8370880: [lworld] Split ciObjArrayKlass into ciRefArrayKlass and ciFlatArrayKlass

[merykitty]

Hi,

This PR splits ciObjArrayKlass into ciRefArrayKlass and ciFlatArrayKlass, aligns the hierarchy with the corresponding types of the VM.

Please kindly review, thanks a lot.

---

Progress

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

Issue

• JDK-8370880: [lworld] Split ciObjArrayKlass into ciRefArrayKlass and ciFlatArrayKlass (Enhancement - P2)

Reviewers

• Tobias Hartmann (@TobiHartmann - Committer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/valhalla.git pull/1739/head:pull/1739
$ git checkout pull/1739

Update a local copy of the PR:
$ git checkout pull/1739
$ git pull https://git.openjdk.org/valhalla.git pull/1739/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 1739

View PR using the GUI difftool:
$ git pr show -t 1739

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/valhalla/pull/1739.diff

Using Webrev

Link to Webrev Comment

[bridgekeeper]

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

[openjdk]

@merykitty 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:

8370880: [lworld] Split ciObjArrayKlass into ciRefArrayKlass and ciFlatArrayKlass

Reviewed-by: thartmann

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 no new commits pushed to the lworld branch. If another commit should be pushed before you perform the /integrate command, your PR will be automatically rebased. If you prefer to avoid any potential automatic rebasing, please check the documentation for the /integrate command for further details.

➡️ To integrate this PR with the above commit message to the lworld branch, type /integrate in a new comment.

[mlbridge]

Webrevs

• 03: Full (8e286e8d)
• 02: Full (cddbcb5b)
• 01: Full - Incremental (be614362)
• 00: Full (b401f2ff)

[merykitty]

There are still a lot of refactoring to be done regarding moving elements from ciArrayKlass to ciObjArrayKlass since now it is the new common ancestor of all arrays of objects.

[TobiHartmann]

Thank you for working on this, Quan-Anh! Testing looks good except for this:

compiler/valhalla/inlinetypes/TestArrayNullMarkers.java#AVF-nAVF
-XX:+IgnoreUnrecognizedVMOptions -XX:-TieredCompilation -XX:-DoEscapeAnalysis -XX:+AlwaysIncrementalInline

# A fatal error has been detected by the Java Runtime Environment:
#
#  Internal Error (/workspace/open/src/hotspot/share/ci/ciMetadata.hpp:105), pid=418142, tid=418159
#  assert(is_flat_array_klass()) failed: bad cast
#
# JRE version: Java(TM) SE Runtime Environment (26.0) (fastdebug build 26-jep401ea2-2025-11-17-0811237.tobias.hartmann.valhalla2)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (fastdebug 26-jep401ea2-2025-11-17-0811237.tobias.hartmann.valhalla2, mixed mode, compressed oops, compressed class ptrs, g1 gc, linux-amd64)
# Problematic frame:
# V  [libjvm.so+0x1ca876a]  TypeAryPtr::flat_layout_helper() const+0x7a

Current CompileTask:
C2:2887  380 % !b        compiler.valhalla.inlinetypes.TestArrayNullMarkers::main @ 698 (2967 bytes)

Stack: [0x00007f5639cfe000,0x00007f5639dfe000],  sp=0x00007f5639df8240,  free space=1000k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
V  [libjvm.so+0x1ca876a]  TypeAryPtr::flat_layout_helper() const+0x7a  (ciMetadata.hpp:105)
V  [libjvm.so+0x102a8a9]  GraphKit::get_layout_helper(Node*, int&)+0x299  (graphKit.cpp:4100)
V  [libjvm.so+0x102cd43]  GraphKit::new_array(Node*, Node*, int, Node**, bool, Node*)+0x43  (graphKit.cpp:4345)
V  [libjvm.so+0x158b712]  LibraryCallKit::inline_newArray(bool, bool)+0x242  (library_call.cpp:4904)
V  [libjvm.so+0x15b2c6a]  LibraryIntrinsic::generate(JVMState*)+0x22a  (library_call.cpp:130)
V  [libjvm.so+0xdbc50f]  Parse::do_call()+0xebf  (doCall.cpp:771)
V  [libjvm.so+0x18fee88]  Parse::do_one_bytecode()+0x4c8  (parse2.cpp:3526)
V  [libjvm.so+0x18e6007]  Parse::do_one_block()+0x357  (parse1.cpp:1703)
V  [libjvm.so+0x18e7300]  Parse::do_all_blocks()+0x130  (parse1.cpp:760)
V  [libjvm.so+0x18eae81]  Parse::Parse(JVMState*, ciMethod*, float)+0xea1  (parse1.cpp:664)
V  [libjvm.so+0x9e1975]  ParseGenerator::generate(JVMState*)+0x135  (callGenerator.cpp:99)
V  [libjvm.so+0x9e6ba3]  CallGenerator::do_late_inline_helper()+0xa03  (callGenerator.cpp:751)
V  [libjvm.so+0xbdc968]  Compile::inline_incrementally_one()+0x1b8  (compile.cpp:2614)
V  [libjvm.so+0xbde0ab]  Compile::inline_incrementally(PhaseIterGVN&)+0x38b  (compile.cpp:2707)
V  [libjvm.so+0xbe0c9e]  Compile::Optimize()+0x47e  (compile.cpp:2843)
V  [libjvm.so+0xbe45c5]  Compile::Compile(ciEnv*, ciMethod*, int, Options, DirectiveSet*)+0x1da5  (compile.cpp:879)
V  [libjvm.so+0x9de490]  C2Compiler::compile_method(ciEnv*, ciMethod*, int, bool, DirectiveSet*)+0x4e0  (c2compiler.cpp:149)
V  [libjvm.so+0xbf3e20]  CompileBroker::invoke_compiler_on_method(CompileTask*)+0x780  (compileBroker.cpp:2345)
V  [libjvm.so+0xbf5680]  CompileBroker::compiler_thread_loop()+0x530  (compileBroker.cpp:1989)
V  [libjvm.so+0x118d15b]  JavaThread::thread_main_inner()+0x13b  (javaThread.cpp:772)
V  [libjvm.so+0x1c68ee6]  Thread::call_run()+0xb6  (thread.cpp:243)
V  [libjvm.so+0x1893bc8]  thread_native_entry(Thread*)+0x128  (os_linux.cpp:883)

[merykitty]

Thanks for your testing. Unfortunately, the issue cuts pretty deep. We have multiple potential inconsistencies when dealing with _klass such as when trying to obtain a flat TypeAryPtr from a non-flat TypeAryPtr. I think it can be addressed as follow-up issues. To fix the immediate crash, I choose to fix a particular issue, that is we can only compute the _klass if we have klass_is_exact.

[merykitty]

Too bad it seems pretty hard to walk on the thin line right now. I think it is better to tighten the type system around arrays first.

[TobiHartmann]

Right, I think so too, the latest version hits more issues :(

[openjdk]

@merykitty this pull request can not be integrated into lworld due to one or more merge conflicts. To resolve these merge conflicts and update this pull request you can run the following commands in the local repository for your personal fork:

git checkout ciRefArray
git fetch https://git.openjdk.org/valhalla.git lworld
git merge FETCH_HEAD
# resolve conflicts and follow the instructions given by git merge
git commit -m "Merge lworld"
git push

[merykitty]

I think this PR is ready for review again.

[TobiHartmann]

Looks good to me otherwise. Thanks!

[TobiHartmann]

UseCompressedOops is always true here, right? It's checked in line 3685. Also above.

[merykitty]

You are right, I have removed those check.

[TobiHartmann]

Thanks, looks good!

[merykitty]

Thanks a lot for your reviews!

/integrate

[openjdk]

Going to push as commit 9aff0ae.

[openjdk]

@merykitty Pushed as commit 9aff0ae.

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