Re: [PATCH v7 11/13] KVM: x86: Implement Intel Processor Trace MSRs read/write

[Alexander Shishkin <alexander.shishkin@linux.intel.com>]

On Thu, May 03, 2018 at 08:08:41PM +0800, Luwei Kang wrote:
> From: Chao Peng <chao.p.peng@linux.intel.com>
> 
> Implement Intel Processor Trace MSRs read/write.

There needs to be a commit message here.

> Signed-off-by: Chao Peng <chao.p.peng@linux.intel.com>
> Signed-off-by: Luwei Kang <luwei.kang@intel.com>
> ---
>  arch/x86/include/asm/intel_pt.h |   8 ++
>  arch/x86/kvm/vmx.c              | 163 ++++++++++++++++++++++++++++++++++++++++
>  arch/x86/kvm/x86.c              |  33 +++++++-
>  3 files changed, 203 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/include/asm/intel_pt.h b/arch/x86/include/asm/intel_pt.h
> index 43ad260..dc0f3f0 100644
> --- a/arch/x86/include/asm/intel_pt.h
> +++ b/arch/x86/include/asm/intel_pt.h
> @@ -5,6 +5,14 @@
>  #define PT_CPUID_LEAVES		2
>  #define PT_CPUID_REGS_NUM	4 /* number of regsters (eax, ebx, ecx, edx) */
>  
> +#define MSR_IA32_RTIT_STATUS_MASK (~(RTIT_STATUS_FILTEREN | \
> +		RTIT_STATUS_CONTEXTEN | RTIT_STATUS_TRIGGEREN | \
> +		RTIT_STATUS_ERROR | RTIT_STATUS_STOPPED | \
> +		RTIT_STATUS_BYTECNT))
> +
> +#define MSR_IA32_RTIT_OUTPUT_BASE_MASK \
> +	(~((1UL << cpuid_query_maxphyaddr(vcpu)) - 1) | 0x7f)

How does this macro make sense in the intel_pt.h? It also relies on vcpu
being in the scope.

>  enum pt_mode {
>  	PT_MODE_SYSTEM = 0,
>  	PT_MODE_HOST,
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index 3ed02a8..2a29ab9 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -2769,6 +2769,77 @@ static void vmx_set_interrupt_shadow(struct kvm_vcpu *vcpu, int mask)
>  		vmcs_write32(GUEST_INTERRUPTIBILITY_INFO, interruptibility);
>  }
>  
> +static int vmx_rtit_ctl_check(struct kvm_vcpu *vcpu, u64 data)
> +{
> +	struct vcpu_vmx *vmx = to_vmx(vcpu);
> +	unsigned long value;
> +
> +	/*
> +	 * Any MSR write that attempts to change bits marked reserved will
> +	 * case a #GP fault.
> +	 */
> +	if (data & vmx->pt_desc.ctl_bitmask)
> +		return 1;
> +
> +	/*
> +	 * Any attempt to modify IA32_RTIT_CTL while TraceEn is set will
> +	 * result in a #GP unless the same write also clears TraceEn.
> +	 */
> +	if ((vmx->pt_desc.guest.ctl & RTIT_CTL_TRACEEN) &&
> +		((vmx->pt_desc.guest.ctl ^ data) & ~RTIT_CTL_TRACEEN))
> +		return 1;
> +
> +	/*
> +	 * WRMSR to IA32_RTIT_CTL that sets TraceEn but clears this bit
> +	 * and FabricEn would cause #GP, if
> +	 * CPUID.(EAX=14H, ECX=0):ECX.SNGLRGNOUT[bit 2] = 0
> +	 */
> +	if ((data & RTIT_CTL_TRACEEN) && !(data & RTIT_CTL_TOPA) &&
> +		!(data & RTIT_CTL_FABRIC_EN) &&
> +		!__pt_cap_get(vmx->pt_desc.caps, PT_CAP_single_range_output))

You seem to be doing a lot of __pt_cap_get()s on each wrmsr. Did you consider
decoding the capabilities once and storing the decoded values instead, so
that in functions like these you can access them by

   if (vmx->pt_desc.caps[PT_CAP_single_range_output]) ...

?

Regards,
--
Alex