On Wed, Jun 10, 2020 at 5:08 PM Michael S. Tsirkin [off-list ref] wrote:

On Wed, Jun 10, 2020 at 04:29:29PM +0200, Eugenio Pérez wrote:

quoted

On Wed, 2020-06-10 at 07:36 -0400, Michael S. Tsirkin wrote:

quoted

As testing shows no performance change, switch to that now.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Link: https://lore.kernel.org/r/20200401183118.8334-3-eperezma@redhat.com (local)
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
 drivers/vhost/test.c  |   2 +-
 drivers/vhost/vhost.c | 318 ++++++++----------------------------------
 drivers/vhost/vhost.h |   7 +-
 3 files changed, 65 insertions(+), 262 deletions(-)

diff --git a/drivers/vhost/test.c b/drivers/vhost/test.c
index 0466921f4772..7d69778aaa26 100644
--- a/drivers/vhost/test.c
+++ b/drivers/vhost/test.c

@@ -119,7 +119,7 @@ static int vhost_test_open(struct inode *inode, struct file *f)
    dev = &n->dev;
    vqs[VHOST_TEST_VQ] = &n->vqs[VHOST_TEST_VQ];
    n->vqs[VHOST_TEST_VQ].handle_kick = handle_vq_kick;
-   vhost_dev_init(dev, vqs, VHOST_TEST_VQ_MAX, UIO_MAXIOV,
+   vhost_dev_init(dev, vqs, VHOST_TEST_VQ_MAX, UIO_MAXIOV + 64,
                   VHOST_TEST_PKT_WEIGHT, VHOST_TEST_WEIGHT, true, NULL);

    f->private_data = n;

diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c
index 11433d709651..28f324fd77df 100644
--- a/drivers/vhost/vhost.c
+++ b/drivers/vhost/vhost.c

@@ -304,6 +304,7 @@ static void vhost_vq_reset(struct vhost_dev *dev,
 {
    vq->num = 1;
    vq->ndescs = 0;
+   vq->first_desc = 0;
    vq->desc = NULL;
    vq->avail = NULL;
    vq->used = NULL;

@@ -372,6 +373,11 @@ static int vhost_worker(void *data)
    return 0;
 }

+static int vhost_vq_num_batch_descs(struct vhost_virtqueue *vq)
+{
+   return vq->max_descs - UIO_MAXIOV;
+}
+
 static void vhost_vq_free_iovecs(struct vhost_virtqueue *vq)
 {
    kfree(vq->descs);

@@ -394,6 +400,9 @@ static long vhost_dev_alloc_iovecs(struct vhost_dev *dev)
    for (i = 0; i < dev->nvqs; ++i) {
            vq = dev->vqs[i];
            vq->max_descs = dev->iov_limit;
+           if (vhost_vq_num_batch_descs(vq) < 0) {
+                   return -EINVAL;
+           }
            vq->descs = kmalloc_array(vq->max_descs,
                                      sizeof(*vq->descs),
                                      GFP_KERNEL);

@@ -1610,6 +1619,7 @@ long vhost_vring_ioctl(struct vhost_dev *d, unsigned int ioctl, void __user *arg
            vq->last_avail_idx = s.num;
            /* Forget the cached index value. */
            vq->avail_idx = vq->last_avail_idx;
+           vq->ndescs = vq->first_desc = 0;

This is not needed if it is done in vhost_vq_set_backend, as far as I can tell.

Actually, maybe it is even better to move `vq->avail_idx = vq->last_avail_idx;` line to vhost_vq_set_backend, it is part
of the backend "set up" procedure, isn't it?

I tested with virtio_test + batch tests sent in
https://lkml.kernel.org/lkml/20200418102217.32327-1-eperezma@redhat.com/T/.

Ow did I forget to merge them for rc1?  Should I have? Maybe Linus won't
yell to hard at me if I merge them after rc1.

quoted

I append here what I'm proposing in case it is clearer this way.

Thanks!

diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c
index 4d198994e7be..809ad2cd2879 100644
--- a/drivers/vhost/vhost.c
+++ b/drivers/vhost/vhost.c

@@ -1617,9 +1617,6 @@ long vhost_vring_ioctl(struct vhost_dev *d, unsigned int ioctl, void __user *arg
                      break;
              }
              vq->last_avail_idx = s.num;
-             /* Forget the cached index value. */
-             vq->avail_idx = vq->last_avail_idx;
-             vq->ndescs = vq->first_desc = 0;
              break;
      case VHOST_GET_VRING_BASE:
              s.index = idx;

diff --git a/drivers/vhost/vhost.h b/drivers/vhost/vhost.h
index fed36af5c444..f4902dc808e4 100644
--- a/drivers/vhost/vhost.h
+++ b/drivers/vhost/vhost.h

@@ -258,6 +258,7 @@ static inline void vhost_vq_set_backend(struct vhost_virtqueue *vq,
                                      void *private_data)
 {
      vq->private_data = private_data;
+     vq->avail_idx = vq->last_avail_idx;
      vq->ndescs = 0;
      vq->first_desc = 0;
 }

Seems like a nice cleanup, though it's harmless right?

Fields ndescs and first_descs are supposed to be updated outside, that
was the intention but maybe I forgot to delete it here, not sure.

Regarding avail_idx, the whole change has been tested with vhost_test
and for vhost needs to have a backend to modify it already so it seems
safe to me.