[Spice-devel] changing the timing of spice client linking in migration (RHBZ #725009)

[Gerd Hoffmann]

On 08/17/11 10:05, Christophe Fergeau wrote:
> On Tue, Aug 16, 2011 at 06:54:58PM -0400, Marc-André Lureau wrote:
>>> qemu
>>> =====
>>> ui/spice-core::migration_state_notifier should handle MIG_STATE_ACTIVE
>>> (for migration start) and MIG_STATE_ERROR/CANCELLED by calling
>>> spice_server_migrate_start, and spice_server_migrate_end,
>>> respectively.
>>> These callbacks are currently declared in spice-experimental.h.
>>
>> Contrary to Christophe, I don't think we should be afraid of using those
>> functions which have not been supported and used since quite some time,
>> afaik.
>
> Just to make things clear, I'm not afraid of this change, I was asking
> whether we should be careful when doing it :) I don't have a strong opinion
> on whether we should do it or not.

I don't see problems with it.  spice_server_migrate_start() + end() are 
there for the long-term plan to re-add seamless migration.  start() 
returns -1 since day one as it isn't implemented yet.  So we can simply 
make qemu fallback to switch-host in case start() fails.

cheers,
   Gerd