[PATCH net v1] tcp: Return user_mss for TCP_MAXSEG in CLOSE/LISTEN state if user_mss set

[PATCH net v1] tcp: Return user_mss for TCP_MAXSEG in CLOSE/LISTEN state if user_mss set

[Cambda Zhu]

This patch replaces the tp->mss_cache check in getting TCP_MAXSEG
with tp->rx_opt.user_mss check for CLOSE/LISTEN sock. Since
tp->mss_cache is initialized with TCP_MSS_DEFAULT, checking if
it's zero is probably a bug.

With this change, getting TCP_MAXSEG before connecting will return
default MSS normally, and return user_mss if user_mss is set.

Fixes: 0c409e85f0ac ("Import 2.3.41pre2")
Reported-by: Jack Yang <mingliang@linux.alibaba.com>
Suggested-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/netdev/CANn89i+3kL9pYtkxkwxwNMzvC_w3LNUum_2=3u+UyLBmGmifHA@mail.gmail.com/#t
Signed-off-by: Cambda Zhu <cambda@linux.alibaba.com>
Link: https://lore.kernel.org/netdev/14D45862-36EA-4076-974C-EA67513C92F6@linux.alibaba.com/
---
v1:
- Return default MSS if user_mss not set for backwards compatibility.
- Send patch to net instead of net-next, with Fixes tag.
- Add Eric's tags.
---
 net/ipv4/tcp.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 4d6392c16b7a..3e01a58724b8 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -4081,7 +4081,8 @@ int do_tcp_getsockopt(struct sock *sk, int level,
 	switch (optname) {
 	case TCP_MAXSEG:
 		val = tp->mss_cache;
-		if (!val && ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
+		if (tp->rx_opt.user_mss &&
+		    ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
 			val = tp->rx_opt.user_mss;
 		if (tp->repair)
 			val = tp->rx_opt.mss_clamp;
-- 
2.16.6

Re: [PATCH net v1] tcp: Return user_mss for TCP_MAXSEG in CLOSE/LISTEN state if user_mss set

[Cambda Zhu]

> On May 24, 2023, at 16:33, Cambda Zhu <cambda@linux.alibaba.com> wrote:
> 
> This patch replaces the tp->mss_cache check in getting TCP_MAXSEG
> with tp->rx_opt.user_mss check for CLOSE/LISTEN sock. Since
> tp->mss_cache is initialized with TCP_MSS_DEFAULT, checking if
> it's zero is probably a bug.
> 
> With this change, getting TCP_MAXSEG before connecting will return
> default MSS normally, and return user_mss if user_mss is set.
> 
> Fixes: 0c409e85f0ac ("Import 2.3.41pre2")
> Reported-by: Jack Yang <mingliang@linux.alibaba.com>
> Suggested-by: Eric Dumazet <edumazet@google.com>
> Link: https://lore.kernel.org/netdev/CANn89i+3kL9pYtkxkwxwNMzvC_w3LNUum_2=3u+UyLBmGmifHA@mail.gmail.com/#t
> Signed-off-by: Cambda Zhu <cambda@linux.alibaba.com>
> Link: https://lore.kernel.org/netdev/14D45862-36EA-4076-974C-EA67513C92F6@linux.alibaba.com/
> ---
> v1:
> - Return default MSS if user_mss not set for backwards compatibility.
> - Send patch to net instead of net-next, with Fixes tag.
> - Add Eric's tags.
> ---
> net/ipv4/tcp.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
> index 4d6392c16b7a..3e01a58724b8 100644
> --- a/net/ipv4/tcp.c
> +++ b/net/ipv4/tcp.c
> @@ -4081,7 +4081,8 @@ int do_tcp_getsockopt(struct sock *sk, int level,
> switch (optname) {
> case TCP_MAXSEG:
> val = tp->mss_cache;
> - if (!val && ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
> + if (tp->rx_opt.user_mss &&
> +    ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
> val = tp->rx_opt.user_mss;
> if (tp->repair)
> val = tp->rx_opt.mss_clamp;
> -- 
> 2.16.6

I see netdev/verify_fixes check failed for the commit 0c409e85f0ac ("Import 2.3.41pre2").
The commit is from:
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git

Should I remove the Fixes tag?

Thanks!

Cambda

Re: [PATCH net v1] tcp: Return user_mss for TCP_MAXSEG in CLOSE/LISTEN state if user_mss set

[Jason Xing]

On Wed, May 24, 2023 at 5:11 PM Cambda Zhu <cambda@linux.alibaba.com> wrote:
>
>
> > On May 24, 2023, at 16:33, Cambda Zhu <cambda@linux.alibaba.com> wrote:
> >
> > This patch replaces the tp->mss_cache check in getting TCP_MAXSEG
> > with tp->rx_opt.user_mss check for CLOSE/LISTEN sock. Since
> > tp->mss_cache is initialized with TCP_MSS_DEFAULT, checking if
> > it's zero is probably a bug.
> >
> > With this change, getting TCP_MAXSEG before connecting will return
> > default MSS normally, and return user_mss if user_mss is set.
> >
> > Fixes: 0c409e85f0ac ("Import 2.3.41pre2")
> > Reported-by: Jack Yang <mingliang@linux.alibaba.com>
> > Suggested-by: Eric Dumazet <edumazet@google.com>
> > Link: https://lore.kernel.org/netdev/CANn89i+3kL9pYtkxkwxwNMzvC_w3LNUum_2=3u+UyLBmGmifHA@mail.gmail.com/#t
> > Signed-off-by: Cambda Zhu <cambda@linux.alibaba.com>
> > Link: https://lore.kernel.org/netdev/14D45862-36EA-4076-974C-EA67513C92F6@linux.alibaba.com/
> > ---
> > v1:
> > - Return default MSS if user_mss not set for backwards compatibility.
> > - Send patch to net instead of net-next, with Fixes tag.
> > - Add Eric's tags.
> > ---
> > net/ipv4/tcp.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
> > index 4d6392c16b7a..3e01a58724b8 100644
> > --- a/net/ipv4/tcp.c
> > +++ b/net/ipv4/tcp.c
> > @@ -4081,7 +4081,8 @@ int do_tcp_getsockopt(struct sock *sk, int level,
> > switch (optname) {
> > case TCP_MAXSEG:
> > val = tp->mss_cache;
> > - if (!val && ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
> > + if (tp->rx_opt.user_mss &&
> > +    ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
> > val = tp->rx_opt.user_mss;
> > if (tp->repair)
> > val = tp->rx_opt.mss_clamp;
> > --
> > 2.16.6
>
> I see netdev/verify_fixes check failed for the commit 0c409e85f0ac ("Import 2.3.41pre2").
> The commit is from:
> git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
>
> Should I remove the Fixes tag?

Please using the correct Fixes tag as below because we cannot trace it
(0c409e85f0ac) in the current tree:

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")

Probably you have to send the v3 version. Let's wait to hear more
opinions about this patch.

Otherwise it looks good to me. Please feel free to add:
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>

Thanks,
Jason

>
> Thanks!
>
> Cambda

Re: [PATCH net v1] tcp: Return user_mss for TCP_MAXSEG in CLOSE/LISTEN state if user_mss set

[Simon Horman]

On Wed, May 24, 2023 at 05:10:54PM +0800, Cambda Zhu wrote:
> 
> > On May 24, 2023, at 16:33, Cambda Zhu <cambda@linux.alibaba.com> wrote:
> > 
> > This patch replaces the tp->mss_cache check in getting TCP_MAXSEG
> > with tp->rx_opt.user_mss check for CLOSE/LISTEN sock. Since
> > tp->mss_cache is initialized with TCP_MSS_DEFAULT, checking if
> > it's zero is probably a bug.
> > 
> > With this change, getting TCP_MAXSEG before connecting will return
> > default MSS normally, and return user_mss if user_mss is set.
> > 
> > Fixes: 0c409e85f0ac ("Import 2.3.41pre2")
> > Reported-by: Jack Yang <mingliang@linux.alibaba.com>
> > Suggested-by: Eric Dumazet <edumazet@google.com>
> > Link: https://lore.kernel.org/netdev/CANn89i+3kL9pYtkxkwxwNMzvC_w3LNUum_2=3u+UyLBmGmifHA@mail.gmail.com/#t
> > Signed-off-by: Cambda Zhu <cambda@linux.alibaba.com>
> > Link: https://lore.kernel.org/netdev/14D45862-36EA-4076-974C-EA67513C92F6@linux.alibaba.com/
> > ---
> > v1:
> > - Return default MSS if user_mss not set for backwards compatibility.
> > - Send patch to net instead of net-next, with Fixes tag.
> > - Add Eric's tags.
> > ---
> > net/ipv4/tcp.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
> > index 4d6392c16b7a..3e01a58724b8 100644
> > --- a/net/ipv4/tcp.c
> > +++ b/net/ipv4/tcp.c
> > @@ -4081,7 +4081,8 @@ int do_tcp_getsockopt(struct sock *sk, int level,
> > switch (optname) {
> > case TCP_MAXSEG:
> > val = tp->mss_cache;
> > - if (!val && ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
> > + if (tp->rx_opt.user_mss &&
> > +    ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
> > val = tp->rx_opt.user_mss;
> > if (tp->repair)
> > val = tp->rx_opt.mss_clamp;
> > -- 
> > 2.16.6
> 
> I see netdev/verify_fixes check failed for the commit 0c409e85f0ac ("Import 2.3.41pre2").
> The commit is from:
> git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
> 
> Should I remove the Fixes tag?

Hi Cambda Zhu,

AFAIK, that is a tree of historical versions that predate git history,
which began with version v2.6.12-rc2.

If this is a fix for a but ghat has been present since before v2.6.12-rc2
then I think standard practice is to use the initial commit of
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git

That is:

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")

N.B.: I did not check to see if the bug you are fixing has been present
since then. But in any case, the fixes tag in your patch is not correct.

-- 
pw-bot: cr