From patchwork Wed Jun 7 00:43:09 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Sean Christopherson X-Patchwork-Id: 13269873 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 633C4C7EE37 for ; Wed, 7 Jun 2023 00:43:21 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S240052AbjFGAnT (ORCPT ); Tue, 6 Jun 2023 20:43:19 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:46990 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S239980AbjFGAnR (ORCPT ); Tue, 6 Jun 2023 20:43:17 -0400 Received: from mail-pl1-x64a.google.com (mail-pl1-x64a.google.com [IPv6:2607:f8b0:4864:20::64a]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 8098510F8 for ; Tue, 6 Jun 2023 17:43:16 -0700 (PDT) Received: by mail-pl1-x64a.google.com with SMTP id d9443c01a7336-1b03f2663caso25937485ad.1 for ; Tue, 06 Jun 2023 17:43:16 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20221208; t=1686098596; x=1688690596; h=cc:to:from:subject:message-id:references:mime-version:in-reply-to :date:reply-to:from:to:cc:subject:date:message-id:reply-to; bh=9+MS+yTKdJPUgAK9JMZ2lmcpf9TYn8U5kmreyrww4/k=; b=GGY+hlFot4AUKOibBQ139XChSkWEIYm+AJuH7ziOV3noh8d0xYGkgJheiWHDDf0Bed v9HBFZtZe4i3+/JpP0y5mcmRM90VCE64QPtXY6VRx6RGq1qcOOuXjYT0jA4Kj1t2WRgP I7CQwNOG28qJewg2eW7reUEZeizEeQ3gnhyc6tuBtXpyohzy/ROZV+bAkWvzOSz8fuB0 BERmi1KuOFmKn4OBfXrtV0eYz0H+yLMJRRkYwMj1LKV4YBX+UpG5Y0rs2D0CHay5djXA Q8CZmM+wz4focbAiFmowqNKrW6ZuzUN7GyXf21kBC31fvSbhbA7J1tzMq2oqTCcyutwo Xztg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1686098596; x=1688690596; h=cc:to:from:subject:message-id:references:mime-version:in-reply-to :date:reply-to:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=9+MS+yTKdJPUgAK9JMZ2lmcpf9TYn8U5kmreyrww4/k=; b=E8lq5COljjvtmfk33HRSUTBYMD/GLnPueE9twkwWamTw6QlLzQmbnnHMMcnIw5jT/Z s8CFJx4ANWgyLCBxlF/cyZUl9Dn3YGZd3LRMNNcyftjEC35Nf1YEbvwqXj1SuUgEAWpF QQvNw3hyIGjoDPjyu0KEaFZW+ung1SMBUxaPYvdZeAmI3Z30h2Rzb1ZjBiwsydFs3tWz q9cbtuWFLFis6Q+4dt6irt+3C/2WKxONzSWGZ4/zUP10MapfePUtxo6vJhKoMKx0uK/V QIich2gXfSdYZmOAX3HWaLdi6R3OZTfM3xwMqNoFhcRn9EpG/8ddESiemLCtssbk4drY aBRA== X-Gm-Message-State: AC+VfDxHjV1jRFLSTb+wjMBxpEOel+NMdJ5OoaVoQxHQsbO8tSBOcAwv cu3GUT4QNkkOmXHNj9+S3KY6zJulNpA= X-Google-Smtp-Source: ACHHUZ7CqvxrAff/0HHKCH+uMhqE9Ff8aPZiF/5N6ugn3nZaMd2KZSl8LkQmYc8KxfHAqBACT8z0dkPsMf8= X-Received: from zagreus.c.googlers.com ([fda3:e722:ac3:cc00:7f:e700:c0a8:5c37]) (user=seanjc job=sendgmr) by 2002:a17:902:cecf:b0:1b1:86ad:332d with SMTP id d15-20020a170902cecf00b001b186ad332dmr1234293plg.12.1686098596054; Tue, 06 Jun 2023 17:43:16 -0700 (PDT) Reply-To: Sean Christopherson Date: Tue, 6 Jun 2023 17:43:09 -0700 In-Reply-To: <20230607004311.1420507-1-seanjc@google.com> Mime-Version: 1.0 References: <20230607004311.1420507-1-seanjc@google.com> X-Mailer: git-send-email 2.41.0.162.gfafddb0af9-goog Message-ID: <20230607004311.1420507-2-seanjc@google.com> Subject: [PATCH 1/2] KVM: x86: Snapshot host's MSR_IA32_ARCH_CAPABILITIES From: Sean Christopherson To: Sean Christopherson , Paolo Bonzini Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org, Chao Gao , Xiaoyao Li , Pawan Gupta Precedence: bulk List-ID: X-Mailing-List: kvm@vger.kernel.org Snapshot the host's MSR_IA32_ARCH_CAPABILITIES, if it's supported, instead of reading the MSR every time KVM wants to query the host state, e.g. when initializing the default value during vCPU creation. The paths that query ARCH_CAPABILITIES aren't particularly performance sensitive, but creating vCPUs is a frequent enough operation that burning 8 bytes is a good trade-off. Alternatively, KVM could add a field in kvm_caps and thus skip the on-demand calculations entirely, but a pure snapshot isn't possible due to the way KVM handles the l1tf_vmx_mitigation module param. And unlike the other "supported" fields in kvm_caps, KVM doesn't enforce the "supported" value, i.e. KVM treats ARCH_CAPABILITIES like a CPUID leaf and lets userspace advertise whatever it wants. Those problems are solvable, but it's not clear there is real benefit versus snapshotting the host value, and grabbing the host value will allow additional cleanup of KVM's FB_CLEAR_CTRL code. Link: https://lore.kernel.org/all/20230524061634.54141-2-chao.gao@intel.com Cc: Chao Gao Cc: Xiaoyao Li Signed-off-by: Sean Christopherson Reviewed-by: Xiaoyao Li Reviewed-by: Chao Gao --- arch/x86/kvm/vmx/vmx.c | 22 ++++++---------------- arch/x86/kvm/x86.c | 13 +++++++------ arch/x86/kvm/x86.h | 1 + 3 files changed, 14 insertions(+), 22 deletions(-) diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c index 2d9d155691a7..42d1148f933c 100644 --- a/arch/x86/kvm/vmx/vmx.c +++ b/arch/x86/kvm/vmx/vmx.c @@ -255,14 +255,9 @@ static int vmx_setup_l1d_flush(enum vmx_l1d_flush_state l1tf) return 0; } - if (boot_cpu_has(X86_FEATURE_ARCH_CAPABILITIES)) { - u64 msr; - - rdmsrl(MSR_IA32_ARCH_CAPABILITIES, msr); - if (msr & ARCH_CAP_SKIP_VMENTRY_L1DFLUSH) { - l1tf_vmx_mitigation = VMENTER_L1D_FLUSH_NOT_REQUIRED; - return 0; - } + if (host_arch_capabilities & ARCH_CAP_SKIP_VMENTRY_L1DFLUSH) { + l1tf_vmx_mitigation = VMENTER_L1D_FLUSH_NOT_REQUIRED; + return 0; } /* If set to auto use the default l1tf mitigation method */ @@ -373,15 +368,10 @@ static int vmentry_l1d_flush_get(char *s, const struct kernel_param *kp) static void vmx_setup_fb_clear_ctrl(void) { - u64 msr; - - if (boot_cpu_has(X86_FEATURE_ARCH_CAPABILITIES) && + if ((host_arch_capabilities & ARCH_CAP_FB_CLEAR_CTRL) && !boot_cpu_has_bug(X86_BUG_MDS) && - !boot_cpu_has_bug(X86_BUG_TAA)) { - rdmsrl(MSR_IA32_ARCH_CAPABILITIES, msr); - if (msr & ARCH_CAP_FB_CLEAR_CTRL) - vmx_fb_clear_ctrl_available = true; - } + !boot_cpu_has_bug(X86_BUG_TAA)) + vmx_fb_clear_ctrl_available = true; } static __always_inline void vmx_disable_fb_clear(struct vcpu_vmx *vmx) diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 7c7be4815eaa..7c2e796fa460 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -237,6 +237,9 @@ EXPORT_SYMBOL_GPL(enable_apicv); u64 __read_mostly host_xss; EXPORT_SYMBOL_GPL(host_xss); +u64 __read_mostly host_arch_capabilities; +EXPORT_SYMBOL_GPL(host_arch_capabilities); + const struct _kvm_stats_desc kvm_vm_stats_desc[] = { KVM_GENERIC_VM_STATS(), STATS_DESC_COUNTER(VM, mmu_shadow_zapped), @@ -1612,12 +1615,7 @@ static bool kvm_is_immutable_feature_msr(u32 msr) static u64 kvm_get_arch_capabilities(void) { - u64 data = 0; - - if (boot_cpu_has(X86_FEATURE_ARCH_CAPABILITIES)) { - rdmsrl(MSR_IA32_ARCH_CAPABILITIES, data); - data &= KVM_SUPPORTED_ARCH_CAP; - } + u64 data = host_arch_capabilities & KVM_SUPPORTED_ARCH_CAP; /* * If nx_huge_pages is enabled, KVM's shadow paging will ensure that @@ -9492,6 +9490,9 @@ static int __kvm_x86_vendor_init(struct kvm_x86_init_ops *ops) kvm_init_pmu_capability(ops->pmu_ops); + if (boot_cpu_has(X86_FEATURE_ARCH_CAPABILITIES)) + rdmsrl(MSR_IA32_ARCH_CAPABILITIES, host_arch_capabilities); + r = ops->hardware_setup(); if (r != 0) goto out_mmu_exit; diff --git a/arch/x86/kvm/x86.h b/arch/x86/kvm/x86.h index 82e3dafc5453..1e7be1f6ab29 100644 --- a/arch/x86/kvm/x86.h +++ b/arch/x86/kvm/x86.h @@ -323,6 +323,7 @@ fastpath_t handle_fastpath_set_msr_irqoff(struct kvm_vcpu *vcpu); extern u64 host_xcr0; extern u64 host_xss; +extern u64 host_arch_capabilities; extern struct kvm_caps kvm_caps; From patchwork Wed Jun 7 00:43:10 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Sean Christopherson X-Patchwork-Id: 13269874 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 95F5EC7EE29 for ; Wed, 7 Jun 2023 00:43:27 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S240241AbjFGAnZ (ORCPT ); Tue, 6 Jun 2023 20:43:25 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:47038 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S239981AbjFGAnV (ORCPT ); Tue, 6 Jun 2023 20:43:21 -0400 Received: from mail-yw1-x1149.google.com (mail-yw1-x1149.google.com [IPv6:2607:f8b0:4864:20::1149]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id C0EE519A0 for ; Tue, 6 Jun 2023 17:43:19 -0700 (PDT) Received: by mail-yw1-x1149.google.com with SMTP id 00721157ae682-568ae92e492so112176217b3.3 for ; Tue, 06 Jun 2023 17:43:19 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20221208; t=1686098599; x=1688690599; h=cc:to:from:subject:message-id:references:mime-version:in-reply-to :date:reply-to:from:to:cc:subject:date:message-id:reply-to; bh=iWpx+RfSpqGCjC/645vo3rjaj4FpsSPwYJ169RAHthE=; b=pxetOLQocTqwrPnV3EV/w5idyhiG8XYmDrXsHQd2U14s8p5u2EVCgIgk+8I7dqiP93 aOmgTbqkfnkFBItWeLUE9Ow9WgnuTiUuSudnCyfK99s4mHpzPT41YEILu+uZ1ZLpste+ obA95rKvnjRhTB4EeAZMn5VWUGX/TQqhqGdwDSRbo4frz/xTu3GGfK1BVHyimWV/rUcT xhNPGlhvhL3IyT1I8eNtErpOoK1MozEEPby6B2okd253BCEY0Bsr/DBfGQLUwAGZ9ppT oB6J2x+3FOG/A80GYb6uAFqSmGu8sl8qGMxvFPDvQPdkNZFgFyBNgkuBGTxMZoSSDelM NEzA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1686098599; x=1688690599; h=cc:to:from:subject:message-id:references:mime-version:in-reply-to :date:reply-to:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=iWpx+RfSpqGCjC/645vo3rjaj4FpsSPwYJ169RAHthE=; b=WRG8FaC6uSNalPGF00Rgo1qQf7PIG40X9hbbOz3Jni5Bd77iMiCjMezWuqTRezTXEv MqKCGxGkvgON564pZpPHjTqefaEk9Yuk969T101ju9TWX/efC5d8234y7JHtNy8MLzXn WGobLLPjsoSu4ZmKAgLGYckaHZu0zcRdJcPh+K+6C7jEgExQC576LNULbVsXc/fECSlJ Lw38Oqj3/w26YdSvQ/SJL35BZ6n+7vbDGcBWE6dZusF+YrXopwg5YufMK8wbx0LkjVKT ay0Acjd+H/21+zsV6FERhcqxEMp49WFPPr1L/fZ4LvhnyZGJiIqPoXr3EcxvEEKAI6Ug KHNg== X-Gm-Message-State: AC+VfDw2NaUApPsGQM2NdRdRgGkGojMx1/Ayew1dYIENaCEZcN64R+Xn LXwU8M14s79/S4TXnggoyWkhnyCwgsY= X-Google-Smtp-Source: ACHHUZ6WDGc6dSyAYqCWbMICZuWNSAvkeozfOxaiyprE0dooA/9MBteFyjCTI2p5nz/KcLZhnd8fMUGpMOE= X-Received: from zagreus.c.googlers.com ([fda3:e722:ac3:cc00:7f:e700:c0a8:5c37]) (user=seanjc job=sendgmr) by 2002:a81:a7c3:0:b0:568:ee83:d87d with SMTP id e186-20020a81a7c3000000b00568ee83d87dmr1857439ywh.5.1686098599082; Tue, 06 Jun 2023 17:43:19 -0700 (PDT) Reply-To: Sean Christopherson Date: Tue, 6 Jun 2023 17:43:10 -0700 In-Reply-To: <20230607004311.1420507-1-seanjc@google.com> Mime-Version: 1.0 References: <20230607004311.1420507-1-seanjc@google.com> X-Mailer: git-send-email 2.41.0.162.gfafddb0af9-goog Message-ID: <20230607004311.1420507-3-seanjc@google.com> Subject: [PATCH 2/2] KVM: VMX: Drop unnecessary vmx_fb_clear_ctrl_available "cache" From: Sean Christopherson To: Sean Christopherson , Paolo Bonzini Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org, Chao Gao , Xiaoyao Li , Pawan Gupta Precedence: bulk List-ID: X-Mailing-List: kvm@vger.kernel.org Now that KVM snapshots the host's MSR_IA32_ARCH_CAPABILITIES, drop the similar snapshot/cache of whether or not KVM is allowed to manipulate ARCH_CAPABILITIES.FB_CLEAR_CTRL. The motivation for the cache was presumably to avoid the RDMSR, e.g. boot_cpu_has_bug() is quite cheap, and modifying the vCPU's MSR_IA32_ARCH_CAPABILITIES is an infrequent option and a relatively slow path. Cc: Pawan Gupta Signed-off-by: Sean Christopherson Reviewed-by: Pawan Gupta Reviewed-by: Xiaoyao Li --- arch/x86/kvm/vmx/vmx.c | 17 +++-------------- 1 file changed, 3 insertions(+), 14 deletions(-) diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c index 42d1148f933c..17003660138a 100644 --- a/arch/x86/kvm/vmx/vmx.c +++ b/arch/x86/kvm/vmx/vmx.c @@ -237,9 +237,6 @@ static const struct { #define L1D_CACHE_ORDER 4 static void *vmx_l1d_flush_pages; -/* Control for disabling CPU Fill buffer clear */ -static bool __read_mostly vmx_fb_clear_ctrl_available; - static int vmx_setup_l1d_flush(enum vmx_l1d_flush_state l1tf) { struct page *page; @@ -366,14 +363,6 @@ static int vmentry_l1d_flush_get(char *s, const struct kernel_param *kp) return sprintf(s, "%s\n", vmentry_l1d_param[l1tf_vmx_mitigation].option); } -static void vmx_setup_fb_clear_ctrl(void) -{ - if ((host_arch_capabilities & ARCH_CAP_FB_CLEAR_CTRL) && - !boot_cpu_has_bug(X86_BUG_MDS) && - !boot_cpu_has_bug(X86_BUG_TAA)) - vmx_fb_clear_ctrl_available = true; -} - static __always_inline void vmx_disable_fb_clear(struct vcpu_vmx *vmx) { u64 msr; @@ -399,7 +388,9 @@ static __always_inline void vmx_enable_fb_clear(struct vcpu_vmx *vmx) static void vmx_update_fb_clear_dis(struct kvm_vcpu *vcpu, struct vcpu_vmx *vmx) { - vmx->disable_fb_clear = vmx_fb_clear_ctrl_available; + vmx->disable_fb_clear = (host_arch_capabilities & ARCH_CAP_FB_CLEAR_CTRL) && + !boot_cpu_has_bug(X86_BUG_MDS) && + !boot_cpu_has_bug(X86_BUG_TAA); /* * If guest will not execute VERW, there is no need to set FB_CLEAR_DIS @@ -8580,8 +8571,6 @@ static int __init vmx_init(void) if (r) goto err_l1d_flush; - vmx_setup_fb_clear_ctrl(); - for_each_possible_cpu(cpu) { INIT_LIST_HEAD(&per_cpu(loaded_vmcss_on_cpu, cpu));