Well, the channel is disabled for some reason, I suspect by the OF.
   It's just a bit to reset in the PCI config space; but the driver doesn't touch
   it because usually the user has explicitely disabled it in the BIOS
   (yes, I know, it's PC centric, but we have to deal with it too).
   
   We could have the driver clear the bit for native mode (in this case there's
   no risk of conficts with another add-on board), based on a compile-time
   option. Would this be enouth ?


hmm... i think this is not what is up.


i just removed the test that was causing the probe to fail.  now i get:

	pciide0: secondary channel configured to native-PCI mode
	pciide0: disabling secondary channel (no drives)

and if i try to use `probe-ide' (ide version of the classic `probe-scsi',
of course), at any point after booting netbsd (including after doing a
`reset-all'), the machine hangs and i need to power cycle it, at which
point the probe works again.  actually, i just tried this again, and the
probe-ide didn't hang.  it didn't probe the 2nd channel at all, though,
like it normally does.  as if it also thought channel was disabled.


it suggests to me that netbsd is doing something wrong still.  :(


.mrg.