2016-04-22 13:19+0200, Igor Mammedov: > On Fri, 22 Apr 2016 11:25:38 +0200 > Greg Kurz <gkurz@xxxxxxxxxxxxxxxxxx> wrote: >> On Thu, 21 Apr 2016 19:36:11 +0200 >> Radim Krčmář <rkrcmar@xxxxxxxxxx> wrote: >> > 2016-04-21 18:45+0200, Greg Kurz: >> > > On Thu, 21 Apr 2016 18:00:19 +0200 >> > > Radim Krčmář <rkrcmar@xxxxxxxxxx> wrote: >> > >> 2016-04-21 16:20+0200, Greg Kurz: >[...] >> > > maybe later >> > > if we have other scenarios where vcpu ids need to cross the limit ? >> > >> > x86 is going to have that soon too -- vcpu_id will be able to range from >> > 0 to 2^32-1 (or 2^31), but MAX_CPUS related data structures probably >> > won't be improved to actually scale, so MAX_CPUS will remain lower. >> > > That's not true, x86 is going to stick with KVM_MAX_VCPUS/qemu's max_cpus, > the only thing that is going to change is that max supported APIC ID > value will be in range 0 to 2^32-1 vs current 8bit one > and since APIC ID is not vcpu_id so it won't affect vcpu_id. I wish it wasn't. vcpu_id is the initial APIC ID -- at least the spec says so and KVM code behaves like that (QEMU does too). It doesn't have to be so, though, KVM_SET_LAPIC provides the interface to set APIC ID. We'd decouple these two and change some related things. (And add yet another cap for that? :]) I'll see what would be really needed, thanks.
