RFR: 8329570: G1: Excessive is_obj_dead_cond calls in verification

Stefan Karlsson stefank at openjdk.org
Wed Apr 3 17:50:10 UTC 2024


On Wed, 3 Apr 2024 17:31:09 GMT, Axel Boldt-Christmas <aboldtch at openjdk.org> wrote:

>> Honestly, I think we are overthinking this. The original patch gives us a major bang for the buck, and it has a benefit of being obvious and thus very backportable. I suggest we go with the current version, and then do any other followups.
>
> I can clarify my position. I completely agree with @shipilev, hence my approval of still stands. I think the current fix is a very reasonable solution to the problem.
> 
> But my initial thought / reaction when looking at the code was "why is the closure have mutable state" and "the check agains `G1MaxVerifyFailures` looks strange".

Yeah, I think the suggestion to keeping the null-check in do_oop_work was overthinking things. But, since that seemed important I proposed an alternative solution that I think actually is cleaner and clearer than what the current code was doing

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

PR Review Comment: https://git.openjdk.org/jdk/pull/18595#discussion_r1550203693


More information about the hotspot-gc-dev mailing list