Events happen in raid arrays, such as device failure and resync. It would be nice to know when they happen. polling /proc/mdstat is one option, but it isn't perfect.
When handling multipath arrays, failures need to be dealt with quickly (rebalance/ retest path). Polling frequently is a waste.
I developed some code to allow event notification through 'poll' or 'select' on mdstat. It is fairly course, but is a usable start. This was done with Doug Ledford because redhat needed it.
This should be review and submitted for inclusion in 2.6.