Il 21/02/2013 18:42, Juan Quintela ha scritto:
> Paolo Bonzini <[email protected]> wrote:
>> Accessing s->state outside the big QEMU lock will simplify a bit the
>> locking/unlocking of the iothread lock.
>>
>> Signed-off-by: Paolo Bonzini <[email protected]>
>
> Reviewed-by: Juan Quintela <[email protected]>
>
> We compensate the locking removal you did on the previous patch.
I thought this was thread-safe at any step (and it took a while to reach
that state :)), did I miss something?
>> + __sync_val_compare_and_swap(&s->state, MIG_STATE_ACTIVE,
>> MIG_STATE_ERROR);
>
> This can be done later, but can we change this to a macro/inline
> function:
>
> inline void migration_set_state(int state)
> {
> __sync_val_compare_and_swap(&s->state, MIG_STATE_ACTIVE, state);
> }
>
> or something like that?
Yes, especially so that we can add a tracepoint as in the
recently-posted series.
Paolo
