[libvirt] An implementation bug in qemuMigrationWaitForCompletion that introduces an unnecessary sleep of 50 ms when a migration completes

Michal Privoznik mprivozn at redhat.com
Mon Apr 13 07:56:22 UTC 2015

On 10.04.2015 09:49, Michal Privoznik wrote:
> On 10.04.2015 00:32, Xing Lin wrote:
>> <snip/>
> Yeah, I guess it's very unlikely that migration will complete within
> first 50ms after issuing the command on the monitor (in which case we
> would again wait pointlessly). The other approach would be to leave the
> sleep() where it is, but enclose it in if (jobInfo->type ==
> VIR_DOMAIN_JOB_UNBOUNDED) {}. I have no preference over these two
> versions though. So ACK to the patch of yours. However, I'll postpone
> merging the patch for a while and let others express their feelings.

As promised, pushed now.


More information about the libvir-list mailing list