Received: by 2002:ac0:bc90:0:0:0:0:0 with SMTP id a16csp3363064img; Mon, 25 Mar 2019 08:48:06 -0700 (PDT) X-Google-Smtp-Source: APXvYqx4NGzx1l0bsSA3t3m+lP8ErBFpXW4nl70ktASOnwnN/TNnMQ9dDMZ6wm8VUtbR554jlZXq X-Received: by 2002:a62:39d6:: with SMTP id u83mr24705415pfj.161.1553528886458; Mon, 25 Mar 2019 08:48:06 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1553528886; cv=none; d=google.com; s=arc-20160816; b=Fu7bHFmqYLo5fLloxO8IPQ/zwN0Y7U78YkHYPLI9czz5GZ/JSCZGJ6d21KbKSHjR0q TfxCKO3H9ZaAkZSewoMVGXW2i/1dLPopM/6DP7/vS9cNOIrUFZ10bgDMvqywowygVkqZ wJ+MARWDIkPMnm0xVNYboB8pI2mlHYjEtNT4oU/d5DGAy3rCp/8EUPzQVxEMAAabmyBJ T9AsdMPQDt7tLDn3c3w4d3zygdrpP+fwf01xv+nS+AY0zW3eOGVKu4r5AMkHSBCSNWIe h8FXStHr4lItdd7q6XUcWGb5RH6mec123F6MTGHOaMLvsGkPhpue5EnkBONlzaTd7CSC OHYg== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=list-id:precedence:sender:user-agent:in-reply-to :content-disposition:mime-version:references:message-id:subject:cc :to:from:date; bh=PRV+o37w44E2ank+r45+jgX+90l8W6P0ea7/vZW3ilQ=; b=AmR8kwIdqym+k96lbG184z0MzbbDXx1t/sOPZHY3QOHlk1ue7CBsCX3yDYyUV9soi9 iyKl2pXSnCgaEji8JSvDympL37OSRaycUKvD+HEcIvk/5skbJbZsAKdwkrN+mOej8YaE 23+jtEPvd+ugDFhnpJjkf4Kax3w/PXsrsZv2Q5KmPGnmhrzZR+rX3zC1xLvEX4p0vTXS AHRPGce+vxehv2Y3OkxERWwFXSXDU72HV5lHauo1VltTprZiCXY98LDIGEM7xrr6RIrf LkJ0bwghFUkFeIo21GSM78W1GtsQWba6qT7mSQ/dabg+XJt4HBC6u0zhYfTmXkbOB8Nn IJUA== ARC-Authentication-Results: i=1; mx.google.com; spf=pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org; dmarc=fail (p=NONE sp=NONE dis=NONE) header.from=intel.com Return-Path: Received: from vger.kernel.org (vger.kernel.org. [209.132.180.67]) by mx.google.com with ESMTP id k17si13390733pfa.181.2019.03.25.08.47.50; Mon, 25 Mar 2019 08:48:06 -0700 (PDT) Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67; Authentication-Results: mx.google.com; spf=pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org; dmarc=fail (p=NONE sp=NONE dis=NONE) header.from=intel.com Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1729429AbfCYPrG (ORCPT + 99 others); Mon, 25 Mar 2019 11:47:06 -0400 Received: from mga02.intel.com ([134.134.136.20]:43920 "EHLO mga02.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726203AbfCYPrG (ORCPT ); Mon, 25 Mar 2019 11:47:06 -0400 X-Amp-Result: UNSCANNABLE X-Amp-File-Uploaded: False Received: from orsmga008.jf.intel.com ([10.7.209.65]) by orsmga101.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 25 Mar 2019 08:47:05 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.60,269,1549958400"; d="scan'208";a="128509561" Received: from sjchrist-coffee.jf.intel.com (HELO linux.intel.com) ([10.54.74.181]) by orsmga008.jf.intel.com with ESMTP; 25 Mar 2019 08:47:05 -0700 Date: Mon, 25 Mar 2019 08:47:05 -0700 From: Sean Christopherson To: Xiaoyao Li Cc: Paolo Bonzini , Radim =?utf-8?B?S3LEjW3DocWZ?= , linux-kernel@vger.kernel.org, kvm@vger.kernel.org, Thomas Gleixner , Ingo Molnar , Borislav Petkov , "H. Peter Anvin" , x86@kernel.org, chao.gao@intel.com Subject: Re: [PATCH v3 2/2] x86/vmx: optimize MSR_MISC_FEATURES_ENABLES switch Message-ID: <20190325154705.GD31069@linux.intel.com> References: <20190325080650.19896-1-xiaoyao.li@linux.intel.com> <20190325080650.19896-3-xiaoyao.li@linux.intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20190325080650.19896-3-xiaoyao.li@linux.intel.com> User-Agent: Mutt/1.5.24 (2015-08-30) Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, Mar 25, 2019 at 04:06:50PM +0800, Xiaoyao Li wrote: > KVM needs to switch MSR_MISC_FEATURES_ENABLES between host and guest in > every pcpu/vcpu context switch. Since WRMSR is expensive, this patch tries > to save cycles by avoiding WRMSR MSR_MISC_FEATURES_ENABLES whenever possible. > > If host's value is zero, nothing needs to be done, since guest can use > kvm emulated cpuid faulting. > > If host's value is non-zero, it need not clear MSR_MISC_FEATURES_ENABLES > unconditionally. We can use hardware cpuid faulting if guest's > value is equal to host'value, thus avoid WRMSR > MSR_MISC_FEATURES_ENABLES. > > Since hardware cpuid faulting takes higher priority than CPUID vm exit, > it should be updated to hardware while guest wrmsr and hardware cpuid > faulting is used for guest. > > Signed-off-by: Xiaoyao Li > --- > arch/x86/include/asm/kvm_host.h | 2 ++ > arch/x86/kvm/vmx/vmx.c | 13 ++++++++++--- > arch/x86/kvm/x86.c | 15 ++++++++++++--- > 3 files changed, 24 insertions(+), 6 deletions(-) > > diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h > index 2c53df4a5a2a..9bcb444b903d 100644 > --- a/arch/x86/include/asm/kvm_host.h > +++ b/arch/x86/include/asm/kvm_host.h > @@ -1343,6 +1343,8 @@ void kvm_lmsw(struct kvm_vcpu *vcpu, unsigned long msw); > void kvm_get_cs_db_l_bits(struct kvm_vcpu *vcpu, int *db, int *l); > int kvm_set_xcr(struct kvm_vcpu *vcpu, u32 index, u64 xcr); > > +int kvm_supported_msr_misc_features_enables(struct kvm_vcpu *vcpu, u64 data); > + > int kvm_get_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr); > int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr); > > diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c > index 65aa947947ba..73bb11f74b36 100644 > --- a/arch/x86/kvm/vmx/vmx.c > +++ b/arch/x86/kvm/vmx/vmx.c > @@ -1035,10 +1035,10 @@ static void vmx_prepare_guest_misc_features_enables(struct vcpu_vmx *vmx) > { > u64 msrval = this_cpu_read(msr_misc_features_shadow); > > - if (!msrval) > + if (!msrval || msrval == vmx->vcpu.arch.msr_misc_features_enables) > return; > > - wrmsrl(MSR_MISC_FEATURES_ENABLES, 0ULL); > + wrmsrl(MSR_MISC_FEATURES_ENABLES, vmx->vcpu.arch.msr_misc_features_enables); > } > > void vmx_prepare_switch_to_guest(struct kvm_vcpu *vcpu) > @@ -1136,7 +1136,7 @@ static void vmx_load_host_misc_features_enables(struct vcpu_vmx *vmx) > { > u64 msrval = this_cpu_read(msr_misc_features_shadow); > > - if (!msrval) > + if (!msrval || msrval == vmx->vcpu.arch.msr_misc_features_enables) > return; > > wrmsrl(MSR_MISC_FEATURES_ENABLES, msrval); > @@ -2043,6 +2043,13 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info) > else > vmx->pt_desc.guest.addr_a[index / 2] = data; > break; > + case MSR_MISC_FEATURES_ENABLES: > + if (!kvm_supported_msr_misc_features_enables(vcpu, data)) > + return 1; > + if (this_cpu_read(msr_misc_features_shadow) && vmx->loaded_cpu_state) > + wrmsrl(MSR_MISC_FEATURES_ENABLES, data); > + vcpu->arch.msr_misc_features_enables = data; > + break; > case MSR_TSC_AUX: > if (!msr_info->host_initiated && > !guest_cpuid_has(vcpu, X86_FEATURE_RDTSCP)) > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c > index ad1df965574e..d2af90422a51 100644 > --- a/arch/x86/kvm/x86.c > +++ b/arch/x86/kvm/x86.c > @@ -2449,6 +2449,17 @@ static void record_steal_time(struct kvm_vcpu *vcpu) > &vcpu->arch.st.steal, sizeof(struct kvm_steal_time)); > } > > +int kvm_supported_msr_misc_features_enables(struct kvm_vcpu *vcpu, u64 data) The name kvm_supported_msr_misc_features_enables() is a bit confusing, e.g. I expected it to return a bitmask of the supported features. Maybe kvm_misc_features_enables_msr_valid() to be consistent with vmx_feature_control_msr_valid()? > +{ > + if (data & ~MSR_MISC_FEATURES_ENABLES_CPUID_FAULT || > + (data & MSR_MISC_FEATURES_ENABLES_CPUID_FAULT && > + !supports_cpuid_fault(vcpu))) > + return 0; > + else > + return 1; No need for the if-else, you can simply do: return (data & ~MSR_MISC_FEATURES_ENABLES_CPUID_FAULT) || (data & MSR_MISC_FEATURES_ENABLES_CPUID_FAULT && !supports_cpuid_fault(vcpu)); You could even shorten it to: return (data & ~MSR_MISC_FEATURES_ENABLES_CPUID_FAULT) || (data && !supports_cpuid_fault(vcpu)); although that might be getting a bit too cute. > +} > +EXPORT_SYMBOL_GPL(kvm_supported_msr_misc_features_enables); > + > int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info) > { > bool pr = false; > @@ -2669,9 +2680,7 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info) > vcpu->arch.msr_platform_info = data; > break; > case MSR_MISC_FEATURES_ENABLES: > - if (data & ~MSR_MISC_FEATURES_ENABLES_CPUID_FAULT || > - (data & MSR_MISC_FEATURES_ENABLES_CPUID_FAULT && > - !supports_cpuid_fault(vcpu))) > + if (!kvm_supported_msr_misc_features_enables(vcpu, data)) > return 1; > vcpu->arch.msr_misc_features_enables = data; > break; > -- > 2.19.1 >