[jira] Closed: (HARMONY-4626) [drlvm][thread] The creation and killing processes of threads are not atomic with respect to each other

["weldon washburn (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HARMONY-4626?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

weldon washburn closed HARMONY-4626.
------------------------------------

    Resolution: Fixed

thread lifecycle commits have fixed this bug

> [drlvm][thread] The creation and killing processes of threads are not atomic with respect to each other
> -------------------------------------------------------------------------------------------------------
>
>                 Key: HARMONY-4626
>                 URL: https://issues.apache.org/jira/browse/HARMONY-4626
>             Project: Harmony
>          Issue Type: Bug
>          Components: DRLVM
>         Environment: Win/Linux
>            Reporter: Li-Gang Wang
>
> The creation and killing of threads are supposed to be serialized. But we observed opposite behaviors in GC. In GCv5, a lock is maintained to protect the mutator list which incorporates all mutators in VM. In fact, if the creation and killing of threads are serialized, this lock is unnecessary. We tried this by commenting out the lock and unlock operations in functions mutator_initialize() and mutator_destruct() and ran thread.SmallStackThreadTest in smoke test. There are intermittent failures showing that when a mutator is trying to destruct itself it can not find itself in the mutator list.

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