This is a note to let you know that I've just added the patch titled
drm/xe: Only use reserved BCS instances for usm migrate exec queue
to the 6.8-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
drm-xe-only-use-reserved-bcs-instances-for-usm-migra.patch
and it can be found in the queue-6.8 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@xxxxxxxxxxxxxxx> know about it.
commit 2da6594eb5a33a254705a6c17ffcb6665e2e268c
Author: Matthew Brost <matthew.brost@xxxxxxxxx>
Date: Mon Apr 15 12:04:53 2024 -0700
drm/xe: Only use reserved BCS instances for usm migrate exec queue
[ Upstream commit c8ea2c31f5ea437199b239d76ad5db27343edb0c ]
The GuC context scheduling queue is 2 entires deep, thus it is possible
for a migration job to be stuck behind a fault if migration exec queue
shares engines with user jobs. This can deadlock as the migrate exec
queue is required to service page faults. Avoid deadlock by only using
reserved BCS instances for usm migrate exec queue.
Fixes: a043fbab7af5 ("drm/xe/pvc: Use fast copy engines as migrate engine on PVC")
Cc: Matt Roper <matthew.d.roper@xxxxxxxxx>
Cc: Niranjana Vishwanathapura <niranjana.vishwanathapura@xxxxxxxxx>
Signed-off-by: Matthew Brost <matthew.brost@xxxxxxxxx>
Link: https://patchwork.freedesktop.org/patch/msgid/20240415190453.696553-2-matthew.brost@xxxxxxxxx
Reviewed-by: Brian Welty <brian.welty@xxxxxxxxx>
(cherry picked from commit 04f4a70a183a688a60fe3882d6e4236ea02cfc67)
Signed-off-by: Thomas Hellström <thomas.hellstrom@xxxxxxxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
diff --git a/drivers/gpu/drm/xe/xe_migrate.c b/drivers/gpu/drm/xe/xe_migrate.c
index 7a6ad3469d748..0f77305c47c88 100644
--- a/drivers/gpu/drm/xe/xe_migrate.c
+++ b/drivers/gpu/drm/xe/xe_migrate.c
@@ -32,7 +32,6 @@
#include "xe_sync.h"
#include "xe_trace.h"
#include "xe_vm.h"
-#include "xe_wa.h"
/**
* struct xe_migrate - migrate context.
@@ -288,10 +287,6 @@ static int xe_migrate_prepare_vm(struct xe_tile *tile, struct xe_migrate *m,
}
/*
- * Due to workaround 16017236439, odd instance hardware copy engines are
- * faster than even instance ones.
- * This function returns the mask involving all fast copy engines and the
- * reserved copy engine to be used as logical mask for migrate engine.
* Including the reserved copy engine is required to avoid deadlocks due to
* migrate jobs servicing the faults gets stuck behind the job that faulted.
*/
@@ -305,8 +300,7 @@ static u32 xe_migrate_usm_logical_mask(struct xe_gt *gt)
if (hwe->class != XE_ENGINE_CLASS_COPY)
continue;
- if (!XE_WA(gt, 16017236439) ||
- xe_gt_is_usm_hwe(gt, hwe) || hwe->instance & 1)
+ if (xe_gt_is_usm_hwe(gt, hwe))
logical_mask |= BIT(hwe->logical_instance);
}
@@ -357,6 +351,10 @@ struct xe_migrate *xe_migrate_init(struct xe_tile *tile)
if (!hwe || !logical_mask)
return ERR_PTR(-EINVAL);
+ /*
+ * XXX: Currently only reserving 1 (likely slow) BCS instance on
+ * PVC, may want to revisit if performance is needed.
+ */
m->q = xe_exec_queue_create(xe, vm, logical_mask, 1, hwe,
EXEC_QUEUE_FLAG_KERNEL |
EXEC_QUEUE_FLAG_PERMANENT |
