Verily did "Justin T. Gibbs" write:

>>>>The problem applies to any busmastering ISA device.

>>> And some broken VLB and EISA cards too. The 14F is a VLB card, right?

>>I don't know about VLB, but the problem does not apply to EISA, since
>>EISA has 32-bits of address.  If there'd been a problem with EISA,
>>then the previous cvs.netbsd.org would have never worked :-)

> Did you miss the "broken" in my comment?  VLB and EISA should not have the
> problem, but some cards and/or motherboard chipsets, do.  Once such
> example is the bt445S.

Hmm, ok, but...

I have a bt757C, which typically fits the "broken EISA" category since it
looks a lot like an ISA card at the driver level (hence the delay for
real eisa glue for bt0).  Can't speak authoratatively on the mobo chipset,
it's a Micronics PCI/EISA 486 job (rare, yes, and the PCI sux on it).

Anyway, with 64M of mem, I haven't seen a DMA problem.

Take it for what it's worth.  *shrug*

-- Rob
----------------------------------------
Internet: windsor@pobox.com      Life: Rob@Sunnyvale.California.USA.Earth
"Da Web": http://pobox.com/~windsor/

"Ain't much distance 'tween a pat on the back and a kick in the pants."
    -- David Lee Roth (1990), `The Dogtown Shuffle'