[oe] question meta-toolchain on x86_64 for armv5te

[Khem Raj]

On Tue, Nov 10, 2009 at 8:03 AM, Johannes Maier <jm at huber-online.com> wrote:
> Hi at all,
>
> I'm new to openembedded and I didn't find any answers to my questions so I hope to find here some help.
>
> I get an error while building the meta-toolchain on x86_64 for armv5te.
>
> So I wanted to ask if it is in general better to use a 32bit system for such tasks since I don't get any error there?
>
> The error is as follows:
>
> angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/config -DLOCALEDIR="\"/usr/local/angstrom/arm/share/locale\"" -DHAVE_CONFIG_H -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/../include/opcode -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/../readline/.. -I../bfd -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/../bfd -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/../include -I../libdecnumber -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/../libdecnumber   -DMI_OUT=1 -DTUI=1
> -isystem/home/test/oe/stuff/opene
> mbedded/tmp/staging/x86_64-linux/usr/include -Wall -Wdeclaration-after-statement -Wpointer-arith -Wformat-nonliteral -Wno-pointer-sign -Wno-unused -Wno-switch -Wno-char-subscripts -Werror /home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/value.c
> | ccache gcc -c -isystem/home/test/oe/stuff/openembedded/tmp/staging/x86_64-linux/usr/include -O2   -I. -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/config -DLOCALEDIR="\"/usr/local/angstrom/arm/share/locale\"" -DHAVE_CONFIG_H -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/../include/opcode -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/../readline/.. -I../bfd -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/../bfd -I/home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/../include -I../libdecnumber
> -I/home/test/oe/stuff/
> openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/../libdecnumber   -DMI_OUT=1 -DTUI=1 -isystem/home/test/oe/stuff/openembedded/tmp/staging/x86_64-linux/usr/include -Wall -Wdeclaration-after-statement -Wpointer-arith -Wformat-nonliteral -Wno-pointer-sign -Wno-unused -Wno-switch -Wno-char-subscripts -Werror /home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/eval.c
> | cc1: warnings being treated as errors
> | /home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/eval.c: In function 'evaluate_subexp_standard':
> | /home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/eval.c:1705: error: 'subscript_array' may be used uninitialized in this function
> | /home/test/oe/stuff/openembedded/tmp/work/x86_64-armv5te-sdk-angstrom-linux-gnueabi/gdb-cross-sdk-6.8-r2/gdb-6.8/gdb/eval.c:1644: note: 'subscript_array' was declared here

you might be using gcc 4.4 on your build box.
gcc 4.4 spits more warnings like what you see. You can do couple of
things. add --disable-werror to gdb recipes but
I would not recommend that. Secondly you can create a patch to gdb/eval.c
Modify line 1644 of gdb/eval.c to
int subscript_array[MAX_FORTRAN_DIMS] = {0};

and this should make gcc 4.4 happy.

-Khem
Thanks