On 2026-06-19 3:55 AM, Andrey Drobyshev wrote:
> During CPR migration, instead of reopening /dev/vhost-vsock device on
> each initialization, let's make sure the already open FD is preserved and
> properly reused on a subsequent initialization.
>
> Signed-off-by: Andrey Drobyshev <[email protected]>
> ---
> hw/virtio/vhost-vsock.c | 25 ++++++++++++++++++++-----
> 1 file changed, 20 insertions(+), 5 deletions(-)
>
> diff --git a/hw/virtio/vhost-vsock.c b/hw/virtio/vhost-vsock.c
> index da244eb1657..ea9a7d2149d 100644
> --- a/hw/virtio/vhost-vsock.c
> +++ b/hw/virtio/vhost-vsock.c
> @@ -20,6 +20,7 @@
> #include "hw/core/qdev-properties.h"
> #include "hw/virtio/vhost-vsock.h"
> #include "monitor/monitor.h"
> +#include "migration/cpr.h"
>
> static void vhost_vsock_get_config(VirtIODevice *vdev, uint8_t *config)
> {
> @@ -126,6 +127,7 @@ static void vhost_vsock_device_realize(DeviceState *dev,
> Error **errp)
> VHostVSockCommon *vvc = VHOST_VSOCK_COMMON(dev);
> VirtIODevice *vdev = VIRTIO_DEVICE(dev);
> VHostVSock *vsock = VHOST_VSOCK(dev);
> + DeviceState *proxy = qdev_get_parent_bus(DEVICE(vsock))->parent;
> int vhostfd;
> int ret;
>
> @@ -141,7 +143,8 @@ static void vhost_vsock_device_realize(DeviceState *dev,
> Error **errp)
> }
>
> if (vsock->conf.vhostfd) {
> - vhostfd = monitor_fd_param(monitor_cur(), vsock->conf.vhostfd, errp);
> + vhostfd = cpr_get_fd_param(proxy->id, vsock->conf.vhostfd, 0,
> + errp);
> if (vhostfd == -1) {
> error_prepend(errp, "vhost-vsock: unable to parse vhostfd: ");
> return;
> @@ -151,10 +154,20 @@ static void vhost_vsock_device_realize(DeviceState
> *dev, Error **errp)
> return;
> }
> } else {
> - vhostfd = open("/dev/vhost-vsock", O_RDWR);
I think this patchset also implicitly fixes the duplicate vhost thread issue
noticed when testing vhost-scsi.
With vhost-scsi, I noticed that if QEMU opens the vhost device fd without
O_CLOEXEC, the fd may survive cpr-exec unexpectedly. Then the new QEMU opens
another vhost fd, and we can end up with duplicate vhost threads.
I did not see the same issue when the fd was opened outside QEMU and passed in
via QMP getfd, probably because QEMU sets close-on-exec on fds received that
way.
With this patchset, vhost-vsock no longer opens a new fd on the incoming side.
It restores the fd from CPR state, so the "duplicate vhost threads" issue should
not happen for this path.
I am still thinking about whether it is worth sending a temporary patch to add
O_CLOEXEC to these vhost device opens, or whether we should enhance vhost-scsi
in the same way as this patchset instead.
Thank you very much!
Dongli Zhang
> - if (vhostfd < 0) {
> - error_setg_file_open(errp, errno, "/dev/vhost-vsock");
> - return;
> + if (cpr_is_incoming()) {
> + vhostfd = cpr_find_fd(proxy->id, 0);
> + if (vhostfd < 0) {
> + error_setg(errp, "vhost-vsock: could not restore vhost FD "
> + "for %s", proxy->id);
> + return;
> + }
> + } else {
> + vhostfd = open("/dev/vhost-vsock", O_RDWR);
> + if (vhostfd < 0) {
> + error_setg_file_open(errp, errno, "/dev/vhost-vsock");
> + return;
> + }
> + cpr_save_fd(proxy->id, 0, vhostfd);
> }
>
> if (!qemu_set_blocking(vhostfd, false, errp)) {
> @@ -193,10 +206,12 @@ static void vhost_vsock_device_unrealize(DeviceState
> *dev)
> {
> VHostVSockCommon *vvc = VHOST_VSOCK_COMMON(dev);
> VirtIODevice *vdev = VIRTIO_DEVICE(dev);
> + DeviceState *proxy = qdev_get_parent_bus(dev)->parent;
>
> /* This will stop vhost backend if appropriate. */
> vhost_vsock_set_status(vdev, 0);
>
> + cpr_delete_fd(proxy->id, 0);
> vhost_dev_cleanup(&vvc->vhost_dev);
> vhost_vsock_common_unrealize(vdev);
> }
