Mark Jackson <mpfj-l...@mimc.co.uk> wrote: > The functions could also return (architecture dependant) a "remapped" > address to be used, then we could support:-
Right, and that is the important part: If I'm allowed to return a remapped address, this API will be trivial to implement on AVR32. If not, it will be quite difficult (and make everything larger and slower). Your idea is good; it's mostly similar to what map_physmem() does today, but perhaps the name is wrong, and I suppose it should probably flush the caches in addition to just remapping the address. Haavard _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot