thorpej@shagadelic.org wrote:
> > How about this patch? (the problem is device_unit() abuse)
> > With this patch, both fdc0 at isa0 and fdc1 at acpi0 work properly.
> 
> How about setting a property for this in device_register() instead?   
> Let's the the NVRAM-reading stuff out of fd.c completely!

I agree it's our goal, but I'm afraid it won't happen
until we have real MI (not isa specific) fdc(4) driver...
My patch is just an "interim compromise."
---
Izumi Tsutsui