This JIRA is no longer relevant. The IN_NATIVE states are used for native calls, so long as those native calls are not JNI calls (for example - any I/O calls that go through our own "faster" Filesystem API end up being IN_NATIVE rather than IN_JNI). Thus, the naming scheme is accurate. It may not be perfect, given that the semantics of our "native" calls are different than our "JNI" calls, but it does not seem to be worth a fix at this point in time.
This JIRA is no longer relevant. The IN_NATIVE states are used for native calls, so long as those native calls are not JNI calls (for example - any I/O calls that go through our own "faster" Filesystem API end up being IN_NATIVE rather than IN_JNI). Thus, the naming scheme is accurate. It may not be perfect, given that the semantics of our "native" calls are different than our "JNI" calls, but it does not seem to be worth a fix at this point in time.