RFR: JDK-8262068: Improve G1 Full GC by skipping compaction for regions with high survival ratio

[Stefan Johansson]

On Fri, 5 Mar 2021 09:27:30 GMT, Hamlin Li <mli at openjdk.org> wrote:

>> Even if the re-run passes this seems to be a real problem, probably just a bit intermittent. Looking at the test output you can see there is a segmentation fault in:
>> `G1FullGCPrepareTask::G1CalculatePointersClosure::prepare_for_skipping_compaction(...)`
>> 
>> Running the test with debug builds might make the issue reproduce on more platforms and give hints on what the problem is.
>> 
>> Hope this helps,
>> Stefan
>
> Thanks Stefan, we're investigating. (currently, we don't have mac env)
> BTW, before it's reproduced locally, is there any way to get more information from the online test result? I can only find the "View raw logs".

I don't know for sure, I tried downloading the "log archive" but it did not include the hs_err-file which would have been very useful. I can kick of an internal run to see if it reproduces here.

-------------

PR: https://git.openjdk.java.net/jdk/pull/2760