Return to BSD News archive
Newsgroups: comp.unix.bsd Path: sserve!manuel.anu.edu.au!munnari.oz.au!uunet!zaphod.mps.ohio-state.edu!rpi!think.com!snorkelwacker.mit.edu!thunder.mcrcim.mcgill.edu!sifon!michaelm From: michaelm@pike.ee.mcgill.ca (Michael Moscovitch) Subject: [386bsd] wd driver device probe woes. Message-ID: <1992Oct18.213715.2805@sifon.cc.mcgill.ca> Summary: IDE drive not found on boot, looks like timeout problem. Keywords: 386bsd, IDE, wd Sender: news@sifon.cc.mcgill.ca Nntp-Posting-Host: pike.ee.mcgill.ca Organization: McGill University - VLSI Laboratory, Montreal, CANADA. Date: Sun, 18 Oct 1992 21:37:15 GMT Lines: 38 I have 2 IDE drives in my 386/33 machine. One for 386bsd and one for DOS. I decided that I wanted to swap things so that 386bsd was on the bigger drive. That's when the fun started. I have had 286bsd 0.0 and 0.1 running sucessfully on a 125 meg drive (ST1144A). I am unable to get it to boot on a WD AP4200. The wd device is not found when I boot the boot floppy. Unfortunatly I had to overwrite the 386bsd disk with the contents of the other one. Finding myself in this awkward situation, I decided to try and find out why it would not recognize the drive. I thought it might be either a timing problem or a problem with the drive parameters. Luckily, I had printed the source for the wd driver earlier and I began reading it. I decided that I should write a dos program to test out some of the functions performed by the probe and attach routines. Anyway, I hacked those sections together and seemed to have found my problem. The DIAGNOSE command in wdprobe seems to take quite a bit of time. As well, the reset seems to take more time than the delay specified. The wdcommand routine was timing out. I found that I needed about 3 times the timeout value. Since the timeout value is cumulative within wdcommand so it needs to account for the time to wait after the reset and the time for the DIAGNOSE command. Has anyone experienced this problem also? Is there an existing patch and/or bootable that fixes it? I tried my own kernel that had many patches (earlier than the patch kits) and the unofficial/newbootables/fixit.fs.debug (I think). If there are no existing bootables out there, I will have to reload onto the small disk and set everything up to recompile the kernel, all the while moving all my dos stuff back and forth to the the other disk. -- <Mike> michaelm@pike.ee.mcgill.ca