linux-kernel - Re: [PATCH net-next 3/3] net: tcp: check timeout by icsk->icsk_timeout in tcp_retransmit

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CADxym3Y7kF9rT+PL6SE_tZFi2Q1CXqhz9e66F1kckcYtMmEkuQ@mail.gmail.com>
Date:   Mon, 31 Jul 2023 17:49:10 +0800
From:   Menglong Dong <menglong8.dong@...il.com>
To:     Neal Cardwell <ncardwell@...gle.com>,
        Eric Dumazet <edumazet@...gle.com>
Cc:     davem@...emloft.net, kuba@...nel.org, pabeni@...hat.com,
        dsahern@...nel.org, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Menglong Dong <imagedong@...cent.com>,
        Yuchung Cheng <ycheng@...gle.com>,
        Soheil Hassas Yeganeh <soheil@...gle.com>
Subject: Re: [PATCH net-next 3/3] net: tcp: check timeout by
 icsk->icsk_timeout in tcp_retransmit_timer()

On Sat, Jul 29, 2023 at 1:15 AM Neal Cardwell <ncardwell@...gle.com> wrote:
>
> On Fri, Jul 28, 2023 at 7:25 AM Neal Cardwell <ncardwell@...gle.com> wrote:
> >
> > On Fri, Jul 28, 2023 at 1:50 AM Eric Dumazet <edumazet@...gle.com> wrote:
> > >
> > > On Fri, Jul 28, 2023 at 8:25 AM Menglong Dong <menglong8.dong@...il.com> wrote:
> > > >
> > > > On Fri, Jul 28, 2023 at 12:44 PM Neal Cardwell <ncardwell@...gle.com> wrote:
> > > > >
> > > > > On Thu, Jul 27, 2023 at 7:57 PM Menglong Dong <menglong8.dong@...il.com> wrote:
> > > > > >
> > > > > > On Fri, Jul 28, 2023 at 3:31 AM Eric Dumazet <edumazet@...gle.com> wrote:
> > > > > > >
> > > > > > > On Thu, Jul 27, 2023 at 2:52 PM <menglong8.dong@...il.com> wrote:
> > > > > > > >
> > > > > > > > From: Menglong Dong <imagedong@...cent.com>
> > > > > > > >
> > > > > > > > In tcp_retransmit_timer(), a window shrunk connection will be regarded
> > > > > > > > as timeout if 'tcp_jiffies32 - tp->rcv_tstamp > TCP_RTO_MAX'. This is not
> > > > > > > > right all the time.
> > > > > > > >
> > > > > > > > The retransmits will become zero-window probes in tcp_retransmit_timer()
> > > > > > > > if the 'snd_wnd==0'. Therefore, the icsk->icsk_rto will come up to
> > > > > > > > TCP_RTO_MAX sooner or later.
> > > > > > > >
> > > > > > > > However, the timer is not precise enough, as it base on timer wheel.
> > > > > > > > Sorry that I am not good at timer, but I know the concept of time-wheel.
> > > > > > > > The longer of the timer, the rougher it will be. So the timeout is not
> > > > > > > > triggered after TCP_RTO_MAX, but 122877ms as I tested.
> > > > > > > >
> > > > > > > > Therefore, 'tcp_jiffies32 - tp->rcv_tstamp > TCP_RTO_MAX' is always true
> > > > > > > > once the RTO come up to TCP_RTO_MAX.
> > > > > > > >
> > > > > > > > Fix this by replacing the 'tcp_jiffies32' with '(u32)icsk->icsk_timeout',
> > > > > > > > which is exact the timestamp of the timeout.
> > > > > > > >
> > > > > > > > Signed-off-by: Menglong Dong <imagedong@...cent.com>
> > > > > > > > ---
> > > > > > > >  net/ipv4/tcp_timer.c | 6 +++++-
> > > > > > > >  1 file changed, 5 insertions(+), 1 deletion(-)
> > > > > > > >
> > > > > > > > diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
> > > > > > > > index 470f581eedd4..3a20db15a186 100644
> > > > > > > > --- a/net/ipv4/tcp_timer.c
> > > > > > > > +++ b/net/ipv4/tcp_timer.c
> > > > > > > > @@ -511,7 +511,11 @@ void tcp_retransmit_timer(struct sock *sk)
> > > > > > > >                                             tp->snd_una, tp->snd_nxt);
> > > > > > > >                 }
> > > > > > > >  #endif
> > > > > > > > -               if (tcp_jiffies32 - tp->rcv_tstamp > TCP_RTO_MAX) {
> > > > > > > > +               /* It's a little rough here, we regard any valid packet that
> > > > > > > > +                * update tp->rcv_tstamp as the reply of the retransmitted
> > > > > > > > +                * packet.
> > > > > > > > +                */
> > > > > > > > +               if ((u32)icsk->icsk_timeout - tp->rcv_tstamp > TCP_RTO_MAX) {
> > > > > > > >                         tcp_write_err(sk);
> > > > > > > >                         goto out;
> > > > > > > >                 }
>
> One potential pre-existing issue with this logic: if the connection is
> restarting from idle, then tp->rcv_tstamp could already be a long time
> (minutes or hours) in the past even on the first RTO, in which case
> the very first RTO that found a zero tp->snd_wnd  would find this
> check returns true, and would destroy the connection immediately. This
> seems extremely brittle.
>
> AFAICT it would be safer to replace this logic with a call to the
> standard tcp_write_timeout() logic that has a more robust check to see
> if the connection should be destroyed.

How about we check it with:

icsk->icsk_timeout - max(tp->retrans_stamp, tp->rcv_tstamp) > TCP_RTO_MAX

?

Therefore, we don't need to do a lot of modification. I have to
say that the way we check here is rough, and a simple loss
of the retransmitted packet or the ACK can cause the die
of the socket.

Maybe we need a more perfect way here? which may introduce
a certain amount of modification.

BTW, should I split out this patch from this series?

>
> neal