[jdk21u-dev] RFR: 8323782: Race: Thread::interrupt vs. AbstractInterruptibleChannel.begin

[Richard Reingruber]

On Thu, 4 Apr 2024 14:21:17 GMT, Richard Reingruber <rrich at openjdk.org> wrote:

> Hi,
> 
> this PR contains a backport of commit https://github.com/openjdk/jdk/commit/4018b2b19629ddb8cd7a56e064dfef371f23e5fa.
> 
> I'd like to do the backport because jdk 21 is affected by the issue. The included test hangs without the change.
> Besides it will reduce the noise in jck tests we're doing regularily (the test api/java_net/ServerSocket/AcceptInterruptibleTests_PlatformThreads fails sometimes because of the issue).
> 
> The original fix does not apply due to renamig of a local variable. Also the context differs because there's no `postInterrupt` method to be called.
> Nevertheless I'd consider the risk low because of the small size of the fix.
> 
> Tested with the included test.
> 
> The fix passed our CI testing: JTReg tests: tier1-4 of hotspot and jdk. All of Langtools and jaxp. JCK, SPECjvm2008, SPECjbb2015, Renaissance Suite, and SAP specific tests (also with ParallelGC).
> Testing was done with fastdebug builds on the main platforms and also on Linux/PPC64le.

I'll switch this back to draft in order to do the jdk 22 backport first.

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

PR Comment: https://git.openjdk.org/jdk21u-dev/pull/451#issuecomment-2040146181