2009-04-23 07:42:52

by Chris Wright

[permalink] [raw]
Subject: [patch 040/100] dm snapshot: avoid having two exceptions for the same chunk

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

From: Mikulas Patocka <[email protected]>

upstream commit: 35bf659b008e83e725dcd30f542e38461dbb867c

We need to check if the exception was completed after dropping the lock.

After regaining the lock, __find_pending_exception checks if the exception
was already placed into &s->pending hash.

But we don't check if the exception was already completed and placed into
&s->complete hash. If the process waiting in alloc_pending_exception was
delayed at this point because of a scheduling latency and the exception
was meanwhile completed, we'd miss that and allocate another pending
exception for already completed chunk.

It would lead to a situation where two records for the same chunk exist
and potential data corruption because multiple snapshot I/Os to the
affected chunk could be redirected to different locations in the
snapshot.

Cc: [email protected]
Signed-off-by: Mikulas Patocka <[email protected]>
Signed-off-by: Alasdair G Kergon <[email protected]>
Signed-off-by: Chris Wright <[email protected]>
---
drivers/md/dm-snap.c | 13 +++++++++++++
1 file changed, 13 insertions(+)

--- a/drivers/md/dm-snap.c
+++ b/drivers/md/dm-snap.c
@@ -1080,6 +1080,13 @@ static int snapshot_map(struct dm_target
goto out_unlock;
}

+ e = lookup_exception(&s->complete, chunk);
+ if (e) {
+ free_pending_exception(pe);
+ remap_exception(s, e, bio, chunk);
+ goto out_unlock;
+ }
+
pe = __find_pending_exception(s, pe, chunk);
if (!pe) {
__invalidate_snapshot(s, -ENOMEM);
@@ -1226,6 +1233,12 @@ static int __origin_write(struct list_he
goto next_snapshot;
}

+ e = lookup_exception(&snap->complete, chunk);
+ if (e) {
+ free_pending_exception(pe);
+ goto next_snapshot;
+ }
+
pe = __find_pending_exception(snap, pe, chunk);
if (!pe) {
__invalidate_snapshot(snap, -ENOMEM);