diff options
Diffstat (limited to 'Documentation/hid')
-rw-r--r-- | Documentation/hid/hid-bpf.rst | 173 | ||||
-rw-r--r-- | Documentation/hid/index.rst | 1 | ||||
-rw-r--r-- | Documentation/hid/intel-ish-hid.rst | 29 | ||||
-rw-r--r-- | Documentation/hid/intel-thc-hid.rst | 568 |
4 files changed, 687 insertions, 84 deletions
diff --git a/Documentation/hid/hid-bpf.rst b/Documentation/hid/hid-bpf.rst index 0765b3298ecf..5939eeafb361 100644 --- a/Documentation/hid/hid-bpf.rst +++ b/Documentation/hid/hid-bpf.rst @@ -129,19 +129,37 @@ When a BPF program needs to emit input events, it needs to talk with the HID protocol, and rely on the HID kernel processing to translate the HID data into input events. +In-tree HID-BPF programs and ``udev-hid-bpf`` +============================================= + +Official device fixes are shipped in the kernel tree as source in the +``drivers/hid/bpf/progs`` directory. This allows to add selftests to them in +``tools/testing/selftests/hid``. + +However, the compilation of these objects is not part of a regular kernel compilation +given that they need an external tool to be loaded. This tool is currently +`udev-hid-bpf <https://libevdev.pages.freedesktop.org/udev-hid-bpf/index.html>`_. + +For convenience, that external repository duplicates the files from here in +``drivers/hid/bpf/progs`` into its own ``src/bpf/stable`` directory. This allows +distributions to not have to pull the entire kernel source tree to ship and package +those HID-BPF fixes. ``udev-hid-bpf`` also has capabilities of handling multiple +objects files depending on the kernel the user is running. + Available types of programs =========================== -HID-BPF is built "on top" of BPF, meaning that we use tracing method to +HID-BPF is built "on top" of BPF, meaning that we use bpf struct_ops method to declare our programs. HID-BPF has the following attachment types available: -1. event processing/filtering with ``SEC("fmod_ret/hid_bpf_device_event")`` in libbpf +1. event processing/filtering with ``SEC("struct_ops/hid_device_event")`` in libbpf 2. actions coming from userspace with ``SEC("syscall")`` in libbpf -3. change of the report descriptor with ``SEC("fmod_ret/hid_bpf_rdesc_fixup")`` in libbpf +3. change of the report descriptor with ``SEC("struct_ops/hid_rdesc_fixup")`` or + ``SEC("struct_ops.s/hid_rdesc_fixup")`` in libbpf -A ``hid_bpf_device_event`` is calling a BPF program when an event is received from +A ``hid_device_event`` is calling a BPF program when an event is received from the device. Thus we are in IRQ context and can act on the data or notify userspace. And given that we are in IRQ context, we can not talk back to the device. @@ -149,37 +167,42 @@ A ``syscall`` means that userspace called the syscall ``BPF_PROG_RUN`` facility. This time, we can do any operations allowed by HID-BPF, and talking to the device is allowed. -Last, ``hid_bpf_rdesc_fixup`` is different from the others as there can be only one +Last, ``hid_rdesc_fixup`` is different from the others as there can be only one BPF program of this type. This is called on ``probe`` from the driver and allows to -change the report descriptor from the BPF program. Once a ``hid_bpf_rdesc_fixup`` +change the report descriptor from the BPF program. Once a ``hid_rdesc_fixup`` program has been loaded, it is not possible to overwrite it unless the program which inserted it allows us by pinning the program and closing all of its fds pointing to it. +Note that ``hid_rdesc_fixup`` can be declared as sleepable (``SEC("struct_ops.s/hid_rdesc_fixup")``). + + Developer API: ============== -User API data structures available in programs: ------------------------------------------------ +Available ``struct_ops`` for HID-BPF: +------------------------------------- .. kernel-doc:: include/linux/hid_bpf.h + :identifiers: hid_bpf_ops -Available tracing functions to attach a HID-BPF program: --------------------------------------------------------- -.. kernel-doc:: drivers/hid/bpf/hid_bpf_dispatch.c - :functions: hid_bpf_device_event hid_bpf_rdesc_fixup +User API data structures available in programs: +----------------------------------------------- -Available API that can be used in all HID-BPF programs: -------------------------------------------------------- +.. kernel-doc:: include/linux/hid_bpf.h + :identifiers: hid_bpf_ctx + +Available API that can be used in all HID-BPF struct_ops programs: +------------------------------------------------------------------ .. kernel-doc:: drivers/hid/bpf/hid_bpf_dispatch.c - :functions: hid_bpf_get_data + :identifiers: hid_bpf_get_data -Available API that can be used in syscall HID-BPF programs: ------------------------------------------------------------ +Available API that can be used in syscall HID-BPF programs or in sleepable HID-BPF struct_ops programs: +------------------------------------------------------------------------------------------------------- .. kernel-doc:: drivers/hid/bpf/hid_bpf_dispatch.c - :functions: hid_bpf_attach_prog hid_bpf_hw_request hid_bpf_hw_output_report hid_bpf_input_report hid_bpf_allocate_context hid_bpf_release_context + :identifiers: hid_bpf_hw_request hid_bpf_hw_output_report hid_bpf_input_report hid_bpf_try_input_report hid_bpf_allocate_context hid_bpf_release_context General overview of a HID-BPF program ===================================== @@ -222,20 +245,21 @@ This allows the following: Effect of a HID-BPF program --------------------------- -For all HID-BPF attachment types except for :c:func:`hid_bpf_rdesc_fixup`, several eBPF -programs can be attached to the same device. +For all HID-BPF attachment types except for :c:func:`hid_rdesc_fixup`, several eBPF +programs can be attached to the same device. If a HID-BPF struct_ops has a +:c:func:`hid_rdesc_fixup` while another is already attached to the device, the +kernel will return `-EINVAL` when attaching the struct_ops. -Unless ``HID_BPF_FLAG_INSERT_HEAD`` is added to the flags while attaching the -program, the new program is appended at the end of the list. -``HID_BPF_FLAG_INSERT_HEAD`` will insert the new program at the beginning of the -list which is useful for e.g. tracing where we need to get the unprocessed events -from the device. +Unless ``BPF_F_BEFORE`` is added to the flags while attaching the program, the new +program is appended at the end of the list. +``BPF_F_BEFORE`` will insert the new program at the beginning of the list which is +useful for e.g. tracing where we need to get the unprocessed events from the device. -Note that if there are multiple programs using the ``HID_BPF_FLAG_INSERT_HEAD`` flag, +Note that if there are multiple programs using the ``BPF_F_BEFORE`` flag, only the most recently loaded one is actually the first in the list. -``SEC("fmod_ret/hid_bpf_device_event")`` -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +``SEC("struct_ops/hid_device_event")`` +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Whenever a matching event is raised, the eBPF programs are called one after the other and are working on the same data buffer. @@ -258,17 +282,17 @@ with, userspace needs to refer to the device by its unique system id (the last 4 in the sysfs path: ``/sys/bus/hid/devices/xxxx:yyyy:zzzz:0000``). To retrieve a context associated with the device, the program must call -:c:func:`hid_bpf_allocate_context` and must release it with :c:func:`hid_bpf_release_context` +hid_bpf_allocate_context() and must release it with hid_bpf_release_context() before returning. Once the context is retrieved, one can also request a pointer to kernel memory with -:c:func:`hid_bpf_get_data`. This memory is big enough to support all input/output/feature +hid_bpf_get_data(). This memory is big enough to support all input/output/feature reports of the given device. -``SEC("fmod_ret/hid_bpf_rdesc_fixup")`` -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +``SEC("struct_ops/hid_rdesc_fixup")`` +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -The ``hid_bpf_rdesc_fixup`` program works in a similar manner to -``.report_fixup`` of ``struct hid_driver``. +The ``hid_rdesc_fixup`` program works in a similar manner to ``.report_fixup`` +of ``struct hid_driver``. When the device is probed, the kernel sets the data buffer of the context with the content of the report descriptor. The memory associated with that buffer is @@ -277,33 +301,31 @@ content of the report descriptor. The memory associated with that buffer is The eBPF program can modify the data buffer at-will and the kernel uses the modified content and size as the report descriptor. -Whenever a ``SEC("fmod_ret/hid_bpf_rdesc_fixup")`` program is attached (if no -program was attached before), the kernel immediately disconnects the HID device -and does a reprobe. +Whenever a struct_ops containing a ``SEC("struct_ops/hid_rdesc_fixup")`` program +is attached (if no program was attached before), the kernel immediately disconnects +the HID device and does a reprobe. -In the same way, when the ``SEC("fmod_ret/hid_bpf_rdesc_fixup")`` program is -detached, the kernel issues a disconnect on the device. +In the same way, when this struct_ops is detached, the kernel issues a disconnect +on the device. There is no ``detach`` facility in HID-BPF. Detaching a program happens when -all the user space file descriptors pointing at a program are closed. +all the user space file descriptors pointing at a HID-BPF struct_ops link are closed. Thus, if we need to replace a report descriptor fixup, some cooperation is required from the owner of the original report descriptor fixup. -The previous owner will likely pin the program in the bpffs, and we can then +The previous owner will likely pin the struct_ops link in the bpffs, and we can then replace it through normal bpf operations. Attaching a bpf program to a device =================================== -``libbpf`` does not export any helper to attach a HID-BPF program. -Users need to use a dedicated ``syscall`` program which will call -``hid_bpf_attach_prog(hid_id, program_fd, flags)``. +We now use standard struct_ops attachment through ``bpf_map__attach_struct_ops()``. +But given that we need to attach a struct_ops to a dedicated HID device, the caller +must set ``hid_id`` in the struct_ops map before loading the program in the kernel. ``hid_id`` is the unique system ID of the HID device (the last 4 numbers in the sysfs path: ``/sys/bus/hid/devices/xxxx:yyyy:zzzz:0000``) -``progam_fd`` is the opened file descriptor of the program to attach. - -``flags`` is of type ``enum hid_bpf_attach_flags``. +One can also set ``flags``, which is of type ``enum hid_bpf_attach_flags``. We can not rely on hidraw to bind a BPF program to a HID device. hidraw is an artefact of the processing of the HID device, and is not stable. Some drivers @@ -358,32 +380,15 @@ For that, we can create a basic skeleton for our BPF program:: extern __u8 *hid_bpf_get_data(struct hid_bpf_ctx *ctx, unsigned int offset, const size_t __sz) __ksym; - extern int hid_bpf_attach_prog(unsigned int hid_id, int prog_fd, u32 flags) __ksym; struct { __uint(type, BPF_MAP_TYPE_RINGBUF); __uint(max_entries, 4096 * 64); } ringbuf SEC(".maps"); - struct attach_prog_args { - int prog_fd; - unsigned int hid; - unsigned int flags; - int retval; - }; - - SEC("syscall") - int attach_prog(struct attach_prog_args *ctx) - { - ctx->retval = hid_bpf_attach_prog(ctx->hid, - ctx->prog_fd, - ctx->flags); - return 0; - } - __u8 current_value = 0; - SEC("?fmod_ret/hid_bpf_device_event") + SEC("struct_ops/hid_device_event") int BPF_PROG(filter_switch, struct hid_bpf_ctx *hid_ctx) { __u8 *data = hid_bpf_get_data(hid_ctx, 0 /* offset */, 192 /* size */); @@ -407,37 +412,37 @@ For that, we can create a basic skeleton for our BPF program:: return 0; } -To attach ``filter_switch``, userspace needs to call the ``attach_prog`` syscall -program first:: + SEC(".struct_ops.link") + struct hid_bpf_ops haptic_tablet = { + .hid_device_event = (void *)filter_switch, + }; + + +To attach ``haptic_tablet``, userspace needs to set ``hid_id`` first:: static int attach_filter(struct hid *hid_skel, int hid_id) { - int err, prog_fd; - int ret = -1; - struct attach_prog_args args = { - .hid = hid_id, - }; - DECLARE_LIBBPF_OPTS(bpf_test_run_opts, tattrs, - .ctx_in = &args, - .ctx_size_in = sizeof(args), - ); + int err, link_fd; - args.prog_fd = bpf_program__fd(hid_skel->progs.filter_switch); + hid_skel->struct_ops.haptic_tablet->hid_id = hid_id; + err = hid__load(skel); + if (err) + return err; - prog_fd = bpf_program__fd(hid_skel->progs.attach_prog); - - err = bpf_prog_test_run_opts(prog_fd, &tattrs); - if (err) - return err; + link_fd = bpf_map__attach_struct_ops(hid_skel->maps.haptic_tablet); + if (!link_fd) { + fprintf(stderr, "can not attach HID-BPF program: %m\n"); + return -1; + } - return args.retval; /* the fd of the created bpf_link */ + return link_fd; /* the fd of the created bpf_link */ } Our userspace program can now listen to notifications on the ring buffer, and is awaken only when the value changes. When the userspace program doesn't need to listen to events anymore, it can just -close the returned fd from :c:func:`attach_filter`, which will tell the kernel to +close the returned bpf link from :c:func:`attach_filter`, which will tell the kernel to detach the program from the HID device. Of course, in other use cases, the userspace program can also pin the fd to the diff --git a/Documentation/hid/index.rst b/Documentation/hid/index.rst index af02cf7cfa82..baf156b44b58 100644 --- a/Documentation/hid/index.rst +++ b/Documentation/hid/index.rst @@ -18,4 +18,5 @@ Human Interface Devices (HID) hid-alps intel-ish-hid + intel-thc-hid amd-sfh-hid diff --git a/Documentation/hid/intel-ish-hid.rst b/Documentation/hid/intel-ish-hid.rst index 55cbaa719a79..2adc174fb576 100644 --- a/Documentation/hid/intel-ish-hid.rst +++ b/Documentation/hid/intel-ish-hid.rst @@ -404,6 +404,35 @@ For more detailed information, please refer to the flow descriptions provided be | ISHTP Driver | | ISH Bootloader | +---------------+ +-----------------+ +Vendor Custom Firmware Loading +^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + +The firmware running inside ISH can be provided by Intel or developed by vendors using the Firmware Development Kit (FDK) provided by Intel. +Intel will upstream the Intel-built firmware to the ``linux-firmware.git`` repository, located under the path ``intel/ish/``. For the Lunar Lake platform, the Intel-built ISH firmware will be named ``ish_lnlm.bin``. +Vendors who wish to upstream their custom firmware should follow these guidelines for naming their firmware files: + +- The firmware filename should use one of the following patterns: + + - ``ish_${intel_plat_gen}_${SYS_VENDOR_CRC32}_${PRODUCT_NAME_CRC32}_${PRODUCT_SKU_CRC32}.bin`` + - ``ish_${intel_plat_gen}_${SYS_VENDOR_CRC32}_${PRODUCT_SKU_CRC32}.bin`` + - ``ish_${intel_plat_gen}_${SYS_VENDOR_CRC32}_${PRODUCT_NAME_CRC32}.bin`` + - ``ish_${intel_plat_gen}_${SYS_VENDOR_CRC32}.bin`` + +- ``${intel_plat_gen}`` indicates the Intel platform generation (e.g., ``lnlm`` for Lunar Lake) and must not exceed 8 characters in length. +- ``${SYS_VENDOR_CRC32}`` is the CRC32 checksum of the ``sys_vendor`` value from the DMI field ``DMI_SYS_VENDOR``. +- ``${PRODUCT_NAME_CRC32}`` is the CRC32 checksum of the ``product_name`` value from the DMI field ``DMI_PRODUCT_NAME``. +- ``${PRODUCT_SKU_CRC32}`` is the CRC32 checksum of the ``product_sku`` value from the DMI field ``DMI_PRODUCT_SKU``. + +During system boot, the ISH Linux driver will attempt to load the firmware in the following order, prioritizing custom firmware with more precise matching patterns: + +1. ``intel/ish/ish_${intel_plat_gen}_${SYS_VENDOR_CRC32}_${PRODUCT_NAME_CRC32}_${PRODUCT_SKU_CRC32}.bin`` +2. ``intel/ish/ish_${intel_plat_gen}_${SYS_VENDOR_CRC32}_${PRODUCT_SKU_CRC32}.bin`` +3. ``intel/ish/ish_${intel_plat_gen}_${SYS_VENDOR_CRC32}_${PRODUCT_NAME_CRC32}.bin`` +4. ``intel/ish/ish_${intel_plat_gen}_${SYS_VENDOR_CRC32}.bin`` +5. ``intel/ish/ish_${intel_plat_gen}.bin`` + +The driver will load the first matching firmware and skip the rest. If no matching firmware is found, it will proceed to the next pattern in the specified order. If all searches fail, the default Intel firmware, listed last in the order above, will be loaded. + ISH Debugging ------------- diff --git a/Documentation/hid/intel-thc-hid.rst b/Documentation/hid/intel-thc-hid.rst new file mode 100644 index 000000000000..dc9250787fc5 --- /dev/null +++ b/Documentation/hid/intel-thc-hid.rst @@ -0,0 +1,568 @@ +.. SPDX-License-Identifier: GPL-2.0 + +================================= +Intel Touch Host Controller (THC) +================================= + +Touch Host Controller is the name of the IP block in PCH that interface with Touch Devices (ex: +touchscreen, touchpad etc.). It is comprised of 3 key functional blocks: + +- A natively half-duplex Quad I/O capable SPI master +- Low latency I2C interface to support HIDI2C compliant devices +- A HW sequencer with RW DMA capability to system memory + +It has a single root space IOSF Primary interface that supports transactions to/from touch devices. +Host driver configures and controls the touch devices over THC interface. THC provides high +bandwidth DMA services to the touch driver and transfers the HID report to host system main memory. + +Hardware sequencer within the THC is responsible for transferring (via DMA) data from touch devices +into system memory. A ring buffer is used to avoid data loss due to asynchronous nature of data +consumption (by host) in relation to data production (by touch device via DMA). + +Unlike other common SPI/I2C controllers, THC handles the HID device data interrupt and reset +signals directly. + +1. Overview +=========== + +1.1 THC software/hardware stack +------------------------------- + +Below diagram illustrates the high-level architecture of THC software/hardware stack, which is fully +capable of supporting HIDSPI/HIDI2C protocol in Linux OS. + +:: + + ---------------------------------------------- + | +-----------------------------------+ | + | | Input Device | | + | +-----------------------------------+ | + | +-----------------------------------+ | + | | HID Multi-touch Driver | | + | +-----------------------------------+ | + | +-----------------------------------+ | + | | HID Core | | + | +-----------------------------------+ | + | +-----------------------------------+ | + | | THC QuickSPI/QuickI2C Driver | | + | +-----------------------------------+ | + | +-----------------------------------+ | + | | THC Hardware Driver | | + | +-----------------------------------+ | + | +----------------+ +----------------+ | + | SW | PCI Bus Driver | | ACPI Resource | | + | +----------------+ +----------------+ | + ---------------------------------------------- + ---------------------------------------------- + | +-----------------------------------+ | + | HW | PCI Bus | | + | +-----------------------------------+ | + | +-----------------------------------+ | + | | THC Controller | | + | +-----------------------------------+ | + | +-----------------------------------+ | + | | Touch IC | | + | +-----------------------------------+ | + ---------------------------------------------- + +Touch IC (TIC), also as known as the Touch devices (touchscreen or touchpad). The discrete analog +components that sense and transfer either discrete touch data or heatmap data in the form of HID +reports over the SPI/I2C bus to the THC Controller on the host. + +THC Host Controller, which is a PCI device HBA (host bus adapter), integrated into the PCH, that +serves as a bridge between the Touch ICs and the host. + +THC Hardware Driver, provides THC hardware operation APIs for above QuickSPI/QuickI2C driver, it +accesses THC MMIO registers to configure and control THC hardware. + +THC QuickSPI/QuickI2C driver, also as known as HIDSPI/HIDI2C driver, is registered as a HID +low-level driver that manages the THC Controller and implements HIDSPI/HIDI2C protocol. + + +1.2 THC hardware diagram +------------------------ +Below diagram shows THC hardware components:: + + --------------------------------- + | THC Controller | + | +---------------------------+ | + | | PCI Config Space | | + | +---------------------------+ | + | +---------------------------+ | + | + MMIO Registers | | + | +---------------------------+ | + +---------------+ | +------------+ +------------+ | + | System Memory +---+--+ DMA | | PIO | | + +---------------+ | +------------+ +------------+ | + | +---------------------------+ | + | | HW Sequencer | | + | +---------------------------+ | + | +------------+ +------------+ | + | | SPI/I2C | | GPIO | | + | | Controller | | Controller | | + | +------------+ +------------+ | + --------------------------------- + +As THC is exposed as a PCI devices, so it has standard PCI config space registers for PCI +enumeration and configuration. + +MMIO Registers, which provide registers access for driver to configure and control THC hardware, +the registers include several categories: Interrupt status and control, DMA configure, +PIO (Programmed I/O, defined in section 3.2) status and control, SPI bus configure, I2C subIP +status and control, reset status and control... + +THC provides two ways for driver to communicate with external Touch ICs: PIO and DMA. +PIO can let driver manually write/read data to/from Touch ICs, instead, THC DMA can +automatically write/read data without driver involved. + +HW Sequencer includes THC major logic, it gets instruction from MMIO registers to control +SPI bus and I2C bus to finish a bus data transaction, it also can automatically handle +Touch ICs interrupt and start DMA receive/send data from/to Touch ICs according to interrupt +type. That means THC HW Sequencer understands HIDSPI/HIDI2C transfer protocol, and handle +the communication without driver involved, what driver needs to do is just configure the THC +properly, and prepare the formatted data packet or handle received data packet. + +As THC supports HIDSPI/HIDI2C protocols, it has SPI controller and I2C subIP in it to expose +SPI bus and I2C bus. THC also integrates a GPIO controller to provide interrupt line support +and reset line support. + +2. THC Hardware Interface +========================= + +2.1 Host Interface +------------------ + +THC is exposed as "PCI Digitizer device" to the host. The PCI product and device IDs are +changed from different generations of processors. So the source code which enumerates drivers +needs to update from generation to generation. + + +2.2 Device Interface +-------------------- + +THC supports two types of bus for Touch IC connection: Enhanced SPI bus and I2C bus. + +2.2.1 SPI Port +~~~~~~~~~~~~~~ + +When PORT_TYPE = 00b in MMIO registers, THC uses SPI interfaces to communicate with external +Touch IC. THC enhanced SPI Bus supports different SPI modes: standard Single IO mode, +Dual IO mode and Quad IO mode. + +In Single IO mode, THC drives MOSI line to send data to Touch ICs, and receives data from Touch +ICs data from MISO line. In Dual IO mode, THC drivers MOSI and MISO both for data sending, and +also receives the data on both line. In Quad IO mode, there are other two lines (IO2 and IO3) +are added, THC drives MOSI (IO0), MISO (IO1), IO2 and IO3 at the same time for data sending, and +also receives the data on those 4 lines. Driver needs to configure THC in different mode by +setting different opcode. + +Beside IO mode, driver also needs to configure SPI bus speed. THC supports up to 42MHz SPI clock +on Intel Lunar Lake platform. + +For THC sending data to Touch IC, the data flow on SPI bus:: + + | --------------------THC sends---------------------------------| + <8Bits OPCode><24Bits Slave Address><Data><Data><Data>........... + +For THC receiving data from Touch IC, the data flow on SPI bus:: + + | ---------THC Sends---------------||-----Touch IC sends--------| + <8Bits OPCode><24Bits Slave Address><Data><Data><Data>........... + +2.2.2 I2C Port +~~~~~~~~~~~~~~ + +THC also integrates I2C controller in it, it's called I2C SubSystem. When PORT_TYPE = 01, THC +is configured to I2C mode. Comparing to SPI mode which can be configured through MMIO registers +directly, THC needs to use PIO read (by setting SubIP read opcode) to I2C subIP APB registers' +value and use PIO write (by setting SubIP write opcode) to do a write operation. + +2.2.3 GPIO interface +~~~~~~~~~~~~~~~~~~~~ + +THC also includes two GPIO pins, one for interrupt and the other for device reset control. + +Interrupt line can be configured to either level triggered or edge triggered by setting MMIO +Control register. + +Reset line is controlled by BIOS (or EFI) through ACPI _RST method, driver needs to call this +device ACPI _RST method to reset touch IC during initialization. + +3. High level concept +===================== + +3.1 Opcode +---------- + +Opcode (operation code) is used to tell THC or Touch IC what the operation will be, such as PIO +read or PIO write. + +When THC is configured to SPI mode, opcodes are used for determining the read/write IO mode. +There are some OPCode examples for SPI IO mode: + +======= ============================== +opcode Corresponding SPI command +======= ============================== +0x0B Read Single I/O +0x02 Write Single I/O +0xBB Read Dual I/O +0xB2 Write Dual I/O +0xEB Read Quad I/O +0xE2 Write Quad I/O +======= ============================== + +In general, different touch IC has different OPCode definition. According to HIDSPI +protocol whitepaper, those OPCodes are defined in device ACPI table, and driver needs to +query those information through OS ACPI APIs during driver initialization, then configures +THC MMIO OPCode registers with correct setting. + +When THC is working in I2C mode, opcodes are used to tell THC what's the next PIO type: +I2C SubIP APB register read, I2C SubIP APB register write, I2C touch IC device read, +I2C touch IC device write, I2C touch IC device write followed by read. + +Here are the THC pre-defined opcodes for I2C mode: + +======= =================================================== =========== +opcode Corresponding I2C command Address +======= =================================================== =========== +0x12 Read I2C SubIP APB internal registers 0h - FFh +0x13 Write I2C SubIP APB internal registers 0h - FFh +0x14 Read external Touch IC through I2C bus N/A +0x18 Write external Touch IC through I2C bus N/A +0x1C Write then read external Touch IC through I2C bus N/A +======= =================================================== =========== + +3.2 PIO +------- + +THC provides a programmed I/O (PIO) access interface for the driver to access the touch IC's +configuration registers, or access I2C subIP's configuration registers. To use PIO to perform +I/O operations, driver should pre-program PIO control registers and PIO data registers and kick +off the sequencing cycle. THC uses different PIO opcodes to distinguish different PIO +operations (PIO read/write/write followed by read). + +If there is a Sequencing Cycle In Progress and an attempt is made to program any of the control, +address, or data register the cycle is blocked and a sequence error will be encountered. + +A status bit indicates when the cycle has completed allowing the driver to know when read results +can be checked and/or when to initiate a new command. If enabled, the cycle done assertion can +interrupt driver with an interrupt. + +Because THC only has 16 FIFO registers for PIO, so all the data transfer through PIO shouldn't +exceed 64 bytes. + +As DMA needs max packet size for transferring configuration, and the max packet size information +always in HID device descriptor which needs THC driver to read it out from HID Device (Touch IC). +So PIO typical use case is, before DMA initialization, write RESET command (PIO write), read +RESET response (PIO read or PIO write followed by read), write Power ON command (PIO write), read +device descriptor (PIO read). + +For how to issue a PIO operation, here is the steps which driver needs follow: + +- Program read/write data size in THC_SS_BC. +- Program I/O target address in THC_SW_SEQ_DATA0_ADDR. +- If write, program the write data in THC_SW_SEQ_DATA0..THC_SW_SEQ_DATAn. +- Program the PIO opcode in THC_SS_CMD. +- Set TSSGO = 1 to start the PIO write sequence. +- If THC_SS_CD_IE = 1, SW will receives a MSI when the PIO is completed. +- If read, read out the data in THC_SW_SEQ_DATA0..THC_SW_SEQ_DATAn. + +3.3 DMA +------- + +THC has 4 DMA channels: Read DMA1, Read DMA2, Write DMA and Software DMA. + +3.3.1 Read DMA Channel +~~~~~~~~~~~~~~~~~~~~~~ + +THC has two Read DMA engines: 1st RxDMA (RxDMA1) and 2nd RxDMA (RxDMA2). RxDMA1 is reserved for +raw data mode. RxDMA2 is used for HID data mode and it is the RxDMA engine currently driver uses +for HID input report data retrieval. + +RxDMA's typical use case is auto receiving the data from Touch IC. Once RxDMA is enabled by +software, THC will start auto-handling receiving logic. + +For SPI mode, THC RxDMA sequence is: when Touch IC triggers a interrupt to THC, THC reads out +report header to identify what's the report type, and what's the report length, according to +above information, THC reads out report body to internal FIFO and start RxDMA coping the data +to system memory. After that, THC update interrupt cause register with report type, and update +RxDMA PRD table read pointer, then trigger a MSI interrupt to notify driver RxDMA finishing +data receiving. + +For I2C mode, THC RxDMA's behavior is a little bit different, because of HIDI2C protocol difference +with HIDSPI protocol, RxDMA only be used to receive input report. The sequence is, when Touch IC +triggers a interrupt to THC, THC first reads out 2 bytes from input report address to determine the +packet length, then use this packet length to start a DMA reading from input report address for +input report data. After that, THC update RxDMA PRD table read pointer, then trigger a MSI interrupt +to notify driver input report data is ready in system memory. + +All above sequence is hardware automatically handled, all driver needs to do is configure RxDMA and +waiting for interrupt ready then read out the data from system memory. + +3.3.2 Software DMA channel +~~~~~~~~~~~~~~~~~~~~~~~~~~ + +THC supports a software triggered RxDMA mode to read the touch data from touch IC. This SW RxDMA +is the 3rd THC RxDMA engine with the similar functionalities as the existing two RxDMAs, the only +difference is this SW RxDMA is triggered by software, and RxDMA2 is triggered by external Touch IC +interrupt. It gives a flexibility to software driver to use RxDMA read Touch IC data in any time. + +Before software starts a SW RxDMA, it shall stop the 1st and 2nd RxDMA, clear PRD read/write pointer +and quiesce the device interrupt (THC_DEVINT_QUIESCE_HW_STS = 1), other operations are the same with +RxDMA. + +3.3.3 Write DMA Channel +~~~~~~~~~~~~~~~~~~~~~~~ + +THC has one write DMA engine, which can be used for sending data to Touch IC automatically. +According to HIDSPI and HIDI2C protocol, every time only one command can be sent to touch IC, and +before last command is completely handled, next command cannot be sent, THC write DMA engine only +supports single PRD table. + +What driver needs to do is, preparing PRD table and DMA buffer, then copy data to DMA buffer and +update PRD table with buffer address and buffer length, then start write DMA. THC will +automatically send the data to touch IC, and trigger a DMA completion interrupt once transferring +is done. + +3.4 PRD +------- + +Physical Region Descriptor (PRD) provides the memory mapping description for THC DMAs. + +3.4.1 PRD table and entry +~~~~~~~~~~~~~~~~~~~~~~~~~ + +In order to improve physical DMA memory usage, modern drivers trend to allocate a virtually +contiguous, but physically fragmented buffer of memory for each data buffer. Linux OS also +provide SGL (scatter gather list) APIs to support this usage. + +THC uses PRD table (physical region descriptor) to support the corresponding OS kernel +SGL that describes the virtual to physical buffer mapping. + +:: + + ------------------------ -------------- -------------- + | PRD table base address +----+ PRD table #1 +-----+ PRD Entry #1 | + ------------------------ -------------- -------------- + -------------- + | PRD Entry #2 | + -------------- + -------------- + | PRD Entry #n | + -------------- + +The read DMA engine supports multiple PRD tables held within a circular buffer that allow the THC +to support multiple data buffers from the Touch IC. This allows host SW to arm the Read DMA engine +with multiple buffers, allowing the Touch IC to send multiple data frames to the THC without SW +interaction. This capability is required when the CPU processes touch frames slower than the +Touch IC can send them. + +To simplify the design, SW assumes worst-case memory fragmentation. Therefore,each PRD table shall +contain the same number of PRD entries, allowing for a global register (per Touch IC) to hold the +number of PRD-entries per PRD table. + +SW allocates up to 128 PRD tables per Read DMA engine as specified in the THC_M_PRT_RPRD_CNTRL.PCD +register field. The number of PRD tables should equal the number of data buffers. + +Max OS memory fragmentation will be at a 4KB boundary, thus to address 1MB of virtually contiguous +memory 256 PRD entries are required for a single PRD Table. SW writes the number of PRD entries +for each PRD table in the THC_M_PRT_RPRD_CNTRL.PTEC register field. The PRD entry's length must be +multiple of 4KB except for the last entry in a PRD table. + +SW allocates all the data buffers and PRD tables only once at host initialization. + +3.4.2 PRD Write pointer and read pointer +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +As PRD tables are organized as a Circular Buffer (CB), a read pointer and a write pointer for a CB +are needed. + +DMA HW consumes the PRD tables in the CB, one PRD entry at a time until the EOP bit is found set +in a PRD entry. At this point HW increments the PRD read pointer. Thus, the read pointer points +to the PRD which the DMA engine is currently processing. This pointer rolls over once the circular +buffer's depth has been traversed with bit[7] the Rollover bit. E.g. if the DMA CB depth is equal +to 4 entries (0011b), then the read pointers will follow this pattern (HW is required to honor +this behavior): 00h 01h 02h 03h 80h 81h 82h 83h 00h 01h ... + +The write pointer is updated by SW. The write pointer points to location in the DMA CB, where the +next PRD table is going to be stored. SW needs to ensure that this pointer rolls over once the +circular buffer's depth has been traversed with Bit[7] as the rollover bit. E.g. if the DMA CB +depth is equal to 5 entries (0100b), then the write pointers will follow this pattern (SW is +required to honor this behavior): 00h 01h 02h 03h 04h 80h 81h 82h 83h 84h 00h 01h .. + +3.4.3 PRD descriptor structure +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +Intel THC uses PRD entry descriptor for every PRD entry. Every PRD entry descriptor occupies +128 bits memories: + +=================== ======== =============================================== +struct field bit(s) description +=================== ======== =============================================== +dest_addr 53..0 destination memory address, as every entry + is 4KB, ignore lowest 10 bits of address. +reserved1 54..62 reserved +int_on_completion 63 completion interrupt enable bit, if this bit + set it means THC will trigger a completion + interrupt. This bit is set by SW driver. +len 87..64 how many bytes of data in this entry. +end_of_prd 88 end of PRD table bit, if this bit is set, + it means this entry is last entry in this PRD + table. This bit is set by SW driver. +hw_status 90..89 HW status bits +reserved2 127..91 reserved +=================== ======== =============================================== + +And one PRD table can include up to 256 PRD entries, as every entries is 4K bytes, so every +PRD table can describe 1M bytes memory. + +.. code-block:: c + + struct thc_prd_table { + struct thc_prd_entry entries[PRD_ENTRIES_NUM]; + }; + +In general, every PRD table means one HID touch data packet. Every DMA engine can support +up to 128 PRD tables (except write DMA, write DMA only has one PRD table). SW driver is responsible +to get max packet length from touch IC, and use this max packet length to create PRD entries for +each PRD table. + +4. HIDSPI support (QuickSPI) +============================ + +Intel THC is total compatible with HIDSPI protocol, THC HW sequenser can accelerate HIDSPI +protocol transferring. + +4.1 Reset Flow +-------------- + +- Call ACPI _RST method to reset Touch IC device. +- Read the reset response from TIC through PIO read. +- Issue a command to retrieve device descriptor from Touch IC through PIO write. +- Read the device descriptor from Touch IC through PIO read. +- If the device descriptor is valid, allocate DMA buffers and configure all DMA channels. +- Issue a command to retrieve report descriptor from Touch IC through DMA. + +4.2 Input Report Data Flow +-------------------------- + +Basic Flow: + +- Touch IC interrupts the THC Controller using an in-band THC interrupt. +- THC Sequencer reads the input report header by transmitting read approval as a signal + to the Touch IC to prepare for host to read from the device. +- THC Sequencer executes a Input Report Body Read operation corresponding to the value + reflected in “Input Report Length” field of the Input Report Header. +- THC DMA engine begins fetching data from the THC Sequencer and writes to host memory + at PRD entry 0 for the current CB PRD table entry. This process continues until the + THC Sequencer signals all data has been read or the THC DMA Read Engine reaches the + end of it's last PRD entry (or both). +- The THC Sequencer checks for the “Last Fragment Flag” bit in the Input Report Header. + If it is clear, the THC Sequencer enters an idle state. +- If the “Last Fragment Flag” bit is enabled the THC Sequencer enters End-of-Frame Processing. + +THC Sequencer End of Frame Processing: + +- THC DMA engine increments the read pointer of the Read PRD CB, sets EOF interrupt status + in RxDMA2 register (THC_M_PRT_READ_DMA_INT_STS_2). +- If THC EOF interrupt is enabled by the driver in the control register (THC_M_PRT_READ_DMA_CNTRL_2), + generates interrupt to software. + +Sequence of steps to read data from RX DMA buffer: + +- THC QuickSPI driver checks CB write Ptr and CB read Ptr to identify if any data frame in DMA + circular buffers. +- THC QuickSPI driver gets first unprocessed PRD table. +- THC QuickSPI driver scans all PRD entries in this PRD table to calculate the total frame size. +- THC QuickSPI driver copies all frame data out. +- THC QuickSPI driver checks the data type according to input report body, and calls related + callbacks to process the data. +- THC QuickSPI driver updates write Ptr. + +4.3 Output Report Data Flow +--------------------------- + +Generic Output Report Flow: + +- HID core calls raw_request callback with a request to THC QuickSPI driver. +- THC QuickSPI Driver converts request provided data into the output report packet and copies it + to THC's write DMA buffer. +- Start TxDMA to complete the write operation. + +5. HIDI2C support (QuickI2C) +============================ + +5.1 Reset Flow +-------------- + +- Read device descriptor from Touch IC device through PIO write followed by read. +- If the device descriptor is valid, allocate DMA buffers and configure all DMA channels. +- Use PIO or TxDMA to write a SET_POWER request to TIC's command register, and check if the + write operation is successfully completed. +- Use PIO or TxDMA to write a RESET request to TIC's command register. If the write operation + is successfully completed, wait for reset response from TIC. +- Use SWDMA to read report descriptor through TIC's report descriptor register. + +5.2 Input Report Data Flow +-------------------------- + +Basic Flow: + +- Touch IC asserts the interrupt indicating that it has an interrupt to send to HOST. + THC Sequencer issues a READ request over the I2C bus. The HIDI2C device returns the + first 2 bytes from the HIDI2C device which contains the length of the received data. +- THC Sequencer continues the Read operation as per the size of data indicated in the + length field. +- THC DMA engine begins fetching data from the THC Sequencer and writes to host memory + at PRD entry 0 for the current CB PRD table entry. THC writes 2Bytes for length field + plus the remaining data to RxDMA buffer. This process continues until the THC Sequencer + signals all data has been read or the THC DMA Read Engine reaches the end of it's last + PRD entry (or both). +- THC Sequencer enters End-of-Input Report Processing. +- If the device has no more input reports to send to the host, it de-asserts the interrupt + line. For any additional input reports, device keeps the interrupt line asserted and + steps 1 through 4 in the flow are repeated. + +THC Sequencer End of Input Report Processing: + +- THC DMA engine increments the read pointer of the Read PRD CB, sets EOF interrupt status + in RxDMA 2 register (THC_M_PRT_READ_DMA_INT_STS_2). +- If THC EOF interrupt is enabled by the driver in the control register + (THC_M_PRT_READ_DMA_CNTRL_2), generates interrupt to software. + +Sequence of steps to read data from RX DMA buffer: + +- THC QuickI2C driver checks CB write Ptr and CB read Ptr to identify if any data frame in DMA + circular buffers. +- THC QuickI2C driver gets first unprocessed PRD table. +- THC QuickI2C driver scans all PRD entries in this PRD table to calculate the total frame size. +- THC QuickI2C driver copies all frame data out. +- THC QuickI2C driver call hid_input_report to send the input report content to HID core, which + includes Report ID + Report Data Content (remove the length field from the original report + data). +- THC QuickI2C driver updates write Ptr. + +5.3 Output Report Data Flow +--------------------------- + +Generic Output Report Flow: + +- HID core call THC QuickI2C raw_request callback. +- THC QuickI2C uses PIO or TXDMA to write a SET_REPORT request to TIC's command register. Report + type in SET_REPORT should be set to Output. +- THC QuickI2C programs TxDMA buffer with TX Data to be written to TIC's data register. The first + 2 bytes should indicate the length of the report followed by the report contents including + Report ID. + +6. THC Debugging +================ + +To debug THC, event tracing mechanism is used. To enable debug logs:: + + echo 1 > /sys/kernel/debug/tracing/events/intel_thc/enable + cat /sys/kernel/debug/tracing/trace + +7. Reference +============ +- HIDSPI: https://download.microsoft.com/download/c/a/0/ca07aef3-3e10-4022-b1e9-c98cea99465d/HidSpiProtocolSpec.pdf +- HIDI2C: https://download.microsoft.com/download/7/d/d/7dd44bb7-2a7a-4505-ac1c-7227d3d96d5b/hid-over-i2c-protocol-spec-v1-0.docx |