RFR: 8263108: Class initialization deadlock in java.lang.constant [v2]

[Jaikiran Pai]

On Wed, 10 Mar 2021 06:36:58 GMT, Jaikiran Pai <jpai at openjdk.org> wrote:

>> Your code change Looks reasonable  to me. Although i am not export in this area but what i observed is the test case which was attached by original submitter passes null to DynamicConstantDesc as follows "DynamicConstantDesc.of(null)". If you pass valid argument like(ConstantDescs.BSM_INVOKE) no deadlock.
>
>> but what i observed is the test case which was attached by original submitter passes null to DynamicConstantDesc as follows "DynamicConstantDesc.of(null)". If you pass valid argument like(ConstantDescs.BSM_INVOKE) no deadlock.
> 
> Hello Vyom,
> 
> The deadlock is reproducible with non-null params too. For example, the test case in this PR consistently reproduces this deadlock when the fix in the source code isn't applied. The `null` in the original report IMO was used just to explain the issue.

Any reviews please?

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

PR: https://git.openjdk.java.net/jdk/pull/2893