[PATCH 2/4] drm/xe/guc: Don't ban LR VM exec queues on PM suspend

[PATCH 2/4] drm/xe/guc: Don't ban LR VM exec queues on PM suspend

[Thomas Hellström]

When xe_guc_submit_stop() is called during an S3/S4 suspend or GT reset,
guc_exec_queue_stop() bans any user exec queue that has a job which has
started but not yet completed.  For normal (non-LR) exec queues this is
the correct behaviour: a started-but-incomplete job at reset time may
indicate a hung workload.

For exec queues attached to Long Running (LR) VMs the same condition is
always true during normal operation: LR jobs are designed to run
indefinitely and are never "completed" in the DRM scheduler sense —
they are preempted and resumed via the preempt-fence mechanism.
Banning such an exec queue on PM suspend permanently prevents the job
from restarting after resume, causing the userspace compute workload to
fail silently.

Fix this by skipping the ban for LR VM exec queues when a system
suspend or hibernation is in progress (pm_sleep_transition_in_progress()).
On GT reset the ban logic is preserved: a hung LR workload should still
be caught.

Fixes: f6375fb3aa94 ("drm/xe: Track LR jobs in DRM scheduler pending list")
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Tomasz Lis <tomasz.lis@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: <stable@vger.kernel.org> # v6.19+
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Assisted-by: GitHub_Copilot:claude-sonnet-4.6
---
 drivers/gpu/drm/xe/xe_guc_exec_queue_types.h |  8 ++++----
 drivers/gpu/drm/xe/xe_guc_submit.c           | 11 ++++++++++-
 2 files changed, 14 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_guc_exec_queue_types.h b/drivers/gpu/drm/xe/xe_guc_exec_queue_types.h
index 8ee76f958dc2..1207d51cf770 100644
--- a/drivers/gpu/drm/xe/xe_guc_exec_queue_types.h
+++ b/drivers/gpu/drm/xe/xe_guc_exec_queue_types.h
@@ -50,10 +50,10 @@ struct xe_guc_exec_queue {
 	/** @suspend_pending: a suspend of the exec_queue is pending */
 	bool suspend_pending;
 	/**
-	 * @suspend_count: number of active suspend requests, protected by
-	 * @sched.msg_lock. The exec_queue is kept suspended as long as this
-	 * is non-zero. Transitions 0->1 send the SUSPEND message; transitions
-	 * 1->0 send the RESUME message.
+	 * @suspend_count: Reference count of active suspend requests. The
+	 * exec_queue remains suspended while this is non-zero, allowing
+	 * multiple concurrent callers to independently hold a suspend without
+	 * prematurely re-enabling the queue. Protected by @sched.msg_lock.
 	 */
 	int suspend_count;
 	/**
diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c
index 50b622cf0c30..d1111b80fbed 100644
--- a/drivers/gpu/drm/xe/xe_guc_submit.c
+++ b/drivers/gpu/drm/xe/xe_guc_submit.c
@@ -9,6 +9,7 @@
 #include <linux/bitmap.h>
 #include <linux/circ_buf.h>
 #include <linux/dma-fence-array.h>
+#include <linux/suspend.h>
 
 #include <drm/drm_managed.h>
 
@@ -2274,8 +2275,16 @@ static void guc_exec_queue_stop(struct xe_guc *guc, struct xe_exec_queue *q)
 	 * Ban any engine (aside from kernel and engines used for VM ops) with a
 	 * started but not complete job or if a job has gone through a GT reset
 	 * more than twice.
+	 *
+	 * LR VM exec queues are excluded from this ban during PM suspend: their
+	 * jobs are intentionally long-running and are preempted and resumed via
+	 * the preempt-fence mechanism. Banning them on PM suspend would
+	 * permanently prevent the job from restarting after resume.
+	 * On GT reset however we do want to ban them, as that may indicate a
+	 * genuinely hung workload.
 	 */
-	if (!(q->flags & (EXEC_QUEUE_FLAG_KERNEL | EXEC_QUEUE_FLAG_VM))) {
+	if (!(q->flags & (EXEC_QUEUE_FLAG_KERNEL | EXEC_QUEUE_FLAG_VM)) &&
+	    !(q->vm && xe_vm_in_lr_mode(q->vm) && pm_sleep_transition_in_progress())) {
 		struct xe_sched_job *job = xe_sched_first_pending_job(sched);
 		bool ban = false;
 
-- 
2.54.0