This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] BZ #19590: Fixed build of shared objects that use libmvec.so functions
- From: Andrew Senkevich <andrew dot n dot senkevich at gmail dot com>
- To: "H.J. Lu" <hjl dot tools at gmail dot com>
- Cc: Joseph Myers <joseph at codesourcery dot com>, libc-alpha <libc-alpha at sourceware dot org>
- Date: Wed, 17 Feb 2016 18:22:01 +0300
- Subject: Re: [PATCH] BZ #19590: Fixed build of shared objects that use libmvec.so functions
- Authentication-results: sourceware.org; auth=none
- References: <CAMXFM3tML81iuKQMKRU-T4Fw0+=sYk0q_BNavMGagt21VcYvzQ at mail dot gmail dot com> <CAMe9rOp7BF2avdWbGTbsxuYQV_rmXSxzDUAwz+nHK-GhWqPKJA at mail dot gmail dot com> <CAMXFM3sG90jn3Em-REfaqtj3OGAjh51OcO=yn1qHqJ4aStkPrg at mail dot gmail dot com> <alpine dot DEB dot 2 dot 10 dot 1602111641250 dot 29940 at digraph dot polyomino dot org dot uk> <CAMXFM3uxH=0DHnwikPjs2AobQ0kxOqKg+j=CuOWX=RLB_i8hxg at mail dot gmail dot com> <CAMe9rOoWy+hqiCOKqFx0nOFVRT_kBETJ5hYE+cY6pFWUke=tkw at mail dot gmail dot com> <CAMXFM3vG1DNELfGaOOoUvRDGgrGWL4m3M8+5ngPg8RPnCLqUog at mail dot gmail dot com> <alpine dot DEB dot 2 dot 10 dot 1602161347240 dot 26462 at digraph dot polyomino dot org dot uk> <CAMXFM3tXQi0wRFNN1gyu-T3C7g_EWPxMfhK4wexADGtcPojNZg at mail dot gmail dot com> <CAMe9rOqP0=iC8g0SiUUQ0Yy3dC=6y6hrTMrzKPTEmGZVAsMZJw at mail dot gmail dot com> <CAMXFM3ubCPM3tcuW2hZK785fxGdZuhVRaQNLr84Y-MW+uVEZ3g at mail dot gmail dot com> <CAMe9rOosDZ3k_YLK2+v0Wy=3G5_on0-Khhk-k-R6350kJLk61A at mail dot gmail dot com>
2016-02-17 18:04 GMT+03:00 H.J. Lu <hjl.tools@gmail.com>:
> On Wed, Feb 17, 2016 at 6:37 AM, Andrew Senkevich
> <andrew.n.senkevich@gmail.com> wrote:
>> 2016-02-17 17:29 GMT+03:00 H.J. Lu <hjl.tools@gmail.com>:
>>> On Wed, Feb 17, 2016 at 6:14 AM, Andrew Senkevich
>>> <andrew.n.senkevich@gmail.com> wrote:
>>>> 2016-02-16 16:49 GMT+03:00 Joseph Myers <joseph@codesourcery.com>:
>>>>> On Tue, 16 Feb 2016, Andrew Senkevich wrote:
>>>>>
>>>>>> Here is patch with tests.
>>>>>
>>>>> This is the wrong approach for tests. Tests for this should not be
>>>>> testing implementation details about aliases, and so should not be
>>>>> creating any wrappers at all. They should be testing vectorizable calls
>>>>> to the scalar functions, compiled several times with different options
>>>>> into both executables and shared libraries.
>>>>
>>>> Please look at attached version.
>>>
>>> There may be 2 issues:
>>>
>>> 1. GCC 4.X may not support pragma simd.
>>> 2. GCC X may call those aliases.
>>>
>>> I think test-libmvec-alias-mod.c should be in assembly.
>>
>> For 1. we use -Wno-unknown-pragmas in make rules.
>
> Then those tests will be skipped.
>
>>> 2. GCC X may call those aliases.
>>
>> Do you mean GCC X may don't call those aliases?
>> Yes, but then it will use not alias version and test will be also work.
>
> Then those tests aren't effective for GCC X.
But it looks enough suitable, if compiler can't generate alias name no
needed to test that alias. Non alias name will be used.
We can add other vector functions to this test also to test all of
them, not only versions with aliases.
--
WBR,
Andrew