RFR: 8274245: sun/tools/jmap/BasicJMapTest.java Mutex rank failures [v2]

[David Holmes]

On Fri, 24 Sep 2021 01:38:21 GMT, Coleen Phillimore <coleenp at openjdk.org> wrote:

>> This test fails due to Mutex rank assertions that I just added.  The new mutexes in the heapDumper needs to be ranked 'nosafepoint' and 'nosafepoint-1' since they don't cause safepoint checks.
>> This code also holds two no-safepoint checking locks and waits on the second.  This doesn't break rank ordering but does run afoul of an old assert that we had in the code, that is no longer has meaning.
>> 
>> The test still fails because of other bugs like JDK-8274196 but it doesn't get the mutex rank assertions anymore.
>> 
>> Testing with tier 5 and 6, where the failure was observed, in progress.
>
> Coleen Phillimore has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Remove tests for special wait assert.

I think the assert still has merit but only for a JavaThread, as it means we are doing a wait on a monitor with no safepoint checking and that could potentially delay a safepoint arbitrarily. But obviously it is fine for NJTs to coordinate with each other (wait & notify) using a nosafepoint Monitor.

David

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

PR: https://git.openjdk.java.net/jdk/pull/5669