RFR: 8319137: release _object in ObjectMonitor dtor to avoid races [v2]

[Aleksey Shipilev]

On Mon, 20 Nov 2023 16:58:37 GMT, Patricio Chilano Mateo <pchilanomate at openjdk.org> wrote:

> So with the current changes this fix is almost a backout except for the extra branch I removed in the loop given that now only the deflater thread can call this, and the fact that the delete loop was factored out into a separate static method. Can't this be considered a backout and rename it as such?

Probably not worth the hassle of doing the backout and resolve the conflicts there. But it should be clearly said somewhere (e.g. both in PR body and the JBS issue) that the issue is with [JDK-8256302](https://bugs.openjdk.org/browse/JDK-8256302) and the fix is the effective backout of [JDK-8256302](https://bugs.openjdk.org/browse/JDK-8256302).

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

PR Comment: https://git.openjdk.org/jdk/pull/16738#issuecomment-1819473242