[PATCH] KVM: SEV: Add lock subtyping in sev_lock_two_vms so lockdep doesn't report false dependencies

[PATCH] KVM: SEV: Add lock subtyping in sev_lock_two_vms so lockdep doesn't report false dependencies

[Wanpeng Li]

From: Wanpeng Li <wanpengli@tencent.com>

Both source and dest vms' kvm->locks are held in sev_lock_two_vms, 
we should mark one with different subtype to avoid false positives 
from lockdep.

Fixes: c9d61dcb0bc26 (KVM: SEV: accept signals in sev_lock_two_vms)
Reported-by: Yiru Xu <xyru1999@gmail.com>
Tested-by: Jinrong Liang <cloudliang@tencent.com>
Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
---
 arch/x86/kvm/svm/sev.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c
index 7656a2c..be28831 100644
--- a/arch/x86/kvm/svm/sev.c
+++ b/arch/x86/kvm/svm/sev.c
@@ -1565,7 +1565,7 @@ static int sev_lock_two_vms(struct kvm *dst_kvm, struct kvm *src_kvm)
 	r = -EINTR;
 	if (mutex_lock_killable(&dst_kvm->lock))
 		goto release_src;
-	if (mutex_lock_killable(&src_kvm->lock))
+	if (mutex_lock_killable_nested(&src_kvm->lock, SINGLE_DEPTH_NESTING))
 		goto unlock_dst;
 	return 0;
 
-- 
2.7.4

Re: [PATCH] KVM: SEV: Add lock subtyping in sev_lock_two_vms so lockdep doesn't report false dependencies

[Paolo Bonzini]

On 1/5/22 07:41, Wanpeng Li wrote:
> From: Wanpeng Li <wanpengli@tencent.com>
> 
> Both source and dest vms' kvm->locks are held in sev_lock_two_vms,
> we should mark one with different subtype to avoid false positives
> from lockdep.
> 
> Fixes: c9d61dcb0bc26 (KVM: SEV: accept signals in sev_lock_two_vms)
> Reported-by: Yiru Xu <xyru1999@gmail.com>
> Tested-by: Jinrong Liang <cloudliang@tencent.com>
> Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
> ---
>   arch/x86/kvm/svm/sev.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/x86/kvm/svm/sev.c b/arch/x86/kvm/svm/sev.c
> index 7656a2c..be28831 100644
> --- a/arch/x86/kvm/svm/sev.c
> +++ b/arch/x86/kvm/svm/sev.c
> @@ -1565,7 +1565,7 @@ static int sev_lock_two_vms(struct kvm *dst_kvm, struct kvm *src_kvm)
>   	r = -EINTR;
>   	if (mutex_lock_killable(&dst_kvm->lock))
>   		goto release_src;
> -	if (mutex_lock_killable(&src_kvm->lock))
> +	if (mutex_lock_killable_nested(&src_kvm->lock, SINGLE_DEPTH_NESTING))
>   		goto unlock_dst;
>   	return 0;
>   

Queued, thanks.

Paolo