KVM: x86: hyper-v: Avoid writing to TSC page without an active vCPU

The following WARN is triggered from kvm_vm_ioctl_set_clock():
 WARNING: CPU: 10 PID: 579353 at arch/x86/kvm/../../../virt/kvm/kvm_main.c:3161 mark_page_dirty_in_slot+0x6c/0x80 [kvm]
 ...
 CPU: 10 PID: 579353 Comm: qemu-system-x86 Tainted: G        W  O      5.16.0.stable #20
 Hardware name: LENOVO 20UF001CUS/20UF001CUS, BIOS R1CET65W(1.34 ) 06/17/2021
 RIP: 0010:mark_page_dirty_in_slot+0x6c/0x80 [kvm]
 ...
 Call Trace:
  <TASK>
  ? kvm_write_guest+0x114/0x120 [kvm]
  kvm_hv_invalidate_tsc_page+0x9e/0xf0 [kvm]
  kvm_arch_vm_ioctl+0xa26/0xc50 [kvm]
  ? schedule+0x4e/0xc0
  ? __cond_resched+0x1a/0x50
  ? futex_wait+0x166/0x250
  ? __send_signal+0x1f1/0x3d0
  kvm_vm_ioctl+0x747/0xda0 [kvm]
  ...

The WARN was introduced by commit 03c0304a86bc ("KVM: Warn if
mark_page_dirty() is called without an active vCPU") but the change seems
to be correct (unlike Hyper-V TSC page update mechanism). In fact, there's
no real need to actually write to guest memory to invalidate TSC page, this
can be done by the first vCPU which goes through kvm_guest_time_update().

Reported-by: Maxim Levitsky <mlevitsk@redhat.com>
Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20220407201013.963226-1-vkuznets@redhat.com>

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 92843fcdc1cfa30b34480884ed653ddb7ce6e3a2..e0c0f0e1f754c11fa44374a9e4536a6cbcaeb630 100644 (file)
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -974,12 +974,10 @@ enum hv_tsc_page_status {
        HV_TSC_PAGE_UNSET = 0,
        /* TSC page MSR was written by the guest, update pending */
        HV_TSC_PAGE_GUEST_CHANGED,
-       /* TSC page MSR was written by KVM userspace, update pending */
+       /* TSC page update was triggered from the host side */
        HV_TSC_PAGE_HOST_CHANGED,
        /* TSC page was properly set up and is currently active  */
        HV_TSC_PAGE_SET,
-       /* TSC page is currently being updated and therefore is inactive */
-       HV_TSC_PAGE_UPDATING,
        /* TSC page was set up with an inaccessible GPA */
        HV_TSC_PAGE_BROKEN,
 };

diff --git a/arch/x86/kvm/hyperv.c b/arch/x86/kvm/hyperv.c
index 123b677111c58411e38e7fd90aa50c83e0291102..46f9dfb6046947ac62a2c86ef18e0b6ef591a2f4 100644 (file)
--- a/arch/x86/kvm/hyperv.c
+++ b/arch/x86/kvm/hyperv.c
@@ -1135,11 +1135,13 @@ void kvm_hv_setup_tsc_page(struct kvm *kvm,
        BUILD_BUG_ON(sizeof(tsc_seq) != sizeof(hv->tsc_ref.tsc_sequence));
        BUILD_BUG_ON(offsetof(struct ms_hyperv_tsc_page, tsc_sequence) != 0);
 
+       mutex_lock(&hv->hv_lock);
+
        if (hv->hv_tsc_page_status == HV_TSC_PAGE_BROKEN ||
+           hv->hv_tsc_page_status == HV_TSC_PAGE_SET ||
            hv->hv_tsc_page_status == HV_TSC_PAGE_UNSET)
-               return;
+               goto out_unlock;
 
-       mutex_lock(&hv->hv_lock);
        if (!(hv->hv_tsc_page & HV_X64_MSR_TSC_REFERENCE_ENABLE))
                goto out_unlock;
 
@@ -1201,45 +1203,19 @@ out_unlock:
        mutex_unlock(&hv->hv_lock);
 }
 
-void kvm_hv_invalidate_tsc_page(struct kvm *kvm)
+void kvm_hv_request_tsc_page_update(struct kvm *kvm)
 {
        struct kvm_hv *hv = to_kvm_hv(kvm);
-       u64 gfn;
-       int idx;
-
-       if (hv->hv_tsc_page_status == HV_TSC_PAGE_BROKEN ||
-           hv->hv_tsc_page_status == HV_TSC_PAGE_UNSET ||
-           tsc_page_update_unsafe(hv))
-               return;
 
        mutex_lock(&hv->hv_lock);
 
-       if (!(hv->hv_tsc_page & HV_X64_MSR_TSC_REFERENCE_ENABLE))
-               goto out_unlock;
-
-       /* Preserve HV_TSC_PAGE_GUEST_CHANGED/HV_TSC_PAGE_HOST_CHANGED states */
-       if (hv->hv_tsc_page_status == HV_TSC_PAGE_SET)
-               hv->hv_tsc_page_status = HV_TSC_PAGE_UPDATING;
+       if (hv->hv_tsc_page_status == HV_TSC_PAGE_SET &&
+           !tsc_page_update_unsafe(hv))
+               hv->hv_tsc_page_status = HV_TSC_PAGE_HOST_CHANGED;
 
-       gfn = hv->hv_tsc_page >> HV_X64_MSR_TSC_REFERENCE_ADDRESS_SHIFT;
-
-       hv->tsc_ref.tsc_sequence = 0;
-
-       /*
-        * Take the srcu lock as memslots will be accessed to check the gfn
-        * cache generation against the memslots generation.
-        */
-       idx = srcu_read_lock(&kvm->srcu);
-       if (kvm_write_guest(kvm, gfn_to_gpa(gfn),
-                           &hv->tsc_ref, sizeof(hv->tsc_ref.tsc_sequence)))
-               hv->hv_tsc_page_status = HV_TSC_PAGE_BROKEN;
-       srcu_read_unlock(&kvm->srcu, idx);
-
-out_unlock:
        mutex_unlock(&hv->hv_lock);
 }
 
-
 static bool hv_check_msr_access(struct kvm_vcpu_hv *hv_vcpu, u32 msr)
 {
        if (!hv_vcpu->enforce_cpuid)

diff --git a/arch/x86/kvm/hyperv.h b/arch/x86/kvm/hyperv.h
index e19c00ee9ab33c14974af99d4a65db9432d0a2a1..da2737f2a956c2ab054af924b83012e4960997f6 100644 (file)
--- a/arch/x86/kvm/hyperv.h
+++ b/arch/x86/kvm/hyperv.h
@@ -137,7 +137,7 @@ void kvm_hv_process_stimers(struct kvm_vcpu *vcpu);
 
 void kvm_hv_setup_tsc_page(struct kvm *kvm,
                           struct pvclock_vcpu_time_info *hv_clock);
-void kvm_hv_invalidate_tsc_page(struct kvm *kvm);
+void kvm_hv_request_tsc_page_update(struct kvm *kvm);
 
 void kvm_hv_init_vm(struct kvm *kvm);
 void kvm_hv_destroy_vm(struct kvm *kvm);

diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index de49a88df1c2eb2d116e0abbca4e81001baa5c25..547ba00ef64fc3d12f2e3457221b2b96685688e0 100644 (file)
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -2901,7 +2901,7 @@ static void kvm_end_pvclock_update(struct kvm *kvm)
 
 static void kvm_update_masterclock(struct kvm *kvm)
 {
-       kvm_hv_invalidate_tsc_page(kvm);
+       kvm_hv_request_tsc_page_update(kvm);
        kvm_start_pvclock_update(kvm);
        pvclock_update_vm_gtod_copy(kvm);
        kvm_end_pvclock_update(kvm);
@@ -3113,8 +3113,7 @@ static int kvm_guest_time_update(struct kvm_vcpu *v)
                                       offsetof(struct compat_vcpu_info, time));
        if (vcpu->xen.vcpu_time_info_set)
                kvm_setup_pvclock_page(v, &vcpu->xen.vcpu_time_info_cache, 0);
-       if (!v->vcpu_idx)
-               kvm_hv_setup_tsc_page(v->kvm, &vcpu->hv_clock);
+       kvm_hv_setup_tsc_page(v->kvm, &vcpu->hv_clock);
        return 0;
 }
 
@@ -6241,7 +6240,7 @@ static int kvm_vm_ioctl_set_clock(struct kvm *kvm, void __user *argp)
        if (data.flags & ~KVM_CLOCK_VALID_FLAGS)
                return -EINVAL;
 
-       kvm_hv_invalidate_tsc_page(kvm);
+       kvm_hv_request_tsc_page_update(kvm);
        kvm_start_pvclock_update(kvm);
        pvclock_update_vm_gtod_copy(kvm);