| Age | Commit message (Collapse) | Author |
|---|
|
Existing driver checks for alternate clock if devm_clk_get() fails
and returns error code for last clock failure. If xilinx_uartps is
called before clock driver, devm_clk_get() returns -EPROBE_DEFER.
In this case, probe should not check for alternate clock as main
clock is already present in DTS and return -EPROBE_DEFER only.
This patch fixes it by not checking for alternate clock when main
clock get returns -EPROBE_DEFER.
Signed-off-by: Rajan Vaja <rajan.vaja@xilinx.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
There is no point in whitelisting a register that the user then cannot
write to, so check the register exists before merging such patches.
v2: Mark SLICE_COMMON_ECO_CHICKEN1 [731c] as write-only
v3: Use different variables for different meanings!
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dale B Stimson <dale.b.stimson@intel.com>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190301140404.26690-6-chris@chris-wilson.co.uk
Link: https://patchwork.freedesktop.org/patch/msgid/20190301160108.19039-1-chris@chris-wilson.co.uk
|
|
As we did lots of change at vim2m driver, let's take the
opportunity and make checkpatch happier, addressing the
errors/warnings that makes sense.
While here, increment driver's version.
No functional changes.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
There's no reason why this driver should use BUG(). Instead,
just properly handle issue, returning an error code where
pertinent.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Pull drm fixes from Dave Airlie:
"Three final fixes, one for a feature that is new in this kernel, one
bochs fix for qemu riscv and one atomic modesetting fix.
I've left a few of the other late fixes until next as I didn't want to
throw in anything that wasn't really necessary"
* tag 'drm-fixes-2019-03-01' of git://anongit.freedesktop.org/drm/drm:
drm/bochs: Fix the ID mismatch error
drm: Block fb changes for async plane updates
drm/amd/display: Use vrr friendly pageflip throttling in DC.
|
|
When in passthrough mode, copy the entire line at once, in
order to make it faster (if not HFLIP).
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Add an horizontal linear scaler using Breseham algorithm in
order to speep up its calculus.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Handling any Y,Cr,Cb formats require some extra logic, as it
handles a group of two pixels. That's easy while we don't do
horizontal scaling.
However, doing horizontal scaling with such formats would require
a lot more code, in order to avoid distortions, as, if it scales
to two non-consecutive points, the logic would need to read 4
points in order to properly convert to RGB.
As this is just a test driver, and we want fast algorithms,
let's just get rid of this format as an output one.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
When resolutions are different, the expected behavior is to
scale the image. Implement a vertical scaler as the first step.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
The vim2m driver doesn't enforce that the capture and output
buffers would have the same size. Do the right thing if the
buffers are different, zeroing the buffer before writing,
ensuring that lines will be aligned and it won't write past
the buffer area.
This is a temporary fix.
A proper fix is to either implement a simple scaler at vim2m,
or to better define the behaviour of M2M transform drivers
at V4L2 API with regards to its capability of scaling the
image or not.
In any case, such changes would deserve a separate patch
anyway, as it would imply on some behavoral change.
Also, as we have an actual bug of writing data at wrong
places, let's fix this here, and add a mental note that
we need to properly address it.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
The icl wm1+ underrun w/a has been added to the spec. It changed
slightly from the previous incarnation by requiring that we mirror
the lines watermark and the ignore lines bit from WM0 into WM1.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228173639.18422-1-ville.syrjala@linux.intel.com
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Tested-by: Clint Taylor <Clinton.A.Taylor@intel.com>
|
|
The only real restriction at vim2m is that width should be
multiple of two, as the copy routine always copy two pixels
each time.
However, Bayer formats are defined as having a 2x2 matrix.
So, odd vertical numbers would cause color distortions at the
last line. So, it makes sense to use step 2 for vertical alignment
on Bayer.
With this patch, the reported formats for video capture will
be:
[0]: 'RGBP' (16-bit RGB 5-6-5)
Size: Stepwise 32x32 - 640x480 with step 2/1
[1]: 'RGBR' (16-bit RGB 5-6-5 BE)
Size: Stepwise 32x32 - 640x480 with step 2/1
[2]: 'RGB3' (24-bit RGB 8-8-8)
Size: Stepwise 32x32 - 640x480 with step 2/1
[3]: 'BGR3' (24-bit BGR 8-8-8)
Size: Stepwise 32x32 - 640x480 with step 2/1
[4]: 'YUYV' (YUYV 4:2:2)
Size: Stepwise 32x32 - 640x480 with step 2/1
[5]: 'BA81' (8-bit Bayer BGBG/GRGR)
Size: Stepwise 32x32 - 640x480 with step 2/2
[6]: 'GBRG' (8-bit Bayer GBGB/RGRG)
Size: Stepwise 32x32 - 640x480 with step 2/2
[7]: 'GRBG' (8-bit Bayer GRGR/BGBG)
Size: Stepwise 32x32 - 640x480 with step 2/2
[8]: 'RGGB' (8-bit Bayer RGRG/GBGB)
Size: Stepwise 32x32 - 640x480 with step 2/2
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
As we do alignments for width, expose it via V4L2 API.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
The copy logic assumes that the data width is multiple of two,
as this is needed in order to support YUYV.
There's no reason to force it to be 8-pixel aligned, as 2-pixel
alignment is enough.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
1) Use two levels for debug:
- level 1: setup stuff
- level 2: add queue/dequeue messages
2) Better display the debug output, translating buffer
type, fourcc and making some messages clearer.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
The vim2m device is interesting to simulate a webcam. As most
sensors are arranged using bayer formats, the best is to support
to output data using those formats.
So, add support for them.
All 4 8-bit bayer formats tested with:
$ qvidcap -p &
$ v4l2-ctl --stream-mmap --stream-out-mmap --stream-to-host localhost --stream-lossless --stream-out-hor-speed 1 -v pixelformat=RGGB
It was tested also with GStreamer with:
$ gst-validate-1.0 filesrc location=some_video.mp4 ! qtdemux ! avdec_h264 ! videoconvert ! videoscale ! v4l2convert disable-passthrough=1 extra-controls="s,horizontal_flip=0,vertical_flip=0" ! bayer2rgb ! videoconvert ! xvimagesink
For all possible HFLIP/VFLIP values.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/hyperv/linux into char-misc-next
Sasha writes:
1. Exopsing counters for state changes of channel ring buffers; this is
useful to investigate performance issues. By Kimberly Brown.
2. Switching to the new generic UUID API, by Andy Shevchenko.
* tag 'hyperv-next-signed' of git://git.kernel.org/pub/scm/linux/kernel/git/hyperv/linux:
Drivers: hv: vmbus: Expose counters for interrupts and full conditions
vmbus: Switch to use new generic UUID API
|
|
The dasd_eckd_restore_device() function calls dasd_generic_read_dev_chars
with a temporary buffer on the stack. With CONFIG_VMAP_STACK=y this is
a vmalloc address but dasd_generic_restore_device() uses the address of
the buffer as I/O address. Circumvent this by using the already allocated
cqr->data buffer for the RDC data.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
|
|
Fix the following sparse warning:
drivers/pci/controller/pci-aardvark.c:469:28: warning:
symbol 'advk_pci_bridge_emul_ops' was not declared. Should it be static?
Fixes: 8a3ebd8de328 ("PCI: aardvark: Implement emulated root PCI bridge config space")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
|
|
Those typos were left over from codespell check, on
my first pass or belong to code added after the time I
ran it.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
A simple mutex used for guarding the flow of requests in and out of the
timeline. In the short-term, it will be used only to guard the addition
of requests into the timeline, taken on alloc and released on commit so
that only one caller can construct a request into the timeline
(important as the seqno and ring pointers must be serialised). This will
be used by observers to ensure that the seqno/hwsp is stable. Later,
when we have reduced retiring to only operate on a single timeline at a
time, we can then use the mutex as the sole guard required for retiring.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190301110547.14758-2-chris@chris-wilson.co.uk
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Commit 60d2fa0dad06 ("fbdev: omap2: no need to check return value of
debugfs_create functions") changed the declaration of the return value
of function dss_debugfs_create_file() and the following two warnings
appeared:
drivers/video/fbdev/omap2/omapfb/dss/core.c: In function ‘dss_debugfs_create_file’:
drivers/video/fbdev/omap2/omapfb/dss/core.c:139:9: warning: ‘return’ with a value,
in function returning void
return 0;
^
drivers/video/fbdev/omap2/omapfb/dss/core.c: In function ‘omap_dss_probe’:
drivers/video/fbdev/omap2/omapfb/dss/core.c:172:6: warning: unused variable ‘r’
[-Wunused-variable]
int r;
^
Rework so function dss_debugfs_create_file() that is declared to return
void don't 'return 0' and remove the declaration of the unused variable
'r'.
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Acked-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Acked-by: Andrzej Pietrasiewicz <andrzejtp2010@gmail.com>
Reviewed-by: Benoit Parrot <bparrot@ti.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Reviewed-by: Lad, Prabhakar <prabhakar.csengg@gmail.com>
Acked-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Houlong Wei <houlong.wei@mediatek.com>
Reviewed-by: Yong Deng <yong.deng@magewell.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
VI planes support coarse scaling which helps to overcome VI scaler
limitations. While exact working of coarse scaling isn't known, it seems
that it just skips programmed amount of rows and columns. This is
especially useful for downscaling very big planes (4K down to 1080p).
Horizontal coarse scaling is currently used to fit one line to VI scaler
buffer.
Vertical coarse scaling is used to assure that VI scaler is actually
capable of processing framebuffer in one frame time.
Signed-off-by: Jernej Skrabec <jernej.skrabec@siol.net>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228200329.11128-4-jernej.skrabec@siol.net
|
|
While all RGB scalers have maximum line size of 2048, some YUV scalers
have maximum line size of 2048 and some have line size of 4096.
Since there is no rule for that, add a quirk.
Signed-off-by: Jernej Skrabec <jernej.skrabec@siol.net>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228200329.11128-3-jernej.skrabec@siol.net
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Use codespell to fix lots of typos over frontends.
Manually verified to avoid false-positives.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Sugaya Taichi <sugaya.taichi@socionext.com> explains:
Here is the series of patches the initial support for SC2000(M10V) of
Milbeaut SoCs. "M10V" is the internal name of SC2000, so commonly used in
source code.
SC2000 is a SoC of the Milbeaut series. equipped with a DSP optimized for
computer vision. It also features advanced functionalities such as 360-degree,
real-time spherical stitching with multi cameras, image stabilization for
without mechanical gimbals, and rolling shutter correction. More detail is
below:
https://www.socionext.com/en/products/assp/milbeaut/SC2000.html
Specifications for developers are below:
- Quad-core 32bit Cortex-A7 on ARMv7-A architecture
- NEON support
- DSP
- GPU
- MAX 3GB DDR3
- Cortex-M0 for power control
- NAND Flash Interface
- SD UHS-I
- SD UHS-II
- SDIO
- USB2.0 HOST / Device
- USB3.0 HOST / Device
- PCI express Gen2
- Ethernet Engine
- I2C
- UART
- SPI
- PWM
Support is quite minimal for now, since it only includes timer, clock,
pictrl and serial controller drivers, so we can only boot to userspace
through initramfs. Support for the other peripherals will come eventually.
* milbeaut/newsoc:
ARM: multi_v7_defconfig: add ARCH_MILBEAUT and ARCH_MILBEAUT_M10V
ARM: configs: Add Milbeaut M10V defconfig
ARM: dts: milbeaut: Add device tree set for the Milbeaut M10V board
clocksource/drivers/timer-milbeaut: Introduce timer for Milbeaut SoCs
dt-bindings: timer: Add Milbeaut M10V timer description
ARM: milbeaut: Add basic support for Milbeaut m10v SoC
dt-bindings: Add documentation for Milbeaut SoCs
dt-bindings: arm: Add SMP enable-method for Milbeaut
dt-bindings: sram: milbeaut: Add binding for Milbeaut smp-sram
Link: https://lore.kernel.org/linux-arm-kernel/1551243056-10521-1-git-send-email-sugaya.taichi@socionext.com/
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
Remove code that is not called anywhere, just
remove the internal ops.
Signed-off-by: Rui Miguel Silva <rui.silva@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
When an rc device is created, we do not know what key codes it will
support, since a new keymap might be loaded at some later point. So,
we set all keybit in the input device.
The uevent for the input device includes all the key codes, of which
there are now 768. This overflows the size of the uevent
(UEVENT_BUFFER_SIZE) and no event is generated.
Revert for now until we figure out a different solution.
This reverts commit fec225a04330d0f222d24feb5bea045526031675.
Cc: <stable@vger.kernel.org> # 4.20+
Reported-by: Christian Holpert <christian@holpert.de>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
|
|
Add timer driver for Milbeaut SoCs series.
The timer has two 32-bit width down counters, one of which is configured
as a clockevent device and the other is configured as a clock source.
Signed-off-by: Sugaya Taichi <sugaya.taichi@socionext.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/leo/linux into arm/drivers
NXP/FSL SoC driver updates for v5.1 take4
DPIO driver
- Add support for cache stashing and enable it in dpaa2-eth driver
GUTS driver
- Make fsl_guts_get_svr() API internal in favor of more generic
soc_device_match()
* tag 'soc-fsl-next-v5.1-4' of git://git.kernel.org/pub/scm/linux/kernel/git/leo/linux:
dpaa2-eth: configure the cache stashing amount on a queue
soc: fsl: dpio: configure cache stashing destination
soc: fsl: dpio: enable frame data cache stashing per software portal
soc: fsl: guts: make fsl_guts_get_svr() static
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
https://git.linaro.org/people/jens.wiklander/linux-tee into arm/drivers
OP-TEE driver
- dual license for optee_msg.h and optee_smc.h
Generic
- add cancellation support to client interface
* tag 'tee-misc-for-v5.1' of https://git.linaro.org/people/jens.wiklander/linux-tee:
tee: optee: update optee_msg.h and optee_smc.h to dual license
tee: add cancellation support to client interface
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
https://git.linaro.org/people/jens.wiklander/linux-tee into arm/fixes
OP-TEE driver
- add missing of_node_put after of_device_is_available
* tag 'tee-fix-for-v5.0' of https://git.linaro.org/people/jens.wiklander/linux-tee:
tee: optee: add missing of_node_put after of_device_is_available
|
|
This patch fix compilation error in 32-bit ARM architecture regarding
division of 2 64-bit variables.
Use the kernel do_div() macro, which is implemented per architecture, for
doing these divisions instead of using the / operator.
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Since 7c5925afbc58 (PCI: dwc: Move MSI IRQs allocation to IRQ domains
hierarchical API) the MSI init claims one of the controller IRQs as a
chained IRQ line for the MSI controller. On some designs, like the i.MX6,
this line is shared with a PCIe legacy IRQ. When the line is claimed for
the MSI domain, any device trying to use this legacy IRQs will fail to
request this IRQ line.
As MSI and legacy IRQs are already mutually exclusive on the DWC core,
as the core won't forward any legacy IRQs once any MSI has been enabled,
users wishing to use legacy IRQs already need to explictly disable MSI
support (usually via the pci=nomsi kernel commandline option). To avoid
any issues with MSI conflicting with legacy IRQs, just skip all of the
DWC MSI initalization, including the IRQ line claim, when MSI is disabled.
Fixes: 7c5925afbc58 ("PCI: dwc: Move MSI IRQs allocation to IRQ domains hierarchical API")
Tested-by: Tim Harvey <tharvey@gateworks.com>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
Cc: stable@vger.kernel.org
|
|
Remove the duplicate implementation of cpumask_to_vpset() and use the
shared implementation. Export hv_max_vp_index, which is required by
cpumask_to_vpset().
Signed-off-by: Maya Nakamura <m.maya.nakamura@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Tested-by: Vitaly Kuznetsov <vkuznets@redhat.com>
|
|
Remove a duplicate definition of VP set (hv_vp_set) and use the common
definition (hv_vpset) that is used in other places.
Change the order of the members in struct hv_pcibus_device so that the
declaration of retarget_msi_interrupt_params is the last member. Struct
hv_vpset, which contains a flexible array, is nested two levels deep in
struct hv_pcibus_device via retarget_msi_interrupt_params.
Add a comment that retarget_msi_interrupt_params should be the last
member of struct hv_pcibus_device.
Signed-off-by: Maya Nakamura <m.maya.nakamura@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Tested-by: Vitaly Kuznetsov <vkuznets@redhat.com>
|
|
Because Hyper-V requires that hypercall arguments be aligned on an 8
byte boundary, add __aligned(8) to struct retarget_msi_interrupt.
Link: https://lore.kernel.org/lkml/87k1hlqlby.fsf@vitty.brq.redhat.com/
Signed-off-by: Maya Nakamura <m.maya.nakamura@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
|
|
The PCIE_AXI_WINDOW0 register defines the inbound window size for
requests coming from PCI endpoints. Requests outside of this window will
be treated as unsupported requests.
Enlarge this window size from 2^31 to 2^33 to support a 8GB address
space (which gives endpoints DMA access to full 4GB DRAM address range
- physical DRAM starts at 0x40000000).
Reported-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Honghui Zhang <honghui.zhang@mediatek.com>
[lorenzo.pieralisi@arm.com: updated commit log]
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
|
|
Mediatek's HW assigns a MMIO address range (typically starts from
0x20000000 to 0x2fffffff for both mt2712 and mt7622) for PCI usage.
This MMIO address space represents the address space that can
be allocated to PCI devices through Base Address Registers.
Even though the full MMIO address range is available to be allocated, it
should be enabled by the PCIE_AHB_TRANS_BASE register in the host
controller and the size that is enabled is determined by AHB2PCIE_SIZE
bits in this register.
Owing to a bug in the MMIO window size computation, current code does
not enable the full size of the available MMIO address range in the
PCI host controller; if the PCI devices BARs requested size exceeds the
size enabled through the PCIE_AHB_TRANS_BASE register the requests
targeting the disabled address address space will be blocked by the root
complex causing a system error.
Existing code has never run into a system error in production because
even half of the enabled MMIO range (128MB) is big enough for typical
devices BAR requests (4MB) but the full MMIO address range should
be enabled regardless.
Fix the MMIO window size computation by using resource_size(mem) instead
of mem->end - mem->start.
Since the MMIO window size for both MT2712 and MT7622 is 0x10000000,
this change will update the parameter passed to fls() from 0xfffffff to
0x10000000 and calculate the whole memory mapped IO range size
correctly.
Detected through coccinelle semantic patch (and related warning):
scripts/coccinelle/api/resource_size.cocci:
pcie-mediatek.c:720:13-16: WARNING: Suspicious code. resource_size is maybe missing with mem
Signed-off-by: Honghui Zhang <honghui.zhang@mediatek.com>
[lorenzo.pieralisi@arm.com: rewrote the commit log]
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
|
