Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:6 as permitted sender) client-ip=2620:137:e000::3:6;
Message-ID: <3b41286f-d2b0-5fdf-88ef-1e63f302f9c8@arm.com>
Date:   Thu, 5 Oct 2023 19:05:01 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:102.0)
 Gecko/20100101 Thunderbird/102.15.1
Subject: Re: [PATCH v2 5/6] arm64: KVM: Write TRFCR value on guest switch with
 nVHE
To:     James Clark <james.clark@arm.com>, coresight@lists.linaro.org,
        linux-arm-kernel@lists.infradead.org, kvmarm@lists.linux.dev,
        broonie@kernel.org, maz@kernel.org
Cc:     Oliver Upton <oliver.upton@linux.dev>,
        James Morse <james.morse@arm.com>,
        Zenghui Yu <yuzenghui@huawei.com>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Will Deacon <will@kernel.org>,
        Mike Leach <mike.leach@linaro.org>,
        Leo Yan <leo.yan@linaro.org>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Anshuman Khandual <anshuman.khandual@arm.com>,
        Rob Herring <robh@kernel.org>,
        Jintack Lim <jintack.lim@linaro.org>,
        Akihiko Odaki <akihiko.odaki@daynix.com>,
        Fuad Tabba <tabba@google.com>, Joey Gouly <joey.gouly@arm.com>,
        linux-kernel@vger.kernel.org
References: <20231005125757.649345-1-james.clark@arm.com>
 <20231005125757.649345-6-james.clark@arm.com>
From:   Suzuki K Poulose <suzuki.poulose@arm.com>
In-Reply-To: <20231005125757.649345-6-james.clark@arm.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 05/10/2023 13:57, James Clark wrote:
> The guest value for TRFCR requested by the Coresight driver is saved
> in sysregs[TRFCR_EL1]. On guest switch this value needs to be written to
> the register. Currently TRFCR is only modified when we want to disable
> trace completely in guests due to an issue with TRBE. Expand the
> __debug_save_trace() function to always write to the register if a
> different value for guests is required, but also keep the existing TRBE
> disable behavior if that's required.
> 
> The TRFCR restore function remains functionally the same, except a value
> of 0 doesn't mean "don't restore" anymore. Now that we save both guest
> and host values the register is restored any time the guest and host
> values differ.
> 
> Signed-off-by: James Clark <james.clark@arm.com>
> ---
>   arch/arm64/include/asm/kvm_hyp.h   |  6 ++-
>   arch/arm64/kvm/debug.c             | 13 +++++-
>   arch/arm64/kvm/hyp/nvhe/debug-sr.c | 63 ++++++++++++++++++------------
>   arch/arm64/kvm/hyp/nvhe/switch.c   |  4 +-
>   4 files changed, 57 insertions(+), 29 deletions(-)
> 
> diff --git a/arch/arm64/include/asm/kvm_hyp.h b/arch/arm64/include/asm/kvm_hyp.h
> index 37e238f526d7..0383fd3d60b5 100644
> --- a/arch/arm64/include/asm/kvm_hyp.h
> +++ b/arch/arm64/include/asm/kvm_hyp.h
> @@ -103,8 +103,10 @@ void __debug_switch_to_guest(struct kvm_vcpu *vcpu);
>   void __debug_switch_to_host(struct kvm_vcpu *vcpu);
>   
>   #ifdef __KVM_NVHE_HYPERVISOR__
> -void __debug_save_host_buffers_nvhe(struct kvm_cpu_context *host_ctxt);
> -void __debug_restore_host_buffers_nvhe(struct kvm_cpu_context *host_ctxt);
> +void __debug_save_host_buffers_nvhe(struct kvm_cpu_context *host_ctxt,
> +				    struct kvm_cpu_context *guest_ctxt);
> +void __debug_restore_host_buffers_nvhe(struct kvm_cpu_context *host_ctxt,
> +				       struct kvm_cpu_context *guest_ctxt);
>   #endif
>   
>   void __fpsimd_save_state(struct user_fpsimd_state *fp_regs);
> diff --git a/arch/arm64/kvm/debug.c b/arch/arm64/kvm/debug.c
> index 19e722359154..d949dd354464 100644
> --- a/arch/arm64/kvm/debug.c
> +++ b/arch/arm64/kvm/debug.c
> @@ -337,10 +337,21 @@ void kvm_arch_vcpu_load_debug_state_flags(struct kvm_vcpu *vcpu)
>   	    !(read_sysreg_s(SYS_PMBIDR_EL1) & BIT(PMBIDR_EL1_P_SHIFT)))
>   		vcpu_set_flag(vcpu, DEBUG_STATE_SAVE_SPE);
>   
> -	/* Check if we have TRBE implemented and available at the host */
> +	/*
> +	 * Check if we have TRBE implemented and available at the host. If it's
> +	 * in use at the time of guest switch it will need to be disabled and
> +	 * then restored.
> +	 */
>   	if (cpuid_feature_extract_unsigned_field(dfr0, ID_AA64DFR0_EL1_TraceBuffer_SHIFT) &&
>   	    !(read_sysreg_s(SYS_TRBIDR_EL1) & TRBIDR_EL1_P))
>   		vcpu_set_flag(vcpu, DEBUG_STATE_SAVE_TRFCR);

As per A3.1 Armv9-A architecture extensions (DDI 0487J.a), FEAT_TRBE
mandates FEAT_TRF. So, we could check FEAT_TRF and if we have a hit, 
skip the TRBE checks. But, having read the code below, it looks like
we need separate flags for TRFCR and TRBE.
	
> +	/*
> +	 * Also save TRFCR on nVHE if FEAT_TRF (TraceFilt) exists. This will be
> +	 * done in cases where use of TRBE doesn't completely disable trace and
> +	 * handles the exclude_host/exclude_guest rules of the trace session.
> +	 */
> +	if (cpuid_feature_extract_unsigned_field(dfr0, ID_AA64DFR0_EL1_TraceFilt_SHIFT))
> +		vcpu_set_flag(vcpu, DEBUG_STATE_SAVE_TRFCR);
	
>   }
>   
>   void kvm_arch_vcpu_put_debug_state_flags(struct kvm_vcpu *vcpu)
> diff --git a/arch/arm64/kvm/hyp/nvhe/debug-sr.c b/arch/arm64/kvm/hyp/nvhe/debug-sr.c
> index 128a57dddabf..c6252029c277 100644
> --- a/arch/arm64/kvm/hyp/nvhe/debug-sr.c
> +++ b/arch/arm64/kvm/hyp/nvhe/debug-sr.c
> @@ -51,42 +51,56 @@ static void __debug_restore_spe(struct kvm_cpu_context *host_ctxt)
>   	write_sysreg_s(ctxt_sys_reg(host_ctxt, PMSCR_EL1), SYS_PMSCR_EL1);
>   }
>   
> -static void __debug_save_trace(struct kvm_cpu_context *host_ctxt)
> +/*
> + * Save TRFCR and disable trace completely if TRBE is being used, otherwise
> + * apply required guest TRFCR value.
> + */
> +static void __debug_save_trace(struct kvm_cpu_context *host_ctxt,
> +			       struct kvm_cpu_context *guest_ctxt)
>   {
> -	ctxt_sys_reg(host_ctxt, TRFCR_EL1) = 0;
> +	ctxt_sys_reg(host_ctxt, TRFCR_EL1) = read_sysreg_s(SYS_TRFCR_EL1);
>   
>   	/* Check if the TRBE is enabled */
> -	if (!(read_sysreg_s(SYS_TRBLIMITR_EL1) & TRBLIMITR_EL1_E))
> -		return;
> -	/*
> -	 * Prohibit trace generation while we are in guest.
> -	 * Since access to TRFCR_EL1 is trapped, the guest can't
> -	 * modify the filtering set by the host.
> -	 */
> -	ctxt_sys_reg(host_ctxt, TRFCR_EL1) = read_sysreg_s(SYS_TRFCR_EL1);
> -	write_sysreg_s(0, SYS_TRFCR_EL1);
> -	isb();
> -	/* Drain the trace buffer to memory */
> -	tsb_csync();
> +	if (read_sysreg_s(SYS_TRBLIMITR_EL1) & TRBLIMITR_EL1_E) {

This is problematic. At this point, we are not sure if TRBE is available
or not (e.g. we could be on a v8.4 CPU or a v9.0 with TRBE disabled by
higher EL). May be we need to add a separate flag to indicate the
presence of TRBE.

Suzuki

> +		/*
> +		 * Prohibit trace generation while we are in guest. Since access
> +		 * to TRFCR_EL1 is trapped, the guest can't modify the filtering
> +		 * set by the host.
> +		 */
> +		ctxt_sys_reg(guest_ctxt, TRFCR_EL1) = 0;
> +		write_sysreg_s(0, SYS_TRFCR_EL1);
> +		isb();
> +		/* Drain the trace buffer to memory */
> +		tsb_csync();
> +	} else {
> +		/*
> +		 * Not using TRBE, so guest trace works. Apply the guest filters
> +		 * provided by the Coresight driver, if different.
> +		 */
> +		if (ctxt_sys_reg(host_ctxt, TRFCR_EL1) !=
> +		    ctxt_sys_reg(guest_ctxt, TRFCR_EL1))
> +			write_sysreg_s(ctxt_sys_reg(guest_ctxt, TRFCR_EL1),
> +				       SYS_TRFCR_EL1);
> +	}
>   }
>   
> -static void __debug_restore_trace(struct kvm_cpu_context *host_ctxt)
> +static void __debug_restore_trace(struct kvm_cpu_context *host_ctxt,
> +				  struct kvm_cpu_context *guest_ctxt)
>   {
> -	if (!ctxt_sys_reg(host_ctxt, TRFCR_EL1))
> -		return;
> -
>   	/* Restore trace filter controls */
> -	write_sysreg_s(ctxt_sys_reg(host_ctxt, TRFCR_EL1), SYS_TRFCR_EL1);
> +	if (ctxt_sys_reg(host_ctxt, TRFCR_EL1) != ctxt_sys_reg(guest_ctxt, TRFCR_EL1))
> +		write_sysreg_s(ctxt_sys_reg(host_ctxt, TRFCR_EL1), SYS_TRFCR_EL1);
>   }
>   
> -void __debug_save_host_buffers_nvhe(struct kvm_cpu_context *host_ctxt)
> +void __debug_save_host_buffers_nvhe(struct kvm_cpu_context *host_ctxt,
> +				    struct kvm_cpu_context *guest_ctxt)
>   {
>   	/* Disable and flush SPE data generation */
>   	if (vcpu_get_flag(host_ctxt->__hyp_running_vcpu, DEBUG_STATE_SAVE_SPE))
>   		__debug_save_spe(host_ctxt);
> -	/* Disable and flush Self-Hosted Trace generation */
> +
>   	if (vcpu_get_flag(host_ctxt->__hyp_running_vcpu, DEBUG_STATE_SAVE_TRFCR))
> -		__debug_save_trace(host_ctxt);
> +		__debug_save_trace(host_ctxt, guest_ctxt);
>   }
>   
>   void __debug_switch_to_guest(struct kvm_vcpu *vcpu)
> @@ -94,12 +108,13 @@ void __debug_switch_to_guest(struct kvm_vcpu *vcpu)
>   	__debug_switch_to_guest_common(vcpu);
>   }
>   
> -void __debug_restore_host_buffers_nvhe(struct kvm_cpu_context *host_ctxt)
> +void __debug_restore_host_buffers_nvhe(struct kvm_cpu_context *host_ctxt,
> +				       struct kvm_cpu_context *guest_ctxt)
>   {
>   	if (vcpu_get_flag(host_ctxt->__hyp_running_vcpu, DEBUG_STATE_SAVE_SPE))
>   		__debug_restore_spe(host_ctxt);
>   	if (vcpu_get_flag(host_ctxt->__hyp_running_vcpu, DEBUG_STATE_SAVE_TRFCR))
> -		__debug_restore_trace(host_ctxt);
> +		__debug_restore_trace(host_ctxt, guest_ctxt);
>   }
>   
>   void __debug_switch_to_host(struct kvm_vcpu *vcpu)
> diff --git a/arch/arm64/kvm/hyp/nvhe/switch.c b/arch/arm64/kvm/hyp/nvhe/switch.c
> index c8f15e4dab19..55207ec31bd3 100644
> --- a/arch/arm64/kvm/hyp/nvhe/switch.c
> +++ b/arch/arm64/kvm/hyp/nvhe/switch.c
> @@ -276,7 +276,7 @@ int __kvm_vcpu_run(struct kvm_vcpu *vcpu)
>   	 * translation regime to EL2 (via MDCR_EL2_E2PB == 0) and
>   	 * before we load guest Stage1.
>   	 */
> -	__debug_save_host_buffers_nvhe(host_ctxt);
> +	__debug_save_host_buffers_nvhe(host_ctxt, guest_ctxt);
>   
>   	/*
>   	 * We're about to restore some new MMU state. Make sure
> @@ -343,7 +343,7 @@ int __kvm_vcpu_run(struct kvm_vcpu *vcpu)
>   	 * This must come after restoring the host sysregs, since a non-VHE
>   	 * system may enable SPE here and make use of the TTBRs.
>   	 */
> -	__debug_restore_host_buffers_nvhe(host_ctxt);
> +	__debug_restore_host_buffers_nvhe(host_ctxt, guest_ctxt);
>   
>   	if (pmu_switch_needed)
>   		__pmu_switch_to_host(vcpu);