[PATCH net-next] net: do not write to msg_get_inq in caller

[PATCH net-next] net: do not write to msg_get_inq in caller

[Willem de Bruijn]

From: Willem de Bruijn <willemb@google.com>

msg_get_inq is an input field from caller to callee. Don't set it in
the callee, as the caller may not clear it on struct reuse.

This is a kernel-internal variant of msghdr only, and the only user
does reinitialize the field. So this is not critical.

But it is more robust to avoid the write, and slightly simpler code.

Callers set msg_get_inq to request the input queue length to be
returned in msg_inq. This is equivalent to but independent from the
SO_INQ request to return that same info as a cmsg (tp->recvmsg_inq).
To reduce branching in the hot path the second also sets the msg_inq.
That is WAI.

This is a small follow-on to commit 4d1442979e4a ("af_unix: don't
post cmsg for SO_INQ unless explicitly asked for"), which fixed the
inverse.

Also collapse two branches using a bitwise or.

Link: https://lore.kernel.org/netdev/willemdebruijn.kernel.24d8030f7a3de@gmail.com/
Signed-off-by: Willem de Bruijn <willemb@google.com>
---
 net/ipv4/tcp.c     | 8 +++-----
 net/unix/af_unix.c | 8 +++-----
 2 files changed, 6 insertions(+), 10 deletions(-)

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index f035440c475a..d5319ebe2452 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -2652,10 +2652,8 @@ static int tcp_recvmsg_locked(struct sock *sk, struct msghdr *msg, size_t len,
 	if (sk->sk_state == TCP_LISTEN)
 		goto out;
 
-	if (tp->recvmsg_inq) {
+	if (tp->recvmsg_inq)
 		*cmsg_flags = TCP_CMSG_INQ;
-		msg->msg_get_inq = 1;
-	}
 	timeo = sock_rcvtimeo(sk, flags & MSG_DONTWAIT);
 
 	/* Urgent data needs to be handled specially. */
@@ -2929,10 +2927,10 @@ int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int flags,
 	ret = tcp_recvmsg_locked(sk, msg, len, flags, &tss, &cmsg_flags);
 	release_sock(sk);
 
-	if ((cmsg_flags || msg->msg_get_inq) && ret >= 0) {
+	if ((cmsg_flags | msg->msg_get_inq) && ret >= 0) {
 		if (cmsg_flags & TCP_CMSG_TS)
 			tcp_recv_timestamp(msg, sk, &tss);
-		if (msg->msg_get_inq) {
+		if ((cmsg_flags & TCP_CMSG_INQ) | msg->msg_get_inq) {
 			msg->msg_inq = tcp_inq_hint(sk);
 			if (cmsg_flags & TCP_CMSG_INQ)
 				put_cmsg(msg, SOL_TCP, TCP_CM_INQ,
diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
index a7ca74653d94..d0511225799b 100644
--- a/net/unix/af_unix.c
+++ b/net/unix/af_unix.c
@@ -2904,7 +2904,6 @@ static int unix_stream_read_generic(struct unix_stream_read_state *state,
 	unsigned int last_len;
 	struct unix_sock *u;
 	int copied = 0;
-	bool do_cmsg;
 	int err = 0;
 	long timeo;
 	int target;
@@ -2930,9 +2929,6 @@ static int unix_stream_read_generic(struct unix_stream_read_state *state,
 
 	u = unix_sk(sk);
 
-	do_cmsg = READ_ONCE(u->recvmsg_inq);
-	if (do_cmsg)
-		msg->msg_get_inq = 1;
 redo:
 	/* Lock the socket to prevent queue disordering
 	 * while sleeps in memcpy_tomsg
@@ -3090,9 +3086,11 @@ static int unix_stream_read_generic(struct unix_stream_read_state *state,
 
 	mutex_unlock(&u->iolock);
 	if (msg) {
+		bool do_cmsg = READ_ONCE(u->recvmsg_inq);
+
 		scm_recv_unix(sock, msg, &scm, flags);
 
-		if (msg->msg_get_inq && (copied ?: err) >= 0) {
+		if ((do_cmsg | msg->msg_get_inq) && (copied ?: err) >= 0) {
 			msg->msg_inq = READ_ONCE(u->inq_len);
 			if (do_cmsg)
 				put_cmsg(msg, SOL_SOCKET, SCM_INQ,
-- 
2.52.0.351.gbe84eed79e-goog

Re: [PATCH net-next] net: do not write to msg_get_inq in caller

[Eric Dumazet]

On Mon, Jan 5, 2026 at 5:33 PM Willem de Bruijn
<willemdebruijn.kernel@gmail.com> wrote:
>
> From: Willem de Bruijn <willemb@google.com>
>
> msg_get_inq is an input field from caller to callee. Don't set it in
> the callee, as the caller may not clear it on struct reuse.
>
> This is a kernel-internal variant of msghdr only, and the only user
> does reinitialize the field. So this is not critical.
>
> But it is more robust to avoid the write, and slightly simpler code.
>
> Callers set msg_get_inq to request the input queue length to be
> returned in msg_inq. This is equivalent to but independent from the
> SO_INQ request to return that same info as a cmsg (tp->recvmsg_inq).
> To reduce branching in the hot path the second also sets the msg_inq.
> That is WAI.
>
> This is a small follow-on to commit 4d1442979e4a ("af_unix: don't
> post cmsg for SO_INQ unless explicitly asked for"), which fixed the
> inverse.
>
> Also collapse two branches using a bitwise or.
>
> Link: https://lore.kernel.org/netdev/willemdebruijn.kernel.24d8030f7a3de@gmail.com/
> Signed-off-by: Willem de Bruijn <willemb@google.com>
> ---

Patch looks sane to me, but the title is a bit confusing, I guess you meant

"net: do not write to msg_get_inq in callee" ?

Also, unix_stream_read_generic() is currently potentially adding a NULL deref
if u->recvmsg_inq is non zero, but msg is NULL ?

If this is the case  we need a Fixes: tag.

Re: [PATCH net-next] net: do not write to msg_get_inq in caller

[Jens Axboe]

On 1/5/26 10:21 AM, Eric Dumazet wrote:
> Also, unix_stream_read_generic() is currently potentially adding a
> NULL deref if u->recvmsg_inq is non zero, but msg is NULL ?
> 
> If this is the case  we need a Fixes: tag.

Good catch - might make sense to queue this patch for 6.19, then?

-- 
Jens Axboe

Re: [PATCH net-next] net: do not write to msg_get_inq in caller

[Willem de Bruijn]

Eric Dumazet wrote:
> On Mon, Jan 5, 2026 at 5:33 PM Willem de Bruijn
> <willemdebruijn.kernel@gmail.com> wrote:
> >
> > From: Willem de Bruijn <willemb@google.com>
> >
> > msg_get_inq is an input field from caller to callee. Don't set it in
> > the callee, as the caller may not clear it on struct reuse.
> >
> > This is a kernel-internal variant of msghdr only, and the only user
> > does reinitialize the field. So this is not critical.
> >
> > But it is more robust to avoid the write, and slightly simpler code.
> >
> > Callers set msg_get_inq to request the input queue length to be
> > returned in msg_inq. This is equivalent to but independent from the
> > SO_INQ request to return that same info as a cmsg (tp->recvmsg_inq).
> > To reduce branching in the hot path the second also sets the msg_inq.
> > That is WAI.
> >
> > This is a small follow-on to commit 4d1442979e4a ("af_unix: don't
> > post cmsg for SO_INQ unless explicitly asked for"), which fixed the
> > inverse.
> >
> > Also collapse two branches using a bitwise or.
> >
> > Link: https://lore.kernel.org/netdev/willemdebruijn.kernel.24d8030f7a3de@gmail.com/
> > Signed-off-by: Willem de Bruijn <willemb@google.com>
> > ---
> 
> Patch looks sane to me, but the title is a bit confusing, I guess you meant
> 
> "net: do not write to msg_get_inq in callee" ?

Indeed, thanks. Will fix.

> 
> Also, unix_stream_read_generic() is currently potentially adding a NULL deref
> if u->recvmsg_inq is non zero, but msg is NULL ?
> 
> If this is the case  we need a Fixes: tag.

Oh good point. state->msg can be NULL as of commit 2b514574f7e8 ("net:
af_unix: implement splice for stream af_unix sockets"). That commit
mentions "we mostly have to deal with a non-existing struct msghdr
argument".

Okay. Will resubmit to net with a Fixes tag (after the usual 24hrs).

Thanks!

Re: [PATCH net-next] net: do not write to msg_get_inq in caller

[Jens Axboe]

On 1/5/26 10:42 AM, Willem de Bruijn wrote:
> Eric Dumazet wrote:
>> On Mon, Jan 5, 2026 at 5:33?PM Willem de Bruijn
>> <willemdebruijn.kernel@gmail.com> wrote:
>>>
>>> From: Willem de Bruijn <willemb@google.com>
>>>
>>> msg_get_inq is an input field from caller to callee. Don't set it in
>>> the callee, as the caller may not clear it on struct reuse.
>>>
>>> This is a kernel-internal variant of msghdr only, and the only user
>>> does reinitialize the field. So this is not critical.
>>>
>>> But it is more robust to avoid the write, and slightly simpler code.
>>>
>>> Callers set msg_get_inq to request the input queue length to be
>>> returned in msg_inq. This is equivalent to but independent from the
>>> SO_INQ request to return that same info as a cmsg (tp->recvmsg_inq).
>>> To reduce branching in the hot path the second also sets the msg_inq.
>>> That is WAI.
>>>
>>> This is a small follow-on to commit 4d1442979e4a ("af_unix: don't
>>> post cmsg for SO_INQ unless explicitly asked for"), which fixed the
>>> inverse.
>>>
>>> Also collapse two branches using a bitwise or.
>>>
>>> Link: https://lore.kernel.org/netdev/willemdebruijn.kernel.24d8030f7a3de@gmail.com/
>>> Signed-off-by: Willem de Bruijn <willemb@google.com>
>>> ---
>>
>> Patch looks sane to me, but the title is a bit confusing, I guess you meant
>>
>> "net: do not write to msg_get_inq in callee" ?
> 
> Indeed, thanks. Will fix.
> 
>>
>> Also, unix_stream_read_generic() is currently potentially adding a NULL deref
>> if u->recvmsg_inq is non zero, but msg is NULL ?
>>
>> If this is the case  we need a Fixes: tag.
> 
> Oh good point. state->msg can be NULL as of commit 2b514574f7e8 ("net:
> af_unix: implement splice for stream af_unix sockets"). That commit
> mentions "we mostly have to deal with a non-existing struct msghdr
> argument".

Worth noting that this is currently not possible, as io_uring should
be the only one setting ->recvmsg_inq and it would not do that via
splice. Should still be fixed of course.

-- 
Jens Axboe

Re: [PATCH net-next] net: do not write to msg_get_inq in caller

[Willem de Bruijn]

Jens Axboe wrote:
> On 1/5/26 10:42 AM, Willem de Bruijn wrote:
> > Eric Dumazet wrote:
> >> On Mon, Jan 5, 2026 at 5:33?PM Willem de Bruijn
> >> <willemdebruijn.kernel@gmail.com> wrote:
> >>>
> >>> From: Willem de Bruijn <willemb@google.com>
> >>>
> >>> msg_get_inq is an input field from caller to callee. Don't set it in
> >>> the callee, as the caller may not clear it on struct reuse.
> >>>
> >>> This is a kernel-internal variant of msghdr only, and the only user
> >>> does reinitialize the field. So this is not critical.
> >>>
> >>> But it is more robust to avoid the write, and slightly simpler code.
> >>>
> >>> Callers set msg_get_inq to request the input queue length to be
> >>> returned in msg_inq. This is equivalent to but independent from the
> >>> SO_INQ request to return that same info as a cmsg (tp->recvmsg_inq).
> >>> To reduce branching in the hot path the second also sets the msg_inq.
> >>> That is WAI.
> >>>
> >>> This is a small follow-on to commit 4d1442979e4a ("af_unix: don't
> >>> post cmsg for SO_INQ unless explicitly asked for"), which fixed the
> >>> inverse.
> >>>
> >>> Also collapse two branches using a bitwise or.
> >>>
> >>> Link: https://lore.kernel.org/netdev/willemdebruijn.kernel.24d8030f7a3de@gmail.com/
> >>> Signed-off-by: Willem de Bruijn <willemb@google.com>
> >>> ---
> >>
> >> Patch looks sane to me, but the title is a bit confusing, I guess you meant
> >>
> >> "net: do not write to msg_get_inq in callee" ?
> > 
> > Indeed, thanks. Will fix.
> > 
> >>
> >> Also, unix_stream_read_generic() is currently potentially adding a NULL deref
> >> if u->recvmsg_inq is non zero, but msg is NULL ?
> >>
> >> If this is the case  we need a Fixes: tag.
> > 
> > Oh good point. state->msg can be NULL as of commit 2b514574f7e8 ("net:
> > af_unix: implement splice for stream af_unix sockets"). That commit
> > mentions "we mostly have to deal with a non-existing struct msghdr
> > argument".
> 
> Worth noting that this is currently not possible, as io_uring should
> be the only one setting ->recvmsg_inq and it would not do that via
> splice. Should still be fixed of course.

recvmsg_inq is written from setsockopt SO_INQ. Do you mean
msg_get_inq?

I think this is reachable with a setsockopt + splice:

        do_cmsg = READ_ONCE(u->recvmsg_inq);
        if (do_cmsg)
                msg->msg_get_inq = 1;

Re: [PATCH net-next] net: do not write to msg_get_inq in caller

[Eric Dumazet]

On Mon, Jan 5, 2026 at 6:57 PM Willem de Bruijn
<willemdebruijn.kernel@gmail.com> wrote:
>
> Jens Axboe wrote:
> > On 1/5/26 10:42 AM, Willem de Bruijn wrote:
> > > Eric Dumazet wrote:
> > >> On Mon, Jan 5, 2026 at 5:33?PM Willem de Bruijn
> > >> <willemdebruijn.kernel@gmail.com> wrote:
> > >>>
> > >>> From: Willem de Bruijn <willemb@google.com>
> > >>>
> > >>> msg_get_inq is an input field from caller to callee. Don't set it in
> > >>> the callee, as the caller may not clear it on struct reuse.
> > >>>
> > >>> This is a kernel-internal variant of msghdr only, and the only user
> > >>> does reinitialize the field. So this is not critical.
> > >>>
> > >>> But it is more robust to avoid the write, and slightly simpler code.
> > >>>
> > >>> Callers set msg_get_inq to request the input queue length to be
> > >>> returned in msg_inq. This is equivalent to but independent from the
> > >>> SO_INQ request to return that same info as a cmsg (tp->recvmsg_inq).
> > >>> To reduce branching in the hot path the second also sets the msg_inq.
> > >>> That is WAI.
> > >>>
> > >>> This is a small follow-on to commit 4d1442979e4a ("af_unix: don't
> > >>> post cmsg for SO_INQ unless explicitly asked for"), which fixed the
> > >>> inverse.
> > >>>
> > >>> Also collapse two branches using a bitwise or.
> > >>>
> > >>> Link: https://lore.kernel.org/netdev/willemdebruijn.kernel.24d8030f7a3de@gmail.com/
> > >>> Signed-off-by: Willem de Bruijn <willemb@google.com>
> > >>> ---
> > >>
> > >> Patch looks sane to me, but the title is a bit confusing, I guess you meant
> > >>
> > >> "net: do not write to msg_get_inq in callee" ?
> > >
> > > Indeed, thanks. Will fix.
> > >
> > >>
> > >> Also, unix_stream_read_generic() is currently potentially adding a NULL deref
> > >> if u->recvmsg_inq is non zero, but msg is NULL ?
> > >>
> > >> If this is the case  we need a Fixes: tag.
> > >
> > > Oh good point. state->msg can be NULL as of commit 2b514574f7e8 ("net:
> > > af_unix: implement splice for stream af_unix sockets"). That commit
> > > mentions "we mostly have to deal with a non-existing struct msghdr
> > > argument".
> >
> > Worth noting that this is currently not possible, as io_uring should
> > be the only one setting ->recvmsg_inq and it would not do that via
> > splice. Should still be fixed of course.
>
> recvmsg_inq is written from setsockopt SO_INQ. Do you mean
> msg_get_inq?
>
> I think this is reachable with a setsockopt + splice:
>
>         do_cmsg = READ_ONCE(u->recvmsg_inq);
>         if (do_cmsg)
>                 msg->msg_get_inq = 1;

It is a bit strange that syzbot did not find it yet.

Re: [PATCH net-next] net: do not write to msg_get_inq in caller

[Jens Axboe]

On 1/5/26 10:57 AM, Willem de Bruijn wrote:
> Jens Axboe wrote:
>> On 1/5/26 10:42 AM, Willem de Bruijn wrote:
>>> Eric Dumazet wrote:
>>>> On Mon, Jan 5, 2026 at 5:33?PM Willem de Bruijn
>>>> <willemdebruijn.kernel@gmail.com> wrote:
>>>>>
>>>>> From: Willem de Bruijn <willemb@google.com>
>>>>>
>>>>> msg_get_inq is an input field from caller to callee. Don't set it in
>>>>> the callee, as the caller may not clear it on struct reuse.
>>>>>
>>>>> This is a kernel-internal variant of msghdr only, and the only user
>>>>> does reinitialize the field. So this is not critical.
>>>>>
>>>>> But it is more robust to avoid the write, and slightly simpler code.
>>>>>
>>>>> Callers set msg_get_inq to request the input queue length to be
>>>>> returned in msg_inq. This is equivalent to but independent from the
>>>>> SO_INQ request to return that same info as a cmsg (tp->recvmsg_inq).
>>>>> To reduce branching in the hot path the second also sets the msg_inq.
>>>>> That is WAI.
>>>>>
>>>>> This is a small follow-on to commit 4d1442979e4a ("af_unix: don't
>>>>> post cmsg for SO_INQ unless explicitly asked for"), which fixed the
>>>>> inverse.
>>>>>
>>>>> Also collapse two branches using a bitwise or.
>>>>>
>>>>> Link: https://lore.kernel.org/netdev/willemdebruijn.kernel.24d8030f7a3de@gmail.com/
>>>>> Signed-off-by: Willem de Bruijn <willemb@google.com>
>>>>> ---
>>>>
>>>> Patch looks sane to me, but the title is a bit confusing, I guess you meant
>>>>
>>>> "net: do not write to msg_get_inq in callee" ?
>>>
>>> Indeed, thanks. Will fix.
>>>
>>>>
>>>> Also, unix_stream_read_generic() is currently potentially adding a NULL deref
>>>> if u->recvmsg_inq is non zero, but msg is NULL ?
>>>>
>>>> If this is the case  we need a Fixes: tag.
>>>
>>> Oh good point. state->msg can be NULL as of commit 2b514574f7e8 ("net:
>>> af_unix: implement splice for stream af_unix sockets"). That commit
>>> mentions "we mostly have to deal with a non-existing struct msghdr
>>> argument".
>>
>> Worth noting that this is currently not possible, as io_uring should
>> be the only one setting ->recvmsg_inq and it would not do that via
>> splice. Should still be fixed of course.
> 
> recvmsg_inq is written from setsockopt SO_INQ. Do you mean
> msg_get_inq?
> 
> I think this is reachable with a setsockopt + splice:
> 
>         do_cmsg = READ_ONCE(u->recvmsg_inq);
>         if (do_cmsg)
>                 msg->msg_get_inq = 1;

Indeed you are right, I mixed up the two...

-- 
Jens Axboe

Re: [PATCH net-next] net: do not write to msg_get_inq in caller

[Jens Axboe]

On 1/5/26 9:32 AM, Willem de Bruijn wrote:
> From: Willem de Bruijn <willemb@google.com>
> 
> msg_get_inq is an input field from caller to callee. Don't set it in
> the callee, as the caller may not clear it on struct reuse.
> 
> This is a kernel-internal variant of msghdr only, and the only user
> does reinitialize the field. So this is not critical.
> 
> But it is more robust to avoid the write, and slightly simpler code.
> 
> Callers set msg_get_inq to request the input queue length to be
> returned in msg_inq. This is equivalent to but independent from the
> SO_INQ request to return that same info as a cmsg (tp->recvmsg_inq).
> To reduce branching in the hot path the second also sets the msg_inq.
> That is WAI.
> 
> This is a small follow-on to commit 4d1442979e4a ("af_unix: don't
> post cmsg for SO_INQ unless explicitly asked for"), which fixed the
> inverse.
> 
> Also collapse two branches using a bitwise or.

Reviewed-by: Jens Axboe <axboe@kernel.dk>

I also ran my usual testing, and as expected, looks fine as well.

-- 
Jens Axboe