public inbox for netdev@vger.kernel.org
 help / color / mirror / Atom feed
From: Jakub Kicinski <kuba@kernel.org>
To: edumazet@google.com
Cc: kuniyu@google.com, ncardwell@google.com, netdev@vger.kernel.org,
	davem@davemloft.net, pabeni@redhat.com, andrew+netdev@lunn.ch,
	horms@kernel.org
Subject: Re: [PATCH net-next] tcp: try to defer / return acked skbs to originating CPU
Date: Sat, 17 Jan 2026 09:10:10 -0800	[thread overview]
Message-ID: <20260117091010.3133de3b@kernel.org> (raw)
In-Reply-To: <20260117164255.785751-1-kuba@kernel.org>

On Sat, 17 Jan 2026 08:42:55 -0800 Jakub Kicinski wrote:
> Running a memcache-like workload under production(ish) load
> on a 300 thread AMD machine we see ~3% of CPU time spent
> in kmem_cache_free() via tcp_ack(), freeing skbs from rtx queue.
> This workloads pins workers away from softirq CPU so
> the Tx skbs are pretty much always allocated on a different
> CPU than where the ACKs arrive. Try to use the defer skb free
> queue to return the skbs back to where they came from.
> This results in a ~4% performance improvement for the workload.

In the interest of full transparency the performance testing was
done on a 6.13-ish kernel. But I don't see anything that'd make
the situation better upstream..

  reply	other threads:[~2026-01-17 17:10 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-01-17 16:42 [PATCH net-next] tcp: try to defer / return acked skbs to originating CPU Jakub Kicinski
2026-01-17 17:10 ` Jakub Kicinski [this message]
2026-01-17 18:16 ` Eric Dumazet
2026-01-17 23:03   ` Jakub Kicinski
2026-01-18 12:15     ` Eric Dumazet
2026-01-19 17:04       ` Jakub Kicinski
2026-01-29 23:04         ` Jakub Kicinski
2026-01-29 23:10           ` Jakub Kicinski
2026-02-16 16:06           ` Eric Dumazet
2026-02-16 17:49             ` Jakub Kicinski
2026-02-16 17:58               ` Eric Dumazet
2026-02-16 18:11                 ` Jakub Kicinski
2026-02-16 18:16                   ` Eric Dumazet
2026-02-17 21:50                     ` Jakub Kicinski
2026-02-17 21:56                       ` Eric Dumazet

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260117091010.3133de3b@kernel.org \
    --to=kuba@kernel.org \
    --cc=andrew+netdev@lunn.ch \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=horms@kernel.org \
    --cc=kuniyu@google.com \
    --cc=ncardwell@google.com \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox