[PATCH 2/2] KVM: VMX: Extend VMX's #AC handding

[Xiaoyao Li <xiaoyao.li@intel.com>]

There are two types of #AC can be generated in Intel CPUs:
 1. legacy alignment check #AC;
 2. split lock #AC;

Legacy alignment check #AC can be injected to guest if guest has enabled
alignemnet check.

When host enables split lock detection, i.e., split_lock_detect!=off,
guest will receive an unexpected #AC when there is a split_lock happens in
guest since KVM doesn't virtualize this feature to guest.

Since the old guests lack split_lock #AC handler and may have split lock
buges. To make guest survive from split lock, applying the similar policy
as host's split lock detect configuration:
 - host split lock detect is sld_warn:
   warning the split lock happened in guest, and disabling split lock
   detect around VM-enter;
 - host split lock detect is sld_fatal:
   forwarding #AC to userspace. (Usually userspace dump the #AC
   exception and kill the guest).

Note, if sld_warn and SMT is enabled, the split lock in guest's vcpu
leads the disabling of split lock detect on the sibling CPU thread during
the vcpu is running.

Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
---
 arch/x86/include/asm/cpu.h  |  1 +
 arch/x86/kernel/cpu/intel.c |  6 ++++++
 arch/x86/kvm/vmx/vmx.c      | 42 ++++++++++++++++++++++++++++++++++---
 arch/x86/kvm/vmx/vmx.h      |  3 +++
 4 files changed, 49 insertions(+), 3 deletions(-)

diff --git a/arch/x86/include/asm/cpu.h b/arch/x86/include/asm/cpu.h
index 167d0539e0ad..b46262afa6c1 100644
--- a/arch/x86/include/asm/cpu.h
+++ b/arch/x86/include/asm/cpu.h
@@ -52,6 +52,7 @@ extern enum split_lock_detect_state get_split_lock_detect_state(void);
 extern void __init cpu_set_core_cap_bits(struct cpuinfo_x86 *c);
 extern void switch_to_sld(unsigned long tifn);
 extern bool handle_user_split_lock(struct pt_regs *regs, long error_code);
+extern void split_lock_detect_set(bool on);
 #else
 static inline enum split_lock_detect_state get_split_lock_detect_state(void)
 {
diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c
index 2f9c48e91caf..889469b54b5a 100644
--- a/arch/x86/kernel/cpu/intel.c
+++ b/arch/x86/kernel/cpu/intel.c
@@ -1124,6 +1124,12 @@ void switch_to_sld(unsigned long tifn)
 	__sld_msr_set(!(tifn & _TIF_SLD));
 }
 
+void split_lock_detect_set(bool on)
+{
+	__sld_msr_set(on);
+}
+EXPORT_SYMBOL_GPL(split_lock_detect_set);
+
 #define SPLIT_LOCK_CPU(model) {X86_VENDOR_INTEL, 6, model, X86_FEATURE_ANY}
 
 /*
diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index cdb4bf50ee14..402a9152c6ee 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -4553,6 +4553,12 @@ static int handle_machine_check(struct kvm_vcpu *vcpu)
 	return 1;
 }
 
+static bool guest_cpu_alignment_check_enabled(struct kvm_vcpu *vcpu)
+{
+	return vmx_get_cpl(vcpu) == 3 && kvm_read_cr0_bits(vcpu, X86_CR0_AM) &&
+	       (kvm_get_rflags(vcpu) & X86_EFLAGS_AC);
+}
+
 static int handle_exception_nmi(struct kvm_vcpu *vcpu)
 {
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
@@ -4618,9 +4624,6 @@ static int handle_exception_nmi(struct kvm_vcpu *vcpu)
 		return handle_rmode_exception(vcpu, ex_no, error_code);
 
 	switch (ex_no) {
-	case AC_VECTOR:
-		kvm_queue_exception_e(vcpu, AC_VECTOR, error_code);
-		return 1;
 	case DB_VECTOR:
 		dr6 = vmcs_readl(EXIT_QUALIFICATION);
 		if (!(vcpu->guest_debug &
@@ -4649,6 +4652,29 @@ static int handle_exception_nmi(struct kvm_vcpu *vcpu)
 		kvm_run->debug.arch.pc = vmcs_readl(GUEST_CS_BASE) + rip;
 		kvm_run->debug.arch.exception = ex_no;
 		break;
+	case AC_VECTOR:
+		/*
+		 * Inject #AC back to guest only when legacy alignment check
+		 * enabled.
+		 * Otherwise, it must be a split-lock #AC.
+		 *  - If sld_state == sld_warn, it can let guest survive by
+		 *    setting the vcpu's diasble_split_lock_detect to true so
+		 *    that it will toggle MSR_TEST.SPLIT_LOCK_DETECT bit during
+		 *    every following VM Entry and Exit;
+		 *  - If sld_state == sld_fatal, it forwards #AC to userspace;
+		 */
+		if (guest_cpu_alignment_check_enabled(vcpu) ||
+		    WARN_ON(get_split_lock_detect_state() == sld_off)) {
+			kvm_queue_exception_e(vcpu, AC_VECTOR, error_code);
+			return 1;
+		}
+		if (get_split_lock_detect_state() == sld_warn) {
+			pr_warn("kvm: split lock #AC happened in %s [%d]\n",
+				current->comm, current->pid);
+			vmx->disable_split_lock_detect = true;
+			return 1;
+		}
+		/* fall through*/
 	default:
 		kvm_run->exit_reason = KVM_EXIT_EXCEPTION;
 		kvm_run->ex.exception = ex_no;
@@ -6511,6 +6537,11 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu)
 	 */
 	x86_spec_ctrl_set_guest(vmx->spec_ctrl, 0);
 
+	if (static_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT) &&
+	    !test_tsk_thread_flag(current, TIF_SLD) &&
+	    unlikely(vmx->disable_split_lock_detect))
+		split_lock_detect_set(false);
+
 	/* L1D Flush includes CPU buffer clear to mitigate MDS */
 	if (static_branch_unlikely(&vmx_l1d_should_flush))
 		vmx_l1d_flush(vcpu);
@@ -6545,6 +6576,11 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu)
 
 	x86_spec_ctrl_restore_host(vmx->spec_ctrl, 0);
 
+	if (static_cpu_has(X86_FEATURE_SPLIT_LOCK_DETECT) &&
+	    !test_tsk_thread_flag(current, TIF_SLD) &&
+	    unlikely(vmx->disable_split_lock_detect))
+		split_lock_detect_set(true);
+
 	/* All fields are clean at this point */
 	if (static_branch_unlikely(&enable_evmcs))
 		current_evmcs->hv_clean_fields |=
diff --git a/arch/x86/kvm/vmx/vmx.h b/arch/x86/kvm/vmx/vmx.h
index 7f42cf3dcd70..912eba66c5d5 100644
--- a/arch/x86/kvm/vmx/vmx.h
+++ b/arch/x86/kvm/vmx/vmx.h
@@ -274,6 +274,9 @@ struct vcpu_vmx {
 
 	bool req_immediate_exit;
 
+	/* Disable split-lock detection when running the vCPU */
+	bool disable_split_lock_detect;
+
 	/* Support for PML */
 #define PML_ENTITY_NUM		512
 	struct page *pml_pg;
-- 
2.23.0