md: protect against crash upon fsync on ro array

commit bbfa57c upstream.

If an fsync occurs on a read-only array, we need to send a
completion for the IO and may not increment the active IO count.
Otherwise, we hit a bug trace and can't stop the MD array anymore.

By advice of Christoph Hellwig we return success upon a flush
request but we return -EROFS for other writes.
We detect flush requests by checking if the bio has zero sectors.

Signed-off-by: Sebastian Riemer <sebastian.riemer@profitbricks.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Ben Hutchings <ben@decadent.org.uk>
Cc: NeilBrown <neilb@suse.de>
Reported-by: Ben Hutchings <ben@decadent.org.uk>
Acked-by: Paul Menzel <paulepanter@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers/md/md.c

@@ -299,6 +299,10 @@ static int md_make_request(struct request_queue *q, struct bio *bio)
 		bio_io_error(bio);
 		return 0;
 	}
+	if (mddev->ro == 1 && unlikely(rw == WRITE)) {
+		bio_endio(bio, bio_sectors(bio) == 0 ? 0 : -EROFS);
+		return 0;
+	}
 	smp_rmb(); /* Ensure implications of  'active' are visible */
 	rcu_read_lock();
 	if (mddev->suspended) {