Trap, Trap, Trap

Could it be that the Ide controller and Sound Card are using the
same DMA channel?  How should these be set/adjusted?

You should press ALT-F2 during boot, when the white box shows up on the upper left corner (quite early). Then Warp will show you which driver it's currently loading. This may give you an idea which driver causes your installation to trap.

Hello Figueiredo!

Your answer is very helpful, but what bothers me is that the
errors often occur when I am running only IBM Warp4 and no apps.  
Are parts of Warp considered still under development?

I think I may have discovered the answer to some of the
problems.

First, the AWARD BIOS allows the user to configure the HD
parameters in one of FOUR ways.  I changed the access setting
from NORMAL to LBA mode.  Don't ask me how the system still
access the drive correctly but it does.

Second, I pulled out the network, modem, and Windows Sound
System Cards.

Third, I removed the 2 8 meg simms.

Fourth, and I really think this is where the problem lies, I
installed a zipfile I found on IBM Testcase called goodide.zip.  
This file contains a modified IBM1S506.ADD file and another
modified file that I forget.  After installing this, the system
became much more stable.

The only problem so far is that the system sometimes hangs
(cursor locks up, screen doesn't repaint properly always,
problems with PMMERGE.DLL) WITHOUT a trap message.

I reinstalled the NetWork Card w/o any noticeable instability.

I reinstalled the 8 meg simms, which seemed to make the system a
bit less stable, but NO TRAPS.

I reinstalled the Sound Card and the system locked up (frozen
cursor) within 5 minutes.  Here I also noticed TWO lines in the
config.sys for 2 different forms of Business Audio SYS drivers,
so I commented out one line and rebooted.  Much more stable than
at first,

HOWEVER, I had put a special line in config.sys to dump trap
errors and I got a TRAP dump involving InJoy dialler which was
open and doing an FTP through Netscape (6/97 release) at the
time.  I can't include the dump since I am at home now, but it
seems that INJoy tried to access some space on the C: (FAT)
drive (OS/2 is on J:, HPFS) that shouldn't have been accessed.  
Would this be a problem with the code in InJoy (1.1)?

The trap info is pretty nice; do you have the skill to read and
analyze it?  I sure don't!

I appreciate your time and insight.  I think you should follow
up on the IBM1S506.ADD driver.  If IBM is working on it, that
indicates there is probably a real problem with it in some
configurations.  The driver now properly reports in verbose mode
which it DID NOT do previously, even though I had the command
line set properly.

One more question, is there a keystroke that one can use to
PAUSE the bootup screen output, or is that info recorded in a
logfile?

Thanks again,

mark bruffey
mbruffey@voicenet.com

That Value s/b 200 not 20 points.

Hi,
it seems you managed to sort out your problem.

I would like to make some comments on two points.

1. RAM. Be carefull with different types of RAM in the same machine. Many problems have been traced to RAMs with different speed or from different manufacturers.

2. Hard Disk Partition Setting. In fact OS/2 seems to have the capability to access a Disk even when the BIOS LBA/NORMAL/LARGE has been changed. I would say that during installation or bootup it uses BIOS settings but after it uses an autonomous mode. This may lead to very bad surprises in the future, namely when running FDISK. My way of dealing with these questions has been to rely on PartitionMagic (FDISK replacement) as a checkup. There has been times when PartitioMagic reports erros in the partition table of some disk and informs that "this is a know BUG of FDISK".

About your other questions I have never seen software without bugs, and sure OS/2 is no exception. The key is to make incremental changes/updates to keep track of things.

Regards.

Thanks again.  I ripped the HD out of that
sytem and put it in a Zenith p200 system.  
To present it seems to run fine, so as you
suggest, I suspect HWare on the MB/RAM  
too!

The error #0160:fff54c8c  Internal Processing Error is from a piece of OS/2 code, calling the function:  _IOGetDiskSpaceInfo + 0x58.   Which would deal as you suspected with your hard drive or disk-space.  (I have a trap analyzer...)

To find out the culprit, add these two lines to your CONFIG.SYS after you make a backup of it:
SET SHELLEXCEPTIONHANDLER=OFF
SET SHAPIEXCEPTIONHANDLER=OFF

These two lines will turn off the OS/2 exception handlers and when a bad driver or piece of code goes amuck, it will cause OS/2 to grab as much data as it can and throw it on the screen.

The two areas you should be most concerned with will be the
CS:EIP and CSLIM codes.  This will tell us where the error occurs in OS/2 Code.

TRAP C = Stack Fault, usually Ring 3.
TRAP D = General Protection Fault, usually failure of a driver in memory causing this. Usually accompanied by TRAP E:
TRAP E = Page Fault, Memory not being handled properly.
If TRAP E and TRAP D happen at the same time, a TRAP 8 usually occurs.  (This is usually due to a network driver failing to bind.

TRAP 6 = Illegal or Invalid OpCode, The processor tried to execute an unreserved invalid opcode, contact software support. (That's the official reason), and that was already answered for you earlier.

(Also, I have a listing of the 256 TRAPS if it will help you, let me know.)  

Used to work at IBM in Austin.

Thanks, Azazeel.  I would love to have the list of trap errors as you offered.  Could you email it to me as a *.zip file?

Thanks,

Mark
mbruffey@voicenet.com