Add a function GetLockConflicts() to lock.c to report xacts holding

locks that would conflict with a specified lock request, without
actually trying to get that lock.  Use this instead of the former ad hoc
method of doing the first wait step in CREATE INDEX CONCURRENTLY.
Fixes problem with undetected deadlock and in many cases will allow the
index creation to proceed sooner than it otherwise could've.  Per
discussion with Greg Stark.

2 files changed, 118 insertions, 24 deletions

diff --git a/src/backend/commands/indexcmds.c b/src/backend/commands/indexcmds.c
index e96235c6f53..81246768bc9 100644
--- a/src/backend/commands/indexcmds.c
+++ b/src/backend/commands/indexcmds.c
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/commands/indexcmds.c,v 1.147 2006/08/25 04:06:48 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/commands/indexcmds.c,v 1.148 2006/08/27 19:14:34 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -119,8 +119,11 @@ DefineIndex(RangeVar *heapRelation,
 	Datum		reloptions;
 	IndexInfo  *indexInfo;
 	int			numberOfAttributes;
+	List	   *old_xact_list;
+	ListCell   *lc;
 	uint32		ixcnt;
 	LockRelId	heaprelid;
+	LOCKTAG		heaplocktag;
 	Snapshot	snapshot;
 	Relation pg_index;
 	HeapTuple indexTuple;
@@ -466,33 +469,26 @@ DefineIndex(RangeVar *heapRelation,
 	CommitTransactionCommand();
 	StartTransactionCommand();
 
-	/* Establish transaction snapshot ... else GetLatestSnapshot complains */
-	(void) GetTransactionSnapshot();
-
 	/*
 	 * Now we must wait until no running transaction could have the table open
-	 * with the old list of indexes.  If we can take an exclusive lock then
-	 * there are none now and anybody who opens it later will get the new
-	 * index in their relcache entry.  Alternatively, if our Xmin reaches our
-	 * own (new) transaction then we know no transactions that started before
-	 * the index was visible are left anyway.
+	 * with the old list of indexes.  To do this, inquire which xacts currently
+	 * would conflict with ShareLock on the table -- ie, which ones have
+	 * a lock that permits writing the table.  Then wait for each of these
+	 * xacts to commit or abort.  Note we do not need to worry about xacts
+	 * that open the table for writing after this point; they will see the
+	 * new index when they open it.
+	 *
+	 * Note: GetLockConflicts() never reports our own xid,
+	 * hence we need not check for that.
 	 */
-	for (;;)
-	{
-		CHECK_FOR_INTERRUPTS();
+	SET_LOCKTAG_RELATION(heaplocktag, heaprelid.dbId, heaprelid.relId);
+	old_xact_list = GetLockConflicts(&heaplocktag, ShareLock);
 
-		if (ConditionalLockRelationOid(relationId, ExclusiveLock))
-		{
-			/* Release the lock right away to avoid blocking anyone */
-			UnlockRelationOid(relationId, ExclusiveLock);
-			break;
-		}
-
-		if (TransactionIdEquals(GetLatestSnapshot()->xmin,
-								GetTopTransactionId()))
-			break;
+	foreach(lc, old_xact_list)
+	{
+		TransactionId xid = lfirst_xid(lc);
 
-		pg_usleep(1000000L);		/* 1 sec */
+		XactLockTableWait(xid);
 	}
 
 	/*
diff --git a/src/backend/storage/lmgr/lock.c b/src/backend/storage/lmgr/lock.c
index 7fe011a2baf..c59999e20ca 100644
--- a/src/backend/storage/lmgr/lock.c
+++ b/src/backend/storage/lmgr/lock.c
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/storage/lmgr/lock.c,v 1.171 2006/08/19 01:36:28 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/storage/lmgr/lock.c,v 1.172 2006/08/27 19:14:34 tgl Exp $
  *
  * NOTES
  *	  A lock table is a shared memory hash table.  When
@@ -1686,6 +1686,104 @@ LockReassignCurrentOwner(void)
 
 
 /*
+ * GetLockConflicts
+ *		Get a list of TransactionIds of xacts currently holding locks
+ *		that would conflict with the specified lock/lockmode.
+ *		xacts merely awaiting such a lock are NOT reported.
+ *
+ * Of course, the result could be out of date by the time it's returned,
+ * so use of this function has to be thought about carefully.
+ *
+ * Only top-level XIDs are reported.  Note we never include the current xact
+ * in the result list, since an xact never blocks itself.
+ */
+List *
+GetLockConflicts(const LOCKTAG *locktag, LOCKMODE lockmode)
+{
+	List	   *result = NIL;
+	LOCKMETHODID lockmethodid = locktag->locktag_lockmethodid;
+	LockMethod	lockMethodTable;
+	LOCK	   *lock;
+	LOCKMASK	conflictMask;
+	SHM_QUEUE  *procLocks;
+	PROCLOCK   *proclock;
+	uint32		hashcode;
+	LWLockId	partitionLock;
+
+	if (lockmethodid <= 0 || lockmethodid >= lengthof(LockMethods))
+		elog(ERROR, "unrecognized lock method: %d", lockmethodid);
+	lockMethodTable = LockMethods[lockmethodid];
+	if (lockmode <= 0 || lockmode > lockMethodTable->numLockModes)
+		elog(ERROR, "unrecognized lock mode: %d", lockmode);
+
+	/*
+	 * Look up the lock object matching the tag.
+	 */
+	hashcode = LockTagHashCode(locktag);
+	partitionLock = LockHashPartitionLock(hashcode);
+
+	LWLockAcquire(partitionLock, LW_SHARED);
+
+	lock = (LOCK *) hash_search_with_hash_value(LockMethodLockHash,
+												(void *) locktag,
+												hashcode,
+												HASH_FIND,
+												NULL);
+	if (!lock)
+	{
+		/*
+		 * If the lock object doesn't exist, there is nothing holding a
+		 * lock on this lockable object.
+		 */
+		LWLockRelease(partitionLock);
+		return NIL;
+	}
+
+	/*
+	 * Examine each existing holder (or awaiter) of the lock.
+	 */
+	conflictMask = lockMethodTable->conflictTab[lockmode];
+
+	procLocks = &(lock->procLocks);
+
+	proclock = (PROCLOCK *) SHMQueueNext(procLocks, procLocks,
+										 offsetof(PROCLOCK, lockLink));
+
+	while (proclock)
+	{
+		if (conflictMask & proclock->holdMask)
+		{
+			PGPROC *proc = proclock->tag.myProc;
+
+			/* A backend never blocks itself */
+			if (proc != MyProc)
+			{
+				/* Fetch xid just once - see GetNewTransactionId */
+				TransactionId xid = proc->xid;
+
+				/*
+				 * Race condition: during xact commit/abort we zero out
+				 * PGPROC's xid before we mark its locks released.  If we see
+				 * zero in the xid field, assume the xact is in process of
+				 * shutting down and act as though the lock is already
+				 * released.
+				 */
+				if (TransactionIdIsValid(xid))
+					result = lappend_xid(result, xid);
+			}
+		}
+
+		proclock = (PROCLOCK *) SHMQueueNext(procLocks, &proclock->lockLink,
+											 offsetof(PROCLOCK, lockLink));
+	}
+
+	LWLockRelease(partitionLock);
+
+	return result;
+}
+
+
+/*
  * AtPrepare_Locks
  *		Do the preparatory work for a PREPARE: make 2PC state file records
  *		for all locks currently held.