From: Vlad Yasevich <vyasevic@redhat.com>
To: netdev@vger.kernel.org
Cc: davem@davemloft.net, eric.dumazet@gmail.com
Subject: [PATCH V2 00/14] Always build GSO/GRO functionality into the kernel
Date: Thu, 15 Nov 2012 13:49:09 -0500 [thread overview]
Message-ID: <1353005363-6974-1-git-send-email-vyasevic@redhat.com> (raw)
This patch series is a revision suggested by Eric to solve the problem where
a host without IPv6 support drops GSO frames from the guest.
The problem is that GSO/GRO support is per protocol, and when said protocol
is not loaded or is disabled, packets attempting to go through GSO/GRO code paths
are dropped. This causes retransmissions and a two orders of magnitude drop in
performance.
Prior attempt to solve the problem simply enabled enough GSO/GRO functionality
for IPv6 protocol when IPv6 was diabled. This did not solve the problem when
the protocol was not build in or was blacklisted.
To solve the problem, it was suggested that we separate GSO/GRO callback
registration from packet processing registrations. That way
GSO/GRO callbacks can be built into the kernel and always be there.
This patch series attempts to do just that.
* Patches 1 and 2 split the GSO/GRO handlers from packet_type and convert
to the new structure.
* Patches 3, 4 and 5 do the same thing for net_protocol structure.
* The rest of the patches try to incrementally move the IPv6 GSO/GRO
code out of the module and into the static kernel build. Some IPv6
helper functions also had to move as well.
* The last patch just moves the offload callbacks into its own function.
Changes since V1:
- Removed dev pointer from the packet_offload, since it's not used.
- Added a erroneousely removed EXPORT_SYMBOL
- Fix build issues with different states of IPv6 support.
Thanks
-vlad
Vlad Yasevich (14):
net: Add generic packet offload infrastructure.
net: Switch to using the new packet offload infrustructure
net: Add net protocol offload registration infrustructure
ipv6: Add new offload registration infrastructure.
ipv4: Switch to using the new offload infrastructure.
ipv6: Switch to using new offload infrastructure.
ipv6: Separate ipv6 offload support
ipv6: Separate tcp offload functionality
ipv6: Separate out UDP offload functionality
ipv6: Move exthdr offload support into separate file
ipv6: Update ipv6 static library with newly needed functions
ipv4: Pull GSO registration out of inet_init()
ipv6: Pull IPv6 GSO registration out of the module
net: Remove code duplication between offload structures
include/linux/netdevice.h | 20 +++-
include/net/ip6_checksum.h | 35 ++++++
include/net/protocol.h | 31 +++---
net/core/dev.c | 107 +++++++++++++++--
net/ipv4/af_inet.c | 84 +++++++++----
net/ipv4/protocol.c | 21 ++++
net/ipv6/Makefile | 7 +-
net/ipv6/af_inet6.c | 240 -------------------------------------
net/ipv6/exthdrs.c | 52 +--------
net/ipv6/exthdrs_core.c | 44 +++++++
net/ipv6/exthdrs_offload.c | 41 +++++++
net/ipv6/ip6_offload.c | 282 ++++++++++++++++++++++++++++++++++++++++++++
net/ipv6/ip6_offload.h | 18 +++
net/ipv6/ip6_output.c | 65 ----------
net/ipv6/output_core.c | 76 ++++++++++++
net/ipv6/protocol.c | 25 ++++
net/ipv6/tcp_ipv6.c | 107 +----------------
net/ipv6/tcpv6_offload.c | 95 +++++++++++++++
net/ipv6/udp.c | 94 ---------------
net/ipv6/udp_offload.c | 119 +++++++++++++++++++
20 files changed, 949 insertions(+), 614 deletions(-)
create mode 100644 net/ipv6/exthdrs_offload.c
create mode 100644 net/ipv6/ip6_offload.c
create mode 100644 net/ipv6/ip6_offload.h
create mode 100644 net/ipv6/output_core.c
create mode 100644 net/ipv6/tcpv6_offload.c
create mode 100644 net/ipv6/udp_offload.c
--
1.7.7.6
next reply other threads:[~2012-11-15 18:49 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-11-15 18:49 Vlad Yasevich [this message]
2012-11-15 18:49 ` [PATCH V2 01/14] net: Add generic packet offload infrastructure Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 02/14] net: Switch to using the new packet offload infrustructure Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 03/14] net: Add net protocol offload registration infrustructure Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 04/14] ipv6: Add new offload registration infrastructure Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 05/14] ipv4: Switch to using the new offload infrastructure Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 06/14] ipv6: Switch to using " Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 07/14] ipv6: Separate ipv6 offload support Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 08/14] ipv6: Separate tcp offload functionality Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 09/14] ipv6: Separate out UDP " Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 10/14] ipv6: Move exthdr offload support into separate file Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 11/14] ipv6: Update ipv6 static library with newly needed functions Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 12/14] ipv4: Pull GSO registration out of inet_init() Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 13/14] ipv6: Pull IPv6 GSO registration out of the module Vlad Yasevich
2012-11-15 18:49 ` [PATCH V2 14/14] net: Remove code duplication between offload structures Vlad Yasevich
2012-11-15 22:42 ` [PATCH V2 00/14] Always build GSO/GRO functionality into the kernel David Miller
2012-11-15 22:48 ` Vlad Yasevich
2012-11-15 22:53 ` David Miller
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=1353005363-6974-1-git-send-email-vyasevic@redhat.com \
--to=vyasevic@redhat.com \
--cc=davem@davemloft.net \
--cc=eric.dumazet@gmail.com \
--cc=netdev@vger.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).