Comment 3 for bug 1233383

trx_mutex also needs to be taken care of.

=== modified file 'Percona-Server/storage/innobase/handler/ha_innodb.cc'
--- Percona-Server/storage/innobase/handler/ha_innodb.cc 2013-10-02 09:14:44 +0000
+++ Percona-Server/storage/innobase/handler/ha_innodb.cc 2013-10-02 20:17:33 +0000
@@ -17273,7 +17273,11 @@
    if (wait_lock) {
     WSREP_DEBUG("canceling wait lock");
     victim_trx->lock.was_chosen_as_deadlock_victim= TRUE;
+ lock_mutex_enter();
+ trx_mutex_enter(victim_trx);
     lock_cancel_waiting_and_release(wait_lock);
+ trx_mutex_exit(victim_trx);
+ lock_mutex_exit();
    }

    wsrep_thd_awake(thd, signal);

fixes it. (Though there is another crash but not related)