Re: [WIP/RFC] MIPS registers overhaul

[Andrew Cagney <ac131313 at redhat dot com>]

> Careful.  If the ABI is o32, and FR == 0/..., then there should be
> > only 16 floating point registers in use.  The original MIPS 1, and
> > r5900 ABIs would both allow use of all 32 32 bit floating point
> > registers.
>
>
> I don't know that that is correct, at least about the "original MIPS
> 1" behaviour.

Sigh, yes:  B 3.3 Formatted Operand Layout of [3.2]:

FPU instructions that operate on formatted operand values specify the 
floating-point register (FPR) that holds a value. An FPR is not 
necessarily the same as a CP1 general register because an FPR is 64 bits 
wide; if this is wider than the CP1 general registers, an aligned set of 
adjacent CP1 general registers is used as the FPR. The 32-bit register 
model provides 16 FPRs specified by the even CP1 general register 
numbers. The 64-bit register model provides 32 FPRs, one per CP1 general
register. Operands that are only 32 bits wide (W and S formats), use 
only half the space in an FPR. The FPR organization and the way that 
operand data is stored in them is shown in the following figures. A 
summary of the data transfer instructions can be found in section B 6.1 
on page B-19.

The key bit being that the terminology differentaties between a 32 bit 
CP1 register and a 64 bit FPR.

This also suggests a better way of representing the registers to the user:

MIPS I:
	$cp0, $cp1, ...: 32/64 bit raw co-processor registers
MIPS I:
	$fp0, $fp2, ...: 64 bit co-processor registers
		note that these are little word ordered
		and contain 32 and 64 bit float values
		(This was suggested to me in a second hand
		off line discussion :-()
MIPS III:
	$fp0, $fp1, ...: 64 bit floating point registers

Andrew