/*
* _hash_addovflpage
*
* Add an overflow page to the bucket whose last page is pointed to by 'buf'.
*
* On entry, the caller must hold a pin but no lock on 'buf'. The pin is
* dropped before exiting (we assume the caller is not interested in 'buf'
* anymore). The returned overflow page will be pinned and write-locked;
* it is guaranteed to be empty.
*
* The caller must hold a pin, but no lock, on the metapage buffer.
* That buffer is returned in the same state.
*
* The caller must hold at least share lock on the bucket, to ensure that
* no one else tries to compact the bucket meanwhile. This guarantees that
* 'buf' won't stop being part of the bucket while it's unlocked.
*
* NB: since this could be executed concurrently by multiple processes,
* one should not assume that the returned overflow page will be the
* immediate successor of the originally passed 'buf'. Additional overflow
* pages might have been added to the bucket chain in between.
*/
Buffer
_hash_addovflpage(Relation rel, Buffer metabuf, Buffer buf)
{
Buffer ovflbuf;
Page page;
Page ovflpage;
HashPageOpaque pageopaque;
HashPageOpaque ovflopaque;
/* allocate and lock an empty overflow page */
ovflbuf = _hash_getovflpage(rel, metabuf);
/*
* Write-lock the tail page. It is okay to hold two buffer locks here
* since there cannot be anyone else contending for access to ovflbuf.
*/
_hash_chgbufaccess(rel, buf, HASH_NOLOCK, HASH_WRITE);
/* probably redundant... */
_hash_checkpage(rel, buf, LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
/* loop to find current tail page, in case someone else inserted too */
for (;;)
{
BlockNumber nextblkno;
page = BufferGetPage(buf);
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
nextblkno = pageopaque->hasho_nextblkno;
if (!BlockNumberIsValid(nextblkno))
break;
/* we assume we do not need to write the unmodified page */
_hash_relbuf(rel, buf);
buf = _hash_getbuf(rel, nextblkno, HASH_WRITE, LH_OVERFLOW_PAGE);
}
/* now that we have correct backlink, initialize new overflow page */
ovflpage = BufferGetPage(ovflbuf);
ovflopaque = (HashPageOpaque) PageGetSpecialPointer(ovflpage);
ovflopaque->hasho_prevblkno = BufferGetBlockNumber(buf);
ovflopaque->hasho_nextblkno = InvalidBlockNumber;
ovflopaque->hasho_bucket = pageopaque->hasho_bucket;
ovflopaque->hasho_flag = LH_OVERFLOW_PAGE;
ovflopaque->hasho_page_id = HASHO_PAGE_ID;
MarkBufferDirty(ovflbuf);
/* logically chain overflow page to previous page */
pageopaque->hasho_nextblkno = BufferGetBlockNumber(ovflbuf);
_hash_wrtbuf(rel, buf);
return ovflbuf;
}
/*
* _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;
}
/*
* _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);
}
/*
* _hash_splitbucket -- split 'obucket' into 'obucket' and 'nbucket'
*
* 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, and compress out any free space in the old
* bucket.
*
* The caller must hold exclusive 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.)
*/
static void
_hash_splitbucket(Relation rel,
Buffer metabuf,
Bucket obucket,
Bucket nbucket,
BlockNumber start_oblkno,
BlockNumber start_nblkno,
uint32 maxbucket,
uint32 highmask,
uint32 lowmask)
{
Bucket bucket;
Buffer obuf;
Buffer nbuf;
BlockNumber oblkno;
BlockNumber nblkno;
bool null;
Datum datum;
HashItem hitem;
HashPageOpaque oopaque;
HashPageOpaque nopaque;
IndexTuple itup;
Size itemsz;
OffsetNumber ooffnum;
OffsetNumber noffnum;
OffsetNumber omaxoffnum;
Page opage;
Page npage;
TupleDesc itupdesc = RelationGetDescr(rel);
/*
* It should be okay to simultaneously write-lock pages from each
* bucket, since no one else can be trying to acquire buffer lock
* on pages of either bucket.
*/
oblkno = start_oblkno;
nblkno = start_nblkno;
obuf = _hash_getbuf(rel, oblkno, HASH_WRITE);
nbuf = _hash_getbuf(rel, nblkno, HASH_WRITE);
opage = BufferGetPage(obuf);
npage = BufferGetPage(nbuf);
_hash_checkpage(rel, opage, LH_BUCKET_PAGE);
oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);
/* initialize the new bucket's primary page */
_hash_pageinit(npage, BufferGetPageSize(nbuf));
nopaque = (HashPageOpaque) PageGetSpecialPointer(npage);
nopaque->hasho_prevblkno = InvalidBlockNumber;
nopaque->hasho_nextblkno = InvalidBlockNumber;
nopaque->hasho_bucket = nbucket;
nopaque->hasho_flag = LH_BUCKET_PAGE;
nopaque->hasho_filler = HASHO_FILL;
/*
* 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.
*/
ooffnum = FirstOffsetNumber;
omaxoffnum = PageGetMaxOffsetNumber(opage);
for (;;)
{
/*
* at each iteration through this loop, each of these variables
* should be up-to-date: obuf opage oopaque ooffnum omaxoffnum
*/
/* check if we're at the end of the page */
if (ooffnum > omaxoffnum)
{
/* at end of page, but check for an(other) overflow page */
oblkno = oopaque->hasho_nextblkno;
if (!BlockNumberIsValid(oblkno))
break;
/*
* we ran out of tuples on this particular page, but we
* have more overflow pages; advance to next page.
*/
_hash_wrtbuf(rel, obuf);
obuf = _hash_getbuf(rel, oblkno, HASH_WRITE);
opage = BufferGetPage(obuf);
_hash_checkpage(rel, opage, LH_OVERFLOW_PAGE);
oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);
ooffnum = FirstOffsetNumber;
omaxoffnum = PageGetMaxOffsetNumber(opage);
continue;
}
/*
* Re-hash the tuple to determine which bucket it now belongs in.
*
* It is annoying to call the hash function while holding locks,
* but releasing and relocking the page for each tuple is unappealing
* too.
*/
hitem = (HashItem) PageGetItem(opage, PageGetItemId(opage, ooffnum));
itup = &(hitem->hash_itup);
datum = index_getattr(itup, 1, itupdesc, &null);
Assert(!null);
bucket = _hash_hashkey2bucket(_hash_datum2hashkey(rel, datum),
maxbucket, highmask, lowmask);
if (bucket == nbucket)
{
/*
* 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(hitem->hash_itup)
+ (sizeof(HashItemData) - sizeof(IndexTupleData));
itemsz = MAXALIGN(itemsz);
if (PageGetFreeSpace(npage) < itemsz)
{
/* write out nbuf and drop lock, but keep pin */
_hash_chgbufaccess(rel, nbuf, HASH_WRITE, HASH_NOLOCK);
/* chain to a new overflow page */
nbuf = _hash_addovflpage(rel, metabuf, nbuf);
npage = BufferGetPage(nbuf);
_hash_checkpage(rel, npage, LH_OVERFLOW_PAGE);
/* we don't need nopaque within the loop */
}
noffnum = OffsetNumberNext(PageGetMaxOffsetNumber(npage));
if (PageAddItem(npage, (Item) hitem, itemsz, noffnum, LP_USED)
== InvalidOffsetNumber)
elog(ERROR, "failed to add index item to \"%s\"",
RelationGetRelationName(rel));
/*
* now delete the tuple from the old bucket. after this
* section of code, 'ooffnum' will actually point to the
* ItemId to which we would point if we had advanced it before
* the deletion (PageIndexTupleDelete repacks the ItemId
* array). this also means that 'omaxoffnum' is exactly one
* less than it used to be, so we really can just decrement it
* instead of calling PageGetMaxOffsetNumber.
*/
PageIndexTupleDelete(opage, ooffnum);
omaxoffnum = OffsetNumberPrev(omaxoffnum);
}
else
{
/*
* the tuple stays on this page. we didn't move anything, so
* we didn't delete anything and therefore we don't have to
* change 'omaxoffnum'.
*/
Assert(bucket == obucket);
ooffnum = OffsetNumberNext(ooffnum);
}
}
/*
* We're at the end of the old bucket chain, so we're done partitioning
* the tuples. Before quitting, call _hash_squeezebucket to ensure the
* tuples remaining in the old bucket (including the overflow pages) are
* packed as tightly as possible. The new bucket is already tight.
*/
_hash_wrtbuf(rel, obuf);
_hash_wrtbuf(rel, nbuf);
_hash_squeezebucket(rel, obucket, start_oblkno);
}
/*
* Attempt to expand the hash table by creating one new bucket.
*
* This will silently do nothing if it cannot get the needed locks.
*
* The caller should hold no locks on the hash index.
*
* The caller must hold a pin, but no lock, on the metapage buffer.
* The buffer is returned in the same state.
*/
void
_hash_expandtable(Relation rel, Buffer metabuf)
{
HashMetaPage metap;
Bucket old_bucket;
Bucket new_bucket;
uint32 spare_ndx;
BlockNumber start_oblkno;
BlockNumber start_nblkno;
uint32 maxbucket;
uint32 highmask;
uint32 lowmask;
/*
* Obtain the page-zero lock to assert the right to begin a split
* (see README).
*
* Note: deadlock should be impossible here. Our own backend could only
* be holding bucket sharelocks due to stopped indexscans; those will not
* block other holders of the page-zero lock, who are only interested in
* acquiring bucket sharelocks themselves. Exclusive bucket locks are
* only taken here and in hashbulkdelete, and neither of these operations
* needs any additional locks to complete. (If, due to some flaw in this
* reasoning, we manage to deadlock anyway, it's okay to error out; the
* index will be left in a consistent state.)
*/
_hash_getlock(rel, 0, HASH_EXCLUSIVE);
/* Write-lock the meta page */
_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);
metap = (HashMetaPage) BufferGetPage(metabuf);
_hash_checkpage(rel, (Page) metap, LH_META_PAGE);
/*
* Check to see if split is still needed; someone else might have already
* done one while we waited for the lock.
*
* Make sure this stays in sync with_hash_doinsert()
*/
if (metap->hashm_ntuples <=
(double) metap->hashm_ffactor * (metap->hashm_maxbucket + 1))
goto fail;
/*
* Determine which bucket is to be split, and attempt to lock the old
* bucket. If we can't get the lock, give up.
*
* The lock protects us against other backends, but not against our own
* backend. Must check for active scans separately.
*
* Ideally we would lock the new bucket too before proceeding, but if
* we are about to cross a splitpoint then the BUCKET_TO_BLKNO mapping
* isn't correct yet. For simplicity we update the metapage first and
* then lock. This should be okay because no one else should be trying
* to lock the new bucket yet...
*/
new_bucket = metap->hashm_maxbucket + 1;
old_bucket = (new_bucket & metap->hashm_lowmask);
start_oblkno = BUCKET_TO_BLKNO(metap, old_bucket);
if (_hash_has_active_scan(rel, old_bucket))
goto fail;
if (!_hash_try_getlock(rel, start_oblkno, HASH_EXCLUSIVE))
goto fail;
/*
* Okay to proceed with split. Update the metapage bucket mapping info.
*/
metap->hashm_maxbucket = new_bucket;
if (new_bucket > metap->hashm_highmask)
{
/* Starting a new doubling */
metap->hashm_lowmask = metap->hashm_highmask;
metap->hashm_highmask = new_bucket | metap->hashm_lowmask;
}
/*
* If the split point is increasing (hashm_maxbucket's log base 2
* increases), we need to adjust the hashm_spares[] array and
* hashm_ovflpoint so that future overflow pages will be created beyond
* this new batch of bucket pages.
*
* XXX should initialize new bucket pages to prevent out-of-order
* page creation? Don't wanna do it right here though.
*/
spare_ndx = _hash_log2(metap->hashm_maxbucket + 1);
if (spare_ndx > metap->hashm_ovflpoint)
{
Assert(spare_ndx == metap->hashm_ovflpoint + 1);
metap->hashm_spares[spare_ndx] = metap->hashm_spares[metap->hashm_ovflpoint];
metap->hashm_ovflpoint = spare_ndx;
}
/* now we can compute the new bucket's primary block number */
start_nblkno = BUCKET_TO_BLKNO(metap, new_bucket);
Assert(!_hash_has_active_scan(rel, new_bucket));
if (!_hash_try_getlock(rel, start_nblkno, HASH_EXCLUSIVE))
elog(PANIC, "could not get lock on supposedly new bucket");
/*
* Copy bucket mapping info now; this saves re-accessing the meta page
* inside _hash_splitbucket's inner loop. Note that once we drop the
* split lock, other splits could begin, so these values might be out of
* date before _hash_splitbucket finishes. That's okay, since all it
* needs is to tell which of these two buckets to map hashkeys into.
*/
maxbucket = metap->hashm_maxbucket;
highmask = metap->hashm_highmask;
lowmask = metap->hashm_lowmask;
/* Write out the metapage and drop lock, but keep pin */
_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);
/* Release split lock; okay for other splits to occur now */
_hash_droplock(rel, 0, HASH_EXCLUSIVE);
/* Relocate records to the new bucket */
_hash_splitbucket(rel, metabuf, old_bucket, new_bucket,
start_oblkno, start_nblkno,
maxbucket, highmask, lowmask);
/* Release bucket locks, allowing others to access them */
_hash_droplock(rel, start_oblkno, HASH_EXCLUSIVE);
_hash_droplock(rel, start_nblkno, HASH_EXCLUSIVE);
return;
/* Here if decide not to split or fail to acquire old bucket lock */
fail:
/* We didn't write the metapage, so just drop lock */
_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);
/* Release split lock */
_hash_droplock(rel, 0, HASH_EXCLUSIVE);
}
/*
* Helper function to perform deletion of index entries from a bucket.
*
* This function expects that the caller has acquired a cleanup lock on the
* primary bucket page, and will return with a write lock again held on the
* primary bucket page. The lock won't necessarily be held continuously,
* though, because we'll release it when visiting overflow pages.
*
* It would be very bad if this function cleaned a page while some other
* backend was in the midst of scanning it, because hashgettuple assumes
* that the next valid TID will be greater than or equal to the current
* valid TID. There can't be any concurrent scans in progress when we first
* enter this function because of the cleanup lock we hold on the primary
* bucket page, but as soon as we release that lock, there might be. We
* handle that by conspiring to prevent those scans from passing our cleanup
* scan. To do that, we lock the next page in the bucket chain before
* releasing the lock on the previous page. (This type of lock chaining is
* not ideal, so we might want to look for a better solution at some point.)
*
* We need to retain a pin on the primary bucket to ensure that no concurrent
* split can start.
*/
void
hashbucketcleanup(Relation rel, Bucket cur_bucket, Buffer bucket_buf,
BlockNumber bucket_blkno, BufferAccessStrategy bstrategy,
uint32 maxbucket, uint32 highmask, uint32 lowmask,
double *tuples_removed, double *num_index_tuples,
bool split_cleanup,
IndexBulkDeleteCallback callback, void *callback_state)
{
BlockNumber blkno;
Buffer buf;
Bucket new_bucket PG_USED_FOR_ASSERTS_ONLY = InvalidBucket;
bool bucket_dirty = false;
blkno = bucket_blkno;
buf = bucket_buf;
if (split_cleanup)
new_bucket = _hash_get_newbucket_from_oldbucket(rel, cur_bucket,
lowmask, maxbucket);
/* Scan each page in bucket */
for (;;)
{
HashPageOpaque opaque;
OffsetNumber offno;
OffsetNumber maxoffno;
Buffer next_buf;
Page page;
OffsetNumber deletable[MaxOffsetNumber];
int ndeletable = 0;
bool retain_pin = false;
bool curr_page_dirty = false;
vacuum_delay_point();
page = BufferGetPage(buf);
opaque = (HashPageOpaque) PageGetSpecialPointer(page);
/* Scan each tuple in page */
maxoffno = PageGetMaxOffsetNumber(page);
for (offno = FirstOffsetNumber;
offno <= maxoffno;
offno = OffsetNumberNext(offno))
{
ItemPointer htup;
IndexTuple itup;
Bucket bucket;
bool kill_tuple = false;
itup = (IndexTuple) PageGetItem(page,
PageGetItemId(page, offno));
htup = &(itup->t_tid);
/*
* To remove the dead tuples, we strictly want to rely on results
* of callback function. refer btvacuumpage for detailed reason.
*/
if (callback && callback(htup, callback_state))
{
kill_tuple = true;
if (tuples_removed)
*tuples_removed += 1;
}
else if (split_cleanup)
{
/* delete the tuples that are moved by split. */
bucket = _hash_hashkey2bucket(_hash_get_indextuple_hashkey(itup),
maxbucket,
highmask,
lowmask);
/* mark the item for deletion */
if (bucket != cur_bucket)
{
/*
* We expect tuples to either belong to curent bucket or
* new_bucket. This is ensured because we don't allow
* further splits from bucket that contains garbage. See
* comments in _hash_expandtable.
*/
Assert(bucket == new_bucket);
kill_tuple = true;
}
}
if (kill_tuple)
{
/* mark the item for deletion */
deletable[ndeletable++] = offno;
}
else
{
/* we're keeping it, so count it */
if (num_index_tuples)
*num_index_tuples += 1;
}
}
/* retain the pin on primary bucket page till end of bucket scan */
if (blkno == bucket_blkno)
retain_pin = true;
else
retain_pin = false;
blkno = opaque->hasho_nextblkno;
/*
* Apply deletions, advance to next page and write page if needed.
*/
if (ndeletable > 0)
{
PageIndexMultiDelete(page, deletable, ndeletable);
bucket_dirty = true;
curr_page_dirty = true;
}
/* bail out if there are no more pages to scan. */
if (!BlockNumberIsValid(blkno))
break;
next_buf = _hash_getbuf_with_strategy(rel, blkno, HASH_WRITE,
LH_OVERFLOW_PAGE,
bstrategy);
/*
* release the lock on previous page after acquiring the lock on next
* page
*/
if (curr_page_dirty)
{
if (retain_pin)
_hash_chgbufaccess(rel, buf, HASH_WRITE, HASH_NOLOCK);
else
_hash_wrtbuf(rel, buf);
curr_page_dirty = false;
}
else if (retain_pin)
_hash_chgbufaccess(rel, buf, HASH_READ, HASH_NOLOCK);
else
_hash_relbuf(rel, buf);
buf = next_buf;
}
/*
* lock the bucket page to clear the garbage flag and squeeze the bucket.
* if the current buffer is same as bucket buffer, then we already have
* lock on bucket page.
*/
if (buf != bucket_buf)
{
_hash_relbuf(rel, buf);
_hash_chgbufaccess(rel, bucket_buf, HASH_NOLOCK, HASH_WRITE);
}
/*
* Clear the garbage flag from bucket after deleting the tuples that are
* moved by split. We purposefully clear the flag before squeeze bucket,
* so that after restart, vacuum shouldn't again try to delete the moved
* by split tuples.
*/
if (split_cleanup)
{
HashPageOpaque bucket_opaque;
Page page;
page = BufferGetPage(bucket_buf);
bucket_opaque = (HashPageOpaque) PageGetSpecialPointer(page);
bucket_opaque->hasho_flag &= ~LH_BUCKET_NEEDS_SPLIT_CLEANUP;
}
/*
* If we have deleted anything, try to compact free space. For squeezing
* the bucket, we must have a cleanup lock, else it can impact the
* ordering of tuples for a scan that has started before it.
*/
if (bucket_dirty && IsBufferCleanupOK(bucket_buf))
_hash_squeezebucket(rel, cur_bucket, bucket_blkno, bucket_buf,
bstrategy);
else
_hash_chgbufaccess(rel, bucket_buf, HASH_WRITE, HASH_NOLOCK);
}
/*
* _hash_splitbucket -- split 'obucket' into 'obucket' and 'nbucket'
*
* 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, and compress out any free space in the old
* bucket.
*
* The caller must hold exclusive 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.)
*/
static void
_hash_splitbucket(Relation rel,
Buffer metabuf,
Bucket obucket,
Bucket nbucket,
BlockNumber start_oblkno,
BlockNumber start_nblkno,
uint32 maxbucket,
uint32 highmask,
uint32 lowmask)
{
BlockNumber oblkno;
BlockNumber nblkno;
Buffer obuf;
Buffer nbuf;
Page opage;
Page npage;
HashPageOpaque oopaque;
HashPageOpaque nopaque;
/*
* It should be okay to simultaneously write-lock pages from each bucket,
* since no one else can be trying to acquire buffer lock on pages of
* either bucket.
*/
oblkno = start_oblkno;
obuf = _hash_getbuf(rel, oblkno, HASH_WRITE, LH_BUCKET_PAGE);
opage = BufferGetPage(obuf);
oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);
nblkno = start_nblkno;
nbuf = _hash_getnewbuf(rel, nblkno, MAIN_FORKNUM);
npage = BufferGetPage(nbuf);
/* initialize the new bucket's primary page */
nopaque = (HashPageOpaque) PageGetSpecialPointer(npage);
nopaque->hasho_prevblkno = InvalidBlockNumber;
nopaque->hasho_nextblkno = InvalidBlockNumber;
nopaque->hasho_bucket = nbucket;
nopaque->hasho_flag = LH_BUCKET_PAGE;
nopaque->hasho_page_id = HASHO_PAGE_ID;
/*
* 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 (;;)
{
OffsetNumber ooffnum;
OffsetNumber omaxoffnum;
OffsetNumber deletable[MaxOffsetNumber];
int ndeletable = 0;
/* Scan each tuple in old page */
omaxoffnum = PageGetMaxOffsetNumber(opage);
for (ooffnum = FirstOffsetNumber;
ooffnum <= omaxoffnum;
ooffnum = OffsetNumberNext(ooffnum))
{
IndexTuple itup;
Size itemsz;
Bucket bucket;
/*
* 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));
bucket = _hash_hashkey2bucket(_hash_get_indextuple_hashkey(itup),
maxbucket, highmask, lowmask);
if (bucket == nbucket)
{
/*
* 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(*itup);
itemsz = MAXALIGN(itemsz);
if (PageGetFreeSpace(npage) < itemsz)
{
/* write out nbuf and drop lock, but keep pin */
_hash_chgbufaccess(rel, nbuf, HASH_WRITE, HASH_NOLOCK);
/* chain to a new overflow page */
nbuf = _hash_addovflpage(rel, metabuf, nbuf);
npage = BufferGetPage(nbuf);
/* we don't need nblkno or nopaque within the loop */
}
/*
* Insert tuple on new page, using _hash_pgaddtup to ensure
* correct ordering by hashkey. This is a tad inefficient
* since we may have to shuffle itempointers repeatedly.
* Possible future improvement: accumulate all the items for
* the new page and qsort them before insertion.
*/
(void) _hash_pgaddtup(rel, nbuf, itemsz, itup);
/*
* Mark tuple for deletion from old page.
*/
deletable[ndeletable++] = ooffnum;
}
else
{
/*
* the tuple stays on this page, so nothing to do.
*/
Assert(bucket == obucket);
}
}
oblkno = oopaque->hasho_nextblkno;
/*
* Done scanning this old page. If we moved any tuples, delete them
* from the old page.
*/
if (ndeletable > 0)
{
PageIndexMultiDelete(opage, deletable, ndeletable);
_hash_wrtbuf(rel, obuf);
}
else
_hash_relbuf(rel, obuf);
/* Exit loop if no more overflow pages in old bucket */
if (!BlockNumberIsValid(oblkno))
break;
/* Else, advance to next old page */
obuf = _hash_getbuf(rel, oblkno, HASH_WRITE, 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. Before quitting, call _hash_squeezebucket to ensure the
* tuples remaining in the old bucket (including the overflow pages) are
* packed as tightly as possible. The new bucket is already tight.
*/
_hash_wrtbuf(rel, nbuf);
_hash_squeezebucket(rel, obucket, start_oblkno, NULL);
}
/*
* _hash_doinsert() -- Handle insertion of a single HashItem in the table.
*
* This routine is called by the public interface routines, hashbuild
* and hashinsert. By here, hashitem is completely filled in.
* The datum to be used as a "key" is in the hashitem.
*/
InsertIndexResult
_hash_doinsert(Relation rel, HashItem hitem)
{
Buffer buf;
Buffer metabuf;
HashMetaPage metap;
IndexTuple itup;
BlockNumber itup_blkno;
OffsetNumber itup_off;
InsertIndexResult res;
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.
*/
itup = &(hitem->hash_itup);
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(hitem->hash_itup)
+ (sizeof(HashItemData) - sizeof(IndexTupleData));
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).
*/
_hash_getlock(rel, 0, HASH_SHARE);
/* Read the metapage */
metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ);
metap = (HashMetaPage) BufferGetPage(metabuf);
_hash_checkpage(rel, (Page) metap, LH_META_PAGE);
/*
* 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))));
/*
* 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);
page = BufferGetPage(buf);
_hash_checkpage(rel, page, LH_BUCKET_PAGE);
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);
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);
}
_hash_checkpage(rel, page, LH_OVERFLOW_PAGE);
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
Assert(pageopaque->hasho_bucket == bucket);
}
/* found page with enough space, so add the item here */
itup_off = _hash_pgaddtup(rel, buf, itemsz, hitem);
itup_blkno = BufferGetBlockNumber(buf);
/* 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);
/* 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);
/* Create the return data structure */
res = (InsertIndexResult) palloc(sizeof(InsertIndexResultData));
ItemPointerSet(&(res->pointerData), itup_blkno, itup_off);
return res;
}
/*
* _hash_metapinit() -- Initialize the metadata page of a hash index,
* the initial buckets, and the initial bitmap page.
*
* The initial number of buckets is dependent on num_tuples, an estimate
* of the number of tuples to be loaded into the index initially. The
* chosen number of buckets is returned.
*
* We are fairly cavalier about locking here, since we know that no one else
* could be accessing this index. In particular the rule about not holding
* multiple buffer locks is ignored.
*/
uint32
_hash_metapinit(Relation rel, double num_tuples, ForkNumber forkNum)
{
HashMetaPage metap;
HashPageOpaque pageopaque;
Buffer metabuf;
Buffer buf;
Page pg;
int32 data_width;
int32 item_width;
int32 ffactor;
double dnumbuckets;
uint32 num_buckets;
uint32 log2_num_buckets;
uint32 i;
/* safety check */
if (RelationGetNumberOfBlocksInFork(rel, forkNum) != 0)
elog(ERROR, "cannot initialize non-empty hash index \"%s\"",
RelationGetRelationName(rel));
/*
* Determine the target fill factor (in tuples per bucket) for this index.
* The idea is to make the fill factor correspond to pages about as full
* as the user-settable fillfactor parameter says. We can compute it
* exactly since the index datatype (i.e. uint32 hash key) is fixed-width.
*/
data_width = sizeof(uint32);
item_width = MAXALIGN(sizeof(IndexTupleData)) + MAXALIGN(data_width) +
sizeof(ItemIdData); /* include the line pointer */
ffactor = RelationGetTargetPageUsage(rel, HASH_DEFAULT_FILLFACTOR) / item_width;
/* keep to a sane range */
if (ffactor < 10)
ffactor = 10;
/*
* Choose the number of initial bucket pages to match the fill factor
* given the estimated number of tuples. We round up the result to the
* next power of 2, however, and always force at least 2 bucket pages. The
* upper limit is determined by considerations explained in
* _hash_expandtable().
*/
dnumbuckets = num_tuples / ffactor;
if (dnumbuckets <= 2.0)
num_buckets = 2;
else if (dnumbuckets >= (double) 0x40000000)
num_buckets = 0x40000000;
else
num_buckets = ((uint32) 1) << _hash_log2((uint32) dnumbuckets);
log2_num_buckets = _hash_log2(num_buckets);
Assert(num_buckets == (((uint32) 1) << log2_num_buckets));
Assert(log2_num_buckets < HASH_MAX_SPLITPOINTS);
/*
* We initialize the metapage, the first N bucket pages, and the first
* bitmap page in sequence, using _hash_getnewbuf to cause smgrextend()
* calls to occur. This ensures that the smgr level has the right idea of
* the physical index length.
*/
metabuf = _hash_getnewbuf(rel, HASH_METAPAGE, forkNum);
pg = BufferGetPage(metabuf);
pageopaque = (HashPageOpaque) PageGetSpecialPointer(pg);
pageopaque->hasho_prevblkno = InvalidBlockNumber;
pageopaque->hasho_nextblkno = InvalidBlockNumber;
pageopaque->hasho_bucket = -1;
pageopaque->hasho_flag = LH_META_PAGE;
pageopaque->hasho_page_id = HASHO_PAGE_ID;
metap = HashPageGetMeta(pg);
metap->hashm_magic = HASH_MAGIC;
metap->hashm_version = HASH_VERSION;
metap->hashm_ntuples = 0;
metap->hashm_nmaps = 0;
metap->hashm_ffactor = ffactor;
metap->hashm_bsize = HashGetMaxBitmapSize(pg);
/* find largest bitmap array size that will fit in page size */
for (i = _hash_log2(metap->hashm_bsize); i > 0; --i)
{
if ((1 << i) <= metap->hashm_bsize)
break;
}
Assert(i > 0);
metap->hashm_bmsize = 1 << i;
metap->hashm_bmshift = i + BYTE_TO_BIT;
Assert((1 << BMPG_SHIFT(metap)) == (BMPG_MASK(metap) + 1));
/*
* Label the index with its primary hash support function's OID. This is
* pretty useless for normal operation (in fact, hashm_procid is not used
* anywhere), but it might be handy for forensic purposes so we keep it.
*/
metap->hashm_procid = index_getprocid(rel, 1, HASHPROC);
/*
* We initialize the index with N buckets, 0 .. N-1, occupying physical
* blocks 1 to N. The first freespace bitmap page is in block N+1. Since
* N is a power of 2, we can set the masks this way:
*/
metap->hashm_maxbucket = metap->hashm_lowmask = num_buckets - 1;
metap->hashm_highmask = (num_buckets << 1) - 1;
MemSet(metap->hashm_spares, 0, sizeof(metap->hashm_spares));
MemSet(metap->hashm_mapp, 0, sizeof(metap->hashm_mapp));
/* Set up mapping for one spare page after the initial splitpoints */
metap->hashm_spares[log2_num_buckets] = 1;
metap->hashm_ovflpoint = log2_num_buckets;
metap->hashm_firstfree = 0;
/*
* Release buffer lock on the metapage while we initialize buckets.
* Otherwise, we'll be in interrupt holdoff and the CHECK_FOR_INTERRUPTS
* won't accomplish anything. It's a bad idea to hold buffer locks for
* long intervals in any case, since that can block the bgwriter.
*/
_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);
/*
* Initialize the first N buckets
*/
for (i = 0; i < num_buckets; i++)
{
/* Allow interrupts, in case N is huge */
CHECK_FOR_INTERRUPTS();
buf = _hash_getnewbuf(rel, BUCKET_TO_BLKNO(metap, i), forkNum);
pg = BufferGetPage(buf);
pageopaque = (HashPageOpaque) PageGetSpecialPointer(pg);
pageopaque->hasho_prevblkno = InvalidBlockNumber;
pageopaque->hasho_nextblkno = InvalidBlockNumber;
pageopaque->hasho_bucket = i;
pageopaque->hasho_flag = LH_BUCKET_PAGE;
pageopaque->hasho_page_id = HASHO_PAGE_ID;
_hash_wrtbuf(rel, buf);
}
/* Now reacquire buffer lock on metapage */
_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);
/*
* Initialize first bitmap page
*/
_hash_initbitmap(rel, metap, num_buckets + 1, forkNum);
/* all done */
_hash_wrtbuf(rel, metabuf);
return num_buckets;
}
/*
* hashgetmulti() -- get multiple tuples at once
*
* This is a somewhat generic implementation: it avoids lock reacquisition
* overhead, but there's no smarts about picking especially good stopping
* points such as index page boundaries.
*/
Datum
hashgetmulti(PG_FUNCTION_ARGS)
{
IndexScanDesc scan = (IndexScanDesc) PG_GETARG_POINTER(0);
ItemPointer tids = (ItemPointer) PG_GETARG_POINTER(1);
int32 max_tids = PG_GETARG_INT32(2);
int32 *returned_tids = (int32 *) PG_GETARG_POINTER(3);
HashScanOpaque so = (HashScanOpaque) scan->opaque;
Relation rel = scan->indexRelation;
bool res = true;
int32 ntids = 0;
/*
* We hold pin but not lock on current buffer while outside the hash AM.
* Reacquire the read lock here.
*/
if (BufferIsValid(so->hashso_curbuf))
_hash_chgbufaccess(rel, so->hashso_curbuf, HASH_NOLOCK, HASH_READ);
while (ntids < max_tids)
{
/*
* Start scan, or advance to next tuple.
*/
if (ItemPointerIsValid(&(scan->currentItemData)))
res = _hash_next(scan, ForwardScanDirection);
else
res = _hash_first(scan, ForwardScanDirection);
/*
* Skip killed tuples if asked to.
*/
if (scan->ignore_killed_tuples)
{
while (res)
{
Page page;
OffsetNumber offnum;
offnum = ItemPointerGetOffsetNumber(&(scan->currentItemData));
page = BufferGetPage(so->hashso_curbuf);
if (!ItemIdDeleted(PageGetItemId(page, offnum)))
break;
res = _hash_next(scan, ForwardScanDirection);
}
}
if (!res)
break;
/* Save tuple ID, and continue scanning */
tids[ntids] = scan->xs_ctup.t_self;
ntids++;
}
/* Release read lock on current buffer, but keep it pinned */
if (BufferIsValid(so->hashso_curbuf))
_hash_chgbufaccess(rel, so->hashso_curbuf, HASH_READ, HASH_NOLOCK);
*returned_tids = ntids;
PG_RETURN_BOOL(res);
}
/*
* hashgettuple() -- Get the next tuple in the scan.
*/
Datum
hashgettuple(PG_FUNCTION_ARGS)
{
IndexScanDesc scan = (IndexScanDesc) PG_GETARG_POINTER(0);
ScanDirection dir = (ScanDirection) PG_GETARG_INT32(1);
HashScanOpaque so = (HashScanOpaque) scan->opaque;
Relation rel = scan->indexRelation;
Page page;
OffsetNumber offnum;
bool res;
/*
* We hold pin but not lock on current buffer while outside the hash AM.
* Reacquire the read lock here.
*/
if (BufferIsValid(so->hashso_curbuf))
_hash_chgbufaccess(rel, so->hashso_curbuf, HASH_NOLOCK, HASH_READ);
/*
* If we've already initialized this scan, we can just advance it in the
* appropriate direction. If we haven't done so yet, we call a routine to
* get the first item in the scan.
*/
if (ItemPointerIsValid(&(scan->currentItemData)))
{
/*
* Check to see if we should kill the previously-fetched tuple.
*/
if (scan->kill_prior_tuple)
{
/*
* Yes, so mark it by setting the LP_DELETE bit in the item flags.
*/
offnum = ItemPointerGetOffsetNumber(&(scan->currentItemData));
page = BufferGetPage(so->hashso_curbuf);
PageGetItemId(page, offnum)->lp_flags |= LP_DELETE;
/*
* Since this can be redone later if needed, it's treated the same
* as a commit-hint-bit status update for heap tuples: we mark the
* buffer dirty but don't make a WAL log entry.
*/
SetBufferCommitInfoNeedsSave(so->hashso_curbuf);
}
/*
* Now continue the scan.
*/
res = _hash_next(scan, dir);
}
else
res = _hash_first(scan, dir);
/*
* Skip killed tuples if asked to.
*/
if (scan->ignore_killed_tuples)
{
while (res)
{
offnum = ItemPointerGetOffsetNumber(&(scan->currentItemData));
page = BufferGetPage(so->hashso_curbuf);
if (!ItemIdDeleted(PageGetItemId(page, offnum)))
break;
res = _hash_next(scan, dir);
}
}
/* Release read lock on current buffer, but keep it pinned */
if (BufferIsValid(so->hashso_curbuf))
_hash_chgbufaccess(rel, so->hashso_curbuf, HASH_READ, HASH_NOLOCK);
PG_RETURN_BOOL(res);
}
/*
* _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;
BlockNumber oldblkno = InvalidBuffer;
bool retry = false;
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;
/* Read the metapage */
metabuf = _hash_getbuf(rel, HASH_METAPAGE, HASH_READ, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
/*
* 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);
/* Release metapage lock, but keep pin. */
_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);
/*
* If the previous iteration of this loop locked what is still the
* correct target bucket, we are done. Otherwise, drop any old lock
* and lock what now appears to be the correct bucket.
*/
if (retry)
{
if (oldblkno == blkno)
break;
_hash_droplock(rel, oldblkno, HASH_SHARE);
}
_hash_getlock(rel, blkno, HASH_SHARE);
/*
* 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;
}
/* done with the metapage */
_hash_dropbuf(rel, metabuf);
/* 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;
}
/*
* _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;
Buffer metabuf;
HashMetaPage metap;
BlockNumber blkno;
Page page;
HashPageOpaque pageopaque;
Size itemsz;
bool do_expand;
uint32 hashkey;
Bucket bucket;
/*
* 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 */
/*
* 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));
/*
* 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 %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);
/* 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);
}
/*
* _hash_addovflpage
*
* Add an overflow page to the bucket whose last page is pointed to by 'buf'.
*
* On entry, the caller must hold a pin but no lock on 'buf'. The pin is
* dropped before exiting (we assume the caller is not interested in 'buf'
* anymore). The returned overflow page will be pinned and write-locked;
* it is guaranteed to be empty.
*
* The caller must hold a pin, but no lock, on the metapage buffer.
* That buffer is returned in the same state.
*
* The caller must hold at least share lock on the bucket, to ensure that
* no one else tries to compact the bucket meanwhile. This guarantees that
* 'buf' won't stop being part of the bucket while it's unlocked.
*
* NB: since this could be executed concurrently by multiple processes,
* one should not assume that the returned overflow page will be the
* immediate successor of the originally passed 'buf'. Additional overflow
* pages might have been added to the bucket chain in between.
*/
Buffer
_hash_addovflpage(Relation rel, Buffer metabuf, Buffer buf)
{
Buffer ovflbuf;
Page page;
Page ovflpage;
HashPageOpaque pageopaque;
HashPageOpaque ovflopaque;
/*CS3223*/ /*declare variables*/
HashMetaPage metap;
Bucket bucket;
int i;
int index;
int bitIndexInElement;
uint32 ovflElement;
uint32 *tempPointer;
/* allocate and lock an empty overflow page */
ovflbuf = _hash_getovflpage(rel, metabuf);
/*CS3223*/
metap = HashPageGetMeta(BufferGetPage(metabuf));
/* find bucket number of primary page */
page = BufferGetPage(buf);
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
bucket = pageopaque -> hasho_bucket;
/*
* Write-lock the tail page. It is okay to hold two buffer locks here
* since there cannot be anyone else contending for access to ovflbuf.
*/
_hash_chgbufaccess(rel, buf, HASH_NOLOCK, HASH_WRITE);
/* probably redundant... */
_hash_checkpage(rel, buf, LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
/* loop to find current tail page, in case someone else inserted too */
//for (;;)
/*CS3223*/
while (i>=0)
{
BlockNumber nextblkno;
page = BufferGetPage(buf);
pageopaque = (HashPageOpaque) PageGetSpecialPointer(page);
nextblkno = pageopaque->hasho_nextblkno;
if (!BlockNumberIsValid(nextblkno))
break;
/* we assume we do not need to write the unmodified page */
_hash_relbuf(rel, buf);
buf = _hash_getbuf(rel, nextblkno, HASH_WRITE, LH_OVERFLOW_PAGE);
/*CS3223*/
i++;
}
/* now that we have correct backlink, initialize new overflow page */
ovflpage = BufferGetPage(ovflbuf);
ovflopaque = (HashPageOpaque) PageGetSpecialPointer(ovflpage);
ovflopaque->hasho_prevblkno = BufferGetBlockNumber(buf);
ovflopaque->hasho_nextblkno = InvalidBlockNumber;
ovflopaque->hasho_bucket = pageopaque->hasho_bucket;
ovflopaque->hasho_flag = LH_OVERFLOW_PAGE;
ovflopaque->hasho_page_id = HASHO_PAGE_ID;
MarkBufferDirty(ovflbuf);
/* logically chain overflow page to previous page */
pageopaque->hasho_nextblkno = BufferGetBlockNumber(ovflbuf);
_hash_wrtbuf(rel, buf);
/*CS3223*/
/* if length of the bucket chain is 1, only one ovflpage was added, which means the bucket was not split before */
if (i == 1) {
index = bucket / 32;
bitIndexInElement = bucket % 32;
ovflElement = (uint32)metap->ovflBkts[index];
ovflElement = ovflElement | (1 << bitIndexInElement);
tempPointer = &(metap->ovflBkts[index]);
*tempPointer = ovflElement;
}
return ovflbuf;
}
/*
* Attempt to expand the hash table by creating one new bucket.
*
* This will silently do nothing if it cannot get the needed locks.
*
* The caller should hold no locks on the hash index.
*
* The caller must hold a pin, but no lock, on the metapage buffer.
* The buffer is returned in the same state.
*/
void
_hash_expandtable(Relation rel, Buffer metabuf)
{
HashMetaPage metap;
Bucket old_bucket;
Bucket new_bucket;
uint32 spare_ndx;
BlockNumber start_oblkno;
BlockNumber start_nblkno;
uint32 maxbucket;
uint32 highmask;
uint32 lowmask;
/*
* Obtain the page-zero lock to assert the right to begin a split (see
* README).
*
* Note: deadlock should be impossible here. Our own backend could only be
* holding bucket sharelocks due to stopped indexscans; those will not
* block other holders of the page-zero lock, who are only interested in
* acquiring bucket sharelocks themselves. Exclusive bucket locks are
* only taken here and in hashbulkdelete, and neither of these operations
* needs any additional locks to complete. (If, due to some flaw in this
* reasoning, we manage to deadlock anyway, it's okay to error out; the
* index will be left in a consistent state.)
*/
_hash_getlock(rel, 0, HASH_EXCLUSIVE);
/* Write-lock the meta page */
_hash_chgbufaccess(rel, metabuf, HASH_NOLOCK, HASH_WRITE);
_hash_checkpage(rel, metabuf, LH_META_PAGE);
metap = HashPageGetMeta(BufferGetPage(metabuf));
/*
* Check to see if split is still needed; someone else might have already
* done one while we waited for the lock.
*
* Make sure this stays in sync with _hash_doinsert()
*/
if (metap->hashm_ntuples <=
(double) metap->hashm_ffactor * (metap->hashm_maxbucket + 1))
goto fail;
/*
* Can't split anymore if maxbucket has reached its maximum possible
* value.
*
* Ideally we'd allow bucket numbers up to UINT_MAX-1 (no higher because
* the calculation maxbucket+1 mustn't overflow). Currently we restrict
* to half that because of overflow looping in _hash_log2() and
* insufficient space in hashm_spares[]. It's moot anyway because an
* index with 2^32 buckets would certainly overflow BlockNumber and hence
* _hash_alloc_buckets() would fail, but if we supported buckets smaller
* than a disk block then this would be an independent constraint.
*
* If you change this, see also the maximum initial number of buckets in
* _hash_metapinit().
*/
if (metap->hashm_maxbucket >= (uint32) 0x7FFFFFFE)
goto fail;
/*
* Determine which bucket is to be split, and attempt to lock the old
* bucket. If we can't get the lock, give up.
*
* The lock protects us against other backends, but not against our own
* backend. Must check for active scans separately.
*/
new_bucket = metap->hashm_maxbucket + 1;
old_bucket = (new_bucket & metap->hashm_lowmask);
start_oblkno = BUCKET_TO_BLKNO(metap, old_bucket);
if (_hash_has_active_scan(rel, old_bucket))
goto fail;
if (!_hash_try_getlock(rel, start_oblkno, HASH_EXCLUSIVE))
goto fail;
/*
* Likewise lock the new bucket (should never fail).
*
* Note: it is safe to compute the new bucket's blkno here, even though we
* may still need to update the BUCKET_TO_BLKNO mapping. This is because
* the current value of hashm_spares[hashm_ovflpoint] correctly shows
* where we are going to put a new splitpoint's worth of buckets.
*/
start_nblkno = BUCKET_TO_BLKNO(metap, new_bucket);
if (_hash_has_active_scan(rel, new_bucket))
elog(ERROR, "scan in progress on supposedly new bucket");
if (!_hash_try_getlock(rel, start_nblkno, HASH_EXCLUSIVE))
elog(ERROR, "could not get lock on supposedly new bucket");
/*
* If the split point is increasing (hashm_maxbucket's log base 2
* increases), we need to allocate a new batch of bucket pages.
*/
spare_ndx = _hash_log2(new_bucket + 1);
if (spare_ndx > metap->hashm_ovflpoint)
{
Assert(spare_ndx == metap->hashm_ovflpoint + 1);
/*
* The number of buckets in the new splitpoint is equal to the total
* number already in existence, i.e. new_bucket. Currently this maps
* one-to-one to blocks required, but someday we may need a more
* complicated calculation here.
*/
if (!_hash_alloc_buckets(rel, start_nblkno, new_bucket))
{
/* can't split due to BlockNumber overflow */
_hash_droplock(rel, start_oblkno, HASH_EXCLUSIVE);
_hash_droplock(rel, start_nblkno, HASH_EXCLUSIVE);
goto fail;
}
}
/*
* Okay to proceed with split. Update the metapage bucket mapping info.
*
* Since we are scribbling on the metapage data right in the shared
* buffer, any failure in this next little bit leaves us with a big
* problem: the metapage is effectively corrupt but could get written back
* to disk. We don't really expect any failure, but just to be sure,
* establish a critical section.
*/
START_CRIT_SECTION();
metap->hashm_maxbucket = new_bucket;
if (new_bucket > metap->hashm_highmask)
{
/* Starting a new doubling */
metap->hashm_lowmask = metap->hashm_highmask;
metap->hashm_highmask = new_bucket | metap->hashm_lowmask;
}
/*
* If the split point is increasing (hashm_maxbucket's log base 2
* increases), we need to adjust the hashm_spares[] array and
* hashm_ovflpoint so that future overflow pages will be created beyond
* this new batch of bucket pages.
*/
if (spare_ndx > metap->hashm_ovflpoint)
{
metap->hashm_spares[spare_ndx] = metap->hashm_spares[metap->hashm_ovflpoint];
metap->hashm_ovflpoint = spare_ndx;
}
/* Done mucking with metapage */
END_CRIT_SECTION();
/*
* Copy bucket mapping info now; this saves re-accessing the meta page
* inside _hash_splitbucket's inner loop. Note that once we drop the
* split lock, other splits could begin, so these values might be out of
* date before _hash_splitbucket finishes. That's okay, since all it
* needs is to tell which of these two buckets to map hashkeys into.
*/
maxbucket = metap->hashm_maxbucket;
highmask = metap->hashm_highmask;
lowmask = metap->hashm_lowmask;
/* Write out the metapage and drop lock, but keep pin */
_hash_chgbufaccess(rel, metabuf, HASH_WRITE, HASH_NOLOCK);
/* Release split lock; okay for other splits to occur now */
_hash_droplock(rel, 0, HASH_EXCLUSIVE);
/* Relocate records to the new bucket */
_hash_splitbucket(rel, metabuf, old_bucket, new_bucket,
start_oblkno, start_nblkno,
maxbucket, highmask, lowmask);
/* Release bucket locks, allowing others to access them */
_hash_droplock(rel, start_oblkno, HASH_EXCLUSIVE);
_hash_droplock(rel, start_nblkno, HASH_EXCLUSIVE);
return;
/* Here if decide not to split or fail to acquire old bucket lock */
fail:
/* We didn't write the metapage, so just drop lock */
_hash_chgbufaccess(rel, metabuf, HASH_READ, HASH_NOLOCK);
/* Release split lock */
_hash_droplock(rel, 0, HASH_EXCLUSIVE);
}
/*
* hashgettuple() -- Get the next tuple in the scan.
*/
bool
hashgettuple(IndexScanDesc scan, ScanDirection dir)
{
HashScanOpaque so = (HashScanOpaque) scan->opaque;
Relation rel = scan->indexRelation;
Buffer buf;
Page page;
OffsetNumber offnum;
ItemPointer current;
bool res;
/* Hash indexes are always lossy since we store only the hash code */
scan->xs_recheck = true;
/*
* We hold pin but not lock on current buffer while outside the hash AM.
* Reacquire the read lock here.
*/
if (BufferIsValid(so->hashso_curbuf))
_hash_chgbufaccess(rel, so->hashso_curbuf, HASH_NOLOCK, HASH_READ);
/*
* If we've already initialized this scan, we can just advance it in the
* appropriate direction. If we haven't done so yet, we call a routine to
* get the first item in the scan.
*/
current = &(so->hashso_curpos);
if (ItemPointerIsValid(current))
{
/*
* An insertion into the current index page could have happened while
* we didn't have read lock on it. Re-find our position by looking
* for the TID we previously returned. (Because we hold share lock on
* the bucket, no deletions or splits could have occurred; therefore
* we can expect that the TID still exists in the current index page,
* at an offset >= where we were.)
*/
OffsetNumber maxoffnum;
buf = so->hashso_curbuf;
Assert(BufferIsValid(buf));
page = BufferGetPage(buf);
TestForOldSnapshot(scan->xs_snapshot, rel, page);
maxoffnum = PageGetMaxOffsetNumber(page);
for (offnum = ItemPointerGetOffsetNumber(current);
offnum <= maxoffnum;
offnum = OffsetNumberNext(offnum))
{
IndexTuple itup;
itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, offnum));
if (ItemPointerEquals(&(so->hashso_heappos), &(itup->t_tid)))
break;
}
if (offnum > maxoffnum)
elog(ERROR, "failed to re-find scan position within index \"%s\"",
RelationGetRelationName(rel));
ItemPointerSetOffsetNumber(current, offnum);
/*
* Check to see if we should kill the previously-fetched tuple.
*/
if (scan->kill_prior_tuple)
{
/*
* Yes, so mark it by setting the LP_DEAD state in the item flags.
*/
ItemIdMarkDead(PageGetItemId(page, offnum));
/*
* Since this can be redone later if needed, mark as a hint.
*/
MarkBufferDirtyHint(buf, true);
}
/*
* Now continue the scan.
*/
res = _hash_next(scan, dir);
}
else
res = _hash_first(scan, dir);
/*
* Skip killed tuples if asked to.
*/
if (scan->ignore_killed_tuples)
{
while (res)
{
offnum = ItemPointerGetOffsetNumber(current);
page = BufferGetPage(so->hashso_curbuf);
if (!ItemIdIsDead(PageGetItemId(page, offnum)))
break;
res = _hash_next(scan, dir);
}
}
/* Release read lock on current buffer, but keep it pinned */
if (BufferIsValid(so->hashso_curbuf))
_hash_chgbufaccess(rel, so->hashso_curbuf, HASH_READ, HASH_NOLOCK);
/* Return current heap TID on success */
scan->xs_ctup.t_self = so->hashso_heappos;
return res;
}
/*
* _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);
}
