> The VGA emulation uses a timout...
> (should use DGA here, it's a lot faster, but then this IS
> a first implementation ;))
It uses a SIGALRM handler, but this is not used a timeout in the
sense that it is supposed to abort execution. Thus, it doesn't
hurt if the FMWRUNGUEST ioctl restarts automatically after the
handler has terminated. (As opposed to the old *timeout* handler,
where the whole point of the handler was to abort the FMWRUNGUEST.)
> How exactly do I check this ?
> I can't think of why it would change, anyway...
> The system call shouldn't touch it, really.
This was the bug that Kevin recently noticed: on restart, the *old*
contents of the guest_context were used again, so that guest execution
always restarted at the beginning :-/ It should be fixed now, just
watch out for strange effects (counter skipping backwards etc.) in
case there are still problems ...
Bye,
Ulrich
--
Ulrich Weigand,
IMMD 1, Universitaet Erlangen-Nuernberg,
Martensstr. 3, D-91058 Erlangen, Phone: +49 9131 85-27688
