[v5 1/5] vhost: skip crypto op fetch before vring init

[v5 1/5] vhost: skip crypto op fetch before vring init

[Gowrishankar Muthukrishnan]

Until virtio avail ring is initialized (by VHOST_USER_SET_VRING_ADDR),
worker thread should not try to fetch crypto op, which would lead to
memory fault.

Fixes: 939066d96563 ("vhost/crypto: add public function implementation")
Cc: stable@dpdk.org

Signed-off-by: Gowrishankar Muthukrishnan <gmuthukrishn@marvell.com>
Acked-by: Akhil Goyal <gakhil@marvell.com>
---
 lib/vhost/vhost_crypto.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/lib/vhost/vhost_crypto.c b/lib/vhost/vhost_crypto.c
index 3dc41a3bd5..55ea24710e 100644
--- a/lib/vhost/vhost_crypto.c
+++ b/lib/vhost/vhost_crypto.c
@@ -1580,6 +1580,16 @@ rte_vhost_crypto_fetch_requests(int vid, uint32_t qid,
 
 	vq = dev->virtqueue[qid];
 
+	if (unlikely(vq == NULL)) {
+		VC_LOG_ERR("Invalid virtqueue %u", qid);
+		return 0;
+	}
+
+	if (unlikely(vq->avail == NULL)) {
+		VC_LOG_DBG("Virtqueue ring not yet initialized %u", qid);
+		return 0;
+	}
+
 	avail_idx = *((volatile uint16_t *)&vq->avail->idx);
 	start_idx = vq->last_used_idx;
 	count = avail_idx - start_idx;
-- 
2.25.1

Re: [v5 1/5] vhost: skip crypto op fetch before vring init

[Maxime Coquelin]

On 2/24/25 11:35 AM, Gowrishankar Muthukrishnan wrote:
> Until virtio avail ring is initialized (by VHOST_USER_SET_VRING_ADDR),
> worker thread should not try to fetch crypto op, which would lead to
> memory fault.
> 
> Fixes: 939066d96563 ("vhost/crypto: add public function implementation")
> Cc: stable@dpdk.org
> 
> Signed-off-by: Gowrishankar Muthukrishnan <gmuthukrishn@marvell.com>
> Acked-by: Akhil Goyal <gakhil@marvell.com>
> ---
>   lib/vhost/vhost_crypto.c | 10 ++++++++++
>   1 file changed, 10 insertions(+)
> 
> diff --git a/lib/vhost/vhost_crypto.c b/lib/vhost/vhost_crypto.c
> index 3dc41a3bd5..55ea24710e 100644
> --- a/lib/vhost/vhost_crypto.c
> +++ b/lib/vhost/vhost_crypto.c
> @@ -1580,6 +1580,16 @@ rte_vhost_crypto_fetch_requests(int vid, uint32_t qid,
>   
>   	vq = dev->virtqueue[qid];
>   
> +	if (unlikely(vq == NULL)) {
> +		VC_LOG_ERR("Invalid virtqueue %u", qid);
> +		return 0;
> +	}
> +
> +	if (unlikely(vq->avail == NULL)) {
> +		VC_LOG_DBG("Virtqueue ring not yet initialized %u", qid);
> +		return 0;
> +	}
> +

I think the proper fix to:
1. acquire the access_lock as reader
2. check whether vq is enabled
3. acquire the iotlb lock as a reader
4. check vq's access_ok

You can have a look at rte_vhost_dequeue_burst for example.

>   	avail_idx = *((volatile uint16_t *)&vq->avail->idx);
>   	start_idx = vq->last_used_idx;
>   	count = avail_idx - start_idx;

[v5 3/5] examples/vhost_crypto: fix user callbacks

[Gowrishankar Muthukrishnan]

In order to handle new vhost user connection, use new_connection
and destroy_connection callbacks.

Fixes: f5188211c721 ("examples/vhost_crypto: add sample application")
Cc: stable@dpdk.org

Signed-off-by: Gowrishankar Muthukrishnan <gmuthukrishn@marvell.com>
Acked-by: Akhil Goyal <gakhil@marvell.com>
---
 examples/vhost_crypto/main.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/examples/vhost_crypto/main.c b/examples/vhost_crypto/main.c
index 558c09a60f..b1fe4120b9 100644
--- a/examples/vhost_crypto/main.c
+++ b/examples/vhost_crypto/main.c
@@ -362,8 +362,8 @@ destroy_device(int vid)
 }
 
 static const struct rte_vhost_device_ops virtio_crypto_device_ops = {
-	.new_device =  new_device,
-	.destroy_device = destroy_device,
+	.new_connection =  new_device,
+	.destroy_connection = destroy_device,
 };
 
 static int
-- 
2.25.1