Critical JNI Native method lookup and thread state

[Doerr, Martin]

Hi everybody,

I have a question about the thread state during lookup of Critical JNI Native methods
(in NativeLookup::lookup_critical_style).

What I observed is that the native lookup gets called via SharedRuntime::resolve_static_call_C
and generate_native_wrapper and it seems like we are all the way in state "_thread_in_vm".

This might be expected. However, a call to os::dll_load might take some time due to IO.
So I would expect a transition to state "_thread_in_native" to avoid safepoint issues, but
I didn't find such a transition.

Did I miss anything or is it intended to go all the way in state "_thread_in_vm"?
Or at which place should we transition to "_thread_in_native"?

Kind regards,
Martin

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20121102/f6c93226/attachment.html