Bug#58136: Crash in vio_close at concurrent disconnect and KILL

The problem is a race between a session closing its vio
(i.e. after a COM_QUIT) at the same time it is being killed by
another thread. This could trigger a assertion in vio_close()
as the two threads could end up closing the same vio, at the
same time. This could happen due to the implementation of
SIGNAL_WITH_VIO_CLOSE, which closes the vio of the thread
being killed.

The solution is to serialize the close of the Vio under
LOCK_thd_data, which protects THD data.

No regression test is added as this is essentially a debug
issue and the test case would be quite convoluted as we would
need to synchronize a session that is being killed -- which
is a bit difficult since debug sync points code does not
synchronize killed sessions.

sql/mysqld.cc:
  Drop lock parameter from close_connection, its not necessary
  any more. The newly introduced THD::disconnect method will take
  care of locking.
sql/mysqld.h:
  Change prototype, add a default parameter for the error code.
sql/sql_class.cc:
  In case SIGNAL_WITH_VIO_CLOSE is defined, the active vio is
  closed and cleared. Subsequent calls will only close the vio
  owned by the session.

1 files changed, 3 insertions, 0 deletions

diff --git a/sql/sql_class.h b/sql/sql_class.h
index db20fd73b42..b8e7b685108 100644
--- a/sql/sql_class.h
+++ b/sql/sql_class.h
@@ -2209,6 +2209,9 @@ public:
 #endif
   void awake(THD::killed_state state_to_set);
 
+  /** Disconnect the associated communication endpoint. */
+  void disconnect();
+
 #ifndef MYSQL_CLIENT
   enum enum_binlog_query_type {
     /* The query can be logged in row format or in statement format. */