> 1) Code is compiled with -D686 instead of -D586 or -D486, etc. Therefore,
> some CPU-dependant code for i686 will be generated, depending on #defines
> in the source.
> 2) The code is compiled with -march=i686 instead of -march=i586 or
> -march=i486, etc.
Yep, yesterday night I came to that conclusion too after compiling
2.3.32pre4 as Pentium with TSC, not i686 (that configuration worked fine on
my plain Pentium). It seems to work (must check later), and comparing the
CFLAGS with the 2.2.14pre's (which use -m486, AFAIR). Compiler bug? Weird,
as the same problems have shown up with several compiler versions. And to
find out if the bug bites me is quite costly, so...
I'll still need a simple test case that shows the problem. And some
time. Got some to spare? ;-)
> You said you were using the stock RH6.1 compiler; which compiler is that?
> egcs1.1.2? gcc2.95.2? Or a different version?
Currently egcs-1.1.2-24 from RehHat 6.1, but I've also used gcc-2.95.2 and
several gcc snapshots. Never had trouble with them on my plain Pentium...
> If the problem persists, it could indeed be a bug in i686 specific code
> in the kernel.
Neither alternative sounds too appetizing, to tell the truth.
Thanks!
-- Dr. Horst H. von Brand mailto:vonbrand@inf.utfsm.cl Departamento de Informatica Fono: +56 32 654431 Universidad Tecnica Federico Santa Maria +56 32 654239 Casilla 110-V, Valparaiso, Chile Fax: +56 32 797513- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/