-
Notifications
You must be signed in to change notification settings - Fork 6.1k
8356587: Missing object ID X in pool jdk.types.Method #26458
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: master
Are you sure you want to change the base?
Conversation
👋 Welcome back mgronlun! A progress list of the required criteria for merging this PR into |
@mgronlun This change now passes all automated pre-integration checks. ℹ️ This project also has non-automated pre-integration requirements. Please see the file CONTRIBUTING.md for details. After integration, the commit message for the final commit will be:
You can use pull request commands such as /summary, /contributor and /issue to adjust it as needed. At the time when this comment was updated there had been 75 new commits pushed to the
As there are no conflicts, your changes will automatically be rebased on top of these commits when integrating. If you prefer to avoid this automatic rebasing, please check the documentation for the /integrate command for further details. ➡️ To integrate this PR with the above commit message to the |
Webrevs
|
/label remove hotspot |
@mgronlun |
/label add hotspot-jfr |
@mgronlun |
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:
Address a regression introduced by JDK-835221. 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().
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.
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
Progress
Issue
Reviewers
Reviewing
Using
git
Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/26458/head:pull/26458
$ git checkout pull/26458
Update a local copy of the PR:
$ git checkout pull/26458
$ git pull https://git.openjdk.org/jdk.git pull/26458/head
Using Skara CLI tools
Checkout this PR locally:
$ git pr checkout 26458
View PR using the GUI difftool:
$ git pr show -t 26458
Using diff file
Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/26458.diff
Using Webrev
Link to Webrev Comment