OpenJDK code review request for jmethodID and JNI itable cache race fixes (6419370)
Daniel D. Daugherty
Daniel.Daugherty at Sun.COM
Sat Sep 12 20:22:18 PDT 2009
I'm looking for a couple of OpenJDK code reviewers for my fix to
the following bug:
6419370 4/4 new jmethodID code has tiny holes in synchronization
The synopsis is a bit misleading since two different caches are
- the jmethodID cache
- the JNI itable index cache
Both of these caches are used by the regular system and when JVM/TI
is used so this code is applicable outside my usual Serviceability
arena. That's why this code review request is so wide...
Here is the URL for the OpenJDK code review round 1 webrev:
I've updated the bug with descriptions of the races involved:
See Evaluation note #3 for a walk through that refers to both
old and new code.
I'm targeting this fix for HSX-17-B02 so I'd like to hear back
from at least one reviewer in a week or so...
Thanks, in advance, for any feedback!
More information about the hotspot-runtime-dev