xen-blkback: fix dispatch_rw_block_io() error path

Commit 7708992 ("xen/blkback: Seperate the bio allocation and the bio
submission") consolidated the pendcnt updates to just a single write,
neglecting the fact that the error path relied on it getting set to 1
up front (such that the decrement in __end_block_io_op() would actually
drop the count to zero, triggering the necessary cleanup actions).

Also remove a misleading and a stale (after said commit) comment.

CC: stable@vger.kernel.org
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>

diff --git a/drivers/block/xen-blkback/blkback.c b/drivers/block/xen-blkback/blkback.c
index eaccc22..477a17c 100644 (file)
--- a/drivers/block/xen-blkback/blkback.c
+++ b/drivers/block/xen-blkback/blkback.c
@@ -1001,13 +1001,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
                bio->bi_end_io  = end_block_io_op;
        }
 
-       /*
-        * We set it one so that the last submit_bio does not have to call
-        * atomic_inc.
-        */
        atomic_set(&pending_req->pendcnt, nbio);
-
-       /* Get a reference count for the disk queue and start sending I/O */
        blk_start_plug(&plug);
 
        for (i = 0; i < nbio; i++)
@@ -1035,6 +1029,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
  fail_put_bio:
        for (i = 0; i < nbio; i++)
                bio_put(biolist[i]);
+       atomic_set(&pending_req->pendcnt, 1);
        __end_block_io_op(pending_req, -EINVAL);
        msleep(1); /* back off a bit */
        return -EIO;