Skip to content

Commit

Permalink
drm/amdgpu/fence: Fix oops due to non-matching drm_sched init/fini
Browse files Browse the repository at this point in the history
commit 5ad7bbf upstream.

Currently amdgpu calls drm_sched_fini() from the fence driver sw fini
routine - such function is expected to be called only after the
respective init function - drm_sched_init() - was executed successfully.

Happens that we faced a driver probe failure in the Steam Deck
recently, and the function drm_sched_fini() was called even without
its counter-part had been previously called, causing the following oops:

amdgpu: probe of 0000:04:00.0 failed with error -110
BUG: kernel NULL pointer dereference, address: 0000000000000090
PGD 0 P4D 0
Oops: 0002 [#1] PREEMPT SMP NOPTI
CPU: 0 PID: 609 Comm: systemd-udevd Not tainted 6.2.0-rc3-gpiccoli torvalds#338
Hardware name: Valve Jupiter/Jupiter, BIOS F7A0113 11/04/2022
RIP: 0010:drm_sched_fini+0x84/0xa0 [gpu_sched]
[...]
Call Trace:
 <TASK>
 amdgpu_fence_driver_sw_fini+0xc8/0xd0 [amdgpu]
 amdgpu_device_fini_sw+0x2b/0x3b0 [amdgpu]
 amdgpu_driver_release_kms+0x16/0x30 [amdgpu]
 devm_drm_dev_init_release+0x49/0x70
 [...]

To prevent that, check if the drm_sched was properly initialized for a
given ring before calling its fini counter-part.

Notice ideally we'd use sched.ready for that; such field is set as the latest
thing on drm_sched_init(). But amdgpu seems to "override" the meaning of such
field - in the above oops for example, it was a GFX ring causing the crash, and
the sched.ready field was set to true in the ring init routine, regardless of
the state of the DRM scheduler. Hence, we ended-up using sched.ops as per
Christian's suggestion [0], and also removed the no_scheduler check [1].

[0] https://lore.kernel.org/amd-gfx/984ee981-2906-0eaf-ccec-9f80975cb136@amd.com/
[1] https://lore.kernel.org/amd-gfx/cd0e2994-f85f-d837-609f-7056d5fb7231@amd.com/

Fixes: 067f44c ("drm/amdgpu: avoid over-handle of fence driver fini in s3 test (v2)")
Suggested-by: Christian König <christian.koenig@amd.com>
Cc: Guchun Chen <guchun.chen@amd.com>
Cc: Luben Tuikov <luben.tuikov@amd.com>
Cc: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Signed-off-by: Guilherme G. Piccoli <gpiccoli@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
  • Loading branch information
guilhermepiccoli authored and Sasha Levin committed Feb 14, 2023
1 parent ba6014c commit 0b0928f
Showing 1 changed file with 7 additions and 1 deletion.
8 changes: 7 additions & 1 deletion drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
Original file line number Diff line number Diff line change
Expand Up @@ -579,7 +579,13 @@ void amdgpu_fence_driver_sw_fini(struct amdgpu_device *adev)
if (!ring || !ring->fence_drv.initialized)
continue;

if (!ring->no_scheduler)
/*
* Notice we check for sched.ops since there's some
* override on the meaning of sched.ready by amdgpu.
* The natural check would be sched.ready, which is
* set as drm_sched_init() finishes...
*/
if (ring->sched.ops)
drm_sched_fini(&ring->sched);

for (j = 0; j <= ring->fence_drv.num_fences_mask; ++j)
Expand Down

0 comments on commit 0b0928f

Please sign in to comment.