RFR: 8298059: Linked stack watermarks don't support nesting [v3]

[Robbin Ehn]

On Tue, 6 Dec 2022 13:23:29 GMT, Erik Österlund <eosterlund at openjdk.org> wrote:

>> The KeepStackGCProcessedMark facility used to ensure a thread (possibly remote) remains fully processed in its scope, has not yet had to support nesting. As it turns out, they can now nest after JDK-8294924. So this patch adds support for nesting. Tested ZGC specific tests tier 1-8, general testing 1-3 and manually tested the two tests that failed and got problem listed.
>
> Erik Österlund has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Change from mtInternal to mtThread

EscapeBarrier do `KeepStackGCProcessedMark ksgcpm(deoptee_thread())` while the deoptee/target thread is blocked.
So to me it seem like a concurrent handshake may also create a KeepStackGCProcessedMark.
Thus this lock should be needed ?

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

PR: https://git.openjdk.org/jdk/pull/11530