RFR: 8348976: MemorySegment::reinretpret should be force inlined

[Chen Liang]

On Wed, 29 Jan 2025 16:15:51 GMT, Jorn Vernee <jvernee at openjdk.org> wrote:

> See the JBS issue.
> 
> Mark the implementation of `MemorySegment::reinterpret` with `@ForceInline` so that we prevent `Reflection.getCallerClass` from falling back to a slow stack walk.
> 
> This is a leftover from an earlier performance investigation, in which we observed `reinterpret` not being inlined, causing significant slowdowns.

Marked as reviewed by liach (Reviewer).

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

PR Review: https://git.openjdk.org/jdk/pull/23352#pullrequestreview-2584185394