Track the last TSC offset set for each VM and ensure that the storing of the offset and the reading of the TSC are never preempted by taking a spinlock.
Signed-off-by: Zachary Amsden <[email protected]> --- arch/x86/include/asm/kvm_host.h | 6 +++++- arch/x86/kvm/svm.c | 30 +++++++++++++++++------------- arch/x86/kvm/vmx.c | 21 +++++++-------------- arch/x86/kvm/x86.c | 20 +++++++++++++++++++- arch/x86/kvm/x86.h | 2 ++ 5 files changed, 50 insertions(+), 29 deletions(-) diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 7ec2472..98d4de8 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -410,8 +410,11 @@ struct kvm_arch { gpa_t ept_identity_map_addr; unsigned long irq_sources_bitmap; - u64 vm_init_tsc; s64 kvmclock_offset; + u64 last_tsc_write; + u64 last_tsc_offset; + u64 last_tsc_nsec; + spinlock_t tsc_write_lock; struct kvm_xen_hvm_config xen_hvm_config; @@ -536,6 +539,7 @@ struct kvm_x86_ops { int (*get_lpage_level)(void); bool (*rdtscp_supported)(void); void (*adjust_tsc_offset)(struct kvm_vcpu *vcpu, s64 adjustment); + void (*write_tsc_offset)(struct kvm_vcpu *vcpu, u64 offset); void (*set_supported_cpuid)(u32 func, struct kvm_cpuid_entry2 *entry); diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index ee2cf30..4654507 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -2527,20 +2527,9 @@ static int svm_set_msr(struct kvm_vcpu *vcpu, unsigned ecx, u64 data) struct vcpu_svm *svm = to_svm(vcpu); switch (ecx) { - case MSR_IA32_TSC: { - u64 tsc_offset = data - native_read_tsc(); - u64 g_tsc_offset = 0; - - if (is_nested(svm)) { - g_tsc_offset = svm->vmcb->control.tsc_offset - - svm->nested.hsave->control.tsc_offset; - svm->nested.hsave->control.tsc_offset = tsc_offset; - } - - svm->vmcb->control.tsc_offset = tsc_offset + g_tsc_offset; - + case MSR_IA32_TSC: + guest_write_tsc(vcpu, data); break; - } case MSR_K6_STAR: svm->vmcb->save.star = data; break; @@ -3417,6 +3406,20 @@ static void svm_adjust_tsc_offset(struct kvm_vcpu *vcpu, s64 adjustment) svm->nested.hsave->control.tsc_offset += adjustment; } +static void svm_write_tsc_offset(struct kvm_vcpu *vcpu, u64 offset) +{ + struct vcpu_svm *svm = to_svm(vcpu); + u64 g_tsc_offset = 0; + + if (is_nested(svm)) { + g_tsc_offset = svm->vmcb->control.tsc_offset - + svm->nested.hsave->control.tsc_offset; + svm->nested.hsave->control.tsc_offset = offset; + } + + svm->vmcb->control.tsc_offset = offset + g_tsc_offset; +} + static void svm_fpu_deactivate(struct kvm_vcpu *vcpu) { struct vcpu_svm *svm = to_svm(vcpu); @@ -3503,6 +3506,7 @@ static struct kvm_x86_ops svm_x86_ops = { .set_supported_cpuid = svm_set_supported_cpuid, .adjust_tsc_offset = svm_adjust_tsc_offset, + .write_tsc_offset = svm_write_tsc_offset, }; static int __init svm_init(void) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index a993e67..9b604b8 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -1133,12 +1133,11 @@ static u64 guest_read_tsc(void) } /* - * writes 'guest_tsc' into guest's timestamp counter "register" - * guest_tsc = host_tsc + tsc_offset ==> tsc_offset = guest_tsc - host_tsc + * writes 'offset' into guest's timestamp counter offset register */ -static void guest_write_tsc(u64 guest_tsc, u64 host_tsc) +static void vmx_write_tsc_offset(struct kvm_vcpu *vcpu, u64 offset) { - vmcs_write64(TSC_OFFSET, guest_tsc - host_tsc); + vmcs_write64(TSC_OFFSET, offset); } static void vmx_adjust_tsc_offset(struct kvm_vcpu *vcpu, s64 adjustment) @@ -1217,7 +1216,6 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, u32 msr_index, u64 data) { struct vcpu_vmx *vmx = to_vmx(vcpu); struct shared_msr_entry *msr; - u64 host_tsc; int ret = 0; switch (msr_index) { @@ -1247,8 +1245,7 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, u32 msr_index, u64 data) vmcs_writel(GUEST_SYSENTER_ESP, data); break; case MSR_IA32_TSC: - rdtscll(host_tsc); - guest_write_tsc(data, host_tsc); + guest_write_tsc(vcpu, data); break; case MSR_IA32_CR_PAT: if (vmcs_config.vmentry_ctrl & VM_ENTRY_LOAD_IA32_PAT) { @@ -2503,7 +2500,7 @@ static int vmx_vcpu_setup(struct vcpu_vmx *vmx) { u32 host_sysenter_cs, msr_low, msr_high; u32 junk; - u64 host_pat, tsc_this, tsc_base; + u64 host_pat; unsigned long a; struct desc_ptr dt; int i; @@ -2644,12 +2641,7 @@ static int vmx_vcpu_setup(struct vcpu_vmx *vmx) vmx->vcpu.arch.cr4_guest_owned_bits |= X86_CR4_PGE; vmcs_writel(CR4_GUEST_HOST_MASK, ~vmx->vcpu.arch.cr4_guest_owned_bits); - tsc_base = vmx->vcpu.kvm->arch.vm_init_tsc; - rdtscll(tsc_this); - if (tsc_this < vmx->vcpu.kvm->arch.vm_init_tsc) - tsc_base = tsc_this; - - guest_write_tsc(0, tsc_base); + guest_write_tsc(&vmx->vcpu, 0); return 0; } @@ -4336,6 +4328,7 @@ static struct kvm_x86_ops vmx_x86_ops = { .set_supported_cpuid = vmx_set_supported_cpuid, .adjust_tsc_offset = vmx_adjust_tsc_offset, + .write_tsc_offset = vmx_write_tsc_offset, }; static int __init vmx_init(void) diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 15c7317..ef847ee 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -932,6 +932,24 @@ static inline void kvm_request_guest_time_update(struct kvm_vcpu *v) set_bit(KVM_REQ_CLOCK_SYNC, &v->requests); } +void guest_write_tsc(struct kvm_vcpu *vcpu, u64 data) +{ + struct kvm *kvm = vcpu->kvm; + u64 offset; + + spin_lock(&kvm->arch.tsc_write_lock); + offset = data - native_read_tsc(); + kvm->arch.last_tsc_nsec = get_kernel_ns(); + kvm->arch.last_tsc_write = data; + kvm->arch.last_tsc_offset = offset; + kvm_x86_ops->write_tsc_offset(vcpu, offset); + spin_unlock(&kvm->arch.tsc_write_lock); + + /* Reset of TSC must disable overshoot protection below */ + vcpu->arch.hv_clock.tsc_timestamp = 0; +} +EXPORT_SYMBOL_GPL(guest_write_tsc); + static int kvm_recompute_guest_time(struct kvm_vcpu *v) { unsigned long flags; @@ -5616,7 +5634,7 @@ struct kvm *kvm_arch_create_vm(void) /* Reserve bit 0 of irq_sources_bitmap for userspace irq source */ set_bit(KVM_USERSPACE_IRQ_SOURCE_ID, &kvm->arch.irq_sources_bitmap); - rdtscll(kvm->arch.vm_init_tsc); + spin_lock_init(&kvm->arch.tsc_write_lock); return kvm; } diff --git a/arch/x86/kvm/x86.h b/arch/x86/kvm/x86.h index f4b5445..ce2aff8 100644 --- a/arch/x86/kvm/x86.h +++ b/arch/x86/kvm/x86.h @@ -75,4 +75,6 @@ static inline struct kvm_mem_aliases *kvm_aliases(struct kvm *kvm) void kvm_before_handle_nmi(struct kvm_vcpu *vcpu); void kvm_after_handle_nmi(struct kvm_vcpu *vcpu); +void guest_write_tsc(struct kvm_vcpu *vcpu, u64 data); + #endif -- 1.7.1 -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
