Bug#55091: Server crashes on ADD PARTITION after a failed attempt

In case of failure in ALTER ... PARTITION under LOCK TABLE
the server could crash, due to it had modified the locked
table object, which was not reverted in case of failure,
resulting in a bad table definition used after the failed
command.

Solved by always closing the LOCKED TABLE, even in case
of error.

Note: this is a 5.1-only fix, bug#56172 fixed it in 5.5+

1 files changed, 20 insertions, 1 deletions

diff --git a/sql/sql_partition.cc b/sql/sql_partition.cc
index 76caa2b0c8d..7b0c47865d8 100644
--- a/sql/sql_partition.cc
+++ b/sql/sql_partition.cc
@@ -5935,6 +5935,12 @@ static void alter_partition_lock_handling(ALTER_PARTITION_PARAM_TYPE *lpt)
   if (lpt->thd->locked_tables)
   {
     /*
+      Close the table if open, to remove/destroy the already altered
+      table->part_info object, so that it is not reused.
+    */
+    if (lpt->table->db_stat)
+      abort_and_upgrade_lock_and_close_table(lpt);
+    /*
       When we have the table locked, it is necessary to reopen the table
       since all table objects were closed and removed as part of the
       ALTER TABLE of partitioning structure.
@@ -6436,7 +6442,20 @@ uint fast_alter_partition_table(THD *thd, TABLE *table,
                                  table, table_list, FALSE, NULL,
                                  written_bin_log));
 err:
-  close_thread_tables(thd);
+  if (thd->locked_tables)
+  {
+    /*
+      table->part_info was altered in prep_alter_part_table and must be
+      destroyed and recreated, since otherwise it will be reused, since
+      we are under LOCK TABLE.
+    */
+    alter_partition_lock_handling(lpt);
+  }
+  else
+  {
+    /* Force the table to be closed to avoid reuse of the table->part_info */
+    close_thread_tables(thd);
+  }
   DBUG_RETURN(TRUE);
 }
 #endif