Repair potential deadlock created by recent changes to recycle btree

index pages: when _bt_getbuf asks the FSM for a free index page, it is
possible (and, in some cases, even moderately likely) that the answer
will be the same page that _bt_split is trying to split.  _bt_getbuf
already knew that the returned page might not be free, but it wasn't
prepared for the possibility that even trying to lock the page could
be problematic.  Fix by doing a conditional rather than unconditional
grab of the page lock.

Author: tglsfdc

src/backend/access/nbtree/nbtpage.c

@@ -9,7 +9,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/access/nbtree/nbtpage.c,v 1.69 2003/08/08 21:41:27 momjian Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/access/nbtree/nbtpage.c,v 1.70 2003/08/10 19:48:08 tgl Exp $
  *
  *	NOTES
  *	   Postgres btree pages look like ordinary relation pages.	The opaque
@@ -409,23 +409,47 @@ _bt_getbuf(Relation rel, BlockNumber blkno, int access)
 		 * that the page is still free.  (For example, an already-free
 		 * page could have been re-used between the time the last VACUUM
 		 * scanned it and the time the VACUUM made its FSM updates.)
+		 *
+		 * In fact, it's worse than that: we can't even assume that it's
+		 * safe to take a lock on the reported page.  If somebody else
+		 * has a lock on it, or even worse our own caller does, we could
+		 * deadlock.  (The own-caller scenario is actually not improbable.
+		 * Consider an index on a serial or timestamp column.  Nearly all
+		 * splits will be at the rightmost page, so it's entirely likely
+		 * that _bt_split will call us while holding a lock on the page most
+		 * recently acquired from FSM.  A VACUUM running concurrently with
+		 * the previous split could well have placed that page back in FSM.)
+		 *
+		 * To get around that, we ask for only a conditional lock on the
+		 * reported page.  If we fail, then someone else is using the page,
+		 * and we may reasonably assume it's not free.  (If we happen to be
+		 * wrong, the worst consequence is the page will be lost to use till
+		 * the next VACUUM, which is no big problem.)
 		 */
 		for (;;)
 		{
 			blkno = GetFreeIndexPage(&rel->rd_node);
 			if (blkno == InvalidBlockNumber)
 				break;
 			buf = ReadBuffer(rel, blkno);
-			LockBuffer(buf, access);
-			page = BufferGetPage(buf);
-			if (_bt_page_recyclable(page))
+			if (ConditionalLockBuffer(buf))
 			{
-				/* Okay to use page.  Re-initialize and return it */
-				_bt_pageinit(page, BufferGetPageSize(buf));
-				return buf;
+				page = BufferGetPage(buf);
+				if (_bt_page_recyclable(page))
+				{
+					/* Okay to use page.  Re-initialize and return it */
+					_bt_pageinit(page, BufferGetPageSize(buf));
+					return buf;
+				}
+				elog(DEBUG2, "FSM returned nonrecyclable page");
+				_bt_relbuf(rel, buf);
+			}
+			else
+			{
+				elog(DEBUG2, "FSM returned nonlockable page");
+				/* couldn't get lock, so just drop pin */
+				ReleaseBuffer(buf);
 			}
-			elog(DEBUG2, "FSM returned nonrecyclable page");
-			_bt_relbuf(rel, buf);
 		}
 
 		/*

src/backend/storage/buffer/bufmgr.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $Header: /cvsroot/pgsql/src/backend/storage/buffer/bufmgr.c,v 1.139 2003/08/04 02:40:03 momjian Exp $
+ *	  $Header: /cvsroot/pgsql/src/backend/storage/buffer/bufmgr.c,v 1.140 2003/08/10 19:48:08 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -1936,6 +1936,37 @@ LockBuffer(Buffer buffer, int mode)
 		elog(ERROR, "unrecognized buffer lock mode: %d", mode);
 }
 
+/*
+ * Acquire the cntx_lock for the buffer, but only if we don't have to wait.
+ *
+ * This assumes the caller wants BUFFER_LOCK_EXCLUSIVE mode.
+ */
+bool
+ConditionalLockBuffer(Buffer buffer)
+{
+	BufferDesc *buf;
+
+	Assert(BufferIsValid(buffer));
+	if (BufferIsLocal(buffer))
+		return true;			/* act as though we got it */
+
+	buf = &(BufferDescriptors[buffer - 1]);
+
+	if (LWLockConditionalAcquire(buf->cntx_lock, LW_EXCLUSIVE))
+	{
+		/*
+		 * This is not the best place to set cntxDirty flag (eg indices do
+		 * not always change buffer they lock in excl mode). But please
+		 * remember that it's critical to set cntxDirty *before* logging
+		 * changes with XLogInsert() - see comments in BufferSync().
+		 */
+		buf->cntxDirty = true;
+
+		return true;
+	}
+	return false;
+}
+
 /*
  * LockBufferForCleanup - lock a buffer in preparation for deleting items
  *

src/include/storage/bufmgr.h

@@ -7,7 +7,7 @@
  * Portions Copyright (c) 1996-2003, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $Id: bufmgr.h,v 1.69 2003/08/04 02:40:14 momjian Exp $
+ * $Id: bufmgr.h,v 1.70 2003/08/10 19:48:08 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -180,6 +180,7 @@ extern void SetBufferCommitInfoNeedsSave(Buffer buffer);
 
 extern void UnlockBuffers(void);
 extern void LockBuffer(Buffer buffer, int mode);
+extern bool ConditionalLockBuffer(Buffer buffer);
 extern void LockBufferForCleanup(Buffer buffer);
 
 extern void AbortBufferIO(void);