8366020: Assert that java code is not executed during the AOT assembly phase by matias9927 · Pull Request #30345 · openjdk/jdk

matias9927 · 2026-03-20T19:11:09Z

User code should not be run during the AOT assembly phase and there are already checks in place to assure this. This patch simply adds a debug mode only check. Verified with tier 1-5 tests.

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-8366020: Assert that java code is not executed during the AOT assembly phase (Enhancement - P4)

Reviewers

Vladimir Kozlov (@vnkozlov - Reviewer)
Coleen Phillimore (@coleenp - Reviewer)
Ioi Lam (@iklam - Reviewer)

Contributors

Ioi Lam <iklam@openjdk.org>

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/30345/head:pull/30345
$ git checkout pull/30345

Update a local copy of the PR:
$ git checkout pull/30345
$ git pull https://git.openjdk.org/jdk.git pull/30345/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 30345

View PR using the GUI difftool:
$ git pr show -t 30345

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/30345.diff

Using Webrev

Link to Webrev Comment

…y phase.

…ert_8366020

bridgekeeper · 2026-03-20T19:12:24Z

👋 Welcome back matsaave! 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 · 2026-03-20T19:14:23Z

@matias9927 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:

8366020: Assert that java code is not executed during the AOT assembly phase

Co-authored-by: Ioi Lam <iklam@openjdk.org>
Reviewed-by: kvn, coleenp, iklam

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 183 new commits pushed to the master branch:

7ffc4a4: 8381019: Remove AppContext usage from AccessBridge
79668b0: 8380710: nmethod::finalize_relocations() should be called before ICache::invalidate_range() is called
28e96d7: 8377976: GenShen: Explicit GC requests must clear concurrent gc request cancellation
... and 180 more: https://git.openjdk.org/jdk/compare/3ddfdd94d92135e2038f48913a4c3e0f9754049b...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.

openjdk · 2026-03-20T19:15:15Z

@matias9927 The following label will be automatically applied to this pull request:

hotspot-runtime

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 · 2026-03-20T19:18:36Z

Webrevs

vnkozlov

To avoid confusion may you should use "Java code" instead of "user code" in title and comments.

vnkozlov · 2026-03-20T20:10:08Z

src/hotspot/share/cds/aotClassInitializer.cpp

    return false;
  }

+#ifndef PRODUCT


Use #ifdef ASSERT instead if you guard for debug VM.

matias9927 · 2026-03-23T17:36:49Z

/author add @iklam

openjdk · 2026-03-23T17:39:17Z

@matias9927 add @iklam is not a valid name and email string.

Syntax: /author [set|remove] [@user | openjdk-user | Full Name <email@address>]. For example:

/author set @openjdk-bot
/author set duke
/author set J. Duke <duke@openjdk.org>
/author @openjdk-bot
/author remove @openjdk-bot
/author remove

User names can only be used for users in the census associated with this repository. For other authors you need to supply the full name and email address.

matias9927 · 2026-03-24T19:07:13Z

/contributor add @iklam

openjdk · 2026-03-24T19:10:05Z

@matias9927
Contributor Ioi Lam <iklam@openjdk.org> successfully added.

coleenp · 2026-03-24T20:01:14Z

src/hotspot/share/cds/aotClassInitializer.cpp

+      if (ik->is_interface() && !ik->interface_needs_clinit_execution_as_super()) {
+        // TODO: why are these interfaces marked as initialized??
+      } else {
+        assert(false, "cannot execute user code");


I don't think they're marked as initialized, I think this case doesn't have Java code to execute.

So why can't this be
// Interfaces with default methods and <clinit> also execute Java code.
assert(!ik->is_interface() || !ik->interface_needs_clinit_execution_as_super(), "cannot execute Java code");

Can you capitalize Java?

I think the assert should be:
assert(ik->is_interface() && !ik->interface_needs_clinit_execution_as_super(), "cannot execute Java code in assembly phase")

Yes, that's correct.

Yes, this assert is the same logic, not mine.

iklam · 2026-03-22T22:09:00Z

src/hotspot/share/cds/aotClassInitializer.cpp

+    // in assembly phase.
+    if (ik->class_loader() != nullptr && !ik->is_hidden()) {
+      if (ik->is_interface() && !ik->interface_needs_clinit_execution_as_super()) {
+        // TODO: why are these interfaces marked as initialized??


I think we can use this version that explains why we allow interfaces that are !interface_needs_clinit_execution_as_super().

I also added

check for user code from -Xbootclasspath

more comments about the different conditions that are being checked.

if (!ik->is_initialized() && !ik->is_being_initialized()) { [...] return false; >>>> } else { // If code in ik is executed, then ik must be in the state of being_initialized or // fully_initialized. // // Check that no user code is executed during the assembly phase. Otherwise the user // code may introduce undesirable environment dependencies into the heap image. if (AOTInitTestClass != nullptr || ArchiveHeapTestClass != nullptr) { // OK: if any of the above two flags are set, we allow user code to be executed // in the assembly phase. Note that these flags are strictly for the purpose // of testing HotSpot and are not available in product builds. } else { if (ik->defined_by_boot_loader()) { // We allow boot classes to be AOT-initialized, except for classes from // -Xbootclasspath (cp index >= 1) be AOT-initialized, as such classes may be // provided by the user application. assert(ik->shared_classpath_index() <= 0, "only boot classed loaded from the modules image can be AOT-initialized"); } else { assert(ik->defined_by_platform_loader() || ik->defined_by_app_loader(), "cannot AOT-initialized classed loaded by other loaders"); // Some classes from platform/app loaders are allowed to be initialized during // the assembly phase. if (ik->is_hidden()) { // OK: hidden classes from platform/app loaders need to be AOT-initialized to // support AOT-linking of lambdas. These hidden classes are generated by the // VM and do not contain user code. } else if (ik->is_interface() && !ik->interface_needs_clinit_execution_as_super()) { // OK: ik is an interface used by a lambda. When AOT-linking lambdas, we only // support interfaces that are not interface_needs_clinit_execution_as_super(). // See AOTConstantPoolResolver::check_lambda_metafactory_signature(). } else { ResourceMark rm; fatal("cannot execute user code in class %s", ik->external_name()); } } } } <<<<

iklam · 2026-03-26T19:18:37Z

src/hotspot/share/cds/aotClassInitializer.cpp

+      if (ik->is_interface() && !ik->interface_needs_clinit_execution_as_super()) {
+        // TODO: why are these interfaces marked as initialized??
+      } else {
+        assert(false, "cannot execute user code");


Yes, that's correct.

coleenp

Okay, this makes sense with all the comments.

vnkozlov

LGTM

matias9927 added 2 commits March 13, 2026 11:22

8366020: Assert that user code is not executed during the AOT assembl…

2873f72

…y phase.

Merge branch 'master' of github.com:openjdk/jdk into aot_assembly_ass…

bcdd186

…ert_8366020

openjdk bot added the hotspot-runtime hotspot-runtime-dev@openjdk.org label Mar 20, 2026

openjdk bot added the rfr Pull request is ready for review label Mar 20, 2026

vnkozlov reviewed Mar 20, 2026

View reviewed changes

matias9927 changed the title ~~8366020: Assert that user code is not executed during the AOT assembly phase.~~ 8366020: Assert that java code is not executed during the AOT assembly phase Mar 20, 2026

Vladimir comments

5886827

coleenp reviewed Mar 24, 2026

View reviewed changes

Coleen comments

b5f04ed

iklam reviewed Mar 26, 2026

View reviewed changes

Ioi comments

d59e2f1

coleenp approved these changes Mar 27, 2026

View reviewed changes

vnkozlov approved these changes Mar 27, 2026

View reviewed changes

openjdk bot added the ready Pull request is ready to be integrated label Mar 27, 2026

iklam approved these changes Mar 27, 2026

View reviewed changes

Conversation

matias9927 commented Mar 20, 2026 • edited by openjdk bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Progress

Issue

Reviewers

Contributors

Reviewing

Uh oh!

bridgekeeper bot commented Mar 20, 2026

Uh oh!

openjdk bot commented Mar 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

openjdk bot commented Mar 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mlbridge bot commented Mar 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Webrevs

Uh oh!

vnkozlov left a comment

Choose a reason for hiding this comment

Uh oh!

vnkozlov Mar 20, 2026

Choose a reason for hiding this comment

Uh oh!

matias9927 commented Mar 23, 2026

Uh oh!

openjdk bot commented Mar 23, 2026

Uh oh!

matias9927 commented Mar 24, 2026

Uh oh!

openjdk bot commented Mar 24, 2026

Uh oh!

coleenp Mar 24, 2026

Choose a reason for hiding this comment

Uh oh!

matias9927 Mar 26, 2026

Choose a reason for hiding this comment

Uh oh!

iklam Mar 26, 2026

Choose a reason for hiding this comment

Uh oh!

coleenp Mar 27, 2026

Choose a reason for hiding this comment

Uh oh!

iklam Mar 22, 2026

Choose a reason for hiding this comment

Uh oh!

iklam Mar 26, 2026

Choose a reason for hiding this comment

Uh oh!

coleenp left a comment

Choose a reason for hiding this comment

Uh oh!

vnkozlov left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

4 participants

matias9927 commented Mar 20, 2026 •

edited by openjdk bot

Loading

openjdk bot commented Mar 20, 2026 •

edited

Loading

openjdk bot commented Mar 20, 2026 •

edited

Loading

mlbridge bot commented Mar 20, 2026 •

edited

Loading