A. The process concerned is accessing an area of memory that SQL Server needs to protect against multiple access - typically on SMP servers. To this end it locks that area of memory (via a spin lock) so that any other access to it gets paused. In concept it is very similar to row/page locking in a database, but is used for internal memory areas within a program. There is plenty of documentation in the Win32 SDK/MSDN about spin locks and their usage.

If the SPINLOOP process does not give up control then it is very likely that SQL Server will become unresponsive. It is unlikely that you will be able to KILL a process in this state and you will probably have to kill SQL Server itself (use KILL command from the NT Resource Kit)

If you are suffering a lot of degradation due to spinloops, then you will need to raise a paid fault call with Microsoft PSS to resolve it.