git://git.onelab.eu / sliver-openvswitch.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 4d2a0f3) | patch
ofproto-dpif: Rate limit calls to facet_learn().
authorEthan Jackson <ethan@nicira.com>
Fri, 22 Mar 2013 02:40:49 +0000 (19:40 -0700)
committerEthan Jackson <ethan@nicira.com>
Wed, 27 Mar 2013 17:39:45 +0000 (10:39 -0700)
commit6cf474d792d3197bdf92e49d4215c8bc942d6e15
treec2e831c03e5fbe729f4114a87e168800e1f1b96atree | snapshot
parent4d2a0f3984ec9ab76467bb2ee06476513a4d8083commit | diff
ofproto-dpif: Rate limit calls to facet_learn().

In the TCP_CRR benchmark, ovs-vswitchd spends so much time in
update_stats() that it has a significant impact on flow setup
performance.  Further work is needed in this area, but for now,
simply rate limiting facet_learn() has a roughly 10% improvement
with complex flow tables.

Signed-off-by: Ethan Jackson <ethan@nicira.com>
ofproto/ofproto-dpif.c diff | blob | history
sliver-openvswitch