PCI Subsystem Rework for X.org 7.1 Proposal

[Egbert Eich]

Jesse Barnes writes:
 > 
 > I'd have to see examples.  It seems like, for the most part, they 
 > shouldn't be necessary in userspace, since the kernel will provide us 
 > with directly dereferenceable pointers to MMIO space.  If we end up 
 > having to write a bunch of support ourselves for a different back end 
 > though, then they might be necessary.  IOW, whether or not they should 
 > be implemented depends on whether or not we care about the reason why 
 > they exist anymore, which I'm not sure of.
 > 

These things were mostly required for pre EV6 alphas.
These platforms required special access patterns to memory
if you needed to do accesses to 8 or 16 bit wide registers.
The macros provided a more or less convenient way to handle 
this situation.
However this was not the only reason that we added this.
Certain platforms required some mechansim to trigger data
writes accross the bus. It is possible to add this to the
macros for that specific platform and thus provide a generic 
way to do that without requiring the driver author to know
how to trigger these for a specific platform.

Cheers,
	Egbert.