From varadam at openjdk.org  Wed Feb  4 10:10:53 2026
From: varadam at openjdk.org (Varada M)
Date: Wed, 4 Feb 2026 10:10:53 GMT
Subject: RFR: 7904115: Fix for AIX test case failures due to incorrect
 alignment for double and pointer [v7]
In-Reply-To: <K4imBMgVlY725idEe3TU4OBPb4Aw5MaXO9dB8ekZoho=.8a496b76-2f6d-40bb-a4c6-3313e96d3cbe@github.com>
References: <K4imBMgVlY725idEe3TU4OBPb4Aw5MaXO9dB8ekZoho=.8a496b76-2f6d-40bb-a4c6-3313e96d3cbe@github.com>
Message-ID: <vgmGBx18V346ZwqVZTFIsNB0KATdG2ZSTgrtK3Q5BXs=.cc977e62-69eb-416d-a423-46b62a72cf7c@github.com>

> Total of 10 test failures observed on AIX:
> jtreg/generator/nestedTypes/TestNestedTypesUnsupported.java 
> jtreg/generator/test8246400/LibTest8246400Test.java 
> jtreg/generator/test8258605/LibTest8258605Test.java 
> jtreg/generator/test8261511/Test8261511.java 
> jtreg/generator/testStruct/LibStructTest.java 
> testng/org/openjdk/jextract/test/toolprovider/ConstantsTest.java 
> testng/org/openjdk/jextract/test/toolprovider/IncompleteArrayTest.java 
> testng/org/openjdk/jextract/test/toolprovider/Test8240811.java 
> testng/org/openjdk/jextract/test/toolprovider/TestClassGeneration.java 
> testng/org/openjdk/jextract/test/toolprovider/nestedAnonOffset/TestNestedAnonOffset.java
> 
> This PR fixes AIX specific layout generation issues related to incorrect alignment double and pointer types.
> 1. Structs containing double fields fail with:
>     i. Unsupported layout: 4%D8
>     ii. Invalid alignment constraint for member layout
>     double in AIX structs has size 8 but alignment 4 (except as first field). AIX specific handling for C_DOUBLE computes the     correct alignment.
> 
> 2. Clang was detected as 32-bit due to missing -m64 during macro extraction, causing inconsistent macros. This caused jextract to interpret pointer constants incorrectly, leading to failures like:
> expected [-1] but found [4294967295]
> 
> 3. TestNestedAnonOffset.java test failed on AIX because it also expects more padding similar to platforms like windows and linux
> 
> 
> After the patch jtreg tests passes successfully.
> 
> JBS: [CODETOOLS-7904115](https://bugs.openjdk.org/browse/CODETOOLS-7904115)

Varada M has updated the pull request incrementally with one additional commit since the last revision:

  7904115: Fix for AIX test case failures due to incorrect alignment for double and pointer

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

Changes:
  - all: https://git.openjdk.org/jextract/pull/296/files
  - new: https://git.openjdk.org/jextract/pull/296/files/a3dbfaf9..49aebd79

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=jextract&pr=296&range=06
 - incr: https://webrevs.openjdk.org/?repo=jextract&pr=296&range=05-06

  Stats: 30 lines in 7 files changed: 8 ins; 14 del; 8 mod
  Patch: https://git.openjdk.org/jextract/pull/296.diff
  Fetch: git fetch https://git.openjdk.org/jextract.git pull/296/head:pull/296

PR: https://git.openjdk.org/jextract/pull/296

From varadam at openjdk.org  Wed Feb  4 10:10:55 2026
From: varadam at openjdk.org (Varada M)
Date: Wed, 4 Feb 2026 10:10:55 GMT
Subject: RFR: 7904115: Fix for AIX test case failures due to incorrect
 alignment for double and pointer [v4]
In-Reply-To: <lqgkBy5GN_P1627T1jq3d55DMY3vTN6qncq83gfhWOk=.86edea51-5492-47aa-b569-80a7ba256117@github.com>
References: <K4imBMgVlY725idEe3TU4OBPb4Aw5MaXO9dB8ekZoho=.8a496b76-2f6d-40bb-a4c6-3313e96d3cbe@github.com>
 <Ol9Nka1U9CaClD4QNO8D-seWo7sO7_RUswvnPUl6m2k=.48a6db9b-9639-49c2-8148-593ab4b3a7b1@github.com>
 <lqgkBy5GN_P1627T1jq3d55DMY3vTN6qncq83gfhWOk=.86edea51-5492-47aa-b569-80a7ba256117@github.com>
Message-ID: <T6PnWaxQGtQONYdssSJGxspJtywk8M4e_hVa68IeaeE=.1cb4c908-5cd0-4abb-b12e-ba86444696b2@github.com>

On Fri, 12 Dec 2025 17:03:56 GMT, Jorn Vernee <jvernee at openjdk.org> wrote:

>> Varada M has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   added field alignment and natural alignment argument
>
> Ok, the current approach means that jextract would only work when the C code is using natural alignment. If you're happy with that, I can accept the patch. But, all the of the changes I suggested are not needed in that case.
> 
> I really think you should try to adjust `expectedAlign` in StructBuilder to get the right alignment (based on Maurizio's comment here: https://github.com/openjdk/jextract/pull/296#issuecomment-3564203711)

Thank you @JornVernee and @mcimadamore for the earlier clarifications.
I'm adding a consolidated explanation of the issue and current understanding.

Failure case 1 :  Invalid alignment constraint for member layout: D8(d) [in thread ?MainThread?]
For this failing scenario, [AllTypes struct](https://github.com/openjdk/jextract/blob/master/test/jtreg/generator/testStruct/struct.h#L41), double is currently being aligned to 8 bytes by taking the struct alignment. This leads to cases where offset % align becomes invalid (eg: 52 % 8 != 0), which results in the exception. The correct approach here should be to calculate and respect the field alignment.

Failure case 2 : expected [D8(d)] but found [4%D8(d)] & unsupported layout 4%D8
1)
union bar2 {
? ?float f;
? ?double d;
};

private static final GroupLayout $LAYOUT = MemoryLayout.unionLayout(
?   ? ? name_collision_h.C_FLOAT.withName("f"),
? ? ? ? MemoryLayout.valueLayout(8, java.nio.ByteOrder.nativeOrder()).withByteAlignment(4).withName("d")
? ? ).withName("bar2");

 2)

?* struct Vector {
?* ? ? double x;
?* ? ? double y;
?* } Vector;
?* 

? ? private static final GroupLayout $LAYOUT = MemoryLayout.structLayout(
? ? ? ? test8246400_h.align(test8246400_h.C_DOUBLE, 4).withName("x"),
? ? ? ? test8246400_h.align(test8246400_h.C_DOUBLE, 4).withName("y")
? ? ).withName("Vector");
? ? 
In the above cases, C_DOUBLE is aligned to 4 bytes but we can?t alter the alignment of [C_DOUBLE layout](https://github.com/openjdk/jextract/blob/master/src/main/java/org/openjdk/jextract/impl/ClassSourceBuilder.java#L262) (Thank you Jorn for clarifying that).  modifying the fundamental alignment of C_DOUBLE would be incorrect and conflicts with the ABI expectations. Therefore, the safer and correct fix would be to adjust the specific test case to be fixed not to use C_DOUBLE.

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

PR Comment: https://git.openjdk.org/jextract/pull/296#issuecomment-3846498472

From nbenalla at openjdk.org  Wed Feb  4 16:40:48 2026
From: nbenalla at openjdk.org (Nizar Benalla)
Date: Wed, 4 Feb 2026 16:40:48 GMT
Subject: RFR: 7904128: Add a struct reinterpret helper that keeps scope
Message-ID: <jUGkI333WzC8Tgk1CyF3r31GQB9kSG3TDeewfYrpB3M=.b1802c34-6712-440d-9040-58c539bf4cd9@github.com>

Please review this change to add a new reinterpret method that keeps scope, I have removed one of the old methods and only kept one with arena and cleanup, and one without.

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

Commit messages:
 - add new reinterpret method that keeps scope

Changes: https://git.openjdk.org/jextract/pull/300/files
  Webrev: https://webrevs.openjdk.org/?repo=jextract&pr=300&range=00
  Issue: https://bugs.openjdk.org/browse/CODETOOLS-7904128
  Stats: 24 lines in 3 files changed: 13 ins; 0 del; 11 mod
  Patch: https://git.openjdk.org/jextract/pull/300.diff
  Fetch: git fetch https://git.openjdk.org/jextract.git pull/300/head:pull/300

PR: https://git.openjdk.org/jextract/pull/300

From mcimadamore at openjdk.org  Wed Feb  4 16:53:09 2026
From: mcimadamore at openjdk.org (Maurizio Cimadamore)
Date: Wed, 4 Feb 2026 16:53:09 GMT
Subject: RFR: 7904128: Add a struct reinterpret helper that keeps scope
In-Reply-To: <jUGkI333WzC8Tgk1CyF3r31GQB9kSG3TDeewfYrpB3M=.b1802c34-6712-440d-9040-58c539bf4cd9@github.com>
References: <jUGkI333WzC8Tgk1CyF3r31GQB9kSG3TDeewfYrpB3M=.b1802c34-6712-440d-9040-58c539bf4cd9@github.com>
Message-ID: <iXecKFndoxsB1GzMPcyrnPJhcLPWuoY679UqgXGqGwc=.d99dc0cd-eb91-45a6-92af-4900fa151230@github.com>

On Wed, 4 Feb 2026 16:27:04 GMT, Nizar Benalla <nbenalla at openjdk.org> wrote:

> Please review this change to add a new reinterpret method that keeps scope, I have removed one of the old methods and only kept one with arena and cleanup, and one without.

In general, the change seems good. I'm a bit unsure re. compatibility impact -- the test changes shows that there's some code out there that might stop working. If we want to play extra safe we could keep all existing methods, and add a new one.

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

PR Review: https://git.openjdk.org/jextract/pull/300#pullrequestreview-3752138451

From varadam at openjdk.org  Thu Feb  5 09:39:19 2026
From: varadam at openjdk.org (Varada M)
Date: Thu, 5 Feb 2026 09:39:19 GMT
Subject: RFR: 7904115: Fix for AIX test case failures due to incorrect
 alignment for double and pointer [v8]
In-Reply-To: <K4imBMgVlY725idEe3TU4OBPb4Aw5MaXO9dB8ekZoho=.8a496b76-2f6d-40bb-a4c6-3313e96d3cbe@github.com>
References: <K4imBMgVlY725idEe3TU4OBPb4Aw5MaXO9dB8ekZoho=.8a496b76-2f6d-40bb-a4c6-3313e96d3cbe@github.com>
Message-ID: <YQaBO8CzgfAB7aS7ZMLhAH4NZxyCUh7lJtockpsr10g=.9d7e7ee9-71f3-4b65-9609-c5e5d4b78356@github.com>

> Total of 10 test failures observed on AIX:
> jtreg/generator/nestedTypes/TestNestedTypesUnsupported.java 
> jtreg/generator/test8246400/LibTest8246400Test.java 
> jtreg/generator/test8258605/LibTest8258605Test.java 
> jtreg/generator/test8261511/Test8261511.java 
> jtreg/generator/testStruct/LibStructTest.java 
> testng/org/openjdk/jextract/test/toolprovider/ConstantsTest.java 
> testng/org/openjdk/jextract/test/toolprovider/IncompleteArrayTest.java 
> testng/org/openjdk/jextract/test/toolprovider/Test8240811.java 
> testng/org/openjdk/jextract/test/toolprovider/TestClassGeneration.java 
> testng/org/openjdk/jextract/test/toolprovider/nestedAnonOffset/TestNestedAnonOffset.java
> 
> This PR fixes AIX specific layout generation issues related to incorrect alignment double and pointer types.
> 1. Structs containing double fields fail with:
>     i. Unsupported layout: 4%D8
>     ii. Invalid alignment constraint for member layout
>     double in AIX structs has size 8 but alignment 4 (except as first field). AIX specific handling for C_DOUBLE computes the     correct alignment.
> 
> 2. Clang was detected as 32-bit due to missing -m64 during macro extraction, causing inconsistent macros. This caused jextract to interpret pointer constants incorrectly, leading to failures like:
> expected [-1] but found [4294967295]
> 
> 3. TestNestedAnonOffset.java test failed on AIX because it also expects more padding similar to platforms like windows and linux
> 
> 
> After the patch jtreg tests passes successfully.
> 
> JBS: [CODETOOLS-7904115](https://bugs.openjdk.org/browse/CODETOOLS-7904115)

Varada M has updated the pull request incrementally with one additional commit since the last revision:

  Update examples.h

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

Changes:
  - all: https://git.openjdk.org/jextract/pull/296/files
  - new: https://git.openjdk.org/jextract/pull/296/files/49aebd79..9b4aec96

Webrevs:
 - full: https://webrevs.openjdk.org/?repo=jextract&pr=296&range=07
 - incr: https://webrevs.openjdk.org/?repo=jextract&pr=296&range=06-07

  Stats: 0 lines in 0 files changed: 0 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/jextract/pull/296.diff
  Fetch: git fetch https://git.openjdk.org/jextract.git pull/296/head:pull/296

PR: https://git.openjdk.org/jextract/pull/296