On Sun, Oct 17, 1999 at 09:01:15AM -0700, Mike Cheponis wrote:
> I installed 1.4.1 on an IBM 18G disk into a 2G partition.  Then I gave the
> disk to a friend to install into his machine.  (Both machines are K6-2/450
> MHz w/64 MB, but are otherwise not identical motherboards).
> 
> Oct 16 15:10:37 USB syslogd: restart
> Oct 16 15:10:38 USB /netbsd:  tn 7 sn 54), retrying
> Oct 16 15:10:38 USB /netbsd: wd0: soft error (corrected)
> Oct 16 15:10:38 USB /netbsd: wd0a: DMA error reading fsbn 114288 of 114288-114319 (wd0 bn 114351; cn 113 tn 7 sn 6), retrying
> Oct 16 15:10:38 USB /netbsd: wd0: soft error (corrected)
> Oct 16 15:10:38 USB /netbsd: wd0a: DMA error reading fsbn 85088 of 85088-85103 (wd0 bn 85151; cn 84 tn 7 sn 38), retrying
> Oct 16 15:10:39 USB /netbsd: wd0: soft error (corrected)
> Oct 16 15:10:39 USB /netbsd: wd0a: DMA error reading fsbn 85104 of 85104-85119 (wd0 bn 85167; cn 84 tn 7 sn 54), retrying
> Oct 16 15:10:39 USB /netbsd: wd0: soft error (corrected)
> 
> [... many more of those ...]
> 
> 
> He gets many many of these "soft errors".  But it makes the system unusable.
> 
> It worked perfectly on my machine.

One of my machines, with a FIC motherboard using the VIA MVP3 chipset,
had this problem, with two different UDMA drives.  Another did not,
with the same drives and an identical motherboard and processor.  I
think I also tried swapping the cable.

Changing the "flags" byte for the drives in the kernel config to limit
the UDMA mode to 1 instead of 2 made the problem go away.  Mmmmmm,
crappy IDE electrical specification.  Now you know why the UDMA modes
added that CRC-32 to every command...