From: Pavel Begunkov <asml.silence@gmail.com>
To: Eric Dumazet <edumazet@google.com>
Cc: netdev <netdev@vger.kernel.org>,
"David S . Miller" <davem@davemloft.net>,
Jakub Kicinski <kuba@kernel.org>, Wei Liu <wei.liu@kernel.org>,
Paul Durrant <paul@xen.org>
Subject: Re: [RFC net-next 00/27] net and/or udp optimisations
Date: Mon, 11 Apr 2022 13:04:10 +0100 [thread overview]
Message-ID: <f0a8e790-aff5-db72-d42e-1cbfe711a092@gmail.com> (raw)
In-Reply-To: <CANn89iL0anmfcX1v1NqOQ6xi2VfY7CmiUwav0jNbz6GuSjUspQ@mail.gmail.com>
On 4/6/22 10:44, Eric Dumazet wrote:
> On Sun, Apr 3, 2022 at 6:08 AM Pavel Begunkov <asml.silence@gmail.com> wrote:
>>
>> A mix of various net optimisations, which were mostly discovered during UDP
>> testing. Benchmarked with an io_uring test using 16B UDP/IPv6 over dummy netdev:
>> 2090K vs 2229K tx/s, +6.6%, or in a 4-8% range if not averaging across reboots.
>>
>> 1-3 removes extra atomics and barriers from sock_wfree() mainly benefitting UDP.
>> 4-7 cleans up some zerocopy helpers
>> 8-16 do inlining of ipv6 and generic net pathes
>> 17 is a small nice performance improvement for TCP zerocopy
>> 18-27 refactors UDP to shed some more overhead
>>
> Please send a smaller series first.
Apologies for delays. Ok, I'll split it.
> About inlining everything around, make sure to include performance
> numbers only for the inline parts.
> We can inline everything and make the kernel 4 x time bigger.
> Synthetic benchmarks will still show improvements but in overall, we
> add icache cost that is going to hurt latencies.
I do care about kernel bloating, but I think we can agree that for most
patches inlining is safe. There are 6 such patches (9-12,15,16). Three
of them (9,11,15) only do simple redirecting to another function
skb_csum_hwoffload_help() in 10 has only two callers. I think we can
agree that they're safe to inline.
That leaves ip6_local_out() with ~8 callers and used quite heavily. And
fl6_update_dst() with ~12 users, I don't have exact data but it appears
not everybody uses ip6 options and so the function does nothing. At
least that's true for UDP cases I care about. I think it's justified
to be inlined. Would you prefer these two to be removed?
> I vote that you focus on the other parts first.
>
> Thank you.
>
>> Pavel Begunkov (27):
>> sock: deduplicate ->sk_wmem_alloc check
>> sock: optimise sock_def_write_space send refcounting
>> sock: optimise sock_def_write_space barriers
>> skbuff: drop zero check from skb_zcopy_set
>> skbuff: drop null check from skb_zcopy
>> net: xen: set zc flags only when there is ubuf
>> skbuff: introduce skb_is_zcopy()
>> skbuff: optimise alloc_skb_with_frags()
>> net: inline sock_alloc_send_skb
>> net: inline part of skb_csum_hwoffload_help
>> net: inline skb_zerocopy_iter_dgram
>> ipv6: inline ip6_local_out()
>> ipv6: help __ip6_finish_output() inlining
>> ipv6: refactor ip6_finish_output2()
>> net: inline dev_queue_xmit()
>> ipv6: partially inline fl6_update_dst()
>> tcp: optimise skb_zerocopy_iter_stream()
>> net: optimise ipcm6 cookie init
>> udp/ipv6: refactor udpv6_sendmsg udplite checks
>> udp/ipv6: move pending section of udpv6_sendmsg
>> udp/ipv6: prioritise the ip6 path over ip4 checks
>> udp/ipv6: optimise udpv6_sendmsg() daddr checks
>> udp/ipv6: optimise out daddr reassignment
>> udp/ipv6: clean up udpv6_sendmsg's saddr init
>> ipv6: refactor opts push in __ip6_make_skb()
>> ipv6: improve opt-less __ip6_make_skb()
>> ipv6: clean up ip6_setup_cork
>>
>> drivers/net/xen-netback/interface.c | 3 +-
>> include/linux/netdevice.h | 27 ++++-
>> include/linux/skbuff.h | 102 +++++++++++++-----
>> include/net/ipv6.h | 37 ++++---
>> include/net/sock.h | 10 +-
>> net/core/datagram.c | 2 -
>> net/core/datagram.h | 15 ---
>> net/core/dev.c | 28 ++---
>> net/core/skbuff.c | 59 ++++-------
>> net/core/sock.c | 50 +++++++--
>> net/ipv4/ip_output.c | 10 +-
>> net/ipv4/tcp.c | 5 +-
>> net/ipv6/datagram.c | 4 +-
>> net/ipv6/exthdrs.c | 15 ++-
>> net/ipv6/ip6_output.c | 88 ++++++++--------
>> net/ipv6/output_core.c | 12 ---
>> net/ipv6/raw.c | 8 +-
>> net/ipv6/udp.c | 158 +++++++++++++---------------
>> net/l2tp/l2tp_ip6.c | 8 +-
>> 19 files changed, 339 insertions(+), 302 deletions(-)
>> delete mode 100644 net/core/datagram.h
>>
>> --
>> 2.35.1
>>
--
Pavel Begunkov
prev parent reply other threads:[~2022-04-11 12:04 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-04-03 13:06 [RFC net-next 00/27] net and/or udp optimisations Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 01/27] sock: deduplicate ->sk_wmem_alloc check Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 02/27] sock: optimise sock_def_write_space send refcounting Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 03/27] sock: optimise sock_def_write_space barriers Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 04/27] skbuff: drop zero check from skb_zcopy_set Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 05/27] skbuff: drop null check from skb_zcopy Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 06/27] net: xen: set zc flags only when there is ubuf Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 07/27] skbuff: introduce skb_is_zcopy() Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 08/27] skbuff: optimise alloc_skb_with_frags() Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 09/27] net: inline sock_alloc_send_skb Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 10/27] net: inline part of skb_csum_hwoffload_help Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 11/27] net: inline skb_zerocopy_iter_dgram Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 12/27] ipv6: inline ip6_local_out() Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 13/27] ipv6: help __ip6_finish_output() inlining Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 14/27] ipv6: refactor ip6_finish_output2() Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 15/27] net: inline dev_queue_xmit() Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 16/27] ipv6: partially inline fl6_update_dst() Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 17/27] tcp: optimise skb_zerocopy_iter_stream() Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 18/27] net: optimise ipcm6 cookie init Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 19/27] udp/ipv6: refactor udpv6_sendmsg udplite checks Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 20/27] udp/ipv6: move pending section of udpv6_sendmsg Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 21/27] udp/ipv6: prioritise the ip6 path over ip4 checks Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 22/27] udp/ipv6: optimise udpv6_sendmsg() daddr checks Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 23/27] udp/ipv6: optimise out daddr reassignment Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 24/27] udp/ipv6: clean up udpv6_sendmsg's saddr init Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 25/27] ipv6: refactor opts push in __ip6_make_skb() Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 26/27] ipv6: improve opt-less __ip6_make_skb() Pavel Begunkov
2022-04-03 13:06 ` [PATCH net-next 27/27] ipv6: clean up ip6_setup_cork Pavel Begunkov
2022-04-06 9:44 ` [RFC net-next 00/27] net and/or udp optimisations Eric Dumazet
2022-04-11 12:04 ` Pavel Begunkov [this message]
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=f0a8e790-aff5-db72-d42e-1cbfe711a092@gmail.com \
--to=asml.silence@gmail.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=kuba@kernel.org \
--cc=netdev@vger.kernel.org \
--cc=paul@xen.org \
--cc=wei.liu@kernel.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: 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;
as well as URLs for NNTP newsgroup(s).