[oe] libvpx causing gcc ICE when building for qemux86-64

[Khem Raj]

On Mon, Nov 21, 2011 at 12:30 PM, Martin Jansa <martin.jansa at gmail.com> wrote:
> On Fri, Nov 04, 2011 at 03:20:28PM +0100, Martin Jansa wrote:
>> On Fri, Nov 04, 2011 at 03:18:06PM +0100, Martin Jansa wrote:
>> > Hi, with latest gcc-4.6 from meta-oe I get this ICE:
>> >
>> > OE @ ~/shr-core/tmp/work/x86_64-oe-linux/libvpx-0.9.5-r0.0/libvpx-v0.9.5 $ x86_64-oe-linux-gcc    -m64 --sysroot=/OE/shr-core/tmp/sysroots/qemux86-64 -O2 -pipe -g -feliminate-unused-debug-types -fPIC -O3 -Wall -Wdeclaration-after-statement -Wdisabled-optimization -Wpointer-arith -Wtype-limits -Wcast-qual -Wno-unused -I. -I"/OE/shr-core/tmp/work/x86_64-oe-linux/libvpx-0.9.5-r0.0/libvpx-v0.9.5" -I/OE/shr-core/tmp/work/x86_64-oe-linux/libvpx-0.9.5-r0.0/libvpx-v0.9.5/vp8 -I/OE/shr-core/tmp/work/x86_64-oe-linux/libvpx-0.9.5-r0.0/libvpx-v0.9.5/vp8 -c -o md5_utils.c.o md5_utils.c
>> > md5_utils.c: In function 'byteSwap':
>> > md5_utils.c:30:1: internal compiler error: Segmentation fault
>> > Please submit a full bug report,
>> > with preprocessed source if appropriate.
>> > See <http://gcc.gnu.org/bugs.html> for instructions.
>>
>> And dropping -O3 is enough to make it compile again.
>
> FWIW: Just tested with newer gcc from oe-core and the problem is still there.

Its a regression caused by one of linaro patches I suppose. I will see
what I can find.

-Khem