Re: [RFC PATCH 13/27] KVM: arm64: Enable access to sanitized CPU features at EL2

[David Brazdil]

> +int copy_ftr_reg(u32 id, struct arm64_ftr_reg *dst)
> +{
> +	struct arm64_ftr_reg *regp = get_arm64_ftr_reg(id);
> +
> +	if (!regp)
> +		return -EINVAL;
> +
> +	memcpy(dst, regp, sizeof(*regp));
> +
> +	return 0;
> +}
> +
>  #define read_sysreg_case(r)	\
>  	case r:		return read_sysreg_s(r)
>  
> diff --git a/arch/arm64/kernel/image-vars.h b/arch/arm64/kernel/image-vars.h
> index dd8ccc9efb6a..c35d768672eb 100644
> --- a/arch/arm64/kernel/image-vars.h
> +++ b/arch/arm64/kernel/image-vars.h
> @@ -116,6 +116,8 @@ __kvm_nvhe___memcpy			= __kvm_nvhe___pi_memcpy;
>  __kvm_nvhe___memset			= __kvm_nvhe___pi_memset;
>  #endif
>  
> +_kvm_nvhe___flush_dcache_area		= __kvm_nvhe___pi___flush_dcache_area;
> +

Could you help my understand why we need this?
* Why do we need PI routines in the first place? Would my series that fixes
  relocations in hyp code remove the need?
* You added these aliases for the string routines because you were worried
  somebody would change the implementation in arch/arm64/lib, right? But this
  cache flush function is defined in hyp/nvhe. So why do we need to point to
  the PI alias if we control the implementation?