In the virtio blk vDPA live migration use case, for the target VM,
before the live migration process, QEMU will set call fd to all
queues of vDPA back-end. QEMU and vDPA back-end stand by
until live migration starts. During live migration process,
QEMU sets kick fd and new call fd. However, after the kick fd
is set to the vDPA back-end, the vDPA back-end configures device
and data path starts. The new call fd will cause some kind of
"re-configuration", this kind of "re-configuration" cause IO drop.
After this patch, vDPA back-end configures device after kick fd
and call fd are well set and make sure no IO drops.
This patch only impact virtio blk vDPA device and does not impact
net device.
Signed-off-by: Andy Pei <andy....@intel.com>
Signed-off-by: Huang Wei <wei.hu...@intel.com>
---
lib/vhost/vhost_user.c | 13 ++++++++++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/lib/vhost/vhost_user.c b/lib/vhost/vhost_user.c
index 9169cf5..14ff266 100644
--- a/lib/vhost/vhost_user.c
+++ b/lib/vhost/vhost_user.c
@@ -2983,6 +2983,7 @@ static int is_vring_iotlb(struct virtio_net *dev,
uint32_t vdpa_type = 0;
uint32_t request;
uint32_t i;
+ uint16_t blk_call_fd;
dev = get_device(vid);
if (dev == NULL)
@@ -3210,9 +3211,15 @@ static int is_vring_iotlb(struct virtio_net *dev,
if (!vdpa_dev)
goto out;
- if (vdpa_type == RTE_VHOST_VDPA_DEVICE_TYPE_BLK
- && request != VHOST_USER_SET_VRING_CALL)
- goto out;
+ if (vdpa_type == RTE_VHOST_VDPA_DEVICE_TYPE_BLK) {
+ if (request == VHOST_USER_SET_VRING_CALL) {
+ blk_call_fd = ctx.msg.payload.u64 &
VHOST_USER_VRING_IDX_MASK;
+ if (blk_call_fd != dev->nr_vring - 1)
+ goto out;
+ } else {
+ goto out;
+ }
+ }
if (!(dev->flags & VIRTIO_DEV_VDPA_CONFIGURED)) {
if (vdpa_dev->ops->dev_conf(dev->vid))
--
1.8.3.1
