Question about soft-fp

[Ralf Baechle]

On Mon, Jan 10, 2000 at 03:37:43PM +0100, Andreas Jaeger wrote:

> Since the soft-fp stuff is now in, how can I use it?  Which ports do
> benefit from it?  I'm interested in using the long double support for
> Mips.  Is this possible?
> 
> What is still missing with soft-fp?
> 
> Could anybody enlighten me, please?

The MIPS ABI defines:

	float		32-bit IEEE single precission
	double		64-bit IEEE double precission
	long double	64-bit IEEE double precission

I think having optional higher precission would be nice to have but it would
have to be disabled unless explicitly enabled by some special compiler
option.

As an aside, the RISC/OS and IRIX kernels have the necessary hooks for
triple and quad precission fp in their fp support code.  Seems that
there was no sufficient interest in actually implementing that.

  Ralf