Thread (1 message) 1 message, 1 author, 2021-12-09

Re: [PATCH v1 6/7] vdpa: Add support for querying control virtqueue index

From: Jason Wang <jasowang@redhat.com>
Date: 2021-12-09 05:45:18

Possibly related (same subject, not in this thread)

On Thu, Dec 9, 2021 at 4:15 AM Eli Cohen [off-list ref] wrote:
Add netlink attribute and callback function to query the control VQ
index of a device.

Example:

$ vdpa dev config show vdpa-a
  vdpa-a: mac 00:00:00:00:88:88 link up link_announce false max_vq_pairs 5 \
  mtu 9000 ctrl_vq_idx 10

I still wonder about the motivation for this. And as discussed, the
ctrl_vq_idx varies depending on whether MQ is negotiated.

Thanks
quoted hunk
Signed-off-by: Eli Cohen <redacted>
---
v0 -> v1:
1. Use logic defined in the spec to deduce virtqueue index.

 drivers/vdpa/vdpa.c       | 25 +++++++++++++++++++++++++
 include/uapi/linux/vdpa.h |  1 +
 2 files changed, 26 insertions(+)
diff --git a/drivers/vdpa/vdpa.c b/drivers/vdpa/vdpa.c
index 3bf016e03512..b4d4b8a7ca4e 100644
--- a/drivers/vdpa/vdpa.c
+++ b/drivers/vdpa/vdpa.c
@@ -712,6 +712,27 @@ static int vdpa_nl_cmd_dev_get_dumpit(struct sk_buff *msg, struct netlink_callba
        return msg->len;
 }

+static int vdpa_dev_net_ctrl_vq_fill(struct vdpa_device *vdev,
+                                    struct sk_buff *msg,
+                                    struct virtio_net_config *config,
+                                    u64 features)
+{
+       u16 N;
+
+       /* control VQ index, if available, is deduced according to the logic
+        * described in the virtio spec in section 5.1.2
+        */
+       if (!(features & BIT_ULL(VIRTIO_NET_F_CTRL_VQ)))
+               return 0;
+
+       if (features & BIT_ULL(VIRTIO_NET_F_MQ))
+               N = le16_to_cpu(config->max_virtqueue_pairs);
+       else
+               N = 1;
+
+       return nla_put_u16(msg, VDPA_ATTR_DEV_CTRL_VQ_IDX, 2 * N);
+}
+
 static int vdpa_dev_net_mq_config_fill(struct vdpa_device *vdev,
                                       struct sk_buff *msg, u64 features,
                                       const struct virtio_net_config *config)
@@ -730,6 +751,7 @@ static int vdpa_dev_net_config_fill(struct vdpa_device *vdev, struct sk_buff *ms
        struct virtio_net_config config = {};
        u64 features;
        u16 val_u16;
+       int err;

        vdpa_get_config(vdev, 0, &config, sizeof(config));
@@ -746,6 +768,9 @@ static int vdpa_dev_net_config_fill(struct vdpa_device *vdev, struct sk_buff *ms
                return -EMSGSIZE;

        features = vdev->config->get_driver_features(vdev);
+       err = vdpa_dev_net_ctrl_vq_fill(vdev, msg, &config, features);
+       if (err)
+               return err;

        return vdpa_dev_net_mq_config_fill(vdev, msg, features, &config);
 }
diff --git a/include/uapi/linux/vdpa.h b/include/uapi/linux/vdpa.h
index a252f06f9dfd..2e3a7f89f42d 100644
--- a/include/uapi/linux/vdpa.h
+++ b/include/uapi/linux/vdpa.h
@@ -34,6 +34,7 @@ enum vdpa_attr {
        VDPA_ATTR_DEV_MAX_VQS,                  /* u32 */
        VDPA_ATTR_DEV_MAX_VQ_SIZE,              /* u16 */
        VDPA_ATTR_DEV_MIN_VQ_SIZE,              /* u16 */
+       VDPA_ATTR_DEV_CTRL_VQ_IDX,              /* u16 */

        VDPA_ATTR_DEV_NET_CFG_MACADDR,          /* binary */
        VDPA_ATTR_DEV_NET_STATUS,               /* u8 */
--
2.33.1
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help