You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@harmony.apache.org by "Sergey Salishev (JIRA)" <ji...@apache.org> on 2007/05/31 15:54:15 UTC
[jira] Created: (HARMONY-4014) [drlvm][thread]100% CPU consumption
in MONITORENTER
[drlvm][thread]100% CPU consumption in MONITORENTER
---------------------------------------------------
Key: HARMONY-4014
URL: https://issues.apache.org/jira/browse/HARMONY-4014
Project: Harmony
Issue Type: Bug
Components: DRLVM
Reporter: Sergey Salishev
The attache test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HARMONY-4014) [drlvm][thread]100% CPU
consumption in MONITORENTER
Posted by "Vasily Zakharov (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-4014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12563157#action_12563157 ]
Vasily Zakharov commented on HARMONY-4014:
------------------------------------------
Pavel,
Thanks you very much, this workaround patch shows this problem is also the cause of 100% CPU consumption in Geronimo.
> [drlvm][thread]100% CPU consumption in MONITORENTER
> ---------------------------------------------------
>
> Key: HARMONY-4014
> URL: https://issues.apache.org/jira/browse/HARMONY-4014
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Reporter: Sergey Salishev
> Assignee: weldon washburn
> Priority: Minor
> Attachments: H-4014-workaround.patch, Test.java
>
>
> The attached test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
> This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HARMONY-4014) [drlvm][thread]100% CPU consumption
in MONITORENTER
Posted by "Sergey Salishev (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-4014?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sergey Salishev updated HARMONY-4014:
-------------------------------------
Attachment: Test.java
> [drlvm][thread]100% CPU consumption in MONITORENTER
> ---------------------------------------------------
>
> Key: HARMONY-4014
> URL: https://issues.apache.org/jira/browse/HARMONY-4014
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Reporter: Sergey Salishev
> Attachments: Test.java
>
>
> The attache test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
> This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HARMONY-4014) [drlvm][thread]100% CPU
consumption in MONITORENTER
Posted by "Pavel Rebriy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-4014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12563171#action_12563171 ]
Pavel Rebriy commented on HARMONY-4014:
---------------------------------------
The second workaround is better because doesn't lead to crash :)
> [drlvm][thread]100% CPU consumption in MONITORENTER
> ---------------------------------------------------
>
> Key: HARMONY-4014
> URL: https://issues.apache.org/jira/browse/HARMONY-4014
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Reporter: Sergey Salishev
> Assignee: weldon washburn
> Priority: Minor
> Attachments: H-4014-workaround.patch, H-4014-workaround.patch, Test.java
>
>
> The attached test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
> This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HARMONY-4014) [drlvm][thread]100% CPU consumption
in MONITORENTER
Posted by "Pavel Rebriy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-4014?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Pavel Rebriy updated HARMONY-4014:
----------------------------------
Attachment: H-4014-workaround.patch
This is workaround for attached test.
With this patch test doesn't take up all CPU.
> [drlvm][thread]100% CPU consumption in MONITORENTER
> ---------------------------------------------------
>
> Key: HARMONY-4014
> URL: https://issues.apache.org/jira/browse/HARMONY-4014
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Reporter: Sergey Salishev
> Assignee: weldon washburn
> Priority: Minor
> Attachments: H-4014-workaround.patch, Test.java
>
>
> The attached test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
> This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HARMONY-4014) [drlvm][thread]100% CPU consumption
in MONITORENTER
Posted by "Sergey Salishev (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-4014?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sergey Salishev updated HARMONY-4014:
-------------------------------------
Description:
The attached test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
was:
The attache test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
> [drlvm][thread]100% CPU consumption in MONITORENTER
> ---------------------------------------------------
>
> Key: HARMONY-4014
> URL: https://issues.apache.org/jira/browse/HARMONY-4014
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Reporter: Sergey Salishev
> Attachments: Test.java
>
>
> The attached test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
> This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (HARMONY-4014) [drlvm][thread]100% CPU consumption
in MONITORENTER
Posted by "weldon washburn (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-4014?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
weldon washburn reassigned HARMONY-4014:
----------------------------------------
Assignee: weldon washburn
> [drlvm][thread]100% CPU consumption in MONITORENTER
> ---------------------------------------------------
>
> Key: HARMONY-4014
> URL: https://issues.apache.org/jira/browse/HARMONY-4014
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Reporter: Sergey Salishev
> Assignee: weldon washburn
> Attachments: Test.java
>
>
> The attached test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
> This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HARMONY-4014) [drlvm][thread]100% CPU
consumption in MONITORENTER
Posted by "Sergey Salishev (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-4014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12538746 ]
Sergey Salishev commented on HARMONY-4014:
------------------------------------------
The problem is still reproducible on 587111 Harmony snapshot.
> [drlvm][thread]100% CPU consumption in MONITORENTER
> ---------------------------------------------------
>
> Key: HARMONY-4014
> URL: https://issues.apache.org/jira/browse/HARMONY-4014
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Reporter: Sergey Salishev
> Assignee: weldon washburn
> Priority: Minor
> Attachments: Test.java
>
>
> The attached test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
> This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HARMONY-4014) [drlvm][thread]100% CPU consumption
in MONITORENTER
Posted by "Pavel Rebriy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-4014?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Pavel Rebriy updated HARMONY-4014:
----------------------------------
Attachment: H-4014-workaround.patch
An updated patch.
> [drlvm][thread]100% CPU consumption in MONITORENTER
> ---------------------------------------------------
>
> Key: HARMONY-4014
> URL: https://issues.apache.org/jira/browse/HARMONY-4014
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Reporter: Sergey Salishev
> Assignee: weldon washburn
> Priority: Minor
> Attachments: H-4014-workaround.patch, H-4014-workaround.patch, Test.java
>
>
> The attached test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
> This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HARMONY-4014) [drlvm][thread]100% CPU
consumption in MONITORENTER
Posted by "Tim Ellison (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-4014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12536962 ]
Tim Ellison commented on HARMONY-4014:
--------------------------------------
Is this issue still valid given the latest thread changes? It is still listed as blocking a critical issue.
> [drlvm][thread]100% CPU consumption in MONITORENTER
> ---------------------------------------------------
>
> Key: HARMONY-4014
> URL: https://issues.apache.org/jira/browse/HARMONY-4014
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Reporter: Sergey Salishev
> Assignee: weldon washburn
> Priority: Minor
> Attachments: Test.java
>
>
> The attached test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
> This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HARMONY-4014) [drlvm][thread]100% CPU consumption
in MONITORENTER
Posted by "weldon washburn (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HARMONY-4014?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
weldon washburn updated HARMONY-4014:
-------------------------------------
Priority: Minor (was: Major)
> [drlvm][thread]100% CPU consumption in MONITORENTER
> ---------------------------------------------------
>
> Key: HARMONY-4014
> URL: https://issues.apache.org/jira/browse/HARMONY-4014
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Reporter: Sergey Salishev
> Assignee: weldon washburn
> Priority: Minor
> Attachments: Test.java
>
>
> The attached test eats 100% of cpu time on Harmony while on RI it's 0% as it's blocked in IO op.
> This is caused by busy wait in monitor enter when trying to inflate thin lock which is held by thread blocked in OS from other thread.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.