From owner-freebsd-mobile Sun Jan 25 10:25:26 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id KAA01459 for freebsd-mobile-outgoing; Sun, 25 Jan 1998 10:25:26 -0800 (PST) (envelope-from owner-freebsd-mobile@FreeBSD.ORG) Received: from cesit1.unifi.it (cesit1.unifi.it [150.217.1.31]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id KAA01454 for ; Sun, 25 Jan 1998 10:25:24 -0800 (PST) (envelope-from ugo@dsi.UNIFI.IT) Received: from aguirre.dsi.unifi.it by CESIT1.UNIFI.IT (PMDF V5.0-4 #3688) id <01ISSZU09TPC000XWR@CESIT1.UNIFI.IT> for freebsd-mobile@FreeBSD.ORG; Sun, 25 Jan 1998 19:25:18 +0100 (MET) Received: from dsi.unifi.it (com1.unifi.it) by aguirre.dsi.unifi.it (4.1/SMI-4.1) id AA15246; Sun, 25 Jan 1998 19:30:55 +0100 Received: from pegasus.home.net (pegasus.home.net [192.168.1.3]) by dsi.unifi.it (8.8.8/8.8.8) with ESMTP id RAA02014; Sun, 25 Jan 1998 17:17:19 +0100 (MET envelope-from ugo) Received: (from ugo@localhost) by pegasus.home.net (8.8.8/8.8.7) id RAA00278; Sun, 25 Jan 1998 17:17:19 +0100 (MET) Date: Sun, 25 Jan 1998 17:17:18 +0100 (MET) From: Ugo Paternostro Subject: Re: Cirrus Logic PD6729 In-reply-to: <199801241819.LAA25873@mt.sri.com> To: Nate Williams Cc: freebsd-mobile@FreeBSD.ORG Message-id: Organization: Not an organization MIME-version: 1.0 X-Mailer: XFMail 1.2 [p0] on FreeBSD Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 8bit X-Priority: 3 (Normal) Sender: owner-freebsd-mobile@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > Did you re-install the includes? I'm compiling stuff with DDB and I > don't have any problems. Yes: yesterday I went to Gianmarco Giovannelli's home (he said me you know him...) and did a "make reinstall" from his -current world. I fixed the includes, and broke ps, w, and the shutdown itself (the kernel now panics every time you do a shutdown while unmounting the disks) :( No problem, I will sup the sources myself and do a "make world" ASAP. I will have to use a NFS disk to make the world, because my laptop has not enough space, but it is the only machine here with the 3.0... *sigh* I think I will beat the "slowest make world record" :) BTW, any chance to build 3.0 on top of 2.2.5 without notching 2.2.5 itself? > want to post this kind of information to -stable and see if they know of > bugs in the current system. To -stable? Not to -current? > (Note, I'm also having problems getting a back-trace as well when it > crashes mounting my DOS partition...) Well, so *you* should post that information... ;-) >> after accessing location 0x10. Please note that the IP points to that >> location. > > Something is hosed up. What about interrupt vector initialization? Where does irq 10 vector point at startup? Today I investigated further on the problem, and I verified a suspect that I had: after the first (cold) boot, the PCMCIA card and the controller are not disabled while rebooting. I proved this making a broadcast ping from another machine on my network. As with many other cards, my 3c589d blinks the led every time a packet is seen on the Ethernet cable (I'm using a BNC connection). Well, the led is blinking after "shutdown -r", while the machine is doing the POST and while it is loading the kernel. So, when the kernel attached the pccard, the pcic is yet configured with enabled IO mem and interrupts, and *probably* the panic i get is simply an irq without a handler. Moreover, I disabled the IO & MEM mapping and the interrupts inside pcic_probe routine, and... tadaaaaa... now it seems to work. A few more tests, with a more decent installation, and I will send you the patches. BTW, wouldn't it be better to disable the controller and to power off the PCMCIA card(s) at shutdown time, just like W*ndows 95 (sorry) does? I gave a look at the kernel sources, and saw that it is calling a series of handlers to do house keeping just before suicide. You could attach yourself an handler to shut off the pccard things. > Ahh, can you disable APM just for kicks to see if that makes things any > better? Yes, no: yes, I can disable APM; no, unfortunately it didn't make things better... > Nate Bye, UP