home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.os2.misc
- Path: sparky!uunet!mcsun!Germany.EU.net!olymp!rs1.thch.uni-bonn.de!greve
- From: greve@rs1.thch.uni-bonn.de (Thomas Greve)
- Subject: Re: "DOS from Drive A" fails with "Illegal Inst
- Message-ID: <1992Nov19.150634.9183@olymp.informatik.uni-bonn.de>
- Sender: usenet@olymp.informatik.uni-bonn.de
- Organization: University of Bonn, phys. Chemistry
- References: <1dpubcINNh9u@myall.awadi.com.au> <1dsug0INNmjf@myall.awadi.com.au>
- Date: Thu, 19 Nov 1992 15:06:34 GMT
- Lines: 60
-
- In article <1dsug0INNmjf@myall.awadi.com.au> dhichens@awadi.com.au writes:
- >> Has anybody come across the problem of "DOS from Drive A"
- >> failing with an "Illegal Instruction" error? The problem occured both
- >> before and after installing the service pack.
-
- > The solution was to set the `HW_ROM_TO_RAM' setting under `DOS
- > settings' to on and reduce the `DOS_RMSIZE' setting to 624K to make
- > room for the ROM mapping. The ROM BIOS settings were then accessed
- > from (much faster) RAM which eliminated the problem. It was suggested
- > that I do this for all DOS sessions even though I have not had this
- > problem with normal DOS sessions.
- Hm. This seems to be a bit, say, misleading to me. I also had (yes,
- past tense :-) this sys3176 error, and even worse, a total crash of
- the system ('... internal error ... contact IBM ...').
-
- But my BIOS ROM is shadowed by chipset on bootup (so no need for
- HW_RAM_to_ROM) and it always passes the power on self test (with
- checksum!). So anyone with a clone AT should be able to leave
- HW_RAM_to_ROM disabled, as he can enable shadowing by the chip set
- (you probably can't on true blue PS's ;-).
-
- This was due to the following:
- o i had rem'd out the vdpmi.sys in config.sys -- i don't run
- windoze
- o i had set _dpmi_ to _auto_
-
- Either re-enabling vdpmi.sys or setting _dpmi_ to _disabled_ improved
- the situation from total crash (with following hard boot and chkdsk --
- *this* is annoying...) to `sys3176'. But no working DOS session yet.
-
- Another interesting thing was the error, that occurred when i tried a
- `DOS from drive A': without fsfilter: no problem, with fsfilter:
- freeze of the desktop, hard boot required.
-
- Interesting, because the image of this very disk created by VMDisk
- booted happily. The problem was: `DOS from drive A' *must* have
- INT_during_IO *disabled*! You'll (this is just guessing) run into bad
- old reentrance problems otherwise -- as you have a bad old DOS
- running...
-
- A DOS session with all memory features (EMS, XMS, DPMI, DOS hi, UMB)
- disabled was the first one that worked reliably. If this does not work
- for you, i might have a look on my other DOS_Settings, as i've changed
- virtually all of them...
-
- (From memory:
- o HW_Timer on
- o FastPaste on
- o Refresh_Emulation off
- o no memory enabled/disabled [except mine enabled ;-]
- o ...)
-
- Now i have a DOS_from_A session running nicely (with HIMEM -- take
- that one from \os2\mdos! -- and fsfilter) and a NW lite server running
- in it. A box in a box ;-).
- --
- - Thomas
-
- greve@rs1.thch.uni-bonn.de
- unt145@dbnrhrz1
-