Re: [PATCH v8 02/10] drm/xe/guc_submit: Introduce guc_exec_queue_reinit_kernel()

[Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>]

On 6/3/2026 3:17 AM, Raag Jadav wrote:
> In preparation of usecases which require re-initializing GuC submission
> after PCIe FLR, introduce guc_exec_queue_reinit_kernel() helper. This will
> restore kernel queues which might have been killed before PCIe FLR.
>
> Signed-off-by: Raag Jadav <raag.jadav@intel.com>
> Tested-by: Lukasz Laguna <lukasz.laguna@intel.com>
> ---
> v4: Teardown exec queues instead of mangling scheduler pending list (Matthew Brost)
> v5: Re-initialize kernel queues through submission backend (Matthew Brost)
> v8: Introduce xe_sched_reinit() (Daniele)
>      Maintain timeout per scheduler instance (Daniele)
>      s/reinit/reinit_kernel (Daniele)
> ---
>   drivers/gpu/drm/xe/xe_exec_queue_types.h    |  2 ++
>   drivers/gpu/drm/xe/xe_execlist.c            |  6 ++++++
>   drivers/gpu/drm/xe/xe_gpu_scheduler.c       |  6 ++++++
>   drivers/gpu/drm/xe/xe_gpu_scheduler.h       |  1 +
>   drivers/gpu/drm/xe/xe_gpu_scheduler_types.h |  5 +++++
>   drivers/gpu/drm/xe/xe_guc_submit.c          | 12 ++++++++++++
>   6 files changed, 32 insertions(+)
>
> diff --git a/drivers/gpu/drm/xe/xe_exec_queue_types.h b/drivers/gpu/drm/xe/xe_exec_queue_types.h
> index 2f5ccf294675..cca1ded68db7 100644
> --- a/drivers/gpu/drm/xe/xe_exec_queue_types.h
> +++ b/drivers/gpu/drm/xe/xe_exec_queue_types.h
> @@ -274,6 +274,8 @@ struct xe_exec_queue {
>   struct xe_exec_queue_ops {
>   	/** @init: Initialize exec queue for submission backend */
>   	int (*init)(struct xe_exec_queue *q);
> +	/** @reinit_kernel: Re-initialize kernel queue for submission backend */
> +	void (*reinit_kernel)(struct xe_exec_queue *q);
>   	/** @kill: Kill inflight submissions for backend */
>   	void (*kill)(struct xe_exec_queue *q);
>   	/** @fini: Undoes the init() for submission backend */
> diff --git a/drivers/gpu/drm/xe/xe_execlist.c b/drivers/gpu/drm/xe/xe_execlist.c
> index 9fb99c038ea8..f67b55782993 100644
> --- a/drivers/gpu/drm/xe/xe_execlist.c
> +++ b/drivers/gpu/drm/xe/xe_execlist.c
> @@ -403,6 +403,11 @@ static void execlist_exec_queue_destroy_async(struct work_struct *w)
>   	xe_exec_queue_fini(q);
>   }
>   
> +static void execlist_exec_queue_reinit_kernel(struct xe_exec_queue *q)
> +{
> +	/* NIY */
> +}
> +
>   static void execlist_exec_queue_kill(struct xe_exec_queue *q)
>   {
>   	/* NIY */
> @@ -466,6 +471,7 @@ static bool execlist_exec_queue_active(struct xe_exec_queue *q)
>   
>   static const struct xe_exec_queue_ops execlist_exec_queue_ops = {
>   	.init = execlist_exec_queue_init,
> +	.reinit_kernel = execlist_exec_queue_reinit_kernel,
>   	.kill = execlist_exec_queue_kill,
>   	.fini = execlist_exec_queue_fini,
>   	.destroy = execlist_exec_queue_destroy,
> diff --git a/drivers/gpu/drm/xe/xe_gpu_scheduler.c b/drivers/gpu/drm/xe/xe_gpu_scheduler.c
> index 67d8ce368486..1e38cfa19a58 100644
> --- a/drivers/gpu/drm/xe/xe_gpu_scheduler.c
> +++ b/drivers/gpu/drm/xe/xe_gpu_scheduler.c
> @@ -76,6 +76,7 @@ int xe_sched_init(struct xe_gpu_scheduler *sched,
>   	};
>   
>   	sched->ops = xe_ops;
> +	sched->timeout = timeout;
>   	spin_lock_init(&sched->msg_lock);
>   	INIT_LIST_HEAD(&sched->msgs);
>   	INIT_WORK(&sched->work_process_msg, xe_sched_process_msg_work);
> @@ -83,6 +84,11 @@ int xe_sched_init(struct xe_gpu_scheduler *sched,
>   	return drm_sched_init(&sched->base, &args);
>   }
>   
> +void xe_sched_reinit(struct xe_gpu_scheduler *sched)
> +{
> +	sched->base.timeout = sched->timeout;

IMO long term we should probably have a more generic drm_sched_reinit 
call in the common code, so that if new fields are added that need to be 
updated on re-init they are automatically handled, but we can probably 
do that as a follow-up.

> +}
> +
>   void xe_sched_fini(struct xe_gpu_scheduler *sched)
>   {
>   	xe_sched_submission_stop(sched);
> diff --git a/drivers/gpu/drm/xe/xe_gpu_scheduler.h b/drivers/gpu/drm/xe/xe_gpu_scheduler.h
> index 664c2db56af3..22e24b4b7171 100644
> --- a/drivers/gpu/drm/xe/xe_gpu_scheduler.h
> +++ b/drivers/gpu/drm/xe/xe_gpu_scheduler.h
> @@ -17,6 +17,7 @@ int xe_sched_init(struct xe_gpu_scheduler *sched,
>   		  long timeout, struct workqueue_struct *timeout_wq,
>   		  atomic_t *score, const char *name,
>   		  struct device *dev);
> +void xe_sched_reinit(struct xe_gpu_scheduler *sched);
>   void xe_sched_fini(struct xe_gpu_scheduler *sched);
>   
>   void xe_sched_submission_start(struct xe_gpu_scheduler *sched);
> diff --git a/drivers/gpu/drm/xe/xe_gpu_scheduler_types.h b/drivers/gpu/drm/xe/xe_gpu_scheduler_types.h
> index 63d9bf92583c..e73fdacca164 100644
> --- a/drivers/gpu/drm/xe/xe_gpu_scheduler_types.h
> +++ b/drivers/gpu/drm/xe/xe_gpu_scheduler_types.h
> @@ -51,6 +51,11 @@ struct xe_gpu_scheduler {
>   	spinlock_t				msg_lock;
>   	/** @work_process_msg: processes messages */
>   	struct work_struct		work_process_msg;
> +	/**
> +	 * @timeout: timeout for the scheduler instance, to be restored
> +	 * during exec queue re-initialization
> +	 */
> +	long				timeout;
>   };
>   
>   #define xe_sched_entity		drm_sched_entity
> diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c
> index ab501513d806..2c514f2bdda7 100644
> --- a/drivers/gpu/drm/xe/xe_guc_submit.c
> +++ b/drivers/gpu/drm/xe/xe_guc_submit.c
> @@ -2040,6 +2040,17 @@ static int guc_exec_queue_init(struct xe_exec_queue *q)
>   	return err;
>   }
>   
> +static void guc_exec_queue_reinit_kernel(struct xe_exec_queue *q)
> +{
> +	struct xe_gpu_scheduler *sched = &q->guc->sched;
> +
> +	xe_gt_assert(guc_to_gt(exec_queue_to_guc(q)), drm_sched_is_stopped(&sched->base));

This assert should be in xe_sched_reinit. with that:

Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>

Daniele

> +	xe_gt_assert(guc_to_gt(exec_queue_to_guc(q)), q->flags & EXEC_QUEUE_FLAG_KERNEL);
> +
> +	atomic_set(&q->guc->state, 0);
> +	xe_sched_reinit(sched);
> +}
> +
>   static void guc_exec_queue_kill(struct xe_exec_queue *q)
>   {
>   	trace_xe_exec_queue_kill(q);
> @@ -2282,6 +2293,7 @@ static bool guc_exec_queue_active(struct xe_exec_queue *q)
>    */
>   static const struct xe_exec_queue_ops guc_exec_queue_ops = {
>   	.init = guc_exec_queue_init,
> +	.reinit_kernel = guc_exec_queue_reinit_kernel,
>   	.kill = guc_exec_queue_kill,
>   	.fini = guc_exec_queue_fini,
>   	.destroy = guc_exec_queue_destroy,