diff options
| author | unknown <mkindahl@dl145h.mysql.com> | 2008-02-14 09:53:01 +0100 |
|---|---|---|
| committer | unknown <mkindahl@dl145h.mysql.com> | 2008-02-14 09:53:01 +0100 |
| commit | 7eee3631e12a2d2cce02b918296293ee0ecbc9c3 (patch) | |
| tree | a37416e423387cb7e561fd2bdc41f32b388203f9 /sql/mysqld.cc | |
| parent | b7ebc60f0f7efe694eefa42a5b7b2b84ac6f9dc1 (diff) | |
| download | mariadb-git-7eee3631e12a2d2cce02b918296293ee0ecbc9c3.tar.gz | |
Fixes to try to handle valgrind warnings identical to those in
BUG#24387, which is closed since long.
sql/mysqld.cc:
Moving my_thread_end() to before pthread_cond_broadcast() since
it might cause other threads to start using resources that are
about to be released, or tries to proceed assuming that the
resources have already been released.
Diffstat (limited to 'sql/mysqld.cc')
| -rw-r--r-- | sql/mysqld.cc | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/sql/mysqld.cc b/sql/mysqld.cc index 08ecc025332..9c4e6f9e2a2 100644 --- a/sql/mysqld.cc +++ b/sql/mysqld.cc @@ -1869,9 +1869,9 @@ bool one_thread_per_connection_end(THD *thd, bool put_in_cache) /* It's safe to broadcast outside a lock (COND... is not deleted here) */ DBUG_PRINT("signal", ("Broadcasting COND_thread_count")); + my_thread_end(); (void) pthread_cond_broadcast(&COND_thread_count); - my_thread_end(); pthread_exit(0); DBUG_RETURN(0); // Impossible } |