add a FILES table that allows the user to run SQL queries on the files used
to store their tables.

Currently supports NDB


sql/ha_berkeley.cc:
  Add fill_files_table
sql/ha_blackhole.cc:
  Add fill_files_table
sql/ha_federated.cc:
  Add fill_files_table
sql/ha_heap.cc:
  Add fill_files_table
sql/ha_innodb.cc:
  Add fill_files_table
sql/ha_myisam.cc:
  Add fill_files_table
sql/ha_myisammrg.cc:
  Add fill_files_table
sql/ha_ndbcluster.cc:
  Add fill_files_table and implementation for NDB disk data DATAFILES and UNDOFILES
sql/ha_partition.cc:
  Add fill_files_table
sql/handler.h:
  Add fill_files_table to handlerton
sql/log.cc:
  Add fill_files_table
sql/mysql_priv.h:
  Add schema_table_store_record as a function that handlertons can call to store rows in INFORMATION_SCHEMA.FILES
sql/sql_show.cc:
  implement the INFORMATION_SCHEMA.FILES table.
  
  Eventually this may move to PERFORMANCE_SCHEMA.
  
  It currently exists to allow users to query disk usage for NDB disk data tables.
storage/csv/ha_tina.cc:
  Add fill_files_table