Subject: Re: ARM ELF toolchain patches
To: Todd Vierling <tv@wasabisystems.com>
From: Richard Earnshaw <rearnsha@buzzard.freeserve.co.uk>
List: port-arm26
Date: 02/21/2001 23:36:04
> On Wed, 21 Feb 2001, Ben Harris wrote:
> 
> : At the moment, we mostly follow GCC's defaults, with the only significant
> : difference being our use of VFP-format doubles rather than FPA-format.
> : This probably makes us different from Linux.  I've no idea how we differ
> : from ARM's view of the world.
> 
> Which, for the purposes of Charles's question, should mean "yes" for integer
> types.
> 
> It might be useful to bring in Richard Earnshaw for an opinion on the double
> floats.

I'm already in (I proposed it :-)

All the signs I see suggest that future ARM developments in the FP area 
are likely to be with the new VFP based co-processors.  These use a 
"natural" endianness model for doubles, rather than the somewhat perverse 
word ordering used on the FPA.

There is only one potential downside of this change and that is that 
support for hardware floating point on the FPA will be limited to single 
precision (unless you are on a big-endian machine -- which we don't have 
anyway).  However, there is only one cpu (ARM7500FE) that has a fpa as 
standard and even on that we currently make no use of the floating point 
coprocessor.

I see no good reason not to make this change, and several why we should.

R.

BTW, the current loss of -mhard-float for FPA code is a temporary issue 
due to the fact that a full patch to support the new VFP format and 
instructions will take a little bit longer to develop.  I fully intend to 
maintain support for the FPA when I do a complete set of patches for VFP 
support.