OSDN Git Service

android-x86/external-libdrm.git
2 years agoandroid: add genrule for generated_static_table_fourcc.h r-x86
Mauro Rossi [Sun, 11 Jul 2021 16:23:52 +0000 (18:23 +0200)]
android: add genrule for generated_static_table_fourcc.h

Fixes the following building error:

external/libdrm/xf86drm.c:158:10: fatal error: 'generated_static_table_fourcc.h' file not found
         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 error generated.

Change-Id: I1b0cac498ed63ebec6e8c03629bbf4a1b6a9618d

2 years agoMerge tag 'libdrm-2.4.107'
Mauro Rossi [Tue, 20 Jul 2021 19:36:41 +0000 (21:36 +0200)]
Merge tag 'libdrm-2.4.107'

libdrm-2.4.107

Change-Id: Ida30e1c76b4f6c1ccfce29c2465efcba15995420

2 years agoBump version to 2.4.107
Bas Nieuwenhuizen [Fri, 2 Jul 2021 01:12:14 +0000 (03:12 +0200)]
Bump version to 2.4.107

Signed-off-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2 years agoamdgpu: update marketing names
Alex Deucher [Mon, 28 Jun 2021 19:48:14 +0000 (15:48 -0400)]
amdgpu: update marketing names

From 21.20 release.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agoREADME.rst: Include some notes about syncing uapi headers
Marius Vlad [Thu, 11 Feb 2021 11:08:05 +0000 (13:08 +0200)]
README.rst: Include some notes about syncing uapi headers

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
2 years agoxf86drm: Add support for decoding AMLOGIC format modifiers
Marius Vlad [Tue, 23 Feb 2021 22:51:45 +0000 (00:51 +0200)]
xf86drm: Add support for decoding AMLOGIC format modifiers

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
2 years agoxf86drm: Add support for decoding AMD format modifiers
Marius Vlad [Wed, 17 Feb 2021 13:59:30 +0000 (15:59 +0200)]
xf86drm: Add support for decoding AMD format modifiers

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
2 years agoxf86drm: Add support for decoding Nvidia format modifiers
Marius Vlad [Mon, 1 Feb 2021 11:45:58 +0000 (13:45 +0200)]
xf86drm: Add support for decoding Nvidia format modifiers

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
2 years agoxf86drm: Add a vendor function to decode the format modifier
Marius Vlad [Mon, 1 Feb 2021 11:39:43 +0000 (13:39 +0200)]
xf86drm: Add a vendor function to decode the format modifier

As format modifiers can be encoded in quite complex forms, the static
table previously added is not sufficient to retrieve, extract and decode
the token formats to a human-readable string.  This patch introduces a
vendor specific callback which could be used to perform an additional
search to match up with vendor encoding scheme, which, will be used
first, before resorting to searching the static table.

With it, add support for decoding the ARM format modifiers.

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
2 years agoxf86drm: Add a human readable representation for format modifiers
Marius Vlad [Mon, 1 Feb 2021 11:23:20 +0000 (13:23 +0200)]
xf86drm: Add a human readable representation for format modifiers

Introduces two new methods to retrieve a human readable representation of a
format modifier:

drmGetFormatModifierName() - returns a format modifier as a string,
from a token modifier
drmGetFormatModifierVendor() - returns the vendor as a string, from a
token modifier

and the fourcc_mod_get_vendor macro that returns the vendor.

New format modifiers added in drm_fourcc.h uapi kernel header should be
sync'ed up with libdrm and should include a human readable
representation for that format modifier, in order to display it
correctly as a string.

That happens with the help of a python script that reads up drm_fourcc
header file and outputs a static table comprised of token modifiers
alongside a vendor table (Suggested-by Simon Ser <contact@emersion.fr>).

The reason for doing it in libdrm is to have a unified place instead of each
user of libdrm having a way to keep track of the format modifiers.

With this patch, modetest has also been modified to make use of it.

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
2 years agoConditionally include <linux/limits.h> and <sys/params.h> on Linux, BSD
Eleni Maria Stea [Sun, 20 Jun 2021 06:33:58 +0000 (09:33 +0300)]
Conditionally include <linux/limits.h> and <sys/params.h> on Linux, BSD

<linux/limits.h> should be included conditionally for Linux only, also
SPECNAMELEN used conditionally when the OS is FreeBSD requires to
include <sys/params.h>.

Signed-off-by: Eleni Maria Stea <elene.mst@gmail.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
2 years ago_WANT_KERNEL_ERRNO must be defined in FreeBSD for ERESTART to be used
Eleni Maria Stea [Mon, 7 Jun 2021 07:24:11 +0000 (10:24 +0300)]
_WANT_KERNEL_ERRNO must be defined in FreeBSD for ERESTART to be used

In FreeBSD's errno.h ERESTART is not defined by default, only when the
user requests the pseudo-errors returned inside the kernel to be
enabled. As a result the block where drmCommandWriteRead is
called returns compile error. Defined _WANT_KERNEL_ERRNO to fix it (see
FreeBSD's /usr/include/errno.h)

Signed-off-by: Eleni Maria Stea <elene.mst@gmail.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
2 years agoinclude <sys/types.h> in xf86drmMode when the OS is FreeBSD
Eleni Maria Stea [Mon, 7 Jun 2021 07:19:18 +0000 (10:19 +0300)]
include <sys/types.h> in xf86drmMode when the OS is FreeBSD

<sys/types.h> need to be included in xf86drmMode.c for type u_int in
<sys/sysctl.h> (that is included when OS is FreeBSD) to be recognized.

Signed-off-by: Eleni Maria Stea <elene.mst@gmail.com>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
2 years agoamdgpu: Add vamgr for capture/replay.
Bas Nieuwenhuizen [Mon, 23 Nov 2020 01:18:05 +0000 (02:18 +0100)]
amdgpu: Add vamgr for capture/replay.

In Vulkan we have extensions to assist with capture in replay in a
world where addresses are returned to the application. This involves
creating buffers at the same VA during replay as they were during
capture.

By itself libdrm_amdgpu already has support for this, but there is
the obvious failure mode that if another buffer is already allocated
at that VA things fail spectacularly. This is an actual issue as
internal buffers, like winsys images or shader binaries also
participate in the same VA allocation.

To avoid this problem applications have to create buffers which
are going to be captured with a flag, and the implementation is to
separate VA allocation for those buffers to reduce the collision risk:

"Implementations are expected to separate such buffers in the GPU address
space so normal allocations will avoid using these addresses. Apps/tools
should avoid mixing app-provided and implementation-provided addresses for
buffers created with VK_BUFFER_CREATE_DEVICE_ADDRESS_CAPTURE_REPLAY_BIT,
to avoid address space allocation conflicts."

This patch implements that by adding a flag for these buffers and allocating
address space from the top of the address range instead of the bottom.

Signed-off-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Christian König <christian.koenig@amd.com>
3 years agotests/amdgpu/hotunplug: Add hotunplug with exported fence
Andrey Grodzovsky [Fri, 4 Jun 2021 20:08:41 +0000 (16:08 -0400)]
tests/amdgpu/hotunplug: Add hotunplug with exported fence

Disconnect device while fence is exported.
Also disable this test for sytem with single GPU.

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alex Deucher alexander.deucher@amd.com
3 years agotests/amdgpu/hotunplug: Add hotunplug with exported bo test
Andrey Grodzovsky [Tue, 1 Jun 2021 18:30:08 +0000 (14:30 -0400)]
tests/amdgpu/hotunplug: Add hotunplug with exported bo test

Disconnect device while BO is exported.

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alex Deucher alexander.deucher@amd.com
3 years agotests/amdgpu/hotunplug: Add unplug with cs test.
Andrey Grodzovsky [Mon, 31 May 2021 18:13:06 +0000 (14:13 -0400)]
tests/amdgpu/hotunplug: Add unplug with cs test.

Same as simple test but while doing cs

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alex Deucher alexander.deucher@amd.com
3 years agotest/amdgpu/hotunplug: Add test suite for GPU unplug
Andrey Grodzovsky [Wed, 26 May 2021 19:15:33 +0000 (15:15 -0400)]
test/amdgpu/hotunplug: Add test suite for GPU unplug

Add plug/unplug device and open/close device file
infrastructure.
Add basic test - unplug device while device file still
open. Close device file afterwards and replug the device.

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alex Deucher alexander.deucher@amd.com
3 years agotest/amdgpu: Add helper functions for hot unplug
Andrey Grodzovsky [Mon, 31 May 2021 15:11:58 +0000 (11:11 -0400)]
test/amdgpu: Add helper functions for hot unplug

Expose close device and add open device wich preserves
test index.

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alex Deucher alexander.deucher@amd.com
3 years agotests/amdgpu: Fix valgrind warning
Andrey Grodzovsky [Tue, 1 Jun 2021 18:56:54 +0000 (14:56 -0400)]
tests/amdgpu: Fix valgrind warning

Struct access after free

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alex Deucher alexander.deucher@amd.com
3 years agoamdgpu: Added product name for E9390,E9560 and E9565 dgpu
Rahul Kumar [Mon, 7 Jun 2021 14:59:19 +0000 (10:59 -0400)]
amdgpu: Added product name for E9390,E9560 and E9565 dgpu

Update marketing names.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agoRevert "tests/amdgpu: fix bo eviction test issue"
Lang Yu [Sat, 8 May 2021 05:55:54 +0000 (13:55 +0800)]
Revert "tests/amdgpu: fix bo eviction test issue"

This reverts commit a5a400c9581c3b91598623603067556b18084c5d.

Bo evict test was disabled by default per below commit.
So still keep it as disabled.

1f6a85cc test/amdgpu: disable bo eviction test by default

Signed-off-by: Lang Yu <Lang.Yu@amd.com>
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Lang Yu <Lang.Yu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
3 years agoMerge tag 'libdrm-2.4.106' into r-x86
Mauro Rossi [Sun, 23 May 2021 20:41:52 +0000 (22:41 +0200)]
Merge tag 'libdrm-2.4.106' into r-x86

libdrm-2.4.106

Change-Id: I7c0039f5f1ebf85d8d7e3868488f39e4223bc7fd

3 years agoMerge tag 'libdrm-2.4.105' into r-x86
Mauro Rossi [Sun, 23 May 2021 20:40:16 +0000 (22:40 +0200)]
Merge tag 'libdrm-2.4.105' into r-x86

libdrm-2.4.105

Change-Id: Icf68fcf833139aff782377ff994061eaf64f4820

3 years agointel: Add support for ADLP
Tejas Upadhyay [Mon, 17 May 2021 05:35:29 +0000 (11:05 +0530)]
intel: Add support for ADLP

Add ADLP platform support and PCIIDs

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Uma Shankar <uma.shankar@intel.com>
Signed-off-by: Pankaj Bharadiya <pankaj.laxminarayan.bharadiya@intel.com>
Signed-off-by: Caz Yokoyama <caz.yokoyama@intel.com>
Signed-off-by: Anusha Srivatsa <anusha.srivatsa@intel.com>
Signed-off-by: Tejas Upadhyay <tejaskumarx.surendrakumar.upadhyay@intel.com>
3 years agoBump version to 2.4.106
Dave Airlie [Tue, 18 May 2021 03:37:56 +0000 (13:37 +1000)]
Bump version to 2.4.106

Signed-off-by: Dave Airlie <airlied@redhat.com>
3 years agonouveau: add debug option to sync pushbuffer submissions
Karol Herbst [Wed, 5 May 2021 12:14:30 +0000 (14:14 +0200)]
nouveau: add debug option to sync pushbuffer submissions

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
3 years agonovueau: document debug flags
Karol Herbst [Wed, 5 May 2021 12:05:52 +0000 (14:05 +0200)]
novueau: document debug flags

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
3 years agonouveau: rework debugging so we can also dump into a file
Karol Herbst [Wed, 5 May 2021 12:01:57 +0000 (14:01 +0200)]
nouveau: rework debugging so we can also dump into a file

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
3 years agonouveau: make debug features accessible in normal builds
Karol Herbst [Wed, 5 May 2021 12:01:29 +0000 (14:01 +0200)]
nouveau: make debug features accessible in normal builds

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
3 years agonouveau: mask push buffer length pushbuf_dump
Karol Herbst [Sun, 3 May 2020 14:21:09 +0000 (16:21 +0200)]
nouveau: mask push buffer length pushbuf_dump

nvc0 sets the NVC0_IB_ENTRY_1_NO_PREFETCH bit on some pushbuffers

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
3 years agonouveau: fix crash in pushbuf_dump with an unmapped bo
Karol Herbst [Sun, 3 May 2020 14:19:53 +0000 (16:19 +0200)]
nouveau: fix crash in pushbuf_dump with an unmapped bo

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
3 years agoci: use `base-devel` tag of archlinux image instead of `base` and then installing...
Eric Engestrom [Thu, 29 Apr 2021 08:15:11 +0000 (10:15 +0200)]
ci: use `base-devel` tag of archlinux image instead of `base` and then installing `base-devel` after

Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Simon Ser <contact@emersion.fr>
3 years agoci: use archlinux/archlinux docker image instead of deprecated and now removed archli...
Eric Engestrom [Thu, 29 Apr 2021 08:13:43 +0000 (10:13 +0200)]
ci: use archlinux/archlinux docker image instead of deprecated and now removed archlinux/base

Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Simon Ser <contact@emersion.fr>
3 years agoRevert "xf86drmMode: set FB_MODIFIERS flag when modifiers are supplied"
Bas Nieuwenhuizen [Wed, 21 Apr 2021 19:31:06 +0000 (21:31 +0200)]
Revert "xf86drmMode: set FB_MODIFIERS flag when modifiers are supplied"

This reverts commit b362850689d1b0048b7f4641cc236128b5a43773.

This breaks when the kernel driver does not support modifiers and the
application properly zeroes the modifiers.

Acked-by: Simon Ser <contact@emersion.fr>
3 years agotest/amdgpu: Add emit mem sync flag for test IB
Jinzhou Su [Mon, 19 Apr 2021 06:04:22 +0000 (14:04 +0800)]
test/amdgpu: Add emit mem sync flag for test IB

In syncobj test, 3 threads will be created. Sometimes
the first gfx IB and the third sdma IB will use same
physical page. There will be risk that sdma engine will
read gfx IB in the same physical page. So better to flush
the cache before commit the sdma IB.

Signed-off-by: Jinzhou Su <Jinzhou.Su@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotest/amdgpu: use tmz ids to check whether enable security tests
Huang Rui [Fri, 9 Apr 2021 10:07:16 +0000 (18:07 +0800)]
test/amdgpu: use tmz ids to check whether enable security tests

Using tmz ids that reported from kernel to decide whether enable
security tests.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
3 years agotests/amdgpu/vcn: update to support aldebaran
James Zhu [Mon, 12 Apr 2021 12:44:59 +0000 (08:44 -0400)]
tests/amdgpu/vcn: update to support aldebaran

VCN is supported after AI family Arcturus.

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
3 years agotests/amdgpu:retire asic_id check on unsupported cases
Feifei Xu [Wed, 25 Nov 2020 09:20:33 +0000 (17:20 +0800)]
tests/amdgpu:retire asic_id check on unsupported cases

Retire the asic_id check for AI family.

Signed-off-by: Feifei Xu <Feifei.Xu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
3 years agotests/amdgpu: update gfx9 BufferCopy/BufferClear
Feifei Xu [Thu, 1 Apr 2021 08:33:00 +0000 (16:33 +0800)]
tests/amdgpu: update gfx9 BufferCopy/BufferClear

buffer_load/store_format_xyzw require 64bit vgpr_a[2].
The original parameter is one u32. Modify the shader binary to
fit the 64bit parameter.

Signed-off-by: Feifei Xu <Feifei.Xu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Tested-by: Gang Long <Gang.Long@amd.com>
3 years agoBump version to 2.4.105
Leo Liu [Wed, 7 Apr 2021 13:54:11 +0000 (09:54 -0400)]
Bump version to 2.4.105

Signed-off-by: Leo Liu <leo.liu@amd.com>
3 years agoamdgpu: add function of INFO ioctl for querying video caps
Leo Liu [Tue, 19 Jan 2021 21:51:02 +0000 (16:51 -0500)]
amdgpu: add function of INFO ioctl for querying video caps

via the newly added uapi/amdgpu_drm interface

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
3 years agoamdgpu: sync up amdgpu_drm.h with latest from kernel
Leo Liu [Tue, 23 Feb 2021 16:22:42 +0000 (11:22 -0500)]
amdgpu: sync up amdgpu_drm.h with latest from kernel

From drm-next:

commit 2cbcb78c9ee5520c8d836c7ff57d1b60ebe8e9b7
Merge: 06debd6e1b28 8c44390d8872
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Fri Mar 26 15:52:01 2021 +0100

    Merge tag 'amd-drm-next-5.13-2021-03-23' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

    amd-drm-next-5.13-2021-03-23:

    amdgpu:
    ...

    UAPI:
    - amdgpu: Add a new INFO ioctl interface to query video capabilities
      rather than hardcoding them in userspace.  This allows us to provide
      fine grained asic capabilities (e.g., if a particular part is
      bandwidth limited, we can limit the capabilities).  Proposed userspace:
      https://gitlab.freedesktop.org/leoliu/drm/-/commits/info_video_caps
      https://gitlab.freedesktop.org/leoliu/mesa/-/commits/info_video_caps
    ...

    Danvet: A bunch of conflicts all over, but it seems to compile ... I
    did put the call to dc_allow_idle_optimizations() on a single line
    since it looked a bit too jarring to be left alone.

Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
    From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210324040147.1990338-1-alexander.deucher@amd.com
Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
3 years agoxf86drmMode: set FB_MODIFIERS flag when modifiers are supplied
Simon Ser [Fri, 5 Mar 2021 15:07:58 +0000 (16:07 +0100)]
xf86drmMode: set FB_MODIFIERS flag when modifiers are supplied

The kernel will always return EINVAL if modifiers are supplied but
the flag DRM_MODE_FB_MODIFIERS isn't set. That's a pretty nice
footgun.

Be a little more helpful and set the flag if the user has supplied
a modifier array.

Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
3 years agotests/amdgpu: fix bo eviction test issue
Lang Yu [Mon, 29 Mar 2021 07:52:51 +0000 (15:52 +0800)]
tests/amdgpu: fix bo eviction test issue

On Raven2/Picasso, the default VRAM size is 2048M,
and the default GTT size is 3072M. If max_allocation
of VRAM exceeds half of GTT size, GTT memory can't
hold evicted bo from VRAM and bo in itself at the
same time. Then amdgpu_cs_list_validate will failed
with "Not enough memory for command submission" error.

NOTE:
The installed DRAM should be larger than 8GB,
if the VRAM size is 2048M.

Signed-off-by: Lang Yu <Lang.Yu@amd.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
3 years agodrm/tests/amdgpu: fix Metadata test failed issue
Lang Yu [Wed, 24 Mar 2021 04:18:41 +0000 (12:18 +0800)]
drm/tests/amdgpu: fix Metadata test failed issue

The unit of size_metadata is one byte not four bytes.
Enable Metadata test.

Signed-off-by: Lang Yu <Lang.Yu@amd.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
3 years agoxf86drmMode: introduce drmModeGetPropertyType
Simon Ser [Thu, 25 Feb 2021 11:19:43 +0000 (12:19 +0100)]
xf86drmMode: introduce drmModeGetPropertyType

We already have drm_property_type_is, but it's needlessly complicated
and doesn't cover all use-cases (requires the caller to provide a
type).

Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
3 years agotest/amdgpu: remove static varible in Syncobj test
Jinzhou Su [Mon, 22 Mar 2021 05:44:30 +0000 (05:44 +0000)]
test/amdgpu: remove static varible in Syncobj test

In syncobj test, wait thread and signal thread create
simultaneously. The ptr for GFX IB and SDMA IP should be
operated separately. With static, there will be risk that
GFX NOP is in SDMA IB or SDMA NOP is in GFX IB, then GFX or
SDMA hang caused.

Signed-off-by: Jinzhou Su <Jinzhou.Su@amd.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
3 years agointel: Keep libdrm working without pread/pwrite ioctls
Ashutosh Dixit [Tue, 22 Sep 2020 00:34:16 +0000 (17:34 -0700)]
intel: Keep libdrm working without pread/pwrite ioctls

The general direction at this time is to phase out pread/write ioctls and
not support them in future products. The ioctls have already been disabled
in i915 for future products. This means libdrm must handle the absence of
these ioctls. This patch does this by modifying drm_intel_gem_bo_subdata()
and drm_intel_gem_bo_get_subdata() to do the read/write using the
pread/pwrite ioctls first but when these ioctls are unavailable fall back
to doing the read/write using a combination of mmap and memcpy.

A similar solution was added to igt-gpu-tools in commit
ad5eb02eb3 ("lib/ioctl_wrappers: Keep IGT working without pread/pwrite
ioctls").

Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
3 years agomeson: use library() instead of shared_library().
Fang Tan [Tue, 9 Mar 2021 08:14:36 +0000 (16:14 +0800)]
meson: use library() instead of shared_library().

This allows users to select the library type (static or shared)
using the Meson -Ddefault_library built-in option.

Issue: https://gitlab.freedesktop.org/mesa/drm/-/issues/45

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Fang Tan <tanfang@uniontech.com>
3 years agoxf86drm: fix null pointer deref in drmGetBufInfo
Alistair Delva [Tue, 2 Mar 2021 16:18:06 +0000 (08:18 -0800)]
xf86drm: fix null pointer deref in drmGetBufInfo

If info.count is large, drmMalloc() / alloca() may fail, and the
resulting null pointer is not null checked before dereference.

Issue: https://gitlab.freedesktop.org/mesa/drm/-/issues/62

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alistair Delva <adelva@google.com>
3 years agointel: Add support for JSL
Tejas Upadhyay [Mon, 1 Mar 2021 09:52:48 +0000 (09:52 +0000)]
intel: Add support for JSL

Add the PCI ID import for JSL.

V1 - Indentation
Signed-off-by: Tejas Upadhyay <tejaskumarx.surendrakumar.upadhyay@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
3 years agoxf86drm: cap number of reported devices by drmGetDevice(2)
Emil Velikov [Fri, 5 Feb 2021 00:41:13 +0000 (00:41 +0000)]
xf86drm: cap number of reported devices by drmGetDevice(2)

Do as the documentation says - when devices non NULL, cap the reported
devices to max_devices. Otherwise we risk out-of-bound access
for users of the API.

v2:
 - Fix this w/o breaking the API

v3:
 - Drop local variables, flip inverted conditional (Simon)

Issue: https://gitlab.freedesktop.org/mesa/drm/-/issues/56
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Simon Ser <contact@emersion.fr>
3 years agoRevert "xf86drm: cap number of reported devices by drmGetDevice(2)"
Emil Velikov [Thu, 11 Feb 2021 21:38:31 +0000 (21:38 +0000)]
Revert "xf86drm: cap number of reported devices by drmGetDevice(2)"

This reverts commit 8cb12a2528d795c45bba5f03b3486b4040fb0f45.

The commit fixed the OOB, yet it broke drmDevices2(0, NULL, 0) - aka we
did not return the total devices list.

Reviewed-by: Simon Ser <contact@emersion.fr>
3 years agoxf86drm: warn about GEM handle reference counting
Simon Ser [Thu, 21 Jan 2021 10:11:38 +0000 (11:11 +0100)]
xf86drm: warn about GEM handle reference counting

Users need to be careful when using drmPrimeHandleToFD or
drmPrimeFDToHandle directly. Mention GBM as a solution.

See [1] for an example mistake.

[1]: https://gitlab.freedesktop.org/drm/nouveau/-/issues/43#note_772661

Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
3 years agoxf86drmMode: add drmIsKMS
Simon Ser [Tue, 23 Feb 2021 17:19:33 +0000 (18:19 +0100)]
xf86drmMode: add drmIsKMS

If a device has a primary node, it doesn't necessarily mean it's
suitable for KMS usage. For instance, render-only drivers also
expose primary nodes.

The check is extracted from Weston [1].

The motivation for this new function is two-fold:

- Avoid an unnecessary GETRESOURCES call. To check whether a
  primary node is suitable for KMS, we don't actually need to
  retrieve the object IDs we just need to check the counts.
- Avoid confusion in user-space and make sure user-space implements
  the check properly. For instance, wlroots doesn't [2]: it uses
  drmGetVersion which succeeds with render-only drivers.

[1]: https://gitlab.freedesktop.org/wayland/weston/-/blob/master/libweston/backend-drm/drm.c#L2689
[2]: https://github.com/swaywm/wlroots/blob/a290d7a78dc36275e24e54f84570f37a66dc67a4/backend/session/session.c#L268

Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
3 years agoamdgpu_drm: sync up with the latest amdgpu_drm.h based on drm-next (https://cgit...
Leo Liu [Sun, 21 Feb 2021 21:13:51 +0000 (16:13 -0500)]
amdgpu_drm: sync up with the latest amdgpu_drm.h based on drm-next (https://cgit.freedesktop.org/drm/drm)

What are these headers?
Adding currently missing stuff from https://cgit.freedesktop.org/drm/drm/tree/include/uapi/drm/amdgpu_drm.h based on
the latest commit there:

commit f730f39eb981af249d57336b47cfe3925632a7fd (HEAD -> drm-next, tag: drm-next-2021-02-19, origin/drm-next, origin/HEAD)
Merge: 4f8ad4045b38 81ce8f04aa96
Author: Dave Airlie <airlied@redhat.com>
Date:   Fri Feb 19 13:54:29 2021 +1000

    Merge tag 'drm-intel-next-fixes-2021-02-18' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

Which headers go where?
From https://cgit.freedesktop.org/drm/drm/tree/include/uapi/drm/amdgpu_drm.h to
https://cgit.freedesktop.org/mesa/drm/tree/include/drm/amdgpu_drm.h

When and which headers to update?
If the kernel uapi drm header changes, the header here should be sync-ed.

When and how to update these files
The steps for generating this patch:

 - Switch to freedesktop drm-next kernel branch (https://cgit.freedesktop.org/drm/drm);
 - Install the headers via `make headers_install';
 - Copy from kernel "include/uapi/drm/amdgpu_drm.h" to libdrm "include/drm/amdgpu_drm.h";
 - generate the patch;

The commits from drm-next (https://cgit.freedesktop.org/drm/drm) are:

Mauro Carvalho Chehab (1)
c45dd3bda1c809eb120452597097e14a96b58c1f drm/amdgpu: fix some kernel-doc markups

Huang Rui(3)
6fbcb00c7984fa7d49af2c361453c0397cdea400 drm/amdgpu: add TOC firmware definition
1e483203965bdab466af0739c1edf7da07da241d drm/amdgpu: add uapi to define van gogh memory type
f7b2cdb23abf62bc3d33c2e0b0009a09412ff475 drm/amdgpu: add uapi to define van gogh series

Pierre-Eric Pelloux-Prayer(1)
16c642ec3fe9a144fbe1e97dc56f13a6308f1381 drm/amdgpu: new ids flag for tmz (v2)

Yong Zhao(1)
130c88931f6cbdb4513d307b4a13fcfff08a8041 drm/amdgpu: Improve the MTYPE comments

Signed-off-by: Leo Liu <leo.liu@amd.com>
3 years agointel: add INTEL_ADLS_IDS to the pciids list
Tejas Upadhyay [Thu, 18 Feb 2021 10:39:49 +0000 (10:39 +0000)]
intel: add INTEL_ADLS_IDS to the pciids list

This enables drm_intel_bufmgr on ADLS

Signed-off-by: Tejas Upadhyay <tejaskumarx.surendrakumar.upadhyay@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
3 years agointel: sync i915_pciids.h with kernel
Tejas Upadhyay [Thu, 18 Feb 2021 10:12:28 +0000 (10:12 +0000)]
intel: sync i915_pciids.h with kernel

Align with kernel commits:

0883d63b19bb ("drm/i915/adl_s: Add ADL-S platform info and PCI ids")
04057a1afc75 ("drm/i915: Sort EHL/JSL PCI IDs")
0e8e272f1368 ("drm/i915/ehl: Remove invalid PCI ID")
605f9c290c1a ("drm/i915: Sort ICL PCI IDs")
514dc424ce4f ("drm/i915: Sort CNL PCI IDs")
32d4ec9a1681 ("drm/i915: Sort CFL PCI IDs")
df3478af1d73 ("drm/i915: Sort CML PCI IDs")
cd988984cbea ("drm/i915: Sort KBL PCI IDs")
b04d36f73771 ("drm/i915: Sort SKL PCI IDs")
9c0b2d30441b ("drm/i915: Sort HSW PCI IDs")
79033a0a7898 ("drm/i915: Ocd the HSW PCI ID hex numbers")
cfb3db8fdae2 ("drm/i915: Try to fix the SKL GT3/4 vs. GT3e/4e comments")
03e399020cd2 ("drm/i915: Add SKL GT1.5 PCI IDs")
812f044df08c ("drm/i915: Reclassify SKL 0x1923 and 0x1927 as ULT")
194909a32aed ("drm/i915: Reclassify SKL 0x192a as GT3")
82e84284ab7d ("drm/i915: Update Haswell PCI IDs")
24ea098b7c0d ("drm/i915/jsl: Split EHL/JSL platform info and PCI ids")
b50b7991b739 ("drm/i915/dg1: add more PCI ids")
d452bd091e16 ("drm/i915: break TGL pci-ids in GT 1 & 2")
f2bde2546b81 ("drm/i915: Remove dubious Valleyview PCI IDs")
0883d63b19bb ("drm/i915/adl_s: Add ADL-S platform info and PCI ids")
04057a1afc75 ("drm/i915: Sort EHL/JSL PCI IDs")
0e8e272f1368 ("drm/i915/ehl: Remove invalid PCI ID")
605f9c290c1a ("drm/i915: Sort ICL PCI IDs")
514dc424ce4f ("drm/i915: Sort CNL PCI IDs")
32d4ec9a1681 ("drm/i915: Sort CFL PCI IDs")
df3478af1d73 ("drm/i915: Sort CML PCI IDs")
cd988984cbea ("drm/i915: Sort KBL PCI IDs")
b04d36f73771 ("drm/i915: Sort SKL PCI IDs")
9c0b2d30441b ("drm/i915: Sort HSW PCI IDs")
79033a0a7898 ("drm/i915: Ocd the HSW PCI ID hex numbers")
cfb3db8fdae2 ("drm/i915: Try to fix the SKL GT3/4 vs. GT3e/4e comments")
03e399020cd2 ("drm/i915: Add SKL GT1.5 PCI IDs")
812f044df08c ("drm/i915: Reclassify SKL 0x1923 and 0x1927 as ULT")
194909a32aed ("drm/i915: Reclassify SKL 0x192a as GT3")
82e84284ab7d ("drm/i915: Update Haswell PCI IDs")
24ea098b7c0d ("drm/i915/jsl: Split EHL/JSL platform info and PCI ids")
b50b7991b739 ("drm/i915/dg1: add more PCI ids")
d452bd091e16 ("drm/i915: break TGL pci-ids in GT 1 & 2")

Signed-off-by: Tejas Upadhyay <tejaskumarx.surendrakumar.upadhyay@intel.com>
Reviewed-by: Landwerlin, Lionel G <lionel.g.landwerlin@intel.com>
3 years agoamdgpu: update marketing names
Alex Deucher [Mon, 15 Feb 2021 15:02:15 +0000 (10:02 -0500)]
amdgpu: update marketing names

From 20.45 release.

Acked-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agoAvoid some compiler errors for tests/util/pattern.c
Victor Hugo Vianna Silva [Wed, 10 Feb 2021 18:16:10 +0000 (18:16 +0000)]
Avoid some compiler errors for tests/util/pattern.c

- Remove one unused variable.
- Convert two int-s into 'unsigned int'.
Motivated by a failed build of Chromium.

Reviewed-by: Emil Velikov <emil.l.velikov@gmail.com>
Signed-off-by: Victor Hugo Vianna Silva <victor.vianna10@gmail.com>
3 years agoxf86drm: cap number of reported devices by drmGetDevice(2)
Emil Velikov [Fri, 5 Feb 2021 00:41:13 +0000 (00:41 +0000)]
xf86drm: cap number of reported devices by drmGetDevice(2)

Do as the documentation says - cap the number of reported devices to the
requested amount - aka max_devices. Otherwise we risk out-of-bound access
for users of the API.

Issue: https://gitlab.freedesktop.org/mesa/drm/-/issues/56
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Reviewed-by: Simon Ser <contact@emersion.fr>
3 years agoMerge tag 'libdrm-2.4.104' into r-x86
Mauro Rossi [Fri, 5 Feb 2021 18:31:46 +0000 (19:31 +0100)]
Merge tag 'libdrm-2.4.104' into r-x86

libdrm-2.4.104

Change-Id: Ie8d817b3104f05d94d8c3f7258b7f23437a7eeec

3 years agotests/amdgpu/vcn: clean abundant codes
Sonny Jiang [Fri, 22 Jan 2021 01:44:03 +0000 (20:44 -0500)]
tests/amdgpu/vcn: clean abundant codes

Remove useless codes.

Signed-off-by: Sonny Jiang <sonny.jiang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
3 years agotests/amdgpu: add vcn test support for dimgrey_cavefish
James Zhu [Wed, 29 Jul 2020 16:55:52 +0000 (12:55 -0400)]
tests/amdgpu: add vcn test support for dimgrey_cavefish

add dimgrey_cavefish chip id in vcn test

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
3 years agotests/amdgpu: add vcn test support for navy_flounder
Tao Zhou [Tue, 23 Jun 2020 06:02:00 +0000 (14:02 +0800)]
tests/amdgpu: add vcn test support for navy_flounder

add navy_flounder chip id in vcn test

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Jiansong Chen <Jiansong.Chen@amd.com>
3 years agotests/util: Add mxsfb-drm driver
Fabio Estevam [Wed, 30 Dec 2020 18:29:33 +0000 (15:29 -0300)]
tests/util: Add mxsfb-drm driver

Add an entry for the "mxsfb-drm" driver, so that the test utilities
work with the mxsfb driver without passing the -M argument.

Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
3 years agomeson: Also search for rst2man.py
Heiko Becker [Fri, 15 Jan 2021 20:11:41 +0000 (21:11 +0100)]
meson: Also search for rst2man.py

That's what upstream docutils installs by default.

Signed-off-by: Heiko Becker <heirecka@exherbo.org>
3 years agoUse dep_rt in amdgpu/meson.build
Valentin Churavy [Tue, 17 Nov 2020 19:51:04 +0000 (19:51 +0000)]
Use dep_rt in amdgpu/meson.build

The amdgpu implementation uses `clock_gettime` so it needs to check whether it needs to link
against `-librt`.

Signed-off-by: Valentin Churavy <v.churavy@gmail.com>
3 years agoBump version to 2.4.104
Simon Ser [Mon, 11 Jan 2021 16:18:20 +0000 (17:18 +0100)]
Bump version to 2.4.104

Signed-off-by: Simon Ser <contact@emersion.fr>
3 years agoheaders: drm: Sync with drm-next
Antonin Décimo [Mon, 16 Nov 2020 11:26:56 +0000 (12:26 +0100)]
headers: drm: Sync with drm-next

Generated using make headers_install from the drm-next
tree - git://anongit.freedesktop.org/drm/drm
branch - drm-next
commit - b10733527bfd864605c33ab2e9a886eec317ec39

The changes were as follows (shortlog from
14d2bd53a47a7e1cb3e03d00a6b952734cf90f3f):

core: (drm_mode.h)

Alexander A. Klimov (1):
      drm: Replace HTTP links with HTTPS ones

Noralf Trønnes (1):
      drm: Add SPI connector type

Oleg Vasilev (1):
      drm: report dp downstream port type as a subconnector property

Simon Ser (1):
      drm: document that blobs are ref'counted

Uma Shankar (3):
      drm: Add HDR source metadata property
      drm: Fixed doc warnings in drm uapi header
      drm: Fix docbook warnings in hdr metadata helper structures

core: (drm_fourcc.h)

Adam Jackson (1):
      drm/fourcc: Fix undefined left shift in DRM_FORMAT_BIG_ENDIAN macros

Bas Nieuwenhuizen (2):
      drm/fourcc: Add AMD DRM modifiers.
      drm/fourcc: Fix modifier field mask for AMD modifiers.

Ben Davis (2):
      drm: drm_fourcc: add NV15, Q410, Q401 YUV formats
      drm: drm_fourcc: Add uncompressed AFBC modifier

Brian Starkey (1):
      drm: drm_fourcc: Add generic alias for 16_16_TILE modifier

Dave Airlie (1):
      Merge tag 'amd-drm-next-5.11-2020-11-05' of git://people.freedesktop.org/~agd5f/linux into drm-next

Dhinakaran Pandiyan (2):
      drm/framebuffer: Format modifier for Intel Gen-12 render compression
      drm/framebuffer: Format modifier for Intel Gen-12 media compression

James Jones (1):
      drm: Generalized NV Block Linear DRM format mod

Maarten Lankhorst (1):
      Backmerge remote-tracking branch 'drm/drm-next' into drm-misc-next

Matteo Franchin (1):
      drm/fourcc: Add AXBXGXRX106106106106 format

Mika Kahola (1):
      uapi/drm/drm_fourcc.h: Note on platform specificity for format modifiers

Neil Armstrong (2):
      drm/fourcc: Add modifier definitions for describing Amlogic Video Framebuffer Compression
      drm/fourcc: fix Amlogic Video Framebuffer Compression macro

Raymond Smith (1):
      drm/fourcc: Add Arm 16x16 block modifier

Simon Ser (4):
      drm/fourcc: document modifier uniqueness requirements
      drm: deprecate DRM_FORMAT_MOD_NONE
      drm/fourcc: add table describing AMD modifiers bit layout
      drm/fourcc: fix AMD modifiers PACKERS field doc

Signed-off-by: Antonin Décimo <antonin.decimo@gmail.com>
3 years agoRemove outdated comments about stdint.h
Simon Ser [Sat, 12 Dec 2020 11:39:36 +0000 (12:39 +0100)]
Remove outdated comments about stdint.h

We include stdint.h unconditionally in the header. We don't require
users to include it manually before xf86drmMode.h.

Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
3 years agoRemove definitions duplicated from drm_mode.h
Simon Ser [Fri, 11 Dec 2020 21:23:00 +0000 (22:23 +0100)]
Remove definitions duplicated from drm_mode.h

I don't exactly know why these were duplicated before. Maybe libdrm
didn't always vendored drm_mode.h from the kernel? In any case, we now
do, so instead of having copy-pasted definitions, just include our
vendored version which cannot be outdated.

Contrary to what the comment says, drm.h doesn't include drm_mode.h, so
we need to add the include.

Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
3 years agotests/amdgpu: Fix on FreeBSD
Emmanuel Vadot [Wed, 4 Nov 2020 15:13:56 +0000 (16:13 +0100)]
tests/amdgpu: Fix on FreeBSD

FreeBSD have endian.h under the sys directory.

Signed-off-by: Emmanuel Vadot <manu@FreeBSD.org>
3 years agotests/amdgpu: Fix a typo
Luben Tuikov [Mon, 30 Nov 2020 22:07:39 +0000 (17:07 -0500)]
tests/amdgpu: Fix a typo

Fix a typo: "TZM" --> "TMZ"

Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotests/modetest: remove bracket in dump_connectors()
Nicolas Caramelli [Tue, 24 Nov 2020 05:35:28 +0000 (06:35 +0100)]
tests/modetest: remove bracket in dump_connectors()

Signed-off-by: Nicolas Caramelli <caramelli.devel@gmail.com>
3 years agoxf86drmMode.h: use ANSI C99 arrays
Simon Ser [Thu, 26 Nov 2020 16:20:04 +0000 (17:20 +0100)]
xf86drmMode.h: use ANSI C99 arrays

This avoids the use of a GNU-specific extension in public headers. Also
see [1].

[1]: https://gitlab.freedesktop.org/mesa/drm/-/merge_requests/80#note_707458

Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>
3 years agoDocument drmModeConnection
Simon Ser [Wed, 18 Nov 2020 09:58:59 +0000 (10:58 +0100)]
Document drmModeConnection

Signed-off-by: Simon Ser <contact@emersion.fr>
3 years agoman: convert to reStructuredText
Simon Ser [Wed, 3 Jun 2020 09:58:07 +0000 (11:58 +0200)]
man: convert to reStructuredText

DocBook makes it hard to write and maintain docs. Hopefully
reStructuredText can make this less painful.

The man pages were converted from DocBook to reStructuredText via
Pandoc:

    pandoc -s -f docbook -t rst -o man/drm.7.rst man/drm.xml

And then manual editing to fixup e.g. references to other man pages. To
compare the result with the DocBook version, this command was used:

    rst2man man/drm-kms.7.rst | man -l -

Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Eric Engestrom <eric@engestrom.ch>
3 years agotests/etnaviv_2d_test: check whether the rendering is correct
Lubomir Rintel [Tue, 1 Dec 2020 20:38:29 +0000 (21:38 +0100)]
tests/etnaviv_2d_test: check whether the rendering is correct

Instead of always dumping the rendered picture, check whether it matches
the expectations. This makes more sense for automated testing.

Retain the ability to dump the picture instead of checking it when a
file name is given as an argument. This also removes use of a hardcoded
file name in a world writable directory, which is an unsafe thing to
do anyway.

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
3 years agotests/etnaviv_2d_test: pick the 2D core
Lubomir Rintel [Tue, 1 Dec 2020 20:38:28 +0000 (21:38 +0100)]
tests/etnaviv_2d_test: pick the 2D core

Run the test on a core capable of 2D rendering instead of hardcoding to
core zero.

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
3 years agotests/etnaviv_2d_test: explain the errors
Lubomir Rintel [Tue, 1 Dec 2020 20:38:27 +0000 (21:38 +0100)]
tests/etnaviv_2d_test: explain the errors

Just so that it's obvious what failed and why.

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
3 years agoandroid: build modetest as cc_binary module
Mauro Rossi [Thu, 19 Nov 2020 09:16:10 +0000 (10:16 +0100)]
android: build modetest as cc_binary module

[Why]
cc_test does not allow to correctly install the executable

[How]
cc_binary module is used instead of cc_test

3 years agotests/amdgpu/vcn: update to not use asic_id for Renoir
Thong Thai [Mon, 2 Nov 2020 15:15:07 +0000 (10:15 -0500)]
tests/amdgpu/vcn: update to not use asic_id for Renoir

Signed-off-by: Thong Thai <thong.thai@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
3 years agoBump version to 2.4.103
Dave Airlie [Wed, 4 Nov 2020 03:19:09 +0000 (13:19 +1000)]
Bump version to 2.4.103

Signed-off-by: Dave Airlie <airlied@redhat.com>
3 years agotests/util: Add imx-dcss driver
Lucas Stach [Fri, 20 Mar 2020 17:27:59 +0000 (18:27 +0100)]
tests/util: Add imx-dcss driver

This makes the test utilities work with the i.MX DCSS driver without the
necessity of using the -M argument.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
3 years agoxf86drm.c: Use integer logarithm.
Paul Gofman [Mon, 26 Oct 2020 13:11:20 +0000 (16:11 +0300)]
xf86drm.c: Use integer logarithm.

log() is affected by FP control word and can provide inaccurate result.

Fixes Killer Instinct under Wine not being able to find AMD vulkan
device.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Paul Gofman <pgofman@codeweavers.com>
3 years agotests/amdgpu: disable VCN test if no VCN ring available(v2)
Tianci.Yin [Fri, 23 Oct 2020 06:56:11 +0000 (14:56 +0800)]
tests/amdgpu: disable VCN test if no VCN ring available(v2)

If KMD has no VCN support, remove the VCN test.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tianci.Yin <tianci.yin@amd.com>
3 years agotests: add komeda to list of modules to look for for testing
Carsten Haitzler [Fri, 16 Oct 2020 10:00:15 +0000 (11:00 +0100)]
tests: add komeda to list of modules to look for for testing

komeda is one of the supported GPUs in the kernel tree so this adds it
to libdrm modules to look for in tests.

Signed-off-by: Carsten Haitzler <carsten.haitzler@arm.com>
3 years agoamdgpu: only enable security tests on raven family
Alex Deucher [Wed, 27 May 2020 21:33:10 +0000 (17:33 -0400)]
amdgpu: only enable security tests on raven family

It's the only asic with support at the moment.

Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotests/amdgpu: Secure bounce test (v4)
Luben Tuikov [Thu, 6 Feb 2020 00:40:46 +0000 (19:40 -0500)]
tests/amdgpu: Secure bounce test (v4)

Implement secure bounce test. Steps implemented
as outlined by Christian K.

v2: Remove gpu_info; add comment describing
    the purpose and steps of the test.
v3: Parameterize "secure" in amdgpu_bo_lcopy() and
    amdgpu_bo_move(). Set them both to 0.
    Allocate buffer Bob to be non-TMZ.
v4: Fix an off-by-one bug which was causing
    the test to segfault.

Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotests/amdgpu: Remove forward declarations
Luben Tuikov [Thu, 21 May 2020 19:32:08 +0000 (15:32 -0400)]
tests/amdgpu: Remove forward declarations

Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotest/amdgpu: enable security suite tests
Aaron Liu [Mon, 18 Nov 2019 08:49:47 +0000 (16:49 +0800)]
test/amdgpu: enable security suite tests

This patch enables security suite tests.

Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotest/amdgpu: add drm version checking for security suite
Aaron Liu [Mon, 18 Nov 2019 08:46:11 +0000 (16:46 +0800)]
test/amdgpu: add drm version checking for security suite

Adding drm version checking for security suite.
drm version need to be at least 3.37.

Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotests/amdgpu: add test to submit a sdma command with secure context
Aaron Liu [Fri, 15 Nov 2019 01:46:32 +0000 (09:46 +0800)]
tests/amdgpu: add test to submit a sdma command with secure context

This patch add test to submit a sdma command with secure context.

Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotests/amdgpu: add atomic dma command to verify the secure buffer (v2)
Aaron Liu [Fri, 15 Nov 2019 01:42:03 +0000 (09:42 +0800)]
tests/amdgpu: add atomic dma command to verify the secure buffer (v2)

DMA's atomic behavir is unlike GFX,If the comparing data is not
equal to destination data,
For GFX, loop again till gfx timeout(system hang).
For DMA, loop again till timer expired and then send interrupt.
So testcase can't use interrupt mechanism.
We take another way to verify. When the comparing data is not
equal to destination data, overwrite the source data to the destination
buffer. Otherwise, original destination data unchanged.
So if the bo_cpu data is overwritten, the result is passed.

Steps:
1. use linear write packet to write 0xdeadbeaf to secure buffer,
2. use atmoic packet and ATOMIC_CMPSWAP_RTN_32 opcode to compare
the cmp_data(0xdeadbeaf) to the written data which has been encrypted.

v2: add the case of (dest_data != cmp_data).

Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotests/amdgpu: add test to submit a gfx command with secure context
Aaron Liu [Thu, 14 Nov 2019 08:36:53 +0000 (16:36 +0800)]
tests/amdgpu: add test to submit a gfx command with secure context

This patch is to test the command submission with secure context.

Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotests/amdgpu: add atomic_mem cp_packet to verify the secure buffer
Aaron Liu [Thu, 14 Nov 2019 08:14:06 +0000 (16:14 +0800)]
tests/amdgpu: add atomic_mem cp_packet to verify the secure buffer

Secure buffer is only able to be read with trusted ip block.
So we need use GFX ip to read it back instead of CPU.
Steps:
1. use write_data packet to write 0xdeadbeaf to secure buffer,
2. use atmoic_mem packet and ATOMIC_CMPSWAP_RTN_32 opcode to compare
the cmp_data(0xdeadbeaf) to the written data which has been encrypted.
If the result is equal, then overwrite the src_data(0x12345678) to the
secure buffer and return directly. Otherwise loop again until gfx timeout
and the secure buffer data unchanged.

Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotests/amdgpu: expand secure param for exec_cs_helper (v2)
Aaron Liu [Thu, 14 Nov 2019 07:54:10 +0000 (15:54 +0800)]
tests/amdgpu: expand secure param for exec_cs_helper (v2)

This patch expands secure param for amdgpu_test_exec_cs_helper_raw.
The flag is transfered to kernel with cs.

v2: squash in change from context to IB flag

Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agotests/amdgpu: add device handle as input param for exec_cs_helper and write_linear_he...
Huang Rui [Fri, 16 Aug 2019 11:02:51 +0000 (19:02 +0800)]
tests/amdgpu: add device handle as input param for exec_cs_helper and write_linear_helper (v4)

This patch is to add add device handle as input param for exec_cs_helper
and write_linear_helper.

Because they are needed in security tests.

v2: fix typo that basic tests should be un-secure.
v3: refine the function implementation.
v4: remove amdgpu_cs_ctx_create3 calling.

Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>