linux-2.6-microblaze.git
3 years agoamdgpu/pm: add extra info to SMU msg pre-check failed message
Darren Powell [Fri, 9 Apr 2021 22:28:24 +0000 (18:28 -0400)]
amdgpu/pm: add extra info to SMU msg pre-check failed message

Insert the value of the response to error message emitted when the
SMU msg pre-check failes

Signed-off-by: Darren Powell <darren.powell@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/pm: Update energy_accumulator in gpu metrics
Harish Kasiviswanathan [Fri, 23 Apr 2021 18:57:55 +0000 (14:57 -0400)]
drm/amd/pm: Update energy_accumulator in gpu metrics

Now that it is available.

v2: squash in typecast fix (Alex)

Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/dc: Fix a missing check bug in dm_dp_mst_detect()
Yingjie Wang [Fri, 9 Apr 2021 00:57:20 +0000 (17:57 -0700)]
drm/amd/dc: Fix a missing check bug in dm_dp_mst_detect()

In dm_dp_mst_detect(), We should check whether or not @connector
has been unregistered from userspace. If the connector is unregistered,
we should return disconnected status.

Fixes: 4562236b3bc0 ("drm/amd/dc: Add dc display driver (v2)")
Signed-off-by: Yingjie Wang <wangyingjie55@126.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: Reject non-zero src_y and src_x for video planes
Harry Wentland [Thu, 22 Apr 2021 23:10:52 +0000 (19:10 -0400)]
drm/amd/display: Reject non-zero src_y and src_x for video planes

[Why]
This hasn't been well tested and leads to complete system hangs on DCN1
based systems, possibly others.

The system hang can be reproduced by gesturing the video on the YouTube
Android app on ChromeOS into full screen.

[How]
Reject atomic commits with non-zero drm_plane_state.src_x or src_y values.

v2:
 - Add code comment describing the reason we're rejecting non-zero
   src_x and src_y
 - Drop gerrit Change-Id
 - Add stable CC
 - Based on amd-staging-drm-next

v3: removed trailing whitespace

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Cc: stable@vger.kernel.org
Cc: nicholas.kazlauskas@amd.com
Cc: amd-gfx@lists.freedesktop.org
Cc: alexander.deucher@amd.com
Cc: Roman.Li@amd.com
Cc: hersenxs.wu@amd.com
Cc: danny.wang@amd.com
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Hersen Wu <hersenxs.wu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: fix concurrent VM flushes on Vega/Navi v2
Christian König [Thu, 22 Apr 2021 11:11:39 +0000 (13:11 +0200)]
drm/amdgpu: fix concurrent VM flushes on Vega/Navi v2

Starting with Vega the hardware supports concurrent flushes
of VMID which can be used to implement per process VMID
allocation.

But concurrent flushes are mutual exclusive with back to
back VMID allocations, fix this to avoid a VMID used in
two ways at the same time.

v2: don't set ring to NULL

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Tested-by: James Zhu <James.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: remove AMDGPU_GEM_CREATE_SHADOW flag
Nirmoy Das [Thu, 22 Apr 2021 15:20:20 +0000 (17:20 +0200)]
drm/amdgpu: remove AMDGPU_GEM_CREATE_SHADOW flag

Remove unused AMDGPU_GEM_CREATE_SHADOW flag.
Userspace is never allowed to use this.

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: cleanup amdgpu_bo_create()
Nirmoy Das [Wed, 21 Apr 2021 16:09:46 +0000 (18:09 +0200)]
drm/amdgpu: cleanup amdgpu_bo_create()

Remove shadow bo related code as vm code is creating shadow bo
using proper API. Without shadow bo code, amdgpu_bo_create() is basically
a wrapper around amdgpu_bo_do_create(). So rename amdgpu_bo_do_create()
to amdgpu_bo_create().

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: create shadow bo using amdgpu_bo_create_shadow()
Nirmoy Das [Thu, 22 Apr 2021 14:54:01 +0000 (16:54 +0200)]
drm/amdgpu: create shadow bo using amdgpu_bo_create_shadow()

Shadow BOs are only needed for vm code so call amdgpu_bo_create_shadow()
directly instead of depending on amdgpu_bo_create().

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: remove unused vm context flags
Nirmoy Das [Thu, 22 Apr 2021 14:38:58 +0000 (16:38 +0200)]
drm/amdgpu: remove unused vm context flags

Remove unused AMDGPU_VM_CONTEXT_GFX and AMDGPU_VM_CONTEXT_COMPUTE
flags.

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: cleanup amdgpu_vm_init()
Nirmoy Das [Thu, 22 Apr 2021 10:53:16 +0000 (12:53 +0200)]
drm/amdgpu: cleanup amdgpu_vm_init()

Currently only way to create compute vm is through
amdgpu_vm_make_compute(). So vm_context isn't required
anymore for amdgpu_vm_init().

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: expose amdgpu_bo_create_shadow()
Nirmoy Das [Wed, 21 Apr 2021 15:27:46 +0000 (17:27 +0200)]
drm/amdgpu: expose amdgpu_bo_create_shadow()

Exposed amdgpu_bo_create_shadow() will be needed
for amdgpu_vm handling.

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: fix coding style and documentation in amdgpu_vram_mgr.c
Christian König [Fri, 12 Feb 2021 09:46:30 +0000 (10:46 +0100)]
drm/amdgpu: fix coding style and documentation in amdgpu_vram_mgr.c

No functional changes, just cleaning up some leftovers and improve
documentation.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Nirmoy Das <nirmoy.aiemd@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: fix coding style and documentation in amdgpu_gtt_mgr.c
Christian König [Thu, 11 Feb 2021 15:32:04 +0000 (16:32 +0100)]
drm/amdgpu: fix coding style and documentation in amdgpu_gtt_mgr.c

Avoid the forward define, fix coding style, add documentation.

No functional change.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Nirmoy Das <nirmoy.aiemd@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: remove redundant initialization to variable r
Colin Ian King [Thu, 22 Apr 2021 12:44:52 +0000 (13:44 +0100)]
drm/amdkfd: remove redundant initialization to variable r

The variable r is being initialized with a value that is never read
and it is being updated later with a new value. The initialization is
redundant and can be removed.

Addresses-Coverity: ("Unused value")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: fix uint32 variable compared to less than zero
Colin Ian King [Thu, 22 Apr 2021 12:31:58 +0000 (13:31 +0100)]
drm/amdkfd: fix uint32 variable compared to less than zero

Currently the call to kfd_process_gpuidx_from_gpuid is returning an
int value and this is being assigned to a uint32_t variable gpuidx
and this is being checked for a negative error return which is always
going to be false. Fix this by making gpuidx an int32_t. This makes
gpuidx also type consistent with the use of gpuidx from the callers.

Addresses-Coverity: ("Unsigned compared against 0")
Fixes: cda0f85bfa5e ("drm/amdkfd: refine migration policy with xnack on")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: set attribute access for default ranges
Alex Sierra [Wed, 21 Apr 2021 18:54:09 +0000 (13:54 -0500)]
drm/amdkfd: set attribute access for default ranges

Attribute access value for default ranges is set, based on
process xnack on/off.
XNACK ON has GPU access attribute for unregistered ranges through page
fault. While XNACK OFF has no access attribute for unregistered ranges.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: svm ranges creation for unregistered memory
Alex Sierra [Mon, 12 Apr 2021 18:35:18 +0000 (13:35 -0500)]
drm/amdkfd: svm ranges creation for unregistered memory

SVM ranges are created for unregistered memory, triggered
by page faults. These ranges are migrated/mapped to
GPU VRAM memory.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: extend xnack limit page fault timeout
Alex Sierra [Mon, 12 Apr 2021 18:34:57 +0000 (13:34 -0500)]
drm/amdgpu: extend xnack limit page fault timeout

Extending this timeout will prevent IH from storm interrupts coming
from SDMA while a page fault is active. Currently, on Aldebaran,
handling that many interrupts can take a lot of CPU time
(up to 4 seconds).
This eventually causes timeouts in other tasks.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: disable gfx ras by default in aldebaran
Hawking Zhang [Thu, 22 Apr 2021 13:58:08 +0000 (21:58 +0800)]
drm/amdgpu: disable gfx ras by default in aldebaran

aldebaran gfx ras is still under development

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Dennis Li <Dennis.Li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: add per-vmid-debug map_process_support
Jonathan Kim [Tue, 2 Mar 2021 16:28:36 +0000 (11:28 -0500)]
drm/amdkfd: add per-vmid-debug map_process_support

In order to support multi-process debugging, HWS PM4 packet
MAP_PROCESS requires an extension of 5 DWORDS to support targeting of
per-vmid SPI debug control registers as well as watch points per process.

Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/amdgpu: add cgls
Kenneth Feng [Thu, 22 Apr 2021 09:31:04 +0000 (17:31 +0800)]
drm/amd/amdgpu: add cgls

enable cgls to improve the runtime power efficiency.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: optimize gfx ras features flag clean
Stanley.Yang [Mon, 19 Apr 2021 03:33:10 +0000 (11:33 +0800)]
drm/amdgpu: optimize gfx ras features flag clean

Signed-off-by: Stanley.Yang <Stanley.Yang@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: Enable SDMA MGCG for Vangogh
Jinzhou Su [Wed, 21 Apr 2021 02:59:13 +0000 (10:59 +0800)]
drm/amdgpu: Enable SDMA MGCG for Vangogh

Add flags AMD_CG_SUPPORT_SDMA_MGCG for Vangogh.
Start to open sdma mgcg from firmware version 70.

Signed-off-by: Jinzhou Su <Jinzhou.Su@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: add support for ras init flags
John Clements [Wed, 21 Apr 2021 07:01:54 +0000 (15:01 +0800)]
drm/amdgpu: add support for ras init flags

conditionally configure ras for dgpu mode or poison propogation mode

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: refine gprs init shaders to check coverage
Dennis Li [Wed, 21 Apr 2021 04:50:22 +0000 (12:50 +0800)]
drm/amdgpu: refine gprs init shaders to check coverage

Add codes to check whether all SIMDs are covered, make sure that all
GPRs are initialized.

Signed-off-by: Dennis Li <Dennis.Li@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/amdgpu/amdgpu_cs: Repair some function naming disparity
Lee Jones [Fri, 16 Apr 2021 14:37:20 +0000 (15:37 +0100)]
drm/amd/amdgpu/amdgpu_cs: Repair some function naming disparity

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c:685: warning: expecting prototype for cs_parser_fini(). Prototype was for amdgpu_cs_parser_fini() instead
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c:1502: warning: expecting prototype for amdgpu_cs_wait_all_fence(). Prototype was for amdgpu_cs_wait_all_fences() instead
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c:1656: warning: expecting prototype for amdgpu_cs_find_bo_va(). Prototype was for amdgpu_cs_find_mapping() instead

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: Jerome Glisse <glisse@freedesktop.org>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Cc: linux-media@vger.kernel.org
Cc: linaro-mm-sig@lists.linaro.org
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/amdgpu/amdgpu_ring: Provide description for 'sched_score'
Lee Jones [Fri, 16 Apr 2021 14:37:18 +0000 (15:37 +0100)]
drm/amd/amdgpu/amdgpu_ring: Provide description for 'sched_score'

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c:169: warning: Function parameter or member 'sched_score' not described in 'amdgpu_ring_init'

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Cc: linux-media@vger.kernel.org
Cc: linaro-mm-sig@lists.linaro.org
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/amdgpu/amdgpu_ttm: Fix incorrectly documented function 'amdgpu_ttm_copy_mem_t...
Lee Jones [Fri, 16 Apr 2021 14:37:17 +0000 (15:37 +0100)]
drm/amd/amdgpu/amdgpu_ttm: Fix incorrectly documented function 'amdgpu_ttm_copy_mem_to_mem()'

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c:311: warning: expecting prototype for amdgpu_copy_ttm_mem_to_mem(). Prototype was for amdgpu_ttm_copy_mem_to_mem() instead

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: Jerome Glisse <glisse@freedesktop.org>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Cc: linux-media@vger.kernel.org
Cc: linaro-mm-sig@lists.linaro.org
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/amdgpu/amdgpu_gart: Correct a couple of function names in the docs
Lee Jones [Fri, 16 Apr 2021 14:37:16 +0000 (15:37 +0100)]
drm/amd/amdgpu/amdgpu_gart: Correct a couple of function names in the docs

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c:73: warning: expecting prototype for amdgpu_dummy_page_init(). Prototype was for amdgpu_gart_dummy_page_init() instead
 drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c:96: warning: expecting prototype for amdgpu_dummy_page_fini(). Prototype was for amdgpu_gart_dummy_page_fini() instead

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Nirmoy Das <nirmoy.das@amd.com>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Reviewed-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/amdgpu/amdgpu_fence: Provide description for 'sched_score'
Lee Jones [Fri, 16 Apr 2021 14:37:14 +0000 (15:37 +0100)]
drm/amd/amdgpu/amdgpu_fence: Provide description for 'sched_score'

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c:444: warning: Function parameter or member 'sched_score' not described in 'amdgpu_fence_driver_init_ring'

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: Jerome Glisse <glisse@freedesktop.org>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Cc: linux-media@vger.kernel.org
Cc: linaro-mm-sig@lists.linaro.org
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/radeon/radeon_device: Provide function name in kernel-doc header
Lee Jones [Fri, 16 Apr 2021 14:37:10 +0000 (15:37 +0100)]
drm/radeon/radeon_device: Provide function name in kernel-doc header

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/radeon/radeon_device.c:1101: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/amdgpu/amdgpu_device: Remove unused variable 'r'
Lee Jones [Fri, 16 Apr 2021 14:37:11 +0000 (15:37 +0100)]
drm/amd/amdgpu/amdgpu_device: Remove unused variable 'r'

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c: In function ‘amdgpu_device_suspend’:
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:3733:6: warning: variable ‘r’ set but not used [-Wunused-but-set-variable]

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Cc: linux-media@vger.kernel.org
Cc: linaro-mm-sig@lists.linaro.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: Add CONFIG_HSA_AMD_SVM
Felix Kuehling [Mon, 29 Mar 2021 22:49:12 +0000 (18:49 -0400)]
drm/amdkfd: Add CONFIG_HSA_AMD_SVM

Control whether to build SVM support into amdgpu with a Kconfig option.
This makes it easier to disable it in production kernels if this new
feature causes problems in production environments.

Use "depends on" instead of "select" for DEVICE_PRIVATE, as is
recommended for visible options.

Reviewed-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: Add SVM API support capability bits
Philip Yang [Thu, 10 Dec 2020 02:51:50 +0000 (21:51 -0500)]
drm/amdkfd: Add SVM API support capability bits

SVMAPISupported property added to HSA_CAPABILITY, the value match
HSA_CAPABILITY defined in Thunk spec:

SVMAPISupported: it will not be supported on older kernels that don't
have HMM or on systems with GFXv8 or older GPUs without support for
48-bit virtual addresses.

CoherentHostAccess property added to HSA_MEMORYPROPERTY, the value match
HSA_MEMORYPROPERTY defined in Thunk spec:

CoherentHostAccess: whether or not device memory can be coherently
accessed by the host CPU.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: multiple gpu migrate vram to vram
Felix Kuehling [Thu, 25 Feb 2021 04:57:33 +0000 (23:57 -0500)]
drm/amdkfd: multiple gpu migrate vram to vram

If prefetch range to gpu with acutal location is another gpu, or GPU
retry fault restore pages to migrate the range with acutal location is
gpu, then migrate from one gpu to another gpu.

Use system memory as bridge because sdma engine may not able to access
another gpu vram, use sdma of source gpu to migrate to system memory,
then use sdma of destination gpu to migrate from system memory to gpu.

Print out gpuid or gpuidx in debug messages.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: add svm range validate timestamp
Felix Kuehling [Thu, 25 Feb 2021 04:55:27 +0000 (23:55 -0500)]
drm/amdkfd: add svm range validate timestamp

With xnack on, add validate timestamp in order to handle GPU vm fault
from multiple GPUs.

If GPU retry fault need migrate the range to the best restore location,
use range validate timestamp to record system timestamp after range is
restored to update GPU page table.

Because multiple pages of same range have multiple retry fault, define
AMDGPU_SVM_RANGE_RETRY_FAULT_PENDING to the long time period that
pending retry fault may still comes after page table update, to skip
duplicate retry fault of same range.

If difference between system timestamp and range last validate timestamp
is bigger than AMDGPU_SVM_RANGE_RETRY_FAULT_PENDING, that means the
retry fault is from another GPU, then continue to handle retry fault
recover.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: refine migration policy with xnack on
Felix Kuehling [Thu, 25 Feb 2021 04:46:28 +0000 (23:46 -0500)]
drm/amdkfd: refine migration policy with xnack on

With xnack on, GPU vm fault handler decide the best restore location,
then migrate range to the best restore location and update GPU mapping
to recover the GPU vm fault.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: add svm_bo eviction to enable_signal cb
Alex Sierra [Tue, 28 Jul 2020 21:03:05 +0000 (16:03 -0500)]
drm/amdgpu: add svm_bo eviction to enable_signal cb

Add to amdgpu_amdkfd_fence.enable_signal callback, support
for svm_bo fence eviction.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: svm bo enable_signal call condition
Alex Sierra [Tue, 28 Jul 2020 21:04:41 +0000 (16:04 -0500)]
drm/amdgpu: svm bo enable_signal call condition

[why]
To support svm bo eviction mechanism.

[how]
If the BO crated has AMDGPU_AMDKFD_CREATE_SVM_BO flag set,
enable_signal callback will be called inside amdgpu_evict_flags.
This also causes gutting of the BO by removing all placements,
so that TTM won't actually do an eviction. Instead it will discard
the memory held by the BO. This is needed for HMM migration to user
mode system memory pages.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: add svm_bo eviction mechanism support
Felix Kuehling [Thu, 25 Feb 2021 04:24:53 +0000 (23:24 -0500)]
drm/amdkfd: add svm_bo eviction mechanism support

svm_bo eviction mechanism is different from regular BOs.
Every SVM_BO created contains one eviction fence and one
worker item for eviction process.
SVM_BOs can be attached to one or more pranges.
For SVM_BO eviction mechanism, TTM will start to call
enable_signal callback for every SVM_BO until VRAM space
is available.
Here, all the ttm_evict calls are synchronous, this guarantees
that each eviction has completed and the fence has signaled before
it returns.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: add param bit flag to create SVM BOs
Alex Sierra [Tue, 28 Jul 2020 18:38:29 +0000 (13:38 -0500)]
drm/amdgpu: add param bit flag to create SVM BOs

Add CREATE_SVM_BO define bit for SVM BOs.
Another define flag was moved to concentrate these
KFD type flags in one include file.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: add svm_bo reference for eviction fence
Alex Sierra [Tue, 28 Jul 2020 18:35:32 +0000 (13:35 -0500)]
drm/amdkfd: add svm_bo reference for eviction fence

[why]
As part of the SVM functionality, the eviction mechanism used for
SVM_BOs is different. This mechanism uses one eviction fence per prange,
instead of one fence per kfd_process.

[how]
A svm_bo reference to amdgpu_amdkfd_fence to allow differentiate between
SVM_BO or regular BO evictions. This also include modifications to set the
reference at the fence creation call.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: SVM API call to restore page tables
Alex Sierra [Tue, 12 May 2020 03:01:24 +0000 (22:01 -0500)]
drm/amdkfd: SVM API call to restore page tables

Use SVM API to restore page tables when retry fault and
compute context are enabled.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: page table restore through svm API
Felix Kuehling [Thu, 25 Feb 2021 04:02:21 +0000 (23:02 -0500)]
drm/amdkfd: page table restore through svm API

Page table restore implementation in SVM API. This is called from
the fault handler at amdgpu_vm. To update page tables through
the page fault retry IH.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: enable 48-bit IH timestamp counter
Alex Sierra [Fri, 15 Jan 2021 23:03:18 +0000 (17:03 -0600)]
drm/amdgpu: enable 48-bit IH timestamp counter

By default this timestamp is 32 bit counter. It gets
overflowed in around 10 minutes.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: invalidate tables on page retry fault
Felix Kuehling [Thu, 25 Feb 2021 03:34:11 +0000 (22:34 -0500)]
drm/amdkfd: invalidate tables on page retry fault

GPU page tables are invalidated by unmapping prange directly at
the mmu notifier, when page fault retry is enabled through
amdgpu_noretry global parameter. The restore page table is
performed at the page fault handler.

If xnack is on, we update GPU mappings after migration to avoid
unnecessary GPUVM faults.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: HMM migrate vram to ram
Felix Kuehling [Wed, 17 Mar 2021 04:24:12 +0000 (00:24 -0400)]
drm/amdkfd: HMM migrate vram to ram

If CPU page fault happens, HMM pgmap_ops callback migrate_to_ram start
migrate memory from vram to ram in steps:

1. migrate_vma_pages get vram pages, and notify HMM to invalidate the
pages, HMM interval notifier callback evict process queues
2. Allocate system memory pages
3. Use svm copy memory to migrate data from vram to ram
4. migrate_vma_pages copy pages structure from vram pages to ram pages
5. Return VM_FAULT_SIGBUS if migration failed, to notify application
6. migrate_vma_finalize put vram pages, page_free callback free vram
pages and vram nodes
7. Restore work wait for migration is finished, then update GPU page
table mapping to system memory, and resume process queues

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: HMM migrate ram to vram
Felix Kuehling [Thu, 25 Feb 2021 01:40:20 +0000 (20:40 -0500)]
drm/amdkfd: HMM migrate ram to vram

Register svm range with same address and size but perferred_location
is changed from CPU to GPU or from GPU to CPU, trigger migration the svm
range from ram to vram or from vram to ram.

If svm range prefetch location is GPU with flags
KFD_IOCTL_SVM_FLAG_HOST_ACCESS, validate the svm range on ram first,
then migrate it from ram to vram.

After migrating to vram is done, CPU access will have cpu page fault,
page fault handler migrate it back to ram and resume cpu access.

Migration steps:

1. migrate_vma_pages get svm range ram pages, notify the
interval is invalidated and unmap from CPU page table, HMM interval
notifier callback evict process queues
2. Allocate new pages in vram using TTM
3. Use svm copy memory to sdma copy data from ram to vram
4. migrate_vma_pages copy ram pages structure to vram pages structure
5. migrate_vma_finalize put ram pages to free ram pages and memory
6. Restore work wait for migration is finished, then update GPUs page
table mapping to new vram pages, resume process queues

If migrate_vma_setup failed to collect all ram pages of range, retry 3
times until success to start migration.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: copy memory through gart table
Philip Yang [Fri, 7 Feb 2020 22:08:04 +0000 (17:08 -0500)]
drm/amdkfd: copy memory through gart table

Use sdma linear copy to migrate data between ram and vram. The sdma
linear copy command uses kernel buffer function queue to access system
memory through gart table.

Use reserved gart table window 0 to map system page address, and vram
page address is direct mapping. Use the same kernel buffer function to
fill in gart table mapping, so this is serialized with memory copy by
sdma job submit. We only need wait for the last memory copy sdma fence
for larger buffer migration.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: support xgmi same hive mapping
Philip Yang [Tue, 23 Jun 2020 21:06:53 +0000 (17:06 -0400)]
drm/amdkfd: support xgmi same hive mapping

amdgpu_gmc_get_vm_pte use bo_va->is_xgmi same hive information to set
pte flags to update GPU mapping. Add local structure variable bo_va, and
update bo_va.is_xgmi, pass it to mapping->bo_va while mapping to GPU.

Assuming xgmi pstate is hi after boot.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: validate vram svm range from TTM
Felix Kuehling [Thu, 25 Feb 2021 00:56:09 +0000 (19:56 -0500)]
drm/amdkfd: validate vram svm range from TTM

If svm range perfetch location is not zero, use TTM to alloc
amdgpu_bo vram nodes to validate svm range, then map vram nodes to GPUs.

Use offset to sub allocate from the same amdgpu_bo to handle overlap
vram range while adding new range or unmapping range.

svm_bo has ref count to trace the shared ranges. If all ranges of shared
amdgpu_bo are migrated to ram, ref count becomes 0, then amdgpu_bo is
released, all ranges svm_bo is set to NULL.

To migrate range from ram back to vram, allocate the same amdgpu_bo
with previous offset if the range has svm_bo.

If prange migrate to VRAM, no CPU mapping exist, then process exit will
not have unmap callback for this prange to free prange and svm bo. Free
outstanding pranges from svms list before process is freed in
svm_range_list_fini.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: set memory limit to avoid OOM with HMM enabled
Philip Yang [Thu, 15 Apr 2021 21:43:32 +0000 (17:43 -0400)]
drm/amdkfd: set memory limit to avoid OOM with HMM enabled

HMM migration alloc sizeof(struct page) on system memory for each VRAM
page, it is 1GB system memory reserved for 64GB VRAM. To avoid
application OOM, increase system memory used size based on VRAM size of
all GPUs, then application alloc memory will fail if system memory usage
reach the limit.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Oak Zeng <Oak.Zeng@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: register HMM device private zone
Philip Yang [Fri, 7 Feb 2020 20:34:33 +0000 (15:34 -0500)]
drm/amdkfd: register HMM device private zone

Register vram memory as MEMORY_DEVICE_PRIVATE type resource, to
allocate vram backing pages for page migration.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: add ioctl to configure and query xnack retries
Alex Sierra [Thu, 28 May 2020 23:27:05 +0000 (18:27 -0500)]
drm/amdkfd: add ioctl to configure and query xnack retries

Xnack retries are used for page fault recovery. Some AMD chip
families support continuously retry while page table entries are invalid.
The driver must handle the page fault interrupt and fill in a valid entry
for the GPU to continue.

This ioctl allows to enable/disable XNACK retries per KFD process.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: add xnack enabled flag to kfd_process
Alex Sierra [Thu, 28 May 2020 23:03:15 +0000 (18:03 -0500)]
drm/amdkfd: add xnack enabled flag to kfd_process

XNACK mode controls the SQ RETRY_DISABLE setting that determines,
whether recoverable page faults can be supported on GFXv9 hardware.
Only on Aldebaran we can support different processes running with
different XNACK modes. On older chips all processes must use the same
RETRY_DISABLE setting. However, processes not relying on recoverable
page faults can work with RETRY enabled. This means XNACK off is always
available as a fallback so we can use the same mode on all GPUs in a
process.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: Enable retry faults unconditionally on Aldebaran
Felix Kuehling [Thu, 11 Feb 2021 20:57:20 +0000 (15:57 -0500)]
drm/amdgpu: Enable retry faults unconditionally on Aldebaran

This is needed to allow per-process XNACK mode selection in the SQ when
booting with XNACK off by default.

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Philip Yang <Philip.Yang@amd.com>
Tested-by: Alex Sierra <alex.sierra@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: svm range eviction and restore
Felix Kuehling [Wed, 24 Feb 2021 23:47:52 +0000 (18:47 -0500)]
drm/amdkfd: svm range eviction and restore

HMM interval notifier callback notify CPU page table will be updated,
stop process queues if the updated address belongs to svm range
registered in process svms objects tree. Scheduled restore work to
update GPU page table using new pages address in the updated svm range.

The restore worker flushes any deferred work to make sure it restores
an up-to-date svm_range_list.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: map svm range to GPUs
Felix Kuehling [Wed, 24 Feb 2021 23:29:06 +0000 (18:29 -0500)]
drm/amdkfd: map svm range to GPUs

Use amdgpu_vm_bo_update_mapping to update GPU page table to map or unmap
svm range system memory pages address to GPUs.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: export vm update mapping interface
Philip Yang [Mon, 23 Sep 2019 15:53:08 +0000 (11:53 -0400)]
drm/amdgpu: export vm update mapping interface

It will be used by kfd to map svm range to GPU, because svm range does
not have amdgpu_bo and bo_va, cannot use amdgpu_bo_update interface, use
amdgpu vm update interface directly.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: deregister svm range
Philip Yang [Wed, 8 Apr 2020 15:09:45 +0000 (11:09 -0400)]
drm/amdkfd: deregister svm range

When application explicitly call unmap or unmap from mmput when
application exit, driver will receive MMU_NOTIFY_UNMAP event to remove
svm range from process svms object tree and list first, unmap from GPUs
(in the following patch).

Split the svm ranges to handle partial unmapping of svm ranges. To
avoid deadlocks, updating MMU notifiers, range lists and interval trees
is done in a deferred worker. New child ranges are attached to their
parent range's child_list until the worker can update the
svm_range_list. svm_range_set_attr flushes deferred work and takes the
mmap_write_lock to guarantee that it has an up-to-date svm_range_list.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: validate svm range system memory
Philip Yang [Sat, 15 Feb 2020 20:02:49 +0000 (15:02 -0500)]
drm/amdkfd: validate svm range system memory

Use HMM to get system memory pages address, which will be used to
map to GPUs or migrate to vram.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: support larger svm range allocation
Philip Yang [Tue, 30 Mar 2021 18:02:53 +0000 (14:02 -0400)]
drm/amdkfd: support larger svm range allocation

For larger range allocation, if hmm_range_fault return -EBUSY, set retry
timeout based on 1 second for every 512MB, this is safe timeout value.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: add common HMM get pages function
Philip Yang [Tue, 25 Feb 2020 02:17:30 +0000 (21:17 -0500)]
drm/amdgpu: add common HMM get pages function

Move the HMM get pages function from amdgpu_ttm and to amdgpu_mn. This
common function will be used by new svm APIs.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: add svm ioctl GET_ATTR op
Philip Yang [Sun, 16 Feb 2020 17:42:00 +0000 (12:42 -0500)]
drm/amdkfd: add svm ioctl GET_ATTR op

Get the intersection of attributes over all memory in the given
range

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: register svm range
Philip Yang [Thu, 6 Feb 2020 18:43:56 +0000 (13:43 -0500)]
drm/amdkfd: register svm range

svm range structure stores the range start address, size, attributes,
flags, prefetch location and gpu bitmap which indicates which GPU this
range maps to. Same virtual address is shared by CPU and GPUs.

Process has svm range list which uses both interval tree and list to
store all svm ranges registered by the process. Interval tree is used by
GPU vm fault handler and CPU page fault handler to get svm range
structure from the specific address. List is used to scan all ranges in
eviction restore work.

No overlap range interval [start, last] exist in svms object interval
tree. If process registers new range which has overlap with old range,
the old range split into 2 ranges depending on the overlap happens at
head or tail part of old range.

Apply attributes preferred location, prefetch location, mapping flags,
migration granularity to svm range, store mapping gpu index into bitmap.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: add svm ioctl API
Philip Yang [Wed, 5 Feb 2020 22:59:15 +0000 (17:59 -0500)]
drm/amdkfd: add svm ioctl API

Add svm (shared virtual memory) ioctl data structure and API definition.

The svm ioctl API is designed to be extensible in the future. All
operations are provided by a single IOCTL to preserve ioctl number
space. The arguments structure ends with a variable size array of
attributes that can be used to set or get one or multiple attributes.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: helper to convert gpu id and idx
Alex Sierra [Wed, 8 Apr 2020 04:08:45 +0000 (23:08 -0500)]
drm/amdkfd: helper to convert gpu id and idx

svm range uses gpu bitmap to store which GPU svm range maps to.
Application pass driver gpu id to specify GPU, the helper is needed to
convert gpu id to gpu bitmap idx.

Access through kfd_process_device pointers array from kfd_process.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: Remove verify_access shortcut for KFD BOs
Felix Kuehling [Wed, 7 Apr 2021 22:48:09 +0000 (18:48 -0400)]
drm/amdgpu: Remove verify_access shortcut for KFD BOs

This shortcut is no longer needed with access managed properly by KFD.

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: Allow access for mmapping KFD BOs
Felix Kuehling [Wed, 7 Apr 2021 22:46:26 +0000 (18:46 -0400)]
drm/amdkfd: Allow access for mmapping KFD BOs

DRM render node file handles are used for CPU mapping of BOs using mmap
by the Thunk. It uses the DRM render node of the GPU where the BO was
allocated.

DRM allows mmap access automatically when it creates a GEM handle for a
BO. KFD BOs don't have GEM handles, so KFD needs to manage access
manually. Use drm_vma_node_allow to allow user mode to mmap BOs allocated
with kfd_ioctl_alloc_memory_of_gpu through the DRM render node that was
used in the kfd_ioctl_acquire_vm call for the same GPU.

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Philip Yang <philip.yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: Use drm_priv to pass VM from KFD to amdgpu
Felix Kuehling [Wed, 7 Apr 2021 22:19:58 +0000 (18:19 -0400)]
drm/amdkfd: Use drm_priv to pass VM from KFD to amdgpu

amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu needs the drm_priv to allow mmap
to access the BO through the corresponding file descriptor. The VM can
also be extracted from drm_priv, so drm_priv can replace the vm parameter
in the kfd2kgd interface.

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Philip Yang <philip.yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu/gmc9: remove dummy read workaround for newer chips
Alex Deucher [Fri, 16 Apr 2021 21:35:58 +0000 (17:35 -0400)]
drm/amdgpu/gmc9: remove dummy read workaround for newer chips

Aldebaran has a hw fix so no longer requires the workaround.

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: Add mem sync flag for IB allocated by SA
Jinzhou Su [Tue, 20 Apr 2021 08:17:14 +0000 (16:17 +0800)]
drm/amdgpu: Add mem sync flag for IB allocated by SA

The buffer of SA bo will be used by many cases. So it's better
to invalidate the cache of indirect buffer allocated by SA before
commit the IB.

Signed-off-by: Jinzhou Su <Jinzhou.Su@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: Fix SDMA RAS error reporting on Aldebaran
Mukul Joshi [Wed, 24 Mar 2021 15:51:35 +0000 (11:51 -0400)]
drm/amdgpu: Fix SDMA RAS error reporting on Aldebaran

Fix the following issues with SDMA RAS error reporting:
1. Read the EDC_COUNTER2 register also to fetch error counts
   for all sub-blocks in SDMA.
2. SDMA RAS on Aldebaran suports single-bit uncorrectable errors
   only. So, report error count in UE count instead of CE count.

Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-By: John Clements <John.Clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: Reset RAS error count and status regs
Mukul Joshi [Wed, 24 Mar 2021 15:36:33 +0000 (11:36 -0400)]
drm/amdgpu: Reset RAS error count and status regs

Reset the RAS error count and error status registers after
reading to prevent over reporting error counts on Aldebaran.

Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-By: John Clements <John.Clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agoRevert "drm/amdgpu: workaround the TMR MC address issue (v2)"
Oak Zeng [Thu, 11 Mar 2021 17:17:51 +0000 (11:17 -0600)]
Revert "drm/amdgpu: workaround the TMR MC address issue (v2)"

This reverts commit 2f055097daef498da57552f422f49de50a1573e6.
2f055097daef498da57552f422f49de50a1573e6 was a driver workaround
when PSP firmware was not ready. Now the PSP fw is ready so we
revert this driver workaround.

Signed-off-by: Oak Zeng <Oak.Zeng@amd.com>
Reviewed-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: 3.2.132
Aric Cyr [Sun, 11 Apr 2021 22:32:58 +0000 (18:32 -0400)]
drm/amd/display: 3.2.132

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: [FW Promotion] Release 0.0.62
Anthony Koo [Sat, 10 Apr 2021 14:03:54 +0000 (10:03 -0400)]
drm/amd/display: [FW Promotion] Release 0.0.62

Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: add helper for enabling mst stream features
Bing Guo [Mon, 5 Apr 2021 20:51:24 +0000 (16:51 -0400)]
drm/amd/display: add helper for enabling mst stream features

[Why]
Some MST devices uses different method to enable mst
specific stream features.

[How]
Add dm_helpers_mst_enable_stream features. This can be
modified later when we are ready to implement those features.

Signed-off-by: Bing Guo <bing.guo@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: Report Proper Quantization Range in AVI Infoframe
Dillon Varone [Fri, 9 Apr 2021 15:27:49 +0000 (11:27 -0400)]
drm/amd/display: Report Proper Quantization Range in AVI Infoframe

[Why?]
When a monitor does not set both QS and QY bits, DC does not
set Q0, Q1, QY0 and QY1 bits in AVI infoframe. Setting RGB bits
should be separate from setting YCC bits.

[How?]
Separate logic for setting RGB and YCC quantization range bits
in the AVI infoframe.

Signed-off-by: Dillon Varone <dillon.varone@amd.com>
Reviewed-by: Chris Park <Chris.Park@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: Fix call to pass bpp in 16ths of a bit
Dillon Varone [Fri, 9 Apr 2021 16:58:14 +0000 (12:58 -0400)]
drm/amd/display: Fix call to pass bpp in 16ths of a bit

[Why & How?]
Call to dc_dsc_compute_bandwidth_range should have min and max bpp
in 16ths of a bit.  Multiply min and max bpp from policy.

Signed-off-by: Dillon Varone <dillon.varone@amd.com>
Reviewed-by: Eryk Brol <Eryk.Brol@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: Fixed typo in function name.
David Galiffi [Thu, 8 Apr 2021 03:30:23 +0000 (23:30 -0400)]
drm/amd/display: Fixed typo in function name.

[How & Why]
Changed "prsent" to "present".

Signed-off-by: David Galiffi <David.Galiffi@amd.com>
Reviewed-by: Chris Park <Chris.Park@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: Always poll for rxstatus in authenticate
Nicholas Kazlauskas [Thu, 8 Apr 2021 16:21:47 +0000 (12:21 -0400)]
drm/amd/display: Always poll for rxstatus in authenticate

[Why]
Requirement from the spec - we shouldn't be potentially exiting out
early based on encryption status.

[How]
Drop the calls from HDCP1 and HDCP2 execution that exit out early
based on link encryption status.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: Add link rate optimization logs for ILR
Michael Strauss [Tue, 6 Apr 2021 16:20:51 +0000 (12:20 -0400)]
drm/amd/display: Add link rate optimization logs for ILR

[Why&How]
Add logs to verify ILR optimization behaviour on boot

Signed-off-by: Michael Strauss <michael.strauss@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: Unconditionally clear training pattern set after lt
Wesley Chalmers [Mon, 5 Apr 2021 23:35:37 +0000 (19:35 -0400)]
drm/amd/display: Unconditionally clear training pattern set after lt

[WHY]
While Link Training is being performed,
and the LTTPRs are in Non-LTTPR or LTTPR Transparent mode,
any DPCD registers besides those used for Link Training are not to be
accessed.

The spec defines the link training registers as DP_TRAINING_PATTERN_SET
(102h) to DP_TRAINING_LANE3_SET (106h), and DP_LANE0_1_STATUS (202h)
to DP_ADJUST_REQUEST_LANE2_3 (207h).

[HOW]
Move the current write to DPCD Address DP_LINK_TRAINING_PATTERN_SET out
of its conditional block.

Signed-off-by: Wesley Chalmers <Wesley.Chalmers@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: Fix FreeSync when RGB MPO in use
Aric Cyr [Wed, 17 Mar 2021 17:57:49 +0000 (13:57 -0400)]
drm/amd/display: Fix FreeSync when RGB MPO in use

[WHY]
We should skip programming manual trigger on non-primary planes when MPO is
enabled.

[HOW]
Implement an explicit mechanism for skipping manual trigger programming
for planes that shouldn't cause the frame to end.

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Acked-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: treat memory as a single-channel for asymmetric memory v2
Hugo Hu [Wed, 20 Jan 2021 07:54:11 +0000 (15:54 +0800)]
drm/amd/display: treat memory as a single-channel for asymmetric memory v2

Previous change had been reverted since it caused hang.
Remake change to avoid defect.

[Why]
1. Driver use umachannelnumber to calculate watermarks for stutter.
In asymmetric memory config, the actual bandwidth is less than
dual-channel. The bandwidth should be the same as single-channel.
2. We found single rank dimm need additional delay time for stutter.

[How]
Get information from each DIMM. Treat memory config as a single-channel
for asymmetric memory in bandwidth calculating.
Add additional delay time for single rank dimm.

Fixes: b8720ed0b87d32 ("drm/amd/display: System black screen hangs on driver load")
Signed-off-by: Hugo Hu <hugo.hu@amd.com>
Reviewed-by: Sung Lee <Sung.Lee@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: removed unused function dc_link_reallocate_mst_payload.
Robin Singh [Mon, 5 Apr 2021 15:27:24 +0000 (11:27 -0400)]
drm/amd/display: removed unused function dc_link_reallocate_mst_payload.

[Why]
Found that dc_link_reallocate_mst_payload is not used anymore
in any of the use case scenario.

[How]
removed dc_link_reallocate_mst_payload function definition
and declaration.

Signed-off-by: Robin Singh <robin.singh@amd.com>
Reviewed-by: Harry Wentland <Harry.Wentland@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: disable seamless boot for external DP
Anthony Wang [Mon, 5 Apr 2021 21:25:25 +0000 (17:25 -0400)]
drm/amd/display: disable seamless boot for external DP

[Why]
Primary feature use case is with eDP panels.

[How]
Fail seamless boot validation if display is not an eDP panel.

Signed-off-by: Anthony Wang <anthony1.wang@amd.com>
Reviewed-by: Martin Leung <Martin.Leung@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: add handling for hdcp2 rx id list validation
Dingchen (David) Zhang [Mon, 25 Jan 2021 23:05:50 +0000 (18:05 -0500)]
drm/amd/display: add handling for hdcp2 rx id list validation

[why]
the current implementation of hdcp2 rx id list validation does not
have handler/checker for invalid message status, e.g. HMAC, the V
parameter calculated from PSP not matching the V prime from Rx.

[how]
return a generic FAILURE for any message status not SUCCESS or
REVOKED.

Signed-off-by: Dingchen (David) Zhang <dingchen.zhang@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: update hdcp display using correct CP type.
Dingchen (David) Zhang [Fri, 8 Jan 2021 22:32:47 +0000 (17:32 -0500)]
drm/amd/display: update hdcp display using correct CP type.

[why]
currently we enforce to update hdcp display using TYPE0, but there
is case that connector CP type prop be TYPE1 instead of type0.

[how]
using the drm prop of CP type of the connector as input argument.

Signed-off-by: Dingchen (David) Zhang <dingchen.zhang@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: Add DSC check to seamless boot validation
Anthony Wang [Mon, 5 Apr 2021 21:13:58 +0000 (17:13 -0400)]
drm/amd/display: Add DSC check to seamless boot validation

[Why & How]
We want to immediately fail seamless boot validation if DSC is active,
as VBIOS currently does not support DSC timings. Add a check for
the relevant flag in dc_validate_seamless_boot_timing.

Signed-off-by: Anthony Wang <anthony1.wang@amd.com>
Reviewed-by: Martin Leung <Martin.Leung@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amd/display: fixed divide by zero kernel crash during dsc enablement
Robin Singh [Tue, 15 Dec 2020 00:14:48 +0000 (19:14 -0500)]
drm/amd/display: fixed divide by zero kernel crash during dsc enablement

[why]
During dsc enable, a divide by zero condition triggered the
kernel crash.

[how]
An IGT test, which enable the DSC, was crashing at the time of
restore the default dsc status, becaue of h_totals value
becoming 0. So add a check before divide condition. If h_total
is zero, gracefully ignore and set the default value.

kernel panic log:

[  128.758827] divide error: 0000 [#1] PREEMPT SMP NOPTI
[  128.762714] CPU: 5 PID: 4562 Comm: amd_dp_dsc Tainted: G        W         5.4.19-android-x86_64 #1
[  128.769728] Hardware name: ADVANCED MICRO DEVICES, INC. Mauna/Mauna, BIOS WMN0B13N Nov 11 2020
[  128.777695] RIP: 0010:hubp2_vready_at_or_After_vsync+0x37/0x7a [amdgpu]
[  128.785707] Code: 80 02 00 00 48 89 f3 48 8b 7f 08 b ......
[  128.805696] RSP: 0018:ffffad8f82d43628 EFLAGS: 00010246
......
[  128.857707] CR2: 00007106d8465000 CR3: 0000000426530000 CR4: 0000000000140ee0
[  128.865695] Call Trace:
[  128.869712] hubp3_setup+0x1f/0x7f [amdgpu]
[  128.873705] dcn20_update_dchubp_dpp+0xc8/0x54a [amdgpu]
[  128.877706] dcn20_program_front_end_for_ctx+0x31d/0x463 [amdgpu]
[  128.885706] dc_commit_state+0x3d2/0x658 [amdgpu]
[  128.889707] amdgpu_dm_atomic_commit_tail+0x4b3/0x1e7c [amdgpu]
[  128.897699] ? dm_read_reg_func+0x41/0xb5 [amdgpu]
[  128.901707] ? dm_read_reg_func+0x41/0xb5 [amdgpu]
[  128.905706] ? __is_insn_slot_addr+0x43/0x48
[  128.909706] ? fill_plane_buffer_attributes+0x29e/0x3dc [amdgpu]
[  128.917705] ? dm_plane_helper_prepare_fb+0x255/0x284 [amdgpu]
[  128.921700] ? usleep_range+0x7c/0x7c
[  128.925705] ? preempt_count_sub+0xf/0x18
[  128.929706] ? _raw_spin_unlock_irq+0x13/0x24
[  128.933732] ? __wait_for_common+0x11e/0x18f
[  128.937705] ? _raw_spin_unlock_irq+0x13/0x24
[  128.941706] ? __wait_for_common+0x11e/0x18f
[  128.945705] commit_tail+0x8b/0xd2 [drm_kms_helper]
[  128.949707] drm_atomic_helper_commit+0xd8/0xf5 [drm_kms_helper]
[  128.957706] amdgpu_dm_atomic_commit+0x337/0x360 [amdgpu]
[  128.961705] ? drm_atomic_check_only+0x543/0x68d [drm]
[  128.969705] ? drm_atomic_set_property+0x760/0x7af [drm]
[  128.973704] ? drm_mode_atomic_ioctl+0x6f3/0x85a [drm]
[  128.977705] drm_mode_atomic_ioctl+0x6f3/0x85a [drm]
[  128.985705] ? drm_atomic_set_property+0x7af/0x7af [drm]
[  128.989706] drm_ioctl_kernel+0x82/0xda [drm]
[  128.993706] drm_ioctl+0x225/0x319 [drm]
[  128.997707] ? drm_atomic_set_property+0x7af/0x7af [drm]
[  129.001706] ? preempt_count_sub+0xf/0x18
[  129.005713] amdgpu_drm_ioctl+0x4b/0x76 [amdgpu]
[  129.009705] vfs_ioctl+0x1d/0x2a
[  129.013705] do_vfs_ioctl+0x419/0x43d
[  129.017707] ksys_ioctl+0x52/0x71
[  129.021707] __x64_sys_ioctl+0x16/0x19
[  129.025706] do_syscall_64+0x78/0x85
[  129.029705] entry_SYSCALL_64_after_hwframe+0x44/0xa9

Signed-off-by: Robin Singh <robin.singh@amd.com>
Reviewed-by: Harry Wentland <Harry.Wentland@amd.com>
Reviewed-by: Robin Singh <Robin.Singh@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: fix GCR_GENERAL_CNTL offset for dimgrey_cavefish
Jiansong Chen [Mon, 19 Apr 2021 08:33:22 +0000 (16:33 +0800)]
drm/amdgpu: fix GCR_GENERAL_CNTL offset for dimgrey_cavefish

dimgrey_cavefish has similar gc_10_3 ip with sienna_cichlid,
so follow its registers offset setting.

Signed-off-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: resolve erroneous gfx_v9_4_2 prints
John Clements [Mon, 19 Apr 2021 03:23:07 +0000 (11:23 +0800)]
drm/amdgpu: resolve erroneous gfx_v9_4_2 prints

resolve bug on aldebaran where gfx error counts will
print on driver load when there are no errors present

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: fix a error injection failed issue
Dennis Li [Fri, 16 Apr 2021 14:41:11 +0000 (22:41 +0800)]
drm/amdgpu: fix a error injection failed issue

because "sscanf(str, "retire_page")" always return 0, if application use
the raw data for error injection, it always wrongly falls into "op ==
3". Change to use strstr instead.

Signed-off-by: Dennis Li <Dennis.Li@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: only harvest gcea/mmea error status in aldebaran
Hawking Zhang [Fri, 16 Apr 2021 09:30:12 +0000 (17:30 +0800)]
drm/amdgpu: only harvest gcea/mmea error status in aldebaran

In aldebaran, driver only needs to harvest SDP
RdRspStatus, WrRspStatus and first parity error
on RdRsp data. Check error type before harvest
error information.

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Stanley Yang <Stanley.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: only harvest gcea/mmea error status in arcturus
Hawking Zhang [Fri, 16 Apr 2021 09:34:13 +0000 (17:34 +0800)]
drm/amdgpu: only harvest gcea/mmea error status in arcturus

SDP RdRspStatus/WrRspStatus or first parity error on
RdRsp data can cause system fatal error in arcturus.
GPU will be freezed in such case.

Driver needs to harvest these error information before
reset the GPU. Check error type to avoid harvest normal
gcea/mmea information.

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Stanley Yang <Stanley.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: enable tmz on renoir asics
Huang Rui [Wed, 14 Apr 2021 10:45:54 +0000 (18:45 +0800)]
drm/amdgpu: enable tmz on renoir asics

The tmz functions are verified on renoir chips as well. So enable it by
default.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Tested-by: Lang Yu <Lang.Yu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdgpu: correct default gfx wdt timeout setting
Hawking Zhang [Fri, 16 Apr 2021 06:44:27 +0000 (14:44 +0800)]
drm/amdgpu: correct default gfx wdt timeout setting

When gfx wdt was configured to fatal_disable, the
timeout period should be configured to 0x0 (timeout
disabled)

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Dennis Li <Dennis.Li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
3 years agodrm/amdkfd: add edc error interrupt handle for poison propogate mode
Dennis Li [Wed, 14 Apr 2021 11:00:34 +0000 (19:00 +0800)]
drm/amdkfd: add edc error interrupt handle for poison propogate mode

In poison progogate mode, when driver receive the edc error interrupt
from SQ, driver should kill the process by pasid which is using the
poison data, and then trigger GPU reset.

Signed-off-by: Dennis Li <Dennis.Li@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>