block: make disk_block_events() properly wait for work cancellation

disk_block_events() should guarantee that the event work is not in
flight on return and once blocked it shouldn't issue further
cancellations.

Because there was no synchronization between the first blocker doing
cancel_delayed_work_sync() and the following blockers, the following
blockers could finish before cancellation was complete, which broke
both guarantees - event work could be in flight and cancellation could
happen after return.

This bug triggered WARN_ON_ONCE() in disk_clear_events() reported in
bug#34662.

  https://bugzilla.kernel.org/show_bug.cgi?id=34662

Fix it by adding an outer mutex which protects both block count
manipulation and work cancellation.

-v2: Use outer mutex instead of bit waitqueue per Linus.

Signed-off-by: Tejun Heo <tj@kernel.org>
Tested-by: Sitsofe Wheeler <sitsofe@yahoo.com>
Reported-by: Sitsofe Wheeler <sitsofe@yahoo.com>
Reported-by: Borislav Petkov <bp@alien8.de>
Reported-by: Meelis Roos <mroos@linux.ee>
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

1 files changed, 10 insertions, 0 deletions

diff --git a/block/genhd.c b/block/genhd.c
index ab0731d8976d..3608289c8ecd 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -1371,6 +1371,7 @@ struct disk_events {
         struct gendisk          *disk;          /* the associated disk */
         spinlock_t              lock;
 
+        struct mutex            block_mutex;    /* protects blocking */
         int                     block;          /* event blocking depth */
         unsigned int            pending;        /* events already sent out */
         unsigned int            clearing;       /* events being cleared */
@@ -1438,12 +1439,20 @@ void disk_block_events(struct gendisk *disk)
         if (!ev)
                 return;
 
+        /*
+         * Outer mutex ensures that the first blocker completes canceling
+         * the event work before further blockers are allowed to finish.
+         */
+        mutex_lock(&ev->block_mutex);
+
         spin_lock_irqsave(&ev->lock, flags);
         cancel = !ev->block++;
         spin_unlock_irqrestore(&ev->lock, flags);
 
         if (cancel)
                 cancel_delayed_work_sync(&disk->ev->dwork);
+
+        mutex_unlock(&ev->block_mutex);
 }
 
 static void __disk_unblock_events(struct gendisk *disk, bool check_now)
@@ -1751,6 +1760,7 @@ static void disk_add_events(struct gendisk *disk)
         INIT_LIST_HEAD(&ev->node);
         ev->disk = disk;
         spin_lock_init(&ev->lock);
+        mutex_init(&ev->block_mutex);
         ev->block = 1;
         ev->poll_msecs = -1;
         INIT_DELAYED_WORK(&ev->dwork, disk_events_workfn);