Skip to contentSkip to conversation

Vhost-vdpa Shadow Virtqueue _F_CTRL_RX_EXTRA commands support

RFC PATCH Superseded Hawkins Jiawei 1 participant 3 messages Jul 3, 2023 Networking

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover Vhost-vdpa Shadow Virtqueue _F_CTRL_RX_EXTRA commands support

This series enables shadowed CVQ to intercept rx commands related to VIRTIO_NET_F_CTRL_RX_EXTRA feature through shadowed CVQ, update the virtio NIC device model so qemu send it in a migration, and the restore of that rx state in the destination.

To test this patch series, one should modify the n->parent_obj.guest_features value in vhost_vdpa_net_load_rx() using gdb, as the linux virtio-net driver does not currently support the VIRTIO_NET_F_CTRL_RX_EXTRA feature.

Note that this patch should be based on [1], which has not been merged yet. I will submit the new viersion patch after it is merged.

[1]. https://lists.nongnu.org/archive/html/qemu-devel/2023-06/msg06559.html

Changelog
=========
v2:
- avoid sending CVQ command in default state suggested by Eugenio

v1: https://lists.nongnu.org/archive/html/qemu-devel/2023-06/msg04956.html

Hawkins Jiawei (2):
vdpa: Restore packet receive filtering state relative with
_F_CTRL_RX_EXTRA feature
vdpa: Allow VIRTIO_NET_F_CTRL_RX_EXTRA in SVQ

 net/vhost-vdpa.c | 83 ++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 83 insertions(+)
Signature · 1 line
2.25.1
#1 lore raw

RFC PATCH v2 1/2 vdpa: Restore packet receive filtering state relative with _F_CTRL_RX_EXTRA feature

This patch refactors vhost_vdpa_net_load_rx() to restore the packet receive filtering state in relation to VIRTIO_NET_F_CTRL_RX_EXTRA feature at device's startup.

---
v2:
- avoid sending CVQ command in default state suggested by Eugenio

v1: https://lists.nongnu.org/archive/html/qemu-devel/2023-06/msg04957.html

 net/vhost-vdpa.c | 82 ++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 82 insertions(+)
net/vhost-vdpa.c +82 -0
index 0410a52043..4919e18208 100644
@@ -890,6 +890,88 @@ static int vhost_vdpa_net_load_rx(VhostVDPAState *s,
         }
     }
 
+    if (virtio_vdev_has_feature(&n->parent_obj, VIRTIO_NET_F_CTRL_RX_EXTRA)) {
+        /* Load the all-unicast mode */
+        on = n->alluni;
+        if (on != 0) {
+            /*
+             * According to virtio_net_reset(), device turns all-unicast mode
+             * off by default.
+             *
+             * Therefore, there is no need to send this CVQ command if the
+             * driver also sets all-unicast mode off, which aligns with
+             * the device's defaults.
+             *
+             * Note that the device's defaults can mismatch the driver's
+             * configuration only at live migration.
+             */
+            r = vhost_vdpa_net_load_rx_mode(s, VIRTIO_NET_CTRL_RX_ALLUNI, on);
+            if (r < 0) {
+                return r;
+            }
+        }
+
+        /* Load the non-multicast mode */
+        on = n->nomulti;
+        if (on != 0) {
+            /*
+             * According to virtio_net_reset(), device turns non-multicast mode
+             * off by default.
+             *
+             * Therefore, there is no need to send this CVQ command if the
+             * driver also sets non-multicast mode off, which aligns with
+             * the device's defaults.
+             *
+             * Note that the device's defaults can mismatch the driver's
+             * configuration only at live migration.
+             */
+            r = vhost_vdpa_net_load_rx_mode(s, VIRTIO_NET_CTRL_RX_NOMULTI, on);
+            if (r < 0) {
+                return r;
+            }
+        }
+
+        /* Load the non-unicast mode */
+        on = n->nouni;
+        if (on != 0) {
+            /*
+             * According to virtio_net_reset(), device turns non-unicast mode
+             * off by default.
+             *
+             * Therefore, there is no need to send this CVQ command if the
+             * driver also sets non-unicast mode off, which aligns with
+             * the device's defaults.
+             *
+             * Note that the device's defaults can mismatch the driver's
+             * configuration only at live migration.
+             */
+            r = vhost_vdpa_net_load_rx_mode(s, VIRTIO_NET_CTRL_RX_NOUNI, on);
+            if (r < 0) {
+                return r;
+            }
+        }
+
+        /* Load the non-broadcast mode */
+        on = n->nobcast;
+        if (on != 0) {
+            /*
+             * According to virtio_net_reset(), device turns non-broadcast mode
+             * off by default.
+             *
+             * Therefore, there is no need to send this CVQ command if the
+             * driver also sets non-broadcast mode off, which aligns with
+             * the device's defaults.
+             *
+             * Note that the device's defaults can mismatch the driver's
+             * configuration only at live migration.
+             */
+            r = vhost_vdpa_net_load_rx_mode(s, VIRTIO_NET_CTRL_RX_NOBCAST, on);
+            if (r < 0) {
+                return r;
+            }
+        }
+    }
+
     return 0;
 }
 
Signature · 1 line
2.25.1
#2 lore raw

RFC PATCH v2 2/2 vdpa: Allow VIRTIO_NET_F_CTRL_RX_EXTRA in SVQ

In reply to Hawkins Jiawei

Enable SVQ with VIRTIO_NET_F_CTRL_RX_EXTRA feature.

---
 net/vhost-vdpa.c | 1 +
 1 file changed, 1 insertion(+)
net/vhost-vdpa.c +1 -0
index 4919e18208..e10a0128c3 100644
@@ -105,6 +105,7 @@ static const uint64_t vdpa_svq_device_features =
     BIT_ULL(VIRTIO_NET_F_STATUS) |
     BIT_ULL(VIRTIO_NET_F_CTRL_VQ) |
     BIT_ULL(VIRTIO_NET_F_CTRL_RX) |
+    BIT_ULL(VIRTIO_NET_F_CTRL_RX_EXTRA) |
     BIT_ULL(VIRTIO_NET_F_MQ) |
     BIT_ULL(VIRTIO_F_ANY_LAYOUT) |
     BIT_ULL(VIRTIO_NET_F_CTRL_MAC_ADDR) |
Signature · 1 line
2.25.1