Re: x86 release builds are slow

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

On Thu, May 01, 2008 at 03:08:24AM +0000, Steven M. Bellovin wrote:
> On Wed, 30 Apr 2008 21:49:56 -0500
> M Graff <explorer%flame.org@localhost> wrote:
> 
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> > 
> > Andrew Doran wrote:
> > 
> > | For xen that might not be possible because GRUB is used. I don't
> > know. If it
> > | can't be done, what are the major reasons for having seperate DOM0
> > and DOMU
> > | kernels?
> > 
> > DOM0 has all the real devices?  :)
> > 
> > Seriously, when you multiply the overhead of unneeded devices by
> > combining DOM0 and DOMU, won't that increase memory overhead quickly?
> > 
> > Seems to me... if you want to make things build faster overall, we
> > need to start making the kernel more dynamic and load modules, not
> > compiled in (and therefore compiled N times) code.
> > 
> That would be a nice direction.  However, there's another approach that
> may be worth talking about: have a single repository for shared .o
> file, and build+link kernels with a bit of unique source plus the
> shared .o files.  (BSD/OS worked that way, since it was mostly not a
> source distribution.)

I don't think this would work for native x86 vs Xen kernels. There's too
much difference in header files (constants, macros and inline functions).

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