Re: [PATCH 2/2] KVM: arm64: Ensure vgic_ready() is ordered against MMIO registration

[Oliver Upton <oliver.upton@linux.dev>]

On Thu, Oct 17, 2024 at 08:58:30AM +0100, Marc Zyngier wrote:
> On Thu, 17 Oct 2024 01:19:47 +0100, Oliver Upton <oliver.upton@linux.dev> wrote:
> I thought we would benefit from the removal of the early-out at the
> beginning of the function, specially given that this is only called on
> the first run of any vcpu -- and we can probably afford to take the
> locks on such a rare event. My impression is that we'd gain in
> robustness and maintainability.
> 
> This can come as a later patch if we agree that this is valuable.

I should've mentioned my rationale a bit more here, thanks for queuing
the patch anyway. Prolonged blackout for live migration is a problem for
larger VMs, and I was too chicken to have all vCPUs pile up on slots_lock
and config_lock.

Too lazy to prove this is/isn't an issue, so I figured I'd preserve the
existing flow as much as possible. If we get this wrong again (which
wouldn't be surprising) then let's drop the early out and just take the
locks.

Seem reasonable?

-- 
Thanks,
Oliver