Path: utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!mailrus!ames!pasteur!agate!labrea!decwrl!sun!pitstop!sundc!seismo!uunet!steinmetz!davidsen From: davidsen@steinmetz.ge.com (William E. Davidsen Jr) Newsgroups: comp.unix.xenix Subject: Re: Xenix 386 v2.3 bug? Keywords: Xenix,Unix,80386 Message-ID: <12837@steinmetz.ge.com> Date: 19 Dec 88 23:37:27 GMT References: <40@maxx.UUCP> <8411@alice.UUCP> <401@impch.UUCP> Reply-To: davidsen@crdos1.UUCP (bill davidsen) Organization: General Electric CRD, Schenectady, NY Lines: 34 In article <401@impch.UUCP> egon@impch.UUCP (Lukas Knobloch) writes: | | Hiya | | From time to time we've got problems with our System it means: | | panic:Memory failure - parity error I may be able to help. I had a similar problem, and it only occurred when I was doing floppy disk i/o, therefore no memory test could (or would) find it. After some extensive detective work it turned out the memory failed only when the CPU and DMA channel were both hitting the same memory. I started swapping memory, to no avail. If I took out any 1/2MB it went away, but replacing the memory brought it back. I finally found that I had a slow support chip on the memory board, and when I put more than 1.5MB on it, it failed. Driving the extra load of more chips, the card just couldn't keep up. At 2.5 it was really unreliable. After trying to find the chip, I decided to just put the extra memory on another board and forget it. So far everything is running just fine, although I will probably go to faster chips when the prices come down. If the problem is not associated with floppy i/o you have some other problem, and the memory could be at fault as well as the support chips, but if the memory test runs and the o/s doesn't, it's usually a slow chip, either memory or support. Hope this helps you (or somebody), the manufacturer was ready to swear that the problem was in Xenix. -- bill davidsen (wedu@ge-crd.arpa) {uunet | philabs}!steinmetz!crdos1!davidsen "Stupidity, like virtue, is its own reward" -me