Re: [PATCH 0/3] tools/hvmloader: Decouple APIC IDs from vCPU IDs

[Roger Pau Monné <roger.pau@xxxxxxxxxx>]

On Tue, Jan 28, 2025 at 04:33:39PM +0000, Alejandro Vallejo wrote:
> The hypervisor, hvmloader and the toolstack currently engage in a shared
> assumption that for every vCPU apicid == 2 * vcpuid. This series removes such
> assumption from hvmloader, by making it read the APIC ID of each vCPU and
> storing it for later use.
> 
> The last patch prevents writing an MP Tables should we have vCPUs that can not
> be represented there. That's at the moment dead code because all vCPUs are
> currently representable in 8 bits. This will inavitably stop being true in the
> future after we increase the maximum number of guest vCPUs.

While I'm fine with the MP Table change, should it also come together
with a patch that introduces the code to create x2APIC entries in
libacpi construct_madt() helper? (and bumping the MADT revision, as
I'm quite sure version 2 didn't have x2APIC entries in the
specification).

Otherwise the MP Table change seems like a red herring, because the
MADT created by libacpi will also be incorrect and APIC IDs will wrap in
local APIC entries, just like it would on MP Tables.

Thanks, Roger.