2010/7/19 Tim Blechmann <[email protected]>:
>> >> Here's a backtrace http://pastebin.ca/1903514 , not sure if it's got
>> >> the info you need
>> >
>> > yes ... seems to be an alignment issue ... unfortunately, gcc has some
>> > issues, with implementing the attribute aligned ... it somehow seems to
>> > be a feature, they once implemented, but never really supported :/
>> >
>> > can you try to tweak the -mpreferred-stack-boundary and -mincoming-stack-
>> > boundary options? in the long run, i will probably need to re-implement
>> > the cache_aligned_array class, but i won't be able to do that today ...
>>
>> -mpreferred-stack-boundary=2 -mincoming-stack-boundary=2
>>
>> Program received signal SIGSEGV, Segmentation fault.
>> 0x08178333 in _mm_load_ps (this=0xbfffe2a0, data=0xbfffe348) at
>> /usr/lib/gcc/i486-linux-gnu/4.4.3/include/xmmintrin.h:899
>> 899       return (__m128) *(__v4sf *)__P;
>>
>> -mpreferred-stack-boundary=3 -mincoming-stack-boundary=3
>>
>> Program received signal SIGSEGV, Segmentation fault.
>> 0x08178375 in _mm_load_ps (this=0xbfffe340, data=0xbfffe3ec) at
>> /usr/lib/gcc/i486-linux-gnu/4.4.3/include/xmmintrin.h:899
>> 899       return (__m128) *(__v4sf *)__P;
>>
>> -mpreferred-stack-boundary=4 -mincoming-stack-boundary=4
>>
>> Program received signal SIGSEGV, Segmentation fault.
>> 0x08178419 in _mm_load_ps (this=0xbfffe210, data=0xbfffe2dc) at
>> /usr/lib/gcc/i486-linux-gnu/4.4.3/include/xmmintrin.h:899
>> 899       return (__m128) *(__v4sf *)__P;
>>
>>  -mpreferred-stack-boundary=5 -mincoming-stack-boundary=5
>>
>> Program received signal SIGSEGV, Segmentation fault.
>> 0x081784c3 in _mm_load_ps (this=0xbfffdea0, data=0xbfffdfdc) at
>> /usr/lib/gcc/i486-linux-gnu/4.4.3/include/xmmintrin.h:899
>> 899       return (__m128) *(__v4sf *)__P;
>>
>>  -mpreferred-stack-boundary=6 -mincoming-stack-boundary=6
>>
>> Program received signal SIGSEGV, Segmentation fault.
>> 0x08178c45 in _mm_load_ps (this=0xbfffd720, data=0xbfffd97c) at
>> /usr/lib/gcc/i486-linux-gnu/4.4.3/include/xmmintrin.h:899
>> 899       return (__m128) *(__v4sf *)__P;
>
> none of the `data' arguments is a correctly aligned pointer :/

oy veh, and I'm not even cross-compiling yet...

Dan
_______________________________________________
nova-dev mailing list
[email protected]
http://klingt.org/cgi-bin/mailman/listinfo/nova-dev
http://tim.klingt.org/nova

Reply via email to