diff mbox series

KVM: nVMX: fixes for preemption timer migration

Message ID 20200709171507.1819-1-pbonzini@redhat.com (mailing list archive)
State New, archived
Headers show
Series KVM: nVMX: fixes for preemption timer migration | expand

Commit Message

Paolo Bonzini July 9, 2020, 5:15 p.m. UTC
Commit 850448f35aaf ("KVM: nVMX: Fix VMX preemption timer migration",
2020-06-01) accidentally broke nVMX live migration from older version
by changing the userspace ABI.  Restore it and, while at it, ensure
that vmx->nested.has_preemption_timer_deadline is always initialized
according to the KVM_STATE_VMX_PREEMPTION_TIMER_DEADLINE flag.

Cc: Makarand Sonare <makarandsonare@google.com>
Fixes: 850448f35aaf ("KVM: nVMX: Fix VMX preemption timer migration")
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 arch/x86/include/uapi/asm/kvm.h | 5 +++--
 arch/x86/kvm/vmx/nested.c       | 3 ++-
 2 files changed, 5 insertions(+), 3 deletions(-)

Comments

Jim Mattson July 9, 2020, 5:23 p.m. UTC | #1
On Thu, Jul 9, 2020 at 10:15 AM Paolo Bonzini <pbonzini@redhat.com> wrote:
>
> Commit 850448f35aaf ("KVM: nVMX: Fix VMX preemption timer migration",
> 2020-06-01) accidentally broke nVMX live migration from older version
> by changing the userspace ABI.  Restore it and, while at it, ensure
> that vmx->nested.has_preemption_timer_deadline is always initialized
> according to the KVM_STATE_VMX_PREEMPTION_TIMER_DEADLINE flag.
>
> Cc: Makarand Sonare <makarandsonare@google.com>
> Fixes: 850448f35aaf ("KVM: nVMX: Fix VMX preemption timer migration")
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>  arch/x86/include/uapi/asm/kvm.h | 5 +++--
>  arch/x86/kvm/vmx/nested.c       | 3 ++-
>  2 files changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/arch/x86/include/uapi/asm/kvm.h b/arch/x86/include/uapi/asm/kvm.h
> index 17c5a038f42d..0780f97c1850 100644
> --- a/arch/x86/include/uapi/asm/kvm.h
> +++ b/arch/x86/include/uapi/asm/kvm.h
> @@ -408,14 +408,15 @@ struct kvm_vmx_nested_state_data {
>  };
>
>  struct kvm_vmx_nested_state_hdr {
> -       __u32 flags;
>         __u64 vmxon_pa;
>         __u64 vmcs12_pa;
> -       __u64 preemption_timer_deadline;
>
>         struct {
>                 __u16 flags;
>         } smm;
> +
> +       __u32 flags;
> +       __u64 preemption_timer_deadline;
>  };
>
>  struct kvm_svm_nested_state_data {
> diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c
> index b26655104d4a..3fc2411edc92 100644
> --- a/arch/x86/kvm/vmx/nested.c
> +++ b/arch/x86/kvm/vmx/nested.c
> @@ -6180,7 +6180,8 @@ static int vmx_set_nested_state(struct kvm_vcpu *vcpu,
>                 vmx->nested.has_preemption_timer_deadline = true;
>                 vmx->nested.preemption_timer_deadline =
>                         kvm_state->hdr.vmx.preemption_timer_deadline;
> -       }
> +       } else
> +               vmx->nested.has_preemption_timer_deadline = false;

Doesn't the coding standard require braces around the else clause?

Reviewed-by: Jim Mattson <jmattson@google.com>
Paolo Bonzini July 9, 2020, 5:26 p.m. UTC | #2
On 09/07/20 19:23, Jim Mattson wrote:
> On Thu, Jul 9, 2020 at 10:15 AM Paolo Bonzini <pbonzini@redhat.com> wrote:
>>
>> Commit 850448f35aaf ("KVM: nVMX: Fix VMX preemption timer migration",
>> 2020-06-01) accidentally broke nVMX live migration from older version
>> by changing the userspace ABI.  Restore it and, while at it, ensure
>> that vmx->nested.has_preemption_timer_deadline is always initialized
>> according to the KVM_STATE_VMX_PREEMPTION_TIMER_DEADLINE flag.
>>
>> Cc: Makarand Sonare <makarandsonare@google.com>
>> Fixes: 850448f35aaf ("KVM: nVMX: Fix VMX preemption timer migration")
>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>> ---
>>  arch/x86/include/uapi/asm/kvm.h | 5 +++--
>>  arch/x86/kvm/vmx/nested.c       | 3 ++-
>>  2 files changed, 5 insertions(+), 3 deletions(-)
>>
>> diff --git a/arch/x86/include/uapi/asm/kvm.h b/arch/x86/include/uapi/asm/kvm.h
>> index 17c5a038f42d..0780f97c1850 100644
>> --- a/arch/x86/include/uapi/asm/kvm.h
>> +++ b/arch/x86/include/uapi/asm/kvm.h
>> @@ -408,14 +408,15 @@ struct kvm_vmx_nested_state_data {
>>  };
>>
>>  struct kvm_vmx_nested_state_hdr {
>> -       __u32 flags;
>>         __u64 vmxon_pa;
>>         __u64 vmcs12_pa;
>> -       __u64 preemption_timer_deadline;
>>
>>         struct {
>>                 __u16 flags;
>>         } smm;
>> +
>> +       __u32 flags;
>> +       __u64 preemption_timer_deadline;
>>  };
>>
>>  struct kvm_svm_nested_state_data {
>> diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c
>> index b26655104d4a..3fc2411edc92 100644
>> --- a/arch/x86/kvm/vmx/nested.c
>> +++ b/arch/x86/kvm/vmx/nested.c
>> @@ -6180,7 +6180,8 @@ static int vmx_set_nested_state(struct kvm_vcpu *vcpu,
>>                 vmx->nested.has_preemption_timer_deadline = true;
>>                 vmx->nested.preemption_timer_deadline =
>>                         kvm_state->hdr.vmx.preemption_timer_deadline;
>> -       }
>> +       } else
>> +               vmx->nested.has_preemption_timer_deadline = false;
> 
> Doesn't the coding standard require braces around the else clause?

Yes, indeed.

Paolo

> Reviewed-by: Jim Mattson <jmattson@google.com>
>
Bandan Das July 9, 2020, 5:36 p.m. UTC | #3
Jim Mattson <jmattson@google.com> writes:

> On Thu, Jul 9, 2020 at 10:15 AM Paolo Bonzini <pbonzini@redhat.com> wrote:
>>
>> Commit 850448f35aaf ("KVM: nVMX: Fix VMX preemption timer migration",
>> 2020-06-01) accidentally broke nVMX live migration from older version
>> by changing the userspace ABI.  Restore it and, while at it, ensure
>> that vmx->nested.has_preemption_timer_deadline is always initialized
>> according to the KVM_STATE_VMX_PREEMPTION_TIMER_DEADLINE flag.
>>
>> Cc: Makarand Sonare <makarandsonare@google.com>
>> Fixes: 850448f35aaf ("KVM: nVMX: Fix VMX preemption timer migration")
>> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
>> ---
>>  arch/x86/include/uapi/asm/kvm.h | 5 +++--
>>  arch/x86/kvm/vmx/nested.c       | 3 ++-
>>  2 files changed, 5 insertions(+), 3 deletions(-)
>>
>> diff --git a/arch/x86/include/uapi/asm/kvm.h b/arch/x86/include/uapi/asm/kvm.h
>> index 17c5a038f42d..0780f97c1850 100644
>> --- a/arch/x86/include/uapi/asm/kvm.h
>> +++ b/arch/x86/include/uapi/asm/kvm.h
>> @@ -408,14 +408,15 @@ struct kvm_vmx_nested_state_data {
>>  };
>>
>>  struct kvm_vmx_nested_state_hdr {
>> -       __u32 flags;
>>         __u64 vmxon_pa;
>>         __u64 vmcs12_pa;
>> -       __u64 preemption_timer_deadline;
>>
>>         struct {
>>                 __u16 flags;
>>         } smm;
>> +
>> +       __u32 flags;
>> +       __u64 preemption_timer_deadline;
>>  };
>>
Oops!

>>  struct kvm_svm_nested_state_data {
>> diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c
>> index b26655104d4a..3fc2411edc92 100644
>> --- a/arch/x86/kvm/vmx/nested.c
>> +++ b/arch/x86/kvm/vmx/nested.c
>> @@ -6180,7 +6180,8 @@ static int vmx_set_nested_state(struct kvm_vcpu *vcpu,
>>                 vmx->nested.has_preemption_timer_deadline = true;
>>                 vmx->nested.preemption_timer_deadline =
>>                         kvm_state->hdr.vmx.preemption_timer_deadline;
>> -       }
>> +       } else
>> +               vmx->nested.has_preemption_timer_deadline = false;
>
> Doesn't the coding standard require braces around the else clause?
>
I think so... for if/else where at least one of them is multiline.

> Reviewed-by: Jim Mattson <jmattson@google.com>

Looks good to me, 
Reviewed-by: Bandan Das <bsd@redhat.com>
diff mbox series

Patch

diff --git a/arch/x86/include/uapi/asm/kvm.h b/arch/x86/include/uapi/asm/kvm.h
index 17c5a038f42d..0780f97c1850 100644
--- a/arch/x86/include/uapi/asm/kvm.h
+++ b/arch/x86/include/uapi/asm/kvm.h
@@ -408,14 +408,15 @@  struct kvm_vmx_nested_state_data {
 };
 
 struct kvm_vmx_nested_state_hdr {
-	__u32 flags;
 	__u64 vmxon_pa;
 	__u64 vmcs12_pa;
-	__u64 preemption_timer_deadline;
 
 	struct {
 		__u16 flags;
 	} smm;
+
+	__u32 flags;
+	__u64 preemption_timer_deadline;
 };
 
 struct kvm_svm_nested_state_data {
diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c
index b26655104d4a..3fc2411edc92 100644
--- a/arch/x86/kvm/vmx/nested.c
+++ b/arch/x86/kvm/vmx/nested.c
@@ -6180,7 +6180,8 @@  static int vmx_set_nested_state(struct kvm_vcpu *vcpu,
 		vmx->nested.has_preemption_timer_deadline = true;
 		vmx->nested.preemption_timer_deadline =
 			kvm_state->hdr.vmx.preemption_timer_deadline;
-	}
+	} else
+		vmx->nested.has_preemption_timer_deadline = false;
 
 	if (nested_vmx_check_controls(vcpu, vmcs12) ||
 	    nested_vmx_check_host_state(vcpu, vmcs12) ||