Ah, after reading Mike's post I see that I am confusing devfs and devfsd (the devfs daemon). But the conclusion is the same; something that makes /dev/hda8 exist isn't getting started.
Perhaps another thing you could try would be (when dropped to a shell during boot failure) to see if the long/proper name of the device exists. Poke around in /dev, see if there's anything in there. Look for something like: /dev/ide/host0/bus0/target0/lun0/part8 or similar. (I got that from my gentoo laptop; looks like I do have devfs after all, lol! I just never bothered to look as it's not my main machine). Obviously some of the numbers may be different. Anyhow, if you could still access that then you could point fsck at that directly. Or mount. Or anything you do manually. Or you could try creating a symlink /dev/hda8 that points to it. If you can get your filesystem mounted, then you can go fix whatever's breaking devfsd (looks to be console.perms, so a good idea would be to reverse the change you made the other day and see what happens ;-) Good luck. Cheers, Gareth