Re: glibc-2.1 on PowerPC cannot disable floating-point

[Geoff Keating <geoffk@ozemail.com.au>]

> Cc: Franz Sirl <Franz.Sirl-kernel@lauterbach.com>
> Date: Wed, 10 Feb 1999 18:25:52 -0500
> From: David Edelsohn <dje@watson.ibm.com>

> 	Was there a reason that the PowerPC port did not utilize the "fpu"
> structure of other ports (e.g., alpha, arm, i386, m68k, and sparc)?  Can
> the floating-point functionality be moved into an "fpu" subdirectory so
> that it follows the structure of other ports and can be disabled?

Originally, I thought that the FPU was an essential part of the
powerpc architecture, and that it would have to be emulated if it did
not exist in hardware.

A fpu subdirectory is planned for glibc 2.2.

It is particularly complicated because the GCC floating-point emulator
doesn't emulate the powerpc multiply-add instruction, so anything that
depends on that must be moved into the fpu subdirectory too.  This
amounts to basically all the powerpc math stuff.

-- 
Geoffrey Keating <geoffk@ozemail.com.au>