Commit 51b33ea3 authored by aelkin/elkin@koti.dsl.inet.fi's avatar aelkin/elkin@koti.dsl.inet.fi
Browse files

Bug #34305 show slave status handling segfaults when slave io is about 

    to leave

The artifact was caused by
a flaw in concurrent accessing the slave's io thd by
the io itself and a handling show slave status thread.
Namely, show_master_info did not acquire mi->run_lock mutex that is
specified for mi->io_thd member.

Fixed with deploying the mutex locking and unlocking. The mutex is kept
short time and without interleaving with mi->data_lock mutex.

Todo: to report and fix an issue with 
    sys_var_slave_skip_counter::{methods} 
seem to acquire incorrectly
     active_mi->rli.run_lock
instead of the specified
     active_mi->rli.data_lock

A test case is difficult to compose, so rpl_packet should continue serving
as the indicator.
parent 69fdf679

sql/slave.cc

+6 −5
Original line number Diff line number Diff line
@@ -2447,14 +2447,15 @@ bool show_master_info(THD* thd, MASTER_INFO* mi) 
    protocol->prepare_for_resend();

  

    /*

      TODO: we read slave_running without run_lock, whereas these variables

      are updated under run_lock and not data_lock. In 5.0 we should lock

      run_lock on top of data_lock (with good order).

      slave_running can be accessed without run_lock but not other

      non-volotile members like mi->io_thd, which is guarded by the mutex.

    */

    pthread_mutex_lock(&mi->run_lock);

    protocol->store(mi->io_thd ? mi->io_thd->proc_info : "", &my_charset_bin);

    pthread_mutex_unlock(&mi->run_lock);



    pthread_mutex_lock(&mi->data_lock);

    pthread_mutex_lock(&mi->rli.data_lock);



    protocol->store(mi->io_thd ? mi->io_thd->proc_info : "", &my_charset_bin);

    protocol->store(mi->host, &my_charset_bin);

    protocol->store(mi->user, &my_charset_bin);

    protocol->store((uint32) mi->port);

sql/slave.h

+2 −2
Original line number Diff line number Diff line
@@ -65,8 +65,8 @@ 
  mi->rli does not either.



  In MASTER_INFO: run_lock, data_lock

  run_lock protects all information about the run state: slave_running, and the

  existence of the I/O thread (to stop/start it, you need this mutex).

  run_lock protects all information about the run state: slave_running, thd

  and the existence of the I/O thread to stop/start it, you need this mutex).

  data_lock protects some moving members of the struct: counters (log name,

  position) and relay log (MYSQL_LOG object).