RFR(M): 8204615: [lworld] C2 support for java.lang.Object methods on value types

Roland Westrelin rwestrel at redhat.com
Thu Jul 5 14:36:01 UTC 2018


> Yes, we deoptimize with the pending exception that was installed by CHECK_THROW_NOSYNC_IMSE if the
> obj is a value type (i.e., always locked) in synchronizer.cpp.

Wouldn't it be simpler to clear the exception, deoptimize and let the
interpreter re-execute the monitorenter and throw the exception?

Roland.


More information about the valhalla-dev mailing list