73 lines
3.0 KiB
Plaintext
73 lines
3.0 KiB
Plaintext
|
============================================================================
|
||
|
|
|
||
|
|
Date: Tue, 30 Jan 96 20:16:56 EET
|
||
|
|
From: kriton!kyrimis@achilles.noc.ntua.gr (Kriton Kyrimis)
|
||
|
|
Subject: old enforcer hit problem fixed!
|
||
|
|
|
||
|
|
Some time ago I reported that I was getting enforcer hits on ramlib whenever
|
||
|
|
starting an ixemul client. Fred was not able to reproduce this problem using
|
||
|
|
a software configuration similar to mine.
|
||
|
|
|
||
|
|
It turns out that this is a known problem for the Fusion-Forty accelerator.
|
||
|
|
The fix is to upgrade the Fusion-Forty's MoveMH utility to version 1.2. I just
|
||
|
|
upgraded, and the problem has disappeared.
|
||
|
|
--
|
||
|
|
Kriton (UUCP: pythia!theseas!kriton!kyrimis)
|
||
|
|
(INTERNET: kyrimis@theseas.softlab.ece.ntua.gr)
|
||
|
|
(WWW: http://www.hpcl.cti.gr/~kyrimis
|
||
|
|
-----
|
||
|
|
"You don't just get up in the morning, eat your cornflakes, and go out and
|
||
|
|
change history."
|
||
|
|
-----
|
||
|
|
|
||
|
|
============================================================================
|
||
|
|
|
||
|
|
Date: Thu, 11 Apr 1996 17:21:32 +0100 (BST)
|
||
|
|
From: Lars Hecking <lhecking@nmrc.ucc.ie>
|
||
|
|
Subject: ixemul 43.0 Makefile.in and trouble-shooting entry
|
||
|
|
|
||
|
|
Hardware: A3000/25 Rev. 9.03, 37.175 ROM, 2+12MB, rev. 04 SCSI chip
|
||
|
|
Quantum LPS240S (id 0), IBM DPES 31080 SCSI-II (id 1)
|
||
|
|
Problem: After installing the new IBM drive, ixemul clients cause serious
|
||
|
|
trouble and don't work properly. Symptoms: gcc chokes on
|
||
|
|
clean, compilable input files with unbalanced endif and similar
|
||
|
|
errors. Diff "finds" differences between identical files. Tar barfs
|
||
|
|
on intact archives.
|
||
|
|
Possible explanation: ixemul.library uses Disable()/Enable() excessively.
|
||
|
|
This may lead to a situation where the SCSI driver is slowed down
|
||
|
|
and the drive has to reselect more often. Reselect is the main
|
||
|
|
trouble source of the WDC33C93-04 chip.
|
||
|
|
|
||
|
|
Solution: Disable reselect for the drive (eg. with RDBFlags from Aminet).
|
||
|
|
Better solution: upgrade to AmigaOS 3.1 ;-)
|
||
|
|
|
||
|
|
============================================================================
|
||
|
|
|
||
|
|
If you use an ixemul.library compiled for the 68000 CPU on an Amiga with an
|
||
|
|
68040 CPU and working FPU, then the IEEESPMul call from mathieeesingbas.library
|
||
|
|
can cause an EMT trap. This function is called indirectly if you multiply
|
||
|
|
two floats. The solution is provided below. Of course, it is even better
|
||
|
|
to just use the ixemul.library for an 68040 with FPU.
|
||
|
|
|
||
|
|
Date: Tue, 7 May 1996 15:27:39 -0700 (MST)
|
||
|
|
From: fnf@ninemoons.com (Fred Fish)
|
||
|
|
Subject: Re: ixemul EMT trap problem revisited
|
||
|
|
|
||
|
|
> > > crashes then I guess IEEESPMul contains a bug. I have mathieeesingbas.library
|
||
|
|
> > > version 37.3.
|
||
|
|
|
||
|
|
This triggered a faint memory of something I saw on aminet:
|
||
|
|
|
||
|
|
MathPatch.lha util/boot 3K 63+Patch for 68040 bug in "mathieeesingbas.
|
||
|
|
|
||
|
|
I installed that and it cured the problem.
|
||
|
|
|
||
|
|
-Fred
|
||
|
|
|
||
|
|
============================================================================
|
||
|
|
|
||
|
|
If you get an "Abort" requester every time you start a Geek Gadgets
|
||
|
|
program, then check for the presence of the multiuser.library on your
|
||
|
|
system and remove it. Apparently this library can cause serious problems
|
||
|
|
with ixemul if there are no volume around that actually use MuFS.
|