From vromero at openjdk.org  Sat Oct  1 19:15:48 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Sat, 1 Oct 2022 19:15:48 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v3]
In-Reply-To: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
References: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
Message-ID: <mLCXI3oinJapSISR4XPp2BX4o3153I8yGJxUQIPsWl8=.1ef36536-699c-4562-b358-0f894a119b24@github.com>

> Value classes constructor should be named <vnew> instead of <init>. This PR can't be pushed without the corresponding support being available on the VM side

Vicente Romero has updated the pull request incrementally with one additional commit since the last revision:

  fixing regression tests

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/771/files
  - new: https://git.openjdk.org/valhalla/pull/771/files/6e2d19c9..6264d345

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=771&range=02
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=771&range=01-02

  Stats: 67 lines in 9 files changed: 9 ins; 38 del; 20 mod
  Patch: https://git.openjdk.org/valhalla/pull/771.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/771/head:pull/771

PR: https://git.openjdk.org/valhalla/pull/771


From thartmann at openjdk.org  Mon Oct  3 11:49:43 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Mon, 3 Oct 2022 11:49:43 GMT
Subject: [lworld] RFR: 8293994: [lworld] Deoptimization from nmethod entry
 barrier breaks scalarized calling convention
Message-ID: <T2jDh5bYxmBfi7ik6kmOFKfqWaP3cxgV0mT_ocda8cY=.c254a606-c49e-4556-a6b7-5ecec43bb280@github.com>

The recent sweeper removal ([JDK-8290025](https://bugs.openjdk.org/browse/JDK-8290025)) always enabled nmethod entry barriers. This revealed an issue where deoptimization from the slow path (`-XX:+DeoptimizeNMethodBarriersALot`) breaks the scalarized calling convention because once deoptimization is triggered at nmethod entry via `BarrierSetNMethod::deoptimize`, we already converted the calling convention and can't easily go back. Also with method handle invocation logic, we can not tell in `SharedRuntime::handle_wrong_method` which calling convention the callee nmethod converted to.

After evaluating several ideas, I found that the cleanest solution is to have an own nmethod entry barrier per entry point right before converting calling conventions.

While debugging, I discovered similar issues with the class init barriers, that I'll address in a follow-up fix ([JDK-8294013](https://bugs.openjdk.org/browse/JDK-8294013)).

Thanks,
Tobias

-------------

Commit messages:
 - Missing aarch64 fix
 - Fixed native methods
 - Fixed initialization of orig_pc
 - 8293994: [lworld] Deoptimization from nmethod entry barrier breaks scalarized calling convention

Changes: https://git.openjdk.org/valhalla/pull/773/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=773&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8293994
  Stats: 335 lines in 22 files changed: 171 ins; 98 del; 66 mod
  Patch: https://git.openjdk.org/valhalla/pull/773.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/773/head:pull/773

PR: https://git.openjdk.org/valhalla/pull/773


From thartmann at openjdk.org  Mon Oct  3 12:05:59 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Mon, 3 Oct 2022 12:05:59 GMT
Subject: [lworld] Integrated: 8293994: [lworld] Deoptimization from nmethod
 entry barrier breaks scalarized calling convention
In-Reply-To: <T2jDh5bYxmBfi7ik6kmOFKfqWaP3cxgV0mT_ocda8cY=.c254a606-c49e-4556-a6b7-5ecec43bb280@github.com>
References: <T2jDh5bYxmBfi7ik6kmOFKfqWaP3cxgV0mT_ocda8cY=.c254a606-c49e-4556-a6b7-5ecec43bb280@github.com>
Message-ID: <AQ4jRvsTCePdsTMIr_aLWFlxvTwV5rUwQCz6zYu_PZI=.25f64424-c782-4547-ad7c-68c120eb4f0f@github.com>

On Mon, 3 Oct 2022 11:41:25 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

> The recent sweeper removal ([JDK-8290025](https://bugs.openjdk.org/browse/JDK-8290025)) always enabled nmethod entry barriers. This revealed an issue where deoptimization from the slow path (`-XX:+DeoptimizeNMethodBarriersALot`) breaks the scalarized calling convention because once deoptimization is triggered at nmethod entry via `BarrierSetNMethod::deoptimize`, we already converted the calling convention and can't easily go back. Also with method handle invocation logic, we can not tell in `SharedRuntime::handle_wrong_method` which calling convention the callee nmethod converted to.
> 
> After evaluating several ideas, I found that the cleanest solution is to have an own nmethod entry barrier per entry point right before converting calling conventions.
> 
> While debugging, I discovered similar issues with the class init barriers, that I'll address in a follow-up fix ([JDK-8294013](https://bugs.openjdk.org/browse/JDK-8294013)).
> 
> Thanks,
> Tobias

This pull request has now been integrated.

Changeset: ee15e2d9
Author:    Tobias Hartmann <thartmann at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/ee15e2d905edf29ed7660ca057d6cadb69de3f87
Stats:     335 lines in 22 files changed: 171 ins; 98 del; 66 mod

8293994: [lworld] Deoptimization from nmethod entry barrier breaks scalarized calling convention

-------------

PR: https://git.openjdk.org/valhalla/pull/773


From fparain at openjdk.org  Mon Oct  3 15:13:38 2022
From: fparain at openjdk.org (Frederic Parain)
Date: Mon, 3 Oct 2022 15:13:38 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v10]
In-Reply-To: <GhNfK95JhehVSysj3vRYst2anHAoxLHH1QqP-xEsR9E=.456c2076-adb9-4903-94de-707344023a9c@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <GhNfK95JhehVSysj3vRYst2anHAoxLHH1QqP-xEsR9E=.456c2076-adb9-4903-94de-707344023a9c@github.com>
Message-ID: <QktQG65d9CqLJog-nyLQIu5isPBdOt9QAJiIstJLTzM=.1efa3092-16d3-4eaa-9eb2-f1223a8a76ef@github.com>

On Thu, 29 Sep 2022 06:49:13 GMT, David Simms <dsimms at openjdk.org> wrote:

>> Runtime component of -XX:+EnablePrimitiveClasses switch
>
> David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 22 commits:
> 
>  - Merge branch 'lworld' into 8288135
>  - Extra assertion info for tracking spurious issue
>  - supports_inline_types() refers to classfile and is separate from Enable flags
>  - Further testing explicity using flags
>  - Disable EnablePrimitiveClasses by default
>  - Explictly flag more tests
>  - Merge branch 'lworld' into 8288135
>  - jdk_valhalla tests
>  - Merge branch 'lworld' into 8288135
>  - Further test fixes
>  - ... and 12 more: https://git.openjdk.org/valhalla/compare/951544ff...1853e0c6

Thank you for tackling this task.
My review is more focused on the VM side than on the tests.
A few comments to be addressed.

Thank you,

Fred

src/hotspot/share/classfile/classFileParser.cpp line 3328:

> 3326:       recognized_modifiers |= JVM_ACC_MODULE;
> 3327:     }
> 3328:     // JVM_ACC_VALUE, JVM_ACC_PRIMITIVE, and JVM_ACC_IDENTITY are defined depending on version and feature flag

The implementation doesn't seem to match the comment nor the intent of the patch. JVM_ACC_VALUE and JVM_ACC_IDENTITY should be controlled by the class file version (tested with supports_inline_types()), but JVM_ACC_PRIMITIVE should be controlled by the class file version and the EnablePrimitiveClasses flag.

src/hotspot/share/classfile/classFileParser.cpp line 5235:

> 5233:       // Can't enable this check fully until JDK upgrades the bytecode generators.
> 5234:       // For now, compare to class file version 51 so old verifier doesn't see Q signatures.
> 5235:       if ( (_major_version < 51 /* CONSTANT_CLASS_DESCRIPTORS */ ) || (!EnablePrimitiveClasses)) {

Are there a CRs to track the changes requires in the JDK and the status of this check and its dependency to JDK updates?

src/hotspot/share/classfile/classFileParser.cpp line 6183:

> 6181:     recognized_modifiers |= JVM_ACC_MODULE;
> 6182:   }
> 6183:   // Are JVM_ACC_VALUE and JVM_ACC_PRIMITIVE support (version and feature check)

JVM_ACC_PRIMITIVE should be controlled by both the value of supports_inline_types() and EnablePrimitiveClasses.

src/hotspot/share/runtime/arguments.cpp line 3036:

> 3034:   }
> 3035: 
> 3036:   if (!EnableValhalla) {

Why silently change the option when -XX:+EnablePrimitiveClass is specified without -XX:+EnableValhalla? Wouldn't be more informative to throw an error message?

test/hotspot/jtreg/runtime/valhalla/inlinetypes/classfileparser/BadACCValue.java line 27:

> 25:  * @test
> 26:  * @summary test that if a class file has ACC_VALUE or ACC_PRIMITIVE set then it must be run
> 27:  *          with option -XX:+?nableValhalla or -XX:+EnablePrimitiveClasses respectively.

Looks like there's an accent on the E of -XX:+?nableValhalla.

-------------

Changes requested by fparain (Committer).

PR: https://git.openjdk.org/valhalla/pull/727


From dsimms at openjdk.org  Tue Oct  4 07:13:46 2022
From: dsimms at openjdk.org (David Simms)
Date: Tue, 4 Oct 2022 07:13:46 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v10]
In-Reply-To: <QktQG65d9CqLJog-nyLQIu5isPBdOt9QAJiIstJLTzM=.1efa3092-16d3-4eaa-9eb2-f1223a8a76ef@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <GhNfK95JhehVSysj3vRYst2anHAoxLHH1QqP-xEsR9E=.456c2076-adb9-4903-94de-707344023a9c@github.com>
 <QktQG65d9CqLJog-nyLQIu5isPBdOt9QAJiIstJLTzM=.1efa3092-16d3-4eaa-9eb2-f1223a8a76ef@github.com>
Message-ID: <wpForaa5R3odXxR9u5xDvvGFjIoa2k7zYjTG-cm4hLk=.f0406034-cb17-4091-baa7-e89dbd0a9b1b@github.com>

On Mon, 3 Oct 2022 14:09:29 GMT, Frederic Parain <fparain at openjdk.org> wrote:

>> David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 22 commits:
>> 
>>  - Merge branch 'lworld' into 8288135
>>  - Extra assertion info for tracking spurious issue
>>  - supports_inline_types() refers to classfile and is separate from Enable flags
>>  - Further testing explicity using flags
>>  - Disable EnablePrimitiveClasses by default
>>  - Explictly flag more tests
>>  - Merge branch 'lworld' into 8288135
>>  - jdk_valhalla tests
>>  - Merge branch 'lworld' into 8288135
>>  - Further test fixes
>>  - ... and 12 more: https://git.openjdk.org/valhalla/compare/951544ff...1853e0c6
>
> src/hotspot/share/classfile/classFileParser.cpp line 3328:
> 
>> 3326:       recognized_modifiers |= JVM_ACC_MODULE;
>> 3327:     }
>> 3328:     // JVM_ACC_VALUE, JVM_ACC_PRIMITIVE, and JVM_ACC_IDENTITY are defined depending on version and feature flag
> 
> The implementation doesn't seem to match the comment nor the intent of the patch. JVM_ACC_VALUE and JVM_ACC_IDENTITY should be controlled by the class file version (tested with supports_inline_types()), but JVM_ACC_PRIMITIVE should be controlled by the class file version and the EnablePrimitiveClasses flag.

Agree the comment doesn't quite match. Let me check the behavior is correct:

Currently (in this patch); 
- if the classfile is new enough to be aware of the existence (`supports_inline_types()`) of new flags, the are placed in the recognized set, and later feature checks produce errors to use the appropriate switches
- older classfiles we simply ignore modifiers that were not defined at the time of that version, if one took a new primitive, and downgraded it's version, the VM will quietly ignore these modifiers (but probably take issue with other classfile features, like static factory with special naming)
- `ACC_SUPER `and `ACC_IDENTITY` overlap requires special handling based on both version and feature flag (`EnableValhalla`)

Is this your view too ?

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From dsimms at openjdk.org  Tue Oct  4 07:21:32 2022
From: dsimms at openjdk.org (David Simms)
Date: Tue, 4 Oct 2022 07:21:32 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v10]
In-Reply-To: <wpForaa5R3odXxR9u5xDvvGFjIoa2k7zYjTG-cm4hLk=.f0406034-cb17-4091-baa7-e89dbd0a9b1b@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <GhNfK95JhehVSysj3vRYst2anHAoxLHH1QqP-xEsR9E=.456c2076-adb9-4903-94de-707344023a9c@github.com>
 <QktQG65d9CqLJog-nyLQIu5isPBdOt9QAJiIstJLTzM=.1efa3092-16d3-4eaa-9eb2-f1223a8a76ef@github.com>
 <wpForaa5R3odXxR9u5xDvvGFjIoa2k7zYjTG-cm4hLk=.f0406034-cb17-4091-baa7-e89dbd0a9b1b@github.com>
Message-ID: <s7VFkbIO0DszPhR3juyjX0n7y4SkOlaZ8558R2fHyfk=.29b4bc2a-acd7-4953-8a77-bedb9a9bc171@github.com>

On Tue, 4 Oct 2022 07:10:03 GMT, David Simms <dsimms at openjdk.org> wrote:

>> src/hotspot/share/classfile/classFileParser.cpp line 3328:
>> 
>>> 3326:       recognized_modifiers |= JVM_ACC_MODULE;
>>> 3327:     }
>>> 3328:     // JVM_ACC_VALUE, JVM_ACC_PRIMITIVE, and JVM_ACC_IDENTITY are defined depending on version and feature flag
>> 
>> The implementation doesn't seem to match the comment nor the intent of the patch. JVM_ACC_VALUE and JVM_ACC_IDENTITY should be controlled by the class file version (tested with supports_inline_types()), but JVM_ACC_PRIMITIVE should be controlled by the class file version and the EnablePrimitiveClasses flag.
>
> Agree the comment doesn't quite match. Let me check the behavior is correct:
> 
> Currently (in this patch); 
> - if the classfile is new enough to be aware of the existence (`supports_inline_types()`) of new flags, the are placed in the recognized set, and later feature checks produce errors to use the appropriate switches
> - older classfiles we simply ignore modifiers that were not defined at the time of that version, if one took a new primitive, and downgraded it's version, the VM will quietly ignore these modifiers (but probably take issue with other classfile features, like static factory with special naming)
> - `ACC_SUPER `and `ACC_IDENTITY` overlap requires special handling based on both version and feature flag (`EnableValhalla`)
> 
> Is this your view too ?

Comment removed, it seems superfluous (the code describes itself).

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From dsimms at openjdk.org  Tue Oct  4 08:22:36 2022
From: dsimms at openjdk.org (David Simms)
Date: Tue, 4 Oct 2022 08:22:36 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v10]
In-Reply-To: <QktQG65d9CqLJog-nyLQIu5isPBdOt9QAJiIstJLTzM=.1efa3092-16d3-4eaa-9eb2-f1223a8a76ef@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <GhNfK95JhehVSysj3vRYst2anHAoxLHH1QqP-xEsR9E=.456c2076-adb9-4903-94de-707344023a9c@github.com>
 <QktQG65d9CqLJog-nyLQIu5isPBdOt9QAJiIstJLTzM=.1efa3092-16d3-4eaa-9eb2-f1223a8a76ef@github.com>
Message-ID: <BvZ_M8fzWaa3LrHlO0lLhtGQ0D1aka-t8cYPlRWuWoY=.bd9fd670-5e4a-49a9-875f-bd021dc0555a@github.com>

On Mon, 3 Oct 2022 14:13:13 GMT, Frederic Parain <fparain at openjdk.org> wrote:

>> David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 22 commits:
>> 
>>  - Merge branch 'lworld' into 8288135
>>  - Extra assertion info for tracking spurious issue
>>  - supports_inline_types() refers to classfile and is separate from Enable flags
>>  - Further testing explicity using flags
>>  - Disable EnablePrimitiveClasses by default
>>  - Explictly flag more tests
>>  - Merge branch 'lworld' into 8288135
>>  - jdk_valhalla tests
>>  - Merge branch 'lworld' into 8288135
>>  - Further test fixes
>>  - ... and 12 more: https://git.openjdk.org/valhalla/compare/951544ff...1853e0c6
>
> src/hotspot/share/runtime/arguments.cpp line 3036:
> 
>> 3034:   }
>> 3035: 
>> 3036:   if (!EnableValhalla) {
> 
> Why silently change the option when -XX:+EnablePrimitiveClass is specified without -XX:+EnableValhalla? Wouldn't be more informative to throw an error message?

Agreed, added error message

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From dsimms at openjdk.org  Tue Oct  4 08:31:18 2022
From: dsimms at openjdk.org (David Simms)
Date: Tue, 4 Oct 2022 08:31:18 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v10]
In-Reply-To: <QktQG65d9CqLJog-nyLQIu5isPBdOt9QAJiIstJLTzM=.1efa3092-16d3-4eaa-9eb2-f1223a8a76ef@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <GhNfK95JhehVSysj3vRYst2anHAoxLHH1QqP-xEsR9E=.456c2076-adb9-4903-94de-707344023a9c@github.com>
 <QktQG65d9CqLJog-nyLQIu5isPBdOt9QAJiIstJLTzM=.1efa3092-16d3-4eaa-9eb2-f1223a8a76ef@github.com>
Message-ID: <_kxByALqJ7hdpc_w98bEphJZT7_k4d73aF_A8CmJE3M=.b5eb5277-17e3-43fb-84d6-828e216052c7@github.com>

On Mon, 3 Oct 2022 14:21:22 GMT, Frederic Parain <fparain at openjdk.org> wrote:

>> David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 22 commits:
>> 
>>  - Merge branch 'lworld' into 8288135
>>  - Extra assertion info for tracking spurious issue
>>  - supports_inline_types() refers to classfile and is separate from Enable flags
>>  - Further testing explicity using flags
>>  - Disable EnablePrimitiveClasses by default
>>  - Explictly flag more tests
>>  - Merge branch 'lworld' into 8288135
>>  - jdk_valhalla tests
>>  - Merge branch 'lworld' into 8288135
>>  - Further test fixes
>>  - ... and 12 more: https://git.openjdk.org/valhalla/compare/951544ff...1853e0c6
>
> src/hotspot/share/classfile/classFileParser.cpp line 5235:
> 
>> 5233:       // Can't enable this check fully until JDK upgrades the bytecode generators.
>> 5234:       // For now, compare to class file version 51 so old verifier doesn't see Q signatures.
>> 5235:       if ( (_major_version < 51 /* CONSTANT_CLASS_DESCRIPTORS */ ) || (!EnablePrimitiveClasses)) {
> 
> Are there a CRs to track the changes requires in the JDK and the status of this check and its dependency to JDK updates?

I can add a reference to [JDK-8270852](https://bugs.openjdk.org/browse/JDK-8270852)

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From dsimms at openjdk.org  Tue Oct  4 08:34:52 2022
From: dsimms at openjdk.org (David Simms)
Date: Tue, 4 Oct 2022 08:34:52 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v10]
In-Reply-To: <QktQG65d9CqLJog-nyLQIu5isPBdOt9QAJiIstJLTzM=.1efa3092-16d3-4eaa-9eb2-f1223a8a76ef@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <GhNfK95JhehVSysj3vRYst2anHAoxLHH1QqP-xEsR9E=.456c2076-adb9-4903-94de-707344023a9c@github.com>
 <QktQG65d9CqLJog-nyLQIu5isPBdOt9QAJiIstJLTzM=.1efa3092-16d3-4eaa-9eb2-f1223a8a76ef@github.com>
Message-ID: <EBf-jiBM77unQ1AFJPGQgkygzjXngF0ZBfTnCa80NVk=.75bcf427-fffc-4e93-b442-134a31773d55@github.com>

On Mon, 3 Oct 2022 14:30:54 GMT, Frederic Parain <fparain at openjdk.org> wrote:

>> David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 22 commits:
>> 
>>  - Merge branch 'lworld' into 8288135
>>  - Extra assertion info for tracking spurious issue
>>  - supports_inline_types() refers to classfile and is separate from Enable flags
>>  - Further testing explicity using flags
>>  - Disable EnablePrimitiveClasses by default
>>  - Explictly flag more tests
>>  - Merge branch 'lworld' into 8288135
>>  - jdk_valhalla tests
>>  - Merge branch 'lworld' into 8288135
>>  - Further test fixes
>>  - ... and 12 more: https://git.openjdk.org/valhalla/compare/951544ff...1853e0c6
>
> src/hotspot/share/classfile/classFileParser.cpp line 6183:
> 
>> 6181:     recognized_modifiers |= JVM_ACC_MODULE;
>> 6182:   }
>> 6183:   // Are JVM_ACC_VALUE and JVM_ACC_PRIMITIVE support (version and feature check)
> 
> JVM_ACC_PRIMITIVE should be controlled by both the value of supports_inline_types() and EnablePrimitiveClasses.

Reverted comment to "defined for class version" `// JVM_ACC_VALUE and JVM_ACC_PRIMITIVE are defined for supported version`

> test/hotspot/jtreg/runtime/valhalla/inlinetypes/classfileparser/BadACCValue.java line 27:
> 
>> 25:  * @test
>> 26:  * @summary test that if a class file has ACC_VALUE or ACC_PRIMITIVE set then it must be run
>> 27:  *          with option -XX:+?nableValhalla or -XX:+EnablePrimitiveClasses respectively.
> 
> Looks like there's an accent on the E of -XX:+?nableValhalla.

Oh yeah, nice spot

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From dsimms at openjdk.org  Tue Oct  4 09:41:19 2022
From: dsimms at openjdk.org (David Simms)
Date: Tue, 4 Oct 2022 09:41:19 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v11]
In-Reply-To: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
Message-ID: <Gg15FmjHr9kUCPdShgnP7TzJ0RlSkkwShTFhN5IuIJY=.7889e629-0183-4c56-98ea-44adf7018baf@github.com>

> Runtime component of -XX:+EnablePrimitiveClasses switch

David Simms has updated the pull request incrementally with one additional commit since the last revision:

  Review comments applied

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/727/files
  - new: https://git.openjdk.org/valhalla/pull/727/files/1853e0c6..75113462

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=727&range=10
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=727&range=09-10

  Stats: 8 lines in 3 files changed: 2 ins; 1 del; 5 mod
  Patch: https://git.openjdk.org/valhalla/pull/727.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/727/head:pull/727

PR: https://git.openjdk.org/valhalla/pull/727


From mchung at openjdk.org  Tue Oct  4 18:43:56 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Tue, 4 Oct 2022 18:43:56 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v11]
In-Reply-To: <Gg15FmjHr9kUCPdShgnP7TzJ0RlSkkwShTFhN5IuIJY=.7889e629-0183-4c56-98ea-44adf7018baf@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <Gg15FmjHr9kUCPdShgnP7TzJ0RlSkkwShTFhN5IuIJY=.7889e629-0183-4c56-98ea-44adf7018baf@github.com>
Message-ID: <uCGc4DSDNy5MnxtjrNAN3WrpwBRY5tZu0T87SQXktKc=.386c87c7-3fd0-42d9-92d2-b82962a19d8c@github.com>

On Tue, 4 Oct 2022 09:41:19 GMT, David Simms <dsimms at openjdk.org> wrote:

>> Runtime component of -XX:+EnablePrimitiveClasses switch
>
> David Simms has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Review comments applied

test/jdk/java/lang/invoke/VarHandles/X-VarHandleTestMethodHandleAccess.java.template line 29:

> 27:  * @test
> 28:  * @compile -XDenablePrimitiveClasses Point.java Value.java VarHandleTestMethodHandleAccess$Type$.java
> 29:  * @run testng/othervm -XX:+EnableValhalla -XX:+EnablePrimitiveClasses -Diters=2000 VarHandleTestMethodHandleAccess$Type$

X-VarHandleTestXXX.java.template changes look wrong and maybe merge issue.  These files were updated recently.

For VarHandles tests, you only need to modify X-VarHandleTestXXX.java.template files to add `-XX:+EnablePrimitiveClasses` flag and then run `test/jdk/java/lang/invoke/VarHandles/generate-vh-tests.sh` that will regenerate the `test/jdk/java/lang/invoke/VarHandles/VarHandle*` tests.

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From mchung at openjdk.org  Tue Oct  4 19:01:25 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Tue, 4 Oct 2022 19:01:25 GMT
Subject: [lworld] RFR: 8294797: [lworld] update javadoc to remove references to
 primitive classes
Message-ID: <bXgNfbFOR8UE5mwxPLNUOJRNdtYkFxwlTGGkZu2UcFE=.5c04614c-c30e-4990-8f89-88328c58572a@github.com>

This reverts the javadoc changes for primitive classes and a couple leftover changes due to the previous design that is no longer applicable.

-------------

Commit messages:
 - 8294797: [lworld] update javadoc to remove references to primitive classes

Changes: https://git.openjdk.org/valhalla/pull/774/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=774&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8294797
  Stats: 81 lines in 13 files changed: 13 ins; 38 del; 30 mod
  Patch: https://git.openjdk.org/valhalla/pull/774.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/774/head:pull/774

PR: https://git.openjdk.org/valhalla/pull/774


From rriggs at openjdk.org  Tue Oct  4 19:22:38 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Tue, 4 Oct 2022 19:22:38 GMT
Subject: [lworld] RFR: 8294678: [lworld] Update AccessFlags tests to include
 ACC_IDENTITY and ACC_VALUE
Message-ID: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>

The tests for Class.accessFlags should be updated to check ACC_IDENTITY and ACC_VALUE. 
Update ClassAccessFlagTest and ObjectMethods test to check consistency between
Class.isIdentity() and Class.isValue() methods with the Class accessFlags().

-------------

Commit messages:
 - 8294678: [lworld] Update AccessFlags tests to include ACC_IDENTITY and ACC_VALUE

Changes: https://git.openjdk.org/valhalla/pull/775/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=775&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8294678
  Stats: 48 lines in 3 files changed: 30 ins; 0 del; 18 mod
  Patch: https://git.openjdk.org/valhalla/pull/775.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/775/head:pull/775

PR: https://git.openjdk.org/valhalla/pull/775


From mchung at openjdk.org  Tue Oct  4 20:07:46 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Tue, 4 Oct 2022 20:07:46 GMT
Subject: [lworld] RFR: 8294678: [lworld] Update AccessFlags tests to
 include ACC_IDENTITY and ACC_VALUE
In-Reply-To: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
References: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
Message-ID: <tGcPH6wjF95M_VG9ZlztqMQkQAV6HF3iBypBO-aTp8o=.36446e54-3a48-499f-b931-976a9008f142@github.com>

On Tue, 4 Oct 2022 19:16:50 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

> The tests for Class.accessFlags should be updated to check ACC_IDENTITY and ACC_VALUE. 
> Update ClassAccessFlagTest and ObjectMethods test to check consistency between
> Class.isIdentity() and Class.isValue() methods with the Class accessFlags().

src/java.base/share/classes/java/lang/Class.java line 1522:

> 1520:         final int mask = ((location == AccessFlag.Location.CLASS) ?
> 1521:                 getClassAccessFlagsRaw() : getModifiers()) |
> 1522:                 (isArray() ? Modifier.IDENTITY : 0);

For an array class `a`, `a.accessFlags()` and `AccessFlag.maskToAccessFlags(a.getModifiers(), Location.CLASS)` won't return the same `AccessFlag` set then.   The access flags for array type is special anyway.    It would be less confusing if `Class::accessFlags()` is consistent with `Class::getModifiers()`.

-------------

PR: https://git.openjdk.org/valhalla/pull/775


From mchung at openjdk.org  Tue Oct  4 20:12:02 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Tue, 4 Oct 2022 20:12:02 GMT
Subject: [lworld] RFR: 8294797: [lworld] update javadoc to remove
 references to primitive classes [v2]
In-Reply-To: <bXgNfbFOR8UE5mwxPLNUOJRNdtYkFxwlTGGkZu2UcFE=.5c04614c-c30e-4990-8f89-88328c58572a@github.com>
References: <bXgNfbFOR8UE5mwxPLNUOJRNdtYkFxwlTGGkZu2UcFE=.5c04614c-c30e-4990-8f89-88328c58572a@github.com>
Message-ID: <2tKW0IgDjcPu8OvHCyFEqs9zU5Lh5cAo6Z2j0SVOG-M=.c57be758-bc04-4469-9e9e-6101ddc743ec@github.com>

> This reverts the javadoc changes for primitive classes and a couple leftover changes due to the previous design that is no longer applicable.

Mandy Chung has refreshed the contents of this pull request, and previous commits have been removed. Incremental views are not available.

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/774/files
  - new: https://git.openjdk.org/valhalla/pull/774/files/7c2794d4..ee15e2d9

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=774&range=01
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=774&range=00-01

  Stats: 81 lines in 13 files changed: 38 ins; 13 del; 30 mod
  Patch: https://git.openjdk.org/valhalla/pull/774.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/774/head:pull/774

PR: https://git.openjdk.org/valhalla/pull/774


From mchung at openjdk.org  Tue Oct  4 20:12:03 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Tue, 4 Oct 2022 20:12:03 GMT
Subject: [lworld] Withdrawn: 8294797: [lworld] update javadoc to remove
 references to primitive classes
In-Reply-To: <bXgNfbFOR8UE5mwxPLNUOJRNdtYkFxwlTGGkZu2UcFE=.5c04614c-c30e-4990-8f89-88328c58572a@github.com>
References: <bXgNfbFOR8UE5mwxPLNUOJRNdtYkFxwlTGGkZu2UcFE=.5c04614c-c30e-4990-8f89-88328c58572a@github.com>
Message-ID: <IEJyxZHWCoynzFLRhjuN8P0LeqUBq03ETbEJrbtqwBo=.f3a73e3d-4e94-4d56-a88d-2cb3005726b1@github.com>

On Tue, 4 Oct 2022 18:53:23 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> This reverts the javadoc changes for primitive classes and a couple leftover changes due to the previous design that is no longer applicable.

This pull request has been closed without being integrated.

-------------

PR: https://git.openjdk.org/valhalla/pull/774


From mchung at openjdk.org  Tue Oct  4 20:14:17 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Tue, 4 Oct 2022 20:14:17 GMT
Subject: [lworld] RFR: 8294797: [lworld] update javadoc to remove references to
 primitive classes
Message-ID: <D4HB8FXUS5U-VC86isP1VN2vxb65dOYkcHcDu_vppbw=.bcc99e2d-2d65-493b-9194-89fd07142719@github.com>

This reverts the javadoc changes for primitive classes and a couple leftover changes due to the previous design that is no longer applicable.

-------------

Commit messages:
 - 8294797: [lworld] update javadoc to remove references to primitive classes

Changes: https://git.openjdk.org/valhalla/pull/776/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=776&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8294797
  Stats: 81 lines in 13 files changed: 13 ins; 38 del; 30 mod
  Patch: https://git.openjdk.org/valhalla/pull/776.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/776/head:pull/776

PR: https://git.openjdk.org/valhalla/pull/776


From rriggs at openjdk.org  Tue Oct  4 20:42:19 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Tue, 4 Oct 2022 20:42:19 GMT
Subject: [lworld] RFR: 8294678: [lworld] Update AccessFlags tests to
 include ACC_IDENTITY and ACC_VALUE [v2]
In-Reply-To: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
References: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
Message-ID: <uuoAm-C08-Rbal79SrT4tE8bmlQbsnRj2oolkBK8UV8=.68855c7d-af29-4870-9fbe-361d8163865c@github.com>

> The tests for Class.accessFlags should be updated to check ACC_IDENTITY and ACC_VALUE. 
> Update ClassAccessFlagTest and ObjectMethods test to check consistency between
> Class.isIdentity() and Class.isValue() methods with the Class accessFlags().

Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:

  Update ObjectMethods test

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/775/files
  - new: https://git.openjdk.org/valhalla/pull/775/files/8515d97e..16c95593

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=775&range=01
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=775&range=00-01

  Stats: 7 lines in 1 file changed: 0 ins; 0 del; 7 mod
  Patch: https://git.openjdk.org/valhalla/pull/775.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/775/head:pull/775

PR: https://git.openjdk.org/valhalla/pull/775


From rriggs at openjdk.org  Tue Oct  4 20:54:36 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Tue, 4 Oct 2022 20:54:36 GMT
Subject: [lworld] RFR: 8294678: [lworld] Update AccessFlags tests to
 include ACC_IDENTITY and ACC_VALUE [v2]
In-Reply-To: <tGcPH6wjF95M_VG9ZlztqMQkQAV6HF3iBypBO-aTp8o=.36446e54-3a48-499f-b931-976a9008f142@github.com>
References: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
 <tGcPH6wjF95M_VG9ZlztqMQkQAV6HF3iBypBO-aTp8o=.36446e54-3a48-499f-b931-976a9008f142@github.com>
Message-ID: <MaxUIDk3wNFljRjOGBuDQBlDyP9s4LTR8Byxn2euYtQ=.745262b3-6287-44f8-8a81-697047050837@github.com>

On Tue, 4 Oct 2022 20:03:59 GMT, Mandy Chung <mchung at openjdk.org> wrote:

>> Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Update ObjectMethods test
>
> src/java.base/share/classes/java/lang/Class.java line 1522:
> 
>> 1520:         final int mask = ((location == AccessFlag.Location.CLASS) ?
>> 1521:                 getClassAccessFlagsRaw() : getModifiers()) |
>> 1522:                 (isArray() ? Modifier.IDENTITY : 0);
> 
> For an array class `a`, `a.accessFlags()` and `AccessFlag.maskToAccessFlags(a.getModifiers(), Location.CLASS)` won't return the same `AccessFlag` set then.   The access flags for array type is special anyway.    It would be less confusing if `Class::accessFlags()` is consistent with `Class::getModifiers()`.

My goal was to have the Class.accessFlags() be consistent with Class methods `isIdentity()` and `isValue()`.
Unless we prevail in adding ACC_IDENTITY and ACC_VALUE to Class.getModifiers() there will be an inconsistency among the APIs.  `getModifiers()`, though the oldest, has the least flexibility.  The goal of the AccessFlags API was to be more flexible, correct, and extensible.  It is a question of what the end state should be and how to get there.
(The check for isArray() to select the INNER_CLASS location was from the original AccessFlag PR.)

-------------

PR: https://git.openjdk.org/valhalla/pull/775


From dsimms at openjdk.org  Wed Oct  5 06:26:34 2022
From: dsimms at openjdk.org (David Simms)
Date: Wed, 5 Oct 2022 06:26:34 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v11]
In-Reply-To: <uCGc4DSDNy5MnxtjrNAN3WrpwBRY5tZu0T87SQXktKc=.386c87c7-3fd0-42d9-92d2-b82962a19d8c@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <Gg15FmjHr9kUCPdShgnP7TzJ0RlSkkwShTFhN5IuIJY=.7889e629-0183-4c56-98ea-44adf7018baf@github.com>
 <uCGc4DSDNy5MnxtjrNAN3WrpwBRY5tZu0T87SQXktKc=.386c87c7-3fd0-42d9-92d2-b82962a19d8c@github.com>
Message-ID: <fFD_0xF0t3Clm1dv0y-ryp8joQR8zGLF_HJjEvajLvs=.c13fd44e-2270-4f81-a603-2255ea50bfa5@github.com>

On Tue, 4 Oct 2022 18:40:20 GMT, Mandy Chung <mchung at openjdk.org> wrote:

>> David Simms has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Review comments applied
>
> test/jdk/java/lang/invoke/VarHandles/X-VarHandleTestMethodHandleAccess.java.template line 29:
> 
>> 27:  * @test
>> 28:  * @compile -XDenablePrimitiveClasses Point.java Value.java VarHandleTestMethodHandleAccess$Type$.java
>> 29:  * @run testng/othervm -XX:+EnableValhalla -XX:+EnablePrimitiveClasses -Diters=2000 VarHandleTestMethodHandleAccess$Type$
> 
> X-VarHandleTestXXX.java.template changes look wrong and maybe merge issue.  These files were updated recently.
> 
> For VarHandles tests, you only need to modify X-VarHandleTestXXX.java.template files to add `-XX:+EnablePrimitiveClasses` flag and then run `test/jdk/java/lang/invoke/VarHandles/generate-vh-tests.sh` that will regenerate the `test/jdk/java/lang/invoke/VarHandles/VarHandle*` tests.

So the problem here is SPP. AFAIK, SPP doesn't support boolean expressions, and `X-VarHandleTestMethodHandleAccess.java.template` refers to both `Value` and `Point `.

I'm sure there are reasons why the build systems' C preprocessor isn't being used, at build time, and we remove generated source from the repository, but that feels out of scope for this change.

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From dsimms at openjdk.org  Wed Oct  5 06:28:47 2022
From: dsimms at openjdk.org (David Simms)
Date: Wed, 5 Oct 2022 06:28:47 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v12]
In-Reply-To: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
Message-ID: <ZDO-bjoRhsRRn9uwU0lxaNZ-DdZBKLZx-evVYl_D2mI=.24f27649-868f-4620-a493-8757eb03a93d@github.com>

> Runtime component of -XX:+EnablePrimitiveClasses switch

David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 24 commits:

 - Merge branch 'lworld' into 8288135
 - Review comments applied
 - Merge branch 'lworld' into 8288135
 - Extra assertion info for tracking spurious issue
 - supports_inline_types() refers to classfile and is separate from Enable flags
 - Further testing explicity using flags
 - Disable EnablePrimitiveClasses by default
 - Explictly flag more tests
 - Merge branch 'lworld' into 8288135
 - jdk_valhalla tests
 - ... and 14 more: https://git.openjdk.org/valhalla/compare/ee15e2d9...92e5253d

-------------

Changes: https://git.openjdk.org/valhalla/pull/727/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=727&range=11
  Stats: 1365 lines in 245 files changed: 836 ins; 65 del; 464 mod
  Patch: https://git.openjdk.org/valhalla/pull/727.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/727/head:pull/727

PR: https://git.openjdk.org/valhalla/pull/727


From thartmann at openjdk.org  Wed Oct  5 11:19:12 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Wed, 5 Oct 2022 11:19:12 GMT
Subject: [lworld] RFR: 8294013: [lworld] Missing class init barriers in c2i
 adapter
Message-ID: <H0jgr2Qy5VsX6GUpfwMVu3ZYdqJIhsjqJ4ejLW3E24c=.a117661e-7906-4796-8345-5942b5619c85@github.com>

Class initialization barriers for static methods are missing in the non-scalarized entry points of nmethods and c2i adapters. I extended the corresponding test to cover this case.

Thanks,
Tobias

-------------

Commit messages:
 - 8294013: [lworld] Missing class init barriers in c2i adapter

Changes: https://git.openjdk.org/valhalla/pull/778/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=778&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8294013
  Stats: 238 lines in 8 files changed: 89 ins; 76 del; 73 mod
  Patch: https://git.openjdk.org/valhalla/pull/778.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/778/head:pull/778

PR: https://git.openjdk.org/valhalla/pull/778


From thartmann at openjdk.org  Wed Oct  5 11:34:46 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Wed, 5 Oct 2022 11:34:46 GMT
Subject: [lworld] Integrated: 8294013: [lworld] Missing class init barriers in
 c2i adapter
In-Reply-To: <H0jgr2Qy5VsX6GUpfwMVu3ZYdqJIhsjqJ4ejLW3E24c=.a117661e-7906-4796-8345-5942b5619c85@github.com>
References: <H0jgr2Qy5VsX6GUpfwMVu3ZYdqJIhsjqJ4ejLW3E24c=.a117661e-7906-4796-8345-5942b5619c85@github.com>
Message-ID: <VsR58KmXaP4FKyrggIy332CiItKJ9qWnLHK1qfAqSJc=.adf31239-1682-40db-8454-ca2797c0fb0a@github.com>

On Wed, 5 Oct 2022 11:11:27 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

> Class initialization barriers for static methods are missing in the non-scalarized entry points of nmethods and c2i adapters. I extended the corresponding test to cover this case.
> 
> Thanks,
> Tobias

This pull request has now been integrated.

Changeset: 9932689c
Author:    Tobias Hartmann <thartmann at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/9932689c343ddfe3d5586ce6c3817f2f96d54258
Stats:     238 lines in 8 files changed: 89 ins; 76 del; 73 mod

8294013: [lworld] Missing class init barriers in c2i adapter

-------------

PR: https://git.openjdk.org/valhalla/pull/778


From fparain at openjdk.org  Wed Oct  5 12:03:38 2022
From: fparain at openjdk.org (Frederic Parain)
Date: Wed, 5 Oct 2022 12:03:38 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v12]
In-Reply-To: <ZDO-bjoRhsRRn9uwU0lxaNZ-DdZBKLZx-evVYl_D2mI=.24f27649-868f-4620-a493-8757eb03a93d@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <ZDO-bjoRhsRRn9uwU0lxaNZ-DdZBKLZx-evVYl_D2mI=.24f27649-868f-4620-a493-8757eb03a93d@github.com>
Message-ID: <KxXNjpvIR83iaNKBzLc3Mw4FVfqyjMf_MzJeHbJxGPM=.690d7226-e9e1-4a97-98b6-51735584c709@github.com>

On Wed, 5 Oct 2022 06:28:47 GMT, David Simms <dsimms at openjdk.org> wrote:

>> Runtime component of -XX:+EnablePrimitiveClasses switch
>
> David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 24 commits:
> 
>  - Merge branch 'lworld' into 8288135
>  - Review comments applied
>  - Merge branch 'lworld' into 8288135
>  - Extra assertion info for tracking spurious issue
>  - supports_inline_types() refers to classfile and is separate from Enable flags
>  - Further testing explicity using flags
>  - Disable EnablePrimitiveClasses by default
>  - Explictly flag more tests
>  - Merge branch 'lworld' into 8288135
>  - jdk_valhalla tests
>  - ... and 14 more: https://git.openjdk.org/valhalla/compare/ee15e2d9...92e5253d

Thanks for the changes.
However, recognized modifiers still need a fix, see proposed code.

Fred

src/hotspot/share/classfile/classFileParser.cpp line 3328:

> 3326:       recognized_modifiers |= JVM_ACC_MODULE;
> 3327:     }
> 3328:     if (supports_inline_types()) {

Suggestion:
if (supports_inline_types() {
  recognized_modifiers |= JVM_ACC_VALUE | JVM_ACC_IDENTITY;
  if (EnablePrimitiveClasses) {
    recognized_modifiers |= JVM_ACC_PRIMITIVE;
  }
}

-------------

Changes requested by fparain (Committer).

PR: https://git.openjdk.org/valhalla/pull/727


From fparain at openjdk.org  Wed Oct  5 12:33:40 2022
From: fparain at openjdk.org (Frederic Parain)
Date: Wed, 5 Oct 2022 12:33:40 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v12]
In-Reply-To: <ZDO-bjoRhsRRn9uwU0lxaNZ-DdZBKLZx-evVYl_D2mI=.24f27649-868f-4620-a493-8757eb03a93d@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <ZDO-bjoRhsRRn9uwU0lxaNZ-DdZBKLZx-evVYl_D2mI=.24f27649-868f-4620-a493-8757eb03a93d@github.com>
Message-ID: <7qQzYCZoCVIV9jok9e-Nz84KvXWe5d-qrzto2P3AjR0=.aaad566f-de54-4f7a-ab55-94c9d9daa20a@github.com>

On Wed, 5 Oct 2022 06:28:47 GMT, David Simms <dsimms at openjdk.org> wrote:

>> Runtime component of -XX:+EnablePrimitiveClasses switch
>
> David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 24 commits:
> 
>  - Merge branch 'lworld' into 8288135
>  - Review comments applied
>  - Merge branch 'lworld' into 8288135
>  - Extra assertion info for tracking spurious issue
>  - supports_inline_types() refers to classfile and is separate from Enable flags
>  - Further testing explicity using flags
>  - Disable EnablePrimitiveClasses by default
>  - Explictly flag more tests
>  - Merge branch 'lworld' into 8288135
>  - jdk_valhalla tests
>  - ... and 14 more: https://git.openjdk.org/valhalla/compare/ee15e2d9...92e5253d

Marked as reviewed by fparain (Committer).

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From rriggs at openjdk.org  Wed Oct  5 17:24:49 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Wed, 5 Oct 2022 17:24:49 GMT
Subject: [lworld] RFR: 8294678: [lworld] Update AccessFlags tests to
 include ACC_IDENTITY and ACC_VALUE [v3]
In-Reply-To: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
References: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
Message-ID: <tj409DQHoG0clEqkpZsYz9u_QZpydK5rkTtwn7GLF04=.5f1a220d-1c85-476f-9986-4d35b68bab27@github.com>

> The tests for Class.accessFlags should be updated to check ACC_IDENTITY and ACC_VALUE. 
> Update ClassAccessFlagTest and ObjectMethods test to check consistency between
> Class.isIdentity() and Class.isValue() methods with the Class accessFlags().

Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:

  Add test cases for interfaces and class with explicit identity and value class declarations

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/775/files
  - new: https://git.openjdk.org/valhalla/pull/775/files/16c95593..ef450fbe

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=775&range=02
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=775&range=01-02

  Stats: 37 lines in 1 file changed: 36 ins; 0 del; 1 mod
  Patch: https://git.openjdk.org/valhalla/pull/775.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/775/head:pull/775

PR: https://git.openjdk.org/valhalla/pull/775


From rriggs at openjdk.org  Wed Oct  5 17:44:34 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Wed, 5 Oct 2022 17:44:34 GMT
Subject: [lworld] RFR: 8294678: [lworld] Update AccessFlags tests to
 include ACC_IDENTITY and ACC_VALUE [v3]
In-Reply-To: <MaxUIDk3wNFljRjOGBuDQBlDyP9s4LTR8Byxn2euYtQ=.745262b3-6287-44f8-8a81-697047050837@github.com>
References: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
 <tGcPH6wjF95M_VG9ZlztqMQkQAV6HF3iBypBO-aTp8o=.36446e54-3a48-499f-b931-976a9008f142@github.com>
 <MaxUIDk3wNFljRjOGBuDQBlDyP9s4LTR8Byxn2euYtQ=.745262b3-6287-44f8-8a81-697047050837@github.com>
Message-ID: <RIdRNyk7jjgTKo6u-3sC3LZTG2DKJkK9cWFW-FhmrD8=.7daac6aa-1884-4394-9e24-72bd805f36b3@github.com>

On Tue, 4 Oct 2022 20:50:46 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

>> src/java.base/share/classes/java/lang/Class.java line 1522:
>> 
>>> 1520:         final int mask = ((location == AccessFlag.Location.CLASS) ?
>>> 1521:                 getClassAccessFlagsRaw() : getModifiers()) |
>>> 1522:                 (isArray() ? Modifier.IDENTITY : 0);
>> 
>> For an array class `a`, `a.accessFlags()` and `AccessFlag.maskToAccessFlags(a.getModifiers(), Location.CLASS)` won't return the same `AccessFlag` set then.   The access flags for array type is special anyway.    It would be less confusing if `Class::accessFlags()` is consistent with `Class::getModifiers()`.
>
> My goal was to have the Class.accessFlags() be consistent with Class methods `isIdentity()` and `isValue()`.
> Unless we prevail in adding ACC_IDENTITY and ACC_VALUE to Class.getModifiers() there will be an inconsistency among the APIs.  `getModifiers()`, though the oldest, has the least flexibility.  The goal of the AccessFlags API was to be more flexible, correct, and extensible.  It is a question of what the end state should be and how to get there.
> (The check for isArray() to select the INNER_CLASS location was from the original AccessFlag PR.)

I'd like to commit this PR as is on this point and propose a separate PR to add ACC_IDENTITY and ACC_VALUE to the modifiers for arrays to be able to judge the impact and get further review.
https://bugs.openjdk.org/browse/JDK-8294866

-------------

PR: https://git.openjdk.org/valhalla/pull/775


From mchung at openjdk.org  Wed Oct  5 18:38:44 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Wed, 5 Oct 2022 18:38:44 GMT
Subject: [lworld] RFR: 8294678: [lworld] Update AccessFlags tests to
 include ACC_IDENTITY and ACC_VALUE [v3]
In-Reply-To: <RIdRNyk7jjgTKo6u-3sC3LZTG2DKJkK9cWFW-FhmrD8=.7daac6aa-1884-4394-9e24-72bd805f36b3@github.com>
References: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
 <tGcPH6wjF95M_VG9ZlztqMQkQAV6HF3iBypBO-aTp8o=.36446e54-3a48-499f-b931-976a9008f142@github.com>
 <MaxUIDk3wNFljRjOGBuDQBlDyP9s4LTR8Byxn2euYtQ=.745262b3-6287-44f8-8a81-697047050837@github.com>
 <RIdRNyk7jjgTKo6u-3sC3LZTG2DKJkK9cWFW-FhmrD8=.7daac6aa-1884-4394-9e24-72bd805f36b3@github.com>
Message-ID: <LZyDrmZlZyAkTmH6FF1ktrpGfjAm1hbQ0n7hQJQ_iKE=.aad24d02-b5a1-44d0-8d3b-1925f0744061@github.com>

On Wed, 5 Oct 2022 17:41:12 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

>> My goal was to have the Class.accessFlags() be consistent with Class methods `isIdentity()` and `isValue()`.
>> Unless we prevail in adding ACC_IDENTITY and ACC_VALUE to Class.getModifiers() there will be an inconsistency among the APIs.  `getModifiers()`, though the oldest, has the least flexibility.  The goal of the AccessFlags API was to be more flexible, correct, and extensible.  It is a question of what the end state should be and how to get there.
>> (The check for isArray() to select the INNER_CLASS location was from the original AccessFlag PR.)
>
> I'd like to commit this PR as is on this point and propose a separate PR to add ACC_IDENTITY and ACC_VALUE to the modifiers for arrays to be able to judge the impact and get further review.
> https://bugs.openjdk.org/browse/JDK-8294866

I agree that we should discuss the modifiers for arrays in a follow-on issue.   I suggest this PR to keep  `Class::getModifiers` and `Class::accessFlags` consistent to match the specification.    JDK-8294866 will do the spec change altogether.

-------------

PR: https://git.openjdk.org/valhalla/pull/775


From mchung at openjdk.org  Wed Oct  5 18:55:39 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Wed, 5 Oct 2022 18:55:39 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v11]
In-Reply-To: <fFD_0xF0t3Clm1dv0y-ryp8joQR8zGLF_HJjEvajLvs=.c13fd44e-2270-4f81-a603-2255ea50bfa5@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <Gg15FmjHr9kUCPdShgnP7TzJ0RlSkkwShTFhN5IuIJY=.7889e629-0183-4c56-98ea-44adf7018baf@github.com>
 <uCGc4DSDNy5MnxtjrNAN3WrpwBRY5tZu0T87SQXktKc=.386c87c7-3fd0-42d9-92d2-b82962a19d8c@github.com>
 <fFD_0xF0t3Clm1dv0y-ryp8joQR8zGLF_HJjEvajLvs=.c13fd44e-2270-4f81-a603-2255ea50bfa5@github.com>
Message-ID: <a1gwgysaykrZLAyacsggPPStf45T__IpabXI9dJ1iQ4=.7c62ea62-2967-4c2e-be5b-53f6f4c4178a@github.com>

On Wed, 5 Oct 2022 06:22:29 GMT, David Simms <dsimms at openjdk.org> wrote:

>> test/jdk/java/lang/invoke/VarHandles/X-VarHandleTestMethodHandleAccess.java.template line 29:
>> 
>>> 27:  * @test
>>> 28:  * @compile -XDenablePrimitiveClasses Point.java Value.java VarHandleTestMethodHandleAccess$Type$.java
>>> 29:  * @run testng/othervm -XX:+EnableValhalla -XX:+EnablePrimitiveClasses -Diters=2000 VarHandleTestMethodHandleAccess$Type$
>> 
>> X-VarHandleTestXXX.java.template changes look wrong and maybe merge issue.  These files were updated recently.
>> 
>> For VarHandles tests, you only need to modify X-VarHandleTestXXX.java.template files to add `-XX:+EnablePrimitiveClasses` flag and then run `test/jdk/java/lang/invoke/VarHandles/generate-vh-tests.sh` that will regenerate the `test/jdk/java/lang/invoke/VarHandles/VarHandle*` tests.
>
> So the problem here is SPP. AFAIK, SPP doesn't support boolean expressions, and `X-VarHandleTestMethodHandleAccess.java.template` refers to both `Value` and `Point `.
> 
> I'm sure there are reasons why the build systems' C preprocessor isn't being used, at build time, and we remove generated source from the repository, but that feels out of scope for this change.

I think there may be confusion due to the name `Value`.   Both `Point` and `Value` class are value classes.   `-KValue`  implies using primitive class.   That was done in LW1 or LW2 before the value objects JEP.
 It's on my TODO to update these tests for value objects.
 
It is okay to remove `if[Point]` case that was added to reduce `iters` from 20000 to 2000.  If you want to remove `if[Point]`, use `-Diters=20000` instead.

Suggestion:

 * @run testng/othervm -XX:+EnableValhalla -XX:+EnablePrimitiveClasses -Diters=20000 VarHandleTestMethodHandleAccess$Type$

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From mchung at openjdk.org  Wed Oct  5 18:55:41 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Wed, 5 Oct 2022 18:55:41 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v12]
In-Reply-To: <ZDO-bjoRhsRRn9uwU0lxaNZ-DdZBKLZx-evVYl_D2mI=.24f27649-868f-4620-a493-8757eb03a93d@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <ZDO-bjoRhsRRn9uwU0lxaNZ-DdZBKLZx-evVYl_D2mI=.24f27649-868f-4620-a493-8757eb03a93d@github.com>
Message-ID: <Jge-e4StMVUBuUzdp9IXedUl8LMUt-xyjDBwrt61Y_8=.d13da925-b44a-4134-b62d-d860abeecf5b@github.com>

On Wed, 5 Oct 2022 06:28:47 GMT, David Simms <dsimms at openjdk.org> wrote:

>> Runtime component of -XX:+EnablePrimitiveClasses switch
>
> David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 24 commits:
> 
>  - Merge branch 'lworld' into 8288135
>  - Review comments applied
>  - Merge branch 'lworld' into 8288135
>  - Extra assertion info for tracking spurious issue
>  - supports_inline_types() refers to classfile and is separate from Enable flags
>  - Further testing explicity using flags
>  - Disable EnablePrimitiveClasses by default
>  - Explictly flag more tests
>  - Merge branch 'lworld' into 8288135
>  - jdk_valhalla tests
>  - ... and 14 more: https://git.openjdk.org/valhalla/compare/ee15e2d9...92e5253d

test/jdk/java/lang/invoke/VarHandles/X-VarHandleTestMethodType.java.template line 29:

> 27:  * @test
> 28:  * @bug 8156486
> 29: #if[Point]

This change intends to make `@compile....` to be conditional.   So it can simply do:

Suggestion:

#if[Point]
     * @compile -XDenablePrimitiveClasses Point.java Value.java VarHandleTestMethodType$Type$.java
#end[Point]


No need to duplicate all `@run` statements.

test/jdk/valhalla/valuetypes/QTypeDescriptorTest.java line 30:

> 28:  * @compile -XDenablePrimitiveClasses QTypeDescriptorTest.java
> 29:  * @run testng/othervm -XX:+EnableValhalla -XX:+EnablePrimitiveClasses QTypeDescriptorTest
> 30:  * @compile -XDenablePrimitiveClasses QTypeDescriptorTest.java

This `@compile` seems duplicated.   Since you are on this file, do u mind taking it out?

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From dsimms at openjdk.org  Wed Oct  5 18:55:41 2022
From: dsimms at openjdk.org (David Simms)
Date: Wed, 5 Oct 2022 18:55:41 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v12]
In-Reply-To: <Jge-e4StMVUBuUzdp9IXedUl8LMUt-xyjDBwrt61Y_8=.d13da925-b44a-4134-b62d-d860abeecf5b@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
 <ZDO-bjoRhsRRn9uwU0lxaNZ-DdZBKLZx-evVYl_D2mI=.24f27649-868f-4620-a493-8757eb03a93d@github.com>
 <Jge-e4StMVUBuUzdp9IXedUl8LMUt-xyjDBwrt61Y_8=.d13da925-b44a-4134-b62d-d860abeecf5b@github.com>
Message-ID: <cncTNUCImItM32Za2ckNTpvvqUjpzDaWtoJyxDRaqkI=.c20bbf28-75b0-4a9a-b346-000b1ec1daa3@github.com>

On Wed, 5 Oct 2022 18:46:31 GMT, Mandy Chung <mchung at openjdk.org> wrote:

>> David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 24 commits:
>> 
>>  - Merge branch 'lworld' into 8288135
>>  - Review comments applied
>>  - Merge branch 'lworld' into 8288135
>>  - Extra assertion info for tracking spurious issue
>>  - supports_inline_types() refers to classfile and is separate from Enable flags
>>  - Further testing explicity using flags
>>  - Disable EnablePrimitiveClasses by default
>>  - Explictly flag more tests
>>  - Merge branch 'lworld' into 8288135
>>  - jdk_valhalla tests
>>  - ... and 14 more: https://git.openjdk.org/valhalla/compare/ee15e2d9...92e5253d
>
> test/jdk/java/lang/invoke/VarHandles/X-VarHandleTestMethodType.java.template line 29:
> 
>> 27:  * @test
>> 28:  * @bug 8156486
>> 29: #if[Point]
> 
> This change intends to make `@compile....` to be conditional.   So it can simply do:
> 
> Suggestion:
> 
> #if[Point]
>      * @compile -XDenablePrimitiveClasses Point.java Value.java VarHandleTestMethodType$Type$.java
> #end[Point]
> 
> 
> No need to duplicate all `@run` statements.

Check

> test/jdk/valhalla/valuetypes/QTypeDescriptorTest.java line 30:
> 
>> 28:  * @compile -XDenablePrimitiveClasses QTypeDescriptorTest.java
>> 29:  * @run testng/othervm -XX:+EnableValhalla -XX:+EnablePrimitiveClasses QTypeDescriptorTest
>> 30:  * @compile -XDenablePrimitiveClasses QTypeDescriptorTest.java
> 
> This `@compile` seems duplicated.   Since you are on this file, do u mind taking it out?

NP

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From vromero at openjdk.org  Wed Oct  5 19:15:43 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Wed, 5 Oct 2022 19:15:43 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v4]
In-Reply-To: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
References: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
Message-ID: <e6_zF5JyfXI9cvpKiSM-gPl47mFnG0gBPOMMyjySFg4=.9254cd72-06d0-4e99-b08c-66d46c34ca8f@github.com>

> Value classes constructor should be named <vnew> instead of <init>. This PR can't be pushed without the corresponding support being available on the VM side

Vicente Romero has updated the pull request incrementally with one additional commit since the last revision:

  javac is generating <vnew> method for abstract value classes

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/771/files
  - new: https://git.openjdk.org/valhalla/pull/771/files/6264d345..eaa6b66d

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=771&range=03
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=771&range=02-03

  Stats: 34 lines in 10 files changed: 4 ins; 16 del; 14 mod
  Patch: https://git.openjdk.org/valhalla/pull/771.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/771/head:pull/771

PR: https://git.openjdk.org/valhalla/pull/771


From rriggs at openjdk.org  Wed Oct  5 20:24:35 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Wed, 5 Oct 2022 20:24:35 GMT
Subject: [lworld] RFR: 8294678: [lworld] Update AccessFlags tests to
 include ACC_IDENTITY and ACC_VALUE [v4]
In-Reply-To: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
References: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
Message-ID: <QpXRsXXc3DXR9gyb_vbupbYOamkHiGmU1QxTzDrRifg=.b44faa79-e7a5-4116-a36f-10756479b2d6@github.com>

> The tests for Class.accessFlags should be updated to check ACC_IDENTITY and ACC_VALUE. 
> Update ClassAccessFlagTest and ObjectMethods test to check consistency between
> Class.isIdentity() and Class.isValue() methods with the Class accessFlags().

Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:

  Revert changes to AccessFlags that would return IDENTITY for array classes.

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/775/files
  - new: https://git.openjdk.org/valhalla/pull/775/files/ef450fbe..07774305

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=775&range=03
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=775&range=02-03

  Stats: 12 lines in 3 files changed: 1 ins; 2 del; 9 mod
  Patch: https://git.openjdk.org/valhalla/pull/775.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/775/head:pull/775

PR: https://git.openjdk.org/valhalla/pull/775


From rriggs at openjdk.org  Wed Oct  5 21:19:27 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Wed, 5 Oct 2022 21:19:27 GMT
Subject: [lworld] RFR: 8294678: [lworld] Update AccessFlags tests to
 include ACC_IDENTITY and ACC_VALUE [v4]
In-Reply-To: <LZyDrmZlZyAkTmH6FF1ktrpGfjAm1hbQ0n7hQJQ_iKE=.aad24d02-b5a1-44d0-8d3b-1925f0744061@github.com>
References: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
 <tGcPH6wjF95M_VG9ZlztqMQkQAV6HF3iBypBO-aTp8o=.36446e54-3a48-499f-b931-976a9008f142@github.com>
 <MaxUIDk3wNFljRjOGBuDQBlDyP9s4LTR8Byxn2euYtQ=.745262b3-6287-44f8-8a81-697047050837@github.com>
 <RIdRNyk7jjgTKo6u-3sC3LZTG2DKJkK9cWFW-FhmrD8=.7daac6aa-1884-4394-9e24-72bd805f36b3@github.com>
 <LZyDrmZlZyAkTmH6FF1ktrpGfjAm1hbQ0n7hQJQ_iKE=.aad24d02-b5a1-44d0-8d3b-1925f0744061@github.com>
Message-ID: <B4A9oc3z2sDmcYOQtbd9lNIpDv6HX5wlWItOBXg2w9Q=.c2724076-3e84-432b-a3f6-194376464cc4@github.com>

On Wed, 5 Oct 2022 18:36:29 GMT, Mandy Chung <mchung at openjdk.org> wrote:

>> I'd like to commit this PR as is on this point and propose a separate PR to add ACC_IDENTITY and ACC_VALUE to the modifiers for arrays to be able to judge the impact and get further review.
>> https://bugs.openjdk.org/browse/JDK-8294866
>
> I agree that we should discuss the modifiers for arrays in a follow-on issue.   I suggest this PR to keep  `Class::getModifiers` and `Class::accessFlags` consistent to match the specification.    JDK-8294866 will do the spec change altogether.

done

-------------

PR: https://git.openjdk.org/valhalla/pull/775


From mchung at openjdk.org  Wed Oct  5 22:42:43 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Wed, 5 Oct 2022 22:42:43 GMT
Subject: [lworld] RFR: 8294678: [lworld] Update AccessFlags tests to
 include ACC_IDENTITY and ACC_VALUE [v4]
In-Reply-To: <QpXRsXXc3DXR9gyb_vbupbYOamkHiGmU1QxTzDrRifg=.b44faa79-e7a5-4116-a36f-10756479b2d6@github.com>
References: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
 <QpXRsXXc3DXR9gyb_vbupbYOamkHiGmU1QxTzDrRifg=.b44faa79-e7a5-4116-a36f-10756479b2d6@github.com>
Message-ID: <zqPd1koq1OXNync7VsT9ZPc8gRrbNVl5sStbwOy1wV4=.bc02f9de-59f7-4653-9126-85281eb03245@github.com>

On Wed, 5 Oct 2022 20:24:35 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

>> The tests for Class.accessFlags should be updated to check ACC_IDENTITY and ACC_VALUE. 
>> Update ClassAccessFlagTest and ObjectMethods test to check consistency between
>> Class.isIdentity() and Class.isValue() methods with the Class accessFlags().
>
> Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Revert changes to AccessFlags that would return IDENTITY for array classes.

Marked as reviewed by mchung (Committer).

-------------

PR: https://git.openjdk.org/valhalla/pull/775


From vromero at openjdk.org  Thu Oct  6 03:27:04 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Thu, 6 Oct 2022 03:27:04 GMT
Subject: [lworld] Integrated: 8294774: add more regression tests for value
 objects
Message-ID: <KuUGdE2CDe67aQI-Y8EkvgG_lidz4eR3aAszLkC2EBk=.7f0f9ba9-d4b8-46ff-8291-13abd230e30b@github.com>

additional test coverage for value objects

-------------

Commit messages:
 - adding tests for compiler keys in the not-yet file
 - updating compiler error key
 - 8294774: add more regression tests for value objects

Changes: https://git.openjdk.org/valhalla/pull/777/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=777&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8294774
  Stats: 559 lines in 20 files changed: 469 ins; 37 del; 53 mod
  Patch: https://git.openjdk.org/valhalla/pull/777.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/777/head:pull/777

PR: https://git.openjdk.org/valhalla/pull/777


From vromero at openjdk.org  Thu Oct  6 03:27:05 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Thu, 6 Oct 2022 03:27:05 GMT
Subject: [lworld] Integrated: 8294774: add more regression tests for value
 objects
In-Reply-To: <KuUGdE2CDe67aQI-Y8EkvgG_lidz4eR3aAszLkC2EBk=.7f0f9ba9-d4b8-46ff-8291-13abd230e30b@github.com>
References: <KuUGdE2CDe67aQI-Y8EkvgG_lidz4eR3aAszLkC2EBk=.7f0f9ba9-d4b8-46ff-8291-13abd230e30b@github.com>
Message-ID: <0T0LzOLsx14o2hTgNW7WjNfd_jYW3Cyke349NjVgDsc=.ea56ae35-a6e3-43d7-8285-a3914f2c9687@github.com>

On Tue, 4 Oct 2022 20:59:26 GMT, Vicente Romero <vromero at openjdk.org> wrote:

> additional test coverage for value objects

This pull request has now been integrated.

Changeset: 6d4ad250
Author:    Vicente Romero <vromero at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/6d4ad25077f07b978919b0fb40771cd84cc5bbed
Stats:     559 lines in 20 files changed: 469 ins; 37 del; 53 mod

8294774: add more regression tests for value objects

-------------

PR: https://git.openjdk.org/valhalla/pull/777


From rriggs at openjdk.org  Thu Oct  6 14:20:20 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Thu, 6 Oct 2022 14:20:20 GMT
Subject: [lworld] Integrated: 8294678: [lworld] Update AccessFlags tests to
 include ACC_IDENTITY and ACC_VALUE
In-Reply-To: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
References: <49z9wFHJu4vjRaf8K1FSAnErUwrDpJ8IfWW-Vups-SI=.a96a670b-ad69-43c3-957f-6bccdac826f1@github.com>
Message-ID: <wqey1fLOGWoEFS5Zo66Co_z_r9AO7g-1OEvGcLEkEhM=.0669727c-2843-40d6-8bc8-d94b0e80a2fe@github.com>

On Tue, 4 Oct 2022 19:16:50 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

> The tests for Class.accessFlags should be updated to check ACC_IDENTITY and ACC_VALUE. 
> Update ClassAccessFlagTest and ObjectMethods test to check consistency between
> Class.isIdentity() and Class.isValue() methods with the Class accessFlags().

This pull request has now been integrated.

Changeset: 866adb67
Author:    Roger Riggs <rriggs at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/866adb676c5a22133e5dcb6cad84da2a104844b2
Stats:     82 lines in 3 files changed: 65 ins; 0 del; 17 mod

8294678: [lworld] Update AccessFlags tests to include ACC_IDENTITY and ACC_VALUE

Reviewed-by: mchung

-------------

PR: https://git.openjdk.org/valhalla/pull/775


From rriggs at openjdk.org  Thu Oct  6 16:40:34 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Thu, 6 Oct 2022 16:40:34 GMT
Subject: [lworld] RFR: 8294797: [lworld] update javadoc to remove
 references to primitive classes
In-Reply-To: <D4HB8FXUS5U-VC86isP1VN2vxb65dOYkcHcDu_vppbw=.bcc99e2d-2d65-493b-9194-89fd07142719@github.com>
References: <D4HB8FXUS5U-VC86isP1VN2vxb65dOYkcHcDu_vppbw=.bcc99e2d-2d65-493b-9194-89fd07142719@github.com>
Message-ID: <jZ0Gk0XxC90WHKAjdnN7M75re5K2i_guz3lqBvUg88Y=.1659a712-9740-4a61-a338-d7562430378a@github.com>

On Tue, 4 Oct 2022 20:08:58 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> This reverts the javadoc changes for primitive classes and a couple leftover changes due to the previous design that is no longer applicable.

src/java.base/share/classes/java/lang/ref/PhantomReference.java line 94:

> 92:      * @param q the queue with which the reference is to be registered,
> 93:      *          or {@code null} if registration is not required
> 94:      * @throws IdentityException if the referent is an instance of an

Instead this should refer to isIdentity(), the implementation uses Objects.isIdentityObject(referent).
Perhaps:
@throws IdentityException if the referent is not an {@link Objects#isIdentityObject(Object o) identity object}.

The same applied to SoftReference and WeakReference. (below)

-------------

PR: https://git.openjdk.org/valhalla/pull/776


From vromero at openjdk.org  Thu Oct  6 17:15:34 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Thu, 6 Oct 2022 17:15:34 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v5]
In-Reply-To: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
References: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
Message-ID: <5pyq7z0DGN68O4Dsg6LYTGeJPVAUnVK6kl8B4x0A0PQ=.ecc75164-1619-431a-a27f-9d0d0db43f67@github.com>

> Value classes constructor should be named <vnew> instead of <init>. This PR can't be pushed without the corresponding support being available on the VM side

Vicente Romero has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains five commits:

 - Merge branch 'lworld' into JDK-8247569
 - javac is generating <vnew> method for abstract value classes
 - fixing regression tests
 - removing experimental code
 - 8247569: [lworld] Align with spec for <vnew> methods

-------------

Changes: https://git.openjdk.org/valhalla/pull/771/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=771&range=04
  Stats: 293 lines in 47 files changed: 87 ins; 59 del; 147 mod
  Patch: https://git.openjdk.org/valhalla/pull/771.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/771/head:pull/771

PR: https://git.openjdk.org/valhalla/pull/771


From mchung at openjdk.org  Thu Oct  6 17:38:34 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Thu, 6 Oct 2022 17:38:34 GMT
Subject: [lworld] RFR: 8294797: [lworld] update javadoc to remove
 references to primitive classes [v2]
In-Reply-To: <D4HB8FXUS5U-VC86isP1VN2vxb65dOYkcHcDu_vppbw=.bcc99e2d-2d65-493b-9194-89fd07142719@github.com>
References: <D4HB8FXUS5U-VC86isP1VN2vxb65dOYkcHcDu_vppbw=.bcc99e2d-2d65-493b-9194-89fd07142719@github.com>
Message-ID: <2JYHyvfIcKoVYmT-pGMABqxhjqWMaYEV0ii5meOfkZk=.869ec76f-c613-4088-88a9-1dea0e103f6d@github.com>

> This reverts the javadoc changes for primitive classes and a couple leftover changes due to the previous design that is no longer applicable.

Mandy Chung has updated the pull request with a new target base due to a merge or a rebase. The incremental webrev excludes the unrelated changes brought in by the merge/rebase. The pull request contains three additional commits since the last revision:

 - update @throws for references
 - Merge branch 'lworld' of https://github.com/openjdk/valhalla into javadoc-fix
 - 8294797: [lworld] update javadoc to remove references to primitive classes

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/776/files
  - new: https://git.openjdk.org/valhalla/pull/776/files/7c2794d4..f0d6634b

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=776&range=01
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=776&range=00-01

  Stats: 891 lines in 35 files changed: 624 ins; 113 del; 154 mod
  Patch: https://git.openjdk.org/valhalla/pull/776.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/776/head:pull/776

PR: https://git.openjdk.org/valhalla/pull/776


From mchung at openjdk.org  Thu Oct  6 17:38:34 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Thu, 6 Oct 2022 17:38:34 GMT
Subject: [lworld] RFR: 8294797: [lworld] update javadoc to remove
 references to primitive classes [v2]
In-Reply-To: <jZ0Gk0XxC90WHKAjdnN7M75re5K2i_guz3lqBvUg88Y=.1659a712-9740-4a61-a338-d7562430378a@github.com>
References: <D4HB8FXUS5U-VC86isP1VN2vxb65dOYkcHcDu_vppbw=.bcc99e2d-2d65-493b-9194-89fd07142719@github.com>
 <jZ0Gk0XxC90WHKAjdnN7M75re5K2i_guz3lqBvUg88Y=.1659a712-9740-4a61-a338-d7562430378a@github.com>
Message-ID: <AGFtWmRghErmd1ZC0kyAH1M8OS-9tPCSiRmZxFoZOsU=.967c5ac3-a6be-488b-b787-ef224264cf0b@github.com>

On Thu, 6 Oct 2022 16:34:13 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

>> Mandy Chung has updated the pull request with a new target base due to a merge or a rebase. The incremental webrev excludes the unrelated changes brought in by the merge/rebase. The pull request contains three additional commits since the last revision:
>> 
>>  - update @throws for references
>>  - Merge branch 'lworld' of https://github.com/openjdk/valhalla into javadoc-fix
>>  - 8294797: [lworld] update javadoc to remove references to primitive classes
>
> src/java.base/share/classes/java/lang/ref/PhantomReference.java line 94:
> 
>> 92:      * @param q the queue with which the reference is to be registered,
>> 93:      *          or {@code null} if registration is not required
>> 94:      * @throws IdentityException if the referent is an instance of an
> 
> Instead this should refer to isIdentity(), the implementation uses Objects.isIdentityObject(referent).
> Perhaps:
> @throws IdentityException if the referent is not an {@link Objects#isIdentityObject(Object o) identity object}.
> 
> The same applied to SoftReference and WeakReference. (below)

updated.

-------------

PR: https://git.openjdk.org/valhalla/pull/776


From rriggs at openjdk.org  Thu Oct  6 17:42:14 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Thu, 6 Oct 2022 17:42:14 GMT
Subject: [lworld] RFR: 8294797: [lworld] update javadoc to remove
 references to primitive classes [v2]
In-Reply-To: <2JYHyvfIcKoVYmT-pGMABqxhjqWMaYEV0ii5meOfkZk=.869ec76f-c613-4088-88a9-1dea0e103f6d@github.com>
References: <D4HB8FXUS5U-VC86isP1VN2vxb65dOYkcHcDu_vppbw=.bcc99e2d-2d65-493b-9194-89fd07142719@github.com>
 <2JYHyvfIcKoVYmT-pGMABqxhjqWMaYEV0ii5meOfkZk=.869ec76f-c613-4088-88a9-1dea0e103f6d@github.com>
Message-ID: <jC3b3ion2NnGx8FhB2C7xgvQP3OmJQuLJFXEQPqjQAU=.28b0cd6c-604b-4606-ac2f-ee9f41916d8a@github.com>

On Thu, 6 Oct 2022 17:38:34 GMT, Mandy Chung <mchung at openjdk.org> wrote:

>> This reverts the javadoc changes for primitive classes and a couple leftover changes due to the previous design that is no longer applicable.
>
> Mandy Chung has updated the pull request with a new target base due to a merge or a rebase. The incremental webrev excludes the unrelated changes brought in by the merge/rebase. The pull request contains three additional commits since the last revision:
> 
>  - update @throws for references
>  - Merge branch 'lworld' of https://github.com/openjdk/valhalla into javadoc-fix
>  - 8294797: [lworld] update javadoc to remove references to primitive classes

Marked as reviewed by rriggs (Committer).

-------------

PR: https://git.openjdk.org/valhalla/pull/776


From mchung at openjdk.org  Thu Oct  6 18:13:32 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Thu, 6 Oct 2022 18:13:32 GMT
Subject: [lworld] Integrated: 8294797: [lworld] update javadoc to remove
 references to primitive classes
In-Reply-To: <D4HB8FXUS5U-VC86isP1VN2vxb65dOYkcHcDu_vppbw=.bcc99e2d-2d65-493b-9194-89fd07142719@github.com>
References: <D4HB8FXUS5U-VC86isP1VN2vxb65dOYkcHcDu_vppbw=.bcc99e2d-2d65-493b-9194-89fd07142719@github.com>
Message-ID: <nN0dLwPzcIVV8skwSHYFZDT5g6rn3aHn5hRhkg-SH_4=.fa772dcd-e348-4c0b-af18-6fd7db7e17c4@github.com>

On Tue, 4 Oct 2022 20:08:58 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> This reverts the javadoc changes for primitive classes and a couple leftover changes due to the previous design that is no longer applicable.

This pull request has now been integrated.

Changeset: 2b06f9ed
Author:    Mandy Chung <mchung at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/2b06f9ed8b64317ea6f4dafb3559c16a6d7c9c95
Stats:     88 lines in 14 files changed: 14 ins; 38 del; 36 mod

8294797: [lworld] update javadoc to remove references to primitive classes

Reviewed-by: rriggs

-------------

PR: https://git.openjdk.org/valhalla/pull/776


From vromero at openjdk.org  Fri Oct  7 00:10:07 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Fri, 7 Oct 2022 00:10:07 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v6]
In-Reply-To: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
References: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
Message-ID: <IeS566OKQQBtEOz06dT2wZX6Wjm1Q6M9ZPkCq_nE5CI=.a4f8cfa9-095a-4ed3-94a7-9aca2241ed52@github.com>

> Value classes constructor should be named <vnew> instead of <init>. This PR can't be pushed without the corresponding support being available on the VM side

Vicente Romero has updated the pull request incrementally with one additional commit since the last revision:

  adding regression tests

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/771/files
  - new: https://git.openjdk.org/valhalla/pull/771/files/f984f661..7f01be1a

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=771&range=05
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=771&range=04-05

  Stats: 108 lines in 1 file changed: 108 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/valhalla/pull/771.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/771/head:pull/771

PR: https://git.openjdk.org/valhalla/pull/771


From vromero at openjdk.org  Fri Oct  7 00:23:18 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Fri, 7 Oct 2022 00:23:18 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v7]
In-Reply-To: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
References: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
Message-ID: <OmsJR8KNIKd1CyM_F-vwHFOcB_PVDW540WcKz5MkwVQ=.48b16cd0-be89-4281-a904-bf864d432bc9@github.com>

> Value classes constructor should be named <vnew> instead of <init>. This PR can't be pushed without the corresponding support being available on the VM side

Vicente Romero has updated the pull request incrementally with one additional commit since the last revision:

  adding missing comment

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/771/files
  - new: https://git.openjdk.org/valhalla/pull/771/files/7f01be1a..b4d69291

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=771&range=06
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=771&range=05-06

  Stats: 1 line in 1 file changed: 1 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/valhalla/pull/771.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/771/head:pull/771

PR: https://git.openjdk.org/valhalla/pull/771


From mchung at openjdk.org  Fri Oct  7 01:03:14 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Fri, 7 Oct 2022 01:03:14 GMT
Subject: [lworld] RFR: 8270852: [lworld] bytecodes spinned at runtime should
 set version to 65535 if preview feature enabled
Message-ID: <A-a5kqNeJVz2HgnfPWj07sFU3uA07udSDqqzXtF7UQY=.5131573c-4100-43f5-8901-fd62f016610c@github.com>

This patch prepares for lworld to switch to a preview feature.  Also fix the generation of lambda classes and proxy classes to have `Preload` attribute for value classes (not only primitive classes).

-------------

Commit messages:
 - cleanup import
 - 8270852: [lworld] bytecodes spinned at runtime should set minor_version to 65535 if preview feature enabled

Changes: https://git.openjdk.org/valhalla/pull/779/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=779&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8270852
  Stats: 17 lines in 2 files changed: 9 ins; 3 del; 5 mod
  Patch: https://git.openjdk.org/valhalla/pull/779.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/779/head:pull/779

PR: https://git.openjdk.org/valhalla/pull/779


From dsimms at openjdk.org  Fri Oct  7 06:46:55 2022
From: dsimms at openjdk.org (David Simms)
Date: Fri, 7 Oct 2022 06:46:55 GMT
Subject: [lworld] RFR: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses [v13]
In-Reply-To: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
Message-ID: <dcSUX_Olz0kjhjjT0jFsBCnN7k0VMYOd0MrpU8rT1rw=.d259cf87-b5a4-4d42-8d5e-a7550d26ac96@github.com>

> Runtime component of -XX:+EnablePrimitiveClasses switch

David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 26 commits:

 - Merge branch 'lworld' into 8288135
 - Review comments applied
 - Merge branch 'lworld' into 8288135
 - Review comments applied
 - Merge branch 'lworld' into 8288135
 - Extra assertion info for tracking spurious issue
 - supports_inline_types() refers to classfile and is separate from Enable flags
 - Further testing explicity using flags
 - Disable EnablePrimitiveClasses by default
 - Explictly flag more tests
 - ... and 16 more: https://git.openjdk.org/valhalla/compare/2b06f9ed...b6e45f89

-------------

Changes: https://git.openjdk.org/valhalla/pull/727/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=727&range=12
  Stats: 1357 lines in 245 files changed: 836 ins; 57 del; 464 mod
  Patch: https://git.openjdk.org/valhalla/pull/727.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/727/head:pull/727

PR: https://git.openjdk.org/valhalla/pull/727


From dsimms at openjdk.org  Fri Oct  7 07:52:51 2022
From: dsimms at openjdk.org (David Simms)
Date: Fri, 7 Oct 2022 07:52:51 GMT
Subject: [lworld] Integrated: 8288135: [lworld] Implement HotSpot flag
 -XX:+EnablePrimitiveClasses
In-Reply-To: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
References: <gxLOvBEKy5ylQgQd5xqzOCS2M1wy_Q5oQLHAb0_hZnU=.bf1fe095-4b06-4da3-85fb-4c4068765942@github.com>
Message-ID: <_GIgjK2iRtgCtqxoJRQ-yPi78C3S3c1yqXB8iAJlJhc=.9e655c08-c7c2-42fc-84db-888df601e588@github.com>

On Thu, 25 Aug 2022 13:16:40 GMT, David Simms <dsimms at openjdk.org> wrote:

> Runtime component of -XX:+EnablePrimitiveClasses switch

This pull request has now been integrated.

Changeset: 6a27741b
Author:    David Simms <dsimms at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/6a27741b79020009871da2a6065048902685ae31
Stats:     1357 lines in 245 files changed: 836 ins; 57 del; 464 mod

8288135: [lworld] Implement HotSpot flag -XX:+EnablePrimitiveClasses

Reviewed-by: fparain

-------------

PR: https://git.openjdk.org/valhalla/pull/727


From thartmann at openjdk.org  Fri Oct  7 12:37:32 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Fri, 7 Oct 2022 12:37:32 GMT
Subject: [lworld] RFR: 8293542: [lworld] Merge InlineTypeNode and
 InlineTypePtrNode in C2
Message-ID: <AgTR9xtsivUW_arPrMBQZxLUyBXN-LRh0ol8EEjfPAc=.e13a86f5-cdb5-475d-8d05-4d2232afaa05@github.com>

This is a large refactoring, removing the need to have both an `InlineTypeNode` as well as an `InlineTypePtrNode` in C2, significantly simplifying the IR. The change also includes several bug fixes that could not be easily extracted into an independent change.

More cleanups will follow with [JDK-8293800](https://bugs.openjdk.org/browse/JDK-8293800).

Thanks,
Tobias

-------------

Commit messages:
 - Minor fix for flat arrays
 - Merge branch 'lworld' into JDK-8293542
 - Merge branch 'lworld' into JDK-8293542
 - More fixes
 - More cleanups
 - 8293542: [lworld] Merge InlineTypeNode and InlineTypePtrNode in C2

Changes: https://git.openjdk.org/valhalla/pull/780/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=780&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8293542
  Stats: 810 lines in 33 files changed: 177 ins; 348 del; 285 mod
  Patch: https://git.openjdk.org/valhalla/pull/780.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/780/head:pull/780

PR: https://git.openjdk.org/valhalla/pull/780


From thartmann at openjdk.org  Fri Oct  7 13:04:18 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Fri, 7 Oct 2022 13:04:18 GMT
Subject: [lworld] RFR: [lworld] Added missing -XX:+EnableValhalla to
 ClassInitBarrier test
Message-ID: <7erJhPHAsmdWWdG7r5l2m-Li0sNInLGTtN9pIiAEkeM=.e27635d2-979c-49f0-88ab-4b54a98f8ab1@github.com>

Forgot to push this with https://github.com/openjdk/valhalla/pull/778.

Thanks,
Tobias

-------------

Commit messages:
 - [lworld] Added missing -XX:+EnableValhalla to ClassInitBarrier test

Changes: https://git.openjdk.org/valhalla/pull/781/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=781&range=00
  Stats: 14 lines in 1 file changed: 0 ins; 0 del; 14 mod
  Patch: https://git.openjdk.org/valhalla/pull/781.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/781/head:pull/781

PR: https://git.openjdk.org/valhalla/pull/781


From thartmann at openjdk.org  Fri Oct  7 13:07:35 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Fri, 7 Oct 2022 13:07:35 GMT
Subject: [lworld] Integrated: [lworld] Added missing -XX:+EnableValhalla to
 ClassInitBarrier test
In-Reply-To: <7erJhPHAsmdWWdG7r5l2m-Li0sNInLGTtN9pIiAEkeM=.e27635d2-979c-49f0-88ab-4b54a98f8ab1@github.com>
References: <7erJhPHAsmdWWdG7r5l2m-Li0sNInLGTtN9pIiAEkeM=.e27635d2-979c-49f0-88ab-4b54a98f8ab1@github.com>
Message-ID: <b3PxxF5HYI49SeYPKlSy9mvV-0Yv1Z4TR0MTqqyrLIw=.a130fce2-9888-499c-982f-abb558bb912e@github.com>

On Fri, 7 Oct 2022 12:50:49 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

> Forgot to push this with https://github.com/openjdk/valhalla/pull/778.
> 
> Thanks,
> Tobias

This pull request has now been integrated.

Changeset: d2fa68bb
Author:    Tobias Hartmann <thartmann at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/d2fa68bb04bf8747133807e11907e362f0104e76
Stats:     14 lines in 1 file changed: 0 ins; 0 del; 14 mod

[lworld] Added missing -XX:+EnableValhalla to ClassInitBarrier test

-------------

PR: https://git.openjdk.org/valhalla/pull/781


From mchung at openjdk.org  Sat Oct  8 19:20:21 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Sat, 8 Oct 2022 19:20:21 GMT
Subject: [lworld] RFR: 8294927: [lworld] LambdaMetafactory class should reject
 identity and value superinterface
Message-ID: <NkavgvY5a42Ph8XpwWHJNOceaZaJOlfNfQz5YgQcLag=.ee5fbdcd-6ad0-4ab3-80f8-b0f94ea98eae@github.com>

This PR adds the validation in LambdaMetafactory to reject if the return type of the given factory type is not an identity interface and not a value interface.   This also fixes a bug in `Class::isIdentity` to return false for interfaces if Valhalla is not enabled.

-------------

Commit messages:
 - Fix Class::isIdentity to return false for interfaces if Valhalla is disabled
 - Merge branch 'lworld' of https://github.com/openjdk/valhalla into lmf-validate
 - 8294927: [lworld] LambdaMetafactory class should reject identity and value superinterface

Changes: https://git.openjdk.org/valhalla/pull/783/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=783&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8294927
  Stats: 258 lines in 11 files changed: 251 ins; 0 del; 7 mod
  Patch: https://git.openjdk.org/valhalla/pull/783.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/783/head:pull/783

PR: https://git.openjdk.org/valhalla/pull/783


From rriggs at openjdk.org  Sat Oct  8 19:22:02 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Sat, 8 Oct 2022 19:22:02 GMT
Subject: [lworld] RFR: 8293551: [lworld] Remove obsolete interfaces
 __primitive__ and __value__, and NonTearable
Message-ID: <OgJR-Lq_8UZmEEpfa-NKZTGg8zw7POlULwGTm7J-uow=.855d8563-60e5-4d6e-99a8-87b639ba045f@github.com>

The java.lang interfaces __primitive__ and __value__ as well as java.lang.NonTearable are no longer needed for value objects.

The files are deleted and tests updated.

-------------

Commit messages:
 - 8293551: [lworld] Remove obsolete interfaces __primitive__ and __value__, and NonTearable

Changes: https://git.openjdk.org/valhalla/pull/782/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=782&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8293551
  Stats: 152 lines in 6 files changed: 1 ins; 150 del; 1 mod
  Patch: https://git.openjdk.org/valhalla/pull/782.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/782/head:pull/782

PR: https://git.openjdk.org/valhalla/pull/782


From thartmann at openjdk.org  Mon Oct 10 08:58:37 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Mon, 10 Oct 2022 08:58:37 GMT
Subject: [lworld] Integrated: 8293542: [lworld] Merge InlineTypeNode and
 InlineTypePtrNode in C2
In-Reply-To: <AgTR9xtsivUW_arPrMBQZxLUyBXN-LRh0ol8EEjfPAc=.e13a86f5-cdb5-475d-8d05-4d2232afaa05@github.com>
References: <AgTR9xtsivUW_arPrMBQZxLUyBXN-LRh0ol8EEjfPAc=.e13a86f5-cdb5-475d-8d05-4d2232afaa05@github.com>
Message-ID: <GzYGDksvyZx3moHx3oLOySTjSw6U5ZpX9taA73aUQJo=.5f50963b-a9b5-4683-9c0d-eb9685c48ae1@github.com>

On Fri, 7 Oct 2022 12:22:10 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

> This is a large refactoring, removing the need to have both an `InlineTypeNode` as well as an `InlineTypePtrNode` in C2, significantly simplifying the IR. The change also includes several bug fixes that could not be easily extracted into an independent change.
> 
> More cleanups will follow with [JDK-8293800](https://bugs.openjdk.org/browse/JDK-8293800).
> 
> Thanks,
> Tobias

This pull request has now been integrated.

Changeset: 7c569cd2
Author:    Tobias Hartmann <thartmann at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/7c569cd2eaeb6a514638fa2483d867c9f17af69d
Stats:     810 lines in 33 files changed: 177 ins; 348 del; 285 mod

8293542: [lworld] Merge InlineTypeNode and InlineTypePtrNode in C2

-------------

PR: https://git.openjdk.org/valhalla/pull/780


From dsimms at openjdk.org  Mon Oct 10 14:51:41 2022
From: dsimms at openjdk.org (David Simms)
Date: Mon, 10 Oct 2022 14:51:41 GMT
Subject: [lworld] RFR: Merge jdk
Message-ID: <gi7bp3w0LKcwDSBQ4Rmvy_Khdc59Z7JVb44ZSPgI7Yw=.a4e5a28b-e4bc-4e6a-922c-3e93a38a6279@github.com>

Merge tag 'jdk-20+18' into lworld_merge_jdk_20_18
Added tag jdk-20+18 for changeset 0ec18382b

# Conflicts:
#	src/hotspot/share/ci/ciSignature.cpp
#	src/hotspot/share/opto/callnode.cpp
#	src/hotspot/share/opto/callnode.hpp
#	src/hotspot/share/opto/graphKit.cpp
#	src/hotspot/share/opto/memnode.cpp
#	src/hotspot/share/runtime/sharedRuntime.cpp
#	src/java.base/share/classes/jdk/internal/reflect/MethodAccessorGenerator.java

-------------

Commit messages:
 - (JDK-8295066) Revert "8242115: C2 SATB barriers are not safepoint-safe"
 - Trailing whitespace
 - Merge branch 'lworld' into lworld_merge_jdk_20_18
 - Merge tag 'jdk-20+18' into lworld_merge_jdk_20_18
 - 8294869: Correct failure of RemovedJDKInternals.java after JDK-8294618
 - 8294397: Replace StringBuffer with StringBuilder within java.text
 - 8294734: Redundant override in AES implementation
 - 8294618: Update openjdk.java.net => openjdk.org
 - 8294840: langtools OptionalDependencyTest.java use File.pathSeparator
 - 8289925: Shared code shouldn't reference the platform specific method frame::interpreter_frame_last_sp()
 - ... and 97 more: https://git.openjdk.org/valhalla/compare/7c569cd2...36781877

The webrevs contain the adjustments done while merging with regards to each parent branch:
 - lworld: https://webrevs.openjdk.org/?repo=valhalla&pr=784&range=00.0
 - jdk: https://webrevs.openjdk.org/?repo=valhalla&pr=784&range=00.1

Changes: https://git.openjdk.org/valhalla/pull/784/files
  Stats: 12855 lines in 388 files changed: 7926 ins; 3377 del; 1552 mod
  Patch: https://git.openjdk.org/valhalla/pull/784.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/784/head:pull/784

PR: https://git.openjdk.org/valhalla/pull/784


From dsimms at openjdk.org  Mon Oct 10 14:53:26 2022
From: dsimms at openjdk.org (David Simms)
Date: Mon, 10 Oct 2022 14:53:26 GMT
Subject: [lworld] Integrated: Merge jdk
In-Reply-To: <gi7bp3w0LKcwDSBQ4Rmvy_Khdc59Z7JVb44ZSPgI7Yw=.a4e5a28b-e4bc-4e6a-922c-3e93a38a6279@github.com>
References: <gi7bp3w0LKcwDSBQ4Rmvy_Khdc59Z7JVb44ZSPgI7Yw=.a4e5a28b-e4bc-4e6a-922c-3e93a38a6279@github.com>
Message-ID: <ed4UMwabwFhjdt-tL-4JaXB7BsmcwDTF_woG4BAy6cQ=.51280831-eadc-43df-8a59-540ad68bb09e@github.com>

On Mon, 10 Oct 2022 09:34:36 GMT, David Simms <dsimms at openjdk.org> wrote:

> Merge tag 'jdk-20+18' into lworld_merge_jdk_20_18
> Added tag jdk-20+18 for changeset 0ec18382b
> 
> # Conflicts:
> #	src/hotspot/share/ci/ciSignature.cpp
> #	src/hotspot/share/opto/callnode.cpp
> #	src/hotspot/share/opto/callnode.hpp
> #	src/hotspot/share/opto/graphKit.cpp
> #	src/hotspot/share/opto/memnode.cpp
> #	src/hotspot/share/runtime/sharedRuntime.cpp
> #	src/java.base/share/classes/jdk/internal/reflect/MethodAccessorGenerator.java

This pull request has now been integrated.

Changeset: 67bae7e7
Author:    David Simms <dsimms at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/67bae7e716ac674ee859f3c97e12d8ad102a8267
Stats:     12855 lines in 388 files changed: 7926 ins; 3377 del; 1552 mod

Merge jdk

Merge tag 'jdk-20+18'

-------------

PR: https://git.openjdk.org/valhalla/pull/784


From mchung at openjdk.org  Mon Oct 10 16:24:59 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Mon, 10 Oct 2022 16:24:59 GMT
Subject: [lworld] RFR: 8293551: [lworld] Remove obsolete interfaces
 __primitive__ and __value__, and NonTearable
In-Reply-To: <OgJR-Lq_8UZmEEpfa-NKZTGg8zw7POlULwGTm7J-uow=.855d8563-60e5-4d6e-99a8-87b639ba045f@github.com>
References: <OgJR-Lq_8UZmEEpfa-NKZTGg8zw7POlULwGTm7J-uow=.855d8563-60e5-4d6e-99a8-87b639ba045f@github.com>
Message-ID: <Hi1YOUFf16w8Ma4i5suwuq5XMLX6Sg7RAKxx4CNjwKU=.8ce1b1f4-787d-426c-9909-d42f6711253b@github.com>

On Fri, 7 Oct 2022 21:05:35 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

> The java.lang interfaces __primitive__ and __value__ as well as java.lang.NonTearable are no longer needed for value objects.
> 
> The files are deleted and tests updated.

Marked as reviewed by mchung (Committer).

-------------

PR: https://git.openjdk.org/valhalla/pull/782


From rriggs at openjdk.org  Mon Oct 10 22:53:56 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Mon, 10 Oct 2022 22:53:56 GMT
Subject: [lworld] RFR: 8270852: [lworld] bytecodes spinned at runtime
 should set minor_version to 65535 if preview feature enabled
In-Reply-To: <A-a5kqNeJVz2HgnfPWj07sFU3uA07udSDqqzXtF7UQY=.5131573c-4100-43f5-8901-fd62f016610c@github.com>
References: <A-a5kqNeJVz2HgnfPWj07sFU3uA07udSDqqzXtF7UQY=.5131573c-4100-43f5-8901-fd62f016610c@github.com>
Message-ID: <S524xc2ea4Z5n-XFIYwBWuZj8_VlkoSTnO08Y-l59ts=.3cf0020b-400a-4492-a085-35b9b27b9f5b@github.com>

On Fri, 7 Oct 2022 00:56:39 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> This patch prepares for lworld to switch to a preview feature.  Also fix the generation of lambda classes and proxy classes to have `Preload` attribute for value classes (not only primitive classes).

Marked as reviewed by rriggs (Committer).

src/java.base/share/classes/java/lang/invoke/InnerClassLambdaMetafactory.java line 645:

> 643:                 c = c.getComponentType();
> 644:             }
> 645:             return (c.isValue() && !PrimitiveClass.isPrimitiveClass(c)) || PrimitiveClass.isPrimitiveValueType(c);

Perhaps there should be a method that computes this compound expression?

-------------

PR: https://git.openjdk.org/valhalla/pull/779


From mchung at openjdk.org  Mon Oct 10 22:53:57 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Mon, 10 Oct 2022 22:53:57 GMT
Subject: [lworld] RFR: 8270852: [lworld] bytecodes spinned at runtime
 should set minor_version to 65535 if preview feature enabled
In-Reply-To: <S524xc2ea4Z5n-XFIYwBWuZj8_VlkoSTnO08Y-l59ts=.3cf0020b-400a-4492-a085-35b9b27b9f5b@github.com>
References: <A-a5kqNeJVz2HgnfPWj07sFU3uA07udSDqqzXtF7UQY=.5131573c-4100-43f5-8901-fd62f016610c@github.com>
 <S524xc2ea4Z5n-XFIYwBWuZj8_VlkoSTnO08Y-l59ts=.3cf0020b-400a-4492-a085-35b9b27b9f5b@github.com>
Message-ID: <25KwFhzugXTzIEgpdNVVZaHJbE-R862ndOeZmCXn0Ig=.58731bcf-978a-444b-af80-89bc457853f2@github.com>

On Mon, 10 Oct 2022 22:30:57 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

>> This patch prepares for lworld to switch to a preview feature.  Also fix the generation of lambda classes and proxy classes to have `Preload` attribute for value classes (not only primitive classes).
>
> src/java.base/share/classes/java/lang/invoke/InnerClassLambdaMetafactory.java line 645:
> 
>> 643:                 c = c.getComponentType();
>> 644:             }
>> 645:             return (c.isValue() && !PrimitiveClass.isPrimitiveClass(c)) || PrimitiveClass.isPrimitiveValueType(c);
> 
> Perhaps there should be a method that computes this compound expression?

This compound expression should not be needed.  Preload attribute is generated for a value class including primitive class.  It's a leftover when we added support for value objects.

-------------

PR: https://git.openjdk.org/valhalla/pull/779


From rriggs at openjdk.org  Mon Oct 10 22:54:06 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Mon, 10 Oct 2022 22:54:06 GMT
Subject: [lworld] RFR: 8294927: [lworld] LambdaMetafactory class should
 reject identity and value superinterface
In-Reply-To: <NkavgvY5a42Ph8XpwWHJNOceaZaJOlfNfQz5YgQcLag=.ee5fbdcd-6ad0-4ab3-80f8-b0f94ea98eae@github.com>
References: <NkavgvY5a42Ph8XpwWHJNOceaZaJOlfNfQz5YgQcLag=.ee5fbdcd-6ad0-4ab3-80f8-b0f94ea98eae@github.com>
Message-ID: <RQgNrHstjmEmHBcM_w3tL6T-eLENHxR10wx6JbPL4GY=.4ecf5dee-9174-4386-96a4-d43ec9b8de53@github.com>

On Sat, 8 Oct 2022 02:12:08 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> This PR adds the validation in LambdaMetafactory to reject if the return type of the given factory type is not an identity interface and not a value interface.   This also fixes a bug in `Class::isIdentity` to return false for interfaces if Valhalla is not enabled.

Marked as reviewed by rriggs (Committer).

-------------

PR: https://git.openjdk.org/valhalla/pull/783


From mchung at openjdk.org  Mon Oct 10 23:37:41 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Mon, 10 Oct 2022 23:37:41 GMT
Subject: [lworld] Integrated: 8294927: [lworld] LambdaMetafactory class should
 reject identity and value superinterface
In-Reply-To: <NkavgvY5a42Ph8XpwWHJNOceaZaJOlfNfQz5YgQcLag=.ee5fbdcd-6ad0-4ab3-80f8-b0f94ea98eae@github.com>
References: <NkavgvY5a42Ph8XpwWHJNOceaZaJOlfNfQz5YgQcLag=.ee5fbdcd-6ad0-4ab3-80f8-b0f94ea98eae@github.com>
Message-ID: <mUe8GX2ppugyDBaV2e9dHMgW1gYIOiTOCYtp9ZBp7jY=.d31e1298-409f-440d-9dda-c4fefdf1dcac@github.com>

On Sat, 8 Oct 2022 02:12:08 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> This PR adds the validation in LambdaMetafactory to reject if the return type of the given factory type is not an identity interface and not a value interface.   This also fixes a bug in `Class::isIdentity` to return false for interfaces if Valhalla is not enabled.

This pull request has now been integrated.

Changeset: 14c71a4d
Author:    Mandy Chung <mchung at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/14c71a4d6f26b41c641799e41685854755d9964e
Stats:     258 lines in 11 files changed: 251 ins; 0 del; 7 mod

8294927: [lworld] LambdaMetafactory class should reject identity and value superinterface

Reviewed-by: rriggs

-------------

PR: https://git.openjdk.org/valhalla/pull/783


From rriggs at openjdk.org  Tue Oct 11 14:59:39 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Tue, 11 Oct 2022 14:59:39 GMT
Subject: [lworld] Integrated: 8293551: [lworld] Remove obsolete interfaces
 __primitive__ and __value__, and NonTearable
In-Reply-To: <OgJR-Lq_8UZmEEpfa-NKZTGg8zw7POlULwGTm7J-uow=.855d8563-60e5-4d6e-99a8-87b639ba045f@github.com>
References: <OgJR-Lq_8UZmEEpfa-NKZTGg8zw7POlULwGTm7J-uow=.855d8563-60e5-4d6e-99a8-87b639ba045f@github.com>
Message-ID: <wJbD3abXtz6jLivWuLQU2TRMH-aY-VKb_AtJOcAY85w=.b965680b-7eb9-4137-bd04-07a1d61f1f77@github.com>

On Fri, 7 Oct 2022 21:05:35 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

> The java.lang interfaces __primitive__ and __value__ as well as java.lang.NonTearable are no longer needed for value objects.
> 
> The files are deleted and tests updated.

This pull request has now been integrated.

Changeset: 8b2d9896
Author:    Roger Riggs <rriggs at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/8b2d9896b60c18070e37b471f2c35c20c63987c6
Stats:     152 lines in 6 files changed: 1 ins; 150 del; 1 mod

8293551: [lworld] Remove obsolete interfaces __primitive__ and __value__, and NonTearable

Reviewed-by: mchung

-------------

PR: https://git.openjdk.org/valhalla/pull/782


From mchung at openjdk.org  Tue Oct 11 16:30:20 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Tue, 11 Oct 2022 16:30:20 GMT
Subject: [lworld] Integrated: 8270852: [lworld] bytecodes spinned at runtime
 should set minor_version to 65535 if preview feature enabled
In-Reply-To: <A-a5kqNeJVz2HgnfPWj07sFU3uA07udSDqqzXtF7UQY=.5131573c-4100-43f5-8901-fd62f016610c@github.com>
References: <A-a5kqNeJVz2HgnfPWj07sFU3uA07udSDqqzXtF7UQY=.5131573c-4100-43f5-8901-fd62f016610c@github.com>
Message-ID: <yrBGtnrlkZXhnWlb1Zq8KQjdGvl_6AoxvoKrP48ZMhc=.ff63c8f5-02d0-492a-84ad-a7abfd21fe47@github.com>

On Fri, 7 Oct 2022 00:56:39 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> This patch prepares for lworld to switch to a preview feature.  Also fix the generation of lambda classes and proxy classes to have `Preload` attribute for value classes (not only primitive classes).

This pull request has now been integrated.

Changeset: cf05d5c1
Author:    Mandy Chung <mchung at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/cf05d5c19640daaf55beee84334784325a897b7c
Stats:     17 lines in 2 files changed: 9 ins; 3 del; 5 mod

8270852: [lworld] bytecodes spinned at runtime should set minor_version to 65535 if preview feature enabled

Reviewed-by: rriggs

-------------

PR: https://git.openjdk.org/valhalla/pull/779


From rriggs at openjdk.org  Tue Oct 11 19:06:41 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Tue, 11 Oct 2022 19:06:41 GMT
Subject: [lworld] RFR: 8294860: [lworld] Add java.util.Objects.isValueObject
Message-ID: <zU5FCYVUaRAuB8jT0G2kbFTsqN6xkC0drXEMb75u91s=.88ce20d7-aee9-4b5e-9250-8aba9ac8842c@github.com>

Add method `java.util.Objects.isValueObject(Object obj)`.

The complement to `Objects.isIdentityObject(Object obj)` this method return true if the object is a value object.

Add the cases to the existing ObjectMethods test.

-------------

Commit messages:
 - 8294860: [lworld] Add java.util.Objects.isValueObject

Changes: https://git.openjdk.org/valhalla/pull/786/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=786&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8294860
  Stats: 15 lines in 2 files changed: 15 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/valhalla/pull/786.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/786/head:pull/786

PR: https://git.openjdk.org/valhalla/pull/786


From mchung at openjdk.org  Tue Oct 11 22:33:37 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Tue, 11 Oct 2022 22:33:37 GMT
Subject: [lworld] RFR: 8280979: [lworld] Rename PrimitiveObjectMethods to
 ValueObjectMethods
Message-ID: <1nByP6vfIscIfRj5db86poJie9ZKYbdxo56_JGEmoM4=.db191e6a-1a02-4e7d-956d-37e5b1058d89@github.com>

This PR does the renaming to match the Value Objects JEP.

-------------

Commit messages:
 - 8280979: Rename PrimitiveObjectMethods to ValueObjectMethods

Changes: https://git.openjdk.org/valhalla/pull/787/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=787&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8280979
  Stats: 999 lines in 14 files changed: 494 ins; 479 del; 26 mod
  Patch: https://git.openjdk.org/valhalla/pull/787.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/787/head:pull/787

PR: https://git.openjdk.org/valhalla/pull/787


From mchung at openjdk.org  Tue Oct 11 22:36:16 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Tue, 11 Oct 2022 22:36:16 GMT
Subject: [lworld] RFR: 8280979: [lworld] Rename PrimitiveObjectMethods to
 ValueObjectMethods [v2]
In-Reply-To: <1nByP6vfIscIfRj5db86poJie9ZKYbdxo56_JGEmoM4=.db191e6a-1a02-4e7d-956d-37e5b1058d89@github.com>
References: <1nByP6vfIscIfRj5db86poJie9ZKYbdxo56_JGEmoM4=.db191e6a-1a02-4e7d-956d-37e5b1058d89@github.com>
Message-ID: <WtoHXIY6zQkl2-Zcq-I5WJilosg9LWTxinEXWQYZtHA=.39ceb9a7-0b94-46f7-9756-5025d6c7b77d@github.com>

> This PR does the renaming to match the Value Objects JEP.

Mandy Chung has updated the pull request incrementally with two additional commits since the last revision:

 - fix typo
 - fix typo

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/787/files
  - new: https://git.openjdk.org/valhalla/pull/787/files/540c4ec4..1d6ab603

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=787&range=01
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=787&range=00-01

  Stats: 4 lines in 1 file changed: 0 ins; 0 del; 4 mod
  Patch: https://git.openjdk.org/valhalla/pull/787.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/787/head:pull/787

PR: https://git.openjdk.org/valhalla/pull/787


From mchung at openjdk.org  Tue Oct 11 22:57:31 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Tue, 11 Oct 2022 22:57:31 GMT
Subject: [lworld] RFR: 8294860: [lworld] Add
 java.util.Objects.isValueObject
In-Reply-To: <zU5FCYVUaRAuB8jT0G2kbFTsqN6xkC0drXEMb75u91s=.88ce20d7-aee9-4b5e-9250-8aba9ac8842c@github.com>
References: <zU5FCYVUaRAuB8jT0G2kbFTsqN6xkC0drXEMb75u91s=.88ce20d7-aee9-4b5e-9250-8aba9ac8842c@github.com>
Message-ID: <WXyeMEP4CWpYju587u5MATEPmC5FIQxeu5D_MgSdRr8=.cf6eebc5-1339-49e3-b40b-a91e590c9912@github.com>

On Tue, 11 Oct 2022 18:51:47 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

> Add method `java.util.Objects.isValueObject(Object obj)`.
> 
> The complement to `Objects.isIdentityObject(Object obj)` this method return true if the object is a value object.
> 
> Add the cases to the existing ObjectMethods test.

Marked as reviewed by mchung (Committer).

src/java.base/share/classes/java/util/Objects.java line 267:

> 265: 
> 266:    /**
> 267:     * {@return {@code true} if the specified object reference is a value object,

Maybe add a link to `{@linkplain Class#isValue value} object`.

"object reference" might cause confusion and consider as pass by reference.    The `java.util.Objects` APIs are inconsistent and some uses "object reference" and some don't.

-------------

PR: https://git.openjdk.org/valhalla/pull/786


From rriggs at openjdk.org  Wed Oct 12 00:26:40 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Wed, 12 Oct 2022 00:26:40 GMT
Subject: [lworld] RFR: 8294860: [lworld] Add
 java.util.Objects.isValueObject [v2]
In-Reply-To: <zU5FCYVUaRAuB8jT0G2kbFTsqN6xkC0drXEMb75u91s=.88ce20d7-aee9-4b5e-9250-8aba9ac8842c@github.com>
References: <zU5FCYVUaRAuB8jT0G2kbFTsqN6xkC0drXEMb75u91s=.88ce20d7-aee9-4b5e-9250-8aba9ac8842c@github.com>
Message-ID: <mpcCfu4altT7V-XaXvbe-5fM9jnlrLcDBp0CEb14HQs=.57ef14e5-5265-4de9-a20c-5c8fbaa27f06@github.com>

> Add method `java.util.Objects.isValueObject(Object obj)`.
> 
> The complement to `Objects.isIdentityObject(Object obj)` this method return true if the object is a value object.
> 
> Add the cases to the existing ObjectMethods test.

Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:

  add link in Objects.isValueObject()

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/786/files
  - new: https://git.openjdk.org/valhalla/pull/786/files/80c6914b..163bca19

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=786&range=01
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=786&range=00-01

  Stats: 1 line in 1 file changed: 0 ins; 0 del; 1 mod
  Patch: https://git.openjdk.org/valhalla/pull/786.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/786/head:pull/786

PR: https://git.openjdk.org/valhalla/pull/786


From dsimms at openjdk.org  Wed Oct 12 08:42:22 2022
From: dsimms at openjdk.org (David Simms)
Date: Wed, 12 Oct 2022 08:42:22 GMT
Subject: [lworld] Withdrawn: 8247569: [lworld] Align with spec for <vnew>
 methods
In-Reply-To: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
References: <pw0pGmlQjRwxctYc8UeN9q9olicRtVQrMJCd3d4DlH4=.ec406270-04f4-4cfc-834c-fa8790fc221f@github.com>
Message-ID: <lts3KvClDBFZy6B5GBVrgNMtOhhLVeamuQ6Dwqd9Pd0=.caee9ed9-09b6-4950-9c10-dec04fc34b6b@github.com>

On Wed, 28 Sep 2022 17:52:34 GMT, Vicente Romero <vromero at openjdk.org> wrote:

> Value classes constructor should be named <vnew> instead of <init>. This PR can't be pushed without the corresponding support being available on the VM side

This pull request has been closed without being integrated.

-------------

PR: https://git.openjdk.org/valhalla/pull/771


From thartmann at openjdk.org  Wed Oct 12 12:51:08 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Wed, 12 Oct 2022 12:51:08 GMT
Subject: [lworld] RFR: 8295152: [lworld] Various C2 compiler issues revealed by
 stress testing
Message-ID: <d8w4GAFr8FaPkee-mGhnLp9EaEs3pvMVl3d67vrIGxQ=.4cbade87-a60d-4005-9624-4d224d9e7484@github.com>

C2 stress testing via aggressive combinations of flags revealed several issues that are fixed by this patch. Expect for one extremely intermittent issue that I'm going to investigate next, the entire (internal) stress job definition now finally passes!

Thanks,
Tobias

-------------

Commit messages:
 - 8295152: [lworld] Various C2 compiler issues revealed by stress testing

Changes: https://git.openjdk.org/valhalla/pull/789/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=789&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8295152
  Stats: 110 lines in 14 files changed: 54 ins; 16 del; 40 mod
  Patch: https://git.openjdk.org/valhalla/pull/789.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/789/head:pull/789

PR: https://git.openjdk.org/valhalla/pull/789


From thartmann at openjdk.org  Wed Oct 12 12:57:27 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Wed, 12 Oct 2022 12:57:27 GMT
Subject: [lworld] Integrated: 8295152: [lworld] Various C2 compiler issues
 revealed by stress testing
In-Reply-To: <d8w4GAFr8FaPkee-mGhnLp9EaEs3pvMVl3d67vrIGxQ=.4cbade87-a60d-4005-9624-4d224d9e7484@github.com>
References: <d8w4GAFr8FaPkee-mGhnLp9EaEs3pvMVl3d67vrIGxQ=.4cbade87-a60d-4005-9624-4d224d9e7484@github.com>
Message-ID: <w7DNA_B3cyyO-U5JDH_EesJZncnH1Cei7tLm3AXnwrE=.4dfb3c0b-5cee-4747-ac03-83581ec4b067@github.com>

On Wed, 12 Oct 2022 12:45:44 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

> C2 stress testing via aggressive combinations of flags revealed several issues that are fixed by this patch. Expect for one extremely intermittent issue that I'm going to investigate next, the entire (internal) stress job definition now finally passes!
> 
> Thanks,
> Tobias

This pull request has now been integrated.

Changeset: 82b63b75
Author:    Tobias Hartmann <thartmann at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/82b63b75a6ca68a4948fcbd0672d3b777b18d255
Stats:     110 lines in 14 files changed: 54 ins; 16 del; 40 mod

8295152: [lworld] Various C2 compiler issues revealed by stress testing

-------------

PR: https://git.openjdk.org/valhalla/pull/789


From jbhateja at openjdk.org  Wed Oct 12 13:20:43 2022
From: jbhateja at openjdk.org (Jatin Bhateja)
Date: Wed, 12 Oct 2022 13:20:43 GMT
Subject: [lworld+vector] RFR: [lworld+vector] Borrowing MultiFields initial
 implementation from https://github.com/fparain/valhalla/tree/multifields.
Message-ID: <z5TkDbU8HyreHkX6OQvGmXjtSXTZgkZpHYO-cHITLsI=.18584e2e-a06a-48b9-b1da-bcc349cff0a0@github.com>

Merging initial Multifields support added by @fparain from his local repository https://github.com/fparain/valhalla/tree/multifields.

Best Regards,
Jatin

-------------

Commit messages:
 - Borrowing MultiFields initial implementation from https://github.com/fparain/valhalla/tree/multifields.

Changes: https://git.openjdk.org/valhalla/pull/790/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=790&range=00
  Stats: 823 lines in 19 files changed: 615 ins; 85 del; 123 mod
  Patch: https://git.openjdk.org/valhalla/pull/790.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/790/head:pull/790

PR: https://git.openjdk.org/valhalla/pull/790


From jbhateja at openjdk.org  Wed Oct 12 13:27:49 2022
From: jbhateja at openjdk.org (Jatin Bhateja)
Date: Wed, 12 Oct 2022 13:27:49 GMT
Subject: [lworld+vector] RFR: [lworld+vector] Borrowing MultiFields initial
 implementation from https://github.com/fparain/valhalla/tree/multifields. [v2]
In-Reply-To: <z5TkDbU8HyreHkX6OQvGmXjtSXTZgkZpHYO-cHITLsI=.18584e2e-a06a-48b9-b1da-bcc349cff0a0@github.com>
References: <z5TkDbU8HyreHkX6OQvGmXjtSXTZgkZpHYO-cHITLsI=.18584e2e-a06a-48b9-b1da-bcc349cff0a0@github.com>
Message-ID: <zbd9krpP5D44xK7LvcNK0Q6nKqrrUkeasGttXYbdZmQ=.27ec7b30-57b2-46ee-b83a-f3a710b63c7d@github.com>

> Merging initial Multifields support added by @fparain from his local repository https://github.com/fparain/valhalla/tree/multifields.
> 
> Best Regards,
> Jatin

Jatin Bhateja has refreshed the contents of this pull request, and previous commits have been removed. Incremental views are not available.

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/790/files
  - new: https://git.openjdk.org/valhalla/pull/790/files/c3a6108e..9039a70b

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=790&range=01
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=790&range=00-01

  Stats: 791 lines in 19 files changed: 53 ins; 583 del; 155 mod
  Patch: https://git.openjdk.org/valhalla/pull/790.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/790/head:pull/790

PR: https://git.openjdk.org/valhalla/pull/790


From jbhateja at openjdk.org  Wed Oct 12 13:27:51 2022
From: jbhateja at openjdk.org (Jatin Bhateja)
Date: Wed, 12 Oct 2022 13:27:51 GMT
Subject: [lworld+vector] Withdrawn: [lworld+vector] Borrowing MultiFields
 initial implementation from
 https://github.com/fparain/valhalla/tree/multifields.
In-Reply-To: <z5TkDbU8HyreHkX6OQvGmXjtSXTZgkZpHYO-cHITLsI=.18584e2e-a06a-48b9-b1da-bcc349cff0a0@github.com>
References: <z5TkDbU8HyreHkX6OQvGmXjtSXTZgkZpHYO-cHITLsI=.18584e2e-a06a-48b9-b1da-bcc349cff0a0@github.com>
Message-ID: <czOo2dVp8tOMK1u_lFECiFgaxQb8BNfXQlHBbX25iMc=.ba4435c8-5f72-4898-b9cc-95fa843a9d36@github.com>

On Wed, 12 Oct 2022 13:13:41 GMT, Jatin Bhateja <jbhateja at openjdk.org> wrote:

> Merging initial Multifields support added by @fparain from his local repository https://github.com/fparain/valhalla/tree/multifields.
> 
> Best Regards,
> Jatin

This pull request has been closed without being integrated.

-------------

PR: https://git.openjdk.org/valhalla/pull/790


From rriggs at openjdk.org  Wed Oct 12 13:54:23 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Wed, 12 Oct 2022 13:54:23 GMT
Subject: [lworld] Integrated: 8294860: [lworld] Add
 java.util.Objects.isValueObject
In-Reply-To: <zU5FCYVUaRAuB8jT0G2kbFTsqN6xkC0drXEMb75u91s=.88ce20d7-aee9-4b5e-9250-8aba9ac8842c@github.com>
References: <zU5FCYVUaRAuB8jT0G2kbFTsqN6xkC0drXEMb75u91s=.88ce20d7-aee9-4b5e-9250-8aba9ac8842c@github.com>
Message-ID: <I74JKDys8SstoUgL_-xsz4FvWuj534DAUd0COYehEHg=.c0b1412d-698c-4b1a-a4b6-8efdf029c7f3@github.com>

On Tue, 11 Oct 2022 18:51:47 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

> Add method `java.util.Objects.isValueObject(Object obj)`.
> 
> The complement to `Objects.isIdentityObject(Object obj)` this method return true if the object is a value object.
> 
> Add the cases to the existing ObjectMethods test.

This pull request has now been integrated.

Changeset: e1c3d839
Author:    Roger Riggs <rriggs at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/e1c3d839c7a9c81bacda90944e3517d594592987
Stats:     15 lines in 2 files changed: 15 ins; 0 del; 0 mod

8294860: [lworld] Add java.util.Objects.isValueObject

Reviewed-by: mchung

-------------

PR: https://git.openjdk.org/valhalla/pull/786


From jbhateja at openjdk.org  Wed Oct 12 14:08:49 2022
From: jbhateja at openjdk.org (Jatin Bhateja)
Date: Wed, 12 Oct 2022 14:08:49 GMT
Subject: [lworld+vector] RFR: [lworld+vector] Borrowing MultiFields initial
 implementation from https://github.com/fparain/valhalla/tree/multifields.
Message-ID: <Wg7b-jzV6GGdhxvkJd_Dcp__GPnzNnGZfefa32ZzAx4=.a1f97bda-1a63-4bd3-b629-91d57c61afbe@github.com>

Merging initial Multifields support added by @fparain in his local repository https://github.com/fparain/valhalla/tree/multifields.

Best Regards,
Jatin

-------------

Commit messages:
 - [lworld+vector] Borrowing MultiFields initial implementation from https://github.com/fparain/valhalla/tree/multifields.

Changes: https://git.openjdk.org/valhalla/pull/792/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=792&range=00
  Stats: 823 lines in 19 files changed: 615 ins; 85 del; 123 mod
  Patch: https://git.openjdk.org/valhalla/pull/792.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/792/head:pull/792

PR: https://git.openjdk.org/valhalla/pull/792


From jbhateja at openjdk.org  Wed Oct 12 15:11:24 2022
From: jbhateja at openjdk.org (Jatin Bhateja)
Date: Wed, 12 Oct 2022 15:11:24 GMT
Subject: [lworld+vector] Integrated: [lworld+vector] Borrowing MultiFields
 initial implementation from
 https://github.com/fparain/valhalla/tree/multifields.
In-Reply-To: <Wg7b-jzV6GGdhxvkJd_Dcp__GPnzNnGZfefa32ZzAx4=.a1f97bda-1a63-4bd3-b629-91d57c61afbe@github.com>
References: <Wg7b-jzV6GGdhxvkJd_Dcp__GPnzNnGZfefa32ZzAx4=.a1f97bda-1a63-4bd3-b629-91d57c61afbe@github.com>
Message-ID: <wUlt1MOodFbuzJPcZf19HSGzA63gOGfEsNlXf4aiEsk=.dbaeaed4-689a-4aa5-b34e-abf7d74038a1@github.com>

On Wed, 12 Oct 2022 14:00:29 GMT, Jatin Bhateja <jbhateja at openjdk.org> wrote:

> Merging initial Multifields support added by @fparain in his local repository https://github.com/fparain/valhalla/tree/multifields.
> 
> Best Regards,
> Jatin

This pull request has now been integrated.

Changeset: ff1901b9
Author:    Jatin Bhateja <jbhateja at openjdk.org>
Committer: Frederic Parain <fparain at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/ff1901b93c3f58de79c04f770b58c83b7c5b0019
Stats:     823 lines in 19 files changed: 615 ins; 85 del; 123 mod

[lworld+vector] Borrowing MultiFields initial implementation from https://github.com/fparain/valhalla/tree/multifields.

Co-authored-by: Frederic Parain <fparain at openjdk.org>

-------------

PR: https://git.openjdk.org/valhalla/pull/792


From mchung at openjdk.org  Wed Oct 12 18:11:32 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Wed, 12 Oct 2022 18:11:32 GMT
Subject: [lworld] RFR: 8280979: [lworld] Rename PrimitiveObjectMethods to
 ValueObjectMethods [v3]
In-Reply-To: <1nByP6vfIscIfRj5db86poJie9ZKYbdxo56_JGEmoM4=.db191e6a-1a02-4e7d-956d-37e5b1058d89@github.com>
References: <1nByP6vfIscIfRj5db86poJie9ZKYbdxo56_JGEmoM4=.db191e6a-1a02-4e7d-956d-37e5b1058d89@github.com>
Message-ID: <8PyAh-pORY6fgQMgePBYNQeSsQ7nxsaUWNOx2VmLCWQ=.09cfcc3f-ea6c-4cc0-8221-7aed2c8b8bf4@github.com>

> This PR does the renaming to match the Value Objects JEP.

Mandy Chung has updated the pull request incrementally with one additional commit since the last revision:

  clean up references to value type

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/787/files
  - new: https://git.openjdk.org/valhalla/pull/787/files/1d6ab603..319b0069

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=787&range=02
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=787&range=01-02

  Stats: 31 lines in 1 file changed: 7 ins; 10 del; 14 mod
  Patch: https://git.openjdk.org/valhalla/pull/787.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/787/head:pull/787

PR: https://git.openjdk.org/valhalla/pull/787


From mchung at openjdk.org  Wed Oct 12 18:17:48 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Wed, 12 Oct 2022 18:17:48 GMT
Subject: [lworld] RFR: 8280979: [lworld] Rename PrimitiveObjectMethods to
 ValueObjectMethods [v4]
In-Reply-To: <1nByP6vfIscIfRj5db86poJie9ZKYbdxo56_JGEmoM4=.db191e6a-1a02-4e7d-956d-37e5b1058d89@github.com>
References: <1nByP6vfIscIfRj5db86poJie9ZKYbdxo56_JGEmoM4=.db191e6a-1a02-4e7d-956d-37e5b1058d89@github.com>
Message-ID: <3CfE4BJeRYiOkda6hHncATO_6OESG60rBG1LRtumLqk=.8047ad01-e794-4182-a2e3-3f4919e0bb11@github.com>

> This PR does the renaming to match the Value Objects JEP.

Mandy Chung has updated the pull request incrementally with one additional commit since the last revision:

  update for built-in primitive type

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/787/files
  - new: https://git.openjdk.org/valhalla/pull/787/files/319b0069..8d473317

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=787&range=03
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=787&range=02-03

  Stats: 3 lines in 1 file changed: 0 ins; 0 del; 3 mod
  Patch: https://git.openjdk.org/valhalla/pull/787.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/787/head:pull/787

PR: https://git.openjdk.org/valhalla/pull/787


From vromero at openjdk.org  Wed Oct 12 19:43:54 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Wed, 12 Oct 2022 19:43:54 GMT
Subject: [lworld] RFR: 8295086: [lworld] follow-up for JDK-8288131
Message-ID: <uwWHqZsU469Jj2mgICVf3XvtCROEzeL31ZM30N_rknA=.29be9b99-caeb-4d8d-8409-f83706e1f04a@github.com>

code related to primitive classes needs to be hidden behind the -XDenablePrimitiveClasses flag, this work is a follow-up of JDK-8288131. Hiding this remaining code behind the flag is not absolutely necessary as it probably won't be executed anyways but it will make easier to located, and extract if needed, the code related with primitive classes only

-------------

Commit messages:
 - Merge branch 'lworld' into JDK-8295086
 - 8295086: [lworld] follow-up for JDK-8288131

Changes: https://git.openjdk.org/valhalla/pull/793/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=793&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8295086
  Stats: 9 lines in 2 files changed: 5 ins; 0 del; 4 mod
  Patch: https://git.openjdk.org/valhalla/pull/793.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/793/head:pull/793

PR: https://git.openjdk.org/valhalla/pull/793


From rriggs at openjdk.org  Wed Oct 12 21:31:36 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Wed, 12 Oct 2022 21:31:36 GMT
Subject: [lworld] RFR: 8280979: [lworld] Rename PrimitiveObjectMethods to
 ValueObjectMethods [v4]
In-Reply-To: <3CfE4BJeRYiOkda6hHncATO_6OESG60rBG1LRtumLqk=.8047ad01-e794-4182-a2e3-3f4919e0bb11@github.com>
References: <1nByP6vfIscIfRj5db86poJie9ZKYbdxo56_JGEmoM4=.db191e6a-1a02-4e7d-956d-37e5b1058d89@github.com>
 <3CfE4BJeRYiOkda6hHncATO_6OESG60rBG1LRtumLqk=.8047ad01-e794-4182-a2e3-3f4919e0bb11@github.com>
Message-ID: <1y_VamSEU8BOxJ7CYDcAmy0ueujOjt0sgXav898q4Cc=.abbfc7e5-9aa8-4f74-b033-2625e25ecc6f@github.com>

On Wed, 12 Oct 2022 18:17:48 GMT, Mandy Chung <mchung at openjdk.org> wrote:

>> This PR does the renaming to match the Value Objects JEP.
>
> Mandy Chung has updated the pull request incrementally with one additional commit since the last revision:
> 
>   update for built-in primitive type

LGTM

-------------

Marked as reviewed by rriggs (Committer).

PR: https://git.openjdk.org/valhalla/pull/787


From rriggs at openjdk.org  Wed Oct 12 22:03:36 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Wed, 12 Oct 2022 22:03:36 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v3]
In-Reply-To: <s1GE2wyyNF6lZnfweudY0xAb4RIamFdqyREZYyaxJ9k=.0cd4ebfc-df1c-46e2-8fe1-fcf295551120@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
 <s1GE2wyyNF6lZnfweudY0xAb4RIamFdqyREZYyaxJ9k=.0cd4ebfc-df1c-46e2-8fe1-fcf295551120@github.com>
Message-ID: <x3Xu7DjIdCSZyoRXIDwU-Cl9Q2j-Eoqzh9sKMjpVDyU=.7c021a7a-c6d5-44c6-8a20-e42df583e625@github.com>

On Thu, 29 Sep 2022 19:45:46 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

>> Specify serialization and deserialization of value objects. Value objects are immutable, Externalizable relies on mutation, Externalizable is not supported for value objects. Revised tests and added tests for graphs of value objects.
>> 
>> Somewhat similar to Record serialization in that back references to the value object store 'null' until the value object is completely deserialized.
>> Similar to regular serialization/deserialization of identity objects.  The serialized form is the same.
>> Deserialization is done by reading fields and storing into a Buffered larval object, converted to a value object when all of the fields have been read and stored.
>> 
>> Externalizable and Value Objects are incompatible, value objects are immutable and Externalizable.readExternal is expected to modify the fields of an allocated object.
>
> Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Remove asserts checking for null in cases where null signals there is no local class
>   and the fields of the class in the stream are to be read and discarded.

The draft of the serialization specification update for value objects is:
http://cr.openjdk.java.net/~rriggs/lworld-serial-spec/

-------------

PR: https://git.openjdk.org/valhalla/pull/761


From mchung at openjdk.org  Wed Oct 12 22:57:24 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Wed, 12 Oct 2022 22:57:24 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v3]
In-Reply-To: <s1GE2wyyNF6lZnfweudY0xAb4RIamFdqyREZYyaxJ9k=.0cd4ebfc-df1c-46e2-8fe1-fcf295551120@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
 <s1GE2wyyNF6lZnfweudY0xAb4RIamFdqyREZYyaxJ9k=.0cd4ebfc-df1c-46e2-8fe1-fcf295551120@github.com>
Message-ID: <BKOpxnrwr3XS69KBFlGbP2KiSX-UYlKn_Ub5uF3TwFI=.86c9435b-3b1b-4b8e-bb3a-a08f78b5516f@github.com>

On Thu, 29 Sep 2022 19:45:46 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

>> Specify serialization and deserialization of value objects. Value objects are immutable, Externalizable relies on mutation, Externalizable is not supported for value objects. Revised tests and added tests for graphs of value objects.
>> 
>> Somewhat similar to Record serialization in that back references to the value object store 'null' until the value object is completely deserialized.
>> Similar to regular serialization/deserialization of identity objects.  The serialized form is the same.
>> Deserialization is done by reading fields and storing into a Buffered larval object, converted to a value object when all of the fields have been read and stored.
>> 
>> Externalizable and Value Objects are incompatible, value objects are immutable and Externalizable.readExternal is expected to modify the fields of an allocated object.
>
> Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Remove asserts checking for null in cases where null signals there is no local class
>   and the fields of the class in the stream are to be read and discarded.

This seems a reasonable way to provide a minimal serialization support for value classes to allow proper solution to be explored.

src/java.base/share/classes/java/io/ObjectStreamClass.java line 2100:

> 2098:                 vals[offsets[i]] = switch (typeCodes[i]) {
> 2099:                     case 'L', '[' ->
> 2100:                             unsafe.isFlattened(f)

Flattened field has `Q` descriptor.  This PR doesn't support primitive class, right?

src/java.base/share/classes/jdk/internal/reflect/ReflectionFactory.java line 335:

> 333:             return constructorToCall;
> 334:         }
> 335:         if (cl.isValue()) {

Perhaps `newConstructorForExternalization` should also throw UOE if the given class is a value class?

-------------

PR: https://git.openjdk.org/valhalla/pull/761


From rriggs at openjdk.org  Thu Oct 13 14:09:31 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Thu, 13 Oct 2022 14:09:31 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v3]
In-Reply-To: <BKOpxnrwr3XS69KBFlGbP2KiSX-UYlKn_Ub5uF3TwFI=.86c9435b-3b1b-4b8e-bb3a-a08f78b5516f@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
 <s1GE2wyyNF6lZnfweudY0xAb4RIamFdqyREZYyaxJ9k=.0cd4ebfc-df1c-46e2-8fe1-fcf295551120@github.com>
 <BKOpxnrwr3XS69KBFlGbP2KiSX-UYlKn_Ub5uF3TwFI=.86c9435b-3b1b-4b8e-bb3a-a08f78b5516f@github.com>
Message-ID: <HuR_2MWT6HW1rg1BClhIHNubUgBuQJgP2YeWPeFWaE0=.fadabfcf-4b55-4190-815c-7bc1838194c8@github.com>

On Wed, 12 Oct 2022 21:25:55 GMT, Mandy Chung <mchung at openjdk.org> wrote:

>> Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Remove asserts checking for null in cases where null signals there is no local class
>>   and the fields of the class in the stream are to be read and discarded.
>
> src/java.base/share/classes/java/io/ObjectStreamClass.java line 2100:
> 
>> 2098:                 vals[offsets[i]] = switch (typeCodes[i]) {
>> 2099:                     case 'L', '[' ->
>> 2100:                             unsafe.isFlattened(f)
> 
> Flattened field has `Q` descriptor.  This PR doesn't support primitive class, right?

As far as I know, a "value type" (aka primitive class) is a value class and the only difference from the Unsafe API point of view is whether it is flattened or not. There may be some additional work for JEP 401.

-------------

PR: https://git.openjdk.org/valhalla/pull/761


From rriggs at openjdk.org  Thu Oct 13 15:05:19 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Thu, 13 Oct 2022 15:05:19 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v3]
In-Reply-To: <HuR_2MWT6HW1rg1BClhIHNubUgBuQJgP2YeWPeFWaE0=.fadabfcf-4b55-4190-815c-7bc1838194c8@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
 <s1GE2wyyNF6lZnfweudY0xAb4RIamFdqyREZYyaxJ9k=.0cd4ebfc-df1c-46e2-8fe1-fcf295551120@github.com>
 <BKOpxnrwr3XS69KBFlGbP2KiSX-UYlKn_Ub5uF3TwFI=.86c9435b-3b1b-4b8e-bb3a-a08f78b5516f@github.com>
 <HuR_2MWT6HW1rg1BClhIHNubUgBuQJgP2YeWPeFWaE0=.fadabfcf-4b55-4190-815c-7bc1838194c8@github.com>
Message-ID: <c49NVmxxSqyz4gGJ4TRj-hFNp8_jEemCsIlskxvrtuE=.70504b02-5216-4cb8-8196-a1b4af6344ed@github.com>

On Thu, 13 Oct 2022 14:07:24 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

>> src/java.base/share/classes/java/io/ObjectStreamClass.java line 2100:
>> 
>>> 2098:                 vals[offsets[i]] = switch (typeCodes[i]) {
>>> 2099:                     case 'L', '[' ->
>>> 2100:                             unsafe.isFlattened(f)
>> 
>> Flattened field has `Q` descriptor.  This PR doesn't support primitive class, right?
>
> As far as I know, a "value type" (aka primitive class) is a value class and the only difference from the Unsafe API point of view is whether it is flattened or not. There may be some additional work for JEP 401.

Also added a few tests for serializable primitive classes.

-------------

PR: https://git.openjdk.org/valhalla/pull/761


From rriggs at openjdk.org  Thu Oct 13 15:15:32 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Thu, 13 Oct 2022 15:15:32 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v4]
In-Reply-To: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
Message-ID: <cQqVqojcFrRITKSdHMl0fpHhRkmU6l6AwNtuiePzPA0=.15156bda-f046-40c2-89ac-15d10f418a4c@github.com>

> Specify serialization and deserialization of value objects. Value objects are immutable, Externalizable relies on mutation, Externalizable is not supported for value objects. Revised tests and added tests for graphs of value objects.
> 
> Somewhat similar to Record serialization in that back references to the value object store 'null' until the value object is completely deserialized.
> Similar to regular serialization/deserialization of identity objects.  The serialized form is the same.
> Deserialization is done by reading fields and storing into a Buffered larval object, converted to a value object when all of the fields have been read and stored.
> 
> Externalizable and Value Objects are incompatible, value objects are immutable and Externalizable.readExternal is expected to modify the fields of an allocated object.

Roger Riggs has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains ten commits:

 - Merge branch 'lworld' into 8281295-lworld-serialize
 - Added throw UOE to ReflectionFactory.newConstructorForExternalization.
   Added tests for serializable primitive value class.
 - Merge branch 'lworld' into 8281295-lworld-serialize
 - fix typos
 - Remove asserts checking for null in cases where null signals there is no local class
   and the fields of the class in the stream are to be read and discarded.
 - javadoc cleanup
 - Merge branch 'lworld' into 8281295-lworld-serialize
 - Merge branch 'lworld' into 8281295-lworld-serialize
 - 8281295: Prototype serialization/deserialization using Unsafe.
   
   Specify serialization and deserialization of value objects.
   Value objects are immutable, Externalizable relies on mutation,
   Externalizable is not supported for value objects.
   Revised tests and added tests for graphs of value objects.

-------------

Changes: https://git.openjdk.org/valhalla/pull/761/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=761&range=03
  Stats: 1107 lines in 9 files changed: 805 ins; 289 del; 13 mod
  Patch: https://git.openjdk.org/valhalla/pull/761.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/761/head:pull/761

PR: https://git.openjdk.org/valhalla/pull/761


From mchung at openjdk.org  Thu Oct 13 16:28:25 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Thu, 13 Oct 2022 16:28:25 GMT
Subject: [lworld] Integrated: 8280979: [lworld] Rename PrimitiveObjectMethods
 to ValueObjectMethods
In-Reply-To: <1nByP6vfIscIfRj5db86poJie9ZKYbdxo56_JGEmoM4=.db191e6a-1a02-4e7d-956d-37e5b1058d89@github.com>
References: <1nByP6vfIscIfRj5db86poJie9ZKYbdxo56_JGEmoM4=.db191e6a-1a02-4e7d-956d-37e5b1058d89@github.com>
Message-ID: <Gpikp2yt6ut4mwrv0E6rbXo1sntQQKeOmTKlTaBTC80=.b156878c-224a-49d2-8f3f-7e7286d07196@github.com>

On Tue, 11 Oct 2022 22:28:06 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> This PR does the renaming to match the Value Objects JEP.

This pull request has now been integrated.

Changeset: 191942b5
Author:    Mandy Chung <mchung at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/191942b54cef9de298646a00d25882bc94e1d4b4
Stats:     996 lines in 14 files changed: 491 ins; 479 del; 26 mod

8280979: [lworld] Rename PrimitiveObjectMethods to ValueObjectMethods

Reviewed-by: rriggs

-------------

PR: https://git.openjdk.org/valhalla/pull/787


From rriggs at openjdk.org  Thu Oct 13 19:53:46 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Thu, 13 Oct 2022 19:53:46 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v5]
In-Reply-To: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
Message-ID: <-OVX9qgNPnWY3P6dyGDAvgzhbLHg57cD6Fzuyg7ictU=.018f7bbf-2031-4367-9fdf-f131d784fbd1@github.com>

> Specify serialization and deserialization of value objects. Value objects are immutable, Externalizable relies on mutation, Externalizable is not supported for value objects. Revised tests and added tests for graphs of value objects.
> 
> Somewhat similar to Record serialization in that back references to the value object store 'null' until the value object is completely deserialized.
> Similar to regular serialization/deserialization of identity objects.  The serialized form is the same.
> Deserialization is done by reading fields and storing into a Buffered larval object, converted to a value object when all of the fields have been read and stored.
> 
> Externalizable and Value Objects are incompatible, value objects are immutable and Externalizable.readExternal is expected to modify the fields of an allocated object.

Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:

  Add primitive class signatures to be treated the same as L types

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/761/files
  - new: https://git.openjdk.org/valhalla/pull/761/files/6e303c79..b51c128f

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=761&range=04
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=761&range=03-04

  Stats: 2 lines in 1 file changed: 0 ins; 0 del; 2 mod
  Patch: https://git.openjdk.org/valhalla/pull/761.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/761/head:pull/761

PR: https://git.openjdk.org/valhalla/pull/761


From duke at openjdk.org  Thu Oct 13 21:41:29 2022
From: duke at openjdk.org (ExE Boss)
Date: Thu, 13 Oct 2022 21:41:29 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v5]
In-Reply-To: <-OVX9qgNPnWY3P6dyGDAvgzhbLHg57cD6Fzuyg7ictU=.018f7bbf-2031-4367-9fdf-f131d784fbd1@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
 <-OVX9qgNPnWY3P6dyGDAvgzhbLHg57cD6Fzuyg7ictU=.018f7bbf-2031-4367-9fdf-f131d784fbd1@github.com>
Message-ID: <9DYcI1n_FlpNZB1Sm72pq75-RdlDzgKsFOKW_mXTK_Q=.1c1cbeef-abaa-48be-a10c-f01218eaeb43@github.com>

On Thu, 13 Oct 2022 19:53:46 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

>> Specify serialization and deserialization of value objects. Value objects are immutable, Externalizable relies on mutation, Externalizable is not supported for value objects. Revised tests and added tests for graphs of value objects.
>> 
>> Somewhat similar to Record serialization in that back references to the value object store 'null' until the value object is completely deserialized.
>> Similar to regular serialization/deserialization of identity objects.  The serialized form is the same.
>> Deserialization is done by reading fields and storing into a Buffered larval object, converted to a value object when all of the fields have been read and stored.
>> 
>> Externalizable and Value Objects are incompatible, value objects are immutable and Externalizable.readExternal is expected to modify the fields of an allocated object.
>
> Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Add primitive class signatures to be treated the same as L types

The?`java.io.Externalizable` interface should?probably be?marked as?an?identity?interface.

-------------

PR: https://git.openjdk.org/valhalla/pull/761


From rriggs at openjdk.org  Thu Oct 13 22:26:19 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Thu, 13 Oct 2022 22:26:19 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v5]
In-Reply-To: <9DYcI1n_FlpNZB1Sm72pq75-RdlDzgKsFOKW_mXTK_Q=.1c1cbeef-abaa-48be-a10c-f01218eaeb43@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
 <-OVX9qgNPnWY3P6dyGDAvgzhbLHg57cD6Fzuyg7ictU=.018f7bbf-2031-4367-9fdf-f131d784fbd1@github.com>
 <9DYcI1n_FlpNZB1Sm72pq75-RdlDzgKsFOKW_mXTK_Q=.1c1cbeef-abaa-48be-a10c-f01218eaeb43@github.com>
Message-ID: <vegZzHYjLNDxPdIZQR2RYKWypWQiHAa8pNrYhzafVcE=.ede6dba1-1c08-4d6f-a3cd-11b18e10496d@github.com>

On Thu, 13 Oct 2022 21:39:15 GMT, ExE Boss <duke at openjdk.org> wrote:

> The `java.io.Externalizable` interface should probably be marked as an identity interface.

That would head-off some misuses, and so far I'm not aware of a viable use case.

-------------

PR: https://git.openjdk.org/valhalla/pull/761


From vromero at openjdk.org  Thu Oct 13 22:54:32 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Thu, 13 Oct 2022 22:54:32 GMT
Subject: [lworld] RFR: 8295086: [lworld] follow-up for JDK-8288131 [v2]
In-Reply-To: <uwWHqZsU469Jj2mgICVf3XvtCROEzeL31ZM30N_rknA=.29be9b99-caeb-4d8d-8409-f83706e1f04a@github.com>
References: <uwWHqZsU469Jj2mgICVf3XvtCROEzeL31ZM30N_rknA=.29be9b99-caeb-4d8d-8409-f83706e1f04a@github.com>
Message-ID: <67vJiSJKkQJLfym_34aWV7PO7Nng8mNx7SaFeMn_9vI=.2546efc8-80c1-428f-8745-1efaf8337de1@github.com>

> code related to primitive classes needs to be hidden behind the -XDenablePrimitiveClasses flag, this work is a follow-up of JDK-8288131. Hiding this remaining code behind the flag is not absolutely necessary as it probably won't be executed anyways but it will make easier to located, and extract if needed, the code related with primitive classes only

Vicente Romero has updated the pull request incrementally with one additional commit since the last revision:

  following on with the changes after merge

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/793/files
  - new: https://git.openjdk.org/valhalla/pull/793/files/d65ccf63..115b7c18

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=793&range=01
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=793&range=00-01

  Stats: 83 lines in 9 files changed: 35 ins; 3 del; 45 mod
  Patch: https://git.openjdk.org/valhalla/pull/793.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/793/head:pull/793

PR: https://git.openjdk.org/valhalla/pull/793


From mchung at openjdk.org  Fri Oct 14 00:18:00 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Fri, 14 Oct 2022 00:18:00 GMT
Subject: [lworld] RFR: 8295234: [lworld] inner class of older classfile
 versions is not recognized as identity class
Message-ID: <TPoR-h0JuoxO3-IsoPGsnMPjWM2C5aBKthV7k6gQp80=.88ff83bb-fec7-4071-b6dc-1dcdbfcca697@github.com>

The modifiers of an inner class uses the access flags in the `InnerClasses`  attribute.   An inner class with an older class file version does not have `ACC_IDENTITY` flag set.   Instead of having `Class::isIdentity`  to depend on `getModifiers()`, this PR proposes to make `Class::isIdentity` a native method implemented by the VM. In addition, `Class::isIdentity` can be made as intrinsics.

I leave `Class::isValue` as is.   A value class must have `ACC_VALUE` flag set.   `Class::getModifiers` is already intrinsified.   We can revisit this in the future when there is a need to make it as a native VM method.

-------------

Commit messages:
 - 8295234: [lworld] inner class of older bytecode versions is not recognized as identity class

Changes: https://git.openjdk.org/valhalla/pull/794/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=794&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8295234
  Stats: 25 lines in 5 files changed: 18 ins; 6 del; 1 mod
  Patch: https://git.openjdk.org/valhalla/pull/794.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/794/head:pull/794

PR: https://git.openjdk.org/valhalla/pull/794


From mchung at openjdk.org  Fri Oct 14 00:55:38 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Fri, 14 Oct 2022 00:55:38 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v5]
In-Reply-To: <-OVX9qgNPnWY3P6dyGDAvgzhbLHg57cD6Fzuyg7ictU=.018f7bbf-2031-4367-9fdf-f131d784fbd1@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
 <-OVX9qgNPnWY3P6dyGDAvgzhbLHg57cD6Fzuyg7ictU=.018f7bbf-2031-4367-9fdf-f131d784fbd1@github.com>
Message-ID: <Shl71_Y8QYKeQfv9eh2QT-bbU5qfi1tHySzDUW7DLuk=.a9227285-3e9a-4702-946e-49a0f6380ca1@github.com>

On Thu, 13 Oct 2022 19:53:46 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

>> Specify serialization and deserialization of value objects. Value objects are immutable, Externalizable relies on mutation, Externalizable is not supported for value objects. Revised tests and added tests for graphs of value objects.
>> 
>> Somewhat similar to Record serialization in that back references to the value object store 'null' until the value object is completely deserialized.
>> Similar to regular serialization/deserialization of identity objects.  The serialized form is the same.
>> Deserialization is done by reading fields and storing into a Buffered larval object, converted to a value object when all of the fields have been read and stored.
>> 
>> Externalizable and Value Objects are incompatible, value objects are immutable and Externalizable.readExternal is expected to modify the fields of an allocated object.
>
> Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Add primitive class signatures to be treated the same as L types

To recap our offline discussion, I would expect `ObjectStreamField::getTypeString` should return `Q-`descriptor if the field type is a primitive class value type (consistent with `Field`).   For example,


class Test implements Serializable {
   SerializablePrimitivePoint p;
   SerializableValuePoint v;
}


where `p` may be a flattened field of a primitive class value type but `v` is a field of a reference type (value class).   In that case, `Field.getType().descriptorString()` on `p`  should return `QSerializablePrimitivePoint;`.

Since this involves the interoperability and future evolution, perhaps add the serialization support for primitive classes as a follow-on issue.   Better to revert the change added for `Q` descriptor and keep this PR for value class only.

-------------

PR: https://git.openjdk.org/valhalla/pull/761


From rriggs at openjdk.org  Fri Oct 14 13:45:41 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Fri, 14 Oct 2022 13:45:41 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v6]
In-Reply-To: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
Message-ID: <g9ryQFEaZ2Dxd4WPml2dxCBpZJq7v3ppofvh341jdsI=.ea6297e5-3b39-446d-a83c-a5e1fbbcd38b@github.com>

> Specify serialization and deserialization of value objects. Value objects are immutable, Externalizable relies on mutation, Externalizable is not supported for value objects. Revised tests and added tests for graphs of value objects.
> 
> Somewhat similar to Record serialization in that back references to the value object store 'null' until the value object is completely deserialized.
> Similar to regular serialization/deserialization of identity objects.  The serialized form is the same.
> Deserialization is done by reading fields and storing into a Buffered larval object, converted to a value object when all of the fields have been read and stored.
> 
> Externalizable and Value Objects are incompatible, value objects are immutable and Externalizable.readExternal is expected to modify the fields of an allocated object.

Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:

  Revert Q type changes, value types (aka primitive classes) are not fully supported

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/761/files
  - new: https://git.openjdk.org/valhalla/pull/761/files/b51c128f..bf1f3a17

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=761&range=05
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=761&range=04-05

  Stats: 2 lines in 1 file changed: 0 ins; 0 del; 2 mod
  Patch: https://git.openjdk.org/valhalla/pull/761.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/761/head:pull/761

PR: https://git.openjdk.org/valhalla/pull/761


From mchung at openjdk.org  Fri Oct 14 16:07:21 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Fri, 14 Oct 2022 16:07:21 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v6]
In-Reply-To: <g9ryQFEaZ2Dxd4WPml2dxCBpZJq7v3ppofvh341jdsI=.ea6297e5-3b39-446d-a83c-a5e1fbbcd38b@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
 <g9ryQFEaZ2Dxd4WPml2dxCBpZJq7v3ppofvh341jdsI=.ea6297e5-3b39-446d-a83c-a5e1fbbcd38b@github.com>
Message-ID: <wAdbAnFcuVoqBHfx9B9IXAp1RER9NFMVG_Zit2Drh48=.25057fcb-e9e5-4c9f-ae20-0b0a1d77dfeb@github.com>

On Fri, 14 Oct 2022 13:45:41 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

>> Specify serialization and deserialization of value objects. Value objects are immutable, Externalizable relies on mutation, Externalizable is not supported for value objects. Revised tests and added tests for graphs of value objects.
>> 
>> Somewhat similar to Record serialization in that back references to the value object store 'null' until the value object is completely deserialized.
>> Similar to regular serialization/deserialization of identity objects.  The serialized form is the same.
>> Deserialization is done by reading fields and storing into a Buffered larval object, converted to a value object when all of the fields have been read and stored.
>> 
>> Externalizable and Value Objects are incompatible, value objects are immutable and Externalizable.readExternal is expected to modify the fields of an allocated object.
>
> Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Revert Q type changes, value types (aka primitive classes) are not fully supported

This looks good as an initial prototype.

src/java.base/share/classes/java/io/ObjectOutputStream.java line 136:

> 134:  * object such as a record.
> 135:  *
> 136:  * Value objects cannot be `java.io.Externalizable`.

Suggestion:

 * Value objects cannot be {@code java.io.Externalizable}.

-------------

Marked as reviewed by mchung (Committer).

PR: https://git.openjdk.org/valhalla/pull/761


From fparain at openjdk.org  Fri Oct 14 17:32:27 2022
From: fparain at openjdk.org (Frederic Parain)
Date: Fri, 14 Oct 2022 17:32:27 GMT
Subject: [lworld] RFR: 8295234: [lworld] inner class of older classfile
 versions is not recognized as identity class
In-Reply-To: <TPoR-h0JuoxO3-IsoPGsnMPjWM2C5aBKthV7k6gQp80=.88ff83bb-fec7-4071-b6dc-1dcdbfcca697@github.com>
References: <TPoR-h0JuoxO3-IsoPGsnMPjWM2C5aBKthV7k6gQp80=.88ff83bb-fec7-4071-b6dc-1dcdbfcca697@github.com>
Message-ID: <dbi6wnf4C1HM8ZvivecUYxPfxpzfPAY87Oook2AqZbs=.137e0bad-efb9-497e-a9e6-b6493278b432@github.com>

On Fri, 14 Oct 2022 00:10:57 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> The modifiers of an inner class uses the access flags in the `InnerClasses`  attribute.   An inner class with an older class file version does not have `ACC_IDENTITY` flag set.   Instead of having `Class::isIdentity`  to depend on `getModifiers()`, this PR proposes to make `Class::isIdentity` a native method implemented by the VM. In addition, `Class::isIdentity` can be made as intrinsics.
> 
> I leave `Class::isValue` as is.   A value class must have `ACC_VALUE` flag set.   `Class::getModifiers` is already intrinsified.   We can revisit this in the future when there is a need to make it as a native VM method.

Looks good to me.

Fred

-------------

Marked as reviewed by fparain (Committer).

PR: https://git.openjdk.org/valhalla/pull/794


From mchung at openjdk.org  Fri Oct 14 19:20:49 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Fri, 14 Oct 2022 19:20:49 GMT
Subject: [lworld] Integrated: 8295234: [lworld] inner class of older classfile
 versions is not recognized as identity class
In-Reply-To: <TPoR-h0JuoxO3-IsoPGsnMPjWM2C5aBKthV7k6gQp80=.88ff83bb-fec7-4071-b6dc-1dcdbfcca697@github.com>
References: <TPoR-h0JuoxO3-IsoPGsnMPjWM2C5aBKthV7k6gQp80=.88ff83bb-fec7-4071-b6dc-1dcdbfcca697@github.com>
Message-ID: <eZpy2MOvBNzDs8g3tQemlEWxlnNSUHDhW2UlMhb6esE=.c9901ccb-fc2f-4d99-b649-7b52882ff3aa@github.com>

On Fri, 14 Oct 2022 00:10:57 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> The modifiers of an inner class uses the access flags in the `InnerClasses`  attribute.   An inner class with an older class file version does not have `ACC_IDENTITY` flag set.   Instead of having `Class::isIdentity`  to depend on `getModifiers()`, this PR proposes to make `Class::isIdentity` a native method implemented by the VM. In addition, `Class::isIdentity` can be made as intrinsics.
> 
> I leave `Class::isValue` as is.   A value class must have `ACC_VALUE` flag set.   `Class::getModifiers` is already intrinsified.   We can revisit this in the future when there is a need to make it as a native VM method.

This pull request has now been integrated.

Changeset: 336ef700
Author:    Mandy Chung <mchung at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/336ef70095ec90b4be47aa29117c943b3b751f84
Stats:     25 lines in 5 files changed: 18 ins; 6 del; 1 mod

8295234: [lworld] inner class of older classfile versions is not recognized as identity class

Reviewed-by: fparain

-------------

PR: https://git.openjdk.org/valhalla/pull/794


From rriggs at openjdk.org  Fri Oct 14 21:22:29 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Fri, 14 Oct 2022 21:22:29 GMT
Subject: [lworld] RFR: 8281295: Prototype serialization/deserialization
 using Unsafe. [v7]
In-Reply-To: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
Message-ID: <-30pdZoORllsw5fAL6uybX0eKnjtufJ2viAv84Qa5fM=.777642fb-45b6-40cf-8ebb-16462a0c514a@github.com>

> Specify serialization and deserialization of value objects. Value objects are immutable, Externalizable relies on mutation, Externalizable is not supported for value objects. Revised tests and added tests for graphs of value objects.
> 
> Somewhat similar to Record serialization in that back references to the value object store 'null' until the value object is completely deserialized.
> Similar to regular serialization/deserialization of identity objects.  The serialized form is the same.
> Deserialization is done by reading fields and storing into a Buffered larval object, converted to a value object when all of the fields have been read and stored.
> 
> Externalizable and Value Objects are incompatible, value objects are immutable and Externalizable.readExternal is expected to modify the fields of an allocated object.

Roger Riggs has updated the pull request incrementally with one additional commit since the last revision:

  Correct formatting of interface Externalizable

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/761/files
  - new: https://git.openjdk.org/valhalla/pull/761/files/bf1f3a17..c14c27a5

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=761&range=06
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=761&range=05-06

  Stats: 1 line in 1 file changed: 0 ins; 0 del; 1 mod
  Patch: https://git.openjdk.org/valhalla/pull/761.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/761/head:pull/761

PR: https://git.openjdk.org/valhalla/pull/761


From rriggs at openjdk.org  Fri Oct 14 21:51:20 2022
From: rriggs at openjdk.org (Roger Riggs)
Date: Fri, 14 Oct 2022 21:51:20 GMT
Subject: [lworld] Integrated: 8281295: Prototype serialization/deserialization
 using Unsafe.
In-Reply-To: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
References: <HWAhL-yZcoYv5Pc6u7T3nUDvns7SM7hSPHH4wsdlx6k=.8a99114e-c33f-4175-baba-4fc22cadc589@github.com>
Message-ID: <d39vBbMnZ9nsGG7pHVlD-nF0D-2-5CmnpN6tJ2YeAb8=.0b29f546-4221-4f9b-a324-187fa473dd0c@github.com>

On Mon, 19 Sep 2022 18:21:31 GMT, Roger Riggs <rriggs at openjdk.org> wrote:

> Specify serialization and deserialization of value objects. Value objects are immutable, Externalizable relies on mutation, Externalizable is not supported for value objects. Revised tests and added tests for graphs of value objects.
> 
> Somewhat similar to Record serialization in that back references to the value object store 'null' until the value object is completely deserialized.
> Similar to regular serialization/deserialization of identity objects.  The serialized form is the same.
> Deserialization is done by reading fields and storing into a Buffered larval object, converted to a value object when all of the fields have been read and stored.
> 
> Externalizable and Value Objects are incompatible, value objects are immutable and Externalizable.readExternal is expected to modify the fields of an allocated object.

This pull request has now been integrated.

Changeset: 143f6b28
Author:    Roger Riggs <rriggs at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/143f6b28b20f0d3ee6c52fad58f5836230bfb112
Stats:     1107 lines in 9 files changed: 805 ins; 289 del; 13 mod

8281295: Prototype serialization/deserialization using Unsafe.

Reviewed-by: mchung

-------------

PR: https://git.openjdk.org/valhalla/pull/761


From thartmann at openjdk.org  Tue Oct 18 14:48:03 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Tue, 18 Oct 2022 14:48:03 GMT
Subject: [lworld] RFR: 8295455: [lworld] find_callee_info_helper hits
 "must have non-abstract attached method" assert
Message-ID: <EEzeTp1y7pm5pFBljdJlh4hcpb3GJvNycKWYXY3LdHk=.1d6d5fd3-b56e-4554-91ef-518e83bb646a@github.com>

I was finally able to reproduce this intermittent issue. The root cause is that the call (re-)resolution code does not correctly detect a C1 caller frame if that one got deoptimized and therefore either hits an assert or uses the wrong callee entry point, leading to incorrect execution.

I also fixed an unrelated issue in the scalarized entry point of a C1 compiled method, that leads to the orig_pc not being zeroed.

Thanks,
Tobias

-------------

Commit messages:
 - 8295455: [lworld] find_callee_info_helper hits "must have non-abstract attached method" assert

Changes: https://git.openjdk.org/valhalla/pull/795/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=795&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8295455
  Stats: 14 lines in 4 files changed: 3 ins; 1 del; 10 mod
  Patch: https://git.openjdk.org/valhalla/pull/795.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/795/head:pull/795

PR: https://git.openjdk.org/valhalla/pull/795


From dsimms at openjdk.org  Tue Oct 18 15:14:08 2022
From: dsimms at openjdk.org (David Simms)
Date: Tue, 18 Oct 2022 15:14:08 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
Message-ID: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>

Merged javac and VM work together, includes VM support (JDK-8294224)

-------------

Commit messages:
 - Fixes for valhalla/valuetypes/Serialization.java
 - Merge branch 'lworld' into vnew_javac_hotspot
 - MethodHandles.java aware of vnew, no longer static/non-static <init>
 - MethodHandles::resolve_MemberName for <vnew>
 - Merge pull request #2 from MrSimms/fixing.failing.regression.tests
 - fixing failing regression tests
 - Merge branch 'vnew' into vnew_javac_hotspot
 - Compile issue
 - Adjust hotspot_valhalla tests
 - Abstract value class constructor is vanilla <init>
 - ... and 13 more: https://git.openjdk.org/valhalla/compare/cf05d5c1...953c2060

Changes: https://git.openjdk.org/valhalla/pull/788/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=788&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8247569
  Stats: 667 lines in 81 files changed: 270 ins; 94 del; 303 mod
  Patch: https://git.openjdk.org/valhalla/pull/788.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/788/head:pull/788

PR: https://git.openjdk.org/valhalla/pull/788


From duke at openjdk.org  Tue Oct 18 15:14:10 2022
From: duke at openjdk.org (ExE Boss)
Date: Tue, 18 Oct 2022 15:14:10 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
In-Reply-To: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
Message-ID: <mS2UZ378TgDeZ_zG3h8PUOSE6gbE9rvJlgpl3AMdSos=.658f5a12-124f-45ce-bd91-4fceadb2066e@github.com>

On Wed, 12 Oct 2022 08:44:41 GMT, David Simms <dsimms at openjdk.org> wrote:

> Merged javac and VM work together, includes VM support (JDK-8294224)

Personally, I?prefer calling?it just?`<new>`, which?would better?support [JDK?8284239].

[JDK?8284239]: https://bugs.openjdk.org/browse/JDK-8284239 "Consider?adding `array::<new>(?)` and?using?it to?compile `T[]::new` method?references"

src/java.base/share/classes/jdk/internal/reflect/ClassFileAssembler.java line 59:

> 57:             classFileVersion += Opcodes.V_PREVIEW;
> 58:         }
> 59:         emitInt(classFileVersion);

Note?that the?reflection implementation intentionally?uses class?file version?49.0 so?that the?`StackMapFrames`?attribute and?other stricter?requirements are?not?necessary. See?[JEP?416] for?details.

https://github.com/openjdk/valhalla/blob/6ae7e4d4aad5712cf2fe6ab9f98dc424fa4170cb/src/hotspot/share/classfile/verifier.hpp#L37-L44

[JEP?416]: https://openjdk.org/jeps/416

-------------

PR: https://git.openjdk.org/valhalla/pull/788


From dsimms at openjdk.org  Tue Oct 18 15:14:12 2022
From: dsimms at openjdk.org (David Simms)
Date: Tue, 18 Oct 2022 15:14:12 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
In-Reply-To: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
Message-ID: <PtaL0zG3X3iiV2qABEYgilUJv30SAKSE5EuqmxvXlvE=.8d3fb02e-0af7-4429-8b57-7e3a0f52cb56@github.com>

On Wed, 12 Oct 2022 08:44:41 GMT, David Simms <dsimms at openjdk.org> wrote:

> Merged javac and VM work together, includes VM support (JDK-8294224)

src/java.base/share/classes/java/lang/constant/DirectMethodHandleDescImpl.java line 67:

> 65:     DirectMethodHandleDescImpl(Kind kind, ClassDesc owner, String name, MethodTypeDesc type) {
> 66:         if (kind == CONSTRUCTOR) {
> 67:             // TODO: missing "value objects", they don't use Q-type desc

Needs fixing to support Value Objects (currently only correct for primitive value types)

-------------

PR: https://git.openjdk.org/valhalla/pull/788


From mchung at openjdk.org  Tue Oct 18 17:38:15 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Tue, 18 Oct 2022 17:38:15 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
In-Reply-To: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
Message-ID: <Vyyp8Ee0aPAOwv6ITeaXvE8PCTl6qkVJEaoVAIAlp0o=.113edad2-cdda-4450-b385-8e9f57758507@github.com>

On Wed, 12 Oct 2022 08:44:41 GMT, David Simms <dsimms at openjdk.org> wrote:

> Merged javac and VM work together, includes VM support (JDK-8294224)

src/java.base/share/classes/java/lang/invoke/MemberName.java line 527:

> 525:     }
> 526:     /** Query whether this member is a constructor. */
> 527:     boolean isConstructor(String name) {

`findConstructor` only finds `<init>`.  To get `<vnew>` method handle, it just uses `Lookup::findStatic` as it's a static method.   This `isConstructor` method would cause confusion.  I can clean this up as a follow-up.

src/java.base/share/classes/java/lang/invoke/MethodHandles.java line 2765:

> 2763:                 throw new NoSuchMethodException("Constructors must have void return type: " + refc.getName());
> 2764:             }
> 2765:             String name = refc.isValue() ? "<vnew>" : "<init>";

`findConstructor` should not find `<vnew>` which is just a static method.

src/java.base/share/classes/java/lang/invoke/MethodHandles.java line 3704:

> 3702:             return (name.startsWith("<") &&
> 3703:             ((name.equals("<init>") && refKind != REF_newInvokeSpecial) ||
> 3704:             ((name.equals("<vnew>") || name.equals("<clinit>")) && refKind != REF_invokeStatic)));

This method should only check for `<vnew>`.   `<clinit>` cannot be looked up.

src/java.base/share/classes/jdk/internal/reflect/MethodAccessorGenerator.java line 95:

> 93:         boolean isStaticFactory = declaringClass.isValue();
> 94:         return (ConstructorAccessor) generate(declaringClass,
> 95:                                               isStaticFactory ? "<vnew>" : "<init>",

Core reflection does not use this method accessor generator by default.   I don't expect this change is needed.  Are you seeing any test failure without this change and other relevant change in jdk.internal.reflect?

test/jdk/valhalla/valuetypes/StaticFactoryMethodHandleTest.java line 85:

> 83:     @Test
> 84:     public void testNoArgStaticFactory() throws Throwable {
> 85:         // test default vnew init factory

Suggestion:

        // test default static vnew factory

-------------

PR: https://git.openjdk.org/valhalla/pull/788


From thartmann at openjdk.org  Wed Oct 19 05:15:20 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Wed, 19 Oct 2022 05:15:20 GMT
Subject: [lworld] Integrated: 8295455: [lworld] find_callee_info_helper hits
 "must have non-abstract attached method" assert
In-Reply-To: <EEzeTp1y7pm5pFBljdJlh4hcpb3GJvNycKWYXY3LdHk=.1d6d5fd3-b56e-4554-91ef-518e83bb646a@github.com>
References: <EEzeTp1y7pm5pFBljdJlh4hcpb3GJvNycKWYXY3LdHk=.1d6d5fd3-b56e-4554-91ef-518e83bb646a@github.com>
Message-ID: <TWxGWpAdnnGNSZ9XSywoacnxtGZCgC8WwfwKStsq0CI=.70e8893f-7485-435b-a49c-23fc82ce4c02@github.com>

On Tue, 18 Oct 2022 14:41:45 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

> I was finally able to reproduce this intermittent issue. The root cause is that the call (re-)resolution code does not correctly detect a C1 caller frame if that one got deoptimized and therefore either hits an assert or uses the wrong callee entry point, leading to incorrect execution.
> 
> I also fixed an unrelated issue in the scalarized entry point of a C1 compiled method, that leads to the orig_pc not being zeroed.
> 
> Thanks,
> Tobias

This pull request has now been integrated.

Changeset: 688186b3
Author:    Tobias Hartmann <thartmann at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/688186b3fd7c7173c83069fe9d4fc87cfce63778
Stats:     14 lines in 4 files changed: 3 ins; 1 del; 10 mod

8295455: [lworld] find_callee_info_helper hits "must have non-abstract attached method" assert

-------------

PR: https://git.openjdk.org/valhalla/pull/795


From dsimms at openjdk.org  Wed Oct 19 08:31:21 2022
From: dsimms at openjdk.org (David Simms)
Date: Wed, 19 Oct 2022 08:31:21 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
In-Reply-To: <PtaL0zG3X3iiV2qABEYgilUJv30SAKSE5EuqmxvXlvE=.8d3fb02e-0af7-4429-8b57-7e3a0f52cb56@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
 <PtaL0zG3X3iiV2qABEYgilUJv30SAKSE5EuqmxvXlvE=.8d3fb02e-0af7-4429-8b57-7e3a0f52cb56@github.com>
Message-ID: <r6WDrnq-_Kl3AIcGfxi-VeBjrsXXR9PuG-_W9uB4-po=.ef78f20d-4f68-491e-8b94-bbff9214becd@github.com>

On Wed, 12 Oct 2022 08:50:10 GMT, David Simms <dsimms at openjdk.org> wrote:

>> Merged javac and VM work together, includes VM support (JDK-8294224)
>
> src/java.base/share/classes/java/lang/constant/DirectMethodHandleDescImpl.java line 67:
> 
>> 65:     DirectMethodHandleDescImpl(Kind kind, ClassDesc owner, String name, MethodTypeDesc type) {
>> 66:         if (kind == CONSTRUCTOR) {
>> 67:             // TODO: missing "value objects", they don't use Q-type desc
> 
> Needs fixing to support Value Objects (currently only correct for primitive value types)

...will be fixed in later PR

-------------

PR: https://git.openjdk.org/valhalla/pull/788


From dsimms at openjdk.org  Wed Oct 19 08:35:18 2022
From: dsimms at openjdk.org (David Simms)
Date: Wed, 19 Oct 2022 08:35:18 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
In-Reply-To: <Vyyp8Ee0aPAOwv6ITeaXvE8PCTl6qkVJEaoVAIAlp0o=.113edad2-cdda-4450-b385-8e9f57758507@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
 <Vyyp8Ee0aPAOwv6ITeaXvE8PCTl6qkVJEaoVAIAlp0o=.113edad2-cdda-4450-b385-8e9f57758507@github.com>
Message-ID: <djNOWwxh0x3utNjroFImziUF1ShlJcwClV-npztU9og=.6b3779dd-01ff-4d11-a365-0c076592fba8@github.com>

On Tue, 18 Oct 2022 17:01:57 GMT, Mandy Chung <mchung at openjdk.org> wrote:

>> Merged javac and VM work together, includes VM support (JDK-8294224)
>
> src/java.base/share/classes/java/lang/invoke/MemberName.java line 527:
> 
>> 525:     }
>> 526:     /** Query whether this member is a constructor. */
>> 527:     boolean isConstructor(String name) {
> 
> `findConstructor` only finds `<init>`.  To get `<vnew>` method handle, it just uses `Lookup::findStatic` as it's a static method.   This `isConstructor` method would cause confusion.  I can clean this up as a follow-up.

Thank you

-------------

PR: https://git.openjdk.org/valhalla/pull/788


From dsimms at openjdk.org  Wed Oct 19 10:30:15 2022
From: dsimms at openjdk.org (David Simms)
Date: Wed, 19 Oct 2022 10:30:15 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v2]
In-Reply-To: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
Message-ID: <q9Z1IdAH0D9Z6xzifBrXclBRD5xVpBroLlYS-vs4SLM=.c68901bf-f092-417e-bab3-5fc882e5fc67@github.com>

> Merged javac and VM work together, includes VM support (JDK-8294224)

David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 26 commits:

 - Post 8281295: revert MethodAccessorGenerator, no longer needed
 - Merge branch 'lworld' into vnew_javac_hotspot
 - mchung: core-libs fixes
 - Fixes for valhalla/valuetypes/Serialization.java
 - Merge branch 'lworld' into vnew_javac_hotspot
 - MethodHandles.java aware of vnew, no longer static/non-static <init>
 - MethodHandles::resolve_MemberName for <vnew>
 - Merge pull request #2 from MrSimms/fixing.failing.regression.tests
   
   fixing failing regression tests
 - fixing failing regression tests
 - Merge branch 'vnew' into vnew_javac_hotspot
 - ... and 16 more: https://git.openjdk.org/valhalla/compare/688186b3...c9d2bb6d

-------------

Changes: https://git.openjdk.org/valhalla/pull/788/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=788&range=01
  Stats: 730 lines in 82 files changed: 319 ins; 103 del; 308 mod
  Patch: https://git.openjdk.org/valhalla/pull/788.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/788/head:pull/788

PR: https://git.openjdk.org/valhalla/pull/788


From dsimms at openjdk.org  Wed Oct 19 10:34:48 2022
From: dsimms at openjdk.org (David Simms)
Date: Wed, 19 Oct 2022 10:34:48 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v3]
In-Reply-To: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
Message-ID: <HKuFatuoSGqJg_hi8Q9jxfr21Vhh-Ot-d_1pRw9fiNs=.dc6112b3-0a1b-412d-9f1f-0a88b9ae479f@github.com>

> Merged javac and VM work together, includes VM support (JDK-8294224)

David Simms has updated the pull request incrementally with one additional commit since the last revision:

  Update test/jdk/valhalla/valuetypes/StaticFactoryMethodHandleTest.java
  
  Co-authored-by: Mandy Chung <mandy.chung at oracle.com>

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/788/files
  - new: https://git.openjdk.org/valhalla/pull/788/files/c9d2bb6d..55e570ee

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=788&range=02
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=788&range=01-02

  Stats: 1 line in 1 file changed: 0 ins; 0 del; 1 mod
  Patch: https://git.openjdk.org/valhalla/pull/788.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/788/head:pull/788

PR: https://git.openjdk.org/valhalla/pull/788


From dsimms at openjdk.org  Wed Oct 19 10:34:50 2022
From: dsimms at openjdk.org (David Simms)
Date: Wed, 19 Oct 2022 10:34:50 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v3]
In-Reply-To: <Vyyp8Ee0aPAOwv6ITeaXvE8PCTl6qkVJEaoVAIAlp0o=.113edad2-cdda-4450-b385-8e9f57758507@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
 <Vyyp8Ee0aPAOwv6ITeaXvE8PCTl6qkVJEaoVAIAlp0o=.113edad2-cdda-4450-b385-8e9f57758507@github.com>
Message-ID: <Jita1KkSf_pRPOyKojtK3cSsnFUkvF6J7wH7BTgkoR8=.4aaec615-f76a-4e28-afcb-d81aa6bd9911@github.com>

On Tue, 18 Oct 2022 17:34:15 GMT, Mandy Chung <mchung at openjdk.org> wrote:

>> David Simms has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Update test/jdk/valhalla/valuetypes/StaticFactoryMethodHandleTest.java
>>   
>>   Co-authored-by: Mandy Chung <mandy.chung at oracle.com>
>
> src/java.base/share/classes/jdk/internal/reflect/MethodAccessorGenerator.java line 95:
> 
>> 93:         boolean isStaticFactory = declaringClass.isValue();
>> 94:         return (ConstructorAccessor) generate(declaringClass,
>> 95:                                               isStaticFactory ? "<vnew>" : "<init>",
> 
> Core reflection does not use this method accessor generator by default.   I don't expect this change is needed.  Are you seeing any test failure without this change and other relevant change in jdk.internal.reflect?

It's gone, JDK-8281295 removed its use

-------------

PR: https://git.openjdk.org/valhalla/pull/788


From thartmann at openjdk.org  Wed Oct 19 11:54:07 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Wed, 19 Oct 2022 11:54:07 GMT
Subject: [lworld] RFR: 8295655: [lworld] AArch64: Deoptimization from return
 value stub fails
Message-ID: <8vWkIqh2c9yyYxjuIxG0pnjBQMklO5-93RHLqAXqkPk=.b6c802fc-792c-4ec9-bb24-6a56d879accc@github.com>

If found this old issue after enabling stress testing on aarch64, similar to x86_64, `set_last_Java_frame` in `generate_return_value_stub` should not set fp on aarch64.

Thanks,
Tobias

-------------

Commit messages:
 - 8295655: [lworld] AArch64: Deoptimization from return value stub fails

Changes: https://git.openjdk.org/valhalla/pull/796/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=796&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8295655
  Stats: 7 lines in 3 files changed: 0 ins; 0 del; 7 mod
  Patch: https://git.openjdk.org/valhalla/pull/796.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/796/head:pull/796

PR: https://git.openjdk.org/valhalla/pull/796


From thartmann at openjdk.org  Wed Oct 19 13:18:47 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Wed, 19 Oct 2022 13:18:47 GMT
Subject: [lworld] Integrated: 8295655: [lworld] AArch64: Deoptimization from
 return value stub fails
In-Reply-To: <8vWkIqh2c9yyYxjuIxG0pnjBQMklO5-93RHLqAXqkPk=.b6c802fc-792c-4ec9-bb24-6a56d879accc@github.com>
References: <8vWkIqh2c9yyYxjuIxG0pnjBQMklO5-93RHLqAXqkPk=.b6c802fc-792c-4ec9-bb24-6a56d879accc@github.com>
Message-ID: <zsl9Aeqh-_2UHBrnvwEfjkJjP2n1dK4g1BMWJg5kpco=.b232a8b2-e837-483b-ad99-36f2cb1af695@github.com>

On Wed, 19 Oct 2022 11:46:37 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

> If found this old issue after enabling stress testing on aarch64, similar to x86_64, `set_last_Java_frame` in `generate_return_value_stub` should not set fp on aarch64.
> 
> Thanks,
> Tobias

This pull request has now been integrated.

Changeset: 6c933c17
Author:    Tobias Hartmann <thartmann at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/6c933c171145f1e28f22456f3556402286b05525
Stats:     7 lines in 3 files changed: 0 ins; 0 del; 7 mod

8295655: [lworld] AArch64: Deoptimization from return value stub fails

-------------

PR: https://git.openjdk.org/valhalla/pull/796


From mchung at openjdk.org  Wed Oct 19 16:14:27 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Wed, 19 Oct 2022 16:14:27 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v3]
In-Reply-To: <r6WDrnq-_Kl3AIcGfxi-VeBjrsXXR9PuG-_W9uB4-po=.ef78f20d-4f68-491e-8b94-bbff9214becd@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
 <PtaL0zG3X3iiV2qABEYgilUJv30SAKSE5EuqmxvXlvE=.8d3fb02e-0af7-4429-8b57-7e3a0f52cb56@github.com>
 <r6WDrnq-_Kl3AIcGfxi-VeBjrsXXR9PuG-_W9uB4-po=.ef78f20d-4f68-491e-8b94-bbff9214becd@github.com>
Message-ID: <o9i2YPBQ6PtWelqJNOlMWzJNsSnw6uPdx1gPvL5LjKw=.6499712b-d63b-4a23-809e-f1de43d49f70@github.com>

On Wed, 19 Oct 2022 08:27:38 GMT, David Simms <dsimms at openjdk.org> wrote:

>> src/java.base/share/classes/java/lang/constant/DirectMethodHandleDescImpl.java line 67:
>> 
>>> 65:     DirectMethodHandleDescImpl(Kind kind, ClassDesc owner, String name, MethodTypeDesc type) {
>>> 66:         if (kind == CONSTRUCTOR) {
>>> 67:             // TODO: missing "value objects", they don't use Q-type desc
>> 
>> Needs fixing to support Value Objects (currently only correct for primitive value types)
>
> ...will be fixed in later PR

My patch I sent to you already fixed this issue.

-------------

PR: https://git.openjdk.org/valhalla/pull/788


From fparain at openjdk.org  Wed Oct 19 18:46:24 2022
From: fparain at openjdk.org (Frederic Parain)
Date: Wed, 19 Oct 2022 18:46:24 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v3]
In-Reply-To: <HKuFatuoSGqJg_hi8Q9jxfr21Vhh-Ot-d_1pRw9fiNs=.dc6112b3-0a1b-412d-9f1f-0a88b9ae479f@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
 <HKuFatuoSGqJg_hi8Q9jxfr21Vhh-Ot-d_1pRw9fiNs=.dc6112b3-0a1b-412d-9f1f-0a88b9ae479f@github.com>
Message-ID: <7SxJvTfFdXqO1RL6i-YTMft7MfCg2yC1S-EzRsoDJuw=.15eb8526-99d8-40cd-b6fd-b68446b9ab8b@github.com>

On Wed, 19 Oct 2022 10:34:48 GMT, David Simms <dsimms at openjdk.org> wrote:

>> Merged javac and VM work together, includes VM support (JDK-8294224)
>
> David Simms has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Update test/jdk/valhalla/valuetypes/StaticFactoryMethodHandleTest.java
>   
>   Co-authored-by: Mandy Chung <mandy.chung at oracle.com>

VM changes look good to me (with one comment on is_object_constructor() and is_static_vnew_factory() methods).

src/hotspot/share/oops/method.cpp line 945:

> 943: // A method named <init>, if non-static, is a classic object constructor.
> 944: bool Method::is_object_constructor() const {
> 945:   return name() == vmSymbols::object_initializer_name() && !is_static();

Do we still need to check that the method is not static?
Now that constructors and static factories have different names, the check of the static flag looks superfluous (applies to the method below as well, and to the comments above the method declaration).

-------------

Marked as reviewed by fparain (Committer).

PR: https://git.openjdk.org/valhalla/pull/788


From mchung at openjdk.org  Wed Oct 19 20:24:29 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Wed, 19 Oct 2022 20:24:29 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v3]
In-Reply-To: <HKuFatuoSGqJg_hi8Q9jxfr21Vhh-Ot-d_1pRw9fiNs=.dc6112b3-0a1b-412d-9f1f-0a88b9ae479f@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
 <HKuFatuoSGqJg_hi8Q9jxfr21Vhh-Ot-d_1pRw9fiNs=.dc6112b3-0a1b-412d-9f1f-0a88b9ae479f@github.com>
Message-ID: <fsxyveP8cVaj4zT1hcaoh1DW51s0YCUL5_9eB-IT3pw=.a0d64cb5-9e53-4109-9fe1-fa96761b691c@github.com>

On Wed, 19 Oct 2022 10:34:48 GMT, David Simms <dsimms at openjdk.org> wrote:

>> Merged javac and VM work together, includes VM support (JDK-8294224)
>
> David Simms has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Update test/jdk/valhalla/valuetypes/StaticFactoryMethodHandleTest.java
>   
>   Co-authored-by: Mandy Chung <mandy.chung at oracle.com>

Looks good.  A couple more adjustment to the spec.

src/java.base/share/classes/java/lang/invoke/MethodHandleInfo.java line 163:

> 161:     /**
> 162:      * Returns the name of the cracked method handle's underlying member.
> 163:      * This is {@code "<init>"} or {@code "<vnew>"} if the underlying member

Suggestion:

     * This is {@code "<init>"} if the underlying member was a constructor,
     * or {@code "<vnew>"} if the underlying member was a value class static
     * factory method, else it is a simple method name or field name.

src/java.compiler/share/classes/javax/lang/model/element/ExecutableElement.java line 143:

> 141:      * {@return the simple name of a constructor, method, or
> 142:      * initializer}  For a constructor, the name {@code "<init>"} or
> 143:      * {@code "<vnew>"} is returned, for a static initializer, the

Suggestion:

     * initializer} For a constructor, the name {@code "<init>"} is
     * returned, for a value class static factory method, the name 
     * {@code "<vnew>"} is returned, for a static initializer, the

-------------

Marked as reviewed by mchung (Committer).

PR: https://git.openjdk.org/valhalla/pull/788


From dsimms at openjdk.org  Thu Oct 20 08:26:43 2022
From: dsimms at openjdk.org (David Simms)
Date: Thu, 20 Oct 2022 08:26:43 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v4]
In-Reply-To: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
Message-ID: <zIQsqBoADAozcgHu6FR_jSsyEEqF9qlMJTbKe3549k8=.f0f11123-ccdc-4ab1-9eab-7f0de8896ada@github.com>

> Merged javac and VM work together, includes VM support (JDK-8294224)

David Simms has updated the pull request incrementally with two additional commits since the last revision:

 - Update src/java.base/share/classes/java/lang/invoke/MethodHandleInfo.java
   
   Co-authored-by: Mandy Chung <mandy.chung at oracle.com>
 - Update src/java.compiler/share/classes/javax/lang/model/element/ExecutableElement.java
   
   Co-authored-by: Mandy Chung <mandy.chung at oracle.com>

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/788/files
  - new: https://git.openjdk.org/valhalla/pull/788/files/55e570ee..8074b30a

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=788&range=03
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=788&range=02-03

  Stats: 5 lines in 2 files changed: 4 ins; 0 del; 1 mod
  Patch: https://git.openjdk.org/valhalla/pull/788.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/788/head:pull/788

PR: https://git.openjdk.org/valhalla/pull/788


From dsimms at openjdk.org  Thu Oct 20 08:31:33 2022
From: dsimms at openjdk.org (David Simms)
Date: Thu, 20 Oct 2022 08:31:33 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v3]
In-Reply-To: <7SxJvTfFdXqO1RL6i-YTMft7MfCg2yC1S-EzRsoDJuw=.15eb8526-99d8-40cd-b6fd-b68446b9ab8b@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
 <HKuFatuoSGqJg_hi8Q9jxfr21Vhh-Ot-d_1pRw9fiNs=.dc6112b3-0a1b-412d-9f1f-0a88b9ae479f@github.com>
 <7SxJvTfFdXqO1RL6i-YTMft7MfCg2yC1S-EzRsoDJuw=.15eb8526-99d8-40cd-b6fd-b68446b9ab8b@github.com>
Message-ID: <4AA_MrTx-TfXieah5dsDtvdE_nszfkhaxCZO9pExfkU=.a9f9f911-0828-41a3-a6b7-fab51355915e@github.com>

On Wed, 19 Oct 2022 18:37:42 GMT, Frederic Parain <fparain at openjdk.org> wrote:

>> David Simms has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   Update test/jdk/valhalla/valuetypes/StaticFactoryMethodHandleTest.java
>>   
>>   Co-authored-by: Mandy Chung <mandy.chung at oracle.com>
>
> src/hotspot/share/oops/method.cpp line 945:
> 
>> 943: // A method named <init>, if non-static, is a classic object constructor.
>> 944: bool Method::is_object_constructor() const {
>> 945:   return name() == vmSymbols::object_initializer_name() && !is_static();
> 
> Do we still need to check that the method is not static?
> Now that constructors and static factories have different names, the check of the static flag looks superfluous (applies to the method below as well, and to the comments above the method declaration).

You are correct, any kind of correctness check has already happened during parsing

-------------

PR: https://git.openjdk.org/valhalla/pull/788


From dsimms at openjdk.org  Thu Oct 20 08:41:33 2022
From: dsimms at openjdk.org (David Simms)
Date: Thu, 20 Oct 2022 08:41:33 GMT
Subject: [lworld] RFR: 8247569: [lworld] Align with spec for <vnew> methods
 [v5]
In-Reply-To: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
Message-ID: <y2hGrGqqakqiDBZA0MqXGNJOnrvbxQxTLUyBUBt6s68=.808aca0f-3c15-4714-be5f-a9e53e655f21@github.com>

> Merged javac and VM work together, includes VM support (JDK-8294224)

David Simms has updated the pull request incrementally with two additional commits since the last revision:

 - github "commit comment" seems to be capable of adding trailing whitespace, fixed
 - Applied fparain comments

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/788/files
  - new: https://git.openjdk.org/valhalla/pull/788/files/8074b30a..2acc29c9

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=788&range=04
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=788&range=03-04

  Stats: 5 lines in 2 files changed: 0 ins; 0 del; 5 mod
  Patch: https://git.openjdk.org/valhalla/pull/788.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/788/head:pull/788

PR: https://git.openjdk.org/valhalla/pull/788


From dsimms at openjdk.org  Thu Oct 20 08:45:04 2022
From: dsimms at openjdk.org (David Simms)
Date: Thu, 20 Oct 2022 08:45:04 GMT
Subject: [lworld] Integrated: 8247569: [lworld] Align with spec for <vnew>
 methods
In-Reply-To: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
References: <oIEdYL7FIGKiJ6up6b0Uss-h81bS4EHiIMDng0K2Ac0=.b312af46-dade-44cc-9640-5b8077eb3005@github.com>
Message-ID: <b7WgPemKPmZzexkl3lk29sb4sUwzT5QyPdfn84-CoE0=.ddd03fc5-f64c-4262-9d15-f669839d9447@github.com>

On Wed, 12 Oct 2022 08:44:41 GMT, David Simms <dsimms at openjdk.org> wrote:

> Merged javac and VM work together, includes VM support (JDK-8294224)

This pull request has now been integrated.

Changeset: cc629c95
Author:    David Simms <dsimms at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/cc629c957ad2c59c539b8fd537a56033febb7d12
Stats:     734 lines in 82 files changed: 323 ins; 103 del; 308 mod

8247569: [lworld] Align with spec for <vnew> methods
8294224: [lworld] VM support for <vnew> factory methods

Co-authored-by: Vicente Romero <vromero at openjdk.org>
Co-authored-by: Harold Seigel <hseigel at openjdk.org>
Co-authored-by: Mandy Chung <mchung at openjdk.org>
Reviewed-by: fparain, mchung

-------------

PR: https://git.openjdk.org/valhalla/pull/788


From thartmann at openjdk.org  Thu Oct 20 12:28:11 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Thu, 20 Oct 2022 12:28:11 GMT
Subject: [lworld] RFR: 8295731: [lworld] C1's linear register allocator spill
 map is incorrect
Message-ID: <CL-yXSa0Bq6XiWh2hUwMdyz-5K4zN_co-nFQw8P4lME=.7513fef4-b55c-4fea-9e45-6cb34ece0820@github.com>

We spuriously assert on AArch64 because the spill location on stack for an 8-byte value is not properly aligned. After some hours of debugging and staring at the Valhalla specific changes, I finally noticed that when moving around the `T_VALUE`/`T_INLINE_TYPE`/`T_PRIMITIVE_OBJECT` basic type definitions, we forgot to adjust the spill map of C1's register allocator accordingly.

Thanks,
Tobias

-------------

Commit messages:
 - 8295731: [lworld] C1's linear register allocator spill map is incorrect

Changes: https://git.openjdk.org/valhalla/pull/797/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=797&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8295731
  Stats: 4 lines in 2 files changed: 0 ins; 0 del; 4 mod
  Patch: https://git.openjdk.org/valhalla/pull/797.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/797/head:pull/797

PR: https://git.openjdk.org/valhalla/pull/797


From thartmann at openjdk.org  Thu Oct 20 14:11:29 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Thu, 20 Oct 2022 14:11:29 GMT
Subject: [lworld] Integrated: 8295731: [lworld] C1's linear register allocator
 spill map is incorrect
In-Reply-To: <CL-yXSa0Bq6XiWh2hUwMdyz-5K4zN_co-nFQw8P4lME=.7513fef4-b55c-4fea-9e45-6cb34ece0820@github.com>
References: <CL-yXSa0Bq6XiWh2hUwMdyz-5K4zN_co-nFQw8P4lME=.7513fef4-b55c-4fea-9e45-6cb34ece0820@github.com>
Message-ID: <Bi2TljNzXp8tw40VG2Un7euRdurgxTuj-OqL4FVzDx0=.3a14f689-4dd6-447b-bbc3-68d2ffeb2bd9@github.com>

On Thu, 20 Oct 2022 12:21:05 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

> We spuriously assert on AArch64 because the spill location on stack for an 8-byte value is not properly aligned. After some hours of debugging and staring at the Valhalla specific changes, I finally noticed that when moving around the `T_VALUE`/`T_INLINE_TYPE`/`T_PRIMITIVE_OBJECT` basic type definitions, we forgot to adjust the spill map of C1's register allocator accordingly.
> 
> Thanks,
> Tobias

This pull request has now been integrated.

Changeset: 4e59de18
Author:    Tobias Hartmann <thartmann at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/4e59de18738484a0258f45cd6a49b029f352b3a2
Stats:     4 lines in 2 files changed: 0 ins; 0 del; 4 mod

8295731: [lworld] C1's linear register allocator spill map is incorrect

-------------

PR: https://git.openjdk.org/valhalla/pull/797


From vromero at openjdk.org  Thu Oct 20 20:15:41 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Thu, 20 Oct 2022 20:15:41 GMT
Subject: [lworld] RFR: 8295086: [lworld] follow-up for JDK-8288131 [v3]
In-Reply-To: <uwWHqZsU469Jj2mgICVf3XvtCROEzeL31ZM30N_rknA=.29be9b99-caeb-4d8d-8409-f83706e1f04a@github.com>
References: <uwWHqZsU469Jj2mgICVf3XvtCROEzeL31ZM30N_rknA=.29be9b99-caeb-4d8d-8409-f83706e1f04a@github.com>
Message-ID: <N5CfgPPWhu3d_lKssKED394Kj7BtYKFyCd6VKQyNxF0=.20851a23-7db2-438a-963a-0973c27890d1@github.com>

> code related to primitive classes needs to be hidden behind the -XDenablePrimitiveClasses flag, this work is a follow-up of JDK-8288131. Hiding this remaining code behind the flag is not absolutely necessary as it probably won't be executed anyways but it will make easier to located, and extract if needed, the code related with primitive classes only

Vicente Romero has updated the pull request incrementally with one additional commit since the last revision:

  additional changes

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/793/files
  - new: https://git.openjdk.org/valhalla/pull/793/files/115b7c18..529d0e01

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=793&range=02
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=793&range=01-02

  Stats: 97 lines in 11 files changed: 22 ins; 23 del; 52 mod
  Patch: https://git.openjdk.org/valhalla/pull/793.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/793/head:pull/793

PR: https://git.openjdk.org/valhalla/pull/793


From vromero at openjdk.org  Thu Oct 20 20:19:56 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Thu, 20 Oct 2022 20:19:56 GMT
Subject: [lworld] RFR: 8295086: [lworld] follow-up for JDK-8288131 [v4]
In-Reply-To: <uwWHqZsU469Jj2mgICVf3XvtCROEzeL31ZM30N_rknA=.29be9b99-caeb-4d8d-8409-f83706e1f04a@github.com>
References: <uwWHqZsU469Jj2mgICVf3XvtCROEzeL31ZM30N_rknA=.29be9b99-caeb-4d8d-8409-f83706e1f04a@github.com>
Message-ID: <X7z6VgIch9zhBKCqw-vMc00uJF63f9gxudE1z0Y3dQc=.136a8b8c-fab5-4add-aafb-aaf79469c917@github.com>

> code related to primitive classes needs to be hidden behind the -XDenablePrimitiveClasses flag, this work is a follow-up of JDK-8288131. Hiding this remaining code behind the flag is not absolutely necessary as it probably won't be executed anyways but it will make easier to located, and extract if needed, the code related with primitive classes only

Vicente Romero has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains five commits:

 - Merge branch 'lworld' into JDK-8295086
 - additional changes
 - following on with the changes after merge
 - Merge branch 'lworld' into JDK-8295086
 - 8295086: [lworld] follow-up for JDK-8288131

-------------

Changes: https://git.openjdk.org/valhalla/pull/793/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=793&range=03
  Stats: 187 lines in 14 files changed: 63 ins; 26 del; 98 mod
  Patch: https://git.openjdk.org/valhalla/pull/793.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/793/head:pull/793

PR: https://git.openjdk.org/valhalla/pull/793


From mchung at openjdk.org  Thu Oct 20 22:50:14 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Thu, 20 Oct 2022 22:50:14 GMT
Subject: [lworld] RFR: 8224057: [lworld] a hidden value class can't have <vnew>
 static factory method
Message-ID: <z0vXbNFDYkYNrFDAJOY8ImGCVnBTo9yuLawXlzhQhrU=.e5f3068b-8340-4d0e-9cbd-91226d969bd2@github.com>

A hidden class has no name.  Rather than declaring a "<vnew>" method, a hidden value class will have to provide a different mechanism to create value instances.

This PR adds a test to verify that.

-------------

Commit messages:
 - 8224057: [lworld] a hidden value class can't have <vnew> static factory method

Changes: https://git.openjdk.org/valhalla/pull/798/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=798&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8224057
  Stats: 221 lines in 3 files changed: 219 ins; 1 del; 1 mod
  Patch: https://git.openjdk.org/valhalla/pull/798.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/798/head:pull/798

PR: https://git.openjdk.org/valhalla/pull/798


From thartmann at openjdk.org  Fri Oct 21 06:03:03 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Fri, 21 Oct 2022 06:03:03 GMT
Subject: [lworld] RFR: 8295735: [lworld] Missing membar after buffer
 initialization in method handle return
Message-ID: <c0-h_aObmXYALjbftF6XDZaGpZnAzjGqtaG6nlloxCg=.9f9eec23-97b4-448c-a25a-2357ae900994@github.com>

A missing membar after buffer initialization in C2's handling of scalarized returns from method handle intrinsics leads to incorrect execution and crashes with compiler/valhalla/inlinetypes/TestBufferTearing.java on aarch64.

Thanks,
Tobias

-------------

Commit messages:
 - Undo incorrect refactoring
 - 8295735: [lworld] Missing membar after buffer initialization in method handle return

Changes: https://git.openjdk.org/valhalla/pull/799/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=799&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8295735
  Stats: 11 lines in 1 file changed: 11 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/valhalla/pull/799.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/799/head:pull/799

PR: https://git.openjdk.org/valhalla/pull/799


From thartmann at openjdk.org  Fri Oct 21 07:15:59 2022
From: thartmann at openjdk.org (Tobias Hartmann)
Date: Fri, 21 Oct 2022 07:15:59 GMT
Subject: [lworld] Integrated: 8295735: [lworld] Missing membar after buffer
 initialization in method handle return
In-Reply-To: <c0-h_aObmXYALjbftF6XDZaGpZnAzjGqtaG6nlloxCg=.9f9eec23-97b4-448c-a25a-2357ae900994@github.com>
References: <c0-h_aObmXYALjbftF6XDZaGpZnAzjGqtaG6nlloxCg=.9f9eec23-97b4-448c-a25a-2357ae900994@github.com>
Message-ID: <2wUIvOznJMyU0efvW2TfZlRULzY0WnQ5g8u0usddf-o=.801db49e-f3e6-46f9-92c9-3a6984a922f8@github.com>

On Fri, 21 Oct 2022 05:55:43 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

> A missing membar after buffer initialization in C2's handling of scalarized returns from method handle intrinsics leads to incorrect execution and crashes with compiler/valhalla/inlinetypes/TestBufferTearing.java on aarch64.
> 
> Thanks,
> Tobias

This pull request has now been integrated.

Changeset: e350088b
Author:    Tobias Hartmann <thartmann at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/e350088b00de212024094c77683a11ac6a53c5a4
Stats:     11 lines in 1 file changed: 11 ins; 0 del; 0 mod

8295735: [lworld] Missing membar after buffer initialization in method handle return

-------------

PR: https://git.openjdk.org/valhalla/pull/799


From mchung at openjdk.org  Fri Oct 21 16:05:14 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Fri, 21 Oct 2022 16:05:14 GMT
Subject: [lworld] Integrated: 8224057: [lworld] a hidden value class can't have
 <vnew> static factory method
In-Reply-To: <z0vXbNFDYkYNrFDAJOY8ImGCVnBTo9yuLawXlzhQhrU=.e5f3068b-8340-4d0e-9cbd-91226d969bd2@github.com>
References: <z0vXbNFDYkYNrFDAJOY8ImGCVnBTo9yuLawXlzhQhrU=.e5f3068b-8340-4d0e-9cbd-91226d969bd2@github.com>
Message-ID: <iZ1TmT1FrQTeXQZ4FoA3mAX4xXET-gltp5lE_FOcUIE=.acbbfbe4-86d6-4833-a8da-51d1f7ddf3ec@github.com>

On Thu, 20 Oct 2022 22:41:53 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> A hidden class has no name.  Rather than declaring a "<vnew>" method, a hidden value class will have to provide a different mechanism to create value instances.
> 
> This PR adds a test to verify that.

This pull request has now been integrated.

Changeset: 3aa9d593
Author:    Mandy Chung <mchung at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/3aa9d593ccc361290b8592e7d487db706c29384b
Stats:     221 lines in 3 files changed: 219 ins; 1 del; 1 mod

8224057: [lworld] a hidden value class can't have <vnew> static factory method

-------------

PR: https://git.openjdk.org/valhalla/pull/798


From vromero at openjdk.org  Sat Oct 22 19:12:35 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Sat, 22 Oct 2022 19:12:35 GMT
Subject: [lworld] RFR: 8295086: [lworld] follow-up for JDK-8288131 [v5]
In-Reply-To: <uwWHqZsU469Jj2mgICVf3XvtCROEzeL31ZM30N_rknA=.29be9b99-caeb-4d8d-8409-f83706e1f04a@github.com>
References: <uwWHqZsU469Jj2mgICVf3XvtCROEzeL31ZM30N_rknA=.29be9b99-caeb-4d8d-8409-f83706e1f04a@github.com>
Message-ID: <C7T2pbJCsgSdDQH5DezpoIW6Aw-AnSxm2WdomxlTv9Y=.ef91db59-2676-4ff1-be39-e03724a4d6f3@github.com>

> code related to primitive classes needs to be hidden behind the -XDenablePrimitiveClasses flag, this work is a follow-up of JDK-8288131. Hiding this remaining code behind the flag is not absolutely necessary as it probably won't be executed anyways but it will make easier to located, and extract if needed, the code related with primitive classes only

Vicente Romero has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains seven commits:

 - Merge branch 'lworld' into JDK-8295086
 - more changes
 - Merge branch 'lworld' into JDK-8295086
 - additional changes
 - following on with the changes after merge
 - Merge branch 'lworld' into JDK-8295086
 - 8295086: [lworld] follow-up for JDK-8288131

-------------

Changes: https://git.openjdk.org/valhalla/pull/793/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=793&range=04
  Stats: 342 lines in 21 files changed: 62 ins; 176 del; 104 mod
  Patch: https://git.openjdk.org/valhalla/pull/793.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/793/head:pull/793

PR: https://git.openjdk.org/valhalla/pull/793


From dsimms at openjdk.org  Mon Oct 24 09:09:43 2022
From: dsimms at openjdk.org (David Simms)
Date: Mon, 24 Oct 2022 09:09:43 GMT
Subject: [lworld] RFR: Merge jdk
Message-ID: <pX2HAXl_HdTxFpeV_J17HA7iSbvdLUDDpl3TrsMB2L0=.90918b73-f9b4-45c6-bb10-ad45cd3b4006@github.com>

Merge tag 'jdk-20+20' into lworld_merge_jdk_20_20

-------------

Commit messages:
 - ZObjArrayAllocator::initialize from 8294900 used default prototype markWord
 - Merge tag 'jdk-20+20' into lworld_merge_jdk_20_20
 - 8294467: Fix sequence-point warnings in Hotspot
 - 8294468: Fix char-subscripts warnings in Hotspot
 - 8295662: jdk/incubator/vector tests fail "assert(VM_Version::supports_avx512vlbw()) failed"
 - 8295668: validate-source failure after JDK-8290011
 - 8295372: CompactNumberFormat handling of number one with decimal part
 - 8295456: (ch) sun.nio.ch.Util::checkBufferPositionAligned gives misleading/incorrect error
 - 8290011: IGV: Remove dead code and cleanup
 - 8290368: Introduce LDAP and RMI protocol-specific object factory filters to JNDI implementation
 - ... and 190 more: https://git.openjdk.org/valhalla/compare/3aa9d593...453a7b45

The webrevs contain the adjustments done while merging with regards to each parent branch:
 - lworld: https://webrevs.openjdk.org/?repo=valhalla&pr=800&range=00.0
 - jdk: https://webrevs.openjdk.org/?repo=valhalla&pr=800&range=00.1

Changes: https://git.openjdk.org/valhalla/pull/800/files
  Stats: 31259 lines in 1203 files changed: 18280 ins; 8098 del; 4881 mod
  Patch: https://git.openjdk.org/valhalla/pull/800.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/800/head:pull/800

PR: https://git.openjdk.org/valhalla/pull/800


From dsimms at openjdk.org  Mon Oct 24 09:22:04 2022
From: dsimms at openjdk.org (David Simms)
Date: Mon, 24 Oct 2022 09:22:04 GMT
Subject: [lworld] RFR: Merge jdk [v2]
In-Reply-To: <pX2HAXl_HdTxFpeV_J17HA7iSbvdLUDDpl3TrsMB2L0=.90918b73-f9b4-45c6-bb10-ad45cd3b4006@github.com>
References: <pX2HAXl_HdTxFpeV_J17HA7iSbvdLUDDpl3TrsMB2L0=.90918b73-f9b4-45c6-bb10-ad45cd3b4006@github.com>
Message-ID: <WNBKyK0yFYrb01_UcnvH4Dx016KwLR50OvcA-8el1rQ=.b3470e3c-d542-4839-8214-d7e5c71b75d4@github.com>

> Merge tag 'jdk-20+20' into lworld_merge_jdk_20_20

David Simms has updated the pull request incrementally with one additional commit since the last revision:

  Compiler warning

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/800/files
  - new: https://git.openjdk.org/valhalla/pull/800/files/453a7b45..e44f8550

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=800&range=01
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=800&range=00-01

  Stats: 9 lines in 1 file changed: 1 ins; 5 del; 3 mod
  Patch: https://git.openjdk.org/valhalla/pull/800.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/800/head:pull/800

PR: https://git.openjdk.org/valhalla/pull/800


From dsimms at openjdk.org  Mon Oct 24 10:22:19 2022
From: dsimms at openjdk.org (David Simms)
Date: Mon, 24 Oct 2022 10:22:19 GMT
Subject: [lworld] RFR: Merge jdk [v3]
In-Reply-To: <pX2HAXl_HdTxFpeV_J17HA7iSbvdLUDDpl3TrsMB2L0=.90918b73-f9b4-45c6-bb10-ad45cd3b4006@github.com>
References: <pX2HAXl_HdTxFpeV_J17HA7iSbvdLUDDpl3TrsMB2L0=.90918b73-f9b4-45c6-bb10-ad45cd3b4006@github.com>
Message-ID: <bzoX5tB34T_9fEPMvrBFgCoUjzJVIA1fXuUHUdklTMU=.e6496949-f73a-4eff-891d-9de39c5ca33a@github.com>

> Merge tag 'jdk-20+20' into lworld_merge_jdk_20_20

David Simms has updated the pull request incrementally with one additional commit since the last revision:

  Compile issue, explicit tmp regs to load_oop_heap

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/800/files
  - new: https://git.openjdk.org/valhalla/pull/800/files/e44f8550..1f0af2a2

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=800&range=02
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=800&range=01-02

  Stats: 9 lines in 3 files changed: 0 ins; 0 del; 9 mod
  Patch: https://git.openjdk.org/valhalla/pull/800.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/800/head:pull/800

PR: https://git.openjdk.org/valhalla/pull/800


From dsimms at openjdk.org  Mon Oct 24 10:35:52 2022
From: dsimms at openjdk.org (David Simms)
Date: Mon, 24 Oct 2022 10:35:52 GMT
Subject: [lworld] RFR: Merge jdk [v4]
In-Reply-To: <pX2HAXl_HdTxFpeV_J17HA7iSbvdLUDDpl3TrsMB2L0=.90918b73-f9b4-45c6-bb10-ad45cd3b4006@github.com>
References: <pX2HAXl_HdTxFpeV_J17HA7iSbvdLUDDpl3TrsMB2L0=.90918b73-f9b4-45c6-bb10-ad45cd3b4006@github.com>
Message-ID: <vhvyc0tm_BMhjbs0AqP4D-ltPpSP-7cpptkl2kEx0ws=.ff856b7e-047e-475a-a59a-b809524529a4@github.com>

> Merge tag 'jdk-20+20' into lworld_merge_jdk_20_20

David Simms has updated the pull request incrementally with one additional commit since the last revision:

  Missed adjusting new assert in merge conflict

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/800/files
  - new: https://git.openjdk.org/valhalla/pull/800/files/1f0af2a2..b272df1f

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=800&range=03
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=800&range=02-03

  Stats: 1 line in 1 file changed: 0 ins; 0 del; 1 mod
  Patch: https://git.openjdk.org/valhalla/pull/800.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/800/head:pull/800

PR: https://git.openjdk.org/valhalla/pull/800


From dsimms at openjdk.org  Mon Oct 24 13:00:54 2022
From: dsimms at openjdk.org (David Simms)
Date: Mon, 24 Oct 2022 13:00:54 GMT
Subject: [lworld] Integrated: Merge jdk
In-Reply-To: <pX2HAXl_HdTxFpeV_J17HA7iSbvdLUDDpl3TrsMB2L0=.90918b73-f9b4-45c6-bb10-ad45cd3b4006@github.com>
References: <pX2HAXl_HdTxFpeV_J17HA7iSbvdLUDDpl3TrsMB2L0=.90918b73-f9b4-45c6-bb10-ad45cd3b4006@github.com>
Message-ID: <ffVIrTp_bN_UaA5y1X1cq9vXHUrdnAizZtG2aD7-ESc=.b82f30ed-c084-4b74-a86d-ed09dc0751f2@github.com>

On Mon, 24 Oct 2022 08:59:30 GMT, David Simms <dsimms at openjdk.org> wrote:

> Merge tag 'jdk-20+20' into lworld_merge_jdk_20_20

This pull request has now been integrated.

Changeset: 3b88153c
Author:    David Simms <dsimms at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/3b88153c50f5b5670bb6bbf756031cbe98c054b8
Stats:     31277 lines in 1205 files changed: 18281 ins; 8103 del; 4893 mod

Merge jdk

Merge tag 'jdk-20+20'

-------------

PR: https://git.openjdk.org/valhalla/pull/800


From vromero at openjdk.org  Mon Oct 24 19:25:03 2022
From: vromero at openjdk.org (Vicente Romero)
Date: Mon, 24 Oct 2022 19:25:03 GMT
Subject: [lworld] Integrated: 8295086: [lworld] follow-up for JDK-8288131
In-Reply-To: <uwWHqZsU469Jj2mgICVf3XvtCROEzeL31ZM30N_rknA=.29be9b99-caeb-4d8d-8409-f83706e1f04a@github.com>
References: <uwWHqZsU469Jj2mgICVf3XvtCROEzeL31ZM30N_rknA=.29be9b99-caeb-4d8d-8409-f83706e1f04a@github.com>
Message-ID: <olpP0YqwXlnoBVfh5PufZBHQC4lhsuWV2qk8IbWdBPI=.0a2ea140-547e-40e8-89c7-c56308f6399d@github.com>

On Wed, 12 Oct 2022 19:35:56 GMT, Vicente Romero <vromero at openjdk.org> wrote:

> code related to primitive classes needs to be hidden behind the -XDenablePrimitiveClasses flag, this work is a follow-up of JDK-8288131. Hiding this remaining code behind the flag is not absolutely necessary as it probably won't be executed anyways but it will make easier to located, and extract if needed, the code related with primitive classes only.
> 
> The initial goal for this one was expanded and I also cleaned up the code, reduced gratuitous differences with master branch, and synchronized the code better with the current spec.

This pull request has now been integrated.

Changeset: eef5d0bd
Author:    Vicente Romero <vromero at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/eef5d0bd79aa97fdb515b5688ba3717ede3a435b
Stats:     342 lines in 21 files changed: 62 ins; 176 del; 104 mod

8295086: [lworld] follow-up for JDK-8288131

-------------

PR: https://git.openjdk.org/valhalla/pull/793


From dsimms at openjdk.org  Thu Oct 27 09:02:12 2022
From: dsimms at openjdk.org (David Simms)
Date: Thu, 27 Oct 2022 09:02:12 GMT
Subject: [lworld] RFR: Merge jdk
Message-ID: <mSwJmPR6pY2EQm8bFZda8Ue2Vsc_NYkY4u9B6lYR_to=.bcca39c9-46f5-4d18-9861-7983e572c326@github.com>

Merge tag 'jdk-20+21' into lworld_merge_jdk_20_21

# Conflicts:
#	src/hotspot/share/oops/constantPool.cpp
#	src/hotspot/share/oops/method.cpp
#	src/hotspot/share/oops/method.hpp
#	src/hotspot/share/runtime/sharedRuntime.cpp
#	src/jdk.compiler/share/classes/com/sun/tools/javac/comp/Lower.java
#	src/jdk.compiler/share/classes/com/sun/tools/javac/comp/Resolve.java

-------------

Commit messages:
 - Merge tag 'jdk-20+21' into lworld_merge_jdk_20_21
 - 8295823: Use enhanced-for cycle instead of Enumeration in java.naming
 - 8293093: NPE in P11KeyStore.getID
 - 8265441: IGV: select block nodes by clicking on it
 - 8295879: JFR: Remove warnings in the jdk.jfr module
 - 8295402: G1: Tighten assertion in G1ScanHRForRegionClosure::scan_heap_roots
 - 8294752: G1: Remove redundant checks in check_obj_during_refinement
 - 8295889: NMT preinit code does not handle allocation errors
 - 8295868: 32-bit Windows build failures after JDK-8294466
 - 8295885: GHA: Bump gcc versions
 - ... and 81 more: https://git.openjdk.org/valhalla/compare/eef5d0bd...ef135072

The webrevs contain the adjustments done while merging with regards to each parent branch:
 - lworld: https://webrevs.openjdk.org/?repo=valhalla&pr=801&range=00.0
 - jdk: https://webrevs.openjdk.org/?repo=valhalla&pr=801&range=00.1

Changes: https://git.openjdk.org/valhalla/pull/801/files
  Stats: 144412 lines in 1016 files changed: 83807 ins; 29949 del; 30656 mod
  Patch: https://git.openjdk.org/valhalla/pull/801.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/801/head:pull/801

PR: https://git.openjdk.org/valhalla/pull/801


From dsimms at openjdk.org  Thu Oct 27 11:28:29 2022
From: dsimms at openjdk.org (David Simms)
Date: Thu, 27 Oct 2022 11:28:29 GMT
Subject: [lworld] RFR: Merge jdk [v2]
In-Reply-To: <mSwJmPR6pY2EQm8bFZda8Ue2Vsc_NYkY4u9B6lYR_to=.bcca39c9-46f5-4d18-9861-7983e572c326@github.com>
References: <mSwJmPR6pY2EQm8bFZda8Ue2Vsc_NYkY4u9B6lYR_to=.bcca39c9-46f5-4d18-9861-7983e572c326@github.com>
Message-ID: <1A83gteqnsBtjJCZvNYp5KDFb0nZU9sF9ZBzbwcFBX8=.4bcd0862-47b4-4cc8-90d4-a051c4564588@github.com>

> Merge tag 'jdk-20+21' into lworld_merge_jdk_20_21
> 
> # Conflicts:
> #	src/hotspot/share/oops/constantPool.cpp
> #	src/hotspot/share/oops/method.cpp
> #	src/hotspot/share/oops/method.hpp
> #	src/hotspot/share/runtime/sharedRuntime.cpp
> #	src/jdk.compiler/share/classes/com/sun/tools/javac/comp/Lower.java
> #	src/jdk.compiler/share/classes/com/sun/tools/javac/comp/Resolve.java

David Simms has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 1459 commits:

 - Merge tag 'jdk-20+21' into lworld_merge_jdk_20_21
   
   Added tag jdk-20+21 for changeset 46e6aee0
   
   # Conflicts:
   #	src/hotspot/share/oops/constantPool.cpp
   #	src/hotspot/share/oops/method.cpp
   #	src/hotspot/share/oops/method.hpp
   #	src/hotspot/share/runtime/sharedRuntime.cpp
   #	src/jdk.compiler/share/classes/com/sun/tools/javac/comp/Lower.java
   #	src/jdk.compiler/share/classes/com/sun/tools/javac/comp/Resolve.java
 - 8295086: [lworld] follow-up for JDK-8288131
 - Merge jdk
   
   Merge tag 'jdk-20+20'
 - 8224057: [lworld] a hidden value class can't have <vnew> static factory method
 - 8295735: [lworld] Missing membar after buffer initialization in method handle return
 - 8295731: [lworld] C1's linear register allocator spill map is incorrect
 - 8247569: [lworld] Align with spec for <vnew> methods
   8294224: [lworld] VM support for <vnew> factory methods
   
   Co-authored-by: Vicente Romero <vromero at openjdk.org>
   Co-authored-by: Harold Seigel <hseigel at openjdk.org>
   Co-authored-by: Mandy Chung <mchung at openjdk.org>
   Reviewed-by: fparain, mchung
 - 8295655: [lworld] AArch64: Deoptimization from return value stub fails
 - 8295455: [lworld] find_callee_info_helper hits "must have non-abstract attached method" assert
 - 8281295: Prototype serialization/deserialization using Unsafe.
   
   Reviewed-by: mchung
 - ... and 1449 more: https://git.openjdk.org/valhalla/compare/46e6aee0...ef135072

-------------

Changes: https://git.openjdk.org/valhalla/pull/801/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=801&range=01
  Stats: 195499 lines in 1605 files changed: 186587 ins; 2175 del; 6737 mod
  Patch: https://git.openjdk.org/valhalla/pull/801.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/801/head:pull/801

PR: https://git.openjdk.org/valhalla/pull/801


From dsimms at openjdk.org  Thu Oct 27 11:28:31 2022
From: dsimms at openjdk.org (David Simms)
Date: Thu, 27 Oct 2022 11:28:31 GMT
Subject: [lworld] Integrated: Merge jdk
In-Reply-To: <mSwJmPR6pY2EQm8bFZda8Ue2Vsc_NYkY4u9B6lYR_to=.bcca39c9-46f5-4d18-9861-7983e572c326@github.com>
References: <mSwJmPR6pY2EQm8bFZda8Ue2Vsc_NYkY4u9B6lYR_to=.bcca39c9-46f5-4d18-9861-7983e572c326@github.com>
Message-ID: <vGvdSiE0eaf1WurNy8oRIqd12N4GMq2wNtoYnIR2hug=.8d2327eb-460e-4f9f-b56d-0783c16ce8f9@github.com>

On Thu, 27 Oct 2022 08:52:53 GMT, David Simms <dsimms at openjdk.org> wrote:

> Merge tag 'jdk-20+21' into lworld_merge_jdk_20_21
> 
> # Conflicts:
> #	src/hotspot/share/oops/constantPool.cpp
> #	src/hotspot/share/oops/method.cpp
> #	src/hotspot/share/oops/method.hpp
> #	src/hotspot/share/runtime/sharedRuntime.cpp
> #	src/jdk.compiler/share/classes/com/sun/tools/javac/comp/Lower.java
> #	src/jdk.compiler/share/classes/com/sun/tools/javac/comp/Resolve.java

This pull request has now been integrated.

Changeset: 21973951
Author:    David Simms <dsimms at openjdk.org>
URL:       https://git.openjdk.org/valhalla/commit/2197395184149a900efa7c95f9bb9cfc926e29d0
Stats:     144412 lines in 1016 files changed: 83807 ins; 29949 del; 30656 mod

Merge jdk

Merge tag 'jdk-20+21'

-------------

PR: https://git.openjdk.org/valhalla/pull/801


From mchung at openjdk.org  Fri Oct 28 17:22:08 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Fri, 28 Oct 2022 17:22:08 GMT
Subject: [lworld] RFR: JDK-8293321: [lworld] substitutability test shall
 support value class with cyclic membership
Message-ID: <RriOj-D3CfiAqQyRPBFO3BmUisDkNsGGOeRAiwHkhVk=.3376d401-14cf-4826-a316-bda56a4349de@github.com>

L-type allows cycles. `acmp` shall support value classes with cyclic membership.  This PR adds the support in the Java implementation of `acmp` (i.e. `ValueObjectMethods::isSubstitutable`). 

This patch includes [John's first cut to make a Y combinator method handle](https://github.com/openjdk/jdk/pull/10155) to enable a method handle to call itself recursively.    PR 10155 is adequate for internal use for now.  A separate RFE will follow to define such an API.

A separate issue [JDK-8296056](https://bugs.openjdk.org/browse/JDK-8296056) tracks the C2 `acmp` support.

-------------

Commit messages:
 - add @ignore test for -Xcomp
 - minor cleanup
 - Support recursive data type for value class
 - Merge branch 'lworld' of https://github.com/openjdk/valhalla into recursive-mh
 - prototype

Changes: https://git.openjdk.org/valhalla/pull/802/files
 Webrev: https://webrevs.openjdk.org/?repo=valhalla&pr=802&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8293321
  Stats: 639 lines in 3 files changed: 623 ins; 7 del; 9 mod
  Patch: https://git.openjdk.org/valhalla/pull/802.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/802/head:pull/802

PR: https://git.openjdk.org/valhalla/pull/802


From jrose at openjdk.org  Fri Oct 28 19:33:03 2022
From: jrose at openjdk.org (John R Rose)
Date: Fri, 28 Oct 2022 19:33:03 GMT
Subject: [lworld] RFR: JDK-8293321: [lworld] substitutability test shall
 support value class with cyclic membership
In-Reply-To: <RriOj-D3CfiAqQyRPBFO3BmUisDkNsGGOeRAiwHkhVk=.3376d401-14cf-4826-a316-bda56a4349de@github.com>
References: <RriOj-D3CfiAqQyRPBFO3BmUisDkNsGGOeRAiwHkhVk=.3376d401-14cf-4826-a316-bda56a4349de@github.com>
Message-ID: <Gkt2Kp6qMMwdVFS_pi7UZe4crvfkslxqDIFHBvlGAgA=.5df39110-da78-452f-9280-9116b3741a17@github.com>

On Fri, 28 Oct 2022 17:15:19 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> L-type allows cycles. `acmp` shall support value classes with cyclic membership.  This PR adds the support in the Java implementation of `acmp` (i.e. `ValueObjectMethods::isSubstitutable`). 
> 
> This patch includes [John's first cut to make a Y combinator method handle](https://github.com/openjdk/jdk/pull/10155) to enable a method handle to call itself recursively.    PR 10155 is adequate for internal use for now.  A separate RFE will follow to define such an API.
> 
> A separate issue [JDK-8296056](https://bugs.openjdk.org/browse/JDK-8296056) tracks the C2 `acmp` support.

Two values `x`, `y` are substitutable if and only if, for every access path `x.a.b.c`, `y.a.b.c` also exists (and vice versa), and moreover the two values of those two accesses are themselves substitutable.  (The call `.getClass()` is one possible path; the rest of the paths are field references, regardless of access protection.)  This definition can be simplified by stipulating that only paths `.a.b.c` which end in a non-value object (primitive, null, id-object) are significant; the equivalence is the same.

In the case of a deeply nested (probably recursively-typed) value, the number of such paths maybe exponentially large (compared with the number of object "nodes").  So a naive exploration of all paths will be infeasible.  For example, a "ladder" of binary nodes where both sides of each "rung" points to the same "ladder" of one less depth, will have 2^D complexity of enumerating all paths.

And cyclic value graphs (which must involve recursive types) have an infinite number of paths, which is even scarier.

That said, this is a known problem with known solutions.  (It is the DFA minimization problem in disguise, aka the RE equivalence problem, and also the type structural equivalence problem which the Algol designers inflicted on themselves.) In practice, in most cases, the naive algorithm will do fine.

https://www.ics.uci.edu/~goodrich/teach/cs162/notes/rm.pdf
https://cs.uwaterloo.ca/~dberry/FTP_SITE/reprints.journals.conferences/BerrySchwartz1979TypeEquiv.pdf

As we discussed offline, I suggest using a relatively native algorithm for a start, one that doesn't "remember where it has already been", but we should run a counter (of node visits), and throw `StackOverflowError` if the counter gets "unreasonably" large.  (Regardless of how many real stack frames were consumed.)   This is allowed by the JVMS, though we frown on it.  We can refine the algorithm over time.  There are lots of ideas in our heads and in the literature for making common cases work better.  By allowing SOE, this becomes a QOS rather than a correctness issue.

We might possibly come up with a solution for the general case that we are comfortable with.  (I can imagine how one would go about it, doing on-the-fly minimization moves on dual DFAs, in the course of a depth-first traversal? but it's tricky.  And the payoff is really small.)  I would want the algorithm to withstand even determined complexity attacks, from value-object graphs crafted by Evil Masterminds.  Since that's not a game I particularly want to play for very long, I think we will also keep the `StackOverflowError` option alive even when we have "smart" algorithms.  This can be tuned by a JVM switch.

For users who take umbrage at SOE when their value graphs get really big, we will have to say, "you're holding it wrong".  No decent language can completely prevent users from harming themselves via some correct but wrong-headed Turing-footgun.  I think this is such a situation.

-------------

PR: https://git.openjdk.org/valhalla/pull/802


From mchung at openjdk.org  Fri Oct 28 21:54:27 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Fri, 28 Oct 2022 21:54:27 GMT
Subject: [lworld] RFR: JDK-8293321: [lworld] substitutability test shall
 support value class with cyclic membership [v2]
In-Reply-To: <RriOj-D3CfiAqQyRPBFO3BmUisDkNsGGOeRAiwHkhVk=.3376d401-14cf-4826-a316-bda56a4349de@github.com>
References: <RriOj-D3CfiAqQyRPBFO3BmUisDkNsGGOeRAiwHkhVk=.3376d401-14cf-4826-a316-bda56a4349de@github.com>
Message-ID: <VkIlsL52zaX6ueT-sWJjiBcv7derE8d8lR_zhsCWBKg=.7402455c-09fe-4f15-a041-2da2403cfd60@github.com>

> L-type allows cycles. `acmp` shall support value classes with cyclic membership.  This PR adds the support in the Java implementation of `acmp` (i.e. `ValueObjectMethods::isSubstitutable`). 
> 
> This patch includes [John's first cut to make a Y combinator method handle](https://github.com/openjdk/jdk/pull/10155) to enable a method handle to call itself recursively.    PR 10155 is adequate for internal use for now.  A separate RFE will follow to define such an API.
> 
> A separate issue [JDK-8296056](https://bugs.openjdk.org/browse/JDK-8296056) tracks the C2 `acmp` support.

Mandy Chung has updated the pull request incrementally with two additional commits since the last revision:

 - Add a test case of large graph throwing SOE
 - Add a node-visit counter and throw SOE if it exceeds the threshold

-------------

Changes:
  - all: https://git.openjdk.org/valhalla/pull/802/files
  - new: https://git.openjdk.org/valhalla/pull/802/files/00757be8..8b1e5508

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=valhalla&pr=802&range=01
 - incr: https://webrevs.openjdk.org/?repo=valhalla&pr=802&range=00-01

  Stats: 192 lines in 2 files changed: 138 ins; 22 del; 32 mod
  Patch: https://git.openjdk.org/valhalla/pull/802.diff
  Fetch: git fetch https://git.openjdk.org/valhalla pull/802/head:pull/802

PR: https://git.openjdk.org/valhalla/pull/802


From mchung at openjdk.org  Fri Oct 28 21:55:43 2022
From: mchung at openjdk.org (Mandy Chung)
Date: Fri, 28 Oct 2022 21:55:43 GMT
Subject: [lworld] RFR: JDK-8293321: [lworld] substitutability test shall
 support value class with cyclic membership
In-Reply-To: <RriOj-D3CfiAqQyRPBFO3BmUisDkNsGGOeRAiwHkhVk=.3376d401-14cf-4826-a316-bda56a4349de@github.com>
References: <RriOj-D3CfiAqQyRPBFO3BmUisDkNsGGOeRAiwHkhVk=.3376d401-14cf-4826-a316-bda56a4349de@github.com>
Message-ID: <JLDAYrPAJ1qDdfs1Rx5Hr0zfGkCaNE01-NMzGNmMYDc=.0205f73e-1425-4cec-a7ee-2dd5029d9039@github.com>

On Fri, 28 Oct 2022 17:15:19 GMT, Mandy Chung <mchung at openjdk.org> wrote:

> L-type allows cycles. `acmp` shall support value classes with cyclic membership.  This PR adds the support in the Java implementation of `acmp` (i.e. `ValueObjectMethods::isSubstitutable`). 
> 
> This patch includes [John's first cut to make a Y combinator method handle](https://github.com/openjdk/jdk/pull/10155) to enable a method handle to call itself recursively.    PR 10155 is adequate for internal use for now.  A separate RFE will follow to define such an API.
> 
> A separate issue [JDK-8296056](https://bugs.openjdk.org/browse/JDK-8296056) tracks the C2 `acmp` support.

Added a counter and throw SOE if it exceeds the threshold for this initial implementation.

-------------

PR: https://git.openjdk.org/valhalla/pull/802