Michael South <msouth@ablecom.net> writes:
> I'm going to take a stab at adding support for my laptop's Cirrus Logic
> PD6832 CardBus controller for PCI. This being NetBSD, it should be done
> The Right Way. There are some issues I'd like opinions on:

Well, first see my other mail.


> * Active projects.  Anybody already working on this stuff?

"Not that I know of" (now that jason's corrected me 8-).


> * Interrupt mappings.  This can be very convoluted, depending on how the
> hardware is wired. The PD6832 has four different interrupt modes. Even
> the PD6729 could be wired for INTA#/INTB# or for IRQx. How should this
> be communicated to the driver?

PCI PCMCIA controllers have appropriately the same set of
configuration issues.  In general, the laptop BIOS sets these types of
parameters up and if you try to change them you run into problems.

CardBus cards shouldn't know much, if anything, about the interrupt
mapping scheme used by their controller, they should just call back
into controller code to set up the mapping.


> * CardBus support.  How should the DMA be integrated into the rest of
> NetBSD?

"In a way similar to the way PCI DMA is."


> * ATA mode, zoomed video, etc. Do we want a gadzillion routines added to
> the pcmcia_chip_functions structure, or should we rethink?

Uh, which, if any, of these are also PCMCIA issues?  (I thought that
at least ATA mode was.)


> * Config options for PCI devices. What about putting such things as IRQ
> type and mask, DMA mask, memory windows, etc. as device options (as IRQ
> and port addresses are for ISA devices)?

These types of things should be done the same way the PCMCIA code
handles them for PCMCIA cards: i.e. the user doesn't specify that type
of thing.



cgd
-- 
Chris Demetriou - cgd@netbsd.org - http://www.netbsd.org/People/pages/cgd.html
Disclaimer: Not speaking for NetBSD, just expressing my own opinion.
Plug: Get your official NetBSD-1.3.2 CDROM set today! http://www.netbsd.com/