[PATCH 1/3] arm64: KVM: Move CPU ID reg trap setup off the world switch path

[christoffer.dall@linaro.org (Christoffer Dall)]

On Fri, Dec 01, 2017 at 03:19:40PM +0000, Dave Martin wrote:
> The HCR_EL2.TID3 flag needs to be set when trapping guest access to
> the CPU ID registers is required.  However, the decision about
> whether to set this bit does not need to be repeated at every
> switch to the guest.
> 
> Instead, it's sufficient to make this decision once and record the
> outcome.
> 
> This patch moves the decision to vcpu_reset_hcr() and records the
> choice made in vcpu->arch.hcr_el2.  The world switch code can then
> load this directly when switching to the guest without the need for
> conditional logic on the critical path.
> 
> Signed-off-by: Dave Martin <Dave.Martin@arm.com>
> Suggested-by: Christoffer Dall <christoffer.dall@linaro.org>
> Cc: Marc Zyngier <marc.zyngier@arm.com>

Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>

> 
> ---
> 
> Note to maintainers: this was discussed on-list [1] prior to the merge
> window, but this patch implementing the agreed decision hasn't been
> posted previously.
> 
> This should be considered a fix for v4.15.

It's actually easier for me to apply this for v4.16 and base my VHE
optimization patches on it.

Thanks,
-Christoffer

> 
> [1] [PATCH v3 02/28] arm64: KVM: Hide unsupported AArch64 CPU features from guests
> http://lists.infradead.org/pipermail/linux-arm-kernel/2017-October/537420.html
> ---
>  arch/arm64/include/asm/kvm_emulate.h | 8 ++++++++
>  arch/arm64/kvm/hyp/switch.c          | 4 ----
>  2 files changed, 8 insertions(+), 4 deletions(-)
> 
> diff --git a/arch/arm64/include/asm/kvm_emulate.h b/arch/arm64/include/asm/kvm_emulate.h
> index 5f28dfa..8ff5aef 100644
> --- a/arch/arm64/include/asm/kvm_emulate.h
> +++ b/arch/arm64/include/asm/kvm_emulate.h
> @@ -52,6 +52,14 @@ static inline void vcpu_reset_hcr(struct kvm_vcpu *vcpu)
>  		vcpu->arch.hcr_el2 |= HCR_E2H;
>  	if (test_bit(KVM_ARM_VCPU_EL1_32BIT, vcpu->arch.features))
>  		vcpu->arch.hcr_el2 &= ~HCR_RW;
> +
> +	/*
> +	 * TID3: trap feature register accesses that we virtualise.
> +	 * For now this is conditional, since no AArch32 feature regs
> +	 * are currently virtualised.
> +	 */
> +	if (vcpu->arch.hcr_el2 & HCR_RW)
> +		vcpu->arch.hcr_el2 |= HCR_TID3;
>  }
>  
>  static inline unsigned long vcpu_get_hcr(struct kvm_vcpu *vcpu)
> diff --git a/arch/arm64/kvm/hyp/switch.c b/arch/arm64/kvm/hyp/switch.c
> index 525c01f..87fd590 100644
> --- a/arch/arm64/kvm/hyp/switch.c
> +++ b/arch/arm64/kvm/hyp/switch.c
> @@ -86,10 +86,6 @@ static void __hyp_text __activate_traps(struct kvm_vcpu *vcpu)
>  		write_sysreg(1 << 30, fpexc32_el2);
>  		isb();
>  	}
> -
> -	if (val & HCR_RW) /* for AArch64 only: */
> -		val |= HCR_TID3; /* TID3: trap feature register accesses */
> -
>  	write_sysreg(val, hcr_el2);
>  
>  	/* Trap on AArch32 cp15 c15 accesses (EL1 or EL0) */
> -- 
> 2.1.4
>