------------------------------------------------------------
revno: 3035.4.1
committer: Davi Arnaut <Davi.Arnaut@Sun.COM>
branch nick: 39897-6.0
timestamp: Thu 2009-01-15 12:17:57 -0200
message:
Bug#39897: lock_multi fails in pushbuild: timeout waiting for processlist
The problem is that relying on the "Table lock" thread state in
its current position to detect that a thread is waiting on a lock
is race prone. The "Table lock" state change happens before the
thread actually tries to grab a lock on a table.
The solution is to move the "Table lock" state so that its set
only when a thread is actually going to wait for a lock. The state
change happens after the thread fails to grab the lock (because it
is owned by other thread) and proceeds to wait on a condition.
This is considered part of work related to WL#4284 "Transactional
DDL locking"
Warning: this patch contains an incompatible change.
When waiting on a lock in thr_lock.c, the server used to display "Locked"
processlist state. After this patch, the state is "Table lock".
The new state was actually intended to be display since year 2002,
when Monty added it. But up until removal of thd->locked boolean
member, this state was ignored by SHOW PROCESSLIST code.
Disabled execution of this test for embedded server until fix for
bug 41971 'Thread state on embedded server is always "Writing to net"'
is back-ported to this tree.
Concurrent execution of statements which require non-table-level
write locks on several instances of the same table (such as
SELECT ... FOR UPDATE which uses same InnoDB table twice or a DML
statement which invokes trigger which tries to update same InnoDB
table directly and through stored function) and statements which
required table-level locks on this table (e.g. LOCK TABLE ... WRITE,
ALTER TABLE, ...) might have resulted in a deadlock.
The problem occured when a thread tried to acquire write lock
(TL_WRITE_ALLOW_WRITE) on the table but had to wait since there was
a pending write lock (TL_WRITE, TL_WRITE_ALLOW_READ) on this table
and we failed to detect that this thread already had another instance
of write lock on it (so in fact we were trying to acquire recursive
lock) because there was also another thread holding write lock on the
table (also TL_WRITE_ALLOW_WRITE). When the latter thread released
its lock neither the first thread nor the thread trying to acquire
TL_WRITE/TL_WRITE_ALLOW_READ were woken up (as table was still write
locked by the first thread) so we ended up with a deadlock.
This patch solves this problem by ensuring that thread which
already has write lock on the table won't wait when it tries
to acquire second write lock on the same table.
