SCOUG-HELP Mailing List Archives

<< Previous Message << >> Next Message >>

Date:	Tue, 30 Jul 2002 12:15:37 PST7
From:	"Steven Levine" <steve53@earthlink.net >
Reply-To:	scoug-help@scoug.com
To:	scoug-help@scoug.com
Subject:	SCOUG-Help: Re: Elusive ECS stability: Why I Can't Salute the Kernel

=====================================================
If you are responding to someone asking for help who
may not be a member of this list, be sure to use the
REPLY TO ALL feature of your email program.
=====================================================

In <3D46FB4B.BF0D8A41@pacbell.net>, on 07/30/02
at 11:47 AM, "J. R. Fox" said:

>the ECS.nl newsgroup. The problem is, when it comes to the "addressing",
>I'm not sure I can recognize the wheat from the chaff. I recall seeing

That not entirely unexpected. :-)

>someone like Scott G. or Irv S. -- in another context -- write something
>like "If CSLIM=ffffffff, then the whole trap screen is probably useless."

Not at all. It just means the the trap did not occur within a 16-bit
driver. The cslim is a handy way of determining which driver trapped,
when it works.

When cslim is all ones, you have to use different methods to determine
which code is trapping.

>conclude with "The System detected an internal processing error at
>location ##0168:fff1eb1c-000e:cb1c." Then, below this, "60000,9084."
>And below that: "07860683." And of course, the ever popular "The System
>is stopped." The ##0168 is always there, as is the 60000,9084 and the
>07860683. The ":fff1ebc" is the one that turns up over and over again in
>the newsgroup threads, and was the Trap-E I wrote down from Kernel
>14_088.

These numbers are the least useful for trap analysis. cs:eip and ss:esp
are much more important. The numbers are your reporting are for where the
kernel halted and are essentially generic. The kernel halts at the same
place for a wide variety of traps.

Newer kernels decode the trap address to a module name when possible. The
reported OSKERNL is a module name. Unfortunately, this does not provide
any additional information.

>I also wrote down one full and one partial Trap-E screen from 14_089e.

Since you not writing down the correct values, why don't you do this.
Get:

http://home.earthlink.net/~steve53/os2diags/DumpTrapScreen.zip

and adjust config.sys as directed. Then use the provided tool extract the
trap screens for you. Saves a lot of writing.

>The first one, from before I disabled C-A-D Cmdr., listed the internal
>proc. error at location ##0168:f18830d1, with 1188, 7023 under that, and
>the constant 07860683 under that. The later Trap, with the same kernel,
>*after* I had disabled C-A-D Cmdr., shows the error at location
>0168:fff1eb73-000e:cb73. And back to 60000, 9084 with the 07860683.

>If any of the register values are key, tell me which ones to watch out
>for.

All the ones you did not list are significant to some level.

>I have the last version, 2.52, and think I picked up that driver. I'll
>double-check on the driver.

I'm pretty sure, you can't even boot without the beta driver, but since I
don't run CAD Cmdr, I'm not entirely sure how the failure shows up.

>How do you go about ditching and re-installing the VDM subsystem ?
>(Assuming this part was not altered by the fixpack.) Selective Install
>covers this ?

No. Uninstall the SDD drivers, using the supplied tool. This should
revert you to VGAGRADD which is the default for MCP/eCS installs. Test
with this. If it still traps, run \os2\setvga.cmd. This should revert to
pure VGA. Test with this. If it still traps, I would do a Selective
Install of DOS VDM support and reapply the MCP2 FP.

>And I'll double-check that. I did scan Config.Sys for anything with
>"kbd" in it. My recollection was that PC replaced several standard
>system files with its own version, whereas C-A-D did less of that.

PC only replaced the keyboard driver. It overrode doscalls.dll.

>We're talking about KBDBASE.SYS here, aren't we ? I'm not even sure

Yes. That's the base keyboard driver that PC replaces. I've always
assumed the CAD does the same. If that's not the case, you'll have to
supply me with the details.

>whether or not C-A-D is supposed to be compatible with eCS, but it has
>been commented out of Config.Sys for several days now.

Commenting out the CAD entries will not restore the IBM driver since
KBDBASE.SYS does not have a CONFIG.SYS entry.

Steven

--
---------------------------------------------------------------------
"Steven Levine" MR2/ICE 2.31a #10183 Warp4/FP15/14.085_W4
www.scoug.com irc.webbnet.org #scoug (Wed 7pm PST)
---------------------------------------------------------------------

=====================================================

To unsubscribe from this list, send an email message
to "steward@scoug.com". In the body of the message,
put the command "unsubscribe scoug-help".

For problems, contact the list owner at
"rollin@scoug.com".

=====================================================

Return to [ 30 | July | 2002 ]

SCOUG, Warp Expo West, and Warpfest are trademarks of the Southern California OS/2 User Group. OS/2, Workplace Shell, and IBM are registered trademarks of International Business Machines Corporation. All other trademarks remain the property of their respective owners.