Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [Intel-xe] [PATCH] drm/xe: fix xe_device_mem_access_get() race
@ 2023-05-03 15:28 Matthew Auld
  2023-05-03 15:30 ` [Intel-xe] ✓ CI.Patch_applied: success for " Patchwork
                   ` (4 more replies)
  0 siblings, 5 replies; 11+ messages in thread
From: Matthew Auld @ 2023-05-03 15:28 UTC (permalink / raw)
  To: intel-xe; +Cc: Rodrigo Vivi

It looks like there is at least one race here, given that the
pm_runtime_suspended() check looks to return false if we are in the
process of suspending the device (RPM_SUSPENDING vs RPM_SUSPENDED). We
later also do xe_pm_runtime_get_if_active(), but since the device is
suspending or has now suspended, this doesn't do anything either.
Following from this we can potentially return from
xe_device_mem_access_get() with the device suspended or about to be,
leading to broken behaviour.

Attempt to fix this by always grabbing the runtime ref when our internal
ref transitions from 0 -> 1, and then wrap the whole thing with a lock
to ensure callers are serialized.

Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/258
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
 drivers/gpu/drm/xe/xe_device.c       | 22 ++++++++++++----------
 drivers/gpu/drm/xe/xe_device_types.h |  5 +++++
 drivers/gpu/drm/xe/xe_pm.c           |  9 ++-------
 drivers/gpu/drm/xe/xe_pm.h           |  2 +-
 4 files changed, 20 insertions(+), 18 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index 45d6e5ff47fd..5f6554bb34d2 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -209,6 +209,8 @@ struct xe_device *xe_device_create(struct pci_dev *pdev,
 
 	xe->ordered_wq = alloc_ordered_workqueue("xe-ordered-wq", 0);
 
+	drmm_mutex_init(&xe->drm, &xe->mem_access.lock);
+
 	err = xe_display_create(xe);
 	if (WARN_ON(err))
 		goto err_put;
@@ -404,26 +406,26 @@ u32 xe_device_ccs_bytes(struct xe_device *xe, u64 size)
 
 void xe_device_mem_access_get(struct xe_device *xe)
 {
-	bool resumed = xe_pm_runtime_resume_if_suspended(xe);
-	int ref = atomic_inc_return(&xe->mem_access.ref);
+	int ref;
 
+	mutex_lock(&xe->mem_access.lock);
+	ref = atomic_inc_return(&xe->mem_access.ref);
 	if (ref == 1)
-		xe->mem_access.hold_rpm = xe_pm_runtime_get_if_active(xe);
-
-	/* The usage counter increased if device was immediately resumed */
-	if (resumed)
-		xe_pm_runtime_put(xe);
+		xe->mem_access.hold_rpm = xe_pm_runtime_resume_and_get(xe);
+	mutex_unlock(&xe->mem_access.lock);
 
 	XE_WARN_ON(ref == S32_MAX);
 }
 
 void xe_device_mem_access_put(struct xe_device *xe)
 {
-	bool hold = xe->mem_access.hold_rpm;
-	int ref = atomic_dec_return(&xe->mem_access.ref);
+	int ref;
 
-	if (!ref && hold)
+	mutex_lock(&xe->mem_access.lock);
+	ref = atomic_dec_return(&xe->mem_access.ref);
+	if (!ref && xe->mem_access.hold_rpm)
 		xe_pm_runtime_put(xe);
+	mutex_unlock(&xe->mem_access.lock);
 
 	XE_WARN_ON(ref < 0);
 }
diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
index 1cb404e48aaa..e8d320f93852 100644
--- a/drivers/gpu/drm/xe/xe_device_types.h
+++ b/drivers/gpu/drm/xe/xe_device_types.h
@@ -257,6 +257,11 @@ struct xe_device {
 	 * triggering additional actions when they occur.
 	 */
 	struct {
+		/**
+		 * @lock: Serialize xe_device_mem_access users, and protect the
+		 * below internal state.
+		 */
+		struct mutex lock;
 		/** @ref: ref count of memory accesses */
 		atomic_t ref;
 		/** @hold_rpm: need to put rpm ref back at the end */
diff --git a/drivers/gpu/drm/xe/xe_pm.c b/drivers/gpu/drm/xe/xe_pm.c
index b7b57f10ba25..b2ffa001e6f7 100644
--- a/drivers/gpu/drm/xe/xe_pm.c
+++ b/drivers/gpu/drm/xe/xe_pm.c
@@ -210,14 +210,9 @@ int xe_pm_runtime_put(struct xe_device *xe)
 	return pm_runtime_put_autosuspend(xe->drm.dev);
 }
 
-/* Return true if resume operation happened and usage count was increased */
-bool xe_pm_runtime_resume_if_suspended(struct xe_device *xe)
+bool xe_pm_runtime_resume_and_get(struct xe_device *xe)
 {
-	/* In case we are suspended we need to immediately wake up */
-	if (pm_runtime_suspended(xe->drm.dev))
-		return !pm_runtime_resume_and_get(xe->drm.dev);
-
-	return false;
+	return !pm_runtime_resume_and_get(xe->drm.dev);
 }
 
 int xe_pm_runtime_get_if_active(struct xe_device *xe)
diff --git a/drivers/gpu/drm/xe/xe_pm.h b/drivers/gpu/drm/xe/xe_pm.h
index 6a885585f653..1b4c15b5e71a 100644
--- a/drivers/gpu/drm/xe/xe_pm.h
+++ b/drivers/gpu/drm/xe/xe_pm.h
@@ -19,7 +19,7 @@ int xe_pm_runtime_suspend(struct xe_device *xe);
 int xe_pm_runtime_resume(struct xe_device *xe);
 int xe_pm_runtime_get(struct xe_device *xe);
 int xe_pm_runtime_put(struct xe_device *xe);
-bool xe_pm_runtime_resume_if_suspended(struct xe_device *xe);
+bool xe_pm_runtime_resume_and_get(struct xe_device *xe);
 int xe_pm_runtime_get_if_active(struct xe_device *xe);
 
 #endif
-- 
2.40.0


^ permalink raw reply related	[flat|nested] 11+ messages in thread
end of thread, other threads:[~2023-05-05 13:22 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-05-03 15:28 [Intel-xe] [PATCH] drm/xe: fix xe_device_mem_access_get() race Matthew Auld
2023-05-03 15:30 ` [Intel-xe] ✓ CI.Patch_applied: success for " Patchwork
2023-05-03 15:31 ` [Intel-xe] ✓ CI.KUnit: " Patchwork
2023-05-03 15:35 ` [Intel-xe] ✓ CI.Build: " Patchwork
2023-05-03 15:58 ` [Intel-xe] ○ CI.BAT: info " Patchwork
2023-05-04  5:31 ` [Intel-xe] [PATCH] " Matthew Brost
2023-05-04 13:39   ` Thomas Hellström
2023-05-04 15:09     ` Rodrigo Vivi
2023-05-04 16:28       ` Matthew Auld
2023-05-05 13:22         ` Rodrigo Vivi
2023-05-04 15:12     ` Matthew Auld

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox