FW: [DMARC FAILURE] [11u] [JFR] RFR 8225797: OldObjectSample event creates unexpected amount of checkpoint data

[Langer, Christoph]

Hi Jaroslav, all,

I’m done with testing your backport patch for JDK-8225797 now.

So, this was your original proposal:
JBS: https://bugs.openjdk.java.net/browse/JDK-8225797
Original Change: https://hg.openjdk.java.net/jdk/jdk/rev/caa25ab47aca
Initial backport Webrev: http://cr.openjdk.java.net/~jbachorik/8225797/webrev.00/

First of all, to make the debug build work, I had to do a very few modifications (incremental diff to your original backport): http://cr.openjdk.java.net/~clanger/webrevs/8225797.fix-debug-incremental.11u/
It was mainly about the ClassLoaderDataGraph_lock and cld->is_unsafe_anonymous() which don’t exist in jdk11 but were used in assertions. So this would only be noticed in debug builds.

So this is the final webrev with the changes worked in: http://cr.openjdk.java.net/~clanger/webrevs/8225797.11u/

Please review.

When pushing this, we also need to push the fix for https://bugs.openjdk.java.net/browse/JDK-8231025 (Incorrect method tag offset for big endian platform), otherwise we’ll see crashes and test failures in big endian platforms like sun_64 or linuxppc64 and AIX. However, this change applies with minor fuzz, so it should just be tagged with a jdk11u-fix-request label.

I also suggest to backport JDK-8231081 (TestMetadataRetention fails due to missing symbol id) immediately after 8225797 but I’ll send out a separate RFR for this.

Best regards
Christoph