[PATCH v2 2/3] drm/xe/xe_guc_submit: Allow lr exec queues to be banned
Matthew Brost
matthew.brost at intel.com
Fri Apr 5 02:17:05 UTC 2024
On Wed, Apr 03, 2024 at 12:19:49PM -0700, Jonathan Cavitt wrote:
> LR queues currently don't get banned during a GT/GuC reset because they
> lack a job. Though they don't have a job to detect the reset status of,
> it's still possible to tell when they should be banned by looking at the
> LRC: if the LRC head and tail don't match, then the exec queue should be
> banned and cleaned up.
>
> This also requires swapping the usage of xe_sched_tdr_queue_imm with
> xe_guc_exec_queue_trigger_cleanup, as the former is specific to non-lr
> exec queues.
>
> Suggested-by: Matthew Brost <matthew.brost at intel.com>
> Signed-off-by: Jonathan Cavitt <jonathan.cavitt at intel.com>
> Reviewed-by: Matthew Brost <matthew.brost at intel.com>
> ---
>
> v2:
> - Fix Subject line
> - Modify change slightly to remove need for "ban" boolean
I prefer the previous version to this version.
Matt
>
> drivers/gpu/drm/xe/xe_guc_submit.c | 12 ++++++++----
> 1 file changed, 8 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c
> index 294f40ab6aaf1..c93ba4bc611b8 100644
> --- a/drivers/gpu/drm/xe/xe_guc_submit.c
> +++ b/drivers/gpu/drm/xe/xe_guc_submit.c
> @@ -1428,12 +1428,16 @@ static void guc_exec_queue_stop(struct xe_guc *guc, struct xe_exec_queue *q)
> if (job) {
> if ((xe_sched_job_started(job) &&
> !xe_sched_job_completed(job)) ||
> - xe_sched_invalidate_job(job, 2)) {
> + xe_sched_invalidate_job(job, 2))
> trace_xe_sched_job_ban(job);
> - set_exec_queue_banned(q);
> - xe_sched_tdr_queue_imm(&q->guc->sched);
> - }
> + else
> + return;
> + } else if (!xe_exec_queue_is_lr(q) ||
> + (xe_lrc_ring_head(q->lrc) == q->lrc->ring.tail)) {
> + return;
> }
> + set_exec_queue_banned(q);
> + xe_guc_exec_queue_trigger_cleanup(q);
> }
> }
>
> --
> 2.25.1
>
More information about the Intel-xe
mailing list