[PATCH v4 05/20] kvm: arm64: Add helper for loading the stage2 setting for a VM

[christoffer.dall@arm.com (Christoffer Dall)]

On Wed, Jul 18, 2018 at 10:18:48AM +0100, Suzuki K Poulose wrote:
> We load the stage2 context of a guest for different operations,
> including running the guest and tlb maintenance on behalf of the
> guest. As of now only the vttbr is private to the guest, but this
> is about to change with IPA per VM. Add a helper to load the stage2
> configuration for a VM, which could do the right thing with the
> future changes.
> 
> Cc: Christoffer Dall <cdall@kernel.org>
> Cc: Marc Zyngier <marc.zyngier@arm.com>
> Reviewed-by: Eric Auger <eric.auger@redhat.com>
> Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
> ---
> Changes since v2:
>  - New patch
> ---
>  arch/arm64/include/asm/kvm_hyp.h | 6 ++++++
>  arch/arm64/kvm/hyp/switch.c      | 2 +-
>  arch/arm64/kvm/hyp/tlb.c         | 4 ++--
>  3 files changed, 9 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/arm64/include/asm/kvm_hyp.h b/arch/arm64/include/asm/kvm_hyp.h
> index 384c343..82f9994 100644
> --- a/arch/arm64/include/asm/kvm_hyp.h
> +++ b/arch/arm64/include/asm/kvm_hyp.h
> @@ -155,5 +155,11 @@ void deactivate_traps_vhe_put(void);
>  u64 __guest_enter(struct kvm_vcpu *vcpu, struct kvm_cpu_context *host_ctxt);
>  void __noreturn __hyp_do_panic(unsigned long, ...);
>  
> +/* Must be called from hyp code running at EL2 */

more importantly than having to run this at EL2, is that it must have
gone through the proper sequence of update_vttbr() and disabling
interrupts to avoid using a stale VMID.

> +static __always_inline void __hyp_text __load_guest_stage2(struct kvm *kvm)
> +{
> +	write_sysreg(kvm->arch.vttbr, vttbr_el2);
> +}
> +
>  #endif /* __ARM64_KVM_HYP_H__ */
>  
> diff --git a/arch/arm64/kvm/hyp/switch.c b/arch/arm64/kvm/hyp/switch.c
> index d496ef5..355fb25 100644
> --- a/arch/arm64/kvm/hyp/switch.c
> +++ b/arch/arm64/kvm/hyp/switch.c
> @@ -195,7 +195,7 @@ void deactivate_traps_vhe_put(void)
>  
>  static void __hyp_text __activate_vm(struct kvm *kvm)
>  {
> -	write_sysreg(kvm->arch.vttbr, vttbr_el2);
> +	__load_guest_stage2(kvm);
>  }
>  
>  static void __hyp_text __deactivate_vm(struct kvm_vcpu *vcpu)
> diff --git a/arch/arm64/kvm/hyp/tlb.c b/arch/arm64/kvm/hyp/tlb.c
> index 131c777..4dbd9c6 100644
> --- a/arch/arm64/kvm/hyp/tlb.c
> +++ b/arch/arm64/kvm/hyp/tlb.c
> @@ -30,7 +30,7 @@ static void __hyp_text __tlb_switch_to_guest_vhe(struct kvm *kvm)
>  	 * bits. Changing E2H is impossible (goodbye TTBR1_EL2), so
>  	 * let's flip TGE before executing the TLB operation.
>  	 */
> -	write_sysreg(kvm->arch.vttbr, vttbr_el2);
> +	__load_guest_stage2(kvm);
>  	val = read_sysreg(hcr_el2);
>  	val &= ~HCR_TGE;
>  	write_sysreg(val, hcr_el2);
> @@ -39,7 +39,7 @@ static void __hyp_text __tlb_switch_to_guest_vhe(struct kvm *kvm)
>  
>  static void __hyp_text __tlb_switch_to_guest_nvhe(struct kvm *kvm)
>  {
> -	write_sysreg(kvm->arch.vttbr, vttbr_el2);
> +	__load_guest_stage2(kvm);
>  	isb();
>  }
>  
> -- 
> 2.7.4
> 

Thanks,
-Christoffer