Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To:     linux-kernel@vger.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        stable@vger.kernel.org,
        Pawan Gupta <pawan.kumar.gupta@linux.intel.com>,
        Borislav Petkov <bp@suse.de>,
        Thomas Gleixner <tglx@linutronix.de>
Subject: [PATCH 5.10 10/11] KVM: x86/speculation: Disable Fill buffer clear within guests
Date:   Tue, 14 Jun 2022 20:40:32 +0200
Message-Id: <20220614183722.415331929@linuxfoundation.org>
In-Reply-To: <20220614183719.878453780@linuxfoundation.org>
References: <20220614183719.878453780@linuxfoundation.org>
User-Agent: quilt/0.66
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>

commit 027bbb884be006b05d9c577d6401686053aa789e upstream

The enumeration of MD_CLEAR in CPUID(EAX=7,ECX=0).EDX{bit 10} is not an
accurate indicator on all CPUs of whether the VERW instruction will
overwrite fill buffers. FB_CLEAR enumeration in
IA32_ARCH_CAPABILITIES{bit 17} covers the case of CPUs that are not
vulnerable to MDS/TAA, indicating that microcode does overwrite fill
buffers.

Guests running in VMM environments may not be aware of all the
capabilities/vulnerabilities of the host CPU. Specifically, a guest may
apply MDS/TAA mitigations when a virtual CPU is enumerated as vulnerable
to MDS/TAA even when the physical CPU is not. On CPUs that enumerate
FB_CLEAR_CTRL the VMM may set FB_CLEAR_DIS to skip overwriting of fill
buffers by the VERW instruction. This is done by setting FB_CLEAR_DIS
during VMENTER and resetting on VMEXIT. For guests that enumerate
FB_CLEAR (explicitly asking for fill buffer clear capability) the VMM
will not use FB_CLEAR_DIS.

Irrespective of guest state, host overwrites CPU buffers before VMENTER
to protect itself from an MMIO capable guest, as part of mitigation for
MMIO Stale Data vulnerabilities.

Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 arch/x86/include/asm/msr-index.h       |    6 ++
 arch/x86/kvm/vmx/vmx.c                 |   69 +++++++++++++++++++++++++++++++++
 arch/x86/kvm/vmx/vmx.h                 |    2 
 arch/x86/kvm/x86.c                     |    3 +
 tools/arch/x86/include/asm/msr-index.h |    6 ++
 5 files changed, 86 insertions(+)

--- a/arch/x86/include/asm/msr-index.h
+++ b/arch/x86/include/asm/msr-index.h
@@ -133,6 +133,11 @@
 						 * VERW clears CPU fill buffer
 						 * even on MDS_NO CPUs.
 						 */
+#define ARCH_CAP_FB_CLEAR_CTRL		BIT(18)	/*
+						 * MSR_IA32_MCU_OPT_CTRL[FB_CLEAR_DIS]
+						 * bit available to control VERW
+						 * behavior.
+						 */
 
 #define MSR_IA32_FLUSH_CMD		0x0000010b
 #define L1D_FLUSH			BIT(0)	/*
@@ -150,6 +155,7 @@
 /* SRBDS support */
 #define MSR_IA32_MCU_OPT_CTRL		0x00000123
 #define RNGDS_MITG_DIS			BIT(0)
+#define FB_CLEAR_DIS			BIT(3)	/* CPU Fill buffer clear disable */
 
 #define MSR_IA32_SYSENTER_CS		0x00000174
 #define MSR_IA32_SYSENTER_ESP		0x00000175
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -226,6 +226,9 @@ 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;
@@ -357,6 +360,60 @@ static int vmentry_l1d_flush_get(char *s
 	return sprintf(s, "%s\n", vmentry_l1d_param[l1tf_vmx_mitigation].option);
 }
 
+static void vmx_setup_fb_clear_ctrl(void)
+{
+	u64 msr;
+
+	if (boot_cpu_has(X86_FEATURE_ARCH_CAPABILITIES) &&
+	    !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;
+	}
+}
+
+static __always_inline void vmx_disable_fb_clear(struct vcpu_vmx *vmx)
+{
+	u64 msr;
+
+	if (!vmx->disable_fb_clear)
+		return;
+
+	rdmsrl(MSR_IA32_MCU_OPT_CTRL, msr);
+	msr |= FB_CLEAR_DIS;
+	wrmsrl(MSR_IA32_MCU_OPT_CTRL, msr);
+	/* Cache the MSR value to avoid reading it later */
+	vmx->msr_ia32_mcu_opt_ctrl = msr;
+}
+
+static __always_inline void vmx_enable_fb_clear(struct vcpu_vmx *vmx)
+{
+	if (!vmx->disable_fb_clear)
+		return;
+
+	vmx->msr_ia32_mcu_opt_ctrl &= ~FB_CLEAR_DIS;
+	wrmsrl(MSR_IA32_MCU_OPT_CTRL, vmx->msr_ia32_mcu_opt_ctrl);
+}
+
+static void vmx_update_fb_clear_dis(struct kvm_vcpu *vcpu, struct vcpu_vmx *vmx)
+{
+	vmx->disable_fb_clear = vmx_fb_clear_ctrl_available;
+
+	/*
+	 * If guest will not execute VERW, there is no need to set FB_CLEAR_DIS
+	 * at VMEntry. Skip the MSR read/write when a guest has no use case to
+	 * execute VERW.
+	 */
+	if ((vcpu->arch.arch_capabilities & ARCH_CAP_FB_CLEAR) ||
+	   ((vcpu->arch.arch_capabilities & ARCH_CAP_MDS_NO) &&
+	    (vcpu->arch.arch_capabilities & ARCH_CAP_TAA_NO) &&
+	    (vcpu->arch.arch_capabilities & ARCH_CAP_PSDP_NO) &&
+	    (vcpu->arch.arch_capabilities & ARCH_CAP_FBSDP_NO) &&
+	    (vcpu->arch.arch_capabilities & ARCH_CAP_SBDR_SSDP_NO)))
+		vmx->disable_fb_clear = false;
+}
+
 static const struct kernel_param_ops vmentry_l1d_flush_ops = {
 	.set = vmentry_l1d_flush_set,
 	.get = vmentry_l1d_flush_get,
@@ -2211,6 +2268,10 @@ static int vmx_set_msr(struct kvm_vcpu *
 			ret = kvm_set_msr_common(vcpu, msr_info);
 	}
 
+	/* FB_CLEAR may have changed, also update the FB_CLEAR_DIS behavior */
+	if (msr_index == MSR_IA32_ARCH_CAPABILITIES)
+		vmx_update_fb_clear_dis(vcpu, vmx);
+
 	return ret;
 }
 
@@ -4483,6 +4544,8 @@ static void vmx_vcpu_reset(struct kvm_vc
 	vpid_sync_context(vmx->vpid);
 	if (init_event)
 		vmx_clear_hlt(vcpu);
+
+	vmx_update_fb_clear_dis(vcpu, vmx);
 }
 
 static void enable_irq_window(struct kvm_vcpu *vcpu)
@@ -6658,6 +6721,8 @@ static noinstr void vmx_vcpu_enter_exit(
 		 kvm_arch_has_assigned_device(vcpu->kvm))
 		mds_clear_cpu_buffers();
 
+	vmx_disable_fb_clear(vmx);
+
 	if (vcpu->arch.cr2 != native_read_cr2())
 		native_write_cr2(vcpu->arch.cr2);
 
@@ -6666,6 +6731,8 @@ static noinstr void vmx_vcpu_enter_exit(
 
 	vcpu->arch.cr2 = native_read_cr2();
 
+	vmx_enable_fb_clear(vmx);
+
 	/*
 	 * VMEXIT disables interrupts (host state), but tracing and lockdep
 	 * have them in state 'on' as recorded before entering guest mode.
@@ -8050,6 +8117,8 @@ static int __init vmx_init(void)
 		return r;
 	}
 
+	vmx_setup_fb_clear_ctrl();
+
 	for_each_possible_cpu(cpu) {
 		INIT_LIST_HEAD(&per_cpu(loaded_vmcss_on_cpu, cpu));
 
--- a/arch/x86/kvm/vmx/vmx.h
+++ b/arch/x86/kvm/vmx/vmx.h
@@ -300,6 +300,8 @@ struct vcpu_vmx {
 	u64 msr_ia32_feature_control;
 	u64 msr_ia32_feature_control_valid_bits;
 	u64 ept_pointer;
+	u64 msr_ia32_mcu_opt_ctrl;
+	bool disable_fb_clear;
 
 	struct pt_desc pt_desc;
 
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -1415,6 +1415,9 @@ static u64 kvm_get_arch_capabilities(voi
 		 */
 	}
 
+	/* Guests don't need to know "Fill buffer clear control" exists */
+	data &= ~ARCH_CAP_FB_CLEAR_CTRL;
+
 	return data;
 }
 
--- a/tools/arch/x86/include/asm/msr-index.h
+++ b/tools/arch/x86/include/asm/msr-index.h
@@ -133,6 +133,11 @@
 						 * VERW clears CPU fill buffer
 						 * even on MDS_NO CPUs.
 						 */
+#define ARCH_CAP_FB_CLEAR_CTRL		BIT(18)	/*
+						 * MSR_IA32_MCU_OPT_CTRL[FB_CLEAR_DIS]
+						 * bit available to control VERW
+						 * behavior.
+						 */
 
 #define MSR_IA32_FLUSH_CMD		0x0000010b
 #define L1D_FLUSH			BIT(0)	/*
@@ -150,6 +155,7 @@
 /* SRBDS support */
 #define MSR_IA32_MCU_OPT_CTRL		0x00000123
 #define RNGDS_MITG_DIS			BIT(0)
+#define FB_CLEAR_DIS			BIT(3)	/* CPU Fill buffer clear disable */
 
 #define MSR_IA32_SYSENTER_CS		0x00000174
 #define MSR_IA32_SYSENTER_ESP		0x00000175