On 2/25/25 11:18, Nilay Shroff wrote:

On 2/25/25 1:28 PM, Hannes Reinecke wrote:

quoted

On 2/24/25 14:30, Nilay Shroff wrote:

quoted

The bdi->ra_pages could be updated under q->limits_lock because it's
usually calculated from the queue limits by queue_limits_commit_update.
So protect reading/writing the sysfs attribute read_ahead_kb using
q->limits_lock instead of q->sysfs_lock.

Signed-off-by: Nilay Shroff <redacted>
---
   block/blk-sysfs.c | 16 ++++++++++------
   1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c
index 8f47d9f30fbf..228f81a9060f 100644
--- a/block/blk-sysfs.c
+++ b/block/blk-sysfs.c

@@ -93,9 +93,9 @@ static ssize_t queue_ra_show(struct gendisk *disk, char *page)
   {
       ssize_t ret;
   -    mutex_lock(&disk->queue->sysfs_lock);
+    mutex_lock(&disk->queue->limits_lock);
       ret = queue_var_show(disk->bdi->ra_pages << (PAGE_SHIFT - 10), page);
-    mutex_unlock(&disk->queue->sysfs_lock);
+    mutex_unlock(&disk->queue->limits_lock);
         return ret;
   }

@@ -111,12 +111,15 @@ queue_ra_store(struct gendisk *disk, const char *page, size_t count)
       ret = queue_var_store(&ra_kb, page, count);
       if (ret < 0)
           return ret;
-
-    mutex_lock(&q->sysfs_lock);
+    /*
+     * ->ra_pages is protected by ->limits_lock because it is usually
+     * calculated from the queue limits by queue_limits_commit_update.
+     */
+    mutex_lock(&q->limits_lock);
       memflags = blk_mq_freeze_queue(q);
       disk->bdi->ra_pages = ra_kb >> (PAGE_SHIFT - 10);
+    mutex_unlock(&q->limits_lock);
       blk_mq_unfreeze_queue(q, memflags);
-    mutex_unlock(&q->sysfs_lock);
   

Cf my comments to the previous patch: Ordering.

Here we take the lock _before_ 'freeze', with the previous patch we took
the lock _after_ 'freeze'.
Why?

Yes this is ->limits_lock which is different from ->elevator_lock. The ->limits_lock
is used by atomic update APIs queue_limits_start_update() and helpers. Here, the
order we follow is : acquire ->limits_lock followed by queue-freeze.

So even here in sysfs attribute store method we follow the same locking order.

Ah. Okay.

Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich