[PATCH] KVM/arm64: Don't emulate a PMU for 32-bit guests if feature not set

Marc Zyngier maz at kernel.org
Wed Apr 27 14:10:01 PDT 2022


On Wed, 27 Apr 2022 10:45:39 +0100,
Alexandru Elisei <alexandru.elisei at arm.com> wrote:
> 
> Hi,
> 
> On Wed, Apr 27, 2022 at 12:57:57AM -0700, Oliver Upton wrote:
> > Hi Alex,
> > 
> > Seems to me all the AArch32 feature register trap logic should come
> > later on as there's a nonzero chance I introduced a bug :) Shall we
> > stop the bleeding w/ your originally proposed patch? Doesn't seem any
> > more objectionable than what we're already doing.
> 
> I am leaning towards merging this patch to prevent people from seeing the
> splat, and when the AArch32 ID reg series gets merged it can be reverted.
> But in the end it's up to Marc to decide what he prefers.

Yeah, I'd like to plug this ASAP. I'll queue the workaround for 5.18,
and we can rework that the proper way for 5.19.

Thanks to both for working on it together, much appreciated.

	M.

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



More information about the linux-arm-kernel mailing list