Hi Rudolf,
As requested, I'm sending the current diff for my tree. It's a wild scattering of half baked things. I'll describe the important stuff here - the rest of the patch is just junk:
* Windows really wants fadt->smi_cmd, fadt->acpi_enable, and fadt->acpi_disable to be non zero.
* The coreboot acpi_facs_t table has incorrect pad (it has 33, but it should be 31). I'm not sure if this confuses Windows, or if it causes other acpi tables to be aligned in a way that Windows doesn't like.
* Windows needs a _CRS definition for PCI0. I grabbed one from qemu after applying a patch proposed on the qemu mailing list:
http://lists.nongnu.org/archive/html/qemu-devel/2008-08/txtXJHrYIW0Lp.txt
* Windows install needs an ACPI KBD definition - otherwise you can't use the keyboard during the install. Again, I just copied it (along with MOU, RTC, FD0) from qemu.
I've also disabled coreboot from running the vga rom. I do want coreboot to copy the rom into ram - but only so seabios can run it. I've also hacked coreboot so that it puts the pir, mptable, and acpi tables into high memory - this ensures that they wont get overwritten when the seabios payload is deployed. (Seabios will find the tables that need to be in 0xf0000 and relocate them.)
-Kevin
-
Kevin O'Connor