Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!constellation!osuunx.ucc.okstate.edu!moe.ksu.ksu.edu!crcnis1.unl.edu!wupost!howland.reston.ans.net!darwin.sura.net!udel!sbcs.sunysb.edu!stark.UUCP!gene From: gene@cs.sunysb.edu!stark (Gene Stark) Newsgroups: comp.os.386bsd.questions Subject: Re: KERNEL PANIC ANYONE Date: 19 Jul 93 14:35:00 Organization: Gene Stark's home system Lines: 44 Message-ID: <GENE.93Jul19143500@stark.uucp> References: <22222@durer.cme.nist.gov> <1993Jul12.180752.29982@prepress.com> <GENE.93Jul14074659@stark.uucp> <CA8wL9.7G4@veda.is> <GENE.93Jul17074743@stark.uucp> <22brtkINNi99@kralizec.zeta.org.au> NNTP-Posting-Host: stark.uucp In-reply-to: bde@kralizec.zeta.org.au's message of 19 Jul 1993 01:57:40 +1000 In article <22brtkINNi99@kralizec.zeta.org.au> bde@kralizec.zeta.org.au (Bruce Evans) writes: [Lots of previous quotes about bootstrap and kernel bzeroing BSS] Maybe that problem I noticed with the DMA bounce buffer at 0xFE098000? Together with the extra 7 pages for the initial page table, this would make the limit 0xFE091000. Now I really don't know what's going on. In trying to upgrade to patchkit 0.2.4 I started getting crashes when I installed the new interrupt code. The problem appeared to be triggered by my TW-523 driver, because the system crashed when I first tried to use that device. (Bruce, you know about this because I assumed it had something to do with the interrupt code, and I subjected you to a barrage of E-mail messages :-)) After a long weekend of debugging, the best I could determine was that a stray pointer somewhere was causing memory to be trashed, leading to a crash *much later* when a corrupted stack caused a return from "hardclock" to jump into never-never land. I did get a chance to appreciate in detail Bruce's very nice work on the interrupt code, though. I had noticed the "kernel must bzero" messages coming from the bootstrap program, but I had ignored them up until that point. It occurred to me that a nonzero initialization of a BSS pointer variable could produce problems of the type that I was seeing. I rebuilt my kernel with some inessential stuff left out so that it loads under 0xFE090000, and the system has been up ever since. I saw the comments in locore.s, and I assumed that I had correctly surmised the problem. Now when I take a closer look, I realize that I don't have the foggiest idea. I don't know about the bounce buffer stuff. Where are those defined? I don't have any SCSI device drivers loaded, and I only have 8MB of memory. I haven't yet dug in to find out what is going on with the page tables, GDT, etc., and I don't have ready access to the proper hardware manuals to really understand what is going on. So, I still conclude that there is a very subtle problem with loading kernels in a certain size range, but I really am not very sure what is causing it. - Gene Stark -- stark@cs.sunysb.edu