(especially in case of bug like BUG#6732 "FLUSH TABLES WITH READ LOCK + COMMIT makes next FLUSH...LOCK hang forever").
sql/lock.cc:
making FLUSH TABLES WITH READ LOCK killable while it's waiting for running commits to finish
sql/mysql_priv.h:
prototype change
sql/sql_parse.cc:
now it's possible that make_global_read_lock_block_commit fails (killed)