This is a long story, sorry about that. The NetBSD version in question
is 1.3 release.

I had a spare Pentium 75 CPU lying around so I thought I'd upgrade my
486 NetBSD system at home. I went and bought an AOpen AP58 motherboard
and a single 16 MB EDO RAM SIMM (I already had one). I also bought a
Seagate ST32122A (2.1GB) IDE disk (in addition to the "old" ST31276A
(1.2GB) I already had NetBSD on).

After installing the motherboard the system wouldn't boot. It passed the
memory test and the disk probe, and then it just hung (no setup summary
screen).

After much fiddling around I discovered that disconnecting the IDE
cables would make the machine boot from floppy (NetBSD or DOS).

I checked the BIOS HD information, but the new motherboard reported the
exact same geometry on my boot disk as the old 486 had. This is the same
geometry as the NetBSD kernel reports.

Since the AP58 has a SiS 5582 chipset I thought that might be the
problem. I also suspected that the board could be defective. My
pusher^H^H^H^H^Hsupplier of PC hardware loaned me another AP58 board as
well as an AP5T (Intel 430TX PCIset). Same problem.

I've tried upgrading the BIOS to the latest version, tried moving the
disks around (primary/secondary IDE, master/slave) and even tried
reinstalling the 486 motherboard just to verify that the disks are OK.

In desperation I reinstalled the 486 yet again, booted NetBSD, labeled
the 2.1GB disk (installed as wd1) without installing bootblocks and
copied the contents of the 1.2GB to the 2.1GB disk using dump piped to
restore. That was when I got the "wdc0(1): lost interrupt" message
reported earlier, but that's another story.

Finally (still using the 486), I booted from a DOS 6.2 diskette and did
an fdisk/mbr on my primary (1.2GB) disk. I reinstalled the Pentium board
with only the "old" 1.2GB disk and it still didn't boot (it hung after
the disk probe) as usual. I removed the 1.2GB and installed the new
2.1GB disk instead and it passed the disk probe and the setup summary
screen. It didn't boot since there were no bootblocks. I booted from a
NetBSD install disk and installed bootblocks. I thought my troubles were
over and rebooted. Nope. Again, the machine hung after the disk probe.

As a last resort I kidnapped two known-to-work 32MB EDO RAM SIMMs from
another system and tried them on the AP5T. Still no go.

Summary:

As long as there is (or has been) NetBSD bootblocks on _any_ HD
connected to _either_ IDE controller on either an AP58 or AP5T
motherboard it won't boot.

Anyone else seen this?

Additional info: Yes, even putting the NetBSD disk as slave (with a
CD-ROM as master) on the Secondary IDE controller will prevent the
system from booting, even though there is a blank disk (the ST32122A) as
master on the Primary IDE controller. I tried this before I copied
NetBSD to the ST32122A.

Additional info 2: I'm running NetBSD 1.3 release on an AOpen AP65
(Intel 440FX PCIset) Pentium PRO system at work with no problems.


Suggestions?

Gunnar

--
Gunnar Helliesen   | Bergen IT Consult AS  | NetBSD/VAX on a uVAX II
Systems Consultant | Bergen, Norway        | '86 Jaguar Sovereign 4.2
gunnar@bitcon.no   | http://www.bitcon.no/ | '73 Mercedes 280 (240D)