Up to now I have the OS working, in a mod'ed form (I hard coded the priority stuff to suit my app in the timer ISR), and two tasks attached to it. It's very tight for RAM, but it fits so far. In an 8052 would be no problem.

But I have a few speed issues. My board is a MIDI processing device, when it gets a bunch of bytes all at once it's not keeping up. MIDI is 31250 bauds, or a byte every 330us or so. When several complete message arrives, my code is not keeping up. I have some restruring to do to try to improve that, or I might try running at 24MHz with another chip.

But the OS is quite decent. I only have it compiling with Franklin at the moment. But it definitely does the trick.