[ovs-dev] [PATCH v4] netdev-dpdk: Reset queue number for vhost devices on vm shutdown.

Stokes, Ian ian.stokes at intel.com
Thu Jun 27 14:37:23 UTC 2019


> On 27/06/2019 15:17, Ian Stokes wrote:
> > On 6/27/2019 10:43 AM, David Marchand wrote:
> >> Rather than poll all disabled queues and waste some memory for vms that
> >> have been shutdown, we can reconfigure when receiving a destroy
> >> connection notification from the vhost library.
> >>
> >> $ while true; do
> >>    ovs-appctl dpif-netdev/pmd-rxq-show |awk '
> >>    /port: / {
> >>      tot++;
> >>      if ($5 == "(enabled)") {
> >>        en++;
> >>      }
> >>    }
> >>    END {
> >>      print "total: " tot ", enabled: " en
> >>    }'
> >>    sleep 1
> >> done
> >>
> >> total: 66, enabled: 66
> >> total: 6, enabled: 2
> >>
> >> This change requires a fix in the DPDK vhost library, so bump the
> minimal
> >> required version to 18.11.2.
> >
> > So in testing this I was seeing an error with the following
> >
> > 2019-06-27T10:00:08Z|00057|dpdk|INFO|VHOST_CONFIG: vhost peer closed
> > 2019-06-27T10:00:08Z|00058|dpdk|ERR|VHOST_CONFIG: (0) device not found.
> > 2019-06-27T10:00:08Z|00059|netdev_dpdk|INFO|vHost Device '' not found
> >
> > It wasn't until I realized I was testing with DPDK 18.11.2.rc1 that it
> > dawned on me the required fix wasn't there...doh!
> >
> 
> Is that version a typo? There is only a couple of unrelated patches
> between between 18.11.2-rc1 and 18.11.2 version

You may be right, just looking at the package I had on the board was dpdk-stable-18.11.2-rc1, but it could have been the case I was pointing to 18.11.1 if the patch was part of 18.11.2.rc1. I've been chopping and changing so much this past week it was more likely a mistake on my part.

Ian



More information about the dev mailing list