[comp.sys.ibm.pc] Is Bios shadowing and multiple 16-bit cards a problem?

allred@ut-emx.UUCP (Kevin L. Allred) (07/12/89)

I recall reading some where about confilicts between the BIOS's of two
16-bit cards when using RAM shadowing.  Since I'm getting a system the
supports BIOS shadowing, and I may get a 16 bit VGA card and a 16 bit
HD controller, I want to make sure I won't end up causing myself
grief.  Can anyone tell me whether there might be a problem or not?

-- 

	Kevin Allred
	allred@emx.cc.utexas.edu
	allred@ut-emx.UUCP

chasm@attctc.DALLAS.TX.US (Charles Marslett) (07/13/89)

In article <15099@ut-emx.UUCP>, allred@ut-emx.UUCP (Kevin L. Allred) writes:
> I recall reading some where about confilicts between the BIOS's of two
> 16-bit cards when using RAM shadowing.  Since I'm getting a system the
> supports BIOS shadowing, and I may get a 16 bit VGA card and a 16 bit
> HD controller, I want to make sure I won't end up causing myself
> grief.  Can anyone tell me whether there might be a problem or not?

The shadowing of the BIOS actually helps if it is done right, because the
real problem arises when a 16-bit access is made to the BIOS of one card
(with an 8-bit BIOS interface) and the other card jumps on the bus and
tells the motherboard hardware that the addressed card is a 16-bit card
(this is a lie, understandable when you look at the bus timing to do real
16-bit transfers at bus speeds greater than 6 MHz, but still a lie).

If the shadowing is done by copying one byte at a time on boot up, all cards
should be compatible with each other (except for the really broken ones ;^).
If shadowning is done by copying one word at a time (oops!), it may make no
difference, or it may exacerbate the problem.

To avoid the problem entirely, just make sure that all the cards you use have
either an 8-bit BIOS or a 16-bit BIOS, then it doesn't matter if the addresses
get confused (when the 16-bit transfer line is yanked, that is).  Here you
may have a problem though -- Tseng based VGAs look like they have 16-bit
BIOSes, but they don't.  STB RapidMeg EMS cards can run in 8-bit, 16-bit,
or dynamic modes (depending on the mother board implementation and other
cards resident in the C0000-DFFFF area).  Video 7 VRAM cards usually have
16-bit BIOS access, but on Everex boxes they run "fast" 8-bit accesses
and occasionally crash the floppy (figure that one out!).  Some 16-bit
disk controllers use a BIOS, some don't, etc.  The best bet is to make sure you can
return anything you buy, and have a viable 8-bit and 16-bit version of
each.  Mix and match (you may strike it lucky, or the BIOS shadowing code
may be done right) and it will work the first time!

Charles Marslett
> -- 
> 
> 	Kevin Allred
> 	allred@emx.cc.utexas.edu
> 	allred@ut-emx.UUCP


===========================================================================
Charles Marslett
STB Systems, Inc.  <== Apply all standard disclaimers
Wordmark Systems   <== No disclaimers required -- that's just me
chasm@killer.dallas.tx.us

jdm@hodge.UUCP (jdm) (07/19/89)

In article <15099@ut-emx.UUCP>, allred@ut-emx.UUCP (Kevin L. Allred) writes:
> I recall reading some where about confilicts between the BIOS's of two
> 16-bit cards when using RAM shadowing.


	I too am looking into 386 motherboards and have been asking 386
	motherboard retailers about this very thing.  It seems that when the
	problem was first discovered it happened on a wide variety of
	motherboards, not just on one make or model.  Very quickly it was
	discovered that all these motherboards with the flakey Shadow RAM
	had one thing in common--the Phoenix BIOS.  It seems the Shadow
	RAM is a thing implemented in the BIOS of the 386 (hence the reason
	why some 386 boards do not have Shadow RAM).  Phoenix's version
	of Shadow RAM get very flakey when you have 2 or more 16-bit cards
	in your system each with their own BIOS (i.e. VGA and drive card).

	Stick to Award or Quatel BIOS.  They are the only ones I have yet
	found that work 100% with Novell.  I've fixed many a problems by
	replacing the AMI or Phoenix BIOS with Award.


-- 

"I'm an anthropologist, not a computer systems architect, damit!"

jdm@hodge.cts.com [uunet zardoz]!hodge!jdm

James D. Murray, Ethnounixologist
Hodge Computer Research Corporation
1588 North Batavia Street 
Orange, California 92667  USA

TEL: (714) 998-7750	Ask for James
FAX: (714) 921-8038	Wait for the carrier