Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1610495 > unrolled thread

[PATCH v2] KVM: x86: cleanup the page tracking SRCU instance

Back to article view | Back to linux.kernel

Contents

  [PATCH v2] KVM: x86: cleanup the page tracking SRCU instance Paolo Bonzini <pbonzini@redhat.com> - 2017-03-28 11:50 +0200
    Re: [PATCH v2] KVM: x86: cleanup the page tracking SRCU instance Xiao Guangrong <xiaoguangrong.eric@gmail.com> - 2017-03-28 13:40 +0200
      Re: [PATCH v2] KVM: x86: cleanup the page tracking SRCU instance Paolo Bonzini <pbonzini@redhat.com> - 2017-03-28 14:10 +0200

#1610495 — [PATCH v2] KVM: x86: cleanup the page tracking SRCU instance

SRCU uses a delayed work item.  Skip cleaning it up, and
the result is use-after-free in the work item callbacks.

Reported-by: Dmitry Vyukov <dvyukov@google.com>
Suggested-by: Dmitry Vyukov <dvyukov@google.com>
Cc: stable@vger.kernel.org
Fixes: 0eb05bf290cfe8610d9680b49abef37febd1c38a
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 arch/x86/include/asm/kvm_page_track.h | 1 +
 arch/x86/kvm/page_track.c             | 8 ++++++++
 arch/x86/kvm/x86.c                    | 1 +
 3 files changed, 10 insertions(+)

diff --git a/arch/x86/include/asm/kvm_page_track.h b/arch/x86/include/asm/kvm_page_track.h
index d74747b031ec..c4eda791f877 100644
--- a/arch/x86/include/asm/kvm_page_track.h
+++ b/arch/x86/include/asm/kvm_page_track.h
@@ -46,6 +46,7 @@ struct kvm_page_track_notifier_node {
 };
 
 void kvm_page_track_init(struct kvm *kvm);
+void kvm_page_track_cleanup(struct kvm *kvm);
 
 void kvm_page_track_free_memslot(struct kvm_memory_slot *free,
 				 struct kvm_memory_slot *dont);
diff --git a/arch/x86/kvm/page_track.c b/arch/x86/kvm/page_track.c
index 37942e419c32..60168cdd0546 100644
--- a/arch/x86/kvm/page_track.c
+++ b/arch/x86/kvm/page_track.c
@@ -160,6 +160,14 @@ bool kvm_page_track_is_active(struct kvm_vcpu *vcpu, gfn_t gfn,
 	return !!ACCESS_ONCE(slot->arch.gfn_track[mode][index]);
 }
 
+void kvm_page_track_cleanup(struct kvm *kvm)
+{
+	struct kvm_page_track_notifier_head *head;
+
+	head = &kvm->arch.track_notifier_head;
+	cleanup_srcu_struct(&head->track_srcu);
+}
+
 void kvm_page_track_init(struct kvm *kvm)
 {
 	struct kvm_page_track_notifier_head *head;
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 64697fe475c3..f2379673912a 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -8070,6 +8070,7 @@ void kvm_arch_sync_events(struct kvm *kvm)
 	cancel_delayed_work_sync(&kvm->arch.kvmclock_update_work);
 	kvm_free_all_assigned_devices(kvm);
 	kvm_free_pit(kvm);
+	kvm_page_track_cleanup(kvm);
 }
 
 int __x86_set_memory_region(struct kvm *kvm, int id, gpa_t gpa, u32 size)
-- 
1.8.3.1

[toc] | [next] | [standalone]

#1610602

On 28/03/2017 5:40 PM, Paolo Bonzini wrote:

>  void kvm_page_track_init(struct kvm *kvm)
>  {
>  	struct kvm_page_track_notifier_head *head;
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 64697fe475c3..f2379673912a 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -8070,6 +8070,7 @@ void kvm_arch_sync_events(struct kvm *kvm)
>  	cancel_delayed_work_sync(&kvm->arch.kvmclock_update_work);
>  	kvm_free_all_assigned_devices(kvm);
>  	kvm_free_pit(kvm);
> +	kvm_page_track_cleanup(kvm);
>  }

Moving it to kvm_arch_destroy_vm() is better as the init
function is called in kvm_arch_init_vm().

Otherwise it looks great to me:
Reviewed-by: Xiao Guangrong <xiaoguangrong.eric@gmail.com>

Thanks for the fix.

[toc] | [prev] | [next] | [standalone]

#1610628

On 28/03/2017 13:25, Xiao Guangrong wrote:
> 
> 
> On 28/03/2017 5:40 PM, Paolo Bonzini wrote:
> 
>>  void kvm_page_track_init(struct kvm *kvm)
>>  {
>>      struct kvm_page_track_notifier_head *head;
>> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
>> index 64697fe475c3..f2379673912a 100644
>> --- a/arch/x86/kvm/x86.c
>> +++ b/arch/x86/kvm/x86.c
>> @@ -8070,6 +8070,7 @@ void kvm_arch_sync_events(struct kvm *kvm)
>>      cancel_delayed_work_sync(&kvm->arch.kvmclock_update_work);
>>      kvm_free_all_assigned_devices(kvm);
>>      kvm_free_pit(kvm);
>> +    kvm_page_track_cleanup(kvm);
>>  }
> 
> Moving it to kvm_arch_destroy_vm() is better as the init
> function is called in kvm_arch_init_vm().

That's fine too, thanks!.

Paolo

> Otherwise it looks great to me:
> Reviewed-by: Xiao Guangrong <xiaoguangrong.eric@gmail.com>
> 
> Thanks for the fix.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web