Red Hat Bugzilla – Bug 665360
vhost-net/kvm lacks fixes/optimizations in net-next as of Dec 23
Last modified: 2013-01-09 18:26:46 EST
Description of problem: vhost-net and kvm modules in upstream have several optimizations that give about 15-20% gain on small packet bandwidth, and fix minor bugs. The list of patches as of today: kvm: fast-path msi injection with irqfd vhost: get/put_user -> __get/__put_user vhost: copy_to_user -> __copy_to_user vhost: fix log ctx signalling vhost: fix return code for log_access_ok() vhost-net: batch use/unuse mm drivers/vhost/vhost.c: delete double assignment vhost: put mm after thread stop vhost: stop worker only if created vhost: replace vhost_workqueue with per-vhost kthread Version-Release number of selected component (if applicable): kernel-2.6.32-92.el6
This request was evaluated by Red Hat Product Management for inclusion in a Red Hat Enterprise Linux maintenance release. Product Management has requested further review of this request by Red Hat Engineering, for potential inclusion in a Red Hat Enterprise Linux Update release for currently deployed products. This request is not yet committed for inclusion in an Update release.
Patch(es) available on kernel-2.6.32-112.el6
Created attachment 490700 [details] vhost-net rhel6.0 vs rhel6.1 Attach the test result of vhost-net rhel6.1/rhel6.0. See high improvment with small packet under guest <->exhost ,and guest <-> exguest scenarios compared with rhel6.1 vhost-net ,but see obvious performance degradation with UDP (R) and TCP (S) under guest <-> host scenario compared with rhel6.1 vhost-net.
Created attachment 493117 [details] vhost-net rhel6.0 vs rhel6.1 megabyte/cpu attach megabyte/cpu sheet base on comment #7's sheet. only focus on samll packets (< MTU) ,rhel6.1 vhost get high improvment in most of the scenarios/protocols expect for TCP (S) from host to guest compared with rhel6.0 vhost.
(In reply to comment #9) > Created attachment 493117 [details] > vhost-net rhel6.0 vs rhel6.1 megabyte/cpu > > attach megabyte/cpu sheet base on comment #7's sheet. only focus on samll > packets (< MTU) ,rhel6.1 vhost get high improvment in most of the > scenarios/protocols expect for TCP (S) from host to guest compared with rhel6.0 > vhost. hi mst what's your comments? as comment #9 ,could i verify it as pass or something else?
to comment 10: yes, I think it's ok.
base on comment #9 #11, change the bug status to verified.
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2011-0542.html