[fixed] Fix the linux yield logic

[CuteAlien]

Yeah, that was the original question - does nanosleep(0) give up the processor. I've searched for an answer to that rather long yesterday and I think it doesn't and we really need to use 1. It guarantees only that it gives up the processor until sleep-time is over, the behavior for 0 is undefined - I even stumbled upon some thread on kernel mailing list where it seems that people had run into the same thing when this behavior was changed at some time.