RFR (XS): 8217765: Internal Error (javaCalls.cpp:61) guarantee(thread->can_call_java()) failed

David Holmes david.holmes at oracle.com
Thu Feb 21 00:39:35 UTC 2019


On 21/02/2019 1:12 am, coleen.phillimore at oracle.com wrote:
> This looks good (with corrections to logging suggestions made by Robbin).

Thanks Coleen!

David

> thanks,
> Coleen
> 
> On 2/19/19 7:53 PM, David Holmes wrote:
>> Bug: https://bugs.openjdk.java.net/browse/JDK-8217765
>> webrev: http://cr.openjdk.java.net/~dholmes/8217765/webrev.13/
>>
>> A JDK 11 user hit what should be an "impossible" code path in the new 
>> nestmate logic, whereby a compiler thread attempted to load a class 
>> and failed because it can't execute Java code. We have not been able 
>> to get enough information to determine exactly how this arises so I 
>> have simply put together a defensive fix that skips the attempted load 
>> if in a thread that can't run Java - the same technique that is 
>> already used in another part of the nestmate logic.
>>
>> There's no regression test as I cannot determine a way to make this 
>> code path execute. But the submitter has tested it with their 
>> application and their problem has been resolved.
>>
>> Other testing: mach5 tiers 1-3
>>
>> Thanks,
>> David
> 


More information about the hotspot-runtime-dev mailing list