RFR: 8295159: DSO created with -ffast-math breaks Java floating-point arithmetic [v8]

[Thomas Stuefe]

On Wed, 11 Oct 2023 13:36:01 GMT, Andrew Haley <aph at openjdk.org> wrote:

>> A bug in GCC causes shared libraries linked with -ffast-math to disable denormal arithmetic. This breaks Java's floating-point semantics.
>> 
>> The bug is https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55522
>> 
>> One solution is to save and restore the floating-point control word around System.loadLibrary(). This isn't perfect, because some shared library might load another shared library at runtime, but it's a lot better than what we do now. 
>> 
>> However, this fix is not complete. `dlopen()` is called from many places in the JDK. I guess the best thing to do is find and wrap them all. I'd like to hear people's opinions.
>
> Andrew Haley has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 28 commits:
> 
>  - Merge branch 'JDK-8295159' of https://github.com/theRealAph/jdk into JDK-8295159
>  - Fix LLVM
>  - Give x32 bug its own ID.
>  - cleanup
>  - Fix conditional compilation
>  - Remove x32 handling
>  - Stash x86-32 changes
>  - MacOS
>  - AArch64
>  - x86-32 changes
>  - ... and 18 more: https://git.openjdk.org/jdk/compare/cef9fff0...c56adbd9

src/hotspot/os/bsd/os_bsd.cpp line 976:

> 974: // same architecture as Hotspot is running on
> 975: 
> 976: void *os::Bsd::dlopen_helper(const char *filename, int mode) {

I thought BSD is switching to clang.

src/hotspot/os/bsd/os_bsd.cpp line 1001:

> 999:     static const volatile double thresh
> 1000:       = jdouble_cast(0x0000000000000003); // 0x0.0000000000003p-1022;
> 1001:     if (unity + thresh == unity || -unity - thresh == -unity) {

Could this expression happen to be precomputed by the compiler at build time? Maybe make the parts volatile?

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

PR Review Comment: https://git.openjdk.org/jdk/pull/10661#discussion_r1355064475
PR Review Comment: https://git.openjdk.org/jdk/pull/10661#discussion_r1355059684