You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@zookeeper.apache.org by "hong (Jira)" <ji...@apache.org> on 2021/05/20 03:28:00 UTC
[jira] [Updated] (ZOOKEEPER-4299) List head do not initialize
before using it in zoo_amulti that will cause segment errors.
[ https://issues.apache.org/jira/browse/ZOOKEEPER-4299?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
hong updated ZOOKEEPER-4299:
----------------------------
Summary: List head do not initialize before using it in zoo_amulti that will cause segment errors. (was: list head do not initialize before us it in zoo_amulti that will cause segment errors.)
> List head do not initialize before using it in zoo_amulti that will cause segment errors.
> -----------------------------------------------------------------------------------------
>
> Key: ZOOKEEPER-4299
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-4299
> Project: ZooKeeper
> Issue Type: Bug
> Components: c client
> Affects Versions: 3.5.4, 3.7.0, 3.6.2
> Reporter: hong
> Priority: Blocker
> Original Estimate: 2h
> Remaining Estimate: 2h
>
> int zoo_amulti(zhandle_t *zh, int count, const zoo_op_t *ops,
> zoo_op_result_t *results, void_completion_t completion, const void *data)
> {
> struct RequestHeader h = \{get_xid(), ZOO_MULTI_OP};
> struct MultiHeader mh = \{-1, 1, -1};
> struct oarchive *oa = create_buffer_oarchive();
> {color:#FF0000}completion_head_t clist = \{ 0 }; // not initialize for cond or lock{color}
> int rc = serialize_RequestHeader(oa, "header", &h);
> ....
>
> {color:#FF0000}queue_completion(&clist, entry, 0); //queue it will segment errors{color}
>
>
> // do lock or unlock which have not been initialized!!
> static void queue_completion(completion_head_t *list, completion_list_t *c,
> int add_to_front)
> {
> lock_completion_list(list);
> queue_completion_nolock(list, c, add_to_front);
> unlock_completion_list(list);
> }
> // oh my god!!
> int unlock_completion_list(completion_head_t *l)
> {
> p_thread_cond_broadcast(&l->cond);
> return p_thread_mutex_unlock(&l->lock);
> }
--
This message was sent by Atlassian Jira
(v8.3.4#803005)