*BSD News Article 22919


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!munnari.oz.au!constellation!osuunx.ucc.okstate.edu!moe.ksu.ksu.edu!vixen.cso.uiuc.edu!howland.reston.ans.net!europa.eng.gtefsd.com!avdms8.msfc.nasa.gov!sol.ctr.columbia.edu!news.kei.com!world!ksr!jfw
From: jfw@ksr.com (John F. Woods)
Newsgroups: comp.os.386bsd.questions
Subject: Re: The reason for stray interrupts
Message-ID: <34175@ksr.com>
Date: 27 Oct 93 12:33:49 EDT
References: <2ais9gINN2t8@xs4all.hacktic.nl> <wilko.751660652@spoetnix.idca.tds.philips.nl>
Sender: news@ksr.com
Lines: 29

wilko@idca.tds.philips.nl (Wilko Bulte) writes:
>ptuomola@hacktic.nl (Petri Tuomola) writes:
>>Many people have been asking why their machines display "ISA strayintr 7"
>>messages. This is an extract from /sys/arch/i386/isa/isa.c:
>>	/* for some reason, we get bursts of intr #7, even if not enabled! */
>>	/*
>>	 * Well the reason you got bursts of intr #7 is because someone
>>	 * raised an interrupt line and dropped it before the 8259 could
>>	 * prioritize it.  This is documented in the intel data book.  This
>>	 * means you have BAD hardware!  I have changed this so that only
>>	 * the first 5 get logged, then it quits logging them, and puts
>>	 * out a special message. rgrimes 3/25/1993
>I've read this piece of comment once in the source, and I don't think it is
>very sensible. One could argue that the 8259 is braindead (which is possibly
>true ;-) but this problem also occurs on machines with chipsets which 
>integrate the intr controller and so have no actual 8259 chip.

This is, however, the PC world, where correcting the hardware mistakes of
the past is viewed with grave suspicion; those integrated controllers may
well deliberately duplicate this misbehavior...  ;-)

>I have run *BSD on various systems, and I have still to find a system that
>does not complain about stray interrupts. And these systems were *not*
>bad hardware, they have run both AT&T and SCO Unix without any problems.
>(No comments on the virtues/vices of Sys V systems please).

I think I've seen it only once or twice on my NetBSD system, when the system
was having distinct hardware problems.  (It's a 486/33, so it ought to be
fast enough to uncover speed-related problems.)