8327997: G1: Move G1ScanClosureBase::reference_iteration_mode to subclass

[albertnetymk]

Simple moving a method from super class to sub class and some documentation.

---

Progress

• Change must be properly reviewed (1 review required, with at least 1 Reviewer)
• Change must not contain extraneous whitespace
• Commit message must refer to an issue

Issue

• JDK-8327997: G1: Move G1ScanClosureBase::reference_iteration_mode to subclass (Enhancement - P4)

Reviewers

• Guoxiong Li (@lgxbslgx - Committer) ⚠️ Review applies to 5513c4a6
• Thomas Schatzl (@tschatzl - Reviewer)
• Stefan Johansson (@kstefanj - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/18244/head:pull/18244
$ git checkout pull/18244

Update a local copy of the PR:
$ git checkout pull/18244
$ git pull https://git.openjdk.org/jdk.git pull/18244/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 18244

View PR using the GUI difftool:
$ git pr show -t 18244

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/18244.diff

Webrev

Link to Webrev Comment

[bridgekeeper]

👋 Welcome back ayang! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

[openjdk]

@albertnetymk The following label will be automatically applied to this pull request:

• hotspot-gc

When this pull request is ready to be reviewed, an "RFR" email will be sent to the corresponding mailing list. If you would like to change these labels, use the /label pull request command.

[mlbridge]

Webrevs

• 01: Full (5a154212)
• 00: Full (5513c4a6)

[lgxbslgx]

Looks good. The root cause of such movement is that we don't create an specific object of the class G1ScanClosureBase. So we can override the method reference_iteration_mode in its subclasses instead of G1ScanClosureBase itself.

If we need to write so much comments to explain such movement, I suggest that we should write more comments in src/hotspot/share/memory/iterator.hpp::ReferenceIterationMode to explain the related things clearer.

[openjdk]

@albertnetymk 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:

8327997: G1: Move G1ScanClosureBase::reference_iteration_mode to subclass

Reviewed-by: gli, tschatzl, sjohanss

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 1 new commit pushed to the master branch:

• 1281e18: 8325613: CTW: Stale method cleanup requires GC after Sweeper removal

Please see this link for an up-to-date comparison between the source branch of this pull request and the master branch.
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 master branch, type /integrate in a new comment.

[tschatzl]

I think the comment is okay, but maybe should put more emphasis on that specifying this iteration mode allows the compiler to (more easily) compile out the check in try_discover.

Also since this isn't the only place this (and similar) optimizations are made elsewhere too using this getter, I kind of agree with @lgxbslgx that the impact of the selection should probably be documented at the enum level (the move of the override into the leaf class here is fine with me).

Just something like "// Selection of one or the other option helps the compiler to remove unnecessary, known beforehand behavior for a given oop closure at compile time" seems sufficient to me there instead of this fairly long comment.
The exact changes can be easily gathered by searching the usages of the enum.

[albertnetymk]

I dropped the comment. This optimization, afaics, is only used/implemented in G1. Moving the comment to the enum level can be confusing, since its impact might not be visible (or even positive) for other collectors. (I inspected the generated asm of g1Remset with and without this optimization and they are indeed different -- the one without the optimization is shorter (in terms of LOC), but it's unclear which one is faster.)

[tschatzl]

I can see multiple references to the DO_FIELDS value in Hotspot code in that getter in several GCs, but only grepped for it without looking into detail about the reasons.

[openjdk]

@albertnetymk Please do not rebase or force-push to an active PR as it invalidates existing review comments. Note for future reference, the bots always squash all changes into a single commit automatically as part of the integration. See OpenJDK Developers’ Guide for more information.

[albertnetymk]

Thanks for review.

/integrate

[openjdk]

Going to push as commit 954c50e.
Since your change was applied there have been 7 commits pushed to the master branch:

• 6f8b0a3: 8327799: JFR view: the "Park Until" field of jdk.ThreadPark is invalid if the parking method is not absolute
• 1d34b74: 8321021: RISC-V: C2 VectorUCastB2X
• 3b9255e: 8325851: Hide PassFailJFrame.Builder constructor
• ad0f329: 8327787: Convert javax/swing/border/Test4129681.java applet test to main
• a43c3cc: 8327826: Convert javax/swing/border/Test4243289.java applet test to main
• 11a3673: 8328110: Allow simultaneous use of PassFailJFrame with split UI and additional windows
• 1281e18: 8325613: CTW: Stale method cleanup requires GC after Sweeper removal

Your commit was automatically rebased without conflicts.

[openjdk]

@albertnetymk Pushed as commit 954c50e.

💡 You may see a message that your pull request was closed with unmerged commits. This can be safely ignored.