11 Dec
2015
11 Dec
'15
4:23 p.m.
On Thu, 2015-12-10 at 12:44 -0500, John Ferlan wrote:
Calling qemuDomainAdjustMaxMemLock() more than necessary shouldn't really cause any harm, but adding a check for
hostdev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI
around the call is even better.
Would that address your concerns?
You mean of course before the pesky "virDomainHostdevDefFree(hostdev)"?
You could perhaps do it after the qemuDomainRemovePCIHostDevice in the switch or within that qemuDomainRemovePCIHostDevice code.
I have now posted v2: https://www.redhat.com/archives/libvir-list/2015-December/msg00454.html Cheers. -- Andrea Bolognani Software Engineer - Virtualization Team