summaryrefslogtreecommitdiffstats
path: root/drivers/block
diff options
context:
space:
mode:
authorConstantine Sapuntzakis <csapuntz@gmail.com>2006-06-23 02:06:08 -0700
committerLinus Torvalds <torvalds@g5.osdl.org>2006-06-23 07:43:09 -0700
commiteefe85ee350ece1bdf3e9a61e941d3394a528660 (patch)
tree7ed531a7946a6d715622a7d39be9042c03de961d /drivers/block
parent98bd34eaf1a7d1f2ed9c4e5d3a9664d3dcdd2159 (diff)
downloadkernel-crypto-eefe85ee350ece1bdf3e9a61e941d3394a528660.tar.gz
kernel-crypto-eefe85ee350ece1bdf3e9a61e941d3394a528660.tar.xz
kernel-crypto-eefe85ee350ece1bdf3e9a61e941d3394a528660.zip
[PATCH] drivers/block/loop.c: don't return garbage if LOOP_SET_STATUS not called
While writing a version of losetup, I ran into the problem that the loop device was returning total garbage. It turns out the problem was that this losetup was only issuing the LOOP_SET_FD ioctl and not issuing a subsequent LOOP_SET_STATUS ioctl. This losetup didn't have any special status to set, so it left out the call. The deeper cause is that loop_set_fd sets the transfer function to NULL, which causes no transfer to happen lo_do_transfer. This patch fixes the problem by setting transfer to transfer_none in loop_set_fd. Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'drivers/block')
-rw-r--r--drivers/block/loop.c2
1 files changed, 1 insertions, 1 deletions
diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 9c3b94e8f03..3c74ea729fc 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -818,7 +818,7 @@ static int loop_set_fd(struct loop_device *lo, struct file *lo_file,
lo->lo_device = bdev;
lo->lo_flags = lo_flags;
lo->lo_backing_file = file;
- lo->transfer = NULL;
+ lo->transfer = transfer_none;
lo->ioctl = NULL;
lo->lo_sizelimit = 0;
lo->old_gfp_mask = mapping_gfp_mask(mapping);