> Perhaps a good way to analyze what's going wrong is to add some kind
> of logging to the VM system; allocate (say) a 1MB buffer in kernel
> memory (wired, presumably) and log each decision by the VM system.
> Then, when you have driven the system into the pathological state,
> freeze the log and dump it out.  (Recognizing that automatically might
> be hard.)

Um, this may be a stupid idea, but when designing custom hardware I
often dump debug data into serial port.  Maybe you could just print
_short_ messages (single characters, if possible) to serial port at
different locations in the code and then capture this data using
another machine.  If you print mostly 1-char messages, you can get
11520 of those out in one second (at 115200 bps).

How does that sound?

  -jm

-- 

                     ---> http://www.jmp.fi/~jmarin/ <---