You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-dev@hadoop.apache.org by "Chris Douglas (JIRA)" <ji...@apache.org> on 2009/01/13 01:07:59 UTC

[jira] Updated: (HADOOP-4671) Make some "flag" booleans volatile because they are set/read in different threads

     [ https://issues.apache.org/jira/browse/HADOOP-4671?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Chris Douglas updated HADOOP-4671:
----------------------------------

    Attachment: 4671-0.patch

Fixed the preceding. If "voTask" refers to Task.ValuesIterator, I suspect this is a false positive. The static analysis tool Peter used may not have detected that hasNext is updated in the function called in the loop, though it is not updated in the loop body. The iterator should not be shared between threads, anyway.

> Make some "flag" booleans volatile because they are set/read in different threads
> ---------------------------------------------------------------------------------
>
>                 Key: HADOOP-4671
>                 URL: https://issues.apache.org/jira/browse/HADOOP-4671
>             Project: Hadoop Core
>          Issue Type: Bug
>          Components: dfs
>            Reporter: Owen O'Malley
>             Fix For: 0.20.0
>
>         Attachments: 4671-0.patch
>
>
> Peter Veentjer	has identified the following variables as being tracked in a loop without assignments or synchronization.
> AgentControllerSocketListener.closing
> HttpConnector.stopMe
> IndexUpdateReducer.closed
> SecundaryNodeName.shouldRun
> voTask.hasNext

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.