RFR: Refactor budgeting to make the logic cleaner

[Roman Kennke]

On Tue, 28 Jun 2022 13:46:12 GMT, Kelvin Nilsen <kdnilsen at openjdk.org> wrote:

>> src/hotspot/share/gc/shenandoah/shenandoahHeap.inline.hpp line 689:
>> 
>>> 687: 
>>> 688: void ShenandoahHeap::set_previous_promotion(size_t promoted_bytes) {
>>> 689:   shenandoah_assert_heaplocked();
>> 
>> Is this related and/or relevant?
>
> We can remove that.  It was "somewhat" related to the historical progression that eventually resulted in this refactoring code.  At one point in my development, I had accidentally tried to set_previous_promotion() without having the heap lock, and the results were unpredictable.

Hmm ok. No, leave it there. If it checks an important condition, then it should be useful.

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

PR: https://git.openjdk.org/shenandoah/pull/148