On 03/14/13 17:36, Hans de Goede wrote:
> From: Alon Levy <al...@redhat.com>
>
> The target has not seen the guest_connected event via
> spice_chr_guest_open or spice_chr_write, and so spice server wrongly
> assumes there is no agent active, while the client continues to send
> motion events only by the agent channel, which the server ignores. The
> net effect is that the mouse is static in the guest.
>
> By registering the interface on post load spice server will pass on the
> agent messages fixing the mouse behavior after migration.
> +static VMStateDescription spice_chr_vmstate = {
> + .name = "spice-chr",
> + .version_id = 1,
> + .minimum_version_id = 1,
> + .post_load = spice_chr_post_load,
> + .fields = (VMStateField[]) {
> + VMSTATE_UINT32(guest_open, SpiceCharDriver),
> + VMSTATE_END_OF_LIST()
> + },
> +};
> + vmstate_register(NULL, -1, &spice_chr_vmstate, s);
> +
That is a showstopper. If there are two of these there is no reliable
way to figure which is which.
Do we actually need that in the first place? I assume virtio-serial
will migrate the guest-open state anyway. So if we can use that somehow
we don't have to live migrate the chardev state ...
cheers,
Gerd
