You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Andrey N. Gura (Jira)" <ji...@apache.org> on 2019/10/08 12:15:00 UTC

[jira] [Updated] (IGNITE-12238) RobinHoodBackwardShiftHashMap works incorrectly on big endian architectures

     [ https://issues.apache.org/jira/browse/IGNITE-12238?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Andrey N. Gura updated IGNITE-12238:
------------------------------------
    Description: 
{{RobinHoodBackwardShiftHashMap}} has bug that can be reproduced only on big endinan architectures. In order to reproduce the problem run the following tests:

* {{RobinHoodBackwardShiftHashMapTest.testCollisionOnRemove}}
* {{RobinHoodBackwardShiftHashMapTest.testRandomOpsPutRemove}}

The problem is {{setIdealBucket()}} method writes {{long}} value to the offheap memory, while {{getIdealBucket()}} reads {{int}} value. For little endian architectures it works because meaningful 4 bytes will written first  to the memory and leading zero bytes will be rewriteen by the next operation. On big endian architecture always 4 zero bytes will be written to the memory.

  was:
{{RobinHoodBackwardShiftHashMap}} has bug that can be reproduced only on big endinan architectures. In order to reproduce the problem run the following tests:

* {{RobinHoodBackwardShiftHashMapTest.testCollisionOnRemove}}
* {{testRandomOpsPutRemove}}

The problem is {{setIdealBucket()}} method writes {{long}} value to the offheap memory, while {{getIdealBucket()}} reads {{int}} value. For little endian architectures it works because meaningful 4 bytes will written first  to the memory and leading zero bytes will be rewriteen by the next operation. On big endian architecture always 4 zero bytes will be written to the memory.


> RobinHoodBackwardShiftHashMap works incorrectly on big endian architectures
> ---------------------------------------------------------------------------
>
>                 Key: IGNITE-12238
>                 URL: https://issues.apache.org/jira/browse/IGNITE-12238
>             Project: Ignite
>          Issue Type: Bug
>            Reporter: Andrey N. Gura
>            Assignee: Andrey N. Gura
>            Priority: Major
>             Fix For: 2.8
>
>          Time Spent: 0.5h
>  Remaining Estimate: 0h
>
> {{RobinHoodBackwardShiftHashMap}} has bug that can be reproduced only on big endinan architectures. In order to reproduce the problem run the following tests:
> * {{RobinHoodBackwardShiftHashMapTest.testCollisionOnRemove}}
> * {{RobinHoodBackwardShiftHashMapTest.testRandomOpsPutRemove}}
> The problem is {{setIdealBucket()}} method writes {{long}} value to the offheap memory, while {{getIdealBucket()}} reads {{int}} value. For little endian architectures it works because meaningful 4 bytes will written first  to the memory and leading zero bytes will be rewriteen by the next operation. On big endian architecture always 4 zero bytes will be written to the memory.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)