My config was actually just the unmodified P2B build originally, but not built with the coreboot toolchain. As for the modified config, it was just P2B with the naming changed to P2-99.
Looking at the picture of the P2B board in the wiki again, the P2-99 may actually just be the same board with one of the ram sockets unpopulated and the 440ZX northbridge chip.
I haven't been able to get flashrom to find the flash chip, despite it recognizing the chipset and using compatible eeproms - I flashed them with coreboot on a different board after all.
Running a git build of flashrom didn't help either, still couldn't find the flash.
It's got an AS97127F, not an AS99127F. I tried to use the trick you linked from the P3B-F, but i2cset just reported a failure to write. I had previously tried getting flashrom to use some other board enables and that didn't work either.
I attached a dmesg log with the ACPI errors. I don't know whether it's related to a missing MBRS table as you said. It also shows "[Firmware Bug]: the BIOS has corrupted hw-PMU resources (MSR 186 is 40002e)" after a reboot. It did that with every build I've tried and not specific to the ACPI patched version.
I just made a clean unmoddified build for P2B with the coreboot toolchain. Might only get to flashing it tommorrow though.
I did get signed in on gerrit, so I should be able to run board_status on it.
If you get the early cbmem init patch uploaded I'll try to test that right away as well.