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

[Langer, Christoph]

Hi Jaroslav,

now that we have an 11.0.7 branch, I started looking at this. First of all, I agree that 8225797 is something we'll want to have in 11u eventually, though it's quite invasive with all its predecessors.

I imported all your patches in the order below. Some of them apply with fuzz. The whole set does not even build on linux_x86. I had to remove the final patch for JDK-8225797 to get it built. I'll run the other patches through our regression test system and will hopefully be able to share some results on Monday.

However, I think that we should review and push each fix one by one, in the order they apply. And we should have a separate review thread for each fix. So I'd like to ask you to start sending out a review mail for JDK-8209850 as a first step.

Thanks & Best regards
Christoph

> -----Original Message-----
> From: jdk-updates-dev <jdk-updates-dev-bounces at openjdk.java.net> On
> Behalf Of Jaroslav Bachorík
> Sent: Freitag, 15. November 2019 16:46
> To: jdk-updates-dev <jdk-updates-dev at openjdk.java.net>
> Subject: [DMARC FAILURE] [11u] [JFR] RFR 8225797: OldObjectSample event
> creates unexpected amount of checkpoint data
> 
> Hi,
> 
> I would like to get this non-trivial JFR related backport reviewed. This
> backport is required because OldObjectSample event in the current (JDK 11)
> form will cause unlimited growth of the recording thus making this event
> unsuitable in production memory leak detection - which is its main purpose.
> 
> The backport for 8225797 requires several other changes to be backported as
> well. I am listing the changes and the related webrevs here in order to
> have all the pieces in one place. The webrevs, unfortunately, are showing
> cumulative changes, so please, disregard that information. Each webrev is
> generated for that particular commit.
> 
> The following tests were run successfully:
> - jdk_tier1
> - jdk_tier2
> - jdk_jfr
> - jdk_core
> 
> ---
> 
> Prerequisites:
> # 8209850 : Allow NamedThreads to use GlobalCounter critical sections
> - applied almost cleanly with only minimal changes to account for slightly
> different code structure (patch diff -
> http://cr.openjdk.java.net/~jbachorik/8209850/webrev.00/patch.diff)
> JIRA .   : https://bugs.openjdk.java.net/browse/JDK-8209850
> Webrev : http://cr.openjdk.java.net/~jbachorik/8209850/webrev.00/
> 
> # 8209976: Improve iteration over non-JavaThreads
> - applied almost cleanly with only minimal changes to account for slightly
> different code structure  (patch diff -
> http://cr.openjdk.java.net/~jbachorik/8209976/webrev.00/patch.diff)
> - JIRA . : https://bugs.openjdk.java.net/browse/JDK-8209976
> - Webrev : http://cr.openjdk.java.net/~jbachorik/8209976/webrev.00/
> <http://cr.openjdk.java.net/~jbachorik/8209850/webrev.00/>
> 
> # 8210024: JFR calls virtual is_Java_thread from ~Thread()
> - applied almost cleanly with only minimal changes to account for slightly
> different code structure  (patch diff -
> http://cr.openjdk.java.net/~jbachorik/8210024/webrev.00/patch.diff)
> - JIRA . : https://bugs.openjdk.java.net/browse/JDK-8210024
> - Webrev : http://cr.openjdk.java.net/~jbachorik/8210024/webrev.00/
> 
> # 8214850: Rename vm_operations.?pp files to vmOperations.?pp files
> - applied almost cleanly with only minimal changes to account for slightly
> different code structure  (patch diff -
> http://cr.openjdk.java.net/~jbachorik/8214850/webrev.00/patch.diff)
> - JIRA . : https://bugs.openjdk.java.net/browse/JDK-8214850
> - Webrev : http://cr.openjdk.java.net/~jbachorik/8214850/webrev.00/
> 
> # 8209802: Garbage collectors should register JFR types themselves to avoid
> build errors.
> - applied almost cleanly with only minimal changes to account for slightly
> different code locations for g1 sources (patch diff -
> http://cr.openjdk.java.net/~jbachorik/8209802/webrev.00/patch.diff)
> - JIRA . : https://bugs.openjdk.java.net/browse/JDK-8209802
> - Webrev : http://cr.openjdk.java.net/~jbachorik/8209802/webrev.00/
> 
> 
> Backport:
> # 8225797: OldObjectSample event creates unexpected amount of
> checkpoint
> data
> - the original patch had to be accommodated to the JDK 11 status of JFR to
> minimize the number of prerequisite backports and therefore it is slightly
> more complex (patch diff -
> http://cr.openjdk.java.net/~jbachorik/8225797/webrev.00/patch.diff)
> - JIRA . : https://bugs.openjdk.java.net/browse/JDK-8225797
> - Webrev : http://cr.openjdk.java.net/~jbachorik/8225797/webrev.00/
> 
> 
> Thanks!
> 
> -JB-