[PATCH] drm/xe: Fix taking invalid lock on wedge
Lucas De Marchi
lucas.demarchi at intel.com
Thu Apr 3 05:38:05 UTC 2025
If device wedges on e.g. GuC upload, the submission is not yet enabled
and the state is not even initialized. Protect the wedge call so it does
nothing in this case. It fixes the following splat:
[] xe 0000:bf:00.0: [drm] device wedged, needs recovery
[] ------------[ cut here ]------------
[] DEBUG_LOCKS_WARN_ON(lock->magic != lock)
[] WARNING: CPU: 48 PID: 312 at kernel/locking/mutex.c:564 __mutex_lock+0x8a1/0xe60
...
[] RIP: 0010:__mutex_lock+0x8a1/0xe60
[] mutex_lock_nested+0x1b/0x30
[] xe_guc_submit_wedge+0x80/0x2b0 [xe]
Signed-off-by: Lucas De Marchi <lucas.demarchi at intel.com>
---
drivers/gpu/drm/xe/xe_guc_submit.c | 9 +++++++++
drivers/gpu/drm/xe/xe_guc_types.h | 5 +++++
2 files changed, 14 insertions(+)
diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c
index 31bc2022bfc2d..813c3c0bb2500 100644
--- a/drivers/gpu/drm/xe/xe_guc_submit.c
+++ b/drivers/gpu/drm/xe/xe_guc_submit.c
@@ -300,6 +300,8 @@ int xe_guc_submit_init(struct xe_guc *guc, unsigned int num_ids)
primelockdep(guc);
+ guc->submission_state.initialized = true;
+
return drmm_add_action_or_reset(&xe->drm, guc_submit_fini, guc);
}
@@ -834,6 +836,13 @@ void xe_guc_submit_wedge(struct xe_guc *guc)
xe_gt_assert(guc_to_gt(guc), guc_to_xe(guc)->wedged.mode);
+ /*
+ * If device is being wedged even before submission_state is
+ * initialized, there's nothing to do here.
+ */
+ if (!guc->submission_state.initialized)
+ return;
+
err = devm_add_action_or_reset(guc_to_xe(guc)->drm.dev,
guc_submit_wedged_fini, guc);
if (err) {
diff --git a/drivers/gpu/drm/xe/xe_guc_types.h b/drivers/gpu/drm/xe/xe_guc_types.h
index 63bac64429a5d..1fde7614fcc52 100644
--- a/drivers/gpu/drm/xe/xe_guc_types.h
+++ b/drivers/gpu/drm/xe/xe_guc_types.h
@@ -89,6 +89,11 @@ struct xe_guc {
struct mutex lock;
/** @submission_state.enabled: submission is enabled */
bool enabled;
+ /**
+ * @submission_state.initialized: mark when submission state is
+ * even initialized - before that not even the lock is valid
+ */
+ bool initialized;
/** @submission_state.fini_wq: submit fini wait queue */
wait_queue_head_t fini_wq;
} submission_state;
More information about the Intel-xe
mailing list