>  Problem found and fix provided by David A. Holland, PR/36610.
 >  Actually, relock is not needed here, and mutex would be unlocked
 >  only on nfs_sigintr() fail case.

I assumed from the context that it wasn't good to call nfs_sigintr
holding the mutex, so I preserved that behavior in my patch... oh
well, guess I should have checked more deeply. :-/

-- 
   - David A. Holland / dholland@eecs.harvard.edu