A few nits. Looks like there has been some churn here so this patch
will need to be rebased.

On Wed, Aug 13, 2025 at 1:01 PM Vladimir Sementsov-Ogievskiy
<[email protected]> wrote:
>
> Add logic to transfer virtio notifiers through migration channel
> for vhost backend migration case.
>
> Signed-off-by: Vladimir Sementsov-Ogievskiy <[email protected]>
> ---
>  hw/virtio/virtio-bus.c     |  2 +-
>  hw/virtio/virtio.c         | 74 ++++++++++++++++++++++++++++++++++++--
>  include/hw/virtio/virtio.h |  2 ++
>  3 files changed, 75 insertions(+), 3 deletions(-)
>
> diff --git a/hw/virtio/virtio-bus.c b/hw/virtio/virtio-bus.c
> index c7e3941b1e..c1848144a2 100644
> --- a/hw/virtio/virtio-bus.c
> +++ b/hw/virtio/virtio-bus.c
> @@ -286,7 +286,7 @@ int virtio_bus_set_host_notifier(VirtioBusState *bus, int 
> n, bool assign)
>          return -ENOSYS;
>      }
>
> -    if (assign) {
> +    if (assign && !virtio_is_vhost_migrating_backend(vdev)) {
>          r = event_notifier_init(notifier, 1);
>          if (r < 0) {
>              error_report("%s: unable to init event notifier: %s (%d)",
> diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
> index 10891f0e0c..87c243edad 100644
> --- a/hw/virtio/virtio.c
> +++ b/hw/virtio/virtio.c
> @@ -26,6 +26,7 @@
>  #include "hw/virtio/virtio.h"
>  #include "hw/virtio/vhost.h"
>  #include "migration/qemu-file-types.h"
> +#include "migration/qemu-file.h"
>  #include "qemu/atomic.h"
>  #include "hw/virtio/virtio-bus.h"
>  #include "hw/qdev-properties.h"
> @@ -2992,6 +2993,7 @@ int virtio_save(VirtIODevice *vdev, QEMUFile *f)
>      VirtioDeviceClass *vdc = VIRTIO_DEVICE_GET_CLASS(vdev);
>      uint32_t guest_features_lo = (vdev->guest_features & 0xffffffff);
>      int i;
> +    bool migrating_backend = virtio_is_vhost_migrating_backend(vdev);
>
>      if (k->save_config) {
>          k->save_config(qbus->parent, f);
> @@ -3025,11 +3027,23 @@ int virtio_save(VirtIODevice *vdev, QEMUFile *f)
>           */
>          qemu_put_be64(f, vdev->vq[i].vring.desc);
>          qemu_put_be16s(f, &vdev->vq[i].last_avail_idx);
> +
> +        if (migrating_backend) {
> +            qemu_file_put_fd(f,
> +                             
> event_notifier_get_fd(&vdev->vq[i].host_notifier));
> +            qemu_file_put_fd(
> +                f, event_notifier_get_fd(&vdev->vq[i].guest_notifier));
> +        }
> +
>          if (k->save_queue) {
>              k->save_queue(qbus->parent, i, f);
>          }
>      }
>
> +    if (migrating_backend) {
> +        qemu_file_put_fd(f, event_notifier_get_fd(&vdev->config_notifier));
> +    }
> +
>      if (vdc->save != NULL) {
>          vdc->save(vdev, f);
>      }
> @@ -3235,6 +3249,8 @@ virtio_load(VirtIODevice *vdev, QEMUFile *f, int 
> version_id)
>      BusState *qbus = qdev_get_parent_bus(DEVICE(vdev));
>      VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(qbus);
>      VirtioDeviceClass *vdc = VIRTIO_DEVICE_GET_CLASS(vdev);
> +    Error *local_err = NULL;
> +    bool migrating_backend = virtio_is_vhost_migrating_backend(vdev);
>
>      /*
>       * We poison the endianness to ensure it does not get used before
> @@ -3304,6 +3320,13 @@ virtio_load(VirtIODevice *vdev, QEMUFile *f, int 
> version_id)
>          vdev->vq[i].signalled_used_valid = false;
>          vdev->vq[i].notification = true;
>
> +        if (migrating_backend) {
> +            event_notifier_init_fd(&vdev->vq[i].host_notifier,
> +                                   qemu_file_get_fd(f));
> +            event_notifier_init_fd(&vdev->vq[i].guest_notifier,
> +                                   qemu_file_get_fd(f));
> +        }
> +
>          if (!vdev->vq[i].vring.desc && vdev->vq[i].last_avail_idx) {
>              error_report("VQ %d address 0x0 "
>                           "inconsistent with Host index 0x%x",
> @@ -3317,6 +3340,10 @@ virtio_load(VirtIODevice *vdev, QEMUFile *f, int 
> version_id)
>          }
>      }
>
> +    if (migrating_backend) {

nit: spurious spaces after &vdev->config_notifier

> +        event_notifier_init_fd(&vdev->config_notifier   , 
> qemu_file_get_fd(f));
> +    }
> +
>      virtio_notify_vector(vdev, VIRTIO_NO_VECTOR);
>
>      if (vdc->load != NULL) {
> @@ -3333,6 +3360,19 @@ virtio_load(VirtIODevice *vdev, QEMUFile *f, int 
> version_id)
>          }
>      }
>
> +    if (migrating_backend) {
> +        /*
> +         * On vhost backend migration, device do load host_features from
> +         * migration stream. So update host_features.
> +         */
> +        vdev->host_features = vdc->get_features(vdev, vdev->host_features,
> +                                                &local_err);
> +        if (local_err) {
> +            error_report_err(local_err);
> +            return -EINVAL;
> +        }
> +    }
> +
>      /* Subsections */
>      ret = vmstate_load_state(f, &vmstate_virtio, vdev, 1);
>      if (ret) {
> @@ -3394,6 +3434,18 @@ virtio_load(VirtIODevice *vdev, QEMUFile *f, int 
> version_id)
>                  continue;
>              }
>
> +            if (migrating_backend) {
> +                /*

"prior to backend migration"?



> +                 * Indices are not synced prior backend migration (as we 
> don't
> +                 * stop vrings by GET_VRING_BASE). No reason to sync them 
> now,
> +                 * and do any checks.
> +                 */
> +                vdev->vq[i].used_idx = 0;
> +                vdev->vq[i].shadow_avail_idx = 0;
> +                vdev->vq[i].inuse = 0;
> +                continue;
> +            }
> +
>              nheads = vring_avail_idx(&vdev->vq[i]) - 
> vdev->vq[i].last_avail_idx;
>              /* Check it isn't doing strange things with descriptor numbers. 
> */
>              if (nheads > vdev->vq[i].vring.num) {
> @@ -3762,8 +3814,9 @@ int virtio_queue_set_guest_notifier(VirtIODevice *vdev, 
> int n, bool assign,
>      EventNotifierHandler *read_fn = is_config ?
>          virtio_config_guest_notifier_read :
>          virtio_queue_guest_notifier_read;
> +    bool migrating_backend = virtio_is_vhost_migrating_backend(vdev);
>
> -    if (assign) {
> +    if (assign && !migrating_backend) {
>          int r = event_notifier_init(notifier, 0);
>          if (r < 0) {
>              return r;
> @@ -3773,7 +3826,7 @@ int virtio_queue_set_guest_notifier(VirtIODevice *vdev, 
> int n, bool assign,
>      event_notifier_set_handler(notifier,
>                                 (assign && !with_irqfd) ? read_fn : NULL);
>
> -    if (!assign) {
> +    if (!assign && !migrating_backend) {
>          /* Test and clear notifier before closing it,*/
>          /* in case poll callback didn't have time to run. */
>          read_fn(notifier);
> @@ -4392,6 +4445,23 @@ done:
>      return element;
>  }
>
> +bool virtio_is_vhost_migrating_backend(VirtIODevice *vdev)
> +{
> +    VirtioDeviceClass *vdc = VIRTIO_DEVICE_GET_CLASS(vdev);
> +    struct vhost_dev *hdev;
> +
> +    if (!vdc->get_vhost) {
> +        return false;
> +    }
> +
> +    hdev = vdc->get_vhost(vdev);
> +    if (!hdev) {
> +        return false;
> +    }
> +
> +    return hdev->migrating_backend;
> +}
> +
>  static const TypeInfo virtio_device_info = {
>      .name = TYPE_VIRTIO_DEVICE,
>      .parent = TYPE_DEVICE,
> diff --git a/include/hw/virtio/virtio.h b/include/hw/virtio/virtio.h
> index 9a4a0a94aa..f94a7e5895 100644
> --- a/include/hw/virtio/virtio.h
> +++ b/include/hw/virtio/virtio.h
> @@ -238,6 +238,8 @@ struct VirtioDeviceClass {
>      bool (*skip_vhost_migration_log)(VirtIODevice *vdev);
>  };
>
> +bool virtio_is_vhost_migrating_backend(VirtIODevice *vdev);
> +
>  void virtio_instance_init_common(Object *proxy_obj, void *data,
>                                   size_t vdev_size, const char *vdev_name);
>
> --
> 2.48.1
>
>

Reply via email to