Re: [PATCH v2 1/9] crypto: virtio: Add spinlock protection with virtqueue notification

[Bibo Mao]

On 2025/12/5 上午9:21, Jason Wang wrote:
On Thu, Dec 4, 2025 at 7:22 PM Bibo Mao <maob...@loongson.cn> wrote:

When VM boots with one virtio-crypto PCI device and builtin backend,
run openssl benchmark command with multiple processes, such as
   openssl speed -evp aes-128-cbc -engine afalg  -seconds 10 -multi 32

openssl processes will hangup and there is error reported like this:
  virtio_crypto virtio0: dataq.0:id 3 is not a head!

It seems that the data virtqueue need protection when it is handled
for virtio done notification. If the spinlock protection is added
in virtcrypto_done_task(), openssl benchmark with multiple processes
works well.

Fixes: fed93fb62e05 ("crypto: virtio - Handle dataq logic with tasklet")
Cc: sta...@vger.kernel.org
Signed-off-by: Bibo Mao <maob...@loongson.cn>
---
  drivers/crypto/virtio/virtio_crypto_core.c | 5 +++++
  1 file changed, 5 insertions(+)

diff --git a/drivers/crypto/virtio/virtio_crypto_core.c 
b/drivers/crypto/virtio/virtio_crypto_core.c
index 3d241446099c..ccc6b5c1b24b 100644
--- a/drivers/crypto/virtio/virtio_crypto_core.c
+++ b/drivers/crypto/virtio/virtio_crypto_core.c
@@ -75,15 +75,20 @@ static void virtcrypto_done_task(unsigned long data)
         struct data_queue *data_vq = (struct data_queue *)data;
         struct virtqueue *vq = data_vq->vq;
         struct virtio_crypto_request *vc_req;
+       unsigned long flags;
         unsigned int len;

+       spin_lock_irqsave(&data_vq->lock, flags);
         do {
                 virtqueue_disable_cb(vq);
                 while ((vc_req = virtqueue_get_buf(vq, &len)) != NULL) {
+                       spin_unlock_irqrestore(&data_vq->lock, flags);
                         if (vc_req->alg_cb)
                                 vc_req->alg_cb(vc_req, len);
+                       spin_lock_irqsave(&data_vq->lock, flags);
                 }
         } while (!virtqueue_enable_cb(vq));
+       spin_unlock_irqrestore(&data_vq->lock, flags);
  }

Another thing that needs to care:

There seems to be a redundant virtqueue_kick() in
virtio_crypto_skcipher_crypt_req() which is out of the protection of
the spinlock.

I think we can simply remote that?
yes, there is redundant virtqueue_kick() in function 
virtio_crypto_skcipher_crypt_req().

Will remove one in next version.

Regards
Bibo Mao

Thanks


  static void virtcrypto_dataq_callback(struct virtqueue *vq)
--
2.39.3