8295159: DSO created with -ffast-math breaks Java floating-point arithmetic

[bulasevich]

Hi all,

backport https://bugs.openjdk.org/browse/JDK-8295159 to fix broken Java floating-point arithmetic.

Minor merge conflicts in os_linux.cpp and os_bsd.cpp have been resolved manually. The conflict is with the NativeLibraryLoadEvent tracing applied on the mainline around the dlopen() call: openjdk/jdk#16903

The fix passed tier1-3 jtreg tests.

Thank you

---

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
• JDK-8295159 needs maintainer approval

Issue

• JDK-8295159: DSO created with -ffast-math breaks Java floating-point arithmetic (Bug - P4 - Requested)

Reviewers

• Paul Hohensee (@phohensee - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk21u-dev.git pull/1111/head:pull/1111
$ git checkout pull/1111

Update a local copy of the PR:
$ git checkout pull/1111
$ git pull https://git.openjdk.org/jdk21u-dev.git pull/1111/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 1111

View PR using the GUI difftool:
$ git pr show -t 1111

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk21u-dev/pull/1111.diff

Using Webrev

Link to Webrev Comment

[bridgekeeper]

👋 Welcome back bulasevich! 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 change is not yet ready to be integrated.
See the Progress checklist in the description for automated requirements.

[openjdk]

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

[mlbridge]

Webrevs

• 00: Full (161842e2)

[openjdk]

⚠️ @bulasevich This change is now ready for you to apply for maintainer approval. This can be done directly in each associated issue or by using the /approval command.

[bulasevich]

/approval request
I ask for backport approval for stability reasons. Wrong floating-point arithmetic can put the application in a very incorrect state. The change is simple and safe.

[openjdk]

@bulasevich
8295159: The approval request has been created successfully.

[GoeLin]

Hi @bulasevich , do we need any of the 6 related issues listed in the JBS issue?

[bulasevich]

Hi @bulasevich , do we need any of the 6 related issues listed in the JBS issue?

Hi @GoeLin

I do not think we need them:

1. JDK-6487931 JVM reports EXCEPTION_FLT_STACK_CHECK when calling a Java method through JNI
   Resolved: 2011-03-07
2. JDK-6550813 Crash because of FPU control word being modified by native code (win32)
   Resolved: 2011-04-25
3. JDK-8317810 DSO created with -ffast-math breaks Java floating-point arithmetic
   Unresolved. It is a JBS record for further similar change for x86 platform.
4. JDK-8321017 Record in JFR that IEEE rounding mode was corrupted by loading a library
   JFR improvement. I do not think it is necessary to backport.
5. JDK-8319708 Assertion 'fsetenv didn't work' in jdk tier4 tests after 8295159 on Linux aarch64 RHEL9.3
   Closed (User error)