RFR: 8255582: Introduce SemaphoreLock and SemaphoreLocker

[Stefan Karlsson]

On Thu, 29 Oct 2020 13:48:28 GMT, Robbin Ehn <rehn at openjdk.org> wrote:

> > I'm closing this PR, given the other available alternatives.
> > Though I must say, I don't think it's fair to use an "imminent rewrite of the lock ranking" as a motivation to push back on my wish to use some kind of locking that doesn't force me to fake a lock order when the code intend to be a leaf operation. A lock ranking rewrite has been discussed for over a decade now.
> 
> The problem is that people think they are leaf, but they are really not.
> If they truly where leaf, shouldn't setting lock rank to 'event' always work?
> 
> EDIT:
> What think we have an issue with is non-reentrant code, you are not a leaf but you know the code paths in the exclusive code region can never call this 'module' again. Which leads to "give me a rank that works" I don't care.
> Correct ?

Yes, you might be right.

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

PR: https://git.openjdk.java.net/jdk/pull/927