[jdk25u-dev] RFR: 8358529: GenShen: Heuristics do not respond to changes in SoftMaxHeapSize

[Aleksey Shipilev]

On Tue, 6 Jan 2026 02:11:38 GMT, Rui Li <duke at openjdk.org> wrote:

> 3 backports in this pr:
> - [JDK-8358529](https://bugs.openjdk.org/browse/JDK-8358529) (For addressing cherry pick conflicts caused by 8372543)
> - [JDK-8372543](https://bugs.openjdk.org/browse/JDK-8372543) (The main backport I'm hoping to backport)
> - [JDK-8373789](https://bugs.openjdk.org/browse/JDK-8373789) (8372543 missed an import. This is the bugfix)
> 
> Even with JDK-8358529, the backport is still not clean for JDK-8372543. It's missing [JDK-8365880](https://bugs.openjdk.org/browse/JDK-8365880) and more. Since the conflicts were only an assertion conflicts ([sample](https://github.com/openjdk/jdk/blame/b1e8c4e030f42ea3146b2502c9ab030bc79a8147/src/hotspot/share/gc/shenandoah/shenandoahFreeSet.cpp#L3065)), I'm skipping the complicated backport rabbit hole and picking the latest assertions. The other two backports (8358529, 8373789) are clean.
> 
> Testing: GHA passed.

I say do [JDK-8358529](https://bugs.openjdk.org/browse/JDK-8358529) separately. As it stands now, there is no record at all that [JDK-8372543](https://bugs.openjdk.org/browse/JDK-8372543) is going to be fixed; the PR only references JDK-8358529. So the PR metadata is completely off. Since you need to do fixes anyway, I say do clean backport of JDK-8358529 first, then do follow-ups.

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

PR Comment: https://git.openjdk.org/jdk25u-dev/pull/132#issuecomment-3744105785