[OpenBIOS] question for the experts on PII vs. Celeron startup

[Thomas J. Merritt]

|<><><><><> Original message from Ronald G Minnich  <><><><><>
|I have been having trouble getting PII to work with memory from powerup. I
|just now realized that I'm seeing basically the same problem on the Intel
|and SiS motherboards. Engineers at SiS, using the same linuxbios romimage
|as I am using, has no troubles on the motherboard BUT: they are using
|Celeron. (see the linuxbios web page: I can still get the intel
|motherboard up with linuxbios, but I have to let the flash recovery code
|run first by twiddling the jumper).
|
|The symptom is that during memory init, at some point the processor seems
|to take a trap during a memory write. It is as though some piece of
|hardware between the processor and memory is unhappy :-)
|
|I'm working with SiS to try a different motherboard. But: I'm now
|wondering if I'm encountering some kind of PII L2 Cache init problem on
|the Slot 1 cards I'm using here. I'm very suspicious that I'm seeing a
|particular kind of problem on my Slot 1 PII systems that SiS is not seeing
|using a Celeron, and that I'm seeing the same kind of problem on both
|Intel and SiS mainboards.
|
|Any comments on this? Any thoughts? Anyone know if anything special
|happens for L2 cache on a Slot 1 card at powerup? It would probably look
|like an IN or OUT to an undocumented address (undocumented since this is
|Intel we're talking about here).
|
|ron
|p.s. linuxbios: http://www.acl.lanl.gov/linuxbios

Ron, in going through the SuperMicro P6DGE BIOS they definitely write a number
of reserved bits and in the 440GX chip.  It is hard to know exactly what
needs to be done to get the board to talk to memory.  What we have
observed is that the processor doesn't trap, but that it doesn't read
back from memory what was written.  After we think we have memory
turned on we do a 

	push	%eax
	pop	%ebx
	cmp	%eax,%ebx
1:
	bne	1b

The PII then spins.  I've yet to find anyone at Intel that will give me
the time of day.  SuperMicro was initially helpful but they have returned
my most recent calls.

TJ Merritt
tjm at codegen.com
1-415-834-9111
-
To unsubscribe: send mail to majordomo at freiburg.linux.de
with 'unsubscribe openbios' in the body of the message