kernel-packages team mailing list archive

[Chris J Arges <1201869@xxxxxxxxxxxxxxxxxx>]

A bisection reveals that the following patch solves the issue:
8093315a91340bca52549044975d8c7f673b28a1 veth: extend device features

However it relies on:
2681128f0ced8aa4e66f221197e183cc16d244fe veth: reduce stat overhead
For a clean cherry-pick.

And after building and testing the following bug needed to be fixed with:
d0e2c55e7c940a3ee91e9e23a2683b593690f1e9 veth: avoid a NULL deref in veth_stats_one

A test build is available here that solves the issue backported to 3.5/3.8:
http://people.canonical.com/~arges/lp1201869/

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1201869

Title:
  poor networking throughput across an OpenStack Neutron router on
  3.5/3.8 kernels

Status in “linux” package in Ubuntu:
  Fix Released
Status in “linux” source package in Quantal:
  In Progress
Status in “linux” source package in Raring:
  In Progress

Bug description:
  OpenStack Neutron does IP forwarding through a network namespace. A
  veth pair is used to connect into the namespace. The veth pair appears
  to be the bottleneck, independent of network namespace. In newer
  versions of Linux (3.9) throughput is much higher.

  This has been confirmed on kernels from 3.5.x-3.8.x. (Quantal and
  Raring lts backports)

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1201869/+subscriptions