On 7/31/23 09:29, Martin Kletzander wrote: > On Fri, Jul 28, 2023 at 04:40:50PM +0200, Michal Privoznik wrote: >> Signed-off-by: Michal Privoznik <mprivozn@xxxxxxxxxx> >> --- >> NEWS.rst | 17 +++++++++++++++++ >> 1 file changed, 17 insertions(+) >> >> diff --git a/NEWS.rst b/NEWS.rst >> index 1de8314a61..0dc909c849 100644 >> --- a/NEWS.rst >> +++ b/NEWS.rst >> @@ -33,8 +33,25 @@ v9.6.0 (unreleased) >> ``/disk/target@bus='scsi'``) supports the ``removable`` attribute at >> ``/disk/target@removable```. >> >> + * qemu: Add NUMA node automatically for memory hotplug >> + >> + When enabling memory hotplug, libvirt required at least one guest >> NUMA to >> + be specified in the domain XML. With this release, libvirt adds it >> + automatically when needed. >> + > > I think it sounds better if we say what libvirt does since this release > rather > than what used to happen before. I would suggest something like: > > Users no longer need to specify guest NUMA node in the domain XML > during > hotplug, libvirt automatically adds one when it is missing. Good point. But we need better phrasing, because "memory hotplug" is basically two things: enabling it in domain XML (e.g. by specifying <maxMemory/> or <maxMemory slots=''/> - as examined by virDomainDefHasMemoryHotplug()), and then there's actual hotplug whilst the domain is running. The NUMA node is added only for the first part. If domain is running we can not add a NUMA node, obviously. Let me fix that in v2, just like the rest of your (valuable) comments. Thanks! Michal
