8287917: System.loadLibrary does not work on Big Sur if JDK is built with macOS SDK 10.15 and earlier

[jmtd]

This pull request contains a backport of commit fe807217 from the openjdk/jdk repository.

The commit being backported was authored by Yoshiki Sato on 15 Jun 2022 and was reviewed by Mandy Chung.

I wish to backport it for Oracle parity.

It does not apply cleanly due to a renamed file in jdk master subsequent to 11u:

src/java.base/macosx/classes/{java/lang => jdk/internal/loader}/ClassLoaderHelper.java

Manual testing results forthcoming; I'll add them as a comment on this PR.

---

Progress

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

Issue

• JDK-8287917: System.loadLibrary does not work on Big Sur if JDK is built with macOS SDK 10.15 and earlier

Reviewers

• Christoph Langer (@RealCLanger - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk11u-dev pull/1298/head:pull/1298
$ git checkout pull/1298

Update a local copy of the PR:
$ git checkout pull/1298
$ git pull https://git.openjdk.org/jdk11u-dev pull/1298/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 1298

View PR using the GUI difftool:
$ git pr show -t 1298

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk11u-dev/pull/1298.diff

[bridgekeeper]

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

[openjdk]

This backport pull request has now been updated with issue from the original commit.

[jmtd]

Test passes after manually building the native component (as with the 17u backport):

 $  pushd test/jdk/java/lang/RuntimeTests/loadLibrary/exeLibraryCache
 $  cc  exeLibraryCache.c -o LibraryCache
 $  popd
 $  $JT_HOME/bin/jtreg -jdk:build/macosx-x86_64-server-release/images/jdk  \
 -nativepath:test/jdk/java/lang/RuntimeTests/loadLibrary/exeLibraryCache  \
test/jdk/java/lang/RuntimeTests/loadLibrary/exeLibraryCache/LibraryFromCache.java

[mlbridge]

Webrevs

• 00: Full (618c6f95)

[bridgekeeper]

@jmtd This pull request has been inactive for more than 4 weeks and will be automatically closed if another 4 weeks passes without any activity. To avoid this, simply add a new comment to the pull request. Feel free to ask for assistance if you need help with progressing this pull request towards integration!

[jmtd]

Ping comment

This requires a review. It has been some time since I did it, so I'll check it's still clean Tomorrow.

[RealCLanger]

Hi @jmtd, sorry, we kind of overlooked this obviously. As Oracle has backported this to 11.0.17, would you mind rebasing this PR to jdk11u? I can then review, test & approve it. Thanks!

[jmtd]

Hi @jmtd, sorry, we kind of overlooked this obviously. As Oracle has backported this to 11.0.17, would you mind rebasing this PR to jdk11u? I can then review, test & approve it. Thanks!

Sure, I'm happy to do that. It will mean opening a new PR since it's a different GitHub repository. That's not a problem.

Before I do that I must first resolve a testing issue locally. The test is being skipped on my machine due to "Test skipped. Library BLAS not found"; with both jdk11u and jdk11u-dev, pre and post this patch. I'll figure that out first.

[jmtd]

@RealCLanger, the patch (after resolving the original conflict, i.e., the commit as in the PR today) applies clean to jdk11u and the test passes.

For belt and braces, I really wanted to try and verify things by proving the test fail prior to applying the patch; unfortunately, my macOS development machine has updated itself since I did the backport:

I initially had macOS Big Sur and SDK version 10.14; at the time I did the backport with these versions I could see the test fail.

I now have SDK version MacOSX11.3.sdk. With this, the native component (exeLibraryCache.c) successfully loads the library libblas.dylib before and after the patch. I've attempted to install an older SDK but I don't know enough yet about how this works, because what I've tried (SDKROOT=/path/to/old/sdk cc ...) the resultant binary is bit-for-bit identical.

To tackle this situation for future macOS work I'm going to investigate setting up some VMs of specific macOS and XCode versions.

Side-note: The test harness runs twice, once to try and load libblas.dylib and another time to load libBLAS.dylib. The second attempt never succeeds for me in any configuration, but the consequence is a test skip, rather than a failure. I might follow this up in jdk/master (same story there)

I leave it up to you whether we should push this now or keep investigating.

Thanks

[RealCLanger]

@jmtd ok. I also don't have an easy possiblity to test on such old OS and SDK releases. I suggest, since this only affects rather old scenarios, we just take this PR into jdk11u-dev for the January update. It's not worth transferring it to jdk11u now and maybe risking something during rampdown.

I'll see if I can review it here.

[RealCLanger]

LGTM. Please flag for approval.

[openjdk]

@jmtd This change now passes all automated pre-integration checks.

After integration, the commit message for the final commit will be:

8287917: System.loadLibrary does not work on Big Sur if JDK is built with macOS SDK 10.15 and earlier

Reviewed-by: clanger

You can use pull request commands such as /summary, /contributor and /issue to adjust it as needed.

At the time when this comment was updated there had been 216 new commits pushed to the master branch:

• 0ba5a81: 8276841: Add support for Visual Studio 2022
• 131eab6: 8292541: [Metrics] Reported memory limit may exceed physical machine memory
• f75b74d: 6782021: It is not possible to read local computer certificates with the SunMSCAPI provider
• 78ba1e5: 8247676: vcruntime140_1.dll is not needed on 32-bit Windows
• 2548ed3: 8242468: VS2019 build missing vcruntime140_1.dll
• 75b3e45: 8239708: Split basics.m4 into basic.m4 and util.m4
• de5b4b4: 8217032: Check pandoc capabilities in configure
• de91323: 8292083: Detected container memory limit may exceed physical machine memory
• 6e30d22: 8292887: Bump update version for OpenJDK: jdk-11.0.18
• 2f0f34c: 8291570: [TESTBUG] Part of JDK-8250984 absent from 11u
• ... and 206 more: https://git.openjdk.org/jdk11u-dev/compare/edd32ef60b3e3c5002505c4fcf88d6c81e94d355...master

As there are no conflicts, your changes will automatically be rebased on top of these commits when integrating. If you prefer to avoid this automatic rebasing, please check the documentation for the /integrate command for further details.

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

[jmtd]

/integrate

[openjdk]

Going to push as commit 9767c4c.
Since your change was applied there have been 217 commits pushed to the master branch:

• c249ca9: 8268893: jcmd to trim the glibc heap
• 0ba5a81: 8276841: Add support for Visual Studio 2022
• 131eab6: 8292541: [Metrics] Reported memory limit may exceed physical machine memory
• f75b74d: 6782021: It is not possible to read local computer certificates with the SunMSCAPI provider
• 78ba1e5: 8247676: vcruntime140_1.dll is not needed on 32-bit Windows
• 2548ed3: 8242468: VS2019 build missing vcruntime140_1.dll
• 75b3e45: 8239708: Split basics.m4 into basic.m4 and util.m4
• de5b4b4: 8217032: Check pandoc capabilities in configure
• de91323: 8292083: Detected container memory limit may exceed physical machine memory
• 6e30d22: 8292887: Bump update version for OpenJDK: jdk-11.0.18
• ... and 207 more: https://git.openjdk.org/jdk11u-dev/compare/edd32ef60b3e3c5002505c4fcf88d6c81e94d355...master

Your commit was automatically rebased without conflicts.

[openjdk]

@jmtd Pushed as commit 9767c4c.

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