[jira] [Commented] (TS-1678) Unnecessary lock releasing in RecRegisterStat()

["James Peach (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/TS-1678?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13567827#comment-13567827 ] 

James Peach commented on TS-1678:
---------------------------------

Yes I agree. I think that the 'release_record_lock' parameter to register_record() is unnecessarily complicated. Can you also simplify or remove that?
                
> Unnecessary lock releasing in RecRegisterStat()
> -----------------------------------------------
>
>                 Key: TS-1678
>                 URL: https://issues.apache.org/jira/browse/TS-1678
>             Project: Traffic Server
>          Issue Type: Improvement
>          Components: Stats
>            Reporter: Yunkai Zhang
>
> It seems that lock releasing in the following code is unnecessary and dangerous in some situation.
> {code}
> //-------------------------------------------------------------------------
> // RecRegisterStat
> //-------------------------------------------------------------------------
> RecRecord *
> RecRegisterStat(RecT rec_type, const char *name, RecDataT data_type, RecData data_default, RecPersistT persist_type)
> {
>   RecRecord *r = NULL;
>   ink_rwlock_wrlock(&g_records_rwlock);
>   if ((r = register_record(rec_type, name, data_type, data_default, false)) != NULL) {
>     r->stat_meta.persist_type = persist_type;
>     rec_mutex_release(&(r->lock));  //dangerous!!
>   } else {
>     ink_debug_assert(!"Can't register record!");
>   }
>   ink_rwlock_unlock(&g_records_rwlock);
>   return r;
> }
> {code}

--
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