diff options
| author | Jens Axboe <axboe@kernel.dk> | 2013-04-08 10:12:11 +0200 |
|---|---|---|
| committer | Jens Axboe <axboe@kernel.dk> | 2013-04-08 10:12:11 +0200 |
| commit | c2fccc1c9f7c81700cbac2120a4ad5441dd37004 (patch) | |
| tree | 196f7a05401e936bfaa368721acebf34cf2e0dc1 /block | |
| parent | c66bb3f075cfe2d17b2427e96e043622db02759c (diff) | |
Revert "loop: cleanup partitions when detaching loop device"
This reverts commit 8761a3dc1f07b163414e2215a2cadbb4cfe2a107.
There are situations where the destruction path is called
with the bdev->bd_mutex already held, which then deadlocks in
loop_clr_fd(). The normal partition cleanup does a trylock()
on the mutex, but it'd be nice to have a more bullet proof
method in loop. So punt this more involved fix to the next
merge window, and just back out this buggy fix for now.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Diffstat (limited to 'block')
| -rw-r--r-- | block/partition-generic.c | 1 |
1 files changed, 0 insertions, 1 deletions
diff --git a/block/partition-generic.c b/block/partition-generic.c index ae95ee6a58aa..789cdea05893 100644 --- a/block/partition-generic.c +++ b/block/partition-generic.c @@ -257,7 +257,6 @@ void delete_partition(struct gendisk *disk, int partno) hd_struct_put(part); } -EXPORT_SYMBOL(delete_partition); static ssize_t whole_disk_show(struct device *dev, struct device_attribute *attr, char *buf) |