Datum
gistendscan(PG_FUNCTION_ARGS)
{
IndexScanDesc scan = (IndexScanDesc) PG_GETARG_POINTER(0);
GISTScanOpaque so;
so = (GISTScanOpaque) scan->opaque;
if (so != NULL)
{
gistfreestack(so->stack);
gistfreestack(so->markstk);
if (so->giststate != NULL)
freeGISTstate(so->giststate);
/* drop pins on buffers -- we aren't holding any locks */
if (BufferIsValid(so->curbuf))
ReleaseBuffer(so->curbuf);
if (BufferIsValid(so->markbuf))
ReleaseBuffer(so->markbuf);
MemoryContextDelete(so->tempCxt);
pfree(scan->opaque);
}
PG_RETURN_VOID();
}
/*
* gistinsert -- wrapper for GiST tuple insertion.
*
* This is the public interface routine for tuple insertion in GiSTs.
* It doesn't do any work; just locks the relation and passes the buck.
*/
Datum
gistinsert(PG_FUNCTION_ARGS)
{
Relation r = (Relation) PG_GETARG_POINTER(0);
Datum *values = (Datum *) PG_GETARG_POINTER(1);
bool *isnull = (bool *) PG_GETARG_POINTER(2);
ItemPointer ht_ctid = (ItemPointer) PG_GETARG_POINTER(3);
#ifdef NOT_USED
Relation heapRel = (Relation) PG_GETARG_POINTER(4);
bool checkUnique = PG_GETARG_BOOL(5);
#endif
IndexTuple itup;
GISTSTATE giststate;
MemoryContext oldCtx;
MemoryContext insertCtx;
insertCtx = createTempGistContext();
oldCtx = MemoryContextSwitchTo(insertCtx);
initGISTstate(&giststate, r);
itup = gistFormTuple(&giststate, r,
values, isnull, true /* size is currently bogus */ );
itup->t_tid = *ht_ctid;
gistdoinsert(r, itup, 0, &giststate);
/* cleanup */
freeGISTstate(&giststate);
MemoryContextSwitchTo(oldCtx);
MemoryContextDelete(insertCtx);
PG_RETURN_BOOL(true);
}
/*
* gistinsert -- wrapper for GiST tuple insertion.
*
* This is the public interface routine for tuple insertion in GiSTs.
* It doesn't do any work; just locks the relation and passes the buck.
*/
bool
gistinsert(Relation r, Datum *values, bool *isnull,
ItemPointer ht_ctid, Relation heapRel,
IndexUniqueCheck checkUnique)
{
IndexTuple itup;
GISTSTATE *giststate;
MemoryContext oldCxt;
giststate = initGISTstate(r);
/*
* We use the giststate's scan context as temp context too. This means
* that any memory leaked by the support functions is not reclaimed until
* end of insert. In most cases, we aren't going to call the support
* functions very many times before finishing the insert, so this seems
* cheaper than resetting a temp context for each function call.
*/
oldCxt = MemoryContextSwitchTo(giststate->tempCxt);
itup = gistFormTuple(giststate, r,
values, isnull, true /* size is currently bogus */ );
itup->t_tid = *ht_ctid;
gistdoinsert(r, itup, 0, giststate);
/* cleanup */
MemoryContextSwitchTo(oldCxt);
freeGISTstate(giststate);
return false;
}
void
gistendscan(IndexScanDesc scan)
{
GISTScanOpaque so = (GISTScanOpaque) scan->opaque;
/*
* freeGISTstate is enough to clean up everything made by gistbeginscan,
* as well as the queueCxt if there is a separate context for it.
*/
freeGISTstate(so->giststate);
}
Datum
gistendscan(PG_FUNCTION_ARGS)
{
IndexScanDesc scan = (IndexScanDesc) PG_GETARG_POINTER(0);
GISTScanOpaque so = (GISTScanOpaque) scan->opaque;
/*
* freeGISTstate is enough to clean up everything made by gistbeginscan,
* as well as the queueCxt if there is a separate context for it.
*/
freeGISTstate(so->giststate);
PG_RETURN_VOID();
}
/*
* gistinsert -- wrapper for GiST tuple insertion.
*
* This is the public interface routine for tuple insertion in GiSTs.
* It doesn't do any work; just locks the relation and passes the buck.
*/
Datum
gistinsert(PG_FUNCTION_ARGS)
{
Relation r = (Relation) PG_GETARG_POINTER(0);
Datum *values = (Datum *) PG_GETARG_POINTER(1);
bool *isnull = (bool *) PG_GETARG_POINTER(2);
ItemPointer ht_ctid = (ItemPointer) PG_GETARG_POINTER(3);
#ifdef NOT_USED
Relation heapRel = (Relation) PG_GETARG_POINTER(4);
IndexUniqueCheck checkUnique = (IndexUniqueCheck) PG_GETARG_INT32(5);
#endif
IndexTuple itup;
GISTSTATE *giststate;
MemoryContext oldCxt;
giststate = initGISTstate(r);
/*
* We use the giststate's scan context as temp context too. This means
* that any memory leaked by the support functions is not reclaimed until
* end of insert. In most cases, we aren't going to call the support
* functions very many times before finishing the insert, so this seems
* cheaper than resetting a temp context for each function call.
*/
oldCxt = MemoryContextSwitchTo(giststate->tempCxt);
itup = gistFormTuple(giststate, r,
values, isnull, true /* size is currently bogus */ );
itup->t_tid = *ht_ctid;
gistdoinsert(r, itup, 0, giststate);
/* cleanup */
MemoryContextSwitchTo(oldCxt);
freeGISTstate(giststate);
PG_RETURN_BOOL(false);
}
/*
* Main entry point to GiST index build. Initially calls insert over and over,
* but switches to more efficient buffering build algorithm after a certain
* number of tuples (unless buffering mode is disabled).
*/
Datum
gistbuild(PG_FUNCTION_ARGS)
{
Relation heap = (Relation) PG_GETARG_POINTER(0);
Relation index = (Relation) PG_GETARG_POINTER(1);
IndexInfo *indexInfo = (IndexInfo *) PG_GETARG_POINTER(2);
IndexBuildResult *result;
double reltuples;
GISTBuildState buildstate;
Buffer buffer;
Page page;
MemoryContext oldcxt = CurrentMemoryContext;
int fillfactor;
buildstate.indexrel = index;
if (index->rd_options)
{
/* Get buffering mode from the options string */
GiSTOptions *options = (GiSTOptions *) index->rd_options;
char *bufferingMode = (char *) options + options->bufferingModeOffset;
if (strcmp(bufferingMode, "on") == 0)
buildstate.bufferingMode = GIST_BUFFERING_STATS;
else if (strcmp(bufferingMode, "off") == 0)
buildstate.bufferingMode = GIST_BUFFERING_DISABLED;
else
buildstate.bufferingMode = GIST_BUFFERING_AUTO;
fillfactor = options->fillfactor;
}
else
{
/*
* By default, switch to buffering mode when the index grows too large
* to fit in cache.
*/
buildstate.bufferingMode = GIST_BUFFERING_AUTO;
fillfactor = GIST_DEFAULT_FILLFACTOR;
}
/* Calculate target amount of free space to leave on pages */
buildstate.freespace = BLCKSZ * (100 - fillfactor) / 100;
/*
* We expect to be called exactly once for any index relation. If that's
* not the case, big trouble's what we have.
*/
if (RelationGetNumberOfBlocks(index) != 0)
elog(ERROR, "index \"%s\" already contains data",
RelationGetRelationName(index));
/* no locking is needed */
buildstate.giststate = initGISTstate(index);
/*
* Create a temporary memory context that is reset once for each tuple
* processed. (Note: we don't bother to make this a child of the
* giststate's scanCxt, so we have to delete it separately at the end.)
*/
buildstate.giststate->tempCxt = createTempGistContext();
/* initialize the root page */
buffer = gistNewBuffer(index);
Assert(BufferGetBlockNumber(buffer) == GIST_ROOT_BLKNO);
page = BufferGetPage(buffer);
START_CRIT_SECTION();
GISTInitBuffer(buffer, F_LEAF);
MarkBufferDirty(buffer);
if (RelationNeedsWAL(index))
{
XLogRecPtr recptr;
XLogRecData rdata;
rdata.data = (char *) &(index->rd_node);
rdata.len = sizeof(RelFileNode);
rdata.buffer = InvalidBuffer;
rdata.next = NULL;
recptr = XLogInsert(RM_GIST_ID, XLOG_GIST_CREATE_INDEX, &rdata);
PageSetLSN(page, recptr);
PageSetTLI(page, ThisTimeLineID);
}
else
PageSetLSN(page, gistGetFakeLSN(heap));
UnlockReleaseBuffer(buffer);
END_CRIT_SECTION();
/* build the index */
buildstate.indtuples = 0;
buildstate.indtuplesSize = 0;
/*
* Do the heap scan.
*/
reltuples = IndexBuildHeapScan(heap, index, indexInfo, true,
gistBuildCallback, (void *) &buildstate);
/*
* If buffering was used, flush out all the tuples that are still in the
* buffers.
*/
if (buildstate.bufferingMode == GIST_BUFFERING_ACTIVE)
{
elog(DEBUG1, "all tuples processed, emptying buffers");
gistEmptyAllBuffers(&buildstate);
gistFreeBuildBuffers(buildstate.gfbb);
}
/* okay, all heap tuples are indexed */
MemoryContextSwitchTo(oldcxt);
MemoryContextDelete(buildstate.giststate->tempCxt);
freeGISTstate(buildstate.giststate);
/*
* Return statistics
*/
result = (IndexBuildResult *) palloc(sizeof(IndexBuildResult));
result->heap_tuples = reltuples;
result->index_tuples = (double) buildstate.indtuples;
PG_RETURN_POINTER(result);
}
Datum
gistvacuumcleanup(PG_FUNCTION_ARGS)
{
IndexVacuumInfo *info = (IndexVacuumInfo *) PG_GETARG_POINTER(0);
GistBulkDeleteResult *stats = (GistBulkDeleteResult *) PG_GETARG_POINTER(1);
Relation rel = info->index;
BlockNumber npages,
blkno;
BlockNumber totFreePages,
nFreePages,
*freePages,
maxFreePages;
BlockNumber lastBlock = GIST_ROOT_BLKNO,
lastFilledBlock = GIST_ROOT_BLKNO;
bool needLock;
/* Set up all-zero stats if gistbulkdelete wasn't called */
if (stats == NULL)
{
stats = (GistBulkDeleteResult *) palloc0(sizeof(GistBulkDeleteResult));
/* use heap's tuple count */
Assert(info->num_heap_tuples >= 0);
stats->std.num_index_tuples = info->num_heap_tuples;
/*
* XXX the above is wrong if index is partial. Would it be OK to just
* return NULL, or is there work we must do below?
*/
}
/* gistVacuumUpdate may cause hard work */
if (info->vacuum_full)
{
GistVacuum gv;
ArrayTuple res;
/* note: vacuum.c already acquired AccessExclusiveLock on index */
gv.index = rel;
initGISTstate(&(gv.giststate), rel);
gv.opCtx = createTempGistContext();
gv.result = stats;
gv.strategy = info->strategy;
/* walk through the entire index for update tuples */
res = gistVacuumUpdate(&gv, GIST_ROOT_BLKNO, false);
/* cleanup */
if (res.itup)
{
int i;
for (i = 0; i < res.ituplen; i++)
pfree(res.itup[i]);
pfree(res.itup);
}
freeGISTstate(&(gv.giststate));
MemoryContextDelete(gv.opCtx);
}
else if (stats->needFullVacuum)
ereport(NOTICE,
(errmsg("index \"%s\" needs VACUUM FULL or REINDEX to finish crash recovery",
RelationGetRelationName(rel))));
/*
* If vacuum full, we already have exclusive lock on the index. Otherwise,
* need lock unless it's local to this backend.
*/
if (info->vacuum_full)
needLock = false;
else
needLock = !RELATION_IS_LOCAL(rel);
/* try to find deleted pages */
if (needLock)
LockRelationForExtension(rel, ExclusiveLock);
npages = RelationGetNumberOfBlocks(rel);
if (needLock)
UnlockRelationForExtension(rel, ExclusiveLock);
maxFreePages = npages;
if (maxFreePages > MaxFSMPages)
maxFreePages = MaxFSMPages;
totFreePages = nFreePages = 0;
freePages = (BlockNumber *) palloc(sizeof(BlockNumber) * maxFreePages);
for (blkno = GIST_ROOT_BLKNO + 1; blkno < npages; blkno++)
{
Buffer buffer;
Page page;
vacuum_delay_point();
buffer = ReadBufferWithStrategy(rel, blkno, info->strategy);
LockBuffer(buffer, GIST_SHARE);
page = (Page) BufferGetPage(buffer);
if (PageIsNew(page) || GistPageIsDeleted(page))
{
if (nFreePages < maxFreePages)
freePages[nFreePages++] = blkno;
totFreePages++;
}
else
lastFilledBlock = blkno;
UnlockReleaseBuffer(buffer);
}
lastBlock = npages - 1;
if (info->vacuum_full && nFreePages > 0)
{ /* try to truncate index */
int i;
for (i = 0; i < nFreePages; i++)
if (freePages[i] >= lastFilledBlock)
{
totFreePages = nFreePages = i;
break;
}
if (lastBlock > lastFilledBlock)
RelationTruncate(rel, lastFilledBlock + 1);
stats->std.pages_removed = lastBlock - lastFilledBlock;
}
RecordIndexFreeSpace(&rel->rd_node, totFreePages, nFreePages, freePages);
pfree(freePages);
/* return statistics */
stats->std.pages_free = totFreePages;
if (needLock)
LockRelationForExtension(rel, ExclusiveLock);
stats->std.num_pages = RelationGetNumberOfBlocks(rel);
if (needLock)
UnlockRelationForExtension(rel, ExclusiveLock);
PG_RETURN_POINTER(stats);
}
/*
* Routine to build an index. Basically calls insert over and over.
*
* XXX: it would be nice to implement some sort of bulk-loading
* algorithm, but it is not clear how to do that.
*/
Datum
gistbuild(PG_FUNCTION_ARGS)
{
MIRROREDLOCK_BUFMGR_DECLARE;
Relation heap = (Relation) PG_GETARG_POINTER(0);
Relation index = (Relation) PG_GETARG_POINTER(1);
IndexInfo *indexInfo = (IndexInfo *) PG_GETARG_POINTER(2);
IndexBuildResult *result;
double reltuples;
GISTBuildState buildstate;
Buffer buffer;
Page page;
/*
* We expect to be called exactly once for any index relation. If that's
* not the case, big trouble's what we have.
*/
if (RelationGetNumberOfBlocks(index) != 0)
elog(ERROR, "index \"%s\" already contains data",
RelationGetRelationName(index));
/* no locking is needed */
initGISTstate(&buildstate.giststate, index);
// -------- MirroredLock ----------
MIRROREDLOCK_BUFMGR_LOCK;
/* initialize the root page */
buffer = gistNewBuffer(index);
Assert(BufferGetBlockNumber(buffer) == GIST_ROOT_BLKNO);
page = BufferGetPage(buffer);
START_CRIT_SECTION();
GISTInitBuffer(buffer, F_LEAF);
MarkBufferDirty(buffer);
if (!index->rd_istemp)
{
XLogRecPtr recptr;
XLogRecData *rdata;
rdata = formCreateRData(index);
recptr = XLogInsert(RM_GIST_ID, XLOG_GIST_CREATE_INDEX, rdata);
PageSetLSN(page, recptr);
PageSetTLI(page, ThisTimeLineID);
}
else
PageSetLSN(page, XLogRecPtrForTemp);
UnlockReleaseBuffer(buffer);
MIRROREDLOCK_BUFMGR_UNLOCK;
// -------- MirroredLock ----------
END_CRIT_SECTION();
/* build the index */
buildstate.numindexattrs = indexInfo->ii_NumIndexAttrs;
buildstate.indtuples = 0;
/*
* create a temporary memory context that is reset once for each tuple
* inserted into the index
*/
buildstate.tmpCtx = createTempGistContext();
/* do the heap scan */
reltuples = IndexBuildScan(heap, index, indexInfo,
gistbuildCallback, (void *) &buildstate);
/* okay, all heap tuples are indexed */
MemoryContextDelete(buildstate.tmpCtx);
freeGISTstate(&buildstate.giststate);
/*
* Return statistics
*/
result = (IndexBuildResult *) palloc(sizeof(IndexBuildResult));
result->heap_tuples = reltuples;
result->index_tuples = buildstate.indtuples;
PG_RETURN_POINTER(result);
}
/*
* gistinsert -- wrapper for GiST tuple insertion.
*
* This is the public interface routine for tuple insertion in GiSTs.
* It doesn't do any work; just locks the relation and passes the buck.
*/
Datum
gistinsert(PG_FUNCTION_ARGS)
{
Relation r = (Relation) PG_GETARG_POINTER(0);
Datum *datum = (Datum *) PG_GETARG_POINTER(1);
char *nulls = (char *) PG_GETARG_POINTER(2);
ItemPointer ht_ctid = (ItemPointer) PG_GETARG_POINTER(3);
#ifdef NOT_USED
Relation heapRel = (Relation) PG_GETARG_POINTER(4);
bool checkUnique = PG_GETARG_BOOL(5);
#endif
InsertIndexResult res;
IndexTuple itup;
GISTSTATE giststate;
GISTENTRY tmpentry;
int i;
bool compvec[INDEX_MAX_KEYS];
/*
* Since GIST is not marked "amconcurrent" in pg_am, caller should
* have acquired exclusive lock on index relation. We need no locking
* here.
*/
/* GiST cannot index tuples with leading NULLs */
if (nulls[0] == 'n')
{
res = NULL;
PG_RETURN_POINTER(res);
}
initGISTstate(&giststate, r);
/* immediately compress keys to normalize */
for (i = 0; i < r->rd_att->natts; i++)
{
if (nulls[i] == 'n')
{
datum[i] = (Datum) 0;
compvec[i] = FALSE;
}
else
{
gistcentryinit(&giststate, i, &tmpentry, datum[i],
NULL, NULL, (OffsetNumber) 0,
-1 /* size is currently bogus */ , TRUE, FALSE);
if (datum[i] != tmpentry.key && !(isAttByVal(&giststate, i)))
compvec[i] = TRUE;
else
compvec[i] = FALSE;
datum[i] = tmpentry.key;
}
}
itup = index_formtuple(giststate.tupdesc, datum, nulls);
itup->t_tid = *ht_ctid;
res = (InsertIndexResult) palloc(sizeof(InsertIndexResultData));
gistdoinsert(r, itup, &res, &giststate);
for (i = 0; i < r->rd_att->natts; i++)
if (compvec[i] == TRUE)
pfree(DatumGetPointer(datum[i]));
pfree(itup);
freeGISTstate(&giststate);
PG_RETURN_POINTER(res);
}
/*
* routine to build an index. Basically calls insert over and over
*/
Datum
gistbuild(PG_FUNCTION_ARGS)
{
Relation heap = (Relation) PG_GETARG_POINTER(0);
Relation index = (Relation) PG_GETARG_POINTER(1);
IndexInfo *indexInfo = (IndexInfo *) PG_GETARG_POINTER(2);
double reltuples;
GISTBuildState buildstate;
Buffer buffer;
/* no locking is needed */
initGISTstate(&buildstate.giststate, index);
/*
* We expect to be called exactly once for any index relation. If
* that's not the case, big trouble's what we have.
*/
if (RelationGetNumberOfBlocks(index) != 0)
elog(ERROR, "index \"%s\" already contains data",
RelationGetRelationName(index));
/* initialize the root page */
buffer = ReadBuffer(index, P_NEW);
GISTInitBuffer(buffer, F_LEAF);
WriteBuffer(buffer);
/* build the index */
buildstate.numindexattrs = indexInfo->ii_NumIndexAttrs;
buildstate.indtuples = 0;
/* do the heap scan */
reltuples = IndexBuildHeapScan(heap, index, indexInfo,
gistbuildCallback, (void *) &buildstate);
/* okay, all heap tuples are indexed */
/*
* Since we just counted the tuples in the heap, we update its stats
* in pg_class to guarantee that the planner takes advantage of the
* index we just created. But, only update statistics during normal
* index definitions, not for indices on system catalogs created
* during bootstrap processing. We must close the relations before
* updating statistics to guarantee that the relcache entries are
* flushed when we increment the command counter in UpdateStats(). But
* we do not release any locks on the relations; those will be held
* until end of transaction.
*/
if (IsNormalProcessingMode())
{
Oid hrelid = RelationGetRelid(heap);
Oid irelid = RelationGetRelid(index);
heap_close(heap, NoLock);
index_close(index);
UpdateStats(hrelid, reltuples);
UpdateStats(irelid, buildstate.indtuples);
}
freeGISTstate(&buildstate.giststate);
#ifdef GISTDEBUG
gist_dumptree(index, 0, GISTP_ROOT, 0);
#endif
PG_RETURN_VOID();
}
