示例#1
0
文件: hashinsert.c 项目: 50wu/gpdb
/*
 *	_hash_doinsert() -- Handle insertion of a single index tuple.
 *
 *		This routine is called by the public interface routines, hashbuild
 *		and hashinsert.  By here, itup is completely filled in.
 */
void
_hash_doinsert(Relation rel, IndexTuple itup)
{
	MIRROREDLOCK_BUFMGR_DECLARE;

	Buffer		buf;
	Buffer		metabuf;
	HashMetaPage metap;
	BlockNumber blkno;
	Page		page;
	HashPageOpaque pageopaque;
	Size		itemsz;
	bool		do_expand;
	uint32		hashkey;
	Bucket		bucket;
	Datum		datum;
	bool		isnull;

	/*
	 * Compute the hash key for the item.  We do this first so as not to need
	 * to hold any locks while running the hash function.
	 */
	if (rel->rd_rel->relnatts != 1)
		elog(ERROR, "hash indexes support only one index key");
	datum = index_getattr(itup, 1, RelationGetDescr(rel), &isnull);
	Assert(!isnull);
	hashkey = _hash_datum2hashkey(rel, datum);

	/* compute item size too */
	itemsz = IndexTupleDSize(*itup);
	itemsz = MAXALIGN(itemsz);	/* be safe, PageAddItem will do this but we
								 * need to be consistent */

	/*
	 * Acquire shared split lock so we can compute the target bucket safely
	 * (see README).
	 */

	 // -------- MirroredLock ----------
	 MIRROREDLOCK_BUFMGR_LOCK;

	_hash_getlock(rel, 0, HASH_SHARE);

	/* Read the metapage */
	metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
	metap = (HashMetaPage) BufferGetPage(metabuf);

	/*
	 * Check whether the item can fit on a hash page at all. (Eventually, we
	 * ought to try to apply TOAST methods if not.)  Note that at this point,
	 * itemsz doesn't include the ItemId.
	 */
	if (itemsz > HashMaxItemSize((Page) metap))
		ereport(ERROR,
				(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
				 errmsg("index row size %lu exceeds hash maximum %lu",
						(unsigned long) itemsz,
						(unsigned long) HashMaxItemSize((Page) metap)),
			errhint("Values larger than a buffer page cannot be indexed.")));

	/*
	 * Compute the target bucket number, and convert to block number.
	 */
	bucket = _hash_hashkey2bucket(hashkey,
								  metap->hashm_maxbucket,
								  metap->hashm_highmask,
								  metap->hashm_lowmask);

	blkno = BUCKET_TO_BLKNO(metap, bucket);

	/* release lock on metapage, but keep pin since we'll need it again */
	_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);

	/*
	 * Acquire share lock on target bucket; then we can release split lock.
	 */
	_hash_getlock(rel, blkno, HASH_SHARE);

	_hash_droplock(rel, 0, HASH_SHARE);

	/* Fetch the primary bucket page for the bucket */
	buf = _hash_getbuf(rel, blkno, HASH_WRITE, LH_BUCKET_PAGE);
	page = BufferGetPage(buf);
	pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
	Assert(pageopaque->hasho_bucket == bucket);

	/* Do the insertion */
	while (PageGetFreeSpace(page) < itemsz)
	{
		/*
		 * no space on this page; check for an overflow page
		 */
		BlockNumber nextblkno = pageopaque->hasho_nextblkno;

		if (BlockNumberIsValid(nextblkno))
		{
			/*
			 * ovfl page exists; go get it.  if it doesn't have room, we'll
			 * find out next pass through the loop test above.
			 */
			_hash_relbuf(rel, buf);
			buf = _hash_getbuf(rel, nextblkno, HASH_WRITE, LH_OVERFLOW_PAGE);
			page = BufferGetPage(buf);
		}
		else
		{
			/*
			 * we're at the end of the bucket chain and we haven't found a
			 * page with enough room.  allocate a new overflow page.
			 */

			/* release our write lock without modifying buffer */
			_hash_chgbufaccess(rel, buf, HASH_READ, HASH_NOLOCK);

			/* chain to a new overflow page */
			buf = _hash_addovflpage(rel, metabuf, buf);
			page = BufferGetPage(buf);

			/* should fit now, given test above */
			Assert(PageGetFreeSpace(page) >= itemsz);
		}
		pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
		Assert(pageopaque->hasho_flag == LH_OVERFLOW_PAGE);
		Assert(pageopaque->hasho_bucket == bucket);
	}

	/* found page with enough space, so add the item here */
	(void) _hash_pgaddtup(rel, buf, itemsz, itup);

	/* write and release the modified page */
	_hash_wrtbuf(rel, buf);

	/* We can drop the bucket lock now */
	_hash_droplock(rel, blkno, HASH_SHARE);

	/*
	 * Write-lock the metapage so we can increment the tuple count. After
	 * incrementing it, check to see if it's time for a split.
	 */
	_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);

	metap->hashm_ntuples += 1;

	/* Make sure this stays in sync with _hash_expandtable() */
	do_expand = metap->hashm_ntuples >
		(double) metap->hashm_ffactor * (metap->hashm_maxbucket + 1);

	/* Write out the metapage and drop lock, but keep pin */
	_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);

	MIRROREDLOCK_BUFMGR_UNLOCK;
	// -------- MirroredLock ----------

	/* Attempt to split if a split is needed */
	if (do_expand)
		_hash_expandtable(rel, metabuf);

	/* Finally drop our pin on the metapage */
	_hash_dropbuf(rel, metabuf);
}
示例#2
0
/*
 *	_hash_first() -- Find the first item in a scan.
 *
 *		Find the first item in the index that
 *		satisfies the qualification associated with the scan descriptor. On
 *		success, the page containing the current index tuple is read locked
 *		and pinned, and the scan's opaque data entry is updated to
 *		include the buffer.
 */
bool
_hash_first(IndexScanDesc scan, ScanDirection dir)
{
	Relation	rel = scan->indexRelation;
	HashScanOpaque so = (HashScanOpaque) scan->opaque;
	ScanKey		cur;
	uint32		hashkey;
	Bucket		bucket;
	BlockNumber blkno;
	Buffer		buf;
	Buffer		metabuf;
	Page		page;
	HashPageOpaque opaque;
	HashMetaPage metap;
	IndexTuple	itup;
	ItemPointer current;
	OffsetNumber offnum;

	pgstat_count_index_scan(rel);

	current = &(so->hashso_curpos);
	ItemPointerSetInvalid(current);

	/*
	 * We do not support hash scans with no index qualification, because we
	 * would have to read the whole index rather than just one bucket. That
	 * creates a whole raft of problems, since we haven't got a practical way
	 * to lock all the buckets against splits or compactions.
	 */
	if (scan->numberOfKeys < 1)
		ereport(ERROR,
				(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
				 errmsg("hash indexes do not support whole-index scans")));

	/* There may be more than one index qual, but we hash only the first */
	cur = &scan->keyData[0];

	/* We support only single-column hash indexes */
	Assert(cur->sk_attno == 1);
	/* And there's only one operator strategy, too */
	Assert(cur->sk_strategy == HTEqualStrategyNumber);

	/*
	 * If the constant in the index qual is NULL, assume it cannot match any
	 * items in the index.
	 */
	if (cur->sk_flags & SK_ISNULL)
		return false;

	/*
	 * Okay to compute the hash key.  We want to do this before acquiring any
	 * locks, in case a user-defined hash function happens to be slow.
	 *
	 * If scankey operator is not a cross-type comparison, we can use the
	 * cached hash function; otherwise gotta look it up in the catalogs.
	 *
	 * We support the convention that sk_subtype == InvalidOid means the
	 * opclass input type; this is a hack to simplify life for ScanKeyInit().
	 */
	if (cur->sk_subtype == rel->rd_opcintype[0] ||
		cur->sk_subtype == InvalidOid)
		hashkey = _hash_datum2hashkey(rel, cur->sk_argument);
	else
		hashkey = _hash_datum2hashkey_type(rel, cur->sk_argument,
										   cur->sk_subtype);

	so->hashso_sk_hash = hashkey;

	/*
	 * Acquire shared split lock so we can compute the target bucket safely
	 * (see README).
	 */
	_hash_getlock(rel, 0, HASH_SHARE);

	/* Read the metapage */
	metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
	metap = HashPageGetMeta(BufferGetPage(metabuf));

	/*
	 * Compute the target bucket number, and convert to block number.
	 */
	bucket = _hash_hashkey2bucket(hashkey,
								  metap->yetToSplitBkts,
								  metap->hashm_highmask,
								  metap->hashm_lowmask); /* CS3223 - change metap->hashm_maxbucket to metap->yetToSplitBkts */

	blkno = BUCKET_TO_BLKNO(metap, bucket);

	/* done with the metapage */
	_hash_relbuf(rel, metabuf);

	/*
	 * Acquire share lock on target bucket; then we can release split lock.
	 */
	_hash_getlock(rel, blkno, HASH_SHARE);

	_hash_droplock(rel, 0, HASH_SHARE);

	/* Update scan opaque state to show we have lock on the bucket */
	so->hashso_bucket = bucket;
	so->hashso_bucket_valid = true;
	so->hashso_bucket_blkno = blkno;

	/* Fetch the primary bucket page for the bucket */
	buf = _hash_getbuf(rel, blkno, HASH_READ, LH_BUCKET_PAGE);
	page = BufferGetPage(buf);
	opaque = (HashPageOpaque) PageGetSpecialPointer(page);
	Assert(opaque->hasho_bucket == bucket);

	/* If a backwards scan is requested, move to the end of the chain */
	if (ScanDirectionIsBackward(dir))
	{
		while (BlockNumberIsValid(opaque->hasho_nextblkno))
			_hash_readnext(rel, &buf, &page, &opaque);
	}

	/* Now find the first tuple satisfying the qualification */
	if (!_hash_step(scan, &buf, dir))
		return false;

	/* if we're here, _hash_step found a valid tuple */
	offnum = ItemPointerGetOffsetNumber(current);
	_hash_checkpage(rel, buf, LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
	page = BufferGetPage(buf);
	itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, offnum));
	so->hashso_heappos = itup->t_tid;

	return true;
}
示例#3
0
/* ------------------------------------------------
 * hash_bitmap_info()
 *
 * Get bitmap information for a particular overflow page
 *
 * Usage: SELECT * FROM hash_bitmap_info('con_hash_index'::regclass, 5);
 * ------------------------------------------------
 */
Datum
hash_bitmap_info(PG_FUNCTION_ARGS)
{
	Oid			indexRelid = PG_GETARG_OID(0);
	uint64		ovflblkno = PG_GETARG_INT64(1);
	HashMetaPage metap;
	Buffer		metabuf,
				mapbuf;
	BlockNumber bitmapblkno;
	Page		mappage;
	bool		bit = false;
	TupleDesc	tupleDesc;
	Relation	indexRel;
	uint32		ovflbitno;
	int32		bitmappage,
				bitmapbit;
	HeapTuple	tuple;
	int			i,
				j;
	Datum		values[3];
	bool		nulls[3];
	uint32	   *freep;

	if (!superuser())
		ereport(ERROR,
				(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
				 (errmsg("must be superuser to use raw page functions"))));

	indexRel = index_open(indexRelid, AccessShareLock);

	if (!IS_HASH(indexRel))
		elog(ERROR, "relation \"%s\" is not a hash index",
			 RelationGetRelationName(indexRel));

	if (RELATION_IS_OTHER_TEMP(indexRel))
		ereport(ERROR,
				(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
				 errmsg("cannot access temporary tables of other sessions")));

	if (ovflblkno >= RelationGetNumberOfBlocks(indexRel))
		ereport(ERROR,
				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
				 errmsg("block number " UINT64_FORMAT " is out of range for relation \"%s\"",
						ovflblkno, RelationGetRelationName(indexRel))));

	/* Read the metapage so we can determine which bitmap page to use */
	metabuf = _hash_getbuf(indexRel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
	metap = HashPageGetMeta(BufferGetPage(metabuf));

	/*
	 * Reject attempt to read the bit for a metapage or bitmap page; this is
	 * only meaningful for overflow pages.
	 */
	if (ovflblkno == 0)
		ereport(ERROR,
				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
				 errmsg("invalid overflow block number %u",
						(BlockNumber) ovflblkno)));
	for (i = 0; i < metap->hashm_nmaps; i++)
		if (metap->hashm_mapp[i] == ovflblkno)
			ereport(ERROR,
					(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
					 errmsg("invalid overflow block number %u",
							(BlockNumber) ovflblkno)));

	/*
	 * Identify overflow bit number.  This will error out for primary bucket
	 * pages, and we've already rejected the metapage and bitmap pages above.
	 */
	ovflbitno = _hash_ovflblkno_to_bitno(metap, (BlockNumber) ovflblkno);

	bitmappage = ovflbitno >> BMPG_SHIFT(metap);
	bitmapbit = ovflbitno & BMPG_MASK(metap);

	if (bitmappage >= metap->hashm_nmaps)
		ereport(ERROR,
				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
				 errmsg("invalid overflow block number %u",
						(BlockNumber) ovflblkno)));

	bitmapblkno = metap->hashm_mapp[bitmappage];

	_hash_relbuf(indexRel, metabuf);

	/* Check the status of bitmap bit for overflow page */
	mapbuf = _hash_getbuf(indexRel, bitmapblkno, HASH_READ, LH_BITMAP_PAGE);
	mappage = BufferGetPage(mapbuf);
	freep = HashPageGetBitmap(mappage);

	bit = ISSET(freep, bitmapbit) != 0;

	_hash_relbuf(indexRel, mapbuf);
	index_close(indexRel, AccessShareLock);

	/* Build a tuple descriptor for our result type */
	if (get_call_result_type(fcinfo, NULL, &tupleDesc) != TYPEFUNC_COMPOSITE)
		elog(ERROR, "return type must be a row type");
	tupleDesc = BlessTupleDesc(tupleDesc);

	MemSet(nulls, 0, sizeof(nulls));

	j = 0;
	values[j++] = Int64GetDatum((int64) bitmapblkno);
	values[j++] = Int32GetDatum(bitmapbit);
	values[j++] = BoolGetDatum(bit);

	tuple = heap_form_tuple(tupleDesc, values, nulls);

	PG_RETURN_DATUM(HeapTupleGetDatum(tuple));
}
示例#4
0
/*
 *	_hash_getovflpage()
 *
 *	Find an available overflow page and return it.  The returned buffer
 *	is pinned and write-locked, and has had _hash_pageinit() applied,
 *	but it is caller's responsibility to fill the special space.
 *
 * The caller must hold a pin, but no lock, on the metapage buffer.
 * That buffer is left in the same state at exit.
 */
static Buffer
_hash_getovflpage(Relation rel, Buffer metabuf)
{
	HashMetaPage metap;
	Buffer		mapbuf = 0;
	Buffer		newbuf;
	BlockNumber blkno;
	uint32		orig_firstfree;
	uint32		splitnum;
	uint32	   *freep = NULL;
	uint32		max_ovflpg;
	uint32		bit;
	uint32		first_page;
	uint32		last_bit;
	uint32		last_page;
	uint32		i,
				j;

	/* Get exclusive lock on the meta page */
	_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);

	_hash_checkpage(rel, metabuf, LH_META_PAGE);
	metap = HashPageGetMeta(BufferGetPage(metabuf));

	/* start search at hashm_firstfree */
	orig_firstfree = metap->hashm_firstfree;
	first_page = orig_firstfree >> BMPG_SHIFT(metap);
	bit = orig_firstfree & BMPG_MASK(metap);
	i = first_page;
	j = bit / BITS_PER_MAP;
	bit &= ~(BITS_PER_MAP - 1);

	/* outer loop iterates once per bitmap page */
	for (;;)
	{
		BlockNumber mapblkno;
		Page		mappage;
		uint32		last_inpage;

		/* want to end search with the last existing overflow page */
		splitnum = metap->hashm_ovflpoint;
		max_ovflpg = metap->hashm_spares[splitnum] - 1;
		last_page = max_ovflpg >> BMPG_SHIFT(metap);
		last_bit = max_ovflpg & BMPG_MASK(metap);

		if (i > last_page)
			break;

		Assert(i < metap->hashm_nmaps);
		mapblkno = metap->hashm_mapp[i];

		if (i == last_page)
			last_inpage = last_bit;
		else
			last_inpage = BMPGSZ_BIT(metap) - 1;

		/* Release exclusive lock on metapage while reading bitmap page */
		_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);

		mapbuf = _hash_getbuf(rel, mapblkno, HASH_WRITE, LH_BITMAP_PAGE);
		mappage = BufferGetPage(mapbuf);
		freep = HashPageGetBitmap(mappage);

		for (; bit <= last_inpage; j++, bit += BITS_PER_MAP)
		{
			if (freep[j] != ALL_SET)
				goto found;
		}

		/* No free space here, try to advance to next map page */
		_hash_relbuf(rel, mapbuf);
		i++;
		j = 0;					/* scan from start of next map page */
		bit = 0;

		/* Reacquire exclusive lock on the meta page */
		_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);
	}

	/*
	 * No free pages --- have to extend the relation to add an overflow page.
	 * First, check to see if we have to add a new bitmap page too.
	 */
	if (last_bit == (uint32) (BMPGSZ_BIT(metap) - 1))
	{
		/*
		 * We create the new bitmap page with all pages marked "in use".
		 * Actually two pages in the new bitmap's range will exist
		 * immediately: the bitmap page itself, and the following page which
		 * is the one we return to the caller.  Both of these are correctly
		 * marked "in use".  Subsequent pages do not exist yet, but it is
		 * convenient to pre-mark them as "in use" too.
		 */
		bit = metap->hashm_spares[splitnum];
		_hash_initbitmap(rel, metap, bitno_to_blkno(metap, bit), MAIN_FORKNUM);
		metap->hashm_spares[splitnum]++;
	}
	else
	{
		/*
		 * Nothing to do here; since the page will be past the last used page,
		 * we know its bitmap bit was preinitialized to "in use".
		 */
	}

	/* Calculate address of the new overflow page */
	bit = metap->hashm_spares[splitnum];
	blkno = bitno_to_blkno(metap, bit);

	/*
	 * Fetch the page with _hash_getnewbuf to ensure smgr's idea of the
	 * relation length stays in sync with ours.  XXX It's annoying to do this
	 * with metapage write lock held; would be better to use a lock that
	 * doesn't block incoming searches.
	 */
	newbuf = _hash_getnewbuf(rel, blkno, MAIN_FORKNUM);

	metap->hashm_spares[splitnum]++;

	/*
	 * Adjust hashm_firstfree to avoid redundant searches.  But don't risk
	 * changing it if someone moved it while we were searching bitmap pages.
	 */
	if (metap->hashm_firstfree == orig_firstfree)
		metap->hashm_firstfree = bit + 1;

	/* Write updated metapage and release lock, but not pin */
	_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);

	return newbuf;

found:
	/* convert bit to bit number within page */
	bit += _hash_firstfreebit(freep[j]);

	/* mark page "in use" in the bitmap */
	SETBIT(freep, bit);
	_hash_wrtbuf(rel, mapbuf);

	/* Reacquire exclusive lock on the meta page */
	_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);

	/* convert bit to absolute bit number */
	bit += (i << BMPG_SHIFT(metap));

	/* Calculate address of the recycled overflow page */
	blkno = bitno_to_blkno(metap, bit);

	/*
	 * Adjust hashm_firstfree to avoid redundant searches.  But don't risk
	 * changing it if someone moved it while we were searching bitmap pages.
	 */
	if (metap->hashm_firstfree == orig_firstfree)
	{
		metap->hashm_firstfree = bit + 1;

		/* Write updated metapage and release lock, but not pin */
		_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);
	}
	else
	{
		/* We didn't change the metapage, so no need to write */
		_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);
	}

	/* Fetch, init, and return the recycled page */
	return _hash_getinitbuf(rel, blkno);
}
示例#5
0
/*
 *	_hash_freeovflpage() -
 *
 *	Remove this overflow page from its bucket's chain, and mark the page as
 *	free.  On entry, ovflbuf is write-locked; it is released before exiting.
 *
 *	Since this function is invoked in VACUUM, we provide an access strategy
 *	parameter that controls fetches of the bucket pages.
 *
 *	Returns the block number of the page that followed the given page
 *	in the bucket, or InvalidBlockNumber if no following page.
 *
 *	NB: caller must not hold lock on metapage, nor on either page that's
 *	adjacent in the bucket chain.  The caller had better hold exclusive lock
 *	on the bucket, too.
 */
BlockNumber
_hash_freeovflpage(Relation rel, Buffer ovflbuf,
				   BufferAccessStrategy bstrategy)
{
	HashMetaPage metap;
	Buffer		metabuf;
	Buffer		mapbuf;
	BlockNumber ovflblkno;
	BlockNumber prevblkno;
	BlockNumber blkno;
	BlockNumber nextblkno;
	HashPageOpaque ovflopaque;
	Page		ovflpage;
	Page		mappage;
	uint32	   *freep;
	uint32		ovflbitno;
	int32		bitmappage,
				bitmapbit;
	Bucket bucket PG_USED_FOR_ASSERTS_ONLY;

	/* Get information from the doomed page */
	_hash_checkpage(rel, ovflbuf, LH_OVERFLOW_PAGE);
	ovflblkno = BufferGetBlockNumber(ovflbuf);
	ovflpage = BufferGetPage(ovflbuf);
	ovflopaque = (HashPageOpaque) PageGetSpecialPointer(ovflpage);
	nextblkno = ovflopaque->hasho_nextblkno;
	prevblkno = ovflopaque->hasho_prevblkno;
	bucket = ovflopaque->hasho_bucket;

	/*
	 * Zero the page for debugging's sake; then write and release it. (Note:
	 * if we failed to zero the page here, we'd have problems with the Assert
	 * in _hash_pageinit() when the page is reused.)
	 */
	MemSet(ovflpage, 0, BufferGetPageSize(ovflbuf));
	_hash_wrtbuf(rel, ovflbuf);

	/*
	 * Fix up the bucket chain.  this is a doubly-linked list, so we must fix
	 * up the bucket chain members behind and ahead of the overflow page being
	 * deleted.  No concurrency issues since we hold exclusive lock on the
	 * entire bucket.
	 */
	if (BlockNumberIsValid(prevblkno))
	{
		Buffer		prevbuf = _hash_getbuf_with_strategy(rel,
														 prevblkno,
														 HASH_WRITE,
										   LH_BUCKET_PAGE | LH_OVERFLOW_PAGE,
														 bstrategy);
		Page		prevpage = BufferGetPage(prevbuf);
		HashPageOpaque prevopaque = (HashPageOpaque) PageGetSpecialPointer(prevpage);

		Assert(prevopaque->hasho_bucket == bucket);
		prevopaque->hasho_nextblkno = nextblkno;
		_hash_wrtbuf(rel, prevbuf);
	}
	if (BlockNumberIsValid(nextblkno))
	{
		Buffer		nextbuf = _hash_getbuf_with_strategy(rel,
														 nextblkno,
														 HASH_WRITE,
														 LH_OVERFLOW_PAGE,
														 bstrategy);
		Page		nextpage = BufferGetPage(nextbuf);
		HashPageOpaque nextopaque = (HashPageOpaque) PageGetSpecialPointer(nextpage);

		Assert(nextopaque->hasho_bucket == bucket);
		nextopaque->hasho_prevblkno = prevblkno;
		_hash_wrtbuf(rel, nextbuf);
	}

	/* Note: bstrategy is intentionally not used for metapage and bitmap */

	/* Read the metapage so we can determine which bitmap page to use */
	metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
	metap = HashPageGetMeta(BufferGetPage(metabuf));

	/* Identify which bit to set */
	ovflbitno = blkno_to_bitno(metap, ovflblkno);

	bitmappage = ovflbitno >> BMPG_SHIFT(metap);
	bitmapbit = ovflbitno & BMPG_MASK(metap);

	if (bitmappage >= metap->hashm_nmaps)
		elog(ERROR, "invalid overflow bit number %u", ovflbitno);
	blkno = metap->hashm_mapp[bitmappage];

	/* Release metapage lock while we access the bitmap page */
	_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);

	/* Clear the bitmap bit to indicate that this overflow page is free */
	mapbuf = _hash_getbuf(rel, blkno, HASH_WRITE, LH_BITMAP_PAGE);
	mappage = BufferGetPage(mapbuf);
	freep = HashPageGetBitmap(mappage);
	Assert(ISSET(freep, bitmapbit));
	CLRBIT(freep, bitmapbit);
	_hash_wrtbuf(rel, mapbuf);

	/* Get write-lock on metapage to update firstfree */
	_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);

	/* if this is now the first free page, update hashm_firstfree */
	if (ovflbitno < metap->hashm_firstfree)
	{
		metap->hashm_firstfree = ovflbitno;
		_hash_wrtbuf(rel, metabuf);
	}
	else
	{
		/* no need to change metapage */
		_hash_relbuf(rel, metabuf);
	}

	return nextblkno;
}
示例#6
0
/*
 *	_hash_finish_split() -- Finish the previously interrupted split operation
 *
 * To complete the split operation, we form the hash table of TIDs in new
 * bucket which is then used by split operation to skip tuples that are
 * already moved before the split operation was previously interrupted.
 *
 * The caller must hold a pin, but no lock, on the metapage and old bucket's
 * primary page buffer.  The buffers are returned in the same state.  (The
 * metapage is only touched if it becomes necessary to add or remove overflow
 * pages.)
 */
void
_hash_finish_split(Relation rel, Buffer metabuf, Buffer obuf, Bucket obucket,
				   uint32 maxbucket, uint32 highmask, uint32 lowmask)
{
	HASHCTL		hash_ctl;
	HTAB	   *tidhtab;
	Buffer		bucket_nbuf = InvalidBuffer;
	Buffer		nbuf;
	Page		npage;
	BlockNumber nblkno;
	BlockNumber bucket_nblkno;
	HashPageOpaque npageopaque;
	Bucket		nbucket;
	bool		found;

	/* Initialize hash tables used to track TIDs */
	memset(&hash_ctl, 0, sizeof(hash_ctl));
	hash_ctl.keysize = sizeof(ItemPointerData);
	hash_ctl.entrysize = sizeof(ItemPointerData);
	hash_ctl.hcxt = CurrentMemoryContext;

	tidhtab =
		hash_create("bucket ctids",
					256,		/* arbitrary initial size */
					&hash_ctl,
					HASH_ELEM | HASH_BLOBS | HASH_CONTEXT);

	bucket_nblkno = nblkno = _hash_get_newblock_from_oldbucket(rel, obucket);

	/*
	 * Scan the new bucket and build hash table of TIDs
	 */
	for (;;)
	{
		OffsetNumber noffnum;
		OffsetNumber nmaxoffnum;

		nbuf = _hash_getbuf(rel, nblkno, HASH_READ,
							LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);

		/* remember the primary bucket buffer to acquire cleanup lock on it. */
		if (nblkno == bucket_nblkno)
			bucket_nbuf = nbuf;

		npage = BufferGetPage(nbuf);
		npageopaque = (HashPageOpaque) PageGetSpecialPointer(npage);

		/* Scan each tuple in new page */
		nmaxoffnum = PageGetMaxOffsetNumber(npage);
		for (noffnum = FirstOffsetNumber;
			 noffnum <= nmaxoffnum;
			 noffnum = OffsetNumberNext(noffnum))
		{
			IndexTuple	itup;

			/* Fetch the item's TID and insert it in hash table. */
			itup = (IndexTuple) PageGetItem(npage,
											PageGetItemId(npage, noffnum));

			(void) hash_search(tidhtab, &itup->t_tid, HASH_ENTER, &found);

			Assert(!found);
		}

		nblkno = npageopaque->hasho_nextblkno;

		/*
		 * release our write lock without modifying buffer and ensure to
		 * retain the pin on primary bucket.
		 */
		if (nbuf == bucket_nbuf)
			LockBuffer(nbuf, BUFFER_LOCK_UNLOCK);
		else
			_hash_relbuf(rel, nbuf);

		/* Exit loop if no more overflow pages in new bucket */
		if (!BlockNumberIsValid(nblkno))
			break;
	}

	/*
	 * Conditionally get the cleanup lock on old and new buckets to perform
	 * the split operation.  If we don't get the cleanup locks, silently give
	 * up and next insertion on old bucket will try again to complete the
	 * split.
	 */
	if (!ConditionalLockBufferForCleanup(obuf))
	{
		hash_destroy(tidhtab);
		return;
	}
	if (!ConditionalLockBufferForCleanup(bucket_nbuf))
	{
		LockBuffer(obuf, BUFFER_LOCK_UNLOCK);
		hash_destroy(tidhtab);
		return;
	}

	npage = BufferGetPage(bucket_nbuf);
	npageopaque = (HashPageOpaque) PageGetSpecialPointer(npage);
	nbucket = npageopaque->hasho_bucket;

	_hash_splitbucket(rel, metabuf, obucket,
					  nbucket, obuf, bucket_nbuf, tidhtab,
					  maxbucket, highmask, lowmask);

	_hash_dropbuf(rel, bucket_nbuf);
	hash_destroy(tidhtab);
}
示例#7
0
/*
 * _hash_splitbucket -- split 'obucket' into 'obucket' and 'nbucket'
 *
 * This routine is used to partition the tuples between old and new bucket and
 * is used to finish the incomplete split operations.  To finish the previously
 * interrupted split operation, the caller needs to fill htab.  If htab is set,
 * then we skip the movement of tuples that exists in htab, otherwise NULL
 * value of htab indicates movement of all the tuples that belong to the new
 * bucket.
 *
 * We are splitting a bucket that consists of a base bucket page and zero
 * or more overflow (bucket chain) pages.  We must relocate tuples that
 * belong in the new bucket.
 *
 * The caller must hold cleanup locks on both buckets to ensure that
 * no one else is trying to access them (see README).
 *
 * The caller must hold a pin, but no lock, on the metapage buffer.
 * The buffer is returned in the same state.  (The metapage is only
 * touched if it becomes necessary to add or remove overflow pages.)
 *
 * Split needs to retain pin on primary bucket pages of both old and new
 * buckets till end of operation.  This is to prevent vacuum from starting
 * while a split is in progress.
 *
 * In addition, the caller must have created the new bucket's base page,
 * which is passed in buffer nbuf, pinned and write-locked.  The lock will be
 * released here and pin must be released by the caller.  (The API is set up
 * this way because we must do _hash_getnewbuf() before releasing the metapage
 * write lock.  So instead of passing the new bucket's start block number, we
 * pass an actual buffer.)
 */
static void
_hash_splitbucket(Relation rel,
				  Buffer metabuf,
				  Bucket obucket,
				  Bucket nbucket,
				  Buffer obuf,
				  Buffer nbuf,
				  HTAB *htab,
				  uint32 maxbucket,
				  uint32 highmask,
				  uint32 lowmask)
{
	Buffer		bucket_obuf;
	Buffer		bucket_nbuf;
	Page		opage;
	Page		npage;
	HashPageOpaque oopaque;
	HashPageOpaque nopaque;
	OffsetNumber itup_offsets[MaxIndexTuplesPerPage];
	IndexTuple	itups[MaxIndexTuplesPerPage];
	Size		all_tups_size = 0;
	int			i;
	uint16		nitups = 0;

	bucket_obuf = obuf;
	opage = BufferGetPage(obuf);
	oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);

	bucket_nbuf = nbuf;
	npage = BufferGetPage(nbuf);
	nopaque = (HashPageOpaque) PageGetSpecialPointer(npage);

	/*
	 * Partition the tuples in the old bucket between the old bucket and the
	 * new bucket, advancing along the old bucket's overflow bucket chain and
	 * adding overflow pages to the new bucket as needed.  Outer loop iterates
	 * once per page in old bucket.
	 */
	for (;;)
	{
		BlockNumber oblkno;
		OffsetNumber ooffnum;
		OffsetNumber omaxoffnum;

		/* Scan each tuple in old page */
		omaxoffnum = PageGetMaxOffsetNumber(opage);
		for (ooffnum = FirstOffsetNumber;
			 ooffnum <= omaxoffnum;
			 ooffnum = OffsetNumberNext(ooffnum))
		{
			IndexTuple	itup;
			Size		itemsz;
			Bucket		bucket;
			bool		found = false;

			/* skip dead tuples */
			if (ItemIdIsDead(PageGetItemId(opage, ooffnum)))
				continue;

			/*
			 * Before inserting a tuple, probe the hash table containing TIDs
			 * of tuples belonging to new bucket, if we find a match, then
			 * skip that tuple, else fetch the item's hash key (conveniently
			 * stored in the item) and determine which bucket it now belongs
			 * in.
			 */
			itup = (IndexTuple) PageGetItem(opage,
											PageGetItemId(opage, ooffnum));

			if (htab)
				(void) hash_search(htab, &itup->t_tid, HASH_FIND, &found);

			if (found)
				continue;

			bucket = _hash_hashkey2bucket(_hash_get_indextuple_hashkey(itup),
										  maxbucket, highmask, lowmask);

			if (bucket == nbucket)
			{
				IndexTuple	new_itup;

				/*
				 * make a copy of index tuple as we have to scribble on it.
				 */
				new_itup = CopyIndexTuple(itup);

				/*
				 * mark the index tuple as moved by split, such tuples are
				 * skipped by scan if there is split in progress for a bucket.
				 */
				new_itup->t_info |= INDEX_MOVED_BY_SPLIT_MASK;

				/*
				 * insert the tuple into the new bucket.  if it doesn't fit on
				 * the current page in the new bucket, we must allocate a new
				 * overflow page and place the tuple on that page instead.
				 */
				itemsz = IndexTupleDSize(*new_itup);
				itemsz = MAXALIGN(itemsz);

				if (PageGetFreeSpaceForMultipleTuples(npage, nitups + 1) < (all_tups_size + itemsz))
				{
					/*
					 * Change the shared buffer state in critical section,
					 * otherwise any error could make it unrecoverable.
					 */
					START_CRIT_SECTION();

					_hash_pgaddmultitup(rel, nbuf, itups, itup_offsets, nitups);
					MarkBufferDirty(nbuf);
					/* log the split operation before releasing the lock */
					log_split_page(rel, nbuf);

					END_CRIT_SECTION();

					/* drop lock, but keep pin */
					LockBuffer(nbuf, BUFFER_LOCK_UNLOCK);

					/* be tidy */
					for (i = 0; i < nitups; i++)
						pfree(itups[i]);
					nitups = 0;
					all_tups_size = 0;

					/* chain to a new overflow page */
					nbuf = _hash_addovflpage(rel, metabuf, nbuf, (nbuf == bucket_nbuf) ? true : false);
					npage = BufferGetPage(nbuf);
					nopaque = (HashPageOpaque) PageGetSpecialPointer(npage);
				}

				itups[nitups++] = new_itup;
				all_tups_size += itemsz;
			}
			else
			{
				/*
				 * the tuple stays on this page, so nothing to do.
				 */
				Assert(bucket == obucket);
			}
		}

		oblkno = oopaque->hasho_nextblkno;

		/* retain the pin on the old primary bucket */
		if (obuf == bucket_obuf)
			LockBuffer(obuf, BUFFER_LOCK_UNLOCK);
		else
			_hash_relbuf(rel, obuf);

		/* Exit loop if no more overflow pages in old bucket */
		if (!BlockNumberIsValid(oblkno))
		{
			/*
			 * Change the shared buffer state in critical section, otherwise
			 * any error could make it unrecoverable.
			 */
			START_CRIT_SECTION();

			_hash_pgaddmultitup(rel, nbuf, itups, itup_offsets, nitups);
			MarkBufferDirty(nbuf);
			/* log the split operation before releasing the lock */
			log_split_page(rel, nbuf);

			END_CRIT_SECTION();

			if (nbuf == bucket_nbuf)
				LockBuffer(nbuf, BUFFER_LOCK_UNLOCK);
			else
				_hash_relbuf(rel, nbuf);

			/* be tidy */
			for (i = 0; i < nitups; i++)
				pfree(itups[i]);
			break;
		}

		/* Else, advance to next old page */
		obuf = _hash_getbuf(rel, oblkno, HASH_READ, LH_OVERFLOW_PAGE);
		opage = BufferGetPage(obuf);
		oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);
	}

	/*
	 * We're at the end of the old bucket chain, so we're done partitioning
	 * the tuples.  Mark the old and new buckets to indicate split is
	 * finished.
	 *
	 * To avoid deadlocks due to locking order of buckets, first lock the old
	 * bucket and then the new bucket.
	 */
	LockBuffer(bucket_obuf, BUFFER_LOCK_EXCLUSIVE);
	opage = BufferGetPage(bucket_obuf);
	oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);

	LockBuffer(bucket_nbuf, BUFFER_LOCK_EXCLUSIVE);
	npage = BufferGetPage(bucket_nbuf);
	nopaque = (HashPageOpaque) PageGetSpecialPointer(npage);

	START_CRIT_SECTION();

	oopaque->hasho_flag &= ~LH_BUCKET_BEING_SPLIT;
	nopaque->hasho_flag &= ~LH_BUCKET_BEING_POPULATED;

	/*
	 * After the split is finished, mark the old bucket to indicate that it
	 * contains deletable tuples.  We will clear split-cleanup flag after
	 * deleting such tuples either at the end of split or at the next split
	 * from old bucket or at the time of vacuum.
	 */
	oopaque->hasho_flag |= LH_BUCKET_NEEDS_SPLIT_CLEANUP;

	/*
	 * now write the buffers, here we don't release the locks as caller is
	 * responsible to release locks.
	 */
	MarkBufferDirty(bucket_obuf);
	MarkBufferDirty(bucket_nbuf);

	if (RelationNeedsWAL(rel))
	{
		XLogRecPtr	recptr;
		xl_hash_split_complete xlrec;

		xlrec.old_bucket_flag = oopaque->hasho_flag;
		xlrec.new_bucket_flag = nopaque->hasho_flag;

		XLogBeginInsert();

		XLogRegisterData((char *) &xlrec, SizeOfHashSplitComplete);

		XLogRegisterBuffer(0, bucket_obuf, REGBUF_STANDARD);
		XLogRegisterBuffer(1, bucket_nbuf, REGBUF_STANDARD);

		recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_SPLIT_COMPLETE);

		PageSetLSN(BufferGetPage(bucket_obuf), recptr);
		PageSetLSN(BufferGetPage(bucket_nbuf), recptr);
	}

	END_CRIT_SECTION();

	/*
	 * If possible, clean up the old bucket.  We might not be able to do this
	 * if someone else has a pin on it, but if not then we can go ahead.  This
	 * isn't absolutely necessary, but it reduces bloat; if we don't do it
	 * now, VACUUM will do it eventually, but maybe not until new overflow
	 * pages have been allocated.  Note that there's no need to clean up the
	 * new bucket.
	 */
	if (IsBufferCleanupOK(bucket_obuf))
	{
		LockBuffer(bucket_nbuf, BUFFER_LOCK_UNLOCK);
		hashbucketcleanup(rel, obucket, bucket_obuf,
						  BufferGetBlockNumber(bucket_obuf), NULL,
						  maxbucket, highmask, lowmask, NULL, NULL, true,
						  NULL, NULL);
	}
	else
	{
		LockBuffer(bucket_nbuf, BUFFER_LOCK_UNLOCK);
		LockBuffer(bucket_obuf, BUFFER_LOCK_UNLOCK);
	}
}
示例#8
0
/*
 *	_hash_doinsert() -- Handle insertion of a single index tuple.
 *
 *		This routine is called by the public interface routines, hashbuild
 *		and hashinsert.  By here, itup is completely filled in.
 */
void
_hash_doinsert(Relation rel, IndexTuple itup)
{
	Buffer		buf = InvalidBuffer;
	Buffer		bucket_buf;
	Buffer		metabuf;
	HashMetaPage metap;
	BlockNumber blkno;
	BlockNumber oldblkno;
	bool		retry;
	Page		page;
	HashPageOpaque pageopaque;
	Size		itemsz;
	bool		do_expand;
	uint32		hashkey;
	Bucket		bucket;
	uint32		maxbucket;
	uint32		highmask;
	uint32		lowmask;

	/*
	 * Get the hash key for the item (it's stored in the index tuple itself).
	 */
	hashkey = _hash_get_indextuple_hashkey(itup);

	/* compute item size too */
	itemsz = IndexTupleDSize(*itup);
	itemsz = MAXALIGN(itemsz);	/* be safe, PageAddItem will do this but we
								 * need to be consistent */

restart_insert:
	/* Read the metapage */
	metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
	metap = HashPageGetMeta(BufferGetPage(metabuf));

	/*
	 * Check whether the item can fit on a hash page at all. (Eventually, we
	 * ought to try to apply TOAST methods if not.)  Note that at this point,
	 * itemsz doesn't include the ItemId.
	 *
	 * XXX this is useless code if we are only storing hash keys.
	 */
	if (itemsz > HashMaxItemSize((Page) metap))
		ereport(ERROR,
				(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
				 errmsg("index row size %zu exceeds hash maximum %zu",
						itemsz, HashMaxItemSize((Page) metap)),
			errhint("Values larger than a buffer page cannot be indexed.")));

	oldblkno = InvalidBlockNumber;
	retry = false;

	/*
	 * Loop until we get a lock on the correct target bucket.
	 */
	for (;;)
	{
		/*
		 * Compute the target bucket number, and convert to block number.
		 */
		bucket = _hash_hashkey2bucket(hashkey,
									  metap->hashm_maxbucket,
									  metap->hashm_highmask,
									  metap->hashm_lowmask);

		blkno = BUCKET_TO_BLKNO(metap, bucket);

		/*
		 * Copy bucket mapping info now; refer the comment in
		 * _hash_expandtable where we copy this information before calling
		 * _hash_splitbucket to see why this is okay.
		 */
		maxbucket = metap->hashm_maxbucket;
		highmask = metap->hashm_highmask;
		lowmask = metap->hashm_lowmask;

		/* Release metapage lock, but keep pin. */
		_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);

		/*
		 * If the previous iteration of this loop locked the primary page of
		 * what is still the correct target bucket, we are done.  Otherwise,
		 * drop any old lock before acquiring the new one.
		 */
		if (retry)
		{
			if (oldblkno == blkno)
				break;
			_hash_relbuf(rel, buf);
		}

		/* Fetch and lock the primary bucket page for the target bucket */
		buf = _hash_getbuf(rel, blkno, HASH_WRITE, LH_BUCKET_PAGE);

		/*
		 * Reacquire metapage lock and check that no bucket split has taken
		 * place while we were awaiting the bucket lock.
		 */
		_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_READ);
		oldblkno = blkno;
		retry = true;
	}

	/* remember the primary bucket buffer to release the pin on it at end. */
	bucket_buf = buf;

	page = BufferGetPage(buf);
	pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
	Assert(pageopaque->hasho_bucket == bucket);

	/*
	 * If this bucket is in the process of being split, try to finish the
	 * split before inserting, because that might create room for the
	 * insertion to proceed without allocating an additional overflow page.
	 * It's only interesting to finish the split if we're trying to insert
	 * into the bucket from which we're removing tuples (the "old" bucket),
	 * not if we're trying to insert into the bucket into which tuples are
	 * being moved (the "new" bucket).
	 */
	if (H_BUCKET_BEING_SPLIT(pageopaque) && IsBufferCleanupOK(buf))
	{
		/* release the lock on bucket buffer, before completing the split. */
		_hash_chgbufaccess(rel, buf, HASH_READ, HASH_NOLOCK);

		_hash_finish_split(rel, metabuf, buf, pageopaque->hasho_bucket,
						   maxbucket, highmask, lowmask);

		/* release the pin on old and meta buffer.  retry for insert. */
		_hash_dropbuf(rel, buf);
		_hash_dropbuf(rel, metabuf);
		goto restart_insert;
	}

	/* Do the insertion */
	while (PageGetFreeSpace(page) < itemsz)
	{
		/*
		 * no space on this page; check for an overflow page
		 */
		BlockNumber nextblkno = pageopaque->hasho_nextblkno;

		if (BlockNumberIsValid(nextblkno))
		{
			/*
			 * ovfl page exists; go get it.  if it doesn't have room, we'll
			 * find out next pass through the loop test above.  we always
			 * release both the lock and pin if this is an overflow page, but
			 * only the lock if this is the primary bucket page, since the pin
			 * on the primary bucket must be retained throughout the scan.
			 */
			if (buf != bucket_buf)
				_hash_relbuf(rel, buf);
			else
				_hash_chgbufaccess(rel, buf, HASH_READ, HASH_NOLOCK);
			buf = _hash_getbuf(rel, nextblkno, HASH_WRITE, LH_OVERFLOW_PAGE);
			page = BufferGetPage(buf);
		}
		else
		{
			/*
			 * we're at the end of the bucket chain and we haven't found a
			 * page with enough room.  allocate a new overflow page.
			 */

			/* release our write lock without modifying buffer */
			_hash_chgbufaccess(rel, buf, HASH_READ, HASH_NOLOCK);

			/* chain to a new overflow page */
			buf = _hash_addovflpage(rel, metabuf, buf, (buf == bucket_buf) ? true : false);
			page = BufferGetPage(buf);

			/* should fit now, given test above */
			Assert(PageGetFreeSpace(page) >= itemsz);
		}
		pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
		Assert(pageopaque->hasho_flag == LH_OVERFLOW_PAGE);
		Assert(pageopaque->hasho_bucket == bucket);
	}

	/* found page with enough space, so add the item here */
	(void) _hash_pgaddtup(rel, buf, itemsz, itup);

	/*
	 * dirty and release the modified page.  if the page we modified was an
	 * overflow page, we also need to separately drop the pin we retained on
	 * the primary bucket page.
	 */
	MarkBufferDirty(buf);
	_hash_relbuf(rel, buf);
	if (buf != bucket_buf)
		_hash_dropbuf(rel, bucket_buf);

	/*
	 * Write-lock the metapage so we can increment the tuple count. After
	 * incrementing it, check to see if it's time for a split.
	 */
	_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);

	metap->hashm_ntuples += 1;

	/* Make sure this stays in sync with _hash_expandtable() */
	do_expand = metap->hashm_ntuples >
		(double) metap->hashm_ffactor * (metap->hashm_maxbucket + 1);

	/* Write out the metapage and drop lock, but keep pin */
	_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);

	/* Attempt to split if a split is needed */
	if (do_expand)
		_hash_expandtable(rel, metabuf);

	/* Finally drop our pin on the metapage */
	_hash_dropbuf(rel, metabuf);
}
示例#9
0
/*
 *	_hash_first() -- Find the first item in a scan.
 *
 *		Find the first item in the index that
 *		satisfies the qualification associated with the scan descriptor. On
 *		success, the page containing the current index tuple is read locked
 *		and pinned, and the scan's opaque data entry is updated to
 *		include the buffer.
 */
bool _hash_first(struct index_scan *scan, enum scandir dir)
{
	struct relation *rel = scan->indexRelation;
	struct hash_scan_opaque_data *so = (struct hash_scan_opaque_data *)scan->opaque;
	struct scankey *cur;
	uint32 hashkey;
	bucket_t bucket;
	block_t blkno;
	buf_id_t buf;
	buf_id_t metabuf;
	page_p page;
	struct hash_page *opaque;
	struct hash_meta_page_data *metap;
	struct index_tuple *itup;
	struct item_ptr *current;
	item_id_t offnum;

	stat_index_scan(rel);

	current = &(so->hashso_curpos);
	ITEM_PTR_SET_INVALID(current);

	/*
	 * We do not support hash scans with no index qualification, because we
	 * would have to read the whole index rather than just one bucket. That
	 * creates a whole raft of problems, since we haven't got a practical way
	 * to lock all the buckets against splits or compactions.
	 */
	if (scan->numberOfKeys < 1) {
		ereport(ERROR, (
		errcode(E_FEATURE_NOT_SUPPORTED),
		errmsg("hash indexes do not support whole-index scans")));
	}

	/* There may be more than one index qual, but we hash only the first */
	cur = &scan->keyData[0];

	/* We support only single-column hash indexes */
	ASSERT(cur->sk_attno == 1);

	/* And there's only one operator strategy, too */
	ASSERT(cur->sk_strategy == HT_EQ_STRATEGY_NR);

	/*
	 * If the constant in the index qual is NULL, assume it cannot match any
	 * items in the index.
	 */
	if (cur->sk_flags & SK_ISNULL)
		return false;

	/*
	 * Okay to compute the hash key.  We want to do this before acquiring any
	 * locks, in case a user-defined hash function happens to be slow.
	 *
	 * If scankey operator is not a cross-type comparison, we can use the
	 * cached hash function; otherwise gotta look it up in the catalogs.
	 *
	 * We support the convention that sk_subtype == INVALID_OID means the
	 * opclass input type; this is a hack to simplify life for scankey_init().
	 */
	if (cur->sk_subtype == rel->rd_opcintype[0]
		|| cur->sk_subtype == INVALID_OID)
		hashkey = _hash_datum2hashkey(rel, cur->sk_argument);
	else
		hashkey = _hash_datum2hashkey_type(rel, cur->sk_argument, cur->sk_subtype);

	so->hashso_sk_hash = hashkey;

	/*
	 * Acquire shared split lock so we can compute the target bucket safely
	 * (see README).
	 */
	_hash_getlock(rel, 0, HASH_SHARE);

	/* Read the metapage */
	metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
	metap = HASH_PAGE_GET_META(BUF_PAGE(metabuf));

	/*
	 * Compute the target bucket number, and convert to block number.
	 */
	bucket = _hash_hashkey2bucket(
		hashkey,
		metap->hashm_maxbucket,	
		metap->hashm_highmask,
		metap->hashm_lowmask);

	blkno = BUCKET_TO_BLKNO(metap, bucket);

	/* done with the metapage */
	_hash_relbuf(rel, metabuf);

	/*
	 * Acquire share lock on target bucket; then we can release split lock.
	 */
	_hash_getlock(rel, blkno, HASH_SHARE);
	_hash_droplock(rel, 0, HASH_SHARE);

	/* Update scan opaque state to show we have lock on the bucket */
	so->hashso_bucket = bucket;
	so->hashso_bucket_valid = true;
	so->hashso_bucket_blkno = blkno;

	/* Fetch the primary bucket page for the bucket */
	buf = _hash_getbuf(rel, blkno, HASH_READ, LH_BUCKET_PAGE);
	page = BUF_PAGE(buf);
	opaque = (struct hash_page*) PAGE_SPECIAL_PTR(page);
	ASSERT(opaque->hasho_bucket == bucket);

	/* If a backwards scan is requested, move to the end of the chain */
	if (SCANDIR_BACKWARD(dir)) {
		while (BLK_NR_VALID(opaque->hasho_nextblkno))
			_hash_readnext(rel, &buf, &page, &opaque);
	}

	/* Now find the first tuple satisfying the qualification */
	if (!_hash_step(scan, &buf, dir))
		return false;

	/* if we're here, _hash_step found a valid tuple */
	offnum = ITEM_PTR_OFFSET(current);
	_hash_checkpage(rel, buf, LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
	page = BUF_PAGE(buf);
	itup = (struct index_tuple *)PAGE_GET_ITEM(page, PAGE_ITEM_ID(page, offnum));
	so->hashso_heappos = itup->t_tid;

	return true;
}
示例#10
0
/*
 * Bulk deletion of all index entries pointing to a set of heap tuples.
 * The set of target tuples is specified via a callback routine that tells
 * whether any given heap tuple (identified by ItemPointer) is being deleted.
 *
 * Result: a palloc'd struct containing statistical info for VACUUM displays.
 */
Datum
hashbulkdelete(PG_FUNCTION_ARGS)
{
	IndexVacuumInfo *info = (IndexVacuumInfo *) PG_GETARG_POINTER(0);
	IndexBulkDeleteResult *stats = (IndexBulkDeleteResult *) PG_GETARG_POINTER(1);
	IndexBulkDeleteCallback callback = (IndexBulkDeleteCallback) PG_GETARG_POINTER(2);
	void	   *callback_state = (void *) PG_GETARG_POINTER(3);
	Relation	rel = info->index;
	double		tuples_removed;
	double		num_index_tuples;
	double		orig_ntuples;
	Bucket		orig_maxbucket;
	Bucket		cur_maxbucket;
	Bucket		cur_bucket;
	Buffer		metabuf;
	HashMetaPage metap;
	HashMetaPageData local_metapage;

	tuples_removed = 0;
	num_index_tuples = 0;

	/*
	 * Read the metapage to fetch original bucket and tuple counts.  Also, we
	 * keep a copy of the last-seen metapage so that we can use its
	 * hashm_spares[] values to compute bucket page addresses.	This is a bit
	 * hokey but perfectly safe, since the interesting entries in the spares
	 * array cannot change under us; and it beats rereading the metapage for
	 * each bucket.
	 */
	metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
	metap = HashPageGetMeta(BufferGetPage(metabuf));
	orig_maxbucket = metap->hashm_maxbucket;
	orig_ntuples = metap->hashm_ntuples;
	memcpy(&local_metapage, metap, sizeof(local_metapage));
	_hash_relbuf(rel, metabuf);

	/* Scan the buckets that we know exist */
	cur_bucket = 0;
	cur_maxbucket = orig_maxbucket;

loop_top:
	while (cur_bucket <= cur_maxbucket)
	{
		BlockNumber bucket_blkno;
		BlockNumber blkno;
		bool		bucket_dirty = false;

		/* Get address of bucket's start page */
		bucket_blkno = BUCKET_TO_BLKNO(&local_metapage, cur_bucket);

		/* Exclusive-lock the bucket so we can shrink it */
		_hash_getlock(rel, bucket_blkno, HASH_EXCLUSIVE);

		/* Shouldn't have any active scans locally, either */
		if (_hash_has_active_scan(rel, cur_bucket))
			elog(ERROR, "hash index has active scan during VACUUM");

		/* Scan each page in bucket */
		blkno = bucket_blkno;
		while (BlockNumberIsValid(blkno))
		{
			Buffer		buf;
			Page		page;
			HashPageOpaque opaque;
			OffsetNumber offno;
			OffsetNumber maxoffno;
			OffsetNumber deletable[MaxOffsetNumber];
			int			ndeletable = 0;

			vacuum_delay_point();

			buf = _hash_getbuf_with_strategy(rel, blkno, HASH_WRITE,
										   LH_BUCKET_PAGE | LH_OVERFLOW_PAGE,
											 info->strategy);
			page = BufferGetPage(buf);
			opaque = (HashPageOpaque) PageGetSpecialPointer(page);
			Assert(opaque->hasho_bucket == cur_bucket);

			/* Scan each tuple in page */
			maxoffno = PageGetMaxOffsetNumber(page);
			for (offno = FirstOffsetNumber;
				 offno <= maxoffno;
				 offno = OffsetNumberNext(offno))
			{
				IndexTuple	itup;
				ItemPointer htup;

				itup = (IndexTuple) PageGetItem(page,
												PageGetItemId(page, offno));
				htup = &(itup->t_tid);
				if (callback(htup, callback_state))
				{
					/* mark the item for deletion */
					deletable[ndeletable++] = offno;
					tuples_removed += 1;
				}
				else
					num_index_tuples += 1;
			}

			/*
			 * Apply deletions and write page if needed, advance to next page.
			 */
			blkno = opaque->hasho_nextblkno;

			if (ndeletable > 0)
			{
				PageIndexMultiDelete(page, deletable, ndeletable);
				_hash_wrtbuf(rel, buf);
				bucket_dirty = true;
			}
			else
				_hash_relbuf(rel, buf);
		}

		/* If we deleted anything, try to compact free space */
		if (bucket_dirty)
			_hash_squeezebucket(rel, cur_bucket, bucket_blkno,
								info->strategy);

		/* Release bucket lock */
		_hash_droplock(rel, bucket_blkno, HASH_EXCLUSIVE);

		/* Advance to next bucket */
		cur_bucket++;
	}

	/* Write-lock metapage and check for split since we started */
	metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_WRITE, LH_META_PAGE);
	metap = HashPageGetMeta(BufferGetPage(metabuf));

	if (cur_maxbucket != metap->hashm_maxbucket)
	{
		/* There's been a split, so process the additional bucket(s) */
		cur_maxbucket = metap->hashm_maxbucket;
		memcpy(&local_metapage, metap, sizeof(local_metapage));
		_hash_relbuf(rel, metabuf);
		goto loop_top;
	}

	/* Okay, we're really done.  Update tuple count in metapage. */

	if (orig_maxbucket == metap->hashm_maxbucket &&
		orig_ntuples == metap->hashm_ntuples)
	{
		/*
		 * No one has split or inserted anything since start of scan, so
		 * believe our count as gospel.
		 */
		metap->hashm_ntuples = num_index_tuples;
	}
	else
	{
		/*
		 * Otherwise, our count is untrustworthy since we may have
		 * double-scanned tuples in split buckets.	Proceed by dead-reckoning.
		 * (Note: we still return estimated_count = false, because using this
		 * count is better than not updating reltuples at all.)
		 */
		if (metap->hashm_ntuples > tuples_removed)
			metap->hashm_ntuples -= tuples_removed;
		else
			metap->hashm_ntuples = 0;
		num_index_tuples = metap->hashm_ntuples;
	}

	_hash_wrtbuf(rel, metabuf);

	/* return statistics */
	if (stats == NULL)
		stats = (IndexBulkDeleteResult *) palloc0(sizeof(IndexBulkDeleteResult));
	stats->estimated_count = false;
	stats->num_index_tuples = num_index_tuples;
	stats->tuples_removed += tuples_removed;
	/* hashvacuumcleanup will fill in num_pages */

	PG_RETURN_POINTER(stats);
}