tests/amdgpu: Fix unused function warning (v2)

This patch fixes:
-Wunused-function

v2: Always enable amdgpu_ras_test().

Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>

tests/amdgpu: Fix various warnings (v2)

This patch fixes the following warnings:
-Wformat=
-Wmaybe-uninitialized
-Wmisleading-indentation
-Wstringop-truncation
-Wunused-function
-Wunused-variable

It also removes forward declarations and moves
global functions to the bottom, keeping locals
at the top, in ras_tests.c.

v2: Fix compilation.

Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>

Revert "tests/amdgpu: Fix various warnings"

This reverts commit fb1634583f3ba22c67cad0df7022b6ac48a40c56.

Revert "tests/amdgpu: Fix unused function warning (v2)"

This reverts commit 4ff499cd85f86a7b5b28f3449de2fbad1b91c795.

Revert "tests/amdgpu: Fix buffer overflow (v3)"

This reverts commit 680542ce086f5d0ba70331f8d350edfae20c534f.

tests/amdgpu: Proper format for "-l"

Proper format for command line option "-l",
listing the supported and unsupported tests:

1) Add an aligned column header.

2) Align all fields into columns.

3) Fixed length fields, come before the last
column, which is a variable length field.

4) Variable length field, which is the name of the
test, goes in the last column.

5) If a suite is disabled, do not iterate over its
tests, as they'd naturally be all disabled.

Now the output looks like this:
$sudo ./amdgpu_test -l
 What: ID:   Status: Name
Suite:  1:  ENABLED: Basic Tests
 Test:  1:  ENABLED: Query Info Test
 Test:  2:  ENABLED: Userptr Test
 Test:  3: DISABLED: bo eviction Test
 Test:  4:  ENABLED: Command submission Test (GFX)
 Test:  5:  ENABLED: Command submission Test (Compute)
 Test:  6:  ENABLED: Command submission Test (Multi-Fence)
 Test:  7:  ENABLED: Command submission Test (SDMA)
 Test:  8:  ENABLED: SW semaphore Test
 Test:  9: DISABLED: Sync dependency Test
 Test: 10: DISABLED: Dispatch Test (Compute)
 Test: 11: DISABLED: Dispatch Test (GFX)
 Test: 12: DISABLED: Draw Test
 Test: 13: DISABLED: GPU reset Test
Suite:  2:  ENABLED: BO Tests
 Test:  1:  ENABLED: Export/Import
 Test:  2: DISABLED: Metadata
 Test:  3:  ENABLED: CPU map/unmap
 Test:  4:  ENABLED: Memory alloc Test
 Test:  5:  ENABLED: Memory fail alloc Test
 Test:  6:  ENABLED: Find bo by CPU mapping
Suite:  3: DISABLED: CS Tests
Suite:  4: DISABLED: VCE Tests
Suite:  5:  ENABLED: VCN Tests
 Test:  1:  ENABLED: VCN DEC create
 Test:  2:  ENABLED: VCN DEC decode
 Test:  3:  ENABLED: VCN DEC destroy
 Test:  4:  ENABLED: VCN ENC create
 Test:  5:  ENABLED: VCN ENC decode
 Test:  6:  ENABLED: VCN ENC destroy
Suite:  6: DISABLED: UVD ENC Tests
Suite:  7: DISABLED: Deadlock Tests
Suite:  8:  ENABLED: VM Tests
 Test:  1:  ENABLED: resere vmid test
 Test:  2:  ENABLED: unaligned map
 Test:  3:  ENABLED: vm mapping test
Suite:  9: DISABLED: RAS Tests
Suite: 10:  ENABLED: SYNCOBJ TIMELINE Tests
 Test:  1:  ENABLED: syncobj timeline test
$_

Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>

tests/amdgpu: Fix buffer overflow (v3)

This patch fixes the following warning:
-Wformat-overflow=

v2: Use the correct strlcat(3).
v3: Use strncat(3) and remove libbsd dependency.

Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>

tests/amdgpu: Fix unused function warning (v2)

This patch fixes:
-Wunused-function

v2: Always enable amdgpu_ras_test().

Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>

tests/amdgpu: Fix various warnings

This patch fixes the following warnings:
-Wformat=
-Wmaybe-uninitialized
-Wmisleading-indentation
-Wstringop-truncation
-Wunused-function
-Wunused-variable

It also removes forward declarations and moves
global functions to the bottom, keeping locals
at the top, in ras_tests.c.

Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>

tests/amdgpu: fix a amdgpu_test hang issue on some platforms.

The computer ring test name mis-match in different files,
thus may be set with TRUE on wrong platforms.

Change-Id: I0b918ff8faf08c9c9f1ad55f4dcd18f66b956901
Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>

meson: Replace 'config.h' with config_file

This fixes an issue with libdrm failing to build when used as a meson
subproject. Using 'config.h' directly will cause it to possibly refer to
the wrong file.

By using `@0@.format(config_file)`, it will be transformed into the
correct relative path, e.g. `./config.h` in normal build,
`./subprojects/libdrm/config.h` in subproject build.

Signed-off-by: Scott Anderson <scott@anderso.nz>
Acked-by: Eric Engestrom <eric.engestrom@intel.com>

add a minimal .gitignore back

It was entirely deleted along with autotools, but adding this simple one
will cover most people's needs.

Suggested-by: Lucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>

gitlab-ci: rename build folder to simply `build`

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>

intel: sync i915_pciids.h with kernel

Changes:
651cc835d5f6 ("drm/i915: Add new EHL/JSL PCI ids")
b6a8781a447c ("drm/i915/cml: Remove unsupport PCI ID")
8717c6b7414f ("drm/i915/cml: Separate U series pci id from origianl list.")

v2: added the latest CML changes

Cc: James Ausmus <james.ausmus@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>

modetest: Use floating vrefresh while dumping mode

Add function to derive floating value of vertical
refresh rate from drm mode using pixel clock,
horizontal total size and vertical total size.

Use this function to find suitable mode having vrefresh
value which is matching with user provided vrefresh value.

If user doesn't provide any vrefresh value in args then
update vertical refresh rate value in pipe args using this
function.

Also use this function for printing floating vrefresh while
dumping all available modes.

This will give more accurate picture to user for available modes
differentiated by floating vertical refresh rate and help user
select more appropriate mode using suitable refresh rate value.

V4:
1) While setting mode, print mode name and vrefresh using struct
   drmModeModeInfo instead of struct pipe_args.
2) Revert back to using a float value instead of float *
   for vrefresh arg in connector_find_mode().

V3:
1) Change name of function used to derive refresh rate.

V2:
1) Don't use inline function for deriving refresh rate from mode.
2) If requested mode not found, print refresh rate only
   if user had provided it in args.

Signed-off-by: Devarsh Thakkar <devarsh.thakkar@xilinx.com>
Reviewed-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

tests/amdgpu: add gfx ring bad slow draw test

for gfx9

Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

tests/amdgpu: add gfx ring draw hang test

for gfx9

Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

tests/amdgpu: add bad slow dispatch test

add gfx/compute bad slow dispatch test for gfx9

Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

tests/amdgpu: add dispatch hang test

add compute/gfx dispatch hang test for gfx9

Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

tests/amdgpu: update draw test for gfx9

Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>

tests/amdgpu: enable dispatch/draw tests for Renoir

It can run dispatch/draw tests on new renoir chips. So it needs to
enable dispatch/draw tests for Renoir again.

Change-Id: I3a72a4bbfe0fc663ee0e3e58d8e9c304f513e568
Signed-off-by: changzhu <Changfeng.Zhu@amd.com>
Reviewed-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>

xf86atomic: require CAS support in libatomic_ops

Since AO_compare_and_swap_full() is used by libdrm, AO_REQUIRE_CAS
must be defined before including <atomic_ops.h> so that we are sure
that CAS support will be provided. This is necessary to make sure that
the AO_compare_and_swap_full() function will be provided on all
architectures, including the ones that don't have built-in CAS support
such as SPARCv8.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Acked-by: Eric Engestrom <eric.engestrom@intel.com>

Add ARM support into xf86drm.h

This provides support for Xorg interface.  Without this the vivante
samples will hang during close requiring a reboot

[Adapted from yocto project]
Upstream-Status: Pending
Signed-off-by: Lauren Post <lauren.post@freescale.com>
Signed-off-by: Evan Kotara <evan.kotara@freescale.com>
[Thomas: change CAS code to only be used on ARMv6/ARMv7, and not
ARMv4/ARMv5, which don't support ldrex/strex. If no CAS implementation
is provided libdrm falls back to a system call for locking/unlocking.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Acked-by: Eric Engestrom <eric.engestrom@intel.com>
---
Changes v1 -> v2:
  - add comment explaining exclusion of ARMv4/ARMv5 and lower

meson.build: fix intel atomics detection

Use the stronger compiler.link() test (instead of the weaker
compiler.compile()) to fix the intel atomics detection.

Fixes false positive in case of sparc compile (buildroot toolchain).

Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Reviewed-by: Eric Engestrom <eric.engestrom@intel.com>

amdgpu: add new marketing names from 19.30

Add new marketing names.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

gitlab-ci: update to current ci-templates master

To workaround skopeo issues.

Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>

tests/nouveau/threaded: adapt ioctl signature

POSIX says ioctl() has the signature (int, int, ...) but glibc has decided to
use (int, unsigned long int, ...) instead.

Use a #ifdef to adapt the replacement function as appropriate.

Signed-off-by: Ross Burton <ross.burton@intel.com>
[Taken from https://raw.githubusercontent.com/openembedded/openembedded-core/master/meta/recipes-graphics/drm/libdrm/musl-ioctl.patch]
Signed-off-by: Peter Seiderer <ps.report@gmx.net>
Reviewed-by: Eric Engestrom <eric.engestrom@intel.com>

tests/amdgpu/basic_tests.c: change BUFFER_SIZE used in Userptr Test adjust to PAGE_SIZE

Userptr Test will fail on PAGE_SIZE bigger than BUFFER_SIZE(8 * 1024)

Signed-off-by: xinxu <xinxu@loongson.cn>
(cherry picked from commit fb7dfdc5fb58795365b70117c3eb625f2edb8f06)

modetest: Add support for setting mode having floating vertical refresh rate

For the scenario where user may require to modeset with a mode
supporting a fractional value for vertical refresh-rate,
appropriate mode can be selected by searching for mode
having matching fractional vertical refresh rate using
below equation.

vrefresh = (1000 * pixel clock) / (htotal * vtotal) Hz.

We do this way since driver doesn't return float value of vrefresh
as it use int for vrefresh in struct drm_mode_info, but we can derive
the actual value using pixel clock, horizontal total size and
vertical total size values.

So for e.g. if user want to select mode having 59.94 Hz as refresh rate
then with this patch it be can done as shown in below command,
given there is an appropriate mode is available :

modetest -M xlnx -s 39:1920x1080-59.94@BG24 -v

NOTE: Above command was tested on xilinx DRM driver with DP
monitor which was supporting mode having 59.94 Hz refresh rate.

V2: Update commit message
V3: Update with below changes as per review comments :
  1) Use epsilon for vrefresh comparison
  2) Use implicit type-casting wherever possible
V4: Keep patch version history on main commit message

Signed-off-by: Devarsh Thakkar <devarsh.thakkar@xilinx.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

gitlab-ci: add PowerPC build

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>

gitlab-ci: add aarch64 & armhf builds

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>

gitlab-ci: add x86 (32 bits) build

Suggested-by: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>

gitlab-ci: set up cross build infra

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>

gitlab-ci: drop arch build down to daily builds

Suggested-by: Michel Dänzer <michel@daenzer.net>
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

Fix missing stdlib includes in xf86drmMode.h

Including xf86drmMode.h results in undefined references to uint32_t
and ssize_t. Include the stdlib headers that define them to allow the
file to be included without xf86drm.h.

Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Eric Engestrom <eric.engestrom@intel.com>

meson: drop old symbols check environment

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

tegra: convert to new symbols check

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

radeon: convert to new symbols check

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

omap: convert to new symbols check

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

nouveau: convert to new symbols check

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

libkms: convert to new symbols check

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

intel: convert to new symbols check

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

freedreno: convert to new symbols check

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

exynos: convert to new symbols check

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

etnaviv: convert to new symbols check

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

amdgpu: convert to new symbols check

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

meson: import Mesa's symbols check script

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

freedreno: drop leftover symbol from the export list

Fixes: 09cbccff558988c57e52 ("freedreno: remove deprecated ringmarker API")
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

libdrm: wrap new flexible syncobj query interface v2

v2: nit-picks fix

Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: Christian König <Christian.Koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
For the xf86drm.[ch] part : Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>

sync up drm.h

    a) delta: drm: use pad as flags in drm_syncobj_timeline_array.
    b) Generated using make headers_install.
    c) Generated from origin/drm-misc-next commit 949561eb85bcee10248e7da51d44a0325d5e0d1b"

Signed-off-by: Chunming Zhou <david1.zhou@amd.com>

gitlab-ci: pre-build containers to improve CI run time and reliability

Shaves about 50% off the build time on both debian and arch builds.
(yeah, I know, it's very small anyway compared to mesa, but we might
 want to add more things in the future)

This also makes the build no longer dependent on external websites:
once the image is build, only fdo-internal services are used.

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>

Revert "Revert "libdrm: remove autotools support""

The external tooling issue has been fixed, so we can delete autotools
again :)

Bump the version to 2.4.100 for autotools

Revert "libdrm: remove autotools support"

This reverts commit f057dc91e93ae21e11ab48a26127d569972f3eae.

Bump the version to 2.4.100

amdgpu: add amdgpu_cs_query_reset_state2 for AMDGPU_CTX_OP_QUERY_STATE2

This is a better GPU reset query.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>

include: update amdgpu_drm.h

Generated from kernel commit:
    815fb4c9d7da862 "drm/amdgpu: return tcc_disabled_mask to userspace"

Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>

test/amdgpu: don't free unused bo handle

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Eric Engestrom <eric.engestrom@intel.com>

libdrm: remove autotools support

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>

RELEASING: update instructions to use meson instead of autotools

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Acked-by: Dylan Baker <dylan@pnwbakers.com>
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>

*-symbols-check: let meson figure out how to execute the scripts

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

*-symbols-check: use normal shell over bash

None of the tests are bash specific. Tested with bash, zsh, dash, mksh
and ksh.

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Tested-by: Niclas Zeising <zeising@daemonic.se>
Reviewed-by: Niclas Zeising <zeising@daemonic.se>
Reviewed-by: Eric Engestrom <eric.engestrom@intel.com>

meson.build: Fix header detection on FreeBSD

FreeBSD requires sys/types.h for sys/sysctl.h, add it as part of the
includes when checking for headers.
Instead of splitting out the check for sys/sysctl.h from the other
header checks, just add sys/types.h to all header checks.

v2 [Emil]
 - add inline comment
 - drop bash/sh hunk

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Signed-off-by: Niclas Zeising <zeising@daemonic.se>
Reviewed-by: Eric Engestrom <eric.engestrom@intel.com>

meson.build: Fix typo

Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Eric Engestrom <eric.engestrom@intel.com>

tests/amdgpu: add the missing deactivation case for dispatch test

Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

amdgpu: add ras feature capability check in inject test

When running ras inject test, it's needed to be aligned
with kernel's ras enablement.

Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

amdgpu: add ras inject unit test

Both UMC and GFX ras single_correctable
inject tests are added.

Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

amdgpu: delete test configuration file

Json package dependence is removed from amdgpu_test,
so this json configuration file is not needed any more.

Suggested-by: Christian König <christian.koenig@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

amdgpu: remove json package dependence

Except CUnit library, no additional external
library should be needed when compiling amdgpu_test.
This will keep this binary self containing.

Suggested-by: Christian König <christian.koenig@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

meson: fix sys/mkdev.h detection on Solaris

On Solaris, sys/sysmacros.h has long-deprecated copies of major() & minor()
but not makedev().
sys/mkdev.h has all three and is the preferred choice.

Let's make sure we check for all 3 major(), minor() and makedev().

Fixes build failure with error:
../xf86drm.c: In function ‘drmOpenMinor’:
../xf86drm.c:454:30: error: implicit declaration of function ‘makedev’ [-Werror=implicit-function-declaration]
  454 |         return drmOpenDevice(makedev(DRM_MAJOR, minor), minor, type);
      |                              ^~~~~~~

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Tested-by: Alan Coopersmith <alan.coopersmith@oracle.com>

intel: sync i915_pciids.h with kernel

Add the new CML PCI IDS.

Align with kernel commit:
bfc4c359b2822 ("drm/i915/cml: Add Missing PCI IDs")

This is in sync with kernel header as of:
0747590267e7 ("drm-tip: 2019y-08m-30d-18h-03m-18s UTC integration manifest")

Cc: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Anusha Srivatsa <anusha.srivatsa@intel.com>

amdgpu: add umc ras inject test configuration

Both umc single_correctable and multi_uncorrectable
inject types are added.

Signed-off-by: Dennis Li <dennis.li@amd.com>
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

tests/amdgpu/ras: refine ras inject test

Ras inject test framework is invalid with original codes,
so refine it to make it work on top of kernel ras inject
feature enablement.

Signed-off-by: Dennis Li <dennis.li@amd.com>
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

amdgpu: add gfx ras inject configuration file

This configuration file will be picked up when
running gfx ras inject tests by amdgpu_test tool.
For the time being, only add those tests that are
successfully trafficked. In addition, this file
can also be modified by user to add or delete ras
inject unit tests for different IP blocks/subblocks.

Signed-off-by: Dennis Li <dennis.li@amd.com>
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

tests/amdgpu: divide dispatch test into compute and gfx

for better clarification

v2: accordingly change dispatch_test caller in gpu_reset test

Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Flora Cui <flora.cui@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

tests/amdgpu: disable reset test for now

ASIC hang randomly.

Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

intel: Add support for EHL

Add the PCI ID import for EHL.

Cc: James Ausmus <james.ausmus@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>

intel: add the TGL 12 PCI IDs and macros

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>

intel: sync i915_pciids.h with kernel

Straight copy from the kernel file, aligned with drm-intel-next-queued
commit cb823ed9915b ("drm/i915/gt: Use intel_gt as the primary object
for handling resets")

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>

tests/amdgpu: add gpu reset test

1. perform gpu reset
2. perform dispatch test to verify gpu reset to a good state

Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>

tests/amdgpu: fix for dispatch/draw test

1. skip test if there's no desired ring
2. clear shader buffer
3. update command buffer for gfx9

Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>

xf86drm: dedupe drmGetDeviceName() logic

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

xf86drm: use max size of drm node name instead of arbitrary size

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

xf86drm: dedupe `#define`s

Adapted from a local patch carried by DragonFlyBSD:
https://github.com/DragonFlyBSD/DPorts/blob/bc056f88f7e4d468d8c9751f831a47b5ae1326e3/graphics/libdrm/files/patch-xf86drm.h

Patch is sadly uncredited (a bot authored the commit), so I can't credit
the author here either.

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>

xf86drm: open correct render node on non-linux

drm render nodes have the same major as drm primary devices but offset
the minor by a base of 128.

I expected the name of the device to have numbering starting at 0 when
these non-linux codepaths were added (before OpenBSD had render nodes).

Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Acked-by: Eric Engestrom <eric.engestrom@intel.com>

xf86drm: test for render nodes before primary nodes

Unlike Linux the OpenBSD primary "drm" device name is substring of the
"drmR" render node device name and strncmp() tests resulted in render
nodes being flagged as primary nodes.

Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Reviewed-by: Eric Engestrom <eric.engestrom@intel.com>
Acked-by: Eric Engestrom <eric.engestrom@intel.com>

tests/util: fix incorrect memset argument order

Make it actually clear the LUT.

Reported-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Dave Airlie <airlied@redhat.com>

Bump version to 2.4.99

amdgpu: Rename fd_mutex/list to dev_mutex/list

Seems to better reflect what they're for.

Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>

amdgpu: Add BO handle to table in amdgpu_bo_create

Simplifies its callers.

dev->bo_table_mutex is now always held when amdgpu_bo_create is called
(this was already the case in amdgpu_bo_import).

Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>

amdgpu: Pass file descriptor directly to amdgpu_close_kms_handle

And propagate drmIoctl's return value.

This allows replacing all remaining open-coded DRM_IOCTL_GEM_CLOSE
ioctl calls with amdgpu_close_kms_handle calls.

Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>

etnaviv: Fix double-free in etna_bo_cache_free()

The following situation can happen in a multithreaded OpenGL application.
A BO is submitted from etna_cmd_stream #1 with flags set for read.
A BO is submitted from etna_cmd_stream #2 with flags set for write.
This triggers a flush on stream #1 and clears the BO's current_stream
pointer. If at this point, stream #2 attempts to queue BO again, which
does happen, the BO will be added to the submit list twice. The Linux
kernel driver correctly detects this and warns about it with "BO at
index %u already on submit list" kernel message.

However, when cleaning the BO cache in etna_bo_cache_free(), the BO
which was submitted twice will also be free()d twice, this triggering
a glibc double free detector.

The fix is easy, even if the BO does not have current_stream set,
iterate over current streams' list of BOs before adding the BO to it
and verify that the BO is not yet there.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Christian Gmeiner <christian.gmeiner@gmail.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Wladimir J. van der Laan <laanwj@gmail.com>

amdgpu: Update amdgpu_bo_handle_type_kms_noimport documentation

To reflect current reality.

Reviewed-by: Christian König <christian.koenig@amd.com>

amdgpu: Move union declaration to top of amdgpu_cs_ctx_override_priority

Avoids compiler warning:

../../amdgpu/amdgpu_cs.c: In function 'amdgpu_cs_ctx_override_priority':
../../amdgpu/amdgpu_cs.c:155:2: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement]
  union drm_amdgpu_sched args;
  ^~~~~

Reviewed-by: Christian König <christian.koenig@amd.com>

etnaviv: drop etna_bo_from_handle symbol

There is no implementation and also no users, so there is no point
in keeping it in the API.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>

util: fix include path for drm_mode.h

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>

modetest: add FP16 format support

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

modetest: add the ability to specify fill patterns on the commandline

Instead of hacking the binary every time, we can now specify directly.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

modetest: add C8 support to generate SMPTE pattern

This includes logic to configure the LUT accordingly.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

modetest: add an add_property_optional variant that does not print errors

As new features are added and others are declared to be legacy, it's
nice to be able to implement fallbacks. As such, create a
property-setting variant that does not generate errors which can very
well be entirely expected.

Will be used for gamma control in a future change.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>