RFR(XS) [8u] 8153267: nmethod's exception cache not multi-thread safe

[David Buck]

Hi!
(explicitly CCing everyone involved with the original JDK 9 review)

Please have a look at this very straightforward backport to 8u-dev of 
Martin's fix for 8153267.

8153267 bug report:   https://bugs.openjdk.java.net/browse/JDK-8153267

8153267 JDK 9 push: 
http://hg.openjdk.java.net/jdk9/jdk9/hotspot/rev/95b8ce0f1228

8u backport webrev:   http://cr.openjdk.java.net/~dbuck/8153267.0_jdk8u/

The only non-trivial difference between Martin's JDK 9 push and my 
backport is that I also included what remained of the change for 
8143897. 8143897 was pushed before 8153267 but then completely 
superseded by 8153267. Specifically, I included the addition of the 
"index" local (as opposed to the multiple calls to count()) from 8143897 
as it will keep both code lines (8 and 9) better in-sync and also makes 
the code more readable in my opinion. So you can think of the above 
webrev as a combined delta of both 8153267 and 8153267.

8143897 bug report:  https://bugs.openjdk.java.net/browse/JDK-8143897

8143897 jdk 9 push: 
http://hg.openjdk.java.net/jdk9/jdk9/hotspot/rev/f918c20107d9

All standard JPRT tests (default and "hotspot" testsets) run and passed.

Cheers,
-Buck