> -----Original Message-----
> From: Michael S. Tsirkin [mailto:m...@redhat.com]
> Sent: Friday, December 1, 2023 10:51 PM
> To: Gonglei (Arei) <arei.gong...@huawei.com>
> Cc: Herbert Xu <herb...@gondor.apana.org.au>; linux-crypto@vger.kernel.org;
> Halil Pasic <pa...@linux.ibm.com>; Jason Wang <jasow...@redhat.com>;
> virtualizat...@lists.linux-foundation.org; linux-ker...@vger.kernel.org;
> wangyangxin <wangyangx...@huawei.com>
> Subject: Re: [PATCH] crypto: virtio-crypto: Handle dataq logic with tasklet
>
> On Mon, Nov 20, 2023 at 11:49:45AM +0000, Gonglei (Arei) wrote:
> > Doing ipsec produces a spinlock recursion warning.
> > This is due to crypto_finalize_request() being called in the upper half.
> > Move virtual data queue processing of virtio-crypto driver to tasklet.
> >
> > Fixes: dbaf0624ffa57 ("crypto: add virtio-crypto driver")
> > Reported-by: Halil Pasic <pa...@linux.ibm.com>
> > Signed-off-by: wangyangxin <wangyangx...@huawei.com>
> > Signed-off-by: Gonglei <arei.gong...@huawei.com>
> > ---
> > drivers/crypto/virtio/virtio_crypto_common.h | 2 ++
> > drivers/crypto/virtio/virtio_crypto_core.c | 23 +++++++++++++----------
> > 2 files changed, 15 insertions(+), 10 deletions(-)
> >
> > diff --git a/drivers/crypto/virtio/virtio_crypto_common.h
> > b/drivers/crypto/virtio/virtio_crypto_common.h
> > index 59a4c02..5c17c6e 100644
> > --- a/drivers/crypto/virtio/virtio_crypto_common.h
> > +++ b/drivers/crypto/virtio/virtio_crypto_common.h
> > @@ -10,6 +10,7 @@
> > #include <linux/virtio.h>
> > #include <linux/crypto.h>
> > #include <linux/spinlock.h>
> > +#include <linux/interrupt.h>
> > #include <crypto/aead.h>
> > #include <crypto/aes.h>
> > #include <crypto/engine.h>
> > @@ -28,6 +29,7 @@ struct data_queue {
> > char name[32];
> >
> > struct crypto_engine *engine;
> > + struct tasklet_struct done_task;
> > };
> >
> > struct virtio_crypto {
> > diff --git a/drivers/crypto/virtio/virtio_crypto_core.c
> > b/drivers/crypto/virtio/virtio_crypto_core.c
> > index 1198bd3..e747f4f 100644
> > --- a/drivers/crypto/virtio/virtio_crypto_core.c
> > +++ b/drivers/crypto/virtio/virtio_crypto_core.c
> > @@ -72,27 +72,28 @@ int virtio_crypto_ctrl_vq_request(struct virtio_crypto
> *vcrypto, struct scatterl
> > return 0;
> > }
> >
> > -static void virtcrypto_dataq_callback(struct virtqueue *vq)
> > +static void virtcrypto_done_task(unsigned long data)
> > {
> > - struct virtio_crypto *vcrypto = vq->vdev->priv;
> > + 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;
> > - unsigned int qid = vq->index;
> >
> > - spin_lock_irqsave(&vcrypto->data_vq[qid].lock, flags);
> > do {
> > virtqueue_disable_cb(vq);
> > while ((vc_req = virtqueue_get_buf(vq, &len)) != NULL) {
> > - spin_unlock_irqrestore(
> > - &vcrypto->data_vq[qid].lock, flags);
> > if (vc_req->alg_cb)
> > vc_req->alg_cb(vc_req, len);
> > - spin_lock_irqsave(
> > - &vcrypto->data_vq[qid].lock, flags);
> > }
> > } while (!virtqueue_enable_cb(vq));
> > - spin_unlock_irqrestore(&vcrypto->data_vq[qid].lock, flags);
> > +}
> > +
> > +static void virtcrypto_dataq_callback(struct virtqueue *vq) {
> > + struct virtio_crypto *vcrypto = vq->vdev->priv;
> > + struct data_queue *dq = &vcrypto->data_vq[vq->index];
> > +
> > + tasklet_schedule(&dq->done_task);
> > }
> >
>
> Don't we then need to wait for tasklet to complete on device remove?
>
>
Yes, we should do it. V2 will be sent soon.
Regards,
-Gonglei
> > static int virtcrypto_find_vqs(struct virtio_crypto *vi) @@ -150,6
> > +151,8 @@ static int virtcrypto_find_vqs(struct virtio_crypto *vi)
> > ret = -ENOMEM;
> > goto err_engine;
> > }
> > + tasklet_init(&vi->data_vq[i].done_task, virtcrypto_done_task,
> > + (unsigned long)&vi->data_vq[i]);
> > }
> >
> > kfree(names);
> > --
> > 1.8.3.1
> >
> >
>
