Skip to contentSkip to conversation

block: mark GFP_NOIO around sysfs ->store()

PATCH In review Ming Lei 2 participants 4 messages Jan 13, 2025

Subsystems (2)
lore
Share & tools
lore

Ming Lei

author
#0 lore raw

PATCH v2 1/1 block: mark GFP_NOIO around sysfs ->store()

sysfs ->store is called with queue freezed, meantime we have several ->store() callbacks(update_nr_requests, wbt, scheduler) to allocate memory with GFP_KERNEL which may run into direct reclaim code path, then potential deadlock can be caused.

Fix the issue by marking NOIO around sysfs ->store()

---
V2:
- add Closes & Reviewed-by & Fixes tag

 block/blk-sysfs.c | 3 +++
 1 file changed, 3 insertions(+)
block/blk-sysfs.c +3 -0
index e828be777206..e09b455874bf 100644
@@ -681,6 +681,7 @@ queue_attr_store(struct kobject *kobj, struct attribute *attr,
 	struct queue_sysfs_entry *entry = to_queue(attr);
 	struct gendisk *disk = container_of(kobj, struct gendisk, queue_kobj);
 	struct request_queue *q = disk->queue;
+	unsigned int noio_flag;
 	ssize_t res;
 
 	if (!entry->store_limit && !entry->store)
@@ -711,7 +712,9 @@ queue_attr_store(struct kobject *kobj, struct attribute *attr,
 
 	mutex_lock(&q->sysfs_lock);
 	blk_mq_freeze_queue(q);
+	noio_flag = memalloc_noio_save();
 	res = entry->store(disk, page, length);
+	memalloc_noio_restore(noio_flag);
 	blk_mq_unfreeze_queue(q);
 	mutex_unlock(&q->sysfs_lock);
 	return res;
Signature · 1 line
2.44.0
#1 lore raw

Hi!

Quoted 7 lines · Ming Lei

sysfs ->store is called with queue freezed, meantime we have several ->store() callbacks(update_nr_requests, wbt, scheduler) to allocate memory with GFP_KERNEL which may run into direct reclaim code path, then potential deadlock can be caused.

Fix the issue by marking NOIO around sysfs ->store()

Quoted 7 lines

Closes:
https://lore.kernel.org/linux-block/[email protected]/

Fixes bd166ef183c2 ("blk-mq-sched: add framework for MQ capable IO

schedulers")

Does this fix also the #2 lockdep splat in that email? Thanks,

Thomas

author
#2 lore raw
Quoted 18 lines · Thomas Hellström

Hi!

Quoted 7 lines · Ming Lei

sysfs ->store is called with queue freezed, meantime we have several ->store() callbacks(update_nr_requests, wbt, scheduler) to allocate memory with GFP_KERNEL which may run into direct reclaim code path, then potential deadlock can be caused.

Fix the issue by marking NOIO around sysfs ->store()

Quoted 7 lines

Closes:
https://lore.kernel.org/linux-block/[email protected]/

Fixes bd166ef183c2 ("blk-mq-sched: add framework for MQ capable IO

schedulers")

Quoted 1 line

Does this fix also the #2 lockdep splat in that email?

No.

The #2 splat fix has been merged to for-6.14/block, and this patch only covers the one reported in the Closes link.

https://lore.kernel.org/linux-block/[email protected]/

Thanks,
Ming

#3 lore raw
Quoted 23 lines · Ming Lei
Quoted 22 lines · Thomas Hellström

Hi!

Quoted 18 lines · Ming Lei

sysfs ->store is called with queue freezed, meantime we have several
->store() callbacks(update_nr_requests, wbt, scheduler) to allocate
memory with GFP_KERNEL which may run into direct reclaim code path,
then potential deadlock can be caused.

Fix the issue by marking NOIO around sysfs ->store()

Closes:
https://lore.kernel.org/linux-block/[email protected]/

Fixes bd166ef183c2 ("blk-mq-sched: add framework for MQ capable

IO
schedulers")

Quoted 7 lines
Quoted 1 line

Does this fix also the #2 lockdep splat in that email?

No.

The #2 splat fix has been merged to for-6.14/block, and this patch only
covers the one reported in the Closes link.

I actually reported two new splats in the Closes link.

(The second was found when using the suggested lockdep priming, but would ofc emerge sooner or later without it). I'm pretty sure Christoph's series was applied when that patch emerged, but I can retry if you want.

Thanks,
Thomas

Quoted 7 lines

https://lore.kernel.org/linux-block/[email protected]/

Thanks,
Ming