🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: comp.bugs.sys5
Article #524

MC68020 assembler bug

#524
From: da
Date: Tue, 07 Apr 1987 18:58
24 lines
1439 bytes
Index: as Motorola 68020 SYSV 

Description:
	System V, R2 & R3 , Motorola 68020  assembler

	Shortsdi() in sgs/as/common/expand1.c or sgs/as/m32/expand1.c
	(wherever you find expand1.c!)
	will return a random value once enough
	span-dependent-instructions are seen.  The message
	     Table overflow: some optimizations lost (SDIs)
	appears when the limit is reached.  Improper assembly is then
	*possible*.

Repeat-By:
	(Too hard to do without modifying assembler to force bug!)
	Temporarily modify expand1.c to return S_SDI always.
	Bsr to extern is assembled wrong as 16 bit offset.
	The assembler only complains if the target is local.
Fix:
	Initialize the local variable sditype to L_SDI.
_________________________________________________________________
David Anderson                ucbvax!mtxinu!sun!hplabs!fortune!da 
Fortune Systems, Belmont California (415)598-4372 hpda!fortune!da
                                                     
Technical Details
Message-ID: <265@fortune.UUCP>
Path: arc1-ca.pugleaf.net!utzoo!utgpu!water!watnot!watmath!clyde!cbatt!ihnp4!fortune!da

Show Full Headers
Complete Article Headers
Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: arc1-ca.pugleaf.net!utzoo!utgpu!water!watnot!watmath!clyde!cbatt!ihnp4!fortune!da
From: da@fortune.UUCP
Newsgroups: comp.bugs.sys5
Subject: MC68020 assembler bug
Message-ID: <265@fortune.UUCP>
Date: Tue, 7-Apr-87 13:58:28 EST
Article-I.D.: fortune.265
Posted: Tue Apr  7 13:58:28 1987
Date-Received: Sat, 11-Apr-87 03:43:52 EST
Organization: Fortune Systems Corp., Belmont CA
Lines: 24
Keywords: MC68020 SysV as