Netdev Archive on lore.kernel.org help / color / mirror / Atom feed
From: Pavel Begunkov <asml.silence@gmail.com> To: netdev@vger.kernel.org, "David S . Miller" <davem@davemloft.net>, Jakub Kicinski <kuba@kernel.org> Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>, David Ahern <dsahern@kernel.org>, Eric Dumazet <edumazet@google.com>, Willem de Bruijn <willemdebruijn.kernel@gmail.com>, linux-kernel@vger.kernel.org, Pavel Begunkov <asml.silence@gmail.com> Subject: [PATCH 00/14] udp optimisation Date: Tue, 11 Jan 2022 01:21:32 +0000 [thread overview] Message-ID: <cover.1641863490.git.asml.silence@gmail.com> (raw) A mainly UDP/IPv6 optimisation patch set. Zerocopy io_uring benchmark over dummy netdev (CPU bound) gives 2068992 -> 2166481 tx/s, which is ~4.7% or over 5% of net layer overhead. Should give similar results for small packet non-zerocopy. - 1/14 and 9/14 remove a get/put dst pair each, so saving 4 atomics per corkless UDP send. - Patches 3-8 optimise iflow handling, in particular removes one 88B memset and one 88B copy. - 10-14 are random improvements, which are not UDP-specific but also beneficial to TCP and others. Pavel Begunkov (14): ipv6: optimise dst referencing ipv6: shuffle up->pending AF_INET bits ipv6: remove daddr temp buffer in __ip6_make_skb ipv6: clean up cork setup/release ipv6: don't zero cork's flowi after use ipv6: pass full cork into __ip6_append_data() ipv6: pass flow in ip6_make_skb together with cork ipv6/udp: don't make extra copies of iflow ipv6: hand dst refs to cork setup skbuff: drop zero check from skb_zcopy_set skbuff: drop null check from skb_zcopy skbuff: optimise alloc_skb_with_frags() net: inline part of skb_csum_hwoffload_help net: inline sock_alloc_send_skb include/linux/netdevice.h | 16 +++++- include/linux/skbuff.h | 45 +++++++++++++--- include/net/ipv6.h | 2 +- include/net/sock.h | 10 +++- net/core/dev.c | 15 ++---- net/core/skbuff.c | 34 +++++------- net/core/sock.c | 7 --- net/ipv4/ip_output.c | 10 ++-- net/ipv4/tcp.c | 5 +- net/ipv6/ip6_output.c | 105 +++++++++++++++++++++----------------- net/ipv6/udp.c | 103 ++++++++++++++++++------------------- 11 files changed, 197 insertions(+), 155 deletions(-) -- 2.34.1
next reply other threads:[~2022-01-11 1:24 UTC|newest] Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top 2022-01-11 1:21 Pavel Begunkov [this message] 2022-01-11 1:21 ` [PATCH 01/14] ipv6: optimise dst referencing Pavel Begunkov 2022-01-11 1:21 ` [PATCH 02/14] ipv6: shuffle up->pending AF_INET bits Pavel Begunkov 2022-01-11 1:21 ` [PATCH 03/14] ipv6: remove daddr temp buffer in __ip6_make_skb Pavel Begunkov 2022-01-11 1:21 ` [PATCH 04/14] ipv6: clean up cork setup/release Pavel Begunkov 2022-01-11 1:21 ` [PATCH 05/14] ipv6: don't zero cork's flowi after use Pavel Begunkov 2022-01-11 1:21 ` [PATCH 06/14] ipv6: pass full cork into __ip6_append_data() Pavel Begunkov 2022-01-11 1:21 ` [PATCH 07/14] ipv6: pass flow in ip6_make_skb together with cork Pavel Begunkov 2022-01-11 1:21 ` [PATCH 08/14] ipv6/udp: don't make extra copies of iflow Pavel Begunkov 2022-01-11 1:21 ` [PATCH 09/14] ipv6: hand dst refs to cork setup Pavel Begunkov 2022-01-11 15:39 ` Willem de Bruijn 2022-01-11 15:57 ` Pavel Begunkov 2022-01-11 17:11 ` Paolo Abeni 2022-01-11 20:39 ` Pavel Begunkov 2022-01-12 11:15 ` Paolo Abeni 2022-01-12 16:49 ` Pavel Begunkov 2022-01-11 1:21 ` [PATCH 10/14] skbuff: drop zero check from skb_zcopy_set Pavel Begunkov 2022-01-11 1:21 ` [PATCH 11/14] skbuff: drop null check from skb_zcopy Pavel Begunkov 2022-01-11 1:21 ` [PATCH 12/14] skbuff: optimise alloc_skb_with_frags() Pavel Begunkov 2022-01-11 1:21 ` [PATCH 13/14] net: inline part of skb_csum_hwoffload_help Pavel Begunkov 2022-01-11 9:24 ` David Laight 2022-01-11 16:59 ` Pavel Begunkov 2022-01-11 17:25 ` David Laight 2022-01-11 20:48 ` Pavel Begunkov 2022-01-12 2:41 ` David Laight 2022-01-12 16:43 ` Pavel Begunkov 2022-01-11 1:21 ` [PATCH 14/14] net: inline sock_alloc_send_skb Pavel Begunkov
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=cover.1641863490.git.asml.silence@gmail.com \ --to=asml.silence@gmail.com \ --cc=davem@davemloft.net \ --cc=dsahern@kernel.org \ --cc=edumazet@google.com \ --cc=kuba@kernel.org \ --cc=linux-kernel@vger.kernel.org \ --cc=netdev@vger.kernel.org \ --cc=willemdebruijn.kernel@gmail.com \ --cc=yoshfuji@linux-ipv6.org \ /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: linkBe 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; as well as URLs for NNTP newsgroup(s).