Re: [PATCH] fanotify: fix use-after-free of file range info

Jan Kara <[email protected]>
Newsgroups org.kernel.vger.linux-fsdevel,org.kernel.vger.linux-kernel,org.kernel.vger.stable
Message-ID <lkaqsnjtsfqbnlm456mxpq4k3hchwpmd4rnwydj2xektrnem7f@gwz5xjiduxfg>
On Thu 30-07-26 16:58:01, Chengfeng Ye wrote:
> fsnotify_pre_content() builds its file_range on the triggering task's
> stack. fanotify_alloc_perm_event() saves a pointer to range.pos in the
> heap-allocated permission event so copy_range_info_to_user() can report
> the offset later.
> 
> The event reader can set the event state to FAN_EVENT_REPORTED and then
> sleep while preparing the file descriptor. If a signal interrupts the
> triggering task at that point, fanotify_get_response() changes the state
> to FAN_EVENT_CANCELED and returns. This unwinds the file_range stack
> frame while the reader still owns the event. The reader then dereferences
> pevent->ppos and copies the stale stack value to userspace.
> 
> KASAN reported:
> 
>   BUG: KASAN: use-after-free in fanotify_read+0x293e/0x2970
>   Read of size 8 at addr ffff88811434fc50 by task fanotify_inotif/95
>   Call Trace:
>    fanotify_read+0x293e/0x2970
>    vfs_read+0x177/0xa20
>    ksys_read+0xf7/0x1c0
>    do_syscall_64+0xf9/0x540
>    entry_SYSCALL_64_after_hwframe+0x77/0x7f
> 
> Copy the range position into the permission event and make ppos refer to
> that event-owned value. The event remains alive until the reader finishes,
> so the reported offset no longer depends on the triggering task's stack.
> 
> Fixes: 870499bc1d4d ("fanotify: report file range info with pre-content events")
> Cc: [email protected]
> Signed-off-by: Chengfeng Ye <[email protected]>

Thanks for report! That's indeed a nasty bug. FWIW I always disliked how we
passed pointer to pos around instead of pos itself only to be able to
indicate it is not defined. Since loff_t is actually signed, Amir, do you
remember why we didn't just define (loff_t)-1 as "range not provided"
message instead of passing around ppos?

I think getting rid of passing ppos around and storing it to the permission
event would be preferrable over the hybrid approach in this patch.

								Honza

> ---
>  fs/notify/fanotify/fanotify.c | 3 ++-
>  fs/notify/fanotify/fanotify.h | 3 ++-
>  2 files changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/notify/fanotify/fanotify.c b/fs/notify/fanotify/fanotify.c
> index a3555bebad63..c97d1be70310 100644
> --- a/fs/notify/fanotify/fanotify.c
> +++ b/fs/notify/fanotify/fanotify.c
> @@ -601,7 +601,8 @@ static struct fanotify_event *fanotify_alloc_perm_event(const void *data,
>  	pevent->state = FAN_EVENT_INIT;
>  	pevent->path = *path;
>  	/* NULL ppos means no range info */
> -	pevent->ppos = range ? &range->pos : NULL;
> +	pevent->pos = range ? range->pos : 0;
> +	pevent->ppos = range ? &pevent->pos : NULL;
>  	pevent->count = range ? range->count : 0;
>  	path_get(path);
>  
> diff --git a/fs/notify/fanotify/fanotify.h b/fs/notify/fanotify/fanotify.h
> index a0619e7694d5..c964df6c0514 100644
> --- a/fs/notify/fanotify/fanotify.h
> +++ b/fs/notify/fanotify/fanotify.h
> @@ -438,7 +438,8 @@ FANOTIFY_ME(struct fanotify_event *event)
>  struct fanotify_perm_event {
>  	struct fanotify_event fae;
>  	struct path path;
> -	const loff_t *ppos;		/* optional file range info */
> +	loff_t pos;
> +	const loff_t *ppos;		/* &pos if range info is available */
>  	size_t count;
>  	u32 response;			/* userspace answer to the event */
>  	unsigned short state;		/* state of the event */
> -- 
> 2.43.0
> 
-- 
Jan Kara <[email protected]>
SUSE Labs, CR
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.