Date: Thu, 23 Sep 1999 21:19:00 -0700 From: Dann Lunsford <dann@greycat.com> To: Eric Lee Green <elgreen@iname.com>, hardware@freebsd.org Subject: Re: Secondary IDE-PCI interface invisible Message-ID: <37EAFBB4.5D0B9C31@greycat.com> References: <37E71CF0.7EFE9E0B@greycat.com> <99092319090601.00682@ehome.local.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Eric Lee Green wrote: > > On Mon, 20 Sep 1999, Dann Lunsford wrote: > > I'm currently running a DFI P5BV3+ motherboard, with 128 megs and a K6-2 > > 350. > > Freebsd 3.3-STABLE. uname -a output follows: > > > > FreeBSD bigphred.greycat.com 3.3-STABLE FreeBSD 3.3-STABLE #1: Mon Sep > > 20 06:13:34 PDT 1999 > > root@bigphred.greycat.com:/usr/src/sys/compile/PHRED i386 > > > > Problem: NOTHING I've tried can make the secondary on-board IDE > > interface > > visible to FreeBSD. The BIOS detects a drive on the secondary, but > > boot just says "Sorry." Output of dmesg follows > > I had a similar problem. The problem is caused by FreeBSD adhering strictly to > the EIDE standard. Other operating systems, such as Linux and Windows 98, allow > misconfigured hard drives to still operate. > Hmm. Sounds real familiar. Had some problems with VMS like that. Rather have the standard, though; makes things a lot simpler in the long run. > In my particular instance, I had removed an IDE CDROM from my secondary > on-board IDE channel and replaced it with a SCSI Yamaha 4-4-16 CD-RW drive. The > hard drive on that channel was a Western Digital Caviar 3.2gb EIDE hard drive. > The WD was strapped as master. If you read the WD manual, though, it says that > if it is the only device on the bus, that it must have NO straps -- i.e., be > configured as neither master nor slave. Also note that if it is the only device > on the bus, it must be on the last connector on the cable, else you can get > some problems due to lack of termination. Blast. That's exactly what this drive is: A WD 3.2g Caviar. OK, you've convinced me. Hunted down the install instructions, yup, no straps if only device. Blast (I'd use something stronger, but I'm too tired.). > > Removing the jumper on the WD made it accessible to FreeBSD, and all still runs > well even unto today. > > Note that this is specific to particular brands of hard drive. For example, my > 8.4gb IBM EIDE hard drive did not care when I removed the 1gb IDE hard drive > that was secondary on that channel. > > And finally, the way I figured out what was happening was by going through wd.c > and putting a LOT of debugging statements until I found the routine that was > returning an erro. If that error had been reported rather than just returned > as an error constant, the problem and its solution would have been obvious. > Unfortunately, wd.c does not properly report errors. Hopefully the re-written > IDE driver for FreeBSD 4.0 will properly report errors, rather than always > reporting a generic "could not detect" error. What's wierd is that enabling "options ATAPI" in the kernel also fixes (at least, for me) the problem. The way you went about it was what I was planning for this weekend; thanks for releasing the results of your labor, saved me a bunch of work. I owe you one. Sheesh. Will recompile without ATAPI and remove jumper, just to see. Guess I read a lot of code this weekend. Know any good refs on PCI? Thanks a *LOT*. Dann Lunsford To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hardware" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?37EAFBB4.5D0B9C31>