July 2025 - hotspot-jfr-dev

Re: RFR: 8364258: ThreadGroup constant pool serialization is not normalized [v4]
by Markus Grönlund 31 Jul '25

31 Jul '25

> Greetings, > > The following change set normalizes the ThreadGroup constants being serialized to the .jfr binary file. > > It does this in two parts: > > 1. Re-inserts a scavenging scheme to remove stale (unloaded) ThreadGroups; some logic that we used to have in place, but may have been lost as part of the major refactoring around JFR Event Streaming, or even earlier. > 2. Introduces a per-epoch "is_serialized" scheme, building on top of JFR epoch generations. > > Problem manifestation is a very high number of duplicated TGs (jdk.types.ThreadGroup) as part of JFR checkpoints. > > Testing: jdk_jfr > > Thanks > Markus Markus Grönlund has updated the pull request incrementally with one additional commit since the last revision: extend is_serialized scheme to JavaThreads ------------- Changes: - all: https://git.openjdk.org/jdk/pull/26558/files - new: https://git.openjdk.org/jdk/pull/26558/files/3434343a..e973f67b Webrevs: - full: https://webrevs.openjdk.org/?repo=jdk&pr=26558&range=03 - incr: https://webrevs.openjdk.org/?repo=jdk&pr=26558&range=02-03 Stats: 38 lines in 3 files changed: 29 ins; 0 del; 9 mod Patch: https://git.openjdk.org/jdk/pull/26558.diff Fetch: git fetch https://git.openjdk.org/jdk.git pull/26558/head:pull/26558 PR: https://git.openjdk.org/jdk/pull/26558

1 0

Re: RFR: 8364258: ThreadGroup constant pool serialization is not normalized [v3]
by Markus Grönlund 31 Jul '25

31 Jul '25

> Greetings, > > The following change set normalizes the ThreadGroup constants being serialized to the .jfr binary file. > > It does this in two parts: > > 1. Re-inserts a scavenging scheme to remove stale (unloaded) ThreadGroups; some logic that we used to have in place, but may have been lost as part of the major refactoring around JFR Event Streaming, or even earlier. > 2. Introduces a per-epoch "is_serialized" scheme, building on top of JFR epoch generations. > > Problem manifestation is a very high number of duplicated TGs (jdk.types.ThreadGroup) as part of JFR checkpoints. > > Testing: jdk_jfr > > Thanks > Markus Markus Grönlund has updated the pull request incrementally with one additional commit since the last revision: remove unused setter because of intrusive list ------------- Changes: - all: https://git.openjdk.org/jdk/pull/26558/files - new: https://git.openjdk.org/jdk/pull/26558/files/8c025caf..3434343a Webrevs: - full: https://webrevs.openjdk.org/?repo=jdk&pr=26558&range=02 - incr: https://webrevs.openjdk.org/?repo=jdk&pr=26558&range=01-02 Stats: 1 line in 1 file changed: 0 ins; 1 del; 0 mod Patch: https://git.openjdk.org/jdk/pull/26558.diff Fetch: git fetch https://git.openjdk.org/jdk.git pull/26558/head:pull/26558 PR: https://git.openjdk.org/jdk/pull/26558

1 0

Integrated: 8364090: Dump JFR recording on CrashOnOutOfMemoryError
by Yasumasa Suenaga 31 Jul '25

31 Jul '25

On Fri, 25 Jul 2025 00:52:29 GMT, Yasumasa Suenaga <ysuenaga(a)openjdk.org> wrote: > JFR emergency dump would happen when OOM was thrown. However it would not contain most recent `OldObjectSample` events emitted by LeakProfiler. > > I [reported this issue in past](https://mail.openjdk.org/pipermail/hotspot-jfr-dev/2019-January/00038…, and it seems to be difficult to fix soon, and also JDK codebase has been updated in several years. It brings us to fix this issue easier than past. > > So I propose again to emit the events from LeakProfiler when OOM happened. > This change passed `jdk_jfr` tests on Linux x64 (excepts TestHeapSummaryEventPSParOld.java reported in [JDK-8364082](https://bugs.openjdk.org/browse/JDK-8364082)) > > Related email thread: https://mail.openjdk.org/pipermail/hotspot-jfr-dev/2025-July/008007.html This pull request has now been integrated. Changeset: 8ed214f3 Author: Yasumasa Suenaga <ysuenaga(a)openjdk.org> URL: https://git.openjdk.org/jdk/commit/8ed214f3b1864ea0095d05497f782ce4131836d4 Stats: 138 lines in 7 files changed: 123 ins; 3 del; 12 mod 8364090: Dump JFR recording on CrashOnOutOfMemoryError Reviewed-by: egahlin, stuefe ------------- PR: https://git.openjdk.org/jdk/pull/26468

1 0

Re: RFR: 8364090: Dump JFR recording on CrashOnOutOfMemoryError [v6]
by Yasumasa Suenaga 31 Jul '25

31 Jul '25

> JFR emergency dump would happen when OOM was thrown. However it would not contain most recent `OldObjectSample` events emitted by LeakProfiler. > > I [reported this issue in past](https://mail.openjdk.org/pipermail/hotspot-jfr-dev/2019-January/00038…, and it seems to be difficult to fix soon, and also JDK codebase has been updated in several years. It brings us to fix this issue easier than past. > > So I propose again to emit the events from LeakProfiler when OOM happened. > This change passed `jdk_jfr` tests on Linux x64 (excepts TestHeapSummaryEventPSParOld.java reported in [JDK-8364082](https://bugs.openjdk.org/browse/JDK-8364082)) > > Related email thread: https://mail.openjdk.org/pipermail/hotspot-jfr-dev/2025-July/008007.html Yasumasa Suenaga has updated the pull request incrementally with one additional commit since the last revision: Refactor TestEmergencyDumpAtOOM.java ------------- Changes: - all: https://git.openjdk.org/jdk/pull/26468/files - new: https://git.openjdk.org/jdk/pull/26468/files/a20d56a7..544e7415 Webrevs: - full: https://webrevs.openjdk.org/?repo=jdk&pr=26468&range=05 - incr: https://webrevs.openjdk.org/?repo=jdk&pr=26468&range=04-05 Stats: 11 lines in 1 file changed: 3 ins; 8 del; 0 mod Patch: https://git.openjdk.org/jdk/pull/26468.diff Fetch: git fetch https://git.openjdk.org/jdk.git pull/26468/head:pull/26468 PR: https://git.openjdk.org/jdk/pull/26468

2 2

Re: RFR: 8364258: ThreadGroup constant pool serialization is not normalized [v2]
by Markus Grönlund 30 Jul '25

30 Jul '25

> Greetings, > > The following change set normalizes the ThreadGroup constants being serialized to the .jfr binary file. > > It does this in two parts: > > 1. Re-inserts a scavenging scheme to remove stale (unloaded) ThreadGroups; some logic that we used to have in place, but may have been lost as part of the major refactoring around JFR Event Streaming, or even earlier. > 2. Introduces a per-epoch "is_serialized" scheme, building on top of JFR epoch generations. > > Problem manifestation is a very high number of duplicated TGs (jdk.types.ThreadGroup) as part of JFR checkpoints. > > Testing: jdk_jfr > > Thanks > Markus Markus Grönlund has updated the pull request incrementally with one additional commit since the last revision: const Handle ------------- Changes: - all: https://git.openjdk.org/jdk/pull/26558/files - new: https://git.openjdk.org/jdk/pull/26558/files/cec818ab..8c025caf Webrevs: - full: https://webrevs.openjdk.org/?repo=jdk&pr=26558&range=01 - incr: https://webrevs.openjdk.org/?repo=jdk&pr=26558&range=00-01 Stats: 2 lines in 1 file changed: 0 ins; 0 del; 2 mod Patch: https://git.openjdk.org/jdk/pull/26558.diff Fetch: git fetch https://git.openjdk.org/jdk.git pull/26558/head:pull/26558 PR: https://git.openjdk.org/jdk/pull/26558

1 0

Re: RFR: 8364090: Dump JFR recording on CrashOnOutOfMemoryError [v4]
by Yasumasa Suenaga 30 Jul '25

30 Jul '25

> JFR emergency dump would happen when OOM was thrown. However it would not contain most recent `OldObjectSample` events emitted by LeakProfiler. > > I [reported this issue in past](https://mail.openjdk.org/pipermail/hotspot-jfr-dev/2019-January/00038…, and it seems to be difficult to fix soon, and also JDK codebase has been updated in several years. It brings us to fix this issue easier than past. > > So I propose again to emit the events from LeakProfiler when OOM happened. > This change passed `jdk_jfr` tests on Linux x64 (excepts TestHeapSummaryEventPSParOld.java reported in [JDK-8364082](https://bugs.openjdk.org/browse/JDK-8364082)) > > Related email thread: https://mail.openjdk.org/pipermail/hotspot-jfr-dev/2025-July/008007.html Yasumasa Suenaga has updated the pull request incrementally with one additional commit since the last revision: Add testcase ------------- Changes: - all: https://git.openjdk.org/jdk/pull/26468/files - new: https://git.openjdk.org/jdk/pull/26468/files/6cf417d8..5ecd4e43 Webrevs: - full: https://webrevs.openjdk.org/?repo=jdk&pr=26468&range=03 - incr: https://webrevs.openjdk.org/?repo=jdk&pr=26468&range=02-03 Stats: 111 lines in 1 file changed: 111 ins; 0 del; 0 mod Patch: https://git.openjdk.org/jdk/pull/26468.diff Fetch: git fetch https://git.openjdk.org/jdk.git pull/26468/head:pull/26468 PR: https://git.openjdk.org/jdk/pull/26468

3 7

Re: RFR: 8364090: Dump JFR recording on CrashOnOutOfMemoryError [v5]
by Yasumasa Suenaga 30 Jul '25

30 Jul '25

> JFR emergency dump would happen when OOM was thrown. However it would not contain most recent `OldObjectSample` events emitted by LeakProfiler. > > I [reported this issue in past](https://mail.openjdk.org/pipermail/hotspot-jfr-dev/2019-January/00038…, and it seems to be difficult to fix soon, and also JDK codebase has been updated in several years. It brings us to fix this issue easier than past. > > So I propose again to emit the events from LeakProfiler when OOM happened. > This change passed `jdk_jfr` tests on Linux x64 (excepts TestHeapSummaryEventPSParOld.java reported in [JDK-8364082](https://bugs.openjdk.org/browse/JDK-8364082)) > > Related email thread: https://mail.openjdk.org/pipermail/hotspot-jfr-dev/2025-July/008007.html Yasumasa Suenaga has updated the pull request incrementally with one additional commit since the last revision: Update TestEmergencyDumpAtOOM.java ------------- Changes: - all: https://git.openjdk.org/jdk/pull/26468/files - new: https://git.openjdk.org/jdk/pull/26468/files/5ecd4e43..a20d56a7 Webrevs: - full: https://webrevs.openjdk.org/?repo=jdk&pr=26468&range=04 - incr: https://webrevs.openjdk.org/?repo=jdk&pr=26468&range=03-04 Stats: 58 lines in 1 file changed: 22 ins; 11 del; 25 mod Patch: https://git.openjdk.org/jdk/pull/26468.diff Fetch: git fetch https://git.openjdk.org/jdk.git pull/26468/head:pull/26468 PR: https://git.openjdk.org/jdk/pull/26468

1 0

[jdk25] RFR: 8356587: Missing object ID X in pool jdk.types.Method
by Markus Grönlund 29 Jul '25

29 Jul '25

8356587: Missing object ID X in pool jdk.types.Method ------------- Commit messages: - Backport a34994476e8f4783c9f5a83a9c3db63ad605b323 Changes: https://git.openjdk.org/jdk/pull/26529/files Webrev: https://webrevs.openjdk.org/?repo=jdk&pr=26529&range=00 Issue: https://bugs.openjdk.org/browse/JDK-8356587 Stats: 303 lines in 24 files changed: 129 ins; 101 del; 73 mod Patch: https://git.openjdk.org/jdk/pull/26529.diff Fetch: git fetch https://git.openjdk.org/jdk.git pull/26529/head:pull/26529 PR: https://git.openjdk.org/jdk/pull/26529

2 2

Integrated: 8356587: Missing object ID X in pool jdk.types.Method
by Markus Grönlund 29 Jul '25

29 Jul '25

On Thu, 24 Jul 2025 12:49:36 GMT, Markus Grönlund <mgronlun(a)openjdk.org> wrote: > Greetings, > > The following change set addresses the data loss resulting in the assertion "Missing object ID X in pool jdk.types.Method". > > It involves three components: > > 1. Address a regression introduced by [JDK-835221](https://bugs.openjdk.org/browse/JDK-8352251). By locating JFR_ONLY(Jfr::check_and_process_sample_request(thread);) before the global_poll() in SafepointMechanism::process(), a stacktrace can be captured, and artifacts tagged, during a safepoint. This breaks an invariant as artifacts, i.e. methods, can be tagged in the wrong epoch (also a stacktrace can be stored in the wrong epoch). Must be moved to post global_poll(). > > 2. Retransform/Redefine classes include a non-safe copy of Method trace flags, leading to stale bits being set onto the new Methods. Method trace flags need to be copied, but must be done under a safepoint. > > 3. It seems that there has been an increase in the frequency of issuing calls to InstanceKlass::purge_previous_versions(), making scratch klasses and old methods disappear before JFR gets the chance to serialize also tagged old methods. Therefore, we need to ensure that we always tag the latest version of a Method. > > There is also a cleanup of gratuitous type conversions, from Klass* to InstanceKlass* in the newly introduced MethodTracing subsystem. > > Testing: jdk_jfr, stress testing > > Thank you > Markus This pull request has now been integrated. Changeset: a3499447 Author: Markus Grönlund <mgronlun(a)openjdk.org> URL: https://git.openjdk.org/jdk/commit/a34994476e8f4783c9f5a83a9c3db63ad605b323 Stats: 303 lines in 24 files changed: 129 ins; 101 del; 73 mod 8356587: Missing object ID X in pool jdk.types.Method Reviewed-by: egahlin ------------- PR: https://git.openjdk.org/jdk/pull/26458

1 0

Re: RFR: 8364090: Dump JFR recording on CrashOnOutOfMemoryError [v3]
by Yasumasa Suenaga 28 Jul '25

28 Jul '25

> JFR emergency dump would happen when OOM was thrown. However it would not contain most recent `OldObjectSample` events emitted by LeakProfiler. > > I [reported this issue in past](https://mail.openjdk.org/pipermail/hotspot-jfr-dev/2019-January/00038…, and it seems to be difficult to fix soon, and also JDK codebase has been updated in several years. It brings us to fix this issue easier than past. > > So I propose again to emit the events from LeakProfiler when OOM happened. > This change passed `jdk_jfr` tests on Linux x64 (excepts TestHeapSummaryEventPSParOld.java reported in [JDK-8364082](https://bugs.openjdk.org/browse/JDK-8364082)) > > Related email thread: https://mail.openjdk.org/pipermail/hotspot-jfr-dev/2025-July/008007.html Yasumasa Suenaga has updated the pull request incrementally with one additional commit since the last revision: Add emit_event_shutdown ------------- Changes: - all: https://git.openjdk.org/jdk/pull/26468/files - new: https://git.openjdk.org/jdk/pull/26468/files/ee27149e..6cf417d8 Webrevs: - full: https://webrevs.openjdk.org/?repo=jdk&pr=26468&range=02 - incr: https://webrevs.openjdk.org/?repo=jdk&pr=26468&range=01-02 Stats: 14 lines in 6 files changed: 4 ins; 0 del; 10 mod Patch: https://git.openjdk.org/jdk/pull/26468.diff Fetch: git fetch https://git.openjdk.org/jdk.git pull/26468/head:pull/26468 PR: https://git.openjdk.org/jdk/pull/26468

3 5