Re: interrupts on APs

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Sat, Dec 10, 2011 at 12:29:54PM +0000, Matthias Scheler wrote:
> On Fri, Dec 09, 2011 at 06:10:04PM +0100, Manuel Bouyer wrote:
> > > I've got a Dual Core machine with an Intel E7230 chipset:
> > > 
> > >   HP ProLiant ML110 G4 (1.0)
> > >   [...]
> > >   pchb0: Intel E7230 Host (rev. 0xc0)
> > > 
> > > But I would need a patch for "netbsd-5".
> > 
> > I think it's the same as with -HEAD: just change:
> > #if 0
> >                         if (ci == NULL ||
> >                             ci->ci_nintrhand > lci->ci_nintrhand) {
> >                                 ci = lci;
> >                         }
> > #else
> >                        ci = &cpu_info_primary;
> > #endif
> > 
> > to #if 1 in x86/x86/intr.c
> 
> No, it is not. The above code was only added in revision 1.61 and commented
> out in revision 1.62. The "netbsd-5" branch was forked of version 1.57.

Sorry, I though the attempt to route interrupts to all processors was prior
to netbsd-5.
I don't know if only this part of code is needed, or if something else needs
to be changed ...

Could you still try to boot a -current kernel ? I don't think you need to
run with it for long, in my case a single-user boot was enough to notice
it didn't work (it didn't even boot to single user).

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--