RFR: 8355698: JDK not supporting sleef could cause exception at runtime after JDK-8353786
Vladimir Ivanov
vlivanov at openjdk.org
Wed Apr 30 19:46:26 UTC 2025
On Wed, 30 Apr 2025 11:00:04 GMT, Hamlin Li <mli at openjdk.org> wrote:
> It's arguable if missing entries is a bug, please check [JDK-8355656](https://bugs.openjdk.org/browse/JDK-8355656) for example
Personally, I'm surprised SVML is affected in a similar way because stubs are provided in assembly form, so they are simply linked into a shared library during assembly phase. I can only guess what causes such inconsistencies during JDK build phase. And it just reassures me that current behavior is a source of bugs in the future.
Overall, it still looks like a JDK build issue to me. Hiding problems occurred during the build is not good. If some toolchains can't successfully build the library, the library shouldn't be included in JDK.
> After all we does not support TANH on riscv and arm for now.
JDK code is aware of that and doesn't try to look up corresponding entries. Ideally, it shouldn't be included into the native library, since it increases its size for no good reason. But having unused code is a redundancy and not a bug per se.
(For SVML it's a bit more complicated, because JDK is the only place where sources of SVML stubs live (there's no upstream project/distribution)).
-------------
PR Comment: https://git.openjdk.org/jdk/pull/24914#issuecomment-2843093609
More information about the hotspot-dev
mailing list