[PATCH v2 2/2] md/raid1: serialize overlap io for writemostly disk

In behind mode, overlap bios for writemostly device are queued.
Those overlapped bios need to wait in waitqueue. They will be woken
up once the in-tree bio finishes.

Previously, using wait_event() would wake up all waiters simultaneously,
and they would compete for the tree lock. The bio which gets the lock
first will be handled, so the write sequence cannot be guaranteed.

For example:
bio1(100,200)
bio2(150,200)
bio3(150,300)

The write sequence of fast device is bio1,bio2,bio3. But the write
sequence of slow device could be bio1,bio3,bio2 due to lock competition.
This causes data corruption.

Use prepare_to_wait_exclusive() to ensure FIFO ordering. This adds
waiters to the tail of the waitqueue and wakes them from the head,
guaranteeing that bios are processed in the order they arrived.

Fixes: 3e148a320979 ("md/raid1: fix potential data inconsistency issue with write behind device")
Reviewed-by: Yu Kuai <[email protected]>
Signed-off-by: Xiao Ni <[email protected]>
---
v2: use prepare_to_wait_exclusive
 drivers/md/raid1.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index be2565dee420..a95085fb3015 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -85,12 +85,20 @@ static void wait_for_serialization(struct md_rdev *rdev, struct r1bio *r1_bio)
 	struct serial_info *si;
 	int idx = sector_to_idx(r1_bio->sector);
 	struct serial_in_rdev *serial = &rdev->serial[idx];
+	DEFINE_WAIT(wait);
 
 	if (WARN_ON(!mddev->serial_info_pool))
 		return;
 	si = mempool_alloc(mddev->serial_info_pool, GFP_NOIO);
-	wait_event(serial->serial_io_wait,
-		   check_and_add_serial(rdev, r1_bio, si, idx) == 0);
+
+	for (;;) {
+		prepare_to_wait_exclusive(&serial->serial_io_wait, &wait,
+					  TASK_UNINTERRUPTIBLE);
+		if (check_and_add_serial(rdev, r1_bio, si, idx) == 0)
+			break;
+		schedule();
+	}
+	finish_wait(&serial->serial_io_wait, &wait);
 }
 
 static void remove_serial(struct md_rdev *rdev, sector_t lo, sector_t hi)
-- 
2.50.1 (Apple Git-155)