Bugzilla Bug 131108

+++ This bug was initially created as a clone of Bug #85450 +++

The problem does not depend on the gcc version and is not a gcc bug.  It also
appears with gcc-3.4.5-r1 and glibc-2.3.5 on my system.

The setjmp.S file (among others) contains a variable or label named __i686,
which is #define'd to be 1 if gcc compiles for i686 and up (-march=pentium2
etc.).

The subsequent patch undefines __i686 for assembler files and thus solves the
problem.  It should not harm on any other platform.

--- Makeconfig~ Wed Feb 16 11:50:19 2005
+++ Makeconfig  Sun Nov  6 18:39:08 2005
@@ -364,7 +364,7 @@
 # directives the preprocessor produces.  If you have troubling compiling
 # assembly code, try using -P here to suppress these directives.
 ifndef asm-CPPFLAGS
-asm-CPPFLAGS =
+asm-CPPFLAGS = -U__i686
 endif

 # ELF always supports init/fini sections


Original bug description:

I've applied 'fixes' for similar gcc-4 bugs but the error doesn't even change.
this occurred with fresh stage and a gcc-3.3 box, both p3's. the jan.18 build
idetical ebuilds for p3s on p4 without error and can build nptlonly or nptl for
p4 and prescott.

make[2]: Entering directory
`/var/tmp/portage/glibc-2.3.4.20050125-r1/work/glibc-2.3.4/setjmp'
i686-pc-linux-gnu-gcc ../sysdeps/i386/elf/setjmp.S -c  -I../include -I.
-I/var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-nptl/setjmp
-I.. -I../libio -I../nptl
-I/var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-nptl
-I../sysdeps/i386/elf -I../nptl/sysdeps/unix/sysv/linux/i386/i686
-I../nptl/sysdeps/unix/sysv/linux/i386 -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/i386/i686 -I../nptl/sysdeps/i386
-I../libidn/sysdeps/unix -I../sysdeps/unix/sysv/linux/i386
-I../sysdeps/unix/sysv/linux -I../sysdeps/gnu -I../sysdeps/unix/common
-I../sysdeps/unix/mman -I../sysdeps/unix/inet -I../sysdeps/unix/sysv/i386
-I../sysdeps/unix/sysv -I../sysdeps/unix/i386 -I../sysdeps/unix
-I../sysdeps/posix -I../sysdeps/i386/i686/fpu -I../sysdeps/i386/i686
-I../sysdeps/i386/i486 -I../nptl/sysdeps/i386/i486 -I../sysdeps/i386/fpu
-I../sysdeps/i386 -I../sysdeps/wordsize-32 -I../sysdeps/ieee754/ldbl-96
-I../sysdeps/ieee754/dbl-64 -I../sysdeps/ieee754/flt-32 -I../sysdeps/ieee754
-I../sysdeps/generic/elf -I../sysdeps/generic -nostdinc -isystem
/usr/lib/gcc/i686-pc-linux-gnu/3.4.3-20050110/include -isystem /usr/include
-D_LIBC_REENTRANT -D_LIBC_REENTRANT -include ../include/libc-symbols.h      
-DASSEMBLER  -DGAS_SYNTAX  -Wa,--noexecstack  -o
/var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-nptl/setjmp/setjmp.o
-MD -MP -MF
/var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-nptl/setjmp/setjmp.o.dt
-MT
/var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-nptl/setjmp/setjmp.o
../sysdeps/i386/elf/setjmp.S: Assembler messages:
../sysdeps/i386/elf/setjmp.S:63: Error: junk at end of line, first unrecognized
character is `1'
../sysdeps/i386/elf/setjmp.S:64: Error: junk at end of line, first unrecognized
character is `1'
../sysdeps/i386/elf/setjmp.S:65: Error: junk at end of line, first unrecognized
character is `1'
../sysdeps/i386/elf/setjmp.S:66: Error: unrecognized symbol type ""
../sysdeps/i386/elf/setjmp.S:66: Error: junk at end of line, first unrecognized
character is `1'
../sysdeps/i386/elf/setjmp.S:67: Error: junk at end of line, first unrecognized
character is `1'
../sysdeps/i386/elf/setjmp.S:70: Error: expected comma after name `' in .size
directive
make[2]: ***
[/var/tmp/portage/glibc-2.3.4.20050125-r1/work/build-default-i686-pc-linux-gnu-nptl/setjmp/setjmp.o]
Error 1
make[2]: Leaving directory
`/var/tmp/portage/glibc-2.3.4.20050125-r1/work/glibc-2.3.4/setjmp'
make[1]: *** [setjmp/subdir_lib] Error 2
make[1]: Leaving directory
`/var/tmp/portage/glibc-2.3.4.20050125-r1/work/glibc-2.3.4'
make: *** [all] Error 2

you forgot to post `emerge info`

Created an attachment (id=85386) [details]
Output of emerge --info

As you might see from the versions of the packages, this box is largely
unfinished and yet to be set up completely.  It is in fact a dual P-3 machine.

the question is, why doesnt this bug show up for a lot more people ...

it's because .S files are not compiled with optimization flags (like
-march=i686)

so why is yours being optimized ?  post the output of `emerge -v --info` please

I have compiled gcc --with-arch=pentium3, so pentium3 is its default
architecture and __i686 is always set, also when compiling .S files.

I suppose, most people don't compile glibc with processor specific
optimizations.

> I suppose, most people don't compile glibc with processor specific
> optimizations.

no ... the glibc build system ignores optimizations for .S files since they
dont really make any sense

Created an attachment (id=85500) [details]
1130_all_glibc-2.4-undefine-__i686.patch

try this instead ... seems to work for me

patch added to 2.4-r2

*** Bug 213077 has been marked as a duplicate of this bug. ***

*** Bug 85450 has been marked as a duplicate of this bug. ***

*** Bug 234381 has been marked as a duplicate of this bug. ***

*** Bug 213077 has been marked as a duplicate of this bug. ***