[PATCH] kvm/vfio: Fix potential deadlock problem in vfio

[PATCH] kvm/vfio: Fix potential deadlock problem in vfio

[Wan Jiabing]

Fix following coccicheck warning:
./virt/kvm/vfio.c:258:1-7: preceding lock on line 236

If kvm_vfio_file_iommu_group() failed, code would goto err_fdput with
mutex_lock acquired and then return ret. It might cause potential
deadlock. Move mutex_unlock bellow err_fdput tag to fix it. 

Fixes: d55d9e7a45721 ("kvm/vfio: Store the struct file in the kvm_vfio_group")
Signed-off-by: Wan Jiabing <wanjiabing@vivo.com>
---
 virt/kvm/vfio.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/virt/kvm/vfio.c b/virt/kvm/vfio.c
index 8f9f7fffb96a..ce1b01d02c51 100644
--- a/virt/kvm/vfio.c
+++ b/virt/kvm/vfio.c
@@ -252,8 +252,8 @@ static int kvm_vfio_group_set_spapr_tce(struct kvm_device *dev,
 		break;
 	}
 
-	mutex_unlock(&kv->lock);
 err_fdput:
+	mutex_unlock(&kv->lock);
 	fdput(f);
 	return ret;
 }
-- 
2.36.1

Re: [PATCH] kvm/vfio: Fix potential deadlock problem in vfio

[Jason Gunthorpe]

On Tue, May 17, 2022 at 10:34:41AM +0800, Wan Jiabing wrote:
> Fix following coccicheck warning:
> ./virt/kvm/vfio.c:258:1-7: preceding lock on line 236
> 
> If kvm_vfio_file_iommu_group() failed, code would goto err_fdput with
> mutex_lock acquired and then return ret. It might cause potential
> deadlock. Move mutex_unlock bellow err_fdput tag to fix it. 
> 
> Fixes: d55d9e7a45721 ("kvm/vfio: Store the struct file in the kvm_vfio_group")
> Signed-off-by: Wan Jiabing <wanjiabing@vivo.com>
> ---
>  virt/kvm/vfio.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>

Thanks,
Jason

Re: [PATCH] kvm/vfio: Fix potential deadlock problem in vfio

[Alex Williamson]

On Tue, 17 May 2022 10:34:41 +0800
Wan Jiabing <wanjiabing@vivo.com> wrote:

> Fix following coccicheck warning:
> ./virt/kvm/vfio.c:258:1-7: preceding lock on line 236
> 
> If kvm_vfio_file_iommu_group() failed, code would goto err_fdput with
> mutex_lock acquired and then return ret. It might cause potential
> deadlock. Move mutex_unlock bellow err_fdput tag to fix it. 
> 
> Fixes: d55d9e7a45721 ("kvm/vfio: Store the struct file in the kvm_vfio_group")
> Signed-off-by: Wan Jiabing <wanjiabing@vivo.com>
> ---
>  virt/kvm/vfio.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/virt/kvm/vfio.c b/virt/kvm/vfio.c
> index 8f9f7fffb96a..ce1b01d02c51 100644
> --- a/virt/kvm/vfio.c
> +++ b/virt/kvm/vfio.c
> @@ -252,8 +252,8 @@ static int kvm_vfio_group_set_spapr_tce(struct kvm_device *dev,
>  		break;
>  	}
>  
> -	mutex_unlock(&kv->lock);
>  err_fdput:
> +	mutex_unlock(&kv->lock);
>  	fdput(f);
>  	return ret;
>  }

Applied to vfio next branch for v5.19.  Thanks!

Alex