[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [patches] eglibc builds and tests on mpc8540



Hello, Steven and Jim!

I've spent some time to check out several of the patches you both
mentioned in the last mails.
As far as I understood the code, a lot of them went into gcc-4.2.0,
so it seems to me as the soft-fp support is already there and should
be working.

I.e.
http://gcc.gnu.org/viewcvs/branches/csl/sourcerygxx-4_1/gcc/config/soft-fp/op-common.h?r1=118640&r2=118639&pathrev=118640
is from 2006/11/10 and was applied to gcc-4.2.0, however some lines have
changed again.

So, for me, it looks like as gcc-4.2.0 can be ideally used "as is"
to compile the latest eglibc-svn-trunk.
Is this correct?

Currently, I am stuck with building gcc-4.2.0 with soft-fp.
I read http://www.ecos.sourceware.org/ml/libc-ports/2006-06/msg00024.html
and lots of other stuff and came to the configure options:
        ../$name-$version/configure \
                --prefix=/usr \
                --libexecdir=/usr/lib \
                --enable-languages=c,c++,objc \
                --enable-threads=posix \
                --enable-__cxa_atexit \
                --enable-clocale=gnu \
                --enable-shared \
                --disable-nls \
                --with-x=no \
                --with-cpu=8540 \
                --with-tune=8540 \
                --with-float=softfp \
                --with-long-double-128

Is this configuration okay?

The last compilation using --without-fp instead of --with-float=softfp failed:

/usr/ports/ckcore/gcc-4.2.0-e500/work/src/build/./gcc/xgcc -B/usr/ports/ckcore/gcc-4.2.0-e500/work/src/build/./gcc/ -B/usr/powerpc-unknown-linux-gnu/bin/ -B/usr/powerpc-unknown-linux-gnu/lib/ -isystem /usr/powerpc-unknown-linux-gnu/include -isystem /usr/powerpc-unknown-linux-gnu/sys-include -O2 -O2 -O2 -pipe  -DIN_GCC    -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition  -isystem ./include  -I. -I. -I../../gcc-4.2.0/gcc -I../../gcc-4.2.0/gcc/. -I../../gcc-4.2.0/gcc/../include -I../../gcc-4.2.0/gcc/../libcpp/include  -I../../gcc-4.2.0/gcc/../libdecnumber -I../libdecnumber  -g0 -finhibit-size-directive -fno-inline-functions -fno-exceptions -fno-zero-initialized-in-bss -fno-toplevel-reorder  -msdata=none \
          -c ../../gcc-4.2.0/gcc/crtstuff.c -DCRT_BEGIN \
          -o crtbegin.o
cc1: internal compiler error: Illegal instruction
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
make[3]: *** [crtbegin.o] Error 1

Any ideas?

Well, I feel pretty lost in the mixture of patches and configure switches.
Can some of you please explain, if the following (more or less undocumented)
switches are still valid or necessary or obsolete:

--with-libc-soft-fp=glibc2.4
(found in http://www.ecos.sourceware.org/ml/glibc-bugs/2006-06/msg00056.html)
This actually worries me as I am currently on glibc-2.3.6.
Is that a problem?

--without-fp
Does this switch enable soft-fp on PowerPC automagically?

--with-float=soft
--with-float=softfp

--disable-multilib
(I don't need 64bit, so I would try it. But are float and nof libraries
also considered as a multilib installation?)

--enable-bootstrap

To save some compile-time for testing, I will retry to build gcc without objc.

Seems I have to climb much further up on this steep learning curve. :-/
Thanks for your help so far.
--
Clemens Koller
__________________________________
R&D Imaging Devices
Anagramm GmbH
Rupert-Mayer-Straße 45/1
Linhof Werksgelände
D-81379 München
Tel.089-741518-50
Fax 089-741518-19
http://www.anagramm-technology.com