8380476: Implement JEP: Ahead-of-Time Code Compilation by vnkozlov · Pull Request #30778 · openjdk/jdk

vnkozlov · 2026-04-16T23:09:52Z

Help applications start up and warm up more quickly by shifting dynamic (JIT) method compilation from production runs to training runs, conveying the necessary native code via the AOT cache.

More details in the JEP.

I confirm that I make this contribution in accordance with the OpenJDK Interim AI Policy.

Progress

Change requires CSR request JDK-8380477 to be approved
Change must not contain extraneous whitespace
Commit message must refer to an issue
Change must be properly reviewed (2 reviews required, with at least 1 Reviewer, 1 Author)

Issues

JDK-8380476: Implement JEP: Ahead-of-Time Code Compilation (Sub-task - P4)
JDK-8380477: Implement JEP: Ahead-of-Time Code Compilation (CSR)

Contributors

John R Rose <[email protected]>
Andrew Dinn <[email protected]>
Ashutosh Mehra <[email protected]>
Dan Heidinga <[email protected]>
Ioi Lam <[email protected]>
Vladimir Ivanov <[email protected]>
Ekaterina Pavlova <[email protected]>
Aleksey Shipilev <[email protected]>
Igor Veresov <[email protected]>

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/30778/head:pull/30778
$ git checkout pull/30778

Update a local copy of the PR:
$ git checkout pull/30778
$ git pull https://git.openjdk.org/jdk.git pull/30778/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 30778

View PR using the GUI difftool:
$ git pr show -t 30778

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/30778.diff

bridgekeeper · 2026-04-16T23:11:09Z

👋 Welcome back kvn! 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 · 2026-04-16T23:11:17Z

❗ This change is not yet ready to be integrated.
See the Progress checklist in the description for automated requirements.

openjdk · 2026-04-16T23:12:16Z

@vnkozlov The following labels will be automatically applied to this pull request:

build
hotspot
shenandoah

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

openjdk · 2026-04-16T23:12:19Z

The total number of required reviews for this PR has been set to 2 based on the presence of this label: hotspot. This can be overridden with the /reviewers command.

vnkozlov · 2026-04-17T00:06:53Z

/contributor add @rose00
/contributor add @adinn
/contributor add @ashu-mehra
/contributor add @DanHeidinga
/contributor add @iklam
/contributor add @iwanowww
/contributor add @katyapav
/contributor add @shipilev
/contributor add @veresov

openjdk · 2026-04-17T00:07:53Z

@vnkozlov
Contributor John R Rose <[email protected]> successfully added.

openjdk · 2026-04-17T00:08:38Z

@vnkozlov
Contributor Andrew Dinn <[email protected]> successfully added.

openjdk · 2026-04-17T00:09:12Z

@vnkozlov
Contributor Ashutosh Mehra <[email protected]> successfully added.

openjdk · 2026-04-17T00:09:53Z

@vnkozlov
Contributor Dan Heidinga <[email protected]> successfully added.

openjdk · 2026-04-17T00:10:31Z

@vnkozlov
Contributor Ioi Lam <[email protected]> successfully added.

openjdk · 2026-04-17T00:11:15Z

@vnkozlov
Contributor Vladimir Ivanov <[email protected]> successfully added.

openjdk · 2026-04-17T00:11:48Z

@vnkozlov
Contributor Ekaterina Pavlova <[email protected]> successfully added.

openjdk · 2026-04-17T00:12:24Z

@vnkozlov
Contributor Aleksey Shipilev <[email protected]> successfully added.

openjdk · 2026-04-17T00:13:14Z

@vnkozlov
Contributor Igor Veresov <[email protected]> successfully added.

vnkozlov · 2026-04-17T12:43:39Z

Why I did not hit this in our internal testing?

assert(_extrs_length <= 380) failed: increase size

vnkozlov · 2026-04-17T12:43:50Z

There is also crash on MacOS-x86 with -XX:+UseShenandoahGC during product run.

shipilev · 2026-04-17T13:11:43Z

I can take a look at Shenandoah parts, if you want. We probably just missing a few C1 stubs for it?

vnkozlov · 2026-04-17T13:24:45Z

I can take a look at Shenandoah parts, if you want. We probably just missing a few C1 stubs for it?

Please do. Look on AOTCodeAddressTable::set_c1_stubs_complete() where they are recorded.

shipilev · 2026-04-17T13:26:07Z

I can take a look at Shenandoah parts, if you want. We probably just missing a few C1 stubs for it?

Please do. Look on AOTCodeAddressTable::set_c1_stubs_complete() where they are recorded.

This fixes the sizes assert:

diff --git a/src/hotspot/share/code/aotCodeCache.cpp b/src/hotspot/share/code/aotCodeCache.cpp
index 4a2c931adaa..b09fe7a68cb 100644
--- a/src/hotspot/share/code/aotCodeCache.cpp
+++ b/src/hotspot/share/code/aotCodeCache.cpp
@@ -3355,7 +3355,7 @@ void AOTCodeReader::read_dbg_strings(DbgStrings& dbg_strings, bool use_string_ta
 //   [_stubs_base, _stubs_base + _stubs_max -1],
 //   [_c_str_base, _c_str_base + _c_str_max -1],
 
-#define _extrs_max 380
+#define _extrs_max 400
 #define _stubs_max static_cast<int>(EntryId::NUM_ENTRYIDS)
 
 #define _extrs_base 0

I think your testing was done without Shenandoah built? I see there is a AOTCodeTest failure on macos-aarch64 too, I am trying to reproduce it locally.

vnkozlov · 2026-04-17T13:27:38Z

From my local X86 testing with Z GC and additional output update:

[0.260s][info ][aot,codecache,init] External addresses 368 recorded and closed
...
[0.304s][debug][aot,codecache,init ] External C1 addresses 4 recorded

Which is very near 380 limit. It is possible we generate more intrinsics/stubs on GHA machines.

vnkozlov · 2026-04-17T13:29:42Z

400 could be still small. I will change it to 500

shipilev · 2026-04-17T13:59:03Z

I see there is a AOTCodeTest failure on macos-aarch64 too, I am trying to reproduce it locally.

Aha, for C1, we "just" miss relocations for the runtime entry points in barriers. So once we call through the AOT-saved method, the barrier addresses are not relocated and we jump to garbage locations. Apply this: shenandoah-aot-1.patch.

There is a failure that looks similar in C2, although I am not sure what's wrong there yet. Digging...

vnkozlov · 2026-05-10T16:17:32Z

If you could merge from master, I can kick off another code analyzer run.

Merged

iklam · 2026-05-15T22:53:53Z

+  product(bool, SkipArchiveHeapVerification, false, DIAGNOSTIC,             \
+          "Skip verification of CDS archive heap")                          \


Is this change necessary for AOT code?

No. I missed to remove it from my patch. I will remove it.

…oad AOT code

…de load when -Xcomp is used

vnkozlov · 2026-05-27T23:13:29Z

@shipilev please look on GHA failure in runtime/cds/appcds/aotCode/AOTCodeTest.java#shenandoah test after merge from mainline.

shipilev · 2026-05-28T08:33:15Z

@shipilev please look on GHA failure in runtime/cds/appcds/aotCode/AOTCodeTest.java#shenandoah test after merge from mainline.

I know what this is: https://bugs.openjdk.org/browse/JDK-8385557. Amusingly, we tried to take care of this AOT path in LBE code, so to avoid regressions for your PR here. And we apparently did things right with AOT constants for cset-base/region-shift, but missed the relocations for slow calls! Which is a problem if runtime calls relocate (ASLR?). So LBE passed with AOT code cache PR on Linux x86/AArch64 without problems, and only Mac AArch64 caught fire. Argh. I am going to amend this code in mainline.

Delawen · 2026-06-09T08:56:22Z

Adding some logging for the map files: vnkozlov#1

8380476: Implement JEP: Ahead-of-Time Code Compilation

8380476: Implement JEP: Ahead-of-Time Code Compilation

786d435

openjdk Bot added csr Pull request needs approved CSR before integration hotspot [email protected] build [email protected] shenandoah [email protected] labels Apr 16, 2026

Remove trailing white space

38327a7

Fix Minimal and Zero VM builds

b2091a0

Increase AOT external addresses table size

24f9ccb

vnkozlov added 2 commits May 9, 2026 13:34

Change DisableAOTCode flag value to bitmask. Add test for it.

3215593

Merge master 05-09-2026

c501d24

openjdk Bot removed the merge-conflict Pull request has merge conflict with target branch label May 10, 2026

vnkozlov added 2 commits May 11, 2026 17:48

Merge master 05-11-2026

6249bd3

Fix Minimal VM. Replace some precond with assert.

6325f31

iklam reviewed May 15, 2026

View reviewed changes

vnkozlov added 13 commits May 15, 2026 18:09

Cleanup AOT code load failure path

ef919d3

Remove UseAOTCodeLoadThread flag - always create separate thread to l…

1dd35f1

…oad AOT code

Fix VerifyAOTCode flag code and add test

2862738

Remove SkipArchiveHeapVerification flag

9ec2939

Record AOT code load failure

9db9d66

Code refactoring. Request normal JIT compilation together with AOT co…

f18697b

…de load when -Xcomp is used

Check array klass can be cached during AOT compilation

2071fae

Validate dependencies when loading AOT code with VerifyAOTCode flag on

7f15fdc

Merge master 05-18-2026

865b22f

Remove trailing whitespace

c9ef645

Merge master 05-20-2026

8444419

Cleanup unused code

fb29c59

Add missing addresses to AOTCodeAddressTable

dff06b4

openjdk Bot added the merge-conflict Pull request has merge conflict with target branch label May 22, 2026

Merge master 05-27-2026

1a91dd0

openjdk Bot removed the merge-conflict Pull request has merge conflict with target branch label May 27, 2026

vnkozlov added 2 commits May 29, 2026 12:49

Merge master 05-29-2026

9710da1

Merge master 06-04-2026

47f7faf

vnkozlov mentioned this pull request Jun 6, 2026

8386108: [leyden] Implement JEP: Ahead-of-Time Code Compilation openjdk/leyden#113

Closed

4 tasks

Delawen added a commit to Delawen/leyden-analyzer that referenced this pull request Jun 15, 2026

feat: adding support for openjdk/jdk#30778

a1f6904

8380476: Implement JEP: Ahead-of-Time Code Compilation

		product(bool, SkipArchiveHeapVerification, false, DIAGNOSTIC, \
		"Skip verification of CDS archive heap") \

Conversation

vnkozlov commented Apr 16, 2026 • edited by openjdk Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Progress

Issues

Contributors

Reviewing

Uh oh!

bridgekeeper Bot commented Apr 16, 2026

Uh oh!

openjdk Bot commented Apr 16, 2026

Uh oh!

openjdk Bot commented Apr 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

openjdk Bot commented Apr 16, 2026

Uh oh!

vnkozlov commented Apr 17, 2026

Uh oh!

openjdk Bot commented Apr 17, 2026

Uh oh!

openjdk Bot commented Apr 17, 2026

Uh oh!

openjdk Bot commented Apr 17, 2026

Uh oh!

openjdk Bot commented Apr 17, 2026

Uh oh!

openjdk Bot commented Apr 17, 2026

Uh oh!

openjdk Bot commented Apr 17, 2026

Uh oh!

openjdk Bot commented Apr 17, 2026

Uh oh!

openjdk Bot commented Apr 17, 2026

Uh oh!

openjdk Bot commented Apr 17, 2026

Uh oh!

vnkozlov commented Apr 17, 2026

Uh oh!

vnkozlov commented Apr 17, 2026

Uh oh!

shipilev commented Apr 17, 2026

Uh oh!

vnkozlov commented Apr 17, 2026

Uh oh!

shipilev commented Apr 17, 2026

Uh oh!

vnkozlov commented Apr 17, 2026

Uh oh!

vnkozlov commented Apr 17, 2026

Uh oh!

shipilev commented Apr 17, 2026

Uh oh!

vnkozlov commented May 10, 2026

Uh oh!

iklam May 15, 2026

Choose a reason for hiding this comment

Uh oh!

vnkozlov May 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

vnkozlov commented May 27, 2026

Uh oh!

shipilev commented May 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Delawen commented Jun 9, 2026

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

7 participants

vnkozlov commented Apr 16, 2026 •

edited by openjdk Bot

Loading

openjdk Bot commented Apr 16, 2026 •

edited

Loading

vnkozlov May 16, 2026 •

edited

Loading

shipilev commented May 28, 2026 •

edited

Loading