This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: egcs 2.91.66 ix86 problems...
- To: Richard Henderson <rth at cygnus dot com>
- Subject: Re: egcs 2.91.66 ix86 problems...
- From: Andreas Jaeger <aj at arthur dot rhein-neckar dot de>
- Date: 03 Jul 1999 10:55:12 +0200
- Cc: Richard Henderson <rth at twiddle dot net>, Cristian Gafton <gafton at redhat dot com>, egcs-patches at egcs dot cygnus dot com, drepper at cygnus dot com
On Thu, 1 Jul 1999 20:43:47 Richard Henderson wrote:
Richard > How about this.
Richard >
Richard >
Richard > r~
Richard >
Richard >
Richard >
Richard > * i386.h (CC1_CPU_SPEC): Don't add -march=foo.
Richard > (CPP_486_SPEC, CPP_586_SPEC): Delete.
Richard > (CPP_K6_SPEC, CPP_686_SPEC): Delete.
Richard > (CPP_CPU_DEFAULT_SPEC): Define to __tune_foo__.
Richard > (CPP_CPU_SPEC): Make -mcpu=foo define __tune_foo__. Make -march=bar
Richard > define __bar__ and related symbols.
Richard > (EXTRA_SPECS): Remove deleted specs.
Richard >
I applied your patch to gcc 2.95 [1] and noticed one strange
settings of flags on my i486:
$ /opt/gcc-2.95/bin/gcc -m386 -E -dM math.c |grep 86
[...]
#define __i386__ 1
#define __tune_i386__ 1
#define __i386 1
#define __tune_i486__ 1
#define i386 1
Why is both __tune_i486__ and __tune_i386__ defined? Does this come
from the following specs rule:
*cpp_cpu_default:
-D__tune_i486__
The other -m/-march/-mcpu values I checked look fine.
Andreas
Footnotes:
[1] Yes, I'd like to have this issue resolved asap and for ever;-).
--
Andreas Jaeger aj@arthur.rhein-neckar.de jaeger@informatik.uni-kl.de
for pgp-key finger ajaeger@aixd1.rhrk.uni-kl.de