Re: Sun Blade 2500 won't come up at all - "Last Trap: Corrected ECC Error"

[Eduardo Horvath <eeh%NetBSD.org@localhost>]

On Mon, 4 Dec 2017, Greg Earle wrote:

> Sorry for the sorta-O/T ...
> 
> My Sun Blade 2500 (Silver) NetBSD test box was turned off for a bit.  Just
> tried to fire it up again but it croaks as soon as I turn it on now:
> 
> ERROR: /pci@1e,600000/isa@7/serial@0,1016: Last Trap: Corrected ECC Error
> 
> (I'm using the 9-pin serial port console at 57,600 baud.)
> 
> Pretty much every PROM command elicits this same message.  Bad DIMM(s)?
> 
> I've got 8 x 1 GB DIMMs in it.  Pretty sure they're all identical.
> 
> Do I just yank them all out and start with a pair in slots 2 and 3 to see
> what's wrong?  It's been so long since I touched anything inside that as soon
> as I tried to take the DIMM fan assembly off, both tabs snapped  :-/

Maybe....

It's been a while, but ISTR OBP does not enable the correctable ECC trap.  
So do you have auto-boot? enabled?

My theory (which may be completely wrong) is that you have auto-boot? 
enabled, so OBP loads the kernel.  The kernel installs a bunch of code, 
including debug vectors in OBP, and turns on the ECC trap.

So try sending a break early to stop the autoboot sequence.  If you can do 
that and get the ok prompt to work properly you can use OBP to dump out 
interesting register contents and pin down the bad hardware.

Otherwise, yes, start swapping DIMMs.  (Although it might not be the 
DIMMs....)

Eduardo