[jira] [Commented] (TS-2254) ink_atomic_increment should return the old value

["Leif Hedstrom (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/TS-2254?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13777731#comment-13777731 ] 

Leif Hedstrom commented on TS-2254:
-----------------------------------

This is for the ARM path, right ? Nice catch!
                
> ink_atomic_increment should return the old value
> ------------------------------------------------
>
>                 Key: TS-2254
>                 URL: https://issues.apache.org/jira/browse/TS-2254
>             Project: Traffic Server
>          Issue Type: Bug
>          Components: Core
>            Reporter: Yu Qing
>            Assignee: Yu Qing
>         Attachments: 0001-TS-2254-ink_atomic_increment-should-return-the-old-v.patch
>
>
> lib/ts/ink_atomic.h
> 197 template<>
> 198 inline int64_t
> 199 ink_atomic_increment<int64_t>(pvint64 mem, int64_t value) {
> 200   int64_t curr;
> 201   ink_mutex_acquire(&__global_death);
> 202   curr = *mem;
> 203   *mem = curr + value;
> 204   ink_mutex_release(&__global_death);
> 205   return curr + value;  //SHOULD return curr!
> 206 }
> this function should return the old value (curr, NOT curr + value). it should return same value as  gcc inline function __sync_fetch_and_add.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira