mariadb/mysql-test/include/check_concurrent_insert.inc
Dmitry Lenev 78c6a8ca30 A 5.1-only version of fix for bug #46947 "Embedded SELECT
without FOR UPDATE is causing a lock".

SELECT statements with subqueries referencing InnoDB tables
were acquiring shared locks on rows in these tables when they
were executed in REPEATABLE-READ mode and with statement or
mixed mode binary logging turned on.

This was a regression which were introduced when fixing
bug 39843.

The problem was that for tables belonging to subqueries
parser set TL_READ_DEFAULT as a lock type. In cases when
statement/mixed binary logging at open_tables() time this
type of lock was converted to TL_READ_NO_INSERT lock at
open_tables() time and caused InnoDB engine to acquire
shared locks on reads from these tables. Although in some
cases such behavior was correct (e.g. for subqueries in
DELETE) in case of SELECT it has caused unnecessary locking.

This patch implements minimal version of the fix for the
specific problem described in the bug-report which supposed
to be not too risky for pushing into 5.1 tree.
The 5.5 tree already contains a more appropriate solution
which also addresses other related issues like bug 53921
"Wrong locks for SELECTs used stored functions may lead
to broken SBR".

This patch tries to solve the problem by ensuring that
TL_READ_DEFAULT lock which is set in the parser for
tables participating in subqueries at open_tables()
time is interpreted as TL_READ_NO_INSERT or TL_READ.
TL_READ is used only if we know that this is a SELECT
and that this particular table is not used by a stored
function.

Test coverage is added for both InnoDB and MyISAM.

This patch introduces an "incompatible" change in locking
scheme for subqueries used in SELECT ... FOR UPDATE and
SELECT .. IN SHARE MODE.

In 4.1 (as well as in 5.0 and 5.1 before fix for bug 39843)
the server would use a snapshot InnoDB read for subqueries
in SELECT FOR UPDATE and SELECT .. IN SHARE MODE statements,
regardless of whether the binary log is on or off.

If the user required a different type of read (i.e. locking
read), he/she could request so explicitly by providing FOR
UPDATE/IN SHARE MODE clause for each individual subquery.

The patch for bug 39843 broke this behaviour (which was not
documented or tested), and started to use locking reads for
all subqueries in SELECT ... FOR UPDATE/IN SHARE MODE.
This patch restores 4.1 behaviour.

This patch should be mostly null-merged into 5.5 tree.
2010-05-28 00:07:40 +04:00

96 lines
2.7 KiB
PHP

#
# SUMMARY
# Check if statement reading table '$table' allows concurrent
# inserts in it.
#
# PARAMETERS
# $table Table in which concurrent inserts should be allowed.
# $con_aux1 Name of the first auxiliary connection to be used by this
# script.
# $con_aux2 Name of the second auxiliary connection to be used by this
# script.
# $statement Statement to be checked.
# $restore_table Table which might be modified by statement to be checked
# and thus needs backing up before its execution and
# restoring after it (can be empty).
#
# EXAMPLE
# lock_sync.test
#
--disable_result_log
--disable_query_log
# Reset DEBUG_SYNC facility for safety.
set debug_sync= "RESET";
if (`SELECT '$restore_table' <> ''`)
{
--eval create temporary table t_backup select * from $restore_table;
}
connection $con_aux1;
set debug_sync='after_lock_tables_takes_lock SIGNAL parked WAIT_FOR go';
--send_eval $statement;
connection $con_aux2;
set debug_sync='now WAIT_FOR parked';
--send_eval insert into $table (i) values (0);
--enable_result_log
--enable_query_log
connection default;
# Wait until concurrent insert is successfully executed while
# statement being checked has its tables locked.
# We use wait_condition.inc instead of simply reaping
# concurrent insert here in order to avoid deadlocks if test
# fails and to time out gracefully instead.
let $wait_condition=
select count(*) = 0 from information_schema.processlist
where info = "insert into $table (i) values (0)";
--source include/wait_condition.inc
--disable_result_log
--disable_query_log
if ($success)
{
# Apparently concurrent insert was successfully executed.
# To be safe against wait_condition.inc succeeding due to
# races let us first reap concurrent insert to ensure that
# it has really been successfully executed.
connection $con_aux2;
--reap
connection default;
set debug_sync= 'now SIGNAL go';
connection $con_aux1;
--reap
connection default;
--echo Success: '$statement' allows concurrent inserts into '$table'.
}
if (!$success)
{
# Waiting has timed out. Apparently concurrent insert was blocked.
# So to be able to continue we need to end our statement first.
set debug_sync= 'now SIGNAL go';
connection $con_aux1;
--reap
connection $con_aux2;
--reap
connection default;
--echo Error: '$statement' doesn't allow concurrent inserts into '$table'!
}
--eval delete from $table where i = 0;
if (`SELECT '$restore_table' <> ''`)
{
--eval truncate table $restore_table;
--eval insert into $restore_table select * from t_backup;
drop temporary table t_backup;
}
# Clean-up. Reset DEBUG_SYNC facility after use.
set debug_sync= "RESET";
--enable_result_log
--enable_query_log