When sizing the heap for the next mutator phase, the MMU determines a minimum desired number of eden regions to keep the MMU.

Currently, this number (adding the current survivor regions) overrides the maximum specified by the heap sizing.

This CR is supposed to discuss this behavior in more detail and at least provide logging/JFR event if/when that happens so that the user can manually improve this. (Logging/JFR event may be split out).

Occurs in e.g. specjbb2015 with 10k fixed IR with 1-16gb heap on a recent machine.

It looks that this behavior is intentional currently, so only filing an RFE.