mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Michal Luczaj <mhal@rbox.co>
To: Kuniyuki Iwashima <kuniyu@google.com>,
	Martin KaFai Lau <martin.lau@linux.dev>
Cc: John Fastabend <john.fastabend@gmail.com>,
	Jakub Sitnicki <jakub@cloudflare.com>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	Simon Horman <horms@kernel.org>,
	Daniel Borkmann <daniel@iogearbox.net>,
	netdev@vger.kernel.org, bpf@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH bpf] bpf, sockmap: Fix af_unix null-ptr-deref in proto update
Date: Mon, 2 Feb 2026 16:10:56 +0100	[thread overview]
Message-ID: <b98a40e8-1be2-47f0-8238-92dee8f5a204@rbox.co> (raw)
In-Reply-To: <CAAVpQUAc3S+Ebx2qp-Pbg9qH-Zc=yHdzVJbchCu++V_4XF3DLg@mail.gmail.com>

On 1/31/26 11:06, Kuniyuki Iwashima wrote:
> On Fri, Jan 30, 2026 at 1:30 PM Martin KaFai Lau <martin.lau@linux.dev> wrote:
>>
>> On 1/30/26 3:00 AM, Michal Luczaj wrote:
>>>>> Follow-up to discussion at
>>>>> https://lore.kernel.org/netdev/20240610174906.32921-1-kuniyu@amazon.com/.
>>>>
>>>> It is a long thread to dig. Please summarize the discussion in the
>>>> commit message.
>>>
>>> OK, there we go:
>>>
>>> The root cause of the null-ptr-deref is that unix_stream_connect() sets
>>> sk_state (`WRITE_ONCE(sk->sk_state, TCP_ESTABLISHED)`) _before_ it assigns
>>> a peer (`unix_peer(sk) = newsk`). sk_state == TCP_ESTABLISHED makes
>>> sock_map_sk_state_allowed() believe that socket is properly set up, which
>>> would include having a defined peer.
>>>
>>> In other words, there's a window when you can call
>>> unix_stream_bpf_update_proto() on socket which still has unix_peer(sk) == NULL.
>>>
>>> My initial idea was to simply move peer assignment _before_ the sk_state
>>> update, but the maintainer wasn't interested in changing the
>>> unix_stream_connect() hot path. He suggested taking care of it in the
>>> sockmap code.
> 
> Yes, we already have a memory barrier for unix_peer(sk) there
> (to save sock_hold()/sock_put() in sendmsg(), see 830a1e5c212fb)
> and another one just for sk->sk_state is not worth the unlikely
> case in sockmap by a buggy user.
> 
> 
>>>
>>> My understanding is that users are not supposed to put sockets in a sockmap
>>> when said socket is only half-way through connect() call. Hence `return
>>> -EINVAL` on a missing peer. Now, if users should be allowed to legally race
>>> connect() vs. sockmap update, then I guess we can wait for connect() to
>>> "finalize" e.g. by taking the unix_state_lock(), as discussed below.
> 
> If a user hit the issue, the user must have updated sockmap while the
> user knows connect() had not returned.  Such a user must prepare
> for failures since it could occur before sock_map_sk_state_allowed() too.
> 
> This is a subtle timing issue and I don't think the kernel should be
> friendly to such buggy users by waiting for connect() etc.
> 
> 
>>>
>>>>   From looking at this commit message, if the existing lock_sock held by
>>>> update_elem is not useful for af_unix,
>>>
>>> Right, the existing lock_sock is not useful. update's lock_sock holds
>>> sock::sk_lock, while unix_state_lock() holds unix_sock::lock.
>>
>> It sounds like lock_sock is the incorrect lock to hold for af_unix. Is
>> taking lock_sock in sock_map doing anything useful for af_unix? Should
>> sock_map hold the unix_state_lock instead of lock_sock?
> 
> If sockmap code does not sleep, unix_state_lock can be used there.
> 
> 
>>
>> Other than update_elem, do other lock_sock() usages in sock_map have a
>> similar issue for af_unix?

As for the sockmap, I think that would be it.

In related news, looks like bpf_iter_unix_seq_show() is missing
unix_state_lock(): lock_sock_fast() won't stop unix_release_sock(). E.g.
bpf iterator can grab unix_sock::peer as it is being released.

>>>> it is not clear why a new test
>>>> "!sk_pair" on top of the existing WRITE_ONCE(sk->sk_state...) is a fix.
>>>
>>> "On top"? Just to make sure we're looking at the same thing: above I was
>>> trying to show two parallel flows with unix_peer() fetch in thread-0 and
>>> WRITE_ONCE(sk->sk_state...) and `unix_peer(sk) = newsk` in thread-1.
>>>
>>> It fixes the problem because now update_proto won't call sock_hold(NULL).
>>>
>>>> A minor thing is sock_map_sk_state_allowed doesn't have
>>>> READ_ONCE(sk->sk_state) for sk_is_stream_unix also.
>>>
>>> Ok, I'll add this as a separate patch in v2. Along with the !tcp case of
>>> sock_map_redirect_allowed()?
>>
>> sgtm. thanks.
>>
>>>
>>>> If unix_stream_connect does not hold lock_sock, can unix_state_lock be
>>>> used here? lock_sock has already been taken, update_elem should not be
>>>> the hot path.
>>>
>>> Yes, it can be used, it was proposed in the old thread. In fact, critical
>>> section can be empty; only used to wait for unix_stream_connect() to
>>> release the lock, which would guarantee unix_peer(sk) != NULL by then.
>>>
>>>          if (!psock->sk_pair) {
>>> +               unix_state_lock(sk);
>>> +               unix_state_unlock(sk);
> 
> I don't like this... we had a similar one in recvmsg(MSG_PEEK) path
> for GC with a biiiiiig comment, which I removed in 118f457da9ed .
> 
> 
>>>                  sk_pair = unix_peer(sk);
>>>                  sock_hold(sk_pair);
>>
>> I don't have a strong opinion on waiting or checking NULL. imo, both are
>> not easy to understand. One is sk_state had already been checked earlier
>> under a lock_sock but still needs to check NULL on unix_peer(). Another
>> one is an empty unix_state_[un]lock(). If taking unix_state_lock, may as
>> well just use the existing unix_peer_get(sk).
> 
> Yes, unix_peer_get() can be safely used there (with extra sock_put()).
> 
> 
>> If its return value cannot
>> (?) be NULL, WARN_ON_ONCE() instead of having a special empty
> 
> I suggested WARN_ON_ONCE() because Michal reproduced it with
> mdelay() and I did not think it could occur in real life, but considering
> PREEMPT_RT, it could be real.  So, the current form in this patch looks
> good to me.

FWIW, it reproduces on CONFIG_PREEMPT_RT=n without mdelay().
Please see
https://lore.kernel.org/bpf/20260123-selftest-signal-on-connect-v1-0-b0256e7025b6@rbox.co/

>> lock/unlock pattern here. If the correct lock (unix_state_lock) was held
>> earlier in update_elem, all these would go away.
>>
>> Also, it is not immediately clear why a non-NULL unix_peer(sk) is safe
>> here. From looking around af_unix.c, is it because the sk refcnt is held
>> earlier in update_elem? For unix_stream, unix_peer(sk) will stay valid
>> until unix_release_sock(sk). Am I reading it correctly?
> 
> unix_stream_connect() holds the peer's refcnt, and once unix_peer(sk)
> is set, it and refcnt are not cleared until close()d.  So unix_peer_get() is
> a bit redundant for sane users.


  reply	other threads:[~2026-02-02 15:11 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-01-29 16:47 Michal Luczaj
2026-01-29 19:41 ` Martin KaFai Lau
2026-01-30 11:00   ` Michal Luczaj
2026-01-30 21:29     ` Martin KaFai Lau
2026-01-31 10:06       ` Kuniyuki Iwashima
2026-02-02 15:10         ` Michal Luczaj [this message]
2026-02-03  3:53           ` Martin KaFai Lau
2026-02-03  9:57             ` Michal Luczaj
2026-02-03 19:47               ` Kuniyuki Iwashima
2026-02-04  7:15                 ` Martin KaFai Lau
2026-02-04  7:58                   ` Kuniyuki Iwashima
2026-02-04 15:41                     ` Michal Luczaj
2026-02-04 19:16                       ` Kuniyuki Iwashima
2026-02-04 20:18                         ` Martin KaFai Lau
2026-02-04 19:34                       ` Martin KaFai Lau
2026-02-04 21:09                         ` Kuniyuki Iwashima
2026-02-05  0:55                           ` Martin KaFai Lau
2026-02-05  2:00                             ` Martin KaFai Lau
2026-02-05  7:39                               ` Kuniyuki Iwashima
2026-02-04 23:25                         ` Michal Luczaj
2026-02-05  0:27                           ` Kuniyuki Iwashima
2026-02-05  0:31                           ` Martin KaFai Lau
2026-02-02 19:15         ` Martin KaFai Lau
2026-02-07 14:37           ` Michal Luczaj

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=b98a40e8-1be2-47f0-8238-92dee8f5a204@rbox.co \
    --to=mhal@rbox.co \
    --cc=bpf@vger.kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=horms@kernel.org \
    --cc=jakub@cloudflare.com \
    --cc=john.fastabend@gmail.com \
    --cc=kuba@kernel.org \
    --cc=kuniyu@google.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=martin.lau@linux.dev \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.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®