From owner-freebsd-stable Tue Jan 9 17:20:30 1996 Return-Path: owner-stable Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id RAA25021 for stable-outgoing; Tue, 9 Jan 1996 17:20:30 -0800 (PST) Received: from localhost.cdrom.com (localhost.cdrom.com [127.0.0.1]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id RAA25012 Tue, 9 Jan 1996 17:20:25 -0800 (PST) Message-Id: <199601100120.RAA25012@freefall.freebsd.org> X-Authentication-Warning: freefall.freebsd.org: Host localhost.cdrom.com [127.0.0.1] didn't use HELO protocol To: "Steven G. Kargl" cc: freebsd-stable@freebsd.org, gibbs@freebsd.org Subject: Re: aha1742 and kernel builds are broken In-reply-to: Your message of "Tue, 09 Jan 1996 17:04:36 PST." <199601100104.RAA00239@troutmask.apl.washington.edu> Date: Tue, 09 Jan 1996 17:20:24 -0800 From: "Justin T. Gibbs" Sender: owner-stable@freebsd.org Precedence: bulk >cc'd to gibbs@freebsd.org I read the stable mailing list. >Supped -stable Friday (05 Jan 96) at 0600 pst. > >I noted that src/sys had some changes/updates including >aha1742.c, so I decided to build a new kernel. > >My new config file (NEWKERNEL) contains > >controller isa0 >controller eisa0 >controller ahb0 correct. >dmesg for NEWKERNEL shows > >eisa0: >Probing for devices on the EISA bus >ahb0: at 0x1c00-0x1cff irq 11 on eisa0 slot 1 >ahb0 waiting for scsi devices to settle >(ahb0:0:0): "MAXTOR MXT-540SL J1.5" type 0 fixed SCSI 2 >sd0(ahb0:0:0): Direct-Access 521MB (1067740 512 byte sectors) >(ahb0:1:0): "QUANTUM LIGHTNING 730S 241E" type 0 fixed SCSI 2 >sd1(ahb0:1:0): Direct-Access 699MB (1431760 512 byte sectors) >(ahb0:2:0): "HP HP35480A 1109" type 1 removable SCSI 2 >st0(ahb0:2:0): Sequential-Access density code 0x13, drive empty > >Clearly, the new sys/i386/eisa/aha1742.c is getting the probe wrong. >That is, OLDKERNEL is finding ahb0 at 0x1000-0x10ff and NEWKERNEL >is find ahb0 at 0x1c00-0x1cff. Actually, the old code was wrong. The 1742 only uses the registers from 0xZc00-0xZcff. Looking at the register definitions in the driver file itself will verify this for you if you have any doubts. After all, how do you think it would find the card at all let alone probe any devices on the bus if the base addresses were wrong? >NEWKERNEL will hang the system without a panic or a message in >/var/log/message. The system will not respond to the keyboard at the >console nor will it respond to telneting into the system from a >remote site. If you want to hang the system, it is quite repeatable >with `iozone 32' in /usr/tmp. Sounds like an interrupt problem. I guess I'll have to install the 1742 that's on my shelf and see if I can reproduce this. Can you tell me what other devices are in your system? >One other point, to get NEWKERNEL to config and compile, I had to >edit sys/i386/conf/files.i386 and comment out the i386/isa/aha1742.c >line. Otherwise, you do not pickup the file aha1742.c. Whoops. Looks like I missed that change in my commit. I'll fix that later tonight. >-- >Steven G. Kargl | Phone: 206-685-4677 | >Applied Physics Lab | Fax: 206-543-6785 | >Univ. of Washington |---------------------| >1013 NE 40th St | FreeBSD 2.x-stable | >Seattle, WA 98105 |---------------------| -- Justin T. Gibbs =========================================== FreeBSD: Turning PCs into workstations ===========================================