RFR: 8351280: Mark Assertion Predicates useless instead of replacing them by a constant directly [v4]

[Christian Hagedorn]

On Fri, 7 Mar 2025 09:20:11 GMT, Christian Hagedorn <chagedorn at openjdk.org> wrote:

>> Pushing inside of `mark_useless` has the advantage that one cannot forget pushing.
>
> Forgot to comment on that. With the next PR, I'm updating the elimination of Template Assertion Predicates and thus require to call `mark_useless()` in a first step to mark them all useless. Afterwards, I'm marking those non-useless again that can be found from loops. Thus, we do not want to always add all nodes to the worklist when calling `mark_useless()` if they are not going to be removed.

While working and testing the follow-up PR, I found that I could still do it differently than originally planned. I'm thus changing this to match what we have in `OpaqueInitializedAssertionPredicateNode`. I've also updated `OpaqueMultiversioning` accordingly.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/23941#discussion_r1984879711