[jira] [Resolved] (IMPALA-4658) Potential race if compiler reorders ReachedLimit() usage

["Abhishek Rawat (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/IMPALA-4658?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Abhishek Rawat resolved IMPALA-4658.
------------------------------------
       Resolution: Fixed
    Fix Version/s: Impala 3.3.0

> Potential race if compiler reorders ReachedLimit() usage
> --------------------------------------------------------
>
>                 Key: IMPALA-4658
>                 URL: https://issues.apache.org/jira/browse/IMPALA-4658
>             Project: IMPALA
>          Issue Type: Bug
>          Components: Backend
>    Affects Versions: Impala 2.8.0
>            Reporter: Matthew Jacobs
>            Assignee: Abhishek Rawat
>            Priority: Major
>             Fix For: Impala 3.3.0
>
>
> There is a potential race in our many usages of {{ExecNode::ReachedLimit()}} during execution from a different thread than the one updating rows_returned, e.g. scanner threads. The compiler in theory is free to notice that they're not updating {{rows_returned_}} and hoist the load up out of loops, etc.
> We should consider AtomicInt64 or NoBarrier_Load(...).



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)