From: Miroslav Cimerman <mc@raidlab.org>
To: linux-btrfs@vger.kernel.org
Cc: David Sterba <dsterba@suse.com>, Chris Mason <mason@kernel.org>,
Qu Wenruo <wqu@suse.com>,
linux-kernel@vger.kernel.org
Subject: [PATCH v2] btrfs: raid56: fix use-after-free when waking up the rbio I/O waiter
Date: Mon, 16 Nov 2026 00:19:03 +0000 [thread overview]
Message-ID: <20261116001903.63524-1-mc@raidlab.org> (raw)
The end_io functions of the RAID56 read and write bios decrement
rbio->stripes_pending and call wake_up(&rbio->io_wait) when it drops
to zero. The submitter waits with wait_event(rbio->io_wait,
atomic_read(&rbio->stripes_pending) == 0) and then can drop the last
reference to the rbio, which frees it.
If the submitter checks the condition after the last decrement and
before the wake_up(), it does not sleep at all:
CPU0 (endio, irq) CPU1 (submitter, rmw_rbio)
raid_wait_write_end_io() rmw_rbio()
| |-submit_write_bios()
| | \-submit_bio()
| |
| |-wait_event()
|-bio_put() | ^^^ checks the condition without
| | taking io_wait.lock
|-atomic_dec_and_test() |
| ^^^ stripes_pending hits 0 |
| | ^^^ sees 0, doesn't sleep
| |
| \-rbio_orig_end_io()
| \-kfree()
| ^^^ last rbio reference, rbio
| freed
|
\-wake_up()
\-_raw_spin_lock_irqsave()
^^^ io_wait.lock is in freed memory
Too late.
Even when the submitter sleeps, the wake-up takes it off the wait
queue, so finish_wait() returns without taking the queue's lock while
wake_up() still holds it:
CPU0 (endio, irq) CPU1 (submitter, rmw_rbio)
raid_wait_write_end_io() rmw_rbio()
| \-wait_event()
| |-prepare_to_wait_event()
| \-schedule()
|-atomic_dec_and_test()
| ^^^ stripes_pending hits 0
|
\-wake_up()
\-__wake_up_common_lock()
|-_raw_spin_lock_irqsave()
|-__wake_up_common()
| ^^^ autoremove_wake_function()
| takes the waiter off the
| queue and wakes CPU1 finish_wait()
| |-list_empty_careful()
| | ^^^ entry already removed,
| | returns without taking
| | io_wait.lock
| |
| \-rbio_orig_end_io()
| \-kfree()
| ^^^ rbio freed
|
\-_raw_spin_unlock_irqrestore()
^^^ io_wait.lock is in freed memory
Too late.
The same happens on the read side (raid_wait_read_end_io() and
submit_read_wait_bio_list()) and in the parity scrub. On 6.12.117
(bio_endio() and the end_io function are tail calls, so they do not
show in the trace):
BUG: kernel NULL pointer dereference, address: 0000000000000000
RIP: 0010:__wake_up_common+0x25/0xa0
Call Trace:
<IRQ>
__wake_up+0x31/0x60
blk_update_request+0xfd/0x410
blk_mq_end_request+0x1d/0x1a0
Kernel panic - not syncing: Fatal exception in interrupt
With KASAN on v7.4-rc2 and a test-only delay between the decrement
and the wake-up:
BUG: KASAN: slab-use-after-free in _raw_spin_lock_irqsave+0x76/0xe0
__wake_up+0x1c/0x60
raid_wait_read_end_io+0x132/0x10f0
blk_update_request+0x21e/0xea0
Allocated by task 82:
alloc_rbio.isra.0+0xa8/0xbf0
raid56_parity_recover+0x10a/0x9c0
Freed by task 84:
kfree+0x154/0x420
rbio_orig_end_io+0x1c4/0x340
recover_rbio+0x393/0x590
Wait with a completion instead. complete() wakes the waiter under the
completion's lock and does not touch it afterwards, and
wait_for_completion() returns only after taking that lock, so the
rbio may be freed as soon as the wait returns. The parity scrub waits
for its writes only when finish_parity_scrub() submitted them.
Fixes: d817ce35d24a ("btrfs: raid56: switch recovery path to a single function")
Fixes: 5eb30ee26fa4 ("btrfs: raid56: introduce the main entrance for RMW path")
Fixes: 6bfd0133bee2 ("btrfs: raid56: switch scrub path to use a single function")
Cc: stable@vger.kernel.org # 6.2+
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Miroslav Cimerman <mc@raidlab.org>
---
Changes in v2:
- ASCII race charts (Qu)
- drop the FuzzyRAID sentence and the completion.h include (Qu)
Found with FuzzyRAID, a fault-injection tester for RAID that is not
public yet. The analysis, the fix, the tests and the commit message
were prepared with the help of an LLM. I reviewed everything.
The crash: 4 of 10,711 fault-injection scenarios on a 3-device data
RAID5 in QEMU (6.12.117). With KASAN on v7.4-rc2, a test-only delay
(20 ms between the last decrement and the wake-up, 0.5 ms before the
read wait) shows the use-after-free above in the first pass of reads
over the RAID5 with one device dead. With this patch and the same
delay there is no report and every block reads back correctly.
v1: https://lore.kernel.org/linux-btrfs/20261115232358.58163-1-mc@raidlab.org/
---
fs/btrfs/raid56.c | 30 ++++++++++++++++++++++--------
fs/btrfs/raid56.h | 3 ++-
2 files changed, 24 insertions(+), 9 deletions(-)
diff --git a/fs/btrfs/raid56.c b/fs/btrfs/raid56.c
index a8a8004bb14c..60f270c3a339 100644
--- a/fs/btrfs/raid56.c
+++ b/fs/btrfs/raid56.c
@@ -1019,7 +1019,7 @@ static struct btrfs_raid_bio *alloc_rbio(struct btrfs_fs_info *fs_info,
}
bio_list_init(&rbio->bio_list);
- init_waitqueue_head(&rbio->io_wait);
+ init_completion(&rbio->io_done);
INIT_LIST_HEAD(&rbio->plug_list);
spin_lock_init(&rbio->bio_list_lock);
INIT_LIST_HEAD(&rbio->stripe_cache);
@@ -1619,7 +1619,19 @@ static void raid_wait_read_end_io(struct bio *bio)
bio_put(bio);
if (atomic_dec_and_test(&rbio->stripes_pending))
- wake_up(&rbio->io_wait);
+ complete(&rbio->io_done);
+}
+
+/*
+ * Prepare to wait for the nr bios about to be submitted. With none there is
+ * nothing to wait for.
+ */
+static void rbio_prepare_io_wait(struct btrfs_raid_bio *rbio, int nr)
+{
+ reinit_completion(&rbio->io_done);
+ atomic_set(&rbio->stripes_pending, nr);
+ if (nr == 0)
+ complete(&rbio->io_done);
}
static void submit_read_wait_bio_list(struct btrfs_raid_bio *rbio,
@@ -1627,7 +1639,7 @@ static void submit_read_wait_bio_list(struct btrfs_raid_bio *rbio,
{
struct bio *bio;
- atomic_set(&rbio->stripes_pending, bio_list_size(bio_list));
+ rbio_prepare_io_wait(rbio, bio_list_size(bio_list));
while ((bio = bio_list_pop(bio_list))) {
bio->bi_end_io = raid_wait_read_end_io;
@@ -1640,7 +1652,7 @@ static void submit_read_wait_bio_list(struct btrfs_raid_bio *rbio,
submit_bio(bio);
}
- wait_event(rbio->io_wait, atomic_read(&rbio->stripes_pending) == 0);
+ wait_for_completion_io(&rbio->io_done);
}
static int alloc_rbio_data_pages(struct btrfs_raid_bio *rbio)
@@ -2329,7 +2341,7 @@ static void raid_wait_write_end_io(struct bio *bio)
rbio_update_error_bitmap(rbio, bio);
bio_put(bio);
if (atomic_dec_and_test(&rbio->stripes_pending))
- wake_up(&rbio->io_wait);
+ complete(&rbio->io_done);
}
static void submit_write_bios(struct btrfs_raid_bio *rbio,
@@ -2337,7 +2349,7 @@ static void submit_write_bios(struct btrfs_raid_bio *rbio,
{
struct bio *bio;
- atomic_set(&rbio->stripes_pending, bio_list_size(bio_list));
+ rbio_prepare_io_wait(rbio, bio_list_size(bio_list));
while ((bio = bio_list_pop(bio_list))) {
bio->bi_end_io = raid_wait_write_end_io;
@@ -2444,7 +2456,7 @@ static void rmw_rbio(struct btrfs_raid_bio *rbio)
/* We should have at least one bio assembled. */
ASSERT(bio_list_size(&bio_list));
submit_write_bios(rbio, &bio_list);
- wait_event(rbio->io_wait, atomic_read(&rbio->stripes_pending) == 0);
+ wait_for_completion_io(&rbio->io_done);
/* We may have more errors than our tolerance during the read. */
for (sectornr = 0; sectornr < rbio->stripe_nsectors; sectornr++) {
@@ -2915,7 +2927,9 @@ static void scrub_rbio(struct btrfs_raid_bio *rbio)
* and writeback the good content.
*/
ret = finish_parity_scrub(rbio);
- wait_event(rbio->io_wait, atomic_read(&rbio->stripes_pending) == 0);
+ /* It submitted its writes if and only if it returned 0. */
+ if (ret == 0)
+ wait_for_completion_io(&rbio->io_done);
for (sector_nr = 0; sector_nr < rbio->stripe_nsectors; sector_nr++) {
int found_errors;
diff --git a/fs/btrfs/raid56.h b/fs/btrfs/raid56.h
index 8700910b8dc9..c86ffd655360 100644
--- a/fs/btrfs/raid56.h
+++ b/fs/btrfs/raid56.h
@@ -175,7 +175,8 @@ struct btrfs_raid_bio {
atomic_t stripes_pending;
- wait_queue_head_t io_wait;
+ /* Completed by the last end_io when stripes_pending drops to zero. */
+ struct completion io_done;
/* Bitmap to record which horizontal stripe has data */
unsigned long dbitmap;
base-commit: 3aefd67f08f8ed59ed0fa90619b25856df415031
--
2.53.0
reply other threads:[~2026-11-16 0:19 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261116001903.63524-1-mc@raidlab.org \
--to=mc@raidlab.org \
--cc=dsterba@suse.com \
--cc=linux-btrfs@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mason@kernel.org \
--cc=wqu@suse.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®