[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <656e4758675b9_1bd6e2086f@john.notmuch>
Date: Mon, 04 Dec 2023 13:40:40 -0800
From: John Fastabend <john.fastabend@...il.com>
To: Kuniyuki Iwashima <kuniyu@...zon.com>,
john.fastabend@...il.com
Cc: bpf@...r.kernel.org,
edumazet@...gle.com,
jakub@...udflare.com,
martin.lau@...nel.org,
netdev@...r.kernel.org,
kuniyu@...zon.com
Subject: RE: [PATCH bpf v2 1/2] bpf: syzkaller found null ptr deref in
unix_bpf proto add
Kuniyuki Iwashima wrote:
> From: John Fastabend <john.fastabend@...il.com>
> Date: Fri, 1 Dec 2023 10:01:38 -0800
> > I added logic to track the sock pair for stream_unix sockets so that we
> > ensure lifetime of the sock matches the time a sockmap could reference
> > the sock (see fixes tag). I forgot though that we allow af_unix unconnected
> > sockets into a sock{map|hash} map.
> >
> > This is problematic because previous fixed expected sk_pair() to exist
> > and did not NULL check it. Because unconnected sockets have a NULL
> > sk_pair this resulted in the NULL ptr dereference found by syzkaller.
> >
> > BUG: KASAN: null-ptr-deref in unix_stream_bpf_update_proto+0x72/0x430 net/unix/unix_bpf.c:171
> > Write of size 4 at addr 0000000000000080 by task syz-executor360/5073
> > Call Trace:
> > <TASK>
> > ...
> > sock_hold include/net/sock.h:777 [inline]
> > unix_stream_bpf_update_proto+0x72/0x430 net/unix/unix_bpf.c:171
> > sock_map_init_proto net/core/sock_map.c:190 [inline]
> > sock_map_link+0xb87/0x1100 net/core/sock_map.c:294
> > sock_map_update_common+0xf6/0x870 net/core/sock_map.c:483
> > sock_map_update_elem_sys+0x5b6/0x640 net/core/sock_map.c:577
> > bpf_map_update_value+0x3af/0x820 kernel/bpf/syscall.c:167
> >
> > We considered just checking for the null ptr and skipping taking a ref
> > on the NULL peer sock. But, if the socket is then connected() after
> > being added to the sockmap we can cause the original issue again. So
> > instead this patch blocks adding af_unix sockets that are not in the
> > ESTABLISHED state.
>
> I'm not sure if someone has the unconnected stream socket use case
> though, can't we call additional sock_hold() in connect() by checking
> sk_prot under sk_callback_lock ?
Could be done I guess yes. I'm not sure the utility of it though. I
thought above patch was the simplest solution and didn't require touching
main af_unix code. I don't actually use the sockmap with af_unix
sockets anywhere so maybe someone who is using this can comment if
unconnected is needed?
>From rcu and locking side looks like holding sk_callback_lock would
be sufficient. I was thinking it would require a rcu grace period
or something but seems not.
I guess I could improve original patch if folks want.
.John
Powered by blists - more mailing lists