kvm: nVHE hyp panic at: __kvm_nvhe_kvm_hyp_handle_sysre

[Naresh Kamboju]

On Wed, 12 Feb 2025 at 19:30, Marc Zyngier <maz at kernel.org> wrote:
>
> On Wed, 12 Feb 2025 11:41:38 +0000,
> Dan Carpenter <dan.carpenter at linaro.org> wrote:
> >
> > On Tue, Feb 11, 2025 at 11:36:31AM +0000, Marc Zyngier wrote:
> > > For the crash at hand, which clearly shows nVHE, can you report
> > > whether the following hack fixes it for you?
> > >
> > >     M.
> >
> > No luck, I'm afraid.  It still crashes the same way.
>
> Right. It was one level deeper. The following change fixes it for
> me. YMMV.

Anders applied this patch and tested on rk3399-rock-pi-4b,
and confirmed that the regression is resolved.

Tested-by: Linux Kernel Functional Testing <lkft at linaro.org>

Link:
 - https://lkft.validation.linaro.org/scheduler/job/8119717#L1251

>
>         M.
>
> diff --git a/arch/arm64/kvm/hyp/include/hyp/switch.h b/arch/arm64/kvm/hyp/include/hyp/switch.h
> index f838a45665f26..b899d815d272f 100644
> --- a/arch/arm64/kvm/hyp/include/hyp/switch.h
> +++ b/arch/arm64/kvm/hyp/include/hyp/switch.h
> @@ -501,9 +501,22 @@ static inline bool handle_tx2_tvm(struct kvm_vcpu *vcpu)
>         return true;
>  }
>
> +/* Open-coded version of timer_get_offset() to allow for kern_hyp_va() */
> +static inline u64 hyp_timer_get_offset(struct arch_timer_context *ctxt)
> +{
> +       u64 offset = 0;
> +
> +       if (ctxt->offset.vm_offset)
> +               offset += *kern_hyp_va(ctxt->offset.vm_offset);
> +       if (ctxt->offset.vcpu_offset)
> +               offset += *kern_hyp_va(ctxt->offset.vcpu_offset);
> +
> +       return offset;
> +}
> +
>  static inline u64 compute_counter_value(struct arch_timer_context *ctxt)
>  {
> -       return arch_timer_read_cntpct_el0() - timer_get_offset(ctxt);
> +       return arch_timer_read_cntpct_el0() - hyp_timer_get_offset(ctxt);
>  }
>
>  static bool kvm_handle_cntxct(struct kvm_vcpu *vcpu)
>
> --
> Without deviation from the norm, progress is not possible.

--
Linaro LKFT
https://lkft.linaro.org