stlc45xx needs pretty recent kernel because of wi-fi stack it depends on. It is great we have it but it got dragged down by the rest of the hardware/software being not supported in recent kernel. For stlc45xx being effective it would also need to push everything from nokia kernel into linux-omap or mainline. It could be done now by community too but it takes far more time and effort then doing it by tighter collaboration with upstream when 2.6.21 was current. Still I see N8x0 linux kernel situation to be unfortunate, there are many things that could be done by community now but looks like there is few people with motivation and both time and skills to hack the kernel. I hope this will be better with N900 because OMAP3 is more open. I don't know about any other OMAP2 based device with community interested in maintaining support in linux kernel except us. Fortunately for OMAP3 there is plenty of them.

IMO dsme does not count. DSME got opened by ripping out closed parts and moving them into mce which is still closed. Mostly we got not so interesting (i.e already reverse-engineered) and almost empty shell. Again it is of course better than nothing but still I consider dsme to be 99% of 'too little too late' variety.