lkml.org 
[lkml]   [2021]   [Jul]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH v4] tcp: fix tcp_init_transfer() to not reset icsk_ca_initialized
From
Date
On 7/4/2021 9:52 PM, Neal Cardwell wrote:
>
>
> On Sat, Jul 3, 2021 at 5:49 PM David Miller <davem@davemloft.net
> <mailto:davem@davemloft.net>> wrote:
> >
> > From: Nguyen Dinh Phi <phind.uet@gmail.com <mailto:phind.uet@gmail.com>>
> > Date: Sat,  3 Jul 2021 17:34:17 +0800
> >
> > > diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
> > > index 7d5e59f688de..855ada2be25e 100644
> > > --- a/net/ipv4/tcp_input.c
> > > +++ b/net/ipv4/tcp_input.c
> > > @@ -5922,7 +5922,6 @@ void tcp_init_transfer(struct sock *sk, int
> bpf_op, struct sk_buff *skb)
> > >               tp->snd_cwnd = tcp_init_cwnd(tp, __sk_dst_get(sk));
> > >       tp->snd_cwnd_stamp = tcp_jiffies32;
> > >
> > > -     icsk->icsk_ca_initialized = 0;
> > >       bpf_skops_established(sk, bpf_op, skb);
> > >       if (!icsk->icsk_ca_initialized)
> > >               tcp_init_congestion_control(sk);
> >
> > Don't you have to make the tcp_init_congestion_control() call
> unconditional now?
>
> I think we want to keep it conditional, to avoid double-initialization
> if the BPF code sets the congestion control algorithm and initializes
> it. But that's relatively new and subtle, so it might be nice for this
> patch to add a comment about that, since it's touching this part of the
> code anyway:
>
> -       icsk->icsk_ca_initialized = 0;
>         bpf_skops_established(sk, bpf_op, skb);
> +       /* Initialize congestion control unless a BPF op initialized it
> already: */
>         if (!icsk->icsk_ca_initialized)
>                 tcp_init_congestion_control(sk);
>
> neal
>
I will make a new version with your comment.
Thank you very much for helping.

\
 
 \ /
  Last update: 2021-07-05 18:36    [W:0.053 / U:0.232 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site