Re: bouyer-xen2 branch status

[Thor Lancelot Simon <tls%rek.tjls.com@localhost>]

On Sat, Dec 18, 2004 at 04:42:54PM +0000, Christos Zoulas wrote:
> In article <Pine.GSO.4.61.0412181142570.2082@rfhpc8317>,
> Hubert Feyrer <hubert%feyrer.de@localhost> wrote:
> >On Fri, 17 Dec 2004, Manuel Bouyer wrote:
> >> - find out why xbd will hang at autoconf when I remove the printf in
> >>  xennet_ctrlif_rx()
> >
> >Maybe this is one of the compiler bugs that require the kernel to be built 
> >with -O0 to work, and that was mentioned here before?
> 
> I think that the -O0 is a red herring and someone must have forgotten a
> "volatile" somewhere, or miffed some assembly code.

All the NetBSD/xen kernels on the Project servers are built with -O2
(though they don't specify any CPU-specific optimization flags).

There seems to be a lot of misinformation going around: first the bit
about needing to use a shell script to pad kernels, now this about how
it's necessary to compile with -O0.  I am not sure where it's coming
from.

Christian did mention a problem with kernels built with CPU-specific
optimizations that he said was fixed in the Xen tree.

-- 
 Thor Lancelot Simon                                          
tls%rek.tjls.com@localhost

"The inconsistency is startling, though admittedly, if consistency is to be
 abandoned or transcended, there is no problem."                - Noam Chomsky