ackerman@sadye.uvm.edu (Steve Ackerman) (07/13/90)
We're having a big problem with our DG AViiON 300, and we're hoping someone may have had a similar experience. We're attempting to write a stand-alone program that runs without the DG/UX kernel. Three weeks ago, while writing/debugging the exception handler, our program bombed hard a few times. By hard, I mean the machine needed to be power-cycled, and upon reboot, some of the configuration parameters were trashed. The amount of memory the machine thought it had was off by 1 mb; the boot path was destroyed (empty), and the monitor prompt was back to the default. One time, the machine crashed and never came back up. The LED power-on indicator which comes on when all beginning tests pass never came on, and we never got any messages or SCM prompt. We did, however, get a FATAL MEMORY ERROR message on the console when we intentionally popped out memory module 0, so we knew the console baud rate settings (as well as the POWER supply) were fine. We took this to be just a coincidence that the Machine broke while we were testing our program. We sent it back to be repaired. They (DG) replaced everything in the unit, and returned it. We got it back this Monday. Last night, while working on the same code, the machine broke again. Once we could take as coincidence, but twice? We weren't doing anything device specific -- just trying to set up our own (basic) exception handler. We've called DG, but they said send it back. I would hate to think that our software could damage the DG so much as to require everything to be replaced. I think that our problem could be as simple as screwed up configuration settings (somewhere). Would anyone have a clue to this problem? Thanks in advance! Steve Ackerman ackerman@uvm.edu -- Steven Ackerman (ackerman@uvm.edu)
jkenton@pinocchio.encore.com (Jeff Kenton) (07/13/90)
From article <1567@uvm-gen.UUCP>, by ackerman@sadye.uvm.edu (Steve Ackerman): > > We're having a big problem with our DG AViiON 300, and we're > hoping someone may have had a similar experience. We're attempting to > write a stand-alone program that runs without the DG/UX kernel. > > . . . > > They (DG) replaced everything in the unit, and returned it. > We got it back this Monday. Last night, while working on the same > code, the machine broke again. Once we could take as coincidence, > but twice? We weren't doing anything device specific -- just trying > to set up our own (basic) exception handler. > Exception handlers for the 88000 are easy to get wrong -- there are so many more possible ways to do it wrong than right -- and a broken one could do almost anything. Could what you are seeing be the result of software that does "almost anything"? Otherwise, it's the hardware. But it sounds from your note that the software is at least triggering something wierd. I hope you have a good logic analyzer handy. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - jeff kenton --- temporarily at jkenton@pinocchio.encore.com --- always at (617) 894-4508 --- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -