[jdk11u-dev] RFR: 8224829: AsyncSSLSocketClose.java has timing issue [v2]

[Christoph Langer]

On Mon, 4 Oct 2021 09:53:34 GMT, Martin Doerr <mdoerr at openjdk.org> wrote:

> Right, we need the `recordLock` for this change. Unfortunately, JDK-8221882 is too large for a backport and the part you have taken from it is not small, either. So, we need to be careful when doing more backports in this area. Your version looks good to me.

Thanks for the review! Yes, the alternative would be to refactor the head change to use `synchronized` instead of recordLock but I think for possible future backports in that area I guess it's not a bad idea to use the new style locking. I went through the code quite thoroughly, to make sure I have catched every occurence of synchronization on the Record object and every usage of recordLock. I've spotted no regressions during testing.

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

PR: https://git.openjdk.java.net/jdk11u-dev/pull/446