IS_GEN(dev_priv, 2)) { for (i = 0; i < 7; i++) I915_WRITE(SWF1(i), dev_priv->regfile.saveSWF1[i]); - } else if (HAS_GMCH_DISPLAY(dev_priv)) { + } else if (HAS_GMCH(dev_priv)) { for (i = 0; i < 16; i++) { I915_WRITE(SWF0(i), dev_priv->regfile.saveSWF0[i]); I915_WRITE(SWF1(i), dev_priv->regfile.saveSWF1[i]); diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c index c0cfe7ae2ba5..41313005af42 100644 --- a/drivers/gpu/drm/i915/i915_sysfs.c +++ b/drivers/gpu/drm/i915/i915_sysfs.c @@ -42,11 +42,11 @@ static inline struct drm_i915_private *kdev_minor_to_i915(struct device *kdev) static u32 calc_residency(struct drm_i915_private *dev_priv, i915_reg_t reg) { - u64 res; + intel_wakeref_t wakeref; + u64 res = 0; - intel_runtime_pm_get(dev_priv); - res = intel_rc6_residency_us(dev_priv, reg); - intel_runtime_pm_put(dev_priv); + with_intel_runtime_pm(dev_priv, wakeref) + res = intel_rc6_residency_us(dev_priv, reg); return DIV_ROUND_CLOSEST_ULL(res, 1000); } @@ -258,9 +258,10 @@ static ssize_t gt_act_freq_mhz_show(struct device *kdev, struct device_attribute *attr, char *buf) { struct drm_i915_private *dev_priv = kdev_minor_to_i915(kdev); + intel_wakeref_t wakeref; int ret; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); mutex_lock(&dev_priv->pcu_lock); if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { @@ -274,7 +275,7 @@ static ssize_t gt_act_freq_mhz_show(struct device *kdev, } mutex_unlock(&dev_priv->pcu_lock); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return snprintf(buf, PAGE_SIZE, "%d\n", ret); } @@ -354,6 +355,7 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev, { struct drm_i915_private *dev_priv = kdev_minor_to_i915(kdev); struct intel_rps *rps = &dev_priv->gt_pm.rps; + intel_wakeref_t wakeref; u32 val; ssize_t ret; @@ -361,7 +363,7 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev, if (ret) return ret; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); mutex_lock(&dev_priv->pcu_lock); @@ -371,7 +373,7 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev, val > rps->max_freq || val < rps->min_freq_softlimit) { mutex_unlock(&dev_priv->pcu_lock); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return -EINVAL; } @@ -392,7 +394,7 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev, mutex_unlock(&dev_priv->pcu_lock); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return ret ?: count; } @@ -412,6 +414,7 @@ static ssize_t gt_min_freq_mhz_store(struct device *kdev, { struct drm_i915_private *dev_priv = kdev_minor_to_i915(kdev); struct intel_rps *rps = &dev_priv->gt_pm.rps; + intel_wakeref_t wakeref; u32 val; ssize_t ret; @@ -419,7 +422,7 @@ static ssize_t gt_min_freq_mhz_store(struct device *kdev, if (ret) return ret; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); mutex_lock(&dev_priv->pcu_lock); @@ -429,7 +432,7 @@ static ssize_t gt_min_freq_mhz_store(struct device *kdev, val > rps->max_freq || val > rps->max_freq_softlimit) { mutex_unlock(&dev_priv->pcu_lock); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return -EINVAL; } @@ -446,7 +449,7 @@ static ssize_t gt_min_freq_mhz_store(struct device *kdev, mutex_unlock(&dev_priv->pcu_lock); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return ret ?: count; } diff --git a/drivers/gpu/drm/i915/i915_timeline.c b/drivers/gpu/drm/i915/i915_timeline.c index 4667cc08c416..b2202d2e58a2 100644 --- a/drivers/gpu/drm/i915/i915_timeline.c +++ b/drivers/gpu/drm/i915/i915_timeline.c @@ -9,34 +9,199 @@ #include "i915_timeline.h" #include "i915_syncmap.h" -void i915_timeline_init(struct drm_i915_private *i915, - struct i915_timeline *timeline, - const char *name) +struct i915_timeline_hwsp { + struct i915_vma *vma; + struct list_head free_link; + u64 free_bitmap; +}; + +static inline struct i915_timeline_hwsp * +i915_timeline_hwsp(const struct i915_timeline *tl) +{ + return tl->hwsp_ggtt->private; +} + +static struct i915_vma *__hwsp_alloc(struct drm_i915_private *i915) +{ + struct drm_i915_gem_object *obj; + struct i915_vma *vma; + + obj = i915_gem_object_create_internal(i915, PAGE_SIZE); + if (IS_ERR(obj)) + return ERR_CAST(obj); + + i915_gem_object_set_cache_coherency(obj, I915_CACHE_LLC); + + vma = i915_vma_instance(obj, &i915->ggtt.vm, NULL); + if (IS_ERR(vma)) + i915_gem_object_put(obj); + + return vma; +} + +static struct i915_vma * +hwsp_alloc(struct i915_timeline *timeline, unsigned int *cacheline) { - lockdep_assert_held(&i915->drm.struct_mutex); + struct drm_i915_private *i915 = timeline->i915; + struct i915_gt_timelines *gt = &i915->gt.timelines; + struct i915_timeline_hwsp *hwsp; + + BUILD_BUG_ON(BITS_PER_TYPE(u64) * CACHELINE_BYTES > PAGE_SIZE); + + spin_lock(>->hwsp_lock); + + /* hwsp_free_list only contains HWSP that have available cachelines */ + hwsp = list_first_entry_or_null(>->hwsp_free_list, + typeof(*hwsp), free_link); + if (!hwsp) { + struct i915_vma *vma; + + spin_unlock(>->hwsp_lock); + + hwsp = kmalloc(sizeof(*hwsp), GFP_KERNEL); + if (!hwsp) + return ERR_PTR(-ENOMEM); + + vma = __hwsp_alloc(i915); + if (IS_ERR(vma)) { + kfree(hwsp); + return vma; + } + + vma->private = hwsp; + hwsp->vma = vma; + hwsp->free_bitmap = ~0ull; + + spin_lock(>->hwsp_lock); + list_add(&hwsp->free_link, >->hwsp_free_list); + } + + GEM_BUG_ON(!hwsp->free_bitmap); + *cacheline = __ffs64(hwsp->free_bitmap); + hwsp->free_bitmap &= ~BIT_ULL(*cacheline); + if (!hwsp->free_bitmap) + list_del(&hwsp->free_link); + + spin_unlock(>->hwsp_lock); + + GEM_BUG_ON(hwsp->vma->private != hwsp); + return hwsp->vma; +} + +static void hwsp_free(struct i915_timeline *timeline) +{ + struct i915_gt_timelines *gt = &timeline->i915->gt.timelines; + struct i915_timeline_hwsp *hwsp; + + hwsp = i915_timeline_hwsp(timeline); + if (!hwsp) /* leave global HWSP alone! */ + return; + + spin_lock(>->hwsp_lock); + + /* As a cacheline becomes available, publish the HWSP on the freelist */ + if (!hwsp->free_bitmap) + list_add_tail(&hwsp->free_link, >->hwsp_free_list); + + hwsp->free_bitmap |= BIT_ULL(timeline->hwsp_offset / CACHELINE_BYTES); + + /* And if no one is left using it, give the page back to the system */ + if (hwsp->free_bitmap == ~0ull) { + i915_vma_put(hwsp->vma); + list_del(&hwsp->free_link); + kfree(hwsp); + } + + spin_unlock(>->hwsp_lock); +} + +int i915_timeline_init(struct drm_i915_private *i915, + struct i915_timeline *timeline, + const char *name, + struct i915_vma *hwsp) +{ + void *vaddr; /* * Ideally we want a set of engines on a single leaf as we expect * to mostly be tracking synchronisation between engines. It is not * a huge issue if this is not the case, but we may want to mitigate * any page crossing penalties if they become an issue. + * + * Called during early_init before we know how many engines there are. */ BUILD_BUG_ON(KSYNCMAP < I915_NUM_ENGINES); + timeline->i915 = i915; timeline->name = name; + timeline->pin_count = 0; + timeline->has_initial_breadcrumb = !hwsp; - list_add(&timeline->link, &i915->gt.timelines); + timeline->hwsp_offset = I915_GEM_HWS_SEQNO_ADDR; + if (!hwsp) { + unsigned int cacheline; + + hwsp = hwsp_alloc(timeline, &cacheline); + if (IS_ERR(hwsp)) + return PTR_ERR(hwsp); + + timeline->hwsp_offset = cacheline * CACHELINE_BYTES; + } + timeline->hwsp_ggtt = i915_vma_get(hwsp); + + vaddr = i915_gem_object_pin_map(hwsp->obj, I915_MAP_WB); + if (IS_ERR(vaddr)) { + hwsp_free(timeline); + i915_vma_put(hwsp); + return PTR_ERR(vaddr); + } - /* Called during early_init before we know how many engines there are */ + timeline->hwsp_seqno = + memset(vaddr + timeline->hwsp_offset, 0, CACHELINE_BYTES); timeline->fence_context = dma_fence_context_alloc(1); spin_lock_init(&timeline->lock); - init_request_active(&timeline->last_request, NULL); + INIT_ACTIVE_REQUEST(&timeline->barrier); + INIT_ACTIVE_REQUEST(&timeline->last_request); INIT_LIST_HEAD(&timeline->requests); i915_syncmap_init(&timeline->sync); + + return 0; +} + +void i915_timelines_init(struct drm_i915_private *i915) +{ + struct i915_gt_timelines *gt = &i915->gt.timelines; + + mutex_init(>->mutex); + INIT_LIST_HEAD(>->active_list); + + spin_lock_init(>->hwsp_lock); + INIT_LIST_HEAD(>->hwsp_free_list); + + /* via i915_gem_wait_for_idle() */ + i915_gem_shrinker_taints_mutex(i915, >->mutex); +} + +static void timeline_add_to_active(struct i915_timeline *tl) +{ + struct i915_gt_timelines *gt = &tl->i915->gt.timelines; + + mutex_lock(>->mutex); + list_add(&tl->link, >->active_list); + mutex_unlock(>->mutex); +} + +static void timeline_remove_from_active(struct i915_timeline *tl) +{ + struct i915_gt_timelines *gt = &tl->i915->gt.timelines; + + mutex_lock(>->mutex); + list_del(&tl->link); + mutex_unlock(>->mutex); } /** @@ -51,11 +216,11 @@ void i915_timeline_init(struct drm_i915_private *i915, */ void i915_timelines_park(struct drm_i915_private *i915) { + struct i915_gt_timelines *gt = &i915->gt.timelines; struct i915_timeline *timeline; - lockdep_assert_held(&i915->drm.struct_mutex); - - list_for_each_entry(timeline, &i915->gt.timelines, link) { + mutex_lock(>->mutex); + list_for_each_entry(timeline, >->active_list, link) { /* * All known fences are completed so we can scrap * the current sync point tracking and start afresh, @@ -64,32 +229,88 @@ void i915_timelines_park(struct drm_i915_private *i915) */ i915_syncmap_free(&timeline->sync); } + mutex_unlock(>->mutex); } void i915_timeline_fini(struct i915_timeline *timeline) { + GEM_BUG_ON(timeline->pin_count); GEM_BUG_ON(!list_empty(&timeline->requests)); + GEM_BUG_ON(i915_active_request_isset(&timeline->barrier)); i915_syncmap_free(&timeline->sync); + hwsp_free(timeline); - list_del(&timeline->link); + i915_gem_object_unpin_map(timeline->hwsp_ggtt->obj); + i915_vma_put(timeline->hwsp_ggtt); } struct i915_timeline * -i915_timeline_create(struct drm_i915_private *i915, const char *name) +i915_timeline_create(struct drm_i915_private *i915, + const char *name, + struct i915_vma *global_hwsp) { struct i915_timeline *timeline; + int err; timeline = kzalloc(sizeof(*timeline), GFP_KERNEL); if (!timeline) return ERR_PTR(-ENOMEM); - i915_timeline_init(i915, timeline, name); + err = i915_timeline_init(i915, timeline, name, global_hwsp); + if (err) { + kfree(timeline); + return ERR_PTR(err); + } + kref_init(&timeline->kref); return timeline; } +int i915_timeline_pin(struct i915_timeline *tl) +{ + int err; + + if (tl->pin_count++) + return 0; + GEM_BUG_ON(!tl->pin_count); + + err = i915_vma_pin(tl->hwsp_ggtt, 0, 0, PIN_GLOBAL | PIN_HIGH); + if (err) + goto unpin; + + tl->hwsp_offset = + i915_ggtt_offset(tl->hwsp_ggtt) + + offset_in_page(tl->hwsp_offset); + + timeline_add_to_active(tl); + + return 0; + +unpin: + tl->pin_count = 0; + return err; +} + +void i915_timeline_unpin(struct i915_timeline *tl) +{ + GEM_BUG_ON(!tl->pin_count); + if (--tl->pin_count) + return; + + timeline_remove_from_active(tl); + + /* + * Since this timeline is idle, all bariers upon which we were waiting + * must also be complete and so we can discard the last used barriers + * without loss of information. + */ + i915_syncmap_free(&tl->sync); + + __i915_vma_unpin(tl->hwsp_ggtt); +} + void __i915_timeline_free(struct kref *kref) { struct i915_timeline *timeline = @@ -99,6 +320,16 @@ void __i915_timeline_free(struct kref *kref) kfree(timeline); } +void i915_timelines_fini(struct drm_i915_private *i915) +{ + struct i915_gt_timelines *gt = &i915->gt.timelines; + + GEM_BUG_ON(!list_empty(>->active_list)); + GEM_BUG_ON(!list_empty(>->hwsp_free_list)); + + mutex_destroy(>->mutex); +} + #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) #include "selftests/mock_timeline.c" #include "selftests/i915_timeline.c" diff --git a/drivers/gpu/drm/i915/i915_timeline.h b/drivers/gpu/drm/i915/i915_timeline.h index 38c1e15e927a..7bec7d2e45bf 100644 --- a/drivers/gpu/drm/i915/i915_timeline.h +++ b/drivers/gpu/drm/i915/i915_timeline.h @@ -28,10 +28,14 @@ #include <linux/list.h> #include <linux/kref.h> +#include "i915_active.h" #include "i915_request.h" #include "i915_syncmap.h" #include "i915_utils.h" +struct i915_vma; +struct i915_timeline_hwsp; + struct i915_timeline { u64 fence_context; u32 seqno; @@ -40,6 +44,13 @@ struct i915_timeline { #define TIMELINE_CLIENT 0 /* default subclass */ #define TIMELINE_ENGINE 1 + unsigned int pin_count; + const u32 *hwsp_seqno; + struct i915_vma *hwsp_ggtt; + u32 hwsp_offset; + + bool has_initial_breadcrumb; + /** * List of breadcrumbs associated with GPU requests currently * outstanding. @@ -48,10 +59,10 @@ struct i915_timeline { /* Contains an RCU guarded pointer to the last request. No reference is * held to the request, users must carefully acquire a reference to - * the request using i915_gem_active_get_request_rcu(), or hold the + * the request using i915_active_request_get_request_rcu(), or hold the * struct_mutex. */ - struct i915_gem_active last_request; + struct i915_active_request last_request; /** * We track the most recent seqno that we wait on in every context so @@ -64,15 +75,27 @@ struct i915_timeline { */ struct i915_syncmap *sync; + /** + * Barrier provides the ability to serialize ordering between different + * timelines. + * + * Users can call i915_timeline_set_barrier which will make all + * subsequent submissions to this timeline be executed only after the + * barrier has been completed. + */ + struct i915_active_request barrier; + struct list_head link; const char *name; + struct drm_i915_private *i915; struct kref kref; }; -void i915_timeline_init(struct drm_i915_private *i915, - struct i915_timeline *tl, - const char *name); +int i915_timeline_init(struct drm_i915_private *i915, + struct i915_timeline *tl, + const char *name, + struct i915_vma *hwsp); void i915_timeline_fini(struct i915_timeline *tl); static inline void @@ -95,7 +118,9 @@ i915_timeline_set_subclass(struct i915_timeline *timeline, } struct i915_timeline * -i915_timeline_create(struct drm_i915_private *i915, const char *name); +i915_timeline_create(struct drm_i915_private *i915, + const char *name, + struct i915_vma *global_hwsp); static inline struct i915_timeline * i915_timeline_get(struct i915_timeline *timeline) @@ -134,6 +159,26 @@ static inline bool i915_timeline_sync_is_later(struct i915_timeline *tl, return __i915_timeline_sync_is_later(tl, fence->context, fence->seqno); } +int i915_timeline_pin(struct i915_timeline *tl); +void i915_timeline_unpin(struct i915_timeline *tl); + +void i915_timelines_init(struct drm_i915_private *i915); void i915_timelines_park(struct drm_i915_private *i915); +void i915_timelines_fini(struct drm_i915_private *i915); + +/** + * i915_timeline_set_barrier - orders submission between different timelines + * @timeline: timeline to set the barrier on + * @rq: request after which new submissions can proceed + * + * Sets the passed in request as the serialization point for all subsequent + * submissions on @timeline. Subsequent requests will not be submitted to GPU + * until the barrier has been completed. + */ +static inline int +i915_timeline_set_barrier(struct i915_timeline *tl, struct i915_request *rq) +{ + return i915_active_request_set(&tl->barrier, rq); +} #endif diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h index 43da14f08dc0..eab313c3163c 100644 --- a/drivers/gpu/drm/i915/i915_trace.h +++ b/drivers/gpu/drm/i915/i915_trace.h @@ -752,31 +752,6 @@ trace_i915_request_out(struct i915_request *rq) #endif #endif -TRACE_EVENT(intel_engine_notify, - TP_PROTO(struct intel_engine_cs *engine, bool waiters), - TP_ARGS(engine, waiters), - - TP_STRUCT__entry( - __field(u32, dev) - __field(u16, class) - __field(u16, instance) - __field(u32, seqno) - __field(bool, waiters) - ), - - TP_fast_assign( - __entry->dev = engine->i915->drm.primary->index; - __entry->class = engine->uabi_class; - __entry->instance = engine->instance; - __entry->seqno = intel_engine_get_seqno(engine); - __entry->waiters = waiters; - ), - - TP_printk("dev=%u, engine=%u:%u, seqno=%u, waiters=%u", - __entry->dev, __entry->class, __entry->instance, - __entry->seqno, __entry->waiters) -); - DEFINE_EVENT(i915_request, i915_request_retire, TP_PROTO(struct i915_request *rq), TP_ARGS(rq) diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c index 5b4d78cdb4ca..b713bed20c38 100644 --- a/drivers/gpu/drm/i915/i915_vma.c +++ b/drivers/gpu/drm/i915/i915_vma.c @@ -63,24 +63,22 @@ static void vma_print_allocator(struct i915_vma *vma, const char *reason) #endif -struct i915_vma_active { - struct i915_gem_active base; - struct i915_vma *vma; - struct rb_node node; - u64 timeline; -}; - -static void -__i915_vma_retire(struct i915_vma *vma, struct i915_request *rq) +static void obj_bump_mru(struct drm_i915_gem_object *obj) { - struct drm_i915_gem_object *obj = vma->obj; + struct drm_i915_private *i915 = to_i915(obj->base.dev); - GEM_BUG_ON(!i915_vma_is_active(vma)); - if (--vma->active_count) - return; + spin_lock(&i915->mm.obj_lock); + if (obj->bind_count) + list_move_tail(&obj->mm.link, &i915->mm.bound_list); + spin_unlock(&i915->mm.obj_lock); - GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); - list_move_tail(&vma->vm_link, &vma->vm->inactive_list); + obj->mm.dirty = true; /* be paranoid */ +} + +static void __i915_vma_retire(struct i915_active *ref) +{ + struct i915_vma *vma = container_of(ref, typeof(*vma), active); + struct drm_i915_gem_object *obj = vma->obj; GEM_BUG_ON(!i915_gem_object_is_active(obj)); if (--obj->active_count) @@ -93,16 +91,12 @@ __i915_vma_retire(struct i915_vma *vma, struct i915_request *rq) reservation_object_unlock(obj->resv); } - /* Bump our place on the bound list to keep it roughly in LRU order + /* + * Bump our place on the bound list to keep it roughly in LRU order * so that we don't steal from recently used but inactive objects * (unless we are forced to ofc!) */ - spin_lock(&rq->i915->mm.obj_lock); - if (obj->bind_count) - list_move_tail(&obj->mm.link, &rq->i915->mm.bound_list); - spin_unlock(&rq->i915->mm.obj_lock); - - obj->mm.dirty = true; /* be paranoid */ + obj_bump_mru(obj); if (i915_gem_object_has_active_reference(obj)) { i915_gem_object_clear_active_reference(obj); @@ -110,21 +104,6 @@ __i915_vma_retire(struct i915_vma *vma, struct i915_request *rq) } } -static void -i915_vma_retire(struct i915_gem_active *base, struct i915_request *rq) -{ - struct i915_vma_active *active = - container_of(base, typeof(*active), base); - - __i915_vma_retire(active->vma, rq); -} - -static void -i915_vma_last_retire(struct i915_gem_active *base, struct i915_request *rq) -{ - __i915_vma_retire(container_of(base, struct i915_vma, last_active), rq); -} - static struct i915_vma * vma_create(struct drm_i915_gem_object *obj, struct i915_address_space *vm, @@ -140,10 +119,9 @@ vma_create(struct drm_i915_gem_object *obj, if (vma == NULL) return ERR_PTR(-ENOMEM); - vma->active = RB_ROOT; + i915_active_init(vm->i915, &vma->active, __i915_vma_retire); + INIT_ACTIVE_REQUEST(&vma->last_fence); - init_request_active(&vma->last_active, i915_vma_last_retire); - init_request_active(&vma->last_fence, NULL); vma->vm = vm; vma->ops = &vm->vma_ops; vma->obj = obj; @@ -190,33 +168,56 @@ vma_create(struct drm_i915_gem_object *obj, i915_gem_object_get_stride(obj)); GEM_BUG_ON(!is_power_of_2(vma->fence_alignment)); - /* - * We put the GGTT vma at the start of the vma-list, followed - * by the ppGGTT vma. This allows us to break early when - * iterating over only the GGTT vma for an object, see - * for_each_ggtt_vma() - */ vma->flags |= I915_VMA_GGTT; - list_add(&vma->obj_link, &obj->vma_list); - } else { - list_add_tail(&vma->obj_link, &obj->vma_list); } + spin_lock(&obj->vma.lock); + rb = NULL; - p = &obj->vma_tree.rb_node; + p = &obj->vma.tree.rb_node; while (*p) { struct i915_vma *pos; + long cmp; rb = *p; pos = rb_entry(rb, struct i915_vma, obj_node); - if (i915_vma_compare(pos, vm, view) < 0) + + /* + * If the view already exists in the tree, another thread + * already created a matching vma, so return the older instance + * and dispose of ours. + */ + cmp = i915_vma_compare(pos, vm, view); + if (cmp == 0) { + spin_unlock(&obj->vma.lock); + kmem_cache_free(vm->i915->vmas, vma); + return pos; + } + + if (cmp < 0) p = &rb->rb_right; else p = &rb->rb_left; } rb_link_node(&vma->obj_node, rb, p); - rb_insert_color(&vma->obj_node, &obj->vma_tree); + rb_insert_color(&vma->obj_node, &obj->vma.tree); + + if (i915_vma_is_ggtt(vma)) + /* + * We put the GGTT vma at the start of the vma-list, followed + * by the ppGGTT vma. This allows us to break early when + * iterating over only the GGTT vma for an object, see + * for_each_ggtt_vma() + */ + list_add(&vma->obj_link, &obj->vma.list); + else + list_add_tail(&vma->obj_link, &obj->vma.list); + + spin_unlock(&obj->vma.lock); + + mutex_lock(&vm->mutex); list_add(&vma->vm_link, &vm->unbound_list); + mutex_unlock(&vm->mutex); return vma; @@ -232,7 +233,7 @@ vma_lookup(struct drm_i915_gem_object *obj, { struct rb_node *rb; - rb = obj->vma_tree.rb_node; + rb = obj->vma.tree.rb_node; while (rb) { struct i915_vma *vma = rb_entry(rb, struct i915_vma, obj_node); long cmp; @@ -272,16 +273,18 @@ i915_vma_instance(struct drm_i915_gem_object *obj, { struct i915_vma *vma; - lockdep_assert_held(&obj->base.dev->struct_mutex); GEM_BUG_ON(view && !i915_is_ggtt(vm)); GEM_BUG_ON(vm->closed); + spin_lock(&obj->vma.lock); vma = vma_lookup(obj, vm, view); - if (!vma) + spin_unlock(&obj->vma.lock); + + /* vma_create() will resolve the race if another creates the vma */ + if (unlikely(!vma)) vma = vma_create(obj, vm, view); GEM_BUG_ON(!IS_ERR(vma) && i915_vma_compare(vma, vm, view)); - GEM_BUG_ON(!IS_ERR(vma) && vma_lookup(obj, vm, view) != vma); return vma; } @@ -659,7 +662,9 @@ i915_vma_insert(struct i915_vma *vma, u64 size, u64 alignment, u64 flags) GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); GEM_BUG_ON(!i915_gem_valid_gtt_space(vma, cache_level)); - list_move_tail(&vma->vm_link, &vma->vm->inactive_list); + mutex_lock(&vma->vm->mutex); + list_move_tail(&vma->vm_link, &vma->vm->bound_list); + mutex_unlock(&vma->vm->mutex); if (vma->obj) { struct drm_i915_gem_object *obj = vma->obj; @@ -692,8 +697,10 @@ i915_vma_remove(struct i915_vma *vma) vma->ops->clear_pages(vma); + mutex_lock(&vma->vm->mutex); drm_mm_remove_node(&vma->node); list_move_tail(&vma->vm_link, &vma->vm->unbound_list); + mutex_unlock(&vma->vm->mutex); /* * Since the unbound list is global, only move to that list if @@ -797,23 +804,27 @@ void i915_vma_reopen(struct i915_vma *vma) static void __i915_vma_destroy(struct i915_vma *vma) { struct drm_i915_private *i915 = vma->vm->i915; - struct i915_vma_active *iter, *n; GEM_BUG_ON(vma->node.allocated); GEM_BUG_ON(vma->fence); - GEM_BUG_ON(i915_gem_active_isset(&vma->last_fence)); + GEM_BUG_ON(i915_active_request_isset(&vma->last_fence)); - list_del(&vma->obj_link); + mutex_lock(&vma->vm->mutex); list_del(&vma->vm_link); - if (vma->obj) - rb_erase(&vma->obj_node, &vma->obj->vma_tree); + mutex_unlock(&vma->vm->mutex); + + if (vma->obj) { + struct drm_i915_gem_object *obj = vma->obj; - rbtree_postorder_for_each_entry_safe(iter, n, &vma->active, node) { - GEM_BUG_ON(i915_gem_active_isset(&iter->base)); - kfree(iter); + spin_lock(&obj->vma.lock); + list_del(&vma->obj_link); + rb_erase(&vma->obj_node, &vma->obj->vma.tree); + spin_unlock(&obj->vma.lock); } + i915_active_fini(&vma->active); + kmem_cache_free(i915->vmas, vma); } @@ -897,104 +908,15 @@ static void export_fence(struct i915_vma *vma, reservation_object_unlock(resv); } -static struct i915_gem_active *active_instance(struct i915_vma *vma, u64 idx) -{ - struct i915_vma_active *active; - struct rb_node **p, *parent; - struct i915_request *old; - - /* - * We track the most recently used timeline to skip a rbtree search - * for the common case, under typical loads we never need the rbtree - * at all. We can reuse the last_active slot if it is empty, that is - * after the previous activity has been retired, or if the active - * matches the current timeline. - * - * Note that we allow the timeline to be active simultaneously in - * the rbtree and the last_active cache. We do this to avoid having - * to search and replace the rbtree element for a new timeline, with - * the cost being that we must be aware that the vma may be retired - * twice for the same timeline (as the older rbtree element will be - * retired before the new request added to last_active). - */ - old = i915_gem_active_raw(&vma->last_active, - &vma->vm->i915->drm.struct_mutex); - if (!old || old->fence.context == idx) - goto out; - - /* Move the currently active fence into the rbtree */ - idx = old->fence.context; - - parent = NULL; - p = &vma->active.rb_node; - while (*p) { - parent = *p; - - active = rb_entry(parent, struct i915_vma_active, node); - if (active->timeline == idx) - goto replace; - - if (active->timeline < idx) - p = &parent->rb_right; - else - p = &parent->rb_left; - } - - active = kmalloc(sizeof(*active), GFP_KERNEL); - - /* kmalloc may retire the vma->last_active request (thanks shrinker)! */ - if (unlikely(!i915_gem_active_raw(&vma->last_active, - &vma->vm->i915->drm.struct_mutex))) { - kfree(active); - goto out; - } - - if (unlikely(!active)) - return ERR_PTR(-ENOMEM); - - init_request_active(&active->base, i915_vma_retire); - active->vma = vma; - active->timeline = idx; - - rb_link_node(&active->node, parent, p); - rb_insert_color(&active->node, &vma->active); - -replace: - /* - * Overwrite the previous active slot in the rbtree with last_active, - * leaving last_active zeroed. If the previous slot is still active, - * we must be careful as we now only expect to receive one retire - * callback not two, and so much undo the active counting for the - * overwritten slot. - */ - if (i915_gem_active_isset(&active->base)) { - /* Retire ourselves from the old rq->active_list */ - __list_del_entry(&active->base.link); - vma->active_count--; - GEM_BUG_ON(!vma->active_count); - } - GEM_BUG_ON(list_empty(&vma->last_active.link)); - list_replace_init(&vma->last_active.link, &active->base.link); - active->base.request = fetch_and_zero(&vma->last_active.request); - -out: - return &vma->last_active; -} - int i915_vma_move_to_active(struct i915_vma *vma, struct i915_request *rq, unsigned int flags) { struct drm_i915_gem_object *obj = vma->obj; - struct i915_gem_active *active; lockdep_assert_held(&rq->i915->drm.struct_mutex); GEM_BUG_ON(!drm_mm_node_allocated(&vma->node)); - active = active_instance(vma, rq->fence.context); - if (IS_ERR(active)) - return PTR_ERR(active); - /* * Add a reference if we're newly entering the active list. * The order in which we add operations to the retirement queue is @@ -1003,11 +925,15 @@ int i915_vma_move_to_active(struct i915_vma *vma, * add the active reference first and queue for it to be dropped * *last*. */ - if (!i915_gem_active_isset(active) && !vma->active_count++) { - list_move_tail(&vma->vm_link, &vma->vm->active_list); + if (!vma->active.count) obj->active_count++; + + if (unlikely(i915_active_ref(&vma->active, rq->fence.context, rq))) { + if (!vma->active.count) + obj->active_count--; + return -ENOMEM; } - i915_gem_active_set(active, rq); + GEM_BUG_ON(!i915_vma_is_active(vma)); GEM_BUG_ON(!obj->active_count); @@ -1016,14 +942,14 @@ int i915_vma_move_to_active(struct i915_vma *vma, obj->write_domain = I915_GEM_DOMAIN_RENDER; if (intel_fb_obj_invalidate(obj, ORIGIN_CS)) - i915_gem_active_set(&obj->frontbuffer_write, rq); + __i915_active_request_set(&obj->frontbuffer_write, rq); obj->read_domains = 0; } obj->read_domains |= I915_GEM_GPU_DOMAINS; if (flags & EXEC_OBJECT_NEEDS_FENCE) - i915_gem_active_set(&vma->last_fence, rq); + __i915_active_request_set(&vma->last_fence, rq); export_fence(vma, rq, flags); return 0; @@ -1041,8 +967,6 @@ int i915_vma_unbind(struct i915_vma *vma) */ might_sleep(); if (i915_vma_is_active(vma)) { - struct i915_vma_active *active, *n; - /* * When a closed VMA is retired, it is unbound - eek. * In order to prevent it from being recursively closed, @@ -1058,21 +982,12 @@ int i915_vma_unbind(struct i915_vma *vma) */ __i915_vma_pin(vma); - ret = i915_gem_active_retire(&vma->last_active, - &vma->vm->i915->drm.struct_mutex); + ret = i915_active_wait(&vma->active); if (ret) goto unpin; - rbtree_postorder_for_each_entry_safe(active, n, - &vma->active, node) { - ret = i915_gem_active_retire(&active->base, - &vma->vm->i915->drm.struct_mutex); - if (ret) - goto unpin; - } - - ret = i915_gem_active_retire(&vma->last_fence, - &vma->vm->i915->drm.struct_mutex); + ret = i915_active_request_retire(&vma->last_fence, + &vma->vm->i915->drm.struct_mutex); unpin: __i915_vma_unpin(vma); if (ret) diff --git a/drivers/gpu/drm/i915/i915_vma.h b/drivers/gpu/drm/i915/i915_vma.h index 4f7c1c7599f4..7c742027f866 100644 --- a/drivers/gpu/drm/i915/i915_vma.h +++ b/drivers/gpu/drm/i915/i915_vma.h @@ -34,6 +34,7 @@ #include "i915_gem_fence_reg.h" #include "i915_gem_object.h" +#include "i915_active.h" #include "i915_request.h" enum i915_cache_level; @@ -71,34 +72,45 @@ struct i915_vma { unsigned int open_count; unsigned long flags; /** - * How many users have pinned this object in GTT space. The following - * users can each hold at most one reference: pwrite/pread, execbuffer - * (objects are not allowed multiple times for the same batchbuffer), - * and the framebuffer code. When switching/pageflipping, the - * framebuffer code has at most two buffers pinned per crtc. + * How many users have pinned this object in GTT space. * - * In the worst case this is 1 + 1 + 1 + 2*2 = 7. That would fit into 3 - * bits with absolutely no headroom. So use 4 bits. + * This is a tightly bound, fairly small number of users, so we + * stuff inside the flags field so that we can both check for overflow + * and detect a no-op i915_vma_pin() in a single check, while also + * pinning the vma. + * + * The worst case display setup would have the same vma pinned for + * use on each plane on each crtc, while also building the next atomic + * state and holding a pin for the length of the cleanup queue. In the + * future, the flip queue may be increased from 1. + * Estimated worst case: 3 [qlen] * 4 [max crtcs] * 7 [max planes] = 84 + * + * For GEM, the number of concurrent users for pwrite/pread is + * unbounded. For execbuffer, it is currently one but will in future + * be extended to allow multiple clients to pin vma concurrently. + * + * We also use suballocated pages, with each suballocation claiming + * its own pin on the shared vma. At present, this is limited to + * exclusive cachelines of a single page, so a maximum of 64 possible + * users. */ -#define I915_VMA_PIN_MASK 0xf -#define I915_VMA_PIN_OVERFLOW BIT(5) +#define I915_VMA_PIN_MASK 0xff +#define I915_VMA_PIN_OVERFLOW BIT(8) /** Flags and address space this VMA is bound to */ -#define I915_VMA_GLOBAL_BIND BIT(6) -#define I915_VMA_LOCAL_BIND BIT(7) +#define I915_VMA_GLOBAL_BIND BIT(9) +#define I915_VMA_LOCAL_BIND BIT(10) #define I915_VMA_BIND_MASK (I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND | I915_VMA_PIN_OVERFLOW) -#define I915_VMA_GGTT BIT(8) -#define I915_VMA_CAN_FENCE BIT(9) -#define I915_VMA_CLOSED BIT(10) -#define I915_VMA_USERFAULT_BIT 11 +#define I915_VMA_GGTT BIT(11) +#define I915_VMA_CAN_FENCE BIT(12) +#define I915_VMA_CLOSED BIT(13) +#define I915_VMA_USERFAULT_BIT 14 #define I915_VMA_USERFAULT BIT(I915_VMA_USERFAULT_BIT) -#define I915_VMA_GGTT_WRITE BIT(12) +#define I915_VMA_GGTT_WRITE BIT(15) - unsigned int active_count; - struct rb_root active; - struct i915_gem_active last_active; - struct i915_gem_active last_fence; + struct i915_active active; + struct i915_active_request last_fence; /** * Support different GGTT views into the same object. @@ -141,9 +153,9 @@ i915_vma_instance(struct drm_i915_gem_object *obj, void i915_vma_unpin_and_release(struct i915_vma **p_vma, unsigned int flags); #define I915_VMA_RELEASE_MAP BIT(0) -static inline bool i915_vma_is_active(struct i915_vma *vma) +static inline bool i915_vma_is_active(const struct i915_vma *vma) { - return vma->active_count; + return !i915_active_is_idle(&vma->active); } int __must_check i915_vma_move_to_active(struct i915_vma *vma, @@ -425,7 +437,7 @@ void i915_vma_parked(struct drm_i915_private *i915); * or the list is empty ofc. */ #define for_each_ggtt_vma(V, OBJ) \ - list_for_each_entry(V, &(OBJ)->vma_list, obj_link) \ + list_for_each_entry(V, &(OBJ)->vma.list, obj_link) \ for_each_until(!i915_vma_is_ggtt(V)) #endif diff --git a/drivers/gpu/drm/i915/icl_dsi.c b/drivers/gpu/drm/i915/icl_dsi.c index d9dcee4ec51f..73a7bee24a66 100644 --- a/drivers/gpu/drm/i915/icl_dsi.c +++ b/drivers/gpu/drm/i915/icl_dsi.c @@ -337,9 +337,11 @@ static void gen11_dsi_enable_io_power(struct intel_encoder *encoder) } for_each_dsi_port(port, intel_dsi->ports) { - intel_display_power_get(dev_priv, port == PORT_A ? - POWER_DOMAIN_PORT_DDI_A_IO : - POWER_DOMAIN_PORT_DDI_B_IO); + intel_dsi->io_wakeref[port] = + intel_display_power_get(dev_priv, + port == PORT_A ? + POWER_DOMAIN_PORT_DDI_A_IO : + POWER_DOMAIN_PORT_DDI_B_IO); } } @@ -1125,10 +1127,18 @@ static void gen11_dsi_disable_io_power(struct intel_encoder *encoder) enum port port; u32 tmp; - intel_display_power_put(dev_priv, POWER_DOMAIN_PORT_DDI_A_IO); - - if (intel_dsi->dual_link) - intel_display_power_put(dev_priv, POWER_DOMAIN_PORT_DDI_B_IO); + for_each_dsi_port(port, intel_dsi->ports) { + intel_wakeref_t wakeref; + + wakeref = fetch_and_zero(&intel_dsi->io_wakeref[port]); + if (wakeref) { + intel_display_power_put(dev_priv, + port == PORT_A ? + POWER_DOMAIN_PORT_DDI_A_IO : + POWER_DOMAIN_PORT_DDI_B_IO, + wakeref); + } + } /* set mode to DDI */ for_each_dsi_port(port, intel_dsi->ports) { @@ -1229,13 +1239,15 @@ static bool gen11_dsi_get_hw_state(struct intel_encoder *encoder, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base); - u32 tmp; - enum port port; enum transcoder dsi_trans; + intel_wakeref_t wakeref; + enum port port; bool ret = false; + u32 tmp; - if (!intel_display_power_get_if_enabled(dev_priv, - encoder->power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + encoder->power_domain); + if (!wakeref) return false; for_each_dsi_port(port, intel_dsi->ports) { @@ -1260,7 +1272,7 @@ static bool gen11_dsi_get_hw_state(struct intel_encoder *encoder, ret = tmp & PIPECONF_ENABLE; } out: - intel_display_power_put(dev_priv, encoder->power_domain); + intel_display_power_put(dev_priv, encoder->power_domain, wakeref); return ret; } @@ -1378,6 +1390,7 @@ void icl_dsi_init(struct drm_i915_private *dev_priv) encoder->disable = gen11_dsi_disable; encoder->port = port; encoder->get_config = gen11_dsi_get_config; + encoder->update_pipe = intel_panel_update_backlight; encoder->compute_config = gen11_dsi_compute_config; encoder->get_hw_state = gen11_dsi_get_hw_state; encoder->type = INTEL_OUTPUT_DSI; diff --git a/drivers/gpu/drm/i915/intel_atomic.c b/drivers/gpu/drm/i915/intel_atomic.c index f415ed239184..7cf9290ea34a 100644 --- a/drivers/gpu/drm/i915/intel_atomic.c +++ b/drivers/gpu/drm/i915/intel_atomic.c @@ -48,7 +48,7 @@ int intel_digital_connector_atomic_get_property(struct drm_connector *connector, const struct drm_connector_state *state, struct drm_property *property, - uint64_t *val) + u64 *val) { struct drm_device *dev = connector->dev; struct drm_i915_private *dev_priv = to_i915(dev); @@ -80,7 +80,7 @@ int intel_digital_connector_atomic_get_property(struct drm_connector *connector, int intel_digital_connector_atomic_set_property(struct drm_connector *connector, struct drm_connector_state *state, struct drm_property *property, - uint64_t val) + u64 val) { struct drm_device *dev = connector->dev; struct drm_i915_private *dev_priv = to_i915(dev); diff --git a/drivers/gpu/drm/i915/intel_atomic_plane.c b/drivers/gpu/drm/i915/intel_atomic_plane.c index 79139d496c78..db0965904439 100644 --- a/drivers/gpu/drm/i915/intel_atomic_plane.c +++ b/drivers/gpu/drm/i915/intel_atomic_plane.c @@ -112,41 +112,39 @@ intel_plane_destroy_state(struct drm_plane *plane, } int intel_plane_atomic_check_with_state(const struct intel_crtc_state *old_crtc_state, - struct intel_crtc_state *crtc_state, + struct intel_crtc_state *new_crtc_state, const struct intel_plane_state *old_plane_state, - struct intel_plane_state *intel_state) + struct intel_plane_state *new_plane_state) { - struct drm_plane *plane = intel_state->base.plane; - struct drm_plane_state *state = &intel_state->base; - struct intel_plane *intel_plane = to_intel_plane(plane); + struct intel_plane *plane = to_intel_plane(new_plane_state->base.plane); int ret; - crtc_state->active_planes &= ~BIT(intel_plane->id); - crtc_state->nv12_planes &= ~BIT(intel_plane->id); - intel_state->base.visible = false; + new_crtc_state->active_planes &= ~BIT(plane->id); + new_crtc_state->nv12_planes &= ~BIT(plane->id); + new_plane_state->base.visible = false; - /* If this is a cursor plane, no further checks are needed. */ - if (!intel_state->base.crtc && !old_plane_state->base.crtc) + if (!new_plane_state->base.crtc && !old_plane_state->base.crtc) return 0; - ret = intel_plane->check_plane(crtc_state, intel_state); + ret = plane->check_plane(new_crtc_state, new_plane_state); if (ret) return ret; /* FIXME pre-g4x don't work like this */ - if (state->visible) - crtc_state->active_planes |= BIT(intel_plane->id); + if (new_plane_state->base.visible) + new_crtc_state->active_planes |= BIT(plane->id); - if (state->visible && state->fb->format->format == DRM_FORMAT_NV12) - crtc_state->nv12_planes |= BIT(intel_plane->id); + if (new_plane_state->base.visible && + new_plane_state->base.fb->format->format == DRM_FORMAT_NV12) + new_crtc_state->nv12_planes |= BIT(plane->id); - if (state->visible || old_plane_state->base.visible) - crtc_state->update_planes |= BIT(intel_plane->id); + if (new_plane_state->base.visible || old_plane_state->base.visible) + new_crtc_state->update_planes |= BIT(plane->id); return intel_plane_atomic_calc_changes(old_crtc_state, - &crtc_state->base, + &new_crtc_state->base, old_plane_state, - state); + &new_plane_state->base); } static int intel_plane_atomic_check(struct drm_plane *plane, @@ -313,7 +311,7 @@ int intel_plane_atomic_get_property(struct drm_plane *plane, const struct drm_plane_state *state, struct drm_property *property, - uint64_t *val) + u64 *val) { DRM_DEBUG_KMS("Unknown property [PROP:%d:%s]\n", property->base.id, property->name); @@ -336,7 +334,7 @@ int intel_plane_atomic_set_property(struct drm_plane *plane, struct drm_plane_state *state, struct drm_property *property, - uint64_t val) + u64 val) { DRM_DEBUG_KMS("Unknown property [PROP:%d:%s]\n", property->base.id, property->name); diff --git a/drivers/gpu/drm/i915/intel_audio.c b/drivers/gpu/drm/i915/intel_audio.c index 202a58cf2d9f..de26cd0a5497 100644 --- a/drivers/gpu/drm/i915/intel_audio.c +++ b/drivers/gpu/drm/i915/intel_audio.c @@ -748,7 +748,8 @@ static void i915_audio_component_get_power(struct device *kdev) static void i915_audio_component_put_power(struct device *kdev) { - intel_display_power_put(kdev_to_i915(kdev), POWER_DOMAIN_AUDIO); + intel_display_power_put_unchecked(kdev_to_i915(kdev), + POWER_DOMAIN_AUDIO); } static void i915_audio_component_codec_wake_override(struct device *kdev, diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c index 140c218128cb..b508d8a735e0 100644 --- a/drivers/gpu/drm/i915/intel_bios.c +++ b/drivers/gpu/drm/i915/intel_bios.c @@ -1663,6 +1663,13 @@ init_vbt_missing_defaults(struct drm_i915_private *dev_priv) struct ddi_vbt_port_info *info = &dev_priv->vbt.ddi_port_info[port]; + /* + * VBT has the TypeC mode (native,TBT/USB) and we don't want + * to detect it. + */ + if (intel_port_is_tc(dev_priv, port)) + continue; + info->supports_dvi = (port != PORT_A && port != PORT_E); info->supports_hdmi = info->supports_dvi; info->supports_dp = (port != PORT_E); @@ -1946,6 +1953,15 @@ bool intel_bios_is_port_present(struct drm_i915_private *dev_priv, enum port por }; int i; + if (HAS_DDI(dev_priv)) { + const struct ddi_vbt_port_info *port_info = + &dev_priv->vbt.ddi_port_info[port]; + + return port_info->supports_dp || + port_info->supports_dvi || + port_info->supports_hdmi; + } + /* FIXME maybe deal with port A as well? */ if (WARN_ON(port == PORT_A) || port >= ARRAY_SIZE(port_mapping)) return false; diff --git a/drivers/gpu/drm/i915/intel_breadcrumbs.c b/drivers/gpu/drm/i915/intel_breadcrumbs.c index 4ed7105d7ff5..cacaa1d04d17 100644 --- a/drivers/gpu/drm/i915/intel_breadcrumbs.c +++ b/drivers/gpu/drm/i915/intel_breadcrumbs.c @@ -29,174 +29,146 @@ #define task_asleep(tsk) ((tsk)->state & TASK_NORMAL && !(tsk)->on_rq) -static unsigned int __intel_breadcrumbs_wakeup(struct intel_breadcrumbs *b) +static void irq_enable(struct intel_engine_cs *engine) { - struct intel_wait *wait; - unsigned int result = 0; - - lockdep_assert_held(&b->irq_lock); - - wait = b->irq_wait; - if (wait) { - /* - * N.B. Since task_asleep() and ttwu are not atomic, the - * waiter may actually go to sleep after the check, causing - * us to suppress a valid wakeup. We prefer to reduce the - * number of false positive missed_breadcrumb() warnings - * at the expense of a few false negatives, as it it easy - * to trigger a false positive under heavy load. Enough - * signal should remain from genuine missed_breadcrumb() - * for us to detect in CI. - */ - bool was_asleep = task_asleep(wait->tsk); - - result = ENGINE_WAKEUP_WAITER; - if (wake_up_process(wait->tsk) && was_asleep) - result |= ENGINE_WAKEUP_ASLEEP; - } + if (!engine->irq_enable) + return; - return result; + /* Caller disables interrupts */ + spin_lock(&engine->i915->irq_lock); + engine->irq_enable(engine); + spin_unlock(&engine->i915->irq_lock); } -unsigned int intel_engine_wakeup(struct intel_engine_cs *engine) +static void irq_disable(struct intel_engine_cs *engine) { - struct intel_breadcrumbs *b = &engine->breadcrumbs; - unsigned long flags; - unsigned int result; - - spin_lock_irqsave(&b->irq_lock, flags); - result = __intel_breadcrumbs_wakeup(b); - spin_unlock_irqrestore(&b->irq_lock, flags); - - return result; -} + if (!engine->irq_disable) + return; -static unsigned long wait_timeout(void) -{ - return round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES); + /* Caller disables interrupts */ + spin_lock(&engine->i915->irq_lock); + engine->irq_disable(engine); + spin_unlock(&engine->i915->irq_lock); } -static noinline void missed_breadcrumb(struct intel_engine_cs *engine) +static void __intel_breadcrumbs_disarm_irq(struct intel_breadcrumbs *b) { - if (GEM_SHOW_DEBUG()) { - struct drm_printer p = drm_debug_printer(__func__); + lockdep_assert_held(&b->irq_lock); - intel_engine_dump(engine, &p, - "%s missed breadcrumb at %pS\n", - engine->name, __builtin_return_address(0)); - } + GEM_BUG_ON(!b->irq_enabled); + if (!--b->irq_enabled) + irq_disable(container_of(b, + struct intel_engine_cs, + breadcrumbs)); - set_bit(engine->id, &engine->i915->gpu_error.missed_irq_rings); + b->irq_armed = false; } -static void intel_breadcrumbs_hangcheck(struct timer_list *t) +void intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine) { - struct intel_engine_cs *engine = - from_timer(engine, t, breadcrumbs.hangcheck); struct intel_breadcrumbs *b = &engine->breadcrumbs; - unsigned int irq_count; if (!b->irq_armed) return; - irq_count = READ_ONCE(b->irq_count); - if (b->hangcheck_interrupts != irq_count) { - b->hangcheck_interrupts = irq_count; - mod_timer(&b->hangcheck, wait_timeout()); - return; - } + spin_lock_irq(&b->irq_lock); + if (b->irq_armed) + __intel_breadcrumbs_disarm_irq(b); + spin_unlock_irq(&b->irq_lock); +} - /* We keep the hangcheck timer alive until we disarm the irq, even - * if there are no waiters at present. - * - * If the waiter was currently running, assume it hasn't had a chance - * to process the pending interrupt (e.g, low priority task on a loaded - * system) and wait until it sleeps before declaring a missed interrupt. - * - * If the waiter was asleep (and not even pending a wakeup), then we - * must have missed an interrupt as the GPU has stopped advancing - * but we still have a waiter. Assuming all batches complete within - * DRM_I915_HANGCHECK_JIFFIES [1.5s]! - */ - if (intel_engine_wakeup(engine) & ENGINE_WAKEUP_ASLEEP) { - missed_breadcrumb(engine); - mod_timer(&b->fake_irq, jiffies + 1); - } else { - mod_timer(&b->hangcheck, wait_timeout()); - } +static inline bool __request_completed(const struct i915_request *rq) +{ + return i915_seqno_passed(__hwsp_seqno(rq), rq->fence.seqno); } -static void intel_breadcrumbs_fake_irq(struct timer_list *t) +bool intel_engine_breadcrumbs_irq(struct intel_engine_cs *engine) { - struct intel_engine_cs *engine = - from_timer(engine, t, breadcrumbs.fake_irq); struct intel_breadcrumbs *b = &engine->breadcrumbs; + struct intel_context *ce, *cn; + struct list_head *pos, *next; + LIST_HEAD(signal); - /* - * The timer persists in case we cannot enable interrupts, - * or if we have previously seen seqno/interrupt incoherency - * ("missed interrupt" syndrome, better known as a "missed breadcrumb"). - * Here the worker will wake up every jiffie in order to kick the - * oldest waiter to do the coherent seqno check. - */ + spin_lock(&b->irq_lock); - spin_lock_irq(&b->irq_lock); - if (b->irq_armed && !__intel_breadcrumbs_wakeup(b)) - __intel_engine_disarm_breadcrumbs(engine); - spin_unlock_irq(&b->irq_lock); - if (!b->irq_armed) - return; + if (b->irq_armed && list_empty(&b->signalers)) + __intel_breadcrumbs_disarm_irq(b); - /* If the user has disabled the fake-irq, restore the hangchecking */ - if (!test_bit(engine->id, &engine->i915->gpu_error.missed_irq_rings)) { - mod_timer(&b->hangcheck, wait_timeout()); - return; - } + list_for_each_entry_safe(ce, cn, &b->signalers, signal_link) { + GEM_BUG_ON(list_empty(&ce->signals)); - mod_timer(&b->fake_irq, jiffies + 1); -} + list_for_each_safe(pos, next, &ce->signals) { + struct i915_request *rq = + list_entry(pos, typeof(*rq), signal_link); -static void irq_enable(struct intel_engine_cs *engine) -{ - /* - * FIXME: Ideally we want this on the API boundary, but for the - * sake of testing with mock breadcrumbs (no HW so unable to - * enable irqs) we place it deep within the bowels, at the point - * of no return. - */ - GEM_BUG_ON(!intel_irqs_enabled(engine->i915)); + if (!__request_completed(rq)) + break; - /* Caller disables interrupts */ - if (engine->irq_enable) { - spin_lock(&engine->i915->irq_lock); - engine->irq_enable(engine); - spin_unlock(&engine->i915->irq_lock); + GEM_BUG_ON(!test_bit(I915_FENCE_FLAG_SIGNAL, + &rq->fence.flags)); + clear_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags); + + /* + * We may race with direct invocation of + * dma_fence_signal(), e.g. i915_request_retire(), + * in which case we can skip processing it ourselves. + */ + if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, + &rq->fence.flags)) + continue; + + /* + * Queue for execution after dropping the signaling + * spinlock as the callback chain may end up adding + * more signalers to the same context or engine. + */ + i915_request_get(rq); + list_add_tail(&rq->signal_link, &signal); + } + + /* + * We process the list deletion in bulk, only using a list_add + * (not list_move) above but keeping the status of + * rq->signal_link known with the I915_FENCE_FLAG_SIGNAL bit. + */ + if (!list_is_first(pos, &ce->signals)) { + /* Advance the list to the first incomplete request */ + __list_del_many(&ce->signals, pos); + if (&ce->signals == pos) /* now empty */ + list_del_init(&ce->signal_link); + } } -} -static void irq_disable(struct intel_engine_cs *engine) -{ - /* Caller disables interrupts */ - if (engine->irq_disable) { - spin_lock(&engine->i915->irq_lock); - engine->irq_disable(engine); - spin_unlock(&engine->i915->irq_lock); + spin_unlock(&b->irq_lock); + + list_for_each_safe(pos, next, &signal) { + struct i915_request *rq = + list_entry(pos, typeof(*rq), signal_link); + + dma_fence_signal(&rq->fence); + i915_request_put(rq); } + + return !list_empty(&signal); } -void __intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine) +bool intel_engine_signal_breadcrumbs(struct intel_engine_cs *engine) { - struct intel_breadcrumbs *b = &engine->breadcrumbs; + bool result; - lockdep_assert_held(&b->irq_lock); - GEM_BUG_ON(b->irq_wait); - GEM_BUG_ON(!b->irq_armed); + local_irq_disable(); + result = intel_engine_breadcrumbs_irq(engine); + local_irq_enable(); - GEM_BUG_ON(!b->irq_enabled); - if (!--b->irq_enabled) - irq_disable(engine); + return result; +} - b->irq_armed = false; +static void signal_irq_work(struct irq_work *work) +{ + struct intel_engine_cs *engine = + container_of(work, typeof(*engine), breadcrumbs.irq_work); + + intel_engine_breadcrumbs_irq(engine); } void intel_engine_pin_breadcrumbs_irq(struct intel_engine_cs *engine) @@ -221,646 +193,155 @@ void intel_engine_unpin_breadcrumbs_irq(struct intel_engine_cs *engine) spin_unlock_irq(&b->irq_lock); } -void intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine) -{ - struct intel_breadcrumbs *b = &engine->breadcrumbs; - struct intel_wait *wait, *n; - - if (!b->irq_armed) - return; - - /* - * We only disarm the irq when we are idle (all requests completed), - * so if the bottom-half remains asleep, it missed the request - * completion. - */ - if (intel_engine_wakeup(engine) & ENGINE_WAKEUP_ASLEEP) - missed_breadcrumb(engine); - - spin_lock_irq(&b->rb_lock); - - spin_lock(&b->irq_lock); - b->irq_wait = NULL; - if (b->irq_armed) - __intel_engine_disarm_breadcrumbs(engine); - spin_unlock(&b->irq_lock); - - rbtree_postorder_for_each_entry_safe(wait, n, &b->waiters, node) { - GEM_BUG_ON(!intel_engine_signaled(engine, wait->seqno)); - RB_CLEAR_NODE(&wait->node); - wake_up_process(wait->tsk); - } - b->waiters = RB_ROOT; - - spin_unlock_irq(&b->rb_lock); -} - -static bool use_fake_irq(const struct intel_breadcrumbs *b) -{ - const struct intel_engine_cs *engine = - container_of(b, struct intel_engine_cs, breadcrumbs); - - if (!test_bit(engine->id, &engine->i915->gpu_error.missed_irq_rings)) - return false; - - /* - * Only start with the heavy weight fake irq timer if we have not - * seen any interrupts since enabling it the first time. If the - * interrupts are still arriving, it means we made a mistake in our - * engine->seqno_barrier(), a timing error that should be transient - * and unlikely to reoccur. - */ - return READ_ONCE(b->irq_count) == b->hangcheck_interrupts; -} - -static void enable_fake_irq(struct intel_breadcrumbs *b) -{ - /* Ensure we never sleep indefinitely */ - if (!b->irq_enabled || use_fake_irq(b)) - mod_timer(&b->fake_irq, jiffies + 1); - else - mod_timer(&b->hangcheck, wait_timeout()); -} - -static bool __intel_breadcrumbs_enable_irq(struct intel_breadcrumbs *b) +static void __intel_breadcrumbs_arm_irq(struct intel_breadcrumbs *b) { struct intel_engine_cs *engine = container_of(b, struct intel_engine_cs, breadcrumbs); - struct drm_i915_private *i915 = engine->i915; - bool enabled; lockdep_assert_held(&b->irq_lock); if (b->irq_armed) - return false; + return; - /* The breadcrumb irq will be disarmed on the interrupt after the + /* + * The breadcrumb irq will be disarmed on the interrupt after the * waiters are signaled. This gives us a single interrupt window in * which we can add a new waiter and avoid the cost of re-enabling * the irq. */ b->irq_armed = true; - if (I915_SELFTEST_ONLY(b->mock)) { - /* For our mock objects we want to avoid interaction - * with the real hardware (which is not set up). So - * we simply pretend we have enabled the powerwell - * and the irq, and leave it up to the mock - * implementation to call intel_engine_wakeup() - * itself when it wants to simulate a user interrupt, - */ - return true; - } - - /* Since we are waiting on a request, the GPU should be busy + /* + * Since we are waiting on a request, the GPU should be busy * and should have its own rpm reference. This is tracked * by i915->gt.awake, we can forgo holding our own wakref * for the interrupt as before i915->gt.awake is released (when * the driver is idle) we disarm the breadcrumbs. */ - /* No interrupts? Kick the waiter every jiffie! */ - enabled = false; - if (!b->irq_enabled++ && - !test_bit(engine->id, &i915->gpu_error.test_irq_rings)) { + if (!b->irq_enabled++) irq_enable(engine); - enabled = true; - } - - enable_fake_irq(b); - return enabled; -} - -static inline struct intel_wait *to_wait(struct rb_node *node) -{ - return rb_entry(node, struct intel_wait, node); -} - -static inline void __intel_breadcrumbs_finish(struct intel_breadcrumbs *b, - struct intel_wait *wait) -{ - lockdep_assert_held(&b->rb_lock); - GEM_BUG_ON(b->irq_wait == wait); - - /* - * This request is completed, so remove it from the tree, mark it as - * complete, and *then* wake up the associated task. N.B. when the - * task wakes up, it will find the empty rb_node, discern that it - * has already been removed from the tree and skip the serialisation - * of the b->rb_lock and b->irq_lock. This means that the destruction - * of the intel_wait is not serialised with the interrupt handler - * by the waiter - it must instead be serialised by the caller. - */ - rb_erase(&wait->node, &b->waiters); - RB_CLEAR_NODE(&wait->node); - - if (wait->tsk->state != TASK_RUNNING) - wake_up_process(wait->tsk); /* implicit smp_wmb() */ -} - -static inline void __intel_breadcrumbs_next(struct intel_engine_cs *engine, - struct rb_node *next) -{ - struct intel_breadcrumbs *b = &engine->breadcrumbs; - - spin_lock(&b->irq_lock); - GEM_BUG_ON(!b->irq_armed); - GEM_BUG_ON(!b->irq_wait); - b->irq_wait = to_wait(next); - spin_unlock(&b->irq_lock); - - /* We always wake up the next waiter that takes over as the bottom-half - * as we may delegate not only the irq-seqno barrier to the next waiter - * but also the task of waking up concurrent waiters. - */ - if (next) - wake_up_process(to_wait(next)->tsk); } -static bool __intel_engine_add_wait(struct intel_engine_cs *engine, - struct intel_wait *wait) +void intel_engine_init_breadcrumbs(struct intel_engine_cs *engine) { struct intel_breadcrumbs *b = &engine->breadcrumbs; - struct rb_node **p, *parent, *completed; - bool first, armed; - u32 seqno; - - GEM_BUG_ON(!wait->seqno); - - /* Insert the request into the retirement ordered list - * of waiters by walking the rbtree. If we are the oldest - * seqno in the tree (the first to be retired), then - * set ourselves as the bottom-half. - * - * As we descend the tree, prune completed branches since we hold the - * spinlock we know that the first_waiter must be delayed and can - * reduce some of the sequential wake up latency if we take action - * ourselves and wake up the completed tasks in parallel. Also, by - * removing stale elements in the tree, we may be able to reduce the - * ping-pong between the old bottom-half and ourselves as first-waiter. - */ - armed = false; - first = true; - parent = NULL; - completed = NULL; - seqno = intel_engine_get_seqno(engine); - - /* If the request completed before we managed to grab the spinlock, - * return now before adding ourselves to the rbtree. We let the - * current bottom-half handle any pending wakeups and instead - * try and get out of the way quickly. - */ - if (i915_seqno_passed(seqno, wait->seqno)) { - RB_CLEAR_NODE(&wait->node); - return first; - } - - p = &b->waiters.rb_node; - while (*p) { - parent = *p; - if (wait->seqno == to_wait(parent)->seqno) { - /* We have multiple waiters on the same seqno, select - * the highest priority task (that with the smallest - * task->prio) to serve as the bottom-half for this - * group. - */ - if (wait->tsk->prio > to_wait(parent)->tsk->prio) { - p = &parent->rb_right; - first = false; - } else { - p = &parent->rb_left; - } - } else if (i915_seqno_passed(wait->seqno, - to_wait(parent)->seqno)) { - p = &parent->rb_right; - if (i915_seqno_passed(seqno, to_wait(parent)->seqno)) - completed = parent; - else - first = false; - } else { - p = &parent->rb_left; - } - } - rb_link_node(&wait->node, parent, p); - rb_insert_color(&wait->node, &b->waiters); - - if (first) { - spin_lock(&b->irq_lock); - b->irq_wait = wait; - /* After assigning ourselves as the new bottom-half, we must - * perform a cursory check to prevent a missed interrupt. - * Either we miss the interrupt whilst programming the hardware, - * or if there was a previous waiter (for a later seqno) they - * may be woken instead of us (due to the inherent race - * in the unlocked read of b->irq_seqno_bh in the irq handler) - * and so we miss the wake up. - */ - armed = __intel_breadcrumbs_enable_irq(b); - spin_unlock(&b->irq_lock); - } - - if (completed) { - /* Advance the bottom-half (b->irq_wait) before we wake up - * the waiters who may scribble over their intel_wait - * just as the interrupt handler is dereferencing it via - * b->irq_wait. - */ - if (!first) { - struct rb_node *next = rb_next(completed); - GEM_BUG_ON(next == &wait->node); - __intel_breadcrumbs_next(engine, next); - } - do { - struct intel_wait *crumb = to_wait(completed); - completed = rb_prev(completed); - __intel_breadcrumbs_finish(b, crumb); - } while (completed); - } - - GEM_BUG_ON(!b->irq_wait); - GEM_BUG_ON(!b->irq_armed); - GEM_BUG_ON(rb_first(&b->waiters) != &b->irq_wait->node); + spin_lock_init(&b->irq_lock); + INIT_LIST_HEAD(&b->signalers); - return armed; + init_irq_work(&b->irq_work, signal_irq_work); } -bool intel_engine_add_wait(struct intel_engine_cs *engine, - struct intel_wait *wait) +void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine) { struct intel_breadcrumbs *b = &engine->breadcrumbs; - bool armed; - - spin_lock_irq(&b->rb_lock); - armed = __intel_engine_add_wait(engine, wait); - spin_unlock_irq(&b->rb_lock); - if (armed) - return armed; - - /* Make the caller recheck if its request has already started. */ - return intel_engine_has_started(engine, wait->seqno); -} + unsigned long flags; -static inline bool chain_wakeup(struct rb_node *rb, int priority) -{ - return rb && to_wait(rb)->tsk->prio <= priority; -} + spin_lock_irqsave(&b->irq_lock, flags); -static inline int wakeup_priority(struct intel_breadcrumbs *b, - struct task_struct *tsk) -{ - if (tsk == b->signaler) - return INT_MIN; + if (b->irq_enabled) + irq_enable(engine); else - return tsk->prio; -} - -static void __intel_engine_remove_wait(struct intel_engine_cs *engine, - struct intel_wait *wait) -{ - struct intel_breadcrumbs *b = &engine->breadcrumbs; - - lockdep_assert_held(&b->rb_lock); - - if (RB_EMPTY_NODE(&wait->node)) - goto out; - - if (b->irq_wait == wait) { - const int priority = wakeup_priority(b, wait->tsk); - struct rb_node *next; - - /* We are the current bottom-half. Find the next candidate, - * the first waiter in the queue on the remaining oldest - * request. As multiple seqnos may complete in the time it - * takes us to wake up and find the next waiter, we have to - * wake up that waiter for it to perform its own coherent - * completion check. - */ - next = rb_next(&wait->node); - if (chain_wakeup(next, priority)) { - /* If the next waiter is already complete, - * wake it up and continue onto the next waiter. So - * if have a small herd, they will wake up in parallel - * rather than sequentially, which should reduce - * the overall latency in waking all the completed - * clients. - * - * However, waking up a chain adds extra latency to - * the first_waiter. This is undesirable if that - * waiter is a high priority task. - */ - u32 seqno = intel_engine_get_seqno(engine); - - while (i915_seqno_passed(seqno, to_wait(next)->seqno)) { - struct rb_node *n = rb_next(next); - - __intel_breadcrumbs_finish(b, to_wait(next)); - next = n; - if (!chain_wakeup(next, priority)) - break; - } - } - - __intel_breadcrumbs_next(engine, next); - } else { - GEM_BUG_ON(rb_first(&b->waiters) == &wait->node); - } - - GEM_BUG_ON(RB_EMPTY_NODE(&wait->node)); - rb_erase(&wait->node, &b->waiters); - RB_CLEAR_NODE(&wait->node); + irq_disable(engine); -out: - GEM_BUG_ON(b->irq_wait == wait); - GEM_BUG_ON(rb_first(&b->waiters) != - (b->irq_wait ? &b->irq_wait->node : NULL)); + spin_unlock_irqrestore(&b->irq_lock, flags); } -void intel_engine_remove_wait(struct intel_engine_cs *engine, - struct intel_wait *wait) +void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine) { - struct intel_breadcrumbs *b = &engine->breadcrumbs; - - /* Quick check to see if this waiter was already decoupled from - * the tree by the bottom-half to avoid contention on the spinlock - * by the herd. - */ - if (RB_EMPTY_NODE(&wait->node)) { - GEM_BUG_ON(READ_ONCE(b->irq_wait) == wait); - return; - } - - spin_lock_irq(&b->rb_lock); - __intel_engine_remove_wait(engine, wait); - spin_unlock_irq(&b->rb_lock); } -static void signaler_set_rtpriority(void) +bool i915_request_enable_breadcrumb(struct i915_request *rq) { - struct sched_param param = { .sched_priority = 1 }; - - sched_setscheduler_nocheck(current, SCHED_FIFO, ¶m); -} + struct intel_breadcrumbs *b = &rq->engine->breadcrumbs; -static int intel_breadcrumbs_signaler(void *arg) -{ - struct intel_engine_cs *engine = arg; - struct intel_breadcrumbs *b = &engine->breadcrumbs; - struct i915_request *rq, *n; + GEM_BUG_ON(test_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags)); - /* Install ourselves with high priority to reduce signalling latency */ - signaler_set_rtpriority(); + if (!test_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags)) + return true; - do { - bool do_schedule = true; - LIST_HEAD(list); - u32 seqno; + spin_lock(&b->irq_lock); + if (test_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags) && + !__request_completed(rq)) { + struct intel_context *ce = rq->hw_context; + struct list_head *pos; - set_current_state(TASK_INTERRUPTIBLE); - if (list_empty(&b->signals)) - goto sleep; + __intel_breadcrumbs_arm_irq(b); /* - * We are either woken up by the interrupt bottom-half, - * or by a client adding a new signaller. In both cases, - * the GPU seqno may have advanced beyond our oldest signal. - * If it has, propagate the signal, remove the waiter and - * check again with the next oldest signal. Otherwise we - * need to wait for a new interrupt from the GPU or for - * a new client. + * We keep the seqno in retirement order, so we can break + * inside intel_engine_breadcrumbs_irq as soon as we've passed + * the last completed request (or seen a request that hasn't + * event started). We could iterate the timeline->requests list, + * but keeping a separate signalers_list has the advantage of + * hopefully being much smaller than the full list and so + * provides faster iteration and detection when there are no + * more interrupts required for this context. + * + * We typically expect to add new signalers in order, so we + * start looking for our insertion point from the tail of + * the list. */ - seqno = intel_engine_get_seqno(engine); + list_for_each_prev(pos, &ce->signals) { + struct i915_request *it = + list_entry(pos, typeof(*it), signal_link); - spin_lock_irq(&b->rb_lock); - list_for_each_entry_safe(rq, n, &b->signals, signaling.link) { - u32 this = rq->signaling.wait.seqno; - - GEM_BUG_ON(!rq->signaling.wait.seqno); - - if (!i915_seqno_passed(seqno, this)) + if (i915_seqno_passed(rq->fence.seqno, it->fence.seqno)) break; - - if (likely(this == i915_request_global_seqno(rq))) { - __intel_engine_remove_wait(engine, - &rq->signaling.wait); - - rq->signaling.wait.seqno = 0; - __list_del_entry(&rq->signaling.link); - - if (!test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, - &rq->fence.flags)) { - list_add_tail(&rq->signaling.link, - &list); - i915_request_get(rq); - } - } } - spin_unlock_irq(&b->rb_lock); - - if (!list_empty(&list)) { - local_bh_disable(); - list_for_each_entry_safe(rq, n, &list, signaling.link) { - dma_fence_signal(&rq->fence); - GEM_BUG_ON(!i915_request_completed(rq)); - i915_request_put(rq); - } - local_bh_enable(); /* kick start the tasklets */ - - /* - * If the engine is saturated we may be continually - * processing completed requests. This angers the - * NMI watchdog if we never let anything else - * have access to the CPU. Let's pretend to be nice - * and relinquish the CPU if we burn through the - * entire RT timeslice! - */ - do_schedule = need_resched(); - } - - if (unlikely(do_schedule)) { -sleep: - if (kthread_should_park()) - kthread_parkme(); - - if (unlikely(kthread_should_stop())) - break; + list_add(&rq->signal_link, pos); + if (pos == &ce->signals) /* catch transitions from empty list */ + list_move_tail(&ce->signal_link, &b->signalers); - schedule(); - } - } while (1); - __set_current_state(TASK_RUNNING); - - return 0; -} - -static void insert_signal(struct intel_breadcrumbs *b, - struct i915_request *request, - const u32 seqno) -{ - struct i915_request *iter; - - lockdep_assert_held(&b->rb_lock); - - /* - * A reasonable assumption is that we are called to add signals - * in sequence, as the requests are submitted for execution and - * assigned a global_seqno. This will be the case for the majority - * of internally generated signals (inter-engine signaling). - * - * Out of order waiters triggering random signaling enabling will - * be more problematic, but hopefully rare enough and the list - * small enough that the O(N) insertion sort is not an issue. - */ - - list_for_each_entry_reverse(iter, &b->signals, signaling.link) - if (i915_seqno_passed(seqno, iter->signaling.wait.seqno)) - break; - - list_add(&request->signaling.link, &iter->signaling.link); -} - -bool intel_engine_enable_signaling(struct i915_request *request, bool wakeup) -{ - struct intel_engine_cs *engine = request->engine; - struct intel_breadcrumbs *b = &engine->breadcrumbs; - struct intel_wait *wait = &request->signaling.wait; - u32 seqno; - - /* - * Note that we may be called from an interrupt handler on another - * device (e.g. nouveau signaling a fence completion causing us - * to submit a request, and so enable signaling). As such, - * we need to make sure that all other users of b->rb_lock protect - * against interrupts, i.e. use spin_lock_irqsave. - */ - - /* locked by dma_fence_enable_sw_signaling() (irqsafe fence->lock) */ - GEM_BUG_ON(!irqs_disabled()); - lockdep_assert_held(&request->lock); - - seqno = i915_request_global_seqno(request); - if (!seqno) /* will be enabled later upon execution */ - return true; - - GEM_BUG_ON(wait->seqno); - wait->tsk = b->signaler; - wait->request = request; - wait->seqno = seqno; - - /* - * Add ourselves into the list of waiters, but registering our - * bottom-half as the signaller thread. As per usual, only the oldest - * waiter (not just signaller) is tasked as the bottom-half waking - * up all completed waiters after the user interrupt. - * - * If we are the oldest waiter, enable the irq (after which we - * must double check that the seqno did not complete). - */ - spin_lock(&b->rb_lock); - insert_signal(b, request, seqno); - wakeup &= __intel_engine_add_wait(engine, wait); - spin_unlock(&b->rb_lock); - - if (wakeup) { - wake_up_process(b->signaler); - return !intel_wait_complete(wait); + set_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags); } + spin_unlock(&b->irq_lock); - return true; + return !__request_completed(rq); } -void intel_engine_cancel_signaling(struct i915_request *request) +void i915_request_cancel_breadcrumb(struct i915_request *rq) { - struct intel_engine_cs *engine = request->engine; - struct intel_breadcrumbs *b = &engine->breadcrumbs; + struct intel_breadcrumbs *b = &rq->engine->breadcrumbs; - GEM_BUG_ON(!irqs_disabled()); - lockdep_assert_held(&request->lock); - - if (!READ_ONCE(request->signaling.wait.seqno)) + if (!test_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags)) return; - spin_lock(&b->rb_lock); - __intel_engine_remove_wait(engine, &request->signaling.wait); - if (fetch_and_zero(&request->signaling.wait.seqno)) - __list_del_entry(&request->signaling.link); - spin_unlock(&b->rb_lock); -} - -int intel_engine_init_breadcrumbs(struct intel_engine_cs *engine) -{ - struct intel_breadcrumbs *b = &engine->breadcrumbs; - struct task_struct *tsk; - - spin_lock_init(&b->rb_lock); - spin_lock_init(&b->irq_lock); - - timer_setup(&b->fake_irq, intel_breadcrumbs_fake_irq, 0); - timer_setup(&b->hangcheck, intel_breadcrumbs_hangcheck, 0); - - INIT_LIST_HEAD(&b->signals); - - /* Spawn a thread to provide a common bottom-half for all signals. - * As this is an asynchronous interface we cannot steal the current - * task for handling the bottom-half to the user interrupt, therefore - * we create a thread to do the coherent seqno dance after the - * interrupt and then signal the waitqueue (via the dma-buf/fence). - */ - tsk = kthread_run(intel_breadcrumbs_signaler, engine, - "i915/signal:%d", engine->id); - if (IS_ERR(tsk)) - return PTR_ERR(tsk); - - b->signaler = tsk; - - return 0; -} - -static void cancel_fake_irq(struct intel_engine_cs *engine) -{ - struct intel_breadcrumbs *b = &engine->breadcrumbs; - - del_timer_sync(&b->fake_irq); /* may queue b->hangcheck */ - del_timer_sync(&b->hangcheck); - clear_bit(engine->id, &engine->i915->gpu_error.missed_irq_rings); -} - -void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine) -{ - struct intel_breadcrumbs *b = &engine->breadcrumbs; - unsigned long flags; - - spin_lock_irqsave(&b->irq_lock, flags); - - /* - * Leave the fake_irq timer enabled (if it is running), but clear the - * bit so that it turns itself off on its next wake up and goes back - * to the long hangcheck interval if still required. - */ - clear_bit(engine->id, &engine->i915->gpu_error.missed_irq_rings); + spin_lock(&b->irq_lock); + if (test_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags)) { + struct intel_context *ce = rq->hw_context; - if (b->irq_enabled) - irq_enable(engine); - else - irq_disable(engine); + list_del(&rq->signal_link); + if (list_empty(&ce->signals)) + list_del_init(&ce->signal_link); - spin_unlock_irqrestore(&b->irq_lock, flags); + clear_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags); + } + spin_unlock(&b->irq_lock); } -void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine) +void intel_engine_print_breadcrumbs(struct intel_engine_cs *engine, + struct drm_printer *p) { struct intel_breadcrumbs *b = &engine->breadcrumbs; + struct intel_context *ce; + struct i915_request *rq; - /* The engines should be idle and all requests accounted for! */ - WARN_ON(READ_ONCE(b->irq_wait)); - WARN_ON(!RB_EMPTY_ROOT(&b->waiters)); - WARN_ON(!list_empty(&b->signals)); + if (list_empty(&b->signalers)) + return; - if (!IS_ERR_OR_NULL(b->signaler)) - kthread_stop(b->signaler); + drm_printf(p, "Signals:\n"); - cancel_fake_irq(engine); + spin_lock_irq(&b->irq_lock); + list_for_each_entry(ce, &b->signalers, signal_link) { + list_for_each_entry(rq, &ce->signals, signal_link) { + drm_printf(p, "\t[%llx:%llx%s] @ %dms\n", + rq->fence.context, rq->fence.seqno, + i915_request_completed(rq) ? "!" : + i915_request_started(rq) ? "*" : + "", + jiffies_to_msecs(jiffies - rq->emitted_jiffies)); + } + } + spin_unlock_irq(&b->irq_lock); } - -#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) -#include "selftests/intel_breadcrumbs.c" -#endif diff --git a/drivers/gpu/drm/i915/intel_cdclk.c b/drivers/gpu/drm/i915/intel_cdclk.c index 2021e484a287..15ba950dee00 100644 --- a/drivers/gpu/drm/i915/intel_cdclk.c +++ b/drivers/gpu/drm/i915/intel_cdclk.c @@ -218,7 +218,7 @@ static unsigned int intel_hpll_vco(struct drm_i915_private *dev_priv) }; const unsigned int *vco_table; unsigned int vco; - uint8_t tmp = 0; + u8 tmp = 0; /* FIXME other chipsets? */ if (IS_GM45(dev_priv)) @@ -249,13 +249,13 @@ static void g33_get_cdclk(struct drm_i915_private *dev_priv, struct intel_cdclk_state *cdclk_state) { struct pci_dev *pdev = dev_priv->drm.pdev; - static const uint8_t div_3200[] = { 12, 10, 8, 7, 5, 16 }; - static const uint8_t div_4000[] = { 14, 12, 10, 8, 6, 20 }; - static const uint8_t div_4800[] = { 20, 14, 12, 10, 8, 24 }; - static const uint8_t div_5333[] = { 20, 16, 12, 12, 8, 28 }; - const uint8_t *div_table; + static const u8 div_3200[] = { 12, 10, 8, 7, 5, 16 }; + static const u8 div_4000[] = { 14, 12, 10, 8, 6, 20 }; + static const u8 div_4800[] = { 20, 14, 12, 10, 8, 24 }; + static const u8 div_5333[] = { 20, 16, 12, 12, 8, 28 }; + const u8 *div_table; unsigned int cdclk_sel; - uint16_t tmp = 0; + u16 tmp = 0; cdclk_state->vco = intel_hpll_vco(dev_priv); @@ -330,12 +330,12 @@ static void i965gm_get_cdclk(struct drm_i915_private *dev_priv, struct intel_cdclk_state *cdclk_state) { struct pci_dev *pdev = dev_priv->drm.pdev; - static const uint8_t div_3200[] = { 16, 10, 8 }; - static const uint8_t div_4000[] = { 20, 12, 10 }; - static const uint8_t div_5333[] = { 24, 16, 14 }; - const uint8_t *div_table; + static const u8 div_3200[] = { 16, 10, 8 }; + static const u8 div_4000[] = { 20, 12, 10 }; + static const u8 div_5333[] = { 24, 16, 14 }; + const u8 *div_table; unsigned int cdclk_sel; - uint16_t tmp = 0; + u16 tmp = 0; cdclk_state->vco = intel_hpll_vco(dev_priv); @@ -375,7 +375,7 @@ static void gm45_get_cdclk(struct drm_i915_private *dev_priv, { struct pci_dev *pdev = dev_priv->drm.pdev; unsigned int cdclk_sel; - uint16_t tmp = 0; + u16 tmp = 0; cdclk_state->vco = intel_hpll_vco(dev_priv); @@ -403,8 +403,8 @@ static void gm45_get_cdclk(struct drm_i915_private *dev_priv, static void hsw_get_cdclk(struct drm_i915_private *dev_priv, struct intel_cdclk_state *cdclk_state) { - uint32_t lcpll = I915_READ(LCPLL_CTL); - uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK; + u32 lcpll = I915_READ(LCPLL_CTL); + u32 freq = lcpll & LCPLL_CLK_FREQ_MASK; if (lcpll & LCPLL_CD_SOURCE_FCLK) cdclk_state->cdclk = 800000; @@ -520,6 +520,7 @@ static void vlv_set_cdclk(struct drm_i915_private *dev_priv, { int cdclk = cdclk_state->cdclk; u32 val, cmd = cdclk_state->voltage_level; + intel_wakeref_t wakeref; switch (cdclk) { case 400000: @@ -539,7 +540,7 @@ static void vlv_set_cdclk(struct drm_i915_private *dev_priv, * a system suspend. So grab the PIPE-A domain, which covers * the HW blocks needed for the following programming. */ - intel_display_power_get(dev_priv, POWER_DOMAIN_PIPE_A); + wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_PIPE_A); mutex_lock(&dev_priv->pcu_lock); val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ); @@ -593,7 +594,7 @@ static void vlv_set_cdclk(struct drm_i915_private *dev_priv, vlv_program_pfi_credits(dev_priv); - intel_display_power_put(dev_priv, POWER_DOMAIN_PIPE_A); + intel_display_power_put(dev_priv, POWER_DOMAIN_PIPE_A, wakeref); } static void chv_set_cdclk(struct drm_i915_private *dev_priv, @@ -601,6 +602,7 @@ static void chv_set_cdclk(struct drm_i915_private *dev_priv, { int cdclk = cdclk_state->cdclk; u32 val, cmd = cdclk_state->voltage_level; + intel_wakeref_t wakeref; switch (cdclk) { case 333333: @@ -619,7 +621,7 @@ static void chv_set_cdclk(struct drm_i915_private *dev_priv, * a system suspend. So grab the PIPE-A domain, which covers * the HW blocks needed for the following programming. */ - intel_display_power_get(dev_priv, POWER_DOMAIN_PIPE_A); + wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_PIPE_A); mutex_lock(&dev_priv->pcu_lock); val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ); @@ -637,7 +639,7 @@ static void chv_set_cdclk(struct drm_i915_private *dev_priv, vlv_program_pfi_credits(dev_priv); - intel_display_power_put(dev_priv, POWER_DOMAIN_PIPE_A); + intel_display_power_put(dev_priv, POWER_DOMAIN_PIPE_A, wakeref); } static int bdw_calc_cdclk(int min_cdclk) @@ -670,8 +672,8 @@ static u8 bdw_calc_voltage_level(int cdclk) static void bdw_get_cdclk(struct drm_i915_private *dev_priv, struct intel_cdclk_state *cdclk_state) { - uint32_t lcpll = I915_READ(LCPLL_CTL); - uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK; + u32 lcpll = I915_READ(LCPLL_CTL); + u32 freq = lcpll & LCPLL_CLK_FREQ_MASK; if (lcpll & LCPLL_CD_SOURCE_FCLK) cdclk_state->cdclk = 800000; @@ -698,7 +700,7 @@ static void bdw_set_cdclk(struct drm_i915_private *dev_priv, const struct intel_cdclk_state *cdclk_state) { int cdclk = cdclk_state->cdclk; - uint32_t val; + u32 val; int ret; if (WARN((I915_READ(LCPLL_CTL) & @@ -1081,7 +1083,7 @@ static void skl_set_cdclk(struct drm_i915_private *dev_priv, static void skl_sanitize_cdclk(struct drm_i915_private *dev_priv) { - uint32_t cdctl, expected; + u32 cdctl, expected; /* * check if the pre-os initialized the display @@ -2688,7 +2690,7 @@ static int vlv_hrawclk(struct drm_i915_private *dev_priv) static int g4x_hrawclk(struct drm_i915_private *dev_priv) { - uint32_t clkcfg; + u32 clkcfg; /* hrawclock is 1/4 the FSB frequency */ clkcfg = I915_READ(CLKCFG); diff --git a/drivers/gpu/drm/i915/intel_color.c b/drivers/gpu/drm/i915/intel_color.c index 37fd9ddf762e..71a1f12c6b2a 100644 --- a/drivers/gpu/drm/i915/intel_color.c +++ b/drivers/gpu/drm/i915/intel_color.c @@ -74,12 +74,12 @@ #define ILK_CSC_COEFF_1_0 \ ((7 << 12) | ILK_CSC_COEFF_FP(CTM_COEFF_1_0, 8)) -static bool lut_is_legacy(struct drm_property_blob *lut) +static bool lut_is_legacy(const struct drm_property_blob *lut) { return drm_color_lut_size(lut) == LEGACY_LUT_LENGTH; } -static bool crtc_state_is_legacy_gamma(struct intel_crtc_state *crtc_state) +static bool crtc_state_is_legacy_gamma(const struct intel_crtc_state *crtc_state) { return !crtc_state->base.degamma_lut && !crtc_state->base.ctm && @@ -115,8 +115,8 @@ static u64 *ctm_mult_by_limited(u64 *result, const u64 *input) static void ilk_load_ycbcr_conversion_matrix(struct intel_crtc *crtc) { - int pipe = crtc->pipe; struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + enum pipe pipe = crtc->pipe; I915_WRITE(PIPE_CSC_PREOFF_HI(pipe), 0); I915_WRITE(PIPE_CSC_PREOFF_ME(pipe), 0); @@ -137,13 +137,14 @@ static void ilk_load_ycbcr_conversion_matrix(struct intel_crtc *crtc) I915_WRITE(PIPE_CSC_MODE(pipe), 0); } -static void ilk_load_csc_matrix(struct intel_crtc_state *crtc_state) +static void ilk_load_csc_matrix(const struct intel_crtc_state *crtc_state) { struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - int i, pipe = crtc->pipe; - uint16_t coeffs[9] = { 0, }; bool limited_color_range = false; + enum pipe pipe = crtc->pipe; + u16 coeffs[9] = {}; + int i; /* * FIXME if there's a gamma LUT after the CSC, we should @@ -171,7 +172,7 @@ static void ilk_load_csc_matrix(struct intel_crtc_state *crtc_state) * hardware. */ for (i = 0; i < ARRAY_SIZE(coeffs); i++) { - uint64_t abs_coeff = ((1ULL << 63) - 1) & input[i]; + u64 abs_coeff = ((1ULL << 63) - 1) & input[i]; /* * Clamp input value to min/max supported by @@ -233,7 +234,7 @@ static void ilk_load_csc_matrix(struct intel_crtc_state *crtc_state) I915_WRITE(PIPE_CSC_PREOFF_LO(pipe), 0); if (INTEL_GEN(dev_priv) > 6) { - uint16_t postoff = 0; + u16 postoff = 0; if (limited_color_range) postoff = (16 * (1 << 12) / 255) & 0x1fff; @@ -244,7 +245,7 @@ static void ilk_load_csc_matrix(struct intel_crtc_state *crtc_state) I915_WRITE(PIPE_CSC_MODE(pipe), 0); } else { - uint32_t mode = CSC_MODE_YUV_TO_RGB; + u32 mode = CSC_MODE_YUV_TO_RGB; if (limited_color_range) mode |= CSC_BLACK_SCREEN_OFFSET; @@ -256,20 +257,20 @@ static void ilk_load_csc_matrix(struct intel_crtc_state *crtc_state) /* * Set up the pipe CSC unit on CherryView. */ -static void cherryview_load_csc_matrix(struct intel_crtc_state *crtc_state) +static void cherryview_load_csc_matrix(const struct intel_crtc_state *crtc_state) { - struct drm_device *dev = crtc_state->base.crtc->dev; - struct drm_i915_private *dev_priv = to_i915(dev); - int pipe = to_intel_crtc(crtc_state->base.crtc)->pipe; - uint32_t mode; + struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + enum pipe pipe = crtc->pipe; + u32 mode; if (crtc_state->base.ctm) { - struct drm_color_ctm *ctm = crtc_state->base.ctm->data; - uint16_t coeffs[9] = { 0, }; + const struct drm_color_ctm *ctm = crtc_state->base.ctm->data; + u16 coeffs[9] = {}; int i; for (i = 0; i < ARRAY_SIZE(coeffs); i++) { - uint64_t abs_coeff = + u64 abs_coeff = ((1ULL << 63) - 1) & ctm->matrix[i]; /* Round coefficient. */ @@ -303,25 +304,16 @@ static void cherryview_load_csc_matrix(struct intel_crtc_state *crtc_state) I915_WRITE(CGM_PIPE_MODE(pipe), mode); } -void intel_color_set_csc(struct intel_crtc_state *crtc_state) -{ - struct drm_device *dev = crtc_state->base.crtc->dev; - struct drm_i915_private *dev_priv = to_i915(dev); - - if (dev_priv->display.load_csc_matrix) - dev_priv->display.load_csc_matrix(crtc_state); -} - /* Loads the legacy palette/gamma unit for the CRTC. */ -static void i9xx_load_luts_internal(struct intel_crtc_state *crtc_state, - struct drm_property_blob *blob) +static void i9xx_load_luts_internal(const struct intel_crtc_state *crtc_state, + const struct drm_property_blob *blob) { struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; int i; - if (HAS_GMCH_DISPLAY(dev_priv)) { + if (HAS_GMCH(dev_priv)) { if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DSI)) assert_dsi_pll_enabled(dev_priv); else @@ -329,23 +321,24 @@ static void i9xx_load_luts_internal(struct intel_crtc_state *crtc_state, } if (blob) { - struct drm_color_lut *lut = blob->data; + const struct drm_color_lut *lut = blob->data; + for (i = 0; i < 256; i++) { - uint32_t word = + u32 word = (drm_color_lut_extract(lut[i].red, 8) << 16) | (drm_color_lut_extract(lut[i].green, 8) << 8) | drm_color_lut_extract(lut[i].blue, 8); - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) I915_WRITE(PALETTE(pipe, i), word); else I915_WRITE(LGC_PALETTE(pipe, i), word); } } else { for (i = 0; i < 256; i++) { - uint32_t word = (i << 16) | (i << 8) | i; + u32 word = (i << 16) | (i << 8) | i; - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) I915_WRITE(PALETTE(pipe, i), word); else I915_WRITE(LGC_PALETTE(pipe, i), word); @@ -353,51 +346,37 @@ static void i9xx_load_luts_internal(struct intel_crtc_state *crtc_state, } } -static void i9xx_load_luts(struct intel_crtc_state *crtc_state) +static void i9xx_load_luts(const struct intel_crtc_state *crtc_state) { i9xx_load_luts_internal(crtc_state, crtc_state->base.gamma_lut); } -/* Loads the legacy palette/gamma unit for the CRTC on Haswell. */ -static void haswell_load_luts(struct intel_crtc_state *crtc_state) +static void hsw_color_commit(const struct intel_crtc_state *crtc_state) { struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - bool reenable_ips = false; - - /* - * Workaround : Do not read or write the pipe palette/gamma data while - * GAMMA_MODE is configured for split gamma and IPS_CTL has IPS enabled. - */ - if (IS_HASWELL(dev_priv) && crtc_state->ips_enabled && - (crtc_state->gamma_mode == GAMMA_MODE_MODE_SPLIT)) { - hsw_disable_ips(crtc_state); - reenable_ips = true; - } - - crtc_state->gamma_mode = GAMMA_MODE_MODE_8BIT; - I915_WRITE(GAMMA_MODE(crtc->pipe), GAMMA_MODE_MODE_8BIT); - i9xx_load_luts(crtc_state); + I915_WRITE(GAMMA_MODE(crtc->pipe), crtc_state->gamma_mode); - if (reenable_ips) - hsw_enable_ips(crtc_state); + ilk_load_csc_matrix(crtc_state); } -static void bdw_load_degamma_lut(struct intel_crtc_state *crtc_state) +static void bdw_load_degamma_lut(const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); - enum pipe pipe = to_intel_crtc(crtc_state->base.crtc)->pipe; - uint32_t i, lut_size = INTEL_INFO(dev_priv)->color.degamma_lut_size; + struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + const struct drm_property_blob *degamma_lut = crtc_state->base.degamma_lut; + u32 i, lut_size = INTEL_INFO(dev_priv)->color.degamma_lut_size; + enum pipe pipe = crtc->pipe; I915_WRITE(PREC_PAL_INDEX(pipe), PAL_PREC_SPLIT_MODE | PAL_PREC_AUTO_INCREMENT); - if (crtc_state->base.degamma_lut) { - struct drm_color_lut *lut = crtc_state->base.degamma_lut->data; + if (degamma_lut) { + const struct drm_color_lut *lut = degamma_lut->data; for (i = 0; i < lut_size; i++) { - uint32_t word = + u32 word = drm_color_lut_extract(lut[i].red, 10) << 20 | drm_color_lut_extract(lut[i].green, 10) << 10 | drm_color_lut_extract(lut[i].blue, 10); @@ -406,7 +385,7 @@ static void bdw_load_degamma_lut(struct intel_crtc_state *crtc_state) } } else { for (i = 0; i < lut_size; i++) { - uint32_t v = (i * ((1 << 10) - 1)) / (lut_size - 1); + u32 v = (i * ((1 << 10) - 1)) / (lut_size - 1); I915_WRITE(PREC_PAL_DATA(pipe), (v << 20) | (v << 10) | v); @@ -414,11 +393,13 @@ static void bdw_load_degamma_lut(struct intel_crtc_state *crtc_state) } } -static void bdw_load_gamma_lut(struct intel_crtc_state *crtc_state, u32 offset) +static void bdw_load_gamma_lut(const struct intel_crtc_state *crtc_state, u32 offset) { - struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); - enum pipe pipe = to_intel_crtc(crtc_state->base.crtc)->pipe; - uint32_t i, lut_size = INTEL_INFO(dev_priv)->color.gamma_lut_size; + struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + const struct drm_property_blob *gamma_lut = crtc_state->base.gamma_lut; + u32 i, lut_size = INTEL_INFO(dev_priv)->color.gamma_lut_size; + enum pipe pipe = crtc->pipe; WARN_ON(offset & ~PAL_PREC_INDEX_VALUE_MASK); @@ -427,11 +408,11 @@ static void bdw_load_gamma_lut(struct intel_crtc_state *crtc_state, u32 offset) PAL_PREC_AUTO_INCREMENT | offset); - if (crtc_state->base.gamma_lut) { - struct drm_color_lut *lut = crtc_state->base.gamma_lut->data; + if (gamma_lut) { + const struct drm_color_lut *lut = gamma_lut->data; for (i = 0; i < lut_size; i++) { - uint32_t word = + u32 word = (drm_color_lut_extract(lut[i].red, 10) << 20) | (drm_color_lut_extract(lut[i].green, 10) << 10) | drm_color_lut_extract(lut[i].blue, 10); @@ -449,7 +430,7 @@ static void bdw_load_gamma_lut(struct intel_crtc_state *crtc_state, u32 offset) drm_color_lut_extract(lut[i].blue, 16)); } else { for (i = 0; i < lut_size; i++) { - uint32_t v = (i * ((1 << 10) - 1)) / (lut_size - 1); + u32 v = (i * ((1 << 10) - 1)) / (lut_size - 1); I915_WRITE(PREC_PAL_DATA(pipe), (v << 20) | (v << 10) | v); @@ -462,37 +443,34 @@ static void bdw_load_gamma_lut(struct intel_crtc_state *crtc_state, u32 offset) } /* Loads the palette/gamma unit for the CRTC on Broadwell+. */ -static void broadwell_load_luts(struct intel_crtc_state *crtc_state) +static void broadwell_load_luts(const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); - enum pipe pipe = to_intel_crtc(crtc_state->base.crtc)->pipe; + struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + enum pipe pipe = crtc->pipe; if (crtc_state_is_legacy_gamma(crtc_state)) { - haswell_load_luts(crtc_state); - return; - } - - bdw_load_degamma_lut(crtc_state); - bdw_load_gamma_lut(crtc_state, - INTEL_INFO(dev_priv)->color.degamma_lut_size); - - crtc_state->gamma_mode = GAMMA_MODE_MODE_SPLIT; - I915_WRITE(GAMMA_MODE(pipe), GAMMA_MODE_MODE_SPLIT); - POSTING_READ(GAMMA_MODE(pipe)); + i9xx_load_luts(crtc_state); + } else { + bdw_load_degamma_lut(crtc_state); + bdw_load_gamma_lut(crtc_state, + INTEL_INFO(dev_priv)->color.degamma_lut_size); - /* - * Reset the index, otherwise it prevents the legacy palette to be - * written properly. - */ - I915_WRITE(PREC_PAL_INDEX(pipe), 0); + /* + * Reset the index, otherwise it prevents the legacy palette to be + * written properly. + */ + I915_WRITE(PREC_PAL_INDEX(pipe), 0); + } } -static void glk_load_degamma_lut(struct intel_crtc_state *crtc_state) +static void glk_load_degamma_lut(const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); - enum pipe pipe = to_intel_crtc(crtc_state->base.crtc)->pipe; - const uint32_t lut_size = 33; - uint32_t i; + struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + enum pipe pipe = crtc->pipe; + const u32 lut_size = 33; + u32 i; /* * When setting the auto-increment bit, the hardware seems to @@ -507,7 +485,7 @@ static void glk_load_degamma_lut(struct intel_crtc_state *crtc_state) * different values per channel, so this just loads a linear table. */ for (i = 0; i < lut_size; i++) { - uint32_t v = (i * (1 << 16)) / (lut_size - 1); + u32 v = (i * (1 << 16)) / (lut_size - 1); I915_WRITE(PRE_CSC_GAMC_DATA(pipe), v); } @@ -517,48 +495,49 @@ static void glk_load_degamma_lut(struct intel_crtc_state *crtc_state) I915_WRITE(PRE_CSC_GAMC_DATA(pipe), (1 << 16)); } -static void glk_load_luts(struct intel_crtc_state *crtc_state) +static void glk_load_luts(const struct intel_crtc_state *crtc_state) { - struct drm_device *dev = crtc_state->base.crtc->dev; - struct drm_i915_private *dev_priv = to_i915(dev); - enum pipe pipe = to_intel_crtc(crtc_state->base.crtc)->pipe; + struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + enum pipe pipe = crtc->pipe; glk_load_degamma_lut(crtc_state); if (crtc_state_is_legacy_gamma(crtc_state)) { - haswell_load_luts(crtc_state); - return; - } - - bdw_load_gamma_lut(crtc_state, 0); + i9xx_load_luts(crtc_state); + } else { + bdw_load_gamma_lut(crtc_state, 0); - crtc_state->gamma_mode = GAMMA_MODE_MODE_10BIT; - I915_WRITE(GAMMA_MODE(pipe), GAMMA_MODE_MODE_10BIT); - POSTING_READ(GAMMA_MODE(pipe)); + /* + * Reset the index, otherwise it prevents the legacy palette to be + * written properly. + */ + I915_WRITE(PREC_PAL_INDEX(pipe), 0); + } } -/* Loads the palette/gamma unit for the CRTC on CherryView. */ -static void cherryview_load_luts(struct intel_crtc_state *crtc_state) +static void cherryview_load_luts(const struct intel_crtc_state *crtc_state) { - struct drm_crtc *crtc = crtc_state->base.crtc; - struct drm_i915_private *dev_priv = to_i915(crtc->dev); - enum pipe pipe = to_intel_crtc(crtc)->pipe; - struct drm_color_lut *lut; - uint32_t i, lut_size; - uint32_t word0, word1; + struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + const struct drm_property_blob *gamma_lut = crtc_state->base.gamma_lut; + const struct drm_property_blob *degamma_lut = crtc_state->base.degamma_lut; + enum pipe pipe = crtc->pipe; + + cherryview_load_csc_matrix(crtc_state); if (crtc_state_is_legacy_gamma(crtc_state)) { - /* Turn off degamma/gamma on CGM block. */ - I915_WRITE(CGM_PIPE_MODE(pipe), - (crtc_state->base.ctm ? CGM_PIPE_MODE_CSC : 0)); - i9xx_load_luts_internal(crtc_state, crtc_state->base.gamma_lut); + i9xx_load_luts_internal(crtc_state, gamma_lut); return; } - if (crtc_state->base.degamma_lut) { - lut = crtc_state->base.degamma_lut->data; - lut_size = INTEL_INFO(dev_priv)->color.degamma_lut_size; + if (degamma_lut) { + const struct drm_color_lut *lut = degamma_lut->data; + int i, lut_size = INTEL_INFO(dev_priv)->color.degamma_lut_size; + for (i = 0; i < lut_size; i++) { + u32 word0, word1; + /* Write LUT in U0.14 format. */ word0 = (drm_color_lut_extract(lut[i].green, 14) << 16) | @@ -570,10 +549,13 @@ static void cherryview_load_luts(struct intel_crtc_state *crtc_state) } } - if (crtc_state->base.gamma_lut) { - lut = crtc_state->base.gamma_lut->data; - lut_size = INTEL_INFO(dev_priv)->color.gamma_lut_size; + if (gamma_lut) { + const struct drm_color_lut *lut = gamma_lut->data; + int i, lut_size = INTEL_INFO(dev_priv)->color.gamma_lut_size; + for (i = 0; i < lut_size; i++) { + u32 word0, word1; + /* Write LUT in U0.10 format. */ word0 = (drm_color_lut_extract(lut[i].green, 10) << 16) | @@ -585,11 +567,6 @@ static void cherryview_load_luts(struct intel_crtc_state *crtc_state) } } - I915_WRITE(CGM_PIPE_MODE(pipe), - (crtc_state->base.ctm ? CGM_PIPE_MODE_CSC : 0) | - (crtc_state->base.degamma_lut ? CGM_PIPE_MODE_DEGAMMA : 0) | - (crtc_state->base.gamma_lut ? CGM_PIPE_MODE_GAMMA : 0)); - /* * Also program a linear LUT in the legacy block (behind the * CGM block). @@ -597,40 +574,73 @@ static void cherryview_load_luts(struct intel_crtc_state *crtc_state) i9xx_load_luts_internal(crtc_state, NULL); } -void intel_color_load_luts(struct intel_crtc_state *crtc_state) +void intel_color_load_luts(const struct intel_crtc_state *crtc_state) { - struct drm_device *dev = crtc_state->base.crtc->dev; - struct drm_i915_private *dev_priv = to_i915(dev); + struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); dev_priv->display.load_luts(crtc_state); } +void intel_color_commit(const struct intel_crtc_state *crtc_state) +{ + struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); + + if (dev_priv->display.color_commit) + dev_priv->display.color_commit(crtc_state); +} + +static int check_lut_size(const struct drm_property_blob *lut, int expected) +{ + int len; + + if (!lut) + return 0; + + len = drm_color_lut_size(lut); + if (len != expected) { + DRM_DEBUG_KMS("Invalid LUT size; got %d, expected %d\n", + len, expected); + return -EINVAL; + } + + return 0; +} + int intel_color_check(struct intel_crtc_state *crtc_state) { struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); - size_t gamma_length, degamma_length; + const struct drm_property_blob *gamma_lut = crtc_state->base.gamma_lut; + const struct drm_property_blob *degamma_lut = crtc_state->base.degamma_lut; + int gamma_length, degamma_length; + u32 gamma_tests, degamma_tests; degamma_length = INTEL_INFO(dev_priv)->color.degamma_lut_size; gamma_length = INTEL_INFO(dev_priv)->color.gamma_lut_size; + degamma_tests = INTEL_INFO(dev_priv)->color.degamma_lut_tests; + gamma_tests = INTEL_INFO(dev_priv)->color.gamma_lut_tests; - /* - * We allow both degamma & gamma luts at the right size or - * NULL. - */ - if ((!crtc_state->base.degamma_lut || - drm_color_lut_size(crtc_state->base.degamma_lut) == degamma_length) && - (!crtc_state->base.gamma_lut || - drm_color_lut_size(crtc_state->base.gamma_lut) == gamma_length)) + /* Always allow legacy gamma LUT with no further checking. */ + if (crtc_state_is_legacy_gamma(crtc_state)) { + crtc_state->gamma_mode = GAMMA_MODE_MODE_8BIT; return 0; + } - /* - * We also allow no degamma lut/ctm and a gamma lut at the legacy - * size (256 entries). - */ - if (crtc_state_is_legacy_gamma(crtc_state)) - return 0; + if (check_lut_size(degamma_lut, degamma_length) || + check_lut_size(gamma_lut, gamma_length)) + return -EINVAL; + + if (drm_color_lut_check(degamma_lut, degamma_tests) || + drm_color_lut_check(gamma_lut, gamma_tests)) + return -EINVAL; - return -EINVAL; + if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv)) + crtc_state->gamma_mode = GAMMA_MODE_MODE_10BIT; + else if (INTEL_GEN(dev_priv) >= 9 || IS_BROADWELL(dev_priv)) + crtc_state->gamma_mode = GAMMA_MODE_MODE_SPLIT; + else + crtc_state->gamma_mode = GAMMA_MODE_MODE_8BIT; + + return 0; } void intel_color_init(struct intel_crtc *crtc) @@ -640,18 +650,17 @@ void intel_color_init(struct intel_crtc *crtc) drm_mode_crtc_set_gamma_size(&crtc->base, 256); if (IS_CHERRYVIEW(dev_priv)) { - dev_priv->display.load_csc_matrix = cherryview_load_csc_matrix; dev_priv->display.load_luts = cherryview_load_luts; } else if (IS_HASWELL(dev_priv)) { - dev_priv->display.load_csc_matrix = ilk_load_csc_matrix; - dev_priv->display.load_luts = haswell_load_luts; + dev_priv->display.load_luts = i9xx_load_luts; + dev_priv->display.color_commit = hsw_color_commit; } else if (IS_BROADWELL(dev_priv) || IS_GEN9_BC(dev_priv) || IS_BROXTON(dev_priv)) { - dev_priv->display.load_csc_matrix = ilk_load_csc_matrix; dev_priv->display.load_luts = broadwell_load_luts; + dev_priv->display.color_commit = hsw_color_commit; } else if (IS_GEMINILAKE(dev_priv) || IS_CANNONLAKE(dev_priv)) { - dev_priv->display.load_csc_matrix = ilk_load_csc_matrix; dev_priv->display.load_luts = glk_load_luts; + dev_priv->display.color_commit = hsw_color_commit; } else { dev_priv->display.load_luts = i9xx_load_luts; } diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c index d9fc1601671d..3716b2ee362f 100644 --- a/drivers/gpu/drm/i915/intel_crt.c +++ b/drivers/gpu/drm/i915/intel_crt.c @@ -83,15 +83,17 @@ static bool intel_crt_get_hw_state(struct intel_encoder *encoder, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_crt *crt = intel_encoder_to_crt(encoder); + intel_wakeref_t wakeref; bool ret; - if (!intel_display_power_get_if_enabled(dev_priv, - encoder->power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + encoder->power_domain); + if (!wakeref) return false; ret = intel_crt_port_enabled(dev_priv, crt->adpa_reg, pipe); - intel_display_power_put(dev_priv, encoder->power_domain); + intel_display_power_put(dev_priv, encoder->power_domain, wakeref); return ret; } @@ -629,19 +631,19 @@ static bool intel_crt_detect_ddc(struct drm_connector *connector) } static enum drm_connector_status -intel_crt_load_detect(struct intel_crt *crt, uint32_t pipe) +intel_crt_load_detect(struct intel_crt *crt, u32 pipe) { struct drm_device *dev = crt->base.base.dev; struct drm_i915_private *dev_priv = to_i915(dev); - uint32_t save_bclrpat; - uint32_t save_vtotal; - uint32_t vtotal, vactive; - uint32_t vsample; - uint32_t vblank, vblank_start, vblank_end; - uint32_t dsl; + u32 save_bclrpat; + u32 save_vtotal; + u32 vtotal, vactive; + u32 vsample; + u32 vblank, vblank_start, vblank_end; + u32 dsl; i915_reg_t bclrpat_reg, vtotal_reg, vblank_reg, vsync_reg, pipeconf_reg, pipe_dsl_reg; - uint8_t st00; + u8 st00; enum drm_connector_status status; DRM_DEBUG_KMS("starting load-detect on CRT\n"); @@ -667,7 +669,7 @@ intel_crt_load_detect(struct intel_crt *crt, uint32_t pipe) I915_WRITE(bclrpat_reg, 0x500050); if (!IS_GEN(dev_priv, 2)) { - uint32_t pipeconf = I915_READ(pipeconf_reg); + u32 pipeconf = I915_READ(pipeconf_reg); I915_WRITE(pipeconf_reg, pipeconf | PIPECONF_FORCE_BORDER); POSTING_READ(pipeconf_reg); /* Wait for next Vblank to substitue @@ -688,8 +690,8 @@ intel_crt_load_detect(struct intel_crt *crt, uint32_t pipe) * Yes, this will flicker */ if (vblank_start <= vactive && vblank_end >= vtotal) { - uint32_t vsync = I915_READ(vsync_reg); - uint32_t vsync_start = (vsync & 0xffff) + 1; + u32 vsync = I915_READ(vsync_reg); + u32 vsync_start = (vsync & 0xffff) + 1; vblank_start = vsync_start; I915_WRITE(vblank_reg, @@ -777,6 +779,7 @@ intel_crt_detect(struct drm_connector *connector, struct drm_i915_private *dev_priv = to_i915(connector->dev); struct intel_crt *crt = intel_attached_crt(connector); struct intel_encoder *intel_encoder = &crt->base; + intel_wakeref_t wakeref; int status, ret; struct intel_load_detect_pipe tmp; @@ -785,7 +788,8 @@ intel_crt_detect(struct drm_connector *connector, force); if (i915_modparams.load_detect_test) { - intel_display_power_get(dev_priv, intel_encoder->power_domain); + wakeref = intel_display_power_get(dev_priv, + intel_encoder->power_domain); goto load_detect; } @@ -793,7 +797,8 @@ intel_crt_detect(struct drm_connector *connector, if (dmi_check_system(intel_spurious_crt_detect)) return connector_status_disconnected; - intel_display_power_get(dev_priv, intel_encoder->power_domain); + wakeref = intel_display_power_get(dev_priv, + intel_encoder->power_domain); if (I915_HAS_HOTPLUG(dev_priv)) { /* We can not rely on the HPD pin always being correctly wired @@ -848,7 +853,7 @@ load_detect: } out: - intel_display_power_put(dev_priv, intel_encoder->power_domain); + intel_display_power_put(dev_priv, intel_encoder->power_domain, wakeref); return status; } @@ -858,10 +863,12 @@ static int intel_crt_get_modes(struct drm_connector *connector) struct drm_i915_private *dev_priv = to_i915(dev); struct intel_crt *crt = intel_attached_crt(connector); struct intel_encoder *intel_encoder = &crt->base; - int ret; + intel_wakeref_t wakeref; struct i2c_adapter *i2c; + int ret; - intel_display_power_get(dev_priv, intel_encoder->power_domain); + wakeref = intel_display_power_get(dev_priv, + intel_encoder->power_domain); i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->vbt.crt_ddc_pin); ret = intel_crt_ddc_get_modes(connector, i2c); @@ -873,7 +880,7 @@ static int intel_crt_get_modes(struct drm_connector *connector) ret = intel_crt_ddc_get_modes(connector, i2c); out: - intel_display_power_put(dev_priv, intel_encoder->power_domain); + intel_display_power_put(dev_priv, intel_encoder->power_domain, wakeref); return ret; } diff --git a/drivers/gpu/drm/i915/intel_csr.c b/drivers/gpu/drm/i915/intel_csr.c index a516697bf57d..e8ac04c33e29 100644 --- a/drivers/gpu/drm/i915/intel_csr.c +++ b/drivers/gpu/drm/i915/intel_csr.c @@ -70,50 +70,50 @@ MODULE_FIRMWARE(BXT_CSR_PATH); struct intel_css_header { /* 0x09 for DMC */ - uint32_t module_type; + u32 module_type; /* Includes the DMC specific header in dwords */ - uint32_t header_len; + u32 header_len; /* always value would be 0x10000 */ - uint32_t header_ver; + u32 header_ver; /* Not used */ - uint32_t module_id; + u32 module_id; /* Not used */ - uint32_t module_vendor; + u32 module_vendor; /* in YYYYMMDD format */ - uint32_t date; + u32 date; /* Size in dwords (CSS_Headerlen + PackageHeaderLen + dmc FWsLen)/4 */ - uint32_t size; + u32 size; /* Not used */ - uint32_t key_size; + u32 key_size; /* Not used */ - uint32_t modulus_size; + u32 modulus_size; /* Not used */ - uint32_t exponent_size; + u32 exponent_size; /* Not used */ - uint32_t reserved1[12]; + u32 reserved1[12]; /* Major Minor */ - uint32_t version; + u32 version; /* Not used */ - uint32_t reserved2[8]; + u32 reserved2[8]; /* Not used */ - uint32_t kernel_header_info; + u32 kernel_header_info; } __packed; struct intel_fw_info { - uint16_t reserved1; + u16 reserved1; /* Stepping (A, B, C, ..., *). * is a wildcard */ char stepping; @@ -121,8 +121,8 @@ struct intel_fw_info { /* Sub-stepping (0, 1, ..., *). * is a wildcard */ char substepping; - uint32_t offset; - uint32_t reserved2; + u32 offset; + u32 reserved2; } __packed; struct intel_package_header { @@ -135,14 +135,14 @@ struct intel_package_header { unsigned char reserved[10]; /* Number of valid entries in the FWInfo array below */ - uint32_t num_entries; + u32 num_entries; struct intel_fw_info fw_info[20]; } __packed; struct intel_dmc_header { /* always value would be 0x40403E3E */ - uint32_t signature; + u32 signature; /* DMC binary header length */ unsigned char header_len; @@ -151,30 +151,30 @@ struct intel_dmc_header { unsigned char header_ver; /* Reserved */ - uint16_t dmcc_ver; + u16 dmcc_ver; /* Major, Minor */ - uint32_t project; + u32 project; /* Firmware program size (excluding header) in dwords */ - uint32_t fw_size; + u32 fw_size; /* Major Minor version */ - uint32_t fw_version; + u32 fw_version; /* Number of valid MMIO cycles present. */ - uint32_t mmio_count; + u32 mmio_count; /* MMIO address */ - uint32_t mmioaddr[8]; + u32 mmioaddr[8]; /* MMIO data */ - uint32_t mmiodata[8]; + u32 mmiodata[8]; /* FW filename */ unsigned char dfile[32]; - uint32_t reserved1[2]; + u32 reserved1[2]; } __packed; struct stepping_info { @@ -230,7 +230,7 @@ intel_get_stepping_info(struct drm_i915_private *dev_priv) static void gen9_set_dc_state_debugmask(struct drm_i915_private *dev_priv) { - uint32_t val, mask; + u32 val, mask; mask = DC_STATE_DEBUG_MASK_MEMORY_UP; @@ -257,7 +257,7 @@ static void gen9_set_dc_state_debugmask(struct drm_i915_private *dev_priv) void intel_csr_load_program(struct drm_i915_private *dev_priv) { u32 *payload = dev_priv->csr.dmc_payload; - uint32_t i, fw_size; + u32 i, fw_size; if (!HAS_CSR(dev_priv)) { DRM_ERROR("No CSR support available for this platform\n"); @@ -289,17 +289,17 @@ void intel_csr_load_program(struct drm_i915_private *dev_priv) gen9_set_dc_state_debugmask(dev_priv); } -static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv, - const struct firmware *fw) +static u32 *parse_csr_fw(struct drm_i915_private *dev_priv, + const struct firmware *fw) { struct intel_css_header *css_header; struct intel_package_header *package_header; struct intel_dmc_header *dmc_header; struct intel_csr *csr = &dev_priv->csr; const struct stepping_info *si = intel_get_stepping_info(dev_priv); - uint32_t dmc_offset = CSR_DEFAULT_FW_OFFSET, readcount = 0, nbytes; - uint32_t i; - uint32_t *dmc_payload; + u32 dmc_offset = CSR_DEFAULT_FW_OFFSET, readcount = 0, nbytes; + u32 i; + u32 *dmc_payload; if (!fw) return NULL; @@ -409,6 +409,21 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv, return memcpy(dmc_payload, &fw->data[readcount], nbytes); } +static void intel_csr_runtime_pm_get(struct drm_i915_private *dev_priv) +{ + WARN_ON(dev_priv->csr.wakeref); + dev_priv->csr.wakeref = + intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); +} + +static void intel_csr_runtime_pm_put(struct drm_i915_private *dev_priv) +{ + intel_wakeref_t wakeref __maybe_unused = + fetch_and_zero(&dev_priv->csr.wakeref); + + intel_display_power_put(dev_priv, POWER_DOMAIN_INIT, wakeref); +} + static void csr_load_work_fn(struct work_struct *work) { struct drm_i915_private *dev_priv; @@ -424,8 +439,7 @@ static void csr_load_work_fn(struct work_struct *work) if (dev_priv->csr.dmc_payload) { intel_csr_load_program(dev_priv); - - intel_display_power_put(dev_priv, POWER_DOMAIN_INIT); + intel_csr_runtime_pm_put(dev_priv); DRM_INFO("Finished loading DMC firmware %s (v%u.%u)\n", dev_priv->csr.fw_path, @@ -467,7 +481,7 @@ void intel_csr_ucode_init(struct drm_i915_private *dev_priv) * suspend as runtime suspend *requires* a working CSR for whatever * reason. */ - intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); + intel_csr_runtime_pm_get(dev_priv); if (INTEL_GEN(dev_priv) >= 12) { /* Allow to load fw via parameter using the last known size */ @@ -538,7 +552,7 @@ void intel_csr_ucode_suspend(struct drm_i915_private *dev_priv) /* Drop the reference held in case DMC isn't loaded. */ if (!dev_priv->csr.dmc_payload) - intel_display_power_put(dev_priv, POWER_DOMAIN_INIT); + intel_csr_runtime_pm_put(dev_priv); } /** @@ -558,7 +572,7 @@ void intel_csr_ucode_resume(struct drm_i915_private *dev_priv) * loaded. */ if (!dev_priv->csr.dmc_payload) - intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); + intel_csr_runtime_pm_get(dev_priv); } /** @@ -574,6 +588,7 @@ void intel_csr_ucode_fini(struct drm_i915_private *dev_priv) return; intel_csr_ucode_suspend(dev_priv); + WARN_ON(dev_priv->csr.wakeref); kfree(dev_priv->csr.dmc_payload); } diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index b1ac89b514c1..ca705546a0ab 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -974,7 +974,7 @@ static void intel_wait_ddi_buf_idle(struct drm_i915_private *dev_priv, DRM_ERROR("Timeout waiting for DDI BUF %c idle bit\n", port_name(port)); } -static uint32_t hsw_pll_to_ddi_pll_sel(const struct intel_shared_dpll *pll) +static u32 hsw_pll_to_ddi_pll_sel(const struct intel_shared_dpll *pll) { switch (pll->info->id) { case DPLL_ID_WRPLL1: @@ -995,8 +995,8 @@ static uint32_t hsw_pll_to_ddi_pll_sel(const struct intel_shared_dpll *pll) } } -static uint32_t icl_pll_to_ddi_pll_sel(struct intel_encoder *encoder, - const struct intel_crtc_state *crtc_state) +static u32 icl_pll_to_ddi_clk_sel(struct intel_encoder *encoder, + const struct intel_crtc_state *crtc_state) { const struct intel_shared_dpll *pll = crtc_state->shared_dpll; int clock = crtc_state->port_clock; @@ -1004,10 +1004,11 @@ static uint32_t icl_pll_to_ddi_pll_sel(struct intel_encoder *encoder, switch (id) { default: + /* + * DPLL_ID_ICL_DPLL0 and DPLL_ID_ICL_DPLL1 should not be used + * here, so do warn if this get passed in + */ MISSING_CASE(id); - /* fall through */ - case DPLL_ID_ICL_DPLL0: - case DPLL_ID_ICL_DPLL1: return DDI_CLK_SEL_NONE; case DPLL_ID_ICL_TBTPLL: switch (clock) { @@ -1021,7 +1022,7 @@ static uint32_t icl_pll_to_ddi_pll_sel(struct intel_encoder *encoder, return DDI_CLK_SEL_TBT_810; default: MISSING_CASE(clock); - break; + return DDI_CLK_SEL_NONE; } case DPLL_ID_ICL_MGPLL1: case DPLL_ID_ICL_MGPLL2: @@ -1243,8 +1244,8 @@ static int skl_calc_wrpll_link(struct drm_i915_private *dev_priv, enum intel_dpll_id pll_id) { i915_reg_t cfgcr1_reg, cfgcr2_reg; - uint32_t cfgcr1_val, cfgcr2_val; - uint32_t p0, p1, p2, dco_freq; + u32 cfgcr1_val, cfgcr2_val; + u32 p0, p1, p2, dco_freq; cfgcr1_reg = DPLL_CFGCR1(pll_id); cfgcr2_reg = DPLL_CFGCR2(pll_id); @@ -1305,8 +1306,8 @@ static int skl_calc_wrpll_link(struct drm_i915_private *dev_priv, int cnl_calc_wrpll_link(struct drm_i915_private *dev_priv, enum intel_dpll_id pll_id) { - uint32_t cfgcr0, cfgcr1; - uint32_t p0, p1, p2, dco_freq, ref_clock; + u32 cfgcr0, cfgcr1; + u32 p0, p1, p2, dco_freq, ref_clock; if (INTEL_GEN(dev_priv) >= 11) { cfgcr0 = I915_READ(ICL_DPLL_CFGCR0(pll_id)); @@ -1391,16 +1392,17 @@ static int icl_calc_tbt_pll_link(struct drm_i915_private *dev_priv, static int icl_calc_mg_pll_link(struct drm_i915_private *dev_priv, enum port port) { + enum tc_port tc_port = intel_port_to_tc(dev_priv, port); u32 mg_pll_div0, mg_clktop_hsclkctl; u32 m1, m2_int, m2_frac, div1, div2, refclk; u64 tmp; refclk = dev_priv->cdclk.hw.ref; - mg_pll_div0 = I915_READ(MG_PLL_DIV0(port)); - mg_clktop_hsclkctl = I915_READ(MG_CLKTOP2_HSCLKCTL(port)); + mg_pll_div0 = I915_READ(MG_PLL_DIV0(tc_port)); + mg_clktop_hsclkctl = I915_READ(MG_CLKTOP2_HSCLKCTL(tc_port)); - m1 = I915_READ(MG_PLL_DIV1(port)) & MG_PLL_DIV1_FBPREDIV_MASK; + m1 = I915_READ(MG_PLL_DIV1(tc_port)) & MG_PLL_DIV1_FBPREDIV_MASK; m2_int = mg_pll_div0 & MG_PLL_DIV0_FBDIV_INT_MASK; m2_frac = (mg_pll_div0 & MG_PLL_DIV0_FRACNEN_H) ? (mg_pll_div0 & MG_PLL_DIV0_FBDIV_FRAC_MASK) >> @@ -1471,7 +1473,7 @@ static void icl_ddi_clock_get(struct intel_encoder *encoder, struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); enum port port = encoder->port; int link_clock = 0; - uint32_t pll_id; + u32 pll_id; pll_id = intel_get_shared_dpll_id(dev_priv, pipe_config->shared_dpll); if (intel_port_is_combophy(dev_priv, port)) { @@ -1496,7 +1498,7 @@ static void cnl_ddi_clock_get(struct intel_encoder *encoder, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); int link_clock = 0; - uint32_t cfgcr0; + u32 cfgcr0; enum intel_dpll_id pll_id; pll_id = intel_get_shared_dpll_id(dev_priv, pipe_config->shared_dpll); @@ -1550,7 +1552,7 @@ static void skl_ddi_clock_get(struct intel_encoder *encoder, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); int link_clock = 0; - uint32_t dpll_ctl1; + u32 dpll_ctl1; enum intel_dpll_id pll_id; pll_id = intel_get_shared_dpll_id(dev_priv, pipe_config->shared_dpll); @@ -1739,7 +1741,7 @@ void intel_ddi_set_vc_payload_alloc(const struct intel_crtc_state *crtc_state, struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum transcoder cpu_transcoder = crtc_state->cpu_transcoder; - uint32_t temp; + u32 temp; temp = I915_READ(TRANS_DDI_FUNC_CTL(cpu_transcoder)); if (state == true) @@ -1757,7 +1759,7 @@ void intel_ddi_enable_transcoder_func(const struct intel_crtc_state *crtc_state) enum pipe pipe = crtc->pipe; enum transcoder cpu_transcoder = crtc_state->cpu_transcoder; enum port port = encoder->port; - uint32_t temp; + u32 temp; /* Enable TRANS_DDI_FUNC_CTL for the pipe to work in HDMI mode */ temp = TRANS_DDI_FUNC_ENABLE; @@ -1841,7 +1843,7 @@ void intel_ddi_disable_transcoder_func(const struct intel_crtc_state *crtc_state struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum transcoder cpu_transcoder = crtc_state->cpu_transcoder; i915_reg_t reg = TRANS_DDI_FUNC_CTL(cpu_transcoder); - uint32_t val = I915_READ(reg); + u32 val = I915_READ(reg); val &= ~(TRANS_DDI_FUNC_ENABLE | TRANS_DDI_PORT_MASK | TRANS_DDI_DP_VC_PAYLOAD_ALLOC); val |= TRANS_DDI_PORT_NONE; @@ -1860,12 +1862,14 @@ int intel_ddi_toggle_hdcp_signalling(struct intel_encoder *intel_encoder, { struct drm_device *dev = intel_encoder->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); + intel_wakeref_t wakeref; enum pipe pipe = 0; int ret = 0; - uint32_t tmp; + u32 tmp; - if (WARN_ON(!intel_display_power_get_if_enabled(dev_priv, - intel_encoder->power_domain))) + wakeref = intel_display_power_get_if_enabled(dev_priv, + intel_encoder->power_domain); + if (WARN_ON(!wakeref)) return -ENXIO; if (WARN_ON(!intel_encoder->get_hw_state(intel_encoder, &pipe))) { @@ -1880,7 +1884,7 @@ int intel_ddi_toggle_hdcp_signalling(struct intel_encoder *intel_encoder, tmp &= ~TRANS_DDI_HDCP_SIGNALLING; I915_WRITE(TRANS_DDI_FUNC_CTL(pipe), tmp); out: - intel_display_power_put(dev_priv, intel_encoder->power_domain); + intel_display_power_put(dev_priv, intel_encoder->power_domain, wakeref); return ret; } @@ -1891,13 +1895,15 @@ bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector) struct intel_encoder *encoder = intel_connector->encoder; int type = intel_connector->base.connector_type; enum port port = encoder->port; - enum pipe pipe = 0; enum transcoder cpu_transcoder; - uint32_t tmp; + intel_wakeref_t wakeref; + enum pipe pipe = 0; + u32 tmp; bool ret; - if (!intel_display_power_get_if_enabled(dev_priv, - encoder->power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + encoder->power_domain); + if (!wakeref) return false; if (!encoder->get_hw_state(encoder, &pipe)) { @@ -1939,7 +1945,7 @@ bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector) } out: - intel_display_power_put(dev_priv, encoder->power_domain); + intel_display_power_put(dev_priv, encoder->power_domain, wakeref); return ret; } @@ -1950,6 +1956,7 @@ static void intel_ddi_get_encoder_pipes(struct intel_encoder *encoder, struct drm_device *dev = encoder->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); enum port port = encoder->port; + intel_wakeref_t wakeref; enum pipe p; u32 tmp; u8 mst_pipe_mask; @@ -1957,8 +1964,9 @@ static void intel_ddi_get_encoder_pipes(struct intel_encoder *encoder, *pipe_mask = 0; *is_dp_mst = false; - if (!intel_display_power_get_if_enabled(dev_priv, - encoder->power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + encoder->power_domain); + if (!wakeref) return; tmp = I915_READ(DDI_BUF_CTL(port)); @@ -2029,7 +2037,7 @@ out: "(PHY_CTL %08x)\n", port_name(port), tmp); } - intel_display_power_put(dev_priv, encoder->power_domain); + intel_display_power_put(dev_priv, encoder->power_domain, wakeref); } bool intel_ddi_get_hw_state(struct intel_encoder *encoder, @@ -2126,7 +2134,7 @@ void intel_ddi_disable_pipe_clock(const struct intel_crtc_state *crtc_state) } static void _skl_ddi_set_iboost(struct drm_i915_private *dev_priv, - enum port port, uint8_t iboost) + enum port port, u8 iboost) { u32 tmp; @@ -2145,7 +2153,7 @@ static void skl_ddi_set_iboost(struct intel_encoder *encoder, struct intel_digital_port *intel_dig_port = enc_to_dig_port(&encoder->base); struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); enum port port = encoder->port; - uint8_t iboost; + u8 iboost; if (type == INTEL_OUTPUT_HDMI) iboost = dev_priv->vbt.ddi_port_info[port].hdmi_boost_level; @@ -2659,7 +2667,7 @@ static void icl_ddi_vswing_sequence(struct intel_encoder *encoder, icl_mg_phy_ddi_vswing_sequence(encoder, link_clock, level); } -static uint32_t translate_signal_level(int signal_levels) +static u32 translate_signal_level(int signal_levels) { int i; @@ -2674,9 +2682,9 @@ static uint32_t translate_signal_level(int signal_levels) return 0; } -static uint32_t intel_ddi_dp_level(struct intel_dp *intel_dp) +static u32 intel_ddi_dp_level(struct intel_dp *intel_dp) { - uint8_t train_set = intel_dp->train_set[0]; + u8 train_set = intel_dp->train_set[0]; int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK | DP_TRAIN_PRE_EMPHASIS_MASK); @@ -2701,7 +2709,7 @@ u32 bxt_signal_levels(struct intel_dp *intel_dp) return 0; } -uint32_t ddi_signal_levels(struct intel_dp *intel_dp) +u32 ddi_signal_levels(struct intel_dp *intel_dp) { struct intel_digital_port *dport = dp_to_dig_port(intel_dp); struct drm_i915_private *dev_priv = to_i915(dport->base.base.dev); @@ -2715,8 +2723,8 @@ uint32_t ddi_signal_levels(struct intel_dp *intel_dp) } static inline -uint32_t icl_dpclka_cfgcr0_clk_off(struct drm_i915_private *dev_priv, - enum port port) +u32 icl_dpclka_cfgcr0_clk_off(struct drm_i915_private *dev_priv, + enum port port) { if (intel_port_is_combophy(dev_priv, port)) { return ICL_DPCLKA_CFGCR0_DDI_CLK_OFF(port); @@ -2851,7 +2859,7 @@ static void intel_ddi_clk_select(struct intel_encoder *encoder, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); enum port port = encoder->port; - uint32_t val; + u32 val; const struct intel_shared_dpll *pll = crtc_state->shared_dpll; if (WARN_ON(!pll)) @@ -2862,7 +2870,7 @@ static void intel_ddi_clk_select(struct intel_encoder *encoder, if (IS_ICELAKE(dev_priv)) { if (!intel_port_is_combophy(dev_priv, port)) I915_WRITE(DDI_CLK_SEL(port), - icl_pll_to_ddi_pll_sel(encoder, crtc_state)); + icl_pll_to_ddi_clk_sel(encoder, crtc_state)); } else if (IS_CANNONLAKE(dev_priv)) { /* Configure DPCLKA_CFGCR0 to map the DPLL to the DDI. */ val = I915_READ(DPCLKA_CFGCR0); @@ -3286,7 +3294,8 @@ static void intel_ddi_post_disable_dp(struct intel_encoder *encoder, intel_edp_panel_vdd_on(intel_dp); intel_edp_panel_off(intel_dp); - intel_display_power_put(dev_priv, dig_port->ddi_io_power_domain); + intel_display_power_put_unchecked(dev_priv, + dig_port->ddi_io_power_domain); intel_ddi_clk_disable(encoder); } @@ -3306,7 +3315,8 @@ static void intel_ddi_post_disable_hdmi(struct intel_encoder *encoder, intel_disable_ddi_buf(encoder, old_crtc_state); - intel_display_power_put(dev_priv, dig_port->ddi_io_power_domain); + intel_display_power_put_unchecked(dev_priv, + dig_port->ddi_io_power_domain); intel_ddi_clk_disable(encoder); @@ -3348,7 +3358,7 @@ void intel_ddi_fdi_post_disable(struct intel_encoder *encoder, const struct drm_connector_state *old_conn_state) { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); - uint32_t val; + u32 val; /* * Bspec lists this as both step 13 (before DDI_BUF_CTL disable) @@ -3548,6 +3558,8 @@ static void intel_ddi_update_pipe_dp(struct intel_encoder *encoder, intel_psr_enable(intel_dp, crtc_state); intel_edp_drrs_enable(intel_dp, crtc_state); + + intel_panel_update_backlight(encoder, crtc_state, conn_state); } static void intel_ddi_update_pipe(struct intel_encoder *encoder, @@ -3626,8 +3638,8 @@ intel_ddi_post_pll_disable(struct intel_encoder *encoder, if (intel_crtc_has_dp_encoder(crtc_state) || intel_port_is_tc(dev_priv, encoder->port)) - intel_display_power_put(dev_priv, - intel_ddi_main_link_aux_domain(dig_port)); + intel_display_power_put_unchecked(dev_priv, + intel_ddi_main_link_aux_domain(dig_port)); } void intel_ddi_prepare_link_retrain(struct intel_dp *intel_dp) @@ -3636,7 +3648,7 @@ void intel_ddi_prepare_link_retrain(struct intel_dp *intel_dp) struct drm_i915_private *dev_priv = to_i915(intel_dig_port->base.base.dev); enum port port = intel_dig_port->base.port; - uint32_t val; + u32 val; bool wait = false; if (I915_READ(DP_TP_CTL(port)) & DP_TP_CTL_ENABLE) { diff --git a/drivers/gpu/drm/i915/intel_device_info.h b/drivers/gpu/drm/i915/intel_device_info.h index 957c6527f76b..e8b8661df746 100644 --- a/drivers/gpu/drm/i915/intel_device_info.h +++ b/drivers/gpu/drm/i915/intel_device_info.h @@ -115,7 +115,7 @@ enum intel_ppgtt { func(has_ddi); \ func(has_dp_mst); \ func(has_fbc); \ - func(has_gmch_display); \ + func(has_gmch); \ func(has_hotplug); \ func(has_ipc); \ func(has_overlay); \ @@ -189,6 +189,8 @@ struct intel_device_info { struct color_luts { u16 degamma_lut_size; u16 gamma_lut_size; + u32 degamma_lut_tests; + u32 gamma_lut_tests; } color; }; diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index e0f40be07131..ccb616351bba 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -51,8 +51,17 @@ #include "intel_dsi.h" #include "intel_frontbuffer.h" +#include "intel_drv.h" +#include "intel_dsi.h" +#include "intel_frontbuffer.h" + +#include "i915_drv.h" +#include "i915_gem_clflush.h" +#include "i915_reset.h" +#include "i915_trace.h" + /* Primary plane formats for gen <= 3 */ -static const uint32_t i8xx_primary_formats[] = { +static const u32 i8xx_primary_formats[] = { DRM_FORMAT_C8, DRM_FORMAT_RGB565, DRM_FORMAT_XRGB1555, @@ -60,7 +69,7 @@ static const uint32_t i8xx_primary_formats[] = { }; /* Primary plane formats for gen >= 4 */ -static const uint32_t i965_primary_formats[] = { +static const u32 i965_primary_formats[] = { DRM_FORMAT_C8, DRM_FORMAT_RGB565, DRM_FORMAT_XRGB8888, @@ -69,18 +78,18 @@ static const uint32_t i965_primary_formats[] = { DRM_FORMAT_XBGR2101010, }; -static const uint64_t i9xx_format_modifiers[] = { +static const u64 i9xx_format_modifiers[] = { I915_FORMAT_MOD_X_TILED, DRM_FORMAT_MOD_LINEAR, DRM_FORMAT_MOD_INVALID }; /* Cursor formats */ -static const uint32_t intel_cursor_formats[] = { +static const u32 intel_cursor_formats[] = { DRM_FORMAT_ARGB8888, }; -static const uint64_t cursor_format_modifiers[] = { +static const u64 cursor_format_modifiers[] = { DRM_FORMAT_MOD_LINEAR, DRM_FORMAT_MOD_INVALID }; @@ -496,7 +505,7 @@ static int pnv_calc_dpll_params(int refclk, struct dpll *clock) return clock->dot; } -static uint32_t i9xx_dpll_compute_m(struct dpll *dpll) +static u32 i9xx_dpll_compute_m(struct dpll *dpll) { return 5 * (dpll->m1 + 2) + (dpll->m2 + 2); } @@ -531,8 +540,8 @@ int chv_calc_dpll_params(int refclk, struct dpll *clock) clock->p = clock->p1 * clock->p2; if (WARN_ON(clock->n == 0 || clock->p == 0)) return 0; - clock->vco = DIV_ROUND_CLOSEST_ULL((uint64_t)refclk * clock->m, - clock->n << 22); + clock->vco = DIV_ROUND_CLOSEST_ULL((u64)refclk * clock->m, + clock->n << 22); clock->dot = DIV_ROUND_CLOSEST(clock->vco, clock->p); return clock->dot / 5; @@ -894,7 +903,7 @@ chv_find_best_dpll(const struct intel_limit *limit, struct drm_device *dev = crtc->base.dev; unsigned int best_error_ppm; struct dpll clock; - uint64_t m2; + u64 m2; int found = false; memset(best_clock, 0, sizeof(*best_clock)); @@ -916,7 +925,7 @@ chv_find_best_dpll(const struct intel_limit *limit, clock.p = clock.p1 * clock.p2; - m2 = DIV_ROUND_CLOSEST_ULL(((uint64_t)target * clock.p * + m2 = DIV_ROUND_CLOSEST_ULL(((u64)target * clock.p * clock.n) << 22, refclk * clock.m1); if (m2 > INT_MAX/clock.m1) @@ -1200,17 +1209,19 @@ void assert_pipe(struct drm_i915_private *dev_priv, enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv, pipe); enum intel_display_power_domain power_domain; + intel_wakeref_t wakeref; /* we keep both pipes enabled on 830 */ if (IS_I830(dev_priv)) state = true; power_domain = POWER_DOMAIN_TRANSCODER(cpu_transcoder); - if (intel_display_power_get_if_enabled(dev_priv, power_domain)) { + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (wakeref) { u32 val = I915_READ(PIPECONF(cpu_transcoder)); cur_state = !!(val & PIPECONF_ENABLE); - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); } else { cur_state = false; } @@ -1611,7 +1622,7 @@ static void ironlake_enable_pch_transcoder(const struct intel_crtc_state *crtc_s struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; i915_reg_t reg; - uint32_t val, pipeconf_val; + u32 val, pipeconf_val; /* Make sure PCH DPLL is enabled */ assert_shared_dpll_enabled(dev_priv, crtc_state->shared_dpll); @@ -1699,7 +1710,7 @@ static void ironlake_disable_pch_transcoder(struct drm_i915_private *dev_priv, enum pipe pipe) { i915_reg_t reg; - uint32_t val; + u32 val; /* FDI relies on the transcoder */ assert_fdi_tx_disabled(dev_priv, pipe); @@ -1756,6 +1767,35 @@ enum pipe intel_crtc_pch_transcoder(struct intel_crtc *crtc) return crtc->pipe; } +static u32 intel_crtc_max_vblank_count(const struct intel_crtc_state *crtc_state) +{ + struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); + + /* + * On i965gm the hardware frame counter reads + * zero when the TV encoder is enabled :( + */ + if (IS_I965GM(dev_priv) && + (crtc_state->output_types & BIT(INTEL_OUTPUT_TVOUT))) + return 0; + + if (INTEL_GEN(dev_priv) >= 5 || IS_G4X(dev_priv)) + return 0xffffffff; /* full 32 bit counter */ + else if (INTEL_GEN(dev_priv) >= 3) + return 0xffffff; /* only 24 bits of frame count */ + else + return 0; /* Gen2 doesn't have a hardware frame counter */ +} + +static void intel_crtc_vblank_on(const struct intel_crtc_state *crtc_state) +{ + struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); + + drm_crtc_set_max_vblank_count(&crtc->base, + intel_crtc_max_vblank_count(crtc_state)); + drm_crtc_vblank_on(&crtc->base); +} + static void intel_enable_pipe(const struct intel_crtc_state *new_crtc_state) { struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->base.crtc); @@ -1774,7 +1814,7 @@ static void intel_enable_pipe(const struct intel_crtc_state *new_crtc_state) * a plane. On ILK+ the pipe PLLs are integrated, so we don't * need the check. */ - if (HAS_GMCH_DISPLAY(dev_priv)) { + if (HAS_GMCH(dev_priv)) { if (intel_crtc_has_type(new_crtc_state, INTEL_OUTPUT_DSI)) assert_dsi_pll_enabled(dev_priv); else @@ -1808,7 +1848,7 @@ static void intel_enable_pipe(const struct intel_crtc_state *new_crtc_state) * when it's derived from the timestamps. So let's wait for the * pipe to start properly before we call drm_crtc_vblank_on() */ - if (dev_priv->drm.max_vblank_count == 0) + if (intel_crtc_max_vblank_count(new_crtc_state) == 0) intel_wait_for_pipe_scanline_moving(crtc); } @@ -2026,6 +2066,7 @@ intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, struct drm_device *dev = fb->dev; struct drm_i915_private *dev_priv = to_i915(dev); struct drm_i915_gem_object *obj = intel_fb_obj(fb); + intel_wakeref_t wakeref; struct i915_vma *vma; unsigned int pinctl; u32 alignment; @@ -2049,7 +2090,7 @@ intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, * intel_runtime_pm_put(), so it is correct to wrap only the * pin/unpin/fence and not more. */ - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); atomic_inc(&dev_priv->gpu_error.pending_fb_pin); @@ -2062,7 +2103,7 @@ intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, * complicated than this. For example, Cherryview appears quite * happy to scanout from anywhere within its global aperture. */ - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) pinctl |= PIN_MAPPABLE; vma = i915_gem_object_pin_to_display_plane(obj, @@ -2104,7 +2145,7 @@ intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, err: atomic_dec(&dev_priv->gpu_error.pending_fb_pin); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return vma; } @@ -2375,7 +2416,7 @@ static int intel_fb_offset_to_xy(int *x, int *y, return 0; } -static unsigned int intel_fb_modifier_to_tiling(uint64_t fb_modifier) +static unsigned int intel_fb_modifier_to_tiling(u64 fb_modifier) { switch (fb_modifier) { case I915_FORMAT_MOD_X_TILED: @@ -3163,7 +3204,7 @@ i9xx_plane_max_stride(struct intel_plane *plane, { struct drm_i915_private *dev_priv = to_i915(plane->base.dev); - if (!HAS_GMCH_DISPLAY(dev_priv)) { + if (!HAS_GMCH(dev_priv)) { return 32*1024; } else if (INTEL_GEN(dev_priv) >= 4) { if (modifier == I915_FORMAT_MOD_X_TILED) @@ -3183,28 +3224,38 @@ i9xx_plane_max_stride(struct intel_plane *plane, } } +static u32 i9xx_plane_ctl_crtc(const struct intel_crtc_state *crtc_state) +{ + struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + u32 dspcntr = 0; + + dspcntr |= DISPPLANE_GAMMA_ENABLE; + + if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) + dspcntr |= DISPPLANE_PIPE_CSC_ENABLE; + + if (INTEL_GEN(dev_priv) < 5) + dspcntr |= DISPPLANE_SEL_PIPE(crtc->pipe); + + return dspcntr; +} + static u32 i9xx_plane_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { struct drm_i915_private *dev_priv = to_i915(plane_state->base.plane->dev); - struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); const struct drm_framebuffer *fb = plane_state->base.fb; unsigned int rotation = plane_state->base.rotation; u32 dspcntr; - dspcntr = DISPLAY_PLANE_ENABLE | DISPPLANE_GAMMA_ENABLE; + dspcntr = DISPLAY_PLANE_ENABLE; if (IS_G4X(dev_priv) || IS_GEN(dev_priv, 5) || IS_GEN(dev_priv, 6) || IS_IVYBRIDGE(dev_priv)) dspcntr |= DISPPLANE_TRICKLE_FEED_DISABLE; - if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) - dspcntr |= DISPPLANE_PIPE_CSC_ENABLE; - - if (INTEL_GEN(dev_priv) < 5) - dspcntr |= DISPPLANE_SEL_PIPE(crtc->pipe); - switch (fb->format->format) { case DRM_FORMAT_C8: dspcntr |= DISPPLANE_8BPP; @@ -3332,11 +3383,13 @@ static void i9xx_update_plane(struct intel_plane *plane, struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum i9xx_plane_id i9xx_plane = plane->i9xx_plane; u32 linear_offset; - u32 dspcntr = plane_state->ctl; int x = plane_state->color_plane[0].x; int y = plane_state->color_plane[0].y; unsigned long irqflags; u32 dspaddr_offset; + u32 dspcntr; + + dspcntr = plane_state->ctl | i9xx_plane_ctl_crtc(crtc_state); linear_offset = intel_fb_xy_to_linear(x, y, plane_state, 0); @@ -3396,10 +3449,23 @@ static void i9xx_disable_plane(struct intel_plane *plane, struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum i9xx_plane_id i9xx_plane = plane->i9xx_plane; unsigned long irqflags; + u32 dspcntr; + + /* + * DSPCNTR pipe gamma enable on g4x+ and pipe csc + * enable on ilk+ affect the pipe bottom color as + * well, so we must configure them even if the plane + * is disabled. + * + * On pre-g4x there is no way to gamma correct the + * pipe bottom color but we'll keep on doing this + * anyway. + */ + dspcntr = i9xx_plane_ctl_crtc(crtc_state); spin_lock_irqsave(&dev_priv->uncore.lock, irqflags); - I915_WRITE_FW(DSPCNTR(i9xx_plane), 0); + I915_WRITE_FW(DSPCNTR(i9xx_plane), dspcntr); if (INTEL_GEN(dev_priv) >= 4) I915_WRITE_FW(DSPSURF(i9xx_plane), 0); else @@ -3414,6 +3480,7 @@ static bool i9xx_plane_get_hw_state(struct intel_plane *plane, struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum intel_display_power_domain power_domain; enum i9xx_plane_id i9xx_plane = plane->i9xx_plane; + intel_wakeref_t wakeref; bool ret; u32 val; @@ -3423,7 +3490,8 @@ static bool i9xx_plane_get_hw_state(struct intel_plane *plane, * display power wells. */ power_domain = POWER_DOMAIN_PIPE(plane->pipe); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) return false; val = I915_READ(DSPCNTR(i9xx_plane)); @@ -3436,7 +3504,7 @@ static bool i9xx_plane_get_hw_state(struct intel_plane *plane, *pipe = (val & DISPPLANE_SEL_PIPE_MASK) >> DISPPLANE_SEL_PIPE_SHIFT; - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); return ret; } @@ -3505,7 +3573,7 @@ u32 skl_plane_stride(const struct intel_plane_state *plane_state, return stride / skl_plane_stride_mult(fb, color_plane, rotation); } -static u32 skl_plane_ctl_format(uint32_t pixel_format) +static u32 skl_plane_ctl_format(u32 pixel_format) { switch (pixel_format) { case DRM_FORMAT_C8: @@ -3575,7 +3643,7 @@ static u32 glk_plane_color_ctl_alpha(const struct intel_plane_state *plane_state } } -static u32 skl_plane_ctl_tiling(uint64_t fb_modifier) +static u32 skl_plane_ctl_tiling(u64 fb_modifier) { switch (fb_modifier) { case DRM_FORMAT_MOD_LINEAR: @@ -3634,6 +3702,20 @@ static u32 cnl_plane_ctl_flip(unsigned int reflect) return 0; } +u32 skl_plane_ctl_crtc(const struct intel_crtc_state *crtc_state) +{ + struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); + u32 plane_ctl = 0; + + if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv)) + return plane_ctl; + + plane_ctl |= PLANE_CTL_PIPE_GAMMA_ENABLE; + plane_ctl |= PLANE_CTL_PIPE_CSC_ENABLE; + + return plane_ctl; +} + u32 skl_plane_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { @@ -3648,10 +3730,7 @@ u32 skl_plane_ctl(const struct intel_crtc_state *crtc_state, if (INTEL_GEN(dev_priv) < 10 && !IS_GEMINILAKE(dev_priv)) { plane_ctl |= skl_plane_ctl_alpha(plane_state); - plane_ctl |= - PLANE_CTL_PIPE_GAMMA_ENABLE | - PLANE_CTL_PIPE_CSC_ENABLE | - PLANE_CTL_PLANE_GAMMA_DISABLE; + plane_ctl |= PLANE_CTL_PLANE_GAMMA_DISABLE; if (plane_state->base.color_encoding == DRM_COLOR_YCBCR_BT709) plane_ctl |= PLANE_CTL_YUV_TO_RGB_CSC_FORMAT_BT709; @@ -3676,19 +3755,27 @@ u32 skl_plane_ctl(const struct intel_crtc_state *crtc_state, return plane_ctl; } +u32 glk_plane_color_ctl_crtc(const struct intel_crtc_state *crtc_state) +{ + struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); + u32 plane_color_ctl = 0; + + if (INTEL_GEN(dev_priv) >= 11) + return plane_color_ctl; + + plane_color_ctl |= PLANE_COLOR_PIPE_GAMMA_ENABLE; + plane_color_ctl |= PLANE_COLOR_PIPE_CSC_ENABLE; + + return plane_color_ctl; +} + u32 glk_plane_color_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { - struct drm_i915_private *dev_priv = - to_i915(plane_state->base.plane->dev); const struct drm_framebuffer *fb = plane_state->base.fb; struct intel_plane *plane = to_intel_plane(plane_state->base.plane); u32 plane_color_ctl = 0; - if (INTEL_GEN(dev_priv) < 11) { - plane_color_ctl |= PLANE_COLOR_PIPE_GAMMA_ENABLE; - plane_color_ctl |= PLANE_COLOR_PIPE_CSC_ENABLE; - } plane_color_ctl |= PLANE_COLOR_PLANE_GAMMA_DISABLE; plane_color_ctl |= glk_plane_color_ctl_alpha(plane_state); @@ -3737,7 +3824,7 @@ __intel_display_resume(struct drm_device *dev, } /* ignore any reset values/BIOS leftovers in the WM registers */ - if (!HAS_GMCH_DISPLAY(to_i915(dev))) + if (!HAS_GMCH(to_i915(dev))) to_intel_atomic_state(state)->skip_intermediate_wm = true; ret = drm_atomic_helper_commit_duplicated_state(state, ctx); @@ -3862,6 +3949,30 @@ unlock: clear_bit(I915_RESET_MODESET, &dev_priv->gpu_error.flags); } +static void icl_set_pipe_chicken(struct intel_crtc *crtc) +{ + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + enum pipe pipe = crtc->pipe; + u32 tmp; + + tmp = I915_READ(PIPE_CHICKEN(pipe)); + + /* + * Display WA #1153: icl + * enable hardware to bypass the alpha math + * and rounding for per-pixel values 00 and 0xff + */ + tmp |= PER_PIXEL_ALPHA_BYPASS_EN; + + /* + * W/A for underruns with linear/X-tiled with + * WM1+ disabled. + */ + tmp |= PM_FILL_MAINTAIN_DBUF_FULLNESS; + + I915_WRITE(PIPE_CHICKEN(pipe), tmp); +} + static void intel_update_pipe_config(const struct intel_crtc_state *old_crtc_state, const struct intel_crtc_state *new_crtc_state) { @@ -3896,6 +4007,19 @@ static void intel_update_pipe_config(const struct intel_crtc_state *old_crtc_sta else if (old_crtc_state->pch_pfit.enabled) ironlake_pfit_disable(old_crtc_state); } + + /* + * We don't (yet) allow userspace to control the pipe background color, + * so force it to black, but apply pipe gamma and CSC so that its + * handling will match how we program our planes. + */ + if (INTEL_GEN(dev_priv) >= 9) + I915_WRITE(SKL_BOTTOM_COLOR(crtc->pipe), + SKL_BOTTOM_COLOR_GAMMA_ENABLE | + SKL_BOTTOM_COLOR_CSC_ENABLE); + + if (INTEL_GEN(dev_priv) >= 11) + icl_set_pipe_chicken(crtc); } static void intel_fdi_normal_train(struct intel_crtc *crtc) @@ -4595,7 +4719,7 @@ static void ironlake_pch_transcoder_set_timings(const struct intel_crtc_state *c static void cpt_set_fdi_bc_bifurcation(struct drm_i915_private *dev_priv, bool enable) { - uint32_t temp; + u32 temp; temp = I915_READ(SOUTH_CHICKEN1); if (!!(temp & FDI_BC_BIFURCATION_SELECT) == enable) @@ -5250,7 +5374,7 @@ intel_pre_disable_primary_noatomic(struct drm_crtc *crtc) * event which is after the vblank start event, so we need to have a * wait-for-vblank between disabling the plane and the pipe. */ - if (HAS_GMCH_DISPLAY(dev_priv) && + if (HAS_GMCH(dev_priv) && intel_set_memory_cxsr(dev_priv, false)) intel_wait_for_vblank(dev_priv, pipe); } @@ -5258,18 +5382,36 @@ intel_pre_disable_primary_noatomic(struct drm_crtc *crtc) static bool hsw_pre_update_disable_ips(const struct intel_crtc_state *old_crtc_state, const struct intel_crtc_state *new_crtc_state) { + struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + if (!old_crtc_state->ips_enabled) return false; if (needs_modeset(&new_crtc_state->base)) return true; + /* + * Workaround : Do not read or write the pipe palette/gamma data while + * GAMMA_MODE is configured for split gamma and IPS_CTL has IPS enabled. + * + * Disable IPS before we program the LUT. + */ + if (IS_HASWELL(dev_priv) && + (new_crtc_state->base.color_mgmt_changed || + new_crtc_state->update_pipe) && + new_crtc_state->gamma_mode == GAMMA_MODE_MODE_SPLIT) + return true; + return !new_crtc_state->ips_enabled; } static bool hsw_post_update_enable_ips(const struct intel_crtc_state *old_crtc_state, const struct intel_crtc_state *new_crtc_state) { + struct intel_crtc *crtc = to_intel_crtc(new_crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + if (!new_crtc_state->ips_enabled) return false; @@ -5277,6 +5419,18 @@ static bool hsw_post_update_enable_ips(const struct intel_crtc_state *old_crtc_s return true; /* + * Workaround : Do not read or write the pipe palette/gamma data while + * GAMMA_MODE is configured for split gamma and IPS_CTL has IPS enabled. + * + * Re-enable IPS after the LUT has been programmed. + */ + if (IS_HASWELL(dev_priv) && + (new_crtc_state->base.color_mgmt_changed || + new_crtc_state->update_pipe) && + new_crtc_state->gamma_mode == GAMMA_MODE_MODE_SPLIT) + return true; + + /* * We can't read out IPS on broadwell, assume the worst and * forcibly enable IPS on the first fastset. */ @@ -5387,7 +5541,7 @@ static void intel_pre_plane_update(struct intel_crtc_state *old_crtc_state, * event which is after the vblank start event, so we need to have a * wait-for-vblank between disabling the plane and the pipe. */ - if (HAS_GMCH_DISPLAY(dev_priv) && old_crtc_state->base.active && + if (HAS_GMCH(dev_priv) && old_crtc_state->base.active && pipe_config->disable_cxsr && intel_set_memory_cxsr(dev_priv, false)) intel_wait_for_vblank(dev_priv, crtc->pipe); @@ -5664,6 +5818,7 @@ static void ironlake_crtc_enable(struct intel_crtc_state *pipe_config, * clocks enabled */ intel_color_load_luts(pipe_config); + intel_color_commit(pipe_config); if (dev_priv->display.initial_watermarks != NULL) dev_priv->display.initial_watermarks(old_intel_state, pipe_config); @@ -5673,7 +5828,7 @@ static void ironlake_crtc_enable(struct intel_crtc_state *pipe_config, ironlake_pch_enable(old_intel_state, pipe_config); assert_vblank_disabled(crtc); - drm_crtc_vblank_on(crtc); + intel_crtc_vblank_on(pipe_config); intel_encoders_enable(crtc, pipe_config, old_state); @@ -5718,7 +5873,7 @@ static void icl_pipe_mbus_enable(struct intel_crtc *crtc) { struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; - uint32_t val; + u32 val; val = MBUS_DBOX_A_CREDIT(2); val |= MBUS_DBOX_BW_CREDIT(1); @@ -5738,7 +5893,6 @@ static void haswell_crtc_enable(struct intel_crtc_state *pipe_config, struct intel_atomic_state *old_intel_state = to_intel_atomic_state(old_state); bool psl_clkgate_wa; - u32 pipe_chicken; if (WARN_ON(intel_crtc->active)) return; @@ -5774,8 +5928,6 @@ static void haswell_crtc_enable(struct intel_crtc_state *pipe_config, haswell_set_pipemisc(pipe_config); - intel_color_set_csc(pipe_config); - intel_crtc->active = true; /* Display WA #1180: WaDisableScalarClockGating: glk, cnl */ @@ -5794,17 +5946,10 @@ static void haswell_crtc_enable(struct intel_crtc_state *pipe_config, * clocks enabled */ intel_color_load_luts(pipe_config); + intel_color_commit(pipe_config); - /* - * Display WA #1153: enable hardware to bypass the alpha math - * and rounding for per-pixel values 00 and 0xff - */ - if (INTEL_GEN(dev_priv) >= 11) { - pipe_chicken = I915_READ(PIPE_CHICKEN(pipe)); - if (!(pipe_chicken & PER_PIXEL_ALPHA_BYPASS_EN)) - I915_WRITE_FW(PIPE_CHICKEN(pipe), - pipe_chicken | PER_PIXEL_ALPHA_BYPASS_EN); - } + if (INTEL_GEN(dev_priv) >= 11) + icl_set_pipe_chicken(intel_crtc); intel_ddi_set_pipe_settings(pipe_config); if (!transcoder_is_dsi(cpu_transcoder)) @@ -5827,7 +5972,7 @@ static void haswell_crtc_enable(struct intel_crtc_state *pipe_config, intel_ddi_set_vc_payload_alloc(pipe_config, true); assert_vblank_disabled(crtc); - drm_crtc_vblank_on(crtc); + intel_crtc_vblank_on(pipe_config); intel_encoders_enable(crtc, pipe_config, old_state); @@ -6109,7 +6254,7 @@ static void modeset_put_power_domains(struct drm_i915_private *dev_priv, enum intel_display_power_domain domain; for_each_power_domain(domain, domains) - intel_display_power_put(dev_priv, domain); + intel_display_power_put_unchecked(dev_priv, domain); } static void valleyview_crtc_enable(struct intel_crtc_state *pipe_config, @@ -6139,8 +6284,6 @@ static void valleyview_crtc_enable(struct intel_crtc_state *pipe_config, i9xx_set_pipeconf(pipe_config); - intel_color_set_csc(pipe_config); - intel_crtc->active = true; intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, true); @@ -6160,13 +6303,14 @@ static void valleyview_crtc_enable(struct intel_crtc_state *pipe_config, i9xx_pfit_enable(pipe_config); intel_color_load_luts(pipe_config); + intel_color_commit(pipe_config); dev_priv->display.initial_watermarks(old_intel_state, pipe_config); intel_enable_pipe(pipe_config); assert_vblank_disabled(crtc); - drm_crtc_vblank_on(crtc); + intel_crtc_vblank_on(pipe_config); intel_encoders_enable(crtc, pipe_config, old_state); } @@ -6216,6 +6360,7 @@ static void i9xx_crtc_enable(struct intel_crtc_state *pipe_config, i9xx_pfit_enable(pipe_config); intel_color_load_luts(pipe_config); + intel_color_commit(pipe_config); if (dev_priv->display.initial_watermarks != NULL) dev_priv->display.initial_watermarks(old_intel_state, @@ -6225,7 +6370,7 @@ static void i9xx_crtc_enable(struct intel_crtc_state *pipe_config, intel_enable_pipe(pipe_config); assert_vblank_disabled(crtc); - drm_crtc_vblank_on(crtc); + intel_crtc_vblank_on(pipe_config); intel_encoders_enable(crtc, pipe_config, old_state); } @@ -6356,7 +6501,7 @@ static void intel_crtc_disable_noatomic(struct drm_crtc *crtc, domains = intel_crtc->enabled_power_domains; for_each_power_domain(domain, domains) - intel_display_power_put(dev_priv, domain); + intel_display_power_put_unchecked(dev_priv, domain); intel_crtc->enabled_power_domains = 0; dev_priv->active_crtcs &= ~(1 << intel_crtc->pipe); @@ -6622,9 +6767,9 @@ static bool intel_crtc_supports_double_wide(const struct intel_crtc *crtc) (crtc->pipe == PIPE_A || IS_I915G(dev_priv)); } -static uint32_t ilk_pipe_pixel_rate(const struct intel_crtc_state *pipe_config) +static u32 ilk_pipe_pixel_rate(const struct intel_crtc_state *pipe_config) { - uint32_t pixel_rate; + u32 pixel_rate; pixel_rate = pipe_config->base.adjusted_mode.crtc_clock; @@ -6634,8 +6779,8 @@ static uint32_t ilk_pipe_pixel_rate(const struct intel_crtc_state *pipe_config) */ if (pipe_config->pch_pfit.enabled) { - uint64_t pipe_w, pipe_h, pfit_w, pfit_h; - uint32_t pfit_size = pipe_config->pch_pfit.size; + u64 pipe_w, pipe_h, pfit_w, pfit_h; + u32 pfit_size = pipe_config->pch_pfit.size; pipe_w = pipe_config->pipe_src_w; pipe_h = pipe_config->pipe_src_h; @@ -6650,7 +6795,7 @@ static uint32_t ilk_pipe_pixel_rate(const struct intel_crtc_state *pipe_config) if (WARN_ON(!pfit_w || !pfit_h)) return pixel_rate; - pixel_rate = div_u64((uint64_t) pixel_rate * pipe_w * pipe_h, + pixel_rate = div_u64((u64)pixel_rate * pipe_w * pipe_h, pfit_w * pfit_h); } @@ -6661,7 +6806,7 @@ static void intel_crtc_compute_pixel_rate(struct intel_crtc_state *crtc_state) { struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) /* FIXME calculate proper pipe pixel rate for GMCH pfit */ crtc_state->pixel_rate = crtc_state->base.adjusted_mode.crtc_clock; @@ -6746,7 +6891,7 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc, } static void -intel_reduce_m_n_ratio(uint32_t *num, uint32_t *den) +intel_reduce_m_n_ratio(u32 *num, u32 *den) { while (*num > DATA_LINK_M_N_MASK || *den > DATA_LINK_M_N_MASK) { @@ -6756,7 +6901,7 @@ intel_reduce_m_n_ratio(uint32_t *num, uint32_t *den) } static void compute_m_n(unsigned int m, unsigned int n, - uint32_t *ret_m, uint32_t *ret_n, + u32 *ret_m, u32 *ret_n, bool constant_n) { /* @@ -6771,7 +6916,7 @@ static void compute_m_n(unsigned int m, unsigned int n, else *ret_n = min_t(unsigned int, roundup_pow_of_two(n), DATA_LINK_N_MAX); - *ret_m = div_u64((uint64_t) m * *ret_n, n); + *ret_m = div_u64((u64)m * *ret_n, n); intel_reduce_m_n_ratio(ret_m, ret_n); } @@ -6801,12 +6946,12 @@ static inline bool intel_panel_use_ssc(struct drm_i915_private *dev_priv) && !(dev_priv->quirks & QUIRK_LVDS_SSC_DISABLE); } -static uint32_t pnv_dpll_compute_fp(struct dpll *dpll) +static u32 pnv_dpll_compute_fp(struct dpll *dpll) { return (1 << dpll->n) << 16 | dpll->m2; } -static uint32_t i9xx_dpll_compute_fp(struct dpll *dpll) +static u32 i9xx_dpll_compute_fp(struct dpll *dpll) { return dpll->n << 16 | dpll->m1 << 8 | dpll->m2; } @@ -7362,7 +7507,7 @@ static void intel_set_pipe_timings(const struct intel_crtc_state *crtc_state) enum pipe pipe = crtc->pipe; enum transcoder cpu_transcoder = crtc_state->cpu_transcoder; const struct drm_display_mode *adjusted_mode = &crtc_state->base.adjusted_mode; - uint32_t crtc_vtotal, crtc_vblank_end; + u32 crtc_vtotal, crtc_vblank_end; int vsyncshift = 0; /* We need to be careful not to changed the adjusted mode, for otherwise @@ -7437,7 +7582,7 @@ static void intel_get_pipe_timings(struct intel_crtc *crtc, struct drm_device *dev = crtc->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); enum transcoder cpu_transcoder = pipe_config->cpu_transcoder; - uint32_t tmp; + u32 tmp; tmp = I915_READ(HTOTAL(cpu_transcoder)); pipe_config->base.adjusted_mode.crtc_hdisplay = (tmp & 0xffff) + 1; @@ -7508,7 +7653,7 @@ static void i9xx_set_pipeconf(const struct intel_crtc_state *crtc_state) { struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - uint32_t pipeconf; + u32 pipeconf; pipeconf = 0; @@ -7753,7 +7898,7 @@ static void i9xx_get_pfit_config(struct intel_crtc *crtc, struct intel_crtc_state *pipe_config) { struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - uint32_t tmp; + u32 tmp; if (INTEL_GEN(dev_priv) <= 3 && (IS_I830(dev_priv) || !IS_MOBILE(dev_priv))) @@ -7968,11 +8113,13 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc, { struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum intel_display_power_domain power_domain; - uint32_t tmp; + intel_wakeref_t wakeref; + u32 tmp; bool ret; power_domain = POWER_DOMAIN_PIPE(crtc->pipe); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) return false; pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB; @@ -8073,7 +8220,7 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc, ret = true; out: - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); return ret; } @@ -8247,7 +8394,7 @@ static void ironlake_init_pch_refclk(struct drm_i915_private *dev_priv) static void lpt_reset_fdi_mphy(struct drm_i915_private *dev_priv) { - uint32_t tmp; + u32 tmp; tmp = I915_READ(SOUTH_CHICKEN2); tmp |= FDI_MPHY_IOSFSB_RESET_CTL; @@ -8269,7 +8416,7 @@ static void lpt_reset_fdi_mphy(struct drm_i915_private *dev_priv) /* WaMPhyProgramming:hsw */ static void lpt_program_fdi_mphy(struct drm_i915_private *dev_priv) { - uint32_t tmp; + u32 tmp; tmp = intel_sbi_read(dev_priv, 0x8008, SBI_MPHY); tmp &= ~(0xFF << 24); @@ -8350,7 +8497,7 @@ static void lpt_program_fdi_mphy(struct drm_i915_private *dev_priv) static void lpt_enable_clkout_dp(struct drm_i915_private *dev_priv, bool with_spread, bool with_fdi) { - uint32_t reg, tmp; + u32 reg, tmp; if (WARN(with_fdi && !with_spread, "FDI requires downspread\n")) with_spread = true; @@ -8389,7 +8536,7 @@ static void lpt_enable_clkout_dp(struct drm_i915_private *dev_priv, /* Sequence to disable CLKOUT_DP */ static void lpt_disable_clkout_dp(struct drm_i915_private *dev_priv) { - uint32_t reg, tmp; + u32 reg, tmp; mutex_lock(&dev_priv->sb_lock); @@ -8414,7 +8561,7 @@ static void lpt_disable_clkout_dp(struct drm_i915_private *dev_priv) #define BEND_IDX(steps) ((50 + (steps)) / 5) -static const uint16_t sscdivintphase[] = { +static const u16 sscdivintphase[] = { [BEND_IDX( 50)] = 0x3B23, [BEND_IDX( 45)] = 0x3B23, [BEND_IDX( 40)] = 0x3C23, @@ -8446,7 +8593,7 @@ static const uint16_t sscdivintphase[] = { */ static void lpt_bend_clkout_dp(struct drm_i915_private *dev_priv, int steps) { - uint32_t tmp; + u32 tmp; int idx = BEND_IDX(steps); if (WARN_ON(steps % 5 != 0)) @@ -8512,7 +8659,7 @@ static void ironlake_set_pipeconf(const struct intel_crtc_state *crtc_state) struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; - uint32_t val; + u32 val; val = 0; @@ -8859,7 +9006,7 @@ static void skylake_get_pfit_config(struct intel_crtc *crtc, struct drm_device *dev = crtc->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); struct intel_crtc_scaler_state *scaler_state = &pipe_config->scaler_state; - uint32_t ps_ctrl = 0; + u32 ps_ctrl = 0; int id = -1; int i; @@ -8871,6 +9018,7 @@ static void skylake_get_pfit_config(struct intel_crtc *crtc, pipe_config->pch_pfit.enabled = true; pipe_config->pch_pfit.pos = I915_READ(SKL_PS_WIN_POS(crtc->pipe, i)); pipe_config->pch_pfit.size = I915_READ(SKL_PS_WIN_SZ(crtc->pipe, i)); + scaler_state->scalers[i].in_use = true; break; } } @@ -9015,7 +9163,7 @@ static void ironlake_get_pfit_config(struct intel_crtc *crtc, { struct drm_device *dev = crtc->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); - uint32_t tmp; + u32 tmp; tmp = I915_READ(PF_CTL(crtc->pipe)); @@ -9040,11 +9188,13 @@ static bool ironlake_get_pipe_config(struct intel_crtc *crtc, struct drm_device *dev = crtc->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); enum intel_display_power_domain power_domain; - uint32_t tmp; + intel_wakeref_t wakeref; + u32 tmp; bool ret; power_domain = POWER_DOMAIN_PIPE(crtc->pipe); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) return false; pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB; @@ -9127,7 +9277,7 @@ static bool ironlake_get_pipe_config(struct intel_crtc *crtc, ret = true; out: - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); return ret; } @@ -9167,7 +9317,7 @@ static void assert_can_disable_lcpll(struct drm_i915_private *dev_priv) I915_STATE_WARN(intel_irqs_enabled(dev_priv), "IRQs enabled\n"); } -static uint32_t hsw_read_dcomp(struct drm_i915_private *dev_priv) +static u32 hsw_read_dcomp(struct drm_i915_private *dev_priv) { if (IS_HASWELL(dev_priv)) return I915_READ(D_COMP_HSW); @@ -9175,7 +9325,7 @@ static uint32_t hsw_read_dcomp(struct drm_i915_private *dev_priv) return I915_READ(D_COMP_BDW); } -static void hsw_write_dcomp(struct drm_i915_private *dev_priv, uint32_t val) +static void hsw_write_dcomp(struct drm_i915_private *dev_priv, u32 val) { if (IS_HASWELL(dev_priv)) { mutex_lock(&dev_priv->pcu_lock); @@ -9200,7 +9350,7 @@ static void hsw_write_dcomp(struct drm_i915_private *dev_priv, uint32_t val) static void hsw_disable_lcpll(struct drm_i915_private *dev_priv, bool switch_to_fclk, bool allow_power_down) { - uint32_t val; + u32 val; assert_can_disable_lcpll(dev_priv); @@ -9247,7 +9397,7 @@ static void hsw_disable_lcpll(struct drm_i915_private *dev_priv, */ static void hsw_restore_lcpll(struct drm_i915_private *dev_priv) { - uint32_t val; + u32 val; val = I915_READ(LCPLL_CTL); @@ -9322,7 +9472,7 @@ static void hsw_restore_lcpll(struct drm_i915_private *dev_priv) */ void hsw_enable_pc8(struct drm_i915_private *dev_priv) { - uint32_t val; + u32 val; DRM_DEBUG_KMS("Enabling package C8+\n"); @@ -9338,7 +9488,7 @@ void hsw_enable_pc8(struct drm_i915_private *dev_priv) void hsw_disable_pc8(struct drm_i915_private *dev_priv) { - uint32_t val; + u32 val; DRM_DEBUG_KMS("Disabling package C8+\n"); @@ -9406,7 +9556,7 @@ static void icelake_get_ddi_pll(struct drm_i915_private *dev_priv, if (WARN_ON(!intel_dpll_is_combophy(id))) return; } else if (intel_port_is_tc(dev_priv, port)) { - id = icl_port_to_mg_pll_id(port); + id = icl_tc_port_to_pll_id(intel_port_to_tc(dev_priv, port)); } else { WARN(1, "Invalid port %x\n", port); return; @@ -9460,7 +9610,7 @@ static void haswell_get_ddi_pll(struct drm_i915_private *dev_priv, struct intel_crtc_state *pipe_config) { enum intel_dpll_id id; - uint32_t ddi_pll_sel = I915_READ(PORT_CLK_SEL(port)); + u32 ddi_pll_sel = I915_READ(PORT_CLK_SEL(port)); switch (ddi_pll_sel) { case PORT_CLK_SEL_WRPLL1: @@ -9517,7 +9667,9 @@ static bool hsw_get_transcoder_state(struct intel_crtc *crtc, * XXX: Do intel_display_power_get_if_enabled before reading this (for * consistency and less surprising code; it's in always on power). */ - for_each_set_bit(panel_transcoder, &panel_transcoder_mask, 32) { + for_each_set_bit(panel_transcoder, + &panel_transcoder_mask, + ARRAY_SIZE(INTEL_INFO(dev_priv)->trans_offsets)) { enum pipe trans_pipe; tmp = I915_READ(TRANS_DDI_FUNC_CTL(panel_transcoder)); @@ -9563,6 +9715,8 @@ static bool hsw_get_transcoder_state(struct intel_crtc *crtc, power_domain = POWER_DOMAIN_TRANSCODER(pipe_config->cpu_transcoder); if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) return false; + + WARN_ON(*power_domain_mask & BIT_ULL(power_domain)); *power_domain_mask |= BIT_ULL(power_domain); tmp = I915_READ(PIPECONF(pipe_config->cpu_transcoder)); @@ -9590,6 +9744,8 @@ static bool bxt_get_dsi_transcoder_state(struct intel_crtc *crtc, power_domain = POWER_DOMAIN_TRANSCODER(cpu_transcoder); if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) continue; + + WARN_ON(*power_domain_mask & BIT_ULL(power_domain)); *power_domain_mask |= BIT_ULL(power_domain); /* @@ -9624,7 +9780,7 @@ static void haswell_get_ddi_port_state(struct intel_crtc *crtc, struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); struct intel_shared_dpll *pll; enum port port; - uint32_t tmp; + u32 tmp; tmp = I915_READ(TRANS_DDI_FUNC_CTL(pipe_config->cpu_transcoder)); @@ -9706,7 +9862,9 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc, power_domain = POWER_DOMAIN_PIPE_PANEL_FITTER(crtc->pipe); if (intel_display_power_get_if_enabled(dev_priv, power_domain)) { + WARN_ON(power_domain_mask & BIT_ULL(power_domain)); power_domain_mask |= BIT_ULL(power_domain); + if (INTEL_GEN(dev_priv) >= 9) skylake_get_pfit_config(crtc, pipe_config); else @@ -9736,7 +9894,7 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc, out: for_each_power_domain(power_domain, power_domain_mask) - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put_unchecked(dev_priv, power_domain); return active; } @@ -9757,7 +9915,7 @@ static u32 intel_cursor_base(const struct intel_plane_state *plane_state) base += plane_state->color_plane[0].offset; /* ILK+ do this automagically */ - if (HAS_GMCH_DISPLAY(dev_priv) && + if (HAS_GMCH(dev_priv) && plane_state->base.rotation & DRM_MODE_ROTATE_180) base += (plane_state->base.crtc_h * plane_state->base.crtc_w - 1) * fb->format->cpp[0]; @@ -9870,11 +10028,15 @@ i845_cursor_max_stride(struct intel_plane *plane, return 2048; } +static u32 i845_cursor_ctl_crtc(const struct intel_crtc_state *crtc_state) +{ + return CURSOR_GAMMA_ENABLE; +} + static u32 i845_cursor_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { return CURSOR_ENABLE | - CURSOR_GAMMA_ENABLE | CURSOR_FORMAT_ARGB | CURSOR_STRIDE(plane_state->color_plane[0].stride); } @@ -9944,7 +10106,9 @@ static void i845_update_cursor(struct intel_plane *plane, unsigned int width = plane_state->base.crtc_w; unsigned int height = plane_state->base.crtc_h; - cntl = plane_state->ctl; + cntl = plane_state->ctl | + i845_cursor_ctl_crtc(crtc_state); + size = (height << 12) | width; base = intel_cursor_base(plane_state); @@ -9986,17 +10150,19 @@ static bool i845_cursor_get_hw_state(struct intel_plane *plane, { struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum intel_display_power_domain power_domain; + intel_wakeref_t wakeref; bool ret; power_domain = POWER_DOMAIN_PIPE(PIPE_A); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) return false; ret = I915_READ(CURCNTR(PIPE_A)) & CURSOR_ENABLE; *pipe = PIPE_A; - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); return ret; } @@ -10009,27 +10175,36 @@ i9xx_cursor_max_stride(struct intel_plane *plane, return plane->base.dev->mode_config.cursor_width * 4; } -static u32 i9xx_cursor_ctl(const struct intel_crtc_state *crtc_state, - const struct intel_plane_state *plane_state) +static u32 i9xx_cursor_ctl_crtc(const struct intel_crtc_state *crtc_state) { - struct drm_i915_private *dev_priv = - to_i915(plane_state->base.plane->dev); struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); u32 cntl = 0; - if (IS_GEN(dev_priv, 6) || IS_IVYBRIDGE(dev_priv)) - cntl |= MCURSOR_TRICKLE_FEED_DISABLE; + if (INTEL_GEN(dev_priv) >= 11) + return cntl; - if (INTEL_GEN(dev_priv) <= 10) { - cntl |= MCURSOR_GAMMA_ENABLE; + cntl |= MCURSOR_GAMMA_ENABLE; - if (HAS_DDI(dev_priv)) - cntl |= MCURSOR_PIPE_CSC_ENABLE; - } + if (HAS_DDI(dev_priv)) + cntl |= MCURSOR_PIPE_CSC_ENABLE; if (INTEL_GEN(dev_priv) < 5 && !IS_G4X(dev_priv)) cntl |= MCURSOR_PIPE_SELECT(crtc->pipe); + return cntl; +} + +static u32 i9xx_cursor_ctl(const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state) +{ + struct drm_i915_private *dev_priv = + to_i915(plane_state->base.plane->dev); + u32 cntl = 0; + + if (IS_GEN(dev_priv, 6) || IS_IVYBRIDGE(dev_priv)) + cntl |= MCURSOR_TRICKLE_FEED_DISABLE; + switch (plane_state->base.crtc_w) { case 64: cntl |= MCURSOR_MODE_64_ARGB_AX; @@ -10154,7 +10329,8 @@ static void i9xx_update_cursor(struct intel_plane *plane, unsigned long irqflags; if (plane_state && plane_state->base.visible) { - cntl = plane_state->ctl; + cntl = plane_state->ctl | + i9xx_cursor_ctl_crtc(crtc_state); if (plane_state->base.crtc_h != plane_state->base.crtc_w) fbc_ctl = CUR_FBC_CTL_EN | (plane_state->base.crtc_h - 1); @@ -10219,6 +10395,7 @@ static bool i9xx_cursor_get_hw_state(struct intel_plane *plane, { struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum intel_display_power_domain power_domain; + intel_wakeref_t wakeref; bool ret; u32 val; @@ -10228,7 +10405,8 @@ static bool i9xx_cursor_get_hw_state(struct intel_plane *plane, * display power wells. */ power_domain = POWER_DOMAIN_PIPE(plane->pipe); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) return false; val = I915_READ(CURCNTR(plane->pipe)); @@ -10241,7 +10419,7 @@ static bool i9xx_cursor_get_hw_state(struct intel_plane *plane, *pipe = (val & MCURSOR_PIPE_SELECT_MASK) >> MCURSOR_PIPE_SELECT_SHIFT; - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); return ret; } @@ -10835,8 +11013,11 @@ int intel_plane_atomic_calc_changes(const struct intel_crtc_state *old_crtc_stat * Despite the w/a only being listed for IVB we assume that * the ILK/SNB note has similar ramifications, hence we apply * the w/a on all three platforms. + * + * With experimental results seems this is needed also for primary + * plane, not only sprite plane. */ - if (plane->id == PLANE_SPRITE0 && + if (plane->id != PLANE_CURSOR && (IS_GEN_RANGE(dev_priv, 5, 6) || IS_IVYBRIDGE(dev_priv)) && (turn_on || (!needs_scaling(old_plane_state) && @@ -10981,7 +11162,8 @@ static int intel_crtc_atomic_check(struct drm_crtc *crtc, int ret; bool mode_changed = needs_modeset(crtc_state); - if (mode_changed && !crtc_state->active) + if (INTEL_GEN(dev_priv) < 5 && !IS_G4X(dev_priv) && + mode_changed && !crtc_state->active) pipe_config->update_wm_post = true; if (mode_changed && crtc_state->enable && @@ -10993,7 +11175,7 @@ static int intel_crtc_atomic_check(struct drm_crtc *crtc, return ret; } - if (crtc_state->color_mgmt_changed) { + if (mode_changed || crtc_state->color_mgmt_changed) { ret = intel_color_check(pipe_config); if (ret) return ret; @@ -11292,7 +11474,7 @@ static void intel_dump_pipe_config(struct intel_crtc *crtc, pipe_config->scaler_state.scaler_users, pipe_config->scaler_state.scaler_id); - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) DRM_DEBUG_KMS("gmch pfit: control: 0x%08x, ratios: 0x%08x, lvds border: 0x%08x\n", pipe_config->gmch_pfit.control, pipe_config->gmch_pfit.pgm_ratios, @@ -11404,44 +11586,38 @@ static bool check_digital_port_conflicts(struct drm_atomic_state *state) return ret; } -static void +static int clear_intel_crtc_state(struct intel_crtc_state *crtc_state) { struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); - struct intel_crtc_scaler_state scaler_state; - struct intel_dpll_hw_state dpll_hw_state; - struct intel_shared_dpll *shared_dpll; - struct intel_crtc_wm_state wm_state; - bool force_thru, ips_force_disable; + struct intel_crtc_state *saved_state; + + saved_state = kzalloc(sizeof(*saved_state), GFP_KERNEL); + if (!saved_state) + return -ENOMEM; /* FIXME: before the switch to atomic started, a new pipe_config was * kzalloc'd. Code that depends on any field being zero should be * fixed, so that the crtc_state can be safely duplicated. For now, * only fields that are know to not cause problems are preserved. */ - scaler_state = crtc_state->scaler_state; - shared_dpll = crtc_state->shared_dpll; - dpll_hw_state = crtc_state->dpll_hw_state; - force_thru = crtc_state->pch_pfit.force_thru; - ips_force_disable = crtc_state->ips_force_disable; + saved_state->scaler_state = crtc_state->scaler_state; + saved_state->shared_dpll = crtc_state->shared_dpll; + saved_state->dpll_hw_state = crtc_state->dpll_hw_state; + saved_state->pch_pfit.force_thru = crtc_state->pch_pfit.force_thru; + saved_state->ips_force_disable = crtc_state->ips_force_disable; if (IS_G4X(dev_priv) || IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) - wm_state = crtc_state->wm; + saved_state->wm = crtc_state->wm; /* Keep base drm_crtc_state intact, only clear our extended struct */ BUILD_BUG_ON(offsetof(struct intel_crtc_state, base)); - memset(&crtc_state->base + 1, 0, + memcpy(&crtc_state->base + 1, &saved_state->base + 1, sizeof(*crtc_state) - sizeof(crtc_state->base)); - crtc_state->scaler_state = scaler_state; - crtc_state->shared_dpll = shared_dpll; - crtc_state->dpll_hw_state = dpll_hw_state; - crtc_state->pch_pfit.force_thru = force_thru; - crtc_state->ips_force_disable = ips_force_disable; - if (IS_G4X(dev_priv) || - IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) - crtc_state->wm = wm_state; + kfree(saved_state); + return 0; } static int @@ -11456,7 +11632,9 @@ intel_modeset_pipe_config(struct drm_crtc *crtc, int i; bool retry = true; - clear_intel_crtc_state(pipe_config); + ret = clear_intel_crtc_state(pipe_config); + if (ret) + return ret; pipe_config->cpu_transcoder = (enum transcoder) to_intel_crtc(crtc)->pipe; @@ -11665,6 +11843,23 @@ pipe_config_err(bool adjust, const char *name, const char *format, ...) va_end(args); } +static bool fastboot_enabled(struct drm_i915_private *dev_priv) +{ + if (i915_modparams.fastboot != -1) + return i915_modparams.fastboot; + + /* Enable fastboot by default on Skylake and newer */ + if (INTEL_GEN(dev_priv) >= 9) + return true; + + /* Enable fastboot by default on VLV and CHV */ + if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + return true; + + /* Disabled by default on all others */ + return false; +} + static bool intel_pipe_config_compare(struct drm_i915_private *dev_priv, struct intel_crtc_state *current_config, @@ -11676,6 +11871,11 @@ intel_pipe_config_compare(struct drm_i915_private *dev_priv, (current_config->base.mode.private_flags & I915_MODE_FLAG_INHERITED) && !(pipe_config->base.mode.private_flags & I915_MODE_FLAG_INHERITED); + if (fixup_inherited && !fastboot_enabled(dev_priv)) { + DRM_DEBUG_KMS("initial modeset and fastboot not set\n"); + ret = false; + } + #define PIPE_CONF_CHECK_X(name) do { \ if (current_config->name != pipe_config->name) { \ pipe_config_err(adjust, __stringify(name), \ @@ -12699,8 +12899,7 @@ static int intel_atomic_check(struct drm_device *dev, return ret; } - if (i915_modparams.fastboot && - intel_pipe_config_compare(dev_priv, + if (intel_pipe_config_compare(dev_priv, to_intel_crtc_state(old_crtc_state), pipe_config, true)) { crtc_state->mode_changed = false; @@ -12749,8 +12948,9 @@ static int intel_atomic_prepare_commit(struct drm_device *dev, u32 intel_crtc_get_vblank_counter(struct intel_crtc *crtc) { struct drm_device *dev = crtc->base.dev; + struct drm_vblank_crtc *vblank = &dev->vblank[drm_crtc_index(&crtc->base)]; - if (!dev->max_vblank_count) + if (!vblank->max_vblank_count) return (u32)drm_crtc_accurate_vblank_count(&crtc->base); return dev->driver->get_vblank_counter(dev, crtc->pipe); @@ -12959,6 +13159,7 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state) struct drm_crtc *crtc; struct intel_crtc *intel_crtc; u64 put_domains[I915_MAX_PIPES] = {}; + intel_wakeref_t wakeref = 0; int i; intel_atomic_commit_fence_wait(intel_state); @@ -12966,7 +13167,7 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state) drm_atomic_helper_wait_for_dependencies(state); if (intel_state->modeset) - intel_display_power_get(dev_priv, POWER_DOMAIN_MODESET); + wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_MODESET); for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) { old_intel_crtc_state = to_intel_crtc_state(old_crtc_state); @@ -13009,7 +13210,7 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state) /* FIXME unify this for all platforms */ if (!new_crtc_state->active && - !HAS_GMCH_DISPLAY(dev_priv) && + !HAS_GMCH(dev_priv) && dev_priv->display.initial_watermarks) dev_priv->display.initial_watermarks(intel_state, new_intel_crtc_state); @@ -13063,6 +13264,16 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state) */ drm_atomic_helper_wait_for_flip_done(dev, state); + for_each_new_crtc_in_state(state, crtc, new_crtc_state, i) { + new_intel_crtc_state = to_intel_crtc_state(new_crtc_state); + + if (new_crtc_state->active && + !needs_modeset(new_crtc_state) && + (new_intel_crtc_state->base.color_mgmt_changed || + new_intel_crtc_state->update_pipe)) + intel_color_load_luts(new_intel_crtc_state); + } + /* * Now that the vblank has passed, we can go ahead and program the * optimal watermarks on platforms that need two-step watermark @@ -13103,7 +13314,7 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state) * the culprit. */ intel_uncore_arm_unclaimed_mmio_detection(dev_priv); - intel_display_power_put(dev_priv, POWER_DOMAIN_MODESET); + intel_display_power_put(dev_priv, POWER_DOMAIN_MODESET, wakeref); } /* @@ -13578,19 +13789,16 @@ static void intel_begin_crtc_commit(struct drm_crtc *crtc, intel_atomic_get_new_crtc_state(old_intel_state, intel_crtc); bool modeset = needs_modeset(&intel_cstate->base); - if (!modeset && - (intel_cstate->base.color_mgmt_changed || - intel_cstate->update_pipe)) { - intel_color_set_csc(intel_cstate); - intel_color_load_luts(intel_cstate); - } - /* Perform vblank evasion around commit operation */ intel_pipe_update_start(intel_cstate); if (modeset) goto out; + if (intel_cstate->base.color_mgmt_changed || + intel_cstate->update_pipe) + intel_color_commit(intel_cstate); + if (intel_cstate->update_pipe) intel_update_pipe_config(old_intel_cstate, intel_cstate); else if (INTEL_GEN(dev_priv) >= 9) @@ -13731,8 +13939,8 @@ intel_legacy_cursor_update(struct drm_plane *plane, struct drm_framebuffer *fb, int crtc_x, int crtc_y, unsigned int crtc_w, unsigned int crtc_h, - uint32_t src_x, uint32_t src_y, - uint32_t src_w, uint32_t src_h, + u32 src_x, u32 src_y, + u32 src_w, u32 src_h, struct drm_modeset_acquire_ctx *ctx) { struct drm_i915_private *dev_priv = to_i915(crtc->dev); @@ -14206,7 +14414,7 @@ static int intel_encoder_clones(struct intel_encoder *encoder) return index_mask; } -static bool has_edp_a(struct drm_i915_private *dev_priv) +static bool ilk_has_edp_a(struct drm_i915_private *dev_priv) { if (!IS_MOBILE(dev_priv)) return false; @@ -14220,7 +14428,7 @@ static bool has_edp_a(struct drm_i915_private *dev_priv) return true; } -static bool intel_crt_present(struct drm_i915_private *dev_priv) +static bool intel_ddi_crt_present(struct drm_i915_private *dev_priv) { if (INTEL_GEN(dev_priv) >= 9) return false; @@ -14228,15 +14436,12 @@ static bool intel_crt_present(struct drm_i915_private *dev_priv) if (IS_HSW_ULT(dev_priv) || IS_BDW_ULT(dev_priv)) return false; - if (IS_CHERRYVIEW(dev_priv)) - return false; - if (HAS_PCH_LPT_H(dev_priv) && I915_READ(SFUSE_STRAP) & SFUSE_STRAP_CRT_DISABLED) return false; /* DDI E can't be used if DDI A requires 4 lanes */ - if (HAS_DDI(dev_priv) && I915_READ(DDI_BUF_CTL(PORT_A)) & DDI_A_4_LANES) + if (I915_READ(DDI_BUF_CTL(PORT_A)) & DDI_A_4_LANES) return false; if (!dev_priv->vbt.int_crt_support) @@ -14291,23 +14496,21 @@ static void intel_setup_outputs(struct drm_i915_private *dev_priv) if (!HAS_DISPLAY(dev_priv)) return; - /* - * intel_edp_init_connector() depends on this completing first, to - * prevent the registeration of both eDP and LVDS and the incorrect - * sharing of the PPS. - */ - intel_lvds_init(dev_priv); - - if (intel_crt_present(dev_priv)) - intel_crt_init(dev_priv); - if (IS_ICELAKE(dev_priv)) { intel_ddi_init(dev_priv, PORT_A); intel_ddi_init(dev_priv, PORT_B); intel_ddi_init(dev_priv, PORT_C); intel_ddi_init(dev_priv, PORT_D); intel_ddi_init(dev_priv, PORT_E); - intel_ddi_init(dev_priv, PORT_F); + /* + * On some ICL SKUs port F is not present. No strap bits for + * this, so rely on VBT. + * Work around broken VBTs on SKUs known to have no port F. + */ + if (IS_ICL_WITH_PORT_F(dev_priv) && + intel_bios_is_port_present(dev_priv, PORT_F)) + intel_ddi_init(dev_priv, PORT_F); + icl_dsi_init(dev_priv); } else if (IS_GEN9_LP(dev_priv)) { /* @@ -14323,6 +14526,9 @@ static void intel_setup_outputs(struct drm_i915_private *dev_priv) } else if (HAS_DDI(dev_priv)) { int found; + if (intel_ddi_crt_present(dev_priv)) + intel_crt_init(dev_priv); + /* * Haswell uses DDI functions to detect digital outputs. * On SKL pre-D0 the strap isn't connected, so we assume @@ -14349,16 +14555,23 @@ static void intel_setup_outputs(struct drm_i915_private *dev_priv) * On SKL we don't have a way to detect DDI-E so we rely on VBT. */ if (IS_GEN9_BC(dev_priv) && - (dev_priv->vbt.ddi_port_info[PORT_E].supports_dp || - dev_priv->vbt.ddi_port_info[PORT_E].supports_dvi || - dev_priv->vbt.ddi_port_info[PORT_E].supports_hdmi)) + intel_bios_is_port_present(dev_priv, PORT_E)) intel_ddi_init(dev_priv, PORT_E); } else if (HAS_PCH_SPLIT(dev_priv)) { int found; + + /* + * intel_edp_init_connector() depends on this completing first, + * to prevent the registration of both eDP and LVDS and the + * incorrect sharing of the PPS. + */ + intel_lvds_init(dev_priv); + intel_crt_init(dev_priv); + dpd_is_edp = intel_dp_is_port_edp(dev_priv, PORT_D); - if (has_edp_a(dev_priv)) + if (ilk_has_edp_a(dev_priv)) intel_dp_init(dev_priv, DP_A, PORT_A); if (I915_READ(PCH_HDMIB) & SDVO_DETECTED) { @@ -14384,6 +14597,9 @@ static void intel_setup_outputs(struct drm_i915_private *dev_priv) } else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { bool has_edp, has_port; + if (IS_VALLEYVIEW(dev_priv) && dev_priv->vbt.int_crt_support) + intel_crt_init(dev_priv); + /* * The DP_DETECTED bit is the latched state of the DDC * SDA pin at boot. However since eDP doesn't require DDC @@ -14426,9 +14642,17 @@ static void intel_setup_outputs(struct drm_i915_private *dev_priv) } vlv_dsi_init(dev_priv); - } else if (!IS_GEN(dev_priv, 2) && !IS_PINEVIEW(dev_priv)) { + } else if (IS_PINEVIEW(dev_priv)) { + intel_lvds_init(dev_priv); + intel_crt_init(dev_priv); + } else if (IS_GEN_RANGE(dev_priv, 3, 4)) { bool found = false; + if (IS_MOBILE(dev_priv)) + intel_lvds_init(dev_priv); + + intel_crt_init(dev_priv); + if (I915_READ(GEN3_SDVOB) & SDVO_DETECTED) { DRM_DEBUG_KMS("probing SDVOB\n"); found = intel_sdvo_init(dev_priv, GEN3_SDVOB, PORT_B); @@ -14460,11 +14684,16 @@ static void intel_setup_outputs(struct drm_i915_private *dev_priv) if (IS_G4X(dev_priv) && (I915_READ(DP_D) & DP_DETECTED)) intel_dp_init(dev_priv, DP_D, PORT_D); - } else if (IS_GEN(dev_priv, 2)) - intel_dvo_init(dev_priv); - if (SUPPORTS_TV(dev_priv)) - intel_tv_init(dev_priv); + if (SUPPORTS_TV(dev_priv)) + intel_tv_init(dev_priv); + } else if (IS_GEN(dev_priv, 2)) { + if (IS_I85X(dev_priv)) + intel_lvds_init(dev_priv); + + intel_crt_init(dev_priv); + intel_dvo_init(dev_priv); + } intel_psr_init(dev_priv); @@ -14631,14 +14860,6 @@ static int intel_framebuffer_init(struct intel_framebuffer *intel_fb, drm_helper_mode_fill_fb_struct(&dev_priv->drm, fb, mode_cmd); - if (fb->format->format == DRM_FORMAT_NV12 && - (fb->width < SKL_MIN_YUV_420_SRC_W || - fb->height < SKL_MIN_YUV_420_SRC_H || - (fb->width % 4) != 0 || (fb->height % 4) != 0)) { - DRM_DEBUG_KMS("src dimensions not correct for NV12\n"); - goto err; - } - for (i = 0; i < fb->format->num_planes; i++) { u32 stride_alignment; @@ -14974,7 +15195,7 @@ retry: * intermediate watermarks (since we don't trust the current * watermarks). */ - if (!HAS_GMCH_DISPLAY(dev_priv)) + if (!HAS_GMCH(dev_priv)) intel_state->skip_intermediate_wm = true; ret = intel_atomic_check(dev, state); @@ -15215,7 +15436,7 @@ int intel_modeset_init(struct drm_device *dev) * Note that we need to do this after reconstructing the BIOS fb's * since the watermark calculation done here will use pstate->fb. */ - if (!HAS_GMCH_DISPLAY(dev_priv)) + if (!HAS_GMCH(dev_priv)) sanitize_watermarks(dev); /* @@ -15408,6 +15629,15 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc, plane->base.type != DRM_PLANE_TYPE_PRIMARY) intel_plane_disable_noatomic(crtc, plane); } + + /* + * Disable any background color set by the BIOS, but enable the + * gamma and CSC to match how we program our planes. + */ + if (INTEL_GEN(dev_priv) >= 9) + I915_WRITE(SKL_BOTTOM_COLOR(crtc->pipe), + SKL_BOTTOM_COLOR_GAMMA_ENABLE | + SKL_BOTTOM_COLOR_CSC_ENABLE); } /* Adjust the state of the output pipe according to whether we @@ -15415,7 +15645,7 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc, if (crtc_state->base.active && !intel_crtc_has_encoders(crtc)) intel_crtc_disable_noatomic(&crtc->base, ctx); - if (crtc_state->base.active || HAS_GMCH_DISPLAY(dev_priv)) { + if (crtc_state->base.active || HAS_GMCH(dev_priv)) { /* * We start out with underrun reporting disabled to avoid races. * For correct bookkeeping mark this on active crtcs. @@ -15444,16 +15674,45 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc, } } +static bool has_bogus_dpll_config(const struct intel_crtc_state *crtc_state) +{ + struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); + + /* + * Some SNB BIOSen (eg. ASUS K53SV) are known to misprogram + * the hardware when a high res displays plugged in. DPLL P + * divider is zero, and the pipe timings are bonkers. We'll + * try to disable everything in that case. + * + * FIXME would be nice to be able to sanitize this state + * without several WARNs, but for now let's take the easy + * road. + */ + return IS_GEN(dev_priv, 6) && + crtc_state->base.active && + crtc_state->shared_dpll && + crtc_state->port_clock == 0; +} + static void intel_sanitize_encoder(struct intel_encoder *encoder) { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_connector *connector; + struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc); + struct intel_crtc_state *crtc_state = crtc ? + to_intel_crtc_state(crtc->base.state) : NULL; /* We need to check both for a crtc link (meaning that the * encoder is active and trying to read from a pipe) and the * pipe itself being active. */ - bool has_active_crtc = encoder->base.crtc && - to_intel_crtc(encoder->base.crtc)->active; + bool has_active_crtc = crtc_state && + crtc_state->base.active; + + if (crtc_state && has_bogus_dpll_config(crtc_state)) { + DRM_DEBUG_KMS("BIOS has misprogrammed the hardware. Disabling pipe %c\n", + pipe_name(crtc->pipe)); + has_active_crtc = false; + } connector = intel_encoder_find_connector(encoder); if (connector && !has_active_crtc) { @@ -15464,16 +15723,25 @@ static void intel_sanitize_encoder(struct intel_encoder *encoder) /* Connector is active, but has no active pipe. This is * fallout from our resume register restoring. Disable * the encoder manually again. */ - if (encoder->base.crtc) { - struct drm_crtc_state *crtc_state = encoder->base.crtc->state; + if (crtc_state) { + struct drm_encoder *best_encoder; DRM_DEBUG_KMS("[ENCODER:%d:%s] manually disabled\n", encoder->base.base.id, encoder->base.name); + + /* avoid oopsing in case the hooks consult best_encoder */ + best_encoder = connector->base.state->best_encoder; + connector->base.state->best_encoder = &encoder->base; + if (encoder->disable) - encoder->disable(encoder, to_intel_crtc_state(crtc_state), connector->base.state); + encoder->disable(encoder, crtc_state, + connector->base.state); if (encoder->post_disable) - encoder->post_disable(encoder, to_intel_crtc_state(crtc_state), connector->base.state); + encoder->post_disable(encoder, crtc_state, + connector->base.state); + + connector->base.state->best_encoder = best_encoder; } encoder->base.crtc = NULL; @@ -15505,19 +15773,25 @@ void i915_redisable_vga_power_on(struct drm_i915_private *dev_priv) void i915_redisable_vga(struct drm_i915_private *dev_priv) { - /* This function can be called both from intel_modeset_setup_hw_state or + intel_wakeref_t wakeref; + + /* + * This function can be called both from intel_modeset_setup_hw_state or * at a very early point in our resume sequence, where the power well * structures are not yet restored. Since this function is at a very * paranoid "someone might have enabled VGA while we were not looking" * level, just check if the power well is enabled instead of trying to * follow the "don't touch the power well if we don't need it" policy - * the rest of the driver uses. */ - if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_VGA)) + * the rest of the driver uses. + */ + wakeref = intel_display_power_get_if_enabled(dev_priv, + POWER_DOMAIN_VGA); + if (!wakeref) return; i915_redisable_vga_power_on(dev_priv); - intel_display_power_put(dev_priv, POWER_DOMAIN_VGA); + intel_display_power_put(dev_priv, POWER_DOMAIN_VGA, wakeref); } /* FIXME read out full plane state for all planes */ @@ -15817,12 +16091,13 @@ intel_modeset_setup_hw_state(struct drm_device *dev, struct drm_modeset_acquire_ctx *ctx) { struct drm_i915_private *dev_priv = to_i915(dev); - struct intel_crtc *crtc; struct intel_crtc_state *crtc_state; struct intel_encoder *encoder; + struct intel_crtc *crtc; + intel_wakeref_t wakeref; int i; - intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); + wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); intel_early_display_was(dev_priv); intel_modeset_readout_hw_state(dev); @@ -15838,10 +16113,12 @@ intel_modeset_setup_hw_state(struct drm_device *dev, * waits, so we need vblank interrupts restored beforehand. */ for_each_intel_crtc(&dev_priv->drm, crtc) { + crtc_state = to_intel_crtc_state(crtc->base.state); + drm_crtc_vblank_reset(&crtc->base); - if (crtc->base.state->active) - drm_crtc_vblank_on(&crtc->base); + if (crtc_state->base.active) + intel_crtc_vblank_on(crtc_state); } intel_sanitize_plane_mapping(dev_priv); @@ -15892,7 +16169,7 @@ intel_modeset_setup_hw_state(struct drm_device *dev, modeset_put_power_domains(dev_priv, put_domains); } - intel_display_power_put(dev_priv, POWER_DOMAIN_INIT); + intel_display_power_put(dev_priv, POWER_DOMAIN_INIT, wakeref); intel_fbc_init_pipe_state(dev_priv); } @@ -16115,7 +16392,7 @@ intel_display_capture_error_state(struct drm_i915_private *dev_priv) error->pipe[i].source = I915_READ(PIPESRC(i)); - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) error->pipe[i].stat = I915_READ(PIPESTAT(i)); } diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 6f0bb9a744fb..cf709835fb9a 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -429,7 +429,7 @@ static void intel_dp_set_common_rates(struct intel_dp *intel_dp) } static bool intel_dp_link_params_valid(struct intel_dp *intel_dp, int link_rate, - uint8_t lane_count) + u8 lane_count) { /* * FIXME: we need to synchronize the current link parameters with @@ -449,7 +449,7 @@ static bool intel_dp_link_params_valid(struct intel_dp *intel_dp, int link_rate, static bool intel_dp_can_link_train_fallback_for_edp(struct intel_dp *intel_dp, int link_rate, - uint8_t lane_count) + u8 lane_count) { const struct drm_display_mode *fixed_mode = intel_dp->attached_connector->panel.fixed_mode; @@ -464,7 +464,7 @@ static bool intel_dp_can_link_train_fallback_for_edp(struct intel_dp *intel_dp, } int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp, - int link_rate, uint8_t lane_count) + int link_rate, u8 lane_count) { int index; @@ -572,19 +572,19 @@ intel_dp_mode_valid(struct drm_connector *connector, return MODE_OK; } -uint32_t intel_dp_pack_aux(const uint8_t *src, int src_bytes) +u32 intel_dp_pack_aux(const u8 *src, int src_bytes) { - int i; - uint32_t v = 0; + int i; + u32 v = 0; if (src_bytes > 4) src_bytes = 4; for (i = 0; i < src_bytes; i++) - v |= ((uint32_t) src[i]) << ((3-i) * 8); + v |= ((u32)src[i]) << ((3 - i) * 8); return v; } -static void intel_dp_unpack_aux(uint32_t src, uint8_t *dst, int dst_bytes) +static void intel_dp_unpack_aux(u32 src, u8 *dst, int dst_bytes) { int i; if (dst_bytes > 4) @@ -601,30 +601,39 @@ intel_dp_init_panel_power_sequencer_registers(struct intel_dp *intel_dp, static void intel_dp_pps_init(struct intel_dp *intel_dp); -static void pps_lock(struct intel_dp *intel_dp) +static intel_wakeref_t +pps_lock(struct intel_dp *intel_dp) { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + intel_wakeref_t wakeref; /* * See intel_power_sequencer_reset() why we need * a power domain reference here. */ - intel_display_power_get(dev_priv, - intel_aux_power_domain(dp_to_dig_port(intel_dp))); + wakeref = intel_display_power_get(dev_priv, + intel_aux_power_domain(dp_to_dig_port(intel_dp))); mutex_lock(&dev_priv->pps_mutex); + + return wakeref; } -static void pps_unlock(struct intel_dp *intel_dp) +static intel_wakeref_t +pps_unlock(struct intel_dp *intel_dp, intel_wakeref_t wakeref) { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); mutex_unlock(&dev_priv->pps_mutex); - intel_display_power_put(dev_priv, - intel_aux_power_domain(dp_to_dig_port(intel_dp))); + intel_aux_power_domain(dp_to_dig_port(intel_dp)), + wakeref); + return 0; } +#define with_pps_lock(dp, wf) \ + for ((wf) = pps_lock(dp); (wf); (wf) = pps_unlock((dp), (wf))) + static void vlv_power_sequencer_kick(struct intel_dp *intel_dp) { @@ -634,7 +643,7 @@ vlv_power_sequencer_kick(struct intel_dp *intel_dp) bool pll_enabled, release_cl_override = false; enum dpio_phy phy = DPIO_PHY(pipe); enum dpio_channel ch = vlv_pipe_to_channel(pipe); - uint32_t DP; + u32 DP; if (WARN(I915_READ(intel_dp->output_reg) & DP_PORT_EN, "skipping pipe %c power sequencer kick due to port %c being active\n", @@ -973,30 +982,29 @@ static int edp_notify_handler(struct notifier_block *this, unsigned long code, struct intel_dp *intel_dp = container_of(this, typeof(* intel_dp), edp_notifier); struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); + intel_wakeref_t wakeref; if (!intel_dp_is_edp(intel_dp) || code != SYS_RESTART) return 0; - pps_lock(intel_dp); - - if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { - enum pipe pipe = vlv_power_sequencer_pipe(intel_dp); - i915_reg_t pp_ctrl_reg, pp_div_reg; - u32 pp_div; - - pp_ctrl_reg = PP_CONTROL(pipe); - pp_div_reg = PP_DIVISOR(pipe); - pp_div = I915_READ(pp_div_reg); - pp_div &= PP_REFERENCE_DIVIDER_MASK; - - /* 0x1F write to PP_DIV_REG sets max cycle delay */ - I915_WRITE(pp_div_reg, pp_div | 0x1F); - I915_WRITE(pp_ctrl_reg, PANEL_UNLOCK_REGS | PANEL_POWER_OFF); - msleep(intel_dp->panel_power_cycle_delay); + with_pps_lock(intel_dp, wakeref) { + if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { + enum pipe pipe = vlv_power_sequencer_pipe(intel_dp); + i915_reg_t pp_ctrl_reg, pp_div_reg; + u32 pp_div; + + pp_ctrl_reg = PP_CONTROL(pipe); + pp_div_reg = PP_DIVISOR(pipe); + pp_div = I915_READ(pp_div_reg); + pp_div &= PP_REFERENCE_DIVIDER_MASK; + + /* 0x1F write to PP_DIV_REG sets max cycle delay */ + I915_WRITE(pp_div_reg, pp_div | 0x1F); + I915_WRITE(pp_ctrl_reg, PANEL_UNLOCK_REGS); + msleep(intel_dp->panel_power_cycle_delay); + } } - pps_unlock(intel_dp); - return 0; } @@ -1042,17 +1050,21 @@ intel_dp_check_edp(struct intel_dp *intel_dp) } } -static uint32_t +static u32 intel_dp_aux_wait_done(struct intel_dp *intel_dp) { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); i915_reg_t ch_ctl = intel_dp->aux_ch_ctl_reg(intel_dp); - uint32_t status; + u32 status; bool done; #define C (((status = I915_READ_NOTRACE(ch_ctl)) & DP_AUX_CH_CTL_SEND_BUSY) == 0) done = wait_event_timeout(dev_priv->gmbus_wait_queue, C, msecs_to_jiffies_timeout(10)); + + /* just trace the final value */ + trace_i915_reg_rw(false, ch_ctl, status, sizeof(status), true); + if (!done) DRM_ERROR("dp aux hw did not signal timeout!\n"); #undef C @@ -1060,7 +1072,7 @@ intel_dp_aux_wait_done(struct intel_dp *intel_dp) return status; } -static uint32_t g4x_get_aux_clock_divider(struct intel_dp *intel_dp, int index) +static u32 g4x_get_aux_clock_divider(struct intel_dp *intel_dp, int index) { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); @@ -1074,7 +1086,7 @@ static uint32_t g4x_get_aux_clock_divider(struct intel_dp *intel_dp, int index) return DIV_ROUND_CLOSEST(dev_priv->rawclk_freq, 2000); } -static uint32_t ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index) +static u32 ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index) { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); @@ -1093,7 +1105,7 @@ static uint32_t ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index) return DIV_ROUND_CLOSEST(dev_priv->rawclk_freq, 2000); } -static uint32_t hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index) +static u32 hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index) { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); @@ -1110,7 +1122,7 @@ static uint32_t hsw_get_aux_clock_divider(struct intel_dp *intel_dp, int index) return ilk_get_aux_clock_divider(intel_dp, index); } -static uint32_t skl_get_aux_clock_divider(struct intel_dp *intel_dp, int index) +static u32 skl_get_aux_clock_divider(struct intel_dp *intel_dp, int index) { /* * SKL doesn't need us to program the AUX clock divider (Hardware will @@ -1120,14 +1132,14 @@ static uint32_t skl_get_aux_clock_divider(struct intel_dp *intel_dp, int index) return index ? 0 : 1; } -static uint32_t g4x_get_aux_send_ctl(struct intel_dp *intel_dp, - int send_bytes, - uint32_t aux_clock_divider) +static u32 g4x_get_aux_send_ctl(struct intel_dp *intel_dp, + int send_bytes, + u32 aux_clock_divider) { struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); struct drm_i915_private *dev_priv = to_i915(intel_dig_port->base.base.dev); - uint32_t precharge, timeout; + u32 precharge, timeout; if (IS_GEN(dev_priv, 6)) precharge = 3; @@ -1150,12 +1162,12 @@ static uint32_t g4x_get_aux_send_ctl(struct intel_dp *intel_dp, (aux_clock_divider << DP_AUX_CH_CTL_BIT_CLOCK_2X_SHIFT); } -static uint32_t skl_get_aux_send_ctl(struct intel_dp *intel_dp, - int send_bytes, - uint32_t unused) +static u32 skl_get_aux_send_ctl(struct intel_dp *intel_dp, + int send_bytes, + u32 unused) { struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); - uint32_t ret; + u32 ret; ret = DP_AUX_CH_CTL_SEND_BUSY | DP_AUX_CH_CTL_DONE | @@ -1175,25 +1187,26 @@ static uint32_t skl_get_aux_send_ctl(struct intel_dp *intel_dp, static int intel_dp_aux_xfer(struct intel_dp *intel_dp, - const uint8_t *send, int send_bytes, - uint8_t *recv, int recv_size, + const u8 *send, int send_bytes, + u8 *recv, int recv_size, u32 aux_send_ctl_flags) { struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); struct drm_i915_private *dev_priv = to_i915(intel_dig_port->base.base.dev); i915_reg_t ch_ctl, ch_data[5]; - uint32_t aux_clock_divider; + u32 aux_clock_divider; + intel_wakeref_t wakeref; int i, ret, recv_bytes; - uint32_t status; int try, clock = 0; + u32 status; bool vdd; ch_ctl = intel_dp->aux_ch_ctl_reg(intel_dp); for (i = 0; i < ARRAY_SIZE(ch_data); i++) ch_data[i] = intel_dp->aux_ch_data_reg(intel_dp, i); - pps_lock(intel_dp); + wakeref = pps_lock(intel_dp); /* * We will be called with VDD already enabled for dpcd/edid/oui reads. @@ -1218,6 +1231,8 @@ intel_dp_aux_xfer(struct intel_dp *intel_dp, break; msleep(1); } + /* just trace the final value */ + trace_i915_reg_rw(false, ch_ctl, status, sizeof(status), true); if (try == 3) { static u32 last_status = -1; @@ -1337,7 +1352,7 @@ out: if (vdd) edp_panel_vdd_off(intel_dp, false); - pps_unlock(intel_dp); + pps_unlock(intel_dp, wakeref); return ret; } @@ -1359,7 +1374,7 @@ static ssize_t intel_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg) { struct intel_dp *intel_dp = container_of(aux, struct intel_dp, aux); - uint8_t txbuf[20], rxbuf[20]; + u8 txbuf[20], rxbuf[20]; size_t txsize, rxsize; int ret; @@ -1692,7 +1707,7 @@ int intel_dp_rate_select(struct intel_dp *intel_dp, int rate) } void intel_dp_compute_rate(struct intel_dp *intel_dp, int port_clock, - uint8_t *link_bw, uint8_t *rate_select) + u8 *link_bw, u8 *rate_select) { /* eDP 1.4 rate select method. */ if (intel_dp->use_rate_select) { @@ -2132,7 +2147,7 @@ intel_dp_compute_config(struct intel_encoder *encoder, return ret; } - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) intel_gmch_panel_fitting(intel_crtc, pipe_config, conn_state->scaling_mode); else @@ -2143,7 +2158,7 @@ intel_dp_compute_config(struct intel_encoder *encoder, if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN) return -EINVAL; - if (HAS_GMCH_DISPLAY(dev_priv) && + if (HAS_GMCH(dev_priv) && adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) return -EINVAL; @@ -2207,7 +2222,7 @@ intel_dp_compute_config(struct intel_encoder *encoder, } void intel_dp_set_link_params(struct intel_dp *intel_dp, - int link_rate, uint8_t lane_count, + int link_rate, u8 lane_count, bool link_mst) { intel_dp->link_trained = false; @@ -2469,15 +2484,15 @@ static bool edp_panel_vdd_on(struct intel_dp *intel_dp) */ void intel_edp_panel_vdd_on(struct intel_dp *intel_dp) { + intel_wakeref_t wakeref; bool vdd; if (!intel_dp_is_edp(intel_dp)) return; - pps_lock(intel_dp); - vdd = edp_panel_vdd_on(intel_dp); - pps_unlock(intel_dp); - + vdd = false; + with_pps_lock(intel_dp, wakeref) + vdd = edp_panel_vdd_on(intel_dp); I915_STATE_WARN(!vdd, "eDP port %c VDD already requested on\n", port_name(dp_to_dig_port(intel_dp)->base.port)); } @@ -2516,19 +2531,21 @@ static void edp_panel_vdd_off_sync(struct intel_dp *intel_dp) if ((pp & PANEL_POWER_ON) == 0) intel_dp->panel_power_off_time = ktime_get_boottime(); - intel_display_power_put(dev_priv, - intel_aux_power_domain(intel_dig_port)); + intel_display_power_put_unchecked(dev_priv, + intel_aux_power_domain(intel_dig_port)); } static void edp_panel_vdd_work(struct work_struct *__work) { - struct intel_dp *intel_dp = container_of(to_delayed_work(__work), - struct intel_dp, panel_vdd_work); + struct intel_dp *intel_dp = + container_of(to_delayed_work(__work), + struct intel_dp, panel_vdd_work); + intel_wakeref_t wakeref; - pps_lock(intel_dp); - if (!intel_dp->want_panel_vdd) - edp_panel_vdd_off_sync(intel_dp); - pps_unlock(intel_dp); + with_pps_lock(intel_dp, wakeref) { + if (!intel_dp->want_panel_vdd) + edp_panel_vdd_off_sync(intel_dp); + } } static void edp_panel_vdd_schedule_off(struct intel_dp *intel_dp) @@ -2618,12 +2635,13 @@ static void edp_panel_on(struct intel_dp *intel_dp) void intel_edp_panel_on(struct intel_dp *intel_dp) { + intel_wakeref_t wakeref; + if (!intel_dp_is_edp(intel_dp)) return; - pps_lock(intel_dp); - edp_panel_on(intel_dp); - pps_unlock(intel_dp); + with_pps_lock(intel_dp, wakeref) + edp_panel_on(intel_dp); } @@ -2662,25 +2680,25 @@ static void edp_panel_off(struct intel_dp *intel_dp) intel_dp->panel_power_off_time = ktime_get_boottime(); /* We got a reference when we enabled the VDD. */ - intel_display_power_put(dev_priv, intel_aux_power_domain(dig_port)); + intel_display_power_put_unchecked(dev_priv, intel_aux_power_domain(dig_port)); } void intel_edp_panel_off(struct intel_dp *intel_dp) { + intel_wakeref_t wakeref; + if (!intel_dp_is_edp(intel_dp)) return; - pps_lock(intel_dp); - edp_panel_off(intel_dp); - pps_unlock(intel_dp); + with_pps_lock(intel_dp, wakeref) + edp_panel_off(intel_dp); } /* Enable backlight in the panel power control. */ static void _intel_edp_backlight_on(struct intel_dp *intel_dp) { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); - u32 pp; - i915_reg_t pp_ctrl_reg; + intel_wakeref_t wakeref; /* * If we enable the backlight right away following a panel power @@ -2690,17 +2708,16 @@ static void _intel_edp_backlight_on(struct intel_dp *intel_dp) */ wait_backlight_on(intel_dp); - pps_lock(intel_dp); - - pp = ironlake_get_pp_control(intel_dp); - pp |= EDP_BLC_ENABLE; + with_pps_lock(intel_dp, wakeref) { + i915_reg_t pp_ctrl_reg = _pp_ctrl_reg(intel_dp); + u32 pp; - pp_ctrl_reg = _pp_ctrl_reg(intel_dp); + pp = ironlake_get_pp_control(intel_dp); + pp |= EDP_BLC_ENABLE; - I915_WRITE(pp_ctrl_reg, pp); - POSTING_READ(pp_ctrl_reg); - - pps_unlock(intel_dp); + I915_WRITE(pp_ctrl_reg, pp); + POSTING_READ(pp_ctrl_reg); + } } /* Enable backlight PWM and backlight PP control. */ @@ -2722,23 +2739,21 @@ void intel_edp_backlight_on(const struct intel_crtc_state *crtc_state, static void _intel_edp_backlight_off(struct intel_dp *intel_dp) { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); - u32 pp; - i915_reg_t pp_ctrl_reg; + intel_wakeref_t wakeref; if (!intel_dp_is_edp(intel_dp)) return; - pps_lock(intel_dp); - - pp = ironlake_get_pp_control(intel_dp); - pp &= ~EDP_BLC_ENABLE; - - pp_ctrl_reg = _pp_ctrl_reg(intel_dp); + with_pps_lock(intel_dp, wakeref) { + i915_reg_t pp_ctrl_reg = _pp_ctrl_reg(intel_dp); + u32 pp; - I915_WRITE(pp_ctrl_reg, pp); - POSTING_READ(pp_ctrl_reg); + pp = ironlake_get_pp_control(intel_dp); + pp &= ~EDP_BLC_ENABLE; - pps_unlock(intel_dp); + I915_WRITE(pp_ctrl_reg, pp); + POSTING_READ(pp_ctrl_reg); + } intel_dp->last_backlight_off = jiffies; edp_wait_backlight_off(intel_dp); @@ -2766,12 +2781,12 @@ static void intel_edp_backlight_power(struct intel_connector *connector, bool enable) { struct intel_dp *intel_dp = intel_attached_dp(&connector->base); + intel_wakeref_t wakeref; bool is_enabled; - pps_lock(intel_dp); - is_enabled = ironlake_get_pp_control(intel_dp) & EDP_BLC_ENABLE; - pps_unlock(intel_dp); - + is_enabled = false; + with_pps_lock(intel_dp, wakeref) + is_enabled = ironlake_get_pp_control(intel_dp) & EDP_BLC_ENABLE; if (is_enabled == enable) return; @@ -2988,16 +3003,18 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base); + intel_wakeref_t wakeref; bool ret; - if (!intel_display_power_get_if_enabled(dev_priv, - encoder->power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + encoder->power_domain); + if (!wakeref) return false; ret = intel_dp_port_enabled(dev_priv, intel_dp->output_reg, encoder->port, pipe); - intel_display_power_put(dev_priv, encoder->power_domain); + intel_display_power_put(dev_priv, encoder->power_domain, wakeref); return ret; } @@ -3165,20 +3182,20 @@ static void chv_post_disable_dp(struct intel_encoder *encoder, static void _intel_dp_set_link_train(struct intel_dp *intel_dp, - uint32_t *DP, - uint8_t dp_train_pat) + u32 *DP, + u8 dp_train_pat) { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); enum port port = intel_dig_port->base.port; - uint8_t train_pat_mask = drm_dp_training_pattern_mask(intel_dp->dpcd); + u8 train_pat_mask = drm_dp_training_pattern_mask(intel_dp->dpcd); if (dp_train_pat & train_pat_mask) DRM_DEBUG_KMS("Using DP training pattern TPS%d\n", dp_train_pat & train_pat_mask); if (HAS_DDI(dev_priv)) { - uint32_t temp = I915_READ(DP_TP_CTL(port)); + u32 temp = I915_READ(DP_TP_CTL(port)); if (dp_train_pat & DP_LINK_SCRAMBLING_DISABLE) temp |= DP_TP_CTL_SCRAMBLE_DISABLE; @@ -3277,24 +3294,23 @@ static void intel_enable_dp(struct intel_encoder *encoder, struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base); struct intel_crtc *crtc = to_intel_crtc(pipe_config->base.crtc); - uint32_t dp_reg = I915_READ(intel_dp->output_reg); + u32 dp_reg = I915_READ(intel_dp->output_reg); enum pipe pipe = crtc->pipe; + intel_wakeref_t wakeref; if (WARN_ON(dp_reg & DP_PORT_EN)) return; - pps_lock(intel_dp); - - if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) - vlv_init_panel_power_sequencer(encoder, pipe_config); - - intel_dp_enable_port(intel_dp, pipe_config); + with_pps_lock(intel_dp, wakeref) { + if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + vlv_init_panel_power_sequencer(encoder, pipe_config); - edp_panel_vdd_on(intel_dp); - edp_panel_on(intel_dp); - edp_panel_vdd_off(intel_dp, true); + intel_dp_enable_port(intel_dp, pipe_config); - pps_unlock(intel_dp); + edp_panel_vdd_on(intel_dp); + edp_panel_on(intel_dp); + edp_panel_vdd_off(intel_dp, true); + } if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { unsigned int lane_mask = 0x0; @@ -3497,14 +3513,14 @@ static void chv_dp_post_pll_disable(struct intel_encoder *encoder, * link status information */ bool -intel_dp_get_link_status(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_STATUS_SIZE]) +intel_dp_get_link_status(struct intel_dp *intel_dp, u8 link_status[DP_LINK_STATUS_SIZE]) { return drm_dp_dpcd_read(&intel_dp->aux, DP_LANE0_1_STATUS, link_status, DP_LINK_STATUS_SIZE) == DP_LINK_STATUS_SIZE; } /* These are source-specific values. */ -uint8_t +u8 intel_dp_voltage_max(struct intel_dp *intel_dp) { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); @@ -3523,8 +3539,8 @@ intel_dp_voltage_max(struct intel_dp *intel_dp) return DP_TRAIN_VOLTAGE_SWING_LEVEL_2; } -uint8_t -intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing) +u8 +intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, u8 voltage_swing) { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; @@ -3569,12 +3585,12 @@ intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing) } } -static uint32_t vlv_signal_levels(struct intel_dp *intel_dp) +static u32 vlv_signal_levels(struct intel_dp *intel_dp) { struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; unsigned long demph_reg_value, preemph_reg_value, uniqtranscale_reg_value; - uint8_t train_set = intel_dp->train_set[0]; + u8 train_set = intel_dp->train_set[0]; switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) { case DP_TRAIN_PRE_EMPH_LEVEL_0: @@ -3655,12 +3671,12 @@ static uint32_t vlv_signal_levels(struct intel_dp *intel_dp) return 0; } -static uint32_t chv_signal_levels(struct intel_dp *intel_dp) +static u32 chv_signal_levels(struct intel_dp *intel_dp) { struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base; u32 deemph_reg_value, margin_reg_value; bool uniq_trans_scale = false; - uint8_t train_set = intel_dp->train_set[0]; + u8 train_set = intel_dp->train_set[0]; switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) { case DP_TRAIN_PRE_EMPH_LEVEL_0: @@ -3738,10 +3754,10 @@ static uint32_t chv_signal_levels(struct intel_dp *intel_dp) return 0; } -static uint32_t -g4x_signal_levels(uint8_t train_set) +static u32 +g4x_signal_levels(u8 train_set) { - uint32_t signal_levels = 0; + u32 signal_levels = 0; switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) { case DP_TRAIN_VOLTAGE_SWING_LEVEL_0: @@ -3777,8 +3793,8 @@ g4x_signal_levels(uint8_t train_set) } /* SNB CPU eDP voltage swing and pre-emphasis control */ -static uint32_t -snb_cpu_edp_signal_levels(uint8_t train_set) +static u32 +snb_cpu_edp_signal_levels(u8 train_set) { int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK | DP_TRAIN_PRE_EMPHASIS_MASK); @@ -3805,8 +3821,8 @@ snb_cpu_edp_signal_levels(uint8_t train_set) } /* IVB CPU eDP voltage swing and pre-emphasis control */ -static uint32_t -ivb_cpu_edp_signal_levels(uint8_t train_set) +static u32 +ivb_cpu_edp_signal_levels(u8 train_set) { int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK | DP_TRAIN_PRE_EMPHASIS_MASK); @@ -3841,8 +3857,8 @@ intel_dp_set_signal_levels(struct intel_dp *intel_dp) struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); enum port port = intel_dig_port->base.port; - uint32_t signal_levels, mask = 0; - uint8_t train_set = intel_dp->train_set[0]; + u32 signal_levels, mask = 0; + u8 train_set = intel_dp->train_set[0]; if (IS_GEN9_LP(dev_priv) || INTEL_GEN(dev_priv) >= 10) { signal_levels = bxt_signal_levels(intel_dp); @@ -3881,7 +3897,7 @@ intel_dp_set_signal_levels(struct intel_dp *intel_dp) void intel_dp_program_link_training_pattern(struct intel_dp *intel_dp, - uint8_t dp_train_pat) + u8 dp_train_pat) { struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); struct drm_i915_private *dev_priv = @@ -3898,7 +3914,7 @@ void intel_dp_set_idle_link_train(struct intel_dp *intel_dp) struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); enum port port = intel_dig_port->base.port; - uint32_t val; + u32 val; if (!HAS_DDI(dev_priv)) return; @@ -3933,7 +3949,7 @@ intel_dp_link_down(struct intel_encoder *encoder, struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base); struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->base.crtc); enum port port = encoder->port; - uint32_t DP = intel_dp->DP; + u32 DP = intel_dp->DP; if (WARN_ON(HAS_DDI(dev_priv))) return; @@ -3992,9 +4008,10 @@ intel_dp_link_down(struct intel_encoder *encoder, intel_dp->DP = DP; if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) { - pps_lock(intel_dp); - intel_dp->active_pipe = INVALID_PIPE; - pps_unlock(intel_dp); + intel_wakeref_t wakeref; + + with_pps_lock(intel_dp, wakeref) + intel_dp->active_pipe = INVALID_PIPE; } } @@ -4273,7 +4290,7 @@ intel_dp_get_sink_irq_esi(struct intel_dp *intel_dp, u8 *sink_irq_vector) DP_DPRX_ESI_LEN; } -u16 intel_dp_dsc_get_output_bpp(int link_clock, uint8_t lane_count, +u16 intel_dp_dsc_get_output_bpp(int link_clock, u8 lane_count, int mode_clock, int mode_hdisplay) { u16 bits_per_pixel, max_bpp_small_joiner_ram; @@ -4340,7 +4357,7 @@ u8 intel_dp_dsc_get_slice_count(struct intel_dp *intel_dp, return 0; } /* Also take into account max slice width */ - min_slice_count = min_t(uint8_t, min_slice_count, + min_slice_count = min_t(u8, min_slice_count, DIV_ROUND_UP(mode_hdisplay, max_slice_width)); @@ -4358,11 +4375,11 @@ u8 intel_dp_dsc_get_slice_count(struct intel_dp *intel_dp, return 0; } -static uint8_t intel_dp_autotest_link_training(struct intel_dp *intel_dp) +static u8 intel_dp_autotest_link_training(struct intel_dp *intel_dp) { int status = 0; int test_link_rate; - uint8_t test_lane_count, test_link_bw; + u8 test_lane_count, test_link_bw; /* (DP CTS 1.2) * 4.3.1.11 */ @@ -4395,10 +4412,10 @@ static uint8_t intel_dp_autotest_link_training(struct intel_dp *intel_dp) return DP_TEST_ACK; } -static uint8_t intel_dp_autotest_video_pattern(struct intel_dp *intel_dp) +static u8 intel_dp_autotest_video_pattern(struct intel_dp *intel_dp) { - uint8_t test_pattern; - uint8_t test_misc; + u8 test_pattern; + u8 test_misc; __be16 h_width, v_height; int status = 0; @@ -4456,9 +4473,9 @@ static uint8_t intel_dp_autotest_video_pattern(struct intel_dp *intel_dp) return DP_TEST_ACK; } -static uint8_t intel_dp_autotest_edid(struct intel_dp *intel_dp) +static u8 intel_dp_autotest_edid(struct intel_dp *intel_dp) { - uint8_t test_result = DP_TEST_ACK; + u8 test_result = DP_TEST_ACK; struct intel_connector *intel_connector = intel_dp->attached_connector; struct drm_connector *connector = &intel_connector->base; @@ -4500,16 +4517,16 @@ static uint8_t intel_dp_autotest_edid(struct intel_dp *intel_dp) return test_result; } -static uint8_t intel_dp_autotest_phy_pattern(struct intel_dp *intel_dp) +static u8 intel_dp_autotest_phy_pattern(struct intel_dp *intel_dp) { - uint8_t test_result = DP_TEST_NAK; + u8 test_result = DP_TEST_NAK; return test_result; } static void intel_dp_handle_test_request(struct intel_dp *intel_dp) { - uint8_t response = DP_TEST_NAK; - uint8_t request = 0; + u8 response = DP_TEST_NAK; + u8 request = 0; int status; status = drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_REQUEST, &request); @@ -4597,12 +4614,10 @@ go_again: return ret; } else { - struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); DRM_DEBUG_KMS("failed to get ESI - device may have failed\n"); intel_dp->is_mst = false; - drm_dp_mst_topology_mgr_set_mst(&intel_dp->mst_mgr, intel_dp->is_mst); - /* send a hotplug event */ - drm_kms_helper_hotplug_event(intel_dig_port->base.base.dev); + drm_dp_mst_topology_mgr_set_mst(&intel_dp->mst_mgr, + intel_dp->is_mst); } } return -EINVAL; @@ -4835,8 +4850,8 @@ static enum drm_connector_status intel_dp_detect_dpcd(struct intel_dp *intel_dp) { struct intel_lspcon *lspcon = dp_to_lspcon(intel_dp); - uint8_t *dpcd = intel_dp->dpcd; - uint8_t type; + u8 *dpcd = intel_dp->dpcd; + u8 type; if (lspcon->active) lspcon_resume(lspcon); @@ -5289,7 +5304,7 @@ bool intel_digital_port_connected(struct intel_encoder *encoder) { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); - if (HAS_GMCH_DISPLAY(dev_priv)) { + if (HAS_GMCH(dev_priv)) { if (IS_GM45(dev_priv)) return gm45_digital_port_connected(encoder); else @@ -5370,12 +5385,13 @@ intel_dp_detect(struct drm_connector *connector, enum drm_connector_status status; enum intel_display_power_domain aux_domain = intel_aux_power_domain(dig_port); + intel_wakeref_t wakeref; DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n", connector->base.id, connector->name); WARN_ON(!drm_modeset_is_locked(&dev_priv->drm.mode_config.connection_mutex)); - intel_display_power_get(dev_priv, aux_domain); + wakeref = intel_display_power_get(dev_priv, aux_domain); /* Can't disconnect eDP */ if (intel_dp_is_edp(intel_dp)) @@ -5441,7 +5457,7 @@ intel_dp_detect(struct drm_connector *connector, ret = intel_dp_retrain_link(encoder, ctx); if (ret) { - intel_display_power_put(dev_priv, aux_domain); + intel_display_power_put(dev_priv, aux_domain, wakeref); return ret; } } @@ -5465,7 +5481,7 @@ out: if (status != connector_status_connected && !intel_dp->is_mst) intel_dp_unset_edid(intel_dp); - intel_display_power_put(dev_priv, aux_domain); + intel_display_power_put(dev_priv, aux_domain, wakeref); return status; } @@ -5478,6 +5494,7 @@ intel_dp_force(struct drm_connector *connector) struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev); enum intel_display_power_domain aux_domain = intel_aux_power_domain(dig_port); + intel_wakeref_t wakeref; DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n", connector->base.id, connector->name); @@ -5486,11 +5503,11 @@ intel_dp_force(struct drm_connector *connector) if (connector->status != connector_status_connected) return; - intel_display_power_get(dev_priv, aux_domain); + wakeref = intel_display_power_get(dev_priv, aux_domain); intel_dp_set_edid(intel_dp); - intel_display_power_put(dev_priv, aux_domain); + intel_display_power_put(dev_priv, aux_domain, wakeref); } static int intel_dp_get_modes(struct drm_connector *connector) @@ -5562,14 +5579,15 @@ void intel_dp_encoder_flush_work(struct drm_encoder *encoder) intel_dp_mst_encoder_cleanup(intel_dig_port); if (intel_dp_is_edp(intel_dp)) { + intel_wakeref_t wakeref; + cancel_delayed_work_sync(&intel_dp->panel_vdd_work); /* * vdd might still be enabled do to the delayed vdd off. * Make sure vdd is actually turned off here. */ - pps_lock(intel_dp); - edp_panel_vdd_off_sync(intel_dp); - pps_unlock(intel_dp); + with_pps_lock(intel_dp, wakeref) + edp_panel_vdd_off_sync(intel_dp); if (intel_dp->edp_notifier.notifier_call) { unregister_reboot_notifier(&intel_dp->edp_notifier); @@ -5591,6 +5609,7 @@ static void intel_dp_encoder_destroy(struct drm_encoder *encoder) void intel_dp_encoder_suspend(struct intel_encoder *intel_encoder) { struct intel_dp *intel_dp = enc_to_intel_dp(&intel_encoder->base); + intel_wakeref_t wakeref; if (!intel_dp_is_edp(intel_dp)) return; @@ -5600,9 +5619,8 @@ void intel_dp_encoder_suspend(struct intel_encoder *intel_encoder) * Make sure vdd is actually turned off here. */ cancel_delayed_work_sync(&intel_dp->panel_vdd_work); - pps_lock(intel_dp); - edp_panel_vdd_off_sync(intel_dp); - pps_unlock(intel_dp); + with_pps_lock(intel_dp, wakeref) + edp_panel_vdd_off_sync(intel_dp); } static @@ -5615,7 +5633,7 @@ int intel_dp_hdcp_write_an_aksv(struct intel_digital_port *intel_dig_port, .address = DP_AUX_HDCP_AKSV, .size = DRM_HDCP_KSV_LEN, }; - uint8_t txbuf[HEADER_SIZE + DRM_HDCP_KSV_LEN] = {}, rxbuf[2], reply = 0; + u8 txbuf[HEADER_SIZE + DRM_HDCP_KSV_LEN] = {}, rxbuf[2], reply = 0; ssize_t dpcd_ret; int ret; @@ -5883,6 +5901,7 @@ void intel_dp_encoder_reset(struct drm_encoder *encoder) struct drm_i915_private *dev_priv = to_i915(encoder->dev); struct intel_dp *intel_dp = enc_to_intel_dp(encoder); struct intel_lspcon *lspcon = dp_to_lspcon(intel_dp); + intel_wakeref_t wakeref; if (!HAS_DDI(dev_priv)) intel_dp->DP = I915_READ(intel_dp->output_reg); @@ -5892,18 +5911,19 @@ void intel_dp_encoder_reset(struct drm_encoder *encoder) intel_dp->reset_link_params = true; - pps_lock(intel_dp); - - if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) - intel_dp->active_pipe = vlv_active_pipe(intel_dp); + with_pps_lock(intel_dp, wakeref) { + if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + intel_dp->active_pipe = vlv_active_pipe(intel_dp); - if (intel_dp_is_edp(intel_dp)) { - /* Reinit the power sequencer, in case BIOS did something with it. */ - intel_dp_pps_init(intel_dp); - intel_edp_panel_vdd_sanitize(intel_dp); + if (intel_dp_is_edp(intel_dp)) { + /* + * Reinit the power sequencer, in case BIOS did + * something nasty with it. + */ + intel_dp_pps_init(intel_dp); + intel_edp_panel_vdd_sanitize(intel_dp); + } } - - pps_unlock(intel_dp); } static const struct drm_connector_funcs intel_dp_connector_funcs = { @@ -5936,6 +5956,7 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd) struct intel_dp *intel_dp = &intel_dig_port->dp; struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); enum irqreturn ret = IRQ_NONE; + intel_wakeref_t wakeref; if (long_hpd && intel_dig_port->base.type == INTEL_OUTPUT_EDP) { /* @@ -5958,8 +5979,8 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd) return IRQ_NONE; } - intel_display_power_get(dev_priv, - intel_aux_power_domain(intel_dig_port)); + wakeref = intel_display_power_get(dev_priv, + intel_aux_power_domain(intel_dig_port)); if (intel_dp->is_mst) { if (intel_dp_check_mst_status(intel_dp) == -EINVAL) { @@ -5989,7 +6010,8 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd) put_power: intel_display_power_put(dev_priv, - intel_aux_power_domain(intel_dig_port)); + intel_aux_power_domain(intel_dig_port), + wakeref); return ret; } @@ -6020,7 +6042,7 @@ intel_dp_add_properties(struct intel_dp *intel_dp, struct drm_connector *connect intel_attach_force_audio_property(connector); intel_attach_broadcast_rgb_property(connector); - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) drm_connector_attach_max_bpc_property(connector, 6, 10); else if (INTEL_GEN(dev_priv) >= 5) drm_connector_attach_max_bpc_property(connector, 6, 12); @@ -6029,7 +6051,7 @@ intel_dp_add_properties(struct intel_dp *intel_dp, struct drm_connector *connect u32 allowed_scalers; allowed_scalers = BIT(DRM_MODE_SCALE_ASPECT) | BIT(DRM_MODE_SCALE_FULLSCREEN); - if (!HAS_GMCH_DISPLAY(dev_priv)) + if (!HAS_GMCH(dev_priv)) allowed_scalers |= BIT(DRM_MODE_SCALE_CENTER); drm_connector_attach_scaling_mode_property(connector, allowed_scalers); @@ -6697,8 +6719,9 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp, struct drm_display_mode *downclock_mode = NULL; bool has_dpcd; struct drm_display_mode *scan; - struct edid *edid; enum pipe pipe = INVALID_PIPE; + intel_wakeref_t wakeref; + struct edid *edid; if (!intel_dp_is_edp(intel_dp)) return true; @@ -6718,13 +6741,11 @@ static bool intel_edp_init_connector(struct intel_dp *intel_dp, return false; } - pps_lock(intel_dp); - - intel_dp_init_panel_power_timestamps(intel_dp); - intel_dp_pps_init(intel_dp); - intel_edp_panel_vdd_sanitize(intel_dp); - - pps_unlock(intel_dp); + with_pps_lock(intel_dp, wakeref) { + intel_dp_init_panel_power_timestamps(intel_dp); + intel_dp_pps_init(intel_dp); + intel_edp_panel_vdd_sanitize(intel_dp); + } /* Cache DPCD and EDID for edp. */ has_dpcd = intel_edp_init_dpcd(intel_dp); @@ -6809,9 +6830,8 @@ out_vdd_off: * vdd might still be enabled do to the delayed vdd off. * Make sure vdd is actually turned off here. */ - pps_lock(intel_dp); - edp_panel_vdd_off_sync(intel_dp); - pps_unlock(intel_dp); + with_pps_lock(intel_dp, wakeref) + edp_panel_vdd_off_sync(intel_dp); return false; } @@ -6903,7 +6923,7 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port, drm_connector_init(dev, connector, &intel_dp_connector_funcs, type); drm_connector_helper_add(connector, &intel_dp_connector_helper_funcs); - if (!HAS_GMCH_DISPLAY(dev_priv)) + if (!HAS_GMCH(dev_priv)) connector->interlace_allowed = true; connector->doublescan_allowed = 0; @@ -6985,6 +7005,7 @@ bool intel_dp_init(struct drm_i915_private *dev_priv, intel_encoder->compute_config = intel_dp_compute_config; intel_encoder->get_hw_state = intel_dp_get_hw_state; intel_encoder->get_config = intel_dp_get_config; + intel_encoder->update_pipe = intel_panel_update_backlight; intel_encoder->suspend = intel_dp_encoder_suspend; if (IS_CHERRYVIEW(dev_priv)) { intel_encoder->pre_pll_enable = chv_dp_pre_pll_enable; @@ -7079,7 +7100,10 @@ void intel_dp_mst_resume(struct drm_i915_private *dev_priv) continue; ret = drm_dp_mst_topology_mgr_resume(&intel_dp->mst_mgr); - if (ret) - intel_dp_check_mst_status(intel_dp); + if (ret) { + intel_dp->is_mst = false; + drm_dp_mst_topology_mgr_set_mst(&intel_dp->mst_mgr, + false); + } } } diff --git a/drivers/gpu/drm/i915/intel_dp_link_training.c b/drivers/gpu/drm/i915/intel_dp_link_training.c index 30be0e39bd5f..b59c87daa4f7 100644 --- a/drivers/gpu/drm/i915/intel_dp_link_training.c +++ b/drivers/gpu/drm/i915/intel_dp_link_training.c @@ -24,7 +24,7 @@ #include "intel_drv.h" static void -intel_dp_dump_link_status(const uint8_t link_status[DP_LINK_STATUS_SIZE]) +intel_dp_dump_link_status(const u8 link_status[DP_LINK_STATUS_SIZE]) { DRM_DEBUG_KMS("ln0_1:0x%x ln2_3:0x%x align:0x%x sink:0x%x adj_req0_1:0x%x adj_req2_3:0x%x", @@ -34,17 +34,17 @@ intel_dp_dump_link_status(const uint8_t link_status[DP_LINK_STATUS_SIZE]) static void intel_get_adjust_train(struct intel_dp *intel_dp, - const uint8_t link_status[DP_LINK_STATUS_SIZE]) + const u8 link_status[DP_LINK_STATUS_SIZE]) { - uint8_t v = 0; - uint8_t p = 0; + u8 v = 0; + u8 p = 0; int lane; - uint8_t voltage_max; - uint8_t preemph_max; + u8 voltage_max; + u8 preemph_max; for (lane = 0; lane < intel_dp->lane_count; lane++) { - uint8_t this_v = drm_dp_get_adjust_request_voltage(link_status, lane); - uint8_t this_p = drm_dp_get_adjust_request_pre_emphasis(link_status, lane); + u8 this_v = drm_dp_get_adjust_request_voltage(link_status, lane); + u8 this_p = drm_dp_get_adjust_request_pre_emphasis(link_status, lane); if (this_v > v) v = this_v; @@ -66,9 +66,9 @@ intel_get_adjust_train(struct intel_dp *intel_dp, static bool intel_dp_set_link_train(struct intel_dp *intel_dp, - uint8_t dp_train_pat) + u8 dp_train_pat) { - uint8_t buf[sizeof(intel_dp->train_set) + 1]; + u8 buf[sizeof(intel_dp->train_set) + 1]; int ret, len; intel_dp_program_link_training_pattern(intel_dp, dp_train_pat); @@ -92,7 +92,7 @@ intel_dp_set_link_train(struct intel_dp *intel_dp, static bool intel_dp_reset_link_train(struct intel_dp *intel_dp, - uint8_t dp_train_pat) + u8 dp_train_pat) { memset(intel_dp->train_set, 0, sizeof(intel_dp->train_set)); intel_dp_set_signal_levels(intel_dp); @@ -128,11 +128,11 @@ static bool intel_dp_link_max_vswing_reached(struct intel_dp *intel_dp) static bool intel_dp_link_training_clock_recovery(struct intel_dp *intel_dp) { - uint8_t voltage; + u8 voltage; int voltage_tries, cr_tries, max_cr_tries; bool max_vswing_reached = false; - uint8_t link_config[2]; - uint8_t link_bw, rate_select; + u8 link_config[2]; + u8 link_bw, rate_select; if (intel_dp->prepare_link_retrain) intel_dp->prepare_link_retrain(intel_dp); @@ -186,7 +186,7 @@ intel_dp_link_training_clock_recovery(struct intel_dp *intel_dp) voltage_tries = 1; for (cr_tries = 0; cr_tries < max_cr_tries; ++cr_tries) { - uint8_t link_status[DP_LINK_STATUS_SIZE]; + u8 link_status[DP_LINK_STATUS_SIZE]; drm_dp_link_train_clock_recovery_delay(intel_dp->dpcd); @@ -282,7 +282,7 @@ intel_dp_link_training_channel_equalization(struct intel_dp *intel_dp) { int tries; u32 training_pattern; - uint8_t link_status[DP_LINK_STATUS_SIZE]; + u8 link_status[DP_LINK_STATUS_SIZE]; bool channel_eq = false; training_pattern = intel_dp_training_pattern(intel_dp); diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c index 5d8471f080d9..fb67cd931117 100644 --- a/drivers/gpu/drm/i915/intel_dp_mst.c +++ b/drivers/gpu/drm/i915/intel_dp_mst.c @@ -242,7 +242,7 @@ static void intel_mst_pre_enable_dp(struct intel_encoder *encoder, struct intel_connector *connector = to_intel_connector(conn_state->connector); int ret; - uint32_t temp; + u32 temp; /* MST encoders are bound to a crtc, not to a connector, * force the mapping here for get_hw_state. diff --git a/drivers/gpu/drm/i915/intel_dpio_phy.c b/drivers/gpu/drm/i915/intel_dpio_phy.c index 3c7f10d17658..95cb8b154f87 100644 --- a/drivers/gpu/drm/i915/intel_dpio_phy.c +++ b/drivers/gpu/drm/i915/intel_dpio_phy.c @@ -413,7 +413,7 @@ static void _bxt_ddi_phy_init(struct drm_i915_private *dev_priv, } if (phy_info->rcomp_phy != -1) { - uint32_t grc_code; + u32 grc_code; bxt_phy_wait_grc_done(dev_priv, phy_info->rcomp_phy); @@ -445,7 +445,7 @@ static void _bxt_ddi_phy_init(struct drm_i915_private *dev_priv, void bxt_ddi_phy_uninit(struct drm_i915_private *dev_priv, enum dpio_phy phy) { const struct bxt_ddi_phy_info *phy_info; - uint32_t val; + u32 val; phy_info = bxt_get_phy_info(dev_priv, phy); @@ -515,7 +515,7 @@ bool bxt_ddi_phy_verify_state(struct drm_i915_private *dev_priv, enum dpio_phy phy) { const struct bxt_ddi_phy_info *phy_info; - uint32_t mask; + u32 mask; bool ok; phy_info = bxt_get_phy_info(dev_priv, phy); @@ -567,8 +567,8 @@ bool bxt_ddi_phy_verify_state(struct drm_i915_private *dev_priv, #undef _CHK } -uint8_t -bxt_ddi_phy_calc_lane_lat_optim_mask(uint8_t lane_count) +u8 +bxt_ddi_phy_calc_lane_lat_optim_mask(u8 lane_count) { switch (lane_count) { case 1: @@ -585,7 +585,7 @@ bxt_ddi_phy_calc_lane_lat_optim_mask(uint8_t lane_count) } void bxt_ddi_phy_set_lane_optim_mask(struct intel_encoder *encoder, - uint8_t lane_lat_optim_mask) + u8 lane_lat_optim_mask) { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); enum port port = encoder->port; @@ -610,7 +610,7 @@ void bxt_ddi_phy_set_lane_optim_mask(struct intel_encoder *encoder, } } -uint8_t +u8 bxt_ddi_phy_get_lane_lat_optim_mask(struct intel_encoder *encoder) { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); @@ -618,7 +618,7 @@ bxt_ddi_phy_get_lane_lat_optim_mask(struct intel_encoder *encoder) enum dpio_phy phy; enum dpio_channel ch; int lane; - uint8_t mask; + u8 mask; bxt_port_to_phy_channel(dev_priv, port, &phy, &ch); @@ -739,7 +739,7 @@ void chv_data_lane_soft_reset(struct intel_encoder *encoder, enum dpio_channel ch = vlv_dport_to_channel(enc_to_dig_port(&encoder->base)); struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc); enum pipe pipe = crtc->pipe; - uint32_t val; + u32 val; val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW0(ch)); if (reset) diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.c b/drivers/gpu/drm/i915/intel_dpll_mgr.c index d513ca875c67..0a42d11c4c33 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.c +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.c @@ -247,7 +247,7 @@ intel_find_shared_dpll(struct intel_crtc *crtc, enum intel_dpll_id range_max) { struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - struct intel_shared_dpll *pll; + struct intel_shared_dpll *pll, *unused_pll = NULL; struct intel_shared_dpll_state *shared_dpll; enum intel_dpll_id i; @@ -257,8 +257,11 @@ intel_find_shared_dpll(struct intel_crtc *crtc, pll = &dev_priv->shared_dplls[i]; /* Only want to check enabled timings first */ - if (shared_dpll[i].crtc_mask == 0) + if (shared_dpll[i].crtc_mask == 0) { + if (!unused_pll) + unused_pll = pll; continue; + } if (memcmp(&crtc_state->dpll_hw_state, &shared_dpll[i].hw_state, @@ -273,14 +276,11 @@ intel_find_shared_dpll(struct intel_crtc *crtc, } /* Ok no matching timings, maybe there's a free one? */ - for (i = range_min; i <= range_max; i++) { - pll = &dev_priv->shared_dplls[i]; - if (shared_dpll[i].crtc_mask == 0) { - DRM_DEBUG_KMS("[CRTC:%d:%s] allocated %s\n", - crtc->base.base.id, crtc->base.name, - pll->info->name); - return pll; - } + if (unused_pll) { + DRM_DEBUG_KMS("[CRTC:%d:%s] allocated %s\n", + crtc->base.base.id, crtc->base.name, + unused_pll->info->name); + return unused_pll; } return NULL; @@ -345,9 +345,12 @@ static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_dpll_hw_state *hw_state) { const enum intel_dpll_id id = pll->info->id; - uint32_t val; + intel_wakeref_t wakeref; + u32 val; - if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + POWER_DOMAIN_PLLS); + if (!wakeref) return false; val = I915_READ(PCH_DPLL(id)); @@ -355,7 +358,7 @@ static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv, hw_state->fp0 = I915_READ(PCH_FP0(id)); hw_state->fp1 = I915_READ(PCH_FP1(id)); - intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); + intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS, wakeref); return val & DPLL_VCO_ENABLE; } @@ -487,7 +490,7 @@ static void hsw_ddi_wrpll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { const enum intel_dpll_id id = pll->info->id; - uint32_t val; + u32 val; val = I915_READ(WRPLL_CTL(id)); I915_WRITE(WRPLL_CTL(id), val & ~WRPLL_PLL_ENABLE); @@ -497,7 +500,7 @@ static void hsw_ddi_wrpll_disable(struct drm_i915_private *dev_priv, static void hsw_ddi_spll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { - uint32_t val; + u32 val; val = I915_READ(SPLL_CTL); I915_WRITE(SPLL_CTL, val & ~SPLL_PLL_ENABLE); @@ -509,15 +512,18 @@ static bool hsw_ddi_wrpll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_dpll_hw_state *hw_state) { const enum intel_dpll_id id = pll->info->id; - uint32_t val; + intel_wakeref_t wakeref; + u32 val; - if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + POWER_DOMAIN_PLLS); + if (!wakeref) return false; val = I915_READ(WRPLL_CTL(id)); hw_state->wrpll = val; - intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); + intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS, wakeref); return val & WRPLL_PLL_ENABLE; } @@ -526,15 +532,18 @@ static bool hsw_ddi_spll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll, struct intel_dpll_hw_state *hw_state) { - uint32_t val; + intel_wakeref_t wakeref; + u32 val; - if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + POWER_DOMAIN_PLLS); + if (!wakeref) return false; val = I915_READ(SPLL_CTL); hw_state->spll = val; - intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); + intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS, wakeref); return val & SPLL_PLL_ENABLE; } @@ -630,11 +639,12 @@ static unsigned hsw_wrpll_get_budget_for_freq(int clock) return budget; } -static void hsw_wrpll_update_rnp(uint64_t freq2k, unsigned budget, - unsigned r2, unsigned n2, unsigned p, +static void hsw_wrpll_update_rnp(u64 freq2k, unsigned int budget, + unsigned int r2, unsigned int n2, + unsigned int p, struct hsw_wrpll_rnp *best) { - uint64_t a, b, c, d, diff, diff_best; + u64 a, b, c, d, diff, diff_best; /* No best (r,n,p) yet */ if (best->p == 0) { @@ -693,7 +703,7 @@ static void hsw_ddi_calculate_wrpll(int clock /* in Hz */, unsigned *r2_out, unsigned *n2_out, unsigned *p_out) { - uint64_t freq2k; + u64 freq2k; unsigned p, n2, r2; struct hsw_wrpll_rnp best = { 0, 0, 0 }; unsigned budget; @@ -759,7 +769,7 @@ static struct intel_shared_dpll *hsw_ddi_hdmi_get_dpll(int clock, struct intel_crtc_state *crtc_state) { struct intel_shared_dpll *pll; - uint32_t val; + u32 val; unsigned int p, n2, r2; hsw_ddi_calculate_wrpll(clock * 1000, &r2, &n2, &p); @@ -921,7 +931,7 @@ static void skl_ddi_pll_write_ctrl1(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { const enum intel_dpll_id id = pll->info->id; - uint32_t val; + u32 val; val = I915_READ(DPLL_CTRL1); @@ -986,12 +996,15 @@ static bool skl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll, struct intel_dpll_hw_state *hw_state) { - uint32_t val; + u32 val; const struct skl_dpll_regs *regs = skl_dpll_regs; const enum intel_dpll_id id = pll->info->id; + intel_wakeref_t wakeref; bool ret; - if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + POWER_DOMAIN_PLLS); + if (!wakeref) return false; ret = false; @@ -1011,7 +1024,7 @@ static bool skl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, ret = true; out: - intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); + intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS, wakeref); return ret; } @@ -1020,12 +1033,15 @@ static bool skl_ddi_dpll0_get_hw_state(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll, struct intel_dpll_hw_state *hw_state) { - uint32_t val; const struct skl_dpll_regs *regs = skl_dpll_regs; const enum intel_dpll_id id = pll->info->id; + intel_wakeref_t wakeref; + u32 val; bool ret; - if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + POWER_DOMAIN_PLLS); + if (!wakeref) return false; ret = false; @@ -1041,15 +1057,15 @@ static bool skl_ddi_dpll0_get_hw_state(struct drm_i915_private *dev_priv, ret = true; out: - intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); + intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS, wakeref); return ret; } struct skl_wrpll_context { - uint64_t min_deviation; /* current minimal deviation */ - uint64_t central_freq; /* chosen central freq */ - uint64_t dco_freq; /* chosen dco freq */ + u64 min_deviation; /* current minimal deviation */ + u64 central_freq; /* chosen central freq */ + u64 dco_freq; /* chosen dco freq */ unsigned int p; /* chosen divider */ }; @@ -1065,11 +1081,11 @@ static void skl_wrpll_context_init(struct skl_wrpll_context *ctx) #define SKL_DCO_MAX_NDEVIATION 600 static void skl_wrpll_try_divider(struct skl_wrpll_context *ctx, - uint64_t central_freq, - uint64_t dco_freq, + u64 central_freq, + u64 dco_freq, unsigned int divider) { - uint64_t deviation; + u64 deviation; deviation = div64_u64(10000 * abs_diff(dco_freq, central_freq), central_freq); @@ -1143,21 +1159,21 @@ static void skl_wrpll_get_multipliers(unsigned int p, } struct skl_wrpll_params { - uint32_t dco_fraction; - uint32_t dco_integer; - uint32_t qdiv_ratio; - uint32_t qdiv_mode; - uint32_t kdiv; - uint32_t pdiv; - uint32_t central_freq; + u32 dco_fraction; + u32 dco_integer; + u32 qdiv_ratio; + u32 qdiv_mode; + u32 kdiv; + u32 pdiv; + u32 central_freq; }; static void skl_wrpll_params_populate(struct skl_wrpll_params *params, - uint64_t afe_clock, - uint64_t central_freq, - uint32_t p0, uint32_t p1, uint32_t p2) + u64 afe_clock, + u64 central_freq, + u32 p0, u32 p1, u32 p2) { - uint64_t dco_freq; + u64 dco_freq; switch (central_freq) { case 9600000000ULL: @@ -1223,10 +1239,10 @@ static bool skl_ddi_calculate_wrpll(int clock /* in Hz */, struct skl_wrpll_params *wrpll_params) { - uint64_t afe_clock = clock * 5; /* AFE Clock is 5x Pixel clock */ - uint64_t dco_central_freq[3] = {8400000000ULL, - 9000000000ULL, - 9600000000ULL}; + u64 afe_clock = clock * 5; /* AFE Clock is 5x Pixel clock */ + u64 dco_central_freq[3] = { 8400000000ULL, + 9000000000ULL, + 9600000000ULL }; static const int even_dividers[] = { 4, 6, 8, 10, 12, 14, 16, 18, 20, 24, 28, 30, 32, 36, 40, 42, 44, 48, 52, 54, 56, 60, 64, 66, 68, @@ -1250,7 +1266,7 @@ skl_ddi_calculate_wrpll(int clock /* in Hz */, for (dco = 0; dco < ARRAY_SIZE(dco_central_freq); dco++) { for (i = 0; i < dividers[d].n_dividers; i++) { unsigned int p = dividers[d].list[i]; - uint64_t dco_freq = p * afe_clock; + u64 dco_freq = p * afe_clock; skl_wrpll_try_divider(&ctx, dco_central_freq[dco], @@ -1296,7 +1312,7 @@ static bool skl_ddi_hdmi_pll_dividers(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state, int clock) { - uint32_t ctrl1, cfgcr1, cfgcr2; + u32 ctrl1, cfgcr1, cfgcr2; struct skl_wrpll_params wrpll_params = { 0, }; /* @@ -1333,7 +1349,7 @@ static bool skl_ddi_dp_set_dpll_hw_state(int clock, struct intel_dpll_hw_state *dpll_hw_state) { - uint32_t ctrl1; + u32 ctrl1; /* * See comment in intel_dpll_hw_state to understand why we always use 0 @@ -1435,7 +1451,7 @@ static const struct intel_shared_dpll_funcs skl_ddi_dpll0_funcs = { static void bxt_ddi_pll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { - uint32_t temp; + u32 temp; enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */ enum dpio_phy phy; enum dpio_channel ch; @@ -1556,7 +1572,7 @@ static void bxt_ddi_pll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */ - uint32_t temp; + u32 temp; temp = I915_READ(BXT_PORT_PLL_ENABLE(port)); temp &= ~PORT_PLL_ENABLE; @@ -1579,14 +1595,17 @@ static bool bxt_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_dpll_hw_state *hw_state) { enum port port = (enum port)pll->info->id; /* 1:1 port->PLL mapping */ - uint32_t val; - bool ret; + intel_wakeref_t wakeref; enum dpio_phy phy; enum dpio_channel ch; + u32 val; + bool ret; bxt_port_to_phy_channel(dev_priv, port, &phy, &ch); - if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + POWER_DOMAIN_PLLS); + if (!wakeref) return false; ret = false; @@ -1643,7 +1662,7 @@ static bool bxt_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, ret = true; out: - intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); + intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS, wakeref); return ret; } @@ -1651,12 +1670,12 @@ out: /* bxt clock parameters */ struct bxt_clk_div { int clock; - uint32_t p1; - uint32_t p2; - uint32_t m2_int; - uint32_t m2_frac; + u32 p1; + u32 p2; + u32 m2_int; + u32 m2_frac; bool m2_frac_en; - uint32_t n; + u32 n; int vco; }; @@ -1723,8 +1742,8 @@ static bool bxt_ddi_set_dpll_hw_state(int clock, struct intel_dpll_hw_state *dpll_hw_state) { int vco = clk_div->vco; - uint32_t prop_coef, int_coef, gain_ctl, targ_cnt; - uint32_t lanestagger; + u32 prop_coef, int_coef, gain_ctl, targ_cnt; + u32 lanestagger; if (vco >= 6200000 && vco <= 6700000) { prop_coef = 4; @@ -1873,7 +1892,7 @@ static void intel_ddi_pll_init(struct drm_device *dev) struct drm_i915_private *dev_priv = to_i915(dev); if (INTEL_GEN(dev_priv) < 9) { - uint32_t val = I915_READ(LCPLL_CTL); + u32 val = I915_READ(LCPLL_CTL); /* * The LCPLL register should be turned on by the BIOS. For now @@ -1959,7 +1978,7 @@ static void cnl_ddi_pll_enable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { const enum intel_dpll_id id = pll->info->id; - uint32_t val; + u32 val; /* 1. Enable DPLL power in DPLL_ENABLE. */ val = I915_READ(CNL_DPLL_ENABLE(id)); @@ -2034,7 +2053,7 @@ static void cnl_ddi_pll_disable(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { const enum intel_dpll_id id = pll->info->id; - uint32_t val; + u32 val; /* * 1. Configure DPCLKA_CFGCR0 to turn off the clock for the DDI. @@ -2091,10 +2110,13 @@ static bool cnl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_dpll_hw_state *hw_state) { const enum intel_dpll_id id = pll->info->id; - uint32_t val; + intel_wakeref_t wakeref; + u32 val; bool ret; - if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + POWER_DOMAIN_PLLS); + if (!wakeref) return false; ret = false; @@ -2113,7 +2135,7 @@ static bool cnl_ddi_pll_get_hw_state(struct drm_i915_private *dev_priv, ret = true; out: - intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); + intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS, wakeref); return ret; } @@ -2225,7 +2247,7 @@ cnl_ddi_calculate_wrpll(int clock, struct skl_wrpll_params *wrpll_params) { u32 afe_clock = clock * 5; - uint32_t ref_clock; + u32 ref_clock; u32 dco_min = 7998000; u32 dco_max = 10000000; u32 dco_mid = (dco_min + dco_max) / 2; @@ -2271,7 +2293,7 @@ static bool cnl_ddi_hdmi_pll_dividers(struct intel_crtc *crtc, int clock) { struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); - uint32_t cfgcr0, cfgcr1; + u32 cfgcr0, cfgcr1; struct skl_wrpll_params wrpll_params = { 0, }; cfgcr0 = DPLL_CFGCR0_HDMI_MODE; @@ -2300,7 +2322,7 @@ static bool cnl_ddi_dp_set_dpll_hw_state(int clock, struct intel_dpll_hw_state *dpll_hw_state) { - uint32_t cfgcr0; + u32 cfgcr0; cfgcr0 = DPLL_CFGCR0_SSC_ENABLE; @@ -2517,7 +2539,7 @@ static bool icl_calc_dpll_state(struct intel_crtc_state *crtc_state, struct intel_dpll_hw_state *pll_state) { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); - uint32_t cfgcr0, cfgcr1; + u32 cfgcr0, cfgcr1; struct skl_wrpll_params pll_params = { 0 }; bool ret; @@ -2547,10 +2569,10 @@ static bool icl_calc_dpll_state(struct intel_crtc_state *crtc_state, } int icl_calc_dp_combo_pll_link(struct drm_i915_private *dev_priv, - uint32_t pll_id) + u32 pll_id) { - uint32_t cfgcr0, cfgcr1; - uint32_t pdiv, kdiv, qdiv_mode, qdiv_ratio, dco_integer, dco_fraction; + u32 cfgcr0, cfgcr1; + u32 pdiv, kdiv, qdiv_mode, qdiv_ratio, dco_integer, dco_fraction; const struct skl_wrpll_params *params; int index, n_entries, link_clock; @@ -2617,14 +2639,14 @@ int icl_calc_dp_combo_pll_link(struct drm_i915_private *dev_priv, return link_clock; } -static enum port icl_mg_pll_id_to_port(enum intel_dpll_id id) +static enum tc_port icl_pll_id_to_tc_port(enum intel_dpll_id id) { - return id - DPLL_ID_ICL_MGPLL1 + PORT_C; + return id - DPLL_ID_ICL_MGPLL1; } -enum intel_dpll_id icl_port_to_mg_pll_id(enum port port) +enum intel_dpll_id icl_tc_port_to_pll_id(enum tc_port tc_port) { - return port - PORT_C + DPLL_ID_ICL_MGPLL1; + return tc_port + DPLL_ID_ICL_MGPLL1; } bool intel_dpll_is_combophy(enum intel_dpll_id id) @@ -2633,10 +2655,10 @@ bool intel_dpll_is_combophy(enum intel_dpll_id id) } static bool icl_mg_pll_find_divisors(int clock_khz, bool is_dp, bool use_ssc, - uint32_t *target_dco_khz, + u32 *target_dco_khz, struct intel_dpll_hw_state *state) { - uint32_t dco_min_freq, dco_max_freq; + u32 dco_min_freq, dco_max_freq; int div1_vals[] = {7, 5, 3, 2}; unsigned int i; int div2; @@ -2712,12 +2734,12 @@ static bool icl_calc_mg_pll_state(struct intel_crtc_state *crtc_state, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); int refclk_khz = dev_priv->cdclk.hw.ref; - uint32_t dco_khz, m1div, m2div_int, m2div_rem, m2div_frac; - uint32_t iref_ndiv, iref_trim, iref_pulse_w; - uint32_t prop_coeff, int_coeff; - uint32_t tdc_targetcnt, feedfwgain; - uint64_t ssc_stepsize, ssc_steplen, ssc_steplog; - uint64_t tmp; + u32 dco_khz, m1div, m2div_int, m2div_rem, m2div_frac; + u32 iref_ndiv, iref_trim, iref_pulse_w; + u32 prop_coeff, int_coeff; + u32 tdc_targetcnt, feedfwgain; + u64 ssc_stepsize, ssc_steplen, ssc_steplog; + u64 tmp; bool use_ssc = false; bool is_dp = !intel_crtc_has_type(crtc_state, INTEL_OUTPUT_HDMI); @@ -2740,7 +2762,7 @@ static bool icl_calc_mg_pll_state(struct intel_crtc_state *crtc_state, } m2div_rem = dco_khz % (refclk_khz * m1div); - tmp = (uint64_t)m2div_rem * (1 << 22); + tmp = (u64)m2div_rem * (1 << 22); do_div(tmp, refclk_khz * m1div); m2div_frac = tmp; @@ -2799,11 +2821,11 @@ static bool icl_calc_mg_pll_state(struct intel_crtc_state *crtc_state, } if (use_ssc) { - tmp = (uint64_t)dco_khz * 47 * 32; + tmp = (u64)dco_khz * 47 * 32; do_div(tmp, refclk_khz * m1div * 10000); ssc_stepsize = tmp; - tmp = (uint64_t)dco_khz * 1000; + tmp = (u64)dco_khz * 1000; ssc_steplen = DIV_ROUND_UP_ULL(tmp, 32 * 2 * 32); } else { ssc_stepsize = 0; @@ -2903,7 +2925,10 @@ icl_get_dpll(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state, ret = icl_calc_dpll_state(crtc_state, encoder, clock, &pll_state); } else { - min = icl_port_to_mg_pll_id(port); + enum tc_port tc_port; + + tc_port = intel_port_to_tc(dev_priv, port); + min = icl_tc_port_to_pll_id(tc_port); max = min; ret = icl_calc_mg_pll_state(crtc_state, encoder, clock, &pll_state); @@ -2937,12 +2962,8 @@ static i915_reg_t icl_pll_id_to_enable_reg(enum intel_dpll_id id) return CNL_DPLL_ENABLE(id); else if (id == DPLL_ID_ICL_TBTPLL) return TBT_PLL_ENABLE; - else - /* - * TODO: Make MG_PLL macros use - * tc port id instead of port id - */ - return MG_PLL_ENABLE(icl_mg_pll_id_to_port(id)); + + return MG_PLL_ENABLE(icl_pll_id_to_tc_port(id)); } static bool icl_pll_get_hw_state(struct drm_i915_private *dev_priv, @@ -2950,11 +2971,13 @@ static bool icl_pll_get_hw_state(struct drm_i915_private *dev_priv, struct intel_dpll_hw_state *hw_state) { const enum intel_dpll_id id = pll->info->id; - uint32_t val; - enum port port; + intel_wakeref_t wakeref; bool ret = false; + u32 val; - if (!intel_display_power_get_if_enabled(dev_priv, POWER_DOMAIN_PLLS)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + POWER_DOMAIN_PLLS); + if (!wakeref) return false; val = I915_READ(icl_pll_id_to_enable_reg(id)); @@ -2966,32 +2989,33 @@ static bool icl_pll_get_hw_state(struct drm_i915_private *dev_priv, hw_state->cfgcr0 = I915_READ(ICL_DPLL_CFGCR0(id)); hw_state->cfgcr1 = I915_READ(ICL_DPLL_CFGCR1(id)); } else { - port = icl_mg_pll_id_to_port(id); - hw_state->mg_refclkin_ctl = I915_READ(MG_REFCLKIN_CTL(port)); + enum tc_port tc_port = icl_pll_id_to_tc_port(id); + + hw_state->mg_refclkin_ctl = I915_READ(MG_REFCLKIN_CTL(tc_port)); hw_state->mg_refclkin_ctl &= MG_REFCLKIN_CTL_OD_2_MUX_MASK; hw_state->mg_clktop2_coreclkctl1 = - I915_READ(MG_CLKTOP2_CORECLKCTL1(port)); + I915_READ(MG_CLKTOP2_CORECLKCTL1(tc_port)); hw_state->mg_clktop2_coreclkctl1 &= MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO_MASK; hw_state->mg_clktop2_hsclkctl = - I915_READ(MG_CLKTOP2_HSCLKCTL(port)); + I915_READ(MG_CLKTOP2_HSCLKCTL(tc_port)); hw_state->mg_clktop2_hsclkctl &= MG_CLKTOP2_HSCLKCTL_TLINEDRV_CLKSEL_MASK | MG_CLKTOP2_HSCLKCTL_CORE_INPUTSEL_MASK | MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_MASK | MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_MASK; - hw_state->mg_pll_div0 = I915_READ(MG_PLL_DIV0(port)); - hw_state->mg_pll_div1 = I915_READ(MG_PLL_DIV1(port)); - hw_state->mg_pll_lf = I915_READ(MG_PLL_LF(port)); - hw_state->mg_pll_frac_lock = I915_READ(MG_PLL_FRAC_LOCK(port)); - hw_state->mg_pll_ssc = I915_READ(MG_PLL_SSC(port)); + hw_state->mg_pll_div0 = I915_READ(MG_PLL_DIV0(tc_port)); + hw_state->mg_pll_div1 = I915_READ(MG_PLL_DIV1(tc_port)); + hw_state->mg_pll_lf = I915_READ(MG_PLL_LF(tc_port)); + hw_state->mg_pll_frac_lock = I915_READ(MG_PLL_FRAC_LOCK(tc_port)); + hw_state->mg_pll_ssc = I915_READ(MG_PLL_SSC(tc_port)); - hw_state->mg_pll_bias = I915_READ(MG_PLL_BIAS(port)); + hw_state->mg_pll_bias = I915_READ(MG_PLL_BIAS(tc_port)); hw_state->mg_pll_tdc_coldst_bias = - I915_READ(MG_PLL_TDC_COLDST_BIAS(port)); + I915_READ(MG_PLL_TDC_COLDST_BIAS(tc_port)); if (dev_priv->cdclk.hw.ref == 38400) { hw_state->mg_pll_tdc_coldst_bias_mask = MG_PLL_TDC_COLDST_COLDSTART; @@ -3007,7 +3031,7 @@ static bool icl_pll_get_hw_state(struct drm_i915_private *dev_priv, ret = true; out: - intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS); + intel_display_power_put(dev_priv, POWER_DOMAIN_PLLS, wakeref); return ret; } @@ -3026,7 +3050,7 @@ static void icl_mg_pll_write(struct drm_i915_private *dev_priv, struct intel_shared_dpll *pll) { struct intel_dpll_hw_state *hw_state = &pll->state.hw_state; - enum port port = icl_mg_pll_id_to_port(pll->info->id); + enum tc_port tc_port = icl_pll_id_to_tc_port(pll->info->id); u32 val; /* @@ -3035,41 +3059,41 @@ static void icl_mg_pll_write(struct drm_i915_private *dev_priv, * during the calc/readout phase if the mask depends on some other HW * state like refclk, see icl_calc_mg_pll_state(). */ - val = I915_READ(MG_REFCLKIN_CTL(port)); + val = I915_READ(MG_REFCLKIN_CTL(tc_port)); val &= ~MG_REFCLKIN_CTL_OD_2_MUX_MASK; val |= hw_state->mg_refclkin_ctl; - I915_WRITE(MG_REFCLKIN_CTL(port), val); + I915_WRITE(MG_REFCLKIN_CTL(tc_port), val); - val = I915_READ(MG_CLKTOP2_CORECLKCTL1(port)); + val = I915_READ(MG_CLKTOP2_CORECLKCTL1(tc_port)); val &= ~MG_CLKTOP2_CORECLKCTL1_A_DIVRATIO_MASK; val |= hw_state->mg_clktop2_coreclkctl1; - I915_WRITE(MG_CLKTOP2_CORECLKCTL1(port), val); + I915_WRITE(MG_CLKTOP2_CORECLKCTL1(tc_port), val); - val = I915_READ(MG_CLKTOP2_HSCLKCTL(port)); + val = I915_READ(MG_CLKTOP2_HSCLKCTL(tc_port)); val &= ~(MG_CLKTOP2_HSCLKCTL_TLINEDRV_CLKSEL_MASK | MG_CLKTOP2_HSCLKCTL_CORE_INPUTSEL_MASK | MG_CLKTOP2_HSCLKCTL_HSDIV_RATIO_MASK | MG_CLKTOP2_HSCLKCTL_DSDIV_RATIO_MASK); val |= hw_state->mg_clktop2_hsclkctl; - I915_WRITE(MG_CLKTOP2_HSCLKCTL(port), val); + I915_WRITE(MG_CLKTOP2_HSCLKCTL(tc_port), val); - I915_WRITE(MG_PLL_DIV0(port), hw_state->mg_pll_div0); - I915_WRITE(MG_PLL_DIV1(port), hw_state->mg_pll_div1); - I915_WRITE(MG_PLL_LF(port), hw_state->mg_pll_lf); - I915_WRITE(MG_PLL_FRAC_LOCK(port), hw_state->mg_pll_frac_lock); - I915_WRITE(MG_PLL_SSC(port), hw_state->mg_pll_ssc); + I915_WRITE(MG_PLL_DIV0(tc_port), hw_state->mg_pll_div0); + I915_WRITE(MG_PLL_DIV1(tc_port), hw_state->mg_pll_div1); + I915_WRITE(MG_PLL_LF(tc_port), hw_state->mg_pll_lf); + I915_WRITE(MG_PLL_FRAC_LOCK(tc_port), hw_state->mg_pll_frac_lock); + I915_WRITE(MG_PLL_SSC(tc_port), hw_state->mg_pll_ssc); - val = I915_READ(MG_PLL_BIAS(port)); + val = I915_READ(MG_PLL_BIAS(tc_port)); val &= ~hw_state->mg_pll_bias_mask; val |= hw_state->mg_pll_bias; - I915_WRITE(MG_PLL_BIAS(port), val); + I915_WRITE(MG_PLL_BIAS(tc_port), val); - val = I915_READ(MG_PLL_TDC_COLDST_BIAS(port)); + val = I915_READ(MG_PLL_TDC_COLDST_BIAS(tc_port)); val &= ~hw_state->mg_pll_tdc_coldst_bias_mask; val |= hw_state->mg_pll_tdc_coldst_bias; - I915_WRITE(MG_PLL_TDC_COLDST_BIAS(port), val); + I915_WRITE(MG_PLL_TDC_COLDST_BIAS(tc_port), val); - POSTING_READ(MG_PLL_TDC_COLDST_BIAS(port)); + POSTING_READ(MG_PLL_TDC_COLDST_BIAS(tc_port)); } static void icl_pll_enable(struct drm_i915_private *dev_priv, @@ -3077,7 +3101,7 @@ static void icl_pll_enable(struct drm_i915_private *dev_priv, { const enum intel_dpll_id id = pll->info->id; i915_reg_t enable_reg = icl_pll_id_to_enable_reg(id); - uint32_t val; + u32 val; val = I915_READ(enable_reg); val |= PLL_POWER_ENABLE; @@ -3118,7 +3142,7 @@ static void icl_pll_disable(struct drm_i915_private *dev_priv, { const enum intel_dpll_id id = pll->info->id; i915_reg_t enable_reg = icl_pll_id_to_enable_reg(id); - uint32_t val; + u32 val; /* The first steps are done by intel_ddi_post_disable(). */ diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.h b/drivers/gpu/drm/i915/intel_dpll_mgr.h index a033d8f06d4a..40e8391a92f2 100644 --- a/drivers/gpu/drm/i915/intel_dpll_mgr.h +++ b/drivers/gpu/drm/i915/intel_dpll_mgr.h @@ -138,14 +138,14 @@ enum intel_dpll_id { struct intel_dpll_hw_state { /* i9xx, pch plls */ - uint32_t dpll; - uint32_t dpll_md; - uint32_t fp0; - uint32_t fp1; + u32 dpll; + u32 dpll_md; + u32 fp0; + u32 fp1; /* hsw, bdw */ - uint32_t wrpll; - uint32_t spll; + u32 wrpll; + u32 spll; /* skl */ /* @@ -154,34 +154,33 @@ struct intel_dpll_hw_state { * the register. This allows us to easily compare the state to share * the DPLL. */ - uint32_t ctrl1; + u32 ctrl1; /* HDMI only, 0 when used for DP */ - uint32_t cfgcr1, cfgcr2; + u32 cfgcr1, cfgcr2; /* cnl */ - uint32_t cfgcr0; + u32 cfgcr0; /* CNL also uses cfgcr1 */ /* bxt */ - uint32_t ebb0, ebb4, pll0, pll1, pll2, pll3, pll6, pll8, pll9, pll10, - pcsdw12; + u32 ebb0, ebb4, pll0, pll1, pll2, pll3, pll6, pll8, pll9, pll10, pcsdw12; /* * ICL uses the following, already defined: - * uint32_t cfgcr0, cfgcr1; - */ - uint32_t mg_refclkin_ctl; - uint32_t mg_clktop2_coreclkctl1; - uint32_t mg_clktop2_hsclkctl; - uint32_t mg_pll_div0; - uint32_t mg_pll_div1; - uint32_t mg_pll_lf; - uint32_t mg_pll_frac_lock; - uint32_t mg_pll_ssc; - uint32_t mg_pll_bias; - uint32_t mg_pll_tdc_coldst_bias; - uint32_t mg_pll_bias_mask; - uint32_t mg_pll_tdc_coldst_bias_mask; + * u32 cfgcr0, cfgcr1; + */ + u32 mg_refclkin_ctl; + u32 mg_clktop2_coreclkctl1; + u32 mg_clktop2_hsclkctl; + u32 mg_pll_div0; + u32 mg_pll_div1; + u32 mg_pll_lf; + u32 mg_pll_frac_lock; + u32 mg_pll_ssc; + u32 mg_pll_bias; + u32 mg_pll_tdc_coldst_bias; + u32 mg_pll_bias_mask; + u32 mg_pll_tdc_coldst_bias_mask; }; /** @@ -280,7 +279,7 @@ struct dpll_info { * Inform the state checker that the DPLL is kept enabled even if * not in use by any CRTC. */ - uint32_t flags; + u32 flags; }; /** @@ -343,9 +342,9 @@ void intel_shared_dpll_init(struct drm_device *dev); void intel_dpll_dump_hw_state(struct drm_i915_private *dev_priv, struct intel_dpll_hw_state *hw_state); int icl_calc_dp_combo_pll_link(struct drm_i915_private *dev_priv, - uint32_t pll_id); + u32 pll_id); int cnl_hdmi_pll_ref_clock(struct drm_i915_private *dev_priv); -enum intel_dpll_id icl_port_to_mg_pll_id(enum port port); +enum intel_dpll_id icl_tc_port_to_pll_id(enum tc_port tc_port); bool intel_dpll_is_combophy(enum intel_dpll_id id); #endif /* _INTEL_DPLL_MGR_H_ */ diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index c1759babe1dd..15db41394b9e 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -29,6 +29,7 @@ #include <linux/i2c.h> #include <linux/hdmi.h> #include <linux/sched/clock.h> +#include <linux/stackdepot.h> #include <drm/i915_drm.h> #include "i915_drv.h" #include <drm/drm_crtc.h> @@ -42,6 +43,8 @@ #include <drm/drm_atomic.h> #include <media/cec-notifier.h> +struct drm_printer; + /** * __wait_for - magic wait macro * @@ -210,6 +213,16 @@ struct intel_fbdev { unsigned long vma_flags; async_cookie_t cookie; int preferred_bpp; + + /* Whether or not fbdev hpd processing is temporarily suspended */ + bool hpd_suspended : 1; + /* Set when a hotplug was received while HPD processing was + * suspended + */ + bool hpd_waiting : 1; + + /* Protects hpd_suspended */ + struct mutex hpd_lock; }; struct intel_encoder { @@ -298,13 +311,12 @@ struct intel_panel { /* Connector and platform specific backlight functions */ int (*setup)(struct intel_connector *connector, enum pipe pipe); - uint32_t (*get)(struct intel_connector *connector); - void (*set)(const struct drm_connector_state *conn_state, uint32_t level); + u32 (*get)(struct intel_connector *connector); + void (*set)(const struct drm_connector_state *conn_state, u32 level); void (*disable)(const struct drm_connector_state *conn_state); void (*enable)(const struct intel_crtc_state *crtc_state, const struct drm_connector_state *conn_state); - uint32_t (*hz_to_pwm)(struct intel_connector *connector, - uint32_t hz); + u32 (*hz_to_pwm)(struct intel_connector *connector, u32 hz); void (*power)(struct intel_connector *, bool enable); } backlight; }; @@ -596,7 +608,7 @@ struct intel_initial_plane_config { struct intel_scaler { int in_use; - uint32_t mode; + u32 mode; }; struct intel_crtc_scaler_state { @@ -628,13 +640,15 @@ struct intel_crtc_scaler_state { }; /* drm_mode->private_flags */ -#define I915_MODE_FLAG_INHERITED 1 +#define I915_MODE_FLAG_INHERITED (1<<0) /* Flag to get scanline using frame time stamps */ #define I915_MODE_FLAG_GET_SCANLINE_FROM_TIMESTAMP (1<<1) +/* Flag to use the scanline counter instead of the pixel counter */ +#define I915_MODE_FLAG_USE_SCANLINE_COUNTER (1<<2) struct intel_pipe_wm { struct intel_wm_level wm[5]; - uint32_t linetime; + u32 linetime; bool fbc_wm_enabled; bool pipe_enabled; bool sprites_enabled; @@ -650,7 +664,7 @@ struct skl_plane_wm { struct skl_pipe_wm { struct skl_plane_wm planes[I915_MAX_PLANES]; - uint32_t linetime; + u32 linetime; }; enum vlv_wm_level { @@ -663,7 +677,7 @@ enum vlv_wm_level { struct vlv_wm_state { struct g4x_pipe_wm wm[NUM_VLV_WM_LEVELS]; struct g4x_sr_wm sr[NUM_VLV_WM_LEVELS]; - uint8_t num_levels; + u8 num_levels; bool cxsr; }; @@ -876,13 +890,13 @@ struct intel_crtc_state { /* Used by SDVO (and if we ever fix it, HDMI). */ unsigned pixel_multiplier; - uint8_t lane_count; + u8 lane_count; /* * Used by platforms having DP/HDMI PHY with programmable lane * latency optimization. */ - uint8_t lane_lat_optim_mask; + u8 lane_lat_optim_mask; /* minimum acceptable voltage level */ u8 min_voltage_level; @@ -926,7 +940,7 @@ struct intel_crtc_state { struct intel_crtc_wm_state wm; /* Gamma mode programmed on the pipe */ - uint32_t gamma_mode; + u32 gamma_mode; /* bitmask of visible planes (enum plane_id) */ u8 active_planes; @@ -1012,7 +1026,7 @@ struct intel_plane { enum pipe pipe; bool has_fbc; bool has_ccs; - uint32_t frontbuffer_bit; + u32 frontbuffer_bit; struct { u32 base, cntl, size; @@ -1107,9 +1121,9 @@ enum link_m_n_set { struct intel_dp_compliance_data { unsigned long edid; - uint8_t video_pattern; - uint16_t hdisplay, vdisplay; - uint8_t bpc; + u8 video_pattern; + u16 hdisplay, vdisplay; + u8 bpc; }; struct intel_dp_compliance { @@ -1122,18 +1136,18 @@ struct intel_dp_compliance { struct intel_dp { i915_reg_t output_reg; - uint32_t DP; + u32 DP; int link_rate; - uint8_t lane_count; - uint8_t sink_count; + u8 lane_count; + u8 sink_count; bool link_mst; bool link_trained; bool has_audio; bool reset_link_params; - uint8_t dpcd[DP_RECEIVER_CAP_SIZE]; - uint8_t psr_dpcd[EDP_PSR_RECEIVER_CAP_SIZE]; - uint8_t downstream_ports[DP_MAX_DOWNSTREAM_PORTS]; - uint8_t edp_dpcd[EDP_DISPLAY_CTL_CAP_SIZE]; + u8 dpcd[DP_RECEIVER_CAP_SIZE]; + u8 psr_dpcd[EDP_PSR_RECEIVER_CAP_SIZE]; + u8 downstream_ports[DP_MAX_DOWNSTREAM_PORTS]; + u8 edp_dpcd[EDP_DISPLAY_CTL_CAP_SIZE]; u8 dsc_dpcd[DP_DSC_RECEIVER_CAP_SIZE]; u8 fec_capable; /* source rates */ @@ -1153,7 +1167,7 @@ struct intel_dp { /* sink or branch descriptor */ struct drm_dp_desc desc; struct drm_dp_aux aux; - uint8_t train_set[4]; + u8 train_set[4]; int panel_power_up_delay; int panel_power_down_delay; int panel_power_cycle_delay; @@ -1195,14 +1209,13 @@ struct intel_dp { struct intel_dp_mst_encoder *mst_encoders[I915_MAX_PIPES]; struct drm_dp_mst_topology_mgr mst_mgr; - uint32_t (*get_aux_clock_divider)(struct intel_dp *dp, int index); + u32 (*get_aux_clock_divider)(struct intel_dp *dp, int index); /* * This function returns the value we have to program the AUX_CTL * register with to kick off an AUX transaction. */ - uint32_t (*get_aux_send_ctl)(struct intel_dp *dp, - int send_bytes, - uint32_t aux_clock_divider); + u32 (*get_aux_send_ctl)(struct intel_dp *dp, int send_bytes, + u32 aux_clock_divider); i915_reg_t (*aux_ch_ctl_reg)(struct intel_dp *dp); i915_reg_t (*aux_ch_data_reg)(struct intel_dp *dp, int index); @@ -1236,7 +1249,7 @@ struct intel_digital_port { struct intel_lspcon lspcon; enum irqreturn (*hpd_pulse)(struct intel_digital_port *, bool); bool release_cl2_override; - uint8_t max_lanes; + u8 max_lanes; /* Used for DP and ICL+ TypeC/DP and TypeC/HDMI ports. */ enum aux_ch aux_ch; enum intel_display_power_domain ddi_io_power_domain; @@ -1471,8 +1484,8 @@ void intel_check_cpu_fifo_underruns(struct drm_i915_private *dev_priv); void intel_check_pch_fifo_underruns(struct drm_i915_private *dev_priv); /* i915_irq.c */ -void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask); -void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, uint32_t mask); +void gen5_enable_gt_irq(struct drm_i915_private *dev_priv, u32 mask); +void gen5_disable_gt_irq(struct drm_i915_private *dev_priv, u32 mask); void gen6_mask_pm_irq(struct drm_i915_private *dev_priv, u32 mask); void gen6_unmask_pm_irq(struct drm_i915_private *dev_priv, u32 mask); void gen11_reset_rps_interrupts(struct drm_i915_private *dev_priv); @@ -1535,7 +1548,7 @@ void intel_ddi_set_vc_payload_alloc(const struct intel_crtc_state *crtc_state, void intel_ddi_compute_min_voltage_level(struct drm_i915_private *dev_priv, struct intel_crtc_state *crtc_state); u32 bxt_signal_levels(struct intel_dp *intel_dp); -uint32_t ddi_signal_levels(struct intel_dp *intel_dp); +u32 ddi_signal_levels(struct intel_dp *intel_dp); u8 intel_ddi_dp_voltage_max(struct intel_encoder *encoder); u8 intel_ddi_dp_pre_emphasis_max(struct intel_encoder *encoder, u8 voltage_swing); @@ -1675,11 +1688,11 @@ void intel_cleanup_plane_fb(struct drm_plane *plane, int intel_plane_atomic_get_property(struct drm_plane *plane, const struct drm_plane_state *state, struct drm_property *property, - uint64_t *val); + u64 *val); int intel_plane_atomic_set_property(struct drm_plane *plane, struct drm_plane_state *state, struct drm_property *property, - uint64_t val); + u64 val); int intel_plane_atomic_calc_changes(const struct intel_crtc_state *old_crtc_state, struct drm_crtc_state *crtc_state, const struct intel_plane_state *old_plane_state, @@ -1753,9 +1766,10 @@ static inline u32 intel_plane_ggtt_offset(const struct intel_plane_state *state) u32 glk_plane_color_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state); +u32 glk_plane_color_ctl_crtc(const struct intel_crtc_state *crtc_state); u32 skl_plane_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state); -u32 glk_color_ctl(const struct intel_plane_state *plane_state); +u32 skl_plane_ctl_crtc(const struct intel_crtc_state *crtc_state); u32 skl_plane_stride(const struct intel_plane_state *plane_state, int plane); int skl_check_plane_surface(struct intel_plane_state *plane_state); @@ -1799,10 +1813,10 @@ bool intel_dp_init(struct drm_i915_private *dev_priv, i915_reg_t output_reg, bool intel_dp_init_connector(struct intel_digital_port *intel_dig_port, struct intel_connector *intel_connector); void intel_dp_set_link_params(struct intel_dp *intel_dp, - int link_rate, uint8_t lane_count, + int link_rate, u8 lane_count, bool link_mst); int intel_dp_get_link_train_fallback_values(struct intel_dp *intel_dp, - int link_rate, uint8_t lane_count); + int link_rate, u8 lane_count); void intel_dp_start_link_train(struct intel_dp *intel_dp); void intel_dp_stop_link_train(struct intel_dp *intel_dp); int intel_dp_retrain_link(struct intel_encoder *encoder, @@ -1834,7 +1848,7 @@ int intel_dp_max_lane_count(struct intel_dp *intel_dp); int intel_dp_rate_select(struct intel_dp *intel_dp, int rate); void intel_dp_hot_plug(struct intel_encoder *intel_encoder); void intel_power_sequencer_reset(struct drm_i915_private *dev_priv); -uint32_t intel_dp_pack_aux(const uint8_t *src, int src_bytes); +u32 intel_dp_pack_aux(const u8 *src, int src_bytes); void intel_plane_destroy(struct drm_plane *plane); void intel_edp_drrs_enable(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state); @@ -1847,24 +1861,24 @@ void intel_edp_drrs_flush(struct drm_i915_private *dev_priv, void intel_dp_program_link_training_pattern(struct intel_dp *intel_dp, - uint8_t dp_train_pat); + u8 dp_train_pat); void intel_dp_set_signal_levels(struct intel_dp *intel_dp); void intel_dp_set_idle_link_train(struct intel_dp *intel_dp); -uint8_t +u8 intel_dp_voltage_max(struct intel_dp *intel_dp); -uint8_t -intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing); +u8 +intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, u8 voltage_swing); void intel_dp_compute_rate(struct intel_dp *intel_dp, int port_clock, - uint8_t *link_bw, uint8_t *rate_select); + u8 *link_bw, u8 *rate_select); bool intel_dp_source_supports_hbr2(struct intel_dp *intel_dp); bool intel_dp_source_supports_hbr3(struct intel_dp *intel_dp); bool -intel_dp_get_link_status(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_STATUS_SIZE]); -uint16_t intel_dp_dsc_get_output_bpp(int link_clock, uint8_t lane_count, - int mode_clock, int mode_hdisplay); -uint8_t intel_dp_dsc_get_slice_count(struct intel_dp *intel_dp, int mode_clock, - int mode_hdisplay); +intel_dp_get_link_status(struct intel_dp *intel_dp, u8 link_status[DP_LINK_STATUS_SIZE]); +u16 intel_dp_dsc_get_output_bpp(int link_clock, u8 lane_count, + int mode_clock, int mode_hdisplay); +u8 intel_dp_dsc_get_slice_count(struct intel_dp *intel_dp, int mode_clock, + int mode_hdisplay); /* intel_vdsc.c */ int intel_dp_compute_dsc_params(struct intel_dp *intel_dp, @@ -2023,6 +2037,9 @@ int intel_panel_setup_backlight(struct drm_connector *connector, enum pipe pipe); void intel_panel_enable_backlight(const struct intel_crtc_state *crtc_state, const struct drm_connector_state *conn_state); +void intel_panel_update_backlight(struct intel_encoder *encoder, + const struct intel_crtc_state *crtc_state, + const struct drm_connector_state *conn_state); void intel_panel_disable_backlight(const struct drm_connector_state *old_conn_state); extern struct drm_display_mode *intel_find_panel_downclock( struct drm_i915_private *dev_priv, @@ -2084,6 +2101,7 @@ bool intel_psr_enabled(struct intel_dp *intel_dp); void intel_init_quirks(struct drm_i915_private *dev_priv); /* intel_runtime_pm.c */ +void intel_runtime_pm_init_early(struct drm_i915_private *dev_priv); int intel_power_domains_init(struct drm_i915_private *); void intel_power_domains_cleanup(struct drm_i915_private *dev_priv); void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume); @@ -2106,6 +2124,7 @@ void bxt_display_core_init(struct drm_i915_private *dev_priv, bool resume); void bxt_display_core_uninit(struct drm_i915_private *dev_priv); void intel_runtime_pm_enable(struct drm_i915_private *dev_priv); void intel_runtime_pm_disable(struct drm_i915_private *dev_priv); +void intel_runtime_pm_cleanup(struct drm_i915_private *dev_priv); const char * intel_display_power_domain_str(enum intel_display_power_domain domain); @@ -2113,33 +2132,42 @@ bool intel_display_power_is_enabled(struct drm_i915_private *dev_priv, enum intel_display_power_domain domain); bool __intel_display_power_is_enabled(struct drm_i915_private *dev_priv, enum intel_display_power_domain domain); -void intel_display_power_get(struct drm_i915_private *dev_priv, - enum intel_display_power_domain domain); -bool intel_display_power_get_if_enabled(struct drm_i915_private *dev_priv, +intel_wakeref_t intel_display_power_get(struct drm_i915_private *dev_priv, enum intel_display_power_domain domain); +intel_wakeref_t +intel_display_power_get_if_enabled(struct drm_i915_private *dev_priv, + enum intel_display_power_domain domain); +void intel_display_power_put_unchecked(struct drm_i915_private *dev_priv, + enum intel_display_power_domain domain); +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) void intel_display_power_put(struct drm_i915_private *dev_priv, - enum intel_display_power_domain domain); + enum intel_display_power_domain domain, + intel_wakeref_t wakeref); +#else +#define intel_display_power_put(i915, domain, wakeref) \ + intel_display_power_put_unchecked(i915, domain) +#endif void icl_dbuf_slices_update(struct drm_i915_private *dev_priv, u8 req_slices); static inline void -assert_rpm_device_not_suspended(struct drm_i915_private *dev_priv) +assert_rpm_device_not_suspended(struct drm_i915_private *i915) { - WARN_ONCE(dev_priv->runtime_pm.suspended, + WARN_ONCE(i915->runtime_pm.suspended, "Device suspended during HW access\n"); } static inline void -assert_rpm_wakelock_held(struct drm_i915_private *dev_priv) +assert_rpm_wakelock_held(struct drm_i915_private *i915) { - assert_rpm_device_not_suspended(dev_priv); - WARN_ONCE(!atomic_read(&dev_priv->runtime_pm.wakeref_count), + assert_rpm_device_not_suspended(i915); + WARN_ONCE(!atomic_read(&i915->runtime_pm.wakeref_count), "RPM wakelock ref not held during HW access"); } /** * disable_rpm_wakeref_asserts - disable the RPM assert checks - * @dev_priv: i915 device instance + * @i915: i915 device instance * * This function disable asserts that check if we hold an RPM wakelock * reference, while keeping the device-not-suspended checks still enabled. @@ -2156,14 +2184,14 @@ assert_rpm_wakelock_held(struct drm_i915_private *dev_priv) * enable_rpm_wakeref_asserts(). */ static inline void -disable_rpm_wakeref_asserts(struct drm_i915_private *dev_priv) +disable_rpm_wakeref_asserts(struct drm_i915_private *i915) { - atomic_inc(&dev_priv->runtime_pm.wakeref_count); + atomic_inc(&i915->runtime_pm.wakeref_count); } /** * enable_rpm_wakeref_asserts - re-enable the RPM assert checks - * @dev_priv: i915 device instance + * @i915: i915 device instance * * This function re-enables the RPM assert checks after disabling them with * disable_rpm_wakeref_asserts. It's meant to be used only in special @@ -2173,15 +2201,39 @@ disable_rpm_wakeref_asserts(struct drm_i915_private *dev_priv) * disable_rpm_wakeref_asserts(). */ static inline void -enable_rpm_wakeref_asserts(struct drm_i915_private *dev_priv) +enable_rpm_wakeref_asserts(struct drm_i915_private *i915) { - atomic_dec(&dev_priv->runtime_pm.wakeref_count); + atomic_dec(&i915->runtime_pm.wakeref_count); } -void intel_runtime_pm_get(struct drm_i915_private *dev_priv); -bool intel_runtime_pm_get_if_in_use(struct drm_i915_private *dev_priv); -void intel_runtime_pm_get_noresume(struct drm_i915_private *dev_priv); -void intel_runtime_pm_put(struct drm_i915_private *dev_priv); +intel_wakeref_t intel_runtime_pm_get(struct drm_i915_private *i915); +intel_wakeref_t intel_runtime_pm_get_if_in_use(struct drm_i915_private *i915); +intel_wakeref_t intel_runtime_pm_get_noresume(struct drm_i915_private *i915); + +#define with_intel_runtime_pm(i915, wf) \ + for ((wf) = intel_runtime_pm_get(i915); (wf); \ + intel_runtime_pm_put((i915), (wf)), (wf) = 0) + +#define with_intel_runtime_pm_if_in_use(i915, wf) \ + for ((wf) = intel_runtime_pm_get_if_in_use(i915); (wf); \ + intel_runtime_pm_put((i915), (wf)), (wf) = 0) + +void intel_runtime_pm_put_unchecked(struct drm_i915_private *i915); +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) +void intel_runtime_pm_put(struct drm_i915_private *i915, intel_wakeref_t wref); +#else +#define intel_runtime_pm_put(i915, wref) intel_runtime_pm_put_unchecked(i915) +#endif + +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) +void print_intel_runtime_pm_wakeref(struct drm_i915_private *i915, + struct drm_printer *p); +#else +static inline void print_intel_runtime_pm_wakeref(struct drm_i915_private *i915, + struct drm_printer *p) +{ +} +#endif void chv_phy_powergate_lanes(struct intel_encoder *encoder, bool override, unsigned int mask); @@ -2287,11 +2339,11 @@ void intel_tv_init(struct drm_i915_private *dev_priv); int intel_digital_connector_atomic_get_property(struct drm_connector *connector, const struct drm_connector_state *state, struct drm_property *property, - uint64_t *val); + u64 *val); int intel_digital_connector_atomic_set_property(struct drm_connector *connector, struct drm_connector_state *state, struct drm_property *property, - uint64_t val); + u64 val); int intel_digital_connector_atomic_check(struct drm_connector *conn, struct drm_connector_state *new_state); struct drm_connector_state * @@ -2338,8 +2390,8 @@ int intel_plane_atomic_check_with_state(const struct intel_crtc_state *old_crtc_ /* intel_color.c */ void intel_color_init(struct intel_crtc *crtc); int intel_color_check(struct intel_crtc_state *crtc_state); -void intel_color_set_csc(struct intel_crtc_state *crtc_state); -void intel_color_load_luts(struct intel_crtc_state *crtc_state); +void intel_color_commit(const struct intel_crtc_state *crtc_state); +void intel_color_load_luts(const struct intel_crtc_state *crtc_state); /* intel_lspcon.c */ bool lspcon_init(struct intel_digital_port *intel_dig_port); diff --git a/drivers/gpu/drm/i915/intel_dsi.h b/drivers/gpu/drm/i915/intel_dsi.h index fc7a09049f81..a9a19778dc7f 100644 --- a/drivers/gpu/drm/i915/intel_dsi.h +++ b/drivers/gpu/drm/i915/intel_dsi.h @@ -39,6 +39,7 @@ struct intel_dsi { struct intel_encoder base; struct intel_dsi_host *dsi_hosts[I915_MAX_PORTS]; + intel_wakeref_t io_wakeref[I915_MAX_PORTS]; /* GPIO Desc for CRC based Panel control */ struct gpio_desc *gpio_panel; @@ -172,7 +173,7 @@ int vlv_dsi_pll_compute(struct intel_encoder *encoder, void vlv_dsi_pll_enable(struct intel_encoder *encoder, const struct intel_crtc_state *config); void vlv_dsi_pll_disable(struct intel_encoder *encoder); -u32 vlv_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp, +u32 vlv_dsi_get_pclk(struct intel_encoder *encoder, struct intel_crtc_state *config); void vlv_dsi_reset_clocks(struct intel_encoder *encoder, enum port port); @@ -182,7 +183,7 @@ int bxt_dsi_pll_compute(struct intel_encoder *encoder, void bxt_dsi_pll_enable(struct intel_encoder *encoder, const struct intel_crtc_state *config); void bxt_dsi_pll_disable(struct intel_encoder *encoder); -u32 bxt_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp, +u32 bxt_dsi_get_pclk(struct intel_encoder *encoder, struct intel_crtc_state *config); void bxt_dsi_reset_clocks(struct intel_encoder *encoder, enum port port); diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c index 236cd040f271..49fa43ff02ba 100644 --- a/drivers/gpu/drm/i915/intel_engine_cs.c +++ b/drivers/gpu/drm/i915/intel_engine_cs.c @@ -25,6 +25,7 @@ #include <drm/drm_print.h> #include "i915_drv.h" +#include "i915_reset.h" #include "intel_ringbuffer.h" #include "intel_lrc.h" @@ -457,12 +458,6 @@ cleanup: void intel_engine_write_global_seqno(struct intel_engine_cs *engine, u32 seqno) { intel_write_status_page(engine, I915_GEM_HWS_INDEX, seqno); - - /* After manually advancing the seqno, fake the interrupt in case - * there are any waiters for that seqno. - */ - intel_engine_wakeup(engine); - GEM_BUG_ON(intel_engine_get_seqno(engine) != seqno); } @@ -479,53 +474,67 @@ static void intel_engine_init_execlist(struct intel_engine_cs *engine) GEM_BUG_ON(!is_power_of_2(execlists_num_ports(execlists))); GEM_BUG_ON(execlists_num_ports(execlists) > EXECLIST_MAX_PORTS); - execlists->queue_priority = INT_MIN; + execlists->queue_priority_hint = INT_MIN; execlists->queue = RB_ROOT_CACHED; } -/** - * intel_engines_setup_common - setup engine state not requiring hw access - * @engine: Engine to setup. - * - * Initializes @engine@ structure members shared between legacy and execlists - * submission modes which do not require hardware access. - * - * Typically done early in the submission mode specific engine setup stage. - */ -void intel_engine_setup_common(struct intel_engine_cs *engine) -{ - i915_timeline_init(engine->i915, &engine->timeline, engine->name); - i915_timeline_set_subclass(&engine->timeline, TIMELINE_ENGINE); - - intel_engine_init_execlist(engine); - intel_engine_init_hangcheck(engine); - intel_engine_init_batch_pool(engine); - intel_engine_init_cmd_parser(engine); -} - static void cleanup_status_page(struct intel_engine_cs *engine) { + struct i915_vma *vma; + /* Prevent writes into HWSP after returning the page to the system */ intel_engine_set_hwsp_writemask(engine, ~0u); - if (HWS_NEEDS_PHYSICAL(engine->i915)) { - void *addr = fetch_and_zero(&engine->status_page.page_addr); + vma = fetch_and_zero(&engine->status_page.vma); + if (!vma) + return; - __free_page(virt_to_page(addr)); - } + if (!HWS_NEEDS_PHYSICAL(engine->i915)) + i915_vma_unpin(vma); + + i915_gem_object_unpin_map(vma->obj); + __i915_gem_object_release_unless_active(vma->obj); +} + +static int pin_ggtt_status_page(struct intel_engine_cs *engine, + struct i915_vma *vma) +{ + unsigned int flags; + + flags = PIN_GLOBAL; + if (!HAS_LLC(engine->i915)) + /* + * On g33, we cannot place HWS above 256MiB, so + * restrict its pinning to the low mappable arena. + * Though this restriction is not documented for + * gen4, gen5, or byt, they also behave similarly + * and hang if the HWS is placed at the top of the + * GTT. To generalise, it appears that all !llc + * platforms have issues with us placing the HWS + * above the mappable region (even though we never + * actually map it). + */ + flags |= PIN_MAPPABLE; + else + flags |= PIN_HIGH; - i915_vma_unpin_and_release(&engine->status_page.vma, - I915_VMA_RELEASE_MAP); + return i915_vma_pin(vma, 0, 0, flags); } static int init_status_page(struct intel_engine_cs *engine) { struct drm_i915_gem_object *obj; struct i915_vma *vma; - unsigned int flags; void *vaddr; int ret; + /* + * Though the HWS register does support 36bit addresses, historically + * we have had hangs and corruption reported due to wild writes if + * the HWS is placed above 4G. We only allow objects to be allocated + * in GFP_DMA32 for i965, and no earlier physical address users had + * access to more than 4G. + */ obj = i915_gem_object_create_internal(engine->i915, PAGE_SIZE); if (IS_ERR(obj)) { DRM_ERROR("Failed to allocate status page\n"); @@ -542,59 +551,67 @@ static int init_status_page(struct intel_engine_cs *engine) goto err; } - flags = PIN_GLOBAL; - if (!HAS_LLC(engine->i915)) - /* On g33, we cannot place HWS above 256MiB, so - * restrict its pinning to the low mappable arena. - * Though this restriction is not documented for - * gen4, gen5, or byt, they also behave similarly - * and hang if the HWS is placed at the top of the - * GTT. To generalise, it appears that all !llc - * platforms have issues with us placing the HWS - * above the mappable region (even though we never - * actually map it). - */ - flags |= PIN_MAPPABLE; - else - flags |= PIN_HIGH; - ret = i915_vma_pin(vma, 0, 0, flags); - if (ret) - goto err; - vaddr = i915_gem_object_pin_map(obj, I915_MAP_WB); if (IS_ERR(vaddr)) { ret = PTR_ERR(vaddr); - goto err_unpin; + goto err; } + engine->status_page.addr = memset(vaddr, 0, PAGE_SIZE); engine->status_page.vma = vma; - engine->status_page.ggtt_offset = i915_ggtt_offset(vma); - engine->status_page.page_addr = memset(vaddr, 0, PAGE_SIZE); + + if (!HWS_NEEDS_PHYSICAL(engine->i915)) { + ret = pin_ggtt_status_page(engine, vma); + if (ret) + goto err_unpin; + } + return 0; err_unpin: - i915_vma_unpin(vma); + i915_gem_object_unpin_map(obj); err: i915_gem_object_put(obj); return ret; } -static int init_phys_status_page(struct intel_engine_cs *engine) +/** + * intel_engines_setup_common - setup engine state not requiring hw access + * @engine: Engine to setup. + * + * Initializes @engine@ structure members shared between legacy and execlists + * submission modes which do not require hardware access. + * + * Typically done early in the submission mode specific engine setup stage. + */ +int intel_engine_setup_common(struct intel_engine_cs *engine) { - struct page *page; + int err; - /* - * Though the HWS register does support 36bit addresses, historically - * we have had hangs and corruption reported due to wild writes if - * the HWS is placed above 4G. - */ - page = alloc_page(GFP_KERNEL | __GFP_DMA32 | __GFP_ZERO); - if (!page) - return -ENOMEM; + err = init_status_page(engine); + if (err) + return err; + + err = i915_timeline_init(engine->i915, + &engine->timeline, + engine->name, + engine->status_page.vma); + if (err) + goto err_hwsp; + + i915_timeline_set_subclass(&engine->timeline, TIMELINE_ENGINE); - engine->status_page.page_addr = page_address(page); + intel_engine_init_breadcrumbs(engine); + intel_engine_init_execlist(engine); + intel_engine_init_hangcheck(engine); + intel_engine_init_batch_pool(engine); + intel_engine_init_cmd_parser(engine); return 0; + +err_hwsp: + cleanup_status_page(engine); + return err; } static void __intel_context_unpin(struct i915_gem_context *ctx, @@ -603,6 +620,56 @@ static void __intel_context_unpin(struct i915_gem_context *ctx, intel_context_unpin(to_intel_context(ctx, engine)); } +struct measure_breadcrumb { + struct i915_request rq; + struct i915_timeline timeline; + struct intel_ring ring; + u32 cs[1024]; +}; + +static int measure_breadcrumb_dw(struct intel_engine_cs *engine) +{ + struct measure_breadcrumb *frame; + int dw = -ENOMEM; + + GEM_BUG_ON(!engine->i915->gt.scratch); + + frame = kzalloc(sizeof(*frame), GFP_KERNEL); + if (!frame) + return -ENOMEM; + + if (i915_timeline_init(engine->i915, + &frame->timeline, "measure", + engine->status_page.vma)) + goto out_frame; + + INIT_LIST_HEAD(&frame->ring.request_list); + frame->ring.timeline = &frame->timeline; + frame->ring.vaddr = frame->cs; + frame->ring.size = sizeof(frame->cs); + frame->ring.effective_size = frame->ring.size; + intel_ring_update_space(&frame->ring); + + frame->rq.i915 = engine->i915; + frame->rq.engine = engine; + frame->rq.ring = &frame->ring; + frame->rq.timeline = &frame->timeline; + + dw = i915_timeline_pin(&frame->timeline); + if (dw < 0) + goto out_timeline; + + dw = engine->emit_fini_breadcrumb(&frame->rq, frame->cs) - frame->cs; + + i915_timeline_unpin(&frame->timeline); + +out_timeline: + i915_timeline_fini(&frame->timeline); +out_frame: + kfree(frame); + return dw; +} + /** * intel_engines_init_common - initialize cengine state which might require hw access * @engine: Engine to initialize. @@ -645,21 +712,14 @@ int intel_engine_init_common(struct intel_engine_cs *engine) } } - ret = intel_engine_init_breadcrumbs(engine); - if (ret) + ret = measure_breadcrumb_dw(engine); + if (ret < 0) goto err_unpin_preempt; - if (HWS_NEEDS_PHYSICAL(i915)) - ret = init_phys_status_page(engine); - else - ret = init_status_page(engine); - if (ret) - goto err_breadcrumbs; + engine->emit_fini_breadcrumb_dw = ret; return 0; -err_breadcrumbs: - intel_engine_fini_breadcrumbs(engine); err_unpin_preempt: if (i915->preempt_context) __intel_context_unpin(i915->preempt_context, engine); @@ -799,15 +859,15 @@ u32 intel_calculate_mcr_s_ss_select(struct drm_i915_private *dev_priv) return mcr_s_ss_select; } -static inline uint32_t +static inline u32 read_subslice_reg(struct drm_i915_private *dev_priv, int slice, int subslice, i915_reg_t reg) { - uint32_t mcr_slice_subslice_mask; - uint32_t mcr_slice_subslice_select; - uint32_t default_mcr_s_ss_select; - uint32_t mcr; - uint32_t ret; + u32 mcr_slice_subslice_mask; + u32 mcr_slice_subslice_select; + u32 default_mcr_s_ss_select; + u32 mcr; + u32 ret; enum forcewake_domains fw_domains; if (INTEL_GEN(dev_priv) >= 11) { @@ -913,10 +973,15 @@ void intel_engine_get_instdone(struct intel_engine_cs *engine, static bool ring_is_idle(struct intel_engine_cs *engine) { struct drm_i915_private *dev_priv = engine->i915; + intel_wakeref_t wakeref; bool idle = true; + if (I915_SELFTEST_ONLY(!engine->mmio_base)) + return true; + /* If the whole device is asleep, the engine must be idle */ - if (!intel_runtime_pm_get_if_in_use(dev_priv)) + wakeref = intel_runtime_pm_get_if_in_use(dev_priv); + if (!wakeref) return true; /* First check that no commands are left in the ring */ @@ -928,7 +993,7 @@ static bool ring_is_idle(struct intel_engine_cs *engine) if (INTEL_GEN(dev_priv) > 2 && !(I915_READ_MODE(engine) & MODE_IDLE)) idle = false; - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); return idle; } @@ -952,9 +1017,6 @@ bool intel_engine_is_idle(struct intel_engine_cs *engine) if (!intel_engine_signaled(engine, intel_engine_last_submit(engine))) return false; - if (I915_SELFTEST_ONLY(engine->breadcrumbs.mock)) - return true; - /* Waiting to drain ELSP? */ if (READ_ONCE(engine->execlists.active)) { struct tasklet_struct *t = &engine->execlists.tasklet; @@ -980,10 +1042,7 @@ bool intel_engine_is_idle(struct intel_engine_cs *engine) return false; /* Ring stopped? */ - if (!ring_is_idle(engine)) - return false; - - return true; + return ring_is_idle(engine); } bool intel_engines_are_idle(struct drm_i915_private *dev_priv) @@ -1027,7 +1086,7 @@ bool intel_engine_has_kernel_context(const struct intel_engine_cs *engine) * the last request that remains in the timeline. When idle, it is * the last executed context as tracked by retirement. */ - rq = __i915_gem_active_peek(&engine->timeline.last_request); + rq = __i915_active_request_peek(&engine->timeline.last_request); if (rq) return rq->hw_context == kernel_context; else @@ -1071,10 +1130,8 @@ void intel_engines_sanitize(struct drm_i915_private *i915, bool force) if (!reset_engines(i915) && !force) return; - for_each_engine(engine, i915, id) { - if (engine->reset.reset) - engine->reset.reset(engine, NULL); - } + for_each_engine(engine, i915, id) + intel_engine_reset(engine, false); } /** @@ -1110,7 +1167,7 @@ void intel_engines_park(struct drm_i915_private *i915) } /* Must be reset upon idling, or we may miss the busy wakeup. */ - GEM_BUG_ON(engine->execlists.queue_priority != INT_MIN); + GEM_BUG_ON(engine->execlists.queue_priority_hint != INT_MIN); if (engine->park) engine->park(engine); @@ -1226,10 +1283,14 @@ static void print_request(struct drm_printer *m, x = print_sched_attr(rq->i915, &rq->sched.attr, buf, x, sizeof(buf)); - drm_printf(m, "%s%x%s [%llx:%llx]%s @ %dms: %s\n", + drm_printf(m, "%s%x%s%s [%llx:%llx]%s @ %dms: %s\n", prefix, rq->global_seqno, - i915_request_completed(rq) ? "!" : "", + i915_request_completed(rq) ? "!" : + i915_request_started(rq) ? "*" : + "", + test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, + &rq->fence.flags) ? "+" : "", rq->fence.context, rq->fence.seqno, buf, jiffies_to_msecs(jiffies - rq->emitted_jiffies), @@ -1320,7 +1381,8 @@ static void intel_engine_print_registers(const struct intel_engine_cs *engine, } if (HAS_EXECLISTS(dev_priv)) { - const u32 *hws = &engine->status_page.page_addr[I915_HWS_CSB_BUF0_INDEX]; + const u32 *hws = + &engine->status_page.addr[I915_HWS_CSB_BUF0_INDEX]; unsigned int idx; u8 read, write; @@ -1363,9 +1425,10 @@ static void intel_engine_print_registers(const struct intel_engine_cs *engine, char hdr[80]; snprintf(hdr, sizeof(hdr), - "\t\tELSP[%d] count=%d, ring->start=%08x, rq: ", + "\t\tELSP[%d] count=%d, ring:{start:%08x, hwsp:%08x}, rq: ", idx, count, - i915_ggtt_offset(rq->ring->vma)); + i915_ggtt_offset(rq->ring->vma), + rq->timeline->hwsp_offset); print_request(m, rq, hdr); } else { drm_printf(m, "\t\tELSP[%d] idle\n", idx); @@ -1420,14 +1483,9 @@ void intel_engine_dump(struct intel_engine_cs *engine, struct drm_printer *m, const char *header, ...) { - const int MAX_REQUESTS_TO_SHOW = 8; - struct intel_breadcrumbs * const b = &engine->breadcrumbs; - const struct intel_engine_execlists * const execlists = &engine->execlists; struct i915_gpu_error * const error = &engine->i915->gpu_error; - struct i915_request *rq, *last; - unsigned long flags; - struct rb_node *rb; - int count; + struct i915_request *rq; + intel_wakeref_t wakeref; if (header) { va_list ap; @@ -1477,80 +1535,30 @@ void intel_engine_dump(struct intel_engine_cs *engine, rq->ring->emit); drm_printf(m, "\t\tring->space: 0x%08x\n", rq->ring->space); + drm_printf(m, "\t\tring->hwsp: 0x%08x\n", + rq->timeline->hwsp_offset); print_request_ring(m, rq); } rcu_read_unlock(); - if (intel_runtime_pm_get_if_in_use(engine->i915)) { + wakeref = intel_runtime_pm_get_if_in_use(engine->i915); + if (wakeref) { intel_engine_print_registers(engine, m); - intel_runtime_pm_put(engine->i915); + intel_runtime_pm_put(engine->i915, wakeref); } else { drm_printf(m, "\tDevice is asleep; skipping register dump\n"); } - local_irq_save(flags); - spin_lock(&engine->timeline.lock); - - last = NULL; - count = 0; - list_for_each_entry(rq, &engine->timeline.requests, link) { - if (count++ < MAX_REQUESTS_TO_SHOW - 1) - print_request(m, rq, "\t\tE "); - else - last = rq; - } - if (last) { - if (count > MAX_REQUESTS_TO_SHOW) { - drm_printf(m, - "\t\t...skipping %d executing requests...\n", - count - MAX_REQUESTS_TO_SHOW); - } - print_request(m, last, "\t\tE "); - } - - last = NULL; - count = 0; - drm_printf(m, "\t\tQueue priority: %d\n", execlists->queue_priority); - for (rb = rb_first_cached(&execlists->queue); rb; rb = rb_next(rb)) { - struct i915_priolist *p = rb_entry(rb, typeof(*p), node); - int i; - - priolist_for_each_request(rq, p, i) { - if (count++ < MAX_REQUESTS_TO_SHOW - 1) - print_request(m, rq, "\t\tQ "); - else - last = rq; - } - } - if (last) { - if (count > MAX_REQUESTS_TO_SHOW) { - drm_printf(m, - "\t\t...skipping %d queued requests...\n", - count - MAX_REQUESTS_TO_SHOW); - } - print_request(m, last, "\t\tQ "); - } - - spin_unlock(&engine->timeline.lock); - - spin_lock(&b->rb_lock); - for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) { - struct intel_wait *w = rb_entry(rb, typeof(*w), node); - - drm_printf(m, "\t%s [%d:%c] waiting for %x\n", - w->tsk->comm, w->tsk->pid, - task_state_to_char(w->tsk), - w->seqno); - } - spin_unlock(&b->rb_lock); - local_irq_restore(flags); + intel_execlists_show_requests(engine, m, print_request, 8); drm_printf(m, "HWSP:\n"); - hexdump(m, engine->status_page.page_addr, PAGE_SIZE); + hexdump(m, engine->status_page.addr, PAGE_SIZE); drm_printf(m, "Idle? %s\n", yesno(intel_engine_is_idle(engine))); + + intel_engine_print_breadcrumbs(engine, m); } static u8 user_class_map[] = { diff --git a/drivers/gpu/drm/i915/intel_fbc.c b/drivers/gpu/drm/i915/intel_fbc.c index 3a93cdd1dd91..656e684e7c9a 100644 --- a/drivers/gpu/drm/i915/intel_fbc.c +++ b/drivers/gpu/drm/i915/intel_fbc.c @@ -596,7 +596,7 @@ static bool stride_is_valid(struct drm_i915_private *dev_priv, } static bool pixel_format_is_valid(struct drm_i915_private *dev_priv, - uint32_t pixel_format) + u32 pixel_format) { switch (pixel_format) { case DRM_FORMAT_XRGB8888: diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c index d54e7c0ce93c..376ffe842e26 100644 --- a/drivers/gpu/drm/i915/intel_fbdev.c +++ b/drivers/gpu/drm/i915/intel_fbdev.c @@ -179,8 +179,9 @@ static int intelfb_create(struct drm_fb_helper *helper, const struct i915_ggtt_view view = { .type = I915_GGTT_VIEW_NORMAL, }; - struct fb_info *info; struct drm_framebuffer *fb; + intel_wakeref_t wakeref; + struct fb_info *info; struct i915_vma *vma; unsigned long flags = 0; bool prealloc = false; @@ -211,7 +212,7 @@ static int intelfb_create(struct drm_fb_helper *helper, } mutex_lock(&dev->struct_mutex); - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); /* Pin the GGTT vma for our access via info->screen_base. * This also validates that any existing fb inherited from the @@ -278,7 +279,7 @@ static int intelfb_create(struct drm_fb_helper *helper, ifbdev->vma = vma; ifbdev->vma_flags = flags; - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); mutex_unlock(&dev->struct_mutex); vga_switcheroo_client_fb_set(pdev, info); return 0; @@ -286,7 +287,7 @@ static int intelfb_create(struct drm_fb_helper *helper, out_unpin: intel_unpin_fb_vma(vma, flags); out_unlock: - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); mutex_unlock(&dev->struct_mutex); return ret; } @@ -680,6 +681,7 @@ int intel_fbdev_init(struct drm_device *dev) if (ifbdev == NULL) return -ENOMEM; + mutex_init(&ifbdev->hpd_lock); drm_fb_helper_prepare(dev, &ifbdev->helper, &intel_fb_helper_funcs); if (!intel_fbdev_init_bios(dev, ifbdev)) @@ -753,6 +755,26 @@ void intel_fbdev_fini(struct drm_i915_private *dev_priv) intel_fbdev_destroy(ifbdev); } +/* Suspends/resumes fbdev processing of incoming HPD events. When resuming HPD + * processing, fbdev will perform a full connector reprobe if a hotplug event + * was received while HPD was suspended. + */ +static void intel_fbdev_hpd_set_suspend(struct intel_fbdev *ifbdev, int state) +{ + bool send_hpd = false; + + mutex_lock(&ifbdev->hpd_lock); + ifbdev->hpd_suspended = state == FBINFO_STATE_SUSPENDED; + send_hpd = !ifbdev->hpd_suspended && ifbdev->hpd_waiting; + ifbdev->hpd_waiting = false; + mutex_unlock(&ifbdev->hpd_lock); + + if (send_hpd) { + DRM_DEBUG_KMS("Handling delayed fbcon HPD event\n"); + drm_fb_helper_hotplug_event(&ifbdev->helper); + } +} + void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous) { struct drm_i915_private *dev_priv = to_i915(dev); @@ -774,6 +796,7 @@ void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous */ if (state != FBINFO_STATE_RUNNING) flush_work(&dev_priv->fbdev_suspend_work); + console_lock(); } else { /* @@ -801,17 +824,26 @@ void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous drm_fb_helper_set_suspend(&ifbdev->helper, state); console_unlock(); + + intel_fbdev_hpd_set_suspend(ifbdev, state); } void intel_fbdev_output_poll_changed(struct drm_device *dev) { struct intel_fbdev *ifbdev = to_i915(dev)->fbdev; + bool send_hpd; if (!ifbdev) return; intel_fbdev_sync(ifbdev); - if (ifbdev->vma || ifbdev->helper.deferred_setup) + + mutex_lock(&ifbdev->hpd_lock); + send_hpd = !ifbdev->hpd_suspended; + ifbdev->hpd_waiting = true; + mutex_unlock(&ifbdev->hpd_lock); + + if (send_hpd && (ifbdev->vma || ifbdev->helper.deferred_setup)) drm_fb_helper_hotplug_event(&ifbdev->helper); } diff --git a/drivers/gpu/drm/i915/intel_fifo_underrun.c b/drivers/gpu/drm/i915/intel_fifo_underrun.c index 9b39975c8389..f33de4be4b89 100644 --- a/drivers/gpu/drm/i915/intel_fifo_underrun.c +++ b/drivers/gpu/drm/i915/intel_fifo_underrun.c @@ -127,8 +127,8 @@ static void ironlake_set_fifo_underrun_reporting(struct drm_device *dev, enum pipe pipe, bool enable) { struct drm_i915_private *dev_priv = to_i915(dev); - uint32_t bit = (pipe == PIPE_A) ? DE_PIPEA_FIFO_UNDERRUN : - DE_PIPEB_FIFO_UNDERRUN; + u32 bit = (pipe == PIPE_A) ? + DE_PIPEA_FIFO_UNDERRUN : DE_PIPEB_FIFO_UNDERRUN; if (enable) ilk_enable_display_irq(dev_priv, bit); @@ -140,7 +140,7 @@ static void ivybridge_check_fifo_underruns(struct intel_crtc *crtc) { struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pipe = crtc->pipe; - uint32_t err_int = I915_READ(GEN7_ERR_INT); + u32 err_int = I915_READ(GEN7_ERR_INT); lockdep_assert_held(&dev_priv->irq_lock); @@ -193,8 +193,8 @@ static void ibx_set_fifo_underrun_reporting(struct drm_device *dev, bool enable) { struct drm_i915_private *dev_priv = to_i915(dev); - uint32_t bit = (pch_transcoder == PIPE_A) ? - SDE_TRANSA_FIFO_UNDER : SDE_TRANSB_FIFO_UNDER; + u32 bit = (pch_transcoder == PIPE_A) ? + SDE_TRANSA_FIFO_UNDER : SDE_TRANSB_FIFO_UNDER; if (enable) ibx_enable_display_interrupt(dev_priv, bit); @@ -206,7 +206,7 @@ static void cpt_check_pch_fifo_underruns(struct intel_crtc *crtc) { struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum pipe pch_transcoder = crtc->pipe; - uint32_t serr_int = I915_READ(SERR_INT); + u32 serr_int = I915_READ(SERR_INT); lockdep_assert_held(&dev_priv->irq_lock); @@ -258,7 +258,7 @@ static bool __intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev, old = !crtc->cpu_fifo_underrun_disabled; crtc->cpu_fifo_underrun_disabled = !enable; - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) i9xx_set_fifo_underrun_reporting(dev, pipe, enable, old); else if (IS_GEN_RANGE(dev_priv, 5, 6)) ironlake_set_fifo_underrun_reporting(dev, pipe, enable); @@ -369,7 +369,7 @@ void intel_cpu_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv, return; /* GMCH can't disable fifo underruns, filter them. */ - if (HAS_GMCH_DISPLAY(dev_priv) && + if (HAS_GMCH(dev_priv) && crtc->cpu_fifo_underrun_disabled) return; @@ -421,7 +421,7 @@ void intel_check_cpu_fifo_underruns(struct drm_i915_private *dev_priv) if (crtc->cpu_fifo_underrun_disabled) continue; - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) i9xx_check_fifo_underruns(crtc); else if (IS_GEN(dev_priv, 7)) ivybridge_check_fifo_underruns(crtc); diff --git a/drivers/gpu/drm/i915/intel_gpu_commands.h b/drivers/gpu/drm/i915/intel_gpu_commands.h index 105e2a9e874a..b96a31bc1080 100644 --- a/drivers/gpu/drm/i915/intel_gpu_commands.h +++ b/drivers/gpu/drm/i915/intel_gpu_commands.h @@ -112,7 +112,6 @@ #define MI_MEM_VIRTUAL (1 << 22) /* 945,g33,965 */ #define MI_USE_GGTT (1 << 22) /* g4x+ */ #define MI_STORE_DWORD_INDEX MI_INSTR(0x21, 1) -#define MI_STORE_DWORD_INDEX_SHIFT 2 /* * Official intel docs are somewhat sloppy concerning MI_LOAD_REGISTER_IMM: * - Always issue a MI_NOOP _before_ the MI_LOAD_REGISTER_IMM - otherwise hw diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index 0f1c4f9ebfd8..744220296653 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -192,4 +192,7 @@ static inline void intel_guc_disable_msg(struct intel_guc *guc, u32 mask) spin_unlock_irq(&guc->irq_lock); } +int intel_guc_reset_engine(struct intel_guc *guc, + struct intel_engine_cs *engine); + #endif diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index d3ebdbc0182e..806fdfd7c78a 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -140,6 +140,9 @@ static struct dentry *create_buf_file_callback(const char *filename, buf_file = debugfs_create_file(filename, mode, parent, buf, &relay_file_operations); + if (IS_ERR(buf_file)) + return NULL; + return buf_file; } @@ -436,6 +439,7 @@ static void guc_log_capture_logs(struct intel_guc_log *log) { struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); + intel_wakeref_t wakeref; guc_read_update_log_buffer(log); @@ -443,9 +447,8 @@ static void guc_log_capture_logs(struct intel_guc_log *log) * Generally device is expected to be active only at this * time, so get/put should be really quick. */ - intel_runtime_pm_get(dev_priv); - guc_action_flush_log_complete(guc); - intel_runtime_pm_put(dev_priv); + with_intel_runtime_pm(dev_priv, wakeref) + guc_action_flush_log_complete(guc); } int intel_guc_log_create(struct intel_guc_log *log) @@ -505,7 +508,8 @@ int intel_guc_log_set_level(struct intel_guc_log *log, u32 level) { struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *dev_priv = guc_to_i915(guc); - int ret; + intel_wakeref_t wakeref; + int ret = 0; BUILD_BUG_ON(GUC_LOG_VERBOSITY_MIN != 0); GEM_BUG_ON(!log->vma); @@ -519,16 +523,14 @@ int intel_guc_log_set_level(struct intel_guc_log *log, u32 level) mutex_lock(&dev_priv->drm.struct_mutex); - if (log->level == level) { - ret = 0; + if (log->level == level) goto out_unlock; - } - intel_runtime_pm_get(dev_priv); - ret = guc_action_control_log(guc, GUC_LOG_LEVEL_IS_VERBOSE(level), - GUC_LOG_LEVEL_IS_ENABLED(level), - GUC_LOG_LEVEL_TO_VERBOSITY(level)); - intel_runtime_pm_put(dev_priv); + with_intel_runtime_pm(dev_priv, wakeref) + ret = guc_action_control_log(guc, + GUC_LOG_LEVEL_IS_VERBOSE(level), + GUC_LOG_LEVEL_IS_ENABLED(level), + GUC_LOG_LEVEL_TO_VERBOSITY(level)); if (ret) { DRM_DEBUG_DRIVER("guc_log_control action failed %d\n", ret); goto out_unlock; @@ -601,6 +603,7 @@ void intel_guc_log_relay_flush(struct intel_guc_log *log) { struct intel_guc *guc = log_to_guc(log); struct drm_i915_private *i915 = guc_to_i915(guc); + intel_wakeref_t wakeref; /* * Before initiating the forceful flush, wait for any pending/ongoing @@ -608,9 +611,8 @@ void intel_guc_log_relay_flush(struct intel_guc_log *log) */ flush_work(&log->relay.flush_work); - intel_runtime_pm_get(i915); - guc_action_flush_log(guc); - intel_runtime_pm_put(i915); + with_intel_runtime_pm(i915, wakeref) + guc_action_flush_log(guc); /* GuC would have updated log buffer by now, so capture it */ guc_log_capture_logs(log); diff --git a/drivers/gpu/drm/i915/intel_guc_submission.c b/drivers/gpu/drm/i915/intel_guc_submission.c index ab1c49b106f2..8bc8aa54aa35 100644 --- a/drivers/gpu/drm/i915/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/intel_guc_submission.c @@ -81,6 +81,12 @@ * */ +static inline u32 intel_hws_preempt_done_address(struct intel_engine_cs *engine) +{ + return (i915_ggtt_offset(engine->status_page.vma) + + I915_GEM_HWS_PREEMPT_ADDR); +} + static inline struct i915_priolist *to_priolist(struct rb_node *rb) { return rb_entry(rb, struct i915_priolist, node); @@ -623,6 +629,8 @@ static void inject_preempt_context(struct work_struct *work) EXECLISTS_ACTIVE_PREEMPT); tasklet_schedule(&engine->execlists.tasklet); } + + (void)I915_SELFTEST_ONLY(engine->execlists.preempt_hang.count++); } /* @@ -666,7 +674,7 @@ static void complete_preempt_context(struct intel_engine_cs *engine) execlists_unwind_incomplete_requests(execlists); wait_for_guc_preempt_report(engine); - intel_write_status_page(engine, I915_GEM_HWS_PREEMPT_INDEX, 0); + intel_write_status_page(engine, I915_GEM_HWS_PREEMPT, 0); } /** @@ -731,7 +739,7 @@ static bool __guc_dequeue(struct intel_engine_cs *engine) if (intel_engine_has_preemption(engine)) { struct guc_preempt_work *preempt_work = &engine->i915->guc.preempt_work[engine->id]; - int prio = execlists->queue_priority; + int prio = execlists->queue_priority_hint; if (__execlists_need_preempt(prio, port_prio(port))) { execlists_set_active(execlists, @@ -777,7 +785,8 @@ static bool __guc_dequeue(struct intel_engine_cs *engine) kmem_cache_free(engine->i915->priorities, p); } done: - execlists->queue_priority = rb ? to_priolist(rb)->priority : INT_MIN; + execlists->queue_priority_hint = + rb ? to_priolist(rb)->priority : INT_MIN; if (submit) port_assign(port, last); if (last) @@ -824,7 +833,7 @@ static void guc_submission_tasklet(unsigned long data) } if (execlists_is_active(execlists, EXECLISTS_ACTIVE_PREEMPT) && - intel_read_status_page(engine, I915_GEM_HWS_PREEMPT_INDEX) == + intel_read_status_page(engine, I915_GEM_HWS_PREEMPT) == GUC_PREEMPT_FINISHED) complete_preempt_context(engine); @@ -834,8 +843,7 @@ static void guc_submission_tasklet(unsigned long data) spin_unlock_irqrestore(&engine->timeline.lock, flags); } -static struct i915_request * -guc_reset_prepare(struct intel_engine_cs *engine) +static void guc_reset_prepare(struct intel_engine_cs *engine) { struct intel_engine_execlists * const execlists = &engine->execlists; @@ -861,8 +869,6 @@ guc_reset_prepare(struct intel_engine_cs *engine) */ if (engine->i915->guc.preempt_wq) flush_workqueue(engine->i915->guc.preempt_wq); - - return i915_gem_find_active_request(engine); } /* diff --git a/drivers/gpu/drm/i915/intel_gvt.c b/drivers/gpu/drm/i915/intel_gvt.c index c22b3e18a0f5..1d7d26e4cf14 100644 --- a/drivers/gpu/drm/i915/intel_gvt.c +++ b/drivers/gpu/drm/i915/intel_gvt.c @@ -49,6 +49,9 @@ static bool is_supported_device(struct drm_i915_private *dev_priv) return true; if (IS_BROXTON(dev_priv)) return true; + if (IS_COFFEELAKE(dev_priv)) + return true; + return false; } @@ -105,15 +108,6 @@ int intel_gvt_init(struct drm_i915_private *dev_priv) return -EIO; } - /* - * We're not in host or fail to find a MPT module, disable GVT-g - */ - ret = intel_gvt_init_host(); - if (ret) { - DRM_DEBUG_DRIVER("Not in host or MPT modules not found\n"); - goto bail; - } - ret = intel_gvt_init_device(dev_priv); if (ret) { DRM_DEBUG_DRIVER("Fail to init GVT device\n"); diff --git a/drivers/gpu/drm/i915/intel_hangcheck.c b/drivers/gpu/drm/i915/intel_hangcheck.c index 51e9efec5116..a219c796e56d 100644 --- a/drivers/gpu/drm/i915/intel_hangcheck.c +++ b/drivers/gpu/drm/i915/intel_hangcheck.c @@ -23,6 +23,18 @@ */ #include "i915_drv.h" +#include "i915_reset.h" + +struct hangcheck { + u64 acthd; + u32 seqno; + enum intel_engine_hangcheck_action action; + unsigned long action_timestamp; + int deadlock; + struct intel_instdone instdone; + bool wedged:1; + bool stalled:1; +}; static bool instdone_unchanged(u32 current_instdone, u32 *old_instdone) { @@ -118,25 +130,22 @@ engine_stuck(struct intel_engine_cs *engine, u64 acthd) } static void hangcheck_load_sample(struct intel_engine_cs *engine, - struct intel_engine_hangcheck *hc) + struct hangcheck *hc) { hc->acthd = intel_engine_get_active_head(engine); hc->seqno = intel_engine_get_seqno(engine); } static void hangcheck_store_sample(struct intel_engine_cs *engine, - const struct intel_engine_hangcheck *hc) + const struct hangcheck *hc) { engine->hangcheck.acthd = hc->acthd; engine->hangcheck.seqno = hc->seqno; - engine->hangcheck.action = hc->action; - engine->hangcheck.stalled = hc->stalled; - engine->hangcheck.wedged = hc->wedged; } static enum intel_engine_hangcheck_action hangcheck_get_action(struct intel_engine_cs *engine, - const struct intel_engine_hangcheck *hc) + const struct hangcheck *hc) { if (engine->hangcheck.seqno != hc->seqno) return ENGINE_ACTIVE_SEQNO; @@ -148,7 +157,7 @@ hangcheck_get_action(struct intel_engine_cs *engine, } static void hangcheck_accumulate_sample(struct intel_engine_cs *engine, - struct intel_engine_hangcheck *hc) + struct hangcheck *hc) { unsigned long timeout = I915_ENGINE_DEAD_TIMEOUT; @@ -194,10 +203,6 @@ static void hangcheck_accumulate_sample(struct intel_engine_cs *engine, break; case ENGINE_DEAD: - if (GEM_SHOW_DEBUG()) { - struct drm_printer p = drm_debug_printer("hangcheck"); - intel_engine_dump(engine, &p, "%s\n", engine->name); - } break; default: @@ -268,22 +273,35 @@ static void i915_hangcheck_elapsed(struct work_struct *work) intel_uncore_arm_unclaimed_mmio_detection(dev_priv); for_each_engine(engine, dev_priv, id) { - struct intel_engine_hangcheck hc; + struct hangcheck hc; + + intel_engine_signal_breadcrumbs(engine); hangcheck_load_sample(engine, &hc); hangcheck_accumulate_sample(engine, &hc); hangcheck_store_sample(engine, &hc); - if (engine->hangcheck.stalled) { + if (hc.stalled) { hung |= intel_engine_flag(engine); if (hc.action != ENGINE_DEAD) stuck |= intel_engine_flag(engine); } - if (engine->hangcheck.wedged) + if (hc.wedged) wedged |= intel_engine_flag(engine); } + if (GEM_SHOW_DEBUG() && (hung | stuck)) { + struct drm_printer p = drm_debug_printer("hangcheck"); + + for_each_engine(engine, dev_priv, id) { + if (intel_engine_is_idle(engine)) + continue; + + intel_engine_dump(engine, &p, "%s\n", engine->name); + } + } + if (wedged) { dev_err(dev_priv->drm.dev, "GPU recovery timed out," diff --git a/drivers/gpu/drm/i915/intel_hdcp.c b/drivers/gpu/drm/i915/intel_hdcp.c index 3fcb3b775948..ce7ba3a9c000 100644 --- a/drivers/gpu/drm/i915/intel_hdcp.c +++ b/drivers/gpu/drm/i915/intel_hdcp.c @@ -838,8 +838,8 @@ void intel_hdcp_atomic_check(struct drm_connector *connector, struct drm_connector_state *old_state, struct drm_connector_state *new_state) { - uint64_t old_cp = old_state->content_protection; - uint64_t new_cp = new_state->content_protection; + u64 old_cp = old_state->content_protection; + u64 new_cp = new_state->content_protection; struct drm_crtc_state *crtc_state; if (!new_state->crtc) { diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c index ca195e6203c5..f125a62eba8c 100644 --- a/drivers/gpu/drm/i915/intel_hdmi.c +++ b/drivers/gpu/drm/i915/intel_hdmi.c @@ -1186,15 +1186,17 @@ static bool intel_hdmi_get_hw_state(struct intel_encoder *encoder, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base); + intel_wakeref_t wakeref; bool ret; - if (!intel_display_power_get_if_enabled(dev_priv, - encoder->power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + encoder->power_domain); + if (!wakeref) return false; ret = intel_sdvo_port_enabled(dev_priv, intel_hdmi->hdmi_reg, pipe); - intel_display_power_put(dev_priv, encoder->power_domain); + intel_display_power_put(dev_priv, encoder->power_domain, wakeref); return ret; } @@ -1586,7 +1588,7 @@ intel_hdmi_mode_valid(struct drm_connector *connector, if (hdmi->has_hdmi_sink && !force_dvi) { /* if we can't do 8bpc we may still be able to do 12bpc */ - if (status != MODE_OK && !HAS_GMCH_DISPLAY(dev_priv)) + if (status != MODE_OK && !HAS_GMCH(dev_priv)) status = hdmi_port_clock_valid(hdmi, clock * 3 / 2, true, force_dvi); @@ -1611,7 +1613,7 @@ static bool hdmi_deep_color_possible(const struct intel_crtc_state *crtc_state, &crtc_state->base.adjusted_mode; int i; - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) return false; if (bpc == 10 && INTEL_GEN(dev_priv) < 11) @@ -1890,11 +1892,12 @@ intel_hdmi_set_edid(struct drm_connector *connector) { struct drm_i915_private *dev_priv = to_i915(connector->dev); struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector); + intel_wakeref_t wakeref; struct edid *edid; bool connected = false; struct i2c_adapter *i2c; - intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS); + wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS); i2c = intel_gmbus_get_adapter(dev_priv, intel_hdmi->ddc_bus); @@ -1909,7 +1912,7 @@ intel_hdmi_set_edid(struct drm_connector *connector) intel_hdmi_dp_dual_mode_detect(connector, edid != NULL); - intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS); + intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS, wakeref); to_intel_connector(connector)->detect_edid = edid; if (edid && edid->input & DRM_EDID_INPUT_DIGITAL) { @@ -1931,11 +1934,12 @@ intel_hdmi_detect(struct drm_connector *connector, bool force) struct drm_i915_private *dev_priv = to_i915(connector->dev); struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector); struct intel_encoder *encoder = &hdmi_to_dig_port(intel_hdmi)->base; + intel_wakeref_t wakeref; DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n", connector->base.id, connector->name); - intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS); + wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS); if (IS_ICELAKE(dev_priv) && !intel_digital_port_connected(encoder)) @@ -1947,7 +1951,7 @@ intel_hdmi_detect(struct drm_connector *connector, bool force) status = connector_status_connected; out: - intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS); + intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS, wakeref); if (status != connector_status_connected) cec_notifier_phys_addr_invalidate(intel_hdmi->cec_notifier); @@ -2146,7 +2150,7 @@ intel_hdmi_add_properties(struct intel_hdmi *intel_hdmi, struct drm_connector *c drm_connector_attach_content_type_property(connector); connector->state->picture_aspect_ratio = HDMI_PICTURE_ASPECT_NONE; - if (!HAS_GMCH_DISPLAY(dev_priv)) + if (!HAS_GMCH(dev_priv)) drm_connector_attach_max_bpc_property(connector, 8, 12); } diff --git a/drivers/gpu/drm/i915/intel_hotplug.c b/drivers/gpu/drm/i915/intel_hotplug.c index ae92d6560165..b8937c788f03 100644 --- a/drivers/gpu/drm/i915/intel_hotplug.c +++ b/drivers/gpu/drm/i915/intel_hotplug.c @@ -226,9 +226,10 @@ static void intel_hpd_irq_storm_reenable_work(struct work_struct *work) container_of(work, typeof(*dev_priv), hotplug.reenable_work.work); struct drm_device *dev = &dev_priv->drm; + intel_wakeref_t wakeref; enum hpd_pin pin; - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); spin_lock_irq(&dev_priv->irq_lock); for_each_hpd_pin(pin) { @@ -261,7 +262,7 @@ static void intel_hpd_irq_storm_reenable_work(struct work_struct *work) dev_priv->display.hpd_irq_setup(dev_priv); spin_unlock_irq(&dev_priv->irq_lock); - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); } bool intel_encoder_hotplug(struct intel_encoder *encoder, @@ -469,7 +470,7 @@ void intel_hpd_irq_handler(struct drm_i915_private *dev_priv, * hotplug bits itself. So only WARN about unexpected * interrupts on saner platforms. */ - WARN_ONCE(!HAS_GMCH_DISPLAY(dev_priv), + WARN_ONCE(!HAS_GMCH(dev_priv), "Received HPD interrupt on pin %d although disabled\n", pin); continue; } diff --git a/drivers/gpu/drm/i915/intel_huc.c b/drivers/gpu/drm/i915/intel_huc.c index bc27b691d824..9bd1c9002c2a 100644 --- a/drivers/gpu/drm/i915/intel_huc.c +++ b/drivers/gpu/drm/i915/intel_huc.c @@ -115,14 +115,14 @@ fail: int intel_huc_check_status(struct intel_huc *huc) { struct drm_i915_private *dev_priv = huc_to_i915(huc); - bool status; + intel_wakeref_t wakeref; + bool status = false; if (!HAS_HUC(dev_priv)) return -ENODEV; - intel_runtime_pm_get(dev_priv); - status = I915_READ(HUC_STATUS2) & HUC_FW_VERIFIED; - intel_runtime_pm_put(dev_priv); + with_intel_runtime_pm(dev_priv, wakeref) + status = I915_READ(HUC_STATUS2) & HUC_FW_VERIFIED; return status; } diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c index c6159aff9dc8..5a733e711355 100644 --- a/drivers/gpu/drm/i915/intel_i2c.c +++ b/drivers/gpu/drm/i915/intel_i2c.c @@ -697,12 +697,13 @@ out: static int gmbus_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs, int num) { - struct intel_gmbus *bus = container_of(adapter, struct intel_gmbus, - adapter); + struct intel_gmbus *bus = + container_of(adapter, struct intel_gmbus, adapter); struct drm_i915_private *dev_priv = bus->dev_priv; + intel_wakeref_t wakeref; int ret; - intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS); + wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS); if (bus->force_bit) { ret = i2c_bit_algo.master_xfer(adapter, msgs, num); @@ -714,17 +715,16 @@ gmbus_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs, int num) bus->force_bit |= GMBUS_FORCE_BIT_RETRY; } - intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS); + intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS, wakeref); return ret; } int intel_gmbus_output_aksv(struct i2c_adapter *adapter) { - struct intel_gmbus *bus = container_of(adapter, struct intel_gmbus, - adapter); + struct intel_gmbus *bus = + container_of(adapter, struct intel_gmbus, adapter); struct drm_i915_private *dev_priv = bus->dev_priv; - int ret; u8 cmd = DRM_HDCP_DDC_AKSV; u8 buf[DRM_HDCP_KSV_LEN] = { 0 }; struct i2c_msg msgs[] = { @@ -741,8 +741,10 @@ int intel_gmbus_output_aksv(struct i2c_adapter *adapter) .buf = buf, } }; + intel_wakeref_t wakeref; + int ret; - intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS); + wakeref = intel_display_power_get(dev_priv, POWER_DOMAIN_GMBUS); mutex_lock(&dev_priv->gmbus_mutex); /* @@ -753,7 +755,7 @@ int intel_gmbus_output_aksv(struct i2c_adapter *adapter) ret = do_gmbus_xfer(adapter, msgs, ARRAY_SIZE(msgs), GMBUS_AKSV_SELECT); mutex_unlock(&dev_priv->gmbus_mutex); - intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS); + intel_display_power_put(dev_priv, POWER_DOMAIN_GMBUS, wakeref); return ret; } @@ -821,7 +823,7 @@ int intel_setup_gmbus(struct drm_i915_private *dev_priv) if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) dev_priv->gpio_mmio_base = VLV_DISPLAY_BASE; - else if (!HAS_GMCH_DISPLAY(dev_priv)) + else if (!HAS_GMCH(dev_priv)) /* * Broxton uses the same PCH offsets for South Display Engine, * even though it doesn't have a PCH. diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 1effbf49fa08..5e98fd79bd9d 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -136,6 +136,7 @@ #include <drm/i915_drm.h> #include "i915_drv.h" #include "i915_gem_render_state.h" +#include "i915_reset.h" #include "i915_vgpu.h" #include "intel_lrc_reg.h" #include "intel_mocs.h" @@ -171,6 +172,12 @@ static void execlists_init_reg_state(u32 *reg_state, struct intel_engine_cs *engine, struct intel_ring *ring); +static inline u32 intel_hws_seqno_address(struct intel_engine_cs *engine) +{ + return (i915_ggtt_offset(engine->status_page.vma) + + I915_GEM_HWS_INDEX_ADDR); +} + static inline struct i915_priolist *to_priolist(struct rb_node *rb) { return rb_entry(rb, struct i915_priolist, node); @@ -181,13 +188,90 @@ static inline int rq_prio(const struct i915_request *rq) return rq->sched.attr.priority; } +static int queue_prio(const struct intel_engine_execlists *execlists) +{ + struct i915_priolist *p; + struct rb_node *rb; + + rb = rb_first_cached(&execlists->queue); + if (!rb) + return INT_MIN; + + /* + * As the priolist[] are inverted, with the highest priority in [0], + * we have to flip the index value to become priority. + */ + p = to_priolist(rb); + return ((p->priority + 1) << I915_USER_PRIORITY_SHIFT) - ffs(p->used); +} + static inline bool need_preempt(const struct intel_engine_cs *engine, - const struct i915_request *last, - int prio) + const struct i915_request *rq) +{ + const int last_prio = rq_prio(rq); + + if (!intel_engine_has_preemption(engine)) + return false; + + if (i915_request_completed(rq)) + return false; + + /* + * Check if the current priority hint merits a preemption attempt. + * + * We record the highest value priority we saw during rescheduling + * prior to this dequeue, therefore we know that if it is strictly + * less than the current tail of ESLP[0], we do not need to force + * a preempt-to-idle cycle. + * + * However, the priority hint is a mere hint that we may need to + * preempt. If that hint is stale or we may be trying to preempt + * ourselves, ignore the request. + */ + if (!__execlists_need_preempt(engine->execlists.queue_priority_hint, + last_prio)) + return false; + + /* + * Check against the first request in ELSP[1], it will, thanks to the + * power of PI, be the highest priority of that context. + */ + if (!list_is_last(&rq->link, &engine->timeline.requests) && + rq_prio(list_next_entry(rq, link)) > last_prio) + return true; + + /* + * If the inflight context did not trigger the preemption, then maybe + * it was the set of queued requests? Pick the highest priority in + * the queue (the first active priolist) and see if it deserves to be + * running instead of ELSP[0]. + * + * The highest priority request in the queue can not be either + * ELSP[0] or ELSP[1] as, thanks again to PI, if it was the same + * context, it's priority would not exceed ELSP[0] aka last_prio. + */ + return queue_prio(&engine->execlists) > last_prio; +} + +__maybe_unused static inline bool +assert_priority_queue(const struct intel_engine_execlists *execlists, + const struct i915_request *prev, + const struct i915_request *next) { - return (intel_engine_has_preemption(engine) && - __execlists_need_preempt(prio, rq_prio(last)) && - !i915_request_completed(last)); + if (!prev) + return true; + + /* + * Without preemption, the prev may refer to the still active element + * which we refuse to let go. + * + * Even with preemption, there are times when we think it is better not + * to preempt and leave an ostensibly lower priority request in flight. + */ + if (port_request(execlists->port) == prev) + return true; + + return rq_prio(prev) >= rq_prio(next); } /* @@ -264,7 +348,8 @@ static void unwind_wa_tail(struct i915_request *rq) assert_ring_tail_valid(rq->ring, rq->tail); } -static void __unwind_incomplete_requests(struct intel_engine_cs *engine) +static struct i915_request * +__unwind_incomplete_requests(struct intel_engine_cs *engine) { struct i915_request *rq, *rn, *active = NULL; struct list_head *uninitialized_var(pl); @@ -302,9 +387,12 @@ static void __unwind_incomplete_requests(struct intel_engine_cs *engine) */ if (!(prio & I915_PRIORITY_NEWCLIENT)) { prio |= I915_PRIORITY_NEWCLIENT; + active->sched.attr.priority = prio; list_move_tail(&active->sched.link, i915_sched_lookup_priolist(engine, prio)); } + + return active; } void @@ -435,11 +523,12 @@ static void execlists_submit_ports(struct intel_engine_cs *engine) desc = execlists_update_context(rq); GEM_DEBUG_EXEC(port[n].context_id = upper_32_bits(desc)); - GEM_TRACE("%s in[%d]: ctx=%d.%d, global=%d (fence %llx:%lld) (current %d), prio=%d\n", + GEM_TRACE("%s in[%d]: ctx=%d.%d, global=%d (fence %llx:%lld) (current %d:%d), prio=%d\n", engine->name, n, port[n].context_id, count, rq->global_seqno, rq->fence.context, rq->fence.seqno, + hwsp_seqno(rq), intel_engine_get_seqno(engine), rq_prio(rq)); } else { @@ -511,6 +600,8 @@ static void inject_preempt_context(struct intel_engine_cs *engine) execlists_clear_active(execlists, EXECLISTS_ACTIVE_HWACK); execlists_set_active(execlists, EXECLISTS_ACTIVE_PREEMPT); + + (void)I915_SELFTEST_ONLY(execlists->preempt_hang.count++); } static void complete_preempt_context(struct intel_engine_execlists *execlists) @@ -579,7 +670,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine) if (!execlists_is_active(execlists, EXECLISTS_ACTIVE_HWACK)) return; - if (need_preempt(engine, last, execlists->queue_priority)) { + if (need_preempt(engine, last)) { inject_preempt_context(engine); return; } @@ -612,7 +703,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine) * WaIdleLiteRestore:bdw,skl * Apply the wa NOOPs to prevent * ring:HEAD == rq:TAIL as we resubmit the - * request. See gen8_emit_breadcrumb() for + * request. See gen8_emit_fini_breadcrumb() for * where we prepare the padding after the * end of the request. */ @@ -625,6 +716,8 @@ static void execlists_dequeue(struct intel_engine_cs *engine) int i; priolist_for_each_request_consume(rq, rn, p, i) { + GEM_BUG_ON(!assert_priority_queue(execlists, last, rq)); + /* * Can we combine this request with the current port? * It has to be the same context/ringbuffer and not @@ -684,20 +777,20 @@ done: /* * Here be a bit of magic! Or sleight-of-hand, whichever you prefer. * - * We choose queue_priority such that if we add a request of greater + * We choose the priority hint such that if we add a request of greater * priority than this, we kick the submission tasklet to decide on * the right order of submitting the requests to hardware. We must * also be prepared to reorder requests as they are in-flight on the - * HW. We derive the queue_priority then as the first "hole" in + * HW. We derive the priority hint then as the first "hole" in * the HW submission ports and if there are no available slots, * the priority of the lowest executing request, i.e. last. * * When we do receive a higher priority request ready to run from the - * user, see queue_request(), the queue_priority is bumped to that + * user, see queue_request(), the priority hint is bumped to that * request triggering preemption on the next dequeue (or subsequent * interrupt for secondary ports). */ - execlists->queue_priority = + execlists->queue_priority_hint = port != execlists->port ? rq_prio(last) : INT_MIN; if (submit) { @@ -728,11 +821,12 @@ execlists_cancel_port_requests(struct intel_engine_execlists * const execlists) while (num_ports-- && port_isset(port)) { struct i915_request *rq = port_request(port); - GEM_TRACE("%s:port%u global=%d (fence %llx:%lld), (current %d)\n", + GEM_TRACE("%s:port%u global=%d (fence %llx:%lld), (current %d:%d)\n", rq->engine->name, (unsigned int)(port - execlists->port), rq->global_seqno, rq->fence.context, rq->fence.seqno, + hwsp_seqno(rq), intel_engine_get_seqno(rq->engine)); GEM_BUG_ON(!execlists->active); @@ -816,10 +910,10 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) list_for_each_entry(rq, &engine->timeline.requests, link) { GEM_BUG_ON(!rq->global_seqno); - if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags)) - continue; + if (!i915_request_signaled(rq)) + dma_fence_set_error(&rq->fence, -EIO); - dma_fence_set_error(&rq->fence, -EIO); + i915_request_mark_complete(rq); } /* Flush the queued requests to the timeline list (for retiring). */ @@ -829,9 +923,9 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) priolist_for_each_request_consume(rq, rn, p, i) { list_del_init(&rq->sched.link); - - dma_fence_set_error(&rq->fence, -EIO); __i915_request_submit(rq); + dma_fence_set_error(&rq->fence, -EIO); + i915_request_mark_complete(rq); } rb_erase_cached(&p->node, &execlists->queue); @@ -845,7 +939,7 @@ static void execlists_cancel_requests(struct intel_engine_cs *engine) /* Remaining _unready_ requests will be nop'ed when submitted */ - execlists->queue_priority = INT_MIN; + execlists->queue_priority_hint = INT_MIN; execlists->queue = RB_ROOT_CACHED; GEM_BUG_ON(port_isset(execlists->port)); @@ -868,6 +962,8 @@ static void process_csb(struct intel_engine_cs *engine) const u32 * const buf = execlists->csb_status; u8 head, tail; + lockdep_assert_held(&engine->timeline.lock); + /* * Note that csb_write, csb_status may be either in HWSP or mmio. * When reading from the csb_write mmio register, we have to be @@ -956,12 +1052,13 @@ static void process_csb(struct intel_engine_cs *engine) EXECLISTS_ACTIVE_USER)); rq = port_unpack(port, &count); - GEM_TRACE("%s out[0]: ctx=%d.%d, global=%d (fence %llx:%lld) (current %d), prio=%d\n", + GEM_TRACE("%s out[0]: ctx=%d.%d, global=%d (fence %llx:%lld) (current %d:%d), prio=%d\n", engine->name, port->context_id, count, rq ? rq->global_seqno : 0, rq ? rq->fence.context : 0, rq ? rq->fence.seqno : 0, + rq ? hwsp_seqno(rq) : 0, intel_engine_get_seqno(engine), rq ? rq_prio(rq) : 0); @@ -1045,7 +1142,7 @@ static void execlists_submission_tasklet(unsigned long data) GEM_TRACE("%s awake?=%d, active=%x\n", engine->name, - engine->i915->gt.awake, + !!engine->i915->gt.awake, engine->execlists.active); spin_lock_irqsave(&engine->timeline.lock, flags); @@ -1075,8 +1172,8 @@ static void __submit_queue_imm(struct intel_engine_cs *engine) static void submit_queue(struct intel_engine_cs *engine, int prio) { - if (prio > engine->execlists.queue_priority) { - engine->execlists.queue_priority = prio; + if (prio > engine->execlists.queue_priority_hint) { + engine->execlists.queue_priority_hint = prio; __submit_queue_imm(engine); } } @@ -1169,6 +1266,23 @@ static int __context_pin(struct i915_gem_context *ctx, struct i915_vma *vma) return i915_vma_pin(vma, 0, 0, flags); } +static void +__execlists_update_reg_state(struct intel_engine_cs *engine, + struct intel_context *ce) +{ + u32 *regs = ce->lrc_reg_state; + struct intel_ring *ring = ce->ring; + + regs[CTX_RING_BUFFER_START + 1] = i915_ggtt_offset(ring->vma); + regs[CTX_RING_HEAD + 1] = ring->head; + regs[CTX_RING_TAIL + 1] = ring->tail; + + /* RPCS */ + if (engine->class == RENDER_CLASS) + regs[CTX_R_PWR_CLK_STATE + 1] = gen8_make_rpcs(engine->i915, + &ce->sseu); +} + static struct intel_context * __execlists_context_pin(struct intel_engine_cs *engine, struct i915_gem_context *ctx, @@ -1207,10 +1321,8 @@ __execlists_context_pin(struct intel_engine_cs *engine, GEM_BUG_ON(!intel_ring_offset_valid(ce->ring, ce->ring->head)); ce->lrc_reg_state = vaddr + LRC_STATE_PN * PAGE_SIZE; - ce->lrc_reg_state[CTX_RING_BUFFER_START+1] = - i915_ggtt_offset(ce->ring->vma); - ce->lrc_reg_state[CTX_RING_HEAD + 1] = ce->ring->head; - ce->lrc_reg_state[CTX_RING_TAIL + 1] = ce->ring->tail; + + __execlists_update_reg_state(engine, ce); ce->state->obj->pin_global++; i915_gem_context_get(ctx); @@ -1250,6 +1362,34 @@ execlists_context_pin(struct intel_engine_cs *engine, return __execlists_context_pin(engine, ctx, ce); } +static int gen8_emit_init_breadcrumb(struct i915_request *rq) +{ + u32 *cs; + + GEM_BUG_ON(!rq->timeline->has_initial_breadcrumb); + + cs = intel_ring_begin(rq, 6); + if (IS_ERR(cs)) + return PTR_ERR(cs); + + /* + * Check if we have been preempted before we even get started. + * + * After this point i915_request_started() reports true, even if + * we get preempted and so are no longer running. + */ + *cs++ = MI_ARB_CHECK; + *cs++ = MI_NOOP; + + *cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT; + *cs++ = rq->timeline->hwsp_offset; + *cs++ = 0; + *cs++ = rq->fence.seqno - 1; + + intel_ring_advance(rq, cs); + return 0; +} + static int emit_pdps(struct i915_request *rq) { const struct intel_engine_cs * const engine = rq->engine; @@ -1675,7 +1815,7 @@ static void enable_execlists(struct intel_engine_cs *engine) _MASKED_BIT_DISABLE(STOP_RING)); I915_WRITE(RING_HWS_PGA(engine->mmio_base), - engine->status_page.ggtt_offset); + i915_ggtt_offset(engine->status_page.vma)); POSTING_READ(RING_HWS_PGA(engine->mmio_base)); } @@ -1712,11 +1852,9 @@ static int gen8_init_common_ring(struct intel_engine_cs *engine) return 0; } -static struct i915_request * -execlists_reset_prepare(struct intel_engine_cs *engine) +static void execlists_reset_prepare(struct intel_engine_cs *engine) { struct intel_engine_execlists * const execlists = &engine->execlists; - struct i915_request *request, *active; unsigned long flags; GEM_TRACE("%s: depth<-%d\n", engine->name, @@ -1732,59 +1870,21 @@ execlists_reset_prepare(struct intel_engine_cs *engine) * prevents the race. */ __tasklet_disable_sync_once(&execlists->tasklet); + GEM_BUG_ON(!reset_in_progress(execlists)); + /* And flush any current direct submission. */ spin_lock_irqsave(&engine->timeline.lock, flags); - - /* - * We want to flush the pending context switches, having disabled - * the tasklet above, we can assume exclusive access to the execlists. - * For this allows us to catch up with an inflight preemption event, - * and avoid blaming an innocent request if the stall was due to the - * preemption itself. - */ - process_csb(engine); - - /* - * The last active request can then be no later than the last request - * now in ELSP[0]. So search backwards from there, so that if the GPU - * has advanced beyond the last CSB update, it will be pardoned. - */ - active = NULL; - request = port_request(execlists->port); - if (request) { - /* - * Prevent the breadcrumb from advancing before we decide - * which request is currently active. - */ - intel_engine_stop_cs(engine); - - list_for_each_entry_from_reverse(request, - &engine->timeline.requests, - link) { - if (__i915_request_completed(request, - request->global_seqno)) - break; - - active = request; - } - } - + process_csb(engine); /* drain preemption events */ spin_unlock_irqrestore(&engine->timeline.lock, flags); - - return active; } -static void execlists_reset(struct intel_engine_cs *engine, - struct i915_request *request) +static void execlists_reset(struct intel_engine_cs *engine, bool stalled) { struct intel_engine_execlists * const execlists = &engine->execlists; + struct i915_request *rq; unsigned long flags; u32 *regs; - GEM_TRACE("%s request global=%d, current=%d\n", - engine->name, request ? request->global_seqno : 0, - intel_engine_get_seqno(engine)); - spin_lock_irqsave(&engine->timeline.lock, flags); /* @@ -1799,12 +1899,18 @@ static void execlists_reset(struct intel_engine_cs *engine, execlists_cancel_port_requests(execlists); /* Push back any incomplete requests for replay after the reset. */ - __unwind_incomplete_requests(engine); + rq = __unwind_incomplete_requests(engine); /* Following the reset, we need to reload the CSB read/write pointers */ reset_csb_pointers(&engine->execlists); - spin_unlock_irqrestore(&engine->timeline.lock, flags); + GEM_TRACE("%s seqno=%d, current=%d, stalled? %s\n", + engine->name, + rq ? rq->global_seqno : 0, + intel_engine_get_seqno(engine), + yesno(stalled)); + if (!rq) + goto out_unlock; /* * If the request was innocent, we leave the request in the ELSP @@ -1817,8 +1923,9 @@ static void execlists_reset(struct intel_engine_cs *engine, * and have to at least restore the RING register in the context * image back to the expected values to skip over the guilty request. */ - if (!request || request->fence.error != -EIO) - return; + i915_reset_request(rq, stalled); + if (!stalled) + goto out_unlock; /* * We want a simple context + ring to execute the breadcrumb update. @@ -1828,25 +1935,22 @@ static void execlists_reset(struct intel_engine_cs *engine, * future request will be after userspace has had the opportunity * to recreate its own state. */ - regs = request->hw_context->lrc_reg_state; + regs = rq->hw_context->lrc_reg_state; if (engine->pinned_default_state) { memcpy(regs, /* skip restoring the vanilla PPHWSP */ engine->pinned_default_state + LRC_STATE_PN * PAGE_SIZE, engine->context_size - PAGE_SIZE); } - execlists_init_reg_state(regs, - request->gem_context, engine, request->ring); /* Move the RING_HEAD onto the breadcrumb, past the hanging batch */ - regs[CTX_RING_BUFFER_START + 1] = i915_ggtt_offset(request->ring->vma); - - request->ring->head = intel_ring_wrap(request->ring, request->postfix); - regs[CTX_RING_HEAD + 1] = request->ring->head; + rq->ring->head = intel_ring_wrap(rq->ring, rq->postfix); + intel_ring_update_space(rq->ring); - intel_ring_update_space(request->ring); + execlists_init_reg_state(regs, rq->gem_context, engine, rq->ring); + __execlists_update_reg_state(engine, rq->hw_context); - /* Reset WaIdleLiteRestore:bdw,skl as well */ - unwind_wa_tail(request); +out_unlock: + spin_unlock_irqrestore(&engine->timeline.lock, flags); } static void execlists_reset_finish(struct intel_engine_cs *engine) @@ -1859,6 +1963,7 @@ static void execlists_reset_finish(struct intel_engine_cs *engine) * to sleep before we restart and reload a context. * */ + GEM_BUG_ON(!reset_in_progress(execlists)); if (!RB_EMPTY_ROOT(&execlists->queue.rb_root)) execlists->tasklet.func(execlists->tasklet.data); @@ -2031,53 +2136,62 @@ static int gen8_emit_flush_render(struct i915_request *request, * used as a workaround for not being allowed to do lite * restore with HEAD==TAIL (WaIdleLiteRestore). */ -static void gen8_emit_wa_tail(struct i915_request *request, u32 *cs) +static u32 *gen8_emit_wa_tail(struct i915_request *request, u32 *cs) { /* Ensure there's always at least one preemption point per-request. */ *cs++ = MI_ARB_CHECK; *cs++ = MI_NOOP; request->wa_tail = intel_ring_offset(request, cs); + + return cs; } -static void gen8_emit_breadcrumb(struct i915_request *request, u32 *cs) +static u32 *gen8_emit_fini_breadcrumb(struct i915_request *request, u32 *cs) { /* w/a: bit 5 needs to be zero for MI_FLUSH_DW address. */ BUILD_BUG_ON(I915_GEM_HWS_INDEX_ADDR & (1 << 5)); - cs = gen8_emit_ggtt_write(cs, request->global_seqno, + cs = gen8_emit_ggtt_write(cs, + request->fence.seqno, + request->timeline->hwsp_offset); + + cs = gen8_emit_ggtt_write(cs, + request->global_seqno, intel_hws_seqno_address(request->engine)); + *cs++ = MI_USER_INTERRUPT; *cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE; + request->tail = intel_ring_offset(request, cs); assert_ring_tail_valid(request->ring, request->tail); - gen8_emit_wa_tail(request, cs); + return gen8_emit_wa_tail(request, cs); } -static const int gen8_emit_breadcrumb_sz = 6 + WA_TAIL_DWORDS; -static void gen8_emit_breadcrumb_rcs(struct i915_request *request, u32 *cs) +static u32 *gen8_emit_fini_breadcrumb_rcs(struct i915_request *request, u32 *cs) { - /* We're using qword write, seqno should be aligned to 8 bytes. */ - BUILD_BUG_ON(I915_GEM_HWS_INDEX & 1); - cs = gen8_emit_ggtt_write_rcs(cs, - request->global_seqno, - intel_hws_seqno_address(request->engine), + request->fence.seqno, + request->timeline->hwsp_offset, PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH | PIPE_CONTROL_DEPTH_CACHE_FLUSH | PIPE_CONTROL_DC_FLUSH_ENABLE | PIPE_CONTROL_FLUSH_ENABLE | PIPE_CONTROL_CS_STALL); + cs = gen8_emit_ggtt_write_rcs(cs, + request->global_seqno, + intel_hws_seqno_address(request->engine), + PIPE_CONTROL_CS_STALL); + *cs++ = MI_USER_INTERRUPT; *cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE; request->tail = intel_ring_offset(request, cs); assert_ring_tail_valid(request->ring, request->tail); - gen8_emit_wa_tail(request, cs); + return gen8_emit_wa_tail(request, cs); } -static const int gen8_emit_breadcrumb_rcs_sz = 8 + WA_TAIL_DWORDS; static int gen8_init_rcs_context(struct i915_request *rq) { @@ -2169,8 +2283,8 @@ logical_ring_default_vfuncs(struct intel_engine_cs *engine) engine->request_alloc = execlists_request_alloc; engine->emit_flush = gen8_emit_flush; - engine->emit_breadcrumb = gen8_emit_breadcrumb; - engine->emit_breadcrumb_sz = gen8_emit_breadcrumb_sz; + engine->emit_init_breadcrumb = gen8_emit_init_breadcrumb; + engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb; engine->set_default_submission = intel_execlists_set_default_submission; @@ -2209,10 +2323,14 @@ logical_ring_default_irqs(struct intel_engine_cs *engine) engine->irq_keep_mask = GT_CONTEXT_SWITCH_INTERRUPT << shift; } -static void +static int logical_ring_setup(struct intel_engine_cs *engine) { - intel_engine_setup_common(engine); + int err; + + err = intel_engine_setup_common(engine); + if (err) + return err; /* Intentionally left blank. */ engine->buffer = NULL; @@ -2222,6 +2340,8 @@ logical_ring_setup(struct intel_engine_cs *engine) logical_ring_default_vfuncs(engine); logical_ring_default_irqs(engine); + + return 0; } static int logical_ring_init(struct intel_engine_cs *engine) @@ -2256,10 +2376,10 @@ static int logical_ring_init(struct intel_engine_cs *engine) } execlists->csb_status = - &engine->status_page.page_addr[I915_HWS_CSB_BUF0_INDEX]; + &engine->status_page.addr[I915_HWS_CSB_BUF0_INDEX]; execlists->csb_write = - &engine->status_page.page_addr[intel_hws_csb_write_index(i915)]; + &engine->status_page.addr[intel_hws_csb_write_index(i915)]; reset_csb_pointers(execlists); @@ -2270,13 +2390,14 @@ int logical_render_ring_init(struct intel_engine_cs *engine) { int ret; - logical_ring_setup(engine); + ret = logical_ring_setup(engine); + if (ret) + return ret; /* Override some for render ring. */ engine->init_context = gen8_init_rcs_context; engine->emit_flush = gen8_emit_flush_render; - engine->emit_breadcrumb = gen8_emit_breadcrumb_rcs; - engine->emit_breadcrumb_sz = gen8_emit_breadcrumb_rcs_sz; + engine->emit_fini_breadcrumb = gen8_emit_fini_breadcrumb_rcs; ret = logical_ring_init(engine); if (ret) @@ -2300,27 +2421,59 @@ int logical_render_ring_init(struct intel_engine_cs *engine) int logical_xcs_ring_init(struct intel_engine_cs *engine) { - logical_ring_setup(engine); + int err; + + err = logical_ring_setup(engine); + if (err) + return err; return logical_ring_init(engine); } -static u32 -make_rpcs(struct drm_i915_private *dev_priv) +u32 gen8_make_rpcs(struct drm_i915_private *i915, struct intel_sseu *req_sseu) { - bool subslice_pg = RUNTIME_INFO(dev_priv)->sseu.has_subslice_pg; - u8 slices = hweight8(RUNTIME_INFO(dev_priv)->sseu.slice_mask); - u8 subslices = hweight8(RUNTIME_INFO(dev_priv)->sseu.subslice_mask[0]); + const struct sseu_dev_info *sseu = &RUNTIME_INFO(i915)->sseu; + bool subslice_pg = sseu->has_subslice_pg; + struct intel_sseu ctx_sseu; + u8 slices, subslices; u32 rpcs = 0; /* * No explicit RPCS request is needed to ensure full * slice/subslice/EU enablement prior to Gen9. */ - if (INTEL_GEN(dev_priv) < 9) + if (INTEL_GEN(i915) < 9) return 0; /* + * If i915/perf is active, we want a stable powergating configuration + * on the system. + * + * We could choose full enablement, but on ICL we know there are use + * cases which disable slices for functional, apart for performance + * reasons. So in this case we select a known stable subset. + */ + if (!i915->perf.oa.exclusive_stream) { + ctx_sseu = *req_sseu; + } else { + ctx_sseu = intel_device_default_sseu(i915); + + if (IS_GEN(i915, 11)) { + /* + * We only need subslice count so it doesn't matter + * which ones we select - just turn off low bits in the + * amount of half of all available subslices per slice. + */ + ctx_sseu.subslice_mask = + ~(~0 << (hweight8(ctx_sseu.subslice_mask) / 2)); + ctx_sseu.slice_mask = 0x1; + } + } + + slices = hweight8(ctx_sseu.slice_mask); + subslices = hweight8(ctx_sseu.subslice_mask); + + /* * Since the SScount bitfield in GEN8_R_PWR_CLK_STATE is only three bits * wide and Icelake has up to eight subslices, specfial programming is * needed in order to correctly enable all subslices. @@ -2345,7 +2498,9 @@ make_rpcs(struct drm_i915_private *dev_priv) * subslices are enabled, or a count between one and four on the first * slice. */ - if (IS_GEN(dev_priv, 11) && slices == 1 && subslices >= 4) { + if (IS_GEN(i915, 11) && + slices == 1 && + subslices > min_t(u8, 4, hweight8(sseu->subslice_mask[0]) / 2)) { GEM_BUG_ON(subslices & 1); subslice_pg = false; @@ -2358,10 +2513,10 @@ make_rpcs(struct drm_i915_private *dev_priv) * must make an explicit request through RPCS for full * enablement. */ - if (RUNTIME_INFO(dev_priv)->sseu.has_slice_pg) { + if (sseu->has_slice_pg) { u32 mask, val = slices; - if (INTEL_GEN(dev_priv) >= 11) { + if (INTEL_GEN(i915) >= 11) { mask = GEN11_RPCS_S_CNT_MASK; val <<= GEN11_RPCS_S_CNT_SHIFT; } else { @@ -2386,18 +2541,16 @@ make_rpcs(struct drm_i915_private *dev_priv) rpcs |= GEN8_RPCS_ENABLE | GEN8_RPCS_SS_CNT_ENABLE | val; } - if (RUNTIME_INFO(dev_priv)->sseu.has_eu_pg) { + if (sseu->has_eu_pg) { u32 val; - val = RUNTIME_INFO(dev_priv)->sseu.eu_per_subslice << - GEN8_RPCS_EU_MIN_SHIFT; + val = ctx_sseu.min_eus_per_subslice << GEN8_RPCS_EU_MIN_SHIFT; GEM_BUG_ON(val & ~GEN8_RPCS_EU_MIN_MASK); val &= GEN8_RPCS_EU_MIN_MASK; rpcs |= val; - val = RUNTIME_INFO(dev_priv)->sseu.eu_per_subslice << - GEN8_RPCS_EU_MAX_SHIFT; + val = ctx_sseu.max_eus_per_subslice << GEN8_RPCS_EU_MAX_SHIFT; GEM_BUG_ON(val & ~GEN8_RPCS_EU_MAX_MASK); val &= GEN8_RPCS_EU_MAX_MASK; @@ -2530,8 +2683,7 @@ static void execlists_init_reg_state(u32 *regs, if (rcs) { regs[CTX_LRI_HEADER_2] = MI_LOAD_REGISTER_IMM(1); - CTX_REG(regs, CTX_R_PWR_CLK_STATE, GEN8_R_PWR_CLK_STATE, - make_rpcs(dev_priv)); + CTX_REG(regs, CTX_R_PWR_CLK_STATE, GEN8_R_PWR_CLK_STATE, 0); i915_oa_init_reg_state(engine, ctx, regs); } @@ -2608,7 +2760,7 @@ static int execlists_context_deferred_alloc(struct i915_gem_context *ctx, { struct drm_i915_gem_object *ctx_obj; struct i915_vma *vma; - uint32_t context_size; + u32 context_size; struct intel_ring *ring; struct i915_timeline *timeline; int ret; @@ -2634,7 +2786,7 @@ static int execlists_context_deferred_alloc(struct i915_gem_context *ctx, goto error_deref_obj; } - timeline = i915_timeline_create(ctx->i915, ctx->name); + timeline = i915_timeline_create(ctx->i915, ctx->name, NULL); if (IS_ERR(timeline)) { ret = PTR_ERR(timeline); goto error_deref_obj; @@ -2692,14 +2844,70 @@ void intel_lr_context_resume(struct drm_i915_private *i915) intel_ring_reset(ce->ring, 0); - if (ce->pin_count) { /* otherwise done in context_pin */ - u32 *regs = ce->lrc_reg_state; + if (ce->pin_count) /* otherwise done in context_pin */ + __execlists_update_reg_state(engine, ce); + } + } +} - regs[CTX_RING_HEAD + 1] = ce->ring->head; - regs[CTX_RING_TAIL + 1] = ce->ring->tail; - } +void intel_execlists_show_requests(struct intel_engine_cs *engine, + struct drm_printer *m, + void (*show_request)(struct drm_printer *m, + struct i915_request *rq, + const char *prefix), + unsigned int max) +{ + const struct intel_engine_execlists *execlists = &engine->execlists; + struct i915_request *rq, *last; + unsigned long flags; + unsigned int count; + struct rb_node *rb; + + spin_lock_irqsave(&engine->timeline.lock, flags); + + last = NULL; + count = 0; + list_for_each_entry(rq, &engine->timeline.requests, link) { + if (count++ < max - 1) + show_request(m, rq, "\t\tE "); + else + last = rq; + } + if (last) { + if (count > max) { + drm_printf(m, + "\t\t...skipping %d executing requests...\n", + count - max); + } + show_request(m, last, "\t\tE "); + } + + last = NULL; + count = 0; + if (execlists->queue_priority_hint != INT_MIN) + drm_printf(m, "\t\tQueue priority hint: %d\n", + execlists->queue_priority_hint); + for (rb = rb_first_cached(&execlists->queue); rb; rb = rb_next(rb)) { + struct i915_priolist *p = rb_entry(rb, typeof(*p), node); + int i; + + priolist_for_each_request(rq, p, i) { + if (count++ < max - 1) + show_request(m, rq, "\t\tQ "); + else + last = rq; } } + if (last) { + if (count > max) { + drm_printf(m, + "\t\t...skipping %d queued requests...\n", + count - max); + } + show_request(m, last, "\t\tQ "); + } + + spin_unlock_irqrestore(&engine->timeline.lock, flags); } #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h index f5a5502ecf70..f1aec8a6986f 100644 --- a/drivers/gpu/drm/i915/intel_lrc.h +++ b/drivers/gpu/drm/i915/intel_lrc.h @@ -97,11 +97,21 @@ int logical_xcs_ring_init(struct intel_engine_cs *engine); */ #define LRC_HEADER_PAGES LRC_PPHWSP_PN +struct drm_printer; + struct drm_i915_private; struct i915_gem_context; void intel_lr_context_resume(struct drm_i915_private *dev_priv); - void intel_execlists_set_default_submission(struct intel_engine_cs *engine); +void intel_execlists_show_requests(struct intel_engine_cs *engine, + struct drm_printer *m, + void (*show_request)(struct drm_printer *m, + struct i915_request *rq, + const char *prefix), + unsigned int max); + +u32 gen8_make_rpcs(struct drm_i915_private *i915, struct intel_sseu *ctx_sseu); + #endif /* _INTEL_LRC_H_ */ diff --git a/drivers/gpu/drm/i915/intel_lspcon.c b/drivers/gpu/drm/i915/intel_lspcon.c index 7d15be5932e0..322bdddda164 100644 --- a/drivers/gpu/drm/i915/intel_lspcon.c +++ b/drivers/gpu/drm/i915/intel_lspcon.c @@ -288,12 +288,12 @@ static bool lspcon_parade_fw_ready(struct drm_dp_aux *aux) } static bool _lspcon_parade_write_infoframe_blocks(struct drm_dp_aux *aux, - uint8_t *avi_buf) + u8 *avi_buf) { u8 avi_if_ctrl; u8 block_count = 0; u8 *data; - uint16_t reg; + u16 reg; ssize_t ret; while (block_count < 4) { @@ -335,10 +335,10 @@ static bool _lspcon_parade_write_infoframe_blocks(struct drm_dp_aux *aux, } static bool _lspcon_write_avi_infoframe_parade(struct drm_dp_aux *aux, - const uint8_t *frame, + const u8 *frame, ssize_t len) { - uint8_t avi_if[LSPCON_PARADE_AVI_IF_DATA_SIZE] = {1, }; + u8 avi_if[LSPCON_PARADE_AVI_IF_DATA_SIZE] = {1, }; /* * Parade's frames contains 32 bytes of data, divided @@ -367,13 +367,13 @@ static bool _lspcon_write_avi_infoframe_parade(struct drm_dp_aux *aux, } static bool _lspcon_write_avi_infoframe_mca(struct drm_dp_aux *aux, - const uint8_t *buffer, ssize_t len) + const u8 *buffer, ssize_t len) { int ret; - uint32_t val = 0; - uint32_t retry; - uint16_t reg; - const uint8_t *data = buffer; + u32 val = 0; + u32 retry; + u16 reg; + const u8 *data = buffer; reg = LSPCON_MCA_AVI_IF_WRITE_OFFSET; while (val < len) { @@ -459,7 +459,7 @@ void lspcon_set_infoframes(struct intel_encoder *encoder, { ssize_t ret; union hdmi_infoframe frame; - uint8_t buf[VIDEO_DIP_DATA_SIZE]; + u8 buf[VIDEO_DIP_DATA_SIZE]; struct intel_digital_port *dig_port = enc_to_dig_port(&encoder->base); struct intel_lspcon *lspcon = &dig_port->lspcon; const struct drm_display_mode *adjusted_mode = diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c index 904d16af89a8..b4aa49768e90 100644 --- a/drivers/gpu/drm/i915/intel_lvds.c +++ b/drivers/gpu/drm/i915/intel_lvds.c @@ -94,15 +94,17 @@ static bool intel_lvds_get_hw_state(struct intel_encoder *encoder, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base); + intel_wakeref_t wakeref; bool ret; - if (!intel_display_power_get_if_enabled(dev_priv, - encoder->power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + encoder->power_domain); + if (!wakeref) return false; ret = intel_lvds_port_enabled(dev_priv, lvds_encoder->reg, pipe); - intel_display_power_put(dev_priv, encoder->power_domain); + intel_display_power_put(dev_priv, encoder->power_domain, wakeref); return ret; } @@ -796,26 +798,6 @@ static bool compute_is_dual_link_lvds(struct intel_lvds_encoder *lvds_encoder) return (val & LVDS_CLKB_POWER_MASK) == LVDS_CLKB_POWER_UP; } -static bool intel_lvds_supported(struct drm_i915_private *dev_priv) -{ - /* - * With the introduction of the PCH we gained a dedicated - * LVDS presence pin, use it. - */ - if (HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv)) - return true; - - /* - * Otherwise LVDS was only attached to mobile products, - * except for the inglorious 830gm - */ - if (INTEL_GEN(dev_priv) <= 4 && - IS_MOBILE(dev_priv) && !IS_I830(dev_priv)) - return true; - - return false; -} - /** * intel_lvds_init - setup LVDS connectors on this device * @dev_priv: i915 device @@ -840,9 +822,6 @@ void intel_lvds_init(struct drm_i915_private *dev_priv) u8 pin; u32 allowed_scalers; - if (!intel_lvds_supported(dev_priv)) - return; - /* Skip init on machines we know falsely report LVDS */ if (dmi_check_system(intel_no_lvds)) { WARN(!dev_priv->vbt.int_lvds_support, @@ -908,6 +887,7 @@ void intel_lvds_init(struct drm_i915_private *dev_priv) } intel_encoder->get_hw_state = intel_lvds_get_hw_state; intel_encoder->get_config = intel_lvds_get_config; + intel_encoder->update_pipe = intel_panel_update_backlight; intel_connector->get_hw_state = intel_connector_get_hw_state; intel_connector_attach_encoder(intel_connector, intel_encoder); diff --git a/drivers/gpu/drm/i915/intel_mocs.c b/drivers/gpu/drm/i915/intel_mocs.c index e976c5ce5479..331e7a678fb7 100644 --- a/drivers/gpu/drm/i915/intel_mocs.c +++ b/drivers/gpu/drm/i915/intel_mocs.c @@ -28,48 +28,60 @@ struct drm_i915_mocs_entry { u32 control_value; u16 l3cc_value; + u16 used; }; struct drm_i915_mocs_table { - u32 size; + unsigned int size; + unsigned int n_entries; const struct drm_i915_mocs_entry *table; }; /* Defines for the tables (XXX_MOCS_0 - XXX_MOCS_63) */ -#define LE_CACHEABILITY(value) ((value) << 0) -#define LE_TGT_CACHE(value) ((value) << 2) +#define _LE_CACHEABILITY(value) ((value) << 0) +#define _LE_TGT_CACHE(value) ((value) << 2) #define LE_LRUM(value) ((value) << 4) #define LE_AOM(value) ((value) << 6) #define LE_RSC(value) ((value) << 7) #define LE_SCC(value) ((value) << 8) #define LE_PFM(value) ((value) << 11) #define LE_SCF(value) ((value) << 14) +#define LE_COS(value) ((value) << 15) +#define LE_SSE(value) ((value) << 17) /* Defines for the tables (LNCFMOCS0 - LNCFMOCS31) - two entries per word */ #define L3_ESC(value) ((value) << 0) #define L3_SCC(value) ((value) << 1) -#define L3_CACHEABILITY(value) ((value) << 4) +#define _L3_CACHEABILITY(value) ((value) << 4) /* Helper defines */ #define GEN9_NUM_MOCS_ENTRIES 62 /* 62 out of 64 - 63 & 64 are reserved. */ +#define GEN11_NUM_MOCS_ENTRIES 64 /* 63-64 are reserved, but configured. */ /* (e)LLC caching options */ -#define LE_PAGETABLE 0 -#define LE_UC 1 -#define LE_WT 2 -#define LE_WB 3 - -/* L3 caching options */ -#define L3_DIRECT 0 -#define L3_UC 1 -#define L3_RESERVED 2 -#define L3_WB 3 +#define LE_0_PAGETABLE _LE_CACHEABILITY(0) +#define LE_1_UC _LE_CACHEABILITY(1) +#define LE_2_WT _LE_CACHEABILITY(2) +#define LE_3_WB _LE_CACHEABILITY(3) /* Target cache */ -#define LE_TC_PAGETABLE 0 -#define LE_TC_LLC 1 -#define LE_TC_LLC_ELLC 2 -#define LE_TC_LLC_ELLC_ALT 3 +#define LE_TC_0_PAGETABLE _LE_TGT_CACHE(0) +#define LE_TC_1_LLC _LE_TGT_CACHE(1) +#define LE_TC_2_LLC_ELLC _LE_TGT_CACHE(2) +#define LE_TC_3_LLC_ELLC_ALT _LE_TGT_CACHE(3) + +/* L3 caching options */ +#define L3_0_DIRECT _L3_CACHEABILITY(0) +#define L3_1_UC _L3_CACHEABILITY(1) +#define L3_2_RESERVED _L3_CACHEABILITY(2) +#define L3_3_WB _L3_CACHEABILITY(3) + +#define MOCS_ENTRY(__idx, __control_value, __l3cc_value) \ + [__idx] = { \ + .control_value = __control_value, \ + .l3cc_value = __l3cc_value, \ + .used = 1, \ + } /* * MOCS tables @@ -80,85 +92,147 @@ struct drm_i915_mocs_table { * LNCFCMOCS0 - LNCFCMOCS32 registers. * * These tables are intended to be kept reasonably consistent across - * platforms. However some of the fields are not applicable to all of - * them. + * HW platforms, and for ICL+, be identical across OSes. To achieve + * that, for Icelake and above, list of entries is published as part + * of bspec. * * Entries not part of the following tables are undefined as far as * userspace is concerned and shouldn't be relied upon. For the time - * being they will be implicitly initialized to the strictest caching - * configuration (uncached) to guarantee forwards compatibility with - * userspace programs written against more recent kernels providing - * additional MOCS entries. + * being they will be initialized to PTE. * - * NOTE: These tables MUST start with being uncached and the length - * MUST be less than 63 as the last two registers are reserved - * by the hardware. These tables are part of the kernel ABI and - * may only be updated incrementally by adding entries at the - * end. + * The last two entries are reserved by the hardware. For ICL+ they + * should be initialized according to bspec and never used, for older + * platforms they should never be written to. + * + * NOTE: These tables are part of bspec and defined as part of hardware + * interface for ICL+. For older platforms, they are part of kernel + * ABI. It is expected that, for specific hardware platform, existing + * entries will remain constant and the table will only be updated by + * adding new entries, filling unused positions. */ +#define GEN9_MOCS_ENTRIES \ + MOCS_ENTRY(I915_MOCS_UNCACHED, \ + LE_1_UC | LE_TC_2_LLC_ELLC, \ + L3_1_UC), \ + MOCS_ENTRY(I915_MOCS_PTE, \ + LE_0_PAGETABLE | LE_TC_2_LLC_ELLC | LE_LRUM(3), \ + L3_3_WB) + static const struct drm_i915_mocs_entry skylake_mocs_table[] = { - [I915_MOCS_UNCACHED] = { - /* 0x00000009 */ - .control_value = LE_CACHEABILITY(LE_UC) | - LE_TGT_CACHE(LE_TC_LLC_ELLC) | - LE_LRUM(0) | LE_AOM(0) | LE_RSC(0) | LE_SCC(0) | - LE_PFM(0) | LE_SCF(0), - - /* 0x0010 */ - .l3cc_value = L3_ESC(0) | L3_SCC(0) | L3_CACHEABILITY(L3_UC), - }, - [I915_MOCS_PTE] = { - /* 0x00000038 */ - .control_value = LE_CACHEABILITY(LE_PAGETABLE) | - LE_TGT_CACHE(LE_TC_LLC_ELLC) | - LE_LRUM(3) | LE_AOM(0) | LE_RSC(0) | LE_SCC(0) | - LE_PFM(0) | LE_SCF(0), - /* 0x0030 */ - .l3cc_value = L3_ESC(0) | L3_SCC(0) | L3_CACHEABILITY(L3_WB), - }, - [I915_MOCS_CACHED] = { - /* 0x0000003b */ - .control_value = LE_CACHEABILITY(LE_WB) | - LE_TGT_CACHE(LE_TC_LLC_ELLC) | - LE_LRUM(3) | LE_AOM(0) | LE_RSC(0) | LE_SCC(0) | - LE_PFM(0) | LE_SCF(0), - /* 0x0030 */ - .l3cc_value = L3_ESC(0) | L3_SCC(0) | L3_CACHEABILITY(L3_WB), - }, + GEN9_MOCS_ENTRIES, + MOCS_ENTRY(I915_MOCS_CACHED, + LE_3_WB | LE_TC_2_LLC_ELLC | LE_LRUM(3), + L3_3_WB) }; /* NOTE: the LE_TGT_CACHE is not used on Broxton */ static const struct drm_i915_mocs_entry broxton_mocs_table[] = { - [I915_MOCS_UNCACHED] = { - /* 0x00000009 */ - .control_value = LE_CACHEABILITY(LE_UC) | - LE_TGT_CACHE(LE_TC_LLC_ELLC) | - LE_LRUM(0) | LE_AOM(0) | LE_RSC(0) | LE_SCC(0) | - LE_PFM(0) | LE_SCF(0), - - /* 0x0010 */ - .l3cc_value = L3_ESC(0) | L3_SCC(0) | L3_CACHEABILITY(L3_UC), - }, - [I915_MOCS_PTE] = { - /* 0x00000038 */ - .control_value = LE_CACHEABILITY(LE_PAGETABLE) | - LE_TGT_CACHE(LE_TC_LLC_ELLC) | - LE_LRUM(3) | LE_AOM(0) | LE_RSC(0) | LE_SCC(0) | - LE_PFM(0) | LE_SCF(0), - - /* 0x0030 */ - .l3cc_value = L3_ESC(0) | L3_SCC(0) | L3_CACHEABILITY(L3_WB), - }, - [I915_MOCS_CACHED] = { - /* 0x00000039 */ - .control_value = LE_CACHEABILITY(LE_UC) | - LE_TGT_CACHE(LE_TC_LLC_ELLC) | - LE_LRUM(3) | LE_AOM(0) | LE_RSC(0) | LE_SCC(0) | - LE_PFM(0) | LE_SCF(0), - - /* 0x0030 */ - .l3cc_value = L3_ESC(0) | L3_SCC(0) | L3_CACHEABILITY(L3_WB), - }, + GEN9_MOCS_ENTRIES, + MOCS_ENTRY(I915_MOCS_CACHED, + LE_1_UC | LE_TC_2_LLC_ELLC | LE_LRUM(3), + L3_3_WB) +}; + +#define GEN11_MOCS_ENTRIES \ + /* Base - Uncached (Deprecated) */ \ + MOCS_ENTRY(I915_MOCS_UNCACHED, \ + LE_1_UC | LE_TC_1_LLC, \ + L3_1_UC), \ + /* Base - L3 + LeCC:PAT (Deprecated) */ \ + MOCS_ENTRY(I915_MOCS_PTE, \ + LE_0_PAGETABLE | LE_TC_1_LLC, \ + L3_3_WB), \ + /* Base - L3 + LLC */ \ + MOCS_ENTRY(2, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3), \ + L3_3_WB), \ + /* Base - Uncached */ \ + MOCS_ENTRY(3, \ + LE_1_UC | LE_TC_1_LLC, \ + L3_1_UC), \ + /* Base - L3 */ \ + MOCS_ENTRY(4, \ + LE_1_UC | LE_TC_1_LLC, \ + L3_3_WB), \ + /* Base - LLC */ \ + MOCS_ENTRY(5, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3), \ + L3_1_UC), \ + /* Age 0 - LLC */ \ + MOCS_ENTRY(6, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(1), \ + L3_1_UC), \ + /* Age 0 - L3 + LLC */ \ + MOCS_ENTRY(7, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(1), \ + L3_3_WB), \ + /* Age: Don't Chg. - LLC */ \ + MOCS_ENTRY(8, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(2), \ + L3_1_UC), \ + /* Age: Don't Chg. - L3 + LLC */ \ + MOCS_ENTRY(9, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(2), \ + L3_3_WB), \ + /* No AOM - LLC */ \ + MOCS_ENTRY(10, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_AOM(1), \ + L3_1_UC), \ + /* No AOM - L3 + LLC */ \ + MOCS_ENTRY(11, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_AOM(1), \ + L3_3_WB), \ + /* No AOM; Age 0 - LLC */ \ + MOCS_ENTRY(12, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(1) | LE_AOM(1), \ + L3_1_UC), \ + /* No AOM; Age 0 - L3 + LLC */ \ + MOCS_ENTRY(13, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(1) | LE_AOM(1), \ + L3_3_WB), \ + /* No AOM; Age:DC - LLC */ \ + MOCS_ENTRY(14, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(2) | LE_AOM(1), \ + L3_1_UC), \ + /* No AOM; Age:DC - L3 + LLC */ \ + MOCS_ENTRY(15, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(2) | LE_AOM(1), \ + L3_3_WB), \ + /* Self-Snoop - L3 + LLC */ \ + MOCS_ENTRY(18, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_SSE(3), \ + L3_3_WB), \ + /* Skip Caching - L3 + LLC(12.5%) */ \ + MOCS_ENTRY(19, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_SCC(7), \ + L3_3_WB), \ + /* Skip Caching - L3 + LLC(25%) */ \ + MOCS_ENTRY(20, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_SCC(3), \ + L3_3_WB), \ + /* Skip Caching - L3 + LLC(50%) */ \ + MOCS_ENTRY(21, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_SCC(1), \ + L3_3_WB), \ + /* Skip Caching - L3 + LLC(75%) */ \ + MOCS_ENTRY(22, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_RSC(1) | LE_SCC(3), \ + L3_3_WB), \ + /* Skip Caching - L3 + LLC(87.5%) */ \ + MOCS_ENTRY(23, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3) | LE_RSC(1) | LE_SCC(7), \ + L3_3_WB), \ + /* HW Reserved - SW program but never use */ \ + MOCS_ENTRY(62, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3), \ + L3_1_UC), \ + /* HW Reserved - SW program but never use */ \ + MOCS_ENTRY(63, \ + LE_3_WB | LE_TC_1_LLC | LE_LRUM(3), \ + L3_1_UC) + +static const struct drm_i915_mocs_entry icelake_mocs_table[] = { + GEN11_MOCS_ENTRIES }; /** @@ -178,13 +252,19 @@ static bool get_mocs_settings(struct drm_i915_private *dev_priv, { bool result = false; - if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv) || - IS_ICELAKE(dev_priv)) { + if (IS_ICELAKE(dev_priv)) { + table->size = ARRAY_SIZE(icelake_mocs_table); + table->table = icelake_mocs_table; + table->n_entries = GEN11_NUM_MOCS_ENTRIES; + result = true; + } else if (IS_GEN9_BC(dev_priv) || IS_CANNONLAKE(dev_priv)) { table->size = ARRAY_SIZE(skylake_mocs_table); + table->n_entries = GEN9_NUM_MOCS_ENTRIES; table->table = skylake_mocs_table; result = true; } else if (IS_GEN9_LP(dev_priv)) { table->size = ARRAY_SIZE(broxton_mocs_table); + table->n_entries = GEN9_NUM_MOCS_ENTRIES; table->table = broxton_mocs_table; result = true; } else { @@ -226,6 +306,19 @@ static i915_reg_t mocs_register(enum intel_engine_id engine_id, int index) } } +/* + * Get control_value from MOCS entry taking into account when it's not used: + * I915_MOCS_PTE's value is returned in this case. + */ +static u32 get_entry_control(const struct drm_i915_mocs_table *table, + unsigned int index) +{ + if (table->table[index].used) + return table->table[index].control_value; + + return table->table[I915_MOCS_PTE].control_value; +} + /** * intel_mocs_init_engine() - emit the mocs control table * @engine: The engine for whom to emit the registers. @@ -238,27 +331,23 @@ void intel_mocs_init_engine(struct intel_engine_cs *engine) struct drm_i915_private *dev_priv = engine->i915; struct drm_i915_mocs_table table; unsigned int index; + u32 unused_value; if (!get_mocs_settings(dev_priv, &table)) return; - GEM_BUG_ON(table.size > GEN9_NUM_MOCS_ENTRIES); - - for (index = 0; index < table.size; index++) - I915_WRITE(mocs_register(engine->id, index), - table.table[index].control_value); - - /* - * Ok, now set the unused entries to uncached. These entries - * are officially undefined and no contract for the contents - * and settings is given for these entries. - * - * Entry 0 in the table is uncached - so we are just writing - * that value to all the used entries. - */ - for (; index < GEN9_NUM_MOCS_ENTRIES; index++) - I915_WRITE(mocs_register(engine->id, index), - table.table[0].control_value); + /* Set unused values to PTE */ + unused_value = table.table[I915_MOCS_PTE].control_value; + + for (index = 0; index < table.size; index++) { + u32 value = get_entry_control(&table, index); + + I915_WRITE(mocs_register(engine->id, index), value); + } + + /* All remaining entries are also unused */ + for (; index < table.n_entries; index++) + I915_WRITE(mocs_register(engine->id, index), unused_value); } /** @@ -276,33 +365,32 @@ static int emit_mocs_control_table(struct i915_request *rq, { enum intel_engine_id engine = rq->engine->id; unsigned int index; + u32 unused_value; u32 *cs; - if (WARN_ON(table->size > GEN9_NUM_MOCS_ENTRIES)) + if (GEM_WARN_ON(table->size > table->n_entries)) return -ENODEV; - cs = intel_ring_begin(rq, 2 + 2 * GEN9_NUM_MOCS_ENTRIES); + /* Set unused values to PTE */ + unused_value = table->table[I915_MOCS_PTE].control_value; + + cs = intel_ring_begin(rq, 2 + 2 * table->n_entries); if (IS_ERR(cs)) return PTR_ERR(cs); - *cs++ = MI_LOAD_REGISTER_IMM(GEN9_NUM_MOCS_ENTRIES); + *cs++ = MI_LOAD_REGISTER_IMM(table->n_entries); for (index = 0; index < table->size; index++) { + u32 value = get_entry_control(table, index); + *cs++ = i915_mmio_reg_offset(mocs_register(engine, index)); - *cs++ = table->table[index].control_value; + *cs++ = value; } - /* - * Ok, now set the unused entries to uncached. These entries - * are officially undefined and no contract for the contents - * and settings is given for these entries. - * - * Entry 0 in the table is uncached - so we are just writing - * that value to all the used entries. - */ - for (; index < GEN9_NUM_MOCS_ENTRIES; index++) { + /* All remaining entries are also unused */ + for (; index < table->n_entries; index++) { *cs++ = i915_mmio_reg_offset(mocs_register(engine, index)); - *cs++ = table->table[0].control_value; + *cs++ = unused_value; } *cs++ = MI_NOOP; @@ -311,12 +399,24 @@ static int emit_mocs_control_table(struct i915_request *rq, return 0; } +/* + * Get l3cc_value from MOCS entry taking into account when it's not used: + * I915_MOCS_PTE's value is returned in this case. + */ +static u16 get_entry_l3cc(const struct drm_i915_mocs_table *table, + unsigned int index) +{ + if (table->table[index].used) + return table->table[index].l3cc_value; + + return table->table[I915_MOCS_PTE].l3cc_value; +} + static inline u32 l3cc_combine(const struct drm_i915_mocs_table *table, u16 low, u16 high) { - return table->table[low].l3cc_value | - table->table[high].l3cc_value << 16; + return low | high << 16; } /** @@ -333,38 +433,43 @@ static inline u32 l3cc_combine(const struct drm_i915_mocs_table *table, static int emit_mocs_l3cc_table(struct i915_request *rq, const struct drm_i915_mocs_table *table) { + u16 unused_value; unsigned int i; u32 *cs; - if (WARN_ON(table->size > GEN9_NUM_MOCS_ENTRIES)) + if (GEM_WARN_ON(table->size > table->n_entries)) return -ENODEV; - cs = intel_ring_begin(rq, 2 + GEN9_NUM_MOCS_ENTRIES); + /* Set unused values to PTE */ + unused_value = table->table[I915_MOCS_PTE].l3cc_value; + + cs = intel_ring_begin(rq, 2 + table->n_entries); if (IS_ERR(cs)) return PTR_ERR(cs); - *cs++ = MI_LOAD_REGISTER_IMM(GEN9_NUM_MOCS_ENTRIES / 2); + *cs++ = MI_LOAD_REGISTER_IMM(table->n_entries / 2); + + for (i = 0; i < table->size / 2; i++) { + u16 low = get_entry_l3cc(table, 2 * i); + u16 high = get_entry_l3cc(table, 2 * i + 1); - for (i = 0; i < table->size/2; i++) { *cs++ = i915_mmio_reg_offset(GEN9_LNCFCMOCS(i)); - *cs++ = l3cc_combine(table, 2 * i, 2 * i + 1); + *cs++ = l3cc_combine(table, low, high); } + /* Odd table size - 1 left over */ if (table->size & 0x01) { - /* Odd table size - 1 left over */ + u16 low = get_entry_l3cc(table, 2 * i); + *cs++ = i915_mmio_reg_offset(GEN9_LNCFCMOCS(i)); - *cs++ = l3cc_combine(table, 2 * i, 0); + *cs++ = l3cc_combine(table, low, unused_value); i++; } - /* - * Now set the rest of the table to uncached - use entry 0 as - * this will be uncached. Leave the last pair uninitialised as - * they are reserved by the hardware. - */ - for (; i < GEN9_NUM_MOCS_ENTRIES / 2; i++) { + /* All remaining entries are also unused */ + for (; i < table->n_entries / 2; i++) { *cs++ = i915_mmio_reg_offset(GEN9_LNCFCMOCS(i)); - *cs++ = l3cc_combine(table, 0, 0); + *cs++ = l3cc_combine(table, unused_value, unused_value); } *cs++ = MI_NOOP; @@ -391,26 +496,35 @@ void intel_mocs_init_l3cc_table(struct drm_i915_private *dev_priv) { struct drm_i915_mocs_table table; unsigned int i; + u16 unused_value; if (!get_mocs_settings(dev_priv, &table)) return; - for (i = 0; i < table.size/2; i++) - I915_WRITE(GEN9_LNCFCMOCS(i), l3cc_combine(&table, 2*i, 2*i+1)); + /* Set unused values to PTE */ + unused_value = table.table[I915_MOCS_PTE].l3cc_value; + + for (i = 0; i < table.size / 2; i++) { + u16 low = get_entry_l3cc(&table, 2 * i); + u16 high = get_entry_l3cc(&table, 2 * i + 1); + + I915_WRITE(GEN9_LNCFCMOCS(i), + l3cc_combine(&table, low, high)); + } /* Odd table size - 1 left over */ if (table.size & 0x01) { - I915_WRITE(GEN9_LNCFCMOCS(i), l3cc_combine(&table, 2*i, 0)); + u16 low = get_entry_l3cc(&table, 2 * i); + + I915_WRITE(GEN9_LNCFCMOCS(i), + l3cc_combine(&table, low, unused_value)); i++; } - /* - * Now set the rest of the table to uncached - use entry 0 as - * this will be uncached. Leave the last pair as initialised as - * they are reserved by the hardware. - */ - for (; i < (GEN9_NUM_MOCS_ENTRIES / 2); i++) - I915_WRITE(GEN9_LNCFCMOCS(i), l3cc_combine(&table, 0, 0)); + /* All remaining entries are also unused */ + for (; i < table.n_entries / 2; i++) + I915_WRITE(GEN9_LNCFCMOCS(i), + l3cc_combine(&table, unused_value, unused_value)); } /** diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c index c300e5787b3c..c0df1dbb0069 100644 --- a/drivers/gpu/drm/i915/intel_overlay.c +++ b/drivers/gpu/drm/i915/intel_overlay.c @@ -186,7 +186,7 @@ struct intel_overlay { struct overlay_registers __iomem *regs; u32 flip_addr; /* flip handling */ - struct i915_gem_active last_flip; + struct i915_active_request last_flip; }; static void i830_overlay_clock_gating(struct drm_i915_private *dev_priv, @@ -214,23 +214,23 @@ static void i830_overlay_clock_gating(struct drm_i915_private *dev_priv, static void intel_overlay_submit_request(struct intel_overlay *overlay, struct i915_request *rq, - i915_gem_retire_fn retire) + i915_active_retire_fn retire) { - GEM_BUG_ON(i915_gem_active_peek(&overlay->last_flip, - &overlay->i915->drm.struct_mutex)); - i915_gem_active_set_retire_fn(&overlay->last_flip, retire, - &overlay->i915->drm.struct_mutex); - i915_gem_active_set(&overlay->last_flip, rq); + GEM_BUG_ON(i915_active_request_peek(&overlay->last_flip, + &overlay->i915->drm.struct_mutex)); + i915_active_request_set_retire_fn(&overlay->last_flip, retire, + &overlay->i915->drm.struct_mutex); + __i915_active_request_set(&overlay->last_flip, rq); i915_request_add(rq); } static int intel_overlay_do_wait_request(struct intel_overlay *overlay, struct i915_request *rq, - i915_gem_retire_fn retire) + i915_active_retire_fn retire) { intel_overlay_submit_request(overlay, rq, retire); - return i915_gem_active_retire(&overlay->last_flip, - &overlay->i915->drm.struct_mutex); + return i915_active_request_retire(&overlay->last_flip, + &overlay->i915->drm.struct_mutex); } static struct i915_request *alloc_request(struct intel_overlay *overlay) @@ -351,8 +351,9 @@ static void intel_overlay_release_old_vma(struct intel_overlay *overlay) i915_vma_put(vma); } -static void intel_overlay_release_old_vid_tail(struct i915_gem_active *active, - struct i915_request *rq) +static void +intel_overlay_release_old_vid_tail(struct i915_active_request *active, + struct i915_request *rq) { struct intel_overlay *overlay = container_of(active, typeof(*overlay), last_flip); @@ -360,7 +361,7 @@ static void intel_overlay_release_old_vid_tail(struct i915_gem_active *active, intel_overlay_release_old_vma(overlay); } -static void intel_overlay_off_tail(struct i915_gem_active *active, +static void intel_overlay_off_tail(struct i915_active_request *active, struct i915_request *rq) { struct intel_overlay *overlay = @@ -423,8 +424,8 @@ static int intel_overlay_off(struct intel_overlay *overlay) * We have to be careful not to repeat work forever an make forward progess. */ static int intel_overlay_recover_from_interrupt(struct intel_overlay *overlay) { - return i915_gem_active_retire(&overlay->last_flip, - &overlay->i915->drm.struct_mutex); + return i915_active_request_retire(&overlay->last_flip, + &overlay->i915->drm.struct_mutex); } /* Wait for pending overlay flip and release old frame. @@ -480,8 +481,6 @@ void intel_overlay_reset(struct drm_i915_private *dev_priv) if (!overlay) return; - intel_overlay_release_old_vid(overlay); - overlay->old_xscale = 0; overlay->old_yscale = 0; overlay->crtc = NULL; @@ -1359,7 +1358,7 @@ void intel_overlay_setup(struct drm_i915_private *dev_priv) overlay->contrast = 75; overlay->saturation = 146; - init_request_active(&overlay->last_flip, NULL); + INIT_ACTIVE_REQUEST(&overlay->last_flip); mutex_lock(&dev_priv->drm.struct_mutex); diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c index ee3e0842d542..beca98d2b035 100644 --- a/drivers/gpu/drm/i915/intel_panel.c +++ b/drivers/gpu/drm/i915/intel_panel.c @@ -1087,20 +1087,11 @@ static void pwm_enable_backlight(const struct intel_crtc_state *crtc_state, intel_panel_actually_set_backlight(conn_state, panel->backlight.level); } -void intel_panel_enable_backlight(const struct intel_crtc_state *crtc_state, - const struct drm_connector_state *conn_state) +static void __intel_panel_enable_backlight(const struct intel_crtc_state *crtc_state, + const struct drm_connector_state *conn_state) { struct intel_connector *connector = to_intel_connector(conn_state->connector); - struct drm_i915_private *dev_priv = to_i915(connector->base.dev); struct intel_panel *panel = &connector->panel; - enum pipe pipe = to_intel_crtc(crtc_state->base.crtc)->pipe; - - if (!panel->backlight.present) - return; - - DRM_DEBUG_KMS("pipe %c\n", pipe_name(pipe)); - - mutex_lock(&dev_priv->backlight_lock); WARN_ON(panel->backlight.max == 0); @@ -1117,6 +1108,24 @@ void intel_panel_enable_backlight(const struct intel_crtc_state *crtc_state, panel->backlight.enabled = true; if (panel->backlight.device) panel->backlight.device->props.power = FB_BLANK_UNBLANK; +} + +void intel_panel_enable_backlight(const struct intel_crtc_state *crtc_state, + const struct drm_connector_state *conn_state) +{ + struct intel_connector *connector = to_intel_connector(conn_state->connector); + struct drm_i915_private *dev_priv = to_i915(connector->base.dev); + struct intel_panel *panel = &connector->panel; + enum pipe pipe = to_intel_crtc(crtc_state->base.crtc)->pipe; + + if (!panel->backlight.present) + return; + + DRM_DEBUG_KMS("pipe %c\n", pipe_name(pipe)); + + mutex_lock(&dev_priv->backlight_lock); + + __intel_panel_enable_backlight(crtc_state, conn_state); mutex_unlock(&dev_priv->backlight_lock); } @@ -1203,17 +1212,20 @@ static int intel_backlight_device_get_brightness(struct backlight_device *bd) struct intel_connector *connector = bl_get_data(bd); struct drm_device *dev = connector->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); - u32 hw_level; - int ret; + intel_wakeref_t wakeref; + int ret = 0; - intel_runtime_pm_get(dev_priv); - drm_modeset_lock(&dev->mode_config.connection_mutex, NULL); + with_intel_runtime_pm(dev_priv, wakeref) { + u32 hw_level; - hw_level = intel_panel_get_backlight(connector); - ret = scale_hw_to_user(connector, hw_level, bd->props.max_brightness); + drm_modeset_lock(&dev->mode_config.connection_mutex, NULL); - drm_modeset_unlock(&dev->mode_config.connection_mutex); - intel_runtime_pm_put(dev_priv); + hw_level = intel_panel_get_backlight(connector); + ret = scale_hw_to_user(connector, + hw_level, bd->props.max_brightness); + + drm_modeset_unlock(&dev->mode_config.connection_mutex); + } return ret; } @@ -1484,8 +1496,8 @@ static int lpt_setup_backlight(struct intel_connector *connector, enum pipe unus { struct drm_i915_private *dev_priv = to_i915(connector->base.dev); struct intel_panel *panel = &connector->panel; - u32 pch_ctl1, pch_ctl2, val; - bool alt; + u32 cpu_ctl2, pch_ctl1, pch_ctl2, val; + bool alt, cpu_mode; if (HAS_PCH_LPT(dev_priv)) alt = I915_READ(SOUTH_CHICKEN2) & LPT_PWM_GRANULARITY; @@ -1499,6 +1511,8 @@ static int lpt_setup_backlight(struct intel_connector *connector, enum pipe unus pch_ctl2 = I915_READ(BLC_PWM_PCH_CTL2); panel->backlight.max = pch_ctl2 >> 16; + cpu_ctl2 = I915_READ(BLC_PWM_CPU_CTL2); + if (!panel->backlight.max) panel->backlight.max = get_backlight_max_vbt(connector); @@ -1507,12 +1521,28 @@ static int lpt_setup_backlight(struct intel_connector *connector, enum pipe unus panel->backlight.min = get_backlight_min_vbt(connector); - val = lpt_get_backlight(connector); + panel->backlight.enabled = pch_ctl1 & BLM_PCH_PWM_ENABLE; + + cpu_mode = panel->backlight.enabled && HAS_PCH_LPT(dev_priv) && + !(pch_ctl1 & BLM_PCH_OVERRIDE_ENABLE) && + (cpu_ctl2 & BLM_PWM_ENABLE); + if (cpu_mode) + val = pch_get_backlight(connector); + else + val = lpt_get_backlight(connector); val = intel_panel_compute_brightness(connector, val); panel->backlight.level = clamp(val, panel->backlight.min, panel->backlight.max); - panel->backlight.enabled = pch_ctl1 & BLM_PCH_PWM_ENABLE; + if (cpu_mode) { + DRM_DEBUG_KMS("CPU backlight register was enabled, switching to PCH override\n"); + + /* Write converted CPU PWM value to PCH override register */ + lpt_set_backlight(connector->base.state, panel->backlight.level); + I915_WRITE(BLC_PWM_PCH_CTL1, pch_ctl1 | BLM_PCH_OVERRIDE_ENABLE); + + I915_WRITE(BLC_PWM_CPU_CTL2, cpu_ctl2 & ~BLM_PWM_ENABLE); + } return 0; } @@ -1773,6 +1803,24 @@ static int pwm_setup_backlight(struct intel_connector *connector, return 0; } +void intel_panel_update_backlight(struct intel_encoder *encoder, + const struct intel_crtc_state *crtc_state, + const struct drm_connector_state *conn_state) +{ + struct intel_connector *connector = to_intel_connector(conn_state->connector); + struct drm_i915_private *dev_priv = to_i915(connector->base.dev); + struct intel_panel *panel = &connector->panel; + + if (!panel->backlight.present) + return; + + mutex_lock(&dev_priv->backlight_lock); + if (!panel->backlight.enabled) + __intel_panel_enable_backlight(crtc_state, conn_state); + + mutex_unlock(&dev_priv->backlight_lock); +} + int intel_panel_setup_backlight(struct drm_connector *connector, enum pipe pipe) { struct drm_i915_private *dev_priv = to_i915(connector->dev); diff --git a/drivers/gpu/drm/i915/intel_pipe_crc.c b/drivers/gpu/drm/i915/intel_pipe_crc.c index bdabcfab8090..a8554dc4f196 100644 --- a/drivers/gpu/drm/i915/intel_pipe_crc.c +++ b/drivers/gpu/drm/i915/intel_pipe_crc.c @@ -44,7 +44,7 @@ static const char * const pipe_crc_sources[] = { }; static int i8xx_pipe_crc_ctl_reg(enum intel_pipe_crc_source *source, - uint32_t *val) + u32 *val) { if (*source == INTEL_PIPE_CRC_SOURCE_AUTO) *source = INTEL_PIPE_CRC_SOURCE_PIPE; @@ -120,7 +120,7 @@ static int i9xx_pipe_crc_auto_source(struct drm_i915_private *dev_priv, static int vlv_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, enum pipe pipe, enum intel_pipe_crc_source *source, - uint32_t *val) + u32 *val) { bool need_stable_symbols = false; @@ -165,7 +165,7 @@ static int vlv_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, * - DisplayPort scrambling: used for EMI reduction */ if (need_stable_symbols) { - uint32_t tmp = I915_READ(PORT_DFT2_G4X); + u32 tmp = I915_READ(PORT_DFT2_G4X); tmp |= DC_BALANCE_RESET_VLV; switch (pipe) { @@ -190,7 +190,7 @@ static int vlv_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, static int i9xx_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, enum pipe pipe, enum intel_pipe_crc_source *source, - uint32_t *val) + u32 *val) { bool need_stable_symbols = false; @@ -244,7 +244,7 @@ static int i9xx_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, * - DisplayPort scrambling: used for EMI reduction */ if (need_stable_symbols) { - uint32_t tmp = I915_READ(PORT_DFT2_G4X); + u32 tmp = I915_READ(PORT_DFT2_G4X); WARN_ON(!IS_G4X(dev_priv)); @@ -265,7 +265,7 @@ static int i9xx_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, static void vlv_undo_pipe_scramble_reset(struct drm_i915_private *dev_priv, enum pipe pipe) { - uint32_t tmp = I915_READ(PORT_DFT2_G4X); + u32 tmp = I915_READ(PORT_DFT2_G4X); switch (pipe) { case PIPE_A: @@ -289,7 +289,7 @@ static void vlv_undo_pipe_scramble_reset(struct drm_i915_private *dev_priv, static void g4x_undo_pipe_scramble_reset(struct drm_i915_private *dev_priv, enum pipe pipe) { - uint32_t tmp = I915_READ(PORT_DFT2_G4X); + u32 tmp = I915_READ(PORT_DFT2_G4X); if (pipe == PIPE_A) tmp &= ~PIPE_A_SCRAMBLE_RESET; @@ -304,7 +304,7 @@ static void g4x_undo_pipe_scramble_reset(struct drm_i915_private *dev_priv, } static int ilk_pipe_crc_ctl_reg(enum intel_pipe_crc_source *source, - uint32_t *val) + u32 *val) { if (*source == INTEL_PIPE_CRC_SOURCE_AUTO) *source = INTEL_PIPE_CRC_SOURCE_PIPE; @@ -392,7 +392,7 @@ unlock: static int ivb_pipe_crc_ctl_reg(struct drm_i915_private *dev_priv, enum pipe pipe, enum intel_pipe_crc_source *source, - uint32_t *val, + u32 *val, bool set_wa) { if (*source == INTEL_PIPE_CRC_SOURCE_AUTO) @@ -589,6 +589,7 @@ int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name) struct intel_pipe_crc *pipe_crc = &dev_priv->pipe_crc[crtc->index]; enum intel_display_power_domain power_domain; enum intel_pipe_crc_source source; + intel_wakeref_t wakeref; u32 val = 0; /* shut up gcc */ int ret = 0; @@ -598,7 +599,8 @@ int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name) } power_domain = POWER_DOMAIN_PIPE(crtc->index); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) { + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) { DRM_DEBUG_KMS("Trying to capture CRC while pipe is off\n"); return -EIO; } @@ -624,7 +626,7 @@ int intel_crtc_set_crc_source(struct drm_crtc *crtc, const char *source_name) pipe_crc->skipped = 0; out: - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); return ret; } diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 48c755dc895b..54307f1df6cf 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -483,7 +483,7 @@ static void vlv_get_fifo_size(struct intel_crtc_state *crtc_state) int sprite0_start, sprite1_start; switch (pipe) { - uint32_t dsparb, dsparb2, dsparb3; + u32 dsparb, dsparb2, dsparb3; case PIPE_A: dsparb = I915_READ(DSPARB); dsparb2 = I915_READ(DSPARB2); @@ -516,7 +516,7 @@ static void vlv_get_fifo_size(struct intel_crtc_state *crtc_state) static int i9xx_get_fifo_size(struct drm_i915_private *dev_priv, enum i9xx_plane_id i9xx_plane) { - uint32_t dsparb = I915_READ(DSPARB); + u32 dsparb = I915_READ(DSPARB); int size; size = dsparb & 0x7f; @@ -532,7 +532,7 @@ static int i9xx_get_fifo_size(struct drm_i915_private *dev_priv, static int i830_get_fifo_size(struct drm_i915_private *dev_priv, enum i9xx_plane_id i9xx_plane) { - uint32_t dsparb = I915_READ(DSPARB); + u32 dsparb = I915_READ(DSPARB); int size; size = dsparb & 0x1ff; @@ -549,7 +549,7 @@ static int i830_get_fifo_size(struct drm_i915_private *dev_priv, static int i845_get_fifo_size(struct drm_i915_private *dev_priv, enum i9xx_plane_id i9xx_plane) { - uint32_t dsparb = I915_READ(DSPARB); + u32 dsparb = I915_READ(DSPARB); int size; size = dsparb & 0x7f; @@ -670,9 +670,9 @@ static unsigned int intel_wm_method1(unsigned int pixel_rate, unsigned int cpp, unsigned int latency) { - uint64_t ret; + u64 ret; - ret = (uint64_t) pixel_rate * cpp * latency; + ret = (u64)pixel_rate * cpp * latency; ret = DIV_ROUND_UP_ULL(ret, 10000); return ret; @@ -1092,9 +1092,9 @@ static int g4x_fbc_fifo_size(int level) } } -static uint16_t g4x_compute_wm(const struct intel_crtc_state *crtc_state, - const struct intel_plane_state *plane_state, - int level) +static u16 g4x_compute_wm(const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state, + int level) { struct intel_plane *plane = to_intel_plane(plane_state->base.plane); struct drm_i915_private *dev_priv = to_i915(plane->base.dev); @@ -1191,9 +1191,9 @@ static bool g4x_raw_fbc_wm_set(struct intel_crtc_state *crtc_state, return dirty; } -static uint32_t ilk_compute_fbc_wm(const struct intel_crtc_state *cstate, - const struct intel_plane_state *pstate, - uint32_t pri_val); +static u32 ilk_compute_fbc_wm(const struct intel_crtc_state *cstate, + const struct intel_plane_state *pstate, + u32 pri_val); static bool g4x_raw_plane_wm_compute(struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) @@ -1601,9 +1601,9 @@ static void vlv_setup_wm_latency(struct drm_i915_private *dev_priv) } } -static uint16_t vlv_compute_wm_level(const struct intel_crtc_state *crtc_state, - const struct intel_plane_state *plane_state, - int level) +static u16 vlv_compute_wm_level(const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state, + int level) { struct intel_plane *plane = to_intel_plane(plane_state->base.plane); struct drm_i915_private *dev_priv = to_i915(plane->base.dev); @@ -1971,7 +1971,7 @@ static void vlv_atomic_update_fifo(struct intel_atomic_state *state, spin_lock(&dev_priv->uncore.lock); switch (crtc->pipe) { - uint32_t dsparb, dsparb2, dsparb3; + u32 dsparb, dsparb2, dsparb3; case PIPE_A: dsparb = I915_READ_FW(DSPARB); dsparb2 = I915_READ_FW(DSPARB2); @@ -2265,8 +2265,8 @@ static void i9xx_update_wm(struct intel_crtc *unused_crtc) { struct drm_i915_private *dev_priv = to_i915(unused_crtc->base.dev); const struct intel_watermark_params *wm_info; - uint32_t fwater_lo; - uint32_t fwater_hi; + u32 fwater_lo; + u32 fwater_hi; int cwm, srwm = 1; int fifo_size; int planea_wm, planeb_wm; @@ -2409,7 +2409,7 @@ static void i845_update_wm(struct intel_crtc *unused_crtc) struct drm_i915_private *dev_priv = to_i915(unused_crtc->base.dev); struct intel_crtc *crtc; const struct drm_display_mode *adjusted_mode; - uint32_t fwater_lo; + u32 fwater_lo; int planea_wm; crtc = single_enabled_crtc(dev_priv); @@ -2458,8 +2458,7 @@ static unsigned int ilk_wm_method2(unsigned int pixel_rate, return ret; } -static uint32_t ilk_wm_fbc(uint32_t pri_val, uint32_t horiz_pixels, - uint8_t cpp) +static u32 ilk_wm_fbc(u32 pri_val, u32 horiz_pixels, u8 cpp) { /* * Neither of these should be possible since this function shouldn't be @@ -2476,22 +2475,21 @@ static uint32_t ilk_wm_fbc(uint32_t pri_val, uint32_t horiz_pixels, } struct ilk_wm_maximums { - uint16_t pri; - uint16_t spr; - uint16_t cur; - uint16_t fbc; + u16 pri; + u16 spr; + u16 cur; + u16 fbc; }; /* * For both WM_PIPE and WM_LP. * mem_value must be in 0.1us units. */ -static uint32_t ilk_compute_pri_wm(const struct intel_crtc_state *cstate, - const struct intel_plane_state *pstate, - uint32_t mem_value, - bool is_lp) +static u32 ilk_compute_pri_wm(const struct intel_crtc_state *cstate, + const struct intel_plane_state *pstate, + u32 mem_value, bool is_lp) { - uint32_t method1, method2; + u32 method1, method2; int cpp; if (mem_value == 0) @@ -2519,11 +2517,11 @@ static uint32_t ilk_compute_pri_wm(const struct intel_crtc_state *cstate, * For both WM_PIPE and WM_LP. * mem_value must be in 0.1us units. */ -static uint32_t ilk_compute_spr_wm(const struct intel_crtc_state *cstate, - const struct intel_plane_state *pstate, - uint32_t mem_value) +static u32 ilk_compute_spr_wm(const struct intel_crtc_state *cstate, + const struct intel_plane_state *pstate, + u32 mem_value) { - uint32_t method1, method2; + u32 method1, method2; int cpp; if (mem_value == 0) @@ -2546,9 +2544,9 @@ static uint32_t ilk_compute_spr_wm(const struct intel_crtc_state *cstate, * For both WM_PIPE and WM_LP. * mem_value must be in 0.1us units. */ -static uint32_t ilk_compute_cur_wm(const struct intel_crtc_state *cstate, - const struct intel_plane_state *pstate, - uint32_t mem_value) +static u32 ilk_compute_cur_wm(const struct intel_crtc_state *cstate, + const struct intel_plane_state *pstate, + u32 mem_value) { int cpp; @@ -2566,9 +2564,9 @@ static uint32_t ilk_compute_cur_wm(const struct intel_crtc_state *cstate, } /* Only for WM_LP. */ -static uint32_t ilk_compute_fbc_wm(const struct intel_crtc_state *cstate, - const struct intel_plane_state *pstate, - uint32_t pri_val) +static u32 ilk_compute_fbc_wm(const struct intel_crtc_state *cstate, + const struct intel_plane_state *pstate, + u32 pri_val) { int cpp; @@ -2734,9 +2732,9 @@ static bool ilk_validate_wm_level(int level, DRM_DEBUG_KMS("Cursor WM%d too large %u (max %u)\n", level, result->cur_val, max->cur); - result->pri_val = min_t(uint32_t, result->pri_val, max->pri); - result->spr_val = min_t(uint32_t, result->spr_val, max->spr); - result->cur_val = min_t(uint32_t, result->cur_val, max->cur); + result->pri_val = min_t(u32, result->pri_val, max->pri); + result->spr_val = min_t(u32, result->spr_val, max->spr); + result->cur_val = min_t(u32, result->cur_val, max->cur); result->enable = true; } @@ -2752,9 +2750,9 @@ static void ilk_compute_wm_level(const struct drm_i915_private *dev_priv, const struct intel_plane_state *curstate, struct intel_wm_level *result) { - uint16_t pri_latency = dev_priv->wm.pri_latency[level]; - uint16_t spr_latency = dev_priv->wm.spr_latency[level]; - uint16_t cur_latency = dev_priv->wm.cur_latency[level]; + u16 pri_latency = dev_priv->wm.pri_latency[level]; + u16 spr_latency = dev_priv->wm.spr_latency[level]; + u16 cur_latency = dev_priv->wm.cur_latency[level]; /* WM1+ latency values stored in 0.5us units */ if (level > 0) { @@ -2778,7 +2776,7 @@ static void ilk_compute_wm_level(const struct drm_i915_private *dev_priv, result->enable = true; } -static uint32_t +static u32 hsw_compute_linetime_wm(const struct intel_crtc_state *cstate) { const struct intel_atomic_state *intel_state = @@ -2807,10 +2805,10 @@ hsw_compute_linetime_wm(const struct intel_crtc_state *cstate) } static void intel_read_wm_latency(struct drm_i915_private *dev_priv, - uint16_t wm[8]) + u16 wm[8]) { if (INTEL_GEN(dev_priv) >= 9) { - uint32_t val; + u32 val; int ret, i; int level, max_level = ilk_wm_max_level(dev_priv); @@ -2894,7 +2892,7 @@ static void intel_read_wm_latency(struct drm_i915_private *dev_priv, wm[0] += 1; } else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) { - uint64_t sskpd = I915_READ64(MCH_SSKPD); + u64 sskpd = I915_READ64(MCH_SSKPD); wm[0] = (sskpd >> 56) & 0xFF; if (wm[0] == 0) @@ -2904,14 +2902,14 @@ static void intel_read_wm_latency(struct drm_i915_private *dev_priv, wm[3] = (sskpd >> 20) & 0x1FF; wm[4] = (sskpd >> 32) & 0x1FF; } else if (INTEL_GEN(dev_priv) >= 6) { - uint32_t sskpd = I915_READ(MCH_SSKPD); + u32 sskpd = I915_READ(MCH_SSKPD); wm[0] = (sskpd >> SSKPD_WM0_SHIFT) & SSKPD_WM_MASK; wm[1] = (sskpd >> SSKPD_WM1_SHIFT) & SSKPD_WM_MASK; wm[2] = (sskpd >> SSKPD_WM2_SHIFT) & SSKPD_WM_MASK; wm[3] = (sskpd >> SSKPD_WM3_SHIFT) & SSKPD_WM_MASK; } else if (INTEL_GEN(dev_priv) >= 5) { - uint32_t mltr = I915_READ(MLTR_ILK); + u32 mltr = I915_READ(MLTR_ILK); /* ILK primary LP0 latency is 700 ns */ wm[0] = 7; @@ -2923,7 +2921,7 @@ static void intel_read_wm_latency(struct drm_i915_private *dev_priv, } static void intel_fixup_spr_wm_latency(struct drm_i915_private *dev_priv, - uint16_t wm[5]) + u16 wm[5]) { /* ILK sprite LP0 latency is 1300 ns */ if (IS_GEN(dev_priv, 5)) @@ -2931,7 +2929,7 @@ static void intel_fixup_spr_wm_latency(struct drm_i915_private *dev_priv, } static void intel_fixup_cur_wm_latency(struct drm_i915_private *dev_priv, - uint16_t wm[5]) + u16 wm[5]) { /* ILK cursor LP0 latency is 1300 ns */ if (IS_GEN(dev_priv, 5)) @@ -2953,7 +2951,7 @@ int ilk_wm_max_level(const struct drm_i915_private *dev_priv) static void intel_print_wm_latency(struct drm_i915_private *dev_priv, const char *name, - const uint16_t wm[8]) + const u16 wm[8]) { int level, max_level = ilk_wm_max_level(dev_priv); @@ -2982,7 +2980,7 @@ static void intel_print_wm_latency(struct drm_i915_private *dev_priv, } static bool ilk_increase_wm_latency(struct drm_i915_private *dev_priv, - uint16_t wm[5], uint16_t min) + u16 wm[5], u16 min) { int level, max_level = ilk_wm_max_level(dev_priv); @@ -2991,7 +2989,7 @@ static bool ilk_increase_wm_latency(struct drm_i915_private *dev_priv, wm[0] = max(wm[0], min); for (level = 1; level <= max_level; level++) - wm[level] = max_t(uint16_t, wm[level], DIV_ROUND_UP(min, 5)); + wm[level] = max_t(u16, wm[level], DIV_ROUND_UP(min, 5)); return true; } @@ -3538,7 +3536,7 @@ static void ilk_write_wm_values(struct drm_i915_private *dev_priv, { struct ilk_wm_values *previous = &dev_priv->wm.hw; unsigned int dirty; - uint32_t val; + u32 val; dirty = ilk_compute_wm_dirty(dev_priv, previous, results); if (!dirty) @@ -3636,14 +3634,9 @@ static u8 intel_enabled_dbuf_slices_num(struct drm_i915_private *dev_priv) * FIXME: We still don't have the proper code detect if we need to apply the WA, * so assume we'll always need it in order to avoid underruns. */ -static bool skl_needs_memory_bw_wa(struct intel_atomic_state *state) +static bool skl_needs_memory_bw_wa(struct drm_i915_private *dev_priv) { - struct drm_i915_private *dev_priv = to_i915(state->base.dev); - - if (IS_GEN9_BC(dev_priv) || IS_BROXTON(dev_priv)) - return true; - - return false; + return IS_GEN9_BC(dev_priv) || IS_BROXTON(dev_priv); } static bool @@ -3675,25 +3668,25 @@ intel_enable_sagv(struct drm_i915_private *dev_priv) if (dev_priv->sagv_status == I915_SAGV_ENABLED) return 0; - DRM_DEBUG_KMS("Enabling the SAGV\n"); + DRM_DEBUG_KMS("Enabling SAGV\n"); mutex_lock(&dev_priv->pcu_lock); ret = sandybridge_pcode_write(dev_priv, GEN9_PCODE_SAGV_CONTROL, GEN9_SAGV_ENABLE); - /* We don't need to wait for the SAGV when enabling */ + /* We don't need to wait for SAGV when enabling */ mutex_unlock(&dev_priv->pcu_lock); /* * Some skl systems, pre-release machines in particular, - * don't actually have an SAGV. + * don't actually have SAGV. */ if (IS_SKYLAKE(dev_priv) && ret == -ENXIO) { DRM_DEBUG_DRIVER("No SAGV found on system, ignoring\n"); dev_priv->sagv_status = I915_SAGV_NOT_CONTROLLED; return 0; } else if (ret < 0) { - DRM_ERROR("Failed to enable the SAGV\n"); + DRM_ERROR("Failed to enable SAGV\n"); return ret; } @@ -3712,7 +3705,7 @@ intel_disable_sagv(struct drm_i915_private *dev_priv) if (dev_priv->sagv_status == I915_SAGV_DISABLED) return 0; - DRM_DEBUG_KMS("Disabling the SAGV\n"); + DRM_DEBUG_KMS("Disabling SAGV\n"); mutex_lock(&dev_priv->pcu_lock); /* bspec says to keep retrying for at least 1 ms */ @@ -3724,14 +3717,14 @@ intel_disable_sagv(struct drm_i915_private *dev_priv) /* * Some skl systems, pre-release machines in particular, - * don't actually have an SAGV. + * don't actually have SAGV. */ if (IS_SKYLAKE(dev_priv) && ret == -ENXIO) { DRM_DEBUG_DRIVER("No SAGV found on system, ignoring\n"); dev_priv->sagv_status = I915_SAGV_NOT_CONTROLLED; return 0; } else if (ret < 0) { - DRM_ERROR("Failed to disable the SAGV (%d)\n", ret); + DRM_ERROR("Failed to disable SAGV (%d)\n", ret); return ret; } @@ -3762,7 +3755,7 @@ bool intel_can_enable_sagv(struct drm_atomic_state *state) sagv_block_time_us = 10; /* - * SKL+ workaround: bspec recommends we disable the SAGV when we have + * SKL+ workaround: bspec recommends we disable SAGV when we have * more then one pipe enabled * * If there are no active CRTCs, no additional checks need be performed @@ -3795,7 +3788,7 @@ bool intel_can_enable_sagv(struct drm_atomic_state *state) latency = dev_priv->wm.skl_latency[level]; - if (skl_needs_memory_bw_wa(intel_state) && + if (skl_needs_memory_bw_wa(dev_priv) && plane->base.state->fb->modifier == I915_FORMAT_MOD_X_TILED) latency += 15; @@ -3803,7 +3796,7 @@ bool intel_can_enable_sagv(struct drm_atomic_state *state) /* * If any of the planes on this pipe don't enable wm levels that * incur memory latencies higher than sagv_block_time_us we - * can't enable the SAGV. + * can't enable SAGV. */ if (latency < sagv_block_time_us) return false; @@ -3832,8 +3825,13 @@ static u16 intel_get_ddb_size(struct drm_i915_private *dev_priv, /* * 12GB/s is maximum BW supported by single DBuf slice. + * + * FIXME dbuf slice code is broken: + * - must wait for planes to stop using the slice before powering it off + * - plane straddling both slices is illegal in multi-pipe scenarios + * - should validate we stay within the hw bandwidth limits */ - if (num_active > 1 || total_data_bw >= GBps(12)) { + if (0 && (num_active > 1 || total_data_bw >= GBps(12))) { ddb->enabled_slices = 2; } else { ddb->enabled_slices = 1; @@ -3932,14 +3930,9 @@ static unsigned int skl_cursor_allocation(int num_active) static void skl_ddb_entry_init_from_hw(struct drm_i915_private *dev_priv, struct skl_ddb_entry *entry, u32 reg) { - u16 mask; - if (INTEL_GEN(dev_priv) >= 11) - mask = ICL_DDB_ENTRY_MASK; - else - mask = SKL_DDB_ENTRY_MASK; - entry->start = reg & mask; - entry->end = (reg >> DDB_ENTRY_END_SHIFT) & mask; + entry->start = reg & DDB_ENTRY_MASK; + entry->end = (reg >> DDB_ENTRY_END_SHIFT) & DDB_ENTRY_MASK; if (entry->end) entry->end += 1; @@ -3992,10 +3985,12 @@ void skl_pipe_ddb_get_hw_state(struct intel_crtc *crtc, struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum intel_display_power_domain power_domain; enum pipe pipe = crtc->pipe; + intel_wakeref_t wakeref; enum plane_id plane_id; power_domain = POWER_DOMAIN_PIPE(pipe); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) return; for_each_plane_id_on_crtc(crtc, plane_id) @@ -4004,7 +3999,7 @@ void skl_pipe_ddb_get_hw_state(struct intel_crtc *crtc, &ddb_y[plane_id], &ddb_uv[plane_id]); - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); } void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv, @@ -4034,7 +4029,7 @@ skl_plane_downscale_amount(const struct intel_crtc_state *cstate, const struct intel_plane_state *pstate) { struct intel_plane *plane = to_intel_plane(pstate->base.plane); - uint32_t src_w, src_h, dst_w, dst_h; + u32 src_w, src_h, dst_w, dst_h; uint_fixed_16_16_t fp_w_ratio, fp_h_ratio; uint_fixed_16_16_t downscale_h, downscale_w; @@ -4080,8 +4075,8 @@ skl_pipe_downscale_amount(const struct intel_crtc_state *crtc_state) return pipe_downscale; if (crtc_state->pch_pfit.enabled) { - uint32_t src_w, src_h, dst_w, dst_h; - uint32_t pfit_size = crtc_state->pch_pfit.size; + u32 src_w, src_h, dst_w, dst_h; + u32 pfit_size = crtc_state->pch_pfit.size; uint_fixed_16_16_t fp_w_ratio, fp_h_ratio; uint_fixed_16_16_t downscale_h, downscale_w; @@ -4114,7 +4109,7 @@ int skl_check_pipe_max_pixel_rate(struct intel_crtc *intel_crtc, const struct drm_plane_state *pstate; struct intel_plane_state *intel_pstate; int crtc_clock, dotclk; - uint32_t pipe_max_pixel_rate; + u32 pipe_max_pixel_rate; uint_fixed_16_16_t pipe_downscale; uint_fixed_16_16_t max_downscale = u32_to_fixed16(1); @@ -4170,8 +4165,8 @@ skl_plane_relative_data_rate(const struct intel_crtc_state *cstate, { struct intel_plane *intel_plane = to_intel_plane(intel_pstate->base.plane); - uint32_t data_rate; - uint32_t width = 0, height = 0; + u32 data_rate; + u32 width = 0, height = 0; struct drm_framebuffer *fb; u32 format; uint_fixed_16_16_t down_scale_amount; @@ -4314,15 +4309,15 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, struct intel_crtc *intel_crtc = to_intel_crtc(crtc); struct skl_ddb_entry *alloc = &cstate->wm.skl.ddb; struct skl_plane_wm *wm; - uint16_t alloc_size, start = 0; - uint16_t total[I915_MAX_PLANES] = {}; - uint16_t uv_total[I915_MAX_PLANES] = {}; + u16 alloc_size, start = 0; + u16 total[I915_MAX_PLANES] = {}; + u16 uv_total[I915_MAX_PLANES] = {}; u64 total_data_rate; enum plane_id plane_id; int num_active; u64 plane_data_rate[I915_MAX_PLANES] = {}; u64 uv_plane_data_rate[I915_MAX_PLANES] = {}; - uint16_t blocks = 0; + u32 blocks; int level; /* Clear the partitioning for disabled planes. */ @@ -4374,8 +4369,8 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, continue; wm = &cstate->wm.skl.optimal.planes[plane_id]; - blocks += wm->wm[level].plane_res_b; - blocks += wm->uv_wm[level].plane_res_b; + blocks += wm->wm[level].min_ddb_alloc; + blocks += wm->uv_wm[level].min_ddb_alloc; } if (blocks < alloc_size) { @@ -4416,7 +4411,7 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, extra = min_t(u16, alloc_size, DIV64_U64_ROUND_UP(alloc_size * rate, total_data_rate)); - total[plane_id] = wm->wm[level].plane_res_b + extra; + total[plane_id] = wm->wm[level].min_ddb_alloc + extra; alloc_size -= extra; total_data_rate -= rate; @@ -4427,7 +4422,7 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, extra = min_t(u16, alloc_size, DIV64_U64_ROUND_UP(alloc_size * rate, total_data_rate)); - uv_total[plane_id] = wm->uv_wm[level].plane_res_b + extra; + uv_total[plane_id] = wm->uv_wm[level].min_ddb_alloc + extra; alloc_size -= extra; total_data_rate -= rate; } @@ -4480,7 +4475,7 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, */ for_each_plane_id_on_crtc(intel_crtc, plane_id) { wm = &cstate->wm.skl.optimal.planes[plane_id]; - if (wm->trans_wm.plane_res_b > total[plane_id]) + if (wm->trans_wm.plane_res_b >= total[plane_id]) memset(&wm->trans_wm, 0, sizeof(wm->trans_wm)); } @@ -4494,10 +4489,10 @@ skl_allocate_pipe_ddb(struct intel_crtc_state *cstate, * 2xcdclk is 1350 MHz and the pixel rate should never exceed that. */ static uint_fixed_16_16_t -skl_wm_method1(const struct drm_i915_private *dev_priv, uint32_t pixel_rate, - uint8_t cpp, uint32_t latency, uint32_t dbuf_block_size) +skl_wm_method1(const struct drm_i915_private *dev_priv, u32 pixel_rate, + u8 cpp, u32 latency, u32 dbuf_block_size) { - uint32_t wm_intermediate_val; + u32 wm_intermediate_val; uint_fixed_16_16_t ret; if (latency == 0) @@ -4512,12 +4507,11 @@ skl_wm_method1(const struct drm_i915_private *dev_priv, uint32_t pixel_rate, return ret; } -static uint_fixed_16_16_t skl_wm_method2(uint32_t pixel_rate, - uint32_t pipe_htotal, - uint32_t latency, - uint_fixed_16_16_t plane_blocks_per_line) +static uint_fixed_16_16_t +skl_wm_method2(u32 pixel_rate, u32 pipe_htotal, u32 latency, + uint_fixed_16_16_t plane_blocks_per_line) { - uint32_t wm_intermediate_val; + u32 wm_intermediate_val; uint_fixed_16_16_t ret; if (latency == 0) @@ -4533,8 +4527,8 @@ static uint_fixed_16_16_t skl_wm_method2(uint32_t pixel_rate, static uint_fixed_16_16_t intel_get_linetime_us(const struct intel_crtc_state *cstate) { - uint32_t pixel_rate; - uint32_t crtc_htotal; + u32 pixel_rate; + u32 crtc_htotal; uint_fixed_16_16_t linetime_us; if (!cstate->base.active) @@ -4551,11 +4545,11 @@ intel_get_linetime_us(const struct intel_crtc_state *cstate) return linetime_us; } -static uint32_t +static u32 skl_adjusted_plane_pixel_rate(const struct intel_crtc_state *cstate, const struct intel_plane_state *pstate) { - uint64_t adjusted_pixel_rate; + u64 adjusted_pixel_rate; uint_fixed_16_16_t downscale_amount; /* Shouldn't reach here on disabled planes... */ @@ -4582,10 +4576,7 @@ skl_compute_plane_wm_params(const struct intel_crtc_state *cstate, struct drm_i915_private *dev_priv = to_i915(plane->base.dev); const struct drm_plane_state *pstate = &intel_pstate->base; const struct drm_framebuffer *fb = pstate->fb; - uint32_t interm_pbpl; - struct intel_atomic_state *state = - to_intel_atomic_state(cstate->base.state); - bool apply_memory_bw_wa = skl_needs_memory_bw_wa(state); + u32 interm_pbpl; /* only NV12 format has two planes */ if (color_plane == 1 && fb->format->format != DRM_FORMAT_NV12) { @@ -4621,7 +4612,7 @@ skl_compute_plane_wm_params(const struct intel_crtc_state *cstate, intel_pstate); if (INTEL_GEN(dev_priv) >= 11 && - fb->modifier == I915_FORMAT_MOD_Yf_TILED && wp->cpp == 8) + fb->modifier == I915_FORMAT_MOD_Yf_TILED && wp->cpp == 1) wp->dbuf_block_size = 256; else wp->dbuf_block_size = 512; @@ -4646,7 +4637,7 @@ skl_compute_plane_wm_params(const struct intel_crtc_state *cstate, wp->y_min_scanlines = 4; } - if (apply_memory_bw_wa) + if (skl_needs_memory_bw_wa(dev_priv)) wp->y_min_scanlines *= 2; wp->plane_bytes_per_line = wp->width * wp->cpp; @@ -4678,6 +4669,15 @@ skl_compute_plane_wm_params(const struct intel_crtc_state *cstate, return 0; } +static bool skl_wm_has_lines(struct drm_i915_private *dev_priv, int level) +{ + if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv)) + return true; + + /* The number of lines are ignored for the level 0 watermark. */ + return level > 0; +} + static void skl_compute_plane_wm(const struct intel_crtc_state *cstate, const struct intel_plane_state *intel_pstate, int level, @@ -4687,13 +4687,16 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *cstate, { struct drm_i915_private *dev_priv = to_i915(intel_pstate->base.plane->dev); - uint32_t latency = dev_priv->wm.skl_latency[level]; + u32 latency = dev_priv->wm.skl_latency[level]; uint_fixed_16_16_t method1, method2; uint_fixed_16_16_t selected_result; - uint32_t res_blocks, res_lines; - struct intel_atomic_state *state = - to_intel_atomic_state(cstate->base.state); - bool apply_memory_bw_wa = skl_needs_memory_bw_wa(state); + u32 res_blocks, res_lines, min_ddb_alloc = 0; + + if (latency == 0) { + /* reject it */ + result->min_ddb_alloc = U16_MAX; + return; + } /* Display WA #1141: kbl,cfl */ if ((IS_KABYLAKE(dev_priv) || IS_COFFEELAKE(dev_priv) || @@ -4701,7 +4704,7 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *cstate, dev_priv->ipc_enabled) latency += 4; - if (apply_memory_bw_wa && wp->x_tiled) + if (skl_needs_memory_bw_wa(dev_priv) && wp->x_tiled) latency += 15; method1 = skl_wm_method1(dev_priv, wp->plane_pixel_rate, @@ -4760,9 +4763,32 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *cstate, } } - /* The number of lines are ignored for the level 0 watermark. */ - if (level > 0 && res_lines > 31) + if (INTEL_GEN(dev_priv) >= 11) { + if (wp->y_tiled) { + int extra_lines; + + if (res_lines % wp->y_min_scanlines == 0) + extra_lines = wp->y_min_scanlines; + else + extra_lines = wp->y_min_scanlines * 2 - + res_lines % wp->y_min_scanlines; + + min_ddb_alloc = mul_round_up_u32_fixed16(res_lines + extra_lines, + wp->plane_blocks_per_line); + } else { + min_ddb_alloc = res_blocks + + DIV_ROUND_UP(res_blocks, 10); + } + } + + if (!skl_wm_has_lines(dev_priv, level)) + res_lines = 0; + + if (res_lines > 31) { + /* reject it */ + result->min_ddb_alloc = U16_MAX; return; + } /* * If res_lines is valid, assume we can use this watermark level @@ -4772,6 +4798,8 @@ static void skl_compute_plane_wm(const struct intel_crtc_state *cstate, */ result->plane_res_b = res_blocks; result->plane_res_l = res_lines; + /* Bspec says: value >= plane ddb allocation -> invalid, hence the +1 here */ + result->min_ddb_alloc = max(min_ddb_alloc, res_blocks) + 1; result->plane_en = true; } @@ -4796,24 +4824,19 @@ skl_compute_wm_levels(const struct intel_crtc_state *cstate, } } -static uint32_t +static u32 skl_compute_linetime_wm(const struct intel_crtc_state *cstate) { struct drm_atomic_state *state = cstate->base.state; struct drm_i915_private *dev_priv = to_i915(state->dev); uint_fixed_16_16_t linetime_us; - uint32_t linetime_wm; + u32 linetime_wm; linetime_us = intel_get_linetime_us(cstate); - - if (is_fixed16_zero(linetime_us)) - return 0; - linetime_wm = fixed16_to_u32_round_up(mul_u32_fixed16(8, linetime_us)); - /* Display WA #1135: bxt:ALL GLK:ALL */ - if ((IS_BROXTON(dev_priv) || IS_GEMINILAKE(dev_priv)) && - dev_priv->ipc_enabled) + /* Display WA #1135: BXT:ALL GLK:ALL */ + if (IS_GEN9_LP(dev_priv) && dev_priv->ipc_enabled) linetime_wm /= 2; return linetime_wm; @@ -4825,9 +4848,9 @@ static void skl_compute_transition_wm(const struct intel_crtc_state *cstate, { struct drm_device *dev = cstate->base.crtc->dev; const struct drm_i915_private *dev_priv = to_i915(dev); - uint16_t trans_min, trans_y_tile_min; - const uint16_t trans_amount = 10; /* This is configurable amount */ - uint16_t wm0_sel_res_b, trans_offset_b, res_blocks; + u16 trans_min, trans_y_tile_min; + const u16 trans_amount = 10; /* This is configurable amount */ + u16 wm0_sel_res_b, trans_offset_b, res_blocks; /* Transition WM are not recommended by HW team for GEN9 */ if (INTEL_GEN(dev_priv) <= 9) @@ -4856,8 +4879,8 @@ static void skl_compute_transition_wm(const struct intel_crtc_state *cstate, wm0_sel_res_b = wm->wm[0].plane_res_b - 1; if (wp->y_tiled) { - trans_y_tile_min = (uint16_t) mul_round_up_u32_fixed16(2, - wp->y_tile_minimum); + trans_y_tile_min = + (u16)mul_round_up_u32_fixed16(2, wp->y_tile_minimum); res_blocks = max(wm0_sel_res_b, trans_y_tile_min) + trans_offset_b; } else { @@ -5031,7 +5054,7 @@ static void skl_write_wm_level(struct drm_i915_private *dev_priv, i915_reg_t reg, const struct skl_wm_level *level) { - uint32_t val = 0; + u32 val = 0; if (level->plane_en) { val |= PLANE_WM_EN; @@ -5122,6 +5145,23 @@ static bool skl_plane_wm_equals(struct drm_i915_private *dev_priv, return skl_wm_level_equals(&wm1->trans_wm, &wm2->trans_wm); } +static bool skl_pipe_wm_equals(struct intel_crtc *crtc, + const struct skl_pipe_wm *wm1, + const struct skl_pipe_wm *wm2) +{ + struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); + enum plane_id plane_id; + + for_each_plane_id_on_crtc(crtc, plane_id) { + if (!skl_plane_wm_equals(dev_priv, + &wm1->planes[plane_id], + &wm2->planes[plane_id])) + return false; + } + + return wm1->linetime == wm2->linetime; +} + static inline bool skl_ddb_entries_overlap(const struct skl_ddb_entry *a, const struct skl_ddb_entry *b) { @@ -5148,26 +5188,24 @@ static int skl_update_pipe_wm(struct intel_crtc_state *cstate, struct skl_pipe_wm *pipe_wm, /* out */ bool *changed /* out */) { + struct intel_crtc *crtc = to_intel_crtc(cstate->base.crtc); int ret; ret = skl_build_pipe_wm(cstate, pipe_wm); if (ret) return ret; - if (!memcmp(old_pipe_wm, pipe_wm, sizeof(*pipe_wm))) - *changed = false; - else - *changed = true; + *changed = !skl_pipe_wm_equals(crtc, old_pipe_wm, pipe_wm); return 0; } -static uint32_t +static u32 pipes_modified(struct intel_atomic_state *state) { struct intel_crtc *crtc; struct intel_crtc_state *cstate; - uint32_t i, ret = 0; + u32 i, ret = 0; for_each_new_intel_crtc_in_state(state, crtc, cstate, i) ret |= drm_crtc_mask(&crtc->base); @@ -5268,7 +5306,7 @@ skl_ddb_add_affected_pipes(struct intel_atomic_state *state, bool *changed) const struct drm_i915_private *dev_priv = to_i915(dev); struct intel_crtc *crtc; struct intel_crtc_state *crtc_state; - uint32_t realloc_pipes = pipes_modified(state); + u32 realloc_pipes = pipes_modified(state); int ret, i; /* @@ -5567,7 +5605,7 @@ static void ilk_optimize_watermarks(struct intel_atomic_state *state, mutex_unlock(&dev_priv->wm.wm_mutex); } -static inline void skl_wm_level_from_reg_val(uint32_t val, +static inline void skl_wm_level_from_reg_val(u32 val, struct skl_wm_level *level) { level->plane_en = val & PLANE_WM_EN; @@ -5583,7 +5621,7 @@ void skl_pipe_wm_get_hw_state(struct intel_crtc *crtc, enum pipe pipe = crtc->pipe; int level, max_level; enum plane_id plane_id; - uint32_t val; + u32 val; max_level = ilk_wm_max_level(dev_priv); @@ -5695,7 +5733,7 @@ static void ilk_pipe_wm_get_hw_state(struct intel_crtc *crtc) static void g4x_read_wm_values(struct drm_i915_private *dev_priv, struct g4x_wm_values *wm) { - uint32_t tmp; + u32 tmp; tmp = I915_READ(DSPFW1); wm->sr.plane = _FW_WM(tmp, SR); @@ -5722,7 +5760,7 @@ static void vlv_read_wm_values(struct drm_i915_private *dev_priv, struct vlv_wm_values *wm) { enum pipe pipe; - uint32_t tmp; + u32 tmp; for_each_pipe(dev_priv, pipe) { tmp = I915_READ(VLV_DDL(pipe)); @@ -6204,10 +6242,6 @@ void intel_init_ipc(struct drm_i915_private *dev_priv) */ DEFINE_SPINLOCK(mchdev_lock); -/* Global for IPS driver to get at the current i915 device. Protected by - * mchdev_lock. */ -static struct drm_i915_private *i915_mch_dev; - bool ironlake_set_drps(struct drm_i915_private *dev_priv, u8 val) { u16 rgvswctl; @@ -6670,7 +6704,7 @@ void gen6_rps_boost(struct i915_request *rq, if (!rps->enabled) return; - if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags)) + if (i915_request_signaled(rq)) return; /* Serializes with i915_request_retire() */ @@ -7850,16 +7884,17 @@ static unsigned long __i915_chipset_val(struct drm_i915_private *dev_priv) unsigned long i915_chipset_val(struct drm_i915_private *dev_priv) { - unsigned long val; + intel_wakeref_t wakeref; + unsigned long val = 0; if (!IS_GEN(dev_priv, 5)) return 0; - spin_lock_irq(&mchdev_lock); - - val = __i915_chipset_val(dev_priv); - - spin_unlock_irq(&mchdev_lock); + with_intel_runtime_pm(dev_priv, wakeref) { + spin_lock_irq(&mchdev_lock); + val = __i915_chipset_val(dev_priv); + spin_unlock_irq(&mchdev_lock); + } return val; } @@ -7936,14 +7971,16 @@ static void __i915_update_gfx_val(struct drm_i915_private *dev_priv) void i915_update_gfx_val(struct drm_i915_private *dev_priv) { + intel_wakeref_t wakeref; + if (!IS_GEN(dev_priv, 5)) return; - spin_lock_irq(&mchdev_lock); - - __i915_update_gfx_val(dev_priv); - - spin_unlock_irq(&mchdev_lock); + with_intel_runtime_pm(dev_priv, wakeref) { + spin_lock_irq(&mchdev_lock); + __i915_update_gfx_val(dev_priv); + spin_unlock_irq(&mchdev_lock); + } } static unsigned long __i915_gfx_val(struct drm_i915_private *dev_priv) @@ -7985,18 +8022,34 @@ static unsigned long __i915_gfx_val(struct drm_i915_private *dev_priv) unsigned long i915_gfx_val(struct drm_i915_private *dev_priv) { - unsigned long val; + intel_wakeref_t wakeref; + unsigned long val = 0; if (!IS_GEN(dev_priv, 5)) return 0; - spin_lock_irq(&mchdev_lock); + with_intel_runtime_pm(dev_priv, wakeref) { + spin_lock_irq(&mchdev_lock); + val = __i915_gfx_val(dev_priv); + spin_unlock_irq(&mchdev_lock); + } - val = __i915_gfx_val(dev_priv); + return val; +} - spin_unlock_irq(&mchdev_lock); +static struct drm_i915_private *i915_mch_dev; - return val; +static struct drm_i915_private *mchdev_get(void) +{ + struct drm_i915_private *i915; + + rcu_read_lock(); + i915 = i915_mch_dev; + if (!kref_get_unless_zero(&i915->drm.ref)) + i915 = NULL; + rcu_read_unlock(); + + return i915; } /** @@ -8007,23 +8060,24 @@ unsigned long i915_gfx_val(struct drm_i915_private *dev_priv) */ unsigned long i915_read_mch_val(void) { - struct drm_i915_private *dev_priv; - unsigned long chipset_val, graphics_val, ret = 0; - - spin_lock_irq(&mchdev_lock); - if (!i915_mch_dev) - goto out_unlock; - dev_priv = i915_mch_dev; - - chipset_val = __i915_chipset_val(dev_priv); - graphics_val = __i915_gfx_val(dev_priv); + struct drm_i915_private *i915; + unsigned long chipset_val = 0; + unsigned long graphics_val = 0; + intel_wakeref_t wakeref; - ret = chipset_val + graphics_val; + i915 = mchdev_get(); + if (!i915) + return 0; -out_unlock: - spin_unlock_irq(&mchdev_lock); + with_intel_runtime_pm(i915, wakeref) { + spin_lock_irq(&mchdev_lock); + chipset_val = __i915_chipset_val(i915); + graphics_val = __i915_gfx_val(i915); + spin_unlock_irq(&mchdev_lock); + } - return ret; + drm_dev_put(&i915->drm); + return chipset_val + graphics_val; } EXPORT_SYMBOL_GPL(i915_read_mch_val); @@ -8034,23 +8088,19 @@ EXPORT_SYMBOL_GPL(i915_read_mch_val); */ bool i915_gpu_raise(void) { - struct drm_i915_private *dev_priv; - bool ret = true; - - spin_lock_irq(&mchdev_lock); - if (!i915_mch_dev) { - ret = false; - goto out_unlock; - } - dev_priv = i915_mch_dev; + struct drm_i915_private *i915; - if (dev_priv->ips.max_delay > dev_priv->ips.fmax) - dev_priv->ips.max_delay--; + i915 = mchdev_get(); + if (!i915) + return false; -out_unlock: + spin_lock_irq(&mchdev_lock); + if (i915->ips.max_delay > i915->ips.fmax) + i915->ips.max_delay--; spin_unlock_irq(&mchdev_lock); - return ret; + drm_dev_put(&i915->drm); + return true; } EXPORT_SYMBOL_GPL(i915_gpu_raise); @@ -8062,23 +8112,19 @@ EXPORT_SYMBOL_GPL(i915_gpu_raise); */ bool i915_gpu_lower(void) { - struct drm_i915_private *dev_priv; - bool ret = true; + struct drm_i915_private *i915; - spin_lock_irq(&mchdev_lock); - if (!i915_mch_dev) { - ret = false; - goto out_unlock; - } - dev_priv = i915_mch_dev; - - if (dev_priv->ips.max_delay < dev_priv->ips.min_delay) - dev_priv->ips.max_delay++; + i915 = mchdev_get(); + if (!i915) + return false; -out_unlock: + spin_lock_irq(&mchdev_lock); + if (i915->ips.max_delay < i915->ips.min_delay) + i915->ips.max_delay++; spin_unlock_irq(&mchdev_lock); - return ret; + drm_dev_put(&i915->drm); + return true; } EXPORT_SYMBOL_GPL(i915_gpu_lower); @@ -8089,13 +8135,16 @@ EXPORT_SYMBOL_GPL(i915_gpu_lower); */ bool i915_gpu_busy(void) { - bool ret = false; + struct drm_i915_private *i915; + bool ret; - spin_lock_irq(&mchdev_lock); - if (i915_mch_dev) - ret = i915_mch_dev->gt.awake; - spin_unlock_irq(&mchdev_lock); + i915 = mchdev_get(); + if (!i915) + return false; + ret = i915->gt.awake; + + drm_dev_put(&i915->drm); return ret; } EXPORT_SYMBOL_GPL(i915_gpu_busy); @@ -8108,24 +8157,19 @@ EXPORT_SYMBOL_GPL(i915_gpu_busy); */ bool i915_gpu_turbo_disable(void) { - struct drm_i915_private *dev_priv; - bool ret = true; - - spin_lock_irq(&mchdev_lock); - if (!i915_mch_dev) { - ret = false; - goto out_unlock; - } - dev_priv = i915_mch_dev; - - dev_priv->ips.max_delay = dev_priv->ips.fstart; + struct drm_i915_private *i915; + bool ret; - if (!ironlake_set_drps(dev_priv, dev_priv->ips.fstart)) - ret = false; + i915 = mchdev_get(); + if (!i915) + return false; -out_unlock: + spin_lock_irq(&mchdev_lock); + i915->ips.max_delay = i915->ips.fstart; + ret = ironlake_set_drps(i915, i915->ips.fstart); spin_unlock_irq(&mchdev_lock); + drm_dev_put(&i915->drm); return ret; } EXPORT_SYMBOL_GPL(i915_gpu_turbo_disable); @@ -8154,18 +8198,14 @@ void intel_gpu_ips_init(struct drm_i915_private *dev_priv) { /* We only register the i915 ips part with intel-ips once everything is * set up, to avoid intel-ips sneaking in and reading bogus values. */ - spin_lock_irq(&mchdev_lock); - i915_mch_dev = dev_priv; - spin_unlock_irq(&mchdev_lock); + rcu_assign_pointer(i915_mch_dev, dev_priv); ips_ping_for_i915_load(); } void intel_gpu_ips_teardown(void) { - spin_lock_irq(&mchdev_lock); - i915_mch_dev = NULL; - spin_unlock_irq(&mchdev_lock); + rcu_assign_pointer(i915_mch_dev, NULL); } static void intel_init_emon(struct drm_i915_private *dev_priv) @@ -8504,7 +8544,7 @@ static void g4x_disable_trickle_feed(struct drm_i915_private *dev_priv) static void ilk_init_clock_gating(struct drm_i915_private *dev_priv) { - uint32_t dspclk_gate = ILK_VRHUNIT_CLOCK_GATE_DISABLE; + u32 dspclk_gate = ILK_VRHUNIT_CLOCK_GATE_DISABLE; /* * Required for FBC @@ -8576,7 +8616,7 @@ static void ilk_init_clock_gating(struct drm_i915_private *dev_priv) static void cpt_init_clock_gating(struct drm_i915_private *dev_priv) { int pipe; - uint32_t val; + u32 val; /* * On Ibex Peak and Cougar Point, we need to disable clock @@ -8611,7 +8651,7 @@ static void cpt_init_clock_gating(struct drm_i915_private *dev_priv) static void gen6_check_mch_setup(struct drm_i915_private *dev_priv) { - uint32_t tmp; + u32 tmp; tmp = I915_READ(MCH_SSKPD); if ((tmp & MCH_SSKPD_WM0_MASK) != MCH_SSKPD_WM0_VAL) @@ -8621,7 +8661,7 @@ static void gen6_check_mch_setup(struct drm_i915_private *dev_priv) static void gen6_init_clock_gating(struct drm_i915_private *dev_priv) { - uint32_t dspclk_gate = ILK_VRHUNIT_CLOCK_GATE_DISABLE; + u32 dspclk_gate = ILK_VRHUNIT_CLOCK_GATE_DISABLE; I915_WRITE(ILK_DSPCLK_GATE_D, dspclk_gate); @@ -8715,7 +8755,7 @@ static void gen6_init_clock_gating(struct drm_i915_private *dev_priv) static void gen7_setup_fixed_func_scheduler(struct drm_i915_private *dev_priv) { - uint32_t reg = I915_READ(GEN7_FF_THREAD_MODE); + u32 reg = I915_READ(GEN7_FF_THREAD_MODE); /* * WaVSThreadDispatchOverride:ivb,vlv @@ -8751,7 +8791,7 @@ static void lpt_init_clock_gating(struct drm_i915_private *dev_priv) static void lpt_suspend_hw(struct drm_i915_private *dev_priv) { if (HAS_PCH_LPT_LP(dev_priv)) { - uint32_t val = I915_READ(SOUTH_DSPCLK_GATE_D); + u32 val = I915_READ(SOUTH_DSPCLK_GATE_D); val &= ~PCH_LP_PARTITION_LEVEL_DISABLE; I915_WRITE(SOUTH_DSPCLK_GATE_D, val); @@ -8989,7 +9029,7 @@ static void hsw_init_clock_gating(struct drm_i915_private *dev_priv) static void ivb_init_clock_gating(struct drm_i915_private *dev_priv) { - uint32_t snpcr; + u32 snpcr; I915_WRITE(ILK_DSPCLK_GATE_D, ILK_VRHUNIT_CLOCK_GATE_DISABLE); @@ -9198,7 +9238,7 @@ static void chv_init_clock_gating(struct drm_i915_private *dev_priv) static void g4x_init_clock_gating(struct drm_i915_private *dev_priv) { - uint32_t dspclk_gate; + u32 dspclk_gate; I915_WRITE(RENCLK_GATE_D1, 0); I915_WRITE(RENCLK_GATE_D2, VF_UNIT_CLOCK_GATE_DISABLE | @@ -9448,7 +9488,7 @@ void intel_init_pm(struct drm_i915_private *dev_priv) static inline int gen6_check_mailbox_status(struct drm_i915_private *dev_priv) { - uint32_t flags = + u32 flags = I915_READ_FW(GEN6_PCODE_MAILBOX) & GEN6_PCODE_ERROR_MASK; switch (flags) { @@ -9471,7 +9511,7 @@ static inline int gen6_check_mailbox_status(struct drm_i915_private *dev_priv) static inline int gen7_check_mailbox_status(struct drm_i915_private *dev_priv) { - uint32_t flags = + u32 flags = I915_READ_FW(GEN6_PCODE_MAILBOX) & GEN6_PCODE_ERROR_MASK; switch (flags) { diff --git a/drivers/gpu/drm/i915/intel_psr.c b/drivers/gpu/drm/i915/intel_psr.c index 0f6b2b4702e3..84a0fb981561 100644 --- a/drivers/gpu/drm/i915/intel_psr.c +++ b/drivers/gpu/drm/i915/intel_psr.c @@ -70,17 +70,17 @@ static bool psr_global_enabled(u32 debug) static bool intel_psr2_enabled(struct drm_i915_private *dev_priv, const struct intel_crtc_state *crtc_state) { - /* Disable PSR2 by default for all platforms */ - if (i915_modparams.enable_psr == -1) - return false; - /* Cannot enable DSC and PSR2 simultaneously */ WARN_ON(crtc_state->dsc_params.compression_enable && crtc_state->has_psr2); switch (dev_priv->psr.debug & I915_PSR_DEBUG_MODE_MASK) { + case I915_PSR_DEBUG_DISABLE: case I915_PSR_DEBUG_FORCE_PSR1: return false; + case I915_PSR_DEBUG_DEFAULT: + if (i915_modparams.enable_psr <= 0) + return false; default: return crtc_state->has_psr2; } @@ -230,7 +230,7 @@ void intel_psr_irq_handler(struct drm_i915_private *dev_priv, u32 psr_iir) static bool intel_dp_get_colorimetry_status(struct intel_dp *intel_dp) { - uint8_t dprx = 0; + u8 dprx = 0; if (drm_dp_dpcd_readb(&intel_dp->aux, DP_DPRX_FEATURE_ENUMERATION_LIST, &dprx) != 1) @@ -240,7 +240,7 @@ static bool intel_dp_get_colorimetry_status(struct intel_dp *intel_dp) static bool intel_dp_get_alpm_status(struct intel_dp *intel_dp) { - uint8_t alpm_caps = 0; + u8 alpm_caps = 0; if (drm_dp_dpcd_readb(&intel_dp->aux, DP_RECEIVER_ALPM_CAP, &alpm_caps) != 1) @@ -384,7 +384,7 @@ static void hsw_psr_setup_aux(struct intel_dp *intel_dp) struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); u32 aux_clock_divider, aux_ctl; int i; - static const uint8_t aux_msg[] = { + static const u8 aux_msg[] = { [0] = DP_AUX_NATIVE_WRITE << 4, [1] = DP_SET_POWER >> 8, [2] = DP_SET_POWER & 0xff, diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 26b7274a2d43..7f841dba87b3 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -33,6 +33,7 @@ #include "i915_drv.h" #include "i915_gem_render_state.h" +#include "i915_reset.h" #include "i915_trace.h" #include "intel_drv.h" #include "intel_workarounds.h" @@ -42,17 +43,10 @@ */ #define LEGACY_REQUEST_SIZE 200 -static unsigned int __intel_ring_space(unsigned int head, - unsigned int tail, - unsigned int size) +static inline u32 intel_hws_seqno_address(struct intel_engine_cs *engine) { - /* - * "If the Ring Buffer Head Pointer and the Tail Pointer are on the - * same cacheline, the Head Pointer must not be greater than the Tail - * Pointer." - */ - GEM_BUG_ON(!is_power_of_2(size)); - return (head - tail - CACHELINE_BYTES) & (size - 1); + return (i915_ggtt_offset(engine->status_page.vma) + + I915_GEM_HWS_INDEX_ADDR); } unsigned int intel_ring_update_space(struct intel_ring *ring) @@ -299,7 +293,7 @@ gen6_render_ring_flush(struct i915_request *rq, u32 mode) return 0; } -static void gen6_rcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) +static u32 *gen6_rcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) { /* First we do the gen6_emit_post_sync_nonzero_flush w/a */ *cs++ = GFX_OP_PIPE_CONTROL(4); @@ -319,6 +313,11 @@ static void gen6_rcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) PIPE_CONTROL_DC_FLUSH_ENABLE | PIPE_CONTROL_QW_WRITE | PIPE_CONTROL_CS_STALL); + *cs++ = rq->timeline->hwsp_offset | PIPE_CONTROL_GLOBAL_GTT; + *cs++ = rq->fence.seqno; + + *cs++ = GFX_OP_PIPE_CONTROL(4); + *cs++ = PIPE_CONTROL_QW_WRITE | PIPE_CONTROL_CS_STALL; *cs++ = intel_hws_seqno_address(rq->engine) | PIPE_CONTROL_GLOBAL_GTT; *cs++ = rq->global_seqno; @@ -327,8 +326,9 @@ static void gen6_rcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) rq->tail = intel_ring_offset(rq, cs); assert_ring_tail_valid(rq->ring, rq->tail); + + return cs; } -static const int gen6_rcs_emit_breadcrumb_sz = 14; static int gen7_render_ring_cs_stall_wa(struct i915_request *rq) @@ -409,7 +409,7 @@ gen7_render_ring_flush(struct i915_request *rq, u32 mode) return 0; } -static void gen7_rcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) +static u32 *gen7_rcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) { *cs++ = GFX_OP_PIPE_CONTROL(4); *cs++ = (PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH | @@ -419,6 +419,13 @@ static void gen7_rcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) PIPE_CONTROL_QW_WRITE | PIPE_CONTROL_GLOBAL_GTT_IVB | PIPE_CONTROL_CS_STALL); + *cs++ = rq->timeline->hwsp_offset; + *cs++ = rq->fence.seqno; + + *cs++ = GFX_OP_PIPE_CONTROL(4); + *cs++ = (PIPE_CONTROL_QW_WRITE | + PIPE_CONTROL_GLOBAL_GTT_IVB | + PIPE_CONTROL_CS_STALL); *cs++ = intel_hws_seqno_address(rq->engine); *cs++ = rq->global_seqno; @@ -427,34 +434,52 @@ static void gen7_rcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) rq->tail = intel_ring_offset(rq, cs); assert_ring_tail_valid(rq->ring, rq->tail); + + return cs; } -static const int gen7_rcs_emit_breadcrumb_sz = 6; -static void gen6_xcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) +static u32 *gen6_xcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) { - *cs++ = MI_FLUSH_DW | MI_FLUSH_DW_OP_STOREDW; - *cs++ = intel_hws_seqno_address(rq->engine) | MI_FLUSH_DW_USE_GTT; + GEM_BUG_ON(rq->timeline->hwsp_ggtt != rq->engine->status_page.vma); + GEM_BUG_ON(offset_in_page(rq->timeline->hwsp_offset) != I915_GEM_HWS_SEQNO_ADDR); + + *cs++ = MI_FLUSH_DW | MI_FLUSH_DW_OP_STOREDW | MI_FLUSH_DW_STORE_INDEX; + *cs++ = I915_GEM_HWS_SEQNO_ADDR | MI_FLUSH_DW_USE_GTT; + *cs++ = rq->fence.seqno; + + *cs++ = MI_FLUSH_DW | MI_FLUSH_DW_OP_STOREDW | MI_FLUSH_DW_STORE_INDEX; + *cs++ = I915_GEM_HWS_INDEX_ADDR | MI_FLUSH_DW_USE_GTT; *cs++ = rq->global_seqno; + *cs++ = MI_USER_INTERRUPT; + *cs++ = MI_NOOP; rq->tail = intel_ring_offset(rq, cs); assert_ring_tail_valid(rq->ring, rq->tail); + + return cs; } -static const int gen6_xcs_emit_breadcrumb_sz = 4; #define GEN7_XCS_WA 32 -static void gen7_xcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) +static u32 *gen7_xcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) { int i; - *cs++ = MI_FLUSH_DW | MI_FLUSH_DW_OP_STOREDW; - *cs++ = intel_hws_seqno_address(rq->engine) | MI_FLUSH_DW_USE_GTT; + GEM_BUG_ON(rq->timeline->hwsp_ggtt != rq->engine->status_page.vma); + GEM_BUG_ON(offset_in_page(rq->timeline->hwsp_offset) != I915_GEM_HWS_SEQNO_ADDR); + + *cs++ = MI_FLUSH_DW | MI_FLUSH_DW_OP_STOREDW | MI_FLUSH_DW_STORE_INDEX; + *cs++ = I915_GEM_HWS_SEQNO_ADDR | MI_FLUSH_DW_USE_GTT; + *cs++ = rq->fence.seqno; + + *cs++ = MI_FLUSH_DW | MI_FLUSH_DW_OP_STOREDW | MI_FLUSH_DW_STORE_INDEX; + *cs++ = I915_GEM_HWS_INDEX_ADDR | MI_FLUSH_DW_USE_GTT; *cs++ = rq->global_seqno; for (i = 0; i < GEN7_XCS_WA; i++) { *cs++ = MI_STORE_DWORD_INDEX; - *cs++ = I915_GEM_HWS_INDEX_ADDR; - *cs++ = rq->global_seqno; + *cs++ = I915_GEM_HWS_SEQNO_ADDR; + *cs++ = rq->fence.seqno; } *cs++ = MI_FLUSH_DW; @@ -462,12 +487,12 @@ static void gen7_xcs_emit_breadcrumb(struct i915_request *rq, u32 *cs) *cs++ = 0; *cs++ = MI_USER_INTERRUPT; - *cs++ = MI_NOOP; rq->tail = intel_ring_offset(rq, cs); assert_ring_tail_valid(rq->ring, rq->tail); + + return cs; } -static const int gen7_xcs_emit_breadcrumb_sz = 8 + GEN7_XCS_WA * 3; #undef GEN7_XCS_WA static void set_hwstam(struct intel_engine_cs *engine, u32 mask) @@ -498,12 +523,17 @@ static void set_hws_pga(struct intel_engine_cs *engine, phys_addr_t phys) I915_WRITE(HWS_PGA, addr); } -static void ring_setup_phys_status_page(struct intel_engine_cs *engine) +static struct page *status_page(struct intel_engine_cs *engine) { - struct page *page = virt_to_page(engine->status_page.page_addr); - phys_addr_t phys = PFN_PHYS(page_to_pfn(page)); + struct drm_i915_gem_object *obj = engine->status_page.vma->obj; - set_hws_pga(engine, phys); + GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj)); + return sg_page(obj->mm.pages->sgl); +} + +static void ring_setup_phys_status_page(struct intel_engine_cs *engine) +{ + set_hws_pga(engine, PFN_PHYS(page_to_pfn(status_page(engine)))); set_hwstam(engine, ~0u); } @@ -570,7 +600,7 @@ static void flush_cs_tlb(struct intel_engine_cs *engine) static void ring_setup_status_page(struct intel_engine_cs *engine) { - set_hwsp(engine, engine->status_page.ggtt_offset); + set_hwsp(engine, i915_ggtt_offset(engine->status_page.vma)); set_hwstam(engine, ~0u); flush_cs_tlb(engine); @@ -700,59 +730,87 @@ static int init_ring_common(struct intel_engine_cs *engine) } /* Papering over lost _interrupts_ immediately following the restart */ - intel_engine_wakeup(engine); + intel_engine_queue_breadcrumbs(engine); out: intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); return ret; } -static struct i915_request *reset_prepare(struct intel_engine_cs *engine) +static void reset_prepare(struct intel_engine_cs *engine) { intel_engine_stop_cs(engine); - return i915_gem_find_active_request(engine); } -static void skip_request(struct i915_request *rq) +static void reset_ring(struct intel_engine_cs *engine, bool stalled) { - void *vaddr = rq->ring->vaddr; + struct i915_timeline *tl = &engine->timeline; + struct i915_request *pos, *rq; + unsigned long flags; u32 head; - head = rq->infix; - if (rq->postfix < head) { - memset32(vaddr + head, MI_NOOP, - (rq->ring->size - head) / sizeof(u32)); - head = 0; + rq = NULL; + spin_lock_irqsave(&tl->lock, flags); + list_for_each_entry(pos, &tl->requests, link) { + if (!i915_request_completed(pos)) { + rq = pos; + break; + } } - memset32(vaddr + head, MI_NOOP, (rq->postfix - head) / sizeof(u32)); -} - -static void reset_ring(struct intel_engine_cs *engine, struct i915_request *rq) -{ - GEM_TRACE("%s request global=%d, current=%d\n", - engine->name, rq ? rq->global_seqno : 0, - intel_engine_get_seqno(engine)); + GEM_TRACE("%s seqno=%d, current=%d, stalled? %s\n", + engine->name, + rq ? rq->global_seqno : 0, + intel_engine_get_seqno(engine), + yesno(stalled)); /* - * Try to restore the logical GPU state to match the continuation - * of the request queue. If we skip the context/PD restore, then - * the next request may try to execute assuming that its context - * is valid and loaded on the GPU and so may try to access invalid - * memory, prompting repeated GPU hangs. + * The guilty request will get skipped on a hung engine. * - * If the request was guilty, we still restore the logical state - * in case the next request requires it (e.g. the aliasing ppgtt), - * but skip over the hung batch. + * Users of client default contexts do not rely on logical + * state preserved between batches so it is safe to execute + * queued requests following the hang. Non default contexts + * rely on preserved state, so skipping a batch loses the + * evolution of the state and it needs to be considered corrupted. + * Executing more queued batches on top of corrupted state is + * risky. But we take the risk by trying to advance through + * the queued requests in order to make the client behaviour + * more predictable around resets, by not throwing away random + * amount of batches it has prepared for execution. Sophisticated + * clients can use gem_reset_stats_ioctl and dma fence status + * (exported via sync_file info ioctl on explicit fences) to observe + * when it loses the context state and should rebuild accordingly. * - * If the request was innocent, we try to replay the request with - * the restored context. + * The context ban, and ultimately the client ban, mechanism are safety + * valves if client submission ends up resulting in nothing more than + * subsequent hangs. */ + if (rq) { - /* If the rq hung, jump to its breadcrumb and skip the batch */ - rq->ring->head = intel_ring_wrap(rq->ring, rq->head); - if (rq->fence.error == -EIO) - skip_request(rq); + /* + * Try to restore the logical GPU state to match the + * continuation of the request queue. If we skip the + * context/PD restore, then the next request may try to execute + * assuming that its context is valid and loaded on the GPU and + * so may try to access invalid memory, prompting repeated GPU + * hangs. + * + * If the request was guilty, we still restore the logical + * state in case the next request requires it (e.g. the + * aliasing ppgtt), but skip over the hung batch. + * + * If the request was innocent, we try to replay the request + * with the restored context. + */ + i915_reset_request(rq, stalled); + + GEM_BUG_ON(rq->ring != engine->buffer); + head = rq->head; + } else { + head = engine->buffer->tail; } + engine->buffer->head = intel_ring_wrap(engine->buffer, head); + + spin_unlock_irqrestore(&tl->lock, flags); } static void reset_finish(struct intel_engine_cs *engine) @@ -836,11 +894,10 @@ static void cancel_requests(struct intel_engine_cs *engine) list_for_each_entry(request, &engine->timeline.requests, link) { GEM_BUG_ON(!request->global_seqno); - if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, - &request->fence.flags)) - continue; + if (!i915_request_signaled(request)) + dma_fence_set_error(&request->fence, -EIO); - dma_fence_set_error(&request->fence, -EIO); + i915_request_mark_complete(request); } intel_write_status_page(engine, @@ -862,29 +919,43 @@ static void i9xx_submit_request(struct i915_request *request) intel_ring_set_tail(request->ring, request->tail)); } -static void i9xx_emit_breadcrumb(struct i915_request *rq, u32 *cs) +static u32 *i9xx_emit_breadcrumb(struct i915_request *rq, u32 *cs) { + GEM_BUG_ON(rq->timeline->hwsp_ggtt != rq->engine->status_page.vma); + GEM_BUG_ON(offset_in_page(rq->timeline->hwsp_offset) != I915_GEM_HWS_SEQNO_ADDR); + *cs++ = MI_FLUSH; *cs++ = MI_STORE_DWORD_INDEX; + *cs++ = I915_GEM_HWS_SEQNO_ADDR; + *cs++ = rq->fence.seqno; + + *cs++ = MI_STORE_DWORD_INDEX; *cs++ = I915_GEM_HWS_INDEX_ADDR; *cs++ = rq->global_seqno; *cs++ = MI_USER_INTERRUPT; - *cs++ = MI_NOOP; rq->tail = intel_ring_offset(rq, cs); assert_ring_tail_valid(rq->ring, rq->tail); + + return cs; } -static const int i9xx_emit_breadcrumb_sz = 6; #define GEN5_WA_STORES 8 /* must be at least 1! */ -static void gen5_emit_breadcrumb(struct i915_request *rq, u32 *cs) +static u32 *gen5_emit_breadcrumb(struct i915_request *rq, u32 *cs) { int i; + GEM_BUG_ON(rq->timeline->hwsp_ggtt != rq->engine->status_page.vma); + GEM_BUG_ON(offset_in_page(rq->timeline->hwsp_offset) != I915_GEM_HWS_SEQNO_ADDR); + *cs++ = MI_FLUSH; + *cs++ = MI_STORE_DWORD_INDEX; + *cs++ = I915_GEM_HWS_SEQNO_ADDR; + *cs++ = rq->fence.seqno; + BUILD_BUG_ON(GEN5_WA_STORES < 1); for (i = 0; i < GEN5_WA_STORES; i++) { *cs++ = MI_STORE_DWORD_INDEX; @@ -893,11 +964,13 @@ static void gen5_emit_breadcrumb(struct i915_request *rq, u32 *cs) } *cs++ = MI_USER_INTERRUPT; + *cs++ = MI_NOOP; rq->tail = intel_ring_offset(rq, cs); assert_ring_tail_valid(rq->ring, rq->tail); + + return cs; } -static const int gen5_emit_breadcrumb_sz = GEN5_WA_STORES * 3 + 2; #undef GEN5_WA_STORES static void @@ -1124,6 +1197,10 @@ int intel_ring_pin(struct intel_ring *ring) GEM_BUG_ON(ring->vaddr); + ret = i915_timeline_pin(ring->timeline); + if (ret) + return ret; + flags = PIN_GLOBAL; /* Ring wraparound at offset 0 sometimes hangs. No idea why. */ @@ -1140,28 +1217,32 @@ int intel_ring_pin(struct intel_ring *ring) else ret = i915_gem_object_set_to_cpu_domain(vma->obj, true); if (unlikely(ret)) - return ret; + goto unpin_timeline; } ret = i915_vma_pin(vma, 0, 0, flags); if (unlikely(ret)) - return ret; + goto unpin_timeline; if (i915_vma_is_map_and_fenceable(vma)) addr = (void __force *)i915_vma_pin_iomap(vma); else addr = i915_gem_object_pin_map(vma->obj, map); - if (IS_ERR(addr)) - goto err; + if (IS_ERR(addr)) { + ret = PTR_ERR(addr); + goto unpin_ring; + } vma->obj->pin_global++; ring->vaddr = addr; return 0; -err: +unpin_ring: i915_vma_unpin(vma); - return PTR_ERR(addr); +unpin_timeline: + i915_timeline_unpin(ring->timeline); + return ret; } void intel_ring_reset(struct intel_ring *ring, u32 tail) @@ -1190,6 +1271,8 @@ void intel_ring_unpin(struct intel_ring *ring) ring->vma->obj->pin_global--; i915_vma_unpin(ring->vma); + + i915_timeline_unpin(ring->timeline); } static struct i915_vma * @@ -1500,13 +1583,18 @@ static int intel_init_ring_buffer(struct intel_engine_cs *engine) struct intel_ring *ring; int err; - intel_engine_setup_common(engine); + err = intel_engine_setup_common(engine); + if (err) + return err; - timeline = i915_timeline_create(engine->i915, engine->name); + timeline = i915_timeline_create(engine->i915, + engine->name, + engine->status_page.vma); if (IS_ERR(timeline)) { err = PTR_ERR(timeline); goto err; } + GEM_BUG_ON(timeline->has_initial_breadcrumb); ring = intel_engine_create_ring(engine, timeline, 32 * PAGE_SIZE); i915_timeline_put(timeline); @@ -1526,6 +1614,8 @@ static int intel_init_ring_buffer(struct intel_engine_cs *engine) if (err) goto err_unpin; + GEM_BUG_ON(ring->timeline->hwsp_ggtt != engine->status_page.vma); + return 0; err_unpin: @@ -1858,6 +1948,7 @@ static int ring_request_alloc(struct i915_request *request) int ret; GEM_BUG_ON(!request->hw_context->pin_count); + GEM_BUG_ON(request->timeline->has_initial_breadcrumb); /* * Flush enough space to reduce the likelihood of waiting after @@ -2194,12 +2285,14 @@ static void intel_ring_default_vfuncs(struct drm_i915_private *dev_priv, engine->context_pin = intel_ring_context_pin; engine->request_alloc = ring_request_alloc; - engine->emit_breadcrumb = i9xx_emit_breadcrumb; - engine->emit_breadcrumb_sz = i9xx_emit_breadcrumb_sz; - if (IS_GEN(dev_priv, 5)) { - engine->emit_breadcrumb = gen5_emit_breadcrumb; - engine->emit_breadcrumb_sz = gen5_emit_breadcrumb_sz; - } + /* + * Using a global execution timeline; the previous final breadcrumb is + * equivalent to our next initial bread so we can elide + * engine->emit_init_breadcrumb(). + */ + engine->emit_fini_breadcrumb = i9xx_emit_breadcrumb; + if (IS_GEN(dev_priv, 5)) + engine->emit_fini_breadcrumb = gen5_emit_breadcrumb; engine->set_default_submission = i9xx_set_default_submission; @@ -2228,13 +2321,11 @@ int intel_init_render_ring_buffer(struct intel_engine_cs *engine) if (INTEL_GEN(dev_priv) >= 7) { engine->init_context = intel_rcs_ctx_init; engine->emit_flush = gen7_render_ring_flush; - engine->emit_breadcrumb = gen7_rcs_emit_breadcrumb; - engine->emit_breadcrumb_sz = gen7_rcs_emit_breadcrumb_sz; + engine->emit_fini_breadcrumb = gen7_rcs_emit_breadcrumb; } else if (IS_GEN(dev_priv, 6)) { engine->init_context = intel_rcs_ctx_init; engine->emit_flush = gen6_render_ring_flush; - engine->emit_breadcrumb = gen6_rcs_emit_breadcrumb; - engine->emit_breadcrumb_sz = gen6_rcs_emit_breadcrumb_sz; + engine->emit_fini_breadcrumb = gen6_rcs_emit_breadcrumb; } else if (IS_GEN(dev_priv, 5)) { engine->emit_flush = gen4_render_ring_flush; } else { @@ -2270,13 +2361,10 @@ int intel_init_bsd_ring_buffer(struct intel_engine_cs *engine) engine->emit_flush = gen6_bsd_ring_flush; engine->irq_enable_mask = GT_BSD_USER_INTERRUPT; - if (IS_GEN(dev_priv, 6)) { - engine->emit_breadcrumb = gen6_xcs_emit_breadcrumb; - engine->emit_breadcrumb_sz = gen6_xcs_emit_breadcrumb_sz; - } else { - engine->emit_breadcrumb = gen7_xcs_emit_breadcrumb; - engine->emit_breadcrumb_sz = gen7_xcs_emit_breadcrumb_sz; - } + if (IS_GEN(dev_priv, 6)) + engine->emit_fini_breadcrumb = gen6_xcs_emit_breadcrumb; + else + engine->emit_fini_breadcrumb = gen7_xcs_emit_breadcrumb; } else { engine->emit_flush = bsd_ring_flush; if (IS_GEN(dev_priv, 5)) @@ -2299,13 +2387,10 @@ int intel_init_blt_ring_buffer(struct intel_engine_cs *engine) engine->emit_flush = gen6_ring_flush; engine->irq_enable_mask = GT_BLT_USER_INTERRUPT; - if (IS_GEN(dev_priv, 6)) { - engine->emit_breadcrumb = gen6_xcs_emit_breadcrumb; - engine->emit_breadcrumb_sz = gen6_xcs_emit_breadcrumb_sz; - } else { - engine->emit_breadcrumb = gen7_xcs_emit_breadcrumb; - engine->emit_breadcrumb_sz = gen7_xcs_emit_breadcrumb_sz; - } + if (IS_GEN(dev_priv, 6)) + engine->emit_fini_breadcrumb = gen6_xcs_emit_breadcrumb; + else + engine->emit_fini_breadcrumb = gen7_xcs_emit_breadcrumb; return intel_init_ring_buffer(engine); } @@ -2323,8 +2408,7 @@ int intel_init_vebox_ring_buffer(struct intel_engine_cs *engine) engine->irq_enable = hsw_vebox_irq_enable; engine->irq_disable = hsw_vebox_irq_disable; - engine->emit_breadcrumb = gen7_xcs_emit_breadcrumb; - engine->emit_breadcrumb_sz = gen7_xcs_emit_breadcrumb_sz; + engine->emit_fini_breadcrumb = gen7_xcs_emit_breadcrumb; return intel_init_ring_buffer(engine); } diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 3c1366c58cf3..710ffb221775 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -5,6 +5,7 @@ #include <drm/drm_util.h> #include <linux/hashtable.h> +#include <linux/irq_work.h> #include <linux/seqlock.h> #include "i915_gem_batch_pool.h" @@ -28,12 +29,11 @@ struct i915_sched_attr; * workarounds! */ #define CACHELINE_BYTES 64 -#define CACHELINE_DWORDS (CACHELINE_BYTES / sizeof(uint32_t)) +#define CACHELINE_DWORDS (CACHELINE_BYTES / sizeof(u32)) struct intel_hw_status_page { struct i915_vma *vma; - u32 *page_addr; - u32 ggtt_offset; + u32 *addr; }; #define I915_READ_TAIL(engine) I915_READ(RING_TAIL((engine)->mmio_base)) @@ -120,13 +120,8 @@ struct intel_instdone { struct intel_engine_hangcheck { u64 acthd; u32 seqno; - enum intel_engine_hangcheck_action action; unsigned long action_timestamp; - int deadlock; struct intel_instdone instdone; - struct i915_request *active_request; - bool stalled:1; - bool wedged:1; }; struct intel_ring { @@ -209,6 +204,7 @@ struct i915_priolist { struct st_preempt_hang { struct completion completion; + unsigned int count; bool inject_hang; }; @@ -299,14 +295,18 @@ struct intel_engine_execlists { unsigned int port_mask; /** - * @queue_priority: Highest pending priority. + * @queue_priority_hint: Highest pending priority. * * When we add requests into the queue, or adjust the priority of * executing requests, we compute the maximum priority of those * pending requests. We can then use this value to determine if * we need to preempt the executing requests to service the queue. + * However, since the we may have recorded the priority of an inflight + * request we wanted to preempt but since completed, at the time of + * dequeuing the priority hint may no longer may match the highest + * available request priority. */ - int queue_priority; + int queue_priority_hint; /** * @queue: queue of requests, in priority lists @@ -382,23 +382,14 @@ struct intel_engine_cs { * the overhead of waking that client is much preferred. */ struct intel_breadcrumbs { - spinlock_t irq_lock; /* protects irq_*; irqsafe */ - struct intel_wait *irq_wait; /* oldest waiter by retirement */ + spinlock_t irq_lock; + struct list_head signalers; - spinlock_t rb_lock; /* protects the rb and wraps irq_lock */ - struct rb_root waiters; /* sorted by retirement, priority */ - struct list_head signals; /* sorted by retirement */ - struct task_struct *signaler; /* used for fence signalling */ + struct irq_work irq_work; /* for use from inside irq_lock */ - struct timer_list fake_irq; /* used after a missed interrupt */ - struct timer_list hangcheck; /* detect missed interrupts */ - - unsigned int hangcheck_interrupts; unsigned int irq_enabled; - unsigned int irq_count; - bool irq_armed : 1; - I915_SELFTEST_DECLARE(bool mock : 1); + bool irq_armed; } breadcrumbs; struct { @@ -412,16 +403,17 @@ struct intel_engine_cs { /** * @enable_count: Reference count for the enabled samplers. * - * Index number corresponds to the bit number from @enable. + * Index number corresponds to @enum drm_i915_pmu_engine_sample. */ - unsigned int enable_count[I915_PMU_SAMPLE_BITS]; + unsigned int enable_count[I915_ENGINE_SAMPLE_COUNT]; /** * @sample: Counter values for sampling events. * * Our internal timer stores the current counters in this field. + * + * Index number corresponds to @enum drm_i915_pmu_engine_sample. */ -#define I915_ENGINE_SAMPLE_MAX (I915_SAMPLE_SEMA + 1) - struct i915_pmu_sample sample[I915_ENGINE_SAMPLE_MAX]; + struct i915_pmu_sample sample[I915_ENGINE_SAMPLE_COUNT]; } pmu; /* @@ -445,9 +437,8 @@ struct intel_engine_cs { int (*init_hw)(struct intel_engine_cs *engine); struct { - struct i915_request *(*prepare)(struct intel_engine_cs *engine); - void (*reset)(struct intel_engine_cs *engine, - struct i915_request *rq); + void (*prepare)(struct intel_engine_cs *engine); + void (*reset)(struct intel_engine_cs *engine, bool stalled); void (*finish)(struct intel_engine_cs *engine); } reset; @@ -471,8 +462,10 @@ struct intel_engine_cs { unsigned int dispatch_flags); #define I915_DISPATCH_SECURE BIT(0) #define I915_DISPATCH_PINNED BIT(1) - void (*emit_breadcrumb)(struct i915_request *rq, u32 *cs); - int emit_breadcrumb_sz; + int (*emit_init_breadcrumb)(struct i915_request *rq); + u32 *(*emit_fini_breadcrumb)(struct i915_request *rq, + u32 *cs); + unsigned int emit_fini_breadcrumb_dw; /* Pass the request to the hardware queue (e.g. directly into * the legacy ringbuffer or to the end of an execlist). @@ -600,7 +593,20 @@ intel_engine_has_preemption(const struct intel_engine_cs *engine) static inline bool __execlists_need_preempt(int prio, int last) { - return prio > max(0, last); + /* + * Allow preemption of low -> normal -> high, but we do + * not allow low priority tasks to preempt other low priority + * tasks under the impression that latency for low priority + * tasks does not matter (as much as background throughput), + * so kiss. + * + * More naturally we would write + * prio >= max(0, last); + * except that we wish to prevent triggering preemption at the same + * priority level: the task that is running should remain running + * to preserve FIFO ordering of dependencies. + */ + return prio > max(I915_PRIORITY_NORMAL - 1, last); } static inline void @@ -678,7 +684,7 @@ static inline u32 intel_read_status_page(const struct intel_engine_cs *engine, int reg) { /* Ensure that the compiler doesn't optimize away the load. */ - return READ_ONCE(engine->status_page.page_addr[reg]); + return READ_ONCE(engine->status_page.addr[reg]); } static inline void @@ -691,12 +697,12 @@ intel_write_status_page(struct intel_engine_cs *engine, int reg, u32 value) */ if (static_cpu_has(X86_FEATURE_CLFLUSH)) { mb(); - clflush(&engine->status_page.page_addr[reg]); - engine->status_page.page_addr[reg] = value; - clflush(&engine->status_page.page_addr[reg]); + clflush(&engine->status_page.addr[reg]); + engine->status_page.addr[reg] = value; + clflush(&engine->status_page.addr[reg]); mb(); } else { - WRITE_ONCE(engine->status_page.page_addr[reg], value); + WRITE_ONCE(engine->status_page.addr[reg], value); } } @@ -717,11 +723,13 @@ intel_write_status_page(struct intel_engine_cs *engine, int reg, u32 value) * The area from dword 0x30 to 0x3ff is available for driver usage. */ #define I915_GEM_HWS_INDEX 0x30 -#define I915_GEM_HWS_INDEX_ADDR (I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT) -#define I915_GEM_HWS_PREEMPT_INDEX 0x32 -#define I915_GEM_HWS_PREEMPT_ADDR (I915_GEM_HWS_PREEMPT_INDEX << MI_STORE_DWORD_INDEX_SHIFT) -#define I915_GEM_HWS_SCRATCH_INDEX 0x40 -#define I915_GEM_HWS_SCRATCH_ADDR (I915_GEM_HWS_SCRATCH_INDEX << MI_STORE_DWORD_INDEX_SHIFT) +#define I915_GEM_HWS_INDEX_ADDR (I915_GEM_HWS_INDEX * sizeof(u32)) +#define I915_GEM_HWS_PREEMPT 0x32 +#define I915_GEM_HWS_PREEMPT_ADDR (I915_GEM_HWS_PREEMPT * sizeof(u32)) +#define I915_GEM_HWS_SEQNO 0x40 +#define I915_GEM_HWS_SEQNO_ADDR (I915_GEM_HWS_SEQNO * sizeof(u32)) +#define I915_GEM_HWS_SCRATCH 0x80 +#define I915_GEM_HWS_SCRATCH_ADDR (I915_GEM_HWS_SCRATCH * sizeof(u32)) #define I915_HWS_CSB_BUF0_INDEX 0x10 #define I915_HWS_CSB_WRITE_INDEX 0x1f @@ -824,9 +832,21 @@ intel_ring_set_tail(struct intel_ring *ring, unsigned int tail) return tail; } +static inline unsigned int +__intel_ring_space(unsigned int head, unsigned int tail, unsigned int size) +{ + /* + * "If the Ring Buffer Head Pointer and the Tail Pointer are on the + * same cacheline, the Head Pointer must not be greater than the Tail + * Pointer." + */ + GEM_BUG_ON(!is_power_of_2(size)); + return (head - tail - CACHELINE_BYTES) & (size - 1); +} + void intel_engine_write_global_seqno(struct intel_engine_cs *engine, u32 seqno); -void intel_engine_setup_common(struct intel_engine_cs *engine); +int intel_engine_setup_common(struct intel_engine_cs *engine); int intel_engine_init_common(struct intel_engine_cs *engine); void intel_engine_cleanup_common(struct intel_engine_cs *engine); @@ -884,93 +904,29 @@ static inline bool intel_engine_has_started(struct intel_engine_cs *engine, void intel_engine_get_instdone(struct intel_engine_cs *engine, struct intel_instdone *instdone); -static inline u32 intel_hws_seqno_address(struct intel_engine_cs *engine) -{ - return engine->status_page.ggtt_offset + I915_GEM_HWS_INDEX_ADDR; -} - -static inline u32 intel_hws_preempt_done_address(struct intel_engine_cs *engine) -{ - return engine->status_page.ggtt_offset + I915_GEM_HWS_PREEMPT_ADDR; -} - -/* intel_breadcrumbs.c -- user interrupt bottom-half for waiters */ -int intel_engine_init_breadcrumbs(struct intel_engine_cs *engine); - -static inline void intel_wait_init(struct intel_wait *wait) -{ - wait->tsk = current; - wait->request = NULL; -} - -static inline void intel_wait_init_for_seqno(struct intel_wait *wait, u32 seqno) -{ - wait->tsk = current; - wait->seqno = seqno; -} - -static inline bool intel_wait_has_seqno(const struct intel_wait *wait) -{ - return wait->seqno; -} - -static inline bool -intel_wait_update_seqno(struct intel_wait *wait, u32 seqno) -{ - wait->seqno = seqno; - return intel_wait_has_seqno(wait); -} - -static inline bool -intel_wait_update_request(struct intel_wait *wait, - const struct i915_request *rq) -{ - return intel_wait_update_seqno(wait, i915_request_global_seqno(rq)); -} - -static inline bool -intel_wait_check_seqno(const struct intel_wait *wait, u32 seqno) -{ - return wait->seqno == seqno; -} - -static inline bool -intel_wait_check_request(const struct intel_wait *wait, - const struct i915_request *rq) -{ - return intel_wait_check_seqno(wait, i915_request_global_seqno(rq)); -} +void intel_engine_init_breadcrumbs(struct intel_engine_cs *engine); +void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine); -static inline bool intel_wait_complete(const struct intel_wait *wait) -{ - return RB_EMPTY_NODE(&wait->node); -} +void intel_engine_pin_breadcrumbs_irq(struct intel_engine_cs *engine); +void intel_engine_unpin_breadcrumbs_irq(struct intel_engine_cs *engine); -bool intel_engine_add_wait(struct intel_engine_cs *engine, - struct intel_wait *wait); -void intel_engine_remove_wait(struct intel_engine_cs *engine, - struct intel_wait *wait); -bool intel_engine_enable_signaling(struct i915_request *request, bool wakeup); -void intel_engine_cancel_signaling(struct i915_request *request); +bool intel_engine_signal_breadcrumbs(struct intel_engine_cs *engine); +void intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine); -static inline bool intel_engine_has_waiter(const struct intel_engine_cs *engine) +static inline void +intel_engine_queue_breadcrumbs(struct intel_engine_cs *engine) { - return READ_ONCE(engine->breadcrumbs.irq_wait); + irq_work_queue(&engine->breadcrumbs.irq_work); } -unsigned int intel_engine_wakeup(struct intel_engine_cs *engine); -#define ENGINE_WAKEUP_WAITER BIT(0) -#define ENGINE_WAKEUP_ASLEEP BIT(1) - -void intel_engine_pin_breadcrumbs_irq(struct intel_engine_cs *engine); -void intel_engine_unpin_breadcrumbs_irq(struct intel_engine_cs *engine); - -void __intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine); -void intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine); +bool intel_engine_breadcrumbs_irq(struct intel_engine_cs *engine); void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine); void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine); +void intel_engine_print_breadcrumbs(struct intel_engine_cs *engine, + struct drm_printer *p); + static inline u32 *gen8_emit_pipe_control(u32 *batch, u32 flags, u32 offset) { memset(batch, 0, 6 * sizeof(u32)); @@ -1019,6 +975,13 @@ gen8_emit_ggtt_write(u32 *cs, u32 value, u32 gtt_offset) return cs; } +static inline void intel_engine_reset(struct intel_engine_cs *engine, + bool stalled) +{ + if (engine->reset.reset) + engine->reset.reset(engine, stalled); +} + void intel_engines_sanitize(struct drm_i915_private *i915, bool force); bool intel_engine_is_idle(struct intel_engine_cs *engine); diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c index 9e9501f82f06..a017a4232c0f 100644 --- a/drivers/gpu/drm/i915/intel_runtime_pm.c +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c @@ -29,6 +29,8 @@ #include <linux/pm_runtime.h> #include <linux/vgaarb.h> +#include <drm/drm_print.h> + #include "i915_drv.h" #include "intel_drv.h" @@ -49,6 +51,268 @@ * present for a given platform. */ +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) + +#include <linux/sort.h> + +#define STACKDEPTH 8 + +static noinline depot_stack_handle_t __save_depot_stack(void) +{ + unsigned long entries[STACKDEPTH]; + struct stack_trace trace = { + .entries = entries, + .max_entries = ARRAY_SIZE(entries), + .skip = 1, + }; + + save_stack_trace(&trace); + if (trace.nr_entries && + trace.entries[trace.nr_entries - 1] == ULONG_MAX) + trace.nr_entries--; + + return depot_save_stack(&trace, GFP_NOWAIT | __GFP_NOWARN); +} + +static void __print_depot_stack(depot_stack_handle_t stack, + char *buf, int sz, int indent) +{ + unsigned long entries[STACKDEPTH]; + struct stack_trace trace = { + .entries = entries, + .max_entries = ARRAY_SIZE(entries), + }; + + depot_fetch_stack(stack, &trace); + snprint_stack_trace(buf, sz, &trace, indent); +} + +static void init_intel_runtime_pm_wakeref(struct drm_i915_private *i915) +{ + struct i915_runtime_pm *rpm = &i915->runtime_pm; + + spin_lock_init(&rpm->debug.lock); +} + +static noinline depot_stack_handle_t +track_intel_runtime_pm_wakeref(struct drm_i915_private *i915) +{ + struct i915_runtime_pm *rpm = &i915->runtime_pm; + depot_stack_handle_t stack, *stacks; + unsigned long flags; + + atomic_inc(&rpm->wakeref_count); + assert_rpm_wakelock_held(i915); + + if (!HAS_RUNTIME_PM(i915)) + return -1; + + stack = __save_depot_stack(); + if (!stack) + return -1; + + spin_lock_irqsave(&rpm->debug.lock, flags); + + if (!rpm->debug.count) + rpm->debug.last_acquire = stack; + + stacks = krealloc(rpm->debug.owners, + (rpm->debug.count + 1) * sizeof(*stacks), + GFP_NOWAIT | __GFP_NOWARN); + if (stacks) { + stacks[rpm->debug.count++] = stack; + rpm->debug.owners = stacks; + } else { + stack = -1; + } + + spin_unlock_irqrestore(&rpm->debug.lock, flags); + + return stack; +} + +static void cancel_intel_runtime_pm_wakeref(struct drm_i915_private *i915, + depot_stack_handle_t stack) +{ + struct i915_runtime_pm *rpm = &i915->runtime_pm; + unsigned long flags, n; + bool found = false; + + if (unlikely(stack == -1)) + return; + + spin_lock_irqsave(&rpm->debug.lock, flags); + for (n = rpm->debug.count; n--; ) { + if (rpm->debug.owners[n] == stack) { + memmove(rpm->debug.owners + n, + rpm->debug.owners + n + 1, + (--rpm->debug.count - n) * sizeof(stack)); + found = true; + break; + } + } + spin_unlock_irqrestore(&rpm->debug.lock, flags); + + if (WARN(!found, + "Unmatched wakeref (tracking %lu), count %u\n", + rpm->debug.count, atomic_read(&rpm->wakeref_count))) { + char *buf; + + buf = kmalloc(PAGE_SIZE, GFP_KERNEL); + if (!buf) + return; + + __print_depot_stack(stack, buf, PAGE_SIZE, 2); + DRM_DEBUG_DRIVER("wakeref %x from\n%s", stack, buf); + + stack = READ_ONCE(rpm->debug.last_release); + if (stack) { + __print_depot_stack(stack, buf, PAGE_SIZE, 2); + DRM_DEBUG_DRIVER("wakeref last released at\n%s", buf); + } + + kfree(buf); + } +} + +static int cmphandle(const void *_a, const void *_b) +{ + const depot_stack_handle_t * const a = _a, * const b = _b; + + if (*a < *b) + return -1; + else if (*a > *b) + return 1; + else + return 0; +} + +static void +__print_intel_runtime_pm_wakeref(struct drm_printer *p, + const struct intel_runtime_pm_debug *dbg) +{ + unsigned long i; + char *buf; + + buf = kmalloc(PAGE_SIZE, GFP_KERNEL); + if (!buf) + return; + + if (dbg->last_acquire) { + __print_depot_stack(dbg->last_acquire, buf, PAGE_SIZE, 2); + drm_printf(p, "Wakeref last acquired:\n%s", buf); + } + + if (dbg->last_release) { + __print_depot_stack(dbg->last_release, buf, PAGE_SIZE, 2); + drm_printf(p, "Wakeref last released:\n%s", buf); + } + + drm_printf(p, "Wakeref count: %lu\n", dbg->count); + + sort(dbg->owners, dbg->count, sizeof(*dbg->owners), cmphandle, NULL); + + for (i = 0; i < dbg->count; i++) { + depot_stack_handle_t stack = dbg->owners[i]; + unsigned long rep; + + rep = 1; + while (i + 1 < dbg->count && dbg->owners[i + 1] == stack) + rep++, i++; + __print_depot_stack(stack, buf, PAGE_SIZE, 2); + drm_printf(p, "Wakeref x%lu taken at:\n%s", rep, buf); + } + + kfree(buf); +} + +static noinline void +untrack_intel_runtime_pm_wakeref(struct drm_i915_private *i915) +{ + struct i915_runtime_pm *rpm = &i915->runtime_pm; + struct intel_runtime_pm_debug dbg = {}; + struct drm_printer p; + unsigned long flags; + + assert_rpm_wakelock_held(i915); + if (atomic_dec_and_lock_irqsave(&rpm->wakeref_count, + &rpm->debug.lock, + flags)) { + dbg = rpm->debug; + + rpm->debug.owners = NULL; + rpm->debug.count = 0; + rpm->debug.last_release = __save_depot_stack(); + + spin_unlock_irqrestore(&rpm->debug.lock, flags); + } + if (!dbg.count) + return; + + p = drm_debug_printer("i915"); + __print_intel_runtime_pm_wakeref(&p, &dbg); + + kfree(dbg.owners); +} + +void print_intel_runtime_pm_wakeref(struct drm_i915_private *i915, + struct drm_printer *p) +{ + struct intel_runtime_pm_debug dbg = {}; + + do { + struct i915_runtime_pm *rpm = &i915->runtime_pm; + unsigned long alloc = dbg.count; + depot_stack_handle_t *s; + + spin_lock_irq(&rpm->debug.lock); + dbg.count = rpm->debug.count; + if (dbg.count <= alloc) { + memcpy(dbg.owners, + rpm->debug.owners, + dbg.count * sizeof(*s)); + } + dbg.last_acquire = rpm->debug.last_acquire; + dbg.last_release = rpm->debug.last_release; + spin_unlock_irq(&rpm->debug.lock); + if (dbg.count <= alloc) + break; + + s = krealloc(dbg.owners, dbg.count * sizeof(*s), GFP_KERNEL); + if (!s) + goto out; + + dbg.owners = s; + } while (1); + + __print_intel_runtime_pm_wakeref(p, &dbg); + +out: + kfree(dbg.owners); +} + +#else + +static void init_intel_runtime_pm_wakeref(struct drm_i915_private *i915) +{ +} + +static depot_stack_handle_t +track_intel_runtime_pm_wakeref(struct drm_i915_private *i915) +{ + atomic_inc(&i915->runtime_pm.wakeref_count); + assert_rpm_wakelock_held(i915); + return -1; +} + +static void untrack_intel_runtime_pm_wakeref(struct drm_i915_private *i915) +{ + assert_rpm_wakelock_held(i915); + atomic_dec(&i915->runtime_pm.wakeref_count); +} + +#endif + bool intel_display_power_well_is_enabled(struct drm_i915_private *dev_priv, enum i915_power_well_id power_well_id); @@ -639,10 +903,10 @@ void gen9_sanitize_dc_state(struct drm_i915_private *dev_priv) * back on and register state is restored. This is guaranteed by the MMIO write * to DC_STATE_EN blocking until the state is restored. */ -static void gen9_set_dc_state(struct drm_i915_private *dev_priv, uint32_t state) +static void gen9_set_dc_state(struct drm_i915_private *dev_priv, u32 state) { - uint32_t val; - uint32_t mask; + u32 val; + u32 mask; if (WARN_ON_ONCE(state & ~dev_priv->csr.allowed_dc_mask)) state &= dev_priv->csr.allowed_dc_mask; @@ -1274,7 +1538,7 @@ static void chv_dpio_cmn_power_well_enable(struct drm_i915_private *dev_priv, { enum dpio_phy phy; enum pipe pipe; - uint32_t tmp; + u32 tmp; WARN_ON_ONCE(power_well->desc->id != VLV_DISP_PW_DPIO_CMN_BC && power_well->desc->id != CHV_DISP_PW_DPIO_CMN_D); @@ -1591,18 +1855,19 @@ __intel_display_power_get_domain(struct drm_i915_private *dev_priv, * Any power domain reference obtained by this function must have a symmetric * call to intel_display_power_put() to release the reference again. */ -void intel_display_power_get(struct drm_i915_private *dev_priv, - enum intel_display_power_domain domain) +intel_wakeref_t intel_display_power_get(struct drm_i915_private *dev_priv, + enum intel_display_power_domain domain) { struct i915_power_domains *power_domains = &dev_priv->power_domains; - - intel_runtime_pm_get(dev_priv); + intel_wakeref_t wakeref = intel_runtime_pm_get(dev_priv); mutex_lock(&power_domains->lock); __intel_display_power_get_domain(dev_priv, domain); mutex_unlock(&power_domains->lock); + + return wakeref; } /** @@ -1617,13 +1882,16 @@ void intel_display_power_get(struct drm_i915_private *dev_priv, * Any power domain reference obtained by this function must have a symmetric * call to intel_display_power_put() to release the reference again. */ -bool intel_display_power_get_if_enabled(struct drm_i915_private *dev_priv, - enum intel_display_power_domain domain) +intel_wakeref_t +intel_display_power_get_if_enabled(struct drm_i915_private *dev_priv, + enum intel_display_power_domain domain) { struct i915_power_domains *power_domains = &dev_priv->power_domains; + intel_wakeref_t wakeref; bool is_enabled; - if (!intel_runtime_pm_get_if_in_use(dev_priv)) + wakeref = intel_runtime_pm_get_if_in_use(dev_priv); + if (!wakeref) return false; mutex_lock(&power_domains->lock); @@ -1637,23 +1905,16 @@ bool intel_display_power_get_if_enabled(struct drm_i915_private *dev_priv, mutex_unlock(&power_domains->lock); - if (!is_enabled) - intel_runtime_pm_put(dev_priv); + if (!is_enabled) { + intel_runtime_pm_put(dev_priv, wakeref); + wakeref = 0; + } - return is_enabled; + return wakeref; } -/** - * intel_display_power_put - release a power domain reference - * @dev_priv: i915 device instance - * @domain: power domain to reference - * - * This function drops the power domain reference obtained by - * intel_display_power_get() and might power down the corresponding hardware - * block right away if this is the last reference. - */ -void intel_display_power_put(struct drm_i915_private *dev_priv, - enum intel_display_power_domain domain) +static void __intel_display_power_put(struct drm_i915_private *dev_priv, + enum intel_display_power_domain domain) { struct i915_power_domains *power_domains; struct i915_power_well *power_well; @@ -1671,9 +1932,33 @@ void intel_display_power_put(struct drm_i915_private *dev_priv, intel_power_well_put(dev_priv, power_well); mutex_unlock(&power_domains->lock); +} + +/** + * intel_display_power_put - release a power domain reference + * @dev_priv: i915 device instance + * @domain: power domain to reference + * + * This function drops the power domain reference obtained by + * intel_display_power_get() and might power down the corresponding hardware + * block right away if this is the last reference. + */ +void intel_display_power_put_unchecked(struct drm_i915_private *dev_priv, + enum intel_display_power_domain domain) +{ + __intel_display_power_put(dev_priv, domain); + intel_runtime_pm_put_unchecked(dev_priv); +} - intel_runtime_pm_put(dev_priv); +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) +void intel_display_power_put(struct drm_i915_private *dev_priv, + enum intel_display_power_domain domain, + intel_wakeref_t wakeref) +{ + __intel_display_power_put(dev_priv, domain); + intel_runtime_pm_put(dev_priv, wakeref); } +#endif #define I830_PIPES_POWER_DOMAINS ( \ BIT_ULL(POWER_DOMAIN_PIPE_A) | \ @@ -3043,10 +3328,10 @@ sanitize_disable_power_well_option(const struct drm_i915_private *dev_priv, return 1; } -static uint32_t get_allowed_dc_mask(const struct drm_i915_private *dev_priv, - int enable_dc) +static u32 get_allowed_dc_mask(const struct drm_i915_private *dev_priv, + int enable_dc) { - uint32_t mask; + u32 mask; int requested_dc; int max_dc; @@ -3311,7 +3596,7 @@ static void icl_dbuf_disable(struct drm_i915_private *dev_priv) static void icl_mbus_init(struct drm_i915_private *dev_priv) { - uint32_t val; + u32 val; val = MBUS_ABOX_BT_CREDIT_POOL1(16) | MBUS_ABOX_BT_CREDIT_POOL2(16) | @@ -3622,7 +3907,7 @@ static void chv_phy_control_init(struct drm_i915_private *dev_priv) * current lane status. */ if (cmn_bc->desc->ops->is_enabled(dev_priv, cmn_bc)) { - uint32_t status = I915_READ(DPLL(PIPE_A)); + u32 status = I915_READ(DPLL(PIPE_A)); unsigned int mask; mask = status & DPLL_PORTB_READY_MASK; @@ -3653,7 +3938,7 @@ static void chv_phy_control_init(struct drm_i915_private *dev_priv) } if (cmn_d->desc->ops->is_enabled(dev_priv, cmn_d)) { - uint32_t status = I915_READ(DPIO_PHY_STATUS); + u32 status = I915_READ(DPIO_PHY_STATUS); unsigned int mask; mask = status & DPLL_PORTD_READY_MASK; @@ -3712,7 +3997,7 @@ static void intel_power_domains_verify_state(struct drm_i915_private *dev_priv); /** * intel_power_domains_init_hw - initialize hardware power domain state - * @dev_priv: i915 device instance + * @i915: i915 device instance * @resume: Called from resume code paths or not * * This function initializes the hardware power domain state and enables all @@ -3726,30 +4011,31 @@ static void intel_power_domains_verify_state(struct drm_i915_private *dev_priv); * intel_power_domains_enable()) and must be paired with * intel_power_domains_fini_hw(). */ -void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume) +void intel_power_domains_init_hw(struct drm_i915_private *i915, bool resume) { - struct i915_power_domains *power_domains = &dev_priv->power_domains; + struct i915_power_domains *power_domains = &i915->power_domains; power_domains->initializing = true; - if (IS_ICELAKE(dev_priv)) { - icl_display_core_init(dev_priv, resume); - } else if (IS_CANNONLAKE(dev_priv)) { - cnl_display_core_init(dev_priv, resume); - } else if (IS_GEN9_BC(dev_priv)) { - skl_display_core_init(dev_priv, resume); - } else if (IS_GEN9_LP(dev_priv)) { - bxt_display_core_init(dev_priv, resume); - } else if (IS_CHERRYVIEW(dev_priv)) { + if (IS_ICELAKE(i915)) { + icl_display_core_init(i915, resume); + } else if (IS_CANNONLAKE(i915)) { + cnl_display_core_init(i915, resume); + } else if (IS_GEN9_BC(i915)) { + skl_display_core_init(i915, resume); + } else if (IS_GEN9_LP(i915)) { + bxt_display_core_init(i915, resume); + } else if (IS_CHERRYVIEW(i915)) { mutex_lock(&power_domains->lock); - chv_phy_control_init(dev_priv); + chv_phy_control_init(i915); mutex_unlock(&power_domains->lock); - } else if (IS_VALLEYVIEW(dev_priv)) { + } else if (IS_VALLEYVIEW(i915)) { mutex_lock(&power_domains->lock); - vlv_cmnlane_wa(dev_priv); + vlv_cmnlane_wa(i915); mutex_unlock(&power_domains->lock); - } else if (IS_IVYBRIDGE(dev_priv) || INTEL_GEN(dev_priv) >= 7) - intel_pch_reset_handshake(dev_priv, !HAS_PCH_NOP(dev_priv)); + } else if (IS_IVYBRIDGE(i915) || INTEL_GEN(i915) >= 7) { + intel_pch_reset_handshake(i915, !HAS_PCH_NOP(i915)); + } /* * Keep all power wells enabled for any dependent HW access during @@ -3757,18 +4043,20 @@ void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume) * resources powered until display HW readout is complete. We drop * this reference in intel_power_domains_enable(). */ - intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); + power_domains->wakeref = + intel_display_power_get(i915, POWER_DOMAIN_INIT); + /* Disable power support if the user asked so. */ if (!i915_modparams.disable_power_well) - intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); - intel_power_domains_sync_hw(dev_priv); + intel_display_power_get(i915, POWER_DOMAIN_INIT); + intel_power_domains_sync_hw(i915); power_domains->initializing = false; } /** * intel_power_domains_fini_hw - deinitialize hw power domain state - * @dev_priv: i915 device instance + * @i915: i915 device instance * * De-initializes the display power domain HW state. It also ensures that the * device stays powered up so that the driver can be reloaded. @@ -3777,21 +4065,24 @@ void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume) * intel_power_domains_disable()) and must be paired with * intel_power_domains_init_hw(). */ -void intel_power_domains_fini_hw(struct drm_i915_private *dev_priv) +void intel_power_domains_fini_hw(struct drm_i915_private *i915) { - /* Keep the power well enabled, but cancel its rpm wakeref. */ - intel_runtime_pm_put(dev_priv); + intel_wakeref_t wakeref __maybe_unused = + fetch_and_zero(&i915->power_domains.wakeref); /* Remove the refcount we took to keep power well support disabled. */ if (!i915_modparams.disable_power_well) - intel_display_power_put(dev_priv, POWER_DOMAIN_INIT); + intel_display_power_put_unchecked(i915, POWER_DOMAIN_INIT); + + intel_power_domains_verify_state(i915); - intel_power_domains_verify_state(dev_priv); + /* Keep the power well enabled, but cancel its rpm wakeref. */ + intel_runtime_pm_put(i915, wakeref); } /** * intel_power_domains_enable - enable toggling of display power wells - * @dev_priv: i915 device instance + * @i915: i915 device instance * * Enable the ondemand enabling/disabling of the display power wells. Note that * power wells not belonging to POWER_DOMAIN_INIT are allowed to be toggled @@ -3801,30 +4092,36 @@ void intel_power_domains_fini_hw(struct drm_i915_private *dev_priv) * of display HW readout (which will acquire the power references reflecting * the current HW state). */ -void intel_power_domains_enable(struct drm_i915_private *dev_priv) +void intel_power_domains_enable(struct drm_i915_private *i915) { - intel_display_power_put(dev_priv, POWER_DOMAIN_INIT); + intel_wakeref_t wakeref __maybe_unused = + fetch_and_zero(&i915->power_domains.wakeref); - intel_power_domains_verify_state(dev_priv); + intel_display_power_put(i915, POWER_DOMAIN_INIT, wakeref); + intel_power_domains_verify_state(i915); } /** * intel_power_domains_disable - disable toggling of display power wells - * @dev_priv: i915 device instance + * @i915: i915 device instance * * Disable the ondemand enabling/disabling of the display power wells. See * intel_power_domains_enable() for which power wells this call controls. */ -void intel_power_domains_disable(struct drm_i915_private *dev_priv) +void intel_power_domains_disable(struct drm_i915_private *i915) { - intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); + struct i915_power_domains *power_domains = &i915->power_domains; - intel_power_domains_verify_state(dev_priv); + WARN_ON(power_domains->wakeref); + power_domains->wakeref = + intel_display_power_get(i915, POWER_DOMAIN_INIT); + + intel_power_domains_verify_state(i915); } /** * intel_power_domains_suspend - suspend power domain state - * @dev_priv: i915 device instance + * @i915: i915 device instance * @suspend_mode: specifies the target suspend state (idle, mem, hibernation) * * This function prepares the hardware power domain state before entering @@ -3833,12 +4130,14 @@ void intel_power_domains_disable(struct drm_i915_private *dev_priv) * It must be called with power domains already disabled (after a call to * intel_power_domains_disable()) and paired with intel_power_domains_resume(). */ -void intel_power_domains_suspend(struct drm_i915_private *dev_priv, +void intel_power_domains_suspend(struct drm_i915_private *i915, enum i915_drm_suspend_mode suspend_mode) { - struct i915_power_domains *power_domains = &dev_priv->power_domains; + struct i915_power_domains *power_domains = &i915->power_domains; + intel_wakeref_t wakeref __maybe_unused = + fetch_and_zero(&power_domains->wakeref); - intel_display_power_put(dev_priv, POWER_DOMAIN_INIT); + intel_display_power_put(i915, POWER_DOMAIN_INIT, wakeref); /* * In case of suspend-to-idle (aka S0ix) on a DMC platform without DC9 @@ -3847,10 +4146,10 @@ void intel_power_domains_suspend(struct drm_i915_private *dev_priv, * resources as required and also enable deeper system power states * that would be blocked if the firmware was inactive. */ - if (!(dev_priv->csr.allowed_dc_mask & DC_STATE_EN_DC9) && + if (!(i915->csr.allowed_dc_mask & DC_STATE_EN_DC9) && suspend_mode == I915_DRM_SUSPEND_IDLE && - dev_priv->csr.dmc_payload != NULL) { - intel_power_domains_verify_state(dev_priv); + i915->csr.dmc_payload) { + intel_power_domains_verify_state(i915); return; } @@ -3859,25 +4158,25 @@ void intel_power_domains_suspend(struct drm_i915_private *dev_priv, * power wells if power domains must be deinitialized for suspend. */ if (!i915_modparams.disable_power_well) { - intel_display_power_put(dev_priv, POWER_DOMAIN_INIT); - intel_power_domains_verify_state(dev_priv); + intel_display_power_put_unchecked(i915, POWER_DOMAIN_INIT); + intel_power_domains_verify_state(i915); } - if (IS_ICELAKE(dev_priv)) - icl_display_core_uninit(dev_priv); - else if (IS_CANNONLAKE(dev_priv)) - cnl_display_core_uninit(dev_priv); - else if (IS_GEN9_BC(dev_priv)) - skl_display_core_uninit(dev_priv); - else if (IS_GEN9_LP(dev_priv)) - bxt_display_core_uninit(dev_priv); + if (IS_ICELAKE(i915)) + icl_display_core_uninit(i915); + else if (IS_CANNONLAKE(i915)) + cnl_display_core_uninit(i915); + else if (IS_GEN9_BC(i915)) + skl_display_core_uninit(i915); + else if (IS_GEN9_LP(i915)) + bxt_display_core_uninit(i915); power_domains->display_core_suspended = true; } /** * intel_power_domains_resume - resume power domain state - * @dev_priv: i915 device instance + * @i915: i915 device instance * * This function resume the hardware power domain state during system resume. * @@ -3885,28 +4184,30 @@ void intel_power_domains_suspend(struct drm_i915_private *dev_priv, * intel_power_domains_enable()) and must be paired with * intel_power_domains_suspend(). */ -void intel_power_domains_resume(struct drm_i915_private *dev_priv) +void intel_power_domains_resume(struct drm_i915_private *i915) { - struct i915_power_domains *power_domains = &dev_priv->power_domains; + struct i915_power_domains *power_domains = &i915->power_domains; if (power_domains->display_core_suspended) { - intel_power_domains_init_hw(dev_priv, true); + intel_power_domains_init_hw(i915, true); power_domains->display_core_suspended = false; } else { - intel_display_power_get(dev_priv, POWER_DOMAIN_INIT); + WARN_ON(power_domains->wakeref); + power_domains->wakeref = + intel_display_power_get(i915, POWER_DOMAIN_INIT); } - intel_power_domains_verify_state(dev_priv); + intel_power_domains_verify_state(i915); } #if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) -static void intel_power_domains_dump_info(struct drm_i915_private *dev_priv) +static void intel_power_domains_dump_info(struct drm_i915_private *i915) { - struct i915_power_domains *power_domains = &dev_priv->power_domains; + struct i915_power_domains *power_domains = &i915->power_domains; struct i915_power_well *power_well; - for_each_power_well(dev_priv, power_well) { + for_each_power_well(i915, power_well) { enum intel_display_power_domain domain; DRM_DEBUG_DRIVER("%-25s %d\n", @@ -3921,7 +4222,7 @@ static void intel_power_domains_dump_info(struct drm_i915_private *dev_priv) /** * intel_power_domains_verify_state - verify the HW/SW state for all power wells - * @dev_priv: i915 device instance + * @i915: i915 device instance * * Verify if the reference count of each power well matches its HW enabled * state and the total refcount of the domains it belongs to. This must be @@ -3929,22 +4230,21 @@ static void intel_power_domains_dump_info(struct drm_i915_private *dev_priv) * acquiring reference counts for any power wells in use and disabling the * ones left on by BIOS but not required by any active output. */ -static void intel_power_domains_verify_state(struct drm_i915_private *dev_priv) +static void intel_power_domains_verify_state(struct drm_i915_private *i915) { - struct i915_power_domains *power_domains = &dev_priv->power_domains; + struct i915_power_domains *power_domains = &i915->power_domains; struct i915_power_well *power_well; bool dump_domain_info; mutex_lock(&power_domains->lock); dump_domain_info = false; - for_each_power_well(dev_priv, power_well) { + for_each_power_well(i915, power_well) { enum intel_display_power_domain domain; int domains_count; bool enabled; - enabled = power_well->desc->ops->is_enabled(dev_priv, - power_well); + enabled = power_well->desc->ops->is_enabled(i915, power_well); if ((power_well->count || power_well->desc->always_on) != enabled) DRM_ERROR("power well %s state mismatch (refcount %d/enabled %d)", @@ -3968,7 +4268,7 @@ static void intel_power_domains_verify_state(struct drm_i915_private *dev_priv) static bool dumped; if (!dumped) { - intel_power_domains_dump_info(dev_priv); + intel_power_domains_dump_info(i915); dumped = true; } } @@ -3978,7 +4278,7 @@ static void intel_power_domains_verify_state(struct drm_i915_private *dev_priv) #else -static void intel_power_domains_verify_state(struct drm_i915_private *dev_priv) +static void intel_power_domains_verify_state(struct drm_i915_private *i915) { } @@ -3986,30 +4286,31 @@ static void intel_power_domains_verify_state(struct drm_i915_private *dev_priv) /** * intel_runtime_pm_get - grab a runtime pm reference - * @dev_priv: i915 device instance + * @i915: i915 device instance * * This function grabs a device-level runtime pm reference (mostly used for GEM * code to ensure the GTT or GT is on) and ensures that it is powered up. * * Any runtime pm reference obtained by this function must have a symmetric * call to intel_runtime_pm_put() to release the reference again. + * + * Returns: the wakeref cookie to pass to intel_runtime_pm_put() */ -void intel_runtime_pm_get(struct drm_i915_private *dev_priv) +intel_wakeref_t intel_runtime_pm_get(struct drm_i915_private *i915) { - struct pci_dev *pdev = dev_priv->drm.pdev; + struct pci_dev *pdev = i915->drm.pdev; struct device *kdev = &pdev->dev; int ret; ret = pm_runtime_get_sync(kdev); WARN_ONCE(ret < 0, "pm_runtime_get_sync() failed: %d\n", ret); - atomic_inc(&dev_priv->runtime_pm.wakeref_count); - assert_rpm_wakelock_held(dev_priv); + return track_intel_runtime_pm_wakeref(i915); } /** * intel_runtime_pm_get_if_in_use - grab a runtime pm reference if device in use - * @dev_priv: i915 device instance + * @i915: i915 device instance * * This function grabs a device-level runtime pm reference if the device is * already in use and ensures that it is powered up. It is illegal to try @@ -4018,12 +4319,13 @@ void intel_runtime_pm_get(struct drm_i915_private *dev_priv) * Any runtime pm reference obtained by this function must have a symmetric * call to intel_runtime_pm_put() to release the reference again. * - * Returns: True if the wakeref was acquired, or False otherwise. + * Returns: the wakeref cookie to pass to intel_runtime_pm_put(), evaluates + * as True if the wakeref was acquired, or False otherwise. */ -bool intel_runtime_pm_get_if_in_use(struct drm_i915_private *dev_priv) +intel_wakeref_t intel_runtime_pm_get_if_in_use(struct drm_i915_private *i915) { if (IS_ENABLED(CONFIG_PM)) { - struct pci_dev *pdev = dev_priv->drm.pdev; + struct pci_dev *pdev = i915->drm.pdev; struct device *kdev = &pdev->dev; /* @@ -4033,18 +4335,15 @@ bool intel_runtime_pm_get_if_in_use(struct drm_i915_private *dev_priv) * atm to the late/early system suspend/resume handlers. */ if (pm_runtime_get_if_in_use(kdev) <= 0) - return false; + return 0; } - atomic_inc(&dev_priv->runtime_pm.wakeref_count); - assert_rpm_wakelock_held(dev_priv); - - return true; + return track_intel_runtime_pm_wakeref(i915); } /** * intel_runtime_pm_get_noresume - grab a runtime pm reference - * @dev_priv: i915 device instance + * @i915: i915 device instance * * This function grabs a device-level runtime pm reference (mostly used for GEM * code to ensure the GTT or GT is on). @@ -4058,41 +4357,50 @@ bool intel_runtime_pm_get_if_in_use(struct drm_i915_private *dev_priv) * * Any runtime pm reference obtained by this function must have a symmetric * call to intel_runtime_pm_put() to release the reference again. + * + * Returns: the wakeref cookie to pass to intel_runtime_pm_put() */ -void intel_runtime_pm_get_noresume(struct drm_i915_private *dev_priv) +intel_wakeref_t intel_runtime_pm_get_noresume(struct drm_i915_private *i915) { - struct pci_dev *pdev = dev_priv->drm.pdev; + struct pci_dev *pdev = i915->drm.pdev; struct device *kdev = &pdev->dev; - assert_rpm_wakelock_held(dev_priv); + assert_rpm_wakelock_held(i915); pm_runtime_get_noresume(kdev); - atomic_inc(&dev_priv->runtime_pm.wakeref_count); + return track_intel_runtime_pm_wakeref(i915); } /** * intel_runtime_pm_put - release a runtime pm reference - * @dev_priv: i915 device instance + * @i915: i915 device instance * * This function drops the device-level runtime pm reference obtained by * intel_runtime_pm_get() and might power down the corresponding * hardware block right away if this is the last reference. */ -void intel_runtime_pm_put(struct drm_i915_private *dev_priv) +void intel_runtime_pm_put_unchecked(struct drm_i915_private *i915) { - struct pci_dev *pdev = dev_priv->drm.pdev; + struct pci_dev *pdev = i915->drm.pdev; struct device *kdev = &pdev->dev; - assert_rpm_wakelock_held(dev_priv); - atomic_dec(&dev_priv->runtime_pm.wakeref_count); + untrack_intel_runtime_pm_wakeref(i915); pm_runtime_mark_last_busy(kdev); pm_runtime_put_autosuspend(kdev); } +#if IS_ENABLED(CONFIG_DRM_I915_DEBUG_RUNTIME_PM) +void intel_runtime_pm_put(struct drm_i915_private *i915, intel_wakeref_t wref) +{ + cancel_intel_runtime_pm_wakeref(i915, wref); + intel_runtime_pm_put_unchecked(i915); +} +#endif + /** * intel_runtime_pm_enable - enable runtime pm - * @dev_priv: i915 device instance + * @i915: i915 device instance * * This function enables runtime pm at the end of the driver load sequence. * @@ -4100,9 +4408,9 @@ void intel_runtime_pm_put(struct drm_i915_private *dev_priv) * subordinate display power domains. That is done by * intel_power_domains_enable(). */ -void intel_runtime_pm_enable(struct drm_i915_private *dev_priv) +void intel_runtime_pm_enable(struct drm_i915_private *i915) { - struct pci_dev *pdev = dev_priv->drm.pdev; + struct pci_dev *pdev = i915->drm.pdev; struct device *kdev = &pdev->dev; /* @@ -4124,7 +4432,7 @@ void intel_runtime_pm_enable(struct drm_i915_private *dev_priv) * so the driver's own RPM reference tracking asserts also work on * platforms without RPM support. */ - if (!HAS_RUNTIME_PM(dev_priv)) { + if (!HAS_RUNTIME_PM(i915)) { int ret; pm_runtime_dont_use_autosuspend(kdev); @@ -4142,17 +4450,35 @@ void intel_runtime_pm_enable(struct drm_i915_private *dev_priv) pm_runtime_put_autosuspend(kdev); } -void intel_runtime_pm_disable(struct drm_i915_private *dev_priv) +void intel_runtime_pm_disable(struct drm_i915_private *i915) { - struct pci_dev *pdev = dev_priv->drm.pdev; + struct pci_dev *pdev = i915->drm.pdev; struct device *kdev = &pdev->dev; /* Transfer rpm ownership back to core */ - WARN(pm_runtime_get_sync(&dev_priv->drm.pdev->dev) < 0, + WARN(pm_runtime_get_sync(kdev) < 0, "Failed to pass rpm ownership back to core\n"); pm_runtime_dont_use_autosuspend(kdev); - if (!HAS_RUNTIME_PM(dev_priv)) + if (!HAS_RUNTIME_PM(i915)) pm_runtime_put(kdev); } + +void intel_runtime_pm_cleanup(struct drm_i915_private *i915) +{ + struct i915_runtime_pm *rpm = &i915->runtime_pm; + int count; + + count = atomic_fetch_inc(&rpm->wakeref_count); /* balance untrack */ + WARN(count, + "i915->runtime_pm.wakeref_count=%d on cleanup\n", + count); + + untrack_intel_runtime_pm_wakeref(i915); +} + +void intel_runtime_pm_init_early(struct drm_i915_private *i915) +{ + init_intel_runtime_pm_wakeref(i915); +} diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c index df2d830a7405..e7b0884ba5a5 100644 --- a/drivers/gpu/drm/i915/intel_sdvo.c +++ b/drivers/gpu/drm/i915/intel_sdvo.c @@ -76,7 +76,7 @@ struct intel_sdvo { i915_reg_t sdvo_reg; /* Active outputs controlled by this SDVO output */ - uint16_t controlled_output; + u16 controlled_output; /* * Capabilities of the SDVO device returned by @@ -91,12 +91,12 @@ struct intel_sdvo { * For multiple function SDVO device, * this is for current attached outputs. */ - uint16_t attached_output; + u16 attached_output; /* * Hotplug activation bits for this device */ - uint16_t hotplug_active; + u16 hotplug_active; enum port port; @@ -104,19 +104,19 @@ struct intel_sdvo { bool has_hdmi_audio; /* DDC bus used by this SDVO encoder */ - uint8_t ddc_bus; + u8 ddc_bus; /* * the sdvo flag gets lost in round trip: dtd->adjusted_mode->dtd */ - uint8_t dtd_sdvo_flags; + u8 dtd_sdvo_flags; }; struct intel_sdvo_connector { struct intel_connector base; /* Mark the type of connector */ - uint16_t output_flag; + u16 output_flag; /* This contains all current supported TV format */ u8 tv_format_supported[TV_FORMAT_NUM]; @@ -184,7 +184,7 @@ to_intel_sdvo_connector(struct drm_connector *connector) container_of((conn_state), struct intel_sdvo_connector_state, base.base) static bool -intel_sdvo_output_setup(struct intel_sdvo *intel_sdvo, uint16_t flags); +intel_sdvo_output_setup(struct intel_sdvo *intel_sdvo, u16 flags); static bool intel_sdvo_tv_create_property(struct intel_sdvo *intel_sdvo, struct intel_sdvo_connector *intel_sdvo_connector, @@ -746,9 +746,9 @@ static bool intel_sdvo_get_input_timing(struct intel_sdvo *intel_sdvo, static bool intel_sdvo_create_preferred_input_timing(struct intel_sdvo *intel_sdvo, struct intel_sdvo_connector *intel_sdvo_connector, - uint16_t clock, - uint16_t width, - uint16_t height) + u16 clock, + u16 width, + u16 height) { struct intel_sdvo_preferred_input_timing_args args; @@ -791,9 +791,9 @@ static bool intel_sdvo_set_clock_rate_mult(struct intel_sdvo *intel_sdvo, u8 val static void intel_sdvo_get_dtd_from_mode(struct intel_sdvo_dtd *dtd, const struct drm_display_mode *mode) { - uint16_t width, height; - uint16_t h_blank_len, h_sync_len, v_blank_len, v_sync_len; - uint16_t h_sync_offset, v_sync_offset; + u16 width, height; + u16 h_blank_len, h_sync_len, v_blank_len, v_sync_len; + u16 h_sync_offset, v_sync_offset; int mode_clock; memset(dtd, 0, sizeof(*dtd)); @@ -898,13 +898,13 @@ static bool intel_sdvo_check_supp_encode(struct intel_sdvo *intel_sdvo) } static bool intel_sdvo_set_encode(struct intel_sdvo *intel_sdvo, - uint8_t mode) + u8 mode) { return intel_sdvo_set_value(intel_sdvo, SDVO_CMD_SET_ENCODE, &mode, 1); } static bool intel_sdvo_set_colorimetry(struct intel_sdvo *intel_sdvo, - uint8_t mode) + u8 mode) { return intel_sdvo_set_value(intel_sdvo, SDVO_CMD_SET_COLORIMETRY, &mode, 1); } @@ -913,11 +913,11 @@ static bool intel_sdvo_set_colorimetry(struct intel_sdvo *intel_sdvo, static void intel_sdvo_dump_hdmi_buf(struct intel_sdvo *intel_sdvo) { int i, j; - uint8_t set_buf_index[2]; - uint8_t av_split; - uint8_t buf_size; - uint8_t buf[48]; - uint8_t *pos; + u8 set_buf_index[2]; + u8 av_split; + u8 buf_size; + u8 buf[48]; + u8 *pos; intel_sdvo_get_value(encoder, SDVO_CMD_GET_HBUF_AV_SPLIT, &av_split, 1); @@ -940,11 +940,11 @@ static void intel_sdvo_dump_hdmi_buf(struct intel_sdvo *intel_sdvo) #endif static bool intel_sdvo_write_infoframe(struct intel_sdvo *intel_sdvo, - unsigned if_index, uint8_t tx_rate, - const uint8_t *data, unsigned length) + unsigned int if_index, u8 tx_rate, + const u8 *data, unsigned int length) { - uint8_t set_buf_index[2] = { if_index, 0 }; - uint8_t hbuf_size, tmp[8]; + u8 set_buf_index[2] = { if_index, 0 }; + u8 hbuf_size, tmp[8]; int i; if (!intel_sdvo_set_value(intel_sdvo, @@ -984,7 +984,7 @@ static bool intel_sdvo_set_avi_infoframe(struct intel_sdvo *intel_sdvo, { const struct drm_display_mode *adjusted_mode = &pipe_config->base.adjusted_mode; - uint8_t sdvo_data[HDMI_INFOFRAME_SIZE(AVI)]; + u8 sdvo_data[HDMI_INFOFRAME_SIZE(AVI)]; union hdmi_infoframe frame; int ret; ssize_t len; @@ -1017,7 +1017,7 @@ static bool intel_sdvo_set_tv_format(struct intel_sdvo *intel_sdvo, const struct drm_connector_state *conn_state) { struct intel_sdvo_tv_format format; - uint32_t format_map; + u32 format_map; format_map = 1 << conn_state->tv.mode; memset(&format, 0, sizeof(format)); @@ -1208,7 +1208,7 @@ static void intel_sdvo_update_props(struct intel_sdvo *intel_sdvo, const struct drm_connector_state *conn_state = &sdvo_state->base.base; struct intel_sdvo_connector *intel_sdvo_conn = to_intel_sdvo_connector(conn_state->connector); - uint16_t val; + u16 val; if (intel_sdvo_conn->left) UPDATE_PROPERTY(sdvo_state->tv.overscan_h, OVERSCAN_H); @@ -1692,10 +1692,10 @@ static bool intel_sdvo_get_capabilities(struct intel_sdvo *intel_sdvo, struct in return true; } -static uint16_t intel_sdvo_get_hotplug_support(struct intel_sdvo *intel_sdvo) +static u16 intel_sdvo_get_hotplug_support(struct intel_sdvo *intel_sdvo) { struct drm_i915_private *dev_priv = to_i915(intel_sdvo->base.base.dev); - uint16_t hotplug; + u16 hotplug; if (!I915_HAS_HOTPLUG(dev_priv)) return 0; @@ -1826,7 +1826,7 @@ intel_sdvo_connector_matches_edid(struct intel_sdvo_connector *sdvo, static enum drm_connector_status intel_sdvo_detect(struct drm_connector *connector, bool force) { - uint16_t response; + u16 response; struct intel_sdvo *intel_sdvo = intel_attached_sdvo(connector); struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector); enum drm_connector_status ret; @@ -1977,7 +1977,7 @@ static void intel_sdvo_get_tv_modes(struct drm_connector *connector) struct intel_sdvo *intel_sdvo = intel_attached_sdvo(connector); const struct drm_connector_state *conn_state = connector->state; struct intel_sdvo_sdtv_resolution_request tv_res; - uint32_t reply = 0, format_map = 0; + u32 reply = 0, format_map = 0; int i; DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n", @@ -2062,7 +2062,7 @@ static int intel_sdvo_connector_atomic_get_property(struct drm_connector *connector, const struct drm_connector_state *state, struct drm_property *property, - uint64_t *val) + u64 *val) { struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector); const struct intel_sdvo_connector_state *sdvo_state = to_intel_sdvo_connector_state((void *)state); @@ -2121,7 +2121,7 @@ static int intel_sdvo_connector_atomic_set_property(struct drm_connector *connector, struct drm_connector_state *state, struct drm_property *property, - uint64_t val) + u64 val) { struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector); struct intel_sdvo_connector_state *sdvo_state = to_intel_sdvo_connector_state(state); @@ -2270,7 +2270,7 @@ static const struct drm_encoder_funcs intel_sdvo_enc_funcs = { static void intel_sdvo_guess_ddc_bus(struct intel_sdvo *sdvo) { - uint16_t mask = 0; + u16 mask = 0; unsigned int num_bits; /* @@ -2671,7 +2671,7 @@ err: } static bool -intel_sdvo_output_setup(struct intel_sdvo *intel_sdvo, uint16_t flags) +intel_sdvo_output_setup(struct intel_sdvo *intel_sdvo, u16 flags) { /* SDVO requires XXX1 function may not exist unless it has XXX0 function.*/ @@ -2747,7 +2747,7 @@ static bool intel_sdvo_tv_create_property(struct intel_sdvo *intel_sdvo, { struct drm_device *dev = intel_sdvo->base.base.dev; struct intel_sdvo_tv_format format; - uint32_t format_map, i; + u32 format_map, i; if (!intel_sdvo_set_target_output(intel_sdvo, type)) return false; @@ -2814,7 +2814,7 @@ intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo, struct drm_connector_state *conn_state = connector->state; struct intel_sdvo_connector_state *sdvo_state = to_intel_sdvo_connector_state(conn_state); - uint16_t response, data_value[2]; + u16 response, data_value[2]; /* when horizontal overscan is supported, Add the left/right property */ if (enhancements.overscan_h) { @@ -2925,7 +2925,7 @@ intel_sdvo_create_enhance_property_lvds(struct intel_sdvo *intel_sdvo, { struct drm_device *dev = intel_sdvo->base.base.dev; struct drm_connector *connector = &intel_sdvo_connector->base.base; - uint16_t response, data_value[2]; + u16 response, data_value[2]; ENHANCEMENT(&connector->state->tv, brightness, BRIGHTNESS); @@ -2939,7 +2939,7 @@ static bool intel_sdvo_create_enhance_property(struct intel_sdvo *intel_sdvo, { union { struct intel_sdvo_enhancements_reply reply; - uint16_t response; + u16 response; } enhancements; BUILD_BUG_ON(sizeof(enhancements) != 2); diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c index 8f3982c03925..b56a1a9ad01d 100644 --- a/drivers/gpu/drm/i915/intel_sprite.c +++ b/drivers/gpu/drm/i915/intel_sprite.c @@ -321,8 +321,8 @@ skl_program_scaler(struct intel_plane *plane, &crtc_state->scaler_state.scalers[scaler_id]; int crtc_x = plane_state->base.dst.x1; int crtc_y = plane_state->base.dst.y1; - uint32_t crtc_w = drm_rect_width(&plane_state->base.dst); - uint32_t crtc_h = drm_rect_height(&plane_state->base.dst); + u32 crtc_w = drm_rect_width(&plane_state->base.dst); + u32 crtc_h = drm_rect_height(&plane_state->base.dst); u16 y_hphase, uv_rgb_hphase; u16 y_vphase, uv_rgb_vphase; int hscale, vscale; @@ -477,23 +477,30 @@ skl_program_plane(struct intel_plane *plane, u32 aux_stride = skl_plane_stride(plane_state, 1); int crtc_x = plane_state->base.dst.x1; int crtc_y = plane_state->base.dst.y1; - uint32_t x = plane_state->color_plane[color_plane].x; - uint32_t y = plane_state->color_plane[color_plane].y; - uint32_t src_w = drm_rect_width(&plane_state->base.src) >> 16; - uint32_t src_h = drm_rect_height(&plane_state->base.src) >> 16; + u32 x = plane_state->color_plane[color_plane].x; + u32 y = plane_state->color_plane[color_plane].y; + u32 src_w = drm_rect_width(&plane_state->base.src) >> 16; + u32 src_h = drm_rect_height(&plane_state->base.src) >> 16; struct intel_plane *linked = plane_state->linked_plane; const struct drm_framebuffer *fb = plane_state->base.fb; u8 alpha = plane_state->base.alpha >> 8; + u32 plane_color_ctl = 0; unsigned long irqflags; u32 keymsk, keymax; + plane_ctl |= skl_plane_ctl_crtc(crtc_state); + + if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv)) + plane_color_ctl = plane_state->color_ctl | + glk_plane_color_ctl_crtc(crtc_state); + /* Sizes are 0 based */ src_w--; src_h--; keymax = (key->max_value & 0xffffff) | PLANE_KEYMAX_ALPHA(alpha); - keymsk = key->channel_mask & 0x3ffffff; + keymsk = key->channel_mask & 0x7ffffff; if (alpha < 0xff) keymsk |= PLANE_KEYMSK_ALPHA_ENABLE; @@ -533,8 +540,7 @@ skl_program_plane(struct intel_plane *plane, } if (INTEL_GEN(dev_priv) >= 10 || IS_GEMINILAKE(dev_priv)) - I915_WRITE_FW(PLANE_COLOR_CTL(pipe, plane_id), - plane_state->color_ctl); + I915_WRITE_FW(PLANE_COLOR_CTL(pipe, plane_id), plane_color_ctl); if (fb->format->is_yuv && icl_is_hdr_plane(plane)) icl_program_input_csc(plane, crtc_state, plane_state); @@ -618,17 +624,19 @@ skl_plane_get_hw_state(struct intel_plane *plane, struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum intel_display_power_domain power_domain; enum plane_id plane_id = plane->id; + intel_wakeref_t wakeref; bool ret; power_domain = POWER_DOMAIN_PIPE(plane->pipe); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) return false; ret = I915_READ(PLANE_CTL(plane->pipe, plane_id)) & PLANE_CTL_ENABLE; *pipe = plane->pipe; - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); return ret; } @@ -731,6 +739,11 @@ vlv_update_clrc(const struct intel_plane_state *plane_state) SP_SH_SIN(sh_sin) | SP_SH_COS(sh_cos)); } +static u32 vlv_sprite_ctl_crtc(const struct intel_crtc_state *crtc_state) +{ + return SP_GAMMA_ENABLE; +} + static u32 vlv_sprite_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { @@ -739,7 +752,7 @@ static u32 vlv_sprite_ctl(const struct intel_crtc_state *crtc_state, const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; u32 sprctl; - sprctl = SP_ENABLE | SP_GAMMA_ENABLE; + sprctl = SP_ENABLE; switch (fb->format->format) { case DRM_FORMAT_YUYV: @@ -806,17 +819,19 @@ vlv_update_plane(struct intel_plane *plane, struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum pipe pipe = plane->pipe; enum plane_id plane_id = plane->id; - u32 sprctl = plane_state->ctl; u32 sprsurf_offset = plane_state->color_plane[0].offset; u32 linear_offset; const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; int crtc_x = plane_state->base.dst.x1; int crtc_y = plane_state->base.dst.y1; - uint32_t crtc_w = drm_rect_width(&plane_state->base.dst); - uint32_t crtc_h = drm_rect_height(&plane_state->base.dst); - uint32_t x = plane_state->color_plane[0].x; - uint32_t y = plane_state->color_plane[0].y; + u32 crtc_w = drm_rect_width(&plane_state->base.dst); + u32 crtc_h = drm_rect_height(&plane_state->base.dst); + u32 x = plane_state->color_plane[0].x; + u32 y = plane_state->color_plane[0].y; unsigned long irqflags; + u32 sprctl; + + sprctl = plane_state->ctl | vlv_sprite_ctl_crtc(crtc_state); /* Sizes are 0 based */ crtc_w--; @@ -882,21 +897,36 @@ vlv_plane_get_hw_state(struct intel_plane *plane, struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum intel_display_power_domain power_domain; enum plane_id plane_id = plane->id; + intel_wakeref_t wakeref; bool ret; power_domain = POWER_DOMAIN_PIPE(plane->pipe); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) return false; ret = I915_READ(SPCNTR(plane->pipe, plane_id)) & SP_ENABLE; *pipe = plane->pipe; - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); return ret; } +static u32 ivb_sprite_ctl_crtc(const struct intel_crtc_state *crtc_state) +{ + struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev); + u32 sprctl = 0; + + sprctl |= SPRITE_GAMMA_ENABLE; + + if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) + sprctl |= SPRITE_PIPE_CSC_ENABLE; + + return sprctl; +} + static u32 ivb_sprite_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { @@ -907,14 +937,11 @@ static u32 ivb_sprite_ctl(const struct intel_crtc_state *crtc_state, const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; u32 sprctl; - sprctl = SPRITE_ENABLE | SPRITE_GAMMA_ENABLE; + sprctl = SPRITE_ENABLE; if (IS_IVYBRIDGE(dev_priv)) sprctl |= SPRITE_TRICKLE_FEED_DISABLE; - if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) - sprctl |= SPRITE_PIPE_CSC_ENABLE; - switch (fb->format->format) { case DRM_FORMAT_XBGR8888: sprctl |= SPRITE_FORMAT_RGBX888 | SPRITE_RGB_ORDER_RGBX; @@ -966,20 +993,22 @@ ivb_update_plane(struct intel_plane *plane, { struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum pipe pipe = plane->pipe; - u32 sprctl = plane_state->ctl, sprscale = 0; u32 sprsurf_offset = plane_state->color_plane[0].offset; u32 linear_offset; const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; int crtc_x = plane_state->base.dst.x1; int crtc_y = plane_state->base.dst.y1; - uint32_t crtc_w = drm_rect_width(&plane_state->base.dst); - uint32_t crtc_h = drm_rect_height(&plane_state->base.dst); - uint32_t x = plane_state->color_plane[0].x; - uint32_t y = plane_state->color_plane[0].y; - uint32_t src_w = drm_rect_width(&plane_state->base.src) >> 16; - uint32_t src_h = drm_rect_height(&plane_state->base.src) >> 16; + u32 crtc_w = drm_rect_width(&plane_state->base.dst); + u32 crtc_h = drm_rect_height(&plane_state->base.dst); + u32 x = plane_state->color_plane[0].x; + u32 y = plane_state->color_plane[0].y; + u32 src_w = drm_rect_width(&plane_state->base.src) >> 16; + u32 src_h = drm_rect_height(&plane_state->base.src) >> 16; + u32 sprctl, sprscale = 0; unsigned long irqflags; + sprctl = plane_state->ctl | ivb_sprite_ctl_crtc(crtc_state); + /* Sizes are 0 based */ src_w--; src_h--; @@ -1051,17 +1080,19 @@ ivb_plane_get_hw_state(struct intel_plane *plane, { struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum intel_display_power_domain power_domain; + intel_wakeref_t wakeref; bool ret; power_domain = POWER_DOMAIN_PIPE(plane->pipe); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) return false; ret = I915_READ(SPRCTL(plane->pipe)) & SPRITE_ENABLE; *pipe = plane->pipe; - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); return ret; } @@ -1074,6 +1105,11 @@ g4x_sprite_max_stride(struct intel_plane *plane, return 16384; } +static u32 g4x_sprite_ctl_crtc(const struct intel_crtc_state *crtc_state) +{ + return DVS_GAMMA_ENABLE; +} + static u32 g4x_sprite_ctl(const struct intel_crtc_state *crtc_state, const struct intel_plane_state *plane_state) { @@ -1084,7 +1120,7 @@ static u32 g4x_sprite_ctl(const struct intel_crtc_state *crtc_state, const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; u32 dvscntr; - dvscntr = DVS_ENABLE | DVS_GAMMA_ENABLE; + dvscntr = DVS_ENABLE; if (IS_GEN(dev_priv, 6)) dvscntr |= DVS_TRICKLE_FEED_DISABLE; @@ -1140,20 +1176,22 @@ g4x_update_plane(struct intel_plane *plane, { struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum pipe pipe = plane->pipe; - u32 dvscntr = plane_state->ctl, dvsscale = 0; u32 dvssurf_offset = plane_state->color_plane[0].offset; u32 linear_offset; const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; int crtc_x = plane_state->base.dst.x1; int crtc_y = plane_state->base.dst.y1; - uint32_t crtc_w = drm_rect_width(&plane_state->base.dst); - uint32_t crtc_h = drm_rect_height(&plane_state->base.dst); - uint32_t x = plane_state->color_plane[0].x; - uint32_t y = plane_state->color_plane[0].y; - uint32_t src_w = drm_rect_width(&plane_state->base.src) >> 16; - uint32_t src_h = drm_rect_height(&plane_state->base.src) >> 16; + u32 crtc_w = drm_rect_width(&plane_state->base.dst); + u32 crtc_h = drm_rect_height(&plane_state->base.dst); + u32 x = plane_state->color_plane[0].x; + u32 y = plane_state->color_plane[0].y; + u32 src_w = drm_rect_width(&plane_state->base.src) >> 16; + u32 src_h = drm_rect_height(&plane_state->base.src) >> 16; + u32 dvscntr, dvsscale = 0; unsigned long irqflags; + dvscntr = plane_state->ctl | g4x_sprite_ctl_crtc(crtc_state); + /* Sizes are 0 based */ src_w--; src_h--; @@ -1217,17 +1255,19 @@ g4x_plane_get_hw_state(struct intel_plane *plane, { struct drm_i915_private *dev_priv = to_i915(plane->base.dev); enum intel_display_power_domain power_domain; + intel_wakeref_t wakeref; bool ret; power_domain = POWER_DOMAIN_PIPE(plane->pipe); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) return false; ret = I915_READ(DVSCNTR(plane->pipe)) & DVS_ENABLE; *pipe = plane->pipe; - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); return ret; } @@ -1698,7 +1738,7 @@ out: return ret; } -static const uint32_t g4x_plane_formats[] = { +static const u32 g4x_plane_formats[] = { DRM_FORMAT_XRGB8888, DRM_FORMAT_YUYV, DRM_FORMAT_YVYU, @@ -1706,13 +1746,13 @@ static const uint32_t g4x_plane_formats[] = { DRM_FORMAT_VYUY, }; -static const uint64_t i9xx_plane_format_modifiers[] = { +static const u64 i9xx_plane_format_modifiers[] = { I915_FORMAT_MOD_X_TILED, DRM_FORMAT_MOD_LINEAR, DRM_FORMAT_MOD_INVALID }; -static const uint32_t snb_plane_formats[] = { +static const u32 snb_plane_formats[] = { DRM_FORMAT_XBGR8888, DRM_FORMAT_XRGB8888, DRM_FORMAT_YUYV, @@ -1721,7 +1761,7 @@ static const uint32_t snb_plane_formats[] = { DRM_FORMAT_VYUY, }; -static const uint32_t vlv_plane_formats[] = { +static const u32 vlv_plane_formats[] = { DRM_FORMAT_RGB565, DRM_FORMAT_ABGR8888, DRM_FORMAT_ARGB8888, @@ -1735,7 +1775,7 @@ static const uint32_t vlv_plane_formats[] = { DRM_FORMAT_VYUY, }; -static const uint32_t skl_plane_formats[] = { +static const u32 skl_plane_formats[] = { DRM_FORMAT_C8, DRM_FORMAT_RGB565, DRM_FORMAT_XRGB8888, @@ -1750,7 +1790,7 @@ static const uint32_t skl_plane_formats[] = { DRM_FORMAT_VYUY, }; -static const uint32_t skl_planar_formats[] = { +static const u32 skl_planar_formats[] = { DRM_FORMAT_C8, DRM_FORMAT_RGB565, DRM_FORMAT_XRGB8888, @@ -1766,7 +1806,7 @@ static const uint32_t skl_planar_formats[] = { DRM_FORMAT_NV12, }; -static const uint64_t skl_plane_format_modifiers_noccs[] = { +static const u64 skl_plane_format_modifiers_noccs[] = { I915_FORMAT_MOD_Yf_TILED, I915_FORMAT_MOD_Y_TILED, I915_FORMAT_MOD_X_TILED, @@ -1774,7 +1814,7 @@ static const uint64_t skl_plane_format_modifiers_noccs[] = { DRM_FORMAT_MOD_INVALID }; -static const uint64_t skl_plane_format_modifiers_ccs[] = { +static const u64 skl_plane_format_modifiers_ccs[] = { I915_FORMAT_MOD_Yf_TILED_CCS, I915_FORMAT_MOD_Y_TILED_CCS, I915_FORMAT_MOD_Yf_TILED, diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c index bd5536f0ec92..3924c4944e1f 100644 --- a/drivers/gpu/drm/i915/intel_tv.c +++ b/drivers/gpu/drm/i915/intel_tv.c @@ -306,7 +306,7 @@ struct tv_mode { u32 clock; u16 refresh; /* in millihertz (for precision) */ - u32 oversample; + u8 oversample; u8 hsync_end; u16 hblank_start, hblank_end, htotal; bool progressive : 1, trilevel_sync : 1, component_only : 1; @@ -339,7 +339,6 @@ struct tv_mode { const struct video_levels *composite_levels, *svideo_levels; const struct color_conversion *composite_color, *svideo_color; const u32 *filter_table; - u16 max_srcw; }; @@ -378,8 +377,8 @@ static const struct tv_mode tv_modes[] = { .name = "NTSC-M", .clock = 108000, .refresh = 59940, - .oversample = TV_OVERSAMPLE_8X, - .component_only = 0, + .oversample = 8, + .component_only = false, /* 525 Lines, 60 Fields, 15.734KHz line, Sub-Carrier 3.580MHz */ .hsync_end = 64, .hblank_end = 124, @@ -421,8 +420,8 @@ static const struct tv_mode tv_modes[] = { .name = "NTSC-443", .clock = 108000, .refresh = 59940, - .oversample = TV_OVERSAMPLE_8X, - .component_only = 0, + .oversample = 8, + .component_only = false, /* 525 Lines, 60 Fields, 15.734KHz line, Sub-Carrier 4.43MHz */ .hsync_end = 64, .hblank_end = 124, .hblank_start = 836, .htotal = 857, @@ -463,8 +462,8 @@ static const struct tv_mode tv_modes[] = { .name = "NTSC-J", .clock = 108000, .refresh = 59940, - .oversample = TV_OVERSAMPLE_8X, - .component_only = 0, + .oversample = 8, + .component_only = false, /* 525 Lines, 60 Fields, 15.734KHz line, Sub-Carrier 3.580MHz */ .hsync_end = 64, .hblank_end = 124, @@ -506,8 +505,8 @@ static const struct tv_mode tv_modes[] = { .name = "PAL-M", .clock = 108000, .refresh = 59940, - .oversample = TV_OVERSAMPLE_8X, - .component_only = 0, + .oversample = 8, + .component_only = false, /* 525 Lines, 60 Fields, 15.734KHz line, Sub-Carrier 3.580MHz */ .hsync_end = 64, .hblank_end = 124, @@ -550,8 +549,8 @@ static const struct tv_mode tv_modes[] = { .name = "PAL-N", .clock = 108000, .refresh = 50000, - .oversample = TV_OVERSAMPLE_8X, - .component_only = 0, + .oversample = 8, + .component_only = false, .hsync_end = 64, .hblank_end = 128, .hblank_start = 844, .htotal = 863, @@ -595,8 +594,8 @@ static const struct tv_mode tv_modes[] = { .name = "PAL", .clock = 108000, .refresh = 50000, - .oversample = TV_OVERSAMPLE_8X, - .component_only = 0, + .oversample = 8, + .component_only = false, .hsync_end = 64, .hblank_end = 142, .hblank_start = 844, .htotal = 863, @@ -635,10 +634,10 @@ static const struct tv_mode tv_modes[] = { }, { .name = "480p", - .clock = 107520, + .clock = 108000, .refresh = 59940, - .oversample = TV_OVERSAMPLE_4X, - .component_only = 1, + .oversample = 4, + .component_only = true, .hsync_end = 64, .hblank_end = 122, .hblank_start = 842, .htotal = 857, @@ -659,10 +658,10 @@ static const struct tv_mode tv_modes[] = { }, { .name = "576p", - .clock = 107520, + .clock = 108000, .refresh = 50000, - .oversample = TV_OVERSAMPLE_4X, - .component_only = 1, + .oversample = 4, + .component_only = true, .hsync_end = 64, .hblank_end = 139, .hblank_start = 859, .htotal = 863, @@ -683,10 +682,10 @@ static const struct tv_mode tv_modes[] = { }, { .name = "720p@60Hz", - .clock = 148800, + .clock = 148500, .refresh = 60000, - .oversample = TV_OVERSAMPLE_2X, - .component_only = 1, + .oversample = 2, + .component_only = true, .hsync_end = 80, .hblank_end = 300, .hblank_start = 1580, .htotal = 1649, @@ -707,10 +706,10 @@ static const struct tv_mode tv_modes[] = { }, { .name = "720p@50Hz", - .clock = 148800, + .clock = 148500, .refresh = 50000, - .oversample = TV_OVERSAMPLE_2X, - .component_only = 1, + .oversample = 2, + .component_only = true, .hsync_end = 80, .hblank_end = 300, .hblank_start = 1580, .htotal = 1979, @@ -728,14 +727,13 @@ static const struct tv_mode tv_modes[] = { .burst_ena = false, .filter_table = filter_table, - .max_srcw = 800 }, { .name = "1080i@50Hz", - .clock = 148800, + .clock = 148500, .refresh = 50000, - .oversample = TV_OVERSAMPLE_2X, - .component_only = 1, + .oversample = 2, + .component_only = true, .hsync_end = 88, .hblank_end = 235, .hblank_start = 2155, .htotal = 2639, @@ -758,10 +756,10 @@ static const struct tv_mode tv_modes[] = { }, { .name = "1080i@60Hz", - .clock = 148800, + .clock = 148500, .refresh = 60000, - .oversample = TV_OVERSAMPLE_2X, - .component_only = 1, + .oversample = 2, + .component_only = true, .hsync_end = 88, .hblank_end = 235, .hblank_start = 2155, .htotal = 2199, @@ -782,8 +780,115 @@ static const struct tv_mode tv_modes[] = { .filter_table = filter_table, }, + + { + .name = "1080p@30Hz", + .clock = 148500, + .refresh = 30000, + .oversample = 2, + .component_only = true, + + .hsync_end = 88, .hblank_end = 235, + .hblank_start = 2155, .htotal = 2199, + + .progressive = true, .trilevel_sync = true, + + .vsync_start_f1 = 8, .vsync_start_f2 = 8, + .vsync_len = 10, + + .veq_ena = false, .veq_start_f1 = 0, + .veq_start_f2 = 0, .veq_len = 0, + + .vi_end_f1 = 44, .vi_end_f2 = 44, + .nbr_end = 1079, + + .burst_ena = false, + + .filter_table = filter_table, + }, + + { + .name = "1080p@50Hz", + .clock = 148500, + .refresh = 50000, + .oversample = 1, + .component_only = true, + + .hsync_end = 88, .hblank_end = 235, + .hblank_start = 2155, .htotal = 2639, + + .progressive = true, .trilevel_sync = true, + + .vsync_start_f1 = 8, .vsync_start_f2 = 8, + .vsync_len = 10, + + .veq_ena = false, .veq_start_f1 = 0, + .veq_start_f2 = 0, .veq_len = 0, + + .vi_end_f1 = 44, .vi_end_f2 = 44, + .nbr_end = 1079, + + .burst_ena = false, + + .filter_table = filter_table, + }, + + { + .name = "1080p@60Hz", + .clock = 148500, + .refresh = 60000, + .oversample = 1, + .component_only = true, + + .hsync_end = 88, .hblank_end = 235, + .hblank_start = 2155, .htotal = 2199, + + .progressive = true, .trilevel_sync = true, + + .vsync_start_f1 = 8, .vsync_start_f2 = 8, + .vsync_len = 10, + + .veq_ena = false, .veq_start_f1 = 0, + .veq_start_f2 = 0, .veq_len = 0, + + .vi_end_f1 = 44, .vi_end_f2 = 44, + .nbr_end = 1079, + + .burst_ena = false, + + .filter_table = filter_table, + }, +}; + +struct intel_tv_connector_state { + struct drm_connector_state base; + + /* + * May need to override the user margins for + * gen3 >1024 wide source vertical centering. + */ + struct { + u16 top, bottom; + } margins; + + bool bypass_vfilter; }; +#define to_intel_tv_connector_state(x) container_of(x, struct intel_tv_connector_state, base) + +static struct drm_connector_state * +intel_tv_connector_duplicate_state(struct drm_connector *connector) +{ + struct intel_tv_connector_state *state; + + state = kmemdup(connector->state, sizeof(*state), GFP_KERNEL); + if (!state) + return NULL; + + __drm_atomic_helper_connector_duplicate_state(connector, &state->base); + return &state->base; +} + static struct intel_tv *enc_to_tv(struct intel_encoder *encoder) { return container_of(encoder, struct intel_tv, base); @@ -859,14 +964,215 @@ intel_tv_mode_valid(struct drm_connector *connector, return MODE_CLOCK_RANGE; } +static int +intel_tv_mode_vdisplay(const struct tv_mode *tv_mode) +{ + if (tv_mode->progressive) + return tv_mode->nbr_end + 1; + else + return 2 * (tv_mode->nbr_end + 1); +} + +static void +intel_tv_mode_to_mode(struct drm_display_mode *mode, + const struct tv_mode *tv_mode) +{ + mode->clock = tv_mode->clock / + (tv_mode->oversample >> !tv_mode->progressive); + + /* + * tv_mode horizontal timings: + * + * hsync_end + * | hblank_end + * | | hblank_start + * | | | htotal + * | _______ | + * ____/ \___ + * \__/ \ + */ + mode->hdisplay = + tv_mode->hblank_start - tv_mode->hblank_end; + mode->hsync_start = mode->hdisplay + + tv_mode->htotal - tv_mode->hblank_start; + mode->hsync_end = mode->hsync_start + + tv_mode->hsync_end; + mode->htotal = tv_mode->htotal + 1; + + /* + * tv_mode vertical timings: + * + * vsync_start + * | vsync_end + * | | vi_end nbr_end + * | | | | + * | | _______ + * \__ ____/ \ + * \__/ + */ + mode->vdisplay = intel_tv_mode_vdisplay(tv_mode); + if (tv_mode->progressive) { + mode->vsync_start = mode->vdisplay + + tv_mode->vsync_start_f1 + 1; + mode->vsync_end = mode->vsync_start + + tv_mode->vsync_len; + mode->vtotal = mode->vdisplay + + tv_mode->vi_end_f1 + 1; + } else { + mode->vsync_start = mode->vdisplay + + tv_mode->vsync_start_f1 + 1 + + tv_mode->vsync_start_f2 + 1; + mode->vsync_end = mode->vsync_start + + 2 * tv_mode->vsync_len; + mode->vtotal = mode->vdisplay + + tv_mode->vi_end_f1 + 1 + + tv_mode->vi_end_f2 + 1; + } + + /* TV has it's own notion of sync and other mode flags, so clear them. */ + mode->flags = 0; + + mode->vrefresh = 0; + mode->vrefresh = drm_mode_vrefresh(mode); + + snprintf(mode->name, sizeof(mode->name), + "%dx%d%c (%s)", + mode->hdisplay, mode->vdisplay, + tv_mode->progressive ? 'p' : 'i', + tv_mode->name); +} + +static void intel_tv_scale_mode_horiz(struct drm_display_mode *mode, + int hdisplay, int left_margin, + int right_margin) +{ + int hsync_start = mode->hsync_start - mode->hdisplay + right_margin; + int hsync_end = mode->hsync_end - mode->hdisplay + right_margin; + int new_htotal = mode->htotal * hdisplay / + (mode->hdisplay - left_margin - right_margin); + + mode->clock = mode->clock * new_htotal / mode->htotal; + + mode->hdisplay = hdisplay; + mode->hsync_start = hdisplay + hsync_start * new_htotal / mode->htotal; + mode->hsync_end = hdisplay + hsync_end * new_htotal / mode->htotal; + mode->htotal = new_htotal; +} + +static void intel_tv_scale_mode_vert(struct drm_display_mode *mode, + int vdisplay, int top_margin, + int bottom_margin) +{ + int vsync_start = mode->vsync_start - mode->vdisplay + bottom_margin; + int vsync_end = mode->vsync_end - mode->vdisplay + bottom_margin; + int new_vtotal = mode->vtotal * vdisplay / + (mode->vdisplay - top_margin - bottom_margin); + + mode->clock = mode->clock * new_vtotal / mode->vtotal; + + mode->vdisplay = vdisplay; + mode->vsync_start = vdisplay + vsync_start * new_vtotal / mode->vtotal; + mode->vsync_end = vdisplay + vsync_end * new_vtotal / mode->vtotal; + mode->vtotal = new_vtotal; +} static void intel_tv_get_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config) { + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct drm_display_mode *adjusted_mode = + &pipe_config->base.adjusted_mode; + struct drm_display_mode mode = {}; + u32 tv_ctl, hctl1, hctl3, vctl1, vctl2, tmp; + struct tv_mode tv_mode = {}; + int hdisplay = adjusted_mode->crtc_hdisplay; + int vdisplay = adjusted_mode->crtc_vdisplay; + int xsize, ysize, xpos, ypos; + pipe_config->output_types |= BIT(INTEL_OUTPUT_TVOUT); - pipe_config->base.adjusted_mode.crtc_clock = pipe_config->port_clock; + tv_ctl = I915_READ(TV_CTL); + hctl1 = I915_READ(TV_H_CTL_1); + hctl3 = I915_READ(TV_H_CTL_3); + vctl1 = I915_READ(TV_V_CTL_1); + vctl2 = I915_READ(TV_V_CTL_2); + + tv_mode.htotal = (hctl1 & TV_HTOTAL_MASK) >> TV_HTOTAL_SHIFT; + tv_mode.hsync_end = (hctl1 & TV_HSYNC_END_MASK) >> TV_HSYNC_END_SHIFT; + + tv_mode.hblank_start = (hctl3 & TV_HBLANK_START_MASK) >> TV_HBLANK_START_SHIFT; + tv_mode.hblank_end = (hctl3 & TV_HSYNC_END_MASK) >> TV_HBLANK_END_SHIFT; + + tv_mode.nbr_end = (vctl1 & TV_NBR_END_MASK) >> TV_NBR_END_SHIFT; + tv_mode.vi_end_f1 = (vctl1 & TV_VI_END_F1_MASK) >> TV_VI_END_F1_SHIFT; + tv_mode.vi_end_f2 = (vctl1 & TV_VI_END_F2_MASK) >> TV_VI_END_F2_SHIFT; + + tv_mode.vsync_len = (vctl2 & TV_VSYNC_LEN_MASK) >> TV_VSYNC_LEN_SHIFT; + tv_mode.vsync_start_f1 = (vctl2 & TV_VSYNC_START_F1_MASK) >> TV_VSYNC_START_F1_SHIFT; + tv_mode.vsync_start_f2 = (vctl2 & TV_VSYNC_START_F2_MASK) >> TV_VSYNC_START_F2_SHIFT; + + tv_mode.clock = pipe_config->port_clock; + + tv_mode.progressive = tv_ctl & TV_PROGRESSIVE; + + switch (tv_ctl & TV_OVERSAMPLE_MASK) { + case TV_OVERSAMPLE_8X: + tv_mode.oversample = 8; + break; + case TV_OVERSAMPLE_4X: + tv_mode.oversample = 4; + break; + case TV_OVERSAMPLE_2X: + tv_mode.oversample = 2; + break; + default: + tv_mode.oversample = 1; + break; + } + + tmp = I915_READ(TV_WIN_POS); + xpos = tmp >> 16; + ypos = tmp & 0xffff; + + tmp = I915_READ(TV_WIN_SIZE); + xsize = tmp >> 16; + ysize = tmp & 0xffff; + + intel_tv_mode_to_mode(&mode, &tv_mode); + + DRM_DEBUG_KMS("TV mode:\n"); + drm_mode_debug_printmodeline(&mode); + + intel_tv_scale_mode_horiz(&mode, hdisplay, + xpos, mode.hdisplay - xsize - xpos); + intel_tv_scale_mode_vert(&mode, vdisplay, + ypos, mode.vdisplay - ysize - ypos); + + adjusted_mode->crtc_clock = mode.clock; + if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) + adjusted_mode->crtc_clock /= 2; + + /* pixel counter doesn't work on i965gm TV output */ + if (IS_I965GM(dev_priv)) + adjusted_mode->private_flags |= + I915_MODE_FLAG_USE_SCANLINE_COUNTER; +} + +static bool intel_tv_source_too_wide(struct drm_i915_private *dev_priv, + int hdisplay) +{ + return IS_GEN(dev_priv, 3) && hdisplay > 1024; +} + +static bool intel_tv_vert_scaling(const struct drm_display_mode *tv_mode, + const struct drm_connector_state *conn_state, + int vdisplay) +{ + return tv_mode->crtc_vdisplay - + conn_state->tv.margins.top - + conn_state->tv.margins.bottom != + vdisplay; } static int @@ -874,9 +1180,14 @@ intel_tv_compute_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config, struct drm_connector_state *conn_state) { + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + struct intel_tv_connector_state *tv_conn_state = + to_intel_tv_connector_state(conn_state); const struct tv_mode *tv_mode = intel_tv_mode_find(conn_state); struct drm_display_mode *adjusted_mode = &pipe_config->base.adjusted_mode; + int hdisplay = adjusted_mode->crtc_hdisplay; + int vdisplay = adjusted_mode->crtc_vdisplay; if (!tv_mode) return -EINVAL; @@ -885,17 +1196,136 @@ intel_tv_compute_config(struct intel_encoder *encoder, return -EINVAL; pipe_config->output_format = INTEL_OUTPUT_FORMAT_RGB; - adjusted_mode->crtc_clock = tv_mode->clock; + DRM_DEBUG_KMS("forcing bpc to 8 for TV\n"); pipe_config->pipe_bpp = 8*3; - /* TV has it's own notion of sync and other mode flags, so clear them. */ - adjusted_mode->flags = 0; + pipe_config->port_clock = tv_mode->clock; + + intel_tv_mode_to_mode(adjusted_mode, tv_mode); + drm_mode_set_crtcinfo(adjusted_mode, 0); + + if (intel_tv_source_too_wide(dev_priv, hdisplay) || + !intel_tv_vert_scaling(adjusted_mode, conn_state, vdisplay)) { + int extra, top, bottom; + + extra = adjusted_mode->crtc_vdisplay - vdisplay; + + if (extra < 0) { + DRM_DEBUG_KMS("No vertical scaling for >1024 pixel wide modes\n"); + return -EINVAL; + } + + /* Need to turn off the vertical filter and center the image */ + + /* Attempt to maintain the relative sizes of the margins */ + top = conn_state->tv.margins.top; + bottom = conn_state->tv.margins.bottom; + + if (top + bottom) + top = extra * top / (top + bottom); + else + top = extra / 2; + bottom = extra - top; + + tv_conn_state->margins.top = top; + tv_conn_state->margins.bottom = bottom; + + tv_conn_state->bypass_vfilter = true; + + if (!tv_mode->progressive) { + adjusted_mode->clock /= 2; + adjusted_mode->crtc_clock /= 2; + adjusted_mode->flags |= DRM_MODE_FLAG_INTERLACE; + } + } else { + tv_conn_state->margins.top = conn_state->tv.margins.top; + tv_conn_state->margins.bottom = conn_state->tv.margins.bottom; + + tv_conn_state->bypass_vfilter = false; + } + + DRM_DEBUG_KMS("TV mode:\n"); + drm_mode_debug_printmodeline(adjusted_mode); /* - * FIXME: We don't check whether the input mode is actually what we want - * or whether userspace is doing something stupid. + * The pipe scanline counter behaviour looks as follows when + * using the TV encoder: + * + * time -> + * + * dsl=vtotal-1 | | + * || || + * ___| | ___| | + * / | / | + * / | / | + * dsl=0 ___/ |_____/ | + * | | | | | | + * ^ ^ ^ ^ ^ + * | | | | pipe vblank/first part of tv vblank + * | | | bottom margin + * | | active + * | top margin + * remainder of tv vblank + * + * When the TV encoder is used the pipe wants to run faster + * than expected rate. During the active portion the TV + * encoder stalls the pipe every few lines to keep it in + * check. When the TV encoder reaches the bottom margin the + * pipe simply stops. Once we reach the TV vblank the pipe is + * no longer stalled and it runs at the max rate (apparently + * oversample clock on gen3, cdclk on gen4). Once the pipe + * reaches the pipe vtotal the pipe stops for the remainder + * of the TV vblank/top margin. The pipe starts up again when + * the TV encoder exits the top margin. + * + * To avoid huge hassles for vblank timestamping we scale + * the pipe timings as if the pipe always runs at the average + * rate it maintains during the active period. This also + * gives us a reasonable guesstimate as to the pixel rate. + * Due to the variation in the actual pipe speed the scanline + * counter will give us slightly erroneous results during the + * TV vblank/margins. But since vtotal was selected such that + * it matches the average rate of the pipe during the active + * portion the error shouldn't cause any serious grief to + * vblank timestamps. + * + * For posterity here is the empirically derived formula + * that gives us the maximum length of the pipe vblank + * we can use without causing display corruption. Following + * this would allow us to have a ticking scanline counter + * everywhere except during the bottom margin (there the + * pipe always stops). Ie. this would eliminate the second + * flat portion of the above graph. However this would also + * complicate vblank timestamping as the pipe vtotal would + * no longer match the average rate the pipe runs at during + * the active portion. Hence following this formula seems + * more trouble that it's worth. + * + * if (IS_GEN(dev_priv, 4)) { + * num = cdclk * (tv_mode->oversample >> !tv_mode->progressive); + * den = tv_mode->clock; + * } else { + * num = tv_mode->oversample >> !tv_mode->progressive; + * den = 1; + * } + * max_pipe_vblank_len ~= + * (num * tv_htotal * (tv_vblank_len + top_margin)) / + * (den * pipe_htotal); */ + intel_tv_scale_mode_horiz(adjusted_mode, hdisplay, + conn_state->tv.margins.left, + conn_state->tv.margins.right); + intel_tv_scale_mode_vert(adjusted_mode, vdisplay, + tv_conn_state->margins.top, + tv_conn_state->margins.bottom); + drm_mode_set_crtcinfo(adjusted_mode, 0); + adjusted_mode->name[0] = '\0'; + + /* pixel counter doesn't work on i965gm TV output */ + if (IS_I965GM(dev_priv)) + adjusted_mode->private_flags |= + I915_MODE_FLAG_USE_SCANLINE_COUNTER; return 0; } @@ -986,14 +1416,16 @@ static void intel_tv_pre_enable(struct intel_encoder *encoder, struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_crtc *intel_crtc = to_intel_crtc(pipe_config->base.crtc); struct intel_tv *intel_tv = enc_to_tv(encoder); + const struct intel_tv_connector_state *tv_conn_state = + to_intel_tv_connector_state(conn_state); const struct tv_mode *tv_mode = intel_tv_mode_find(conn_state); - u32 tv_ctl; + u32 tv_ctl, tv_filter_ctl; u32 scctl1, scctl2, scctl3; int i, j; const struct video_levels *video_levels; const struct color_conversion *color_conversion; bool burst_ena; - int xpos = 0x0, ypos = 0x0; + int xpos, ypos; unsigned int xsize, ysize; if (!tv_mode) @@ -1029,7 +1461,21 @@ static void intel_tv_pre_enable(struct intel_encoder *encoder, } tv_ctl |= TV_ENC_PIPE_SEL(intel_crtc->pipe); - tv_ctl |= tv_mode->oversample; + + switch (tv_mode->oversample) { + case 8: + tv_ctl |= TV_OVERSAMPLE_8X; + break; + case 4: + tv_ctl |= TV_OVERSAMPLE_4X; + break; + case 2: + tv_ctl |= TV_OVERSAMPLE_2X; + break; + default: + tv_ctl |= TV_OVERSAMPLE_NONE; + break; + } if (tv_mode->progressive) tv_ctl |= TV_PROGRESSIVE; @@ -1081,19 +1527,20 @@ static void intel_tv_pre_enable(struct intel_encoder *encoder, assert_pipe_disabled(dev_priv, intel_crtc->pipe); /* Filter ctl must be set before TV_WIN_SIZE */ - I915_WRITE(TV_FILTER_CTL_1, TV_AUTO_SCALE); + tv_filter_ctl = TV_AUTO_SCALE; + if (tv_conn_state->bypass_vfilter) + tv_filter_ctl |= TV_V_FILTER_BYPASS; + I915_WRITE(TV_FILTER_CTL_1, tv_filter_ctl); + xsize = tv_mode->hblank_start - tv_mode->hblank_end; - if (tv_mode->progressive) - ysize = tv_mode->nbr_end + 1; - else - ysize = 2*tv_mode->nbr_end + 1; + ysize = intel_tv_mode_vdisplay(tv_mode); - xpos += conn_state->tv.margins.left; - ypos += conn_state->tv.margins.top; + xpos = conn_state->tv.margins.left; + ypos = tv_conn_state->margins.top; xsize -= (conn_state->tv.margins.left + conn_state->tv.margins.right); - ysize -= (conn_state->tv.margins.top + - conn_state->tv.margins.bottom); + ysize -= (tv_conn_state->margins.top + + tv_conn_state->margins.bottom); I915_WRITE(TV_WIN_POS, (xpos<<16)|ypos); I915_WRITE(TV_WIN_SIZE, (xsize<<16)|ysize); @@ -1110,23 +1557,6 @@ static void intel_tv_pre_enable(struct intel_encoder *encoder, I915_WRITE(TV_CTL, tv_ctl); } -static const struct drm_display_mode reported_modes[] = { - { - .name = "NTSC 480i", - .clock = 107520, - .hdisplay = 1280, - .hsync_start = 1368, - .hsync_end = 1496, - .htotal = 1712, - - .vdisplay = 1024, - .vsync_start = 1027, - .vsync_end = 1034, - .vtotal = 1104, - .type = DRM_MODE_TYPE_DRIVER, - }, -}; - static int intel_tv_detect_type(struct intel_tv *intel_tv, struct drm_connector *connector) @@ -1233,16 +1663,18 @@ static void intel_tv_find_better_format(struct drm_connector *connector) const struct tv_mode *tv_mode = intel_tv_mode_find(connector->state); int i; - if ((intel_tv->type == DRM_MODE_CONNECTOR_Component) == - tv_mode->component_only) + /* Component supports everything so we can keep the current mode */ + if (intel_tv->type == DRM_MODE_CONNECTOR_Component) return; + /* If the current mode is fine don't change it */ + if (!tv_mode->component_only) + return; for (i = 0; i < ARRAY_SIZE(tv_modes); i++) { - tv_mode = tv_modes + i; + tv_mode = &tv_modes[i]; - if ((intel_tv->type == DRM_MODE_CONNECTOR_Component) == - tv_mode->component_only) + if (!tv_mode->component_only) break; } @@ -1254,7 +1686,6 @@ intel_tv_detect(struct drm_connector *connector, struct drm_modeset_acquire_ctx *ctx, bool force) { - struct drm_display_mode mode; struct intel_tv *intel_tv = intel_attached_tv(connector); enum drm_connector_status status; int type; @@ -1263,13 +1694,11 @@ intel_tv_detect(struct drm_connector *connector, connector->base.id, connector->name, force); - mode = reported_modes[0]; - if (force) { struct intel_load_detect_pipe tmp; int ret; - ret = intel_get_load_detect_pipe(connector, &mode, &tmp, ctx); + ret = intel_get_load_detect_pipe(connector, NULL, &tmp, ctx); if (ret < 0) return ret; @@ -1293,84 +1722,85 @@ intel_tv_detect(struct drm_connector *connector, } static const struct input_res { - const char *name; - int w, h; + u16 w, h; } input_res_table[] = { - {"640x480", 640, 480}, - {"800x600", 800, 600}, - {"1024x768", 1024, 768}, - {"1280x1024", 1280, 1024}, - {"848x480", 848, 480}, - {"1280x720", 1280, 720}, - {"1920x1080", 1920, 1080}, + { 640, 480 }, + { 800, 600 }, + { 1024, 768 }, + { 1280, 1024 }, + { 848, 480 }, + { 1280, 720 }, + { 1920, 1080 }, }; -/* - * Chose preferred mode according to line number of TV format - */ +/* Choose preferred mode according to line number of TV format */ +static bool +intel_tv_is_preferred_mode(const struct drm_display_mode *mode, + const struct tv_mode *tv_mode) +{ + int vdisplay = intel_tv_mode_vdisplay(tv_mode); + + /* prefer 480 line modes for all SD TV modes */ + if (vdisplay <= 576) + vdisplay = 480; + + return vdisplay == mode->vdisplay; +} + static void -intel_tv_choose_preferred_modes(const struct tv_mode *tv_mode, - struct drm_display_mode *mode_ptr) +intel_tv_set_mode_type(struct drm_display_mode *mode, + const struct tv_mode *tv_mode) { - if (tv_mode->nbr_end < 480 && mode_ptr->vdisplay == 480) - mode_ptr->type |= DRM_MODE_TYPE_PREFERRED; - else if (tv_mode->nbr_end > 480) { - if (tv_mode->progressive == true && tv_mode->nbr_end < 720) { - if (mode_ptr->vdisplay == 720) - mode_ptr->type |= DRM_MODE_TYPE_PREFERRED; - } else if (mode_ptr->vdisplay == 1080) - mode_ptr->type |= DRM_MODE_TYPE_PREFERRED; - } + mode->type = DRM_MODE_TYPE_DRIVER; + + if (intel_tv_is_preferred_mode(mode, tv_mode)) + mode->type |= DRM_MODE_TYPE_PREFERRED; } static int intel_tv_get_modes(struct drm_connector *connector) { - struct drm_display_mode *mode_ptr; + struct drm_i915_private *dev_priv = to_i915(connector->dev); const struct tv_mode *tv_mode = intel_tv_mode_find(connector->state); - int j, count = 0; - u64 tmp; + int i, count = 0; - for (j = 0; j < ARRAY_SIZE(input_res_table); - j++) { - const struct input_res *input = &input_res_table[j]; - unsigned int hactive_s = input->w; - unsigned int vactive_s = input->h; + for (i = 0; i < ARRAY_SIZE(input_res_table); i++) { + const struct input_res *input = &input_res_table[i]; + struct drm_display_mode *mode; - if (tv_mode->max_srcw && input->w > tv_mode->max_srcw) + if (input->w > 1024 && + !tv_mode->progressive && + !tv_mode->component_only) continue; - if (input->w > 1024 && (!tv_mode->progressive - && !tv_mode->component_only)) + /* no vertical scaling with wide sources on gen3 */ + if (IS_GEN(dev_priv, 3) && input->w > 1024 && + input->h > intel_tv_mode_vdisplay(tv_mode)) continue; - mode_ptr = drm_mode_create(connector->dev); - if (!mode_ptr) + mode = drm_mode_create(connector->dev); + if (!mode) continue; - strlcpy(mode_ptr->name, input->name, DRM_DISPLAY_MODE_LEN); - - mode_ptr->hdisplay = hactive_s; - mode_ptr->hsync_start = hactive_s + 1; - mode_ptr->hsync_end = hactive_s + 64; - if (mode_ptr->hsync_end <= mode_ptr->hsync_start) - mode_ptr->hsync_end = mode_ptr->hsync_start + 1; - mode_ptr->htotal = hactive_s + 96; - - mode_ptr->vdisplay = vactive_s; - mode_ptr->vsync_start = vactive_s + 1; - mode_ptr->vsync_end = vactive_s + 32; - if (mode_ptr->vsync_end <= mode_ptr->vsync_start) - mode_ptr->vsync_end = mode_ptr->vsync_start + 1; - mode_ptr->vtotal = vactive_s + 33; - - tmp = mul_u32_u32(tv_mode->refresh, mode_ptr->vtotal); - tmp *= mode_ptr->htotal; - tmp = div_u64(tmp, 1000000); - mode_ptr->clock = (int) tmp; - - mode_ptr->type = DRM_MODE_TYPE_DRIVER; - intel_tv_choose_preferred_modes(tv_mode, mode_ptr); - drm_mode_probed_add(connector, mode_ptr); + + /* + * We take the TV mode and scale it to look + * like it had the expected h/vdisplay. This + * provides the most information to userspace + * about the actual timings of the mode. We + * do ignore the margins though. + */ + intel_tv_mode_to_mode(mode, tv_mode); + if (count == 0) { + DRM_DEBUG_KMS("TV mode:\n"); + drm_mode_debug_printmodeline(mode); + } + intel_tv_scale_mode_horiz(mode, input->w, 0, 0); + intel_tv_scale_mode_vert(mode, input->h, 0, 0); + intel_tv_set_mode_type(mode, tv_mode); + + drm_mode_set_name(mode); + + drm_mode_probed_add(connector, mode); count++; } @@ -1383,7 +1813,7 @@ static const struct drm_connector_funcs intel_tv_connector_funcs = { .destroy = intel_connector_destroy, .fill_modes = drm_helper_probe_single_connector_modes, .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, - .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, + .atomic_duplicate_state = intel_tv_connector_duplicate_state, }; static int intel_tv_atomic_check(struct drm_connector *connector, @@ -1530,11 +1960,15 @@ intel_tv_init(struct drm_i915_private *dev_priv) connector->doublescan_allowed = false; /* Create TV properties then attach current values */ - for (i = 0; i < ARRAY_SIZE(tv_modes); i++) + for (i = 0; i < ARRAY_SIZE(tv_modes); i++) { + /* 1080p50/1080p60 not supported on gen3 */ + if (IS_GEN(dev_priv, 3) && + tv_modes[i].oversample == 1) + break; + tv_format_names[i] = tv_modes[i].name; - drm_mode_create_tv_properties(dev, - ARRAY_SIZE(tv_modes), - tv_format_names); + } + drm_mode_create_tv_properties(dev, i, tv_format_names); drm_object_attach_property(&connector->base, dev->mode_config.tv_mode_property, state->tv.mode); diff --git a/drivers/gpu/drm/i915/intel_uc.c b/drivers/gpu/drm/i915/intel_uc.c index 731b82afe636..e711eb3268bc 100644 --- a/drivers/gpu/drm/i915/intel_uc.c +++ b/drivers/gpu/drm/i915/intel_uc.c @@ -26,6 +26,7 @@ #include "intel_guc_submission.h" #include "intel_guc.h" #include "i915_drv.h" +#include "i915_reset.h" static void guc_free_load_err_log(struct intel_guc *guc); diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c index fff468f17d2d..75646a1e0051 100644 --- a/drivers/gpu/drm/i915/intel_uncore.c +++ b/drivers/gpu/drm/i915/intel_uncore.c @@ -1670,6 +1670,7 @@ int i915_reg_read_ioctl(struct drm_device *dev, struct drm_i915_private *dev_priv = to_i915(dev); struct drm_i915_reg_read *reg = data; struct reg_whitelist const *entry; + intel_wakeref_t wakeref; unsigned int flags; int remain; int ret = 0; @@ -1695,388 +1696,22 @@ int i915_reg_read_ioctl(struct drm_device *dev, flags = reg->offset & (entry->size - 1); - intel_runtime_pm_get(dev_priv); - if (entry->size == 8 && flags == I915_REG_READ_8B_WA) - reg->val = I915_READ64_2x32(entry->offset_ldw, - entry->offset_udw); - else if (entry->size == 8 && flags == 0) - reg->val = I915_READ64(entry->offset_ldw); - else if (entry->size == 4 && flags == 0) - reg->val = I915_READ(entry->offset_ldw); - else if (entry->size == 2 && flags == 0) - reg->val = I915_READ16(entry->offset_ldw); - else if (entry->size == 1 && flags == 0) - reg->val = I915_READ8(entry->offset_ldw); - else - ret = -EINVAL; - intel_runtime_pm_put(dev_priv); - - return ret; -} - -static void gen3_stop_engine(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - const u32 base = engine->mmio_base; - - if (intel_engine_stop_cs(engine)) - DRM_DEBUG_DRIVER("%s: timed out on STOP_RING\n", engine->name); - - I915_WRITE_FW(RING_HEAD(base), I915_READ_FW(RING_TAIL(base))); - POSTING_READ_FW(RING_HEAD(base)); /* paranoia */ - - I915_WRITE_FW(RING_HEAD(base), 0); - I915_WRITE_FW(RING_TAIL(base), 0); - POSTING_READ_FW(RING_TAIL(base)); - - /* The ring must be empty before it is disabled */ - I915_WRITE_FW(RING_CTL(base), 0); - - /* Check acts as a post */ - if (I915_READ_FW(RING_HEAD(base)) != 0) - DRM_DEBUG_DRIVER("%s: ring head not parked\n", - engine->name); -} - -static void i915_stop_engines(struct drm_i915_private *dev_priv, - unsigned int engine_mask) -{ - struct intel_engine_cs *engine; - enum intel_engine_id id; - - if (INTEL_GEN(dev_priv) < 3) - return; - - for_each_engine_masked(engine, dev_priv, engine_mask, id) - gen3_stop_engine(engine); -} - -static bool i915_in_reset(struct pci_dev *pdev) -{ - u8 gdrst; - - pci_read_config_byte(pdev, I915_GDRST, &gdrst); - return gdrst & GRDOM_RESET_STATUS; -} - -static int i915_do_reset(struct drm_i915_private *dev_priv, - unsigned int engine_mask, - unsigned int retry) -{ - struct pci_dev *pdev = dev_priv->drm.pdev; - int err; - - /* Assert reset for at least 20 usec, and wait for acknowledgement. */ - pci_write_config_byte(pdev, I915_GDRST, GRDOM_RESET_ENABLE); - usleep_range(50, 200); - err = wait_for(i915_in_reset(pdev), 500); - - /* Clear the reset request. */ - pci_write_config_byte(pdev, I915_GDRST, 0); - usleep_range(50, 200); - if (!err) - err = wait_for(!i915_in_reset(pdev), 500); - - return err; -} - -static bool g4x_reset_complete(struct pci_dev *pdev) -{ - u8 gdrst; - - pci_read_config_byte(pdev, I915_GDRST, &gdrst); - return (gdrst & GRDOM_RESET_ENABLE) == 0; -} - -static int g33_do_reset(struct drm_i915_private *dev_priv, - unsigned int engine_mask, - unsigned int retry) -{ - struct pci_dev *pdev = dev_priv->drm.pdev; - - pci_write_config_byte(pdev, I915_GDRST, GRDOM_RESET_ENABLE); - return wait_for(g4x_reset_complete(pdev), 500); -} - -static int g4x_do_reset(struct drm_i915_private *dev_priv, - unsigned int engine_mask, - unsigned int retry) -{ - struct pci_dev *pdev = dev_priv->drm.pdev; - int ret; - - /* WaVcpClkGateDisableForMediaReset:ctg,elk */ - I915_WRITE(VDECCLK_GATE_D, - I915_READ(VDECCLK_GATE_D) | VCP_UNIT_CLOCK_GATE_DISABLE); - POSTING_READ(VDECCLK_GATE_D); - - pci_write_config_byte(pdev, I915_GDRST, - GRDOM_MEDIA | GRDOM_RESET_ENABLE); - ret = wait_for(g4x_reset_complete(pdev), 500); - if (ret) { - DRM_DEBUG_DRIVER("Wait for media reset failed\n"); - goto out; - } - - pci_write_config_byte(pdev, I915_GDRST, - GRDOM_RENDER | GRDOM_RESET_ENABLE); - ret = wait_for(g4x_reset_complete(pdev), 500); - if (ret) { - DRM_DEBUG_DRIVER("Wait for render reset failed\n"); - goto out; + with_intel_runtime_pm(dev_priv, wakeref) { + if (entry->size == 8 && flags == I915_REG_READ_8B_WA) + reg->val = I915_READ64_2x32(entry->offset_ldw, + entry->offset_udw); + else if (entry->size == 8 && flags == 0) + reg->val = I915_READ64(entry->offset_ldw); + else if (entry->size == 4 && flags == 0) + reg->val = I915_READ(entry->offset_ldw); + else if (entry->size == 2 && flags == 0) + reg->val = I915_READ16(entry->offset_ldw); + else if (entry->size == 1 && flags == 0) + reg->val = I915_READ8(entry->offset_ldw); + else + ret = -EINVAL; } -out: - pci_write_config_byte(pdev, I915_GDRST, 0); - - I915_WRITE(VDECCLK_GATE_D, - I915_READ(VDECCLK_GATE_D) & ~VCP_UNIT_CLOCK_GATE_DISABLE); - POSTING_READ(VDECCLK_GATE_D); - - return ret; -} - -static int ironlake_do_reset(struct drm_i915_private *dev_priv, - unsigned int engine_mask, - unsigned int retry) -{ - int ret; - - I915_WRITE(ILK_GDSR, ILK_GRDOM_RENDER | ILK_GRDOM_RESET_ENABLE); - ret = intel_wait_for_register(dev_priv, - ILK_GDSR, ILK_GRDOM_RESET_ENABLE, 0, - 500); - if (ret) { - DRM_DEBUG_DRIVER("Wait for render reset failed\n"); - goto out; - } - - I915_WRITE(ILK_GDSR, ILK_GRDOM_MEDIA | ILK_GRDOM_RESET_ENABLE); - ret = intel_wait_for_register(dev_priv, - ILK_GDSR, ILK_GRDOM_RESET_ENABLE, 0, - 500); - if (ret) { - DRM_DEBUG_DRIVER("Wait for media reset failed\n"); - goto out; - } - -out: - I915_WRITE(ILK_GDSR, 0); - POSTING_READ(ILK_GDSR); - return ret; -} - -/* Reset the hardware domains (GENX_GRDOM_*) specified by mask */ -static int gen6_hw_domain_reset(struct drm_i915_private *dev_priv, - u32 hw_domain_mask) -{ - int err; - - /* GEN6_GDRST is not in the gt power well, no need to check - * for fifo space for the write or forcewake the chip for - * the read - */ - __raw_i915_write32(dev_priv, GEN6_GDRST, hw_domain_mask); - - /* Wait for the device to ack the reset requests */ - err = __intel_wait_for_register_fw(dev_priv, - GEN6_GDRST, hw_domain_mask, 0, - 500, 0, - NULL); - if (err) - DRM_DEBUG_DRIVER("Wait for 0x%08x engines reset failed\n", - hw_domain_mask); - - return err; -} - -/** - * gen6_reset_engines - reset individual engines - * @dev_priv: i915 device - * @engine_mask: mask of intel_ring_flag() engines or ALL_ENGINES for full reset - * @retry: the count of of previous attempts to reset. - * - * This function will reset the individual engines that are set in engine_mask. - * If you provide ALL_ENGINES as mask, full global domain reset will be issued. - * - * Note: It is responsibility of the caller to handle the difference between - * asking full domain reset versus reset for all available individual engines. - * - * Returns 0 on success, nonzero on error. - */ -static int gen6_reset_engines(struct drm_i915_private *dev_priv, - unsigned int engine_mask, - unsigned int retry) -{ - struct intel_engine_cs *engine; - const u32 hw_engine_mask[I915_NUM_ENGINES] = { - [RCS] = GEN6_GRDOM_RENDER, - [BCS] = GEN6_GRDOM_BLT, - [VCS] = GEN6_GRDOM_MEDIA, - [VCS2] = GEN8_GRDOM_MEDIA2, - [VECS] = GEN6_GRDOM_VECS, - }; - u32 hw_mask; - - if (engine_mask == ALL_ENGINES) { - hw_mask = GEN6_GRDOM_FULL; - } else { - unsigned int tmp; - - hw_mask = 0; - for_each_engine_masked(engine, dev_priv, engine_mask, tmp) - hw_mask |= hw_engine_mask[engine->id]; - } - - return gen6_hw_domain_reset(dev_priv, hw_mask); -} - -static u32 gen11_lock_sfc(struct drm_i915_private *dev_priv, - struct intel_engine_cs *engine) -{ - u8 vdbox_sfc_access = RUNTIME_INFO(dev_priv)->vdbox_sfc_access; - i915_reg_t sfc_forced_lock, sfc_forced_lock_ack; - u32 sfc_forced_lock_bit, sfc_forced_lock_ack_bit; - i915_reg_t sfc_usage; - u32 sfc_usage_bit; - u32 sfc_reset_bit; - - switch (engine->class) { - case VIDEO_DECODE_CLASS: - if ((BIT(engine->instance) & vdbox_sfc_access) == 0) - return 0; - - sfc_forced_lock = GEN11_VCS_SFC_FORCED_LOCK(engine); - sfc_forced_lock_bit = GEN11_VCS_SFC_FORCED_LOCK_BIT; - - sfc_forced_lock_ack = GEN11_VCS_SFC_LOCK_STATUS(engine); - sfc_forced_lock_ack_bit = GEN11_VCS_SFC_LOCK_ACK_BIT; - - sfc_usage = GEN11_VCS_SFC_LOCK_STATUS(engine); - sfc_usage_bit = GEN11_VCS_SFC_USAGE_BIT; - sfc_reset_bit = GEN11_VCS_SFC_RESET_BIT(engine->instance); - break; - - case VIDEO_ENHANCEMENT_CLASS: - sfc_forced_lock = GEN11_VECS_SFC_FORCED_LOCK(engine); - sfc_forced_lock_bit = GEN11_VECS_SFC_FORCED_LOCK_BIT; - - sfc_forced_lock_ack = GEN11_VECS_SFC_LOCK_ACK(engine); - sfc_forced_lock_ack_bit = GEN11_VECS_SFC_LOCK_ACK_BIT; - - sfc_usage = GEN11_VECS_SFC_USAGE(engine); - sfc_usage_bit = GEN11_VECS_SFC_USAGE_BIT; - sfc_reset_bit = GEN11_VECS_SFC_RESET_BIT(engine->instance); - break; - - default: - return 0; - } - - /* - * Tell the engine that a software reset is going to happen. The engine - * will then try to force lock the SFC (if currently locked, it will - * remain so until we tell the engine it is safe to unlock; if currently - * unlocked, it will ignore this and all new lock requests). If SFC - * ends up being locked to the engine we want to reset, we have to reset - * it as well (we will unlock it once the reset sequence is completed). - */ - I915_WRITE_FW(sfc_forced_lock, - I915_READ_FW(sfc_forced_lock) | sfc_forced_lock_bit); - - if (__intel_wait_for_register_fw(dev_priv, - sfc_forced_lock_ack, - sfc_forced_lock_ack_bit, - sfc_forced_lock_ack_bit, - 1000, 0, NULL)) { - DRM_DEBUG_DRIVER("Wait for SFC forced lock ack failed\n"); - return 0; - } - - if (I915_READ_FW(sfc_usage) & sfc_usage_bit) - return sfc_reset_bit; - - return 0; -} - -static void gen11_unlock_sfc(struct drm_i915_private *dev_priv, - struct intel_engine_cs *engine) -{ - u8 vdbox_sfc_access = RUNTIME_INFO(dev_priv)->vdbox_sfc_access; - i915_reg_t sfc_forced_lock; - u32 sfc_forced_lock_bit; - - switch (engine->class) { - case VIDEO_DECODE_CLASS: - if ((BIT(engine->instance) & vdbox_sfc_access) == 0) - return; - - sfc_forced_lock = GEN11_VCS_SFC_FORCED_LOCK(engine); - sfc_forced_lock_bit = GEN11_VCS_SFC_FORCED_LOCK_BIT; - break; - - case VIDEO_ENHANCEMENT_CLASS: - sfc_forced_lock = GEN11_VECS_SFC_FORCED_LOCK(engine); - sfc_forced_lock_bit = GEN11_VECS_SFC_FORCED_LOCK_BIT; - break; - - default: - return; - } - - I915_WRITE_FW(sfc_forced_lock, - I915_READ_FW(sfc_forced_lock) & ~sfc_forced_lock_bit); -} - -/** - * gen11_reset_engines - reset individual engines - * @dev_priv: i915 device - * @engine_mask: mask of intel_ring_flag() engines or ALL_ENGINES for full reset - * - * This function will reset the individual engines that are set in engine_mask. - * If you provide ALL_ENGINES as mask, full global domain reset will be issued. - * - * Note: It is responsibility of the caller to handle the difference between - * asking full domain reset versus reset for all available individual engines. - * - * Returns 0 on success, nonzero on error. - */ -static int gen11_reset_engines(struct drm_i915_private *dev_priv, - unsigned int engine_mask) -{ - const u32 hw_engine_mask[I915_NUM_ENGINES] = { - [RCS] = GEN11_GRDOM_RENDER, - [BCS] = GEN11_GRDOM_BLT, - [VCS] = GEN11_GRDOM_MEDIA, - [VCS2] = GEN11_GRDOM_MEDIA2, - [VCS3] = GEN11_GRDOM_MEDIA3, - [VCS4] = GEN11_GRDOM_MEDIA4, - [VECS] = GEN11_GRDOM_VECS, - [VECS2] = GEN11_GRDOM_VECS2, - }; - struct intel_engine_cs *engine; - unsigned int tmp; - u32 hw_mask; - int ret; - - BUILD_BUG_ON(VECS2 + 1 != I915_NUM_ENGINES); - - if (engine_mask == ALL_ENGINES) { - hw_mask = GEN11_GRDOM_FULL; - } else { - hw_mask = 0; - for_each_engine_masked(engine, dev_priv, engine_mask, tmp) { - hw_mask |= hw_engine_mask[engine->id]; - hw_mask |= gen11_lock_sfc(dev_priv, engine); - } - } - - ret = gen6_hw_domain_reset(dev_priv, hw_mask); - - if (engine_mask != ALL_ENGINES) - for_each_engine_masked(engine, dev_priv, engine_mask, tmp) - gen11_unlock_sfc(dev_priv, engine); - return ret; } @@ -2184,202 +1819,15 @@ int __intel_wait_for_register(struct drm_i915_private *dev_priv, (reg_value & mask) == value, slow_timeout_ms * 1000, 10, 1000); + /* just trace the final value */ + trace_i915_reg_rw(false, reg, reg_value, sizeof(reg_value), true); + if (out_value) *out_value = reg_value; return ret; } -static int gen8_engine_reset_prepare(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - int ret; - - I915_WRITE_FW(RING_RESET_CTL(engine->mmio_base), - _MASKED_BIT_ENABLE(RESET_CTL_REQUEST_RESET)); - - ret = __intel_wait_for_register_fw(dev_priv, - RING_RESET_CTL(engine->mmio_base), - RESET_CTL_READY_TO_RESET, - RESET_CTL_READY_TO_RESET, - 700, 0, - NULL); - if (ret) - DRM_ERROR("%s: reset request timeout\n", engine->name); - - return ret; -} - -static void gen8_engine_reset_cancel(struct intel_engine_cs *engine) -{ - struct drm_i915_private *dev_priv = engine->i915; - - I915_WRITE_FW(RING_RESET_CTL(engine->mmio_base), - _MASKED_BIT_DISABLE(RESET_CTL_REQUEST_RESET)); -} - -static int reset_engines(struct drm_i915_private *i915, - unsigned int engine_mask, - unsigned int retry) -{ - if (INTEL_GEN(i915) >= 11) - return gen11_reset_engines(i915, engine_mask); - else - return gen6_reset_engines(i915, engine_mask, retry); -} - -static int gen8_reset_engines(struct drm_i915_private *dev_priv, - unsigned int engine_mask, - unsigned int retry) -{ - struct intel_engine_cs *engine; - const bool reset_non_ready = retry >= 1; - unsigned int tmp; - int ret; - - for_each_engine_masked(engine, dev_priv, engine_mask, tmp) { - ret = gen8_engine_reset_prepare(engine); - if (ret && !reset_non_ready) - goto skip_reset; - - /* - * If this is not the first failed attempt to prepare, - * we decide to proceed anyway. - * - * By doing so we risk context corruption and with - * some gens (kbl), possible system hang if reset - * happens during active bb execution. - * - * We rather take context corruption instead of - * failed reset with a wedged driver/gpu. And - * active bb execution case should be covered by - * i915_stop_engines we have before the reset. - */ - } - - ret = reset_engines(dev_priv, engine_mask, retry); - -skip_reset: - for_each_engine_masked(engine, dev_priv, engine_mask, tmp) - gen8_engine_reset_cancel(engine); - - return ret; -} - -typedef int (*reset_func)(struct drm_i915_private *, - unsigned int engine_mask, unsigned int retry); - -static reset_func intel_get_gpu_reset(struct drm_i915_private *dev_priv) -{ - if (!i915_modparams.reset) - return NULL; - - if (INTEL_GEN(dev_priv) >= 8) - return gen8_reset_engines; - else if (INTEL_GEN(dev_priv) >= 6) - return gen6_reset_engines; - else if (IS_GEN(dev_priv, 5)) - return ironlake_do_reset; - else if (IS_G4X(dev_priv)) - return g4x_do_reset; - else if (IS_G33(dev_priv) || IS_PINEVIEW(dev_priv)) - return g33_do_reset; - else if (INTEL_GEN(dev_priv) >= 3) - return i915_do_reset; - else - return NULL; -} - -int intel_gpu_reset(struct drm_i915_private *dev_priv, - const unsigned int engine_mask) -{ - reset_func reset = intel_get_gpu_reset(dev_priv); - unsigned int retry; - int ret; - - GEM_BUG_ON(!engine_mask); - - /* - * We want to perform per-engine reset from atomic context (e.g. - * softirq), which imposes the constraint that we cannot sleep. - * However, experience suggests that spending a bit of time waiting - * for a reset helps in various cases, so for a full-device reset - * we apply the opposite rule and wait if we want to. As we should - * always follow up a failed per-engine reset with a full device reset, - * being a little faster, stricter and more error prone for the - * atomic case seems an acceptable compromise. - * - * Unfortunately this leads to a bimodal routine, when the goal was - * to have a single reset function that worked for resetting any - * number of engines simultaneously. - */ - might_sleep_if(engine_mask == ALL_ENGINES); - - /* - * If the power well sleeps during the reset, the reset - * request may be dropped and never completes (causing -EIO). - */ - intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); - for (retry = 0; retry < 3; retry++) { - - /* - * We stop engines, otherwise we might get failed reset and a - * dead gpu (on elk). Also as modern gpu as kbl can suffer - * from system hang if batchbuffer is progressing when - * the reset is issued, regardless of READY_TO_RESET ack. - * Thus assume it is best to stop engines on all gens - * where we have a gpu reset. - * - * WaKBLVECSSemaphoreWaitPoll:kbl (on ALL_ENGINES) - * - * WaMediaResetMainRingCleanup:ctg,elk (presumably) - * - * FIXME: Wa for more modern gens needs to be validated - */ - i915_stop_engines(dev_priv, engine_mask); - - ret = -ENODEV; - if (reset) { - ret = reset(dev_priv, engine_mask, retry); - GEM_TRACE("engine_mask=%x, ret=%d, retry=%d\n", - engine_mask, ret, retry); - } - if (ret != -ETIMEDOUT || engine_mask != ALL_ENGINES) - break; - - cond_resched(); - } - intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); - - return ret; -} - -bool intel_has_gpu_reset(struct drm_i915_private *dev_priv) -{ - return intel_get_gpu_reset(dev_priv) != NULL; -} - -bool intel_has_reset_engine(struct drm_i915_private *dev_priv) -{ - return (INTEL_INFO(dev_priv)->has_reset_engine && - i915_modparams.reset >= 2); -} - -int intel_reset_guc(struct drm_i915_private *dev_priv) -{ - u32 guc_domain = INTEL_GEN(dev_priv) >= 11 ? GEN11_GRDOM_GUC : - GEN9_GRDOM_GUC; - int ret; - - GEM_BUG_ON(!HAS_GUC(dev_priv)); - - intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL); - ret = gen6_hw_domain_reset(dev_priv, guc_domain); - intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL); - - return ret; -} - bool intel_uncore_unclaimed_mmio(struct drm_i915_private *dev_priv) { return check_for_unclaimed_mmio(dev_priv); diff --git a/drivers/gpu/drm/i915/intel_vdsc.c b/drivers/gpu/drm/i915/intel_vdsc.c index 48537827616f..23abf03736e7 100644 --- a/drivers/gpu/drm/i915/intel_vdsc.c +++ b/drivers/gpu/drm/i915/intel_vdsc.c @@ -1082,6 +1082,6 @@ void intel_dsc_disable(const struct intel_crtc_state *old_crtc_state) I915_WRITE(dss_ctl2_reg, dss_ctl2_val); /* Disable Power wells for VDSC/joining */ - intel_display_power_put(dev_priv, - intel_dsc_power_domain(old_crtc_state)); + intel_display_power_put_unchecked(dev_priv, + intel_dsc_power_domain(old_crtc_state)); } diff --git a/drivers/gpu/drm/i915/intel_workarounds.c b/drivers/gpu/drm/i915/intel_workarounds.c index 3210ad4e08f7..15f4a6dee5aa 100644 --- a/drivers/gpu/drm/i915/intel_workarounds.c +++ b/drivers/gpu/drm/i915/intel_workarounds.c @@ -142,7 +142,8 @@ static void _wa_add(struct i915_wa_list *wal, const struct i915_wa *wa) } static void -__wa_add(struct i915_wa_list *wal, i915_reg_t reg, u32 mask, u32 val) +wa_write_masked_or(struct i915_wa_list *wal, i915_reg_t reg, u32 mask, + u32 val) { struct i915_wa wa = { .reg = reg, @@ -153,16 +154,32 @@ __wa_add(struct i915_wa_list *wal, i915_reg_t reg, u32 mask, u32 val) _wa_add(wal, &wa); } -#define WA_REG(addr, mask, val) __wa_add(wal, (addr), (mask), (val)) +static void +wa_masked_en(struct i915_wa_list *wal, i915_reg_t reg, u32 val) +{ + wa_write_masked_or(wal, reg, val, _MASKED_BIT_ENABLE(val)); +} + +static void +wa_write(struct i915_wa_list *wal, i915_reg_t reg, u32 val) +{ + wa_write_masked_or(wal, reg, ~0, val); +} + +static void +wa_write_or(struct i915_wa_list *wal, i915_reg_t reg, u32 val) +{ + wa_write_masked_or(wal, reg, val, val); +} #define WA_SET_BIT_MASKED(addr, mask) \ - WA_REG(addr, (mask), _MASKED_BIT_ENABLE(mask)) + wa_write_masked_or(wal, (addr), (mask), _MASKED_BIT_ENABLE(mask)) #define WA_CLR_BIT_MASKED(addr, mask) \ - WA_REG(addr, (mask), _MASKED_BIT_DISABLE(mask)) + wa_write_masked_or(wal, (addr), (mask), _MASKED_BIT_DISABLE(mask)) #define WA_SET_FIELD_MASKED(addr, mask, value) \ - WA_REG(addr, (mask), _MASKED_FIELD(mask, value)) + wa_write_masked_or(wal, (addr), (mask), _MASKED_FIELD((mask), (value))) static void gen8_ctx_workarounds_init(struct intel_engine_cs *engine) { @@ -532,6 +549,12 @@ static void icl_ctx_workarounds_init(struct intel_engine_cs *engine) if (IS_ICL_REVID(i915, ICL_REVID_A0, ICL_REVID_A0)) WA_SET_BIT_MASKED(GEN11_COMMON_SLICE_CHICKEN3, GEN11_BLEND_EMB_FIX_DISABLE_IN_RCC); + + /* WaEnableFloatBlendOptimization:icl */ + wa_write_masked_or(wal, + GEN10_CACHE_MODE_SS, + 0, /* write-only, so skip validation */ + _MASKED_BIT_ENABLE(FLOAT_BLEND_OPTIMIZATION_ENABLE)); } void intel_engine_init_ctx_wa(struct intel_engine_cs *engine) @@ -603,43 +626,6 @@ int intel_engine_emit_ctx_wa(struct i915_request *rq) } static void -wa_masked_en(struct i915_wa_list *wal, i915_reg_t reg, u32 val) -{ - struct i915_wa wa = { - .reg = reg, - .mask = val, - .val = _MASKED_BIT_ENABLE(val) - }; - - _wa_add(wal, &wa); -} - -static void -wa_write_masked_or(struct i915_wa_list *wal, i915_reg_t reg, u32 mask, - u32 val) -{ - struct i915_wa wa = { - .reg = reg, - .mask = mask, - .val = val - }; - - _wa_add(wal, &wa); -} - -static void -wa_write(struct i915_wa_list *wal, i915_reg_t reg, u32 val) -{ - wa_write_masked_or(wal, reg, ~0, val); -} - -static void -wa_write_or(struct i915_wa_list *wal, i915_reg_t reg, u32 val) -{ - wa_write_masked_or(wal, reg, val, val); -} - -static void gen9_gt_workarounds_init(struct drm_i915_private *i915, struct i915_wa_list *wal) { /* WaDisableKillLogic:bxt,skl,kbl */ diff --git a/drivers/gpu/drm/i915/selftests/huge_pages.c b/drivers/gpu/drm/i915/selftests/huge_pages.c index 6c10734e948d..a9a2fa35876f 100644 --- a/drivers/gpu/drm/i915/selftests/huge_pages.c +++ b/drivers/gpu/drm/i915/selftests/huge_pages.c @@ -1449,7 +1449,7 @@ static int igt_ppgtt_pin_update(void *arg) * huge-gtt-pages. */ - if (!HAS_FULL_48BIT_PPGTT(dev_priv)) { + if (!ppgtt || !i915_vm_is_48bit(&ppgtt->vm)) { pr_info("48b PPGTT not supported, skipping\n"); return 0; } @@ -1756,6 +1756,7 @@ int i915_gem_huge_page_live_selftests(struct drm_i915_private *dev_priv) }; struct drm_file *file; struct i915_gem_context *ctx; + intel_wakeref_t wakeref; int err; if (!HAS_PPGTT(dev_priv)) { @@ -1771,7 +1772,7 @@ int i915_gem_huge_page_live_selftests(struct drm_i915_private *dev_priv) return PTR_ERR(file); mutex_lock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); ctx = live_context(dev_priv, file); if (IS_ERR(ctx)) { @@ -1785,7 +1786,7 @@ int i915_gem_huge_page_live_selftests(struct drm_i915_private *dev_priv) err = i915_subtests(tests, ctx); out_unlock: - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); mutex_unlock(&dev_priv->drm.struct_mutex); mock_file_free(dev_priv, file); diff --git a/drivers/gpu/drm/i915/selftests/i915_active.c b/drivers/gpu/drm/i915/selftests/i915_active.c new file mode 100644 index 000000000000..337b1f98b923 --- /dev/null +++ b/drivers/gpu/drm/i915/selftests/i915_active.c @@ -0,0 +1,157 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2018 Intel Corporation + */ + +#include "../i915_selftest.h" + +#include "igt_flush_test.h" +#include "lib_sw_fence.h" + +struct live_active { + struct i915_active base; + bool retired; +}; + +static void __live_active_retire(struct i915_active *base) +{ + struct live_active *active = container_of(base, typeof(*active), base); + + active->retired = true; +} + +static int __live_active_setup(struct drm_i915_private *i915, + struct live_active *active) +{ + struct intel_engine_cs *engine; + struct i915_sw_fence *submit; + enum intel_engine_id id; + unsigned int count = 0; + int err = 0; + + submit = heap_fence_create(GFP_KERNEL); + if (!submit) + return -ENOMEM; + + i915_active_init(i915, &active->base, __live_active_retire); + active->retired = false; + + if (!i915_active_acquire(&active->base)) { + pr_err("First i915_active_acquire should report being idle\n"); + err = -EINVAL; + goto out; + } + + for_each_engine(engine, i915, id) { + struct i915_request *rq; + + rq = i915_request_alloc(engine, i915->kernel_context); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + break; + } + + err = i915_sw_fence_await_sw_fence_gfp(&rq->submit, + submit, + GFP_KERNEL); + if (err >= 0) + err = i915_active_ref(&active->base, + rq->fence.context, rq); + i915_request_add(rq); + if (err) { + pr_err("Failed to track active ref!\n"); + break; + } + + count++; + } + + i915_active_release(&active->base); + if (active->retired && count) { + pr_err("i915_active retired before submission!\n"); + err = -EINVAL; + } + if (active->base.count != count) { + pr_err("i915_active not tracking all requests, found %d, expected %d\n", + active->base.count, count); + err = -EINVAL; + } + +out: + i915_sw_fence_commit(submit); + heap_fence_put(submit); + + return err; +} + +static int live_active_wait(void *arg) +{ + struct drm_i915_private *i915 = arg; + struct live_active active; + intel_wakeref_t wakeref; + int err; + + /* Check that we get a callback when requests retire upon waiting */ + + mutex_lock(&i915->drm.struct_mutex); + wakeref = intel_runtime_pm_get(i915); + + err = __live_active_setup(i915, &active); + + i915_active_wait(&active.base); + if (!active.retired) { + pr_err("i915_active not retired after waiting!\n"); + err = -EINVAL; + } + + i915_active_fini(&active.base); + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + err = -EIO; + + intel_runtime_pm_put(i915, wakeref); + mutex_unlock(&i915->drm.struct_mutex); + return err; +} + +static int live_active_retire(void *arg) +{ + struct drm_i915_private *i915 = arg; + struct live_active active; + intel_wakeref_t wakeref; + int err; + + /* Check that we get a callback when requests are indirectly retired */ + + mutex_lock(&i915->drm.struct_mutex); + wakeref = intel_runtime_pm_get(i915); + + err = __live_active_setup(i915, &active); + + /* waits for & retires all requests */ + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + err = -EIO; + + if (!active.retired) { + pr_err("i915_active not retired after flushing!\n"); + err = -EINVAL; + } + + i915_active_fini(&active.base); + intel_runtime_pm_put(i915, wakeref); + mutex_unlock(&i915->drm.struct_mutex); + return err; +} + +int i915_active_live_selftests(struct drm_i915_private *i915) +{ + static const struct i915_subtest tests[] = { + SUBTEST(live_active_wait), + SUBTEST(live_active_retire), + }; + + if (i915_terminally_wedged(&i915->gpu_error)) + return 0; + + return i915_subtests(tests, i915); +} diff --git a/drivers/gpu/drm/i915/selftests/i915_gem.c b/drivers/gpu/drm/i915/selftests/i915_gem.c index bdcc53e15e75..e77b7ed449ae 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem.c +++ b/drivers/gpu/drm/i915/selftests/i915_gem.c @@ -16,9 +16,10 @@ static int switch_to_context(struct drm_i915_private *i915, { struct intel_engine_cs *engine; enum intel_engine_id id; + intel_wakeref_t wakeref; int err = 0; - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); for_each_engine(engine, i915, id) { struct i915_request *rq; @@ -32,7 +33,7 @@ static int switch_to_context(struct drm_i915_private *i915, i915_request_add(rq); } - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); return err; } @@ -65,7 +66,9 @@ static void trash_stolen(struct drm_i915_private *i915) static void simulate_hibernate(struct drm_i915_private *i915) { - intel_runtime_pm_get(i915); + intel_wakeref_t wakeref; + + wakeref = intel_runtime_pm_get(i915); /* * As a final sting in the tail, invalidate stolen. Under a real S4, @@ -76,7 +79,7 @@ static void simulate_hibernate(struct drm_i915_private *i915) */ trash_stolen(i915); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); } static int pm_prepare(struct drm_i915_private *i915) @@ -93,39 +96,39 @@ static int pm_prepare(struct drm_i915_private *i915) static void pm_suspend(struct drm_i915_private *i915) { - intel_runtime_pm_get(i915); - - i915_gem_suspend_gtt_mappings(i915); - i915_gem_suspend_late(i915); + intel_wakeref_t wakeref; - intel_runtime_pm_put(i915); + with_intel_runtime_pm(i915, wakeref) { + i915_gem_suspend_gtt_mappings(i915); + i915_gem_suspend_late(i915); + } } static void pm_hibernate(struct drm_i915_private *i915) { - intel_runtime_pm_get(i915); + intel_wakeref_t wakeref; - i915_gem_suspend_gtt_mappings(i915); + with_intel_runtime_pm(i915, wakeref) { + i915_gem_suspend_gtt_mappings(i915); - i915_gem_freeze(i915); - i915_gem_freeze_late(i915); - - intel_runtime_pm_put(i915); + i915_gem_freeze(i915); + i915_gem_freeze_late(i915); + } } static void pm_resume(struct drm_i915_private *i915) { + intel_wakeref_t wakeref; + /* * Both suspend and hibernate follow the same wakeup path and assume * that runtime-pm just works. */ - intel_runtime_pm_get(i915); - - intel_engines_sanitize(i915, false); - i915_gem_sanitize(i915); - i915_gem_resume(i915); - - intel_runtime_pm_put(i915); + with_intel_runtime_pm(i915, wakeref) { + intel_engines_sanitize(i915, false); + i915_gem_sanitize(i915); + i915_gem_resume(i915); + } } static int igt_gem_suspend(void *arg) diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_coherency.c b/drivers/gpu/drm/i915/selftests/i915_gem_coherency.c index f7392c1ffe75..fd89a5a33c1a 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem_coherency.c +++ b/drivers/gpu/drm/i915/selftests/i915_gem_coherency.c @@ -279,6 +279,7 @@ static int igt_gem_coherency(void *arg) struct drm_i915_private *i915 = arg; const struct igt_coherency_mode *read, *write, *over; struct drm_i915_gem_object *obj; + intel_wakeref_t wakeref; unsigned long count, n; u32 *offsets, *values; int err = 0; @@ -298,7 +299,7 @@ static int igt_gem_coherency(void *arg) values = offsets + ncachelines; mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); for (over = igt_coherency_mode; over->name; over++) { if (!over->set) continue; @@ -376,7 +377,7 @@ static int igt_gem_coherency(void *arg) } } unlock: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); kfree(offsets); return err; diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_context.c b/drivers/gpu/drm/i915/selftests/i915_gem_context.c index d00cdf3c2939..d00d0bb07784 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem_context.c +++ b/drivers/gpu/drm/i915/selftests/i915_gem_context.c @@ -24,9 +24,13 @@ #include <linux/prime_numbers.h> +#include "../i915_reset.h" #include "../i915_selftest.h" #include "i915_random.h" #include "igt_flush_test.h" +#include "igt_live_test.h" +#include "igt_reset.h" +#include "igt_spinner.h" #include "mock_drm.h" #include "mock_gem_device.h" @@ -34,84 +38,6 @@ #define DW_PER_PAGE (PAGE_SIZE / sizeof(u32)) -struct live_test { - struct drm_i915_private *i915; - const char *func; - const char *name; - - unsigned int reset_global; - unsigned int reset_engine[I915_NUM_ENGINES]; -}; - -static int begin_live_test(struct live_test *t, - struct drm_i915_private *i915, - const char *func, - const char *name) -{ - struct intel_engine_cs *engine; - enum intel_engine_id id; - int err; - - t->i915 = i915; - t->func = func; - t->name = name; - - err = i915_gem_wait_for_idle(i915, - I915_WAIT_LOCKED, - MAX_SCHEDULE_TIMEOUT); - if (err) { - pr_err("%s(%s): failed to idle before, with err=%d!", - func, name, err); - return err; - } - - i915->gpu_error.missed_irq_rings = 0; - t->reset_global = i915_reset_count(&i915->gpu_error); - - for_each_engine(engine, i915, id) - t->reset_engine[id] = - i915_reset_engine_count(&i915->gpu_error, engine); - - return 0; -} - -static int end_live_test(struct live_test *t) -{ - struct drm_i915_private *i915 = t->i915; - struct intel_engine_cs *engine; - enum intel_engine_id id; - - if (igt_flush_test(i915, I915_WAIT_LOCKED)) - return -EIO; - - if (t->reset_global != i915_reset_count(&i915->gpu_error)) { - pr_err("%s(%s): GPU was reset %d times!\n", - t->func, t->name, - i915_reset_count(&i915->gpu_error) - t->reset_global); - return -EIO; - } - - for_each_engine(engine, i915, id) { - if (t->reset_engine[id] == - i915_reset_engine_count(&i915->gpu_error, engine)) - continue; - - pr_err("%s(%s): engine '%s' was reset %d times!\n", - t->func, t->name, engine->name, - i915_reset_engine_count(&i915->gpu_error, engine) - - t->reset_engine[id]); - return -EIO; - } - - if (i915->gpu_error.missed_irq_rings) { - pr_err("%s(%s): Missed interrupts on engines %lx\n", - t->func, t->name, i915->gpu_error.missed_irq_rings); - return -EIO; - } - - return 0; -} - static int live_nop_switch(void *arg) { const unsigned int nctx = 1024; @@ -119,8 +45,9 @@ static int live_nop_switch(void *arg) struct intel_engine_cs *engine; struct i915_gem_context **ctx; enum intel_engine_id id; + intel_wakeref_t wakeref; + struct igt_live_test t; struct drm_file *file; - struct live_test t; unsigned long n; int err = -ENODEV; @@ -140,7 +67,7 @@ static int live_nop_switch(void *arg) return PTR_ERR(file); mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); ctx = kcalloc(nctx, sizeof(*ctx), GFP_KERNEL); if (!ctx) { @@ -184,7 +111,7 @@ static int live_nop_switch(void *arg) pr_info("Populated %d contexts on %s in %lluns\n", nctx, engine->name, ktime_to_ns(times[1] - times[0])); - err = begin_live_test(&t, i915, __func__, engine->name); + err = igt_live_test_begin(&t, i915, __func__, engine->name); if (err) goto out_unlock; @@ -232,7 +159,7 @@ static int live_nop_switch(void *arg) break; } - err = end_live_test(&t); + err = igt_live_test_end(&t); if (err) goto out_unlock; @@ -243,7 +170,7 @@ static int live_nop_switch(void *arg) } out_unlock: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); mock_file_free(i915, file); return err; @@ -553,10 +480,10 @@ static int igt_ctx_exec(void *arg) struct drm_i915_private *i915 = arg; struct drm_i915_gem_object *obj = NULL; unsigned long ncontexts, ndwords, dw; + struct igt_live_test t; struct drm_file *file; IGT_TIMEOUT(end_time); LIST_HEAD(objects); - struct live_test t; int err = -ENODEV; /* @@ -574,7 +501,7 @@ static int igt_ctx_exec(void *arg) mutex_lock(&i915->drm.struct_mutex); - err = begin_live_test(&t, i915, __func__, ""); + err = igt_live_test_begin(&t, i915, __func__, ""); if (err) goto out_unlock; @@ -593,6 +520,8 @@ static int igt_ctx_exec(void *arg) } for_each_engine(engine, i915, id) { + intel_wakeref_t wakeref; + if (!engine->context_size) continue; /* No logical context support in HW */ @@ -607,9 +536,9 @@ static int igt_ctx_exec(void *arg) } } - intel_runtime_pm_get(i915); - err = gpu_fill(obj, ctx, engine, dw); - intel_runtime_pm_put(i915); + err = 0; + with_intel_runtime_pm(i915, wakeref) + err = gpu_fill(obj, ctx, engine, dw); if (err) { pr_err("Failed to fill dword %lu [%lu/%lu] with gpu (%s) in ctx %u [full-ppgtt? %s], err=%d\n", ndwords, dw, max_dwords(obj), @@ -642,7 +571,7 @@ static int igt_ctx_exec(void *arg) } out_unlock: - if (end_live_test(&t)) + if (igt_live_test_end(&t)) err = -EIO; mutex_unlock(&i915->drm.struct_mutex); @@ -650,6 +579,469 @@ out_unlock: return err; } +static struct i915_vma *rpcs_query_batch(struct i915_vma *vma) +{ + struct drm_i915_gem_object *obj; + u32 *cmd; + int err; + + if (INTEL_GEN(vma->vm->i915) < 8) + return ERR_PTR(-EINVAL); + + obj = i915_gem_object_create_internal(vma->vm->i915, PAGE_SIZE); + if (IS_ERR(obj)) + return ERR_CAST(obj); + + cmd = i915_gem_object_pin_map(obj, I915_MAP_WB); + if (IS_ERR(cmd)) { + err = PTR_ERR(cmd); + goto err; + } + + *cmd++ = MI_STORE_REGISTER_MEM_GEN8; + *cmd++ = i915_mmio_reg_offset(GEN8_R_PWR_CLK_STATE); + *cmd++ = lower_32_bits(vma->node.start); + *cmd++ = upper_32_bits(vma->node.start); + *cmd = MI_BATCH_BUFFER_END; + + i915_gem_object_unpin_map(obj); + + err = i915_gem_object_set_to_gtt_domain(obj, false); + if (err) + goto err; + + vma = i915_vma_instance(obj, vma->vm, NULL); + if (IS_ERR(vma)) { + err = PTR_ERR(vma); + goto err; + } + + err = i915_vma_pin(vma, 0, 0, PIN_USER); + if (err) + goto err; + + return vma; + +err: + i915_gem_object_put(obj); + return ERR_PTR(err); +} + +static int +emit_rpcs_query(struct drm_i915_gem_object *obj, + struct i915_gem_context *ctx, + struct intel_engine_cs *engine, + struct i915_request **rq_out) +{ + struct i915_request *rq; + struct i915_vma *batch; + struct i915_vma *vma; + int err; + + GEM_BUG_ON(!intel_engine_can_store_dword(engine)); + + vma = i915_vma_instance(obj, &ctx->ppgtt->vm, NULL); + if (IS_ERR(vma)) + return PTR_ERR(vma); + + err = i915_gem_object_set_to_gtt_domain(obj, false); + if (err) + return err; + + err = i915_vma_pin(vma, 0, 0, PIN_USER); + if (err) + return err; + + batch = rpcs_query_batch(vma); + if (IS_ERR(batch)) { + err = PTR_ERR(batch); + goto err_vma; + } + + rq = i915_request_alloc(engine, ctx); + if (IS_ERR(rq)) { + err = PTR_ERR(rq); + goto err_batch; + } + + err = engine->emit_bb_start(rq, batch->node.start, batch->node.size, 0); + if (err) + goto err_request; + + err = i915_vma_move_to_active(batch, rq, 0); + if (err) + goto skip_request; + + err = i915_vma_move_to_active(vma, rq, EXEC_OBJECT_WRITE); + if (err) + goto skip_request; + + i915_gem_object_set_active_reference(batch->obj); + i915_vma_unpin(batch); + i915_vma_close(batch); + + i915_vma_unpin(vma); + + *rq_out = i915_request_get(rq); + + i915_request_add(rq); + + return 0; + +skip_request: + i915_request_skip(rq, err); +err_request: + i915_request_add(rq); +err_batch: + i915_vma_unpin(batch); +err_vma: + i915_vma_unpin(vma); + + return err; +} + +#define TEST_IDLE BIT(0) +#define TEST_BUSY BIT(1) +#define TEST_RESET BIT(2) + +static int +__sseu_prepare(struct drm_i915_private *i915, + const char *name, + unsigned int flags, + struct i915_gem_context *ctx, + struct intel_engine_cs *engine, + struct igt_spinner **spin_out) +{ + int ret = 0; + + if (flags & (TEST_BUSY | TEST_RESET)) { + struct igt_spinner *spin; + struct i915_request *rq; + + spin = kzalloc(sizeof(*spin), GFP_KERNEL); + if (!spin) { + ret = -ENOMEM; + goto out; + } + + ret = igt_spinner_init(spin, i915); + if (ret) + return ret; + + rq = igt_spinner_create_request(spin, ctx, engine, MI_NOOP); + if (IS_ERR(rq)) { + ret = PTR_ERR(rq); + igt_spinner_fini(spin); + kfree(spin); + goto out; + } + + i915_request_add(rq); + + if (!igt_wait_for_spinner(spin, rq)) { + pr_err("%s: Spinner failed to start!\n", name); + igt_spinner_end(spin); + igt_spinner_fini(spin); + kfree(spin); + ret = -ETIMEDOUT; + goto out; + } + + *spin_out = spin; + } + +out: + return ret; +} + +static int +__read_slice_count(struct drm_i915_private *i915, + struct i915_gem_context *ctx, + struct intel_engine_cs *engine, + struct drm_i915_gem_object *obj, + struct igt_spinner *spin, + u32 *rpcs) +{ + struct i915_request *rq = NULL; + u32 s_mask, s_shift; + unsigned int cnt; + u32 *buf, val; + long ret; + + ret = emit_rpcs_query(obj, ctx, engine, &rq); + if (ret) + return ret; + + if (spin) + igt_spinner_end(spin); + + ret = i915_request_wait(rq, I915_WAIT_LOCKED, MAX_SCHEDULE_TIMEOUT); + i915_request_put(rq); + if (ret < 0) + return ret; + + buf = i915_gem_object_pin_map(obj, I915_MAP_WB); + if (IS_ERR(buf)) { + ret = PTR_ERR(buf); + return ret; + } + + if (INTEL_GEN(i915) >= 11) { + s_mask = GEN11_RPCS_S_CNT_MASK; + s_shift = GEN11_RPCS_S_CNT_SHIFT; + } else { + s_mask = GEN8_RPCS_S_CNT_MASK; + s_shift = GEN8_RPCS_S_CNT_SHIFT; + } + + val = *buf; + cnt = (val & s_mask) >> s_shift; + *rpcs = val; + + i915_gem_object_unpin_map(obj); + + return cnt; +} + +static int +__check_rpcs(const char *name, u32 rpcs, int slices, unsigned int expected, + const char *prefix, const char *suffix) +{ + if (slices == expected) + return 0; + + if (slices < 0) { + pr_err("%s: %s read slice count failed with %d%s\n", + name, prefix, slices, suffix); + return slices; + } + + pr_err("%s: %s slice count %d is not %u%s\n", + name, prefix, slices, expected, suffix); + + pr_info("RPCS=0x%x; %u%sx%u%s\n", + rpcs, slices, + (rpcs & GEN8_RPCS_S_CNT_ENABLE) ? "*" : "", + (rpcs & GEN8_RPCS_SS_CNT_MASK) >> GEN8_RPCS_SS_CNT_SHIFT, + (rpcs & GEN8_RPCS_SS_CNT_ENABLE) ? "*" : ""); + + return -EINVAL; +} + +static int +__sseu_finish(struct drm_i915_private *i915, + const char *name, + unsigned int flags, + struct i915_gem_context *ctx, + struct i915_gem_context *kctx, + struct intel_engine_cs *engine, + struct drm_i915_gem_object *obj, + unsigned int expected, + struct igt_spinner *spin) +{ + unsigned int slices = + hweight32(intel_device_default_sseu(i915).slice_mask); + u32 rpcs = 0; + int ret = 0; + + if (flags & TEST_RESET) { + ret = i915_reset_engine(engine, "sseu"); + if (ret) + goto out; + } + + ret = __read_slice_count(i915, ctx, engine, obj, + flags & TEST_RESET ? NULL : spin, &rpcs); + ret = __check_rpcs(name, rpcs, ret, expected, "Context", "!"); + if (ret) + goto out; + + ret = __read_slice_count(i915, kctx, engine, obj, NULL, &rpcs); + ret = __check_rpcs(name, rpcs, ret, slices, "Kernel context", "!"); + +out: + if (spin) + igt_spinner_end(spin); + + if ((flags & TEST_IDLE) && ret == 0) { + ret = i915_gem_wait_for_idle(i915, + I915_WAIT_LOCKED, + MAX_SCHEDULE_TIMEOUT); + if (ret) + return ret; + + ret = __read_slice_count(i915, ctx, engine, obj, NULL, &rpcs); + ret = __check_rpcs(name, rpcs, ret, expected, + "Context", " after idle!"); + } + + return ret; +} + +static int +__sseu_test(struct drm_i915_private *i915, + const char *name, + unsigned int flags, + struct i915_gem_context *ctx, + struct intel_engine_cs *engine, + struct drm_i915_gem_object *obj, + struct intel_sseu sseu) +{ + struct igt_spinner *spin = NULL; + struct i915_gem_context *kctx; + int ret; + + kctx = kernel_context(i915); + if (IS_ERR(kctx)) + return PTR_ERR(kctx); + + ret = __sseu_prepare(i915, name, flags, ctx, engine, &spin); + if (ret) + goto out; + + ret = __i915_gem_context_reconfigure_sseu(ctx, engine, sseu); + if (ret) + goto out; + + ret = __sseu_finish(i915, name, flags, ctx, kctx, engine, obj, + hweight32(sseu.slice_mask), spin); + +out: + if (spin) { + igt_spinner_end(spin); + igt_spinner_fini(spin); + kfree(spin); + } + + kernel_context_close(kctx); + + return ret; +} + +static int +__igt_ctx_sseu(struct drm_i915_private *i915, + const char *name, + unsigned int flags) +{ + struct intel_sseu default_sseu = intel_device_default_sseu(i915); + struct intel_engine_cs *engine = i915->engine[RCS]; + struct drm_i915_gem_object *obj; + struct i915_gem_context *ctx; + struct intel_sseu pg_sseu; + intel_wakeref_t wakeref; + struct drm_file *file; + int ret; + + if (INTEL_GEN(i915) < 9) + return 0; + + if (!RUNTIME_INFO(i915)->sseu.has_slice_pg) + return 0; + + if (hweight32(default_sseu.slice_mask) < 2) + return 0; + + /* + * Gen11 VME friendly power-gated configuration with half enabled + * sub-slices. + */ + pg_sseu = default_sseu; + pg_sseu.slice_mask = 1; + pg_sseu.subslice_mask = + ~(~0 << (hweight32(default_sseu.subslice_mask) / 2)); + + pr_info("SSEU subtest '%s', flags=%x, def_slices=%u, pg_slices=%u\n", + name, flags, hweight32(default_sseu.slice_mask), + hweight32(pg_sseu.slice_mask)); + + file = mock_file(i915); + if (IS_ERR(file)) + return PTR_ERR(file); + + if (flags & TEST_RESET) + igt_global_reset_lock(i915); + + mutex_lock(&i915->drm.struct_mutex); + + ctx = i915_gem_create_context(i915, file->driver_priv); + if (IS_ERR(ctx)) { + ret = PTR_ERR(ctx); + goto out_unlock; + } + + obj = i915_gem_object_create_internal(i915, PAGE_SIZE); + if (IS_ERR(obj)) { + ret = PTR_ERR(obj); + goto out_unlock; + } + + wakeref = intel_runtime_pm_get(i915); + + /* First set the default mask. */ + ret = __sseu_test(i915, name, flags, ctx, engine, obj, default_sseu); + if (ret) + goto out_fail; + + /* Then set a power-gated configuration. */ + ret = __sseu_test(i915, name, flags, ctx, engine, obj, pg_sseu); + if (ret) + goto out_fail; + + /* Back to defaults. */ + ret = __sseu_test(i915, name, flags, ctx, engine, obj, default_sseu); + if (ret) + goto out_fail; + + /* One last power-gated configuration for the road. */ + ret = __sseu_test(i915, name, flags, ctx, engine, obj, pg_sseu); + if (ret) + goto out_fail; + +out_fail: + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + ret = -EIO; + + i915_gem_object_put(obj); + + intel_runtime_pm_put(i915, wakeref); + +out_unlock: + mutex_unlock(&i915->drm.struct_mutex); + + if (flags & TEST_RESET) + igt_global_reset_unlock(i915); + + mock_file_free(i915, file); + + if (ret) + pr_err("%s: Failed with %d!\n", name, ret); + + return ret; +} + +static int igt_ctx_sseu(void *arg) +{ + struct { + const char *name; + unsigned int flags; + } *phase, phases[] = { + { .name = "basic", .flags = 0 }, + { .name = "idle", .flags = TEST_IDLE }, + { .name = "busy", .flags = TEST_BUSY }, + { .name = "busy-reset", .flags = TEST_BUSY | TEST_RESET }, + { .name = "busy-idle", .flags = TEST_BUSY | TEST_IDLE }, + { .name = "reset-idle", .flags = TEST_RESET | TEST_IDLE }, + }; + unsigned int i; + int ret = 0; + + for (i = 0, phase = phases; ret == 0 && i < ARRAY_SIZE(phases); + i++, phase++) + ret = __igt_ctx_sseu(arg, phase->name, phase->flags); + + return ret; +} + static int igt_ctx_readonly(void *arg) { struct drm_i915_private *i915 = arg; @@ -657,11 +1049,11 @@ static int igt_ctx_readonly(void *arg) struct i915_gem_context *ctx; struct i915_hw_ppgtt *ppgtt; unsigned long ndwords, dw; + struct igt_live_test t; struct drm_file *file; I915_RND_STATE(prng); IGT_TIMEOUT(end_time); LIST_HEAD(objects); - struct live_test t; int err = -ENODEV; /* @@ -676,7 +1068,7 @@ static int igt_ctx_readonly(void *arg) mutex_lock(&i915->drm.struct_mutex); - err = begin_live_test(&t, i915, __func__, ""); + err = igt_live_test_begin(&t, i915, __func__, ""); if (err) goto out_unlock; @@ -699,6 +1091,8 @@ static int igt_ctx_readonly(void *arg) unsigned int id; for_each_engine(engine, i915, id) { + intel_wakeref_t wakeref; + if (!intel_engine_can_store_dword(engine)) continue; @@ -713,9 +1107,9 @@ static int igt_ctx_readonly(void *arg) i915_gem_object_set_readonly(obj); } - intel_runtime_pm_get(i915); - err = gpu_fill(obj, ctx, engine, dw); - intel_runtime_pm_put(i915); + err = 0; + with_intel_runtime_pm(i915, wakeref) + err = gpu_fill(obj, ctx, engine, dw); if (err) { pr_err("Failed to fill dword %lu [%lu/%lu] with gpu (%s) in ctx %u [full-ppgtt? %s], err=%d\n", ndwords, dw, max_dwords(obj), @@ -752,7 +1146,7 @@ static int igt_ctx_readonly(void *arg) } out_unlock: - if (end_live_test(&t)) + if (igt_live_test_end(&t)) err = -EIO; mutex_unlock(&i915->drm.struct_mutex); @@ -976,10 +1370,11 @@ static int igt_vm_isolation(void *arg) struct drm_i915_private *i915 = arg; struct i915_gem_context *ctx_a, *ctx_b; struct intel_engine_cs *engine; + intel_wakeref_t wakeref; + struct igt_live_test t; struct drm_file *file; I915_RND_STATE(prng); unsigned long count; - struct live_test t; unsigned int id; u64 vm_total; int err; @@ -998,7 +1393,7 @@ static int igt_vm_isolation(void *arg) mutex_lock(&i915->drm.struct_mutex); - err = begin_live_test(&t, i915, __func__, ""); + err = igt_live_test_begin(&t, i915, __func__, ""); if (err) goto out_unlock; @@ -1022,7 +1417,7 @@ static int igt_vm_isolation(void *arg) GEM_BUG_ON(ctx_b->ppgtt->vm.total != vm_total); vm_total -= I915_GTT_PAGE_SIZE; - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); count = 0; for_each_engine(engine, i915, id) { @@ -1067,9 +1462,9 @@ static int igt_vm_isolation(void *arg) count, RUNTIME_INFO(i915)->num_rings); out_rpm: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); out_unlock: - if (end_live_test(&t)) + if (igt_live_test_end(&t)) err = -EIO; mutex_unlock(&i915->drm.struct_mutex); @@ -1165,6 +1560,7 @@ static int igt_switch_to_kernel_context(void *arg) struct intel_engine_cs *engine; struct i915_gem_context *ctx; enum intel_engine_id id; + intel_wakeref_t wakeref; int err; /* @@ -1175,7 +1571,7 @@ static int igt_switch_to_kernel_context(void *arg) */ mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); ctx = kernel_context(i915); if (IS_ERR(ctx)) { @@ -1200,7 +1596,7 @@ out_unlock: if (igt_flush_test(i915, I915_WAIT_LOCKED)) err = -EIO; - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); kernel_context_close(ctx); @@ -1232,6 +1628,7 @@ int i915_gem_context_live_selftests(struct drm_i915_private *dev_priv) SUBTEST(live_nop_switch), SUBTEST(igt_ctx_exec), SUBTEST(igt_ctx_readonly), + SUBTEST(igt_ctx_sseu), SUBTEST(igt_vm_isolation), }; diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_evict.c b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c index 4365979d8222..32dce7176f63 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem_evict.c +++ b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c @@ -29,11 +29,23 @@ #include "mock_drm.h" #include "mock_gem_device.h" -static int populate_ggtt(struct drm_i915_private *i915) +static void quirk_add(struct drm_i915_gem_object *obj, + struct list_head *objects) { + /* quirk is only for live tiled objects, use it to declare ownership */ + GEM_BUG_ON(obj->mm.quirked); + obj->mm.quirked = true; + list_add(&obj->st_link, objects); +} + +static int populate_ggtt(struct drm_i915_private *i915, + struct list_head *objects) +{ + unsigned long unbound, bound, count; struct drm_i915_gem_object *obj; u64 size; + count = 0; for (size = 0; size + I915_GTT_PAGE_SIZE <= i915->ggtt.vm.total; size += I915_GTT_PAGE_SIZE) { @@ -43,21 +55,36 @@ static int populate_ggtt(struct drm_i915_private *i915) if (IS_ERR(obj)) return PTR_ERR(obj); + quirk_add(obj, objects); + vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, 0); if (IS_ERR(vma)) return PTR_ERR(vma); + + count++; } - if (!list_empty(&i915->mm.unbound_list)) { - size = 0; - list_for_each_entry(obj, &i915->mm.unbound_list, mm.link) - size++; + unbound = 0; + list_for_each_entry(obj, &i915->mm.unbound_list, mm.link) + if (obj->mm.quirked) + unbound++; + if (unbound) { + pr_err("%s: Found %lu objects unbound, expected %u!\n", + __func__, unbound, 0); + return -EINVAL; + } - pr_err("Found %lld objects unbound!\n", size); + bound = 0; + list_for_each_entry(obj, &i915->mm.bound_list, mm.link) + if (obj->mm.quirked) + bound++; + if (bound != count) { + pr_err("%s: Found %lu objects bound, expected %lu!\n", + __func__, bound, count); return -EINVAL; } - if (list_empty(&i915->ggtt.vm.inactive_list)) { + if (list_empty(&i915->ggtt.vm.bound_list)) { pr_err("No objects on the GGTT inactive list!\n"); return -EINVAL; } @@ -67,21 +94,26 @@ static int populate_ggtt(struct drm_i915_private *i915) static void unpin_ggtt(struct drm_i915_private *i915) { + struct i915_ggtt *ggtt = &i915->ggtt; struct i915_vma *vma; - list_for_each_entry(vma, &i915->ggtt.vm.inactive_list, vm_link) - i915_vma_unpin(vma); + mutex_lock(&ggtt->vm.mutex); + list_for_each_entry(vma, &i915->ggtt.vm.bound_list, vm_link) + if (vma->obj->mm.quirked) + i915_vma_unpin(vma); + mutex_unlock(&ggtt->vm.mutex); } -static void cleanup_objects(struct drm_i915_private *i915) +static void cleanup_objects(struct drm_i915_private *i915, + struct list_head *list) { struct drm_i915_gem_object *obj, *on; - list_for_each_entry_safe(obj, on, &i915->mm.unbound_list, mm.link) - i915_gem_object_put(obj); - - list_for_each_entry_safe(obj, on, &i915->mm.bound_list, mm.link) + list_for_each_entry_safe(obj, on, list, st_link) { + GEM_BUG_ON(!obj->mm.quirked); + obj->mm.quirked = false; i915_gem_object_put(obj); + } mutex_unlock(&i915->drm.struct_mutex); @@ -94,11 +126,12 @@ static int igt_evict_something(void *arg) { struct drm_i915_private *i915 = arg; struct i915_ggtt *ggtt = &i915->ggtt; + LIST_HEAD(objects); int err; /* Fill the GGTT with pinned objects and try to evict one. */ - err = populate_ggtt(i915); + err = populate_ggtt(i915, &objects); if (err) goto cleanup; @@ -127,7 +160,7 @@ static int igt_evict_something(void *arg) } cleanup: - cleanup_objects(i915); + cleanup_objects(i915, &objects); return err; } @@ -136,13 +169,14 @@ static int igt_overcommit(void *arg) struct drm_i915_private *i915 = arg; struct drm_i915_gem_object *obj; struct i915_vma *vma; + LIST_HEAD(objects); int err; /* Fill the GGTT with pinned objects and then try to pin one more. * We expect it to fail. */ - err = populate_ggtt(i915); + err = populate_ggtt(i915, &objects); if (err) goto cleanup; @@ -152,6 +186,8 @@ static int igt_overcommit(void *arg) goto cleanup; } + quirk_add(obj, &objects); + vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, 0); if (!IS_ERR(vma) || PTR_ERR(vma) != -ENOSPC) { pr_err("Failed to evict+insert, i915_gem_object_ggtt_pin returned err=%d\n", (int)PTR_ERR(vma)); @@ -160,7 +196,7 @@ static int igt_overcommit(void *arg) } cleanup: - cleanup_objects(i915); + cleanup_objects(i915, &objects); return err; } @@ -172,11 +208,12 @@ static int igt_evict_for_vma(void *arg) .start = 0, .size = 4096, }; + LIST_HEAD(objects); int err; /* Fill the GGTT with pinned objects and try to evict a range. */ - err = populate_ggtt(i915); + err = populate_ggtt(i915, &objects); if (err) goto cleanup; @@ -199,7 +236,7 @@ static int igt_evict_for_vma(void *arg) } cleanup: - cleanup_objects(i915); + cleanup_objects(i915, &objects); return err; } @@ -222,6 +259,7 @@ static int igt_evict_for_cache_color(void *arg) }; struct drm_i915_gem_object *obj; struct i915_vma *vma; + LIST_HEAD(objects); int err; /* Currently the use of color_adjust is limited to cache domains within @@ -237,6 +275,7 @@ static int igt_evict_for_cache_color(void *arg) goto cleanup; } i915_gem_object_set_cache_level(obj, I915_CACHE_LLC); + quirk_add(obj, &objects); vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, I915_GTT_PAGE_SIZE | flags); @@ -252,6 +291,7 @@ static int igt_evict_for_cache_color(void *arg) goto cleanup; } i915_gem_object_set_cache_level(obj, I915_CACHE_LLC); + quirk_add(obj, &objects); /* Neighbouring; same colour - should fit */ vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, @@ -287,7 +327,7 @@ static int igt_evict_for_cache_color(void *arg) cleanup: unpin_ggtt(i915); - cleanup_objects(i915); + cleanup_objects(i915, &objects); ggtt->vm.mm.color_adjust = NULL; return err; } @@ -296,11 +336,12 @@ static int igt_evict_vm(void *arg) { struct drm_i915_private *i915 = arg; struct i915_ggtt *ggtt = &i915->ggtt; + LIST_HEAD(objects); int err; /* Fill the GGTT with pinned objects and try to evict everything. */ - err = populate_ggtt(i915); + err = populate_ggtt(i915, &objects); if (err) goto cleanup; @@ -322,7 +363,7 @@ static int igt_evict_vm(void *arg) } cleanup: - cleanup_objects(i915); + cleanup_objects(i915, &objects); return err; } @@ -336,6 +377,7 @@ static int igt_evict_contexts(void *arg) struct drm_mm_node node; struct reserved *next; } *reserved = NULL; + intel_wakeref_t wakeref; struct drm_mm_node hole; unsigned long count; int err; @@ -355,7 +397,7 @@ static int igt_evict_contexts(void *arg) return 0; mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); /* Reserve a block so that we know we have enough to fit a few rq */ memset(&hole, 0, sizeof(hole)); @@ -400,8 +442,10 @@ static int igt_evict_contexts(void *arg) struct drm_file *file; file = mock_file(i915); - if (IS_ERR(file)) - return PTR_ERR(file); + if (IS_ERR(file)) { + err = PTR_ERR(file); + break; + } count = 0; mutex_lock(&i915->drm.struct_mutex); @@ -464,7 +508,7 @@ out_locked: } if (drm_mm_node_allocated(&hole)) drm_mm_remove_node(&hole); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); return err; @@ -480,14 +524,17 @@ int i915_gem_evict_mock_selftests(void) SUBTEST(igt_overcommit), }; struct drm_i915_private *i915; - int err; + intel_wakeref_t wakeref; + int err = 0; i915 = mock_gem_device(); if (!i915) return -ENOMEM; mutex_lock(&i915->drm.struct_mutex); - err = i915_subtests(tests, i915); + with_intel_runtime_pm(i915, wakeref) + err = i915_subtests(tests, i915); + mutex_unlock(&i915->drm.struct_mutex); drm_dev_put(&i915->drm); diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c index a9ed0ecc94e2..3850ef4a5ec8 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c @@ -275,6 +275,7 @@ static int lowlevel_hole(struct drm_i915_private *i915, for (n = 0; n < count; n++) { u64 addr = hole_start + order[n] * BIT_ULL(size); + intel_wakeref_t wakeref; GEM_BUG_ON(addr + BIT_ULL(size) > vm->total); @@ -293,9 +294,9 @@ static int lowlevel_hole(struct drm_i915_private *i915, mock_vma.node.size = BIT_ULL(size); mock_vma.node.start = addr; - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); vm->insert_entries(vm, &mock_vma, I915_CACHE_NONE, 0); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); } count = n; @@ -1144,6 +1145,7 @@ static int igt_ggtt_page(void *arg) struct drm_i915_private *i915 = arg; struct i915_ggtt *ggtt = &i915->ggtt; struct drm_i915_gem_object *obj; + intel_wakeref_t wakeref; struct drm_mm_node tmp; unsigned int *order, n; int err; @@ -1169,7 +1171,7 @@ static int igt_ggtt_page(void *arg) if (err) goto out_unpin; - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); for (n = 0; n < count; n++) { u64 offset = tmp.start + n * PAGE_SIZE; @@ -1216,7 +1218,7 @@ static int igt_ggtt_page(void *arg) kfree(order); out_remove: ggtt->vm.clear_range(&ggtt->vm, tmp.start, tmp.size); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); drm_mm_remove_node(&tmp); out_unpin: i915_gem_object_unpin_pages(obj); @@ -1235,7 +1237,10 @@ static void track_vma_bind(struct i915_vma *vma) __i915_gem_object_pin_pages(obj); vma->pages = obj->mm.pages; - list_move_tail(&vma->vm_link, &vma->vm->inactive_list); + + mutex_lock(&vma->vm->mutex); + list_move_tail(&vma->vm_link, &vma->vm->bound_list); + mutex_unlock(&vma->vm->mutex); } static int exercise_mock(struct drm_i915_private *i915, @@ -1265,27 +1270,35 @@ static int exercise_mock(struct drm_i915_private *i915, static int igt_mock_fill(void *arg) { - return exercise_mock(arg, fill_hole); + struct i915_ggtt *ggtt = arg; + + return exercise_mock(ggtt->vm.i915, fill_hole); } static int igt_mock_walk(void *arg) { - return exercise_mock(arg, walk_hole); + struct i915_ggtt *ggtt = arg; + + return exercise_mock(ggtt->vm.i915, walk_hole); } static int igt_mock_pot(void *arg) { - return exercise_mock(arg, pot_hole); + struct i915_ggtt *ggtt = arg; + + return exercise_mock(ggtt->vm.i915, pot_hole); } static int igt_mock_drunk(void *arg) { - return exercise_mock(arg, drunk_hole); + struct i915_ggtt *ggtt = arg; + + return exercise_mock(ggtt->vm.i915, drunk_hole); } static int igt_gtt_reserve(void *arg) { - struct drm_i915_private *i915 = arg; + struct i915_ggtt *ggtt = arg; struct drm_i915_gem_object *obj, *on; LIST_HEAD(objects); u64 total; @@ -1298,11 +1311,12 @@ static int igt_gtt_reserve(void *arg) /* Start by filling the GGTT */ for (total = 0; - total + 2*I915_GTT_PAGE_SIZE <= i915->ggtt.vm.total; - total += 2*I915_GTT_PAGE_SIZE) { + total + 2 * I915_GTT_PAGE_SIZE <= ggtt->vm.total; + total += 2 * I915_GTT_PAGE_SIZE) { struct i915_vma *vma; - obj = i915_gem_object_create_internal(i915, 2*PAGE_SIZE); + obj = i915_gem_object_create_internal(ggtt->vm.i915, + 2 * PAGE_SIZE); if (IS_ERR(obj)) { err = PTR_ERR(obj); goto out; @@ -1316,20 +1330,20 @@ static int igt_gtt_reserve(void *arg) list_add(&obj->st_link, &objects); - vma = i915_vma_instance(obj, &i915->ggtt.vm, NULL); + vma = i915_vma_instance(obj, &ggtt->vm, NULL); if (IS_ERR(vma)) { err = PTR_ERR(vma); goto out; } - err = i915_gem_gtt_reserve(&i915->ggtt.vm, &vma->node, + err = i915_gem_gtt_reserve(&ggtt->vm, &vma->node, obj->base.size, total, obj->cache_level, 0); if (err) { pr_err("i915_gem_gtt_reserve (pass 1) failed at %llu/%llu with err=%d\n", - total, i915->ggtt.vm.total, err); + total, ggtt->vm.total, err); goto out; } track_vma_bind(vma); @@ -1347,11 +1361,12 @@ static int igt_gtt_reserve(void *arg) /* Now we start forcing evictions */ for (total = I915_GTT_PAGE_SIZE; - total + 2*I915_GTT_PAGE_SIZE <= i915->ggtt.vm.total; - total += 2*I915_GTT_PAGE_SIZE) { + total + 2 * I915_GTT_PAGE_SIZE <= ggtt->vm.total; + total += 2 * I915_GTT_PAGE_SIZE) { struct i915_vma *vma; - obj = i915_gem_object_create_internal(i915, 2*PAGE_SIZE); + obj = i915_gem_object_create_internal(ggtt->vm.i915, + 2 * PAGE_SIZE); if (IS_ERR(obj)) { err = PTR_ERR(obj); goto out; @@ -1365,20 +1380,20 @@ static int igt_gtt_reserve(void *arg) list_add(&obj->st_link, &objects); - vma = i915_vma_instance(obj, &i915->ggtt.vm, NULL); + vma = i915_vma_instance(obj, &ggtt->vm, NULL); if (IS_ERR(vma)) { err = PTR_ERR(vma); goto out; } - err = i915_gem_gtt_reserve(&i915->ggtt.vm, &vma->node, + err = i915_gem_gtt_reserve(&ggtt->vm, &vma->node, obj->base.size, total, obj->cache_level, 0); if (err) { pr_err("i915_gem_gtt_reserve (pass 2) failed at %llu/%llu with err=%d\n", - total, i915->ggtt.vm.total, err); + total, ggtt->vm.total, err); goto out; } track_vma_bind(vma); @@ -1399,7 +1414,7 @@ static int igt_gtt_reserve(void *arg) struct i915_vma *vma; u64 offset; - vma = i915_vma_instance(obj, &i915->ggtt.vm, NULL); + vma = i915_vma_instance(obj, &ggtt->vm, NULL); if (IS_ERR(vma)) { err = PTR_ERR(vma); goto out; @@ -1411,18 +1426,18 @@ static int igt_gtt_reserve(void *arg) goto out; } - offset = random_offset(0, i915->ggtt.vm.total, + offset = random_offset(0, ggtt->vm.total, 2*I915_GTT_PAGE_SIZE, I915_GTT_MIN_ALIGNMENT); - err = i915_gem_gtt_reserve(&i915->ggtt.vm, &vma->node, + err = i915_gem_gtt_reserve(&ggtt->vm, &vma->node, obj->base.size, offset, obj->cache_level, 0); if (err) { pr_err("i915_gem_gtt_reserve (pass 3) failed at %llu/%llu with err=%d\n", - total, i915->ggtt.vm.total, err); + total, ggtt->vm.total, err); goto out; } track_vma_bind(vma); @@ -1448,7 +1463,7 @@ out: static int igt_gtt_insert(void *arg) { - struct drm_i915_private *i915 = arg; + struct i915_ggtt *ggtt = arg; struct drm_i915_gem_object *obj, *on; struct drm_mm_node tmp = {}; const struct invalid_insert { @@ -1457,8 +1472,8 @@ static int igt_gtt_insert(void *arg) u64 start, end; } invalid_insert[] = { { - i915->ggtt.vm.total + I915_GTT_PAGE_SIZE, 0, - 0, i915->ggtt.vm.total, + ggtt->vm.total + I915_GTT_PAGE_SIZE, 0, + 0, ggtt->vm.total, }, { 2*I915_GTT_PAGE_SIZE, 0, @@ -1488,7 +1503,7 @@ static int igt_gtt_insert(void *arg) /* Check a couple of obviously invalid requests */ for (ii = invalid_insert; ii->size; ii++) { - err = i915_gem_gtt_insert(&i915->ggtt.vm, &tmp, + err = i915_gem_gtt_insert(&ggtt->vm, &tmp, ii->size, ii->alignment, I915_COLOR_UNEVICTABLE, ii->start, ii->end, @@ -1503,11 +1518,12 @@ static int igt_gtt_insert(void *arg) /* Start by filling the GGTT */ for (total = 0; - total + I915_GTT_PAGE_SIZE <= i915->ggtt.vm.total; + total + I915_GTT_PAGE_SIZE <= ggtt->vm.total; total += I915_GTT_PAGE_SIZE) { struct i915_vma *vma; - obj = i915_gem_object_create_internal(i915, I915_GTT_PAGE_SIZE); + obj = i915_gem_object_create_internal(ggtt->vm.i915, + I915_GTT_PAGE_SIZE); if (IS_ERR(obj)) { err = PTR_ERR(obj); goto out; @@ -1521,15 +1537,15 @@ static int igt_gtt_insert(void *arg) list_add(&obj->st_link, &objects); - vma = i915_vma_instance(obj, &i915->ggtt.vm, NULL); + vma = i915_vma_instance(obj, &ggtt->vm, NULL); if (IS_ERR(vma)) { err = PTR_ERR(vma); goto out; } - err = i915_gem_gtt_insert(&i915->ggtt.vm, &vma->node, + err = i915_gem_gtt_insert(&ggtt->vm, &vma->node, obj->base.size, 0, obj->cache_level, - 0, i915->ggtt.vm.total, + 0, ggtt->vm.total, 0); if (err == -ENOSPC) { /* maxed out the GGTT space */ @@ -1538,7 +1554,7 @@ static int igt_gtt_insert(void *arg) } if (err) { pr_err("i915_gem_gtt_insert (pass 1) failed at %llu/%llu with err=%d\n", - total, i915->ggtt.vm.total, err); + total, ggtt->vm.total, err); goto out; } track_vma_bind(vma); @@ -1550,7 +1566,7 @@ static int igt_gtt_insert(void *arg) list_for_each_entry(obj, &objects, st_link) { struct i915_vma *vma; - vma = i915_vma_instance(obj, &i915->ggtt.vm, NULL); + vma = i915_vma_instance(obj, &ggtt->vm, NULL); if (IS_ERR(vma)) { err = PTR_ERR(vma); goto out; @@ -1570,7 +1586,7 @@ static int igt_gtt_insert(void *arg) struct i915_vma *vma; u64 offset; - vma = i915_vma_instance(obj, &i915->ggtt.vm, NULL); + vma = i915_vma_instance(obj, &ggtt->vm, NULL); if (IS_ERR(vma)) { err = PTR_ERR(vma); goto out; @@ -1585,13 +1601,13 @@ static int igt_gtt_insert(void *arg) goto out; } - err = i915_gem_gtt_insert(&i915->ggtt.vm, &vma->node, + err = i915_gem_gtt_insert(&ggtt->vm, &vma->node, obj->base.size, 0, obj->cache_level, - 0, i915->ggtt.vm.total, + 0, ggtt->vm.total, 0); if (err) { pr_err("i915_gem_gtt_insert (pass 2) failed at %llu/%llu with err=%d\n", - total, i915->ggtt.vm.total, err); + total, ggtt->vm.total, err); goto out; } track_vma_bind(vma); @@ -1607,11 +1623,12 @@ static int igt_gtt_insert(void *arg) /* And then force evictions */ for (total = 0; - total + 2*I915_GTT_PAGE_SIZE <= i915->ggtt.vm.total; - total += 2*I915_GTT_PAGE_SIZE) { + total + 2 * I915_GTT_PAGE_SIZE <= ggtt->vm.total; + total += 2 * I915_GTT_PAGE_SIZE) { struct i915_vma *vma; - obj = i915_gem_object_create_internal(i915, 2*I915_GTT_PAGE_SIZE); + obj = i915_gem_object_create_internal(ggtt->vm.i915, + 2 * I915_GTT_PAGE_SIZE); if (IS_ERR(obj)) { err = PTR_ERR(obj); goto out; @@ -1625,19 +1642,19 @@ static int igt_gtt_insert(void *arg) list_add(&obj->st_link, &objects); - vma = i915_vma_instance(obj, &i915->ggtt.vm, NULL); + vma = i915_vma_instance(obj, &ggtt->vm, NULL); if (IS_ERR(vma)) { err = PTR_ERR(vma); goto out; } - err = i915_gem_gtt_insert(&i915->ggtt.vm, &vma->node, + err = i915_gem_gtt_insert(&ggtt->vm, &vma->node, obj->base.size, 0, obj->cache_level, - 0, i915->ggtt.vm.total, + 0, ggtt->vm.total, 0); if (err) { pr_err("i915_gem_gtt_insert (pass 3) failed at %llu/%llu with err=%d\n", - total, i915->ggtt.vm.total, err); + total, ggtt->vm.total, err); goto out; } track_vma_bind(vma); @@ -1664,17 +1681,25 @@ int i915_gem_gtt_mock_selftests(void) SUBTEST(igt_gtt_insert), }; struct drm_i915_private *i915; + struct i915_ggtt ggtt; int err; i915 = mock_gem_device(); if (!i915) return -ENOMEM; + mock_init_ggtt(i915, &ggtt); + mutex_lock(&i915->drm.struct_mutex); - err = i915_subtests(tests, i915); + err = i915_subtests(tests, &ggtt); + mock_device_flush(i915); mutex_unlock(&i915->drm.struct_mutex); + i915_gem_drain_freed_objects(i915); + + mock_fini_ggtt(&ggtt); drm_dev_put(&i915->drm); + return err; } diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_object.c b/drivers/gpu/drm/i915/selftests/i915_gem_object.c index be7ecb66ad11..395ae878e0f7 100644 --- a/drivers/gpu/drm/i915/selftests/i915_gem_object.c +++ b/drivers/gpu/drm/i915/selftests/i915_gem_object.c @@ -308,6 +308,7 @@ static int igt_partial_tiling(void *arg) const unsigned int nreal = 1 << 12; /* largest tile row x2 */ struct drm_i915_private *i915 = arg; struct drm_i915_gem_object *obj; + intel_wakeref_t wakeref; int tiling; int err; @@ -333,7 +334,7 @@ static int igt_partial_tiling(void *arg) } mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); if (1) { IGT_TIMEOUT(end); @@ -444,7 +445,7 @@ next_tiling: ; } out_unlock: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); i915_gem_object_unpin_pages(obj); out: @@ -506,11 +507,13 @@ static void disable_retire_worker(struct drm_i915_private *i915) mutex_lock(&i915->drm.struct_mutex); if (!i915->gt.active_requests++) { - intel_runtime_pm_get(i915); - i915_gem_unpark(i915); - intel_runtime_pm_put(i915); + intel_wakeref_t wakeref; + + with_intel_runtime_pm(i915, wakeref) + i915_gem_unpark(i915); } mutex_unlock(&i915->drm.struct_mutex); + cancel_delayed_work_sync(&i915->gt.retire_work); cancel_delayed_work_sync(&i915->gt.idle_work); } @@ -578,6 +581,8 @@ static int igt_mmap_offset_exhaustion(void *arg) /* Now fill with busy dead objects that we expect to reap */ for (loop = 0; loop < 3; loop++) { + intel_wakeref_t wakeref; + if (i915_terminally_wedged(&i915->gpu_error)) break; @@ -587,10 +592,10 @@ static int igt_mmap_offset_exhaustion(void *arg) goto out; } + err = 0; mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); - err = make_obj_busy(obj); - intel_runtime_pm_put(i915); + with_intel_runtime_pm(i915, wakeref) + err = make_obj_busy(obj); mutex_unlock(&i915->drm.struct_mutex); if (err) { pr_err("[loop %d] Failed to busy the object\n", loop); diff --git a/drivers/gpu/drm/i915/selftests/i915_live_selftests.h b/drivers/gpu/drm/i915/selftests/i915_live_selftests.h index a15713cae3b3..6d766925ad04 100644 --- a/drivers/gpu/drm/i915/selftests/i915_live_selftests.h +++ b/drivers/gpu/drm/i915/selftests/i915_live_selftests.h @@ -12,7 +12,9 @@ selftest(sanitycheck, i915_live_sanitycheck) /* keep first (igt selfcheck) */ selftest(uncore, intel_uncore_live_selftests) selftest(workarounds, intel_workarounds_live_selftests) +selftest(timelines, i915_timeline_live_selftests) selftest(requests, i915_request_live_selftests) +selftest(active, i915_active_live_selftests) selftest(objects, i915_gem_object_live_selftests) selftest(dmabuf, i915_gem_dmabuf_live_selftests) selftest(coherency, i915_gem_coherency_live_selftests) diff --git a/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h b/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h index 1b70208eeea7..88e5ab586337 100644 --- a/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h +++ b/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h @@ -15,8 +15,7 @@ selftest(scatterlist, scatterlist_mock_selftests) selftest(syncmap, i915_syncmap_mock_selftests) selftest(uncore, intel_uncore_mock_selftests) selftest(engine, intel_engine_cs_mock_selftests) -selftest(breadcrumbs, intel_breadcrumbs_mock_selftests) -selftest(timelines, i915_gem_timeline_mock_selftests) +selftest(timelines, i915_timeline_mock_selftests) selftest(requests, i915_request_mock_selftests) selftest(objects, i915_gem_object_mock_selftests) selftest(dmabuf, i915_gem_dmabuf_mock_selftests) diff --git a/drivers/gpu/drm/i915/selftests/i915_random.c b/drivers/gpu/drm/i915/selftests/i915_random.c index 1f415ce47018..716a3f19f030 100644 --- a/drivers/gpu/drm/i915/selftests/i915_random.c +++ b/drivers/gpu/drm/i915/selftests/i915_random.c @@ -41,18 +41,37 @@ u64 i915_prandom_u64_state(struct rnd_state *rnd) return x; } -void i915_random_reorder(unsigned int *order, unsigned int count, - struct rnd_state *state) +void i915_prandom_shuffle(void *arr, size_t elsz, size_t count, + struct rnd_state *state) { - unsigned int i, j; + char stack[128]; + + if (WARN_ON(elsz > sizeof(stack) || count > U32_MAX)) + return; + + if (!elsz || !count) + return; + + /* Fisher-Yates shuffle courtesy of Knuth */ + while (--count) { + size_t swp; + + swp = i915_prandom_u32_max_state(count + 1, state); + if (swp == count) + continue; - for (i = 0; i < count; i++) { - BUILD_BUG_ON(sizeof(unsigned int) > sizeof(u32)); - j = i915_prandom_u32_max_state(count, state); - swap(order[i], order[j]); + memcpy(stack, arr + count * elsz, elsz); + memcpy(arr + count * elsz, arr + swp * elsz, elsz); + memcpy(arr + swp * elsz, stack, elsz); } } +void i915_random_reorder(unsigned int *order, unsigned int count, + struct rnd_state *state) +{ + i915_prandom_shuffle(order, sizeof(*order), count, state); +} + unsigned int *i915_random_order(unsigned int count, struct rnd_state *state) { unsigned int *order, i; diff --git a/drivers/gpu/drm/i915/selftests/i915_random.h b/drivers/gpu/drm/i915/selftests/i915_random.h index 7dffedc501ca..8e1ff9c105b6 100644 --- a/drivers/gpu/drm/i915/selftests/i915_random.h +++ b/drivers/gpu/drm/i915/selftests/i915_random.h @@ -54,4 +54,7 @@ void i915_random_reorder(unsigned int *order, unsigned int count, struct rnd_state *state); +void i915_prandom_shuffle(void *arr, size_t elsz, size_t count, + struct rnd_state *state); + #endif /* !__I915_SELFTESTS_RANDOM_H__ */ diff --git a/drivers/gpu/drm/i915/selftests/i915_request.c b/drivers/gpu/drm/i915/selftests/i915_request.c index 07e557815308..6733dc5b6b4c 100644 --- a/drivers/gpu/drm/i915/selftests/i915_request.c +++ b/drivers/gpu/drm/i915/selftests/i915_request.c @@ -25,8 +25,12 @@ #include <linux/prime_numbers.h> #include "../i915_selftest.h" +#include "i915_random.h" +#include "igt_live_test.h" +#include "lib_sw_fence.h" #include "mock_context.h" +#include "mock_drm.h" #include "mock_gem_device.h" static int igt_add_request(void *arg) @@ -246,93 +250,285 @@ err_context_0: return err; } -int i915_request_mock_selftests(void) +struct smoketest { + struct intel_engine_cs *engine; + struct i915_gem_context **contexts; + atomic_long_t num_waits, num_fences; + int ncontexts, max_batch; + struct i915_request *(*request_alloc)(struct i915_gem_context *, + struct intel_engine_cs *); +}; + +static struct i915_request * +__mock_request_alloc(struct i915_gem_context *ctx, + struct intel_engine_cs *engine) { - static const struct i915_subtest tests[] = { - SUBTEST(igt_add_request), - SUBTEST(igt_wait_request), - SUBTEST(igt_fence_wait), - SUBTEST(igt_request_rewind), - }; - struct drm_i915_private *i915; - int err; + return mock_request(engine, ctx, 0); +} - i915 = mock_gem_device(); - if (!i915) +static struct i915_request * +__live_request_alloc(struct i915_gem_context *ctx, + struct intel_engine_cs *engine) +{ + return i915_request_alloc(engine, ctx); +} + +static int __igt_breadcrumbs_smoketest(void *arg) +{ + struct smoketest *t = arg; + struct mutex * const BKL = &t->engine->i915->drm.struct_mutex; + const unsigned int max_batch = min(t->ncontexts, t->max_batch) - 1; + const unsigned int total = 4 * t->ncontexts + 1; + unsigned int num_waits = 0, num_fences = 0; + struct i915_request **requests; + I915_RND_STATE(prng); + unsigned int *order; + int err = 0; + + /* + * A very simple test to catch the most egregious of list handling bugs. + * + * At its heart, we simply create oodles of requests running across + * multiple kthreads and enable signaling on them, for the sole purpose + * of stressing our breadcrumb handling. The only inspection we do is + * that the fences were marked as signaled. + */ + + requests = kmalloc_array(total, sizeof(*requests), GFP_KERNEL); + if (!requests) return -ENOMEM; - err = i915_subtests(tests, i915); - drm_dev_put(&i915->drm); + order = i915_random_order(total, &prng); + if (!order) { + err = -ENOMEM; + goto out_requests; + } - return err; -} + while (!kthread_should_stop()) { + struct i915_sw_fence *submit, *wait; + unsigned int n, count; -struct live_test { - struct drm_i915_private *i915; - const char *func; - const char *name; + submit = heap_fence_create(GFP_KERNEL); + if (!submit) { + err = -ENOMEM; + break; + } - unsigned int reset_count; -}; + wait = heap_fence_create(GFP_KERNEL); + if (!wait) { + i915_sw_fence_commit(submit); + heap_fence_put(submit); + err = ENOMEM; + break; + } -static int begin_live_test(struct live_test *t, - struct drm_i915_private *i915, - const char *func, - const char *name) -{ - int err; + i915_random_reorder(order, total, &prng); + count = 1 + i915_prandom_u32_max_state(max_batch, &prng); - t->i915 = i915; - t->func = func; - t->name = name; + for (n = 0; n < count; n++) { + struct i915_gem_context *ctx = + t->contexts[order[n] % t->ncontexts]; + struct i915_request *rq; - err = i915_gem_wait_for_idle(i915, - I915_WAIT_LOCKED, - MAX_SCHEDULE_TIMEOUT); - if (err) { - pr_err("%s(%s): failed to idle before, with err=%d!", - func, name, err); - return err; + mutex_lock(BKL); + + rq = t->request_alloc(ctx, t->engine); + if (IS_ERR(rq)) { + mutex_unlock(BKL); + err = PTR_ERR(rq); + count = n; + break; + } + + err = i915_sw_fence_await_sw_fence_gfp(&rq->submit, + submit, + GFP_KERNEL); + + requests[n] = i915_request_get(rq); + i915_request_add(rq); + + mutex_unlock(BKL); + + if (err >= 0) + err = i915_sw_fence_await_dma_fence(wait, + &rq->fence, + 0, + GFP_KERNEL); + + if (err < 0) { + i915_request_put(rq); + count = n; + break; + } + } + + i915_sw_fence_commit(submit); + i915_sw_fence_commit(wait); + + if (!wait_event_timeout(wait->wait, + i915_sw_fence_done(wait), + HZ / 2)) { + struct i915_request *rq = requests[count - 1]; + + pr_err("waiting for %d fences (last %llx:%lld) on %s timed out!\n", + count, + rq->fence.context, rq->fence.seqno, + t->engine->name); + i915_gem_set_wedged(t->engine->i915); + GEM_BUG_ON(!i915_request_completed(rq)); + i915_sw_fence_wait(wait); + err = -EIO; + } + + for (n = 0; n < count; n++) { + struct i915_request *rq = requests[n]; + + if (!test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, + &rq->fence.flags)) { + pr_err("%llu:%llu was not signaled!\n", + rq->fence.context, rq->fence.seqno); + err = -EINVAL; + } + + i915_request_put(rq); + } + + heap_fence_put(wait); + heap_fence_put(submit); + + if (err < 0) + break; + + num_fences += count; + num_waits++; + + cond_resched(); } - i915->gpu_error.missed_irq_rings = 0; - t->reset_count = i915_reset_count(&i915->gpu_error); + atomic_long_add(num_fences, &t->num_fences); + atomic_long_add(num_waits, &t->num_waits); - return 0; + kfree(order); +out_requests: + kfree(requests); + return err; } -static int end_live_test(struct live_test *t) +static int mock_breadcrumbs_smoketest(void *arg) { - struct drm_i915_private *i915 = t->i915; + struct drm_i915_private *i915 = arg; + struct smoketest t = { + .engine = i915->engine[RCS], + .ncontexts = 1024, + .max_batch = 1024, + .request_alloc = __mock_request_alloc + }; + unsigned int ncpus = num_online_cpus(); + struct task_struct **threads; + unsigned int n; + int ret = 0; + + /* + * Smoketest our breadcrumb/signal handling for requests across multiple + * threads. A very simple test to only catch the most egregious of bugs. + * See __igt_breadcrumbs_smoketest(); + */ - i915_retire_requests(i915); + threads = kmalloc_array(ncpus, sizeof(*threads), GFP_KERNEL); + if (!threads) + return -ENOMEM; - if (wait_for(intel_engines_are_idle(i915), 10)) { - pr_err("%s(%s): GPU not idle\n", t->func, t->name); - return -EIO; + t.contexts = + kmalloc_array(t.ncontexts, sizeof(*t.contexts), GFP_KERNEL); + if (!t.contexts) { + ret = -ENOMEM; + goto out_threads; } - if (t->reset_count != i915_reset_count(&i915->gpu_error)) { - pr_err("%s(%s): GPU was reset %d times!\n", - t->func, t->name, - i915_reset_count(&i915->gpu_error) - t->reset_count); - return -EIO; + mutex_lock(&t.engine->i915->drm.struct_mutex); + for (n = 0; n < t.ncontexts; n++) { + t.contexts[n] = mock_context(t.engine->i915, "mock"); + if (!t.contexts[n]) { + ret = -ENOMEM; + goto out_contexts; + } } + mutex_unlock(&t.engine->i915->drm.struct_mutex); + + for (n = 0; n < ncpus; n++) { + threads[n] = kthread_run(__igt_breadcrumbs_smoketest, + &t, "igt/%d", n); + if (IS_ERR(threads[n])) { + ret = PTR_ERR(threads[n]); + ncpus = n; + break; + } - if (i915->gpu_error.missed_irq_rings) { - pr_err("%s(%s): Missed interrupts on engines %lx\n", - t->func, t->name, i915->gpu_error.missed_irq_rings); - return -EIO; + get_task_struct(threads[n]); } - return 0; + msleep(jiffies_to_msecs(i915_selftest.timeout_jiffies)); + + for (n = 0; n < ncpus; n++) { + int err; + + err = kthread_stop(threads[n]); + if (err < 0 && !ret) + ret = err; + + put_task_struct(threads[n]); + } + pr_info("Completed %lu waits for %lu fence across %d cpus\n", + atomic_long_read(&t.num_waits), + atomic_long_read(&t.num_fences), + ncpus); + + mutex_lock(&t.engine->i915->drm.struct_mutex); +out_contexts: + for (n = 0; n < t.ncontexts; n++) { + if (!t.contexts[n]) + break; + mock_context_close(t.contexts[n]); + } + mutex_unlock(&t.engine->i915->drm.struct_mutex); + kfree(t.contexts); +out_threads: + kfree(threads); + + return ret; +} + +int i915_request_mock_selftests(void) +{ + static const struct i915_subtest tests[] = { + SUBTEST(igt_add_request), + SUBTEST(igt_wait_request), + SUBTEST(igt_fence_wait), + SUBTEST(igt_request_rewind), + SUBTEST(mock_breadcrumbs_smoketest), + }; + struct drm_i915_private *i915; + intel_wakeref_t wakeref; + int err = 0; + + i915 = mock_gem_device(); + if (!i915) + return -ENOMEM; + + with_intel_runtime_pm(i915, wakeref) + err = i915_subtests(tests, i915); + + drm_dev_put(&i915->drm); + + return err; } static int live_nop_request(void *arg) { struct drm_i915_private *i915 = arg; struct intel_engine_cs *engine; - struct live_test t; + intel_wakeref_t wakeref; + struct igt_live_test t; unsigned int id; int err = -ENODEV; @@ -342,7 +538,7 @@ static int live_nop_request(void *arg) */ mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); for_each_engine(engine, i915, id) { struct i915_request *request = NULL; @@ -350,7 +546,7 @@ static int live_nop_request(void *arg) IGT_TIMEOUT(end_time); ktime_t times[2] = {}; - err = begin_live_test(&t, i915, __func__, engine->name); + err = igt_live_test_begin(&t, i915, __func__, engine->name); if (err) goto out_unlock; @@ -392,7 +588,7 @@ static int live_nop_request(void *arg) break; } - err = end_live_test(&t); + err = igt_live_test_end(&t); if (err) goto out_unlock; @@ -403,7 +599,7 @@ static int live_nop_request(void *arg) } out_unlock: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); return err; } @@ -478,7 +674,8 @@ static int live_empty_request(void *arg) { struct drm_i915_private *i915 = arg; struct intel_engine_cs *engine; - struct live_test t; + intel_wakeref_t wakeref; + struct igt_live_test t; struct i915_vma *batch; unsigned int id; int err = 0; @@ -489,7 +686,7 @@ static int live_empty_request(void *arg) */ mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); batch = empty_batch(i915); if (IS_ERR(batch)) { @@ -503,7 +700,7 @@ static int live_empty_request(void *arg) unsigned long n, prime; ktime_t times[2] = {}; - err = begin_live_test(&t, i915, __func__, engine->name); + err = igt_live_test_begin(&t, i915, __func__, engine->name); if (err) goto out_batch; @@ -539,7 +736,7 @@ static int live_empty_request(void *arg) break; } - err = end_live_test(&t); + err = igt_live_test_end(&t); if (err) goto out_batch; @@ -553,7 +750,7 @@ out_batch: i915_vma_unpin(batch); i915_vma_put(batch); out_unlock: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); return err; } @@ -637,8 +834,9 @@ static int live_all_engines(void *arg) struct drm_i915_private *i915 = arg; struct intel_engine_cs *engine; struct i915_request *request[I915_NUM_ENGINES]; + intel_wakeref_t wakeref; + struct igt_live_test t; struct i915_vma *batch; - struct live_test t; unsigned int id; int err; @@ -648,9 +846,9 @@ static int live_all_engines(void *arg) */ mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); - err = begin_live_test(&t, i915, __func__, ""); + err = igt_live_test_begin(&t, i915, __func__, ""); if (err) goto out_unlock; @@ -722,7 +920,7 @@ static int live_all_engines(void *arg) request[id] = NULL; } - err = end_live_test(&t); + err = igt_live_test_end(&t); out_request: for_each_engine(engine, i915, id) @@ -731,7 +929,7 @@ out_request: i915_vma_unpin(batch); i915_vma_put(batch); out_unlock: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); return err; } @@ -742,7 +940,8 @@ static int live_sequential_engines(void *arg) struct i915_request *request[I915_NUM_ENGINES] = {}; struct i915_request *prev = NULL; struct intel_engine_cs *engine; - struct live_test t; + intel_wakeref_t wakeref; + struct igt_live_test t; unsigned int id; int err; @@ -753,9 +952,9 @@ static int live_sequential_engines(void *arg) */ mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); - err = begin_live_test(&t, i915, __func__, ""); + err = igt_live_test_begin(&t, i915, __func__, ""); if (err) goto out_unlock; @@ -838,7 +1037,7 @@ static int live_sequential_engines(void *arg) GEM_BUG_ON(!i915_request_completed(request[id])); } - err = end_live_test(&t); + err = igt_live_test_end(&t); out_request: for_each_engine(engine, i915, id) { @@ -860,11 +1059,183 @@ out_request: i915_request_put(request[id]); } out_unlock: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); return err; } +static int +max_batches(struct i915_gem_context *ctx, struct intel_engine_cs *engine) +{ + struct i915_request *rq; + int ret; + + /* + * Before execlists, all contexts share the same ringbuffer. With + * execlists, each context/engine has a separate ringbuffer and + * for the purposes of this test, inexhaustible. + * + * For the global ringbuffer though, we have to be very careful + * that we do not wrap while preventing the execution of requests + * with a unsignaled fence. + */ + if (HAS_EXECLISTS(ctx->i915)) + return INT_MAX; + + rq = i915_request_alloc(engine, ctx); + if (IS_ERR(rq)) { + ret = PTR_ERR(rq); + } else { + int sz; + + ret = rq->ring->size - rq->reserved_space; + i915_request_add(rq); + + sz = rq->ring->emit - rq->head; + if (sz < 0) + sz += rq->ring->size; + ret /= sz; + ret /= 2; /* leave half spare, in case of emergency! */ + } + + return ret; +} + +static int live_breadcrumbs_smoketest(void *arg) +{ + struct drm_i915_private *i915 = arg; + struct smoketest t[I915_NUM_ENGINES]; + unsigned int ncpus = num_online_cpus(); + unsigned long num_waits, num_fences; + struct intel_engine_cs *engine; + struct task_struct **threads; + struct igt_live_test live; + enum intel_engine_id id; + intel_wakeref_t wakeref; + struct drm_file *file; + unsigned int n; + int ret = 0; + + /* + * Smoketest our breadcrumb/signal handling for requests across multiple + * threads. A very simple test to only catch the most egregious of bugs. + * See __igt_breadcrumbs_smoketest(); + * + * On real hardware this time. + */ + + wakeref = intel_runtime_pm_get(i915); + + file = mock_file(i915); + if (IS_ERR(file)) { + ret = PTR_ERR(file); + goto out_rpm; + } + + threads = kcalloc(ncpus * I915_NUM_ENGINES, + sizeof(*threads), + GFP_KERNEL); + if (!threads) { + ret = -ENOMEM; + goto out_file; + } + + memset(&t[0], 0, sizeof(t[0])); + t[0].request_alloc = __live_request_alloc; + t[0].ncontexts = 64; + t[0].contexts = kmalloc_array(t[0].ncontexts, + sizeof(*t[0].contexts), + GFP_KERNEL); + if (!t[0].contexts) { + ret = -ENOMEM; + goto out_threads; + } + + mutex_lock(&i915->drm.struct_mutex); + for (n = 0; n < t[0].ncontexts; n++) { + t[0].contexts[n] = live_context(i915, file); + if (!t[0].contexts[n]) { + ret = -ENOMEM; + goto out_contexts; + } + } + + ret = igt_live_test_begin(&live, i915, __func__, ""); + if (ret) + goto out_contexts; + + for_each_engine(engine, i915, id) { + t[id] = t[0]; + t[id].engine = engine; + t[id].max_batch = max_batches(t[0].contexts[0], engine); + if (t[id].max_batch < 0) { + ret = t[id].max_batch; + mutex_unlock(&i915->drm.struct_mutex); + goto out_flush; + } + /* One ring interleaved between requests from all cpus */ + t[id].max_batch /= num_online_cpus() + 1; + pr_debug("Limiting batches to %d requests on %s\n", + t[id].max_batch, engine->name); + + for (n = 0; n < ncpus; n++) { + struct task_struct *tsk; + + tsk = kthread_run(__igt_breadcrumbs_smoketest, + &t[id], "igt/%d.%d", id, n); + if (IS_ERR(tsk)) { + ret = PTR_ERR(tsk); + mutex_unlock(&i915->drm.struct_mutex); + goto out_flush; + } + + get_task_struct(tsk); + threads[id * ncpus + n] = tsk; + } + } + mutex_unlock(&i915->drm.struct_mutex); + + msleep(jiffies_to_msecs(i915_selftest.timeout_jiffies)); + +out_flush: + num_waits = 0; + num_fences = 0; + for_each_engine(engine, i915, id) { + for (n = 0; n < ncpus; n++) { + struct task_struct *tsk = threads[id * ncpus + n]; + int err; + + if (!tsk) + continue; + + err = kthread_stop(tsk); + if (err < 0 && !ret) + ret = err; + + put_task_struct(tsk); + } + + num_waits += atomic_long_read(&t[id].num_waits); + num_fences += atomic_long_read(&t[id].num_fences); + } + pr_info("Completed %lu waits for %lu fences across %d engines and %d cpus\n", + num_waits, num_fences, RUNTIME_INFO(i915)->num_rings, ncpus); + + mutex_lock(&i915->drm.struct_mutex); + ret = igt_live_test_end(&live) ?: ret; +out_contexts: + mutex_unlock(&i915->drm.struct_mutex); + kfree(t[0].contexts); +out_threads: + kfree(threads); +out_file: + mock_file_free(i915, file); +out_rpm: + intel_runtime_pm_put(i915, wakeref); + + return ret; +} + int i915_request_live_selftests(struct drm_i915_private *i915) { static const struct i915_subtest tests[] = { @@ -872,6 +1243,7 @@ int i915_request_live_selftests(struct drm_i915_private *i915) SUBTEST(live_all_engines), SUBTEST(live_sequential_engines), SUBTEST(live_empty_request), + SUBTEST(live_breadcrumbs_smoketest), }; if (i915_terminally_wedged(&i915->gpu_error)) diff --git a/drivers/gpu/drm/i915/selftests/i915_selftest.c b/drivers/gpu/drm/i915/selftests/i915_selftest.c index 86c54ea37f48..10ef0e636a24 100644 --- a/drivers/gpu/drm/i915/selftests/i915_selftest.c +++ b/drivers/gpu/drm/i915/selftests/i915_selftest.c @@ -197,6 +197,49 @@ int i915_live_selftests(struct pci_dev *pdev) return 0; } +static bool apply_subtest_filter(const char *caller, const char *name) +{ + char *filter, *sep, *tok; + bool result = true; + + filter = kstrdup(i915_selftest.filter, GFP_KERNEL); + for (sep = filter; (tok = strsep(&sep, ","));) { + bool allow = true; + char *sl; + + if (*tok == '!') { + allow = false; + tok++; + } + + if (*tok == '\0') + continue; + + sl = strchr(tok, '/'); + if (sl) { + *sl++ = '\0'; + if (strcmp(tok, caller)) { + if (allow) + result = false; + continue; + } + tok = sl; + } + + if (strcmp(tok, name)) { + if (allow) + result = false; + continue; + } + + result = allow; + break; + } + kfree(filter); + + return result; +} + int __i915_subtests(const char *caller, const struct i915_subtest *st, unsigned int count, @@ -209,6 +252,9 @@ int __i915_subtests(const char *caller, if (signal_pending(current)) return -EINTR; + if (!apply_subtest_filter(caller, st->name)) + continue; + pr_debug(DRIVER_NAME ": Running %s/%s\n", caller, st->name); GEM_TRACE("Running %s/%s\n", caller, st->name); @@ -244,6 +290,7 @@ bool __igt_timeout(unsigned long timeout, const char *fmt, ...) module_param_named(st_random_seed, i915_selftest.random_seed, uint, 0400); module_param_named(st_timeout, i915_selftest.timeout_ms, uint, 0400); +module_param_named(st_filter, i915_selftest.filter, charp, 0400); module_param_named_unsafe(mock_selftests, i915_selftest.mock, int, 0400); MODULE_PARM_DESC(mock_selftests, "Run selftests before loading, using mock hardware (0:disabled [default], 1:run tests then load driver, -1:run tests then exit module)"); diff --git a/drivers/gpu/drm/i915/selftests/i915_timeline.c b/drivers/gpu/drm/i915/selftests/i915_timeline.c index 19f1c6a5c8fb..12ea69b1a1e5 100644 --- a/drivers/gpu/drm/i915/selftests/i915_timeline.c +++ b/drivers/gpu/drm/i915/selftests/i915_timeline.c @@ -4,12 +4,155 @@ * Copyright © 2017-2018 Intel Corporation */ +#include <linux/prime_numbers.h> + #include "../i915_selftest.h" #include "i915_random.h" +#include "igt_flush_test.h" #include "mock_gem_device.h" #include "mock_timeline.h" +static struct page *hwsp_page(struct i915_timeline *tl) +{ + struct drm_i915_gem_object *obj = tl->hwsp_ggtt->obj; + + GEM_BUG_ON(!i915_gem_object_has_pinned_pages(obj)); + return sg_page(obj->mm.pages->sgl); +} + +static unsigned long hwsp_cacheline(struct i915_timeline *tl) +{ + unsigned long address = (unsigned long)page_address(hwsp_page(tl)); + + return (address + tl->hwsp_offset) / CACHELINE_BYTES; +} + +#define CACHELINES_PER_PAGE (PAGE_SIZE / CACHELINE_BYTES) + +struct mock_hwsp_freelist { + struct drm_i915_private *i915; + struct radix_tree_root cachelines; + struct i915_timeline **history; + unsigned long count, max; + struct rnd_state prng; +}; + +enum { + SHUFFLE = BIT(0), +}; + +static void __mock_hwsp_record(struct mock_hwsp_freelist *state, + unsigned int idx, + struct i915_timeline *tl) +{ + tl = xchg(&state->history[idx], tl); + if (tl) { + radix_tree_delete(&state->cachelines, hwsp_cacheline(tl)); + i915_timeline_put(tl); + } +} + +static int __mock_hwsp_timeline(struct mock_hwsp_freelist *state, + unsigned int count, + unsigned int flags) +{ + struct i915_timeline *tl; + unsigned int idx; + + while (count--) { + unsigned long cacheline; + int err; + + tl = i915_timeline_create(state->i915, "mock", NULL); + if (IS_ERR(tl)) + return PTR_ERR(tl); + + cacheline = hwsp_cacheline(tl); + err = radix_tree_insert(&state->cachelines, cacheline, tl); + if (err) { + if (err == -EEXIST) { + pr_err("HWSP cacheline %lu already used; duplicate allocation!\n", + cacheline); + } + i915_timeline_put(tl); + return err; + } + + idx = state->count++ % state->max; + __mock_hwsp_record(state, idx, tl); + } + + if (flags & SHUFFLE) + i915_prandom_shuffle(state->history, + sizeof(*state->history), + min(state->count, state->max), + &state->prng); + + count = i915_prandom_u32_max_state(min(state->count, state->max), + &state->prng); + while (count--) { + idx = --state->count % state->max; + __mock_hwsp_record(state, idx, NULL); + } + + return 0; +} + +static int mock_hwsp_freelist(void *arg) +{ + struct mock_hwsp_freelist state; + const struct { + const char *name; + unsigned int flags; + } phases[] = { + { "linear", 0 }, + { "shuffled", SHUFFLE }, + { }, + }, *p; + unsigned int na; + int err = 0; + + INIT_RADIX_TREE(&state.cachelines, GFP_KERNEL); + state.prng = I915_RND_STATE_INITIALIZER(i915_selftest.random_seed); + + state.i915 = mock_gem_device(); + if (!state.i915) + return -ENOMEM; + + /* + * Create a bunch of timelines and check that their HWSP do not overlap. + * Free some, and try again. + */ + + state.max = PAGE_SIZE / sizeof(*state.history); + state.count = 0; + state.history = kcalloc(state.max, sizeof(*state.history), GFP_KERNEL); + if (!state.history) { + err = -ENOMEM; + goto err_put; + } + + mutex_lock(&state.i915->drm.struct_mutex); + for (p = phases; p->name; p++) { + pr_debug("%s(%s)\n", __func__, p->name); + for_each_prime_number_from(na, 1, 2 * CACHELINES_PER_PAGE) { + err = __mock_hwsp_timeline(&state, na, p->flags); + if (err) + goto out; + } + } + +out: + for (na = 0; na < state.max; na++) + __mock_hwsp_record(&state, na, NULL); + mutex_unlock(&state.i915->drm.struct_mutex); + kfree(state.history); +err_put: + drm_dev_put(&state.i915->drm); + return err; +} + struct __igt_sync { const char *name; u32 seqno; @@ -256,12 +399,331 @@ static int bench_sync(void *arg) return 0; } -int i915_gem_timeline_mock_selftests(void) +int i915_timeline_mock_selftests(void) { static const struct i915_subtest tests[] = { + SUBTEST(mock_hwsp_freelist), SUBTEST(igt_sync), SUBTEST(bench_sync), }; return i915_subtests(tests, NULL); } + +static int emit_ggtt_store_dw(struct i915_request *rq, u32 addr, u32 value) +{ + u32 *cs; + + cs = intel_ring_begin(rq, 4); + if (IS_ERR(cs)) + return PTR_ERR(cs); + + if (INTEL_GEN(rq->i915) >= 8) { + *cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT; + *cs++ = addr; + *cs++ = 0; + *cs++ = value; + } else if (INTEL_GEN(rq->i915) >= 4) { + *cs++ = MI_STORE_DWORD_IMM_GEN4 | MI_USE_GGTT; + *cs++ = 0; + *cs++ = addr; + *cs++ = value; + } else { + *cs++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL; + *cs++ = addr; + *cs++ = value; + *cs++ = MI_NOOP; + } + + intel_ring_advance(rq, cs); + + return 0; +} + +static struct i915_request * +tl_write(struct i915_timeline *tl, struct intel_engine_cs *engine, u32 value) +{ + struct i915_request *rq; + int err; + + lockdep_assert_held(&tl->i915->drm.struct_mutex); /* lazy rq refs */ + + err = i915_timeline_pin(tl); + if (err) { + rq = ERR_PTR(err); + goto out; + } + + rq = i915_request_alloc(engine, engine->i915->kernel_context); + if (IS_ERR(rq)) + goto out_unpin; + + err = emit_ggtt_store_dw(rq, tl->hwsp_offset, value); + i915_request_add(rq); + if (err) + rq = ERR_PTR(err); + +out_unpin: + i915_timeline_unpin(tl); +out: + if (IS_ERR(rq)) + pr_err("Failed to write to timeline!\n"); + return rq; +} + +static struct i915_timeline * +checked_i915_timeline_create(struct drm_i915_private *i915) +{ + struct i915_timeline *tl; + + tl = i915_timeline_create(i915, "live", NULL); + if (IS_ERR(tl)) + return tl; + + if (*tl->hwsp_seqno != tl->seqno) { + pr_err("Timeline created with incorrect breadcrumb, found %x, expected %x\n", + *tl->hwsp_seqno, tl->seqno); + i915_timeline_put(tl); + return ERR_PTR(-EINVAL); + } + + return tl; +} + +static int live_hwsp_engine(void *arg) +{ +#define NUM_TIMELINES 4096 + struct drm_i915_private *i915 = arg; + struct i915_timeline **timelines; + struct intel_engine_cs *engine; + enum intel_engine_id id; + intel_wakeref_t wakeref; + unsigned long count, n; + int err = 0; + + /* + * Create a bunch of timelines and check we can write + * independently to each of their breadcrumb slots. + */ + + timelines = kvmalloc_array(NUM_TIMELINES * I915_NUM_ENGINES, + sizeof(*timelines), + GFP_KERNEL); + if (!timelines) + return -ENOMEM; + + mutex_lock(&i915->drm.struct_mutex); + wakeref = intel_runtime_pm_get(i915); + + count = 0; + for_each_engine(engine, i915, id) { + if (!intel_engine_can_store_dword(engine)) + continue; + + for (n = 0; n < NUM_TIMELINES; n++) { + struct i915_timeline *tl; + struct i915_request *rq; + + tl = checked_i915_timeline_create(i915); + if (IS_ERR(tl)) { + err = PTR_ERR(tl); + goto out; + } + + rq = tl_write(tl, engine, count); + if (IS_ERR(rq)) { + i915_timeline_put(tl); + err = PTR_ERR(rq); + goto out; + } + + timelines[count++] = tl; + } + } + +out: + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + err = -EIO; + + for (n = 0; n < count; n++) { + struct i915_timeline *tl = timelines[n]; + + if (!err && *tl->hwsp_seqno != n) { + pr_err("Invalid seqno stored in timeline %lu, found 0x%x\n", + n, *tl->hwsp_seqno); + err = -EINVAL; + } + i915_timeline_put(tl); + } + + intel_runtime_pm_put(i915, wakeref); + mutex_unlock(&i915->drm.struct_mutex); + + kvfree(timelines); + + return err; +#undef NUM_TIMELINES +} + +static int live_hwsp_alternate(void *arg) +{ +#define NUM_TIMELINES 4096 + struct drm_i915_private *i915 = arg; + struct i915_timeline **timelines; + struct intel_engine_cs *engine; + enum intel_engine_id id; + intel_wakeref_t wakeref; + unsigned long count, n; + int err = 0; + + /* + * Create a bunch of timelines and check we can write + * independently to each of their breadcrumb slots with adjacent + * engines. + */ + + timelines = kvmalloc_array(NUM_TIMELINES * I915_NUM_ENGINES, + sizeof(*timelines), + GFP_KERNEL); + if (!timelines) + return -ENOMEM; + + mutex_lock(&i915->drm.struct_mutex); + wakeref = intel_runtime_pm_get(i915); + + count = 0; + for (n = 0; n < NUM_TIMELINES; n++) { + for_each_engine(engine, i915, id) { + struct i915_timeline *tl; + struct i915_request *rq; + + if (!intel_engine_can_store_dword(engine)) + continue; + + tl = checked_i915_timeline_create(i915); + if (IS_ERR(tl)) { + err = PTR_ERR(tl); + goto out; + } + + rq = tl_write(tl, engine, count); + if (IS_ERR(rq)) { + i915_timeline_put(tl); + err = PTR_ERR(rq); + goto out; + } + + timelines[count++] = tl; + } + } + +out: + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + err = -EIO; + + for (n = 0; n < count; n++) { + struct i915_timeline *tl = timelines[n]; + + if (!err && *tl->hwsp_seqno != n) { + pr_err("Invalid seqno stored in timeline %lu, found 0x%x\n", + n, *tl->hwsp_seqno); + err = -EINVAL; + } + i915_timeline_put(tl); + } + + intel_runtime_pm_put(i915, wakeref); + mutex_unlock(&i915->drm.struct_mutex); + + kvfree(timelines); + + return err; +#undef NUM_TIMELINES +} + +static int live_hwsp_recycle(void *arg) +{ + struct drm_i915_private *i915 = arg; + struct intel_engine_cs *engine; + enum intel_engine_id id; + intel_wakeref_t wakeref; + unsigned long count; + int err = 0; + + /* + * Check seqno writes into one timeline at a time. We expect to + * recycle the breadcrumb slot between iterations and neither + * want to confuse ourselves or the GPU. + */ + + mutex_lock(&i915->drm.struct_mutex); + wakeref = intel_runtime_pm_get(i915); + + count = 0; + for_each_engine(engine, i915, id) { + IGT_TIMEOUT(end_time); + + if (!intel_engine_can_store_dword(engine)) + continue; + + do { + struct i915_timeline *tl; + struct i915_request *rq; + + tl = checked_i915_timeline_create(i915); + if (IS_ERR(tl)) { + err = PTR_ERR(tl); + goto out; + } + + rq = tl_write(tl, engine, count); + if (IS_ERR(rq)) { + i915_timeline_put(tl); + err = PTR_ERR(rq); + goto out; + } + + if (i915_request_wait(rq, + I915_WAIT_LOCKED, + HZ / 5) < 0) { + pr_err("Wait for timeline writes timed out!\n"); + i915_timeline_put(tl); + err = -EIO; + goto out; + } + + if (*tl->hwsp_seqno != count) { + pr_err("Invalid seqno stored in timeline %lu, found 0x%x\n", + count, *tl->hwsp_seqno); + err = -EINVAL; + } + + i915_timeline_put(tl); + count++; + + if (err) + goto out; + + i915_timelines_park(i915); /* Encourage recycling! */ + } while (!__igt_timeout(end_time, NULL)); + } + +out: + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + err = -EIO; + intel_runtime_pm_put(i915, wakeref); + mutex_unlock(&i915->drm.struct_mutex); + + return err; +} + +int i915_timeline_live_selftests(struct drm_i915_private *i915) +{ + static const struct i915_subtest tests[] = { + SUBTEST(live_hwsp_recycle), + SUBTEST(live_hwsp_engine), + SUBTEST(live_hwsp_alternate), + }; + + return i915_subtests(tests, i915); +} diff --git a/drivers/gpu/drm/i915/selftests/i915_vma.c b/drivers/gpu/drm/i915/selftests/i915_vma.c index ffa74290e054..cf1de82741fa 100644 --- a/drivers/gpu/drm/i915/selftests/i915_vma.c +++ b/drivers/gpu/drm/i915/selftests/i915_vma.c @@ -28,6 +28,7 @@ #include "mock_gem_device.h" #include "mock_context.h" +#include "mock_gtt.h" static bool assert_vma(struct i915_vma *vma, struct drm_i915_gem_object *obj, @@ -141,7 +142,8 @@ static int create_vmas(struct drm_i915_private *i915, static int igt_vma_create(void *arg) { - struct drm_i915_private *i915 = arg; + struct i915_ggtt *ggtt = arg; + struct drm_i915_private *i915 = ggtt->vm.i915; struct drm_i915_gem_object *obj, *on; struct i915_gem_context *ctx, *cn; unsigned long num_obj, num_ctx; @@ -245,7 +247,7 @@ static bool assert_pin_einval(const struct i915_vma *vma, static int igt_vma_pin1(void *arg) { - struct drm_i915_private *i915 = arg; + struct i915_ggtt *ggtt = arg; const struct pin_mode modes[] = { #define VALID(sz, fl) { .size = (sz), .flags = (fl), .assert = assert_pin_valid, .string = #sz ", " #fl ", (valid) " } #define __INVALID(sz, fl, check, eval) { .size = (sz), .flags = (fl), .assert = (check), .string = #sz ", " #fl ", (invalid " #eval ")" } @@ -256,30 +258,30 @@ static int igt_vma_pin1(void *arg) VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | 4096), VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | 8192), - VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | (i915->ggtt.mappable_end - 4096)), - VALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | (i915->ggtt.mappable_end - 4096)), - VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | (i915->ggtt.vm.total - 4096)), - - VALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | (i915->ggtt.mappable_end - 4096)), - INVALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | i915->ggtt.mappable_end), - VALID(0, PIN_GLOBAL | PIN_OFFSET_FIXED | (i915->ggtt.vm.total - 4096)), - INVALID(0, PIN_GLOBAL | PIN_OFFSET_FIXED | i915->ggtt.vm.total), + VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | (ggtt->mappable_end - 4096)), + VALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | (ggtt->mappable_end - 4096)), + VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | (ggtt->vm.total - 4096)), + + VALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | (ggtt->mappable_end - 4096)), + INVALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | ggtt->mappable_end), + VALID(0, PIN_GLOBAL | PIN_OFFSET_FIXED | (ggtt->vm.total - 4096)), + INVALID(0, PIN_GLOBAL | PIN_OFFSET_FIXED | ggtt->vm.total), INVALID(0, PIN_GLOBAL | PIN_OFFSET_FIXED | round_down(U64_MAX, PAGE_SIZE)), VALID(4096, PIN_GLOBAL), VALID(8192, PIN_GLOBAL), - VALID(i915->ggtt.mappable_end - 4096, PIN_GLOBAL | PIN_MAPPABLE), - VALID(i915->ggtt.mappable_end, PIN_GLOBAL | PIN_MAPPABLE), - NOSPACE(i915->ggtt.mappable_end + 4096, PIN_GLOBAL | PIN_MAPPABLE), - VALID(i915->ggtt.vm.total - 4096, PIN_GLOBAL), - VALID(i915->ggtt.vm.total, PIN_GLOBAL), - NOSPACE(i915->ggtt.vm.total + 4096, PIN_GLOBAL), + VALID(ggtt->mappable_end - 4096, PIN_GLOBAL | PIN_MAPPABLE), + VALID(ggtt->mappable_end, PIN_GLOBAL | PIN_MAPPABLE), + NOSPACE(ggtt->mappable_end + 4096, PIN_GLOBAL | PIN_MAPPABLE), + VALID(ggtt->vm.total - 4096, PIN_GLOBAL), + VALID(ggtt->vm.total, PIN_GLOBAL), + NOSPACE(ggtt->vm.total + 4096, PIN_GLOBAL), NOSPACE(round_down(U64_MAX, PAGE_SIZE), PIN_GLOBAL), - INVALID(8192, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | (i915->ggtt.mappable_end - 4096)), - INVALID(8192, PIN_GLOBAL | PIN_OFFSET_FIXED | (i915->ggtt.vm.total - 4096)), + INVALID(8192, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | (ggtt->mappable_end - 4096)), + INVALID(8192, PIN_GLOBAL | PIN_OFFSET_FIXED | (ggtt->vm.total - 4096)), INVALID(8192, PIN_GLOBAL | PIN_OFFSET_FIXED | (round_down(U64_MAX, PAGE_SIZE) - 4096)), - VALID(8192, PIN_GLOBAL | PIN_OFFSET_BIAS | (i915->ggtt.mappable_end - 4096)), + VALID(8192, PIN_GLOBAL | PIN_OFFSET_BIAS | (ggtt->mappable_end - 4096)), #if !IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM) /* Misusing BIAS is a programming error (it is not controllable @@ -287,10 +289,10 @@ static int igt_vma_pin1(void *arg) * However, the tests are still quite interesting for checking * variable start, end and size. */ - NOSPACE(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | i915->ggtt.mappable_end), - NOSPACE(0, PIN_GLOBAL | PIN_OFFSET_BIAS | i915->ggtt.vm.total), - NOSPACE(8192, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | (i915->ggtt.mappable_end - 4096)), - NOSPACE(8192, PIN_GLOBAL | PIN_OFFSET_BIAS | (i915->ggtt.vm.total - 4096)), + NOSPACE(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | ggtt->mappable_end), + NOSPACE(0, PIN_GLOBAL | PIN_OFFSET_BIAS | ggtt->vm.total), + NOSPACE(8192, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | (ggtt->mappable_end - 4096)), + NOSPACE(8192, PIN_GLOBAL | PIN_OFFSET_BIAS | (ggtt->vm.total - 4096)), #endif { }, #undef NOSPACE @@ -306,13 +308,13 @@ static int igt_vma_pin1(void *arg) * focusing on error handling of boundary conditions. */ - GEM_BUG_ON(!drm_mm_clean(&i915->ggtt.vm.mm)); + GEM_BUG_ON(!drm_mm_clean(&ggtt->vm.mm)); - obj = i915_gem_object_create_internal(i915, PAGE_SIZE); + obj = i915_gem_object_create_internal(ggtt->vm.i915, PAGE_SIZE); if (IS_ERR(obj)) return PTR_ERR(obj); - vma = checked_vma_instance(obj, &i915->ggtt.vm, NULL); + vma = checked_vma_instance(obj, &ggtt->vm, NULL); if (IS_ERR(vma)) goto out; @@ -403,8 +405,8 @@ static unsigned int rotated_size(const struct intel_rotation_plane_info *a, static int igt_vma_rotate(void *arg) { - struct drm_i915_private *i915 = arg; - struct i915_address_space *vm = &i915->ggtt.vm; + struct i915_ggtt *ggtt = arg; + struct i915_address_space *vm = &ggtt->vm; struct drm_i915_gem_object *obj; const struct intel_rotation_plane_info planes[] = { { .width = 1, .height = 1, .stride = 1 }, @@ -431,7 +433,7 @@ static int igt_vma_rotate(void *arg) * that the page layout within the rotated VMA match our expectations. */ - obj = i915_gem_object_create_internal(i915, max_pages * PAGE_SIZE); + obj = i915_gem_object_create_internal(vm->i915, max_pages * PAGE_SIZE); if (IS_ERR(obj)) goto out; @@ -602,8 +604,8 @@ static bool assert_pin(struct i915_vma *vma, static int igt_vma_partial(void *arg) { - struct drm_i915_private *i915 = arg; - struct i915_address_space *vm = &i915->ggtt.vm; + struct i915_ggtt *ggtt = arg; + struct i915_address_space *vm = &ggtt->vm; const unsigned int npages = 1021; /* prime! */ struct drm_i915_gem_object *obj; const struct phase { @@ -621,7 +623,7 @@ static int igt_vma_partial(void *arg) * we are returned the same VMA when we later request the same range. */ - obj = i915_gem_object_create_internal(i915, npages*PAGE_SIZE); + obj = i915_gem_object_create_internal(vm->i915, npages * PAGE_SIZE); if (IS_ERR(obj)) goto out; @@ -670,7 +672,7 @@ static int igt_vma_partial(void *arg) } count = 0; - list_for_each_entry(vma, &obj->vma_list, obj_link) + list_for_each_entry(vma, &obj->vma.list, obj_link) count++; if (count != nvma) { pr_err("(%s) All partial vma were not recorded on the obj->vma_list: found %u, expected %u\n", @@ -699,7 +701,7 @@ static int igt_vma_partial(void *arg) i915_vma_unpin(vma); count = 0; - list_for_each_entry(vma, &obj->vma_list, obj_link) + list_for_each_entry(vma, &obj->vma.list, obj_link) count++; if (count != nvma) { pr_err("(%s) allocated an extra full vma!\n", p->name); @@ -723,17 +725,24 @@ int i915_vma_mock_selftests(void) SUBTEST(igt_vma_partial), }; struct drm_i915_private *i915; + struct i915_ggtt ggtt; int err; i915 = mock_gem_device(); if (!i915) return -ENOMEM; + mock_init_ggtt(i915, &ggtt); + mutex_lock(&i915->drm.struct_mutex); - err = i915_subtests(tests, i915); + err = i915_subtests(tests, &ggtt); + mock_device_flush(i915); mutex_unlock(&i915->drm.struct_mutex); + i915_gem_drain_freed_objects(i915); + + mock_fini_ggtt(&ggtt); drm_dev_put(&i915->drm); + return err; } - diff --git a/drivers/gpu/drm/i915/selftests/igt_live_test.c b/drivers/gpu/drm/i915/selftests/igt_live_test.c new file mode 100644 index 000000000000..3e902761cd16 --- /dev/null +++ b/drivers/gpu/drm/i915/selftests/igt_live_test.c @@ -0,0 +1,78 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2018 Intel Corporation + */ + +#include "../i915_drv.h" + +#include "../i915_selftest.h" +#include "igt_flush_test.h" +#include "igt_live_test.h" + +int igt_live_test_begin(struct igt_live_test *t, + struct drm_i915_private *i915, + const char *func, + const char *name) +{ + struct intel_engine_cs *engine; + enum intel_engine_id id; + int err; + + lockdep_assert_held(&i915->drm.struct_mutex); + + t->i915 = i915; + t->func = func; + t->name = name; + + err = i915_gem_wait_for_idle(i915, + I915_WAIT_INTERRUPTIBLE | + I915_WAIT_LOCKED, + MAX_SCHEDULE_TIMEOUT); + if (err) { + pr_err("%s(%s): failed to idle before, with err=%d!", + func, name, err); + return err; + } + + t->reset_global = i915_reset_count(&i915->gpu_error); + + for_each_engine(engine, i915, id) + t->reset_engine[id] = + i915_reset_engine_count(&i915->gpu_error, engine); + + return 0; +} + +int igt_live_test_end(struct igt_live_test *t) +{ + struct drm_i915_private *i915 = t->i915; + struct intel_engine_cs *engine; + enum intel_engine_id id; + + lockdep_assert_held(&i915->drm.struct_mutex); + + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + return -EIO; + + if (t->reset_global != i915_reset_count(&i915->gpu_error)) { + pr_err("%s(%s): GPU was reset %d times!\n", + t->func, t->name, + i915_reset_count(&i915->gpu_error) - t->reset_global); + return -EIO; + } + + for_each_engine(engine, i915, id) { + if (t->reset_engine[id] == + i915_reset_engine_count(&i915->gpu_error, engine)) + continue; + + pr_err("%s(%s): engine '%s' was reset %d times!\n", + t->func, t->name, engine->name, + i915_reset_engine_count(&i915->gpu_error, engine) - + t->reset_engine[id]); + return -EIO; + } + + return 0; +} diff --git a/drivers/gpu/drm/i915/selftests/igt_live_test.h b/drivers/gpu/drm/i915/selftests/igt_live_test.h new file mode 100644 index 000000000000..c0e9f99d50de --- /dev/null +++ b/drivers/gpu/drm/i915/selftests/igt_live_test.h @@ -0,0 +1,35 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright © 2019 Intel Corporation + */ + +#ifndef IGT_LIVE_TEST_H +#define IGT_LIVE_TEST_H + +#include "../i915_gem.h" + +struct drm_i915_private; + +struct igt_live_test { + struct drm_i915_private *i915; + const char *func; + const char *name; + + unsigned int reset_global; + unsigned int reset_engine[I915_NUM_ENGINES]; +}; + +/* + * Flush the GPU state before and after the test to ensure that no residual + * code is running on the GPU that may affect this test. Also compare the + * state before and after the test and alert if it unexpectedly changes, + * e.g. if the GPU was reset. + */ +int igt_live_test_begin(struct igt_live_test *t, + struct drm_i915_private *i915, + const char *func, + const char *name); +int igt_live_test_end(struct igt_live_test *t); + +#endif /* IGT_LIVE_TEST_H */ diff --git a/drivers/gpu/drm/i915/selftests/igt_spinner.c b/drivers/gpu/drm/i915/selftests/igt_spinner.c index 0e70df0230b8..9ebd9225684e 100644 --- a/drivers/gpu/drm/i915/selftests/igt_spinner.c +++ b/drivers/gpu/drm/i915/selftests/igt_spinner.c @@ -185,11 +185,6 @@ void igt_spinner_fini(struct igt_spinner *spin) bool igt_wait_for_spinner(struct igt_spinner *spin, struct i915_request *rq) { - if (!wait_event_timeout(rq->execute, - READ_ONCE(rq->global_seqno), - msecs_to_jiffies(10))) - return false; - return !(wait_for_us(i915_seqno_passed(hws_seqno(spin, rq), rq->fence.seqno), 10) && diff --git a/drivers/gpu/drm/i915/selftests/intel_breadcrumbs.c b/drivers/gpu/drm/i915/selftests/intel_breadcrumbs.c deleted file mode 100644 index f03b407fdbe2..000000000000 --- a/drivers/gpu/drm/i915/selftests/intel_breadcrumbs.c +++ /dev/null @@ -1,470 +0,0 @@ -/* - * Copyright © 2016 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS - * IN THE SOFTWARE. - * - */ - -#include "../i915_selftest.h" -#include "i915_random.h" - -#include "mock_gem_device.h" -#include "mock_engine.h" - -static int check_rbtree(struct intel_engine_cs *engine, - const unsigned long *bitmap, - const struct intel_wait *waiters, - const int count) -{ - struct intel_breadcrumbs *b = &engine->breadcrumbs; - struct rb_node *rb; - int n; - - if (&b->irq_wait->node != rb_first(&b->waiters)) { - pr_err("First waiter does not match first element of wait-tree\n"); - return -EINVAL; - } - - n = find_first_bit(bitmap, count); - for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) { - struct intel_wait *w = container_of(rb, typeof(*w), node); - int idx = w - waiters; - - if (!test_bit(idx, bitmap)) { - pr_err("waiter[%d, seqno=%d] removed but still in wait-tree\n", - idx, w->seqno); - return -EINVAL; - } - - if (n != idx) { - pr_err("waiter[%d, seqno=%d] does not match expected next element in tree [%d]\n", - idx, w->seqno, n); - return -EINVAL; - } - - n = find_next_bit(bitmap, count, n + 1); - } - - return 0; -} - -static int check_completion(struct intel_engine_cs *engine, - const unsigned long *bitmap, - const struct intel_wait *waiters, - const int count) -{ - int n; - - for (n = 0; n < count; n++) { - if (intel_wait_complete(&waiters[n]) != !!test_bit(n, bitmap)) - continue; - - pr_err("waiter[%d, seqno=%d] is %s, but expected %s\n", - n, waiters[n].seqno, - intel_wait_complete(&waiters[n]) ? "complete" : "active", - test_bit(n, bitmap) ? "active" : "complete"); - return -EINVAL; - } - - return 0; -} - -static int check_rbtree_empty(struct intel_engine_cs *engine) -{ - struct intel_breadcrumbs *b = &engine->breadcrumbs; - - if (b->irq_wait) { - pr_err("Empty breadcrumbs still has a waiter\n"); - return -EINVAL; - } - - if (!RB_EMPTY_ROOT(&b->waiters)) { - pr_err("Empty breadcrumbs, but wait-tree not empty\n"); - return -EINVAL; - } - - return 0; -} - -static int igt_random_insert_remove(void *arg) -{ - const u32 seqno_bias = 0x1000; - I915_RND_STATE(prng); - struct intel_engine_cs *engine = arg; - struct intel_wait *waiters; - const int count = 4096; - unsigned int *order; - unsigned long *bitmap; - int err = -ENOMEM; - int n; - - mock_engine_reset(engine); - - waiters = kvmalloc_array(count, sizeof(*waiters), GFP_KERNEL); - if (!waiters) - goto out_engines; - - bitmap = kcalloc(DIV_ROUND_UP(count, BITS_PER_LONG), sizeof(*bitmap), - GFP_KERNEL); - if (!bitmap) - goto out_waiters; - - order = i915_random_order(count, &prng); - if (!order) - goto out_bitmap; - - for (n = 0; n < count; n++) - intel_wait_init_for_seqno(&waiters[n], seqno_bias + n); - - err = check_rbtree(engine, bitmap, waiters, count); - if (err) - goto out_order; - - /* Add and remove waiters into the rbtree in random order. At each - * step, we verify that the rbtree is correctly ordered. - */ - for (n = 0; n < count; n++) { - int i = order[n]; - - intel_engine_add_wait(engine, &waiters[i]); - __set_bit(i, bitmap); - - err = check_rbtree(engine, bitmap, waiters, count); - if (err) - goto out_order; - } - - i915_random_reorder(order, count, &prng); - for (n = 0; n < count; n++) { - int i = order[n]; - - intel_engine_remove_wait(engine, &waiters[i]); - __clear_bit(i, bitmap); - - err = check_rbtree(engine, bitmap, waiters, count); - if (err) - goto out_order; - } - - err = check_rbtree_empty(engine); -out_order: - kfree(order); -out_bitmap: - kfree(bitmap); -out_waiters: - kvfree(waiters); -out_engines: - mock_engine_flush(engine); - return err; -} - -static int igt_insert_complete(void *arg) -{ - const u32 seqno_bias = 0x1000; - struct intel_engine_cs *engine = arg; - struct intel_wait *waiters; - const int count = 4096; - unsigned long *bitmap; - int err = -ENOMEM; - int n, m; - - mock_engine_reset(engine); - - waiters = kvmalloc_array(count, sizeof(*waiters), GFP_KERNEL); - if (!waiters) - goto out_engines; - - bitmap = kcalloc(DIV_ROUND_UP(count, BITS_PER_LONG), sizeof(*bitmap), - GFP_KERNEL); - if (!bitmap) - goto out_waiters; - - for (n = 0; n < count; n++) { - intel_wait_init_for_seqno(&waiters[n], n + seqno_bias); - intel_engine_add_wait(engine, &waiters[n]); - __set_bit(n, bitmap); - } - err = check_rbtree(engine, bitmap, waiters, count); - if (err) - goto out_bitmap; - - /* On each step, we advance the seqno so that several waiters are then - * complete (we increase the seqno by increasingly larger values to - * retire more and more waiters at once). All retired waiters should - * be woken and removed from the rbtree, and so that we check. - */ - for (n = 0; n < count; n = m) { - int seqno = 2 * n; - - GEM_BUG_ON(find_first_bit(bitmap, count) != n); - - if (intel_wait_complete(&waiters[n])) { - pr_err("waiter[%d, seqno=%d] completed too early\n", - n, waiters[n].seqno); - err = -EINVAL; - goto out_bitmap; - } - - /* complete the following waiters */ - mock_seqno_advance(engine, seqno + seqno_bias); - for (m = n; m <= seqno; m++) { - if (m == count) - break; - - GEM_BUG_ON(!test_bit(m, bitmap)); - __clear_bit(m, bitmap); - } - - intel_engine_remove_wait(engine, &waiters[n]); - RB_CLEAR_NODE(&waiters[n].node); - - err = check_rbtree(engine, bitmap, waiters, count); - if (err) { - pr_err("rbtree corrupt after seqno advance to %d\n", - seqno + seqno_bias); - goto out_bitmap; - } - - err = check_completion(engine, bitmap, waiters, count); - if (err) { - pr_err("completions after seqno advance to %d failed\n", - seqno + seqno_bias); - goto out_bitmap; - } - } - - err = check_rbtree_empty(engine); -out_bitmap: - kfree(bitmap); -out_waiters: - kvfree(waiters); -out_engines: - mock_engine_flush(engine); - return err; -} - -struct igt_wakeup { - struct task_struct *tsk; - atomic_t *ready, *set, *done; - struct intel_engine_cs *engine; - unsigned long flags; -#define STOP 0 -#define IDLE 1 - wait_queue_head_t *wq; - u32 seqno; -}; - -static bool wait_for_ready(struct igt_wakeup *w) -{ - DEFINE_WAIT(ready); - - set_bit(IDLE, &w->flags); - if (atomic_dec_and_test(w->done)) - wake_up_var(w->done); - - if (test_bit(STOP, &w->flags)) - goto out; - - for (;;) { - prepare_to_wait(w->wq, &ready, TASK_INTERRUPTIBLE); - if (atomic_read(w->ready) == 0) - break; - - schedule(); - } - finish_wait(w->wq, &ready); - -out: - clear_bit(IDLE, &w->flags); - if (atomic_dec_and_test(w->set)) - wake_up_var(w->set); - - return !test_bit(STOP, &w->flags); -} - -static int igt_wakeup_thread(void *arg) -{ - struct igt_wakeup *w = arg; - struct intel_wait wait; - - while (wait_for_ready(w)) { - GEM_BUG_ON(kthread_should_stop()); - - intel_wait_init_for_seqno(&wait, w->seqno); - intel_engine_add_wait(w->engine, &wait); - for (;;) { - set_current_state(TASK_UNINTERRUPTIBLE); - if (i915_seqno_passed(intel_engine_get_seqno(w->engine), - w->seqno)) - break; - - if (test_bit(STOP, &w->flags)) /* emergency escape */ - break; - - schedule(); - } - intel_engine_remove_wait(w->engine, &wait); - __set_current_state(TASK_RUNNING); - } - - return 0; -} - -static void igt_wake_all_sync(atomic_t *ready, - atomic_t *set, - atomic_t *done, - wait_queue_head_t *wq, - int count) -{ - atomic_set(set, count); - atomic_set(ready, 0); - wake_up_all(wq); - - wait_var_event(set, !atomic_read(set)); - atomic_set(ready, count); - atomic_set(done, count); -} - -static int igt_wakeup(void *arg) -{ - I915_RND_STATE(prng); - struct intel_engine_cs *engine = arg; - struct igt_wakeup *waiters; - DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); - const int count = 4096; - const u32 max_seqno = count / 4; - atomic_t ready, set, done; - int err = -ENOMEM; - int n, step; - - mock_engine_reset(engine); - - waiters = kvmalloc_array(count, sizeof(*waiters), GFP_KERNEL); - if (!waiters) - goto out_engines; - - /* Create a large number of threads, each waiting on a random seqno. - * Multiple waiters will be waiting for the same seqno. - */ - atomic_set(&ready, count); - for (n = 0; n < count; n++) { - waiters[n].wq = &wq; - waiters[n].ready = &ready; - waiters[n].set = &set; - waiters[n].done = &done; - waiters[n].engine = engine; - waiters[n].flags = BIT(IDLE); - - waiters[n].tsk = kthread_run(igt_wakeup_thread, &waiters[n], - "i915/igt:%d", n); - if (IS_ERR(waiters[n].tsk)) - goto out_waiters; - - get_task_struct(waiters[n].tsk); - } - - for (step = 1; step <= max_seqno; step <<= 1) { - u32 seqno; - - /* The waiter threads start paused as we assign them a random - * seqno and reset the engine. Once the engine is reset, - * we signal that the threads may begin their wait upon their - * seqno. - */ - for (n = 0; n < count; n++) { - GEM_BUG_ON(!test_bit(IDLE, &waiters[n].flags)); - waiters[n].seqno = - 1 + prandom_u32_state(&prng) % max_seqno; - } - mock_seqno_advance(engine, 0); - igt_wake_all_sync(&ready, &set, &done, &wq, count); - - /* Simulate the GPU doing chunks of work, with one or more - * seqno appearing to finish at the same time. A random number - * of threads will be waiting upon the update and hopefully be - * woken. - */ - for (seqno = 1; seqno <= max_seqno + step; seqno += step) { - usleep_range(50, 500); - mock_seqno_advance(engine, seqno); - } - GEM_BUG_ON(intel_engine_get_seqno(engine) < 1 + max_seqno); - - /* With the seqno now beyond any of the waiting threads, they - * should all be woken, see that they are complete and signal - * that they are ready for the next test. We wait until all - * threads are complete and waiting for us (i.e. not a seqno). - */ - if (!wait_var_event_timeout(&done, - !atomic_read(&done), 10 * HZ)) { - pr_err("Timed out waiting for %d remaining waiters\n", - atomic_read(&done)); - err = -ETIMEDOUT; - break; - } - - err = check_rbtree_empty(engine); - if (err) - break; - } - -out_waiters: - for (n = 0; n < count; n++) { - if (IS_ERR(waiters[n].tsk)) - break; - - set_bit(STOP, &waiters[n].flags); - } - mock_seqno_advance(engine, INT_MAX); /* wakeup any broken waiters */ - igt_wake_all_sync(&ready, &set, &done, &wq, n); - - for (n = 0; n < count; n++) { - if (IS_ERR(waiters[n].tsk)) - break; - - kthread_stop(waiters[n].tsk); - put_task_struct(waiters[n].tsk); - } - - kvfree(waiters); -out_engines: - mock_engine_flush(engine); - return err; -} - -int intel_breadcrumbs_mock_selftests(void) -{ - static const struct i915_subtest tests[] = { - SUBTEST(igt_random_insert_remove), - SUBTEST(igt_insert_complete), - SUBTEST(igt_wakeup), - }; - struct drm_i915_private *i915; - int err; - - i915 = mock_gem_device(); - if (!i915) - return -ENOMEM; - - err = i915_subtests(tests, i915->engine[RCS]); - drm_dev_put(&i915->drm); - - return err; -} diff --git a/drivers/gpu/drm/i915/selftests/intel_guc.c b/drivers/gpu/drm/i915/selftests/intel_guc.c index 32cba4cae31a..c5e0a0e98fcb 100644 --- a/drivers/gpu/drm/i915/selftests/intel_guc.c +++ b/drivers/gpu/drm/i915/selftests/intel_guc.c @@ -137,12 +137,13 @@ static bool client_doorbell_in_sync(struct intel_guc_client *client) static int igt_guc_clients(void *args) { struct drm_i915_private *dev_priv = args; + intel_wakeref_t wakeref; struct intel_guc *guc; int err = 0; GEM_BUG_ON(!HAS_GUC(dev_priv)); mutex_lock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); guc = &dev_priv->guc; if (!guc) { @@ -225,7 +226,7 @@ out: guc_clients_create(guc); guc_clients_enable(guc); unlock: - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); mutex_unlock(&dev_priv->drm.struct_mutex); return err; } @@ -238,13 +239,14 @@ unlock: static int igt_guc_doorbells(void *arg) { struct drm_i915_private *dev_priv = arg; + intel_wakeref_t wakeref; struct intel_guc *guc; int i, err = 0; u16 db_id; GEM_BUG_ON(!HAS_GUC(dev_priv)); mutex_lock(&dev_priv->drm.struct_mutex); - intel_runtime_pm_get(dev_priv); + wakeref = intel_runtime_pm_get(dev_priv); guc = &dev_priv->guc; if (!guc) { @@ -337,7 +339,7 @@ out: guc_client_free(clients[i]); } unlock: - intel_runtime_pm_put(dev_priv); + intel_runtime_pm_put(dev_priv, wakeref); mutex_unlock(&dev_priv->drm.struct_mutex); return err; } diff --git a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c index 0aadbd9c7d56..7b6f3bea9ef8 100644 --- a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c +++ b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c @@ -363,9 +363,7 @@ static int igt_global_reset(void *arg) /* Check that we can issue a global GPU reset */ igt_global_reset_lock(i915); - set_bit(I915_RESET_HANDOFF, &i915->gpu_error.flags); - mutex_lock(&i915->drm.struct_mutex); reset_count = i915_reset_count(&i915->gpu_error); i915_reset(i915, ALL_ENGINES, NULL); @@ -374,9 +372,7 @@ static int igt_global_reset(void *arg) pr_err("No GPU reset recorded!\n"); err = -EINVAL; } - mutex_unlock(&i915->drm.struct_mutex); - GEM_BUG_ON(test_bit(I915_RESET_HANDOFF, &i915->gpu_error.flags)); igt_global_reset_unlock(i915); if (i915_terminally_wedged(&i915->gpu_error)) @@ -388,22 +384,21 @@ static int igt_global_reset(void *arg) static int igt_wedged_reset(void *arg) { struct drm_i915_private *i915 = arg; + intel_wakeref_t wakeref; /* Check that we can recover a wedged device with a GPU reset */ igt_global_reset_lock(i915); - mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); i915_gem_set_wedged(i915); - GEM_BUG_ON(!i915_terminally_wedged(&i915->gpu_error)); - set_bit(I915_RESET_HANDOFF, &i915->gpu_error.flags); + mutex_lock(&i915->drm.struct_mutex); + GEM_BUG_ON(!i915_terminally_wedged(&i915->gpu_error)); i915_reset(i915, ALL_ENGINES, NULL); - GEM_BUG_ON(test_bit(I915_RESET_HANDOFF, &i915->gpu_error.flags)); - - intel_runtime_pm_put(i915); mutex_unlock(&i915->drm.struct_mutex); + + intel_runtime_pm_put(i915, wakeref); igt_global_reset_unlock(i915); return i915_terminally_wedged(&i915->gpu_error) ? -EIO : 0; @@ -454,8 +449,6 @@ static int __igt_reset_engine(struct drm_i915_private *i915, bool active) set_bit(I915_RESET_ENGINE + id, &i915->gpu_error.flags); do { - u32 seqno = intel_engine_get_seqno(engine); - if (active) { struct i915_request *rq; @@ -484,8 +477,6 @@ static int __igt_reset_engine(struct drm_i915_private *i915, bool active) break; } - GEM_BUG_ON(!rq->global_seqno); - seqno = rq->global_seqno - 1; i915_request_put(rq); } @@ -501,16 +492,15 @@ static int __igt_reset_engine(struct drm_i915_private *i915, bool active) break; } - reset_engine_count += active; if (i915_reset_engine_count(&i915->gpu_error, engine) != - reset_engine_count) { - pr_err("%s engine reset %srecorded!\n", - engine->name, active ? "not " : ""); + ++reset_engine_count) { + pr_err("%s engine reset not recorded!\n", + engine->name); err = -EINVAL; break; } - if (!wait_for_idle(engine)) { + if (!i915_reset_flush(i915)) { struct drm_printer p = drm_info_printer(i915->drm.dev); @@ -733,7 +723,6 @@ static int __igt_reset_engines(struct drm_i915_private *i915, set_bit(I915_RESET_ENGINE + id, &i915->gpu_error.flags); do { - u32 seqno = intel_engine_get_seqno(engine); struct i915_request *rq = NULL; if (flags & TEST_ACTIVE) { @@ -761,9 +750,6 @@ static int __igt_reset_engines(struct drm_i915_private *i915, err = -EIO; break; } - - GEM_BUG_ON(!rq->global_seqno); - seqno = rq->global_seqno - 1; } err = i915_reset_engine(engine, NULL); @@ -800,10 +786,9 @@ static int __igt_reset_engines(struct drm_i915_private *i915, reported = i915_reset_engine_count(&i915->gpu_error, engine); reported -= threads[engine->id].resets; - if (reported != (flags & TEST_ACTIVE ? count : 0)) { - pr_err("i915_reset_engine(%s:%s): reset %lu times, but reported %lu, expected %lu reported\n", - engine->name, test_name, count, reported, - (flags & TEST_ACTIVE ? count : 0)); + if (reported != count) { + pr_err("i915_reset_engine(%s:%s): reset %lu times, but reported %lu\n", + engine->name, test_name, count, reported); if (!err) err = -EINVAL; } @@ -902,20 +887,13 @@ static int igt_reset_engines(void *arg) return 0; } -static u32 fake_hangcheck(struct i915_request *rq, u32 mask) +static u32 fake_hangcheck(struct drm_i915_private *i915, u32 mask) { - struct i915_gpu_error *error = &rq->i915->gpu_error; - u32 reset_count = i915_reset_count(error); - - error->stalled_mask = mask; - - /* set_bit() must be after we have setup the backchannel (mask) */ - smp_mb__before_atomic(); - set_bit(I915_RESET_HANDOFF, &error->flags); + u32 count = i915_reset_count(&i915->gpu_error); - wake_up_all(&error->wait_queue); + i915_reset(i915, mask, NULL); - return reset_count; + return count; } static int igt_reset_wait(void *arg) @@ -961,7 +939,7 @@ static int igt_reset_wait(void *arg) goto out_rq; } - reset_count = fake_hangcheck(rq, ALL_ENGINES); + reset_count = fake_hangcheck(i915, ALL_ENGINES); timeout = i915_request_wait(rq, I915_WAIT_LOCKED, 10); if (timeout < 0) { @@ -971,7 +949,6 @@ static int igt_reset_wait(void *arg) goto out_rq; } - GEM_BUG_ON(test_bit(I915_RESET_HANDOFF, &i915->gpu_error.flags)); if (i915_reset_count(&i915->gpu_error) == reset_count) { pr_err("No GPU reset recorded!\n"); err = -EINVAL; @@ -1150,7 +1127,7 @@ static int __igt_reset_evict_vma(struct drm_i915_private *i915, wait_for_completion(&arg.completion); - if (wait_for(waitqueue_active(&rq->execute), 10)) { + if (wait_for(!list_empty(&rq->fence.cb_list), 10)) { struct drm_printer p = drm_info_printer(i915->drm.dev); pr_err("igt/evict_vma kthread did not wait\n"); @@ -1161,7 +1138,7 @@ static int __igt_reset_evict_vma(struct drm_i915_private *i915, } out_reset: - fake_hangcheck(rq, intel_engine_flag(rq->engine)); + fake_hangcheck(rq->i915, intel_engine_flag(rq->engine)); if (tsk) { struct igt_wedge_me w; @@ -1340,12 +1317,7 @@ static int igt_reset_queue(void *arg) goto fini; } - reset_count = fake_hangcheck(prev, ENGINE_MASK(id)); - - i915_reset(i915, ENGINE_MASK(id), NULL); - - GEM_BUG_ON(test_bit(I915_RESET_HANDOFF, - &i915->gpu_error.flags)); + reset_count = fake_hangcheck(i915, ENGINE_MASK(id)); if (prev->fence.error != -EIO) { pr_err("GPU reset not recorded on hanging request [fence.error=%d]!\n", @@ -1564,6 +1536,7 @@ static int igt_atomic_reset_engine(struct intel_engine_cs *engine, pr_err("%s(%s): Failed to start request %llx, at %x\n", __func__, engine->name, rq->fence.seqno, hws_seqno(&h, rq)); + i915_gem_set_wedged(i915); err = -EIO; } @@ -1587,7 +1560,6 @@ out: static void force_reset(struct drm_i915_private *i915) { i915_gem_set_wedged(i915); - set_bit(I915_RESET_HANDOFF, &i915->gpu_error.flags); i915_reset(i915, 0, NULL); } @@ -1600,6 +1572,7 @@ static int igt_atomic_reset(void *arg) { } }; struct drm_i915_private *i915 = arg; + intel_wakeref_t wakeref; int err = 0; /* Check that the resets are usable from atomic context */ @@ -1609,13 +1582,33 @@ static int igt_atomic_reset(void *arg) igt_global_reset_lock(i915); mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); /* Flush any requests before we get started and check basics */ force_reset(i915); if (i915_terminally_wedged(&i915->gpu_error)) goto unlock; + if (intel_has_gpu_reset(i915)) { + const typeof(*phases) *p; + + for (p = phases; p->name; p++) { + GEM_TRACE("intel_gpu_reset under %s\n", p->name); + + p->critical_section_begin(); + err = intel_gpu_reset(i915, ALL_ENGINES); + p->critical_section_end(); + + if (err) { + pr_err("intel_gpu_reset failed under %s\n", + p->name); + goto out; + } + } + + force_reset(i915); + } + if (intel_has_reset_engine(i915)) { struct intel_engine_cs *engine; enum intel_engine_id id; @@ -1636,7 +1629,7 @@ out: force_reset(i915); unlock: - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); igt_global_reset_unlock(i915); @@ -1660,6 +1653,7 @@ int intel_hangcheck_live_selftests(struct drm_i915_private *i915) SUBTEST(igt_handle_error), SUBTEST(igt_atomic_reset), }; + intel_wakeref_t wakeref; bool saved_hangcheck; int err; @@ -1669,8 +1663,9 @@ int intel_hangcheck_live_selftests(struct drm_i915_private *i915) if (i915_terminally_wedged(&i915->gpu_error)) return -EIO; /* we're long past hope of a successful reset */ - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); saved_hangcheck = fetch_and_zero(&i915_modparams.enable_hangcheck); + drain_delayed_work(&i915->gpu_error.hangcheck_work); /* flush param */ err = i915_subtests(tests, i915); @@ -1679,7 +1674,7 @@ int intel_hangcheck_live_selftests(struct drm_i915_private *i915) mutex_unlock(&i915->drm.struct_mutex); i915_modparams.enable_hangcheck = saved_hangcheck; - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); return err; } diff --git a/drivers/gpu/drm/i915/selftests/intel_lrc.c b/drivers/gpu/drm/i915/selftests/intel_lrc.c index 00caaa00f02f..58144e024751 100644 --- a/drivers/gpu/drm/i915/selftests/intel_lrc.c +++ b/drivers/gpu/drm/i915/selftests/intel_lrc.c @@ -4,6 +4,10 @@ * Copyright © 2018 Intel Corporation */ +#include <linux/prime_numbers.h> + +#include "../i915_reset.h" + #include "../i915_selftest.h" #include "igt_flush_test.h" #include "igt_spinner.h" @@ -18,13 +22,14 @@ static int live_sanitycheck(void *arg) struct i915_gem_context *ctx; enum intel_engine_id id; struct igt_spinner spin; + intel_wakeref_t wakeref; int err = -ENOMEM; if (!HAS_LOGICAL_RING_CONTEXTS(i915)) return 0; mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); if (igt_spinner_init(&spin, i915)) goto err_unlock; @@ -65,7 +70,7 @@ err_spin: igt_spinner_fini(&spin); err_unlock: igt_flush_test(i915, I915_WAIT_LOCKED); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); return err; } @@ -77,13 +82,14 @@ static int live_preempt(void *arg) struct igt_spinner spin_hi, spin_lo; struct intel_engine_cs *engine; enum intel_engine_id id; + intel_wakeref_t wakeref; int err = -ENOMEM; if (!HAS_LOGICAL_RING_PREEMPTION(i915)) return 0; mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); if (igt_spinner_init(&spin_hi, i915)) goto err_unlock; @@ -158,7 +164,7 @@ err_spin_hi: igt_spinner_fini(&spin_hi); err_unlock: igt_flush_test(i915, I915_WAIT_LOCKED); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); return err; } @@ -171,13 +177,14 @@ static int live_late_preempt(void *arg) struct intel_engine_cs *engine; struct i915_sched_attr attr = {}; enum intel_engine_id id; + intel_wakeref_t wakeref; int err = -ENOMEM; if (!HAS_LOGICAL_RING_PREEMPTION(i915)) return 0; mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); if (igt_spinner_init(&spin_hi, i915)) goto err_unlock; @@ -251,7 +258,7 @@ err_spin_hi: igt_spinner_fini(&spin_hi); err_unlock: igt_flush_test(i915, I915_WAIT_LOCKED); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); return err; @@ -263,6 +270,243 @@ err_wedged: goto err_ctx_lo; } +struct preempt_client { + struct igt_spinner spin; + struct i915_gem_context *ctx; +}; + +static int preempt_client_init(struct drm_i915_private *i915, + struct preempt_client *c) +{ + c->ctx = kernel_context(i915); + if (!c->ctx) + return -ENOMEM; + + if (igt_spinner_init(&c->spin, i915)) + goto err_ctx; + + return 0; + +err_ctx: + kernel_context_close(c->ctx); + return -ENOMEM; +} + +static void preempt_client_fini(struct preempt_client *c) +{ + igt_spinner_fini(&c->spin); + kernel_context_close(c->ctx); +} + +static int live_suppress_self_preempt(void *arg) +{ + struct drm_i915_private *i915 = arg; + struct intel_engine_cs *engine; + struct i915_sched_attr attr = { + .priority = I915_USER_PRIORITY(I915_PRIORITY_MAX) + }; + struct preempt_client a, b; + enum intel_engine_id id; + intel_wakeref_t wakeref; + int err = -ENOMEM; + + /* + * Verify that if a preemption request does not cause a change in + * the current execution order, the preempt-to-idle injection is + * skipped and that we do not accidentally apply it after the CS + * completion event. + */ + + if (!HAS_LOGICAL_RING_PREEMPTION(i915)) + return 0; + + if (USES_GUC_SUBMISSION(i915)) + return 0; /* presume black blox */ + + mutex_lock(&i915->drm.struct_mutex); + wakeref = intel_runtime_pm_get(i915); + + if (preempt_client_init(i915, &a)) + goto err_unlock; + if (preempt_client_init(i915, &b)) + goto err_client_a; + + for_each_engine(engine, i915, id) { + struct i915_request *rq_a, *rq_b; + int depth; + + engine->execlists.preempt_hang.count = 0; + + rq_a = igt_spinner_create_request(&a.spin, + a.ctx, engine, + MI_NOOP); + if (IS_ERR(rq_a)) { + err = PTR_ERR(rq_a); + goto err_client_b; + } + + i915_request_add(rq_a); + if (!igt_wait_for_spinner(&a.spin, rq_a)) { + pr_err("First client failed to start\n"); + goto err_wedged; + } + + for (depth = 0; depth < 8; depth++) { + rq_b = igt_spinner_create_request(&b.spin, + b.ctx, engine, + MI_NOOP); + if (IS_ERR(rq_b)) { + err = PTR_ERR(rq_b); + goto err_client_b; + } + i915_request_add(rq_b); + + GEM_BUG_ON(i915_request_completed(rq_a)); + engine->schedule(rq_a, &attr); + igt_spinner_end(&a.spin); + + if (!igt_wait_for_spinner(&b.spin, rq_b)) { + pr_err("Second client failed to start\n"); + goto err_wedged; + } + + swap(a, b); + rq_a = rq_b; + } + igt_spinner_end(&a.spin); + + if (engine->execlists.preempt_hang.count) { + pr_err("Preemption recorded x%d, depth %d; should have been suppressed!\n", + engine->execlists.preempt_hang.count, + depth); + err = -EINVAL; + goto err_client_b; + } + + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + goto err_wedged; + } + + err = 0; +err_client_b: + preempt_client_fini(&b); +err_client_a: + preempt_client_fini(&a); +err_unlock: + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + err = -EIO; + intel_runtime_pm_put(i915, wakeref); + mutex_unlock(&i915->drm.struct_mutex); + return err; + +err_wedged: + igt_spinner_end(&b.spin); + igt_spinner_end(&a.spin); + i915_gem_set_wedged(i915); + err = -EIO; + goto err_client_b; +} + +static int live_chain_preempt(void *arg) +{ + struct drm_i915_private *i915 = arg; + struct intel_engine_cs *engine; + struct preempt_client hi, lo; + enum intel_engine_id id; + intel_wakeref_t wakeref; + int err = -ENOMEM; + + /* + * Build a chain AB...BA between two contexts (A, B) and request + * preemption of the last request. It should then complete before + * the previously submitted spinner in B. + */ + + if (!HAS_LOGICAL_RING_PREEMPTION(i915)) + return 0; + + mutex_lock(&i915->drm.struct_mutex); + wakeref = intel_runtime_pm_get(i915); + + if (preempt_client_init(i915, &hi)) + goto err_unlock; + + if (preempt_client_init(i915, &lo)) + goto err_client_hi; + + for_each_engine(engine, i915, id) { + struct i915_sched_attr attr = { + .priority = I915_USER_PRIORITY(I915_PRIORITY_MAX), + }; + int count, i; + + for_each_prime_number_from(count, 1, 32) { /* must fit ring! */ + struct i915_request *rq; + + rq = igt_spinner_create_request(&hi.spin, + hi.ctx, engine, + MI_ARB_CHECK); + if (IS_ERR(rq)) + goto err_wedged; + i915_request_add(rq); + if (!igt_wait_for_spinner(&hi.spin, rq)) + goto err_wedged; + + rq = igt_spinner_create_request(&lo.spin, + lo.ctx, engine, + MI_ARB_CHECK); + if (IS_ERR(rq)) + goto err_wedged; + i915_request_add(rq); + + for (i = 0; i < count; i++) { + rq = i915_request_alloc(engine, lo.ctx); + if (IS_ERR(rq)) + goto err_wedged; + i915_request_add(rq); + } + + rq = i915_request_alloc(engine, hi.ctx); + if (IS_ERR(rq)) + goto err_wedged; + i915_request_add(rq); + engine->schedule(rq, &attr); + + igt_spinner_end(&hi.spin); + if (i915_request_wait(rq, I915_WAIT_LOCKED, HZ / 5) < 0) { + struct drm_printer p = + drm_info_printer(i915->drm.dev); + + pr_err("Failed to preempt over chain of %d\n", + count); + intel_engine_dump(engine, &p, + "%s\n", engine->name); + goto err_wedged; + } + igt_spinner_end(&lo.spin); + } + } + + err = 0; +err_client_lo: + preempt_client_fini(&lo); +err_client_hi: + preempt_client_fini(&hi); +err_unlock: + if (igt_flush_test(i915, I915_WAIT_LOCKED)) + err = -EIO; + intel_runtime_pm_put(i915, wakeref); + mutex_unlock(&i915->drm.struct_mutex); + return err; + +err_wedged: + igt_spinner_end(&hi.spin); + igt_spinner_end(&lo.spin); + i915_gem_set_wedged(i915); + err = -EIO; + goto err_client_lo; +} + static int live_preempt_hang(void *arg) { struct drm_i915_private *i915 = arg; @@ -270,6 +514,7 @@ static int live_preempt_hang(void *arg) struct igt_spinner spin_hi, spin_lo; struct intel_engine_cs *engine; enum intel_engine_id id; + intel_wakeref_t wakeref; int err = -ENOMEM; if (!HAS_LOGICAL_RING_PREEMPTION(i915)) @@ -279,7 +524,7 @@ static int live_preempt_hang(void *arg) return 0; mutex_lock(&i915->drm.struct_mutex); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); if (igt_spinner_init(&spin_hi, i915)) goto err_unlock; @@ -374,7 +619,7 @@ err_spin_hi: igt_spinner_fini(&spin_hi); err_unlock: igt_flush_test(i915, I915_WAIT_LOCKED); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); mutex_unlock(&i915->drm.struct_mutex); return err; } @@ -562,6 +807,7 @@ static int live_preempt_smoke(void *arg) .ncontext = 1024, }; const unsigned int phase[] = { 0, BATCH }; + intel_wakeref_t wakeref; int err = -ENOMEM; u32 *cs; int n; @@ -576,7 +822,7 @@ static int live_preempt_smoke(void *arg) return -ENOMEM; mutex_lock(&smoke.i915->drm.struct_mutex); - intel_runtime_pm_get(smoke.i915); + wakeref = intel_runtime_pm_get(smoke.i915); smoke.batch = i915_gem_object_create_internal(smoke.i915, PAGE_SIZE); if (IS_ERR(smoke.batch)) { @@ -627,7 +873,7 @@ err_ctx: err_batch: i915_gem_object_put(smoke.batch); err_unlock: - intel_runtime_pm_put(smoke.i915); + intel_runtime_pm_put(smoke.i915, wakeref); mutex_unlock(&smoke.i915->drm.struct_mutex); kfree(smoke.contexts); @@ -640,6 +886,8 @@ int intel_execlists_live_selftests(struct drm_i915_private *i915) SUBTEST(live_sanitycheck), SUBTEST(live_preempt), SUBTEST(live_late_preempt), + SUBTEST(live_suppress_self_preempt), + SUBTEST(live_chain_preempt), SUBTEST(live_preempt_hang), SUBTEST(live_preempt_smoke), }; diff --git a/drivers/gpu/drm/i915/selftests/intel_workarounds.c b/drivers/gpu/drm/i915/selftests/intel_workarounds.c index 8b3f3200a3bd..b15c4f26c593 100644 --- a/drivers/gpu/drm/i915/selftests/intel_workarounds.c +++ b/drivers/gpu/drm/i915/selftests/intel_workarounds.c @@ -5,6 +5,7 @@ */ #include "../i915_selftest.h" +#include "../i915_reset.h" #include "igt_flush_test.h" #include "igt_reset.h" @@ -60,10 +61,11 @@ reference_lists_fini(struct drm_i915_private *i915, struct wa_lists *lists) static struct drm_i915_gem_object * read_nonprivs(struct i915_gem_context *ctx, struct intel_engine_cs *engine) { + const u32 base = engine->mmio_base; struct drm_i915_gem_object *result; + intel_wakeref_t wakeref; struct i915_request *rq; struct i915_vma *vma; - const u32 base = engine->mmio_base; u32 srm, *cs; int err; int i; @@ -92,9 +94,9 @@ read_nonprivs(struct i915_gem_context *ctx, struct intel_engine_cs *engine) if (err) goto err_obj; - intel_runtime_pm_get(engine->i915); - rq = i915_request_alloc(engine, ctx); - intel_runtime_pm_put(engine->i915); + rq = ERR_PTR(-ENODEV); + with_intel_runtime_pm(engine->i915, wakeref) + rq = i915_request_alloc(engine, ctx); if (IS_ERR(rq)) { err = PTR_ERR(rq); goto err_pin; @@ -212,7 +214,6 @@ out_put: static int do_device_reset(struct intel_engine_cs *engine) { - set_bit(I915_RESET_HANDOFF, &engine->i915->gpu_error.flags); i915_reset(engine->i915, ENGINE_MASK(engine->id), "live_workarounds"); return 0; } @@ -228,20 +229,22 @@ switch_to_scratch_context(struct intel_engine_cs *engine, { struct i915_gem_context *ctx; struct i915_request *rq; + intel_wakeref_t wakeref; int err = 0; ctx = kernel_context(engine->i915); if (IS_ERR(ctx)) return PTR_ERR(ctx); - intel_runtime_pm_get(engine->i915); - - if (spin) - rq = igt_spinner_create_request(spin, ctx, engine, MI_NOOP); - else - rq = i915_request_alloc(engine, ctx); - - intel_runtime_pm_put(engine->i915); + rq = ERR_PTR(-ENODEV); + with_intel_runtime_pm(engine->i915, wakeref) { + if (spin) + rq = igt_spinner_create_request(spin, + ctx, engine, + MI_NOOP); + else + rq = i915_request_alloc(engine, ctx); + } kernel_context_close(ctx); @@ -273,6 +276,7 @@ static int check_whitelist_across_reset(struct intel_engine_cs *engine, bool want_spin = reset == do_engine_reset; struct i915_gem_context *ctx; struct igt_spinner spin; + intel_wakeref_t wakeref; int err; pr_info("Checking %d whitelisted registers (RING_NONPRIV) [%s]\n", @@ -298,9 +302,8 @@ static int check_whitelist_across_reset(struct intel_engine_cs *engine, if (err) goto out; - intel_runtime_pm_get(i915); - err = reset(engine); - intel_runtime_pm_put(i915); + with_intel_runtime_pm(i915, wakeref) + err = reset(engine); if (want_spin) { igt_spinner_end(&spin); @@ -390,7 +393,7 @@ static int live_gpu_reset_gt_engine_workarounds(void *arg) { struct drm_i915_private *i915 = arg; - struct i915_gpu_error *error = &i915->gpu_error; + intel_wakeref_t wakeref; struct wa_lists lists; bool ok; @@ -400,21 +403,21 @@ live_gpu_reset_gt_engine_workarounds(void *arg) pr_info("Verifying after GPU reset...\n"); igt_global_reset_lock(i915); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); + reference_lists_init(i915, &lists); ok = verify_gt_engine_wa(i915, &lists, "before reset"); if (!ok) goto out; - set_bit(I915_RESET_HANDOFF, &error->flags); i915_reset(i915, ALL_ENGINES, "live_workarounds"); ok = verify_gt_engine_wa(i915, &lists, "after reset"); out: reference_lists_fini(i915, &lists); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); igt_global_reset_unlock(i915); return ok ? 0 : -ESRCH; @@ -429,6 +432,7 @@ live_engine_reset_gt_engine_workarounds(void *arg) struct igt_spinner spin; enum intel_engine_id id; struct i915_request *rq; + intel_wakeref_t wakeref; struct wa_lists lists; int ret = 0; @@ -440,7 +444,8 @@ live_engine_reset_gt_engine_workarounds(void *arg) return PTR_ERR(ctx); igt_global_reset_lock(i915); - intel_runtime_pm_get(i915); + wakeref = intel_runtime_pm_get(i915); + reference_lists_init(i915, &lists); for_each_engine(engine, i915, id) { @@ -496,7 +501,7 @@ live_engine_reset_gt_engine_workarounds(void *arg) err: reference_lists_fini(i915, &lists); - intel_runtime_pm_put(i915); + intel_runtime_pm_put(i915, wakeref); igt_global_reset_unlock(i915); kernel_context_close(ctx); diff --git a/drivers/gpu/drm/i915/selftests/lib_sw_fence.c b/drivers/gpu/drm/i915/selftests/lib_sw_fence.c index b26f07b55d86..2bfa72c1654b 100644 --- a/drivers/gpu/drm/i915/selftests/lib_sw_fence.c +++ b/drivers/gpu/drm/i915/selftests/lib_sw_fence.c @@ -76,3 +76,57 @@ void timed_fence_fini(struct timed_fence *tf) destroy_timer_on_stack(&tf->timer); i915_sw_fence_fini(&tf->fence); } + +struct heap_fence { + struct i915_sw_fence fence; + union { + struct kref ref; + struct rcu_head rcu; + }; +}; + +static int __i915_sw_fence_call +heap_fence_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state) +{ + struct heap_fence *h = container_of(fence, typeof(*h), fence); + + switch (state) { + case FENCE_COMPLETE: + break; + + case FENCE_FREE: + heap_fence_put(&h->fence); + } + + return NOTIFY_DONE; +} + +struct i915_sw_fence *heap_fence_create(gfp_t gfp) +{ + struct heap_fence *h; + + h = kmalloc(sizeof(*h), gfp); + if (!h) + return NULL; + + i915_sw_fence_init(&h->fence, heap_fence_notify); + refcount_set(&h->ref.refcount, 2); + + return &h->fence; +} + +static void heap_fence_release(struct kref *ref) +{ + struct heap_fence *h = container_of(ref, typeof(*h), ref); + + i915_sw_fence_fini(&h->fence); + + kfree_rcu(h, rcu); +} + +void heap_fence_put(struct i915_sw_fence *fence) +{ + struct heap_fence *h = container_of(fence, typeof(*h), fence); + + kref_put(&h->ref, heap_fence_release); +} diff --git a/drivers/gpu/drm/i915/selftests/lib_sw_fence.h b/drivers/gpu/drm/i915/selftests/lib_sw_fence.h index 474aafb92ae1..1f9927e10f3a 100644 --- a/drivers/gpu/drm/i915/selftests/lib_sw_fence.h +++ b/drivers/gpu/drm/i915/selftests/lib_sw_fence.h @@ -39,4 +39,7 @@ struct timed_fence { void timed_fence_init(struct timed_fence *tf, unsigned long expires); void timed_fence_fini(struct timed_fence *tf); +struct i915_sw_fence *heap_fence_create(gfp_t gfp); +void heap_fence_put(struct i915_sw_fence *fence); + #endif /* _LIB_SW_FENCE_H_ */ diff --git a/drivers/gpu/drm/i915/selftests/mock_context.c b/drivers/gpu/drm/i915/selftests/mock_context.c index d937bdff26f9..b646cdcdd602 100644 --- a/drivers/gpu/drm/i915/selftests/mock_context.c +++ b/drivers/gpu/drm/i915/selftests/mock_context.c @@ -45,11 +45,8 @@ mock_context(struct drm_i915_private *i915, INIT_LIST_HEAD(&ctx->handles_list); INIT_LIST_HEAD(&ctx->hw_id_link); - for (n = 0; n < ARRAY_SIZE(ctx->__engine); n++) { - struct intel_context *ce = &ctx->__engine[n]; - - ce->gem_context = ctx; - } + for (n = 0; n < ARRAY_SIZE(ctx->__engine); n++) + intel_context_init(&ctx->__engine[n], ctx, i915->engine[n]); ret = i915_gem_context_pin_hw_id(ctx); if (ret < 0) diff --git a/drivers/gpu/drm/i915/selftests/mock_engine.c b/drivers/gpu/drm/i915/selftests/mock_engine.c index 50e1a0b1af7e..08f0cab02e0f 100644 --- a/drivers/gpu/drm/i915/selftests/mock_engine.c +++ b/drivers/gpu/drm/i915/selftests/mock_engine.c @@ -30,6 +30,52 @@ struct mock_ring { struct i915_timeline timeline; }; +static void mock_timeline_pin(struct i915_timeline *tl) +{ + tl->pin_count++; +} + +static void mock_timeline_unpin(struct i915_timeline *tl) +{ + GEM_BUG_ON(!tl->pin_count); + tl->pin_count--; +} + +static struct intel_ring *mock_ring(struct intel_engine_cs *engine) +{ + const unsigned long sz = PAGE_SIZE / 2; + struct mock_ring *ring; + + ring = kzalloc(sizeof(*ring) + sz, GFP_KERNEL); + if (!ring) + return NULL; + + if (i915_timeline_init(engine->i915, + &ring->timeline, engine->name, + NULL)) { + kfree(ring); + return NULL; + } + + ring->base.size = sz; + ring->base.effective_size = sz; + ring->base.vaddr = (void *)(ring + 1); + ring->base.timeline = &ring->timeline; + + INIT_LIST_HEAD(&ring->base.request_list); + intel_ring_update_space(&ring->base); + + return &ring->base; +} + +static void mock_ring_free(struct intel_ring *base) +{ + struct mock_ring *ring = container_of(base, typeof(*ring), base); + + i915_timeline_fini(&ring->timeline); + kfree(ring); +} + static struct mock_request *first_request(struct mock_engine *engine) { return list_first_entry_or_null(&engine->hw_queue, @@ -37,24 +83,29 @@ static struct mock_request *first_request(struct mock_engine *engine) link); } -static void advance(struct mock_engine *engine, - struct mock_request *request) +static void advance(struct mock_request *request) { list_del_init(&request->link); - mock_seqno_advance(&engine->base, request->base.global_seqno); + intel_engine_write_global_seqno(request->base.engine, + request->base.global_seqno); + i915_request_mark_complete(&request->base); + GEM_BUG_ON(!i915_request_completed(&request->base)); + + intel_engine_queue_breadcrumbs(request->base.engine); } static void hw_delay_complete(struct timer_list *t) { struct mock_engine *engine = from_timer(engine, t, hw_delay); struct mock_request *request; + unsigned long flags; - spin_lock(&engine->hw_lock); + spin_lock_irqsave(&engine->hw_lock, flags); /* Timer fired, first request is complete */ request = first_request(engine); if (request) - advance(engine, request); + advance(request); /* * Also immediately signal any subsequent 0-delay requests, but @@ -66,20 +117,24 @@ static void hw_delay_complete(struct timer_list *t) break; } - advance(engine, request); + advance(request); } - spin_unlock(&engine->hw_lock); + spin_unlock_irqrestore(&engine->hw_lock, flags); } static void mock_context_unpin(struct intel_context *ce) { + mock_timeline_unpin(ce->ring->timeline); i915_gem_context_put(ce->gem_context); } static void mock_context_destroy(struct intel_context *ce) { GEM_BUG_ON(ce->pin_count); + + if (ce->ring) + mock_ring_free(ce->ring); } static const struct intel_context_ops mock_context_ops = { @@ -92,14 +147,26 @@ mock_context_pin(struct intel_engine_cs *engine, struct i915_gem_context *ctx) { struct intel_context *ce = to_intel_context(ctx, engine); + int err = -ENOMEM; + + if (ce->pin_count++) + return ce; - if (!ce->pin_count++) { - i915_gem_context_get(ctx); - ce->ring = engine->buffer; - ce->ops = &mock_context_ops; + if (!ce->ring) { + ce->ring = mock_ring(engine); + if (!ce->ring) + goto err; } + mock_timeline_pin(ce->ring->timeline); + + ce->ops = &mock_context_ops; + i915_gem_context_get(ctx); return ce; + +err: + ce->pin_count = 0; + return ERR_PTR(err); } static int mock_request_alloc(struct i915_request *request) @@ -118,9 +185,9 @@ static int mock_emit_flush(struct i915_request *request, return 0; } -static void mock_emit_breadcrumb(struct i915_request *request, - u32 *flags) +static u32 *mock_emit_breadcrumb(struct i915_request *request, u32 *cs) { + return cs; } static void mock_submit_request(struct i915_request *request) @@ -128,49 +195,20 @@ static void mock_submit_request(struct i915_request *request) struct mock_request *mock = container_of(request, typeof(*mock), base); struct mock_engine *engine = container_of(request->engine, typeof(*engine), base); + unsigned long flags; i915_request_submit(request); GEM_BUG_ON(!request->global_seqno); - spin_lock_irq(&engine->hw_lock); + spin_lock_irqsave(&engine->hw_lock, flags); list_add_tail(&mock->link, &engine->hw_queue); if (mock->link.prev == &engine->hw_queue) { if (mock->delay) mod_timer(&engine->hw_delay, jiffies + mock->delay); else - advance(engine, mock); + advance(mock); } - spin_unlock_irq(&engine->hw_lock); -} - -static struct intel_ring *mock_ring(struct intel_engine_cs *engine) -{ - const unsigned long sz = PAGE_SIZE / 2; - struct mock_ring *ring; - - ring = kzalloc(sizeof(*ring) + sz, GFP_KERNEL); - if (!ring) - return NULL; - - i915_timeline_init(engine->i915, &ring->timeline, engine->name); - - ring->base.size = sz; - ring->base.effective_size = sz; - ring->base.vaddr = (void *)(ring + 1); - ring->base.timeline = &ring->timeline; - - INIT_LIST_HEAD(&ring->base.request_list); - intel_ring_update_space(&ring->base); - - return &ring->base; -} - -static void mock_ring_free(struct intel_ring *base) -{ - struct mock_ring *ring = container_of(base, typeof(*ring), base); - - i915_timeline_fini(&ring->timeline); - kfree(ring); + spin_unlock_irqrestore(&engine->hw_lock, flags); } struct intel_engine_cs *mock_engine(struct drm_i915_private *i915, @@ -189,39 +227,37 @@ struct intel_engine_cs *mock_engine(struct drm_i915_private *i915, engine->base.i915 = i915; snprintf(engine->base.name, sizeof(engine->base.name), "%s", name); engine->base.id = id; - engine->base.status_page.page_addr = (void *)(engine + 1); + engine->base.status_page.addr = (void *)(engine + 1); engine->base.context_pin = mock_context_pin; engine->base.request_alloc = mock_request_alloc; engine->base.emit_flush = mock_emit_flush; - engine->base.emit_breadcrumb = mock_emit_breadcrumb; + engine->base.emit_fini_breadcrumb = mock_emit_breadcrumb; engine->base.submit_request = mock_submit_request; - i915_timeline_init(i915, &engine->base.timeline, engine->base.name); + if (i915_timeline_init(i915, + &engine->base.timeline, + engine->base.name, + NULL)) + goto err_free; i915_timeline_set_subclass(&engine->base.timeline, TIMELINE_ENGINE); intel_engine_init_breadcrumbs(&engine->base); - engine->base.breadcrumbs.mock = true; /* prevent touching HW for irqs */ /* fake hw queue */ spin_lock_init(&engine->hw_lock); timer_setup(&engine->hw_delay, hw_delay_complete, 0); INIT_LIST_HEAD(&engine->hw_queue); - engine->base.buffer = mock_ring(&engine->base); - if (!engine->base.buffer) - goto err_breadcrumbs; - if (IS_ERR(intel_context_pin(i915->kernel_context, &engine->base))) - goto err_ring; + goto err_breadcrumbs; return &engine->base; -err_ring: - mock_ring_free(engine->base.buffer); err_breadcrumbs: intel_engine_fini_breadcrumbs(&engine->base); i915_timeline_fini(&engine->base.timeline); +err_free: kfree(engine); return NULL; } @@ -235,16 +271,14 @@ void mock_engine_flush(struct intel_engine_cs *engine) del_timer_sync(&mock->hw_delay); spin_lock_irq(&mock->hw_lock); - list_for_each_entry_safe(request, rn, &mock->hw_queue, link) { - list_del_init(&request->link); - mock_seqno_advance(&mock->base, request->base.global_seqno); - } + list_for_each_entry_safe(request, rn, &mock->hw_queue, link) + advance(request); spin_unlock_irq(&mock->hw_lock); } void mock_engine_reset(struct intel_engine_cs *engine) { - intel_write_status_page(engine, I915_GEM_HWS_INDEX, 0); + intel_engine_write_global_seqno(engine, 0); } void mock_engine_free(struct intel_engine_cs *engine) @@ -261,8 +295,6 @@ void mock_engine_free(struct intel_engine_cs *engine) __intel_context_unpin(engine->i915->kernel_context, engine); - mock_ring_free(engine->buffer); - intel_engine_fini_breadcrumbs(engine); i915_timeline_fini(&engine->timeline); diff --git a/drivers/gpu/drm/i915/selftests/mock_engine.h b/drivers/gpu/drm/i915/selftests/mock_engine.h index 133d0c21790d..b9cc3a245f16 100644 --- a/drivers/gpu/drm/i915/selftests/mock_engine.h +++ b/drivers/gpu/drm/i915/selftests/mock_engine.h @@ -46,10 +46,4 @@ void mock_engine_flush(struct intel_engine_cs *engine); void mock_engine_reset(struct intel_engine_cs *engine); void mock_engine_free(struct intel_engine_cs *engine); -static inline void mock_seqno_advance(struct intel_engine_cs *engine, u32 seqno) -{ - intel_write_status_page(engine, I915_GEM_HWS_INDEX, seqno); - intel_engine_wakeup(engine); -} - #endif /* !__MOCK_ENGINE_H__ */ diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.c b/drivers/gpu/drm/i915/selftests/mock_gem_device.c index baa3c38919de..14ae46fda49f 100644 --- a/drivers/gpu/drm/i915/selftests/mock_gem_device.c +++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.c @@ -58,8 +58,8 @@ static void mock_device_release(struct drm_device *dev) i915_gem_contexts_lost(i915); mutex_unlock(&i915->drm.struct_mutex); - cancel_delayed_work_sync(&i915->gt.retire_work); - cancel_delayed_work_sync(&i915->gt.idle_work); + drain_delayed_work(&i915->gt.retire_work); + drain_delayed_work(&i915->gt.idle_work); i915_gem_drain_workqueue(i915); mutex_lock(&i915->drm.struct_mutex); @@ -68,13 +68,14 @@ static void mock_device_release(struct drm_device *dev) i915_gem_contexts_fini(i915); mutex_unlock(&i915->drm.struct_mutex); + i915_timelines_fini(i915); + drain_workqueue(i915->wq); i915_gem_drain_freed_objects(i915); mutex_lock(&i915->drm.struct_mutex); - mock_fini_ggtt(i915); + mock_fini_ggtt(&i915->ggtt); mutex_unlock(&i915->drm.struct_mutex); - WARN_ON(!list_empty(&i915->gt.timelines)); destroy_workqueue(i915->wq); @@ -154,15 +155,17 @@ struct drm_i915_private *mock_gem_device(void) pdev->dev.archdata.iommu = (void *)-1; #endif + i915 = (struct drm_i915_private *)(pdev + 1); + pci_set_drvdata(pdev, i915); + + intel_runtime_pm_init_early(i915); + dev_pm_domain_set(&pdev->dev, &pm_domain); pm_runtime_enable(&pdev->dev); pm_runtime_dont_use_autosuspend(&pdev->dev); if (pm_runtime_enabled(&pdev->dev)) WARN_ON(pm_runtime_get_sync(&pdev->dev)); - i915 = (struct drm_i915_private *)(pdev + 1); - pci_set_drvdata(pdev, i915); - err = drm_dev_init(&i915->drm, &mock_driver, &pdev->dev); if (err) { pr_err("Failed to initialise mock GEM device: err=%d\n", err); @@ -186,6 +189,7 @@ struct drm_i915_private *mock_gem_device(void) init_waitqueue_head(&i915->gpu_error.wait_queue); init_waitqueue_head(&i915->gpu_error.reset_queue); + mutex_init(&i915->gpu_error.wedge_mutex); i915->wq = alloc_ordered_workqueue("mock", 0); if (!i915->wq) @@ -223,13 +227,14 @@ struct drm_i915_private *mock_gem_device(void) if (!i915->priorities) goto err_dependencies; - INIT_LIST_HEAD(&i915->gt.timelines); + i915_timelines_init(i915); + INIT_LIST_HEAD(&i915->gt.active_rings); INIT_LIST_HEAD(&i915->gt.closed_vma); mutex_lock(&i915->drm.struct_mutex); - mock_init_ggtt(i915); + mock_init_ggtt(i915, &i915->ggtt); mkwrite_device_info(i915)->ring_mask = BIT(0); i915->kernel_context = mock_context(i915, NULL); @@ -250,6 +255,7 @@ err_context: i915_gem_contexts_fini(i915); err_unlock: mutex_unlock(&i915->drm.struct_mutex); + i915_timelines_fini(i915); kmem_cache_destroy(i915->priorities); err_dependencies: kmem_cache_destroy(i915->dependencies); diff --git a/drivers/gpu/drm/i915/selftests/mock_gtt.c b/drivers/gpu/drm/i915/selftests/mock_gtt.c index 6ae418c76015..cd83929fde8e 100644 --- a/drivers/gpu/drm/i915/selftests/mock_gtt.c +++ b/drivers/gpu/drm/i915/selftests/mock_gtt.c @@ -70,7 +70,7 @@ mock_ppgtt(struct drm_i915_private *i915, ppgtt->vm.total = round_down(U64_MAX, PAGE_SIZE); ppgtt->vm.file = ERR_PTR(-ENODEV); - i915_address_space_init(&ppgtt->vm, i915); + i915_address_space_init(&ppgtt->vm, VM_CLASS_PPGTT); ppgtt->vm.clear_range = nop_clear_range; ppgtt->vm.insert_page = mock_insert_page; @@ -97,11 +97,12 @@ static void mock_unbind_ggtt(struct i915_vma *vma) { } -void mock_init_ggtt(struct drm_i915_private *i915) +void mock_init_ggtt(struct drm_i915_private *i915, struct i915_ggtt *ggtt) { - struct i915_ggtt *ggtt = &i915->ggtt; + memset(ggtt, 0, sizeof(*ggtt)); ggtt->vm.i915 = i915; + ggtt->vm.is_ggtt = true; ggtt->gmadr = (struct resource) DEFINE_RES_MEM(0, 2048 * PAGE_SIZE); ggtt->mappable_end = resource_size(&ggtt->gmadr); @@ -117,14 +118,10 @@ void mock_init_ggtt(struct drm_i915_private *i915) ggtt->vm.vma_ops.set_pages = ggtt_set_pages; ggtt->vm.vma_ops.clear_pages = clear_pages; - i915_address_space_init(&ggtt->vm, i915); - - ggtt->vm.is_ggtt = true; + i915_address_space_init(&ggtt->vm, VM_CLASS_GGTT); } -void mock_fini_ggtt(struct drm_i915_private *i915) +void mock_fini_ggtt(struct i915_ggtt *ggtt) { - struct i915_ggtt *ggtt = &i915->ggtt; - i915_address_space_fini(&ggtt->vm); } diff --git a/drivers/gpu/drm/i915/selftests/mock_gtt.h b/drivers/gpu/drm/i915/selftests/mock_gtt.h index 9a0a833bb545..40d544bde1d5 100644 --- a/drivers/gpu/drm/i915/selftests/mock_gtt.h +++ b/drivers/gpu/drm/i915/selftests/mock_gtt.h @@ -25,8 +25,8 @@ #ifndef __MOCK_GTT_H #define __MOCK_GTT_H -void mock_init_ggtt(struct drm_i915_private *i915); -void mock_fini_ggtt(struct drm_i915_private *i915); +void mock_init_ggtt(struct drm_i915_private *i915, struct i915_ggtt *ggtt); +void mock_fini_ggtt(struct i915_ggtt *ggtt); struct i915_hw_ppgtt * mock_ppgtt(struct drm_i915_private *i915, diff --git a/drivers/gpu/drm/i915/selftests/mock_timeline.c b/drivers/gpu/drm/i915/selftests/mock_timeline.c index dcf3b16f5a07..d2de9ece2118 100644 --- a/drivers/gpu/drm/i915/selftests/mock_timeline.c +++ b/drivers/gpu/drm/i915/selftests/mock_timeline.c @@ -10,11 +10,13 @@ void mock_timeline_init(struct i915_timeline *timeline, u64 context) { + timeline->i915 = NULL; timeline->fence_context = context; spin_lock_init(&timeline->lock); - init_request_active(&timeline->last_request, NULL); + INIT_ACTIVE_REQUEST(&timeline->barrier); + INIT_ACTIVE_REQUEST(&timeline->last_request); INIT_LIST_HEAD(&timeline->requests); i915_syncmap_init(&timeline->sync); @@ -24,5 +26,5 @@ void mock_timeline_init(struct i915_timeline *timeline, u64 context) void mock_timeline_fini(struct i915_timeline *timeline) { - i915_timeline_fini(timeline); + i915_syncmap_free(&timeline->sync); } diff --git a/drivers/gpu/drm/i915/vlv_dsi.c b/drivers/gpu/drm/i915/vlv_dsi.c index 3225621f820c..6403728fe778 100644 --- a/drivers/gpu/drm/i915/vlv_dsi.c +++ b/drivers/gpu/drm/i915/vlv_dsi.c @@ -275,7 +275,7 @@ static int intel_dsi_compute_config(struct intel_encoder *encoder, if (fixed_mode) { intel_fixed_panel_mode(fixed_mode, adjusted_mode); - if (HAS_GMCH_DISPLAY(dev_priv)) + if (HAS_GMCH(dev_priv)) intel_gmch_panel_fitting(crtc, pipe_config, conn_state->scaling_mode); else @@ -289,6 +289,11 @@ static int intel_dsi_compute_config(struct intel_encoder *encoder, /* DSI uses short packets for sync events, so clear mode flags for DSI */ adjusted_mode->flags = 0; + if (intel_dsi->pixel_format == MIPI_DSI_FMT_RGB888) + pipe_config->pipe_bpp = 24; + else + pipe_config->pipe_bpp = 18; + if (IS_GEN9_LP(dev_priv)) { /* Enable Frame time stamp based scanline reporting */ adjusted_mode->private_flags |= @@ -673,6 +678,10 @@ static void intel_dsi_port_enable(struct intel_encoder *encoder, LANE_CONFIGURATION_DUAL_LINK_B : LANE_CONFIGURATION_DUAL_LINK_A; } + + if (intel_dsi->pixel_format != MIPI_DSI_FMT_RGB888) + temp |= DITHERING_ENABLE; + /* assert ip_tg_enable signal */ I915_WRITE(port_ctrl, temp | DPI_ENABLE); POSTING_READ(port_ctrl); @@ -959,13 +968,15 @@ static bool intel_dsi_get_hw_state(struct intel_encoder *encoder, { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base); + intel_wakeref_t wakeref; enum port port; bool active = false; DRM_DEBUG_KMS("\n"); - if (!intel_display_power_get_if_enabled(dev_priv, - encoder->power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, + encoder->power_domain); + if (!wakeref) return false; /* @@ -1021,7 +1032,7 @@ static bool intel_dsi_get_hw_state(struct intel_encoder *encoder, } out_put_power: - intel_display_power_put(dev_priv, encoder->power_domain); + intel_display_power_put(dev_priv, encoder->power_domain, wakeref); return active; } @@ -1057,10 +1068,8 @@ static void bxt_dsi_get_pipe_config(struct intel_encoder *encoder, } fmt = I915_READ(MIPI_DSI_FUNC_PRG(port)) & VID_MODE_FORMAT_MASK; - pipe_config->pipe_bpp = - mipi_dsi_pixel_format_to_bpp( - pixel_format_from_register_bits(fmt)); - bpp = pipe_config->pipe_bpp; + bpp = mipi_dsi_pixel_format_to_bpp( + pixel_format_from_register_bits(fmt)); /* Enable Frame time stamo based scanline reporting */ adjusted_mode->private_flags |= @@ -1198,11 +1207,9 @@ static void intel_dsi_get_config(struct intel_encoder *encoder, if (IS_GEN9_LP(dev_priv)) { bxt_dsi_get_pipe_config(encoder, pipe_config); - pclk = bxt_dsi_get_pclk(encoder, pipe_config->pipe_bpp, - pipe_config); + pclk = bxt_dsi_get_pclk(encoder, pipe_config); } else { - pclk = vlv_dsi_get_pclk(encoder, pipe_config->pipe_bpp, - pipe_config); + pclk = vlv_dsi_get_pclk(encoder, pipe_config); } if (pclk) { @@ -1574,6 +1581,7 @@ vlv_dsi_get_hw_panel_orientation(struct intel_connector *connector) enum drm_panel_orientation orientation; struct intel_plane *plane; struct intel_crtc *crtc; + intel_wakeref_t wakeref; enum pipe pipe; u32 val; @@ -1584,7 +1592,8 @@ vlv_dsi_get_hw_panel_orientation(struct intel_connector *connector) plane = to_intel_plane(crtc->base.primary); power_domain = POWER_DOMAIN_PIPE(pipe); - if (!intel_display_power_get_if_enabled(dev_priv, power_domain)) + wakeref = intel_display_power_get_if_enabled(dev_priv, power_domain); + if (!wakeref) return DRM_MODE_PANEL_ORIENTATION_UNKNOWN; val = I915_READ(DSPCNTR(plane->i9xx_plane)); @@ -1596,7 +1605,7 @@ vlv_dsi_get_hw_panel_orientation(struct intel_connector *connector) else orientation = DRM_MODE_PANEL_ORIENTATION_NORMAL; - intel_display_power_put(dev_priv, power_domain); + intel_display_power_put(dev_priv, power_domain, wakeref); return orientation; } @@ -1624,7 +1633,7 @@ static void intel_dsi_add_properties(struct intel_connector *connector) u32 allowed_scalers; allowed_scalers = BIT(DRM_MODE_SCALE_ASPECT) | BIT(DRM_MODE_SCALE_FULLSCREEN); - if (!HAS_GMCH_DISPLAY(dev_priv)) + if (!HAS_GMCH(dev_priv)) allowed_scalers |= BIT(DRM_MODE_SCALE_CENTER); drm_connector_attach_scaling_mode_property(&connector->base, @@ -1688,6 +1697,7 @@ void vlv_dsi_init(struct drm_i915_private *dev_priv) intel_encoder->post_disable = intel_dsi_post_disable; intel_encoder->get_hw_state = intel_dsi_get_hw_state; intel_encoder->get_config = intel_dsi_get_config; + intel_encoder->update_pipe = intel_panel_update_backlight; intel_connector->get_hw_state = intel_connector_get_hw_state; diff --git a/drivers/gpu/drm/i915/vlv_dsi_pll.c b/drivers/gpu/drm/i915/vlv_dsi_pll.c index a132a8037ecc..954d5a8c4fa7 100644 --- a/drivers/gpu/drm/i915/vlv_dsi_pll.c +++ b/drivers/gpu/drm/i915/vlv_dsi_pll.c @@ -252,20 +252,12 @@ void bxt_dsi_pll_disable(struct intel_encoder *encoder) DRM_ERROR("Timeout waiting for PLL lock deassertion\n"); } -static void assert_bpp_mismatch(enum mipi_dsi_pixel_format fmt, int pipe_bpp) -{ - int bpp = mipi_dsi_pixel_format_to_bpp(fmt); - - WARN(bpp != pipe_bpp, - "bpp match assertion failure (expected %d, current %d)\n", - bpp, pipe_bpp); -} - -u32 vlv_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp, +u32 vlv_dsi_get_pclk(struct intel_encoder *encoder, struct intel_crtc_state *config) { struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base); + int bpp = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format); u32 dsi_clock, pclk; u32 pll_ctl, pll_div; u32 m = 0, p = 0, n; @@ -319,15 +311,12 @@ u32 vlv_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp, dsi_clock = (m * refclk) / (p * n); - /* pixel_format and pipe_bpp should agree */ - assert_bpp_mismatch(intel_dsi->pixel_format, pipe_bpp); - - pclk = DIV_ROUND_CLOSEST(dsi_clock * intel_dsi->lane_count, pipe_bpp); + pclk = DIV_ROUND_CLOSEST(dsi_clock * intel_dsi->lane_count, bpp); return pclk; } -u32 bxt_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp, +u32 bxt_dsi_get_pclk(struct intel_encoder *encoder, struct intel_crtc_state *config) { u32 pclk; @@ -335,12 +324,7 @@ u32 bxt_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp, u32 dsi_ratio; struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base); struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); - - /* Divide by zero */ - if (!pipe_bpp) { - DRM_ERROR("Invalid BPP(0)\n"); - return 0; - } + int bpp = mipi_dsi_pixel_format_to_bpp(intel_dsi->pixel_format); config->dsi_pll.ctrl = I915_READ(BXT_DSI_PLL_CTL); @@ -348,10 +332,7 @@ u32 bxt_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp, dsi_clk = (dsi_ratio * BXT_REF_CLOCK_KHZ) / 2; - /* pixel_format and pipe_bpp should agree */ - assert_bpp_mismatch(intel_dsi->pixel_format, pipe_bpp); - - pclk = DIV_ROUND_CLOSEST(dsi_clk * intel_dsi->lane_count, pipe_bpp); + pclk = DIV_ROUND_CLOSEST(dsi_clk * intel_dsi->lane_count, bpp); DRM_DEBUG_DRIVER("Calculated pclk=%u\n", pclk); return pclk; diff --git a/drivers/gpu/drm/mgag200/mgag200_main.c b/drivers/gpu/drm/mgag200/mgag200_main.c index 79d54103d470..163255099779 100644 --- a/drivers/gpu/drm/mgag200/mgag200_main.c +++ b/drivers/gpu/drm/mgag200/mgag200_main.c @@ -33,7 +33,7 @@ int mgag200_framebuffer_init(struct drm_device *dev, struct drm_gem_object *obj) { int ret; - + drm_helper_mode_fill_fb_struct(dev, &gfb->base, mode_cmd); gfb->obj = obj; ret = drm_framebuffer_init(dev, &gfb->base, &mga_fb_funcs); @@ -318,13 +318,9 @@ int mgag200_dumb_create(struct drm_file *file, static void mgag200_bo_unref(struct mgag200_bo **bo) { - struct ttm_buffer_object *tbo; - if ((*bo) == NULL) return; - - tbo = &((*bo)->bo); - ttm_bo_unref(&tbo); + ttm_bo_put(&((*bo)->bo)); *bo = NULL; } diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.h b/drivers/gpu/drm/nouveau/nouveau_bo.h index 73c48440d4d7..846f4bdec0de 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.h +++ b/drivers/gpu/drm/nouveau/nouveau_bo.h @@ -61,12 +61,14 @@ nouveau_bo_ref(struct nouveau_bo *ref, struct nouveau_bo **pnvbo) return -EINVAL; prev = *pnvbo; - *pnvbo = ref ? nouveau_bo(ttm_bo_reference(&ref->bo)) : NULL; - if (prev) { - struct ttm_buffer_object *bo = &prev->bo; - - ttm_bo_unref(&bo); + if (ref) { + ttm_bo_get(&ref->bo); + *pnvbo = nouveau_bo(&ref->bo); + } else { + *pnvbo = NULL; } + if (prev) + ttm_bo_put(&prev->bo); return 0; } diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c index b56524d343c3..fb028e3b5f51 100644 --- a/drivers/gpu/drm/nouveau/nouveau_gem.c +++ b/drivers/gpu/drm/nouveau/nouveau_gem.c @@ -41,7 +41,6 @@ nouveau_gem_object_del(struct drm_gem_object *gem) { struct nouveau_bo *nvbo = nouveau_gem_object(gem); struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev); - struct ttm_buffer_object *bo = &nvbo->bo; struct device *dev = drm->dev->dev; int ret; @@ -56,7 +55,7 @@ nouveau_gem_object_del(struct drm_gem_object *gem) /* reset filp so nouveau_bo_del_ttm() can test for it */ gem->filp = NULL; - ttm_bo_unref(&bo); + ttm_bo_put(&nvbo->bo); pm_runtime_mark_last_busy(dev); pm_runtime_put_autosuspend(dev); diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c index d587779a80b4..a97294ac96d5 100644 --- a/drivers/gpu/drm/radeon/ci_dpm.c +++ b/drivers/gpu/drm/radeon/ci_dpm.c @@ -5676,7 +5676,7 @@ int ci_dpm_init(struct radeon_device *rdev) u16 data_offset, size; u8 frev, crev; struct ci_power_info *pi; - enum pci_bus_speed speed_cap; + enum pci_bus_speed speed_cap = PCI_SPEED_UNKNOWN; struct pci_dev *root = rdev->pdev->bus->self; int ret; @@ -5685,7 +5685,8 @@ int ci_dpm_init(struct radeon_device *rdev) return -ENOMEM; rdev->pm.dpm.priv = pi; - speed_cap = pcie_get_speed_cap(root); + if (!pci_is_root_bus(rdev->pdev->bus)) + speed_cap = pcie_get_speed_cap(root); if (speed_cap == PCI_SPEED_UNKNOWN) { pi->sys_pcie_mask = 0; } else { diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c index 8fb60b3af015..0a785ef0ab66 100644 --- a/drivers/gpu/drm/radeon/si_dpm.c +++ b/drivers/gpu/drm/radeon/si_dpm.c @@ -6899,7 +6899,7 @@ int si_dpm_init(struct radeon_device *rdev) struct ni_power_info *ni_pi; struct si_power_info *si_pi; struct atom_clock_dividers dividers; - enum pci_bus_speed speed_cap; + enum pci_bus_speed speed_cap = PCI_SPEED_UNKNOWN; struct pci_dev *root = rdev->pdev->bus->self; int ret; @@ -6911,7 +6911,8 @@ int si_dpm_init(struct radeon_device *rdev) eg_pi = &ni_pi->eg; pi = &eg_pi->rv7xx; - speed_cap = pcie_get_speed_cap(root); + if (!pci_is_root_bus(rdev->pdev->bus)) + speed_cap = pcie_get_speed_cap(root); if (speed_cap == PCI_SPEED_UNKNOWN) { si_pi->sys_pcie_mask = 0; } else { diff --git a/drivers/gpu/drm/rcar-du/Kconfig b/drivers/gpu/drm/rcar-du/Kconfig index 225141656e19..7c36e2777a15 100644 --- a/drivers/gpu/drm/rcar-du/Kconfig +++ b/drivers/gpu/drm/rcar-du/Kconfig @@ -4,6 +4,7 @@ config DRM_RCAR_DU depends on DRM && OF depends on ARM || ARM64 depends on ARCH_RENESAS || COMPILE_TEST + imply DRM_RCAR_LVDS select DRM_KMS_HELPER select DRM_KMS_CMA_HELPER select DRM_GEM_CMA_HELPER diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c index 4d1e5098b70e..4cdea14d552f 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c +++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c @@ -28,6 +28,7 @@ #include "rcar_du_plane.h" #include "rcar_du_regs.h" #include "rcar_du_vsp.h" +#include "rcar_lvds.h" static u32 rcar_du_crtc_read(struct rcar_du_crtc *rcrtc, u32 reg) { @@ -659,8 +660,27 @@ static void rcar_du_crtc_atomic_enable(struct drm_crtc *crtc, struct drm_crtc_state *old_state) { struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc); + struct rcar_du_crtc_state *rstate = to_rcar_crtc_state(crtc->state); + struct rcar_du_device *rcdu = rcrtc->group->dev; rcar_du_crtc_get(rcrtc); + + /* + * On D3/E3 the dot clock is provided by the LVDS encoder attached to + * the DU channel. We need to enable its clock output explicitly if + * the LVDS output is disabled. + */ + if (rcdu->info->lvds_clk_mask & BIT(rcrtc->index) && + rstate->outputs == BIT(RCAR_DU_OUTPUT_DPAD0)) { + struct rcar_du_encoder *encoder = + rcdu->encoders[RCAR_DU_OUTPUT_LVDS0 + rcrtc->index]; + const struct drm_display_mode *mode = + &crtc->state->adjusted_mode; + + rcar_lvds_clk_enable(encoder->base.bridge, + mode->clock * 1000); + } + rcar_du_crtc_start(rcrtc); } @@ -668,10 +688,24 @@ static void rcar_du_crtc_atomic_disable(struct drm_crtc *crtc, struct drm_crtc_state *old_state) { struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc); + struct rcar_du_crtc_state *rstate = to_rcar_crtc_state(old_state); + struct rcar_du_device *rcdu = rcrtc->group->dev; rcar_du_crtc_stop(rcrtc); rcar_du_crtc_put(rcrtc); + if (rcdu->info->lvds_clk_mask & BIT(rcrtc->index) && + rstate->outputs == BIT(RCAR_DU_OUTPUT_DPAD0)) { + struct rcar_du_encoder *encoder = + rcdu->encoders[RCAR_DU_OUTPUT_LVDS0 + rcrtc->index]; + + /* + * Disable the LVDS clock output, see + * rcar_du_crtc_atomic_enable(). + */ + rcar_lvds_clk_disable(encoder->base.bridge); + } + spin_lock_irq(&crtc->dev->event_lock); if (crtc->state->event) { drm_crtc_send_vblank_event(crtc, crtc->state->event); diff --git a/drivers/gpu/drm/rcar-du/rcar_du_drv.h b/drivers/gpu/drm/rcar-du/rcar_du_drv.h index 6c187d0bf7c2..1327cd0df90a 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_drv.h +++ b/drivers/gpu/drm/rcar-du/rcar_du_drv.h @@ -22,6 +22,7 @@ struct device; struct drm_device; struct drm_property; struct rcar_du_device; +struct rcar_du_encoder; #define RCAR_DU_FEATURE_CRTC_IRQ_CLOCK BIT(0) /* Per-CRTC IRQ and clock */ #define RCAR_DU_FEATURE_VSP1_SOURCE BIT(1) /* Has inputs from VSP1 */ @@ -81,6 +82,8 @@ struct rcar_du_device { struct rcar_du_crtc crtcs[RCAR_DU_MAX_CRTCS]; unsigned int num_crtcs; + struct rcar_du_encoder *encoders[RCAR_DU_OUTPUT_MAX]; + struct rcar_du_group groups[RCAR_DU_MAX_GROUPS]; struct rcar_du_vsp vsps[RCAR_DU_MAX_VSPS]; diff --git a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c index 369e930750e9..8ee4e762f4e5 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c +++ b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c @@ -30,8 +30,7 @@ static const struct drm_encoder_funcs encoder_funcs = { int rcar_du_encoder_init(struct rcar_du_device *rcdu, enum rcar_du_output output, - struct device_node *enc_node, - struct device_node *con_node) + struct device_node *enc_node) { struct rcar_du_encoder *renc; struct drm_encoder *encoder; @@ -42,6 +41,7 @@ int rcar_du_encoder_init(struct rcar_du_device *rcdu, if (renc == NULL) return -ENOMEM; + rcdu->encoders[output] = renc; renc->output = output; encoder = rcar_encoder_to_drm_encoder(renc); diff --git a/drivers/gpu/drm/rcar-du/rcar_du_encoder.h b/drivers/gpu/drm/rcar-du/rcar_du_encoder.h index 552f2a02e5b5..df9be4524301 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_encoder.h +++ b/drivers/gpu/drm/rcar-du/rcar_du_encoder.h @@ -26,7 +26,6 @@ struct rcar_du_encoder { int rcar_du_encoder_init(struct rcar_du_device *rcdu, enum rcar_du_output output, - struct device_node *enc_node, - struct device_node *con_node); + struct device_node *enc_node); #endif /* __RCAR_DU_ENCODER_H__ */ diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c index 18d21bc1d265..3b7d50a8fb9b 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c @@ -332,17 +332,10 @@ static int rcar_du_encoders_init_one(struct rcar_du_device *rcdu, enum rcar_du_output output, struct of_endpoint *ep) { - struct device_node *connector = NULL; - struct device_node *encoder = NULL; - struct device_node *ep_node = NULL; - struct device_node *entity_ep_node; struct device_node *entity; int ret; - /* - * Locate the connected entity and infer its type from the number of - * endpoints. - */ + /* Locate the connected entity and initialize the encoder. */ entity = of_graph_get_remote_port_parent(ep->local_node); if (!entity) { dev_dbg(rcdu->dev, "unconnected endpoint %pOF, skipping\n", @@ -354,52 +347,17 @@ static int rcar_du_encoders_init_one(struct rcar_du_device *rcdu, dev_dbg(rcdu->dev, "connected entity %pOF is disabled, skipping\n", entity); + of_node_put(entity); return -ENODEV; } - entity_ep_node = of_graph_get_remote_endpoint(ep->local_node); - - for_each_endpoint_of_node(entity, ep_node) { - if (ep_node == entity_ep_node) - continue; - - /* - * We've found one endpoint other than the input, this must - * be an encoder. Locate the connector. - */ - encoder = entity; - connector = of_graph_get_remote_port_parent(ep_node); - of_node_put(ep_node); - - if (!connector) { - dev_warn(rcdu->dev, - "no connector for encoder %pOF, skipping\n", - encoder); - of_node_put(entity_ep_node); - of_node_put(encoder); - return -ENODEV; - } - - break; - } - - of_node_put(entity_ep_node); - - if (!encoder) { - dev_warn(rcdu->dev, - "no encoder found for endpoint %pOF, skipping\n", - ep->local_node); - return -ENODEV; - } - - ret = rcar_du_encoder_init(rcdu, output, encoder, connector); + ret = rcar_du_encoder_init(rcdu, output, entity); if (ret && ret != -EPROBE_DEFER) dev_warn(rcdu->dev, "failed to initialize encoder %pOF on output %u (%d), skipping\n", - encoder, output, ret); + entity, output, ret); - of_node_put(encoder); - of_node_put(connector); + of_node_put(entity); return ret; } diff --git a/drivers/gpu/drm/rcar-du/rcar_lvds.c b/drivers/gpu/drm/rcar-du/rcar_lvds.c index 044c65680ef3..7ef97b2a6eda 100644 --- a/drivers/gpu/drm/rcar-du/rcar_lvds.c +++ b/drivers/gpu/drm/rcar-du/rcar_lvds.c @@ -23,6 +23,7 @@ #include <drm/drm_panel.h> #include <drm/drm_probe_helper.h> +#include "rcar_lvds.h" #include "rcar_lvds_regs.h" struct rcar_lvds; @@ -183,8 +184,9 @@ struct pll_info { static void rcar_lvds_d3_e3_pll_calc(struct rcar_lvds *lvds, struct clk *clk, unsigned long target, struct pll_info *pll, - u32 clksel) + u32 clksel, bool dot_clock_only) { + unsigned int div7 = dot_clock_only ? 1 : 7; unsigned long output; unsigned long fin; unsigned int m_min; @@ -218,9 +220,9 @@ static void rcar_lvds_d3_e3_pll_calc(struct rcar_lvds *lvds, struct clk *clk, * `------------> | | * |/ * - * The /7 divider is optional when the LVDS PLL is used to generate a - * dot clock for the DU RGB output, without using the LVDS encoder. We - * don't support this configuration yet. + * The /7 divider is optional, it is enabled when the LVDS PLL is used + * to drive the LVDS encoder, and disabled when used to generate a dot + * clock for the DU RGB output, without using the LVDS encoder. * * The PLL allowed input frequency range is 12 MHz to 192 MHz. */ @@ -280,7 +282,7 @@ static void rcar_lvds_d3_e3_pll_calc(struct rcar_lvds *lvds, struct clk *clk, * the PLL, followed by a an optional fixed /7 * divider. */ - fout = fvco / (1 << e) / 7; + fout = fvco / (1 << e) / div7; div = DIV_ROUND_CLOSEST(fout, target); diff = abs(fout / div - target); @@ -301,7 +303,7 @@ static void rcar_lvds_d3_e3_pll_calc(struct rcar_lvds *lvds, struct clk *clk, done: output = fin * pll->pll_n / pll->pll_m / (1 << pll->pll_e) - / 7 / pll->div; + / div7 / pll->div; error = (long)(output - target) * 10000 / (long)target; dev_dbg(lvds->dev, @@ -311,17 +313,18 @@ done: pll->pll_m, pll->pll_n, pll->pll_e, pll->div); } -static void rcar_lvds_pll_setup_d3_e3(struct rcar_lvds *lvds, unsigned int freq) +static void __rcar_lvds_pll_setup_d3_e3(struct rcar_lvds *lvds, + unsigned int freq, bool dot_clock_only) { struct pll_info pll = { .diff = (unsigned long)-1 }; u32 lvdpllcr; rcar_lvds_d3_e3_pll_calc(lvds, lvds->clocks.dotclkin[0], freq, &pll, - LVDPLLCR_CKSEL_DU_DOTCLKIN(0)); + LVDPLLCR_CKSEL_DU_DOTCLKIN(0), dot_clock_only); rcar_lvds_d3_e3_pll_calc(lvds, lvds->clocks.dotclkin[1], freq, &pll, - LVDPLLCR_CKSEL_DU_DOTCLKIN(1)); + LVDPLLCR_CKSEL_DU_DOTCLKIN(1), dot_clock_only); rcar_lvds_d3_e3_pll_calc(lvds, lvds->clocks.extal, freq, &pll, - LVDPLLCR_CKSEL_EXTAL); + LVDPLLCR_CKSEL_EXTAL, dot_clock_only); lvdpllcr = LVDPLLCR_PLLON | pll.clksel | LVDPLLCR_CLKOUT | LVDPLLCR_PLLN(pll.pll_n - 1) | LVDPLLCR_PLLM(pll.pll_m - 1); @@ -330,6 +333,9 @@ static void rcar_lvds_pll_setup_d3_e3(struct rcar_lvds *lvds, unsigned int freq) lvdpllcr |= LVDPLLCR_STP_CLKOUTE | LVDPLLCR_OUTCLKSEL | LVDPLLCR_PLLE(pll.pll_e - 1); + if (dot_clock_only) + lvdpllcr |= LVDPLLCR_OCKSEL; + rcar_lvds_write(lvds, LVDPLLCR, lvdpllcr); if (pll.div > 1) @@ -343,6 +349,57 @@ static void rcar_lvds_pll_setup_d3_e3(struct rcar_lvds *lvds, unsigned int freq) rcar_lvds_write(lvds, LVDDIV, 0); } +static void rcar_lvds_pll_setup_d3_e3(struct rcar_lvds *lvds, unsigned int freq) +{ + __rcar_lvds_pll_setup_d3_e3(lvds, freq, false); +} + +/* ----------------------------------------------------------------------------- + * Clock - D3/E3 only + */ + +int rcar_lvds_clk_enable(struct drm_bridge *bridge, unsigned long freq) +{ + struct rcar_lvds *lvds = bridge_to_rcar_lvds(bridge); + int ret; + + if (WARN_ON(!(lvds->info->quirks & RCAR_LVDS_QUIRK_EXT_PLL))) + return -ENODEV; + + dev_dbg(lvds->dev, "enabling LVDS PLL, freq=%luHz\n", freq); + + WARN_ON(lvds->enabled); + + ret = clk_prepare_enable(lvds->clocks.mod); + if (ret < 0) + return ret; + + __rcar_lvds_pll_setup_d3_e3(lvds, freq, true); + + lvds->enabled = true; + return 0; +} +EXPORT_SYMBOL_GPL(rcar_lvds_clk_enable); + +void rcar_lvds_clk_disable(struct drm_bridge *bridge) +{ + struct rcar_lvds *lvds = bridge_to_rcar_lvds(bridge); + + if (WARN_ON(!(lvds->info->quirks & RCAR_LVDS_QUIRK_EXT_PLL))) + return; + + dev_dbg(lvds->dev, "disabling LVDS PLL\n"); + + WARN_ON(!lvds->enabled); + + rcar_lvds_write(lvds, LVDPLLCR, 0); + + clk_disable_unprepare(lvds->clocks.mod); + + lvds->enabled = false; +} +EXPORT_SYMBOL_GPL(rcar_lvds_clk_disable); + /* ----------------------------------------------------------------------------- * Bridge */ @@ -545,7 +602,10 @@ static int rcar_lvds_attach(struct drm_bridge *bridge) return drm_bridge_attach(bridge->encoder, lvds->next_bridge, bridge); - /* Otherwise we have a panel, create a connector. */ + /* Otherwise if we have a panel, create a connector. */ + if (!lvds->panel) + return 0; + ret = drm_connector_init(bridge->dev, connector, &rcar_lvds_conn_funcs, DRM_MODE_CONNECTOR_LVDS); if (ret < 0) @@ -593,7 +653,8 @@ static int rcar_lvds_parse_dt(struct rcar_lvds *lvds) local_output = of_graph_get_endpoint_by_regs(lvds->dev->of_node, 1, 0); if (!local_output) { dev_dbg(lvds->dev, "unconnected port@1\n"); - return -ENODEV; + ret = -ENODEV; + goto done; } /* @@ -643,6 +704,15 @@ done: of_node_put(remote_input); of_node_put(remote); + /* + * On D3/E3 the LVDS encoder provides a clock to the DU, which can be + * used for the DPAD output even when the LVDS output is not connected. + * Don't fail probe in that case as the DU will need the bridge to + * control the clock. + */ + if (lvds->info->quirks & RCAR_LVDS_QUIRK_EXT_PLL) + return ret == -ENODEV ? 0 : ret; + return ret; } @@ -786,6 +856,7 @@ static const struct rcar_lvds_device_info rcar_lvds_r8a77995_info = { static const struct of_device_id rcar_lvds_of_table[] = { { .compatible = "renesas,r8a7743-lvds", .data = &rcar_lvds_gen2_info }, + { .compatible = "renesas,r8a7744-lvds", .data = &rcar_lvds_gen2_info }, { .compatible = "renesas,r8a774c0-lvds", .data = &rcar_lvds_r8a77990_info }, { .compatible = "renesas,r8a7790-lvds", .data = &rcar_lvds_r8a7790_info }, { .compatible = "renesas,r8a7791-lvds", .data = &rcar_lvds_gen2_info }, diff --git a/drivers/gpu/drm/rcar-du/rcar_lvds.h b/drivers/gpu/drm/rcar-du/rcar_lvds.h new file mode 100644 index 000000000000..a709cae1bc32 --- /dev/null +++ b/drivers/gpu/drm/rcar-du/rcar_lvds.h @@ -0,0 +1,27 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * rcar_lvds.h -- R-Car LVDS Encoder + * + * Copyright (C) 2013-2018 Renesas Electronics Corporation + * + * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com) + */ + +#ifndef __RCAR_LVDS_H__ +#define __RCAR_LVDS_H__ + +struct drm_bridge; + +#if IS_ENABLED(CONFIG_DRM_RCAR_LVDS) +int rcar_lvds_clk_enable(struct drm_bridge *bridge, unsigned long freq); +void rcar_lvds_clk_disable(struct drm_bridge *bridge); +#else +static inline int rcar_lvds_clk_enable(struct drm_bridge *bridge, + unsigned long freq) +{ + return -ENOSYS; +} +static inline void rcar_lvds_clk_disable(struct drm_bridge *bridge) { } +#endif /* CONFIG_DRM_RCAR_LVDS */ + +#endif /* __RCAR_LVDS_H__ */ diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c index dbb69063b3d5..19fc601c9eeb 100644 --- a/drivers/gpu/drm/scheduler/sched_main.c +++ b/drivers/gpu/drm/scheduler/sched_main.c @@ -60,8 +60,6 @@ static void drm_sched_process_job(struct dma_fence *f, struct dma_fence_cb *cb); -static void drm_sched_expel_job_unlocked(struct drm_sched_job *s_job); - /** * drm_sched_rq_init - initialize a given run queue struct * @@ -286,8 +284,6 @@ static void drm_sched_job_finish(struct work_struct *work) cancel_delayed_work_sync(&sched->work_tdr); spin_lock_irqsave(&sched->job_list_lock, flags); - /* remove job from ring_mirror_list */ - list_del_init(&s_job->node); /* queue TDR for next job */ drm_sched_start_timeout(sched); spin_unlock_irqrestore(&sched->job_list_lock, flags); @@ -295,22 +291,11 @@ static void drm_sched_job_finish(struct work_struct *work) sched->ops->free_job(s_job); } -static void drm_sched_job_finish_cb(struct dma_fence *f, - struct dma_fence_cb *cb) -{ - struct drm_sched_job *job = container_of(cb, struct drm_sched_job, - finish_cb); - schedule_work(&job->finish_work); -} - static void drm_sched_job_begin(struct drm_sched_job *s_job) { struct drm_gpu_scheduler *sched = s_job->sched; unsigned long flags; - dma_fence_add_callback(&s_job->s_fence->finished, &s_job->finish_cb, - drm_sched_job_finish_cb); - spin_lock_irqsave(&sched->job_list_lock, flags); list_add_tail(&s_job->node, &sched->ring_mirror_list); drm_sched_start_timeout(sched); @@ -335,6 +320,51 @@ static void drm_sched_job_timedout(struct work_struct *work) spin_unlock_irqrestore(&sched->job_list_lock, flags); } + /** + * drm_sched_increase_karma - Update sched_entity guilty flag + * + * @bad: The job guilty of time out + * + * Increment on every hang caused by the 'bad' job. If this exceeds the hang + * limit of the scheduler then the respective sched entity is marked guilty and + * jobs from it will not be scheduled further + */ +void drm_sched_increase_karma(struct drm_sched_job *bad) +{ + int i; + struct drm_sched_entity *tmp; + struct drm_sched_entity *entity; + struct drm_gpu_scheduler *sched = bad->sched; + + /* don't increase @bad's karma if it's from KERNEL RQ, + * because sometimes GPU hang would cause kernel jobs (like VM updating jobs) + * corrupt but keep in mind that kernel jobs always considered good. + */ + if (bad->s_priority != DRM_SCHED_PRIORITY_KERNEL) { + atomic_inc(&bad->karma); + for (i = DRM_SCHED_PRIORITY_MIN; i < DRM_SCHED_PRIORITY_KERNEL; + i++) { + struct drm_sched_rq *rq = &sched->sched_rq[i]; + + spin_lock(&rq->lock); + list_for_each_entry_safe(entity, tmp, &rq->entities, list) { + if (bad->s_fence->scheduled.context == + entity->fence_context) { + if (atomic_read(&bad->karma) > + bad->sched->hang_limit) + if (entity->guilty) + atomic_set(entity->guilty, 1); + break; + } + } + spin_unlock(&rq->lock); + if (&entity->list != &rq->entities) + break; + } + } +} +EXPORT_SYMBOL(drm_sched_increase_karma); + /** * drm_sched_hw_job_reset - stop the scheduler if it contains the bad job * @@ -342,50 +372,42 @@ static void drm_sched_job_timedout(struct work_struct *work) * @bad: bad scheduler job * */ -void drm_sched_hw_job_reset(struct drm_gpu_scheduler *sched, struct drm_sched_job *bad) +void drm_sched_stop(struct drm_gpu_scheduler *sched) { struct drm_sched_job *s_job; - struct drm_sched_entity *entity, *tmp; unsigned long flags; - int i; + struct dma_fence *last_fence = NULL; + kthread_park(sched->thread); + + /* + * Verify all the signaled jobs in mirror list are removed from the ring + * by waiting for the latest job to enter the list. This should insure that + * also all the previous jobs that were in flight also already singaled + * and removed from the list. + */ spin_lock_irqsave(&sched->job_list_lock, flags); list_for_each_entry_reverse(s_job, &sched->ring_mirror_list, node) { if (s_job->s_fence->parent && dma_fence_remove_callback(s_job->s_fence->parent, - &s_job->s_fence->cb)) { + &s_job->cb)) { dma_fence_put(s_job->s_fence->parent); s_job->s_fence->parent = NULL; atomic_dec(&sched->hw_rq_count); + } else { + last_fence = dma_fence_get(&s_job->s_fence->finished); + break; } } spin_unlock_irqrestore(&sched->job_list_lock, flags); - if (bad && bad->s_priority != DRM_SCHED_PRIORITY_KERNEL) { - atomic_inc(&bad->karma); - /* don't increase @bad's karma if it's from KERNEL RQ, - * becuase sometimes GPU hang would cause kernel jobs (like VM updating jobs) - * corrupt but keep in mind that kernel jobs always considered good. - */ - for (i = DRM_SCHED_PRIORITY_MIN; i < DRM_SCHED_PRIORITY_KERNEL; i++ ) { - struct drm_sched_rq *rq = &sched->sched_rq[i]; - - spin_lock(&rq->lock); - list_for_each_entry_safe(entity, tmp, &rq->entities, list) { - if (bad->s_fence->scheduled.context == entity->fence_context) { - if (atomic_read(&bad->karma) > bad->sched->hang_limit) - if (entity->guilty) - atomic_set(entity->guilty, 1); - break; - } - } - spin_unlock(&rq->lock); - if (&entity->list != &rq->entities) - break; - } + if (last_fence) { + dma_fence_wait(last_fence, false); + dma_fence_put(last_fence); } } -EXPORT_SYMBOL(drm_sched_hw_job_reset); + +EXPORT_SYMBOL(drm_sched_stop); /** * drm_sched_job_recovery - recover jobs after a reset @@ -393,18 +415,58 @@ EXPORT_SYMBOL(drm_sched_hw_job_reset); * @sched: scheduler instance * */ -void drm_sched_job_recovery(struct drm_gpu_scheduler *sched) +void drm_sched_start(struct drm_gpu_scheduler *sched, bool full_recovery) { struct drm_sched_job *s_job, *tmp; - bool found_guilty = false; - unsigned long flags; int r; - spin_lock_irqsave(&sched->job_list_lock, flags); + if (!full_recovery) + goto unpark; + + /* + * Locking the list is not required here as the sched thread is parked + * so no new jobs are being pushed in to HW and in drm_sched_stop we + * flushed all the jobs who were still in mirror list but who already + * signaled and removed them self from the list. Also concurrent + * GPU recovers can't run in parallel. + */ + list_for_each_entry_safe(s_job, tmp, &sched->ring_mirror_list, node) { + struct dma_fence *fence = s_job->s_fence->parent; + + if (fence) { + r = dma_fence_add_callback(fence, &s_job->cb, + drm_sched_process_job); + if (r == -ENOENT) + drm_sched_process_job(fence, &s_job->cb); + else if (r) + DRM_ERROR("fence add callback failed (%d)\n", + r); + } else + drm_sched_process_job(NULL, &s_job->cb); + } + + drm_sched_start_timeout(sched); + +unpark: + kthread_unpark(sched->thread); +} +EXPORT_SYMBOL(drm_sched_start); + +/** + * drm_sched_resubmit_jobs - helper to relunch job from mirror ring list + * + * @sched: scheduler instance + * + */ +void drm_sched_resubmit_jobs(struct drm_gpu_scheduler *sched) +{ + struct drm_sched_job *s_job, *tmp; + uint64_t guilty_context; + bool found_guilty = false; + + /*TODO DO we need spinlock here ? */ list_for_each_entry_safe(s_job, tmp, &sched->ring_mirror_list, node) { struct drm_sched_fence *s_fence = s_job->s_fence; - struct dma_fence *fence; - uint64_t guilty_context; if (!found_guilty && atomic_read(&s_job->karma) > sched->hang_limit) { found_guilty = true; @@ -414,31 +476,11 @@ void drm_sched_job_recovery(struct drm_gpu_scheduler *sched) if (found_guilty && s_job->s_fence->scheduled.context == guilty_context) dma_fence_set_error(&s_fence->finished, -ECANCELED); - spin_unlock_irqrestore(&sched->job_list_lock, flags); - fence = sched->ops->run_job(s_job); + s_job->s_fence->parent = sched->ops->run_job(s_job); atomic_inc(&sched->hw_rq_count); - - if (fence) { - s_fence->parent = dma_fence_get(fence); - r = dma_fence_add_callback(fence, &s_fence->cb, - drm_sched_process_job); - if (r == -ENOENT) - drm_sched_process_job(fence, &s_fence->cb); - else if (r) - DRM_ERROR("fence add callback failed (%d)\n", - r); - dma_fence_put(fence); - } else { - if (s_fence->finished.error < 0) - drm_sched_expel_job_unlocked(s_job); - drm_sched_process_job(NULL, &s_fence->cb); - } - spin_lock_irqsave(&sched->job_list_lock, flags); } - drm_sched_start_timeout(sched); - spin_unlock_irqrestore(&sched->job_list_lock, flags); } -EXPORT_SYMBOL(drm_sched_job_recovery); +EXPORT_SYMBOL(drm_sched_resubmit_jobs); /** * drm_sched_job_init - init a scheduler job @@ -552,18 +594,27 @@ drm_sched_select_entity(struct drm_gpu_scheduler *sched) */ static void drm_sched_process_job(struct dma_fence *f, struct dma_fence_cb *cb) { - struct drm_sched_fence *s_fence = - container_of(cb, struct drm_sched_fence, cb); + struct drm_sched_job *s_job = container_of(cb, struct drm_sched_job, cb); + struct drm_sched_fence *s_fence = s_job->s_fence; struct drm_gpu_scheduler *sched = s_fence->sched; + unsigned long flags; + + cancel_delayed_work(&sched->work_tdr); - dma_fence_get(&s_fence->finished); atomic_dec(&sched->hw_rq_count); atomic_dec(&sched->num_jobs); + + spin_lock_irqsave(&sched->job_list_lock, flags); + /* remove job from ring_mirror_list */ + list_del_init(&s_job->node); + spin_unlock_irqrestore(&sched->job_list_lock, flags); + drm_sched_fence_finished(s_fence); trace_drm_sched_process_job(s_fence); - dma_fence_put(&s_fence->finished); wake_up_interruptible(&sched->wake_up_worker); + + schedule_work(&s_job->finish_work); } /** @@ -626,34 +677,22 @@ static int drm_sched_main(void *param) if (fence) { s_fence->parent = dma_fence_get(fence); - r = dma_fence_add_callback(fence, &s_fence->cb, + r = dma_fence_add_callback(fence, &sched_job->cb, drm_sched_process_job); if (r == -ENOENT) - drm_sched_process_job(fence, &s_fence->cb); + drm_sched_process_job(fence, &sched_job->cb); else if (r) DRM_ERROR("fence add callback failed (%d)\n", r); dma_fence_put(fence); - } else { - if (s_fence->finished.error < 0) - drm_sched_expel_job_unlocked(sched_job); - drm_sched_process_job(NULL, &s_fence->cb); - } + } else + drm_sched_process_job(NULL, &sched_job->cb); wake_up(&sched->job_scheduled); } return 0; } -static void drm_sched_expel_job_unlocked(struct drm_sched_job *s_job) -{ - struct drm_gpu_scheduler *sched = s_job->sched; - - spin_lock(&sched->job_list_lock); - list_del_init(&s_job->node); - spin_unlock(&sched->job_list_lock); -} - /** * drm_sched_init - Init a gpu scheduler instance * diff --git a/drivers/gpu/drm/tegra/Makefile b/drivers/gpu/drm/tegra/Makefile index 2e0d6213f6bc..33c463e8d49f 100644 --- a/drivers/gpu/drm/tegra/Makefile +++ b/drivers/gpu/drm/tegra/Makefile @@ -10,6 +10,7 @@ tegra-drm-y := \ dc.o \ output.o \ rgb.o \ + hda.o \ hdmi.o \ mipi-phy.o \ dsi.o \ diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 4b70ce664c41..0c5f1e6a0446 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -92,10 +92,6 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags) return -ENOMEM; if (iommu_present(&platform_bus_type)) { - u64 carveout_start, carveout_end, gem_start, gem_end; - struct iommu_domain_geometry *geometry; - unsigned long order; - tegra->domain = iommu_domain_alloc(&platform_bus_type); if (!tegra->domain) { err = -ENOMEM; @@ -105,27 +101,6 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags) err = iova_cache_get(); if (err < 0) goto domain; - - geometry = &tegra->domain->geometry; - gem_start = geometry->aperture_start; - gem_end = geometry->aperture_end - CARVEOUT_SZ; - carveout_start = gem_end + 1; - carveout_end = geometry->aperture_end; - - order = __ffs(tegra->domain->pgsize_bitmap); - init_iova_domain(&tegra->carveout.domain, 1UL << order, - carveout_start >> order); - - tegra->carveout.shift = iova_shift(&tegra->carveout.domain); - tegra->carveout.limit = carveout_end >> tegra->carveout.shift; - - drm_mm_init(&tegra->mm, gem_start, gem_end - gem_start + 1); - mutex_init(&tegra->mm_lock); - - DRM_DEBUG("IOMMU apertures:\n"); - DRM_DEBUG(" GEM: %#llx-%#llx\n", gem_start, gem_end); - DRM_DEBUG(" Carveout: %#llx-%#llx\n", carveout_start, - carveout_end); } mutex_init(&tegra->clients_lock); @@ -159,6 +134,36 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags) if (err < 0) goto fbdev; + if (tegra->domain) { + u64 carveout_start, carveout_end, gem_start, gem_end; + u64 dma_mask = dma_get_mask(&device->dev); + dma_addr_t start, end; + unsigned long order; + + start = tegra->domain->geometry.aperture_start & dma_mask; + end = tegra->domain->geometry.aperture_end & dma_mask; + + gem_start = start; + gem_end = end - CARVEOUT_SZ; + carveout_start = gem_end + 1; + carveout_end = end; + + order = __ffs(tegra->domain->pgsize_bitmap); + init_iova_domain(&tegra->carveout.domain, 1UL << order, + carveout_start >> order); + + tegra->carveout.shift = iova_shift(&tegra->carveout.domain); + tegra->carveout.limit = carveout_end >> tegra->carveout.shift; + + drm_mm_init(&tegra->mm, gem_start, gem_end - gem_start + 1); + mutex_init(&tegra->mm_lock); + + DRM_DEBUG("IOMMU apertures:\n"); + DRM_DEBUG(" GEM: %#llx-%#llx\n", gem_start, gem_end); + DRM_DEBUG(" Carveout: %#llx-%#llx\n", carveout_start, + carveout_end); + } + if (tegra->hub) { err = tegra_display_hub_prepare(tegra->hub); if (err < 0) @@ -1041,6 +1046,7 @@ int tegra_drm_register_client(struct tegra_drm *tegra, { mutex_lock(&tegra->clients_lock); list_add_tail(&client->list, &tegra->clients); + client->drm = tegra; mutex_unlock(&tegra->clients_lock); return 0; @@ -1051,6 +1057,7 @@ int tegra_drm_unregister_client(struct tegra_drm *tegra, { mutex_lock(&tegra->clients_lock); list_del_init(&client->list); + client->drm = NULL; mutex_unlock(&tegra->clients_lock); return 0; diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h index e2dee5c9c03b..70154c253d45 100644 --- a/drivers/gpu/drm/tegra/drm.h +++ b/drivers/gpu/drm/tegra/drm.h @@ -88,6 +88,7 @@ int tegra_drm_submit(struct tegra_drm_context *context, struct tegra_drm_client { struct host1x_client base; struct list_head list; + struct tegra_drm *drm; unsigned int version; const struct tegra_drm_client_ops *ops; @@ -124,7 +125,7 @@ struct tegra_output { struct drm_panel *panel; struct i2c_adapter *ddc; const struct edid *edid; - struct cec_notifier *notifier; + struct cec_notifier *cec; unsigned int hpd_irq; int hpd_gpio; enum of_gpio_flags hpd_gpio_flags; diff --git a/drivers/gpu/drm/tegra/hda.c b/drivers/gpu/drm/tegra/hda.c new file mode 100644 index 000000000000..94245a18a043 --- /dev/null +++ b/drivers/gpu/drm/tegra/hda.c @@ -0,0 +1,63 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright (C) 2019 NVIDIA Corporation + */ + +#include <linux/bug.h> + +#include <sound/hda_verbs.h> + +#include "hda.h" + +void tegra_hda_parse_format(unsigned int format, struct tegra_hda_format *fmt) +{ + unsigned int mul, div, bits, channels; + + if (format & AC_FMT_TYPE_NON_PCM) + fmt->pcm = false; + else + fmt->pcm = true; + + if (format & AC_FMT_BASE_44K) + fmt->sample_rate = 44100; + else + fmt->sample_rate = 48000; + + mul = (format & AC_FMT_MULT_MASK) >> AC_FMT_MULT_SHIFT; + div = (format & AC_FMT_DIV_MASK) >> AC_FMT_DIV_SHIFT; + + fmt->sample_rate *= (mul + 1) / (div + 1); + + switch (format & AC_FMT_BITS_MASK) { + case AC_FMT_BITS_8: + fmt->bits = 8; + break; + + case AC_FMT_BITS_16: + fmt->bits = 16; + break; + + case AC_FMT_BITS_20: + fmt->bits = 20; + break; + + case AC_FMT_BITS_24: + fmt->bits = 24; + break; + + case AC_FMT_BITS_32: + fmt->bits = 32; + break; + + default: + bits = (format & AC_FMT_BITS_MASK) >> AC_FMT_BITS_SHIFT; + WARN(1, "invalid number of bits: %#x\n", bits); + fmt->bits = 8; + break; + } + + channels = (format & AC_FMT_CHAN_MASK) >> AC_FMT_CHAN_SHIFT; + + /* channels are encoded as n - 1 */ + fmt->channels = channels + 1; +} diff --git a/drivers/gpu/drm/tegra/hda.h b/drivers/gpu/drm/tegra/hda.h new file mode 100644 index 000000000000..77269955a4f2 --- /dev/null +++ b/drivers/gpu/drm/tegra/hda.h @@ -0,0 +1,20 @@ +// SPDX-License-Identifier: MIT +/* + * Copyright (C) 2019 NVIDIA Corporation + */ + +#ifndef DRM_TEGRA_HDA_H +#define DRM_TEGRA_HDA_H 1 + +#include <linux/types.h> + +struct tegra_hda_format { + unsigned int sample_rate; + unsigned int channels; + unsigned int bits; + bool pcm; +}; + +void tegra_hda_parse_format(unsigned int format, struct tegra_hda_format *fmt); + +#endif diff --git a/drivers/gpu/drm/tegra/hdmi.c b/drivers/gpu/drm/tegra/hdmi.c index 78bc7c906e22..47c55974756d 100644 --- a/drivers/gpu/drm/tegra/hdmi.c +++ b/drivers/gpu/drm/tegra/hdmi.c @@ -11,6 +11,7 @@ #include <linux/debugfs.h> #include <linux/gpio.h> #include <linux/hdmi.h> +#include <linux/math64.h> #include <linux/of_device.h> #include <linux/pm_runtime.h> #include <linux/regulator/consumer.h> @@ -20,10 +21,7 @@ #include <drm/drm_crtc.h> #include <drm/drm_probe_helper.h> -#include <sound/hda_verbs.h> - -#include <media/cec-notifier.h> - +#include "hda.h" #include "hdmi.h" #include "drm.h" #include "dc.h" @@ -71,8 +69,7 @@ struct tegra_hdmi { const struct tegra_hdmi_config *config; unsigned int audio_source; - unsigned int audio_sample_rate; - unsigned int audio_channels; + struct tegra_hda_format format; unsigned int pixel_clock; bool stereo; @@ -119,68 +116,11 @@ static inline void tegra_hdmi_writel(struct tegra_hdmi *hdmi, u32 value, } struct tegra_hdmi_audio_config { - unsigned int pclk; unsigned int n; unsigned int cts; unsigned int aval; }; -static const struct tegra_hdmi_audio_config tegra_hdmi_audio_32k[] = { - { 25200000, 4096, 25200, 24000 }, - { 27000000, 4096, 27000, 24000 }, - { 74250000, 4096, 74250, 24000 }, - { 148500000, 4096, 148500, 24000 }, - { 0, 0, 0, 0 }, -}; - -static const struct tegra_hdmi_audio_config tegra_hdmi_audio_44_1k[] = { - { 25200000, 5880, 26250, 25000 }, - { 27000000, 5880, 28125, 25000 }, - { 74250000, 4704, 61875, 20000 }, - { 148500000, 4704, 123750, 20000 }, - { 0, 0, 0, 0 }, -}; - -static const struct tegra_hdmi_audio_config tegra_hdmi_audio_48k[] = { - { 25200000, 6144, 25200, 24000 }, - { 27000000, 6144, 27000, 24000 }, - { 74250000, 6144, 74250, 24000 }, - { 148500000, 6144, 148500, 24000 }, - { 0, 0, 0, 0 }, -}; - -static const struct tegra_hdmi_audio_config tegra_hdmi_audio_88_2k[] = { - { 25200000, 11760, 26250, 25000 }, - { 27000000, 11760, 28125, 25000 }, - { 74250000, 9408, 61875, 20000 }, - { 148500000, 9408, 123750, 20000 }, - { 0, 0, 0, 0 }, -}; - -static const struct tegra_hdmi_audio_config tegra_hdmi_audio_96k[] = { - { 25200000, 12288, 25200, 24000 }, - { 27000000, 12288, 27000, 24000 }, - { 74250000, 12288, 74250, 24000 }, - { 148500000, 12288, 148500, 24000 }, - { 0, 0, 0, 0 }, -}; - -static const struct tegra_hdmi_audio_config tegra_hdmi_audio_176_4k[] = { - { 25200000, 23520, 26250, 25000 }, - { 27000000, 23520, 28125, 25000 }, - { 74250000, 18816, 61875, 20000 }, - { 148500000, 18816, 123750, 20000 }, - { 0, 0, 0, 0 }, -}; - -static const struct tegra_hdmi_audio_config tegra_hdmi_audio_192k[] = { - { 25200000, 24576, 25200, 24000 }, - { 27000000, 24576, 27000, 24000 }, - { 74250000, 24576, 74250, 24000 }, - { 148500000, 24576, 148500, 24000 }, - { 0, 0, 0, 0 }, -}; - static const struct tmds_config tegra20_tmds_config[] = { { /* slow pixel clock modes */ .pclk = 27000000, @@ -418,52 +358,53 @@ static const struct tmds_config tegra124_tmds_config[] = { }, }; -static const struct tegra_hdmi_audio_config * -tegra_hdmi_get_audio_config(unsigned int sample_rate, unsigned int pclk) +static int +tegra_hdmi_get_audio_config(unsigned int audio_freq, unsigned int pix_clock, + struct tegra_hdmi_audio_config *config) { - const struct tegra_hdmi_audio_config *table; - - switch (sample_rate) { - case 32000: - table = tegra_hdmi_audio_32k; - break; - - case 44100: - table = tegra_hdmi_audio_44_1k; - break; - - case 48000: - table = tegra_hdmi_audio_48k; - break; - - case 88200: - table = tegra_hdmi_audio_88_2k; - break; - - case 96000: - table = tegra_hdmi_audio_96k; - break; - - case 176400: - table = tegra_hdmi_audio_176_4k; - break; - - case 192000: - table = tegra_hdmi_audio_192k; - break; - - default: - return NULL; - } - - while (table->pclk) { - if (table->pclk == pclk) - return table; - - table++; + const unsigned int afreq = 128 * audio_freq; + const unsigned int min_n = afreq / 1500; + const unsigned int max_n = afreq / 300; + const unsigned int ideal_n = afreq / 1000; + int64_t min_err = (uint64_t)-1 >> 1; + unsigned int min_delta = -1; + int n; + + memset(config, 0, sizeof(*config)); + config->n = -1; + + for (n = min_n; n <= max_n; n++) { + uint64_t cts_f, aval_f; + unsigned int delta; + int64_t cts, err; + + /* compute aval in 48.16 fixed point */ + aval_f = ((int64_t)24000000 << 16) * n; + do_div(aval_f, afreq); + /* It should round without any rest */ + if (aval_f & 0xFFFF) + continue; + + /* Compute cts in 48.16 fixed point */ + cts_f = ((int64_t)pix_clock << 16) * n; + do_div(cts_f, afreq); + /* Round it to the nearest integer */ + cts = (cts_f & ~0xFFFF) + ((cts_f & BIT(15)) << 1); + + delta = abs(n - ideal_n); + + /* Compute the absolute error */ + err = abs((int64_t)cts_f - cts); + if (err < min_err || (err == min_err && delta < min_delta)) { + config->n = n; + config->cts = cts >> 16; + config->aval = aval_f >> 16; + min_delta = delta; + min_err = err; + } } - return NULL; + return config->n != -1 ? 0 : -EINVAL; } static void tegra_hdmi_setup_audio_fs_tables(struct tegra_hdmi *hdmi) @@ -510,7 +451,7 @@ static void tegra_hdmi_write_aval(struct tegra_hdmi *hdmi, u32 value) unsigned int i; for (i = 0; i < ARRAY_SIZE(regs); i++) { - if (regs[i].sample_rate == hdmi->audio_sample_rate) { + if (regs[i].sample_rate == hdmi->format.sample_rate) { tegra_hdmi_writel(hdmi, value, regs[i].offset); break; } @@ -519,8 +460,9 @@ static void tegra_hdmi_write_aval(struct tegra_hdmi *hdmi, u32 value) static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi) { - const struct tegra_hdmi_audio_config *config; + struct tegra_hdmi_audio_config config; u32 source, value; + int err; switch (hdmi->audio_source) { case HDA: @@ -564,7 +506,7 @@ static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi) * play back system startup sounds early. It is possibly not * needed on Linux at all. */ - if (hdmi->audio_channels == 2) + if (hdmi->format.channels == 2) value = SOR_AUDIO_CNTRL0_INJECT_NULLSMPL; else value = 0; @@ -595,25 +537,28 @@ static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi) tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_AUDIO_SPARE0); } - config = tegra_hdmi_get_audio_config(hdmi->audio_sample_rate, - hdmi->pixel_clock); - if (!config) { + err = tegra_hdmi_get_audio_config(hdmi->format.sample_rate, + hdmi->pixel_clock, &config); + if (err < 0) { dev_err(hdmi->dev, "cannot set audio to %u Hz at %u Hz pixel clock\n", - hdmi->audio_sample_rate, hdmi->pixel_clock); - return -EINVAL; + hdmi->format.sample_rate, hdmi->pixel_clock); + return err; } + dev_dbg(hdmi->dev, "audio: pixclk=%u, n=%u, cts=%u, aval=%u\n", + hdmi->pixel_clock, config.n, config.cts, config.aval); + tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_HDMI_ACR_CTRL); value = AUDIO_N_RESETF | AUDIO_N_GENERATE_ALTERNATE | - AUDIO_N_VALUE(config->n - 1); + AUDIO_N_VALUE(config.n - 1); tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_N); - tegra_hdmi_writel(hdmi, ACR_SUBPACK_N(config->n) | ACR_ENABLE, + tegra_hdmi_writel(hdmi, ACR_SUBPACK_N(config.n) | ACR_ENABLE, HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_HIGH); - tegra_hdmi_writel(hdmi, ACR_SUBPACK_CTS(config->cts), + tegra_hdmi_writel(hdmi, ACR_SUBPACK_CTS(config.cts), HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_LOW); value = SPARE_HW_CTS | SPARE_FORCE_SW_CTS | SPARE_CTS_RESET_VAL(1); @@ -624,7 +569,7 @@ static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi) tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_N); if (hdmi->config->has_hda) - tegra_hdmi_write_aval(hdmi, config->aval); + tegra_hdmi_write_aval(hdmi, config.aval); tegra_hdmi_setup_audio_fs_tables(hdmi); @@ -788,7 +733,7 @@ static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi) return; } - frame.channels = hdmi->audio_channels; + frame.channels = hdmi->format.channels; err = hdmi_audio_infoframe_pack(&frame, buffer, sizeof(buffer)); if (err < 0) { @@ -1590,24 +1535,6 @@ static const struct of_device_id tegra_hdmi_of_match[] = { }; MODULE_DEVICE_TABLE(of, tegra_hdmi_of_match); -static void hda_format_parse(unsigned int format, unsigned int *rate, - unsigned int *channels) -{ - unsigned int mul, div; - - if (format & AC_FMT_BASE_44K) - *rate = 44100; - else - *rate = 48000; - - mul = (format & AC_FMT_MULT_MASK) >> AC_FMT_MULT_SHIFT; - div = (format & AC_FMT_DIV_MASK) >> AC_FMT_DIV_SHIFT; - - *rate = *rate * (mul + 1) / (div + 1); - - *channels = (format & AC_FMT_CHAN_MASK) >> AC_FMT_CHAN_SHIFT; -} - static irqreturn_t tegra_hdmi_irq(int irq, void *data) { struct tegra_hdmi *hdmi = data; @@ -1624,14 +1551,9 @@ static irqreturn_t tegra_hdmi_irq(int irq, void *data) value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH0); if (value & SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID) { - unsigned int sample_rate, channels; - format = value & SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK; - hda_format_parse(format, &sample_rate, &channels); - - hdmi->audio_sample_rate = sample_rate; - hdmi->audio_channels = channels; + tegra_hda_parse_format(format, &hdmi->format); err = tegra_hdmi_setup_audio(hdmi); if (err < 0) { @@ -1665,8 +1587,6 @@ static int tegra_hdmi_probe(struct platform_device *pdev) hdmi->dev = &pdev->dev; hdmi->audio_source = AUTO; - hdmi->audio_sample_rate = 48000; - hdmi->audio_channels = 2; hdmi->stereo = false; hdmi->dvi = false; @@ -1710,10 +1630,6 @@ static int tegra_hdmi_probe(struct platform_device *pdev) return PTR_ERR(hdmi->vdd); } - hdmi->output.notifier = cec_notifier_get(&pdev->dev); - if (hdmi->output.notifier == NULL) - return -ENOMEM; - hdmi->output.dev = &pdev->dev; err = tegra_output_probe(&hdmi->output); @@ -1772,9 +1688,6 @@ static int tegra_hdmi_remove(struct platform_device *pdev) tegra_output_remove(&hdmi->output); - if (hdmi->output.notifier) - cec_notifier_put(hdmi->output.notifier); - return 0; } diff --git a/drivers/gpu/drm/tegra/output.c b/drivers/gpu/drm/tegra/output.c index c662efc7e413..9c2b9dad55c3 100644 --- a/drivers/gpu/drm/tegra/output.c +++ b/drivers/gpu/drm/tegra/output.c @@ -36,7 +36,7 @@ int tegra_output_connector_get_modes(struct drm_connector *connector) else if (output->ddc) edid = drm_get_edid(connector, output->ddc); - cec_notifier_set_phys_addr_from_edid(output->notifier, edid); + cec_notifier_set_phys_addr_from_edid(output->cec, edid); drm_connector_update_edid_property(connector, edid); if (edid) { @@ -73,7 +73,7 @@ tegra_output_connector_detect(struct drm_connector *connector, bool force) } if (status != connector_status_connected) - cec_notifier_phys_addr_invalidate(output->notifier); + cec_notifier_phys_addr_invalidate(output->cec); return status; } @@ -174,11 +174,18 @@ int tegra_output_probe(struct tegra_output *output) disable_irq(output->hpd_irq); } + output->cec = cec_notifier_get(output->dev); + if (!output->cec) + return -ENOMEM; + return 0; } void tegra_output_remove(struct tegra_output *output) { + if (output->cec) + cec_notifier_put(output->cec); + if (gpio_is_valid(output->hpd_gpio)) { free_irq(output->hpd_irq, output); gpio_free(output->hpd_gpio); diff --git a/drivers/gpu/drm/tegra/sor.c b/drivers/gpu/drm/tegra/sor.c index 44feac2a0359..40057106f5f3 100644 --- a/drivers/gpu/drm/tegra/sor.c +++ b/drivers/gpu/drm/tegra/sor.c @@ -19,8 +19,6 @@ #include <soc/tegra/pmc.h> -#include <sound/hda_verbs.h> - #include <drm/drm_atomic_helper.h> #include <drm/drm_dp_helper.h> #include <drm/drm_panel.h> @@ -28,6 +26,7 @@ #include "dc.h" #include "drm.h" +#include "hda.h" #include "sor.h" #include "trace.h" @@ -411,6 +410,8 @@ struct tegra_sor { struct clk *clk_dp; struct clk *clk; + u8 xbar_cfg[5]; + struct drm_dp_aux *aux; struct drm_info_list *debugfs_files; @@ -429,10 +430,7 @@ struct tegra_sor { struct delayed_work scdc; bool scdc_enabled; - struct { - unsigned int sample_rate; - unsigned int channels; - } audio; + struct tegra_hda_format format; }; struct tegra_sor_state { @@ -1818,7 +1816,7 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder) /* XXX not in TRM */ for (value = 0, i = 0; i < 5; i++) - value |= SOR_XBAR_CTRL_LINK0_XSEL(i, sor->soc->xbar_cfg[i]) | + value |= SOR_XBAR_CTRL_LINK0_XSEL(i, sor->xbar_cfg[i]) | SOR_XBAR_CTRL_LINK1_XSEL(i, i); tegra_sor_writel(sor, 0x00000000, SOR_XBAR_POL); @@ -2186,7 +2184,7 @@ static int tegra_sor_hdmi_enable_audio_infoframe(struct tegra_sor *sor) return err; } - frame.channels = sor->audio.channels; + frame.channels = sor->format.channels; err = hdmi_audio_infoframe_pack(&frame, buffer, sizeof(buffer)); if (err < 0) { @@ -2215,7 +2213,7 @@ static void tegra_sor_hdmi_audio_enable(struct tegra_sor *sor) value |= SOR_AUDIO_CNTRL_SOURCE_SELECT(SOURCE_SELECT_HDA); /* inject null samples */ - if (sor->audio.channels != 2) + if (sor->format.channels != 2) value &= ~SOR_AUDIO_CNTRL_INJECT_NULLSMPL; else value |= SOR_AUDIO_CNTRL_INJECT_NULLSMPL; @@ -2246,7 +2244,7 @@ static void tegra_sor_hdmi_audio_enable(struct tegra_sor *sor) value = SOR_HDMI_AUDIO_N_RESET | SOR_HDMI_AUDIO_N_LOOKUP; tegra_sor_writel(sor, value, SOR_HDMI_AUDIO_N); - value = (24000 * 4096) / (128 * sor->audio.sample_rate / 1000); + value = (24000 * 4096) / (128 * sor->format.sample_rate / 1000); tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_0320); tegra_sor_writel(sor, 4096, SOR_AUDIO_NVAL_0320); @@ -2259,15 +2257,15 @@ static void tegra_sor_hdmi_audio_enable(struct tegra_sor *sor) tegra_sor_writel(sor, 20000, SOR_AUDIO_AVAL_1764); tegra_sor_writel(sor, 18816, SOR_AUDIO_NVAL_1764); - value = (24000 * 6144) / (128 * sor->audio.sample_rate / 1000); + value = (24000 * 6144) / (128 * sor->format.sample_rate / 1000); tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_0480); tegra_sor_writel(sor, 6144, SOR_AUDIO_NVAL_0480); - value = (24000 * 12288) / (128 * sor->audio.sample_rate / 1000); + value = (24000 * 12288) / (128 * sor->format.sample_rate / 1000); tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_0960); tegra_sor_writel(sor, 12288, SOR_AUDIO_NVAL_0960); - value = (24000 * 24576) / (128 * sor->audio.sample_rate / 1000); + value = (24000 * 24576) / (128 * sor->format.sample_rate / 1000); tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_1920); tegra_sor_writel(sor, 24576, SOR_AUDIO_NVAL_1920); @@ -2555,7 +2553,7 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder) /* XXX not in TRM */ for (value = 0, i = 0; i < 5; i++) - value |= SOR_XBAR_CTRL_LINK0_XSEL(i, sor->soc->xbar_cfg[i]) | + value |= SOR_XBAR_CTRL_LINK0_XSEL(i, sor->xbar_cfg[i]) | SOR_XBAR_CTRL_LINK1_XSEL(i, i); tegra_sor_writel(sor, 0x00000000, SOR_XBAR_POL); @@ -3176,6 +3174,8 @@ MODULE_DEVICE_TABLE(of, tegra_sor_of_match); static int tegra_sor_parse_dt(struct tegra_sor *sor) { struct device_node *np = sor->dev->of_node; + u32 xbar_cfg[5]; + unsigned int i; u32 value; int err; @@ -3193,25 +3193,18 @@ static int tegra_sor_parse_dt(struct tegra_sor *sor) sor->pad = TEGRA_IO_PAD_HDMI_DP0 + sor->index; } - return 0; -} - -static void tegra_hda_parse_format(unsigned int format, unsigned int *rate, - unsigned int *channels) -{ - unsigned int mul, div; - - if (format & AC_FMT_BASE_44K) - *rate = 44100; - else - *rate = 48000; - - mul = (format & AC_FMT_MULT_MASK) >> AC_FMT_MULT_SHIFT; - div = (format & AC_FMT_DIV_MASK) >> AC_FMT_DIV_SHIFT; - - *rate = *rate * (mul + 1) / (div + 1); + err = of_property_read_u32_array(np, "nvidia,xbar-cfg", xbar_cfg, 5); + if (err < 0) { + /* fall back to default per-SoC XBAR configuration */ + for (i = 0; i < 5; i++) + sor->xbar_cfg[i] = sor->soc->xbar_cfg[i]; + } else { + /* copy cells to SOR XBAR configuration */ + for (i = 0; i < 5; i++) + sor->xbar_cfg[i] = xbar_cfg[i]; + } - *channels = (format & AC_FMT_CHAN_MASK) >> AC_FMT_CHAN_SHIFT; + return 0; } static irqreturn_t tegra_sor_irq(int irq, void *data) @@ -3226,14 +3219,11 @@ static irqreturn_t tegra_sor_irq(int irq, void *data) value = tegra_sor_readl(sor, SOR_AUDIO_HDA_CODEC_SCRATCH0); if (value & SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID) { - unsigned int format, sample_rate, channels; + unsigned int format; format = value & SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK; - tegra_hda_parse_format(format, &sample_rate, &channels); - - sor->audio.sample_rate = sample_rate; - sor->audio.channels = channels; + tegra_hda_parse_format(format, &sor->format); tegra_sor_hdmi_audio_enable(sor); } else { diff --git a/drivers/gpu/drm/tegra/vic.c b/drivers/gpu/drm/tegra/vic.c index d47983deb1cf..39bfed9623de 100644 --- a/drivers/gpu/drm/tegra/vic.c +++ b/drivers/gpu/drm/tegra/vic.c @@ -26,6 +26,7 @@ struct vic_config { const char *firmware; unsigned int version; + bool supports_sid; }; struct vic { @@ -105,6 +106,22 @@ static int vic_boot(struct vic *vic) if (vic->booted) return 0; + if (vic->config->supports_sid) { + struct iommu_fwspec *spec = dev_iommu_fwspec_get(vic->dev); + u32 value; + + value = TRANSCFG_ATT(1, TRANSCFG_SID_FALCON) | + TRANSCFG_ATT(0, TRANSCFG_SID_HW); + vic_writel(vic, value, VIC_TFBIF_TRANSCFG); + + if (spec && spec->num_ids > 0) { + value = spec->ids[0] & 0xffff; + + vic_writel(vic, value, VIC_THI_STREAMID0); + vic_writel(vic, value, VIC_THI_STREAMID1); + } + } + /* setup clockgating registers */ vic_writel(vic, CG_IDLE_CG_DLY_CNT(4) | CG_IDLE_CG_EN | @@ -181,13 +198,6 @@ static int vic_init(struct host1x_client *client) vic->domain = tegra->domain; } - if (!vic->falcon.data) { - vic->falcon.data = tegra; - err = falcon_load_firmware(&vic->falcon); - if (err < 0) - goto detach; - } - vic->channel = host1x_channel_request(client->dev); if (!vic->channel) { err = -ENOMEM; @@ -246,6 +256,30 @@ static const struct host1x_client_ops vic_client_ops = { .exit = vic_exit, }; +static int vic_load_firmware(struct vic *vic) +{ + int err; + + if (vic->falcon.data) + return 0; + + vic->falcon.data = vic->client.drm; + + err = falcon_read_firmware(&vic->falcon, vic->config->firmware); + if (err < 0) + goto cleanup; + + err = falcon_load_firmware(&vic->falcon); + if (err < 0) + goto cleanup; + + return 0; + +cleanup: + vic->falcon.data = NULL; + return err; +} + static int vic_open_channel(struct tegra_drm_client *client, struct tegra_drm_context *context) { @@ -256,19 +290,25 @@ static int vic_open_channel(struct tegra_drm_client *client, if (err < 0) return err; + err = vic_load_firmware(vic); + if (err < 0) + goto rpm_put; + err = vic_boot(vic); - if (err < 0) { - pm_runtime_put(vic->dev); - return err; - } + if (err < 0) + goto rpm_put; context->channel = host1x_channel_get(vic->channel); if (!context->channel) { - pm_runtime_put(vic->dev); - return -ENOMEM; + err = -ENOMEM; + goto rpm_put; } return 0; + +rpm_put: + pm_runtime_put(vic->dev); + return err; } static void vic_close_channel(struct tegra_drm_context *context) @@ -291,6 +331,7 @@ static const struct tegra_drm_client_ops vic_ops = { static const struct vic_config vic_t124_config = { .firmware = NVIDIA_TEGRA_124_VIC_FIRMWARE, .version = 0x40, + .supports_sid = false, }; #define NVIDIA_TEGRA_210_VIC_FIRMWARE "nvidia/tegra210/vic04_ucode.bin" @@ -298,6 +339,7 @@ static const struct vic_config vic_t124_config = { static const struct vic_config vic_t210_config = { .firmware = NVIDIA_TEGRA_210_VIC_FIRMWARE, .version = 0x21, + .supports_sid = false, }; #define NVIDIA_TEGRA_186_VIC_FIRMWARE "nvidia/tegra186/vic04_ucode.bin" @@ -305,6 +347,7 @@ static const struct vic_config vic_t210_config = { static const struct vic_config vic_t186_config = { .firmware = NVIDIA_TEGRA_186_VIC_FIRMWARE, .version = 0x18, + .supports_sid = true, }; #define NVIDIA_TEGRA_194_VIC_FIRMWARE "nvidia/tegra194/vic.bin" @@ -312,6 +355,7 @@ static const struct vic_config vic_t186_config = { static const struct vic_config vic_t194_config = { .firmware = NVIDIA_TEGRA_194_VIC_FIRMWARE, .version = 0x19, + .supports_sid = true, }; static const struct of_device_id vic_match[] = { @@ -372,10 +416,6 @@ static int vic_probe(struct platform_device *pdev) if (err < 0) return err; - err = falcon_read_firmware(&vic->falcon, vic->config->firmware); - if (err < 0) - goto exit_falcon; - platform_set_drvdata(pdev, vic); INIT_LIST_HEAD(&vic->client.base.list); @@ -393,7 +433,6 @@ static int vic_probe(struct platform_device *pdev) err = host1x_client_register(&vic->client.base); if (err < 0) { dev_err(dev, "failed to register host1x client: %d\n", err); - platform_set_drvdata(pdev, NULL); goto exit_falcon; } diff --git a/drivers/gpu/drm/tegra/vic.h b/drivers/gpu/drm/tegra/vic.h index 21844817a7e1..017584340dd6 100644 --- a/drivers/gpu/drm/tegra/vic.h +++ b/drivers/gpu/drm/tegra/vic.h @@ -17,11 +17,20 @@ /* VIC registers */ +#define VIC_THI_STREAMID0 0x00000030 +#define VIC_THI_STREAMID1 0x00000034 + #define NV_PVIC_MISC_PRI_VIC_CG 0x000016d0 #define CG_IDLE_CG_DLY_CNT(val) ((val & 0x3f) << 0) #define CG_IDLE_CG_EN (1 << 6) #define CG_WAKEUP_DLY_CNT(val) ((val & 0xf) << 16) +#define VIC_TFBIF_TRANSCFG 0x00002044 +#define TRANSCFG_ATT(i, v) (((v) & 0x3) << (i * 4)) +#define TRANSCFG_SID_HW 0 +#define TRANSCFG_SID_PHY 1 +#define TRANSCFG_SID_FALCON 2 + /* Firmware offsets */ #define VIC_UCODE_FCE_HEADER_OFFSET (6*4) diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 0ec08394e17a..3f56647cdb35 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -198,19 +198,22 @@ static void ttm_bo_ref_bug(struct kref *list_kref) void ttm_bo_del_from_lru(struct ttm_buffer_object *bo) { + struct ttm_bo_device *bdev = bo->bdev; + bool notify = false; + if (!list_empty(&bo->swap)) { list_del_init(&bo->swap); kref_put(&bo->list_kref, ttm_bo_ref_bug); + notify = true; } if (!list_empty(&bo->lru)) { list_del_init(&bo->lru); kref_put(&bo->list_kref, ttm_bo_ref_bug); + notify = true; } - /* - * TODO: Add a driver hook to delete from - * driver-specific LRU's here. - */ + if (notify && bdev->driver->del_from_lru_notify) + bdev->driver->del_from_lru_notify(bo); } void ttm_bo_del_sub_from_lru(struct ttm_buffer_object *bo) @@ -676,15 +679,6 @@ void ttm_bo_put(struct ttm_buffer_object *bo) } EXPORT_SYMBOL(ttm_bo_put); -void ttm_bo_unref(struct ttm_buffer_object **p_bo) -{ - struct ttm_buffer_object *bo = *p_bo; - - *p_bo = NULL; - ttm_bo_put(bo); -} -EXPORT_SYMBOL(ttm_bo_unref); - int ttm_bo_lock_delayed_workqueue(struct ttm_bo_device *bdev) { return cancel_delayed_work_sync(&bdev->wq); diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c index a1d977fbade5..e86a29a1e51f 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_vm.c +++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c @@ -71,7 +71,7 @@ static vm_fault_t ttm_bo_vm_fault_idle(struct ttm_buffer_object *bo, ttm_bo_get(bo); up_read(&vmf->vma->vm_mm->mmap_sem); (void) dma_fence_wait(bo->moving, true); - ttm_bo_unreserve(bo); + reservation_object_unlock(bo->resv); ttm_bo_put(bo); goto out_unlock; } @@ -131,11 +131,7 @@ static vm_fault_t ttm_bo_vm_fault(struct vm_fault *vmf) * for reserve, and if it fails, retry the fault after waiting * for the buffer to become unreserved. */ - err = ttm_bo_reserve(bo, true, true, NULL); - if (unlikely(err != 0)) { - if (err != -EBUSY) - return VM_FAULT_NOPAGE; - + if (unlikely(!reservation_object_trylock(bo->resv))) { if (vmf->flags & FAULT_FLAG_ALLOW_RETRY) { if (!(vmf->flags & FAULT_FLAG_RETRY_NOWAIT)) { ttm_bo_get(bo); @@ -165,6 +161,8 @@ static vm_fault_t ttm_bo_vm_fault(struct vm_fault *vmf) } if (bdev->driver->fault_reserve_notify) { + struct dma_fence *moving = dma_fence_get(bo->moving); + err = bdev->driver->fault_reserve_notify(bo); switch (err) { case 0: @@ -177,6 +175,13 @@ static vm_fault_t ttm_bo_vm_fault(struct vm_fault *vmf) ret = VM_FAULT_SIGBUS; goto out_unlock; } + + if (bo->moving != moving) { + spin_lock(&bdev->glob->lru_lock); + ttm_bo_move_to_lru_tail(bo, NULL); + spin_unlock(&bdev->glob->lru_lock); + } + dma_fence_put(moving); } /* @@ -291,7 +296,7 @@ static vm_fault_t ttm_bo_vm_fault(struct vm_fault *vmf) out_io_unlock: ttm_mem_io_unlock(man); out_unlock: - ttm_bo_unreserve(bo); + reservation_object_unlock(bo->resv); return ret; } diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c index f7508e907536..4704b2df3688 100644 --- a/drivers/gpu/drm/v3d/v3d_sched.c +++ b/drivers/gpu/drm/v3d/v3d_sched.c @@ -234,18 +234,21 @@ v3d_gpu_reset_for_timeout(struct v3d_dev *v3d, struct drm_sched_job *sched_job) for (q = 0; q < V3D_MAX_QUEUES; q++) { struct drm_gpu_scheduler *sched = &v3d->queue[q].sched; - kthread_park(sched->thread); - drm_sched_hw_job_reset(sched, (sched_job->sched == sched ? - sched_job : NULL)); + drm_sched_stop(sched); + + if(sched_job) + drm_sched_increase_karma(sched_job); } /* get the GPU back into the init state */ v3d_reset(v3d); + for (q = 0; q < V3D_MAX_QUEUES; q++) + drm_sched_resubmit_jobs(sched_job->sched); + /* Unblock schedulers and restart their jobs. */ for (q = 0; q < V3D_MAX_QUEUES; q++) { - drm_sched_job_recovery(&v3d->queue[q].sched); - kthread_unpark(v3d->queue[q].sched.thread); + drm_sched_start(&v3d->queue[q].sched, true); } mutex_unlock(&v3d->reset_lock); diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c index f39a183d59c2..e7e946035027 100644 --- a/drivers/gpu/drm/virtio/virtgpu_object.c +++ b/drivers/gpu/drm/virtio/virtgpu_object.c @@ -28,10 +28,21 @@ static int virtio_gpu_resource_id_get(struct virtio_gpu_device *vgdev, uint32_t *resid) { +#if 0 int handle = ida_alloc(&vgdev->resource_ida, GFP_KERNEL); if (handle < 0) return handle; +#else + static int handle; + + /* + * FIXME: dirty hack to avoid re-using IDs, virglrenderer + * can't deal with that. Needs fixing in virglrenderer, also + * should figure a better way to handle that in the guest. + */ + handle++; +#endif *resid = handle + 1; return 0; @@ -39,7 +50,9 @@ static int virtio_gpu_resource_id_get(struct virtio_gpu_device *vgdev, static void virtio_gpu_resource_id_put(struct virtio_gpu_device *vgdev, uint32_t id) { +#if 0 ida_free(&vgdev->resource_ida, id - 1); +#endif } static void virtio_gpu_ttm_bo_destroy(struct ttm_buffer_object *tbo) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c index 7ce1c2f87d9a..5d5c2bce01f3 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c @@ -534,7 +534,6 @@ static void vmw_user_bo_release(struct ttm_base_object **p_base) { struct vmw_user_buffer_object *vmw_user_bo; struct ttm_base_object *base = *p_base; - struct ttm_buffer_object *bo; *p_base = NULL; @@ -543,8 +542,7 @@ static void vmw_user_bo_release(struct ttm_base_object **p_base) vmw_user_bo = container_of(base, struct vmw_user_buffer_object, prime.base); - bo = &vmw_user_bo->vbo.base; - ttm_bo_unref(&bo); + ttm_bo_put(&vmw_user_bo->vbo.base); } @@ -597,7 +595,6 @@ int vmw_user_bo_alloc(struct vmw_private *dev_priv, struct ttm_base_object **p_base) { struct vmw_user_buffer_object *user_bo; - struct ttm_buffer_object *tmp; int ret; user_bo = kzalloc(sizeof(*user_bo), GFP_KERNEL); @@ -614,7 +611,7 @@ int vmw_user_bo_alloc(struct vmw_private *dev_priv, if (unlikely(ret != 0)) return ret; - tmp = ttm_bo_reference(&user_bo->vbo.base); + ttm_bo_get(&user_bo->vbo.base); ret = ttm_prime_object_init(tfile, size, &user_bo->prime, @@ -623,7 +620,7 @@ int vmw_user_bo_alloc(struct vmw_private *dev_priv, &vmw_user_bo_release, &vmw_user_bo_ref_obj_release); if (unlikely(ret != 0)) { - ttm_bo_unref(&tmp); + ttm_bo_put(&user_bo->vbo.base); goto out_no_base_object; } @@ -911,7 +908,7 @@ int vmw_user_bo_lookup(struct ttm_object_file *tfile, vmw_user_bo = container_of(base, struct vmw_user_buffer_object, prime.base); - (void)ttm_bo_reference(&vmw_user_bo->vbo.base); + ttm_bo_get(&vmw_user_bo->vbo.base); if (p_base) *p_base = base; else diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf.c index 48d1380a952e..70dab55e7888 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf.c @@ -765,7 +765,7 @@ static bool vmw_cmdbuf_try_alloc(struct vmw_cmdbuf_man *man, if (info->done) return true; - + memset(info->node, 0, sizeof(*info->node)); spin_lock(&man->lock); ret = drm_mm_insert_node(&man->mm, info->node, info->page_size); @@ -1276,8 +1276,10 @@ int vmw_cmdbuf_set_pool_size(struct vmw_cmdbuf_man *man, return 0; out_no_map: - if (man->using_mob) - ttm_bo_unref(&man->cmd_space); + if (man->using_mob) { + ttm_bo_put(man->cmd_space); + man->cmd_space = NULL; + } return ret; } @@ -1380,7 +1382,8 @@ void vmw_cmdbuf_remove_pool(struct vmw_cmdbuf_man *man) (void) vmw_cmdbuf_idle(man, false, 10*HZ); if (man->using_mob) { (void) ttm_bo_kunmap(&man->map_obj); - ttm_bo_unref(&man->cmd_space); + ttm_bo_put(man->cmd_space); + man->cmd_space = NULL; } else { dma_free_coherent(&man->dev_priv->dev->pdev->dev, man->size, man->map, man->handle); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index cd607ba9c2fe..accb2fafe2f1 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -1337,18 +1337,15 @@ static inline void vmw_bo_unreference(struct vmw_buffer_object **buf) *buf = NULL; if (tmp_buf != NULL) { - struct ttm_buffer_object *bo = &tmp_buf->base; - - ttm_bo_unref(&bo); + ttm_bo_put(&tmp_buf->base); } } static inline struct vmw_buffer_object * vmw_bo_reference(struct vmw_buffer_object *buf) { - if (ttm_bo_reference(&buf->base)) - return buf; - return NULL; + ttm_bo_get(&buf->base); + return buf; } static inline struct ttm_mem_global *vmw_mem_glob(struct vmw_private *dev_priv) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c b/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c index 7ed179d30ec5..d83cc66e1210 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_mob.c @@ -300,7 +300,8 @@ out_no_setup: &batch->otables[i]); } - ttm_bo_unref(&batch->otable_bo); + ttm_bo_put(batch->otable_bo); + batch->otable_bo = NULL; out_no_bo: return ret; } @@ -365,7 +366,8 @@ static void vmw_otable_batch_takedown(struct vmw_private *dev_priv, vmw_bo_fence_single(bo, NULL); ttm_bo_unreserve(bo); - ttm_bo_unref(&batch->otable_bo); + ttm_bo_put(batch->otable_bo); + batch->otable_bo = NULL; } /* @@ -463,7 +465,8 @@ static int vmw_mob_pt_populate(struct vmw_private *dev_priv, out_unreserve: ttm_bo_unreserve(mob->pt_bo); - ttm_bo_unref(&mob->pt_bo); + ttm_bo_put(mob->pt_bo); + mob->pt_bo = NULL; return ret; } @@ -580,8 +583,10 @@ static void vmw_mob_pt_setup(struct vmw_mob *mob, */ void vmw_mob_destroy(struct vmw_mob *mob) { - if (mob->pt_bo) - ttm_bo_unref(&mob->pt_bo); + if (mob->pt_bo) { + ttm_bo_put(mob->pt_bo); + mob->pt_bo = NULL; + } kfree(mob); } @@ -698,8 +703,10 @@ int vmw_mob_bind(struct vmw_private *dev_priv, out_no_cmd_space: vmw_fifo_resource_dec(dev_priv); - if (pt_set_up) - ttm_bo_unref(&mob->pt_bo); + if (pt_set_up) { + ttm_bo_put(mob->pt_bo); + mob->pt_bo = NULL; + } return -ENOMEM; } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c index 3025bfc001a1..a7c30e567f09 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c @@ -461,7 +461,8 @@ vmw_resource_check_buffer(struct ww_acquire_ctx *ticket, } INIT_LIST_HEAD(&val_list); - val_buf->bo = ttm_bo_reference(&res->backup->base); + ttm_bo_get(&res->backup->base); + val_buf->bo = &res->backup->base; val_buf->num_shared = 0; list_add_tail(&val_buf->head, &val_list); ret = ttm_eu_reserve_buffers(ticket, &val_list, interruptible, NULL); @@ -484,7 +485,8 @@ vmw_resource_check_buffer(struct ww_acquire_ctx *ticket, out_no_validate: ttm_eu_backoff_reservation(ticket, &val_list); out_no_reserve: - ttm_bo_unref(&val_buf->bo); + ttm_bo_put(val_buf->bo); + val_buf->bo = NULL; if (backup_dirty) vmw_bo_unreference(&res->backup); @@ -544,7 +546,8 @@ vmw_resource_backoff_reservation(struct ww_acquire_ctx *ticket, INIT_LIST_HEAD(&val_list); list_add_tail(&val_buf->head, &val_list); ttm_eu_backoff_reservation(ticket, &val_list); - ttm_bo_unref(&val_buf->bo); + ttm_bo_put(val_buf->bo); + val_buf->bo = NULL; } /** diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c index b3f547fc5d3d..e9944ac2e057 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_validation.c @@ -628,8 +628,10 @@ void vmw_validation_unref_lists(struct vmw_validation_context *ctx) struct vmw_validation_bo_node *entry; struct vmw_validation_res_node *val; - list_for_each_entry(entry, &ctx->bo_list, base.head) - ttm_bo_unref(&entry->base.bo); + list_for_each_entry(entry, &ctx->bo_list, base.head) { + ttm_bo_put(entry->base.bo); + entry->base.bo = NULL; + } list_splice_init(&ctx->resource_ctx_list, &ctx->resource_list); list_for_each_entry(val, &ctx->resource_list, head) diff --git a/drivers/gpu/host1x/bus.c b/drivers/gpu/host1x/bus.c index b4c385d4a6af..103fffc1904b 100644 --- a/drivers/gpu/host1x/bus.c +++ b/drivers/gpu/host1x/bus.c @@ -15,8 +15,10 @@ * along with this program. If not, see <http://www.gnu.org/licenses/>. */ +#include <linux/debugfs.h> #include <linux/host1x.h> #include <linux/of.h> +#include <linux/seq_file.h> #include <linux/slab.h> #include <linux/of_device.h> @@ -500,6 +502,36 @@ static void host1x_detach_driver(struct host1x *host1x, mutex_unlock(&host1x->devices_lock); } +static int host1x_devices_show(struct seq_file *s, void *data) +{ + struct host1x *host1x = s->private; + struct host1x_device *device; + + mutex_lock(&host1x->devices_lock); + + list_for_each_entry(device, &host1x->devices, list) { + struct host1x_subdev *subdev; + + seq_printf(s, "%s\n", dev_name(&device->dev)); + + mutex_lock(&device->subdevs_lock); + + list_for_each_entry(subdev, &device->active, list) + seq_printf(s, " %pOFf: %s\n", subdev->np, + dev_name(subdev->client->dev)); + + list_for_each_entry(subdev, &device->subdevs, list) + seq_printf(s, " %pOFf:\n", subdev->np); + + mutex_unlock(&device->subdevs_lock); + } + + mutex_unlock(&host1x->devices_lock); + + return 0; +} +DEFINE_SHOW_ATTRIBUTE(host1x_devices); + /** * host1x_register() - register a host1x controller * @host1x: host1x controller @@ -523,6 +555,9 @@ int host1x_register(struct host1x *host1x) mutex_unlock(&drivers_lock); + debugfs_create_file("devices", S_IRUGO, host1x->debugfs, host1x, + &host1x_devices_fops); + return 0; } diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c index 91df51e631b2..f45b7c69b694 100644 --- a/drivers/gpu/host1x/cdma.c +++ b/drivers/gpu/host1x/cdma.c @@ -41,7 +41,17 @@ * means that the push buffer is full, not empty. */ -#define HOST1X_PUSHBUFFER_SLOTS 512 +/* + * Typically the commands written into the push buffer are a pair of words. We + * use slots to represent each of these pairs and to simplify things. Note the + * strange number of slots allocated here. 512 slots will fit exactly within a + * single memory page. We also need one additional word at the end of the push + * buffer for the RESTART opcode that will instruct the CDMA to jump back to + * the beginning of the push buffer. With 512 slots, this means that we'll use + * 2 memory pages and waste 4092 bytes of the second page that will never be + * used. + */ +#define HOST1X_PUSHBUFFER_SLOTS 511 /* * Clean up push buffer resources @@ -143,7 +153,10 @@ static void host1x_pushbuffer_push(struct push_buffer *pb, u32 op1, u32 op2) WARN_ON(pb->pos == pb->fence); *(p++) = op1; *(p++) = op2; - pb->pos = (pb->pos + 8) & (pb->size - 1); + pb->pos += 8; + + if (pb->pos >= pb->size) + pb->pos -= pb->size; } /* @@ -153,7 +166,10 @@ static void host1x_pushbuffer_push(struct push_buffer *pb, u32 op1, u32 op2) static void host1x_pushbuffer_pop(struct push_buffer *pb, unsigned int slots) { /* Advance the next write position */ - pb->fence = (pb->fence + slots * 8) & (pb->size - 1); + pb->fence += slots * 8; + + if (pb->fence >= pb->size) + pb->fence -= pb->size; } /* @@ -161,7 +177,12 @@ static void host1x_pushbuffer_pop(struct push_buffer *pb, unsigned int slots) */ static u32 host1x_pushbuffer_space(struct push_buffer *pb) { - return ((pb->fence - pb->pos) & (pb->size - 1)) / 8; + unsigned int fence = pb->fence; + + if (pb->fence < pb->pos) + fence += pb->size; + + return (fence - pb->pos) / 8; } /* @@ -210,7 +231,7 @@ unsigned int host1x_cdma_wait_locked(struct host1x_cdma *cdma, cdma->event = event; mutex_unlock(&cdma->lock); - down(&cdma->sem); + wait_for_completion(&cdma->complete); mutex_lock(&cdma->lock); } @@ -218,6 +239,45 @@ unsigned int host1x_cdma_wait_locked(struct host1x_cdma *cdma, } /* + * Sleep (if necessary) until the push buffer has enough free space. + * + * Must be called with the cdma lock held. + */ +int host1x_cdma_wait_pushbuffer_space(struct host1x *host1x, + struct host1x_cdma *cdma, + unsigned int needed) +{ + while (true) { + struct push_buffer *pb = &cdma->push_buffer; + unsigned int space; + + space = host1x_pushbuffer_space(pb); + if (space >= needed) + break; + + trace_host1x_wait_cdma(dev_name(cdma_to_channel(cdma)->dev), + CDMA_EVENT_PUSH_BUFFER_SPACE); + + host1x_hw_cdma_flush(host1x, cdma); + + /* If somebody has managed to already start waiting, yield */ + if (cdma->event != CDMA_EVENT_NONE) { + mutex_unlock(&cdma->lock); + schedule(); + mutex_lock(&cdma->lock); + continue; + } + + cdma->event = CDMA_EVENT_PUSH_BUFFER_SPACE; + + mutex_unlock(&cdma->lock); + wait_for_completion(&cdma->complete); + mutex_lock(&cdma->lock); + } + + return 0; +} +/* * Start timer that tracks the time spent by the job. * Must be called with the cdma lock held. */ @@ -314,7 +374,7 @@ static void update_cdma_locked(struct host1x_cdma *cdma) if (signal) { cdma->event = CDMA_EVENT_NONE; - up(&cdma->sem); + complete(&cdma->complete); } } @@ -323,7 +383,7 @@ void host1x_cdma_update_sync_queue(struct host1x_cdma *cdma, { struct host1x *host1x = cdma_to_host1x(cdma); u32 restart_addr, syncpt_incrs, syncpt_val; - struct host1x_job *job = NULL; + struct host1x_job *job, *next_job = NULL; syncpt_val = host1x_syncpt_load(cdma->timeout.syncpt); @@ -341,40 +401,37 @@ void host1x_cdma_update_sync_queue(struct host1x_cdma *cdma, __func__); list_for_each_entry(job, &cdma->sync_queue, list) { - if (syncpt_val < job->syncpt_end) - break; + if (syncpt_val < job->syncpt_end) { + + if (!list_is_last(&job->list, &cdma->sync_queue)) + next_job = list_next_entry(job, list); + + goto syncpt_incr; + } host1x_job_dump(dev, job); } + /* all jobs have been completed */ + job = NULL; + +syncpt_incr: + /* - * Walk the sync_queue, first incrementing with the CPU syncpts that - * are partially executed (the first buffer) or fully skipped while - * still in the current context (slots are also NOP-ed). + * Increment with CPU the remaining syncpts of a partially executed job. * - * At the point contexts are interleaved, syncpt increments must be - * done inline with the pushbuffer from a GATHER buffer to maintain - * the order (slots are modified to be a GATHER of syncpt incrs). - * - * Note: save in restart_addr the location where the timed out buffer - * started in the PB, so we can start the refetch from there (with the - * modified NOP-ed PB slots). This lets things appear to have completed - * properly for this buffer and resources are freed. + * CDMA will continue execution starting with the next job or will get + * into idle state. */ - - dev_dbg(dev, "%s: perform CPU incr on pending same ctx buffers\n", - __func__); - - if (!list_empty(&cdma->sync_queue)) - restart_addr = job->first_get; + if (next_job) + restart_addr = next_job->first_get; else restart_addr = cdma->last_pos; - /* do CPU increments as long as this context continues */ - list_for_each_entry_from(job, &cdma->sync_queue, list) { - /* different context, gets us out of this loop */ - if (job->client != cdma->timeout.client) - break; + /* do CPU increments for the remaining syncpts */ + if (job) { + dev_dbg(dev, "%s: perform CPU incr on pending buffers\n", + __func__); /* won't need a timeout when replayed */ job->timeout = 0; @@ -389,21 +446,10 @@ void host1x_cdma_update_sync_queue(struct host1x_cdma *cdma, syncpt_incrs, job->syncpt_end, job->num_slots); - syncpt_val += syncpt_incrs; + dev_dbg(dev, "%s: finished sync_queue modification\n", + __func__); } - /* - * The following sumbits from the same client may be dependent on the - * failed submit and therefore they may fail. Force a small timeout - * to make the queue cleanup faster. - */ - - list_for_each_entry_from(job, &cdma->sync_queue, list) - if (job->client == cdma->timeout.client) - job->timeout = min_t(unsigned int, job->timeout, 500); - - dev_dbg(dev, "%s: finished sync_queue modification\n", __func__); - /* roll back DMAGET and start up channel again */ host1x_hw_cdma_resume(host1x, cdma, restart_addr); } @@ -416,7 +462,7 @@ int host1x_cdma_init(struct host1x_cdma *cdma) int err; mutex_init(&cdma->lock); - sema_init(&cdma->sem, 0); + init_completion(&cdma->complete); INIT_LIST_HEAD(&cdma->sync_queue); @@ -510,6 +556,59 @@ void host1x_cdma_push(struct host1x_cdma *cdma, u32 op1, u32 op2) } /* + * Push four words into two consecutive push buffer slots. Note that extra + * care needs to be taken not to split the two slots across the end of the + * push buffer. Otherwise the RESTART opcode at the end of the push buffer + * that ensures processing will restart at the beginning will break up the + * four words. + * + * Blocks as necessary if the push buffer is full. + */ +void host1x_cdma_push_wide(struct host1x_cdma *cdma, u32 op1, u32 op2, + u32 op3, u32 op4) +{ + struct host1x_channel *channel = cdma_to_channel(cdma); + struct host1x *host1x = cdma_to_host1x(cdma); + struct push_buffer *pb = &cdma->push_buffer; + unsigned int needed = 2, extra = 0, i; + unsigned int space = cdma->slots_free; + + if (host1x_debug_trace_cmdbuf) + trace_host1x_cdma_push_wide(dev_name(channel->dev), op1, op2, + op3, op4); + + /* compute number of extra slots needed for padding */ + if (pb->pos + 16 > pb->size) { + extra = (pb->size - pb->pos) / 8; + needed += extra; + } + + host1x_cdma_wait_pushbuffer_space(host1x, cdma, needed); + space = host1x_pushbuffer_space(pb); + + cdma->slots_free = space - needed; + cdma->slots_used += needed; + + /* + * Note that we rely on the fact that this is only used to submit wide + * gather opcodes, which consist of 3 words, and they are padded with + * a NOP to avoid having to deal with fractional slots (a slot always + * represents 2 words). The fourth opcode passed to this function will + * therefore always be a NOP. + * + * This works around a slight ambiguity when it comes to opcodes. For + * all current host1x incarnations the NOP opcode uses the exact same + * encoding (0x20000000), so we could hard-code the value here, but a + * new incarnation may change it and break that assumption. + */ + for (i = 0; i < extra; i++) + host1x_pushbuffer_push(pb, op4, op4); + + host1x_pushbuffer_push(pb, op1, op2); + host1x_pushbuffer_push(pb, op3, op4); +} + +/* * End a cdma submit * Kick off DMA, add job to the sync queue, and a number of slots to be freed * from the pushbuffer. The handles for a submit must all be pinned at the same diff --git a/drivers/gpu/host1x/cdma.h b/drivers/gpu/host1x/cdma.h index e97e17b82370..3a5e0408b8d1 100644 --- a/drivers/gpu/host1x/cdma.h +++ b/drivers/gpu/host1x/cdma.h @@ -20,7 +20,7 @@ #define __HOST1X_CDMA_H #include <linux/sched.h> -#include <linux/semaphore.h> +#include <linux/completion.h> #include <linux/list.h> struct host1x_syncpt; @@ -69,8 +69,8 @@ enum cdma_event { struct host1x_cdma { struct mutex lock; /* controls access to shared state */ - struct semaphore sem; /* signalled when event occurs */ - enum cdma_event event; /* event that sem is waiting for */ + struct completion complete; /* signalled when event occurs */ + enum cdma_event event; /* event that complete is waiting for */ unsigned int slots_used; /* pb slots used in current submit */ unsigned int slots_free; /* pb slots free in current submit */ unsigned int first_get; /* DMAGET value, where submit begins */ @@ -90,6 +90,8 @@ int host1x_cdma_init(struct host1x_cdma *cdma); int host1x_cdma_deinit(struct host1x_cdma *cdma); int host1x_cdma_begin(struct host1x_cdma *cdma, struct host1x_job *job); void host1x_cdma_push(struct host1x_cdma *cdma, u32 op1, u32 op2); +void host1x_cdma_push_wide(struct host1x_cdma *cdma, u32 op1, u32 op2, + u32 op3, u32 op4); void host1x_cdma_end(struct host1x_cdma *cdma, struct host1x_job *job); void host1x_cdma_update(struct host1x_cdma *cdma); void host1x_cdma_peek(struct host1x_cdma *cdma, u32 dmaget, int slot, diff --git a/drivers/gpu/host1x/dev.c b/drivers/gpu/host1x/dev.c index 419d8929a98f..ee3c7b81a29d 100644 --- a/drivers/gpu/host1x/dev.c +++ b/drivers/gpu/host1x/dev.c @@ -120,6 +120,15 @@ static const struct host1x_info host1x05_info = { .dma_mask = DMA_BIT_MASK(34), }; +static const struct host1x_sid_entry tegra186_sid_table[] = { + { + /* VIC */ + .base = 0x1af0, + .offset = 0x30, + .limit = 0x34 + }, +}; + static const struct host1x_info host1x06_info = { .nb_channels = 63, .nb_pts = 576, @@ -127,8 +136,19 @@ static const struct host1x_info host1x06_info = { .nb_bases = 16, .init = host1x06_init, .sync_offset = 0x0, - .dma_mask = DMA_BIT_MASK(34), + .dma_mask = DMA_BIT_MASK(40), .has_hypervisor = true, + .num_sid_entries = ARRAY_SIZE(tegra186_sid_table), + .sid_table = tegra186_sid_table, +}; + +static const struct host1x_sid_entry tegra194_sid_table[] = { + { + /* VIC */ + .base = 0x1af0, + .offset = 0x30, + .limit = 0x34 + }, }; static const struct host1x_info host1x07_info = { @@ -140,6 +160,8 @@ static const struct host1x_info host1x07_info = { .sync_offset = 0x0, .dma_mask = DMA_BIT_MASK(40), .has_hypervisor = true, + .num_sid_entries = ARRAY_SIZE(tegra194_sid_table), + .sid_table = tegra194_sid_table, }; static const struct of_device_id host1x_of_match[] = { @@ -154,6 +176,19 @@ static const struct of_device_id host1x_of_match[] = { }; MODULE_DEVICE_TABLE(of, host1x_of_match); +static void host1x_setup_sid_table(struct host1x *host) +{ + const struct host1x_info *info = host->info; + unsigned int i; + + for (i = 0; i < info->num_sid_entries; i++) { + const struct host1x_sid_entry *entry = &info->sid_table[i]; + + host1x_hypervisor_writel(host, entry->offset, entry->base); + host1x_hypervisor_writel(host, entry->limit, entry->base + 4); + } +} + static int host1x_probe(struct platform_device *pdev) { struct host1x *host; @@ -248,6 +283,8 @@ static int host1x_probe(struct platform_device *pdev) host->group = iommu_group_get(&pdev->dev); if (host->group) { struct iommu_domain_geometry *geometry; + u64 mask = dma_get_mask(host->dev); + dma_addr_t start, end; unsigned long order; err = iova_cache_get(); @@ -275,11 +312,12 @@ static int host1x_probe(struct platform_device *pdev) } geometry = &host->domain->geometry; + start = geometry->aperture_start & mask; + end = geometry->aperture_end & mask; order = __ffs(host->domain->pgsize_bitmap); - init_iova_domain(&host->iova, 1UL << order, - geometry->aperture_start >> order); - host->iova_end = geometry->aperture_end; + init_iova_domain(&host->iova, 1UL << order, start >> order); + host->iova_end = end; } skip_iommu: @@ -316,6 +354,9 @@ skip_iommu: host1x_debug_init(host); + if (host->info->has_hypervisor) + host1x_setup_sid_table(host); + err = host1x_register(host); if (err < 0) goto fail_deinit_intr; diff --git a/drivers/gpu/host1x/dev.h b/drivers/gpu/host1x/dev.h index 36f44ffebe73..05216a7e4830 100644 --- a/drivers/gpu/host1x/dev.h +++ b/drivers/gpu/host1x/dev.h @@ -94,6 +94,12 @@ struct host1x_intr_ops { int (*free_syncpt_irq)(struct host1x *host); }; +struct host1x_sid_entry { + unsigned int base; + unsigned int offset; + unsigned int limit; +}; + struct host1x_info { unsigned int nb_channels; /* host1x: number of channels supported */ unsigned int nb_pts; /* host1x: number of syncpoints supported */ @@ -103,6 +109,8 @@ struct host1x_info { unsigned int sync_offset; /* offset of syncpoint registers */ u64 dma_mask; /* mask of addressable memory */ bool has_hypervisor; /* has hypervisor registers */ + unsigned int num_sid_entries; + const struct host1x_sid_entry *sid_table; }; struct host1x { diff --git a/drivers/gpu/host1x/hw/cdma_hw.c b/drivers/gpu/host1x/hw/cdma_hw.c index ce320534cbed..5d61088db2bb 100644 --- a/drivers/gpu/host1x/hw/cdma_hw.c +++ b/drivers/gpu/host1x/hw/cdma_hw.c @@ -39,8 +39,6 @@ static void push_buffer_init(struct push_buffer *pb) static void cdma_timeout_cpu_incr(struct host1x_cdma *cdma, u32 getptr, u32 syncpt_incrs, u32 syncval, u32 nr_slots) { - struct host1x *host1x = cdma_to_host1x(cdma); - struct push_buffer *pb = &cdma->push_buffer; unsigned int i; for (i = 0; i < syncpt_incrs; i++) @@ -48,18 +46,6 @@ static void cdma_timeout_cpu_incr(struct host1x_cdma *cdma, u32 getptr, /* after CPU incr, ensure shadow is up to date */ host1x_syncpt_load(cdma->timeout.syncpt); - - /* NOP all the PB slots */ - while (nr_slots--) { - u32 *p = (u32 *)(pb->mapped + getptr); - *(p++) = HOST1X_OPCODE_NOP; - *(p++) = HOST1X_OPCODE_NOP; - dev_dbg(host1x->dev, "%s: NOP at %pad+%#x\n", __func__, - &pb->dma, getptr); - getptr = (getptr + 8) & (pb->size - 1); - } - - wmb(); } /* @@ -68,20 +54,31 @@ static void cdma_timeout_cpu_incr(struct host1x_cdma *cdma, u32 getptr, static void cdma_start(struct host1x_cdma *cdma) { struct host1x_channel *ch = cdma_to_channel(cdma); + u64 start, end; if (cdma->running) return; cdma->last_pos = cdma->push_buffer.pos; + start = cdma->push_buffer.dma; + end = cdma->push_buffer.size + 4; host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP, HOST1X_CHANNEL_DMACTRL); /* set base, put and end pointer */ - host1x_ch_writel(ch, cdma->push_buffer.dma, HOST1X_CHANNEL_DMASTART); + host1x_ch_writel(ch, lower_32_bits(start), HOST1X_CHANNEL_DMASTART); +#if HOST1X_HW >= 6 + host1x_ch_writel(ch, upper_32_bits(start), HOST1X_CHANNEL_DMASTART_HI); +#endif host1x_ch_writel(ch, cdma->push_buffer.pos, HOST1X_CHANNEL_DMAPUT); - host1x_ch_writel(ch, cdma->push_buffer.dma + cdma->push_buffer.size + 4, - HOST1X_CHANNEL_DMAEND); +#if HOST1X_HW >= 6 + host1x_ch_writel(ch, 0, HOST1X_CHANNEL_DMAPUT_HI); +#endif + host1x_ch_writel(ch, lower_32_bits(end), HOST1X_CHANNEL_DMAEND); +#if HOST1X_HW >= 6 + host1x_ch_writel(ch, upper_32_bits(end), HOST1X_CHANNEL_DMAEND_HI); +#endif /* reset GET */ host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP | @@ -104,6 +101,7 @@ static void cdma_timeout_restart(struct host1x_cdma *cdma, u32 getptr) { struct host1x *host1x = cdma_to_host1x(cdma); struct host1x_channel *ch = cdma_to_channel(cdma); + u64 start, end; if (cdma->running) return; @@ -113,10 +111,18 @@ static void cdma_timeout_restart(struct host1x_cdma *cdma, u32 getptr) host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP, HOST1X_CHANNEL_DMACTRL); + start = cdma->push_buffer.dma; + end = cdma->push_buffer.size + 4; + /* set base, end pointer (all of memory) */ - host1x_ch_writel(ch, cdma->push_buffer.dma, HOST1X_CHANNEL_DMASTART); - host1x_ch_writel(ch, cdma->push_buffer.dma + cdma->push_buffer.size, - HOST1X_CHANNEL_DMAEND); + host1x_ch_writel(ch, lower_32_bits(start), HOST1X_CHANNEL_DMASTART); +#if HOST1X_HW >= 6 + host1x_ch_writel(ch, upper_32_bits(start), HOST1X_CHANNEL_DMASTART_HI); +#endif + host1x_ch_writel(ch, lower_32_bits(end), HOST1X_CHANNEL_DMAEND); +#if HOST1X_HW >= 6 + host1x_ch_writel(ch, upper_32_bits(end), HOST1X_CHANNEL_DMAEND_HI); +#endif /* set GET, by loading the value in PUT (then reset GET) */ host1x_ch_writel(ch, getptr, HOST1X_CHANNEL_DMAPUT); diff --git a/drivers/gpu/host1x/hw/channel_hw.c b/drivers/gpu/host1x/hw/channel_hw.c index 95ea81172a83..27101c04a827 100644 --- a/drivers/gpu/host1x/hw/channel_hw.c +++ b/drivers/gpu/host1x/hw/channel_hw.c @@ -17,6 +17,7 @@ */ #include <linux/host1x.h> +#include <linux/iommu.h> #include <linux/slab.h> #include <trace/events/host1x.h> @@ -60,15 +61,37 @@ static void trace_write_gather(struct host1x_cdma *cdma, struct host1x_bo *bo, static void submit_gathers(struct host1x_job *job) { struct host1x_cdma *cdma = &job->channel->cdma; +#if HOST1X_HW < 6 + struct device *dev = job->channel->dev; +#endif unsigned int i; for (i = 0; i < job->num_gathers; i++) { struct host1x_job_gather *g = &job->gathers[i]; - u32 op1 = host1x_opcode_gather(g->words); - u32 op2 = g->base + g->offset; + dma_addr_t addr = g->base + g->offset; + u32 op2, op3; + + op2 = lower_32_bits(addr); + op3 = upper_32_bits(addr); + + trace_write_gather(cdma, g->bo, g->offset, g->words); + + if (op3 != 0) { +#if HOST1X_HW >= 6 + u32 op1 = host1x_opcode_gather_wide(g->words); + u32 op4 = HOST1X_OPCODE_NOP; + + host1x_cdma_push_wide(cdma, op1, op2, op3, op4); +#else + dev_err(dev, "invalid gather for push buffer %pad\n", + &addr); + continue; +#endif + } else { + u32 op1 = host1x_opcode_gather(g->words); - trace_write_gather(cdma, g->bo, g->offset, op1 & 0xffff); - host1x_cdma_push(cdma, op1, op2); + host1x_cdma_push(cdma, op1, op2); + } } } @@ -89,6 +112,16 @@ static inline void synchronize_syncpt_base(struct host1x_job *job) HOST1X_UCLASS_LOAD_SYNCPT_BASE_VALUE_F(value)); } +static void host1x_channel_set_streamid(struct host1x_channel *channel) +{ +#if HOST1X_HW >= 6 + struct iommu_fwspec *spec = dev_iommu_fwspec_get(channel->dev->parent); + u32 sid = spec ? spec->ids[0] & 0xffff : 0x7f; + + host1x_ch_writel(channel, sid, HOST1X_CHANNEL_SMMU_STREAMID); +#endif +} + static int channel_submit(struct host1x_job *job) { struct host1x_channel *ch = job->channel; @@ -120,6 +153,8 @@ static int channel_submit(struct host1x_job *job) goto error; } + host1x_channel_set_streamid(ch); + /* begin a CDMA submit */ err = host1x_cdma_begin(&ch->cdma, job); if (err) { diff --git a/drivers/gpu/host1x/hw/host1x06_hardware.h b/drivers/gpu/host1x/hw/host1x06_hardware.h index 3039c92ea605..dd37b10c8d04 100644 --- a/drivers/gpu/host1x/hw/host1x06_hardware.h +++ b/drivers/gpu/host1x/hw/host1x06_hardware.h @@ -22,6 +22,7 @@ #include <linux/types.h> #include <linux/bitops.h> +#include "hw_host1x06_channel.h" #include "hw_host1x06_uclass.h" #include "hw_host1x06_vm.h" #include "hw_host1x06_hypervisor.h" @@ -137,6 +138,11 @@ static inline u32 host1x_opcode_gather_incr(unsigned offset, unsigned count) return (6 << 28) | (offset << 16) | BIT(15) | BIT(14) | count; } +static inline u32 host1x_opcode_gather_wide(unsigned count) +{ + return (12 << 28) | count; +} + #define HOST1X_OPCODE_NOP host1x_opcode_nonincr(0, 0) #endif diff --git a/drivers/gpu/host1x/hw/host1x07_hardware.h b/drivers/gpu/host1x/hw/host1x07_hardware.h index 1353e7ab71dd..9f6da4ee5443 100644 --- a/drivers/gpu/host1x/hw/host1x07_hardware.h +++ b/drivers/gpu/host1x/hw/host1x07_hardware.h @@ -22,6 +22,7 @@ #include <linux/types.h> #include <linux/bitops.h> +#include "hw_host1x07_channel.h" #include "hw_host1x07_uclass.h" #include "hw_host1x07_vm.h" #include "hw_host1x07_hypervisor.h" @@ -137,6 +138,11 @@ static inline u32 host1x_opcode_gather_incr(unsigned offset, unsigned count) return (6 << 28) | (offset << 16) | BIT(15) | BIT(14) | count; } +static inline u32 host1x_opcode_gather_wide(unsigned count) +{ + return (12 << 28) | count; +} + #define HOST1X_OPCODE_NOP host1x_opcode_nonincr(0, 0) #endif diff --git a/drivers/gpu/host1x/hw/hw_host1x06_channel.h b/drivers/gpu/host1x/hw/hw_host1x06_channel.h new file mode 100644 index 000000000000..18ae1c57bbea --- /dev/null +++ b/drivers/gpu/host1x/hw/hw_host1x06_channel.h @@ -0,0 +1,11 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2019 NVIDIA Corporation. + */ + +#ifndef HOST1X_HW_HOST1X06_CHANNEL_H +#define HOST1X_HW_HOST1X06_CHANNEL_H + +#define HOST1X_CHANNEL_SMMU_STREAMID 0x084 + +#endif diff --git a/drivers/gpu/host1x/hw/hw_host1x07_channel.h b/drivers/gpu/host1x/hw/hw_host1x07_channel.h new file mode 100644 index 000000000000..96fa72bbd7ab --- /dev/null +++ b/drivers/gpu/host1x/hw/hw_host1x07_channel.h @@ -0,0 +1,11 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2019 NVIDIA Corporation. + */ + +#ifndef HOST1X_HW_HOST1X07_CHANNEL_H +#define HOST1X_HW_HOST1X07_CHANNEL_H + +#define HOST1X_CHANNEL_SMMU_STREAMID 0x084 + +#endif diff --git a/include/drm/drm_color_mgmt.h b/include/drm/drm_color_mgmt.h index 90ef9996d9a4..d1c662d92ab7 100644 --- a/include/drm/drm_color_mgmt.h +++ b/include/drm/drm_color_mgmt.h @@ -69,4 +69,32 @@ int drm_plane_create_color_properties(struct drm_plane *plane, u32 supported_ranges, enum drm_color_encoding default_encoding, enum drm_color_range default_range); + +/** + * enum drm_color_lut_tests - hw-specific LUT tests to perform + * + * The drm_color_lut_check() function takes a bitmask of the values here to + * determine which tests to apply to a userspace-provided LUT. + */ +enum drm_color_lut_tests { + /** + * @DRM_COLOR_LUT_EQUAL_CHANNELS: + * + * Checks whether the entries of a LUT all have equal values for the + * red, green, and blue channels. Intended for hardware that only + * accepts a single value per LUT entry and assumes that value applies + * to all three color components. + */ + DRM_COLOR_LUT_EQUAL_CHANNELS = BIT(0), + + /** + * @DRM_COLOR_LUT_NON_DECREASING: + * + * Checks whether the entries of a LUT are always flat or increasing + * (never decreasing). + */ + DRM_COLOR_LUT_NON_DECREASING = BIT(1), +}; + +int drm_color_lut_check(const struct drm_property_blob *lut, u32 tests); #endif diff --git a/include/drm/gpu_scheduler.h b/include/drm/gpu_scheduler.h index 47e19796c450..0daca4d8dad9 100644 --- a/include/drm/gpu_scheduler.h +++ b/include/drm/gpu_scheduler.h @@ -138,10 +138,6 @@ struct drm_sched_fence { struct dma_fence finished; /** - * @cb: the callback for the parent fence below. - */ - struct dma_fence_cb cb; - /** * @parent: the fence returned by &drm_sched_backend_ops.run_job * when scheduling the job on hardware. We signal the * &drm_sched_fence.finished fence once parent is signalled. @@ -181,6 +177,7 @@ struct drm_sched_fence *to_drm_sched_fence(struct dma_fence *f); * be scheduled further. * @s_priority: the priority of the job. * @entity: the entity to which this job belongs. + * @cb: the callback for the parent fence in s_fence. * * A job is created by the driver using drm_sched_job_init(), and * should call drm_sched_entity_push_job() once it wants the scheduler @@ -197,6 +194,7 @@ struct drm_sched_job { atomic_t karma; enum drm_sched_priority s_priority; struct drm_sched_entity *entity; + struct dma_fence_cb cb; }; static inline bool drm_sched_invalidate_job(struct drm_sched_job *s_job, @@ -298,9 +296,10 @@ int drm_sched_job_init(struct drm_sched_job *job, void *owner); void drm_sched_job_cleanup(struct drm_sched_job *job); void drm_sched_wakeup(struct drm_gpu_scheduler *sched); -void drm_sched_hw_job_reset(struct drm_gpu_scheduler *sched, - struct drm_sched_job *job); -void drm_sched_job_recovery(struct drm_gpu_scheduler *sched); +void drm_sched_stop(struct drm_gpu_scheduler *sched); +void drm_sched_start(struct drm_gpu_scheduler *sched, bool full_recovery); +void drm_sched_resubmit_jobs(struct drm_gpu_scheduler *sched); +void drm_sched_increase_karma(struct drm_sched_job *bad); bool drm_sched_dependency_optimized(struct dma_fence* fence, struct drm_sched_entity *entity); void drm_sched_fault(struct drm_gpu_scheduler *sched); diff --git a/include/drm/i915_pciids.h b/include/drm/i915_pciids.h index 192667144693..d2fad7b0fcf6 100644 --- a/include/drm/i915_pciids.h +++ b/include/drm/i915_pciids.h @@ -394,6 +394,9 @@ INTEL_VGA_DEVICE(0x3E9A, info) /* SRV GT2 */ /* CFL H */ +#define INTEL_CFL_H_GT1_IDS(info) \ + INTEL_VGA_DEVICE(0x3E9C, info) + #define INTEL_CFL_H_GT2_IDS(info) \ INTEL_VGA_DEVICE(0x3E9B, info), /* Halo GT2 */ \ INTEL_VGA_DEVICE(0x3E94, info) /* Halo GT2 */ @@ -426,6 +429,7 @@ #define INTEL_CFL_IDS(info) \ INTEL_CFL_S_GT1_IDS(info), \ INTEL_CFL_S_GT2_IDS(info), \ + INTEL_CFL_H_GT1_IDS(info), \ INTEL_CFL_H_GT2_IDS(info), \ INTEL_CFL_U_GT2_IDS(info), \ INTEL_CFL_U_GT3_IDS(info), \ @@ -457,9 +461,13 @@ INTEL_VGA_DEVICE(0x8A51, info), \ INTEL_VGA_DEVICE(0x8A5C, info), \ INTEL_VGA_DEVICE(0x8A5D, info), \ + INTEL_VGA_DEVICE(0x8A59, info), \ + INTEL_VGA_DEVICE(0x8A58, info), \ INTEL_VGA_DEVICE(0x8A52, info), \ INTEL_VGA_DEVICE(0x8A5A, info), \ INTEL_VGA_DEVICE(0x8A5B, info), \ + INTEL_VGA_DEVICE(0x8A57, info), \ + INTEL_VGA_DEVICE(0x8A56, info), \ INTEL_VGA_DEVICE(0x8A71, info), \ INTEL_VGA_DEVICE(0x8A70, info) diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h index 3fc4854dce49..49d9cdfc58f2 100644 --- a/include/drm/ttm/ttm_bo_api.h +++ b/include/drm/ttm/ttm_bo_api.h @@ -296,23 +296,6 @@ static inline void ttm_bo_get(struct ttm_buffer_object *bo) } /** - * ttm_bo_reference - reference a struct ttm_buffer_object - * - * @bo: The buffer object. - * - * Returns a refcounted pointer to a buffer object. - * - * This function is deprecated. Use @ttm_bo_get instead. - */ - -static inline struct ttm_buffer_object * -ttm_bo_reference(struct ttm_buffer_object *bo) -{ - ttm_bo_get(bo); - return bo; -} - -/** * ttm_bo_get_unless_zero - reference a struct ttm_buffer_object unless * its refcount has already reached zero. * @bo: The buffer object. @@ -387,17 +370,6 @@ int ttm_bo_validate(struct ttm_buffer_object *bo, void ttm_bo_put(struct ttm_buffer_object *bo); /** - * ttm_bo_unref - * - * @bo: The buffer object. - * - * Unreference and clear a pointer to a buffer object. - * - * This function is deprecated. Use @ttm_bo_put instead. - */ -void ttm_bo_unref(struct ttm_buffer_object **bo); - -/** * ttm_bo_add_to_lru * * @bo: The buffer object. diff --git a/include/drm/ttm/ttm_bo_driver.h b/include/drm/ttm/ttm_bo_driver.h index 56642fbb6f73..cbf3180cb612 100644 --- a/include/drm/ttm/ttm_bo_driver.h +++ b/include/drm/ttm/ttm_bo_driver.h @@ -381,6 +381,15 @@ struct ttm_bo_driver { */ int (*access_memory)(struct ttm_buffer_object *bo, unsigned long offset, void *buf, int len, int write); + + /** + * struct ttm_bo_driver member del_from_lru_notify + * + * @bo: the buffer object deleted from lru + * + * notify driver that a BO was deleted from LRU. + */ + void (*del_from_lru_notify)(struct ttm_buffer_object *bo); }; /** diff --git a/include/trace/events/host1x.h b/include/trace/events/host1x.h index a37ef73092e5..3d340b6f1ea3 100644 --- a/include/trace/events/host1x.h +++ b/include/trace/events/host1x.h @@ -80,6 +80,32 @@ TRACE_EVENT(host1x_cdma_push, __entry->name, __entry->op1, __entry->op2) ); +TRACE_EVENT(host1x_cdma_push_wide, + TP_PROTO(const char *name, u32 op1, u32 op2, u32 op3, u32 op4), + + TP_ARGS(name, op1, op2, op3, op4), + + TP_STRUCT__entry( + __field(const char *, name) + __field(u32, op1) + __field(u32, op2) + __field(u32, op3) + __field(u32, op4) + ), + + TP_fast_assign( + __entry->name = name; + __entry->op1 = op1; + __entry->op2 = op2; + __entry->op3 = op3; + __entry->op4 = op4; + ), + + TP_printk("name=%s, op1=%08x, op2=%08x, op3=%08x op4=%08x", + __entry->name, __entry->op1, __entry->op2, __entry->op3, + __entry->op4) +); + TRACE_EVENT(host1x_cdma_push_gather, TP_PROTO(const char *name, struct host1x_bo *bo, u32 words, u32 offset, void *cmdbuf), diff --git a/include/uapi/drm/amdgpu_drm.h b/include/uapi/drm/amdgpu_drm.h index be84e43c1e19..2acbc8bc465b 100644 --- a/include/uapi/drm/amdgpu_drm.h +++ b/include/uapi/drm/amdgpu_drm.h @@ -523,6 +523,7 @@ struct drm_amdgpu_gem_va { #define AMDGPU_CHUNK_ID_SYNCOBJ_IN 0x04 #define AMDGPU_CHUNK_ID_SYNCOBJ_OUT 0x05 #define AMDGPU_CHUNK_ID_BO_HANDLES 0x06 +#define AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES 0x07 struct drm_amdgpu_cs_chunk { __u32 chunk_id; @@ -565,6 +566,11 @@ union drm_amdgpu_cs { * caches (L2/vL1/sL1/I$). */ #define AMDGPU_IB_FLAG_TC_WB_NOT_INVALIDATE (1 << 3) +/* Set GDS_COMPUTE_MAX_WAVE_ID = DEFAULT before PACKET3_INDIRECT_BUFFER. + * This will reset wave ID counters for the IB. + */ +#define AMDGPU_IB_FLAG_RESET_GDS_MAX_WAVE_ID (1 << 4) + struct drm_amdgpu_cs_chunk_ib { __u32 _pad; /** AMDGPU_IB_FLAG_* */ diff --git a/include/uapi/drm/drm_fourcc.h b/include/uapi/drm/drm_fourcc.h index 0fa360223255..bab20298f422 100644 --- a/include/uapi/drm/drm_fourcc.h +++ b/include/uapi/drm/drm_fourcc.h @@ -595,6 +595,9 @@ extern "C" { * AFBC has several features which may be supported and/or used, which are * represented using bits in the modifier. Not all combinations are valid, * and different devices or use-cases may support different combinations. + * + * Further information on the use of AFBC modifiers can be found in + * Documentation/gpu/afbc.rst */ #define DRM_FORMAT_MOD_ARM_AFBC(__afbc_mode) fourcc_mod_code(ARM, __afbc_mode) diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h index 298b2e197744..397810fa2d33 100644 --- a/include/uapi/drm/i915_drm.h +++ b/include/uapi/drm/i915_drm.h @@ -1486,9 +1486,73 @@ struct drm_i915_gem_context_param { #define I915_CONTEXT_MAX_USER_PRIORITY 1023 /* inclusive */ #define I915_CONTEXT_DEFAULT_PRIORITY 0 #define I915_CONTEXT_MIN_USER_PRIORITY -1023 /* inclusive */ + /* + * When using the following param, value should be a pointer to + * drm_i915_gem_context_param_sseu. + */ +#define I915_CONTEXT_PARAM_SSEU 0x7 __u64 value; }; +/** + * Context SSEU programming + * + * It may be necessary for either functional or performance reason to configure + * a context to run with a reduced number of SSEU (where SSEU stands for Slice/ + * Sub-slice/EU). + * + * This is done by configuring SSEU configuration using the below + * @struct drm_i915_gem_context_param_sseu for every supported engine which + * userspace intends to use. + * + * Not all GPUs or engines support this functionality in which case an error + * code -ENODEV will be returned. + * + * Also, flexibility of possible SSEU configuration permutations varies between + * GPU generations and software imposed limitations. Requesting such a + * combination will return an error code of -EINVAL. + * + * NOTE: When perf/OA is active the context's SSEU configuration is ignored in + * favour of a single global setting. + */ +struct drm_i915_gem_context_param_sseu { + /* + * Engine class & instance to be configured or queried. + */ + __u16 engine_class; + __u16 engine_instance; + + /* + * Unused for now. Must be cleared to zero. + */ + __u32 flags; + + /* + * Mask of slices to enable for the context. Valid values are a subset + * of the bitmask value returned for I915_PARAM_SLICE_MASK. + */ + __u64 slice_mask; + + /* + * Mask of subslices to enable for the context. Valid values are a + * subset of the bitmask value return by I915_PARAM_SUBSLICE_MASK. + */ + __u64 subslice_mask; + + /* + * Minimum/Maximum number of EUs to enable per subslice for the + * context. min_eus_per_subslice must be inferior or equal to + * max_eus_per_subslice. + */ + __u16 min_eus_per_subslice; + __u16 max_eus_per_subslice; + + /* + * Unused for now. Must be cleared to zero. + */ + __u32 rsvd; +}; + enum drm_i915_oa_format { I915_OA_FORMAT_A13 = 1, /* HSW only */ I915_OA_FORMAT_A29, /* HSW only */ |