Bug#307985: libc6: 2.3.5-1 ppc built lacks GLIBC_2.3.3 version in libpthread
> > gcc-3.4 tst-fini1mod.c -c -std=gnu99 -O2 -O3 -Wall -Winline -Wstrict-prototypes -Wwrite-strings -fstrict-aliasing -g -g1 -pipe -mnew-mnemonics
> > -fpic -I../include -I. -I/home/benh/glibc-2.3.5/build-tree/powerpc-nptl/nptl -I.. -I../libio -I/home/benh/glibc-2.3.5/build-tree/powerpc-np
> > tl -I../sysdeps/powerpc/powerpc32/elf -I../sysdeps/powerpc/elf -I../nptl/sysdeps/unix/sysv/linux/powerpc/powerpc32 -I../nptl/sysdeps/unix/sysv/l
> > inux/powerpc -I../nptl/sysdeps/unix/sysv/linux -I../nptl/sysdeps/pthread -I../sysdeps/pthread -I../nptl/sysdeps/unix/sysv -I../nptl/sysdeps/unix
> > -I../nptl/sysdeps/powerpc -I../sysdeps/unix/sysv/linux/powerpc/powerpc32 -I../sysdeps/unix/sysv/linux/powerpc -I../sysdeps/unix/sysv/linux -I..
> > /sysdeps/gnu -I../sysdeps/unix/common -I../sysdeps/unix/mman -I../sysdeps/unix/inet -I../sysdeps/unix/sysv -I../sysdeps/unix/powerpc -I../sysdep
> > s/unix -I../sysdeps/posix -I../sysdeps/powerpc/powerpc32/fpu -I../sysdeps/powerpc/powerpc32 -I../sysdeps/wordsize-32 -I../sysdeps/powerpc/soft-f
> > p -I../sysdeps/powerpc/fpu -I../sysdeps/powerpc -I../sysdeps/ieee754/flt-32 -I../sysdeps/ieee754/dbl-64 -I../sysdeps/ieee754 -I../sysdeps/generi
> > c/elf -I../sysdeps/generic -nostdinc -isystem /usr/lib/gcc/powerpc-linux/3.4.4/include -isystem /home/benh/glibc-2.3.5/debian/include -D_LIBC_RE
> > ENTRANT -D_LIBC_REENTRANT -include ../include/libc-symbols.h -DPIC -DSHARED -DNOT_IN_libc=1 -DNOT_IN_libc -o /home/benh/glibc-2.3.5/build-tre
> > e/powerpc-nptl/nptl/tst-fini1mod.os -MD -MP -MF /home/benh/glibc-2.3.5/build-tree/powerpc-nptl/nptl/tst-fini1mod.os.dt -MT /home/benh/glibc-2.3.
> > 5/build-tree/powerpc-nptl/nptl/tst-fini1mod.os
> > ../elf/tst-execstack-mod.c: In function `tryme':
> > ../elf/tst-execstack-mod.c:24: internal compiler error: in gen_subprogram_die, at dwarf2out.c:10913
> >
> > Looks like our gcc-3.4 isn't quite there yet ...
>
> Huh. I've never seen that one before; yeah, file a bug, I suppose.
Ok, that happens to be a known old gcc-3.4 regression (#16676 in gcc
bugzilla). It is "caused" by the use of -g1 and nested functions. debian
nptl makefile adds -g1 for a reason I don't know. Turning that into -g2
appear to fix it for a couple of files, I'm now trying a full -g2 build.
I'm trying to get gcc folks interested in fixing it, but it seem to be
difficult to get them interested in anything non-4.0 :(
Ben.
Reply to: