From 8d34f6862853b4b67e29b368dfedf7d4c28d694b Mon Sep 17 00:00:00 2001
From: Heikki Linnakangas <heikki.linnakangas@iki.fi>
Date: Tue, 22 Apr 2014 09:50:47 +0300
Subject: [PATCH] Avoid transient bogus page contents when creating a sequence.

Don't use simple_heap_insert to insert the tuple to a sequence relation.
simple_heap_insert creates a heap insertion WAL record, and replaying that
will create a regular heap page without the special area containing the
sequence magic constant, which is wrong for a sequence. That was not a bug
because we always created a sequence WAL record after that, and replaying
that overwrote the bogus heap page, and the transient state could never be
seen by another backend because it was only done when creating a new
sequence relation. But it's simpler and cleaner to avoid that in the first
place.
---
 src/backend/commands/sequence.c | 55 ++++++++++-----------------------
 1 file changed, 16 insertions(+), 39 deletions(-)

diff --git a/src/backend/commands/sequence.c b/src/backend/commands/sequence.c
index ed696be4c4..2829b1e304 100644
--- a/src/backend/commands/sequence.c
+++ b/src/backend/commands/sequence.c
@@ -307,6 +307,7 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
 	Buffer		buf;
 	Page		page;
 	sequence_magic *sm;
+	OffsetNumber offnum;
 
 	/* Initialize first page of relation with special magic number */
 
@@ -319,56 +320,33 @@ fill_seq_with_data(Relation rel, HeapTuple tuple)
 	sm = (sequence_magic *) PageGetSpecialPointer(page);
 	sm->magic = SEQ_MAGIC;
 
-	/* hack: ensure heap_insert will insert on the just-created page */
-	RelationSetTargetBlock(rel, 0);
-
 	/* Now insert sequence tuple */
-	simple_heap_insert(rel, tuple);
 
-	Assert(ItemPointerGetOffsetNumber(&(tuple->t_self)) == FirstOffsetNumber);
+	LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
 
 	/*
-	 * Two special hacks here:
-	 *
-	 * 1. Since VACUUM does not process sequences, we have to force the tuple
+	 * Since VACUUM does not process sequences, we have to force the tuple
 	 * to have xmin = FrozenTransactionId now.	Otherwise it would become
 	 * invisible to SELECTs after 2G transactions.	It is okay to do this
 	 * because if the current transaction aborts, no other xact will ever
 	 * examine the sequence tuple anyway.
-	 *
-	 * 2. Even though heap_insert emitted a WAL log record, we have to emit an
-	 * XLOG_SEQ_LOG record too, since (a) the heap_insert record will not have
-	 * the right xmin, and (b) REDO of the heap_insert record would re-init
-	 * page and sequence magic number would be lost.  This means two log
-	 * records instead of one :-(
 	 */
-	LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
+	HeapTupleHeaderSetXmin(tuple->t_data, FrozenTransactionId);
+	HeapTupleHeaderSetXminFrozen(tuple->t_data);
+	HeapTupleHeaderSetCmin(tuple->t_data, FirstCommandId);
+	HeapTupleHeaderSetXmax(tuple->t_data, InvalidTransactionId);
+	tuple->t_data->t_infomask |= HEAP_XMAX_INVALID;
+	ItemPointerSet(&tuple->t_data->t_ctid, 0, FirstOffsetNumber);
 
 	START_CRIT_SECTION();
 
-	{
-		/*
-		 * Note that the "tuple" structure is still just a local tuple record
-		 * created by heap_form_tuple; its t_data pointer doesn't point at the
-		 * disk buffer.  To scribble on the disk buffer we need to fetch the
-		 * item pointer.  But do the same to the local tuple, since that will
-		 * be the source for the WAL log record, below.
-		 */
-		ItemId		itemId;
-		Item		item;
-
-		itemId = PageGetItemId((Page) page, FirstOffsetNumber);
-		item = PageGetItem((Page) page, itemId);
-
-		HeapTupleHeaderSetXmin((HeapTupleHeader) item, FrozenTransactionId);
-		HeapTupleHeaderSetXminFrozen((HeapTupleHeader) item);
-
-		HeapTupleHeaderSetXmin(tuple->t_data, FrozenTransactionId);
-		HeapTupleHeaderSetXminFrozen(tuple->t_data);
-	}
-
 	MarkBufferDirty(buf);
 
+	offnum = PageAddItem(page, (Item) tuple->t_data, tuple->t_len,
+						 InvalidOffsetNumber, false, false);
+	if (offnum != FirstOffsetNumber)
+		elog(ERROR, "failed to add sequence tuple to page");
+
 	/* XLOG stuff */
 	if (RelationNeedsWAL(rel))
 	{
@@ -1576,9 +1554,8 @@ seq_redo(XLogRecPtr lsn, XLogRecord *record)
 	page = (Page) BufferGetPage(buffer);
 
 	/*
-	 * We must always reinit the page and reinstall the magic number (see
-	 * comments in fill_seq_with_data).  However, since this WAL record type
-	 * is also used for updating sequences, it's possible that a hot-standby
+	 * We always reinit the page.  However, since this WAL record type is
+	 * also used for updating sequences, it's possible that a hot-standby
 	 * backend is examining the page concurrently; so we mustn't transiently
 	 * trash the buffer.  The solution is to build the correct new page
 	 * contents in local workspace and then memcpy into the buffer.  Then only
-- 
2.40.0