Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!not-for-mail From: David Henshaw <dhenshaw@silas.cc.monash.edu.au> Newsgroups: comp.unix.bsd.freebsd.misc Subject: Re: Problem with 2.2.1 Install !!! Date: Tue, 17 Jun 1997 22:51:44 +1000 Organization: Monash University Lines: 38 Message-ID: <33A68860.41C67EA6@silas.cc.monash.edu.au> References: <33A2659D.41C67EA6@silas.cc.monash.edu.au> <33A282BD.31DFF4F5@freebsd.org> <5o47mr$e49@uriah.heep.sax.de> NNTP-Posting-Host: ascendp-int-59.cc.monash.edu.au Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-NNTP-Posting-User: root X-Mailer: Mozilla 3.01Gold (X11; I; FreeBSD 2.2.1-RELEASE i386) To: Joerg Wunsch <joerg_wunsch@uriah.heep.sax.de> Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:43033 > Well, there are even more 3Com drivers :) (vx, eg, el), but he wrote > about a 3C503, so the ed0 is correct. > Thanks for backing me up on that one Joerg :) > > > received the following rather large (and awe inspiring/scary) message. > > > > > > Fatal trap 12 : page fault while in kernel mode > > > fault virtual address = 0xefc00000 > > > fault code = supervisor read, page not present > > > > Hmmm. That's definitely weird - it sure would be nice to know exactly > > which driver triggers this. > > Right after ed0, the probe should continue with ed1. The fault is > inside function kvtop(), a very short function converting a KVA > (kernel virtual address) into a physical address, or simpler call it a > page map read. (It's actually in the inlined call to > pmap_kextract().) > ed1 has been disabled so no probe occurs. Disabling the probe on ed0 fixed the problem. > This means the page tables on your machine are corrupted. Oh well, i > don't say it's not the os's fault, but it's not unlikely a memory > problem either. At any rate, it will require serious investigation > before it's possible to spot this (since the time when the page table > is corrupted and the spot where the crash happens might be very > different). > How can I go about fixing the problem so I can use the ed0 driver ? -- David Henshaw dhenshaw@silas.cc.monash.edu.au