Re: [PATCH v2 01/19] KVM: arm64: Hide generic ptrauth from protected VMs without ptrauth

From: Marc Zyngier

Date: Wed Nov 18 2026 - 05:55:21 EST


On Tue, 17 Nov 2026 19:47:36 +0000,
Fuad Tabba <fuad.tabba@xxxxxxxxx> wrote:
>
> Hi Marc,
>
> On Tue, 17 Nov 2026 19:41:54 +0000, Marc Zyngier <maz@xxxxxxxxxx> wrote:
> [...]
> > Burying a fix in a selftest series is a recipe to make sure it is
> > ignored for a full cycle... Why isn't that with the "massive feature
> > set" series?
>
> It's in this one because pkvm_id_regs depends on it. The test checks
> that GPA and GPI read 0 in a protected VM without ptrauth, and it
> fails without the fix.

But that's a *test*. Nobody runs tests as a production workload.
Delaying a fix to satisfy testing makes no sense to me. If someone
screams that their tests are failing, let them scream, safe in the
knowledge that you have fixed the problem rather than the symptom.

> Would you prefer that I send it on its own, or with the respin of the
> features series?

I'll cherry-pick it directly.

Thanks,

M.

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