Subject: FPE & libm
To: None <port-m68k@NetBSD.ORG>
From: Ken Nakata <firstname.lastname@example.org>
Date: 02/01/1996 18:06:21
Well, this isn't easy thing to say...
I'm thinking of stripping m68k FPE down a bit; especially, I want to
pull the implemented transcendental functions out of FPE.
I'm not proud of what I've done in those functions, and quite frankly
I don't think I'll ever do any better than that, either. I can barely
handle the seminumerical stuff like floating point add/sub/mul/div,
and there's no way I'm going to beat the stability and speed of a
proven technology (= Sun's libm).
Why don't we do the same thing they do for i386 port? I.e. having the
default libm not calling most FP instructions. Of course, it's
*theoretically* slower than calculating everything in the kernel...
We could provide the user with two sets of libm, one directly using
68881 instructions, the other not (= default).
There are still some bugs in FPE, and I'd be feeling much easier and
confident if all I have to debug is something I can understand.
I appreciate any comments.