|
AW: more lock modes based on AbstractQueuedSynchronizer: msg#00001java.jsr.166-concurrency
As far as I understand: * "shared" only means "potentially shared", i.e. two threads possibly operating concurrently "in" the lock. The exclusive mode is more performant since the AQS "knows better" how many waiting threads to wake up * the lock state must be encodable in an int * a lock request must be encodable in an int (why is this? Well, it can't be more information as is encodable in the state ...) So it seems well possible to do what you're thinking about. AQS is very powerful indeed ... Matthias
|
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| Previous by Date: | more lock modes based on AbstractQueuedSynchronizer, Hu, Jinsong |
|---|---|
| Next by Date: | Re: AW: more lock modes based on AbstractQueuedSynchronizer, Doug Lea |
| Previous by Thread: | more lock modes based on AbstractQueuedSynchronizer, Hu, Jinsong |
| Next by Thread: | Re: AW: more lock modes based on AbstractQueuedSynchronizer, Doug Lea |
| Indexes: | [Date] [Thread] [Top] [All Lists] |
| News | FAQ | advertise |