RE: kernel broken using CPUTYPE and 2.95.4

["Robin P. Blanchard" <robin.blanchard@xxxxxxxxxxxxx>]

> -----Original Message-----
> From: kernel-errors@xxxxxxxxxxxxxxxxxxxxxxx 
> [mailto:kernel-errors@xxxxxxxxxxxxxxxxxxxxxxx] On Behalf Of 
> Hiten Pandya
> Sent: Tuesday, May 11, 2004 9:27 AM
> To: kernel@xxxxxxxxxxxxxxxxxxxxxxx
> Subject: Re: kernel broken using CPUTYPE and 2.95.4
> 
> Jeroen Ruigrok/asmodai wrote:
> > -On [20040510 22:12], David Rhodus (drhodus@xxxxxxxxxxx) wrote:
> > 
> >>Well we already have a mudge in for when people try to use -O3 on a 
> >>kernel build, I wonder if we should add this in too ?:-/ 
> Saying that 
> >>even though I don't like messing around with the options a user has 
> >>specified.
> > 
> > 
> > I wholeheartedly agree with you on that David, however in this 
> > specific case I just have to wonder if we should protect the users 
> > against themselves and us from the users.  Yes, I know, I should've 
> > known better, but what's the use in having CPUTYPE in 
> make.conf if we 
> > _known_ it will create binaries which cause strange 
> problems to creep 
> > up?  Might as well just yank it and be done with the headache.
> > 
> > Opinions on this?  Matt?
> 
> 	I know I am not Matt (!) but you can't throw away the
> 	CPUTYPE option just because `p4' doesn't work, while things
> 	like `pentiumpro', `p6', `p3' and such work perfectly well.
> 
> 	Pentium-4 is known to cause problems, this is the same
> 	in FreeBSD-land last time I checked.
> 
> > That is, until I get the compiler done... *whistles*
> > 
> 
> 	You keep us waiting dammnit! *takes out whip* :-)
> 
> 		-Hiten
> 		hmp@xxxxxxxxxxxxx
> 
> 

Well, for the record, setting CPUTYPE=i686 with CCVER=gcc2 yields a broken
kernel for me. Unsetting CPUTYPE yields working kernel.

---------------------------------------
Robin P. Blanchard
Systems Integration Specialist
Georgia Center for Continuing Education
fon: 706.542.2404 < > fax: 706.542.6546
---------------------------------------