[PATCH v2 2/2] drm/xe: Add mutex locking to devcoredump

[John.C.Harrison at Intel.com]

From: John Harrison <John.C.Harrison at Intel.com>

There are now multiple places that can trigger a coredump. Some of
which can happen in parallel. There is already a check against
capturing multiple dumps sequentially, but without locking it doesn't
guarantee to work against concurrent dumps. And if two dumps do happen
in parallel, they can end up doing Bad Things such as one call stack
freeing the data the other call stack is still processing. Which leads
to a crashed kernel.

Further, it is possible for the DRM timeout to expire and trigger a
free of the capture while a user is still reading that capture out
through sysfs. Again leading to dodgy pointer problems.

So, add a mutext lock around the capture, read and free functions to
prevent inteference.

v2: Swap tiny scope spin_lock for larger scope mutex and fix
kernel-doc comment (review feedback from Matthe Brost)

Signed-off-by: John Harrison <John.C.Harrison at Intel.com>
---
 drivers/gpu/drm/xe/xe_devcoredump.c       | 26 +++++++++++++++++++++--
 drivers/gpu/drm/xe/xe_devcoredump_types.h |  4 +++-
 2 files changed, 27 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_devcoredump.c b/drivers/gpu/drm/xe/xe_devcoredump.c
index dd48745a8a46..0621754ddfd2 100644
--- a/drivers/gpu/drm/xe/xe_devcoredump.c
+++ b/drivers/gpu/drm/xe/xe_devcoredump.c
@@ -202,21 +202,29 @@ static ssize_t xe_devcoredump_read(char *buffer, loff_t offset,
 	if (!coredump)
 		return -ENODEV;
 
+	mutex_lock(&coredump->lock);
+
 	ss = &coredump->snapshot;
 
 	/* Ensure delayed work is captured before continuing */
 	flush_work(&ss->work);
 
-	if (!ss->read.buffer)
+	if (!ss->read.buffer) {
+		mutex_unlock(&coredump->lock);
 		return -ENODEV;
+	}
 
-	if (offset >= ss->read.size)
+	if (offset >= ss->read.size) {
+		mutex_unlock(&coredump->lock);
 		return 0;
+	}
 
 	byte_copied = count < ss->read.size - offset ? count :
 		ss->read.size - offset;
 	memcpy(buffer, ss->read.buffer + offset, byte_copied);
 
+	mutex_unlock(&coredump->lock);
+
 	return byte_copied;
 }
 
@@ -228,6 +236,8 @@ static void xe_devcoredump_free(void *data)
 	if (!data || !coredump_to_xe(coredump))
 		return;
 
+	mutex_lock(&coredump->lock);
+
 	cancel_work_sync(&coredump->snapshot.work);
 
 	xe_devcoredump_snapshot_free(&coredump->snapshot);
@@ -238,6 +248,8 @@ static void xe_devcoredump_free(void *data)
 	coredump->captured = false;
 	drm_info(&coredump_to_xe(coredump)->drm,
 		 "Xe device coredump has been deleted.\n");
+
+	mutex_unlock(&coredump->lock);
 }
 
 static void devcoredump_snapshot(struct xe_devcoredump *coredump,
@@ -312,8 +324,11 @@ void xe_devcoredump(struct xe_exec_queue *q, struct xe_sched_job *job, const cha
 	struct xe_devcoredump *coredump = &xe->devcoredump;
 	va_list varg;
 
+	mutex_lock(&coredump->lock);
+
 	if (coredump->captured) {
 		drm_dbg(&xe->drm, "Multiple hangs are occurring, but only the first snapshot was taken\n");
+		mutex_unlock(&coredump->lock);
 		return;
 	}
 
@@ -332,6 +347,7 @@ void xe_devcoredump(struct xe_exec_queue *q, struct xe_sched_job *job, const cha
 	dev_coredumpm_timeout(xe->drm.dev, THIS_MODULE, coredump, 0, GFP_KERNEL,
 			      xe_devcoredump_read, xe_devcoredump_free,
 			      XE_COREDUMP_TIMEOUT_JIFFIES);
+	mutex_unlock(&coredump->lock);
 }
 
 static void xe_driver_devcoredump_fini(void *arg)
@@ -343,6 +359,12 @@ static void xe_driver_devcoredump_fini(void *arg)
 
 int xe_devcoredump_init(struct xe_device *xe)
 {
+	int err;
+
+	err = drmm_mutex_init(&xe->drm, &xe->devcoredump.lock);
+	if (err)
+		return err;
+
 	return devm_add_action_or_reset(xe->drm.dev, xe_driver_devcoredump_fini, &xe->drm);
 }
 
diff --git a/drivers/gpu/drm/xe/xe_devcoredump_types.h b/drivers/gpu/drm/xe/xe_devcoredump_types.h
index e6234e887102..1a1d16a96b2d 100644
--- a/drivers/gpu/drm/xe/xe_devcoredump_types.h
+++ b/drivers/gpu/drm/xe/xe_devcoredump_types.h
@@ -80,7 +80,9 @@ struct xe_devcoredump_snapshot {
  * for reading the information.
  */
 struct xe_devcoredump {
-	/** @captured: The snapshot of the first hang has already been taken. */
+	/** @lock: protects access to entire structure */
+	struct mutex lock;
+	/** @captured: The snapshot of the first hang has already been taken */
 	bool captured;
 	/** @snapshot: Snapshot is captured at time of the first crash */
 	struct xe_devcoredump_snapshot snapshot;
-- 
2.47.0