cachepc-linux

Fork of AMDESE/linux with modifications for CachePC side-channel attack
git clone https://git.sinitax.com/sinitax/cachepc-linux
Log | Files | Refs | README | LICENSE | sfeed.txt

vkms.rst (7337B)


      1.. _vkms:
      2
      3==========================================
      4 drm/vkms Virtual Kernel Modesetting
      5==========================================
      6
      7.. kernel-doc:: drivers/gpu/drm/vkms/vkms_drv.c
      8   :doc: vkms (Virtual Kernel Modesetting)
      9
     10Setup
     11=====
     12
     13The VKMS driver can be setup with the following steps:
     14
     15To check if VKMS is loaded, run::
     16
     17  lsmod | grep vkms
     18
     19This should list the VKMS driver. If no output is obtained, then
     20you need to enable and/or load the VKMS driver.
     21Ensure that the VKMS driver has been set as a loadable module in your
     22kernel config file. Do::
     23
     24  make nconfig
     25
     26  Go to `Device Drivers> Graphics support`
     27
     28  Enable `Virtual KMS (EXPERIMENTAL)`
     29
     30Compile and build the kernel for the changes to get reflected.
     31Now, to load the driver, use::
     32
     33  sudo modprobe vkms
     34
     35On running the lsmod command now, the VKMS driver will appear listed.
     36You can also observe the driver being loaded in the dmesg logs.
     37
     38The VKMS driver has optional features to simulate different kinds of hardware,
     39which are exposed as module options. You can use the `modinfo` command
     40to see the module options for vkms::
     41
     42  modinfo vkms
     43
     44Module options are helpful when testing, and enabling modules
     45can be done while loading vkms. For example, to load vkms with cursor enabled,
     46use::
     47
     48  sudo modprobe vkms enable_cursor=1
     49
     50To disable the driver, use ::
     51
     52  sudo modprobe -r vkms
     53
     54Testing With IGT
     55================
     56
     57The IGT GPU Tools is a test suite used specifically for debugging and
     58development of the DRM drivers.
     59The IGT Tools can be installed from
     60`here <https://gitlab.freedesktop.org/drm/igt-gpu-tools>`_ .
     61
     62The tests need to be run without a compositor, so you need to switch to text
     63only mode. You can do this by::
     64
     65  sudo systemctl isolate multi-user.target
     66
     67To return to graphical mode, do::
     68
     69  sudo systemctl isolate graphical.target
     70
     71Once you are in text only mode, you can run tests using the --device switch
     72or IGT_DEVICE variable to specify the device filter for the driver we want
     73to test. IGT_DEVICE can also be used with the run-test.sh script to run the
     74tests for a specific driver::
     75
     76  sudo ./build/tests/<name of test> --device "sys:/sys/devices/platform/vkms"
     77  sudo IGT_DEVICE="sys:/sys/devices/platform/vkms" ./build/tests/<name of test>
     78  sudo IGT_DEVICE="sys:/sys/devices/platform/vkms" ./scripts/run-tests.sh -t <name of test>
     79
     80For example, to test the functionality of the writeback library,
     81we can run the kms_writeback test::
     82
     83  sudo ./build/tests/kms_writeback --device "sys:/sys/devices/platform/vkms"
     84  sudo IGT_DEVICE="sys:/sys/devices/platform/vkms" ./build/tests/kms_writeback
     85  sudo IGT_DEVICE="sys:/sys/devices/platform/vkms" ./scripts/run-tests.sh -t kms_writeback
     86
     87You can also run subtests if you do not want to run the entire test::
     88
     89  sudo ./build/tests/kms_flip --run-subtest basic-plain-flip --device "sys:/sys/devices/platform/vkms"
     90  sudo IGT_DEVICE="sys:/sys/devices/platform/vkms" ./build/tests/kms_flip --run-subtest basic-plain-flip
     91
     92TODO
     93====
     94
     95If you want to do any of the items listed below, please share your interest
     96with VKMS maintainers.
     97
     98IGT better support
     99------------------
    100
    101Debugging:
    102
    103- kms_plane: some test cases are failing due to timeout on capturing CRC;
    104
    105- kms_flip: when running test cases in sequence, some successful individual
    106  test cases are failing randomly; when individually, some successful test
    107  cases display in the log the following error::
    108
    109  [drm:vkms_prepare_fb [vkms]] ERROR vmap failed: -4
    110
    111Virtual hardware (vblank-less) mode:
    112
    113- VKMS already has support for vblanks simulated via hrtimers, which can be
    114  tested with kms_flip test; in some way, we can say that VKMS already mimics
    115  the real hardware vblank. However, we also have virtual hardware that does
    116  not support vblank interrupt and completes page_flip events right away; in
    117  this case, compositor developers may end up creating a busy loop on virtual
    118  hardware. It would be useful to support Virtual Hardware behavior in VKMS
    119  because this can help compositor developers to test their features in
    120  multiple scenarios.
    121
    122Add Plane Features
    123------------------
    124
    125There's lots of plane features we could add support for:
    126
    127- Clearing primary plane: clear primary plane before plane composition (at the
    128  start) for correctness of pixel blend ops. It also guarantees alpha channel
    129  is cleared in the target buffer for stable crc. [Good to get started]
    130
    131- ARGB format on primary plane: blend the primary plane into background with
    132  translucent alpha.
    133
    134- Support when the primary plane isn't exactly matching the output size: blend
    135  the primary plane into the black background.
    136
    137- Full alpha blending on all planes.
    138
    139- Rotation, scaling.
    140
    141- Additional buffer formats, especially YUV formats for video like NV12.
    142  Low/high bpp RGB formats would also be interesting.
    143
    144- Async updates (currently only possible on cursor plane using the legacy
    145  cursor api).
    146
    147For all of these, we also want to review the igt test coverage and make sure
    148all relevant igt testcases work on vkms. They are good options for internship
    149project.
    150
    151Runtime Configuration
    152---------------------
    153
    154We want to be able to reconfigure vkms instance without having to reload the
    155module. Use/Test-cases:
    156
    157- Hotplug/hotremove connectors on the fly (to be able to test DP MST handling
    158  of compositors).
    159
    160- Configure planes/crtcs/connectors (we'd need some code to have more than 1 of
    161  them first).
    162
    163- Change output configuration: Plug/unplug screens, change EDID, allow changing
    164  the refresh rate.
    165
    166The currently proposed solution is to expose vkms configuration through
    167configfs. All existing module options should be supported through configfs
    168too.
    169
    170Writeback support
    171-----------------
    172
    173- The writeback and CRC capture operations share the use of composer_enabled
    174  boolean to ensure vblanks. Probably, when these operations work together,
    175  composer_enabled needs to refcounting the composer state to proper work.
    176  [Good to get started]
    177
    178- Add support for cloned writeback outputs and related test cases using a
    179  cloned output in the IGT kms_writeback.
    180
    181- As a v4l device. This is useful for debugging compositors on special vkms
    182  configurations, so that developers see what's really going on.
    183
    184Output Features
    185---------------
    186
    187- Variable refresh rate/freesync support. This probably needs prime buffer
    188  sharing support, so that we can use vgem fences to simulate rendering in
    189  testing. Also needs support to specify the EDID.
    190
    191- Add support for link status, so that compositors can validate their runtime
    192  fallbacks when e.g. a Display Port link goes bad.
    193
    194CRC API Improvements
    195--------------------
    196
    197- Optimize CRC computation ``compute_crc()`` and plane blending ``blend()``
    198
    199Atomic Check using eBPF
    200-----------------------
    201
    202Atomic drivers have lots of restrictions which are not exposed to userspace in
    203any explicit form through e.g. possible property values. Userspace can only
    204inquiry about these limits through the atomic IOCTL, possibly using the
    205TEST_ONLY flag. Trying to add configurable code for all these limits, to allow
    206compositors to be tested against them, would be rather futile exercise. Instead
    207we could add support for eBPF to validate any kind of atomic state, and
    208implement a library of different restrictions.
    209
    210This needs a bunch of features (plane compositing, multiple outputs, ...)
    211enabled already to make sense.