Hi !

On Tue, 14 Dec 2004, Thor Lancelot Simon wrote:
> Thinking about this a bit more, I think I should put back the old RB_HALT
> behaviour for domain 0.  If you "halt" domain 0 instead of "reboot"ing
> it, you probably do not want the machine to come back up right away.
>
> What do other people think about this?

I think it's possible to shut down domain0 and leave other domains running,
isn't it? Just besides the sende of this :)

If this is possible, i see a more generic quistion on how about handling
domain0. One might want to set up all domains and then terminate the
"controlling master" if this is somehow possible (don't know if i tried this
with Linux so far).

If it's not possible, i also think we should separate reboot/halt here (i wrote
about this issue in my patch mail). For now i didn't think about the
differentiation between Dom0 and DomU at this place. SCHEDOP_yield() on halt
(looped), SCHEDOP_exit on reboot (as Xen reboots automatically afterwards,
which is also configurable i think)

...Michael

-- 
http://www.unixiron.org/    Home Powered by: (Net|Open|Free)BSD IRIX NonStop-UX
Solaris AIX HP-UX Tru64 MUNIX Ultrix VMS SINIX Dolphin_Unix OpenStep MacOS A/UX