RFR: 8349479: C2: when a Type node becomes dead, make CFG path that uses it unreachable [v3]

[Roland Westrelin]

On Mon, 7 Apr 2025 06:44:54 GMT, Christian Hagedorn <chagedorn at openjdk.org> wrote:

>> I agree that we would need to be consistent and that it makes little sense to add null checks in code that have been around forever and has never caused issues. Maybe we can somehow have igvn itself assert that every node it processes has a set of expected inputs non null? I suppose, every node type would need to define which of its inputs can be null, then.
>
> Yes, that would be a nice verification. Should we file an RFE for that?

I filed https://bugs.openjdk.org/browse/JDK-8353895 for that

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

PR Review Comment: https://git.openjdk.org/jdk/pull/23468#discussion_r2031509907