Virtualization - Feb 2014 - [PATCH net v2] vhost: fix ref cnt checking deadlock

vhost checked the counter within the refcnt before decrementing.  It
really wanted to know that it is the one that has the last reference, as
a way to batch freeing resources a bit more efficiently.

Note: we only let refcount go to 0 on device release.

This works well but we now access the ref counter twice so there's a
race: all users might see a high count and decide to defer freeing
resources.
In the end no one initiates freeing resources until the last reference
is gone (which is on VM shotdown so might happen after a looooong time).

Let's do what we probably should have done straight away:
switch from kref to plain atomic, documenting the
semantics, return the refcount value atomically after decrement,
then use that to avoid the deadlock.

Reported-by: Qin Chuanyu <qinchuanyu at huawei.com>
Signed-off-by: Michael S. Tsirkin <mst at redhat.com>
---

This patch is needed for 3.14 and -stable.

 drivers/vhost/net.c | 41 ++++++++++++++++++++---------------------
 1 file changed, 20 insertions(+), 21 deletions(-)

diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
index 831eb4f..b12176f 100644
--- a/drivers/vhost/net.c
+++ b/drivers/vhost/net.c
@@ -70,7 +70,12 @@ enum {
 };
 
 struct vhost_net_ubuf_ref {
-	struct kref kref;
+	/* refcount follows semantics similar to kref:
+	 *  0: object is released
+	 *  1: no outstanding ubufs
+	 * >1: outstanding ubufs
+	 */
+	atomic_t refcount;
 	wait_queue_head_t wait;
 	struct vhost_virtqueue *vq;
 };
@@ -116,14 +121,6 @@ static void vhost_net_enable_zcopy(int vq)
 	vhost_net_zcopy_mask |= 0x1 << vq;
 }
 
-static void vhost_net_zerocopy_done_signal(struct kref *kref)
-{
-	struct vhost_net_ubuf_ref *ubufs;
-
-	ubufs = container_of(kref, struct vhost_net_ubuf_ref, kref);
-	wake_up(&ubufs->wait);
-}
-
 static struct vhost_net_ubuf_ref *
 vhost_net_ubuf_alloc(struct vhost_virtqueue *vq, bool zcopy)
 {
@@ -134,21 +131,24 @@ vhost_net_ubuf_alloc(struct vhost_virtqueue *vq, bool
zcopy)
 	ubufs = kmalloc(sizeof(*ubufs), GFP_KERNEL);
 	if (!ubufs)
 		return ERR_PTR(-ENOMEM);
-	kref_init(&ubufs->kref);
+	atomic_set(&ubufs->refcount, 1);
 	init_waitqueue_head(&ubufs->wait);
 	ubufs->vq = vq;
 	return ubufs;
 }
 
-static void vhost_net_ubuf_put(struct vhost_net_ubuf_ref *ubufs)
+static int vhost_net_ubuf_put(struct vhost_net_ubuf_ref *ubufs)
 {
-	kref_put(&ubufs->kref, vhost_net_zerocopy_done_signal);
+	int r = atomic_sub_return(1, &ubufs->refcount);
+	if (unlikely(!r))
+		wake_up(&ubufs->wait);
+	return r;
 }
 
 static void vhost_net_ubuf_put_and_wait(struct vhost_net_ubuf_ref *ubufs)
 {
-	kref_put(&ubufs->kref, vhost_net_zerocopy_done_signal);
-	wait_event(ubufs->wait, !atomic_read(&ubufs->kref.refcount));
+	vhost_net_ubuf_put(ubufs);
+	wait_event(ubufs->wait, !atomic_read(&ubufs->refcount));
 }
 
 static void vhost_net_ubuf_put_wait_and_free(struct vhost_net_ubuf_ref *ubufs)
@@ -306,22 +306,21 @@ static void vhost_zerocopy_callback(struct ubuf_info
*ubuf, bool success)
 {
 	struct vhost_net_ubuf_ref *ubufs = ubuf->ctx;
 	struct vhost_virtqueue *vq = ubufs->vq;
-	int cnt = atomic_read(&ubufs->kref.refcount);
+	int cnt;
 
 	/* set len to mark this desc buffers done DMA */
 	vq->heads[ubuf->desc].len = success ?
 		VHOST_DMA_DONE_LEN : VHOST_DMA_FAILED_LEN;
-	vhost_net_ubuf_put(ubufs);
+	cnt = vhost_net_ubuf_put(ubufs);
 
 	/*
 	 * Trigger polling thread if guest stopped submitting new buffers:
-	 * in this case, the refcount after decrement will eventually reach 1
-	 * so here it is 2.
+	 * in this case, the refcount after decrement will eventually reach 1.
 	 * We also trigger polling periodically after each 16 packets
 	 * (the value 16 here is more or less arbitrary, it's tuned to trigger
 	 * less than 10% of times).
 	 */
-	if (cnt <= 2 || !(cnt % 16))
+	if (cnt <= 1 || !(cnt % 16))
 		vhost_poll_queue(&vq->poll);
 }
 
@@ -420,7 +419,7 @@ static void handle_tx(struct vhost_net *net)
 			msg.msg_control = ubuf;
 			msg.msg_controllen = sizeof(ubuf);
 			ubufs = nvq->ubufs;
-			kref_get(&ubufs->kref);
+			atomic_inc(&ubufs->refcount);
 			nvq->upend_idx = (nvq->upend_idx + 1) % UIO_MAXIOV;
 		} else {
 			msg.msg_control = NULL;
@@ -785,7 +784,7 @@ static void vhost_net_flush(struct vhost_net *n)
 		vhost_net_ubuf_put_and_wait(n->vqs[VHOST_NET_VQ_TX].ubufs);
 		mutex_lock(&n->vqs[VHOST_NET_VQ_TX].vq.mutex);
 		n->tx_flush = false;
-		kref_init(&n->vqs[VHOST_NET_VQ_TX].ubufs->kref);
+		atomic_set(&n->vqs[VHOST_NET_VQ_TX].ubufs->refcount, 1);
 		mutex_unlock(&n->vqs[VHOST_NET_VQ_TX].vq.mutex);
 	}
 }
-- 
MST

On 02/13/2014 05:42 PM, Michael S. Tsirkin wrote:
> vhost checked the counter within the refcnt before decrementing.  It
> really wanted to know that it is the one that has the last reference, as
> a way to batch freeing resources a bit more efficiently.
>
> Note: we only let refcount go to 0 on device release.
>
> This works well but we now access the ref counter twice so there's a
> race: all users might see a high count and decide to defer freeing
> resources.
> In the end no one initiates freeing resources until the last reference
> is gone (which is on VM shotdown so might happen after a looooong time).
>
> Let's do what we probably should have done straight away:
> switch from kref to plain atomic, documenting the
> semantics, return the refcount value atomically after decrement,
> then use that to avoid the deadlock.
>
> Reported-by: Qin Chuanyu <qinchuanyu at huawei.com>
> Signed-off-by: Michael S. Tsirkin <mst at redhat.com>
> ---
>
> This patch is needed for 3.14 and -stable.
>
>  drivers/vhost/net.c | 41 ++++++++++++++++++++---------------------
>  1 file changed, 20 insertions(+), 21 deletions(-)
>
> diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c
> index 831eb4f..b12176f 100644
> --- a/drivers/vhost/net.c
> +++ b/drivers/vhost/net.c
> @@ -70,7 +70,12 @@ enum {
>  };
>  
>  struct vhost_net_ubuf_ref {
> -	struct kref kref;
> +	/* refcount follows semantics similar to kref:
> +	 *  0: object is released
> +	 *  1: no outstanding ubufs
> +	 * >1: outstanding ubufs
> +	 */
> +	atomic_t refcount;
>  	wait_queue_head_t wait;
>  	struct vhost_virtqueue *vq;
>  };
> @@ -116,14 +121,6 @@ static void vhost_net_enable_zcopy(int vq)
>  	vhost_net_zcopy_mask |= 0x1 << vq;
>  }
>  
> -static void vhost_net_zerocopy_done_signal(struct kref *kref)
> -{
> -	struct vhost_net_ubuf_ref *ubufs;
> -
> -	ubufs = container_of(kref, struct vhost_net_ubuf_ref, kref);
> -	wake_up(&ubufs->wait);
> -}
> -
>  static struct vhost_net_ubuf_ref *
>  vhost_net_ubuf_alloc(struct vhost_virtqueue *vq, bool zcopy)
>  {
> @@ -134,21 +131,24 @@ vhost_net_ubuf_alloc(struct vhost_virtqueue *vq, bool
zcopy)
>  	ubufs = kmalloc(sizeof(*ubufs), GFP_KERNEL);
>  	if (!ubufs)
>  		return ERR_PTR(-ENOMEM);
> -	kref_init(&ubufs->kref);
> +	atomic_set(&ubufs->refcount, 1);
>  	init_waitqueue_head(&ubufs->wait);
>  	ubufs->vq = vq;
>  	return ubufs;
>  }
>  
> -static void vhost_net_ubuf_put(struct vhost_net_ubuf_ref *ubufs)
> +static int vhost_net_ubuf_put(struct vhost_net_ubuf_ref *ubufs)
>  {
> -	kref_put(&ubufs->kref, vhost_net_zerocopy_done_signal);
> +	int r = atomic_sub_return(1, &ubufs->refcount);
> +	if (unlikely(!r))
> +		wake_up(&ubufs->wait);
> +	return r;
>  }
>  
>  static void vhost_net_ubuf_put_and_wait(struct vhost_net_ubuf_ref *ubufs)
>  {
> -	kref_put(&ubufs->kref, vhost_net_zerocopy_done_signal);
> -	wait_event(ubufs->wait, !atomic_read(&ubufs->kref.refcount));
> +	vhost_net_ubuf_put(ubufs);
> +	wait_event(ubufs->wait, !atomic_read(&ubufs->refcount));
>  }
>  
>  static void vhost_net_ubuf_put_wait_and_free(struct vhost_net_ubuf_ref
*ubufs)
> @@ -306,22 +306,21 @@ static void vhost_zerocopy_callback(struct ubuf_info
*ubuf, bool success)
>  {
>  	struct vhost_net_ubuf_ref *ubufs = ubuf->ctx;
>  	struct vhost_virtqueue *vq = ubufs->vq;
> -	int cnt = atomic_read(&ubufs->kref.refcount);
> +	int cnt;
>  
>  	/* set len to mark this desc buffers done DMA */
>  	vq->heads[ubuf->desc].len = success ?
>  		VHOST_DMA_DONE_LEN : VHOST_DMA_FAILED_LEN;
> -	vhost_net_ubuf_put(ubufs);
> +	cnt = vhost_net_ubuf_put(ubufs);
>  
>  	/*
>  	 * Trigger polling thread if guest stopped submitting new buffers:
> -	 * in this case, the refcount after decrement will eventually reach 1
> -	 * so here it is 2.
> +	 * in this case, the refcount after decrement will eventually reach 1.
>  	 * We also trigger polling periodically after each 16 packets
>  	 * (the value 16 here is more or less arbitrary, it's tuned to
trigger
>  	 * less than 10% of times).
>  	 */
> -	if (cnt <= 2 || !(cnt % 16))
> +	if (cnt <= 1 || !(cnt % 16))
>  		vhost_poll_queue(&vq->poll);
>  }
>  
> @@ -420,7 +419,7 @@ static void handle_tx(struct vhost_net *net)
>  			msg.msg_control = ubuf;
>  			msg.msg_controllen = sizeof(ubuf);
>  			ubufs = nvq->ubufs;
> -			kref_get(&ubufs->kref);
> +			atomic_inc(&ubufs->refcount);
>  			nvq->upend_idx = (nvq->upend_idx + 1) % UIO_MAXIOV;
>  		} else {
>  			msg.msg_control = NULL;
> @@ -785,7 +784,7 @@ static void vhost_net_flush(struct vhost_net *n)
>  		vhost_net_ubuf_put_and_wait(n->vqs[VHOST_NET_VQ_TX].ubufs);
>  		mutex_lock(&n->vqs[VHOST_NET_VQ_TX].vq.mutex);
>  		n->tx_flush = false;
> -		kref_init(&n->vqs[VHOST_NET_VQ_TX].ubufs->kref);
> +		atomic_set(&n->vqs[VHOST_NET_VQ_TX].ubufs->refcount, 1);
>  		mutex_unlock(&n->vqs[VHOST_NET_VQ_TX].vq.mutex);
>  	}
>  }
Acked-by: Jason Wang <jasowang at redhat.com>

From: "Michael S. Tsirkin" <mst at redhat.com>
Date: Thu, 13 Feb 2014 11:42:05 +0200
> vhost checked the counter within the refcnt before decrementing.  It
> really wanted to know that it is the one that has the last reference, as
> a way to batch freeing resources a bit more efficiently.
> 
> Note: we only let refcount go to 0 on device release.
> 
> This works well but we now access the ref counter twice so there's a
> race: all users might see a high count and decide to defer freeing
> resources.
> In the end no one initiates freeing resources until the last reference
> is gone (which is on VM shotdown so might happen after a looooong time).
> 
> Let's do what we probably should have done straight away:
> switch from kref to plain atomic, documenting the
> semantics, return the refcount value atomically after decrement,
> then use that to avoid the deadlock.
> 
> Reported-by: Qin Chuanyu <qinchuanyu at huawei.com>
> Signed-off-by: Michael S. Tsirkin <mst at redhat.com>
> ---
> 
> This patch is needed for 3.14 and -stable.
Applied and queued up for -stable.