Request for Review(xs) - 8155263: DisableStartThread should not be applied to GC worker threads
sangheon.kim at oracle.com
Tue Jul 19 18:11:50 UTC 2016
The change looks good to me too.
On 07/19/2016 10:47 AM, Jon Masamitsu wrote:
> Thanks, Thomas, for the Review.
> I still need one more review, please.
> On 7/12/2016 11:49 AM, Jon Masamitsu wrote:
>> Yes, Thomas, you guessed correctly. And so others don't have to guess.
>> On 7/12/2016 10:37 AM, Jon Masamitsu wrote:
>>> The flag DisableStartThread is defined for Java threads
>>> develop(bool, DisableStartThread, false, \
>>> "Disable starting of additional Java threads " \
>>> "(for debugging only)")
>>> but was being applied to GC threads in one case. I removed
>>> that use.
>>> Tested by comparing the GC threads before and after the
>>> change to verify that the concurrent GC threads were running.
>>> Using +DisableStartThread on the command line cause the Java
>>> threads not to run and under those circumstances a safepoint
>>> is not reached (Java threads never get there). Any of the test
>>> programs I tried hung at the first GC.
More information about the hotspot-gc-dev