> the PNP patch renames request_region because it introduces new concepts of
> hardware-resource management. All request_ functions are piped to a
> central 'request_hw_region()' function, with the appropriate flag. The
> 'IO-region' is just a special flag. If we use the PNP patch, it would be
> very misleading to have 'request_region'. Why should 'request_region'
> be special?
I agree with you that it's a less than ideal state of affairs. But, IMHO,
the way to go is to introduce the new "request_hw_region()" call and its
friends, but keep request_region(), request_irq() and so on around, even
if they're just at this stage stubs that do nothing special and their
use is deprecated. Then, maybe in 2.3, they could be removed. I'm not
particularly emotionally attached to the names we've got, I just feel that
changing them all of a sudden is not a good idea.
phil