Hi,
I spotted that between commits 70d201a40823 and 052d534373b7 my GPU
begins randomly hanging when I open the GNOME shell activity screen.
I found a good reproducing script.
- Launch Elden Ring game
- Continue game (game world should be loaded)
- Press start (windows) button
Here GPU hanged with 99% probability, if GPU not hanged that press
start button several times for ensure.
And founded bad commit is looking so:
f7fe64ad0f22ff034f8ebcfbd7299ee9cc9b57d7 is the first bad commit
commit f7fe64ad0f22ff034f8ebcfbd7299ee9cc9b57d7
Author: Matthew Brost <[email protected]>
Date: Mon Oct 30 20:24:37 2023 -0700
drm/sched: Split free_job into own work item
Rather than call free_job and run_job in same work item have a dedicated
work item for each. This aligns with the design and intended use of work
queues.
v2:
- Test for DMA_FENCE_FLAG_TIMESTAMP_BIT before setting
timestamp in free_job() work item (Danilo)
v3:
- Drop forward dec of drm_sched_select_entity (Boris)
- Return in drm_sched_run_job_work if entity NULL (Boris)
v4:
- Replace dequeue with peek and invert logic (Luben)
- Wrap to 100 lines (Luben)
- Update comments for *_queue / *_queue_if_ready functions (Luben)
v5:
- Drop peek argument, blindly reinit idle (Luben)
- s/drm_sched_free_job_queue_if_ready/drm_sched_free_job_queue_if_done
(Luben)
- Update work_run_job & work_free_job kernel doc (Luben)
v6:
- Do not move drm_sched_select_entity in file (Luben)
Signed-off-by: Matthew Brost <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Luben Tuikov <[email protected]>
Signed-off-by: Luben Tuikov <[email protected]>
drivers/gpu/drm/scheduler/sched_main.c | 146 ++++++++++++++++++++++-----------
include/drm/gpu_scheduler.h | 4 +-
2 files changed, 101 insertions(+), 49 deletions(-)
Unfortunately GPU hangs still occurs even on 6.8-rc1 so why I wrote
here bug report.
GPU: Radeon 7900XTX
CPU: Ryzen 7950X
Full hardware specs are here: https://linux-hardware.org/?probe=9e5edb123e
Also I attach full bisect logs and kernel logs from each bisect step
in archives.
Who could dig into it, please?
--
Best Regards,
Mike Gavrilov.
On Wed, Jan 24, 2024 at 7:19 AM Mikhail Gavrilov
<[email protected]> wrote:
>
> Who could dig into it, please?
You decided to revert it?
https://lkml.org/lkml/2024/1/22/1866
Also I forgot to attach the kernel build .config in the previous
message. I'm going to fix it here.
It may be useful for reproducing my bug script.
--
Best Regards,
Mike Gavrilov.
On 1/24/2024 08:37, Mikhail Gavrilov wrote:
> On Wed, Jan 24, 2024 at 7:19 AM Mikhail Gavrilov
> <[email protected]> wrote:
>>
>> Who could dig into it, please?
>
> You decided to revert it?
> https://lkml.org/lkml/2024/1/22/1866
It's not a straight "git revert" on 6.8-rc1 because of some other
contextual changes.
I posted that as an RFC specifically "in case" that's the direction we
go and don't get a proper solution together.
Matthew also posted a debugging patch here for use with ftrace and the
GPU scheduler events: https://gitlab.freedesktop.org/drm/amd/-/issues/3124
I reproduced it with that as well and posted my ftrace results.
>
> Also I forgot to attach the kernel build .config in the previous
> message. I'm going to fix it here.
> It may be useful for reproducing my bug script.
>