[patch 16/43] dm crypt: wait for endio to complete before destruction

From: Greg KH
Date: Fri Mar 20 2009 - 19:31:58 EST


2.6.28-stable review patch. If anyone has any objections, please let us know.

------------------

From: Milan Broz <mbroz@xxxxxxxxxx>

commit b35f8caa0890169000fec22902290d9a15274cbd upstream.

The following oops has been reported when dm-crypt runs over a loop device.

...
[ 70.381058] Process loop0 (pid: 4268, ti=cf3b2000 task=cf1cc1f0 task.ti=cf3b2000)
...
[ 70.381058] Call Trace:
[ 70.381058] [<d0d76601>] ? crypt_dec_pending+0x5e/0x62 [dm_crypt]
[ 70.381058] [<d0d767b8>] ? crypt_endio+0xa2/0xaa [dm_crypt]
[ 70.381058] [<d0d76716>] ? crypt_endio+0x0/0xaa [dm_crypt]
[ 70.381058] [<c01a2f24>] ? bio_endio+0x2b/0x2e
[ 70.381058] [<d0806530>] ? dec_pending+0x224/0x23b [dm_mod]
[ 70.381058] [<d08066e4>] ? clone_endio+0x79/0xa4 [dm_mod]
[ 70.381058] [<d080666b>] ? clone_endio+0x0/0xa4 [dm_mod]
[ 70.381058] [<c01a2f24>] ? bio_endio+0x2b/0x2e
[ 70.381058] [<c02bad86>] ? loop_thread+0x380/0x3b7
[ 70.381058] [<c02ba8a1>] ? do_lo_send_aops+0x0/0x165
[ 70.381058] [<c013754f>] ? autoremove_wake_function+0x0/0x33
[ 70.381058] [<c02baa06>] ? loop_thread+0x0/0x3b7

When a table is being replaced, it waits for I/O to complete
before destroying the mempool, but the endio function doesn't
call mempool_free() until after completing the bio.

Fix it by swapping the order of those two operations.

The same problem occurs in dm.c with md referenced after dec_pending.
Again, we swap the order.

Signed-off-by: Milan Broz <mbroz@xxxxxxxxxx>
Signed-off-by: Alasdair G Kergon <agk@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxx>

---
drivers/md/dm-crypt.c | 17 ++++++++++-------
drivers/md/dm.c | 32 +++++++++++++++++++-------------
2 files changed, 29 insertions(+), 20 deletions(-)

--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -480,9 +480,12 @@ static int __noflush_suspending(struct m
static void dec_pending(struct dm_io *io, int error)
{
unsigned long flags;
+ int io_error;
+ struct bio *bio;
+ struct mapped_device *md = io->md;

/* Push-back supersedes any I/O errors */
- if (error && !(io->error > 0 && __noflush_suspending(io->md)))
+ if (error && !(io->error > 0 && __noflush_suspending(md)))
io->error = error;

if (atomic_dec_and_test(&io->io_count)) {
@@ -492,25 +495,28 @@ static void dec_pending(struct dm_io *io
* This must be handled before the sleeper on
* suspend queue merges the pushback list.
*/
- spin_lock_irqsave(&io->md->pushback_lock, flags);
- if (__noflush_suspending(io->md))
- bio_list_add(&io->md->pushback, io->bio);
+ spin_lock_irqsave(&md->pushback_lock, flags);
+ if (__noflush_suspending(md))
+ bio_list_add(&md->pushback, io->bio);
else
/* noflush suspend was interrupted. */
io->error = -EIO;
- spin_unlock_irqrestore(&io->md->pushback_lock, flags);
+ spin_unlock_irqrestore(&md->pushback_lock, flags);
}

end_io_acct(io);

- if (io->error != DM_ENDIO_REQUEUE) {
- blk_add_trace_bio(io->md->queue, io->bio,
+ io_error = io->error;
+ bio = io->bio;
+
+ free_io(md, io);
+
+ if (io_error != DM_ENDIO_REQUEUE) {
+ blk_add_trace_bio(md->queue, io->bio,
BLK_TA_COMPLETE);

- bio_endio(io->bio, io->error);
+ bio_endio(bio, io_error);
}
-
- free_io(io->md, io);
}
}

@@ -518,6 +524,7 @@ static void clone_endio(struct bio *bio,
{
int r = 0;
struct dm_target_io *tio = bio->bi_private;
+ struct dm_io *io = tio->io;
struct mapped_device *md = tio->io->md;
dm_endio_fn endio = tio->ti->type->end_io;

@@ -541,15 +548,14 @@ static void clone_endio(struct bio *bio,
}
}

- dec_pending(tio->io, error);
-
/*
* Store md for cleanup instead of tio which is about to get freed.
*/
bio->bi_private = md->bs;

- bio_put(bio);
free_tio(md, tio);
+ bio_put(bio);
+ dec_pending(io, error);
}

static sector_t max_io_len(struct mapped_device *md,
--- a/drivers/md/dm-crypt.c
+++ b/drivers/md/dm-crypt.c
@@ -568,19 +568,22 @@ static void crypt_inc_pending(struct dm_
static void crypt_dec_pending(struct dm_crypt_io *io)
{
struct crypt_config *cc = io->target->private;
+ struct bio *base_bio = io->base_bio;
+ struct dm_crypt_io *base_io = io->base_io;
+ int error = io->error;

if (!atomic_dec_and_test(&io->pending))
return;

- if (likely(!io->base_io))
- bio_endio(io->base_bio, io->error);
+ mempool_free(io, cc->io_pool);
+
+ if (likely(!base_io))
+ bio_endio(base_bio, error);
else {
- if (io->error && !io->base_io->error)
- io->base_io->error = io->error;
- crypt_dec_pending(io->base_io);
+ if (error && !base_io->error)
+ base_io->error = error;
+ crypt_dec_pending(base_io);
}
-
- mempool_free(io, cc->io_pool);
}

/*


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/