RFR(S): JDK-8054194 jstack crash: assert(handle != NULL) failed: JNI handle should not be null

[Dmitry Samersoff]

Staffan,

On 2014-09-24 13:50, Staffan Larsen wrote:
> Wouldn’t “path == NULL” be more idiomatic than “path == 0”?

Agree!

> On a process note, I think it would be better if you push this fix with a new bug and leave the original bug (8054194) open. I think this is better since you are not really fixing 8054194. The new bug could say something like “SA: add value checking in setImageAndSymbolPath()”. 

After the fix we will not get assert(handle != NULL) anymore, so all
DKFL rules become invalid.

Also, it's secondary crash, so we would see jstack exception in jstack
output file rather than in output of test it self.

So I would prefer to push the changes under this CR - as it solves exact
problem mentioned in synopses and DKFL and file a new CR (may be in
hotspot/runtime category)

-Dmitry

> 
> 
> Thanks,
> /Staffan
> 
> On 24 sep 2014, at 11:36, Dmitry Samersoff <dmitry.samersoff at oracle.com> wrote:
> 
>> Hi Everybody,
>>
>> Please review the fix:
>>
>> http://cr.openjdk.java.net/~dsamersoff/JDK-8054194/webrev.02/
>>
>> Time to time, attaching jstack to test being killed by timeout, we get a
>> secondary crash from JVM running jstack. The crash is intermittent and
>> hard to reproduce.
>>
>> This fix doesn't address the root cause (I'm going to file a new CR to
>> have it investigated) but intended to reduce a noise in nightly tests  -
>> we will get meaningful exception from jstack rather than VM crash.
>>
>> -Dmitry
>>
>> -- 
>> Dmitry Samersoff
>> Oracle Java development team, Saint Petersburg, Russia
>> * I would love to change the world, but they won't give me the sources.
> 


-- 
Dmitry Samersoff
Oracle Java development team, Saint Petersburg, Russia
* I would love to change the world, but they won't give me the sources.