[infinispan-issues] [JBoss JIRA] Resolved: (ISPN-318) Locking could optionally be "fair"

Manik Surtani (JIRA) jira-events at lists.jboss.org
Tue Jan 5 11:13:30 EST 2010


     [ https://jira.jboss.org/jira/browse/ISPN-318?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Manik Surtani resolved ISPN-318.
--------------------------------

    Fix Version/s:     (was: 4.1.0.GA)
       Resolution: Rejected


Is this really necessary?

> Locking could optionally be "fair"
> ----------------------------------
>
>                 Key: ISPN-318
>                 URL: https://jira.jboss.org/jira/browse/ISPN-318
>             Project: Infinispan
>          Issue Type: Feature Request
>          Components: Locking and Concurrency
>    Affects Versions: 4.0.0.CR3
>            Reporter: Manik Surtani
>            Assignee: Manik Surtani
>
> Would require a number of changes:
> 1.  A new boolean field and getter/setter in Configuration.LockingType (follow patterns of other boolean flags here to ensure proper XSD and parser generation), perhaps called "useFairLocking", defaults to false.
> 2.  Modify LockManagerImpl to pass this flag into the LockContainer implementations (in LockManagerImpl#startLockManager())
> 3.  Modify the 4 concrete LockContainer implementations to use this flag when creating new locks.  Implementations that use JDK reentrant locks just need to pass the fairness flag into the ReentrantLock constructor.
> 4.  Modify OwnableReentrantLock to handle fairness.  Referring to the JDK's ReentrantLock.FairSync may help.  OwnableReentrantLock combines features of the JDK ReentrantLock and ReentrantLock.NonFairSync into a single class, so perhaps the best approach is to create a new OwnableFairReentrantLock.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        


More information about the infinispan-issues mailing list