Performance of DefaultISMLocking.releaseReadLock()

[Thomas Auinger]

Hello,
 
I am currently profiling a single-threaded Jackrabbit 2.8 application, which is 
reading 1800 nodes into POJOs (we store entity data in JCR).
 
VisualVM sampling tells me, that 57% of all time is spend in method 
org.apache.jackrabbit.core.state.DefaultISMLocking.releaseReadLock()
 
Now while that method does not look very complicated, I was wondering if there 
are known tips how to speed up locking (wiki page says its "TODO") or if 
anybody has helpful thoughts on this.
 
Thanks in advance
 
Tom