Rethink handling of [Prevent|Is]InTransactionBlock in pipeline mode.

tglsfdc · tglsfdc · commit ae47f8a9664a · 2022-12-13T14:23:59.000-05:00
Commits f929441 et al. made IsInTransactionBlock() set the XACT_FLAGS_NEEDIMMEDIATECOMMIT flag before returning "false", on the grounds that that kept its API promises equivalent to those of PreventInTransactionBlock(). This turns out to be a bad idea though, because it allows an ANALYZE in a pipelined series of commands to cause an immediate commit, which is unexpected. Furthermore, if we return "false" then we have another issue, which is that ANALYZE will decide it's allowed to do internal commit-and-start-transaction sequences, thus possibly unexpectedly committing the effects of previous commands in the pipeline. To fix the latter situation, invent another transaction state flag XACT_FLAGS_PIPELINING, which explicitly records the fact that we have executed some extended-protocol command and not yet seen a commit for it. Then, require that flag to not be set before allowing InTransactionBlock() to return "false". Having done that, we can remove its setting of NEEDIMMEDIATECOMMIT without fear of causing problems. This means that the API guarantees of IsInTransactionBlock now diverge from PreventInTransactionBlock, which is mildly annoying, but it seems OK given the very limited usage of IsInTransactionBlock. (In any case, a caller preferring the old behavior could always set NEEDIMMEDIATECOMMIT for itself.) For consistency also require XACT_FLAGS_PIPELINING to not be set in PreventInTransactionBlock. This too is meant to prevent commands such as CREATE DATABASE from silently committing previous commands in a pipeline. Per report from Peter Eisentraut. As before, back-patch to all supported branches (which sadly no longer includes v10). Discussion: https://postgr.es/m/65a899dd-aebc-f667-1d0a-abb89ff3abf8@enterprisedb.com
diff --git a/doc/src/sgml/libpq.sgml b/doc/src/sgml/libpq.sgml
@@ -5024,10 +5024,11 @@ int PQflush(PGconn *conn);
   </para>
 
   <para>
-   While the pipeline API was introduced in
+   While <application>libpq</application>'s pipeline API was introduced in
    <productname>PostgreSQL</productname> 14, it is a client-side feature
    which doesn't require special server support and works on any server
-   that supports the v3 extended query protocol.
+   that supports the v3 extended query protocol.  For more information see
+   <xref linkend="protocol-flow-pipelining"/>.
   </para>
 
   <sect2 id="libpq-pipeline-using">
diff --git a/doc/src/sgml/protocol.sgml b/doc/src/sgml/protocol.sgml
@@ -1093,17 +1093,18 @@ SELCT 1/0;<!-- this typo is intentional -->
     implicit <command>ROLLBACK</command> if they failed.  However, there
     are a few DDL commands (such as <command>CREATE DATABASE</command>)
     that cannot be executed inside a transaction block.  If one of
-    these is executed in a pipeline, it will, upon success, force an
-    immediate commit to preserve database consistency.
-    A Sync immediately following one of these has no effect except to
+    these is executed in a pipeline, it will fail unless it is the first
+    command in the pipeline.  Furthermore, upon success it will force an
+    immediate commit to preserve database consistency.  Thus a Sync
+    immediately following one of these commands has no effect except to
     respond with ReadyForQuery.
    </para>
 
    <para>
     When using this method, completion of the pipeline must be determined
     by counting ReadyForQuery messages and waiting for that to reach the
     number of Syncs sent.  Counting command completion responses is
-    unreliable, since some of the commands may not be executed and thus not
+    unreliable, since some of the commands may be skipped and thus not
     produce a completion message.
    </para>
   </sect2>
diff --git a/src/backend/access/transam/xact.c b/src/backend/access/transam/xact.c
@@ -3426,6 +3426,16 @@ PreventInTransactionBlock(bool isTopLevel, const char *stmtType)
 				 errmsg("%s cannot run inside a subtransaction",
 						stmtType)));
 
+	/*
+	 * inside a pipeline that has started an implicit transaction?
+	 */
+	if (MyXactFlags & XACT_FLAGS_PIPELINING)
+		ereport(ERROR,
+				(errcode(ERRCODE_ACTIVE_SQL_TRANSACTION),
+		/* translator: %s represents an SQL statement name */
+				 errmsg("%s cannot be executed within a pipeline",
+						stmtType)));
+
 	/*
 	 * inside a function call?
 	 */
@@ -3515,9 +3525,11 @@ CheckTransactionBlock(bool isTopLevel, bool throwError, const char *stmtType)
  *	a transaction block than when running as single commands.  ANALYZE is
  *	currently the only example.
  *
- *	If this routine returns "false", then the calling statement is
- *	guaranteed that if it completes without error, its results will be
- *	committed immediately.
+ *	If this routine returns "false", then the calling statement is allowed
+ *	to perform internal transaction-commit-and-start cycles; there is not a
+ *	risk of messing up any transaction already in progress.  (Note that this
+ *	is not the identical guarantee provided by PreventInTransactionBlock,
+ *	since we will not force a post-statement commit.)
  *
  *	isTopLevel: passed down from ProcessUtility to determine whether we are
  *	inside a function.
@@ -3535,20 +3547,16 @@ IsInTransactionBlock(bool isTopLevel)
 	if (IsSubTransaction())
 		return true;
 
+	if (MyXactFlags & XACT_FLAGS_PIPELINING)
+		return true;
+
 	if (!isTopLevel)
 		return true;
 
 	if (CurrentTransactionState->blockState != TBLOCK_DEFAULT &&
 		CurrentTransactionState->blockState != TBLOCK_STARTED)
 		return true;
 
-	/*
-	 * If we tell the caller we're not in a transaction block, then inform
-	 * postgres.c that it had better commit when the statement is done.
-	 * Otherwise our report could be a lie.
-	 */
-	MyXactFlags |= XACT_FLAGS_NEEDIMMEDIATECOMMIT;
-
 	return false;
 }
 
diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c
@@ -2226,6 +2226,12 @@ exec_execute_message(const char *portal_name, long max_rows)
 			 */
 			CommandCounterIncrement();
 
+			/*
+			 * Set XACT_FLAGS_PIPELINING whenever we complete an Execute
+			 * message without immediately committing the transaction.
+			 */
+			MyXactFlags |= XACT_FLAGS_PIPELINING;
+
 			/*
 			 * Disable statement timeout whenever we complete an Execute
 			 * message.  The next protocol message will start a fresh timeout.
@@ -2241,6 +2247,12 @@ exec_execute_message(const char *portal_name, long max_rows)
 		/* Portal run not complete, so send PortalSuspended */
 		if (whereToSendOutput == DestRemote)
 			pq_putemptymessage('s');
+
+		/*
+		 * Set XACT_FLAGS_PIPELINING whenever we suspend an Execute message,
+		 * too.
+		 */
+		MyXactFlags |= XACT_FLAGS_PIPELINING;
 	}
 
 	/*
diff --git a/src/include/access/xact.h b/src/include/access/xact.h
@@ -113,6 +113,13 @@ extern int	MyXactFlags;
  */
 #define XACT_FLAGS_NEEDIMMEDIATECOMMIT			(1U << 2)
 
+/*
+ * XACT_FLAGS_PIPELINING - set when we complete an extended-query-protocol
+ * Execute message.  This is useful for detecting that an implicit transaction
+ * block has been created via pipelining.
+ */
+#define XACT_FLAGS_PIPELINING					(1U << 3)
+
 /*
  *	start- and end-of-transaction callbacks for dynamically loaded modules
  */
