Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!cs.utexas.edu!uunet!tektronix!orca!tekecs!frip!andrew
From: andrew@frip.wv.tek.com (Andrew Klossner)
Newsgroups: comp.arch
Subject: address masking gimmick
Message-ID: <11271@tekecs.GWD.TEK.COM>
Date: 11 Apr 89 19:36:41 GMT
References: <28200296@mcdurb>
Sender: andrew@tekecs.GWD.TEK.COM
Organization: Tektronix, Wilsonville, Oregon
Lines: 19

[]

	"Time yet again for expression of one of my favorite
	architectural "gimmicks": provide a mask that is implicitly
	ANDed with any {instruction,data} address before use -- this
	way you could place tag bits in the low bits, in the high bits,
	or in any bit you wanted (and not have to rely on features like
	"The 68000 only implements 24 address bits", etc.)."

We used this approach in implementing Smalltalk on a 68020-based
workstation.  Everything was fine until a system call in which the
process passed a tagged address to the kernel.  We looked at the work
involved in teaching the kernel to apply the mask to every user address
that it ever sees, and threw up our hands.  The compromise was to mask
the address on a page fault, but to require the application explicitly
to mask addresses passed through syscalls.

  -=- Andrew Klossner   (uunet!tektronix!orca!frip!andrew)      [UUCP]
                        (andrew%frip.wv.tek.com@relay.cs.net)   [ARPA]