8255131: G1CollectedHeap::is_in() returns wrong result

[tschatzl]

Hi all,

please review this change to exchange the broken G1CollectedHeap::is_in() by G1CollectedHeap::is_in_exact() that has the same performance profile and correct.

G1CollectedHeap::is_in() returns true for any region ever committed (and crashes for not yet committed regions), while is_in_exact() (almost) did the right thing. Almost because HeapRegionManager::is_available() would assert if the given pointer is outside the heap and it's not guarded by is_in_reserved.

Test: tier1

Thanks,
Thomas

---

Progress

• Change must not contain extraneous whitespace
• Commit message must refer to an issue
• Change must be properly reviewed

Issue

• JDK-8255131: G1CollectedHeap::is_in() returns wrong result

Reviewers

• Stefan Johansson (@kstefanj - Reviewer) ⚠️ Review applies to 0fa2e85
• Kim Barrett (@kimbarrett - Reviewer)
• Albert Mingkun Yang (@albertnetymk - Author)

Download

$ git fetch https://git.openjdk.java.net/jdk pull/799/head:pull/799
$ git checkout pull/799

[bridgekeeper]

👋 Welcome back tschatzl! 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]

@tschatzl 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 - Incremental (df181bb)
• 00: Full (0fa2e85)

[kstefanj]

Looks good!

[openjdk]

@tschatzl 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:

8255131: G1CollectedHeap::is_in() returns wrong result

Reviewed-by: sjohanss, kbarrett, ayang

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 89 new commits pushed to the master branch:

• 107fb9c: 8255049: Remove support for the hsdis decode_instructions entry point in hotspot
• c1524c5: 8255208: CodeStrings passed to Disassembler::decode are ignored
• 8e5dff0: 8255213: Shenandoah build fails without JFR after JDK-8255036
• fe74f3c: 8255224: x86_32 tests fail with "bad AD file" after JDK-8223051
• 64dc4b1: 8255225: compiler/aot tests fail on Windows with NPE during artifact resolution
• a824781: 8255210: [Vector API] jdk/incubator/vector/Int256VectorTests.java crashes on AVX512 machines
• 2ca7a08: 8255268: 32-bit failures in runtime/Metaspace/elastic
• d8d9197: 8242559: Clean the "libawt_xawt" library from code for headless mode
• ff5f226: 8255086: Update the root locale display names
• 8afdcae: 8255031: Update java/util/prefs/AddNodeChangeListener.java to report more failure info
• ... and 79 more: https://git.openjdk.java.net/jdk/compare/21e67e54604e60f57106be32d48ba75ebf8c1130...master

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.

[kimbarrett]

This line is misindented. But rather than adjusting the indentation, how about instead changing the function to

bool G1CollectedHeap::is_in(const void* p) const {
  return is_in_reserved(p) && _hrm->is_available(addr_to_region((HeapWord*)p));
}

[kstefanj]

I was thinking something similar, so I would approve going with this proposal.

[tschatzl]

Changed as requested.

Had a discussion with ayang about the specification of CollectedHeap::is_in(): it reads
// Returns "TRUE" iff "p" points into the committed areas of the heap. // This method can be expensive so avoid using it in performance critical // code.
but ZGC also checks if p is < top() within the region. This seems a good idea, but needs a change of that "specification" (and more testing although I think G1 fulfills that as well), so I'd like to postpone this change for now.

[tschatzl]

Thanks @kimbarrett @albertnetymk @kstefanj for your reviews!

/integrate

[openjdk]

@tschatzl Since your change was applied there have been 89 commits pushed to the master branch:

• 107fb9c: 8255049: Remove support for the hsdis decode_instructions entry point in hotspot
• c1524c5: 8255208: CodeStrings passed to Disassembler::decode are ignored
• 8e5dff0: 8255213: Shenandoah build fails without JFR after JDK-8255036
• fe74f3c: 8255224: x86_32 tests fail with "bad AD file" after JDK-8223051
• 64dc4b1: 8255225: compiler/aot tests fail on Windows with NPE during artifact resolution
• a824781: 8255210: [Vector API] jdk/incubator/vector/Int256VectorTests.java crashes on AVX512 machines
• 2ca7a08: 8255268: 32-bit failures in runtime/Metaspace/elastic
• d8d9197: 8242559: Clean the "libawt_xawt" library from code for headless mode
• ff5f226: 8255086: Update the root locale display names
• 8afdcae: 8255031: Update java/util/prefs/AddNodeChangeListener.java to report more failure info
• ... and 79 more: https://git.openjdk.java.net/jdk/compare/21e67e54604e60f57106be32d48ba75ebf8c1130...master

Your commit was automatically rebased without conflicts.

Pushed as commit c520469.

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