Mailing-List: contact commits-help@harmony.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@harmony.apache.org
Message-ID: <32924031.1170420726083.JavaMail.jira@brutus>
Date: Fri, 2 Feb 2007 04:52:06 -0800 (PST)
From: "Salikh Zakirov (JIRA)" <jira@apache.org>
To: commits@harmony.apache.org
Subject: [jira] Commented: (HARMONY-2858) [drlvm][vmcore] Safe shutdown of
 native threads enabling...
In-Reply-To: <2892596.1166797044778.JavaMail.jira@brutus>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HARMONY-2858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12469737 ] 

Salikh Zakirov commented on HARMONY-2858:
-----------------------------------------

As far as I can see from the patch, it does following things
1) adds states to the thread library, the BLOCKED state would not allow new thread to start by terminating it silently (as error code from thread_start_proc is effectively ignored)
2) adds new field to the thread block: shutdown callback
3) adds new arguments to hythread_create_ex and hythread_attach_ex to pass shutdown callback
4) adds new interface hythread_cancel_safe() which effectively calls the saved callback function with the argument specified at the thread creation time.
The callback is executed on the same thread that called hythread_cancel_safe().

I think that this particular way of shutting down threads is not sufficiently different from what we can do without introducing new interface. Moreover, there are no current users of the new interface, and even in the GCv4 case, the same shutdown algorithm could have been implemented without introducing new thread functions.
I would say that this patch is not needed. 

> [drlvm][vmcore] Safe shutdown of native threads enabling...
> -----------------------------------------------------------
>
>                 Key: HARMONY-2858
>                 URL: https://issues.apache.org/jira/browse/HARMONY-2858
>             Project: Harmony
>          Issue Type: Bug
>          Components: DRLVM
>            Reporter: Elena Semukhina
>         Assigned To: Gregory Shimansky
>         Attachments: H-2858_native_shutdown_update1.patch
>
>
> Currently, DRLVM doesn't provide an ability to stop native threads in a safe manner. The attached patch fixes this issue. It also contains changes for gcv4.1 which could be used as an example how to use new shutdown interface. DRLVM tests passed on Win2003-32bit & SUSE9-ia32.
>  I've created this issue on behalf of Evgueni Brevnov since he has left for vacation. 

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