I built a 2.6.24.7 kernel almost exactly to Nils his config, had to change some things like USB storage as my main boot medium, and it worked. I think the generic IDE driver loads up, still says 'IDE disabled' but then probes for hdb en hda. hda actually is a Transcend IDE disk (Compact Flash sort of IDE disk, but still).
hdparm -tT reports a whopping 1.87mb/s and I can't enable DMA, operation not permitted, suggesting the chipset driver is missing, which I belive is compiled in.
The more interesting part though is, that I was able to install partitions, grub and actually am booting and running with said 2.6.24.7 kernel and openwrt distro! I attached a new lspci -vvvxxx output, which shows slightly different IDE timings.
In any case, I went trough the CS5536 documentation and also found MSR stuff about enabling the IDE port (or hooking it to flash chips directly) and that there where certain 'balls' that also setup the board a certain way. After checking the pinout of the CS5536 and using the very handy msrutil I found the IDE port is actually connected in IDE mode, and why the bios lists stupid notions about the port being disabled baffles me.
I am now ready I think to start giving coreboot a whirl on this thing. Though I still wonder, if the bios can boot the kernel from the IDE 'drive', I assume coreboot will equally be able to boot it; but if recent kernels refuse to use the disk, I may end up being screwed anyway :S. Well I could always make a big initrd I suppose with everything in it ...
Sorry for ranting on this list :p
Oliver