RFR: 8354996: Reduce dynamic code generation for a single downcall [v2]

[Maurizio Cimadamore]

On Fri, 18 Apr 2025 18:25:54 GMT, Chen Liang <liach at openjdk.org> wrote:

>> Perf numbers for simple main:
>> Linking of `Class::forName0` down from ~152 to ~83
>> 
>> For calling little color management system https://bugs.openjdk.org/browse/JDK-8313344:
>> JNI: ~45
>> baseline panama: ~164
>> patch: ~103
>> 
>> Also see #24705.
>
> Chen Liang has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Simplify FFM linker warmup

src/java.base/share/classes/jdk/internal/foreign/abi/NativeEntryPoint.java line 57:

> 55:             if (!(o instanceof CacheKey other)) return false;
> 56: 
> 57:             return methodType == other.methodType && abi == other.abi && capturedStateMask == other.capturedStateMask

Is this correct? The old code was using `MethodType::equals`, but now we seem to use `==` instead. I suppose the idea is that method types are interned, and their `equals` method is only really meaningful when checking whether a method type with same shape has already been interned?

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

PR Review Comment: https://git.openjdk.org/jdk/pull/24742#discussion_r2063500588