*BSD News Article 5254


Return to BSD News archive

Path: sserve!manuel!munnari.oz.au!uunet!mcsun!uknet!doc.ic.ac.uk!kd
From: kd@sophia.doc.ic.ac.uk (Kostis Dryllerakis)
Newsgroups: comp.unix.bsd
Subject: Re: /386bsd: reject 65340
Message-ID: <19d7v2INN75q@frigate.doc.ic.ac.uk>
Date: 18 Sep 92 18:42:42 GMT
References: <3594@ra.nrl.navy.mil> <1992Sep18.085534@IASTATE.EDU> <rwa.716836288@aupair.cs.athabascau.ca>
Sender: kd@sophia (Kostis Dryllerakis)
Organization: Imperial College, London, U.K.
Lines: 43
NNTP-Posting-Host: sophia.doc.ic.ac.uk


In article <rwa.716836288@aupair.cs.athabascau.ca>, rwa@aupair.cs.athabascau.ca (Ross Alexander) writes:
> jzelle@IASTATE.EDU (James A Zelle) writes:
> >In article <3594@ra.nrl.navy.mil>, deal@hightop.nrl.navy.mil (Richard
> >Deal) writes:
> >>/386bsd: reject65340 messages.  The number is mostly 65340 but is
> >>some times 65342, 65344, 65366.  if I type
> >I'm having a similar problem with a Western Digital Ethernet card,
> >except I think the error says "reject655341" in my case.  The error
> >goes away when the Ethernet cable is disconnected, so it appears that
> >386BSD doesn't like what it sees on the network.  Any ideas?
> 
> This problem arises from having shadow memory enabled in the range of
> addresses (0xD0000 - 0xD03FFF as I remember) that the dual-port ram in
> the WD8013 card corresponds to.  Turn off any shadowing or caching in
> this region via your BIOS EAPROM setup firmware and the problem will
> go away (your mileage may vary :).
> 
	It might not be that easy. We have a NEC 386PC of some age having
the same problem. No CMOS setup is available in ROM (or that is what we
think) and no shadow RAM should appear.
	What seems curious is the fact the the WE card is recoginsed at
boot time as being in IRQ 9 instead of IRQ 2 mentioned in the documentation.
The card WD8003E is set to IRQ2 and -the funny bit- it will not accept a
jumper setting for IRQ 9.


> The errors go away when the cable is disconnected because the WD8013
> stops generating interrupts, and so the kernel stops finding bogus
> stuff in what it thinks is the WD8013 memory (but isn't, because the
> hardware is lying).
> 

Any further ideas for the problem seems that would be greatly appreciated.

Kostis

-------------------------------------------------------------------------------
K J Dryllerakis					Deparment  of  Computing
Logic Programming Group				Imperial College, LONDON
Janet: kd @uk.ac.ic.doc				Tel:  +71 589 5111 x5089
Uucp : kd@icdoc.UUCP
-------------------------------------------------------------------------------