Avoid spurious deadlocks when upgrading a tuple lock

When two (or more) transactions are waiting for transaction T1 to release a
tuple-level lock, and transaction T1 upgrades its lock to a higher level, a
spurious deadlock can be reported among the waiting transactions when T1
finishes.  The simplest example case seems to be:

T1: select id from job where name = 'a' for key share;
Y: select id from job where name = 'a' for update; -- starts waiting for X
Z: select id from job where name = 'a' for key share;
T1: update job set name = 'b' where id = 1;
Z: update job set name = 'c' where id = 1; -- starts waiting for X
T1: rollback;

At this point, transaction Y is rolled back on account of a deadlock: Y
holds the heavyweight tuple lock and is waiting for the Xmax to be released,
while Z holds part of the multixact and tries to acquire the heavyweight
lock (per protocol) and goes to sleep; once X releases its part of the
multixact, Z is awakened only to be put back to sleep on the heavyweight
lock that Y is holding while sleeping.  Kaboom.

This can be avoided by having Z skip the heavyweight lock acquisition.  As
far as I can see, the biggest downside is that if there are multiple Z
transactions, the order in which they resume after X finishes is not
guaranteed.

Backpatch to 9.6.  The patch applies cleanly on 9.5, but the new tests don't
work there (because isolationtester is not smart enough), so I'm not going
to risk it.

Author: Oleksii Kliukin
Discussion: https://postgr.es/m/B9C9D7CD-EB94-4635-91B6-E558ACEC0EC3@hintbits.com

Author: alvherre

src/backend/access/heap/README.tuplock

@@ -36,6 +36,16 @@ do LockTuple as well, if there is any conflict, to ensure that they don't
 starve out waiting exclusive-lockers.  However, if there is not any active
 conflict for a tuple, we don't incur any extra overhead.
 
+We make an exception to the above rule for those lockers that already hold
+some lock on a tuple and attempt to acquire a stronger one on it.  In that
+case, we skip the LockTuple() call even when there are conflicts, provided
+that the target tuple is being locked, updated or deleted by multiple sessions
+concurrently.  Failing to skip the lock would risk a deadlock, e.g., between a
+session that was first to record its weaker lock in the tuple header and would
+be waiting on the LockTuple() call to upgrade to the stronger lock level, and
+another session that has already done LockTuple() and is waiting for the first
+session transaction to release its tuple header-level lock.
+
 We provide four levels of tuple locking strength: SELECT FOR UPDATE obtains an
 exclusive lock which prevents any kind of modification of the tuple. This is
 the lock level that is implicitly taken by DELETE operations, and also by

src/backend/access/heap/heapam.c

@@ -95,7 +95,7 @@ static void GetMultiXactIdHintBits(MultiXactId multi, uint16 *new_infomask,
 static TransactionId MultiXactIdGetUpdateXid(TransactionId xmax,
 											 uint16 t_infomask);
 static bool DoesMultiXactIdConflict(MultiXactId multi, uint16 infomask,
-									LockTupleMode lockmode);
+									LockTupleMode lockmode, bool *current_is_member);
 static void MultiXactIdWait(MultiXactId multi, MultiXactStatus status, uint16 infomask,
 							Relation rel, ItemPointer ctid, XLTW_Oper oper,
 							int *remaining);
@@ -2547,15 +2547,20 @@ heap_delete(Relation relation, ItemPointer tid,
 		 */
 		if (infomask & HEAP_XMAX_IS_MULTI)
 		{
-			/* wait for multixact */
+			bool		current_is_member = false;
+
 			if (DoesMultiXactIdConflict((MultiXactId) xwait, infomask,
-										LockTupleExclusive))
+										LockTupleExclusive, &current_is_member))
 			{
 				LockBuffer(buffer, BUFFER_LOCK_UNLOCK);
 
-				/* acquire tuple lock, if necessary */
-				heap_acquire_tuplock(relation, &(tp.t_self), LockTupleExclusive,
-									 LockWaitBlock, &have_tuple_lock);
+				/*
+				 * Acquire the lock, if necessary (but skip it when we're
+				 * requesting a lock and already have one; avoids deadlock).
+				 */
+				if (!current_is_member)
+					heap_acquire_tuplock(relation, &(tp.t_self), LockTupleExclusive,
+										 LockWaitBlock, &have_tuple_lock);
 
 				/* wait for multixact */
 				MultiXactIdWait((MultiXactId) xwait, MultiXactStatusUpdate, infomask,
@@ -3126,15 +3131,20 @@ heap_update(Relation relation, ItemPointer otid, HeapTuple newtup,
 		{
 			TransactionId update_xact;
 			int			remain;
+			bool		current_is_member = false;
 
 			if (DoesMultiXactIdConflict((MultiXactId) xwait, infomask,
-										*lockmode))
+										*lockmode, &current_is_member))
 			{
 				LockBuffer(buffer, BUFFER_LOCK_UNLOCK);
 
-				/* acquire tuple lock, if necessary */
-				heap_acquire_tuplock(relation, &(oldtup.t_self), *lockmode,
-									 LockWaitBlock, &have_tuple_lock);
+				/*
+				 * Acquire the lock, if necessary (but skip it when we're
+				 * requesting a lock and already have one; avoids deadlock).
+				 */
+				if (!current_is_member)
+					heap_acquire_tuplock(relation, &(oldtup.t_self), *lockmode,
+										 LockWaitBlock, &have_tuple_lock);
 
 				/* wait for multixact */
 				MultiXactIdWait((MultiXactId) xwait, mxact_status, infomask,
@@ -4028,6 +4038,7 @@ heap_lock_tuple(Relation relation, HeapTuple tuple,
 		uint16		infomask;
 		uint16		infomask2;
 		bool		require_sleep;
+		bool		skip_tuple_lock;
 		ItemPointerData t_ctid;
 
 		/* must copy state data before unlocking buffer */
@@ -4053,6 +4064,7 @@ heap_lock_tuple(Relation relation, HeapTuple tuple,
 		if (first_time)
 		{
 			first_time = false;
+			skip_tuple_lock = false;
 
 			if (infomask & HEAP_XMAX_IS_MULTI)
 			{
@@ -4081,6 +4093,21 @@ heap_lock_tuple(Relation relation, HeapTuple tuple,
 						result = TM_Ok;
 						goto out_unlocked;
 					}
+					else
+					{
+						/*
+						 * Disable acquisition of the heavyweight tuple lock.
+						 * Otherwise, when promoting a weaker lock, we might
+						 * deadlock with another locker that has acquired the
+						 * heavyweight tuple lock and is waiting for our
+						 * transaction to finish.
+						 *
+						 * Note that in this case we still need to wait for
+						 * the multixact if required, to avoid acquiring
+						 * conflicting locks.
+						 */
+						skip_tuple_lock = true;
+					}
 				}
 
 				if (members)
@@ -4235,7 +4262,7 @@ heap_lock_tuple(Relation relation, HeapTuple tuple,
 			if (infomask & HEAP_XMAX_IS_MULTI)
 			{
 				if (!DoesMultiXactIdConflict((MultiXactId) xwait, infomask,
-											 mode))
+											 mode, NULL))
 				{
 					/*
 					 * No conflict, but if the xmax changed under us in the
@@ -4312,13 +4339,15 @@ heap_lock_tuple(Relation relation, HeapTuple tuple,
 			/*
 			 * Acquire tuple lock to establish our priority for the tuple, or
 			 * die trying.  LockTuple will release us when we are next-in-line
-			 * for the tuple.  We must do this even if we are share-locking.
+			 * for the tuple.  We must do this even if we are share-locking,
+			 * but not if we already have a weaker lock on the tuple.
 			 *
 			 * If we are forced to "start over" below, we keep the tuple lock;
 			 * this arranges that we stay at the head of the line while
 			 * rechecking tuple state.
 			 */
-			if (!heap_acquire_tuplock(relation, tid, mode, wait_policy,
+			if (!skip_tuple_lock &&
+				!heap_acquire_tuplock(relation, tid, mode, wait_policy,
 									  &have_tuple_lock))
 			{
 				/*
@@ -6516,10 +6545,13 @@ HeapTupleGetUpdateXid(HeapTupleHeader tuple)
  * tuple lock of the given strength?
  *
  * The passed infomask pairs up with the given multixact in the tuple header.
+ *
+ * If current_is_member is not NULL, it is set to 'true' if the current
+ * transaction is a member of the given multixact.
  */
 static bool
 DoesMultiXactIdConflict(MultiXactId multi, uint16 infomask,
-						LockTupleMode lockmode)
+						LockTupleMode lockmode, bool *current_is_member)
 {
 	int			nmembers;
 	MultiXactMember *members;
@@ -6540,15 +6572,24 @@ DoesMultiXactIdConflict(MultiXactId multi, uint16 infomask,
 			TransactionId memxid;
 			LOCKMODE	memlockmode;
 
-			memlockmode = LOCKMODE_from_mxstatus(members[i].status);
+			if (result && (current_is_member == NULL || *current_is_member))
+				break;
 
-			/* ignore members that don't conflict with the lock we want */
-			if (!DoLockModesConflict(memlockmode, wanted))
-				continue;
+			memlockmode = LOCKMODE_from_mxstatus(members[i].status);
 
-			/* ignore members from current xact */
+			/* ignore members from current xact (but track their presence) */
 			memxid = members[i].xid;
 			if (TransactionIdIsCurrentTransactionId(memxid))
+			{
+				if (current_is_member != NULL)
+					*current_is_member = true;
+				continue;
+			}
+			else if (result)
+				continue;
+
+			/* ignore members that don't conflict with the lock we want */
+			if (!DoLockModesConflict(memlockmode, wanted))
 				continue;
 
 			if (ISUPDATE_from_mxstatus(members[i].status))
@@ -6567,10 +6608,11 @@ DoesMultiXactIdConflict(MultiXactId multi, uint16 infomask,
 			/*
 			 * Whatever remains are either live lockers that conflict with our
 			 * wanted lock, and updaters that are not aborted.  Those conflict
-			 * with what we want, so return true.
+			 * with what we want.  Set up to return true, but keep going to
+			 * look for the current transaction among the multixact members,
+			 * if needed.
 			 */
 			result = true;
-			break;
 		}
 		pfree(members);
 	}

src/test/isolation/expected/tuplelock-upgrade-no-deadlock.out

@@ -0,0 +1,150 @@
+Parsed test spec with 3 sessions
+
+starting permutation: s1_share s2_for_update s3_share s3_for_update s1_rollback s3_rollback s2_rollback
+step s1_share: select id from tlu_job where id = 1 for share;
+id             
+
+1              
+step s2_for_update: select id from tlu_job where id = 1 for update; <waiting ...>
+step s3_share: select id from tlu_job where id = 1 for share;
+id             
+
+1              
+step s3_for_update: select id from tlu_job where id = 1 for update; <waiting ...>
+step s1_rollback: rollback;
+step s3_for_update: <... completed>
+id             
+
+1              
+step s3_rollback: rollback;
+step s2_for_update: <... completed>
+id             
+
+1              
+step s2_rollback: rollback;
+
+starting permutation: s1_keyshare s2_for_update s3_keyshare s1_update s3_update s1_rollback s3_rollback s2_rollback
+step s1_keyshare: select id from tlu_job where id = 1 for key share;
+id             
+
+1              
+step s2_for_update: select id from tlu_job where id = 1 for update; <waiting ...>
+step s3_keyshare: select id from tlu_job where id = 1 for key share;
+id             
+
+1              
+step s1_update: update tlu_job set name = 'b' where id = 1;
+step s3_update: update tlu_job set name = 'c' where id = 1; <waiting ...>
+step s1_rollback: rollback;
+step s3_update: <... completed>
+step s3_rollback: rollback;
+step s2_for_update: <... completed>
+id             
+
+1              
+step s2_rollback: rollback;
+
+starting permutation: s1_keyshare s2_for_update s3_keyshare s1_update s3_update s1_commit s3_rollback s2_rollback
+step s1_keyshare: select id from tlu_job where id = 1 for key share;
+id             
+
+1              
+step s2_for_update: select id from tlu_job where id = 1 for update; <waiting ...>
+step s3_keyshare: select id from tlu_job where id = 1 for key share;
+id             
+
+1              
+step s1_update: update tlu_job set name = 'b' where id = 1;
+step s3_update: update tlu_job set name = 'c' where id = 1; <waiting ...>
+step s1_commit: commit;
+step s3_update: <... completed>
+step s3_rollback: rollback;
+step s2_for_update: <... completed>
+id             
+
+1              
+step s2_rollback: rollback;
+
+starting permutation: s1_keyshare s2_for_update s3_keyshare s3_delete s1_rollback s3_rollback s2_rollback
+step s1_keyshare: select id from tlu_job where id = 1 for key share;
+id             
+
+1              
+step s2_for_update: select id from tlu_job where id = 1 for update; <waiting ...>
+step s3_keyshare: select id from tlu_job where id = 1 for key share;
+id             
+
+1              
+step s3_delete: delete from tlu_job where id = 1; <waiting ...>
+step s1_rollback: rollback;
+step s3_delete: <... completed>
+step s3_rollback: rollback;
+step s2_for_update: <... completed>
+id             
+
+1              
+step s2_rollback: rollback;
+
+starting permutation: s1_keyshare s2_for_update s3_keyshare s3_delete s1_rollback s3_commit s2_rollback
+step s1_keyshare: select id from tlu_job where id = 1 for key share;
+id             
+
+1              
+step s2_for_update: select id from tlu_job where id = 1 for update; <waiting ...>
+step s3_keyshare: select id from tlu_job where id = 1 for key share;
+id             
+
+1              
+step s3_delete: delete from tlu_job where id = 1; <waiting ...>
+step s1_rollback: rollback;
+step s3_delete: <... completed>
+step s3_commit: commit;
+step s2_for_update: <... completed>
+id             
+
+step s2_rollback: rollback;
+
+starting permutation: s1_share s2_for_update s3_for_update s1_rollback s2_rollback s3_rollback
+step s1_share: select id from tlu_job where id = 1 for share;
+id             
+
+1              
+step s2_for_update: select id from tlu_job where id = 1 for update; <waiting ...>
+step s3_for_update: select id from tlu_job where id = 1 for update; <waiting ...>
+step s1_rollback: rollback;
+step s2_for_update: <... completed>
+id             
+
+1              
+step s2_rollback: rollback;
+step s3_for_update: <... completed>
+id             
+
+1              
+step s3_rollback: rollback;
+
+starting permutation: s1_share s2_update s3_update s1_rollback s2_rollback s3_rollback
+step s1_share: select id from tlu_job where id = 1 for share;
+id             
+
+1              
+step s2_update: update tlu_job set name = 'b' where id = 1; <waiting ...>
+step s3_update: update tlu_job set name = 'c' where id = 1; <waiting ...>
+step s1_rollback: rollback;
+step s2_update: <... completed>
+step s2_rollback: rollback;
+step s3_update: <... completed>
+step s3_rollback: rollback;
+
+starting permutation: s1_share s2_delete s3_delete s1_rollback s2_rollback s3_rollback
+step s1_share: select id from tlu_job where id = 1 for share;
+id             
+
+1              
+step s2_delete: delete from tlu_job where id = 1; <waiting ...>
+step s3_delete: delete from tlu_job where id = 1; <waiting ...>
+step s1_rollback: rollback;
+step s2_delete: <... completed>
+step s2_rollback: rollback;
+step s3_delete: <... completed>
+step s3_rollback: rollback;

src/test/isolation/isolation_schedule

@@ -49,6 +49,7 @@ test: reindex-concurrently
 test: propagate-lock-delete
 test: tuplelock-conflict
 test: tuplelock-update
+test: tuplelock-upgrade-no-deadlock
 test: freeze-the-dead
 test: nowait
 test: nowait-2