Btrfs: fix hang when failing to submit bio of directIO

The hang is uncoverd by generic/019.

btrfs_endio_direct_write() skips the "finish_ordered_fn" part when it hits
an error, thus those added ordered extents will never get processed, which
block processes that waiting for them via btrfs_start_ordered_extent().

This fixes the above, and meanwhile finish_ordered_fn will do the space
accounting work.

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Tested-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 53afda0..0b9fb81 100644 (file)
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -7872,8 +7872,6 @@ static void btrfs_endio_direct_write(struct bio *bio, int err)
        struct bio *dio_bio;
        int ret;
 
-       if (err)
-               goto out_done;
 again:
        ret = btrfs_dec_test_first_ordered_pending(inode, &ordered,
                                                   &ordered_offset,
@@ -7896,7 +7894,6 @@ out_test:
                ordered = NULL;
                goto again;
        }
-out_done:
        dio_bio = dip->dio_bio;
 
        kfree(dip);