On Tue, May 15, 2007 at 04:24:26PM -0400, Corey Osgood wrote:
I'll post a draft soon, so that people can see which functions are affected. This needs an abuild-run and tests on a bunch of different hardware targets to make sure nothing breaks...
Well, if you want to compare notes at all, I've attached my raminit.c (the file, not a patch).
Oh, that's fine. Those were not the sort of SPD functions I was talking about, we didn't duplicate any work :-) Your functions merely _use_ SPD functions (or spd_read_byte/smbus_read_byte per se), if at all.
I posted a draft patch of my SPD refactoring stuff for preliminary review...
If you could, please maintain the MODEL_440ZX stuff, as that's necessary for 440zx support. MODEL_440ZX will be defined in the mainboard auto.c, once this is all up and running, and those ifdef's are the ONLY differences between the two northbridges. However, there's no way to detect the difference at runtime that I can find, so a define is the only real way.
Hm, why is that? Don't they have some sort of different IDs we can read from some register?
But if that's not possible, a #define is fine too, IMO. It's great that you planned for merged code beween 440BX/ZX/FX/etc.
Maybe we should rename the 'i440bx' directory to 'i440xx' later to make it clear that (later) multiple chipsets will be supported?
Uwe.