lray@civilgate.ce.uiuc.edu (Leland Ray) (03/22/90)
It is a weird quirk of Unix that "bus error" is returned for almost any kind of memory fault. Why this is, I've no idea, perhaps a Unix guru could enlighten us. Anyway, a bus error is 99% of the time a memory fault. You should the tb command (traceback) to obtain a stack traceback that will tell you what line your error is on. Look for all the typical problems, like array indices out of bounds, or incorrectly passing some array. You will likely find it to be a simple programming error. By the way, from time to time a heavily loaded 3500 will experience a real bus error (bus time out). I think it is a characteristic of the WD disk controller. Perhaps I should APR this in. Hmmm....
krowitz%richter@UMIX.CC.UMICH.EDU (David Krowitz) (03/23/90)
Yes! Please DO send in APR's when you find real problems with your Apollo systems! People generally assume that someone else will find and report the problem, or that it's not worth tracking down ... and the problem will persist for years! I've apparently just found a bug in the DN10000's AT-bus interrupt handler while trying to port a GPIO driver for an Ikon-10092 centronics printer interface card to the DN10K. I would have thought that after two years of software development on the DN10K that *somebody* ought to have stumbled across the same condition and would have reported it. -- David Krowitz krowitz@richter.mit.edu (18.83.0.109) krowitz%richter.mit.edu@eddie.mit.edu krowitz%richter.mit.edu@mitvma.bitnet (in order of decreasing preference)