This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: can't build x86-32 libc on x86-64 with mainline binutils
- From: "H. J. Lu" <hjl at lucon dot org>
- To: Alexandre Oliva <aoliva at redhat dot com>
- Cc: libc-alpha at sources dot redhat dot com, binutils at sources dot redhat dot com
- Date: Thu, 2 Mar 2006 10:33:13 -0800
- Subject: Re: can't build x86-32 libc on x86-64 with mainline binutils
- References: <or4q2gsou6.fsf@free.oliva.athome.lsd.ic.unicamp.br>
On Thu, Mar 02, 2006 at 03:06:57PM -0300, Alexandre Oliva wrote:
> Am I the only one having trouble assembling
> glibc/sysdeps/i386/fpu/e_atanh.S with -m32 with binutils mainline
> toplevel-bootstrapped along with GCC trunk on x86_64? I've hit this
> for several days now, maybe several weeks? (time flies)
>
> The assembler fails the sanity check for carry == 0 on
> gas/atof-generic.c:419 on line 40 of the preprocessed asm file. In
> fact, assembling that line by itself produces the same error.
>
> Here's how I configured the toolchain:
>
> ../configure amd64-linux-gnu --enable-languages=all
>
> and here's how I configured glibc:
>
> ../src/configure '--prefix=/usr' '--enable-add-ons=nptl' \
> '--without-cvs' '--with-tls' '--with-headers=/usr/include' \
> '--build=i686-pc-linux-gnu' 'CC=gcc -m32 -march=i686 -U__i686' \
> 'CXX=g++ -m32 -march=i686 -U__i686'
>
> (the ugly -U__i686 was introduced months ago; IIRC without it we'd
> fail to assemble the GOT register set up routines that have __i686 in
> their names in such a way that the preprocessor would expand __i686 to
> 1. I haven't tried without it, I've only just realized it was still
> there)
So mainline assembler doesn't work for you. Can you open a bug report
with a small testcase if you haven't done so? I will look into it.
H.J.