Re: prom-libc status (was Re: putting yaboot and the bootinfoin one file)

["Benjamin Herrenschmidt" <benh@kernel.crashing.org>]

>> What assumptions are you talking about ?
>
>I guess it's just due to lack of documentation, but lots of drivers just
>assume too many hardwired addresses (like, gpio's etc).

Heh, well, tell me about drivers doing that when the actual
information is available elsewhere :) Apple only started to
put GPIO information in the device-tree on recent models,
so for all older ones,GPIOs have to be hard coded (and that's
exactly what darwin does btw). For a long time, for example
the mac-io chip was hardwired at 0xf3000000 and MacOS itself
had some hard coded addresses in it's drivers.

Now, Apple is re-inventing the whell with some ACPI-like
"scripts" they stuff in some OF nodes that you can interpret,
they do that among other for turning on/off the ATA/100 clock
on the windtunnel G4.

One possibility for us would be to write an interpreter for
that and use exclusively that mecanism in the kernel, then
providing a "fixup" file adding those "scripts" to all nodes
of old OFs where they are not present.