nbtpage.c File Reference

#include "postgres.h"
#include "access/nbtree.h"
#include "access/nbtxlog.h"
#include "access/tableam.h"
#include "access/transam.h"
#include "access/xlog.h"
#include "access/xloginsert.h"
#include "common/int.h"
#include "miscadmin.h"
#include "storage/indexfsm.h"
#include "storage/predicate.h"
#include "storage/procarray.h"
#include "utils/injection_point.h"
#include "utils/memdebug.h"
#include "utils/memutils.h"
#include "utils/snapmgr.h"

Include dependency graph for nbtpage.c:

Go to the source code of this file.

Functions
static BTMetaPageData *	_bt_getmeta (Relation rel, Buffer metabuf)
static void	_bt_delitems_delete (Relation rel, Buffer buf, TransactionId snapshotConflictHorizon, bool isCatalogRel, OffsetNumber *deletable, int ndeletable, BTVacuumPosting *updatable, int nupdatable)
static char *	_bt_delitems_update (BTVacuumPosting *updatable, int nupdatable, OffsetNumber *updatedoffsets, Size *updatedbuflen, bool needswal)
static bool	_bt_mark_page_halfdead (Relation rel, Relation heaprel, Buffer leafbuf, BTStack stack)
static bool	_bt_unlink_halfdead_page (Relation rel, Buffer leafbuf, BlockNumber scanblkno, bool *rightsib_empty, BTVacState *vstate)
static bool	_bt_lock_subtree_parent (Relation rel, Relation heaprel, BlockNumber child, BTStack stack, Buffer *subtreeparent, OffsetNumber *poffset, BlockNumber *topparent, BlockNumber *topparentrightsib)
static void	_bt_pendingfsm_add (BTVacState *vstate, BlockNumber target, FullTransactionId safexid)
void	_bt_initmetapage (Page page, BlockNumber rootbknum, uint32 level, bool allequalimage)
void	_bt_upgrademetapage (Page page)
bool	_bt_vacuum_needs_cleanup (Relation rel)
void	_bt_set_cleanup_info (Relation rel, BlockNumber num_delpages)
Buffer	_bt_getroot (Relation rel, Relation heaprel, int access)
Buffer	_bt_gettrueroot (Relation rel)
int	_bt_getrootheight (Relation rel)
void	_bt_metaversion (Relation rel, bool *heapkeyspace, bool *allequalimage)
void	_bt_checkpage (Relation rel, Buffer buf)
Buffer	_bt_getbuf (Relation rel, BlockNumber blkno, int access)
Buffer	_bt_allocbuf (Relation rel, Relation heaprel)
Buffer	_bt_relandgetbuf (Relation rel, Buffer obuf, BlockNumber blkno, int access)
void	_bt_relbuf (Relation rel, Buffer buf)
void	_bt_lockbuf (Relation rel, Buffer buf, int access)
void	_bt_unlockbuf (Relation rel, Buffer buf)
bool	_bt_conditionallockbuf (Relation rel, Buffer buf)
void	_bt_upgradelockbufcleanup (Relation rel, Buffer buf)
void	_bt_pageinit (Page page, Size size)
void	_bt_delitems_vacuum (Relation rel, Buffer buf, OffsetNumber *deletable, int ndeletable, BTVacuumPosting *updatable, int nupdatable)
static int	_bt_delitems_cmp (const void *a, const void *b)
void	_bt_delitems_delete_check (Relation rel, Buffer buf, Relation heapRel, TM_IndexDeleteOp *delstate)
static bool	_bt_leftsib_splitflag (Relation rel, BlockNumber leftsib, BlockNumber target)
static bool	_bt_rightsib_halfdeadflag (Relation rel, BlockNumber leafrightsib)
void	_bt_pagedel (Relation rel, Buffer leafbuf, BTVacState *vstate)
void	_bt_pendingfsm_init (Relation rel, BTVacState *vstate, bool cleanuponly)
void	_bt_pendingfsm_finalize (Relation rel, BTVacState *vstate)

Function Documentation

_bt_allocbuf()

Buffer _bt_allocbuf	(	Relation	rel,
		Relation	heaprel
	)

Definition at line 854 of file nbtpage.c.

{
    Buffer      buf;
    BlockNumber blkno;
    Page        page;
 
    Assert(heaprel != NULL);
 
    /*
     * First see if the FSM knows of any free pages.
     *
     * We can't trust the FSM's report unreservedly; we have to check 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);
        if (blkno == InvalidBlockNumber)
            break;
        buf = ReadBuffer(rel, blkno);
        if (_bt_conditionallockbuf(rel, buf))
        {
            page = BufferGetPage(buf);
 
            /*
             * It's possible to find an all-zeroes page in an index.  For
             * example, a backend might successfully extend the relation one
             * page and then crash before it is able to make a WAL entry for
             * adding the page.  If we find a zeroed page then reclaim it
             * immediately.
             */
            if (PageIsNew(page))
            {
                /* Okay to use page.  Initialize and return it. */
                _bt_pageinit(page, BufferGetPageSize(buf));
                return buf;
            }
 
            if (BTPageIsRecyclable(page, heaprel))
            {
                /*
                 * If we are generating WAL for Hot Standby then create a WAL
                 * record that will allow us to conflict with queries running
                 * on standby, in case they have snapshots older than safexid
                 * value
                 */
                if (RelationNeedsWAL(rel) && XLogStandbyInfoActive())
                {
                    xl_btree_reuse_page xlrec_reuse;
 
                    /*
                     * Note that we don't register the buffer with the record,
                     * because this operation doesn't modify the page (that
                     * already happened, back when VACUUM deleted the page).
                     * This record only exists to provide a conflict point for
                     * Hot Standby.  See record REDO routine comments.
                     */
                    xlrec_reuse.locator = rel->rd_locator;
                    xlrec_reuse.block = blkno;
                    xlrec_reuse.snapshotConflictHorizon = BTPageGetDeleteXid(page);
                    xlrec_reuse.isCatalogRel =
                        RelationIsAccessibleInLogicalDecoding(heaprel);
 
                    XLogBeginInsert();
                    XLogRegisterData(&xlrec_reuse, SizeOfBtreeReusePage);
 
                    XLogInsert(RM_BTREE_ID, XLOG_BTREE_REUSE_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);
        }
    }
 
    /*
     * Extend the relation by one page. Need to use RBM_ZERO_AND_LOCK or we
     * risk a race condition against btvacuumscan --- see comments therein.
     * This forces us to repeat the valgrind request that _bt_lockbuf()
     * otherwise would make, as we can't use _bt_lockbuf() without introducing
     * a race.
     */
    buf = ExtendBufferedRel(BMR_REL(rel), MAIN_FORKNUM, NULL, EB_LOCK_FIRST);
    if (!RelationUsesLocalBuffers(rel))
        VALGRIND_MAKE_MEM_DEFINED(BufferGetPage(buf), BLCKSZ);
 
    /* Initialize the new page before returning it */
    page = BufferGetPage(buf);
    Assert(PageIsNew(page));
    _bt_pageinit(page, BufferGetPageSize(buf));
 
    return buf;
}

References _bt_conditionallockbuf(), _bt_pageinit(), _bt_relbuf(), Assert, BMR_REL, BTPageGetDeleteXid(), BTPageIsRecyclable(), buf, BufferGetPage(), BufferGetPageSize(), DEBUG2, EB_LOCK_FIRST, elog, ExtendBufferedRel(), fb(), GetFreeIndexPage(), InvalidBlockNumber, xl_btree_reuse_page::locator, MAIN_FORKNUM, PageIsNew(), RelationData::rd_locator, ReadBuffer(), RelationIsAccessibleInLogicalDecoding, RelationNeedsWAL, RelationUsesLocalBuffers, ReleaseBuffer(), SizeOfBtreeReusePage, VALGRIND_MAKE_MEM_DEFINED, XLOG_BTREE_REUSE_PAGE, XLogBeginInsert(), XLogInsert(), XLogRegisterData(), and XLogStandbyInfoActive.

Referenced by _bt_getroot(), _bt_newlevel(), and _bt_split().

_bt_checkpage()

void _bt_checkpage	(	Relation	rel,
		Buffer	buf
	)

Definition at line 782 of file nbtpage.c.

{
    Page        page = BufferGetPage(buf);
 
    /*
     * ReadBuffer verifies that every newly-read page passes
     * PageHeaderIsValid, which means it either contains a reasonably sane
     * page header or is all-zero.  We have to defend against the all-zero
     * case, however.
     */
    if (PageIsNew(page))
        ereport(ERROR,
                (errcode(ERRCODE_INDEX_CORRUPTED),
                 errmsg("index \"%s\" contains unexpected zero page at block %u",
                        RelationGetRelationName(rel),
                        BufferGetBlockNumber(buf)),
                 errhint("Please REINDEX it.")));
 
    /*
     * Additionally check that the special area looks sane.
     */
    if (PageGetSpecialSize(page) != MAXALIGN(sizeof(BTPageOpaqueData)))
        ereport(ERROR,
                (errcode(ERRCODE_INDEX_CORRUPTED),
                 errmsg("index \"%s\" contains corrupted page at block %u",
                        RelationGetRelationName(rel),
                        BufferGetBlockNumber(buf)),
                 errhint("Please REINDEX it.")));
}

References buf, BufferGetBlockNumber(), BufferGetPage(), ereport, errcode(), errhint(), errmsg, ERROR, fb(), MAXALIGN, PageGetSpecialSize(), PageIsNew(), and RelationGetRelationName.

Referenced by _bt_getbuf(), _bt_relandgetbuf(), _bt_search_insert(), bt_recheck_sibling_links(), btvacuumpage(), and palloc_btree_page().

_bt_conditionallockbuf()

bool _bt_conditionallockbuf	(	Relation	rel,
		Buffer	buf
	)

Definition at line 1101 of file nbtpage.c.

{
    /* ConditionalLockBuffer() asserts that pin is held by this backend */
    if (!ConditionalLockBuffer(buf))
        return false;
 
    if (!RelationUsesLocalBuffers(rel))
        VALGRIND_MAKE_MEM_DEFINED(BufferGetPage(buf), BLCKSZ);
 
    return true;
}

References buf, BufferGetPage(), ConditionalLockBuffer(), fb(), RelationUsesLocalBuffers, and VALGRIND_MAKE_MEM_DEFINED.

Referenced by _bt_allocbuf(), and _bt_search_insert().

_bt_delitems_cmp()

static int _bt_delitems_cmp ( const void *  a, const void *  b  )

static

Definition at line 1474 of file nbtpage.c.

{
    const TM_IndexDelete *indexdelete1 = a;
    const TM_IndexDelete *indexdelete2 = b;
 
    Assert(indexdelete1->id != indexdelete2->id);
 
    return pg_cmp_s16(indexdelete1->id, indexdelete2->id);
}

References a, Assert, b, fb(), and pg_cmp_s16().

Referenced by _bt_delitems_delete_check().

_bt_delitems_delete()

static void _bt_delitems_delete ( Relation  rel, Buffer  buf, TransactionId  snapshotConflictHorizon, bool  isCatalogRel, OffsetNumber *  deletable, int  ndeletable, BTVacuumPosting *  updatable, int  nupdatable  )

static

Definition at line 1293 of file nbtpage.c.

{
    Page        page = BufferGetPage(buf);
    BTPageOpaque opaque;
    bool        needswal = RelationNeedsWAL(rel);
    char       *updatedbuf = NULL;
    Size        updatedbuflen = 0;
    OffsetNumber updatedoffsets[MaxIndexTuplesPerPage];
    XLogRecPtr  recptr;
 
    /* Shouldn't be called unless there's something to do */
    Assert(ndeletable > 0 || nupdatable > 0);
 
    /* Generate new versions of posting lists without deleted TIDs */
    if (nupdatable > 0)
        updatedbuf = _bt_delitems_update(updatable, nupdatable,
                                         updatedoffsets, &updatedbuflen,
                                         needswal);
 
    /* No ereport(ERROR) until changes are logged */
    START_CRIT_SECTION();
 
    /* Handle updates and deletes just like _bt_delitems_vacuum */
    for (int i = 0; i < nupdatable; i++)
    {
        OffsetNumber updatedoffset = updatedoffsets[i];
        IndexTuple  itup;
        Size        itemsz;
 
        itup = updatable[i]->itup;
        itemsz = MAXALIGN(IndexTupleSize(itup));
        if (!PageIndexTupleOverwrite(page, updatedoffset, itup, itemsz))
            elog(PANIC, "failed to update partially dead item in block %u of index \"%s\"",
                 BufferGetBlockNumber(buf), RelationGetRelationName(rel));
    }
 
    if (ndeletable > 0)
        PageIndexMultiDelete(page, deletable, ndeletable);
 
    /*
     * Unlike _bt_delitems_vacuum, we *must not* clear the vacuum cycle ID at
     * this point.  The VACUUM command alone controls vacuum cycle IDs.
     */
    opaque = BTPageGetOpaque(page);
 
    /*
     * Clear the BTP_HAS_GARBAGE page flag.
     *
     * This flag indicates the presence of LP_DEAD items on the page (though
     * not reliably).  Note that we only rely on it with pg_upgrade'd
     * !heapkeyspace indexes.
     */
    opaque->btpo_flags &= ~BTP_HAS_GARBAGE;
 
    MarkBufferDirty(buf);
 
    /* XLOG stuff */
    if (needswal)
    {
        xl_btree_delete xlrec_delete;
 
        xlrec_delete.snapshotConflictHorizon = snapshotConflictHorizon;
        xlrec_delete.ndeleted = ndeletable;
        xlrec_delete.nupdated = nupdatable;
        xlrec_delete.isCatalogRel = isCatalogRel;
 
        XLogBeginInsert();
        XLogRegisterBuffer(0, buf, REGBUF_STANDARD);
        XLogRegisterData(&xlrec_delete, SizeOfBtreeDelete);
 
        if (ndeletable > 0)
            XLogRegisterBufData(0, deletable,
                                ndeletable * sizeof(OffsetNumber));
 
        if (nupdatable > 0)
        {
            XLogRegisterBufData(0, updatedoffsets,
                                nupdatable * sizeof(OffsetNumber));
            XLogRegisterBufData(0, updatedbuf, updatedbuflen);
        }
 
        recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_DELETE);
    }
    else
        recptr = XLogGetFakeLSN(rel);
 
    PageSetLSN(page, recptr);
 
    END_CRIT_SECTION();
 
    /* can't leak memory here */
    if (updatedbuf != NULL)
        pfree(updatedbuf);
    /* free tuples allocated within _bt_delitems_update() */
    for (int i = 0; i < nupdatable; i++)
        pfree(updatable[i]->itup);
}

References _bt_delitems_update(), Assert, BTPageGetOpaque, BTPageOpaqueData::btpo_flags, buf, BufferGetBlockNumber(), BufferGetPage(), elog, END_CRIT_SECTION, fb(), i, IndexTupleSize(), BTVacuumPostingData::itup, MarkBufferDirty(), MAXALIGN, MaxIndexTuplesPerPage, PageIndexMultiDelete(), PageIndexTupleOverwrite(), PageSetLSN(), PANIC, pfree(), REGBUF_STANDARD, RelationGetRelationName, RelationNeedsWAL, SizeOfBtreeDelete, xl_btree_delete::snapshotConflictHorizon, START_CRIT_SECTION, XLOG_BTREE_DELETE, XLogBeginInsert(), XLogGetFakeLSN(), XLogInsert(), XLogRegisterBufData(), XLogRegisterBuffer(), and XLogRegisterData().

Referenced by _bt_delitems_delete_check().

_bt_delitems_delete_check()

void _bt_delitems_delete_check	(	Relation	rel,
		Buffer	buf,
		Relation	heapRel,
		TM_IndexDeleteOp *	delstate
	)

Definition at line 1523 of file nbtpage.c.

{
    Page        page = BufferGetPage(buf);
    TransactionId snapshotConflictHorizon;
    bool        isCatalogRel;
    OffsetNumber postingidxoffnum = InvalidOffsetNumber;
    int         ndeletable = 0,
                nupdatable = 0;
    OffsetNumber deletable[MaxIndexTuplesPerPage];
    BTVacuumPosting updatable[MaxIndexTuplesPerPage];
 
    /* Use tableam interface to determine which tuples to delete first */
    snapshotConflictHorizon = table_index_delete_tuples(heapRel, delstate);
    isCatalogRel = RelationIsAccessibleInLogicalDecoding(heapRel);
 
    /* Should not WAL-log snapshotConflictHorizon unless it's required */
    if (!XLogStandbyInfoActive())
        snapshotConflictHorizon = InvalidTransactionId;
 
    /*
     * Construct a leaf-page-wise description of what _bt_delitems_delete()
     * needs to do to physically delete index tuples from the page.
     *
     * Must sort deltids array to restore leaf-page-wise order (original order
     * before call to tableam).  This is the order that the loop expects.
     *
     * Note that deltids array might be a lot smaller now.  It might even have
     * no entries at all (with bottom-up deletion caller), in which case there
     * is nothing left to do.
     */
    qsort(delstate->deltids, delstate->ndeltids, sizeof(TM_IndexDelete),
          _bt_delitems_cmp);
    if (delstate->ndeltids == 0)
    {
        Assert(delstate->bottomup);
        return;
    }
 
    /* We definitely have to delete at least one index tuple (or one TID) */
    for (int i = 0; i < delstate->ndeltids; i++)
    {
        TM_IndexStatus *dstatus = delstate->status + delstate->deltids[i].id;
        OffsetNumber idxoffnum = dstatus->idxoffnum;
        ItemId      itemid = PageGetItemId(page, idxoffnum);
        IndexTuple  itup = (IndexTuple) PageGetItem(page, itemid);
        int         nestedi,
                    nitem;
        BTVacuumPosting vacposting;
 
        Assert(OffsetNumberIsValid(idxoffnum));
 
        if (idxoffnum == postingidxoffnum)
        {
            /*
             * This deltid entry is a TID from a posting list tuple that has
             * already been completely processed
             */
            Assert(BTreeTupleIsPosting(itup));
            Assert(ItemPointerCompare(BTreeTupleGetHeapTID(itup),
                                      &delstate->deltids[i].tid) < 0);
            Assert(ItemPointerCompare(BTreeTupleGetMaxHeapTID(itup),
                                      &delstate->deltids[i].tid) >= 0);
            continue;
        }
 
        if (!BTreeTupleIsPosting(itup))
        {
            /* Plain non-pivot tuple */
            Assert(ItemPointerEquals(&itup->t_tid, &delstate->deltids[i].tid));
            if (dstatus->knowndeletable)
                deletable[ndeletable++] = idxoffnum;
            continue;
        }
 
        /*
         * itup is a posting list tuple whose lowest deltids entry (which may
         * or may not be for the first TID from itup) is considered here now.
         * We should process all of the deltids entries for the posting list
         * together now, though (not just the lowest).  Remember to skip over
         * later itup-related entries during later iterations of outermost
         * loop.
         */
        postingidxoffnum = idxoffnum;   /* Remember work in outermost loop */
        nestedi = i;            /* Initialize for first itup deltids entry */
        vacposting = NULL;      /* Describes final action for itup */
        nitem = BTreeTupleGetNPosting(itup);
        for (int p = 0; p < nitem; p++)
        {
            ItemPointer ptid = BTreeTupleGetPostingN(itup, p);
            int         ptidcmp = -1;
 
            /*
             * This nested loop reuses work across ptid TIDs taken from itup.
             * We take advantage of the fact that both itup's TIDs and deltids
             * entries (within a single itup/posting list grouping) must both
             * be in ascending TID order.
             */
            for (; nestedi < delstate->ndeltids; nestedi++)
            {
                TM_IndexDelete *tcdeltid = &delstate->deltids[nestedi];
                TM_IndexStatus *tdstatus = (delstate->status + tcdeltid->id);
 
                /* Stop once we get past all itup related deltids entries */
                Assert(tdstatus->idxoffnum >= idxoffnum);
                if (tdstatus->idxoffnum != idxoffnum)
                    break;
 
                /* Skip past non-deletable itup related entries up front */
                if (!tdstatus->knowndeletable)
                    continue;
 
                /* Entry is first partial ptid match (or an exact match)? */
                ptidcmp = ItemPointerCompare(&tcdeltid->tid, ptid);
                if (ptidcmp >= 0)
                {
                    /* Greater than or equal (partial or exact) match... */
                    break;
                }
            }
 
            /* ...exact ptid match to a deletable deltids entry? */
            if (ptidcmp != 0)
                continue;
 
            /* Exact match for deletable deltids entry -- ptid gets deleted */
            if (vacposting == NULL)
            {
                vacposting = palloc(offsetof(BTVacuumPostingData, deletetids) +
                                    nitem * sizeof(uint16));
                vacposting->itup = itup;
                vacposting->updatedoffset = idxoffnum;
                vacposting->ndeletedtids = 0;
            }
            vacposting->deletetids[vacposting->ndeletedtids++] = p;
        }
 
        /* Final decision on itup, a posting list tuple */
 
        if (vacposting == NULL)
        {
            /* No TIDs to delete from itup -- do nothing */
        }
        else if (vacposting->ndeletedtids == nitem)
        {
            /* Straight delete of itup (to delete all TIDs) */
            deletable[ndeletable++] = idxoffnum;
            /* Turns out we won't need granular information */
            pfree(vacposting);
        }
        else
        {
            /* Delete some (but not all) TIDs from itup */
            Assert(vacposting->ndeletedtids > 0 &&
                   vacposting->ndeletedtids < nitem);
            updatable[nupdatable++] = vacposting;
        }
    }
 
    /* Physically delete tuples (or TIDs) using deletable (or updatable) */
    _bt_delitems_delete(rel, buf, snapshotConflictHorizon, isCatalogRel,
                        deletable, ndeletable, updatable, nupdatable);
 
    /* be tidy */
    for (int i = 0; i < nupdatable; i++)
        pfree(updatable[i]);
}

References _bt_delitems_cmp(), _bt_delitems_delete(), Assert, BTreeTupleGetHeapTID(), BTreeTupleGetMaxHeapTID(), BTreeTupleGetNPosting(), BTreeTupleGetPostingN(), BTreeTupleIsPosting(), buf, BufferGetPage(), fb(), i, TM_IndexStatus::idxoffnum, InvalidOffsetNumber, InvalidTransactionId, ItemPointerCompare(), ItemPointerEquals(), MaxIndexTuplesPerPage, OffsetNumberIsValid, PageGetItem(), PageGetItemId(), palloc(), pfree(), qsort, RelationIsAccessibleInLogicalDecoding, IndexTupleData::t_tid, table_index_delete_tuples(), and XLogStandbyInfoActive.

Referenced by _bt_bottomupdel_pass(), and _bt_simpledel_pass().

_bt_delitems_update()

static char * _bt_delitems_update ( BTVacuumPosting *  updatable, int  nupdatable, OffsetNumber *  updatedoffsets, Size *  updatedbuflen, bool  needswal  )

static

Definition at line 1415 of file nbtpage.c.

{
    char       *updatedbuf = NULL;
    Size        buflen = 0;
 
    /* Shouldn't be called unless there's something to do */
    Assert(nupdatable > 0);
 
    for (int i = 0; i < nupdatable; i++)
    {
        BTVacuumPosting vacposting = updatable[i];
        Size        itemsz;
 
        /* Replace work area IndexTuple with updated version */
        _bt_update_posting(vacposting);
 
        /* Keep track of size of xl_btree_update for updatedbuf in passing */
        itemsz = SizeOfBtreeUpdate + vacposting->ndeletedtids * sizeof(uint16);
        buflen += itemsz;
 
        /* Build updatedoffsets buffer in passing */
        updatedoffsets[i] = vacposting->updatedoffset;
    }
 
    /* XLOG stuff */
    if (needswal)
    {
        Size        offset = 0;
 
        /* Allocate, set final size for caller */
        updatedbuf = palloc(buflen);
        *updatedbuflen = buflen;
        for (int i = 0; i < nupdatable; i++)
        {
            BTVacuumPosting vacposting = updatable[i];
            Size        itemsz;
            xl_btree_update update;
 
            update.ndeletedtids = vacposting->ndeletedtids;
            memcpy(updatedbuf + offset, &update.ndeletedtids,
                   SizeOfBtreeUpdate);
            offset += SizeOfBtreeUpdate;
 
            itemsz = update.ndeletedtids * sizeof(uint16);
            memcpy(updatedbuf + offset, vacposting->deletetids, itemsz);
            offset += itemsz;
        }
    }
 
    return updatedbuf;
}

References _bt_update_posting(), Assert, fb(), i, memcpy(), xl_btree_update::ndeletedtids, palloc(), and SizeOfBtreeUpdate.

Referenced by _bt_delitems_delete(), and _bt_delitems_vacuum().

_bt_delitems_vacuum()

void _bt_delitems_vacuum	(	Relation	rel,
		Buffer	buf,
		OffsetNumber *	deletable,
		int	ndeletable,
		BTVacuumPosting *	updatable,
		int	nupdatable
	)

Definition at line 1162 of file nbtpage.c.

{
    Page        page = BufferGetPage(buf);
    BTPageOpaque opaque;
    bool        needswal = RelationNeedsWAL(rel);
    char       *updatedbuf = NULL;
    Size        updatedbuflen = 0;
    OffsetNumber updatedoffsets[MaxIndexTuplesPerPage];
    XLogRecPtr  recptr;
 
    /* Shouldn't be called unless there's something to do */
    Assert(ndeletable > 0 || nupdatable > 0);
 
    /* Generate new version of posting lists without deleted TIDs */
    if (nupdatable > 0)
        updatedbuf = _bt_delitems_update(updatable, nupdatable,
                                         updatedoffsets, &updatedbuflen,
                                         needswal);
 
    /* No ereport(ERROR) until changes are logged */
    START_CRIT_SECTION();
 
    /*
     * Handle posting tuple updates.
     *
     * Deliberately do this before handling simple deletes.  If we did it the
     * other way around (i.e. WAL record order -- simple deletes before
     * updates) then we'd have to make compensating changes to the 'updatable'
     * array of offset numbers.
     *
     * PageIndexTupleOverwrite() won't unset each item's LP_DEAD bit when it
     * happens to already be set.  It's important that we not interfere with
     * any future simple index tuple deletion operations.
     */
    for (int i = 0; i < nupdatable; i++)
    {
        OffsetNumber updatedoffset = updatedoffsets[i];
        IndexTuple  itup;
        Size        itemsz;
 
        itup = updatable[i]->itup;
        itemsz = MAXALIGN(IndexTupleSize(itup));
        if (!PageIndexTupleOverwrite(page, updatedoffset, itup, itemsz))
            elog(PANIC, "failed to update partially dead item in block %u of index \"%s\"",
                 BufferGetBlockNumber(buf), RelationGetRelationName(rel));
    }
 
    /* Now handle simple deletes of entire tuples */
    if (ndeletable > 0)
        PageIndexMultiDelete(page, deletable, ndeletable);
 
    /*
     * We can clear the vacuum cycle ID since this page has certainly been
     * processed by the current vacuum scan.
     */
    opaque = BTPageGetOpaque(page);
    opaque->btpo_cycleid = 0;
 
    /*
     * Clear the BTP_HAS_GARBAGE page flag.
     *
     * This flag indicates the presence of LP_DEAD items on the page (though
     * not reliably).  Note that we only rely on it with pg_upgrade'd
     * !heapkeyspace indexes.  That's why clearing it here won't usually
     * interfere with simple index tuple deletion.
     */
    opaque->btpo_flags &= ~BTP_HAS_GARBAGE;
 
    MarkBufferDirty(buf);
 
    /* XLOG stuff */
    if (needswal)
    {
        xl_btree_vacuum xlrec_vacuum;
 
        xlrec_vacuum.ndeleted = ndeletable;
        xlrec_vacuum.nupdated = nupdatable;
 
        XLogBeginInsert();
        XLogRegisterBuffer(0, buf, REGBUF_STANDARD);
        XLogRegisterData(&xlrec_vacuum, SizeOfBtreeVacuum);
 
        if (ndeletable > 0)
            XLogRegisterBufData(0, deletable,
                                ndeletable * sizeof(OffsetNumber));
 
        if (nupdatable > 0)
        {
            XLogRegisterBufData(0, updatedoffsets,
                                nupdatable * sizeof(OffsetNumber));
            XLogRegisterBufData(0, updatedbuf, updatedbuflen);
        }
 
        recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_VACUUM);
    }
    else
        recptr = XLogGetFakeLSN(rel);
 
    PageSetLSN(page, recptr);
 
    END_CRIT_SECTION();
 
    /* can't leak memory here */
    if (updatedbuf != NULL)
        pfree(updatedbuf);
    /* free tuples allocated within _bt_delitems_update() */
    for (int i = 0; i < nupdatable; i++)
        pfree(updatable[i]->itup);
}

References _bt_delitems_update(), Assert, BTPageGetOpaque, BTPageOpaqueData::btpo_cycleid, BTPageOpaqueData::btpo_flags, buf, BufferGetBlockNumber(), BufferGetPage(), elog, END_CRIT_SECTION, fb(), i, IndexTupleSize(), BTVacuumPostingData::itup, MarkBufferDirty(), MAXALIGN, MaxIndexTuplesPerPage, xl_btree_vacuum::ndeleted, PageIndexMultiDelete(), PageIndexTupleOverwrite(), PageSetLSN(), PANIC, pfree(), REGBUF_STANDARD, RelationGetRelationName, RelationNeedsWAL, SizeOfBtreeVacuum, START_CRIT_SECTION, XLOG_BTREE_VACUUM, XLogBeginInsert(), XLogGetFakeLSN(), XLogInsert(), XLogRegisterBufData(), XLogRegisterBuffer(), and XLogRegisterData().

Referenced by btvacuumpage().

_bt_getbuf()

Buffer _bt_getbuf	(	Relation	rel,
		BlockNumber	blkno,
		int	access
	)

Definition at line 830 of file nbtpage.c.

{
    Buffer      buf;
 
    Assert(BlockNumberIsValid(blkno));
 
    /* Read an existing block of the relation */
    buf = ReadBuffer(rel, blkno);
    _bt_lockbuf(rel, buf, access);
    _bt_checkpage(rel, buf);
 
    return buf;
}

References _bt_checkpage(), _bt_lockbuf(), Assert, BlockNumberIsValid(), buf, and ReadBuffer().

Referenced by _bt_finish_split(), _bt_getroot(), _bt_getrootheight(), _bt_getstackbuf(), _bt_gettrueroot(), _bt_insertonpg(), _bt_killitems(), _bt_leftsib_splitflag(), _bt_lock_and_validate_left(), _bt_metaversion(), _bt_moveright(), _bt_newlevel(), _bt_pagedel(), _bt_readnextpage(), _bt_rightsib_halfdeadflag(), _bt_set_cleanup_info(), _bt_split(), _bt_unlink_halfdead_page(), and _bt_vacuum_needs_cleanup().

_bt_getmeta()

static BTMetaPageData * _bt_getmeta ( Relation  rel, Buffer  metabuf  )

static

Definition at line 143 of file nbtpage.c.

{
    Page        metapg;
    BTPageOpaque metaopaque;
    BTMetaPageData *metad;
 
    metapg = BufferGetPage(metabuf);
    metaopaque = BTPageGetOpaque(metapg);
    metad = BTPageGetMeta(metapg);
 
    /* sanity-check the metapage */
    if (!P_ISMETA(metaopaque) ||
        metad->btm_magic != BTREE_MAGIC)
        ereport(ERROR,
                (errcode(ERRCODE_INDEX_CORRUPTED),
                 errmsg("index \"%s\" is not a btree",
                        RelationGetRelationName(rel))));
 
    if (metad->btm_version < BTREE_MIN_VERSION ||
        metad->btm_version > BTREE_VERSION)
        ereport(ERROR,
                (errcode(ERRCODE_INDEX_CORRUPTED),
                 errmsg("version mismatch in index \"%s\": file version %d, "
                        "current version %d, minimal supported version %d",
                        RelationGetRelationName(rel),
                        metad->btm_version, BTREE_VERSION, BTREE_MIN_VERSION)));
 
    return metad;
}

References BTPageGetMeta, BTPageGetOpaque, BTREE_MAGIC, BTREE_MIN_VERSION, BTREE_VERSION, BufferGetPage(), ereport, errcode(), errmsg, ERROR, fb(), P_ISMETA, and RelationGetRelationName.

Referenced by _bt_getroot(), _bt_getrootheight(), _bt_gettrueroot(), and _bt_metaversion().

_bt_getroot()

Buffer _bt_getroot	(	Relation	rel,
		Relation	heaprel,
		int	access
	)

Definition at line 347 of file nbtpage.c.

{
    Buffer      metabuf;
    Buffer      rootbuf;
    Page        rootpage;
    BTPageOpaque rootopaque;
    BlockNumber rootblkno;
    uint32      rootlevel;
    BTMetaPageData *metad;
    XLogRecPtr  recptr;
 
    Assert(access == BT_READ || heaprel != NULL);
 
    /*
     * Try to use previously-cached metapage data to find the root.  This
     * normally saves one buffer access per index search, which is a very
     * helpful savings in bufmgr traffic and hence contention.
     */
    if (rel->rd_amcache != NULL)
    {
        metad = (BTMetaPageData *) rel->rd_amcache;
        /* We shouldn't have cached it if any of these fail */
        Assert(metad->btm_magic == BTREE_MAGIC);
        Assert(metad->btm_version >= BTREE_MIN_VERSION);
        Assert(metad->btm_version <= BTREE_VERSION);
        Assert(!metad->btm_allequalimage ||
               metad->btm_version > BTREE_NOVAC_VERSION);
        Assert(metad->btm_root != P_NONE);
 
        rootblkno = metad->btm_fastroot;
        Assert(rootblkno != P_NONE);
        rootlevel = metad->btm_fastlevel;
 
        rootbuf = _bt_getbuf(rel, rootblkno, BT_READ);
        rootpage = BufferGetPage(rootbuf);
        rootopaque = BTPageGetOpaque(rootpage);
 
        /*
         * Since the cache might be stale, we check the page more carefully
         * here than normal.  We *must* check that it's not deleted. If it's
         * not alone on its level, then we reject too --- this may be overly
         * paranoid but better safe than sorry.  Note we don't check P_ISROOT,
         * because that's not set in a "fast root".
         */
        if (!P_IGNORE(rootopaque) &&
            rootopaque->btpo_level == rootlevel &&
            P_LEFTMOST(rootopaque) &&
            P_RIGHTMOST(rootopaque))
        {
            /* OK, accept cached page as the root */
            return rootbuf;
        }
        _bt_relbuf(rel, rootbuf);
        /* Cache is stale, throw it away */
        if (rel->rd_amcache)
            pfree(rel->rd_amcache);
        rel->rd_amcache = NULL;
    }
 
    metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
    metad = _bt_getmeta(rel, metabuf);
 
    /* if no root page initialized yet, do it */
    if (metad->btm_root == P_NONE)
    {
        Page        metapg;
 
        /* If access = BT_READ, caller doesn't want us to create root yet */
        if (access == BT_READ)
        {
            _bt_relbuf(rel, metabuf);
            return InvalidBuffer;
        }
 
        /* trade in our read lock for a write lock */
        _bt_unlockbuf(rel, metabuf);
        _bt_lockbuf(rel, metabuf, BT_WRITE);
 
        /*
         * Race condition:  if someone else initialized the metadata between
         * the time we released the read lock and acquired the write lock, we
         * must avoid doing it again.
         */
        if (metad->btm_root != P_NONE)
        {
            /*
             * Metadata initialized by someone else.  In order to guarantee no
             * deadlocks, we have to release the metadata page and start all
             * over again.  (Is that really true? But it's hardly worth trying
             * to optimize this case.)
             */
            _bt_relbuf(rel, metabuf);
            return _bt_getroot(rel, heaprel, access);
        }
 
        /*
         * Get, initialize, write, and leave a lock of the appropriate type on
         * the new root page.  Since this is the first page in the tree, it's
         * a leaf as well as the root.
         */
        rootbuf = _bt_allocbuf(rel, heaprel);
        rootblkno = BufferGetBlockNumber(rootbuf);
        rootpage = BufferGetPage(rootbuf);
        rootopaque = BTPageGetOpaque(rootpage);
        rootopaque->btpo_prev = rootopaque->btpo_next = P_NONE;
        rootopaque->btpo_flags = (BTP_LEAF | BTP_ROOT);
        rootopaque->btpo_level = 0;
        rootopaque->btpo_cycleid = 0;
        /* Get raw page pointer for metapage */
        metapg = BufferGetPage(metabuf);
 
        /* NO ELOG(ERROR) till meta is updated */
        START_CRIT_SECTION();
 
        /* upgrade metapage if needed */
        if (metad->btm_version < BTREE_NOVAC_VERSION)
            _bt_upgrademetapage(metapg);
 
        metad->btm_root = rootblkno;
        metad->btm_level = 0;
        metad->btm_fastroot = rootblkno;
        metad->btm_fastlevel = 0;
        metad->btm_last_cleanup_num_delpages = 0;
        metad->btm_last_cleanup_num_heap_tuples = -1.0;
 
        MarkBufferDirty(rootbuf);
        MarkBufferDirty(metabuf);
 
        /* XLOG stuff */
        if (RelationNeedsWAL(rel))
        {
            xl_btree_newroot xlrec;
            xl_btree_metadata md;
 
            XLogBeginInsert();
            XLogRegisterBuffer(0, rootbuf, REGBUF_WILL_INIT);
            XLogRegisterBuffer(2, metabuf, REGBUF_WILL_INIT | REGBUF_STANDARD);
 
            Assert(metad->btm_version >= BTREE_NOVAC_VERSION);
            md.version = metad->btm_version;
            md.root = rootblkno;
            md.level = 0;
            md.fastroot = rootblkno;
            md.fastlevel = 0;
            md.last_cleanup_num_delpages = 0;
            md.allequalimage = metad->btm_allequalimage;
 
            XLogRegisterBufData(2, &md, sizeof(xl_btree_metadata));
 
            xlrec.rootblk = rootblkno;
            xlrec.level = 0;
 
            XLogRegisterData(&xlrec, SizeOfBtreeNewroot);
 
            recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_NEWROOT);
        }
        else
            recptr = XLogGetFakeLSN(rel);
 
        PageSetLSN(rootpage, recptr);
        PageSetLSN(metapg, recptr);
 
        END_CRIT_SECTION();
 
        /*
         * swap root write lock for read lock.  There is no danger of anyone
         * else accessing the new root page while it's unlocked, since no one
         * else knows where it is yet.
         */
        _bt_unlockbuf(rel, rootbuf);
        _bt_lockbuf(rel, rootbuf, BT_READ);
 
        /* okay, metadata is correct, release lock on it without caching */
        _bt_relbuf(rel, metabuf);
    }
    else
    {
        rootblkno = metad->btm_fastroot;
        Assert(rootblkno != P_NONE);
        rootlevel = metad->btm_fastlevel;
 
        /*
         * Cache the metapage data for next time
         */
        rel->rd_amcache = MemoryContextAlloc(rel->rd_indexcxt,
                                             sizeof(BTMetaPageData));
        memcpy(rel->rd_amcache, metad, sizeof(BTMetaPageData));
 
        /*
         * We are done with the metapage; arrange to release it via first
         * _bt_relandgetbuf call
         */
        rootbuf = metabuf;
 
        for (;;)
        {
            rootbuf = _bt_relandgetbuf(rel, rootbuf, rootblkno, BT_READ);
            rootpage = BufferGetPage(rootbuf);
            rootopaque = BTPageGetOpaque(rootpage);
 
            if (!P_IGNORE(rootopaque))
                break;
 
            /* it's dead, Jim.  step right one page */
            if (P_RIGHTMOST(rootopaque))
                elog(ERROR, "no live root page found in index \"%s\"",
                     RelationGetRelationName(rel));
            rootblkno = rootopaque->btpo_next;
        }
 
        if (rootopaque->btpo_level != rootlevel)
            elog(ERROR, "root page %u of index \"%s\" has level %u, expected %u",
                 rootblkno, RelationGetRelationName(rel),
                 rootopaque->btpo_level, rootlevel);
    }
 
    /*
     * By here, we have a pin and read lock on the root page, and no lock set
     * on the metadata page.  Return the root page's buffer.
     */
    return rootbuf;
}

References _bt_allocbuf(), _bt_getbuf(), _bt_getmeta(), _bt_getroot(), _bt_lockbuf(), _bt_relandgetbuf(), _bt_relbuf(), _bt_unlockbuf(), _bt_upgrademetapage(), xl_btree_metadata::allequalimage, Assert, BT_READ, BT_WRITE, BTP_LEAF, BTP_ROOT, BTPageGetOpaque, BTREE_MAGIC, BTREE_METAPAGE, BTREE_MIN_VERSION, BTREE_NOVAC_VERSION, BTREE_VERSION, BufferGetBlockNumber(), BufferGetPage(), elog, END_CRIT_SECTION, ERROR, xl_btree_metadata::fastlevel, xl_btree_metadata::fastroot, fb(), InvalidBuffer, xl_btree_metadata::last_cleanup_num_delpages, xl_btree_metadata::level, MarkBufferDirty(), memcpy(), MemoryContextAlloc(), P_IGNORE, P_LEFTMOST, P_NONE, P_RIGHTMOST, PageSetLSN(), pfree(), RelationData::rd_amcache, RelationData::rd_indexcxt, REGBUF_STANDARD, REGBUF_WILL_INIT, RelationGetRelationName, RelationNeedsWAL, xl_btree_metadata::root, SizeOfBtreeNewroot, START_CRIT_SECTION, xl_btree_metadata::version, XLOG_BTREE_NEWROOT, XLogBeginInsert(), XLogGetFakeLSN(), XLogInsert(), XLogRegisterBufData(), XLogRegisterBuffer(), and XLogRegisterData().

Referenced by _bt_get_endpoint(), _bt_getroot(), and _bt_search().

_bt_getrootheight()

int _bt_getrootheight

(

Relation

rel

)

Definition at line 660 of file nbtpage.c.

{
    BTMetaPageData *metad;
 
    if (rel->rd_amcache == NULL)
    {
        Buffer      metabuf;
 
        metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
        metad = _bt_getmeta(rel, metabuf);
 
        /*
         * If there's no root page yet, _bt_getroot() doesn't expect a cache
         * to be made, so just stop here and report the index height is zero.
         * (XXX perhaps _bt_getroot() should be changed to allow this case.)
         */
        if (metad->btm_root == P_NONE)
        {
            _bt_relbuf(rel, metabuf);
            return 0;
        }
 
        /*
         * Cache the metapage data for next time
         */
        rel->rd_amcache = MemoryContextAlloc(rel->rd_indexcxt,
                                             sizeof(BTMetaPageData));
        memcpy(rel->rd_amcache, metad, sizeof(BTMetaPageData));
        _bt_relbuf(rel, metabuf);
    }
 
    /* Get cached page */
    metad = (BTMetaPageData *) rel->rd_amcache;
    /* We shouldn't have cached it if any of these fail */
    Assert(metad->btm_magic == BTREE_MAGIC);
    Assert(metad->btm_version >= BTREE_MIN_VERSION);
    Assert(metad->btm_version <= BTREE_VERSION);
    Assert(!metad->btm_allequalimage ||
           metad->btm_version > BTREE_NOVAC_VERSION);
    Assert(metad->btm_fastroot != P_NONE);
 
    return metad->btm_fastlevel;
}

References _bt_getbuf(), _bt_getmeta(), _bt_relbuf(), Assert, BT_READ, BTREE_MAGIC, BTREE_METAPAGE, BTREE_MIN_VERSION, BTREE_NOVAC_VERSION, BTREE_VERSION, fb(), memcpy(), MemoryContextAlloc(), P_NONE, RelationData::rd_amcache, and RelationData::rd_indexcxt.

Referenced by _bt_insertonpg(), and btgettreeheight().

_bt_gettrueroot()

Buffer _bt_gettrueroot

(

Relation

rel

)

Definition at line 585 of file nbtpage.c.

{
    Buffer      metabuf;
    Buffer      rootbuf;
    Page        rootpage;
    BTPageOpaque rootopaque;
    BlockNumber rootblkno;
    uint32      rootlevel;
    BTMetaPageData *metad;
 
    /*
     * We don't try to use cached metapage data here, since (a) this path is
     * not performance-critical, and (b) if we are here it suggests our cache
     * is out-of-date anyway.  In light of point (b), it's probably safest to
     * actively flush any cached metapage info.
     */
    if (rel->rd_amcache)
        pfree(rel->rd_amcache);
    rel->rd_amcache = NULL;
 
    metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
    metad = _bt_getmeta(rel, metabuf);
 
    /* if no root page initialized yet, fail */
    if (metad->btm_root == P_NONE)
    {
        _bt_relbuf(rel, metabuf);
        return InvalidBuffer;
    }
 
    rootblkno = metad->btm_root;
    rootlevel = metad->btm_level;
 
    /*
     * We are done with the metapage; arrange to release it via first
     * _bt_relandgetbuf call
     */
    rootbuf = metabuf;
 
    for (;;)
    {
        rootbuf = _bt_relandgetbuf(rel, rootbuf, rootblkno, BT_READ);
        rootpage = BufferGetPage(rootbuf);
        rootopaque = BTPageGetOpaque(rootpage);
 
        if (!P_IGNORE(rootopaque))
            break;
 
        /* it's dead, Jim.  step right one page */
        if (P_RIGHTMOST(rootopaque))
            elog(ERROR, "no live root page found in index \"%s\"",
                 RelationGetRelationName(rel));
        rootblkno = rootopaque->btpo_next;
    }
 
    if (rootopaque->btpo_level != rootlevel)
        elog(ERROR, "root page %u of index \"%s\" has level %u, expected %u",
             rootblkno, RelationGetRelationName(rel),
             rootopaque->btpo_level, rootlevel);
 
    return rootbuf;
}

References _bt_getbuf(), _bt_getmeta(), _bt_relandgetbuf(), _bt_relbuf(), BT_READ, BTPageGetOpaque, BTREE_METAPAGE, BufferGetPage(), elog, ERROR, fb(), InvalidBuffer, P_IGNORE, P_NONE, P_RIGHTMOST, pfree(), RelationData::rd_amcache, and RelationGetRelationName.

Referenced by _bt_get_endpoint().

_bt_initmetapage()

void _bt_initmetapage	(	Page	page,
		BlockNumber	rootbknum,
		uint32	level,
		bool	allequalimage
	)

Definition at line 68 of file nbtpage.c.

{
    BTMetaPageData *metad;
    BTPageOpaque metaopaque;
 
    _bt_pageinit(page, BLCKSZ);
 
    metad = BTPageGetMeta(page);
    metad->btm_magic = BTREE_MAGIC;
    metad->btm_version = BTREE_VERSION;
    metad->btm_root = rootbknum;
    metad->btm_level = level;
    metad->btm_fastroot = rootbknum;
    metad->btm_fastlevel = level;
    metad->btm_last_cleanup_num_delpages = 0;
    metad->btm_last_cleanup_num_heap_tuples = -1.0;
    metad->btm_allequalimage = allequalimage;
 
    metaopaque = BTPageGetOpaque(page);
    metaopaque->btpo_flags = BTP_META;
 
    /*
     * Set pd_lower just past the end of the metadata.  This is essential,
     * because without doing so, metadata will be lost if xlog.c compresses
     * the page.
     */
    ((PageHeader) page)->pd_lower =
        ((char *) metad + sizeof(BTMetaPageData)) - (char *) page;
}

References _bt_pageinit(), BTP_META, BTPageGetMeta, BTPageGetOpaque, BTREE_MAGIC, BTREE_VERSION, and fb().

Referenced by _bt_uppershutdown(), and btbuildempty().

_bt_leftsib_splitflag()

static bool _bt_leftsib_splitflag ( Relation  rel, BlockNumber  leftsib, BlockNumber  target  )

static

Definition at line 1705 of file nbtpage.c.

{
    Buffer      buf;
    Page        page;
    BTPageOpaque opaque;
    bool        result;
 
    /* Easy case: No left sibling */
    if (leftsib == P_NONE)
        return false;
 
    buf = _bt_getbuf(rel, leftsib, BT_READ);
    page = BufferGetPage(buf);
    opaque = BTPageGetOpaque(page);
 
    /*
     * If the left sibling was concurrently split, so that its next-pointer
     * doesn't point to the current page anymore, the split that created
     * target must be completed.  Caller can reasonably expect that there will
     * be a downlink to the target page that it can relocate using its stack.
     * (We don't allow splitting an incompletely split page again until the
     * previous split has been completed.)
     */
    result = (opaque->btpo_next == target && P_INCOMPLETE_SPLIT(opaque));
    _bt_relbuf(rel, buf);
 
    return result;
}

References _bt_getbuf(), _bt_relbuf(), BT_READ, BTPageGetOpaque, BTPageOpaqueData::btpo_next, buf, BufferGetPage(), P_INCOMPLETE_SPLIT, P_NONE, and result.

Referenced by _bt_lock_subtree_parent(), and _bt_pagedel().

_bt_lock_subtree_parent()

static bool _bt_lock_subtree_parent ( Relation  rel, Relation  heaprel, BlockNumber  child, BTStack  stack, Buffer *  subtreeparent, OffsetNumber *  poffset, BlockNumber *  topparent, BlockNumber *  topparentrightsib  )

static

Definition at line 2830 of file nbtpage.c.

{
    BlockNumber parent,
                leftsibparent;
    OffsetNumber parentoffset,
                maxoff;
    Buffer      pbuf;
    Page        page;
    BTPageOpaque opaque;
 
    /*
     * Locate the pivot tuple whose downlink points to "child".  Write lock
     * the parent page itself.
     */
    pbuf = _bt_getstackbuf(rel, heaprel, stack, child);
    if (pbuf == InvalidBuffer)
    {
        /*
         * Failed to "re-find" a pivot tuple whose downlink matched our child
         * block number on the parent level -- the index must be corrupt.
         * Don't even try to delete the leafbuf subtree.  Just report the
         * issue and press on with vacuuming the index.
         *
         * Note: _bt_getstackbuf() recovers from concurrent page splits that
         * take place on the parent level.  Its approach is a near-exhaustive
         * linear search.  This also gives it a surprisingly good chance of
         * recovering in the event of a buggy or inconsistent opclass.  But we
         * don't rely on that here.
         */
        ereport(LOG,
                (errcode(ERRCODE_INDEX_CORRUPTED),
                 errmsg_internal("failed to re-find parent key in index \"%s\" for deletion target page %u",
                                 RelationGetRelationName(rel), child)));
        Assert(false);
        return false;
    }
 
    parent = stack->bts_blkno;
    parentoffset = stack->bts_offset;
 
    page = BufferGetPage(pbuf);
    opaque = BTPageGetOpaque(page);
    maxoff = PageGetMaxOffsetNumber(page);
    leftsibparent = opaque->btpo_prev;
 
    /*
     * _bt_getstackbuf() completes page splits on returned parent buffer when
     * required.
     *
     * In general it's a bad idea for VACUUM to use up more disk space, which
     * is why page deletion does not finish incomplete page splits most of the
     * time.  We allow this limited exception because the risk is much lower,
     * and the potential downside of not proceeding is much higher:  A single
     * internal page with the INCOMPLETE_SPLIT flag set might otherwise
     * prevent us from deleting hundreds of empty leaf pages from one level
     * down.
     */
    Assert(!P_INCOMPLETE_SPLIT(opaque));
 
    if (parentoffset < maxoff)
    {
        /*
         * Child is not the rightmost child in parent, so it's safe to delete
         * the subtree whose root/topparent is child page
         */
        *subtreeparent = pbuf;
        *poffset = parentoffset;
        return true;
    }
 
    /*
     * Child is the rightmost child of parent.
     *
     * Since it's the rightmost child of parent, deleting the child (or
     * deleting the subtree whose root/topparent is the child page) is only
     * safe when it's also possible to delete the parent.
     */
    Assert(parentoffset == maxoff);
    if (parentoffset != P_FIRSTDATAKEY(opaque) || P_RIGHTMOST(opaque))
    {
        /*
         * Child isn't parent's only child, or parent is rightmost on its
         * entire level.  Definitely cannot delete any pages.
         */
        _bt_relbuf(rel, pbuf);
        return false;
    }
 
    /*
     * Now make sure that the parent deletion is itself safe by examining the
     * child's grandparent page.  Recurse, passing the parent page as the
     * child page (child's grandparent is the parent on the next level up). If
     * parent deletion is unsafe, then child deletion must also be unsafe (in
     * which case caller cannot delete any pages at all).
     */
    *topparent = parent;
    *topparentrightsib = opaque->btpo_next;
 
    /*
     * Release lock on parent before recursing.
     *
     * It's OK to release page locks on parent before recursive call locks
     * grandparent.  An internal page can only acquire an entry if the child
     * is split, but that cannot happen as long as we still hold a lock on the
     * leafbuf page.
     */
    _bt_relbuf(rel, pbuf);
 
    /*
     * Before recursing, check that the left sibling of parent (if any) is not
     * marked with INCOMPLETE_SPLIT flag first (must do so after we drop the
     * parent lock).
     *
     * Note: We deliberately avoid completing incomplete splits here.
     */
    if (_bt_leftsib_splitflag(rel, leftsibparent, parent))
        return false;
 
    /* Recurse to examine child page's grandparent page */
    return _bt_lock_subtree_parent(rel, heaprel, parent, stack->bts_parent,
                                   subtreeparent, poffset,
                                   topparent, topparentrightsib);
}

References _bt_getstackbuf(), _bt_leftsib_splitflag(), _bt_lock_subtree_parent(), _bt_relbuf(), Assert, BTPageGetOpaque, BTPageOpaqueData::btpo_next, BTPageOpaqueData::btpo_prev, BTStackData::bts_blkno, BTStackData::bts_offset, BTStackData::bts_parent, BufferGetPage(), ereport, errcode(), errmsg_internal(), fb(), InvalidBuffer, LOG, P_FIRSTDATAKEY, P_INCOMPLETE_SPLIT, P_RIGHTMOST, PageGetMaxOffsetNumber(), and RelationGetRelationName.

Referenced by _bt_lock_subtree_parent(), and _bt_mark_page_halfdead().

_bt_lockbuf()

void _bt_lockbuf	(	Relation	rel,
		Buffer	buf,
		int	access
	)

Definition at line 1047 of file nbtpage.c.

{
    /* LockBuffer() asserts that pin is held by this backend */
    LockBuffer(buf, access);
 
    /*
     * It doesn't matter that _bt_unlockbuf() won't get called in the event of
     * an nbtree error (e.g. a unique violation error).  That won't cause
     * Valgrind false positives.
     *
     * The nbtree client requests are superimposed on top of the bufmgr.c
     * buffer pin client requests.  In the event of an nbtree error the buffer
     * will certainly get marked as defined when the backend once again
     * acquires its first pin on the buffer. (Of course, if the backend never
     * touches the buffer again then it doesn't matter that it remains
     * non-accessible to Valgrind.)
     *
     * Note: When an IndexTuple C pointer gets computed using an ItemId read
     * from a page while a lock was held, the C pointer becomes unsafe to
     * dereference forever as soon as the lock is released.  Valgrind can only
     * detect cases where the pointer gets dereferenced with no _current_
     * lock/pin held, though.
     */
    if (!RelationUsesLocalBuffers(rel))
        VALGRIND_MAKE_MEM_DEFINED(BufferGetPage(buf), BLCKSZ);
}

References buf, BufferGetPage(), fb(), LockBuffer(), RelationUsesLocalBuffers, and VALGRIND_MAKE_MEM_DEFINED.

Referenced by _bt_getbuf(), _bt_getroot(), _bt_killitems(), _bt_moveright(), _bt_pagedel(), _bt_relandgetbuf(), _bt_search(), _bt_set_cleanup_info(), _bt_unlink_halfdead_page(), and btvacuumpage().

_bt_mark_page_halfdead()

static bool _bt_mark_page_halfdead ( Relation  rel, Relation  heaprel, Buffer  leafbuf, BTStack  stack  )

static

Definition at line 2102 of file nbtpage.c.

{
    BlockNumber leafblkno;
    BlockNumber leafrightsib;
    BlockNumber topparent;
    BlockNumber topparentrightsib;
    ItemId      itemid;
    Page        page;
    BTPageOpaque opaque;
    Buffer      subtreeparent;
    OffsetNumber poffset;
    OffsetNumber nextoffset;
    IndexTuple  itup;
    IndexTupleData trunctuple;
    XLogRecPtr  recptr;
 
    page = BufferGetPage(leafbuf);
    opaque = BTPageGetOpaque(page);
 
    Assert(!P_RIGHTMOST(opaque) && !P_ISROOT(opaque) &&
           P_ISLEAF(opaque) && !P_IGNORE(opaque) &&
           P_FIRSTDATAKEY(opaque) > PageGetMaxOffsetNumber(page));
    Assert(heaprel != NULL);
 
    /*
     * Save info about the leaf page.
     */
    leafblkno = BufferGetBlockNumber(leafbuf);
    leafrightsib = opaque->btpo_next;
 
    /*
     * Before attempting to lock the parent page, check that the right sibling
     * is not in half-dead state.  A half-dead right sibling would have no
     * downlink in the parent, which would be highly confusing later when we
     * delete the downlink.  It would fail the "right sibling of target page
     * is also the next child in parent page" cross-check below.
     */
    if (_bt_rightsib_halfdeadflag(rel, leafrightsib))
    {
        elog(DEBUG1, "could not delete page %u because its right sibling %u is half-dead",
             leafblkno, leafrightsib);
        return false;
    }
 
    /*
     * We cannot delete a page that is the rightmost child of its immediate
     * parent, unless it is the only child --- in which case the parent has to
     * be deleted too, and the same condition applies recursively to it. We
     * have to check this condition all the way up before trying to delete,
     * and lock the parent of the root of the to-be-deleted subtree (the
     * "subtree parent").  _bt_lock_subtree_parent() locks the subtree parent
     * for us.  We remove the downlink to the "top parent" page (subtree root
     * page) from the subtree parent page below.
     *
     * Initialize topparent to be leafbuf page now.  The final to-be-deleted
     * subtree is often a degenerate one page subtree consisting only of the
     * leafbuf page.  When that happens, the leafbuf page is the final subtree
     * root page/top parent page.
     */
    topparent = leafblkno;
    topparentrightsib = leafrightsib;
    if (!_bt_lock_subtree_parent(rel, heaprel, leafblkno, stack,
                                 &subtreeparent, &poffset,
                                 &topparent, &topparentrightsib))
        return false;
 
    page = BufferGetPage(subtreeparent);
    opaque = BTPageGetOpaque(page);
 
#ifdef USE_ASSERT_CHECKING
 
    /*
     * This is just an assertion because _bt_lock_subtree_parent should have
     * guaranteed tuple has the expected contents
     */
    itemid = PageGetItemId(page, poffset);
    itup = (IndexTuple) PageGetItem(page, itemid);
    Assert(BTreeTupleGetDownLink(itup) == topparent);
#endif
 
    nextoffset = OffsetNumberNext(poffset);
    itemid = PageGetItemId(page, nextoffset);
    itup = (IndexTuple) PageGetItem(page, itemid);
 
    /*
     * Check that the parent-page index items we're about to delete/overwrite
     * in subtree parent page contain what we expect.  This can fail if the
     * index has become corrupt for some reason.  When that happens we back
     * out of deletion of the leafbuf subtree.  (This is just like the case
     * where _bt_lock_subtree_parent() cannot "re-find" leafbuf's downlink.)
     */
    if (BTreeTupleGetDownLink(itup) != topparentrightsib)
    {
        ereport(LOG,
                (errcode(ERRCODE_INDEX_CORRUPTED),
                 errmsg_internal("right sibling %u of block %u is not next child %u of block %u in index \"%s\"",
                                 topparentrightsib, topparent,
                                 BTreeTupleGetDownLink(itup),
                                 BufferGetBlockNumber(subtreeparent),
                                 RelationGetRelationName(rel))));
 
        _bt_relbuf(rel, subtreeparent);
        Assert(false);
        return false;
    }
 
    /*
     * Any insert which would have gone on the leaf block will now go to its
     * right sibling.  In other words, the key space moves right.
     */
    PredicateLockPageCombine(rel, leafblkno, leafrightsib);
 
    /* No ereport(ERROR) until changes are logged */
    START_CRIT_SECTION();
 
    /*
     * Update parent of subtree.  We want to delete the downlink to the top
     * parent page/root of the subtree, and the *following* key.  Easiest way
     * is to copy the right sibling's downlink over the downlink that points
     * to top parent page, and then delete the right sibling's original pivot
     * tuple.
     *
     * Lanin and Shasha make the key space move left when deleting a page,
     * whereas the key space moves right here.  That's why we cannot simply
     * delete the pivot tuple with the downlink to the top parent page.  See
     * nbtree/README.
     */
    page = BufferGetPage(subtreeparent);
    opaque = BTPageGetOpaque(page);
 
    itemid = PageGetItemId(page, poffset);
    itup = (IndexTuple) PageGetItem(page, itemid);
    BTreeTupleSetDownLink(itup, topparentrightsib);
 
    nextoffset = OffsetNumberNext(poffset);
    PageIndexTupleDelete(page, nextoffset);
 
    /*
     * Mark the leaf page as half-dead, and stamp it with a link to the top
     * parent page.  When the leaf page is also the top parent page, the link
     * is set to InvalidBlockNumber.
     */
    page = BufferGetPage(leafbuf);
    opaque = BTPageGetOpaque(page);
    opaque->btpo_flags |= BTP_HALF_DEAD;
 
    Assert(PageGetMaxOffsetNumber(page) == P_HIKEY);
    MemSet(&trunctuple, 0, sizeof(IndexTupleData));
    trunctuple.t_info = sizeof(IndexTupleData);
    if (topparent != leafblkno)
        BTreeTupleSetTopParent(&trunctuple, topparent);
    else
        BTreeTupleSetTopParent(&trunctuple, InvalidBlockNumber);
 
    if (!PageIndexTupleOverwrite(page, P_HIKEY, &trunctuple, IndexTupleSize(&trunctuple)))
        elog(ERROR, "could not overwrite high key in half-dead page");
 
    /* Must mark buffers dirty before XLogInsert */
    MarkBufferDirty(subtreeparent);
    MarkBufferDirty(leafbuf);
 
    /* XLOG stuff */
    if (RelationNeedsWAL(rel))
    {
        xl_btree_mark_page_halfdead xlrec;
 
        xlrec.poffset = poffset;
        xlrec.leafblk = leafblkno;
        if (topparent != leafblkno)
            xlrec.topparent = topparent;
        else
            xlrec.topparent = InvalidBlockNumber;
 
        XLogBeginInsert();
        XLogRegisterBuffer(0, leafbuf, REGBUF_WILL_INIT);
        XLogRegisterBuffer(1, subtreeparent, REGBUF_STANDARD);
 
        page = BufferGetPage(leafbuf);
        opaque = BTPageGetOpaque(page);
        xlrec.leftblk = opaque->btpo_prev;
        xlrec.rightblk = opaque->btpo_next;
 
        XLogRegisterData(&xlrec, SizeOfBtreeMarkPageHalfDead);
 
        recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_MARK_PAGE_HALFDEAD);
    }
    else
        recptr = XLogGetFakeLSN(rel);
 
    page = BufferGetPage(subtreeparent);
    PageSetLSN(page, recptr);
    page = BufferGetPage(leafbuf);
    PageSetLSN(page, recptr);
 
    END_CRIT_SECTION();
 
    _bt_relbuf(rel, subtreeparent);
    return true;
}

References _bt_lock_subtree_parent(), _bt_relbuf(), _bt_rightsib_halfdeadflag(), Assert, BTP_HALF_DEAD, BTPageGetOpaque, BTPageOpaqueData::btpo_flags, BTPageOpaqueData::btpo_next, BTPageOpaqueData::btpo_prev, BTreeTupleGetDownLink(), BTreeTupleSetDownLink(), BTreeTupleSetTopParent(), BufferGetBlockNumber(), BufferGetPage(), DEBUG1, elog, END_CRIT_SECTION, ereport, errcode(), errmsg_internal(), ERROR, fb(), IndexTupleSize(), InvalidBlockNumber, LOG, MarkBufferDirty(), MemSet, OffsetNumberNext, P_FIRSTDATAKEY, P_HIKEY, P_IGNORE, P_ISLEAF, P_ISROOT, P_RIGHTMOST, PageGetItem(), PageGetItemId(), PageGetMaxOffsetNumber(), PageIndexTupleDelete(), PageIndexTupleOverwrite(), PageSetLSN(), xl_btree_mark_page_halfdead::poffset, PredicateLockPageCombine(), REGBUF_STANDARD, REGBUF_WILL_INIT, RelationGetRelationName, RelationNeedsWAL, SizeOfBtreeMarkPageHalfDead, START_CRIT_SECTION, XLOG_BTREE_MARK_PAGE_HALFDEAD, XLogBeginInsert(), XLogGetFakeLSN(), XLogInsert(), XLogRegisterBuffer(), and XLogRegisterData().

Referenced by _bt_pagedel().

_bt_metaversion()

void _bt_metaversion	(	Relation	rel,
		bool *	heapkeyspace,
		bool *	allequalimage
	)

Definition at line 724 of file nbtpage.c.

{
    BTMetaPageData *metad;
 
    if (rel->rd_amcache == NULL)
    {
        Buffer      metabuf;
 
        metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
        metad = _bt_getmeta(rel, metabuf);
 
        /*
         * If there's no root page yet, _bt_getroot() doesn't expect a cache
         * to be made, so just stop here.  (XXX perhaps _bt_getroot() should
         * be changed to allow this case.)
         */
        if (metad->btm_root == P_NONE)
        {
            *heapkeyspace = metad->btm_version > BTREE_NOVAC_VERSION;
            *allequalimage = metad->btm_allequalimage;
 
            _bt_relbuf(rel, metabuf);
            return;
        }
 
        /*
         * Cache the metapage data for next time
         *
         * An on-the-fly version upgrade performed by _bt_upgrademetapage()
         * can change the nbtree version for an index without invalidating any
         * local cache.  This is okay because it can only happen when moving
         * from version 2 to version 3, both of which are !heapkeyspace
         * versions.
         */
        rel->rd_amcache = MemoryContextAlloc(rel->rd_indexcxt,
                                             sizeof(BTMetaPageData));
        memcpy(rel->rd_amcache, metad, sizeof(BTMetaPageData));
        _bt_relbuf(rel, metabuf);
    }
 
    /* Get cached page */
    metad = (BTMetaPageData *) rel->rd_amcache;
    /* We shouldn't have cached it if any of these fail */
    Assert(metad->btm_magic == BTREE_MAGIC);
    Assert(metad->btm_version >= BTREE_MIN_VERSION);
    Assert(metad->btm_version <= BTREE_VERSION);
    Assert(!metad->btm_allequalimage ||
           metad->btm_version > BTREE_NOVAC_VERSION);
    Assert(metad->btm_fastroot != P_NONE);
 
    *heapkeyspace = metad->btm_version > BTREE_NOVAC_VERSION;
    *allequalimage = metad->btm_allequalimage;
}

References _bt_getbuf(), _bt_getmeta(), _bt_relbuf(), Assert, BT_READ, BTREE_MAGIC, BTREE_METAPAGE, BTREE_MIN_VERSION, BTREE_NOVAC_VERSION, BTREE_VERSION, fb(), memcpy(), MemoryContextAlloc(), P_NONE, RelationData::rd_amcache, and RelationData::rd_indexcxt.

Referenced by _bt_first(), _bt_mkscankey(), and bt_index_check_callback().

_bt_pagedel()

void _bt_pagedel	(	Relation	rel,
		Buffer	leafbuf,
		BTVacState *	vstate
	)

Definition at line 1812 of file nbtpage.c.

{
    BlockNumber rightsib;
    bool        rightsib_empty;
    Page        page;
    BTPageOpaque opaque;
 
    /*
     * Save original leafbuf block number from caller.  Only deleted blocks
     * that are <= scanblkno are added to bulk delete stat's pages_deleted
     * count.
     */
    BlockNumber scanblkno = BufferGetBlockNumber(leafbuf);
 
    /*
     * "stack" is a search stack leading (approximately) to the target page.
     * It is initially NULL, but when iterating, we keep it to avoid
     * duplicated search effort.
     *
     * Also, when "stack" is not NULL, we have already checked that the
     * current page is not the right half of an incomplete split, i.e. the
     * left sibling does not have its INCOMPLETE_SPLIT flag set, including
     * when the current target page is to the right of caller's initial page
     * (the scanblkno page).
     */
    BTStack     stack = NULL;
 
    for (;;)
    {
        page = BufferGetPage(leafbuf);
        opaque = BTPageGetOpaque(page);
 
        /*
         * Internal pages are never deleted directly, only as part of deleting
         * the whole subtree all the way down to leaf level.
         *
         * Also check for deleted pages here.  Caller never passes us a fully
         * deleted page.  Only VACUUM can delete pages, so there can't have
         * been a concurrent deletion.  Assume that we reached any deleted
         * page encountered here by following a sibling link, and that the
         * index is corrupt.
         */
        Assert(!P_ISDELETED(opaque));
        if (!P_ISLEAF(opaque) || P_ISDELETED(opaque))
        {
            /*
             * Pre-9.4 page deletion only marked internal pages as half-dead,
             * but now we only use that flag on leaf pages. The old algorithm
             * was never supposed to leave half-dead pages in the tree, it was
             * just a transient state, but it was nevertheless possible in
             * error scenarios. We don't know how to deal with them here. They
             * are harmless as far as searches are considered, but inserts
             * into the deleted keyspace could add out-of-order downlinks in
             * the upper levels. Log a notice, hopefully the admin will notice
             * and reindex.
             */
            if (P_ISHALFDEAD(opaque))
                ereport(LOG,
                        (errcode(ERRCODE_INDEX_CORRUPTED),
                         errmsg("index \"%s\" contains a half-dead internal page",
                                RelationGetRelationName(rel)),
                         errhint("This can be caused by an interrupted VACUUM in version 9.3 or older, before upgrade. Please REINDEX it.")));
 
            if (P_ISDELETED(opaque))
                ereport(LOG,
                        (errcode(ERRCODE_INDEX_CORRUPTED),
                         errmsg_internal("found deleted block %u while following right link from block %u in index \"%s\"",
                                         BufferGetBlockNumber(leafbuf),
                                         scanblkno,
                                         RelationGetRelationName(rel))));
 
            _bt_relbuf(rel, leafbuf);
            return;
        }
 
        /*
         * We can never delete rightmost pages nor root pages.  While at it,
         * check that page is empty, since it's possible that the leafbuf page
         * was empty a moment ago, but has since had some inserts.
         *
         * To keep the algorithm simple, we also never delete an incompletely
         * split page (they should be rare enough that this doesn't make any
         * meaningful difference to disk usage):
         *
         * The INCOMPLETE_SPLIT flag on the page tells us if the page is the
         * left half of an incomplete split, but ensuring that it's not the
         * right half is more complicated.  For that, we have to check that
         * the left sibling doesn't have its INCOMPLETE_SPLIT flag set using
         * _bt_leftsib_splitflag().  On the first iteration, we temporarily
         * release the lock on scanblkno/leafbuf, check the left sibling, and
         * construct a search stack to scanblkno.  On subsequent iterations,
         * we know we stepped right from a page that passed these tests, so
         * it's OK.
         */
        if (P_RIGHTMOST(opaque) || P_ISROOT(opaque) ||
            P_FIRSTDATAKEY(opaque) <= PageGetMaxOffsetNumber(page) ||
            P_INCOMPLETE_SPLIT(opaque))
        {
            /* Should never fail to delete a half-dead page */
            Assert(!P_ISHALFDEAD(opaque));
 
            _bt_relbuf(rel, leafbuf);
            return;
        }
 
        /*
         * First, remove downlink pointing to the page (or a parent of the
         * page, if we are going to delete a taller subtree), and mark the
         * leafbuf page half-dead
         */
        if (!P_ISHALFDEAD(opaque))
        {
            /*
             * We need an approximate pointer to the page's parent page.  We
             * use a variant of the standard search mechanism to search for
             * the page's high key; this will give us a link to either the
             * current parent or someplace to its left (if there are multiple
             * equal high keys, which is possible with !heapkeyspace indexes).
             *
             * Also check if this is the right-half of an incomplete split
             * (see comment above).
             */
            if (!stack)
            {
                BTScanInsert itup_key;
                ItemId      itemid;
                IndexTuple  targetkey;
                BlockNumber leftsib,
                            leafblkno;
                Buffer      sleafbuf;
 
                itemid = PageGetItemId(page, P_HIKEY);
                targetkey = CopyIndexTuple((IndexTuple) PageGetItem(page, itemid));
 
                leftsib = opaque->btpo_prev;
                leafblkno = BufferGetBlockNumber(leafbuf);
 
                /*
                 * To avoid deadlocks, we'd better drop the leaf page lock
                 * before going further.
                 */
                _bt_unlockbuf(rel, leafbuf);
 
                /*
                 * Check that the left sibling of leafbuf (if any) is not
                 * marked with INCOMPLETE_SPLIT flag before proceeding
                 */
                Assert(leafblkno == scanblkno);
                if (_bt_leftsib_splitflag(rel, leftsib, leafblkno))
                {
                    ReleaseBuffer(leafbuf);
                    return;
                }
 
                /*
                 * We need an insertion scan key, so build one.
                 *
                 * _bt_search searches for the leaf page that contains any
                 * matching non-pivot tuples, but we need it to "search" for
                 * the high key pivot from the page that we're set to delete.
                 * Compensate for the mismatch by having _bt_search locate the
                 * last position < equal-to-untruncated-prefix non-pivots.
                 */
                itup_key = _bt_mkscankey(rel, targetkey);
 
                /* Set up a BTLessStrategyNumber-like insertion scan key */
                itup_key->nextkey = false;
                itup_key->backward = true;
                stack = _bt_search(rel, NULL, itup_key, &sleafbuf, BT_READ, true);
                /* won't need a second lock or pin on leafbuf */
                _bt_relbuf(rel, sleafbuf);
 
                /*
                 * Re-lock the leaf page, and start over to use our stack
                 * within _bt_mark_page_halfdead.  We must do it that way
                 * because it's possible that leafbuf can no longer be
                 * deleted.  We need to recheck.
                 *
                 * Note: We can't simply hold on to the sleafbuf lock instead,
                 * because it's barely possible that sleafbuf is not the same
                 * page as leafbuf.  This happens when leafbuf split after our
                 * original lock was dropped, but before _bt_search finished
                 * its descent.  We rely on the assumption that we'll find
                 * leafbuf isn't safe to delete anymore in this scenario.
                 * (Page deletion can cope with the stack being to the left of
                 * leafbuf, but not to the right of leafbuf.)
                 */
                _bt_lockbuf(rel, leafbuf, BT_WRITE);
                continue;
            }
 
            /*
             * See if it's safe to delete the leaf page, and determine how
             * many parent/internal pages above the leaf level will be
             * deleted.  If it's safe then _bt_mark_page_halfdead will also
             * perform the first phase of deletion, which includes marking the
             * leafbuf page half-dead.
             */
            Assert(P_ISLEAF(opaque) && !P_IGNORE(opaque));
            if (!_bt_mark_page_halfdead(rel, vstate->info->heaprel, leafbuf,
                                        stack))
            {
                _bt_relbuf(rel, leafbuf);
                return;
            }
        }
        else
        {
            INJECTION_POINT("nbtree-finish-half-dead-page-vacuum", NULL);
        }
 
        /*
         * Then unlink it from its siblings.  Each call to
         * _bt_unlink_halfdead_page unlinks the topmost page from the subtree,
         * making it shallower.  Iterate until the leafbuf page is deleted.
         */
        rightsib_empty = false;
        Assert(P_ISLEAF(opaque) && P_ISHALFDEAD(opaque));
        while (P_ISHALFDEAD(opaque))
        {
            /* Check for interrupts in _bt_unlink_halfdead_page */
            if (!_bt_unlink_halfdead_page(rel, leafbuf, scanblkno,
                                          &rightsib_empty, vstate))
            {
                /*
                 * _bt_unlink_halfdead_page should never fail, since we
                 * established that deletion is generally safe in
                 * _bt_mark_page_halfdead -- index must be corrupt.
                 *
                 * Note that _bt_unlink_halfdead_page already released the
                 * lock and pin on leafbuf for us.
                 */
                Assert(false);
                return;
            }
        }
 
        Assert(P_ISLEAF(opaque) && P_ISDELETED(opaque));
 
        rightsib = opaque->btpo_next;
 
        _bt_relbuf(rel, leafbuf);
 
        /*
         * Check here, as calling loops will have locks held, preventing
         * interrupts from being processed.
         */
        CHECK_FOR_INTERRUPTS();
 
        /*
         * The page has now been deleted. If its right sibling is completely
         * empty, it's possible that the reason we haven't deleted it earlier
         * is that it was the rightmost child of the parent. Now that we
         * removed the downlink for this page, the right sibling might now be
         * the only child of the parent, and could be removed. It would be
         * picked up by the next vacuum anyway, but might as well try to
         * remove it now, so loop back to process the right sibling.
         *
         * Note: This relies on the assumption that _bt_getstackbuf() will be
         * able to reuse our original descent stack with a different child
         * block (provided that the child block is to the right of the
         * original leaf page reached by _bt_search()). It will even update
         * the descent stack each time we loop around, avoiding repeated work.
         */
        if (!rightsib_empty)
            break;
 
        leafbuf = _bt_getbuf(rel, rightsib, BT_WRITE);
    }
}

References _bt_getbuf(), _bt_leftsib_splitflag(), _bt_lockbuf(), _bt_mark_page_halfdead(), _bt_mkscankey(), _bt_relbuf(), _bt_search(), _bt_unlink_halfdead_page(), _bt_unlockbuf(), Assert, BTScanInsertData::backward, BT_READ, BT_WRITE, BTPageGetOpaque, BTPageOpaqueData::btpo_next, BTPageOpaqueData::btpo_prev, BufferGetBlockNumber(), BufferGetPage(), CHECK_FOR_INTERRUPTS, CopyIndexTuple(), ereport, errcode(), errhint(), errmsg, errmsg_internal(), fb(), INJECTION_POINT, LOG, BTScanInsertData::nextkey, P_FIRSTDATAKEY, P_HIKEY, P_IGNORE, P_INCOMPLETE_SPLIT, P_ISDELETED, P_ISHALFDEAD, P_ISLEAF, P_ISROOT, P_RIGHTMOST, PageGetItem(), PageGetItemId(), PageGetMaxOffsetNumber(), RelationGetRelationName, and ReleaseBuffer().

Referenced by btvacuumpage().

_bt_pageinit()

void _bt_pageinit	(	Page	page,
		Size	size
	)

Definition at line 1137 of file nbtpage.c.

{
    PageInit(page, size, sizeof(BTPageOpaqueData));
}

References PageInit().

Referenced by _bt_allocbuf(), _bt_blnewpage(), _bt_initmetapage(), _bt_restore_meta(), _bt_split(), btree_xlog_mark_page_halfdead(), btree_xlog_newroot(), btree_xlog_split(), and btree_xlog_unlink_page().

_bt_pendingfsm_add()

static void _bt_pendingfsm_add ( BTVacState *  vstate, BlockNumber  target, FullTransactionId  safexid  )

static

Definition at line 3080 of file nbtpage.c.

{
    Assert(vstate->npendingpages <= vstate->bufsize);
    Assert(vstate->bufsize <= vstate->maxbufsize);
 
#ifdef USE_ASSERT_CHECKING
 
    /*
     * Verify an assumption made by _bt_pendingfsm_finalize(): pages from the
     * array will always be in safexid order (since that is the order that we
     * save them in here)
     */
    if (vstate->npendingpages > 0)
    {
        FullTransactionId lastsafexid =
            vstate->pendingpages[vstate->npendingpages - 1].safexid;
 
        Assert(FullTransactionIdFollowsOrEquals(safexid, lastsafexid));
    }
#endif
 
    /*
     * If temp buffer reaches maxbufsize/work_mem capacity then we discard
     * information about this page.
     *
     * Note that this also covers the case where we opted to not use the
     * optimization in _bt_pendingfsm_init().
     */
    if (vstate->npendingpages == vstate->maxbufsize)
        return;
 
    /* Consider enlarging buffer */
    if (vstate->npendingpages == vstate->bufsize)
    {
        int         newbufsize = vstate->bufsize * 2;
 
        /* Respect work_mem */
        if (newbufsize > vstate->maxbufsize)
            newbufsize = vstate->maxbufsize;
 
        vstate->bufsize = newbufsize;
        vstate->pendingpages =
            repalloc(vstate->pendingpages,
                     sizeof(BTPendingFSM) * vstate->bufsize);
    }
 
    /* Save metadata for newly deleted page */
    vstate->pendingpages[vstate->npendingpages].target = target;
    vstate->pendingpages[vstate->npendingpages].safexid = safexid;
    vstate->npendingpages++;
}

References Assert, fb(), FullTransactionIdFollowsOrEquals, and repalloc().

Referenced by _bt_unlink_halfdead_page().

_bt_pendingfsm_finalize()

void _bt_pendingfsm_finalize	(	Relation	rel,
		BTVacState *	vstate
	)

Definition at line 3013 of file nbtpage.c.

{
    IndexBulkDeleteResult *stats = vstate->stats;
    Relation    heaprel = vstate->info->heaprel;
 
    Assert(stats->pages_newly_deleted >= vstate->npendingpages);
    Assert(heaprel != NULL);
 
    if (vstate->npendingpages == 0)
    {
        /* Just free memory when nothing to do */
        if (vstate->pendingpages)
            pfree(vstate->pendingpages);
 
        return;
    }
 
#ifdef DEBUG_BTREE_PENDING_FSM
 
    /*
     * Debugging aid: Sleep for 5 seconds to greatly increase the chances of
     * placing pending pages in the FSM.  Note that the optimization will
     * never be effective without some other backend concurrently consuming an
     * XID.
     */
    pg_usleep(5000000L);
#endif
 
    /*
     * Recompute VACUUM XID boundaries.
     *
     * We don't actually care about the oldest non-removable XID.  Computing
     * the oldest such XID has a useful side-effect that we rely on: it
     * forcibly updates the XID horizon state for this backend.  This step is
     * essential; GlobalVisCheckRemovableFullXid() will not reliably recognize
     * that it is now safe to recycle newly deleted pages without this step.
     */
    GetOldestNonRemovableTransactionId(heaprel);
 
    for (unsigned int i = 0; i < vstate->npendingpages; i++)
    {
        BlockNumber target = vstate->pendingpages[i].target;
        FullTransactionId safexid = vstate->pendingpages[i].safexid;
 
        /*
         * Do the equivalent of checking BTPageIsRecyclable(), but without
         * accessing the page again a second time.
         *
         * Give up on finding the first non-recyclable page -- all later pages
         * must be non-recyclable too, since _bt_pendingfsm_add() adds pages
         * to the array in safexid order.
         */
        if (!GlobalVisCheckRemovableFullXid(heaprel, safexid))
            break;
 
        RecordFreeIndexPage(rel, target);
        stats->pages_free++;
    }
 
    pfree(vstate->pendingpages);
}

References Assert, fb(), GetOldestNonRemovableTransactionId(), GlobalVisCheckRemovableFullXid(), i, IndexBulkDeleteResult::pages_free, IndexBulkDeleteResult::pages_newly_deleted, pfree(), pg_usleep(), and RecordFreeIndexPage().

Referenced by btvacuumscan().

_bt_pendingfsm_init()

void _bt_pendingfsm_init	(	Relation	rel,
		BTVacState *	vstate,
		bool	cleanuponly
	)

Definition at line 2971 of file nbtpage.c.

{
    Size        maxbufsize;
 
    /*
     * Don't bother with optimization in cleanup-only case -- we don't expect
     * any newly deleted pages.  Besides, cleanup-only calls to btvacuumscan()
     * can only take place because this optimization didn't work out during
     * the last VACUUM.
     */
    if (cleanuponly)
        return;
 
    /*
     * Cap maximum size of array so that we always respect work_mem.  Avoid
     * int overflow here.
     */
    vstate->bufsize = 256;
    maxbufsize = (work_mem * (Size) 1024) / sizeof(BTPendingFSM);
    maxbufsize = Min(maxbufsize, MaxAllocSize / sizeof(BTPendingFSM));
    /* BTVacState.maxbufsize has type int */
    maxbufsize = Min(maxbufsize, INT_MAX);
    /* Stay sane with small work_mem */
    maxbufsize = Max(maxbufsize, vstate->bufsize);
    vstate->maxbufsize = (int) maxbufsize;
 
    /* Allocate buffer, indicate that there are currently 0 pending pages */
    vstate->pendingpages = palloc_array(BTPendingFSM, vstate->bufsize);
    vstate->npendingpages = 0;
}

References fb(), Max, MaxAllocSize, Min, palloc_array, and work_mem.

Referenced by btvacuumscan().

_bt_relandgetbuf()

Buffer _bt_relandgetbuf	(	Relation	rel,
		Buffer	obuf,
		BlockNumber	blkno,
		int	access
	)

Definition at line 988 of file nbtpage.c.

{
    Buffer      buf;
 
    Assert(BlockNumberIsValid(blkno));
    if (BufferIsValid(obuf))
    {
        if (BufferGetBlockNumber(obuf) == blkno)
        {
            /* trade in old lock mode for new lock */
            _bt_unlockbuf(rel, obuf);
            buf = obuf;
        }
        else
        {
            /* release lock and pin at once, that's a bit more efficient */
            _bt_relbuf(rel, obuf);
            buf = ReadBuffer(rel, blkno);
        }
    }
    else
        buf = ReadBuffer(rel, blkno);
 
    _bt_lockbuf(rel, buf, access);
    _bt_checkpage(rel, buf);
 
    return buf;
}

References _bt_checkpage(), _bt_lockbuf(), _bt_relbuf(), _bt_unlockbuf(), Assert, BlockNumberIsValid(), buf, BufferGetBlockNumber(), BufferIsValid(), fb(), and ReadBuffer().

Referenced by _bt_check_unique(), _bt_get_endpoint(), _bt_getroot(), _bt_gettrueroot(), _bt_lock_and_validate_left(), _bt_moveright(), _bt_search(), and _bt_stepright().

_bt_relbuf()

void _bt_relbuf	(	Relation	rel,
		Buffer	buf
	)

Definition at line 1024 of file nbtpage.c.

{
    /*
     * Buffer is pinned and locked, which means that it is expected to be
     * defined and addressable.  Check that proactively.
     */
    VALGRIND_CHECK_MEM_IS_DEFINED(BufferGetPage(buf), BLCKSZ);
    if (!RelationUsesLocalBuffers(rel))
        VALGRIND_MAKE_MEM_NOACCESS(BufferGetPage(buf), BLCKSZ);
 
    UnlockReleaseBuffer(buf);
}

References buf, BufferGetPage(), fb(), RelationUsesLocalBuffers, UnlockReleaseBuffer(), VALGRIND_CHECK_MEM_IS_DEFINED, and VALGRIND_MAKE_MEM_NOACCESS.

Referenced by _bt_allocbuf(), _bt_check_unique(), _bt_doinsert(), _bt_drop_lock_and_maybe_pin(), _bt_finish_split(), _bt_getroot(), _bt_getrootheight(), _bt_getstackbuf(), _bt_gettrueroot(), _bt_insert_parent(), _bt_insertonpg(), _bt_killitems(), _bt_leftsib_splitflag(), _bt_lock_and_validate_left(), _bt_lock_subtree_parent(), _bt_mark_page_halfdead(), _bt_metaversion(), _bt_moveright(), _bt_newlevel(), _bt_pagedel(), _bt_readnextpage(), _bt_relandgetbuf(), _bt_rightsib_halfdeadflag(), _bt_search_insert(), _bt_set_cleanup_info(), _bt_split(), _bt_stepright(), _bt_unlink_halfdead_page(), _bt_vacuum_needs_cleanup(), bt_rootdescend(), btvacuumpage(), and pgstat_btree_page().

_bt_rightsib_halfdeadflag()

static bool _bt_rightsib_halfdeadflag ( Relation  rel, BlockNumber  leafrightsib  )

static

Definition at line 1762 of file nbtpage.c.

{
    Buffer      buf;
    Page        page;
    BTPageOpaque opaque;
    bool        result;
 
    Assert(leafrightsib != P_NONE);
 
    buf = _bt_getbuf(rel, leafrightsib, BT_READ);
    page = BufferGetPage(buf);
    opaque = BTPageGetOpaque(page);
 
    Assert(P_ISLEAF(opaque) && !P_ISDELETED(opaque));
    result = P_ISHALFDEAD(opaque);
    _bt_relbuf(rel, buf);
 
    return result;
}

References _bt_getbuf(), _bt_relbuf(), Assert, BT_READ, BTPageGetOpaque, buf, BufferGetPage(), P_ISDELETED, P_ISHALFDEAD, P_ISLEAF, P_NONE, and result.

Referenced by _bt_mark_page_halfdead().

_bt_set_cleanup_info()

void _bt_set_cleanup_info	(	Relation	rel,
		BlockNumber	num_delpages
	)

Definition at line 233 of file nbtpage.c.

{
    Buffer      metabuf;
    Page        metapg;
    BTMetaPageData *metad;
    XLogRecPtr  recptr;
 
    /*
     * On-disk compatibility note: The btm_last_cleanup_num_delpages metapage
     * field started out as a TransactionId field called btm_oldest_btpo_xact.
     * Both "versions" are just uint32 fields.  It was convenient to repurpose
     * the field when we began to use 64-bit XIDs in deleted pages.
     *
     * It's possible that a pg_upgrade'd database will contain an XID value in
     * what is now recognized as the metapage's btm_last_cleanup_num_delpages
     * field.  _bt_vacuum_needs_cleanup() may even believe that this value
     * indicates that there are lots of pages that it needs to recycle, when
     * in reality there are only one or two.  The worst that can happen is
     * that there will be a call to btvacuumscan a little earlier, which will
     * set btm_last_cleanup_num_delpages to a sane value when we're called.
     *
     * Note also that the metapage's btm_last_cleanup_num_heap_tuples field is
     * no longer used as of PostgreSQL 14.  We set it to -1.0 on rewrite, just
     * to be consistent.
     */
    metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
    metapg = BufferGetPage(metabuf);
    metad = BTPageGetMeta(metapg);
 
    /* Don't miss chance to upgrade index/metapage when BTREE_MIN_VERSION */
    if (metad->btm_version >= BTREE_NOVAC_VERSION &&
        metad->btm_last_cleanup_num_delpages == num_delpages)
    {
        /* Usually means index continues to have num_delpages of 0 */
        _bt_relbuf(rel, metabuf);
        return;
    }
 
    /* trade in our read lock for a write lock */
    _bt_unlockbuf(rel, metabuf);
    _bt_lockbuf(rel, metabuf, BT_WRITE);
 
    START_CRIT_SECTION();
 
    /* upgrade meta-page if needed */
    if (metad->btm_version < BTREE_NOVAC_VERSION)
        _bt_upgrademetapage(metapg);
 
    /* update cleanup-related information */
    metad->btm_last_cleanup_num_delpages = num_delpages;
    metad->btm_last_cleanup_num_heap_tuples = -1.0;
    MarkBufferDirty(metabuf);
 
    /* write wal record if needed */
    if (RelationNeedsWAL(rel))
    {
        xl_btree_metadata md;
 
        XLogBeginInsert();
        XLogRegisterBuffer(0, metabuf, REGBUF_WILL_INIT | REGBUF_STANDARD);
 
        Assert(metad->btm_version >= BTREE_NOVAC_VERSION);
        md.version = metad->btm_version;
        md.root = metad->btm_root;
        md.level = metad->btm_level;
        md.fastroot = metad->btm_fastroot;
        md.fastlevel = metad->btm_fastlevel;
        md.last_cleanup_num_delpages = num_delpages;
        md.allequalimage = metad->btm_allequalimage;
 
        XLogRegisterBufData(0, &md, sizeof(xl_btree_metadata));
 
        recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_META_CLEANUP);
    }
    else
        recptr = XLogGetFakeLSN(rel);
 
    PageSetLSN(metapg, recptr);
 
    END_CRIT_SECTION();
 
    _bt_relbuf(rel, metabuf);
}

References _bt_getbuf(), _bt_lockbuf(), _bt_relbuf(), _bt_unlockbuf(), _bt_upgrademetapage(), xl_btree_metadata::allequalimage, Assert, BT_READ, BT_WRITE, BTPageGetMeta, BTREE_METAPAGE, BTREE_NOVAC_VERSION, BufferGetPage(), END_CRIT_SECTION, xl_btree_metadata::fastlevel, xl_btree_metadata::fastroot, fb(), xl_btree_metadata::last_cleanup_num_delpages, xl_btree_metadata::level, MarkBufferDirty(), PageSetLSN(), REGBUF_STANDARD, REGBUF_WILL_INIT, RelationNeedsWAL, xl_btree_metadata::root, START_CRIT_SECTION, xl_btree_metadata::version, XLOG_BTREE_META_CLEANUP, XLogBeginInsert(), XLogGetFakeLSN(), XLogInsert(), XLogRegisterBufData(), and XLogRegisterBuffer().

Referenced by btvacuumcleanup().

_bt_unlink_halfdead_page()

static bool _bt_unlink_halfdead_page ( Relation  rel, Buffer  leafbuf, BlockNumber  scanblkno, bool *  rightsib_empty, BTVacState *  vstate  )

static

Definition at line 2329 of file nbtpage.c.

{
    BlockNumber leafblkno = BufferGetBlockNumber(leafbuf);
    IndexBulkDeleteResult *stats = vstate->stats;
    BlockNumber leafleftsib;
    BlockNumber leafrightsib;
    BlockNumber target;
    BlockNumber leftsib;
    BlockNumber rightsib;
    Buffer      lbuf = InvalidBuffer;
    Buffer      buf;
    Buffer      rbuf;
    Buffer      metabuf = InvalidBuffer;
    Page        metapg = NULL;
    BTMetaPageData *metad = NULL;
    ItemId      itemid;
    Page        page;
    BTPageOpaque opaque;
    FullTransactionId safexid;
    bool        rightsib_is_rightmost;
    uint32      targetlevel;
    IndexTuple  leafhikey;
    BlockNumber leaftopparent;
    XLogRecPtr  recptr;
 
    page = BufferGetPage(leafbuf);
    opaque = BTPageGetOpaque(page);
 
    Assert(P_ISLEAF(opaque) && !P_ISDELETED(opaque) && P_ISHALFDEAD(opaque));
 
    /*
     * Remember some information about the leaf page.
     */
    itemid = PageGetItemId(page, P_HIKEY);
    leafhikey = (IndexTuple) PageGetItem(page, itemid);
    target = BTreeTupleGetTopParent(leafhikey);
    leafleftsib = opaque->btpo_prev;
    leafrightsib = opaque->btpo_next;
 
    _bt_unlockbuf(rel, leafbuf);
 
    INJECTION_POINT("nbtree-leave-page-half-dead", NULL);
 
    /*
     * Check here, as calling loops will have locks held, preventing
     * interrupts from being processed.
     */
    CHECK_FOR_INTERRUPTS();
 
    /* Unlink the current top parent of the subtree */
    if (!BlockNumberIsValid(target))
    {
        /* Target is leaf page (or leaf page is top parent, if you prefer) */
        target = leafblkno;
 
        buf = leafbuf;
        leftsib = leafleftsib;
        targetlevel = 0;
    }
    else
    {
        /* Target is the internal page taken from leaf's top parent link */
        Assert(target != leafblkno);
 
        /* Fetch the block number of the target's left sibling */
        buf = _bt_getbuf(rel, target, BT_READ);
        page = BufferGetPage(buf);
        opaque = BTPageGetOpaque(page);
        leftsib = opaque->btpo_prev;
        targetlevel = opaque->btpo_level;
        Assert(targetlevel > 0);
 
        /*
         * To avoid deadlocks, we'd better drop the target page lock before
         * going further.
         */
        _bt_unlockbuf(rel, buf);
    }
 
    /*
     * We have to lock the pages we need to modify in the standard order:
     * moving right, then up.  Else we will deadlock against other writers.
     *
     * So, first lock the leaf page, if it's not the target.  Then find and
     * write-lock the current left sibling of the target page.  The sibling
     * that was current a moment ago could have split, so we may have to move
     * right.
     */
    if (target != leafblkno)
        _bt_lockbuf(rel, leafbuf, BT_WRITE);
    if (leftsib != P_NONE)
    {
        lbuf = _bt_getbuf(rel, leftsib, BT_WRITE);
        page = BufferGetPage(lbuf);
        opaque = BTPageGetOpaque(page);
        while (P_ISDELETED(opaque) || opaque->btpo_next != target)
        {
            bool        leftsibvalid = true;
 
            /*
             * Before we follow the link from the page that was the left
             * sibling mere moments ago, validate its right link.  This
             * reduces the opportunities for loop to fail to ever make any
             * progress in the presence of index corruption.
             *
             * Note: we rely on the assumption that there can only be one
             * vacuum process running at a time (against the same index).
             */
            if (P_RIGHTMOST(opaque) || P_ISDELETED(opaque) ||
                leftsib == opaque->btpo_next)
                leftsibvalid = false;
 
            leftsib = opaque->btpo_next;
            _bt_relbuf(rel, lbuf);
 
            if (!leftsibvalid)
            {
                /*
                 * This is known to fail in the field; sibling link corruption
                 * is relatively common.  Press on with vacuuming rather than
                 * just throwing an ERROR.
                 */
                ereport(LOG,
                        (errcode(ERRCODE_INDEX_CORRUPTED),
                         errmsg_internal("valid left sibling for deletion target could not be located: "
                                         "left sibling %u of target %u with leafblkno %u and scanblkno %u on level %u of index \"%s\"",
                                         leftsib, target, leafblkno, scanblkno,
                                         targetlevel, RelationGetRelationName(rel))));
 
                /* Must release all pins and locks on failure exit */
                ReleaseBuffer(buf);
                if (target != leafblkno)
                    _bt_relbuf(rel, leafbuf);
 
                return false;
            }
 
            CHECK_FOR_INTERRUPTS();
 
            /* step right one page */
            lbuf = _bt_getbuf(rel, leftsib, BT_WRITE);
            page = BufferGetPage(lbuf);
            opaque = BTPageGetOpaque(page);
        }
    }
    else
        lbuf = InvalidBuffer;
 
    /* Next write-lock the target page itself */
    _bt_lockbuf(rel, buf, BT_WRITE);
    page = BufferGetPage(buf);
    opaque = BTPageGetOpaque(page);
 
    /*
     * Check page is still empty etc, else abandon deletion.  This is just for
     * paranoia's sake; a half-dead page cannot resurrect because there can be
     * only one vacuum process running at a time.
     */
    if (P_RIGHTMOST(opaque) || P_ISROOT(opaque) || P_ISDELETED(opaque))
        elog(ERROR, "target page changed status unexpectedly in block %u of index \"%s\"",
             target, RelationGetRelationName(rel));
 
    if (opaque->btpo_prev != leftsib)
        ereport(ERROR,
                (errcode(ERRCODE_INDEX_CORRUPTED),
                 errmsg_internal("target page left link unexpectedly changed from %u to %u in block %u of index \"%s\"",
                                 leftsib, opaque->btpo_prev, target,
                                 RelationGetRelationName(rel))));
 
    if (target == leafblkno)
    {
        if (P_FIRSTDATAKEY(opaque) <= PageGetMaxOffsetNumber(page) ||
            !P_ISLEAF(opaque) || !P_ISHALFDEAD(opaque))
            elog(ERROR, "target leaf page changed status unexpectedly in block %u of index \"%s\"",
                 target, RelationGetRelationName(rel));
 
        /* Leaf page is also target page: don't set leaftopparent */
        leaftopparent = InvalidBlockNumber;
    }
    else
    {
        IndexTuple  finaldataitem;
 
        if (P_FIRSTDATAKEY(opaque) != PageGetMaxOffsetNumber(page) ||
            P_ISLEAF(opaque))
            elog(ERROR, "target internal page on level %u changed status unexpectedly in block %u of index \"%s\"",
                 targetlevel, target, RelationGetRelationName(rel));
 
        /* Target is internal: set leaftopparent for next call here...  */
        itemid = PageGetItemId(page, P_FIRSTDATAKEY(opaque));
        finaldataitem = (IndexTuple) PageGetItem(page, itemid);
        leaftopparent = BTreeTupleGetDownLink(finaldataitem);
        /* ...except when it would be a redundant pointer-to-self */
        if (leaftopparent == leafblkno)
            leaftopparent = InvalidBlockNumber;
    }
 
    /* No leaftopparent for level 0 (leaf page) or level 1 target */
    Assert(!BlockNumberIsValid(leaftopparent) || targetlevel > 1);
 
    /*
     * And next write-lock the (current) right sibling.
     */
    rightsib = opaque->btpo_next;
    rbuf = _bt_getbuf(rel, rightsib, BT_WRITE);
    page = BufferGetPage(rbuf);
    opaque = BTPageGetOpaque(page);
 
    /*
     * Validate target's right sibling page.  Its left link must point back to
     * the target page.
     */
    if (opaque->btpo_prev != target)
    {
        /*
         * This is known to fail in the field; sibling link corruption is
         * relatively common.  Press on with vacuuming rather than just
         * throwing an ERROR (same approach used for left-sibling's-right-link
         * validation check a moment ago).
         */
        ereport(LOG,
                (errcode(ERRCODE_INDEX_CORRUPTED),
                 errmsg_internal("right sibling's left-link doesn't match: "
                                 "right sibling %u of target %u with leafblkno %u "
                                 "and scanblkno %u spuriously links to non-target %u "
                                 "on level %u of index \"%s\"",
                                 rightsib, target, leafblkno,
                                 scanblkno, opaque->btpo_prev,
                                 targetlevel, RelationGetRelationName(rel))));
 
        /* Must release all pins and locks on failure exit */
        if (BufferIsValid(lbuf))
            _bt_relbuf(rel, lbuf);
        _bt_relbuf(rel, rbuf);
        _bt_relbuf(rel, buf);
        if (target != leafblkno)
            _bt_relbuf(rel, leafbuf);
 
        return false;
    }
 
    rightsib_is_rightmost = P_RIGHTMOST(opaque);
    *rightsib_empty = (P_FIRSTDATAKEY(opaque) > PageGetMaxOffsetNumber(page));
 
    /*
     * If we are deleting the next-to-last page on the target's level, then
     * the rightsib is a candidate to become the new fast root. (In theory, it
     * might be possible to push the fast root even further down, but the odds
     * of doing so are slim, and the locking considerations daunting.)
     *
     * We can safely acquire a lock on the metapage here --- see comments for
     * _bt_newlevel().
     */
    if (leftsib == P_NONE && rightsib_is_rightmost)
    {
        page = BufferGetPage(rbuf);
        opaque = BTPageGetOpaque(page);
        if (P_RIGHTMOST(opaque))
        {
            /* rightsib will be the only one left on the level */
            metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_WRITE);
            metapg = BufferGetPage(metabuf);
            metad = BTPageGetMeta(metapg);
 
            /*
             * The expected case here is btm_fastlevel == targetlevel+1; if
             * the fastlevel is <= targetlevel, something is wrong, and we
             * choose to overwrite it to fix it.
             */
            if (metad->btm_fastlevel > targetlevel + 1)
            {
                /* no update wanted */
                _bt_relbuf(rel, metabuf);
                metabuf = InvalidBuffer;
            }
        }
    }
 
    /*
     * Here we begin doing the deletion.
     */
 
    /* No ereport(ERROR) until changes are logged */
    START_CRIT_SECTION();
 
    /*
     * Update siblings' side-links.  Note the target page's side-links will
     * continue to point to the siblings.  Asserts here are just rechecking
     * things we already verified above.
     */
    if (BufferIsValid(lbuf))
    {
        page = BufferGetPage(lbuf);
        opaque = BTPageGetOpaque(page);
        Assert(opaque->btpo_next == target);
        opaque->btpo_next = rightsib;
    }
    page = BufferGetPage(rbuf);
    opaque = BTPageGetOpaque(page);
    Assert(opaque->btpo_prev == target);
    opaque->btpo_prev = leftsib;
 
    /*
     * If we deleted a parent of the targeted leaf page, instead of the leaf
     * itself, update the leaf to point to the next remaining child in the
     * subtree.
     *
     * Note: We rely on the fact that a buffer pin on the leaf page has been
     * held since leafhikey was initialized.  This is safe, though only
     * because the page was already half-dead at that point.  The leaf page
     * cannot have been modified by any other backend during the period when
     * no lock was held.
     */
    if (target != leafblkno)
        BTreeTupleSetTopParent(leafhikey, leaftopparent);
 
    /*
     * Mark the page itself deleted.  It can be recycled when all current
     * transactions are gone.  Storing GetTopTransactionId() would work, but
     * we're in VACUUM and would not otherwise have an XID.  Having already
     * updated links to the target, ReadNextFullTransactionId() suffices as an
     * upper bound.  Any scan having retained a now-stale link is advertising
     * in its PGPROC an xmin less than or equal to the value we read here.  It
     * will continue to do so, holding back the xmin horizon, for the duration
     * of that scan.
     */
    page = BufferGetPage(buf);
    opaque = BTPageGetOpaque(page);
    Assert(P_ISHALFDEAD(opaque) || !P_ISLEAF(opaque));
 
    /*
     * Store upper bound XID that's used to determine when deleted page is no
     * longer needed as a tombstone
     */
    safexid = ReadNextFullTransactionId();
    BTPageSetDeleted(page, safexid);
    opaque->btpo_cycleid = 0;
 
    /* And update the metapage, if needed */
    if (BufferIsValid(metabuf))
    {
        /* upgrade metapage if needed */
        if (metad->btm_version < BTREE_NOVAC_VERSION)
            _bt_upgrademetapage(metapg);
        metad->btm_fastroot = rightsib;
        metad->btm_fastlevel = targetlevel;
        MarkBufferDirty(metabuf);
    }
 
    /* Must mark buffers dirty before XLogInsert */
    MarkBufferDirty(rbuf);
    MarkBufferDirty(buf);
    if (BufferIsValid(lbuf))
        MarkBufferDirty(lbuf);
    if (target != leafblkno)
        MarkBufferDirty(leafbuf);
 
    /* XLOG stuff */
    if (RelationNeedsWAL(rel))
    {
        xl_btree_unlink_page xlrec;
        xl_btree_metadata xlmeta;
        uint8       xlinfo;
 
        XLogBeginInsert();
 
        XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
        if (BufferIsValid(lbuf))
            XLogRegisterBuffer(1, lbuf, REGBUF_STANDARD);
        XLogRegisterBuffer(2, rbuf, REGBUF_STANDARD);
        if (target != leafblkno)
            XLogRegisterBuffer(3, leafbuf, REGBUF_WILL_INIT);
 
        /* information stored on the target/to-be-unlinked block */
        xlrec.leftsib = leftsib;
        xlrec.rightsib = rightsib;
        xlrec.level = targetlevel;
        xlrec.safexid = safexid;
 
        /* information needed to recreate the leaf block (if not the target) */
        xlrec.leafleftsib = leafleftsib;
        xlrec.leafrightsib = leafrightsib;
        xlrec.leaftopparent = leaftopparent;
 
        XLogRegisterData(&xlrec, SizeOfBtreeUnlinkPage);
 
        if (BufferIsValid(metabuf))
        {
            XLogRegisterBuffer(4, metabuf, REGBUF_WILL_INIT | REGBUF_STANDARD);
 
            Assert(metad->btm_version >= BTREE_NOVAC_VERSION);
            xlmeta.version = metad->btm_version;
            xlmeta.root = metad->btm_root;
            xlmeta.level = metad->btm_level;
            xlmeta.fastroot = metad->btm_fastroot;
            xlmeta.fastlevel = metad->btm_fastlevel;
            xlmeta.last_cleanup_num_delpages = metad->btm_last_cleanup_num_delpages;
            xlmeta.allequalimage = metad->btm_allequalimage;
 
            XLogRegisterBufData(4, &xlmeta, sizeof(xl_btree_metadata));
            xlinfo = XLOG_BTREE_UNLINK_PAGE_META;
        }
        else
            xlinfo = XLOG_BTREE_UNLINK_PAGE;
 
        recptr = XLogInsert(RM_BTREE_ID, xlinfo);
    }
    else
        recptr = XLogGetFakeLSN(rel);
 
    if (BufferIsValid(metabuf))
        PageSetLSN(metapg, recptr);
    page = BufferGetPage(rbuf);
    PageSetLSN(page, recptr);
    page = BufferGetPage(buf);
    PageSetLSN(page, recptr);
    if (BufferIsValid(lbuf))
    {
        page = BufferGetPage(lbuf);
        PageSetLSN(page, recptr);
    }
    if (target != leafblkno)
    {
        page = BufferGetPage(leafbuf);
        PageSetLSN(page, recptr);
    }
 
    END_CRIT_SECTION();
 
    /* release metapage */
    if (BufferIsValid(metabuf))
        _bt_relbuf(rel, metabuf);
 
    /* release siblings */
    if (BufferIsValid(lbuf))
        _bt_relbuf(rel, lbuf);
    _bt_relbuf(rel, rbuf);
 
    /* If the target is not leafbuf, we're done with it now -- release it */
    if (target != leafblkno)
        _bt_relbuf(rel, buf);
 
    /*
     * Maintain pages_newly_deleted, which is simply the number of pages
     * deleted by the ongoing VACUUM operation.
     *
     * Maintain pages_deleted in a way that takes into account how
     * btvacuumpage() will count deleted pages that have yet to become
     * scanblkno -- only count page when it's not going to get that treatment
     * later on.
     */
    stats->pages_newly_deleted++;
    if (target <= scanblkno)
        stats->pages_deleted++;
 
    /*
     * Remember information about the target page (now a newly deleted page)
     * in dedicated vstate space for later.  The page will be considered as a
     * candidate to place in the FSM at the end of the current btvacuumscan()
     * call.
     */
    _bt_pendingfsm_add(vstate, target, safexid);
 
    /* Success - hold on to lock on leafbuf (might also have been target) */
    return true;
}

References _bt_getbuf(), _bt_lockbuf(), _bt_pendingfsm_add(), _bt_relbuf(), _bt_unlockbuf(), _bt_upgrademetapage(), Assert, BlockNumberIsValid(), BT_READ, BT_WRITE, BTPageGetMeta, BTPageGetOpaque, BTPageSetDeleted(), BTPageOpaqueData::btpo_cycleid, BTPageOpaqueData::btpo_level, BTPageOpaqueData::btpo_next, BTPageOpaqueData::btpo_prev, BTREE_METAPAGE, BTREE_NOVAC_VERSION, BTreeTupleGetDownLink(), BTreeTupleGetTopParent(), BTreeTupleSetTopParent(), buf, BufferGetBlockNumber(), BufferGetPage(), BufferIsValid(), CHECK_FOR_INTERRUPTS, elog, END_CRIT_SECTION, ereport, errcode(), errmsg_internal(), ERROR, fb(), INJECTION_POINT, InvalidBlockNumber, InvalidBuffer, LOG, MarkBufferDirty(), P_FIRSTDATAKEY, P_HIKEY, P_ISDELETED, P_ISHALFDEAD, P_ISLEAF, P_ISROOT, P_NONE, P_RIGHTMOST, PageGetItem(), PageGetItemId(), PageGetMaxOffsetNumber(), IndexBulkDeleteResult::pages_deleted, IndexBulkDeleteResult::pages_newly_deleted, PageSetLSN(), ReadNextFullTransactionId(), REGBUF_STANDARD, REGBUF_WILL_INIT, RelationGetRelationName, RelationNeedsWAL, ReleaseBuffer(), SizeOfBtreeUnlinkPage, START_CRIT_SECTION, XLOG_BTREE_UNLINK_PAGE, XLOG_BTREE_UNLINK_PAGE_META, XLogBeginInsert(), XLogGetFakeLSN(), XLogInsert(), XLogRegisterBufData(), XLogRegisterBuffer(), and XLogRegisterData().

Referenced by _bt_pagedel().

_bt_unlockbuf()

void _bt_unlockbuf	(	Relation	rel,
		Buffer	buf
	)

Definition at line 1078 of file nbtpage.c.

{
    /*
     * Buffer is pinned and locked, which means that it is expected to be
     * defined and addressable.  Check that proactively.
     */
    VALGRIND_CHECK_MEM_IS_DEFINED(BufferGetPage(buf), BLCKSZ);
 
    /* LockBuffer() asserts that pin is held by this backend */
    LockBuffer(buf, BUFFER_LOCK_UNLOCK);
 
    if (!RelationUsesLocalBuffers(rel))
        VALGRIND_MAKE_MEM_NOACCESS(BufferGetPage(buf), BLCKSZ);
}

References buf, BUFFER_LOCK_UNLOCK, BufferGetPage(), fb(), LockBuffer(), RelationUsesLocalBuffers, VALGRIND_CHECK_MEM_IS_DEFINED, and VALGRIND_MAKE_MEM_NOACCESS.

Referenced by _bt_drop_lock_and_maybe_pin(), _bt_getroot(), _bt_killitems(), _bt_moveright(), _bt_pagedel(), _bt_readfirstpage(), _bt_relandgetbuf(), _bt_search(), _bt_set_cleanup_info(), and _bt_unlink_halfdead_page().

_bt_upgradelockbufcleanup()

void _bt_upgradelockbufcleanup	(	Relation	rel,
		Buffer	buf
	)

Definition at line 1117 of file nbtpage.c.

{
    /*
     * Buffer is pinned and locked, which means that it is expected to be
     * defined and addressable.  Check that proactively.
     */
    VALGRIND_CHECK_MEM_IS_DEFINED(BufferGetPage(buf), BLCKSZ);
 
    /* LockBuffer() asserts that pin is held by this backend */
    LockBuffer(buf, BUFFER_LOCK_UNLOCK);
    LockBufferForCleanup(buf);
}

References buf, BUFFER_LOCK_UNLOCK, BufferGetPage(), fb(), LockBuffer(), LockBufferForCleanup(), and VALGRIND_CHECK_MEM_IS_DEFINED.

Referenced by btvacuumpage().

_bt_upgrademetapage()

void _bt_upgrademetapage

(

Page

page

)

Definition at line 108 of file nbtpage.c.

{
    BTMetaPageData *metad;
    BTPageOpaque metaopaque PG_USED_FOR_ASSERTS_ONLY;
 
    metad = BTPageGetMeta(page);
    metaopaque = BTPageGetOpaque(page);
 
    /* It must be really a meta page of upgradable version */
    Assert(metaopaque->btpo_flags & BTP_META);
    Assert(metad->btm_version < BTREE_NOVAC_VERSION);
    Assert(metad->btm_version >= BTREE_MIN_VERSION);
 
    /* Set version number and fill extra fields added into version 3 */
    metad->btm_version = BTREE_NOVAC_VERSION;
    metad->btm_last_cleanup_num_delpages = 0;
    metad->btm_last_cleanup_num_heap_tuples = -1.0;
    /* Only a REINDEX can set this field */
    Assert(!metad->btm_allequalimage);
    metad->btm_allequalimage = false;
 
    /* Adjust pd_lower (see _bt_initmetapage() for details) */
    ((PageHeader) page)->pd_lower =
        ((char *) metad + sizeof(BTMetaPageData)) - (char *) page;
}

References Assert, BTP_META, BTPageGetMeta, BTPageGetOpaque, BTREE_MIN_VERSION, BTREE_NOVAC_VERSION, fb(), and PG_USED_FOR_ASSERTS_ONLY.

Referenced by _bt_getroot(), _bt_insertonpg(), _bt_newlevel(), _bt_set_cleanup_info(), and _bt_unlink_halfdead_page().

_bt_vacuum_needs_cleanup()

bool _bt_vacuum_needs_cleanup

(

Relation

rel

)

Definition at line 180 of file nbtpage.c.

{
    Buffer      metabuf;
    Page        metapg;
    BTMetaPageData *metad;
    uint32      btm_version;
    BlockNumber prev_num_delpages;
 
    /*
     * Copy details from metapage to local variables quickly.
     *
     * Note that we deliberately avoid using cached version of metapage here.
     */
    metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_READ);
    metapg = BufferGetPage(metabuf);
    metad = BTPageGetMeta(metapg);
    btm_version = metad->btm_version;
 
    if (btm_version < BTREE_NOVAC_VERSION)
    {
        /*
         * Metapage needs to be dynamically upgraded to store fields that are
         * only present when btm_version >= BTREE_NOVAC_VERSION
         */
        _bt_relbuf(rel, metabuf);
        return true;
    }
 
    prev_num_delpages = metad->btm_last_cleanup_num_delpages;
    _bt_relbuf(rel, metabuf);
 
    /*
     * Trigger cleanup in rare cases where prev_num_delpages exceeds 5% of the
     * total size of the index.  We can reasonably expect (though are not
     * guaranteed) to be able to recycle this many pages if we decide to do a
     * btvacuumscan call during the ongoing btvacuumcleanup.  For further
     * details see the nbtree/README section on placing deleted pages in the
     * FSM.
     */
    if (prev_num_delpages > 0 &&
        prev_num_delpages > RelationGetNumberOfBlocks(rel) / 20)
        return true;
 
    return false;
}

References _bt_getbuf(), _bt_relbuf(), BT_READ, BTPageGetMeta, BTREE_METAPAGE, BTREE_NOVAC_VERSION, BufferGetPage(), fb(), and RelationGetNumberOfBlocks.

Referenced by btvacuumcleanup().