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

Jorn Vernee jvernee at openjdk.org
Thu Jan 30 15:01:27 UTC 2025


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.

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

Commit messages:
 - Force inline MemorySegment::reinterpret

Changes: https://git.openjdk.org/jdk/pull/23352/files
  Webrev: https://webrevs.openjdk.org/?repo=jdk&pr=23352&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8348976
  Stats: 3 lines in 1 file changed: 3 ins; 0 del; 0 mod
  Patch: https://git.openjdk.org/jdk/pull/23352.diff
  Fetch: git fetch https://git.openjdk.org/jdk.git pull/23352/head:pull/23352

PR: https://git.openjdk.org/jdk/pull/23352


More information about the core-libs-dev mailing list