RFR (S): 8135025: Error message is repeated for large value at G1ConcRefinementThreads

[Kim Barrett]

On Sep 8, 2015, at 5:36 PM, Thomas Schatzl <thomas.schatzl at oracle.com> wrote:
> 
>>>   56     *ecode = JNI_ENOMEM;
>>>   57     vm_shutdown_during_initialization("Could not create ConcurrentG1Refine");
>>>   58     return NULL;
>> vm_shutdown_during_initialization() just prints out error message while 
>> vm_exit_during_initialization() finally aborts.
> 
> Okay, that's what I was missing. Thanks for the clarification.

vm_shutdown_during_initialization does more than just print an error message.
However, it doesn’t completely terminate the VM; instead that’s supposed to be
handled by the caller.  I think vm_shutdown_during_initialization might be a
mistake that should be refactored away; I’ve been looking at it as part of the VM
exit processing discussion I’ve been having with Dan and David.  However, I
think this use is consistent with other uses and with what was there before
Sangheon’s proposed changes.  Doing anything about vm_shutdown_during_initialization
should be part of some other RFE.  It would probably be GC-related because
nearly all present uses of that function are in GC initialization code.