[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CANn89i+7FWCX6+-puMj7L9f+=Ji6mV-Oca63k-c7OsEikmFPHQ@mail.gmail.com>
Date: Fri, 3 Nov 2023 11:27:46 +0100
From: Eric Dumazet <edumazet@...gle.com>
To: Jiri Slaby <jirislaby@...nel.org>
Cc: "David S . Miller" <davem@...emloft.net>, Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>, netdev@...r.kernel.org,
Soheil Hassas Yeganeh <soheil@...gle.com>, Neal Cardwell <ncardwell@...gle.com>,
Yuchung Cheng <ycheng@...gle.com>, eric.dumazet@...il.com
Subject: Re: [PATCH net-next] tcp: get rid of sysctl_tcp_adv_win_scale
On Fri, Nov 3, 2023 at 11:14 AM Jiri Slaby <jirislaby@...nel.org> wrote:
>
> On 03. 11. 23, 9:17, Eric Dumazet wrote:
> > What happens if you double /proc/sys/net/ipv4/tcp_wmem and/or
> > /proc/sys/net/ipv4/tcp_rmem first value ?
> > (4096 -> 8192)
>
> No change:
> # cat /proc/sys/net/ipv4/tcp_wmem
> 8192 16384 4194304
>
>
> But this:
> # cat /proc/sys/net/ipv4/tcp_rmem
> 8192 16384 4194304
>
> results in test to run 4.7 s, so even 3 more times slower!
You might try setting tcp_shrink_window sysctl to one, with some side
effects for normal TCP flows.
Powered by blists - more mailing lists