This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 13/14] [x86_64] Vector math functions (sincos and tests)
- From: "Tulio Magno Quites Machado Filho" <tuliom at linux dot vnet dot ibm dot com>
- To: Andrew Senkevich <andrew dot n dot senkevich at gmail dot com>
- Cc: Joseph Myers <joseph at codesourcery dot com>, libc-alpha <libc-alpha at sourceware dot org>, sjmunroe at us dot ibm dot com, Anton Blanchard <anton at samba dot org>
- Cc:
- Date: Fri, 19 Jun 2015 13:21:04 -0300
- Subject: Re: [PATCH 13/14] [x86_64] Vector math functions (sincos and tests)
- Authentication-results: sourceware.org; auth=none
- References: <CAMXFM3t4CRSW2pRELmJMU5bE0eoYfUx7pO2-iLqLiGvgE79d_A at mail dot gmail dot com> <20150619075546 dot 4284b164 at kryten> <CAMXFM3srg-a6+MJQ7btm-wMVUBazEif53EyHbnpv=+RCgAocvA at mail dot gmail dot com> <alpine dot DEB dot 2 dot 10 dot 1506191124350 dot 24907 at digraph dot polyomino dot org dot uk> <CAMXFM3tzxuVj06vQwWWc4JBXd8MAtLJYOia3mpqViY7w6Bt08g at mail dot gmail dot com> <87vbejhlfr dot fsf at totoro dot br dot ibm dot com> <CAMXFM3sKK_0DKx8hucptOHT6x0sRCkJxNDA+vXCRCjgvGaA55w at mail dot gmail dot com>
Andrew Senkevich <andrew.n.senkevich@gmail.com> writes:
> 2015-06-19 16:33 GMT+03:00 Tulio Magno Quites Machado Filho
> <tuliom@linux.vnet.ibm.com>:
>> Andrew Senkevich <andrew.n.senkevich@gmail.com> writes:
>>
>>> 2015-06-19 14:25 GMT+03:00 Joseph Myers <joseph@codesourcery.com>:
>>>> On Fri, 19 Jun 2015, Andrew Senkevich wrote:
>>>>
>>>>> Proposal is fix it so:
>>>>>
>>>>> diff --git a/bits/libm-simd-decl-stubs.h b/bits/libm-simd-decl-stubs.h
>>>>> index ec1fa69..6d0558a 100644
>>>>> --- a/bits/libm-simd-decl-stubs.h
>>>>> +++ b/bits/libm-simd-decl-stubs.h
>>>>> @@ -45,6 +45,10 @@
>>>>> #define __DECL_SIMD_sincosf
>>>>> #define __DECL_SIMD_sincosl
>>>>>
>>>>> +/* This is needed because of definition of sincos
>>>>> + in sysdeps/ieee754/ldbl-opt/s_sin.c. */
>>>>> +# define __DECL_SIMD_sincos_disable
>>>>> +
>>>>
>>>> It would seem better to me to put this in the .c file that needs it,
>>>> rather than in an installed header - it should only be relevant when
>>>> building glibc, not when using the installed library.
>>>
>>> diff --git a/ChangeLog b/ChangeLog
>>> index 7dedfb4..75f926c 100644
>>> --- a/ChangeLog
>>> +++ b/ChangeLog
>>> @@ -1,3 +1,8 @@
>>> +2015-06-19 Andrew Senkevich <andrew.senkevich@intel.com>
>>> +
>>> + * sysdeps/ieee754/ldbl-opt/s_sin.c (__DECL_SIMD_sincos_disable): Added
>>> + empty definition for proper unfolding of __MATHDECL_VEC.
>>> +
>>> 2015-06-19 Stefan Liebler <stli@linux.vnet.ibm.com>
>>>
>>> * benchtests/bench-strcoll.c:
>>> diff --git a/sysdeps/ieee754/ldbl-opt/s_sin.c b/sysdeps/ieee754/ldbl-opt/s_sin.c
>>> index a11d5a3..884bc51 100644
>>> --- a/sysdeps/ieee754/ldbl-opt/s_sin.c
>>> +++ b/sysdeps/ieee754/ldbl-opt/s_sin.c
>>> @@ -1,5 +1,7 @@
>>> /* dbl-64/s_sin.c uses NAN and sincos identifiers internally. */
>>> #define sincos sincos_disable
>>> +/* Stab for proper __MATHDECL_VEC unfolding. */
>>> +#define __DECL_SIMD_sincos_disable
>>> #include <math_ldbl_opt.h>
>>> #undef NAN
>>> #undef sincos
>>>
>>> Ok?
>>
>> It doesn't fix the problem.
>
> Could you please explain why with more details (exact compilation
> command caused fail will the best help)?
Sure!
All the log is available in the community BuildBot at
http://glibc-build.hack.frob.com/waterfall
The log of the build is here:
http://130.211.48.148:8080/builders/glibc-power8-linux/builds/0/steps/annotate/logs/stdio
--
Tulio Magno