On 6/2/20 5:53 AM, Igor Mammedov wrote:
On Mon, 1 Jun 2020 17:14:20 -0300
Daniel Henrique Barboza <danielhb413(a)gmail.com> wrote:
[...]
> An ideal situation would be QEMU to never accept incomplete NUMA
topologies
> in the first place.
At least with your series I can safely drop deprecated incomplete NUMA topologies
on QEMU side (which were producing warnings for a while)
Keep in mind that I'm not enabling this auto-fill feature for all guests across
the board. The code is relying on the availability of QEMU_CAPS_NUMA.
Granted, there are not many 6 year old guests running Libvirt 1.2.6/QEMU 2.2.0
around, and I'm not even sure if we care about supporting this scenario, but I
feel obliged to mention this for the record :)
Thanks,
DHB