On 15.02.2018 15:53, Eric Blake wrote: > On 02/15/2018 08:40 AM, Viktor Mihajlovski wrote: >> On 15.02.2018 15:19, Eric Blake wrote: >>> On 02/15/2018 04:20 AM, Viktor Mihajlovski wrote: >>>> From: Luiz Capitulino <lcapitul...@redhat.com> >>>> >>>> The query-cpus command has an extremely serious side effect: >>>> it always interrupts all running vCPUs so that they can run >>>> ioctl calls. This can cause a huge performance degradation for >>>> some workloads. And most of the information retrieved by the >>>> ioctl calls are not even used by query-cpus. >>>> >>>> This commit introduces a replacement for query-cpus called > >>> You know, we have no back-compat guarantees on HMP. We could make 'info >>> cpu' just ALWAYS call query-cpus-fast, with no HMP counterpart for the >>> slower query-cpus, and without needing a deprecation period. But I'll >>> leave that up to David if that makes more sense. >> Ditching info cpus_fast would make me happy as well, because it would >> cause less headache on the libvirt side of things. > > Why is libvirt using HMP in the first place? Libvirt should always be > using the QMP command, when one exists. > Which it does, but there's still a lot of fallback code including HMP "info cpus". In real life this should of course never be used, because the QMP is always there.
-- Regards, Viktor Mihajlovski