database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-11-03 09:13:20 +03:00
Code Activity

Get rid of page-level locking in btree-s.

Browse Source 
LockBuffer is used to acquire read/write access
to index pages. Pages are released before leaving
index internals.
This commit is contained in:
Vadim B. Mikheev
1999-05-25 18:20:31 +00:00
parent 07842084fe
commit 7d443a85af
7 changed files with 73 additions and 132 deletions
Download Patch File Download Diff File Expand all files Collapse all files

38
src/backend/access/nbtree/nbtinsert.c
View File

@@ -7,7 +7,7 @@
*
*
* IDENTIFICATION
* $Header: /cvsroot/pgsql/src/backend/access/nbtree/nbtinsert.c,v 1.40 1999/05/25 16:07:23 momjian Exp $
* $Header: /cvsroot/pgsql/src/backend/access/nbtree/nbtinsert.c,v 1.41 1999/05/25 18:20:28 vadim Exp $
*
*-------------------------------------------------------------------------
*/
@@ -19,7 +19,6 @@
#include <access/nbtree.h>
#include <access/heapam.h>
#include <access/xact.h>
#include <storage/bufmgr.h>
#include <fmgr.h>
#ifndef HAVE_MEMMOVE
@@ -64,14 +63,11 @@ _bt_doinsert(Relation rel, BTItem btitem, bool index_is_unique, Relation heapRel
/* find the page containing this key */
stack = _bt_search(rel, natts, itup_scankey, &buf);
blkno = BufferGetBlockNumber(buf);
/* trade in our read lock for a write lock */
_bt_relbuf(rel, buf, BT_READ);
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
LockBuffer(buf, BT_WRITE);
l1:
buf = _bt_getbuf(rel, blkno, BT_WRITE);
/*
* If the page was split between the time that we surrendered our read
* lock and acquired our write lock, then this page may no longer be
@@ -81,6 +77,7 @@ l1:
*/
buf = _bt_moveright(rel, buf, natts, itup_scankey, BT_WRITE);
blkno = BufferGetBlockNumber(buf);
/* if we're not allowing duplicates, make sure the key isn't */
/* already in the node */
@@ -99,13 +96,13 @@ l1:
/* key on the page before trying to compare it */
if (!PageIsEmpty(page) && offset <= maxoff)
{
TupleDesc itupdesc;
BTItem cbti;
HeapTupleData htup;
BTPageOpaque opaque;
Buffer nbuf;
BlockNumber blkno;
bool chtup = true;
TupleDesc itupdesc;
BTItem cbti;
HeapTupleData htup;
BTPageOpaque opaque;
Buffer nbuf;
BlockNumber nblkno;
bool chtup = true;
itupdesc = RelationGetDescr(rel);
nbuf = InvalidBuffer;
@@ -157,7 +154,8 @@ l1:
_bt_relbuf(rel, nbuf, BT_READ);
_bt_relbuf(rel, buf, BT_WRITE);
XactLockTableWait(xwait);
goto l1;/* continue from the begin */
buf = _bt_getbuf(rel, blkno, BT_WRITE);
goto l1; /* continue from the begin */
}
elog(ERROR, "Cannot insert a duplicate key into a unique index");
}
@@ -177,12 +175,12 @@ l1:
* min key of the right page is the same, ooh - so
* many dead duplicates...
*/
blkno = opaque->btpo_next;
nblkno = opaque->btpo_next;
if (nbuf != InvalidBuffer)
_bt_relbuf(rel, nbuf, BT_READ);
for (nbuf = InvalidBuffer;;)
for (nbuf = InvalidBuffer; ; )
{
nbuf = _bt_getbuf(rel, blkno, BT_READ);
nbuf = _bt_getbuf(rel, nblkno, BT_READ);
page = BufferGetPage(nbuf);
maxoff = PageGetMaxOffsetNumber(page);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
@@ -193,10 +191,10 @@ l1:
}
else
{ /* Empty or "pseudo"-empty page - get next */
blkno = opaque->btpo_next;
nblkno = opaque->btpo_next;
_bt_relbuf(rel, nbuf, BT_READ);
nbuf = InvalidBuffer;
if (blkno == P_NONE)
if (nblkno == P_NONE)
break;
}
}