diff options
| author |   Thadeu Lima de Souza Cascardo <cascardo@redhat.com> | 2016-09-15 19:11:53 -0300 |
|---|---|---|
| committer |   David S. Miller <davem@davemloft.net> | 2016-09-18 22:14:01 -0400 |
| commit | db74a3335e0f645e3139c80bcfc90feb01d8e304 (patch) | |
| tree | 5adec60da4faa80b188cc0b45639224c0a197db4 /net/sched/sch_fq.c | |
| parent | openvswitch: fix flow stats accounting when node 0 is not possible (diff) | |
| download | linux-dev-db74a3335e0f645e3139c80bcfc90feb01d8e304.tar.xz linux-dev-db74a3335e0f645e3139c80bcfc90feb01d8e304.zip | |
openvswitch: use percpu flow stats
Instead of using flow stats per NUMA node, use it per CPU. When using
megaflows, the stats lock can be a bottleneck in scalability.
On a E5-2690 12-core system, usual throughput went from ~4Mpps to
~15Mpps when forwarding between two 40GbE ports with a single flow
configured on the datapath.
This has been tested on a system with possible CPUs 0-7,16-23. After
module removal, there were no corruption on the slab cache.
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@redhat.com>
Cc: pravin shelar <pshelar@ovn.org>
Acked-by: Pravin B Shelar <pshelar@ovn.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/sched/sch_fq.c')
0 files changed, 0 insertions, 0 deletions