Subject: Re: NetBSD 1.5.2 on IPX boot problems...
To: William Barnett-Lewis <wlewis@mailbag.com>
From: Rob Healey <rhealey@norstar.com>
List: port-sparc
Date: 01/25/2002 18:18:50
> Rick Kelly wrote:
> > 
> > William Barnett-Lewis said:
> > 
> > >The problem is that some of the old IPX's and SS2's had bad cache chips
> > >in the chip set. The only way to tell is if you get an error msg on
> > >booting NetBSD. I was able to get my IPX to use the Weitek by swapping
> > >in a good cache chip from an otherwise dead IPX. It's really a trial and
> > >error kind of thing.
> > 
> > Do you mean this line in dmesg:
> > 
> > cpu0 at mainbus0: cache chip bug; trap page uncached: CY7C601 @ 40 MHz,
> > TMS390C602A FPU 
> 
> That's the one. Bad cache chip that will not play nice with Weitek +
> NetBSD. Kernal panics every time. Took me forever to track down why and
> it didn't help that Mr. OpenBSD didn't believe me because his IPX worked
> right... 
> 
	Ummm, my weitek's show up with the above warning, different CPU
	names though, and work fine as headless servers banging on the
	disks, use as a scanner station and misc SCSI peripherals. One even
	gateways to a private net via a second network card.

	What else is needed to cause the problem? My autoconf line is:

	cpu0 at mainbus0: cache chip bug; trap page uncached: W8601/8701 or 
						MB86903 @ 40 MHz, on-chip FPU
	cpu0: 64K byte write-through, 32 bytes/line, hw flush: cache enabled

	These upgrades probably aren't worth ALOT of money since the just
	clock double the integer unit. Money would probably be better
	spent getting an SS5, SS20, Ultra1 or other SPARC on e-bay...

	-Rob