/*
* Form a non-leaf entry tuple by copying the key data from the given tuple,
* which can be either a leaf or non-leaf entry tuple.
*
* Any posting list in the source tuple is not copied. The specified child
* block number is inserted into t_tid.
*/
static IndexTuple
GinFormInteriorTuple(IndexTuple itup, Page page, BlockNumber childblk)
{
IndexTuple nitup;
if (GinPageIsLeaf(page) && !GinIsPostingTree(itup))
{
/* Tuple contains a posting list, just copy stuff before that */
uint32 origsize = GinGetPostingOffset(itup);
origsize = MAXALIGN(origsize);
nitup = (IndexTuple) palloc(origsize);
memcpy(nitup, itup, origsize);
/* ... be sure to fix the size header field ... */
nitup->t_info &= ~INDEX_SIZE_MASK;
nitup->t_info |= origsize;
}
else
{
/* Copy the tuple as-is */
nitup = (IndexTuple) palloc(IndexTupleSize(itup));
memcpy(nitup, itup, IndexTupleSize(itup));
}
/* Now insert the correct downlink */
GinSetDownlink(nitup, childblk);
return nitup;
}
static bool
entryIsEnoughSpace(GinBtree btree, Buffer buf, OffsetNumber off,
GinBtreeEntryInsertData *insertData)
{
Size releasedsz = 0;
Size addedsz;
Page page = BufferGetPage(buf);
Assert(insertData->entry);
Assert(!GinPageIsData(page));
if (insertData->isDelete)
{
IndexTuple itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, off));
releasedsz = MAXALIGN(IndexTupleSize(itup)) + sizeof(ItemIdData);
}
addedsz = MAXALIGN(IndexTupleSize(insertData->entry)) + sizeof(ItemIdData);
if (PageGetFreeSpace(page) + releasedsz >= addedsz)
return true;
return false;
}
/*
* Place tuple on page and fills WAL record
*/
static void
entryPlaceToPage(GinBtree btree, Buffer buf, OffsetNumber off, XLogRecData **prdata)
{
Page page = BufferGetPage(buf);
static XLogRecData rdata[3];
OffsetNumber placed;
static ginxlogInsert data;
int cnt = 0;
*prdata = rdata;
data.updateBlkno = entryPreparePage(btree, page, off);
placed = PageAddItem(page, (Item) btree->entry, IndexTupleSize(btree->entry), off, false, false);
if (placed != off)
elog(ERROR, "failed to add item to index page in \"%s\"",
RelationGetRelationName(btree->index));
data.node = btree->index->rd_node;
data.blkno = BufferGetBlockNumber(buf);
data.offset = off;
data.nitem = 1;
data.isDelete = btree->isDelete;
data.isData = false;
data.isLeaf = GinPageIsLeaf(page) ? TRUE : FALSE;
/*
* Prevent full page write if child's split occurs. That is needed to
* remove incomplete splits while replaying WAL
*
* data.updateBlkno contains new block number (of newly created right
* page) for recently splited page.
*/
if (data.updateBlkno == InvalidBlockNumber)
{
rdata[0].buffer = buf;
rdata[0].buffer_std = TRUE;
rdata[0].data = NULL;
rdata[0].len = 0;
rdata[0].next = &rdata[1];
cnt++;
}
rdata[cnt].buffer = InvalidBuffer;
rdata[cnt].data = (char *) &data;
rdata[cnt].len = sizeof(ginxlogInsert);
rdata[cnt].next = &rdata[cnt + 1];
cnt++;
rdata[cnt].buffer = InvalidBuffer;
rdata[cnt].data = (char *) btree->entry;
rdata[cnt].len = IndexTupleSize(btree->entry);
rdata[cnt].next = NULL;
btree->entry = NULL;
}
/*
* Place tuple on page and fills WAL record
*
* If the tuple doesn't fit, returns false without modifying the page.
*
* On insertion to an internal node, in addition to inserting the given item,
* the downlink of the existing item at 'off' is updated to point to
* 'updateblkno'.
*/
static GinPlaceToPageRC
entryPlaceToPage(GinBtree btree, Buffer buf, GinBtreeStack *stack,
void *insertPayload, BlockNumber updateblkno,
XLogRecData **prdata, Page *newlpage, Page *newrpage)
{
GinBtreeEntryInsertData *insertData = insertPayload;
Page page = BufferGetPage(buf);
OffsetNumber off = stack->off;
OffsetNumber placed;
int cnt = 0;
/* these must be static so they can be returned to caller */
static XLogRecData rdata[3];
static ginxlogInsertEntry data;
/* quick exit if it doesn't fit */
if (!entryIsEnoughSpace(btree, buf, off, insertData))
{
entrySplitPage(btree, buf, stack, insertPayload, updateblkno,
prdata, newlpage, newrpage);
return SPLIT;
}
START_CRIT_SECTION();
*prdata = rdata;
entryPreparePage(btree, page, off, insertData, updateblkno);
placed = PageAddItem(page,
(Item) insertData->entry,
IndexTupleSize(insertData->entry),
off, false, false);
if (placed != off)
elog(ERROR, "failed to add item to index page in \"%s\"",
RelationGetRelationName(btree->index));
data.isDelete = insertData->isDelete;
data.offset = off;
rdata[cnt].buffer = buf;
rdata[cnt].buffer_std = true;
rdata[cnt].data = (char *) &data;
rdata[cnt].len = offsetof(ginxlogInsertEntry, tuple);
rdata[cnt].next = &rdata[cnt + 1];
cnt++;
rdata[cnt].buffer = buf;
rdata[cnt].buffer_std = true;
rdata[cnt].data = (char *) insertData->entry;
rdata[cnt].len = IndexTupleSize(insertData->entry);
rdata[cnt].next = NULL;
return INSERTED;
}
/*
* Place tuple on page and fills WAL record
*
* If the tuple doesn't fit, returns false without modifying the page.
*
* On insertion to an internal node, in addition to inserting the given item,
* the downlink of the existing item at 'off' is updated to point to
* 'updateblkno'.
*
* On INSERTED, registers the buffer as buffer ID 0, with data.
* On SPLIT, returns rdata that represents the split pages in *prdata.
*/
static GinPlaceToPageRC
entryPlaceToPage(GinBtree btree, Buffer buf, GinBtreeStack *stack,
void *insertPayload, BlockNumber updateblkno,
Page *newlpage, Page *newrpage)
{
GinBtreeEntryInsertData *insertData = insertPayload;
Page page = BufferGetPage(buf, NULL, NULL, BGP_NO_SNAPSHOT_TEST);
OffsetNumber off = stack->off;
OffsetNumber placed;
/* this must be static so it can be returned to caller. */
static ginxlogInsertEntry data;
/* quick exit if it doesn't fit */
if (!entryIsEnoughSpace(btree, buf, off, insertData))
{
entrySplitPage(btree, buf, stack, insertPayload, updateblkno,
newlpage, newrpage);
return SPLIT;
}
START_CRIT_SECTION();
entryPreparePage(btree, page, off, insertData, updateblkno);
placed = PageAddItem(page,
(Item) insertData->entry,
IndexTupleSize(insertData->entry),
off, false, false);
if (placed != off)
elog(ERROR, "failed to add item to index page in \"%s\"",
RelationGetRelationName(btree->index));
if (RelationNeedsWAL(btree->index))
{
data.isDelete = insertData->isDelete;
data.offset = off;
XLogBeginInsert();
XLogRegisterBuffer(0, buf, REGBUF_STANDARD);
XLogRegisterBufData(0, (char *) &data,
offsetof(ginxlogInsertEntry, tuple));
XLogRegisterBufData(0, (char *) insertData->entry,
IndexTupleSize(insertData->entry));
}
return INSERTED;
}
/*
* Inserts only one entry to the index, but it can add more than 1 ItemPointer.
*/
static void
ginEntryInsert(Relation index, GinState *ginstate, Datum value, ItemPointerData *items, uint32 nitem, bool isBuild)
{
GinBtreeData btree;
GinBtreeStack *stack;
IndexTuple itup;
Page page;
prepareEntryScan(&btree, index, value, ginstate);
stack = ginFindLeafPage(&btree, NULL);
page = BufferGetPage(stack->buffer);
if (btree.findItem(&btree, stack))
{
/* found entry */
itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, stack->off));
if (GinIsPostingTree(itup))
{
/* lock root of posting tree */
GinPostingTreeScan *gdi;
BlockNumber rootPostingTree = GinGetPostingTree(itup);
/* release all stack */
LockBuffer(stack->buffer, GIN_UNLOCK);
freeGinBtreeStack(stack);
/* insert into posting tree */
gdi = prepareScanPostingTree(index, rootPostingTree, FALSE);
gdi->btree.isBuild = isBuild;
insertItemPointer(gdi, items, nitem);
return;
}
itup = addItemPointersToTuple(index, ginstate, stack, itup, items, nitem, isBuild);
btree.isDelete = TRUE;
}
else
{
/* We suppose, that tuple can store at list one itempointer */
itup = GinFormTuple(ginstate, value, items, 1);
if (itup == NULL || IndexTupleSize(itup) >= GinMaxItemSize)
elog(ERROR, "huge tuple");
if (nitem > 1)
{
IndexTuple previtup = itup;
itup = addItemPointersToTuple(index, ginstate, stack, previtup, items + 1, nitem - 1, isBuild);
pfree(previtup);
}
}
btree.entry = itup;
ginInsertValue(&btree, stack);
pfree(itup);
}
/*
* Fills new root by rightest values from child.
* Also called from rumxlog, should not use btree
*/
void
rumEntryFillRoot(RumBtree btree, Buffer root, Buffer lbuf, Buffer rbuf,
Page page, Page lpage, Page rpage)
{
IndexTuple itup;
itup = rumPageGetLinkItup(btree, lbuf, lpage);
if (PageAddItem(page, (Item) itup, IndexTupleSize(itup), InvalidOffsetNumber, false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add item to index root page");
pfree(itup);
itup = rumPageGetLinkItup(btree, rbuf, rpage);
if (PageAddItem(page, (Item) itup, IndexTupleSize(itup), InvalidOffsetNumber, false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add item to index root page");
pfree(itup);
}
static void
decodePageSplitRecord(PageSplitRecord *decoded, XLogRecord *record)
{
char *begin = XLogRecGetData(record),
*ptr;
int j,
i = 0;
decoded->data = (gistxlogPageSplit *) begin;
decoded->page = (NewPage *) palloc(sizeof(NewPage) * decoded->data->npage);
ptr = begin + sizeof(gistxlogPageSplit);
for (i = 0; i < decoded->data->npage; i++)
{
Assert(ptr - begin < record->xl_len);
decoded->page[i].header = (gistxlogPage *) ptr;
ptr += sizeof(gistxlogPage);
decoded->page[i].itup = (IndexTuple *)
palloc(sizeof(IndexTuple) * decoded->page[i].header->num);
j = 0;
while (j < decoded->page[i].header->num)
{
Assert(ptr - begin < record->xl_len);
decoded->page[i].itup[j] = (IndexTuple) ptr;
ptr += IndexTupleSize((IndexTuple) ptr);
j++;
}
}
}
/*
* Write XLOG record describing a page update. The update can include any
* number of deletions and/or insertions of tuples on a single index page.
*
* If this update inserts a downlink for a split page, also record that
* the F_FOLLOW_RIGHT flag on the child page is cleared and NSN set.
*
* Note that both the todelete array and the tuples are marked as belonging
* to the target buffer; they need not be stored in XLOG if XLogInsert decides
* to log the whole buffer contents instead.
*/
XLogRecPtr
gistXLogUpdate(Buffer buffer,
OffsetNumber *todelete, int ntodelete,
IndexTuple *itup, int ituplen,
Buffer leftchildbuf)
{
gistxlogPageUpdate xlrec;
int i;
XLogRecPtr recptr;
xlrec.ntodelete = ntodelete;
xlrec.ntoinsert = ituplen;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, sizeof(gistxlogPageUpdate));
XLogRegisterBuffer(0, buffer, REGBUF_STANDARD);
XLogRegisterBufData(0, (char *) todelete, sizeof(OffsetNumber) * ntodelete);
/* new tuples */
for (i = 0; i < ituplen; i++)
XLogRegisterBufData(0, (char *) (itup[i]), IndexTupleSize(itup[i]));
/*
* Include a full page image of the child buf. (only necessary if a
* checkpoint happened since the child page was split)
*/
if (BufferIsValid(leftchildbuf))
XLogRegisterBuffer(1, leftchildbuf, REGBUF_STANDARD);
recptr = XLogInsert(RM_GIST_ID, XLOG_GIST_PAGE_UPDATE);
return recptr;
}
/*
* _hash_formitem -- construct a hash index entry
*/
HashItem
_hash_formitem(IndexTuple itup)
{
int nbytes_hitem;
HashItem hitem;
Size tuplen;
/* disallow nulls in hash keys */
if (IndexTupleHasNulls(itup))
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("hash indexes cannot contain null keys")));
/*
* make a copy of the index tuple (XXX do we still need to copy?)
*
* HashItemData used to have more fields than IndexTupleData, but no
* longer...
*/
tuplen = IndexTupleSize(itup);
nbytes_hitem = tuplen +
(sizeof(HashItemData) - sizeof(IndexTupleData));
hitem = (HashItem) palloc(nbytes_hitem);
memcpy((char *) &(hitem->hash_itup), (char *) itup, tuplen);
return hitem;
}
/*
* Fills new root by rightest values from child.
* Also called from ginxlog, should not use btree
*/
void
entryFillRoot(GinBtree btree, Buffer root, Buffer lbuf, Buffer rbuf)
{
Page page;
IndexTuple itup;
page = BufferGetPage(root);
itup = ginPageGetLinkItup(lbuf);
if (PageAddItem(page, (Item) itup, IndexTupleSize(itup), InvalidOffsetNumber, false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add item to index root page");
itup = ginPageGetLinkItup(rbuf);
if (PageAddItem(page, (Item) itup, IndexTupleSize(itup), InvalidOffsetNumber, false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add item to index root page");
}
/*
* Fills new root by rightest values from child.
* Also called from ginxlog, should not use btree
*/
void
ginEntryFillRoot(GinBtree btree, Page root,
BlockNumber lblkno, Page lpage,
BlockNumber rblkno, Page rpage)
{
IndexTuple itup;
itup = GinFormInteriorTuple(getRightMostTuple(lpage), lpage, lblkno);
if (PageAddItem(root, (Item) itup, IndexTupleSize(itup), InvalidOffsetNumber, false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add item to index root page");
pfree(itup);
itup = GinFormInteriorTuple(getRightMostTuple(rpage), rpage, rblkno);
if (PageAddItem(root, (Item) itup, IndexTupleSize(itup), InvalidOffsetNumber, false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add item to index root page");
pfree(itup);
}
/*-------------------------------------------------------
* bt_page_print_tuples()
*
* Form a tuple describing index tuple at a given offset
* ------------------------------------------------------
*/
static Datum
bt_page_print_tuples(FuncCallContext *fctx, Page page, OffsetNumber offset)
{
char *values[6];
HeapTuple tuple;
ItemId id;
IndexTuple itup;
int j;
int off;
int dlen;
char *dump;
char *ptr;
id = PageGetItemId(page, offset);
if (!ItemIdIsValid(id))
elog(ERROR, "invalid ItemId");
itup = (IndexTuple) PageGetItem(page, id);
j = 0;
values[j++] = psprintf("%d", offset);
values[j++] = psprintf("(%u,%u)",
ItemPointerGetBlockNumberNoCheck(&itup->t_tid),
ItemPointerGetOffsetNumberNoCheck(&itup->t_tid));
values[j++] = psprintf("%d", (int) IndexTupleSize(itup));
values[j++] = psprintf("%c", IndexTupleHasNulls(itup) ? 't' : 'f');
values[j++] = psprintf("%c", IndexTupleHasVarwidths(itup) ? 't' : 'f');
ptr = (char *) itup + IndexInfoFindDataOffset(itup->t_info);
dlen = IndexTupleSize(itup) - IndexInfoFindDataOffset(itup->t_info);
dump = palloc0(dlen * 3 + 1);
values[j] = dump;
for (off = 0; off < dlen; off++)
{
if (off > 0)
*dump++ = ' ';
sprintf(dump, "%02x", *(ptr + off) & 0xff);
dump += 2;
}
tuple = BuildTupleFromCStrings(fctx->attinmeta, values);
return HeapTupleGetDatum(tuple);
}
/*
* Form a tuple for entry tree.
*
* On leaf pages, Index tuple has non-traditional layout. Tuple may contain
* posting list or root blocknumber of posting tree.
* Macros: GinIsPostingTree(itup) / GinSetPostingTree(itup, blkno)
* 1) Posting list
* - itup->t_info & INDEX_SIZE_MASK contains total size of tuple as usual
* - ItemPointerGetBlockNumber(&itup->t_tid) contains original
* size of tuple (without posting list).
* Macros: GinGetOrigSizePosting(itup) / GinSetOrigSizePosting(itup,n)
* - ItemPointerGetOffsetNumber(&itup->t_tid) contains number
* of elements in posting list (number of heap itempointers)
* Macros: GinGetNPosting(itup) / GinSetNPosting(itup,n)
* - After standard part of tuple there is a posting list, ie, array
* of heap itempointers
* Macros: GinGetPosting(itup)
* 2) Posting tree
* - itup->t_info & INDEX_SIZE_MASK contains size of tuple as usual
* - ItemPointerGetBlockNumber(&itup->t_tid) contains block number of
* root of posting tree
* - ItemPointerGetOffsetNumber(&itup->t_tid) contains magic number
* GIN_TREE_POSTING, which distinguishes this from posting-list case
*
* Attributes of an index tuple are different for single and multicolumn index.
* For single-column case, index tuple stores only value to be indexed.
* For multicolumn case, it stores two attributes: column number of value
* and value.
*/
IndexTuple
GinFormTuple(GinState *ginstate, OffsetNumber attnum, Datum key, ItemPointerData *ipd, uint32 nipd)
{
bool isnull[2] = {FALSE, FALSE};
IndexTuple itup;
if (ginstate->oneCol)
itup = index_form_tuple(ginstate->origTupdesc, &key, isnull);
else
{
Datum datums[2];
datums[0] = UInt16GetDatum(attnum);
datums[1] = key;
itup = index_form_tuple(ginstate->tupdesc[attnum - 1], datums, isnull);
}
GinSetOrigSizePosting(itup, IndexTupleSize(itup));
if (nipd > 0)
{
uint32 newsize = MAXALIGN(SHORTALIGN(IndexTupleSize(itup)) + sizeof(ItemPointerData) * nipd);
if (newsize >= INDEX_SIZE_MASK)
return NULL;
if (newsize > TOAST_INDEX_TARGET && nipd > 1)
return NULL;
itup = repalloc(itup, newsize);
/* set new size */
itup->t_info &= ~INDEX_SIZE_MASK;
itup->t_info |= newsize;
if (ipd)
memcpy(GinGetPosting(itup), ipd, sizeof(ItemPointerData) * nipd);
GinSetNPosting(itup, nipd);
}
else
{
GinSetNPosting(itup, 0);
}
return itup;
}
/*
* Check space for itup vector on page
*/
bool
gistnospace(Page page, IndexTuple *itvec, int len, OffsetNumber todelete, Size freespace)
{
unsigned int size = freespace,
deleted = 0;
int i;
for (i = 0; i < len; i++)
size += IndexTupleSize(itvec[i]) + sizeof(ItemIdData);
if (todelete != InvalidOffsetNumber)
{
IndexTuple itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, todelete));
deleted = IndexTupleSize(itup) + sizeof(ItemIdData);
}
return (PageGetFreeSpace(page) + deleted < size);
}
static void
ginRedoInsertListPage(XLogReaderState *record)
{
XLogRecPtr lsn = record->EndRecPtr;
ginxlogInsertListPage *data = (ginxlogInsertListPage *) XLogRecGetData(record);
Buffer buffer;
Page page;
OffsetNumber l,
off = FirstOffsetNumber;
int i,
tupsize;
char *payload;
IndexTuple tuples;
Size totaltupsize;
/* We always re-initialize the page. */
buffer = XLogInitBufferForRedo(record, 0);
page = BufferGetPage(buffer);
GinInitBuffer(buffer, GIN_LIST);
GinPageGetOpaque(page)->rightlink = data->rightlink;
if (data->rightlink == InvalidBlockNumber)
{
/* tail of sublist */
GinPageSetFullRow(page);
GinPageGetOpaque(page)->maxoff = 1;
}
else
{
GinPageGetOpaque(page)->maxoff = 0;
}
payload = XLogRecGetBlockData(record, 0, &totaltupsize);
tuples = (IndexTuple) payload;
for (i = 0; i < data->ntuples; i++)
{
tupsize = IndexTupleSize(tuples);
l = PageAddItem(page, (Item) tuples, tupsize, off, false, false);
if (l == InvalidOffsetNumber)
elog(ERROR, "failed to add item to index page");
tuples = (IndexTuple) (((char *) tuples) + tupsize);
off++;
}
Assert((char *) tuples == payload + totaltupsize);
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
UnlockReleaseBuffer(buffer);
}
/*
* Check space for itup vector on page
*/
static int
gistnospace(Page page, IndexTuple *itvec, int len)
{
unsigned int size = 0;
int i;
for (i = 0; i < len; i++)
size += IndexTupleSize(itvec[i]) + sizeof(ItemIdData);
return (PageGetFreeSpace(page) < size);
}
/*
* Returns new tuple with copied value from source tuple.
* New tuple will not store posting list
*/
static IndexTuple
copyIndexTuple(IndexTuple itup, Page page)
{
IndexTuple nitup;
if (GinPageIsLeaf(page) && !GinIsPostingTree(itup))
{
nitup = (IndexTuple) palloc(MAXALIGN(GinGetOrigSizePosting(itup)));
memcpy(nitup, itup, GinGetOrigSizePosting(itup));
nitup->t_info &= ~INDEX_SIZE_MASK;
nitup->t_info |= GinGetOrigSizePosting(itup);
}
else
{
nitup = (IndexTuple) palloc(MAXALIGN(IndexTupleSize(itup)));
memcpy(nitup, itup, IndexTupleSize(itup));
}
return nitup;
}
/*
* Create a palloc'd copy of an index tuple.
*/
IndexTuple
CopyIndexTuple(IndexTuple source)
{
IndexTuple result;
Size size;
size = IndexTupleSize(source);
result = (IndexTuple) palloc(size);
memcpy(result, source, size);
return result;
}
static void
ginRedoSplitEntry(Page lpage, Page rpage, void *rdata)
{
ginxlogSplitEntry *data = (ginxlogSplitEntry *) rdata;
IndexTuple itup = (IndexTuple) ((char *) rdata + sizeof(ginxlogSplitEntry));
OffsetNumber i;
for (i = 0; i < data->separator; i++)
{
if (PageAddItem(lpage, (Item) itup, IndexTupleSize(itup), InvalidOffsetNumber, false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add item to gin index page");
itup = (IndexTuple) (((char *) itup) + MAXALIGN(IndexTupleSize(itup)));
}
for (i = data->separator; i < data->nitem; i++)
{
if (PageAddItem(rpage, (Item) itup, IndexTupleSize(itup), InvalidOffsetNumber, false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add item to gin index page");
itup = (IndexTuple) (((char *) itup) + MAXALIGN(IndexTupleSize(itup)));
}
}
static void
ginRedoInsertListPage(XLogRecPtr lsn, XLogRecord *record)
{
ginxlogInsertListPage *data = (ginxlogInsertListPage *) XLogRecGetData(record);
Buffer buffer;
Page page;
OffsetNumber l,
off = FirstOffsetNumber;
int i,
tupsize;
IndexTuple tuples = (IndexTuple) (XLogRecGetData(record) + sizeof(ginxlogInsertListPage));
/* If we have a full-page image, restore it and we're done */
if (record->xl_info & XLR_BKP_BLOCK(0))
{
(void) RestoreBackupBlock(lsn, record, 0, false, false);
return;
}
buffer = XLogReadBuffer(data->node, data->blkno, true);
Assert(BufferIsValid(buffer));
page = BufferGetPage(buffer);
GinInitBuffer(buffer, GIN_LIST);
GinPageGetOpaque(page)->rightlink = data->rightlink;
if (data->rightlink == InvalidBlockNumber)
{
/* tail of sublist */
GinPageSetFullRow(page);
GinPageGetOpaque(page)->maxoff = 1;
}
else
{
GinPageGetOpaque(page)->maxoff = 0;
}
for (i = 0; i < data->ntuples; i++)
{
tupsize = IndexTupleSize(tuples);
l = PageAddItem(page, (Item) tuples, tupsize, off, false, false);
if (l == InvalidOffsetNumber)
elog(ERROR, "failed to add item to index page");
tuples = (IndexTuple) (((char *) tuples) + tupsize);
}
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
UnlockReleaseBuffer(buffer);
}
static void
rtnewroot(Relation r, IndexTuple lt, IndexTuple rt)
{
Buffer b;
Page p;
b = ReadBuffer(r, P_ROOT);
RTInitBuffer(b, 0);
p = BufferGetPage(b);
if (PageAddItem(p, (Item) lt, IndexTupleSize(lt),
FirstOffsetNumber,
LP_USED) == InvalidOffsetNumber)
elog(ERROR, "failed to add index item to \"%s\"",
RelationGetRelationName(r));
if (PageAddItem(p, (Item) rt, IndexTupleSize(rt),
OffsetNumberNext(FirstOffsetNumber),
LP_USED) == InvalidOffsetNumber)
elog(ERROR, "failed to add index item to \"%s\"",
RelationGetRelationName(r));
WriteBuffer(b);
}
static bool
entryIsEnoughSpace(GinBtree btree, Buffer buf, OffsetNumber off)
{
Size itupsz = 0;
Page page = BufferGetPage(buf);
Assert(btree->entry);
Assert(!GinPageIsData(page));
if (btree->isDelete)
{
IndexTuple itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, off));
itupsz = MAXALIGN(IndexTupleSize(itup)) + sizeof(ItemIdData);
}
if (PageGetFreeSpace(page) + itupsz >= MAXALIGN(IndexTupleSize(btree->entry)) + sizeof(ItemIdData))
return true;
return false;
}
bool
gistfitpage(IndexTuple *itvec, int len)
{
int i;
Size size = 0;
for (i = 0; i < len; i++)
size += IndexTupleSize(itvec[i]) + sizeof(ItemIdData);
/* TODO: Consider fillfactor */
return (size <= GiSTPageSize);
}
/* FIXME This should do exactly the same checks of lp_flags as in heap.c */
uint32 check_index_tuple(Relation rel, PageHeader header, int block, int i, char *buffer) {
uint32 nerrs = 0;
int j, a, b, c, d;
IndexTuple itup = (IndexTuple)(buffer + header->pd_linp[i].lp_off);
/* FIXME This is used when checking overflowing attributes, but it's not clear what
* exactly this means / how it works. Needs a bit more investigation and maybe a review
* from soneone who really knows the b-tree implementation. */
int dlen = IndexTupleSize(itup) - IndexInfoFindDataOffset(itup->t_info);
ereport(DEBUG2,(errmsg("[%d:%d] off=%d len=%d tid=(%d,%d)", block, (i+1),
header->pd_linp[i].lp_off, header->pd_linp[i].lp_len,
BlockIdGetBlockNumber(&(itup->t_tid.ip_blkid)),
itup->t_tid.ip_posid )));
/* check intersection with other tuples */
/* [A,B] vs [C,D] */
a = header->pd_linp[i].lp_off;
b = header->pd_linp[i].lp_off + header->pd_linp[i].lp_len;
ereport(DEBUG2,(errmsg("[%d:%d] checking intersection with other tuples", block, (i+1))));
for (j = 0; j < i; j++) {
/* FIXME Skip UNUSED/REDIRECT/DEAD tuples */
if (! (header->pd_linp[i].lp_flags == LP_NORMAL)) {
ereport(DEBUG3,(errmsg("[%d:%d] skipped (not LP_NORMAL)", block, (j+1))));
continue;
}
c = header->pd_linp[j].lp_off;
d = header->pd_linp[j].lp_off + header->pd_linp[j].lp_len;
/* [A,C,B] or [A,D,B] or [C,A,D] or [C,B,D] */
if (((a < c) && (c < b)) || ((a < d) && (d < b)) ||
((c < a) && (a < d)) || ((c < b) && (b < d))) {
ereport(WARNING,(errmsg("[%d:%d] intersects with [%d:%d] (%d,%d) vs. (%d,%d)", block, (i+1), block, j, a, b, c, d)));
++nerrs;
}
}
/* check attributes only for tuples with (lp_flags==LP_NORMAL) */
if (header->pd_linp[i].lp_flags == LP_NORMAL) {
nerrs += check_index_tuple_attributes(rel, header, block, i + 1, buffer, dlen);
}
return nerrs;
}
static void
ginRedoInsertListPage(XLogRecPtr lsn, XLogRecord *record)
{
ginxlogInsertListPage *data = (ginxlogInsertListPage *) XLogRecGetData(record);
Buffer buffer;
Page page;
OffsetNumber l,
off = FirstOffsetNumber;
int i,
tupsize;
IndexTuple tuples = (IndexTuple) (XLogRecGetData(record) + sizeof(ginxlogInsertListPage));
/*
* Backup blocks are not used, we always re-initialize the page.
*/
Assert(!(record->xl_info & XLR_BKP_BLOCK_MASK));
buffer = XLogReadBuffer(data->node, data->blkno, true);
Assert(BufferIsValid(buffer));
page = BufferGetPage(buffer);
GinInitBuffer(buffer, GIN_LIST);
GinPageGetOpaque(page)->rightlink = data->rightlink;
if (data->rightlink == InvalidBlockNumber)
{
/* tail of sublist */
GinPageSetFullRow(page);
GinPageGetOpaque(page)->maxoff = 1;
}
else
{
GinPageGetOpaque(page)->maxoff = 0;
}
for (i = 0; i < data->ntuples; i++)
{
tupsize = IndexTupleSize(tuples);
l = PageAddItem(page, (Item) tuples, tupsize, off, false, false);
if (l == InvalidOffsetNumber)
elog(ERROR, "failed to add item to index page");
tuples = (IndexTuple) (((char *) tuples) + tupsize);
off++;
}
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
UnlockReleaseBuffer(buffer);
}
/*
* Workhouse routine for doing insertion into a GiST index. Note that
* this routine assumes it is invoked in a short-lived memory context,
* so it does not bother releasing palloc'd allocations.
*/
static void
gistdoinsert(Relation r, IndexTuple itup, Size freespace, GISTSTATE *giststate)
{
GISTInsertState state;
memset(&state, 0, sizeof(GISTInsertState));
state.itup = (IndexTuple *) palloc(sizeof(IndexTuple));
state.itup[0] = (IndexTuple) palloc(IndexTupleSize(itup));
memcpy(state.itup[0], itup, IndexTupleSize(itup));
state.ituplen = 1;
state.freespace = freespace;
state.r = r;
state.key = itup->t_tid;
state.needInsertComplete = true;
state.stack = (GISTInsertStack *) palloc0(sizeof(GISTInsertStack));
state.stack->blkno = GIST_ROOT_BLKNO;
gistfindleaf(&state, giststate);
gistmakedeal(&state, giststate);
}
IndexTupleData *
gistfillitupvec(IndexTuple *vec, int veclen, int *memlen)
{
char *ptr,
*ret;
int i;
*memlen = 0;
for (i = 0; i < veclen; i++)
*memlen += IndexTupleSize(vec[i]);
ptr = ret = palloc(*memlen);
for (i = 0; i < veclen; i++)
{
memcpy(ptr, vec[i], IndexTupleSize(vec[i]));
ptr += IndexTupleSize(vec[i]);
}
return (IndexTupleData *) ret;
}
/*
* Form a non-leaf entry tuple by copying the key data from the given tuple,
* which can be either a leaf or non-leaf entry tuple.
*
* Any posting list in the source tuple is not copied. The specified child
* block number is inserted into t_tid.
*/
static IndexTuple
RumFormInteriorTuple(RumBtree btree, IndexTuple itup, Page page,
BlockNumber childblk)
{
IndexTuple nitup;
RumNullCategory category;
if (RumPageIsLeaf(page) && !RumIsPostingTree(itup))
{
/* Tuple contains a posting list, just copy stuff before that */
uint32 origsize = RumGetPostingOffset(itup);
origsize = MAXALIGN(origsize);
nitup = (IndexTuple) palloc(origsize);
memcpy(nitup, itup, origsize);
/* ... be sure to fix the size header field ... */
nitup->t_info &= ~INDEX_SIZE_MASK;
nitup->t_info |= origsize;
}
else
{
/* Copy the tuple as-is */
nitup = (IndexTuple) palloc(IndexTupleSize(itup));
memcpy(nitup, itup, IndexTupleSize(itup));
}
/* Now insert the correct downlink */
RumSetDownlink(nitup, childblk);
rumtuple_get_key(btree->rumstate, itup, &category);
if (category != RUM_CAT_NORM_KEY)
{
Assert(IndexTupleHasNulls(itup));
nitup->t_info |= INDEX_NULL_MASK;
RumSetNullCategory(nitup, category);
}
return nitup;
}
SpGistInnerTuple
spgFormInnerTuple(SpGistState *state, bool hasPrefix, Datum prefix, int nNodes, IndexTuple *nodes)
{
SpGistInnerTuple tup;
unsigned int size;
int i;
char *ptr;
size = SGITHDRSZ;
if (hasPrefix)
size += getTypeLength(&state->attPrefixType, prefix);
for(i=0; i<nNodes; i++)
size += IndexTupleSize(nodes[i]);
Assert(size < 0xffff);
tup = palloc0(size);
tup->hasPrefix = !!hasPrefix;
tup->nNodes = nNodes;
tup->size = size;
if (tup->hasPrefix)
memcpyDatum(SGITDATAPTR(tup), &state->attPrefixType, prefix);
ptr= (char*)SGITNODEPTR(tup, state);
for(i=0; i<nNodes; i++)
{
IndexTuple node = nodes[i];
memcpy(ptr, node, IndexTupleSize(node));
ptr += IndexTupleSize(node);
}
return tup;
}