Since 45ecaea73883 ("drm/sched: Partial revert of 'drm/sched: Keep s_fence->parent pointer'") still active jobs aren't put back in the pending list on drm_sched_start(), as they don't have a active parent fence anymore, so if the GPU is still working and the timeout is extended, all currently active jobs will be freed. To avoid prematurely freeing jobs that are still active on the GPU, don't block the scheduler until we are fully committed to actually reset the GPU. Cc: stable@xxxxxxxxxxxxxxx #6.0 Signed-off-by: Lucas Stach <l.stach@xxxxxxxxxxxxxx> --- The behavior change in the scheduler is unfortunate and at least deserves some updated documentation. This change aligns etnaviv with the behavior of other drivers and avoids the issue. --- drivers/gpu/drm/etnaviv/etnaviv_sched.c | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/etnaviv/etnaviv_sched.c b/drivers/gpu/drm/etnaviv/etnaviv_sched.c index 1ae87dfd19c4..35d7c2ef7a57 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_sched.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_sched.c @@ -38,15 +38,12 @@ static enum drm_gpu_sched_stat etnaviv_sched_timedout_job(struct drm_sched_job u32 dma_addr; int change; - /* block scheduler */ - drm_sched_stop(&gpu->sched, sched_job); - /* * If the GPU managed to complete this jobs fence, the timout is * spurious. Bail out. */ if (dma_fence_is_signaled(submit->out_fence)) - goto out_no_timeout; + return DRM_GPU_SCHED_STAT_NOMINAL; /* * If the GPU is still making forward progress on the front-end (which @@ -59,9 +56,12 @@ static enum drm_gpu_sched_stat etnaviv_sched_timedout_job(struct drm_sched_job change < 0 || change > 16) { gpu->hangcheck_dma_addr = dma_addr; gpu->hangcheck_fence = gpu->completed_fence; - goto out_no_timeout; + return DRM_GPU_SCHED_STAT_NOMINAL; } + /* block scheduler */ + drm_sched_stop(&gpu->sched, sched_job); + if(sched_job) drm_sched_increase_karma(sched_job); @@ -73,11 +73,6 @@ static enum drm_gpu_sched_stat etnaviv_sched_timedout_job(struct drm_sched_job drm_sched_start(&gpu->sched, true); return DRM_GPU_SCHED_STAT_NOMINAL; - -out_no_timeout: - /* restart scheduler after GPU is usable again */ - drm_sched_start(&gpu->sched, true); - return DRM_GPU_SCHED_STAT_NOMINAL; } static void etnaviv_sched_free_job(struct drm_sched_job *sched_job) -- 2.39.2