Re: [RFC, PATCH] ARM related deprecations

[Richard Earnshaw <rearnsha at arm dot com>]

On 28/02/12 23:42, John Tytgat wrote:
> In message <4F4D0E64.1020302@arm.com>
>           Richard Earnshaw <rearnsha@arm.com> wrote:
> 
>> On 28/02/12 17:10, Joseph S. Myers wrote:
>>> On Tue, 28 Feb 2012, Richard Earnshaw wrote:
>>>
>>>> Here's an updated patch and a suggested web-page patch:
>>>
>>> uClinux targets do not have -gnu in their target triplets.
>>>
>>> You missed the WinCE port from the list in the webpage patch of ports 
>>> being deprecated for using FPA floating-point format and lacking a modern 
>>> alternative using VFP format.  (The config.gcc patch also got the triplet 
>>> for WinCE wrong - it's arm*-wince-pe* not arm*-*-wince*.  Though ISTR the 
>>> more functional port that hasn't been contributed to FSF GCC uses some 
>>> other triplet.)
>>>
>>
>> Have I ever said how much I hate triplets?
>>
>> Here's an updated version, plus some updates to install.texi to avoid
>> references to the obsolete ports.
>>
>> OK?
> 
> As mentioned/asked in http://gcc.gnu.org/ml/gcc-patches/2011-12/msg00744.html
> our GCC/ARM port for RISC OS which is using for one of its multilib
> variants FPA (to interface with the SharedCLibrary used in RISC OS).
> Other multilib variants are soft-float and VPF.  I would like to keep
> the FPA support in gcc 4.8.

RISC OS support is not (and never has been) part of the FSF source tree.

> 
> I'm wondering if the removal of FPA is really necessary.  Does FPA support
> prevent future ARM improvements in GCC ?
> 

It's been 10 years since I last saw any real interest in the FPA and
even that was for legacy parts.  In fact I think it's been at least 10
years since FPA support was dropped from ARM's commercial compiler.

I've been through the test results for the last three years, the last
suggestion of any testing for a port that uses the FPA instructions was
in February 2009 for the legacy Linux ABI, since that time there has
only been soft-float support tested.

It might sometimes look as though carrying this dead code around has no
cost, but I can assure you that whenever we go into some of the
selection logic that supports picking between FPA, VFP and soft float we
do have to think about what the implications might be for each case.

Furthermore, just having those instructions in the compiler has a cost
in terms of the time it takes to build the compiler and, to a limited
extent, the amount of time it takes to compile an application.

Overall, I can no longer see a valid justification for carrying this
code within GCC.

R.