This is the mail archive of the
mailing list for the GCC project.
Re: 126.gcc broken as of Dec15
- From: Jan Hubicka <jh at suse dot cz>
- To: Diego Novillo <dnovillo at redhat dot com>
- Cc: Jan Hubicka <jh at suse dot cz>, gcc at gcc dot gnu dot org
- Date: Fri, 21 Dec 2001 17:20:08 +0100
- Subject: Re: 126.gcc broken as of Dec15
- References: <20011220170536.A24674@tornado.cygnus.com>
> This patch seems to have broken 126.gcc in SPEC95. It's a
> runtime error triggered by -O3. The resulting 126.gcc binary
> fails to compile c-decl-s.i.
> I still haven't chased the bug, but reverting this patch fixes
> the regression. Both trunk and the cfg-branch are broken.
> Fri Dec 14 21:23:54 CET 2001 Jan Hubicka <email@example.com>
> * config.gcc: Revamp target_cpu_default2 to strings;
> support new x86 variants.
> * i386.c (override_options): Default x86_cpu_string and x86_arch_string
> properly; set prefetch_sse.
> * i386.h (x86_prefetch_sse): Declare.
> (TARGET_PREFETCH_SSE): New.
> (CPP_CPU_DEFAULT_SPEC): Define according to the new macros.
> (TARGET_CPU_DEFAULT_*): New.
This sounds really strange, as the prefetch code is disabled by default.
I will re-check the patch tonight.