I have updated patches [1] sent by David Vrabel last year: [1] https://lists.gnu.org/archive/html/qemu-devel/2018-03/msg05077.html
This change adds a new ???reset device??? feature to the vhost-user protocol and a corresponding VHOST_USER_RESET_DEVICE message to notify vhost-user backends when a virtio scsi device is reset by a guest. It also adds support for this new feature/message in vhost-user-scsi. Now, iff a vhost-user-scsi backend reports that it supports the new "reset device" feature, QEMU will send a VHOST_USER_RESET_DEVICE message when the guest resets the virtio device. Existing backends will be unaffected. Other types vhost-user backends can benefit from using this new message. Those built using libvhost-user, for example, rely on the depricated VHOST_USER_RESET_OWNER message to notify vhost-user backends about device resets and can be updated to use this new supported message. Raphael Raphael Norwitz (2): vhost-user: add VHOST_USER_RESET_DEVICE to reset devices vhost-user-scsi: reset the device if supported docs/interop/vhost-user.rst | 15 +++++++++++++++ hw/scsi/vhost-user-scsi.c | 24 ++++++++++++++++++++++++ hw/virtio/vhost-user.c | 8 +++++++- 3 files changed, 46 insertions(+), 1 deletion(-) -- 1.8.3.1