Merge lp:~sergei.glushchenko/percona-server/5.6-ps-bug1282008 into lp:percona-server/5.6

[Sergei Glushchenko]

merge from 5.5
http://jenkins.percona.com/view/PS%205.6/job/percona-server-5.6-param/533/

[Laurynas Biveinis (laurynas-biveinis)]

Looks good, setting to WiP if it needs a re-merge from 5.5.

review: Needs Information

[Sergei Glushchenko (sergei.glushchenko)]

http://jenkins.percona.com/view/PS%205.6/job/percona-server-5.6-param/546/

[Laurynas Biveinis (laurynas-biveinis)]

The previous MP had a thd_prepare_connection call and this one is an automerge from 5.5?

review: Needs Information

[Sergei Glushchenko (sergei.glushchenko)]

That is weird because I was sure I MP'ed version without thd_prepare_connection. The difference is that thd_prepare_connection invokes lex_start. I had not much luck making it work, but I didn't investigated deeper.

[Laurynas Biveinis (laurynas-biveinis)]

review: Approve

=== modified file 'sql/threadpool_common.cc'
--- sql/threadpool_common.cc	2013-12-02 12:25:50 +0000
+++ sql/threadpool_common.cc	2014-03-15 08:32:12 +0000
@@ -24,6 +24,7 @@
 #include <debug_sync.h>
 #include <threadpool.h>
 #include <global_threads.h>
+#include <probes_mysql.h>
 
 
 /* Threadpool parameters */
@@ -212,7 +213,9 @@
 
   if (!setup_connection_thread_globals(thd))
   {
-    if (!login_connection(thd))
+    bool rc= login_connection(thd);
+    MYSQL_AUDIT_NOTIFY_CONNECTION_CONNECT(thd);
+    if (!rc)
     {
       prepare_new_connection_state(thd);
       
@@ -229,6 +232,9 @@
         thd->m_server_idle= true;
         threadpool_init_net_server_extension(thd);
       }
+
+      MYSQL_CONNECTION_START(thd->thread_id, &thd->security_ctx->priv_user[0],
+                             (char *) thd->security_ctx->host_or_ip);
     }
   }
   worker_context.restore();