ARM: Support for JVMTI notifications from JIT compiler on code generation

[Pavel Tisnovsky]

Andrew Dinn wrote:
> Ok, here is the patch revised in the light of the previous feedback.
> Both the patch /as was/ and /as is/ after this revision have been run on
> a custom kernel tweaked so that oprofile operates correctly and are
> generating coherent profile output. So, this can be considered to be tested.
> 
> One thing no one commented on was the use of the
> JvmtiJavaThreadEventTransition marker and corresponding lack of any
> HandleMark or other markers before calling out to the agent in the added
> method JvmtiExport::post_compiled_method_load().
> 
> Anyone got anything to add on this or other aspects of the patch? Or can
> I go ahead and commit this?
> 
> regards,
> 
> 
> Andrew Dinn
> -----------
> 

Looks ok,

I only saw two typos:

s/corrsponding/corresponding
s/thence/hence

And I have this question:

address_bci_map_add() - does not it need any check for address_bci_map_length value?
Are you sure it will newer be higher that THUMB2_MAX_BYTECODE_SIZE?

Cheers,
Pavel