Re: [PATCH v2] bcache: fix cached_dev.sb_bio use-after-free and crash

On Mon, Mar 23, 2026 at 09:01:19PM +0800, [email protected] wrote:
> From: Mingzhe Zou <[email protected]>
> 
> In our production environment, we have received multiple crash reports
> regarding libceph, which have caught our attention:
> 
> ```
> [6888366.280350] Call Trace:
> [6888366.280452]  blk_update_request+0x14e/0x370
> [6888366.280561]  blk_mq_end_request+0x1a/0x130
> [6888366.280671]  rbd_img_handle_request+0x1a0/0x1b0 [rbd]
> [6888366.280792]  rbd_obj_handle_request+0x32/0x40 [rbd]
> [6888366.280903]  __complete_request+0x22/0x70 [libceph]
> [6888366.281032]  osd_dispatch+0x15e/0xb40 [libceph]
> [6888366.281164]  ? inet_recvmsg+0x5b/0xd0
> [6888366.281272]  ? ceph_tcp_recvmsg+0x6f/0xa0 [libceph]
> [6888366.281405]  ceph_con_process_message+0x79/0x140 [libceph]
> [6888366.281534]  ceph_con_v1_try_read+0x5d7/0xf30 [libceph]
> [6888366.281661]  ceph_con_workfn+0x329/0x680 [libceph]
> ```
> 
> After analyzing the coredump file, we found that the address of dc->sb_bio
> has been freed. We know that cached_dev is only freed when it is stopped.
> 
> Since sb_bio is a part of struct cached_dev, rather than an alloc every time.
> If the device is stopped while writing to the superblock, the released address
> will be accessed at endio.
> 
> This patch hopes to wait for sb_write to complete in cached_dev_free.
> 
> It should be noted that we analyzed the cause of the problem, then tell
> all details to the QWEN and adopted the modifications it made.
> 
> Signed-off-by: Mingzhe Zou <[email protected]>
> 
> ---
> v2: fix the crash caused by not calling closure_init in v1
> ---
>  drivers/md/bcache/super.c | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
> index 64bb38c95895..b76edbaaf4f3 100644
> --- a/drivers/md/bcache/super.c
> +++ b/drivers/md/bcache/super.c
> @@ -1373,6 +1373,13 @@ static CLOSURE_CALLBACK(cached_dev_free)
>  
>  	mutex_unlock(&bch_register_lock);
>  
> +	/*
> +	 * Wait for any pending sb_write to complete before free.
> +	 * The sb_bio is embedded in struct cached_dev, so we must
> +	 * ensure no I/O is in progress.
> +	 */
> +	down(&dc->sb_write_mutex);
> +

I know what you mean. dc->sb_write cannot be access out of bch_write_bdev_super().
But the above down() method is not comfortable IMHO.

Fortunately when cached_dev_free() is called from cached_dev_flush(), kobjs of
bcache device is delted by kobject_del(&d->kobj), there is no chance to call
bch_write_bdev_super() via sysfs interface. And when cached_dev_free() is called,
other code path calling bch_write_bdev_super() won't happen neither.

So a pair of
	down(&dc->sb_write_mutex);
	up(&dc->sb_write_mutex);
might be enough to make sure the last on-flight bch_write_bdev_super() will
complete?



>  	if (dc->sb_disk)
>  		folio_put(virt_to_folio(dc->sb_disk));
>  
> -- 
> 2.34.1