From owner-freebsd-stable Mon May 1 19:53:18 2000 Delivered-To: freebsd-stable@freebsd.org Received: from jason.argos.org (a1-3b058.neo.rr.com [24.93.181.58]) by hub.freebsd.org (Postfix) with ESMTP id C82E037BA3A for ; Mon, 1 May 2000 19:51:51 -0700 (PDT) (envelope-from mike@argos.org) Received: from localhost (mike@localhost) by jason.argos.org (8.9.1/8.9.1) with ESMTP id WAA10165; Mon, 1 May 2000 22:50:28 -0400 Date: Mon, 1 May 2000 22:50:27 -0400 (EDT) From: Mike Nowlin To: Tim Pozar Cc: freebsd-stable@freebsd.org Subject: Re: Geek Port crash In-Reply-To: <20000430140040.A71703@lns.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > I am playing with the same beast (ppi) and am experiancing the same > panics. Did you make any head way? News flash: I just took my test machine back to 4.0-CURRENT as of "Computer D-Day" (December 31, 1999, 23:59:59), and the ppi devices do work on that version - I did have to recompile my test program. I'm gonna bump it up to Feb 1 of this year and see what happens. Stay tuned... (hmmm.. telnet connections to that machine seem to have broken, and my cvsupd on another box just dumped core.... :( Oh, well -- at least today is being consistently bad.) ...CVSup to 2000.02.01.00.00.01... (time passes) hawk:/home/mike# ./geekport panic: nexus_setup_intr: NULL irq resource! syncing disks... 17 15 done Uptime: 31s boom! Obviously, the Feb 1 version is broken - let's try something halfway. ...CVSup to 2000.01.14.00.00.01... (more time passes... If I was running Windoze, my mail program would have blown up by now.) hawk:/home/mike# ./geekport WARNING: driver ppi should register devices with make_dev() (dev_t = "#ppi/0") geek port opened - fd=3 ^C hawk:/home/mike# cc geekport.c -o geekport hawk:/home/mike# ./geekport geek port opened - fd=3 ^C It works, and my little LED doohickey is blinking away. So, the problem developed somewhere between Jan 14 and Feb 1. Looks like there were some changes to sys/dev/ppbus/* on Jan 14 and Jan 23, so I'll pick Jan 21.. ...CVSup to 2000.01.21.00.00.01... (snore...) Edit src/sys/dev/ppbus/if_plip.c Add delta 1.16 2000.01.14.00.17.51 nsouch Edit src/sys/dev/ppbus/immio.c Add delta 1.10 2000.01.14.00.17.51 nsouch Edit src/sys/dev/ppbus/lpbb.c Add delta 1.10 2000.01.14.00.17.52 nsouch Edit src/sys/dev/ppbus/lpt.c Add delta 1.11 2000.01.14.00.17.52 nsouch Checkout src/sys/dev/ppbus/pcfclock.c Edit src/sys/dev/ppbus/ppb_1284.c Add delta 1.10 2000.01.14.00.17.53 nsouch Add delta 1.11 2000.01.14.08.03.14 nsouch Edit src/sys/dev/ppbus/ppb_1284.h Add delta 1.6 2000.01.14.00.17.53 nsouch Add delta 1.7 2000.01.14.08.03.14 nsouch Edit src/sys/dev/ppbus/ppb_base.c Add delta 1.10 2000.01.14.00.17.54 nsouch Edit src/sys/dev/ppbus/ppb_msq.c Add delta 1.8 2000.01.14.00.17.55 nsouch Edit src/sys/dev/ppbus/ppb_msq.h Add delta 1.8 2000.01.14.00.17.55 nsouch Edit src/sys/dev/ppbus/ppbconf.c Add delta 1.16 2000.01.14.00.17.55 nsouch Edit src/sys/dev/ppbus/ppbconf.h Add delta 1.17 2000.01.14.00.17.56 nsouch Checkout src/sys/dev/ppbus/ppbio.h Checkout src/sys/dev/ppbus/ppbus_if.m Edit src/sys/dev/ppbus/ppi.c Add delta 1.18 2000.01.14.00.17.57 nsouch Add delta 1.19 2000.01.14.08.03.15 nsouch Add delta 1.20 2000.01.14.15.54.45 peter Edit src/sys/dev/ppbus/pps.c Add delta 1.23 2000.01.14.00.17.58 nsouch Edit src/sys/dev/ppbus/vpo.c Add delta 1.19 2000.01.14.00.17.58 nsouch Edit src/sys/dev/ppbus/vpoio.c Add delta 1.10 2000.01.14.00.17.59 nsouch Edit src/sys/dev/ppbus/vpoio.h Add delta 1.4 2000.01.14.00.17.59 nsouch (ugh - lots of changes. Of course, it might not be in the ppbus section itself - this is an IRQ error. BTW: I'm using the GENERIC config files.) Dum-da-da-dum! hawk:/home/mike# cc geekport.c -o geekport hawk:/home/mike# ./geekport panic: nexus_setup_intr: NULL irq resource! syncing disks... 62 59 40 34 15 done OK - problem is somewhere between Jan 14 @ midnight and Jan 21 @ midnight... Gut feeling is that it's in that ppbus stuff above. After looking at the sources a little more, it appears that Jan 14 was when the ppbus code was ported over to newbus. A coincidence? I think not. :) A search of the archives and reading the commit logs reveal several reports of this problem, but little in the way of what it might be. I'm gonna play around with it some more, but I could REALLY use some hints, or an answer as to why the ppi device would even require an IRQ in the first place, unless you request one. I know next to nothing about the kernel guts, especially the interrupt processing. please please please?? Thanks - Mike To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message