Re: [PATCH v6 6/9] KVM: x86: lapic: don't allow to change APIC ID unconditionally

[Maxim Levitsky]

On Fri, 2022-02-25 at 15:42 +0000, David Woodhouse wrote:
> On Fri, 2022-02-25 at 17:11 +0200, Maxim Levitsky wrote:
> > On Fri, 2022-02-25 at 14:56 +0000, David Woodhouse wrote:
> > > On Fri, 2022-02-25 at 16:46 +0200, Maxim Levitsky wrote:
> > > > Assuming that this is approved and accepted upstream,
> > > > that is even better that my proposal of doing this
> > > > when APICv is enabled.
> > > > 
> > > > Since now apic id is always read only, now we should not 
> > > > forget to clean up some parts of kvm like kvm_recalculate_apic_map,
> > > > which are not needed anymore
> > > 
> > > Can we also now optimise kvm_get_vcpu_by_id() so it doesn't have to do
> > > a linear search over all the vCPUs when there isn't a 1:1
> > > correspondence with the vCPU index?
> > 
> > I don't think so since vcpu id can still be set by userspace to anything,
> > and this is even used to encode topology in it.
> 
> Yes, but it can only be set at vCPU creation time and it has to be
> unique.
> 
> > However a hash table can still be used there to speed it up regardless of
> > read-only apic id IMHO.
> > 
> > Or, even better than a hash table, I see that KVM already 
> > limits vcpu_id to KVM_MAX_VCPUS * 4 with a comment that only two extra
> > bits of topology are used:
> 
> We already have the kvm_apic_map which provides a fast lookup. The key
> point here is that the APIC ID can't *change* from vcpu->vcpu_id any
> more, so we can actually use the kvm_apic_map for kvm_get_vcpu_by_id()
> now, can't we?
> 
Right! I wrote my response partially when I still assumed that vcpu_id
can be any 32 bit number (thus hash table), 
and later checked that it is capped by KVM_MAX_VCPUS * 4 which isn't a big number, 
plus as I now see in the kvm_recalculate_apic_map
the map is dynamically allocated up to the max apic id.

(technically speaking an array is a hash table).

Now the map would only be needed to be rebuit few times when new vCPUs are added,
and can be used to locate vcpu by its apic id.

I so hope that this patch is accepted so all of this could be done.

Best regards,
	Maxim Levitsky