This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Bootstrap Failure (i686-pc-linux-gnu, --with-arch=pentium4)
- From: Volker Reichelt <reichelt at igpm dot rwth-aachen dot de>
- To: Ranjit Mathew <rmathew at gmail dot com>
- Cc: gcc at gcc dot gnu dot org, bernds at gcc dot gnu dot org
- Date: Fri, 17 Jun 2005 14:02:09 +0200 (CEST)
- Subject: Re: Bootstrap Failure (i686-pc-linux-gnu, --with-arch=pentium4)
> Hi,
>
> For two consecutive days, I have been unable to
> build GCC mainline on i686-pc-linux-gnu:
> /home/ranmath/src/gcc/gcc-20050617/gcc/config/i386/i386.c: In function
> 'ix86_expand_vector_init':
> /home/ranmath/src/gcc/gcc-20050617/gcc/config/i386/i386.c:17080:
> error: insn does not satisfy its constraints:
> (insn 367 343 378 4
> /home/ranmath/src/gcc/gcc-20050617/gcc/config/i386/i386.c:17047 (set
> (reg:QI 138)
> (const_int 0 [0x0])) 45 {*movqi_1} (nil)
> (expr_list:REG_EQUIV (const_int 0 [0x0])
> (nil)))
> /home/ranmath/src/gcc/gcc-20050617/gcc/config/i386/i386.c:17080:
> internal compiler error: in reload_cse_simplify_operands, at
> postreload.c:391
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See <URL:http://gcc.gnu.org/bugs.html> for instructions.
> ----------------------------------------- 8<
> -----------------------------------------
>
> I build with "--with-arch=pentium4 --disable-checking" and that
> might explain why no one has apparently reported it yet.
It has already been reported, it's PR22089 ;-)
http://gcc.gnu.org/PR22089
Nobody patches reload without causing regressions.
Sometimes even looking hard at the code can cause failures. ;-)
> Thanks,
> Ranjit.
Regards,
Volker