JDK7 SIGSEGV with G1 garbage collector

[Vladimir Kozlov]

Punya,

Can I add your test to our regression tests together with the fix?

http://cr.openjdk.java.net/~kvn/8023472/webrev/test/compiler/gcbarriers/G1CrashTest.java.html

I added our copyright header and pointed you as author.

Regards,
Vladimir

On 8/21/13 11:24 AM, Punya Biswal wrote:
> Hi Bengt,
>
> Thanks for the prompt response! Please let me know if I can provide
> anything else that will be useful.
>
> --
> Punya Biswal
> Palantir | Engineer
>
>
> On 8/21/13 6:28 AM, "Bengt Rutisson" <bengt.rutisson at oracle.com> wrote:
>
>>
>> Hi again,
>>
>> This seems to be a bug in the C2 compiler. I've filed JDK-8023472 to
>> track this. I'll come back with a URL once this bug is publicly available.
>>
>> Thanks again for the great bug report!
>> Bengt
>>
>> On 8/21/13 8:52 AM, Bengt Rutisson wrote:
>>>
>>> Hi Punya,
>>>
>>> Thank you for the excellent bug report!
>>>
>>> I cloned the test case and can reproduce the issue both with 7u25
>>> build 15 (that you used) and also with the latest build of JDK8.
>>>
>>> I'll dig some more into what is going on and file a bug. Will ping
>>> this thread when I have more information.
>>>
>>> Bengt
>>>
>>> On 8/21/13 12:05 AM, Punya Biswal wrote:
>>>> [Apologies if this message is delivered twice; I sent it once before I
>>>> realized this was a members-only list.]
>>>>
>>>> We're running into a JVM segfault when using a third-party library
>>>> (Elasticsearch) alongside the G1 garbage collector. This issue has been
>>>> reported elsewhere (for example,
>>>>
>>>> https://urldefense.proofpoint.com/v1/url?u=https://jira.terracotta.org/j
>>>> ira/browse/CDV-1651&k=fDZpZZQMmYwf27OU23GmAQ%3D%3D%0A&r=kTrYN051orSRhyA6
>>>> mqYxbjRIX%2BBCPm7thmzLC79vBeM%3D%0A&m=eQXvXMfzOBqOYQJCPTAI8m0yYLgOXGzbUN
>>>> Y%2FyEd6LjE%3D%0A&s=4517eb7bba43d15b07ba2a16457f9371f3d18509f2b13a22de37
>>>> 6bdb2e19e95f) and might already
>>>> be on
>>>> the OpenJDK roadmap. It's usually associated with the use of GNU Trove.
>>>>
>>>> We've been able to reduce the amount of code required to reproduce the
>>>> crash to a short program that reliably crashes on my computer
>>>> (a MacBook Pro) using only JDK core classes. I've also included the
>>>> hs_err_* diagnostic file; both are at
>>>>
>>>> https://urldefense.proofpoint.com/v1/url?u=https://gist.github.com/punya
>>>> /6287943&k=fDZpZZQMmYwf27OU23GmAQ%3D%3D%0A&r=kTrYN051orSRhyA6mqYxbjRIX%2
>>>> BBCPm7thmzLC79vBeM%3D%0A&m=eQXvXMfzOBqOYQJCPTAI8m0yYLgOXGzbUNY%2FyEd6LjE
>>>> %3D%0A&s=1a3b70ed491698a5a44b0fc5d03329444dbe1b093056d9609cf1ff8c57deb28
>>>> 8 .
>>>>
>>>> Does this reduced repro help understand what's going on?
>>>>
>>>>
>>>> A few points of interest:
>>>>
>>>> * running with -XX:+UseG1GC -Dcount=100000 always crashes
>>>> * running with -XX:+UseG1GC -Xint -Dcount=100000 never crashes
>>>> * running with -Xint -Dcount=100000 never crashes
>>>>
>>>>
>>>> Using -ea and/or -server don't any effect these results. With regard to
>>>> JVM versions,
>>>>
>>>> * it never crashes on JDK 6u51
>>>> * it crashes on JDK 7u25 and JDK 7u43 (prerelease)
>>>> * it never crashes on JDK 8 (prerelease)
>>>>
>>>>
>>>> I'm happy to provide a link to a core dump if that helps.
>>>>
>>>> --
>>>> Punya Biswal
>>>> Palantir | Engineer
>>>> pbiswal at palantir.com | 206 605 2397
>>>
>>