RFR: 8324868: debug agent does not properly handle interrupts of a virtual thread

[Serguei Spitsyn]

On Wed, 28 Feb 2024 05:37:05 GMT, Chris Plummer <cjplummer at openjdk.org> wrote:

>> test/jdk/com/sun/jdi/InterruptHangTest.java line 146:
>> 
>>> 144:         // When in precise mode, make sure that every interrupt sent resulted in
>>> 145:         // an InterruptedException. Note the interruptor always ends up sending
>>> 146:         // one extra interrupt at the end.
>> 
>> Nit: Why do we need this extra interrupt. Can we re-organize the loop to get rid of this extra interrupt? It will make the logics a little bit more simple.
>
> The extra interrupt is because the interruptor will get another shot at interrupting once the interuptee does the following on the last interation (not this is done on every iteration of the main interruptee loop)
> 
>                     synchronized(InterruptHangTarg.sync) {
>                         // Let the interruptor thread know it can start interrupting again
>                         sync.notify();
>                     }
> 
> Avoiding this last notify() would require that it be conditionally avoided on the last iteration of the loop (by comparing ii to INTERRUPTS_EXPECTED), and then as a result set some sort of flag so the PreciseInterruptor doesn't do another Thread.interrupt() after the interruptee wakes it up with the sync.notify(). 
> 
> Having a check in the middle of a loop to see if we are on the last iteration isn't exactly pretty either. I felt having the extra interrupt() was a better choice.

Okay. It makes sense.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/17989#discussion_r1505471458