Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!ccplumb
From: ccplumb@watmath.UUCP
Newsgroups: comp.arch
Subject: An idea I've been kicking around
Message-ID: <12899@watmath.UUCP>
Date: Fri, 17-Apr-87 03:12:15 EST
Article-I.D.: watmath.12899
Posted: Fri Apr 17 03:12:15 1987
Date-Received: Sat, 18-Apr-87 03:18:00 EST
Reply-To: watmath!ccplumb (Colin Plumb)
Distribution: comp
Organization: U of Waterloo, Ontario
Lines: 19
Confusion: U. of Waterloo, Ontario

As many people have told me (by mail... thank you!), I just
invented the IBM 370.  Okay, I guess this proves there are
no *really* bad problems with copying the return address into
a register.

So, I'd like to ask:
Why don't people still do that?  It simplifies the architecture,
but doesn't really change anything.  The stack push is just
delayed until the called routine is entered, which begins with
the usual movem r1/r2/r3/...,-(sp).  The register is restored
again just before the ret.

Why has this been dropped?  Just for 8-bit CPU's which can't hold
the return value internally?
--
	-Colin Plumb (watmath!ccplumb)

Silly quote:
It is better to have tried and failed than never to have failed at all.