[PATCH 1/7] dm cache: fix null-deref with concurrent writes in passthrough mode

[Ming-Hung Tsai]

In passthrough mode, when dm-cache starts to invalidate a cache
entry and bio prison cell lock fails due to concurrent write to
the same cached block, mg->cell remains NULL. The error path in
invalidate_complete() attempts to unlock and free the cell
unconditionally, causing a NULL pointer dereference:

KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007]
CPU: 0 UID: 0 PID: 134 Comm: fio Not tainted 6.19.0-rc7 #3 PREEMPT
RIP: 0010:dm_cell_unlock_v2+0x3f/0x210
<snip>
Call Trace:
 invalidate_complete+0xef/0x430
 map_bio+0x130f/0x1a10
 cache_map+0x320/0x6b0
 __map_bio+0x458/0x510
 dm_submit_bio+0x40e/0x16d0
 __submit_bio+0x419/0x870
<snip>

Reproduce steps:

1. Create a cache device

dmsetup create cmeta --table "0 8192 linear /dev/sdc 0"
dmsetup create cdata --table "0 131072 linear /dev/sdc 8192"
dmsetup create corig --table "0 262144 linear /dev/sdc 262144"
dd if=/dev/zero of=/dev/mapper/cmeta bs=4k count=1 oflag=direct
dmsetup create cache --table "0 262144 cache /dev/mapper/cmeta \
/dev/mapper/cdata /dev/mapper/corig 128 2 metadata2 writethrough smq 0"

2. Promote the first data block into cache

fio --filename=/dev/mapper/cache --name=populate --rw=write --bs=4k \
--direct=1 --size=64k

3. Reload the cache into passthrough mode

dmsetup suspend cache
dmsetup reload cache --table "0 262144 cache /dev/mapper/cmeta \
/dev/mapper/cdata /dev/mapper/corig 128 2 metadata2 passthrough smq 0"
dmsetup resume cache

4. Write to the first cached block concurrently

fio --filename=/dev/mapper/cache --name test --rw=randwrite --bs=4k \
--randrepeat=0 --direct=1 --numjobs=2 --size 64k

Fix by checking if mg->cell is valid before attempting to unlock it.

Fixes: b29d4986d0da ("dm cache: significant rework to leverage 
dm-bio-prison-v2")
Signed-off-by: Ming-Hung Tsai <mt...@redhat.com>
---
 drivers/md/dm-cache-target.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/md/dm-cache-target.c b/drivers/md/dm-cache-target.c
index a10d75a562db..a2e07b124901 100644
--- a/drivers/md/dm-cache-target.c
+++ b/drivers/md/dm-cache-target.c
@@ -1462,8 +1462,10 @@ static void invalidate_complete(struct 
dm_cache_migration *mg, bool success)
        struct cache *cache = mg->cache;
 
        bio_list_init(&bios);
-       if (dm_cell_unlock_v2(cache->prison, mg->cell, &bios))
-               free_prison_cell(cache, mg->cell);
+       if (mg->cell) {
+               if (dm_cell_unlock_v2(cache->prison, mg->cell, &bios))
+                       free_prison_cell(cache, mg->cell);
+       }
 
        if (!success && mg->overwrite_bio)
                bio_io_error(mg->overwrite_bio);
-- 
2.49.0