On Wed, 17 Jun 2026 14:45:00 GMT, Jorn Vernee <[email protected]> wrote:

>> Per Minborg has updated the pull request incrementally with one additional 
>> commit since the last revision:
>> 
>>   Remove class
>
> src/java.base/share/classes/java/lang/Thread.java line 394:
> 
>> 392:      * Returns a pointer to the pooled memory or zero if the pool 
>> cannot be acquired.
>> 393:      */
>> 394:     @ForceInline
> 
> This doesn't look like a good use case for `@ForceInline`. Force inline 
> should generally only be used for methods that 'evaporate' to almost nothing, 
> e.g. because some intrinsic folds away most of the code (i.e. cases where 
> estimates based on the bytecode size of the method are wrong). That doesn't 
> seem to be the case here.

I think it's fine if most of the allocation path is out-of-line, as long as we 
don't return objects across the inlining boundary, which would then not be 
scalarized. But in this case we try to delay the creation of the memory segment 
as long as possible, so we shouldn't have that problem.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/31365#discussion_r3429150101

Reply via email to