This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 1/2] BZ#13743: PowerPC - Add a new header for platformspecific functions
- From: David Miller <davem at davemloft dot net>
- To: munroesj at us dot ibm dot com, munroesj at linux dot vnet dot ibm dot com
- Cc: carlos at systemhalted dot org, thomas at codesourcery dot com, tuliom at linux dot vnet dot ibm dot com, roland at hack dot frob dot com, joseph at codesourcery dot com, ryan dot arnold at gmail dot com, libc-alpha at sourceware dot org, sjmunroe at us dot ibm dot com, drepper at gmail dot com
- Date: Thu, 08 Mar 2012 15:03:32 -0800 (PST)
- Subject: Re: [PATCH 1/2] BZ#13743: PowerPC - Add a new header for platformspecific functions
- References: <1331225818.4861.149.camel@spokane1.rchland.ibm.com><20120308.133023.1654649244698937759.davem@davemloft.net><1331247219.4861.170.camel@spokane1.rchland.ibm.com>
From: Steven Munroe <munroesj@linux.vnet.ibm.com>
Date: Thu, 08 Mar 2012 16:53:38 -0600
> On Thu, 2012-03-08 at 13:30 -0800, David Miller wrote:
>> From: Steven Munroe <munroesj@linux.vnet.ibm.com>
>> Date: Thu, 08 Mar 2012 10:56:58 -0600
>>
>> > On Thu, 2012-03-08 at 11:23 -0500, Carlos O'Donell wrote:
>> >> What stops GCC from installing a header?
>> >>
>> > Well I would have to ask. But the facts are that the existing headers
>> > ppc_asm.h and ppu_intrinsics.h seem to be primarily intended for
>> > internal gcc use, are not published in /usr/include, and are not
>> > documented.
>>
>> We add platform specific headers for GCC to export all the time.
>>
>> x86 has all of the MMX/SSE intrinsic interfaces in header files
>> which GCC exports, and similarly Sparc has a "visintrin.h" header
>> for application access to VIS instructions.
>>
> Yes but is it documented anywhere?
Intel's MMX and SSE intrinsic interfaces are documented in their
own documentation, GCC simply followed the lead.
The Sparc stuff I implemented on my own because Sun's VIS ABI's were
done sloppily, without using vector types, and thus couldn't be
followed strictly by GCC.
But all of this is an irrelevant discussion.
If the documentation is important TO YOU, I'm sure YOU will add it.