> Hi, for as long as I can remember, I've been having this problem. The
> problem is when I do a bunch of disk I/O on my second ATA66 drive, my
> machine totally locks up. It's repeatable, but not always the exact same
> circumstances will cause it. Sometimes it happens when I had been doing
> things with the second ATA66 drive (from here on just referred to as wd1)
> and do something that causes a lot of activity on the first ATA66 drive
> (from here on referred to as wd0), sometimes when moving files around on wd1
> without touching wd0, but basically the best way I can think of to describe
> what leads to it happening is "a bunch of disk I/O."

I had the same problem as well, on an HPT370A (not on-board). The system
behaved strangely but I couldn't locate the bug - just as you describe, it
hanged because of "a bunch of disk IO". Then my second disk died
(with a half of RAID0 partition on it. RAIDframe, not the HPT's "raid"),
which made me trow that HPT away. Everything is ok since then ;).

> I have been looking through the mailing list archives, and I'm not the only
> one who has had this problem, but I don't see any solutions or workarounds
> (other than switch the disks to using PIO, which isn't an option, if I
> wanted PIO I wouldn't have bought ATA66 disks)

Sometimes it hanged even with both drives (primary+secondary master) in PIO.
Completely non-deterministic though.

> I think the most important thing to note here is that this never once
> happened in Windows 98, and the last time I tried Win98 I had the same drive
> on the first ata66 controller, and a maxtor ATA66 on the second. I am 100%
> positive it isn't happening because wd1 is bad, because I have had this
> happen with something like four or five different ATA66 drives up to this
> point.

Maybe you don't have DMA turned on in Windows? I wouldn't use this as
reference, the hardware probably is buggy and the win-drivers may work
it around...

> So ISTM there is some quirk/bug/whatever in the HPT366 on the Abit BE6/BP6
> motherboards relating to using both ATA66 channels that windows handles
> correctly, and NetBSD doesn't? Maybe correctly isn't the word, but handles
> without the system locking up in any case.
> Does anyone know if this is fixable in NetBSD at all?
> I am probably going to be buying a Promise PCI ATA100 TX2 card as soon as I
> can afford it, but I would really like the existing hardware configuration
> to work.

I in your situation would buy new IDE card. I don't know whether the HPT
really is responsible for the death of my disk, but I would believe it...

> FWIW, this happened in FreeBSD as well. I didn't have two ata66 drives in
> here during the short time this box ran Linux, so I never found out if Linux
> would have the same problem, though if what I've found sifting through the
> web for information, it seems that Linux does.

I had the same problem in Linux.

Regards,
	-- Jachym _Freza_ Holecek