GitLab

 - Backport patch from 5.0


sql/mysqld.cc:
  Add else case to set "lower_case_file_system" also when lower_case_table_names are 1

By default lowercase_table_names are 1 on windows, and that makes the initialization of lower_case_filesystem variable to be skipped.
Add an else bracnh that does the initialization.


sql/mysqld.cc:
  Intialize the "lowercase_filesystem" variable also when "lowercase_table_names" is 1

- Adapt it to work with the handlerton class


sql/handler.cc:
  Only create new MERGE handler if merge engine is enabled
sql/mysql_priv.h:
  Use the myisammrg_hton.state field for the have_merge_db option
sql/mysqld.cc:
  Handle the OPT_MERGE (--skip-merge) option case

to avoid the potential security problem.
(see bug #15195: Security Breach with MERGE table)

    Problem:
    mysqld --collation-server=xxx --character-set-server=yyy
    didn't work as expected: collation_server was set not to xxx,
    but to the default collation of character set "yyy".
    
    With different argument order it worked as expected:
    mysqld --character-set-server=yyy --collation-server=yyy 
    
    Fix:
    initializate default_collation_name to 0
    when processing --character-set-server
    only if --collation-server has not been specified
    in command line.



mysql-test/r/ctype_ucs2_def.result:
  Adding test case
mysql-test/t/ctype_ucs2_def-master.opt:
  Specifying variables in reverse order, to cover the bug.
mysql-test/t/ctype_ucs2_def.test:
  Adding test case
sql/mysqld.cc:
  Don't clear default_collation_name when processing 
      --character-set-server if collation has already
      been specified using --collation-server

include/my_libwrap.h:
  Changed includes to the header file.
mysys/my_libwrap.c:
  Added comment and .c file now takes needed includes from
  the corresponding .h file.
sql/mysqld.cc:
  Include this block from my_libwra.h now.
  Moved two variables out of the otherwise
  same block.

sql/mysqld.cc:
  Fix for Bug#18246 "compilation error with tcp_wrapper"
  
  Added wrapper functions.

            read buffer
Setting read buffer to values greater than SSIZE_MAX results in
unexpected behavior.

According to read(2) manual:
If count is greater than SSIZE_MAX, the result is unspecified.

Set upper limit for read_buffer_size and read_rnd_buffer_size to
SSIZE_MAX.


include/my_global.h:
  Define SSIZE_MAX if not defined.
sql/mysqld.cc:
  Set upper limit for read_buffer_size and read_rnd_buffer_size to
  SSIZE_MAX.

Two threads both try a shutdown sequence which creates a race to the
de-init/free of certain resources.

This exists in similar form in the client as 17926: "mysql.exe crashes
when ctrl-c is pressed in windows."


sql/mysqld.cc:
  We have three potential ways of hitting the iceberg:
  - unireg_end()   has basic de-init
  - unireg_abort() has extended de-init
  - main() has a de-init sequence similar to unireg_abort()
  
  In the Windows version of the server, Control-C is handled
  in a different thread from the one main() is in.  The main
  loop is told to end, then unireg_abort() is called.  Its
  de-init and that of main() will then race each other for
  mutex- and cond-var-destroys, free(), and finally exit().
  
  This patch introduces a special case for Windows that will eliminate
  the race by ending the signal-handler via unireg_end() instead if
  SIGINT is signalled.  This seems the least intrusive fix that still
  fixes user-visible behaviour.

Replaced COND_refresh with COND_global_read_lock becasue of a bug in NTPL threads when using different mutexes as arguments to pthread_cond_wait()
The original code caused a hang in FLUSH TABLES WITH READ LOCK in some circumstances because pthread_cond_broadcast() was not delivered to other threads.
This fixes:
Bug#16986: Deadlock condition with MyISAM tables
Bug#20048: FLUSH TABLES WITH READ LOCK causes a deadlock


mysql-test/r/flush.result:
  Added test case for deadlock with FLUSH TABLES WITH READ LOCK
mysql-test/r/lock_multi.result:
  Test for bug in LOCK TABLE + optimize table
mysql-test/t/flush.test:
  Added test case for deadlock with FLUSH TABLES WITH READ LOCK
mysql-test/t/lock_multi.test:
  Test for bug in LOCK TABLE + optimize table
sql/lock.cc:
  Replaced COND_refresh with COND_global_read_lock becasue of a bug in NTPL threads when using different mutexes as arguments to pthread_cond_wait()
  The original code caused a hang in FLUSH TABLES WITH READ LOCK in some circumstances because pthread_cond_broadcast() was not delivered to other threads
sql/mysql_priv.h:
  Added COND_global_read_lock
sql/mysqld.cc:
  Added COND_global_read_lock
sql/sql_base.cc:
  More DBUG statements
  Added a broadcast in remove_table_from_cache() to release any threads waiting in open

sql/mysqld.cc:
  Added --skip-external-locking to the --external-locking help.

 - A segfault occured when the function 'kill_server' called 
   'my_sigset' with signal number 0. 'my_sigset' is a macro which
   uses 'sigaction' to install the signal handler with an invalid
   signal number will on most platforms return EINVAL but yields
   a segfauilt on IRIX 6.5
 - The server crash was detected by mysqld_safe and it was restarted although
   a shutdown was requested. 
 - Semantics of kill_server(0) is not known, leaving it intact


include/my_pthread.h:
  Check return value from sigaction with a DBUG_ASSERT
  Also DBUG_ASSERT if signal number 0 is passed
sql/mysqld.cc:
  Don't call my_sigset if signo is 0

client/mysqlbinlog.cc:
  Now my_end() deallocates DBUG by default, but that fails in mysqlbinlog
  because of global destructors that use DBUG.
dbug/dbug.c:
  Add a facility to deallocate the debug stack, to avoid memory leak warnings
  in Valgrind.
include/my_dbug.h:
  Add a facility to deallocate the debug stack, to avoid memory leak warnings
  in Valgrind.
include/my_sys.h:
  Change my_end() to deallocate DBUG memory by default (can be disabled with
  MY_DONT_FREE_DBUG option).
libmysql/libmysql.c:
  Do not deallocate DBUG during cleanup.
mysys/my_init.c:
  Change my_end() to deallocate DBUG memory by default (can be disabled with
  MY_DONT_FREE_DBUG option).
sql/mysqld.cc:
  Add missing my_thread_end() call, seems to occasionally trigger a memory
  leak (not repeatable).

Bug #19606: ssl variables are not displayed in show variables
Bug #19616: log_queries_not_using_indexes  is not listed in show variables

  Make basedir, datadir, tmpdir, log_queries_not_using_indexes, ssl_ca,
  ssl_capath, ssl_cert, ssl_cipher, and ssl_key all available both from
  SHOW VARIABLES and as @@variables.

  As a side-effect of this change, log_queries_not_using_indexes can
  be changed at runtime (but only globally, not per-connection).


include/sslopt-longopts.h:
  Put options in alphabetical order
include/sslopt-vars.h:
  Allow define of SSL_VARS_NOT_STATIC to prevent variables from not being
  made static.
mysql-test/r/variables.result:
  Add new results
mysql-test/t/variables.test:
  Add new regression tests
sql/mysql_priv.h:
  Add extern for opt_log_queries_not_using_indexes
sql/mysqld.cc:
  Handle opt_log_queries_not_using_indexes as extern, and define
  SSL_VARS_NO_STATIC so they can be accessed outside of mysqld.cc
sql/set_var.cc:
  Handle basedir, datadir, tmpdir, log_queries_not_using_indexes, and
  various ssl settings so that they are accessible as server variables
  and listed in SHOW VARIABLES.
sql/set_var.h:
  Add new sys_var_constr_str_ptr class, for when we have a system variable
  that is only set via the command-line that is a pointer to a string.

Add function 'vio_end' that will cleanup resources allocated by vio and the components it uses.


include/violite.h:
  Import patch warnings.patch
libmysql/libmysql.c:
  Import patch warnings.patch
sql/mysqld.cc:
  Import patch warnings.patch
vio/test-ssl.c:
  Import patch warnings.patch
vio/test-sslclient.c:
  Import patch warnings.patch
vio/test-sslserver.c:
  Import patch warnings.patch
vio/vio.c:
  Import patch warnings.patch
vio/viosslfactories.c:
  Import patch warnings.patch

Removed duplication (not a user-visible change)


sql/mysqld.cc:
  Two of the startup options for mysqld, innodb_commit_concurrency
  and innodb_thread_concurrency, were using the same enumeration
  value. Changed to give innodb_commit_concurrency it's own value.

sql/mysqld.cc:
  Fix a typo.

sql/ha_innodb.cc:
  Import patch foo2
sql/ha_innodb.h:
  Import patch foo2
sql/handler.cc:
  Import patch foo2
sql/handler.h:
  Import patch foo2
sql/mysqld.cc:
  Import patch foo2
sql/set_var.cc:
  Import patch foo2
sql/sql_class.h:
  Import patch foo2
sql/sql_repl.cc:
  Import patch foo2

sql/ha_innodb.cc:
  Import patch foo1
sql/ha_innodb.h:
  Import patch foo1
sql/handler.cc:
  Import patch foo1
sql/handler.h:
  Import patch foo1
sql/mysqld.cc:
  Import patch foo1
sql/set_var.cc:
  Import patch foo1
sql/sql_class.h:
  Import patch foo1
sql/sql_repl.cc:
  Import patch foo1

 - Add variable "have_dynamic_loading" and use it to check if the udf test should be run.


mysql-test/include/have_udf.inc:
  Use variable "have_dynamic_loading" to to determine if there is support for udf's
mysql-test/r/have_udf.require:
  Use variable "have_dynamic_loading" to to determine if there is support for udf's
mysql-test/t/disabled.def:
  Remove udf from disabled
sql/mysql_priv.h:
  Add have_dlopen variable
sql/mysqld.cc:
  Add have_dlopen variable
sql/set_var.cc:
  Add have_dlopen variable and name it "have_dynamic_loading" in "SHOW VARIABLES"

 Applied changes due Heikki Tuuri.


sql/mysqld.cc:
  Set innodb_thread_concurrency default value to 8 (Bug#15868).
   Changes due to Heikki Tuuri.

too many open statements". The patch adds a new global variable
@@max_prepared_stmt_count. This variable limits the total number
of prepared statements in the server. The default value of
@@max_prepared_stmt_count is 16382. 16382 small statements
(a select against 3 tables with GROUP, ORDER and LIMIT) consume 
100MB of RAM. Once this limit has been reached, the server will 
refuse to prepare a new statement and return ER_UNKNOWN_ERROR 
(unfortunately, we can't add new errors to 4.1 without breaking 5.0). The limit is changeable after startup
and can accept any value from 0 to 1 million. In case
the new value of the limit is less than the current
statement count, no new statements can be added, while the old
still can be used. Additionally, the current count of prepared 
statements is now available through a global read-only variable 
@@prepared_stmt_count.


mysql-test/r/ps.result:
  Test results fixed (a test case for Bug#16365)
mysql-test/t/ps.test:
  A test case for Bug#16365 "Prepared Statements: DoS with too many 
  open statements". Also fix statement leaks in other tests.
sql/mysql_priv.h:
  Add declarations for new global variables.
sql/mysqld.cc:
  Add definitions of max_prepared_stmt_count, prepared_stmt_count.
sql/set_var.cc:
  Implement support for @@prepared_stmt_count and 
  @@max_prepared_stmt_count. Currently these variables are queried
  without acquiring LOCK_prepared_stmt_count due to limitations of
  the set_var/sys_var class design. Updates are, however, protected 
  with a lock.
sql/set_var.h:
  New declarations to add support for @@max_prepared_stmt_count.
  Implement a new class, where the lock to be used when updating
  a variable is a parameter.
sql/sql_class.cc:
  Add accounting of the total number of prepared statements in the
  server to the methods of Statement_map.
sql/sql_class.h:
  Add accounting of the total number of prepared statements in the
  server to the methods of Statement_map.
sql/sql_prepare.cc:
  Statement_map::insert will now send a message in case of an
  error.

After FLUSH STATUS max_used_connections was reset to 0, and haven't
been updated while cached threads were reused, until the moment a new
thread was created.

The first suggested fix from original bug report was implemented:

  a) On flushing the status, set max_used_connections to
  threads_connected, not to 0.

  b) Check if it is necessary to increment max_used_connections when
  taking a thread from the cache as well as when creating new threads


mysql-test/r/status.result:
  Add result for bug#15933.
mysql-test/t/status.test:
  Add test case for bug#15933.
  Fixed typo.
sql/mysql_priv.h:
  Add declaration of refresh_status(), which is now external.
sql/mysqld.cc:
  Remove start_cached_thread() (code moved directly into create_new_thread()).
  Add comment for create_new_thread ().
  In create_new_thread() update max_used_connections when creating new thread
  and when reusing the cached one.
  Move refresh_status() from sql/sql_parse.cc here, on refresh set
  max_used_connections to the current number of connections.
sql/sql_parse.cc:
  refresh_status() moved to sql/mysqld.cc.

  ¨MySQL server crashes if you try to access to InnoDB table¨
  crash caused by schizophrenic mysqld - 2 memory locations for logically same function
  with conflicting values.
  Fixed by backporting from 5.1 changes to have_xyz_db declarations.


sql/mysql_priv.h:
  Backport have_xyz_db changes from 5.1
sql/mysqld.cc:
  Backport have_xyz_db changes from 5.1

  Backport option description change from 5.1 to 5.0.


sql/mysqld.cc:
  Backport option description change from 5.1 to 5.0.

Remove duplicate code
Merge common functions
Enforce MySQL coding standard


include/violite.h:
  Cleanup SSL implementation
sql-common/client.c:
  Cleanup SSL implementation
sql/mysql_priv.h:
  Cleanup SSL implementation
sql/mysqld.cc:
  Cleanup SSL implementation
sql/sql_acl.cc:
  Cleanup SSL implementation
vio/vio.c:
  Cleanup SSL implementation
vio/vio_priv.h:
  Cleanup SSL implementation
vio/viossl.c:
  Cleanup SSL implementation
vio/viosslfactories.c:
  Cleanup SSL implementation

   After the ChangeSet 1.1892.20.1 2005/08/24 (Bug #12562) SYSDATE() is not an alias
   of NOW() and is unsafe for replication.
   `SYSDATE()' backward compatible aliasing clashes with the idea #12562 
   fix. To make it safe-replicatable we have to either use RBR or to restore
   the pre-5.0 style.
   --sysdate-is-now command line flag was introduced to provide backward compatibility.


sql/mysqld.cc:
  New option to force SYSDATE's backward compatible with 4.1 aliasing to NOW (not default)
sql/sql_class.h:
  new slot to fill at init time and check at parse
sql/sql_yacc.yy:
  calling NOW's branches when --sysdate-is-now
mysql-test/r/sysdate_is_now.result:
  New BitKeeper file ``mysql-test/r/sysdate_is_now.result''
mysql-test/t/sysdate_is_now-master.opt:
  New BitKeeper file ``mysql-test/t/sysdate_is_now-master.opt''
mysql-test/t/sysdate_is_now.test:
  New BitKeeper file ``mysql-test/t/sysdate_is_now.test''

This patch does 1) fix my build breakage  2) Complete the removal of all symbols which could clash with another parser. 


sql/mysql_priv.h:
  Porting update
sql/mysqld.cc:
  Porting update
sql/sp.cc:
  Porting update
sql/sql_lex.cc:
  Porting update
sql/sql_lex.h:
  Porting update
sql/sql_parse.cc:
  Porting update
sql/sql_prepare.cc:
  Portinng update
sql/sql_trigger.cc:
  Porting update
sql/sql_view.cc:
  Porting update

(Needed for "list of pushes" web page and autopush)


include/mysql.h:
  Fix to embedded server to be able to run tests on it
libmysql/libmysql.c:
  Fix to embedded server to be able to run tests on it
libmysqld/emb_qcache.cc:
  Fix to embedded server to be able to run tests on it
libmysqld/embedded_priv.h:
  Fix to embedded server to be able to run tests on it
libmysqld/lib_sql.cc:
  Fix to embedded server to be able to run tests on it
libmysqld/libmysqld.c:
  Fix to embedded server to be able to run tests on it
mysql-test/mysql-test-run.sh:
  Fix to embedded server to be able to run tests on it
mysql-test/r/binlog.result:
  Updated test for embedded server
mysql-test/r/ctype_cp932.result:
  Updated test for embedded server
mysql-test/r/innodb.result:
  Updated test for embedded server
mysql-test/r/mysqltest.result:
  Updated test for embedded server
mysql-test/r/query_cache.result:
  Updated test for embedded server
mysql-test/r/query_cache_notembedded.result:
  Updated test for embedded server
mysql-test/r/sp-error.result:
  Updated test for embedded server
mysql-test/r/sp.result:
  Updated test for embedded server
mysql-test/r/subselect.result:
  Updated test for embedded server
mysql-test/r/view.result:
  Updated test for embedded server
mysql-test/r/view_grant.result:
  Updated test for embedded server
mysql-test/t/backup.test:
  Updated test for embedded server
mysql-test/t/binlog.test:
  Updated test for embedded server
mysql-test/t/blackhole.test:
  Updated test for embedded server
mysql-test/t/compress.test:
  Updated test for embedded server
mysql-test/t/ctype_cp932.test:
  Updated test for embedded server
mysql-test/t/delayed.test:
  Updated test for embedded server
mysql-test/t/handler.test:
  Updated test for embedded server
mysql-test/t/innodb.test:
  Updated test for embedded server
mysql-test/t/mysql.test:
  Updated test for embedded server
mysql-test/t/mysql_client_test.test:
  Updated test for embedded server
mysql-test/t/mysqltest.test:
  Updated test for embedded server
mysql-test/t/query_cache.test:
  Updated test for embedded server
mysql-test/t/query_cache_notembedded.test:
  Updated test for embedded server
mysql-test/t/read_only.test:
  Updated test for embedded server
mysql-test/t/skip_grants.test:
  Updated test for embedded server
mysql-test/t/sp-destruct.test:
  Updated test for embedded server
mysql-test/t/sp-error.test:
  Updated test for embedded server
mysql-test/t/sp-threads.test:
  Updated test for embedded server
mysql-test/t/sp.test:
  Updated test for embedded server
mysql-test/t/subselect.test:
  Updated test for embedded server
mysql-test/t/temp_table.test:
  Updated test for embedded server
mysql-test/t/view.test:
  Updated test for embedded server
mysql-test/t/view_grant.test:
  Updated test for embedded server
mysql-test/t/wait_timeout.test:
  Updated test for embedded server
mysys/mf_dirname.c:
  Review fix: Don't access data outside of array
mysys/my_bitmap.c:
  Remove compiler warnings
scripts/mysql_fix_privilege_tables.sql:
  Add flush privileges to .sql script so that one doesn't have to reboot mysqld when one runs the mysql_fix_privilege_script
sql-common/client.c:
  Updated test for embedded server
sql/item.cc:
  Remove DBUG_PRINT statement that can cause crashes when running with --debug
sql/mysqld.cc:
  Fix to embedded server to be able to run tests on it
sql/protocol.cc:
  Fix to embedded server to be able to run tests on it
  (Trivial reconstruction of code)
sql/protocol.h:
  Fix to embedded server to be able to run tests on it
sql/sql_base.cc:
  Better comment
sql/sql_class.cc:
  Fix to embedded server to be able to run tests on it
sql/sql_class.h:
  Fix to embedded server to be able to run tests on it
sql/sql_cursor.cc:
  Fix to embedded server to be able to run tests on it
sql/sql_parse.cc:
  Fix to embedded server to be able to run tests on it
  Don't crash for disabled commands when using embedded server
sql/sql_prepare.cc:
  Fix to embedded server to be able to run tests on it
mysql-test/r/ctype_cp932_notembedded.result:
  New BitKeeper file ``mysql-test/r/ctype_cp932_notembedded.result''
mysql-test/r/innodb_notembedded.result:
  New BitKeeper file ``mysql-test/r/innodb_notembedded.result''
mysql-test/r/sp.result.orig:
  New BitKeeper file ``mysql-test/r/sp.result.orig''
mysql-test/r/sp_notembedded.result:
  New BitKeeper file ``mysql-test/r/sp_notembedded.result''
mysql-test/r/subselect_notembedded.result:
  New BitKeeper file ``mysql-test/r/subselect_notembedded.result''
mysql-test/t/ctype_cp932_notembedded.test:
  New BitKeeper file ``mysql-test/t/ctype_cp932_notembedded.test''
mysql-test/t/innodb_notembedded.test:
  New BitKeeper file ``mysql-test/t/innodb_notembedded.test''
mysql-test/t/sp.test.orig:
  New BitKeeper file ``mysql-test/t/sp.test.orig''
mysql-test/t/sp_notembedded.test:
  New BitKeeper file ``mysql-test/t/sp_notembedded.test''
mysql-test/t/subselect_notembedded.test:
  New BitKeeper file ``mysql-test/t/subselect_notembedded.test''

  Fix out of order entries in option struct.


sql/mysqld.cc:
  Fix out of order entries in option struct.

if the function, invoked in a non-binlogged caller (e.g. SELECT, DO), failed half-way on the master,
slave would stop and complain that error code between him and master mismatch. 
To solve this, when a stored function is invoked in a non-binlogged caller (e.g. SELECT, DO), we binlog the function
call as SELECT instead of as DO (see revision comment of sp_head.cc for more).
And: minor wording change in the help text.
This cset will cause conflicts in 5.1, I'll merge.


mysql-test/r/rpl_sp.result:
  result update
mysql-test/t/rpl_sp-slave.opt:
  bug just fixed so option not needed
mysql-test/t/rpl_sp.test:
  test for more half-failed functions with DO and SELECT, to test the bug of this changeset.
  cleanup at the end.
sql/mysqld.cc:
  function -> stored function (change suggested by Paul)
sql/sp_head.cc:
  When a function updates data and is called from a non-binlogged statement (SELECT, DO), we binlog it 
  as SELECT myfunc(), and not DO myfunc() like before.

  Backporting character_set_filesystem from 5.0 to 5.1.


sql/mysqld.cc:
  Backporting character_set_filesystem from 5.0 to 5.1.
sql/set_var.cc:
  Backporting character_set_filesystem from 5.0 to 5.1.
sql/set_var.h:
  Backporting character_set_filesystem from 5.0 to 5.1.
sql/sql_class.cc:
  Backporting character_set_filesystem from 5.0 to 5.1.
sql/sql_class.h:
  Backporting character_set_filesystem from 5.0 to 5.1.
sql/sql_yacc.yy:
  Backporting character_set_filesystem from 5.0 to 5.1.
mysql-test/t/variables.test:
  Backporting character_set_filesystem from 5.0 to 5.1.
mysql-test/r/variables.result:
  Backporting character_set_filesystem from 5.0 to 5.1.

 Make innodb_thread_concurrency 0 by default, and extend the usable
 range from 0 to 1000 (0 disables the thread throttling).


innobase/include/srv0srv.h:
  Remove SRV_CONCURRENCY_THRESHOLD. The magic value is 0 from now on.
innobase/srv/srv0srv.c:
  Remove SRV_CONCURRENCY_THRESHOLD. The magic value is 0 from now on.
mysql-test/r/innodb.result:
  innodb_thread_concurrency is 0..1000 now, and the default is 0.
mysql-test/t/innodb.test:
  innodb_thread_concurrency is 0..1000 now.
sql/ha_innodb.cc:
  Remove SRV_CONCURRENCY_THRESHOLD. The magic value is 0 from now on.
sql/mysqld.cc:
  innodb_thread_concurrency is 0..1000 now, and 0 (the default)
   disables the thread throttling.

 Make innodb_flush_log_at_trx_commit a settable global variable.


innobase/include/srv0srv.h:
  Change the type of srv_flush_log_at_trx_commit.
innobase/srv/srv0srv.c:
  Change the type of srv_flush_log_at_trx_commit.
sql/ha_innodb.cc:
  Remove innobase_flush_log_at_trx_commit.
   (Set srv_flush_log_at_trx_commit directly).
sql/ha_innodb.h:
  Remove innobase_flush_log_at_trx_commit.
   (Set srv_flush_log_at_trx_commit directly).
sql/mysqld.cc:
  Bypass the variable innobase innobase_flush_log_at_trx_commit.
sql/set_var.cc:
  Make innodb_flush_log_at_trx_commit a settable global variable.

Changes to Netware specific mysqld_safe.c


include/config-netware.h:
  NetWare specific change to increase thread stack size.
innobase/os/os0thread.c:
  NetWare specific change to increase thread stack size.
netware/mysqld_safe.c:
  NetWare specific change to make multiple mysqld_safe instances
  work when called through a NCF file.
sql/mysqld.cc:
  NetWare specific change to increase thread stack size.

If the server is started with --port-open-timeout=# 
retry to bind the TCP port if it fails with EADDRINUSE.
Use the new option in mysql-test-run.sh.


mysql-test/mysql-test-run.sh:
  Bug#15591 Occasional test suite abort due to port in use
  Use the new option --port-open-timeout=# for server starts.
  This should fix occasionally late closed TCP ports.
sql/mysqld.cc:
  Bug#15591 Occasional test suite abort due to port in use
  If the server is started with --port-open-timeout=# 
  retry to bind the TCP port if it fails with EADDRINUSE.