[PATCH v2 00/19] KVM: arm64: Rework timer offsetting for fun and profit

Marc Zyngier maz at kernel.org
Thu Mar 23 15:54:36 PDT 2023


On Thu, 23 Mar 2023 22:19:21 +0000,
Colton Lewis <coltonlewis at google.com> wrote:
> 
> Hello Marc,
> 
> I had thought I sent this earlier but our conversation today made me
> aware I hadn't.
> 
> Marc Zyngier <maz at kernel.org> writes:
> 
> > way. Colton reported some other issues with this test, but I cannot
> > reproduce them here, making me think this might be related to CNTPOFF
> > (but again, I don't have such HW at hand).
> 
> I can no longer reproduce any issues on any platform with this version.

That's an interesting outcome. I haven't quite worked out what makes
it fail or fixes it, unless this is related to the fix I have for the
test itself.

> Two minor comments:
> 
> - I'm not sure you ever addressed my comment from v1 asking if you
>   should add CNTPOFF_EL2 to vcpu_sysreg in
>   arch/arm64/include/asm/kvm_host.h

I don't think it makes much sense until the NV patches support ECV,
which we do not expose to the guest at the moment. For the time
being, we present to the guest something that is more or less an
ARMv8.1 implementation. Once these patches are in, we will be able to
look at adding more features.

> 
> - You left the capital O flag in your selftest changes even though it is
>   no longer used.

Ah, in the getopt() call! Well spotted. I'll fix that now.

> Otherwise,
> 
> Reviewed-by: Colton Lewis <coltonlewis at google.com>

Thanks!

	M.

-- 
Without deviation from the norm, progress is not possible.



More information about the linux-arm-kernel mailing list