Am Samstag, 26. Februar 2011, 11:19:23 schrieb Paul Menzel:
One question, to make this message even longer, is the problems Scott is having to boot Windows just due to the board port (lack of documentation(?)) or does the generic new AMD “chipset” code also needs more work, so that it would be better to wait for another board where then coreboot could be ported to more quickly after Scott found the “shortcomings” of the chipset code?
It's most likely our support for various tables (the ACPI family, mptable, pir, ...). These are highly manual for now, and it would reduce the time to market for a board with supported components if these could be autogenerated somehow (and in a way that can be validated on all kinds of OS variants/versions and apply for a lot of boards automatically).
Windows is very picky about these tables, and fails if they're not perfect. Linux is (by necessity) more tolerant with crappy data, as the non-windows parts of ACPI are often downright broken.
Patrick