RFR: 8330051: Small ObjectMonitor spinning code cleanups [v2]

Coleen Phillimore coleenp at openjdk.org
Wed Apr 17 12:32:44 UTC 2024


On Wed, 10 Apr 2024 18:36:42 GMT, Coleen Phillimore <coleenp at openjdk.org> wrote:

>> Please review this patch that refactors some functions in the adaptive spinning code and moves the Knobs to where they're used and can be modified for experiments in the following code.  Also changed 'goto's to a 'break'.  I didn't change the names of TrySpin or TryLock, since all the names in this file are this style.
>> 
>> The only non-cosmetic change is for the pre-spin.  If the TryLock returns that a CAS failed, then the pre-spin ends, which shouldn't make a difference wrt to performance because the pre-spin is only for 10 iterations, but somehow my numbers are a little bit better for both Dacapo pmd and xalan.
>> 
>> Tested with tier1-7.
>
> Coleen Phillimore has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Move first fixed spin after the comment that describes why we do it.

Thank you for the knowledgable review and comments.

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

PR Review: https://git.openjdk.org/jdk/pull/18730#pullrequestreview-2005864150


More information about the hotspot-runtime-dev mailing list