Datum
ltree_in(PG_FUNCTION_ARGS)
{
char *buf = (char *) PG_GETARG_POINTER(0);
char *ptr;
nodeitem *list,
*lptr;
int num = 0,
totallen = 0;
int state = LTPRS_WAITNAME;
ltree *result;
ltree_level *curlevel;
int charlen;
int pos = 0;
ptr = buf;
while (*ptr)
{
charlen = pg_mblen(ptr);
if (charlen == 1 && t_iseq(ptr, '.'))
num++;
ptr += charlen;
}
list = lptr = (nodeitem *) palloc(sizeof(nodeitem) * (num + 1));
ptr = buf;
while (*ptr)
{
charlen = pg_mblen(ptr);
if (state == LTPRS_WAITNAME)
{
if (ISALNUM(ptr))
{
lptr->start = ptr;
lptr->wlen = 0;
state = LTPRS_WAITDELIM;
}
else
UNCHAR;
}
else if (state == LTPRS_WAITDELIM)
{
if (charlen == 1 && t_iseq(ptr, '.'))
{
lptr->len = ptr - lptr->start;
if (lptr->wlen > 255)
ereport(ERROR,
(errcode(ERRCODE_NAME_TOO_LONG),
errmsg("name of level is too long"),
errdetail("Name length is %d, must "
"be < 256, in position %d.",
lptr->wlen, pos)));
totallen += MAXALIGN(lptr->len + LEVEL_HDRSIZE);
lptr++;
state = LTPRS_WAITNAME;
}
else if (!ISALNUM(ptr))
UNCHAR;
}
else
/* internal error */
elog(ERROR, "internal error in parser");
ptr += charlen;
lptr->wlen++;
pos++;
}
if (state == LTPRS_WAITDELIM)
{
lptr->len = ptr - lptr->start;
if (lptr->wlen > 255)
ereport(ERROR,
(errcode(ERRCODE_NAME_TOO_LONG),
errmsg("name of level is too long"),
errdetail("Name length is %d, must "
"be < 256, in position %d.",
lptr->wlen, pos)));
totallen += MAXALIGN(lptr->len + LEVEL_HDRSIZE);
lptr++;
}
else if (!(state == LTPRS_WAITNAME && lptr == list))
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("syntax error"),
errdetail("Unexpected end of line.")));
result = (ltree *) palloc0(LTREE_HDRSIZE + totallen);
SET_VARSIZE(result, LTREE_HDRSIZE + totallen);
result->numlevel = lptr - list;
curlevel = LTREE_FIRST(result);
lptr = list;
while (lptr - list < result->numlevel)
{
curlevel->len = (uint16) lptr->len;
memcpy(curlevel->name, lptr->start, lptr->len);
curlevel = LEVEL_NEXT(curlevel);
lptr++;
}
pfree(list);
PG_RETURN_POINTER(result);
}
/*
* Attempt to read an XLOG record into readRecordBuf.
*/
static bool
ReadRecord(void)
{
char *buffer;
XLogRecord *record;
XLogContRecord *contrecord;
uint32 len,
total_len;
int retries = 0;
restart:
while (logRecOff <= 0 || logRecOff > XLOG_BLCKSZ - SizeOfXLogRecord)
{
/* Need to advance to new page */
if (! readXLogPage())
return false;
logRecOff = XLogPageHeaderSize((XLogPageHeader) pageBuffer);
if ((((XLogPageHeader) pageBuffer)->xlp_info & ~XLP_LONG_HEADER) != 0)
{
printf("Unexpected page info flags %04X at offset %X\n",
((XLogPageHeader) pageBuffer)->xlp_info, logPageOff);
/* Check for a continuation record */
if (((XLogPageHeader) pageBuffer)->xlp_info & XLP_FIRST_IS_CONTRECORD)
{
printf("Skipping unexpected continuation record at offset %X\n",
logPageOff);
contrecord = (XLogContRecord *) (pageBuffer + logRecOff);
logRecOff += MAXALIGN(contrecord->xl_rem_len + SizeOfXLogContRecord);
}
}
}
curRecPtr.xlogid = logId;
curRecPtr.xrecoff = logSeg * XLogSegSize + logPageOff + logRecOff;
record = (XLogRecord *) (pageBuffer + logRecOff);
if (record->xl_len == 0)
{
/* Stop if XLOG_SWITCH was found. */
if (record->xl_rmid == RM_XLOG_ID && record->xl_info == XLOG_SWITCH)
{
dumpXLogRecord(record, false);
return false;
}
printf("ReadRecord: record with zero len at %u/%08X\n",
curRecPtr.xlogid, curRecPtr.xrecoff);
/* Attempt to recover on new page, but give up after a few... */
logRecOff = 0;
if (++retries > 4)
return false;
goto restart;
}
if (record->xl_tot_len < SizeOfXLogRecord + record->xl_len ||
record->xl_tot_len > SizeOfXLogRecord + record->xl_len +
XLR_MAX_BKP_BLOCKS * (sizeof(BkpBlock) + BLCKSZ))
{
printf("invalid record length(expected %u ~ %u, actual %d) at %X/%X\n",
(unsigned int)SizeOfXLogRecord + record->xl_len,
(unsigned int)SizeOfXLogRecord + record->xl_len +
XLR_MAX_BKP_BLOCKS * (sizeof(BkpBlock) + BLCKSZ),
record->xl_tot_len,
curRecPtr.xlogid, curRecPtr.xrecoff);
return false;
}
total_len = record->xl_tot_len;
/*
* Allocate or enlarge readRecordBuf as needed. To avoid useless
* small increases, round its size to a multiple of XLOG_BLCKSZ, and make
* sure it's at least 4*BLCKSZ to start with. (That is enough for all
* "normal" records, but very large commit or abort records might need
* more space.)
*/
if (total_len > readRecordBufSize)
{
uint32 newSize = total_len;
newSize += XLOG_BLCKSZ - (newSize % XLOG_BLCKSZ);
newSize = Max(newSize, 4 * XLOG_BLCKSZ);
if (readRecordBuf)
free(readRecordBuf);
readRecordBuf = (char *) malloc(newSize);
if (!readRecordBuf)
{
readRecordBufSize = 0;
/* We treat this as a "bogus data" condition */
fprintf(stderr, "record length %u at %X/%X too long\n",
total_len, curRecPtr.xlogid, curRecPtr.xrecoff);
return false;
}
readRecordBufSize = newSize;
}
buffer = readRecordBuf;
len = XLOG_BLCKSZ - curRecPtr.xrecoff % XLOG_BLCKSZ; /* available in block */
if (total_len > len)
{
/* Need to reassemble record */
uint32 gotlen = len;
memcpy(buffer, record, len);
record = (XLogRecord *) buffer;
buffer += len;
for (;;)
{
uint32 pageHeaderSize;
if (! readXLogPage())
{
/* XXX ought to be able to advance to new input file! */
fprintf(stderr, "Unable to read continuation page?\n");
dumpXLogRecord(record, true);
return false;
}
if (!(((XLogPageHeader) pageBuffer)->xlp_info & XLP_FIRST_IS_CONTRECORD))
{
printf("ReadRecord: there is no ContRecord flag in logfile %u seg %u off %u\n",
logId, logSeg, logPageOff);
return false;
}
pageHeaderSize = XLogPageHeaderSize((XLogPageHeader) pageBuffer);
contrecord = (XLogContRecord *) (pageBuffer + pageHeaderSize);
if (contrecord->xl_rem_len == 0 ||
total_len != (contrecord->xl_rem_len + gotlen))
{
printf("ReadRecord: invalid cont-record len %u in logfile %u seg %u off %u\n",
contrecord->xl_rem_len, logId, logSeg, logPageOff);
return false;
}
len = XLOG_BLCKSZ - pageHeaderSize - SizeOfXLogContRecord;
if (contrecord->xl_rem_len > len)
{
memcpy(buffer, (char *)contrecord + SizeOfXLogContRecord, len);
gotlen += len;
buffer += len;
continue;
}
memcpy(buffer, (char *) contrecord + SizeOfXLogContRecord,
contrecord->xl_rem_len);
logRecOff = MAXALIGN(pageHeaderSize + SizeOfXLogContRecord + contrecord->xl_rem_len);
break;
}
if (!RecordIsValid(record, curRecPtr))
return false;
return true;
}
/* Record is contained in this page */
memcpy(buffer, record, total_len);
record = (XLogRecord *) buffer;
logRecOff += MAXALIGN(total_len);
if (!RecordIsValid(record, curRecPtr))
return false;
return true;
}
Datum
lquery_in(PG_FUNCTION_ARGS)
{
char *buf = (char *) PG_GETARG_POINTER(0);
char *ptr;
int num = 0,
totallen = 0,
numOR = 0;
int state = LQPRS_WAITLEVEL;
lquery *result;
nodeitem *lptr = NULL;
lquery_level *cur,
*curqlevel,
*tmpql;
lquery_variant *lrptr = NULL;
bool hasnot = false;
bool wasbad = false;
int charlen;
int pos = 0;
ptr = buf;
while (*ptr)
{
charlen = pg_mblen(ptr);
if (charlen == 1)
{
if (t_iseq(ptr, '.'))
num++;
else if (t_iseq(ptr, '|'))
numOR++;
}
ptr += charlen;
}
num++;
curqlevel = tmpql = (lquery_level *) palloc0(ITEMSIZE * num);
ptr = buf;
while (*ptr)
{
charlen = pg_mblen(ptr);
if (state == LQPRS_WAITLEVEL)
{
if (ISALNUM(ptr))
{
GETVAR(curqlevel) = lptr = (nodeitem *) palloc0(sizeof(nodeitem) * (numOR + 1));
lptr->start = ptr;
state = LQPRS_WAITDELIM;
curqlevel->numvar = 1;
}
else if (charlen == 1 && t_iseq(ptr, '!'))
{
GETVAR(curqlevel) = lptr = (nodeitem *) palloc0(sizeof(nodeitem) * (numOR + 1));
lptr->start = ptr + 1;
state = LQPRS_WAITDELIM;
curqlevel->numvar = 1;
curqlevel->flag |= LQL_NOT;
hasnot = true;
}
else if (charlen == 1 && t_iseq(ptr, '*'))
state = LQPRS_WAITOPEN;
else
UNCHAR;
}
else if (state == LQPRS_WAITVAR)
{
if (ISALNUM(ptr))
{
lptr++;
lptr->start = ptr;
state = LQPRS_WAITDELIM;
curqlevel->numvar++;
}
else
UNCHAR;
}
else if (state == LQPRS_WAITDELIM)
{
if (charlen == 1 && t_iseq(ptr, '@'))
{
if (lptr->start == ptr)
UNCHAR;
lptr->flag |= LVAR_INCASE;
curqlevel->flag |= LVAR_INCASE;
}
else if (charlen == 1 && t_iseq(ptr, '*'))
{
if (lptr->start == ptr)
UNCHAR;
lptr->flag |= LVAR_ANYEND;
curqlevel->flag |= LVAR_ANYEND;
}
else if (charlen == 1 && t_iseq(ptr, '%'))
{
if (lptr->start == ptr)
UNCHAR;
lptr->flag |= LVAR_SUBLEXEME;
curqlevel->flag |= LVAR_SUBLEXEME;
}
else if (charlen == 1 && t_iseq(ptr, '|'))
{
lptr->len = ptr - lptr->start -
((lptr->flag & LVAR_SUBLEXEME) ? 1 : 0) -
((lptr->flag & LVAR_INCASE) ? 1 : 0) -
((lptr->flag & LVAR_ANYEND) ? 1 : 0);
if (lptr->wlen > 255)
ereport(ERROR,
(errcode(ERRCODE_NAME_TOO_LONG),
errmsg("name of level is too long"),
errdetail("Name length is %d, must "
"be < 256, in position %d.",
lptr->wlen, pos)));
state = LQPRS_WAITVAR;
}
else if (charlen == 1 && t_iseq(ptr, '.'))
{
lptr->len = ptr - lptr->start -
((lptr->flag & LVAR_SUBLEXEME) ? 1 : 0) -
((lptr->flag & LVAR_INCASE) ? 1 : 0) -
((lptr->flag & LVAR_ANYEND) ? 1 : 0);
if (lptr->wlen > 255)
ereport(ERROR,
(errcode(ERRCODE_NAME_TOO_LONG),
errmsg("name of level is too long"),
errdetail("Name length is %d, must "
"be < 256, in position %d.",
lptr->wlen, pos)));
state = LQPRS_WAITLEVEL;
curqlevel = NEXTLEV(curqlevel);
}
else if (ISALNUM(ptr))
{
if (lptr->flag)
UNCHAR;
}
else
UNCHAR;
}
else if (state == LQPRS_WAITOPEN)
{
if (charlen == 1 && t_iseq(ptr, '{'))
state = LQPRS_WAITFNUM;
else if (charlen == 1 && t_iseq(ptr, '.'))
{
curqlevel->low = 0;
curqlevel->high = 0xffff;
curqlevel = NEXTLEV(curqlevel);
state = LQPRS_WAITLEVEL;
}
else
UNCHAR;
}
else if (state == LQPRS_WAITFNUM)
{
if (charlen == 1 && t_iseq(ptr, ','))
state = LQPRS_WAITSNUM;
else if (t_isdigit(ptr))
{
curqlevel->low = atoi(ptr);
state = LQPRS_WAITND;
}
else
UNCHAR;
}
else if (state == LQPRS_WAITSNUM)
{
if (t_isdigit(ptr))
{
curqlevel->high = atoi(ptr);
state = LQPRS_WAITCLOSE;
}
else if (charlen == 1 && t_iseq(ptr, '}'))
{
curqlevel->high = 0xffff;
state = LQPRS_WAITEND;
}
else
UNCHAR;
}
else if (state == LQPRS_WAITCLOSE)
{
if (charlen == 1 && t_iseq(ptr, '}'))
state = LQPRS_WAITEND;
else if (!t_isdigit(ptr))
UNCHAR;
}
else if (state == LQPRS_WAITND)
{
if (charlen == 1 && t_iseq(ptr, '}'))
{
curqlevel->high = curqlevel->low;
state = LQPRS_WAITEND;
}
else if (charlen == 1 && t_iseq(ptr, ','))
state = LQPRS_WAITSNUM;
else if (!t_isdigit(ptr))
UNCHAR;
}
else if (state == LQPRS_WAITEND)
{
if (charlen == 1 && t_iseq(ptr, '.'))
{
state = LQPRS_WAITLEVEL;
curqlevel = NEXTLEV(curqlevel);
}
else
UNCHAR;
}
else
/* internal error */
elog(ERROR, "internal error in parser");
ptr += charlen;
if (state == LQPRS_WAITDELIM)
lptr->wlen++;
pos++;
}
if (state == LQPRS_WAITDELIM)
{
if (lptr->start == ptr)
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("syntax error"),
errdetail("Unexpected end of line.")));
lptr->len = ptr - lptr->start -
((lptr->flag & LVAR_SUBLEXEME) ? 1 : 0) -
((lptr->flag & LVAR_INCASE) ? 1 : 0) -
((lptr->flag & LVAR_ANYEND) ? 1 : 0);
if (lptr->len == 0)
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("syntax error"),
errdetail("Unexpected end of line.")));
if (lptr->wlen > 255)
ereport(ERROR,
(errcode(ERRCODE_NAME_TOO_LONG),
errmsg("name of level is too long"),
errdetail("Name length is %d, must "
"be < 256, in position %d.",
lptr->wlen, pos)));
}
else if (state == LQPRS_WAITOPEN)
curqlevel->high = 0xffff;
else if (state != LQPRS_WAITEND)
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("syntax error"),
errdetail("Unexpected end of line.")));
curqlevel = tmpql;
totallen = LQUERY_HDRSIZE;
while ((char *) curqlevel - (char *) tmpql < num * ITEMSIZE)
{
totallen += LQL_HDRSIZE;
if (curqlevel->numvar)
{
lptr = GETVAR(curqlevel);
while (lptr - GETVAR(curqlevel) < curqlevel->numvar)
{
totallen += MAXALIGN(LVAR_HDRSIZE + lptr->len);
lptr++;
}
}
else if (curqlevel->low > curqlevel->high)
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("syntax error"),
errdetail("Low limit(%d) is greater than upper(%d).",
curqlevel->low, curqlevel->high)));
curqlevel = NEXTLEV(curqlevel);
}
result = (lquery *) palloc0(totallen);
SET_VARSIZE(result, totallen);
result->numlevel = num;
result->firstgood = 0;
result->flag = 0;
if (hasnot)
result->flag |= LQUERY_HASNOT;
cur = LQUERY_FIRST(result);
curqlevel = tmpql;
while ((char *) curqlevel - (char *) tmpql < num * ITEMSIZE)
{
memcpy(cur, curqlevel, LQL_HDRSIZE);
cur->totallen = LQL_HDRSIZE;
if (curqlevel->numvar)
{
lrptr = LQL_FIRST(cur);
lptr = GETVAR(curqlevel);
while (lptr - GETVAR(curqlevel) < curqlevel->numvar)
{
cur->totallen += MAXALIGN(LVAR_HDRSIZE + lptr->len);
lrptr->len = lptr->len;
lrptr->flag = lptr->flag;
lrptr->val = ltree_crc32_sz(lptr->start, lptr->len);
memcpy(lrptr->name, lptr->start, lptr->len);
lptr++;
lrptr = LVAR_NEXT(lrptr);
}
pfree(GETVAR(curqlevel));
if (cur->numvar > 1 || cur->flag != 0)
wasbad = true;
else if (wasbad == false)
(result->firstgood)++;
}
else
wasbad = true;
curqlevel = NEXTLEV(curqlevel);
cur = LQL_NEXT(cur);
}
pfree(tmpql);
PG_RETURN_POINTER(result);
}
/*
* Insert a tuple to the new relation. This has to track heap_insert
* and its subsidiary functions!
*
* t_self of the tuple is set to the new TID of the tuple. If t_ctid of the
* tuple is invalid on entry, it's replaced with the new TID as well (in
* the inserted data only, not in the caller's copy).
*/
static void
raw_heap_insert(RewriteState state, HeapTuple tup)
{
Page page = state->rs_buffer;
Size pageFreeSpace,
saveFreeSpace;
Size len;
OffsetNumber newoff;
HeapTuple heaptup;
/*
* If the new tuple is too big for storage or contains already toasted
* out-of-line attributes from some other relation, invoke the toaster.
*
* Note: below this point, heaptup is the data we actually intend to store
* into the relation; tup is the caller's original untoasted data.
*/
if (state->rs_new_rel->rd_rel->relkind == RELKIND_TOASTVALUE)
{
/* toast table entries should never be recursively toasted */
Assert(!HeapTupleHasExternal(tup));
heaptup = tup;
}
else if (HeapTupleHasExternal(tup) || tup->t_len > TOAST_TUPLE_THRESHOLD)
heaptup = toast_insert_or_update(state->rs_new_rel, tup, NULL,
HEAP_INSERT_SKIP_FSM |
(state->rs_use_wal ?
0 : HEAP_INSERT_SKIP_WAL));
else
heaptup = tup;
len = MAXALIGN(heaptup->t_len); /* be conservative */
/*
* If we're gonna fail for oversize tuple, do it right away
*/
if (len > MaxHeapTupleSize)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("row is too big: size %lu, maximum size %lu",
(unsigned long) len,
(unsigned long) MaxHeapTupleSize)));
/* Compute desired extra freespace due to fillfactor option */
saveFreeSpace = RelationGetTargetPageFreeSpace(state->rs_new_rel,
HEAP_DEFAULT_FILLFACTOR);
/* Now we can check to see if there's enough free space already. */
if (state->rs_buffer_valid)
{
pageFreeSpace = PageGetHeapFreeSpace(page);
if (len + saveFreeSpace > pageFreeSpace)
{
/* Doesn't fit, so write out the existing page */
/* XLOG stuff */
if (state->rs_use_wal)
log_newpage(&state->rs_new_rel->rd_node,
MAIN_FORKNUM,
state->rs_blockno,
page,
true);
/*
* Now write the page. We say isTemp = true even if it's not a
* temp table, because there's no need for smgr to schedule an
* fsync for this write; we'll do it ourselves in
* end_heap_rewrite.
*/
RelationOpenSmgr(state->rs_new_rel);
PageSetChecksumInplace(page, state->rs_blockno);
smgrextend(state->rs_new_rel->rd_smgr, MAIN_FORKNUM,
state->rs_blockno, (char *) page, true);
state->rs_blockno++;
state->rs_buffer_valid = false;
}
}
if (!state->rs_buffer_valid)
{
/* Initialize a new empty page */
PageInit(page, BLCKSZ, 0);
state->rs_buffer_valid = true;
}
/* And now we can insert the tuple into the page */
newoff = PageAddItem(page, (Item) heaptup->t_data, heaptup->t_len,
InvalidOffsetNumber, false, true);
if (newoff == InvalidOffsetNumber)
elog(ERROR, "failed to add tuple");
/* Update caller's t_self to the actual position where it was stored */
ItemPointerSet(&(tup->t_self), state->rs_blockno, newoff);
/*
* Insert the correct position into CTID of the stored tuple, too, if the
* caller didn't supply a valid CTID.
*/
if (!ItemPointerIsValid(&tup->t_data->t_ctid))
{
ItemId newitemid;
HeapTupleHeader onpage_tup;
newitemid = PageGetItemId(page, newoff);
onpage_tup = (HeapTupleHeader) PageGetItem(page, newitemid);
onpage_tup->t_ctid = tup->t_self;
}
/* If heaptup is a private copy, release it. */
if (heaptup != tup)
heap_freetuple(heaptup);
}
/*
* Update tuple origtup (size origsz), located in offset oldoff of buffer
* oldbuf, to newtup (size newsz) as summary tuple for the page range starting
* at heapBlk. oldbuf must not be locked on entry, and is not locked at exit.
*
* If samepage is true, attempt to put the new tuple in the same page, but if
* there's no room, use some other one.
*
* If the update is successful, return true; the revmap is updated to point to
* the new tuple. If the update is not done for whatever reason, return false.
* Caller may retry the update if this happens.
*/
bool
brin_doupdate(Relation idxrel, BlockNumber pagesPerRange,
BrinRevmap *revmap, BlockNumber heapBlk,
Buffer oldbuf, OffsetNumber oldoff,
const BrinTuple *origtup, Size origsz,
const BrinTuple *newtup, Size newsz,
bool samepage)
{
Page oldpage;
ItemId oldlp;
BrinTuple *oldtup;
Size oldsz;
Buffer newbuf;
bool extended;
Assert(newsz == MAXALIGN(newsz));
/* If the item is oversized, don't bother. */
if (newsz > BrinMaxItemSize)
{
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("index row size %lu exceeds maximum %lu for index \"%s\"",
(unsigned long) newsz,
(unsigned long) BrinMaxItemSize,
RelationGetRelationName(idxrel))));
return false; /* keep compiler quiet */
}
/* make sure the revmap is long enough to contain the entry we need */
brinRevmapExtend(revmap, heapBlk);
if (!samepage)
{
/* need a page on which to put the item */
newbuf = brin_getinsertbuffer(idxrel, oldbuf, newsz, &extended);
if (!BufferIsValid(newbuf))
{
Assert(!extended);
return false;
}
/*
* Note: it's possible (though unlikely) that the returned newbuf is
* the same as oldbuf, if brin_getinsertbuffer determined that the old
* buffer does in fact have enough space.
*/
if (newbuf == oldbuf)
{
Assert(!extended);
newbuf = InvalidBuffer;
}
}
else
{
LockBuffer(oldbuf, BUFFER_LOCK_EXCLUSIVE);
newbuf = InvalidBuffer;
extended = false;
}
oldpage = BufferGetPage(oldbuf);
oldlp = PageGetItemId(oldpage, oldoff);
/*
* Check that the old tuple wasn't updated concurrently: it might have
* moved someplace else entirely ...
*/
if (!ItemIdIsNormal(oldlp))
{
LockBuffer(oldbuf, BUFFER_LOCK_UNLOCK);
/*
* If this happens, and the new buffer was obtained by extending the
* relation, then we need to ensure we don't leave it uninitialized or
* forget about it.
*/
if (BufferIsValid(newbuf))
{
if (extended)
brin_initialize_empty_new_buffer(idxrel, newbuf);
UnlockReleaseBuffer(newbuf);
if (extended)
FreeSpaceMapVacuum(idxrel);
}
return false;
}
oldsz = ItemIdGetLength(oldlp);
oldtup = (BrinTuple *) PageGetItem(oldpage, oldlp);
/*
* ... or it might have been updated in place to different contents.
*/
if (!brin_tuples_equal(oldtup, oldsz, origtup, origsz))
{
LockBuffer(oldbuf, BUFFER_LOCK_UNLOCK);
if (BufferIsValid(newbuf))
{
if (extended)
brin_initialize_empty_new_buffer(idxrel, newbuf);
UnlockReleaseBuffer(newbuf);
if (extended)
FreeSpaceMapVacuum(idxrel);
}
return false;
}
/*
* Great, the old tuple is intact. We can proceed with the update.
*
* If there's enough room in the old page for the new tuple, replace it.
*
* Note that there might now be enough space on the page even though the
* caller told us there isn't, if a concurrent update moved another tuple
* elsewhere or replaced a tuple with a smaller one.
*/
if (((BrinPageFlags(oldpage) & BRIN_EVACUATE_PAGE) == 0) &&
brin_can_do_samepage_update(oldbuf, origsz, newsz))
{
if (BufferIsValid(newbuf))
{
/* as above */
if (extended)
brin_initialize_empty_new_buffer(idxrel, newbuf);
UnlockReleaseBuffer(newbuf);
}
START_CRIT_SECTION();
if (!PageIndexTupleOverwrite(oldpage, oldoff, (Item) newtup, newsz))
elog(ERROR, "failed to replace BRIN tuple");
MarkBufferDirty(oldbuf);
/* XLOG stuff */
if (RelationNeedsWAL(idxrel))
{
xl_brin_samepage_update xlrec;
XLogRecPtr recptr;
uint8 info = XLOG_BRIN_SAMEPAGE_UPDATE;
xlrec.offnum = oldoff;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfBrinSamepageUpdate);
XLogRegisterBuffer(0, oldbuf, REGBUF_STANDARD);
XLogRegisterBufData(0, (char *) newtup, newsz);
recptr = XLogInsert(RM_BRIN_ID, info);
PageSetLSN(oldpage, recptr);
}
END_CRIT_SECTION();
LockBuffer(oldbuf, BUFFER_LOCK_UNLOCK);
if (extended)
FreeSpaceMapVacuum(idxrel);
return true;
}
else if (newbuf == InvalidBuffer)
{
/*
* Not enough space, but caller said that there was. Tell them to
* start over.
*/
LockBuffer(oldbuf, BUFFER_LOCK_UNLOCK);
return false;
}
else
{
/*
* Not enough free space on the oldpage. Put the new tuple on the new
* page, and update the revmap.
*/
Page newpage = BufferGetPage(newbuf);
Buffer revmapbuf;
ItemPointerData newtid;
OffsetNumber newoff;
BlockNumber newblk = InvalidBlockNumber;
Size freespace = 0;
revmapbuf = brinLockRevmapPageForUpdate(revmap, heapBlk);
START_CRIT_SECTION();
/*
* We need to initialize the page if it's newly obtained. Note we
* will WAL-log the initialization as part of the update, so we don't
* need to do that here.
*/
if (extended)
brin_page_init(BufferGetPage(newbuf), BRIN_PAGETYPE_REGULAR);
PageIndexTupleDeleteNoCompact(oldpage, oldoff);
newoff = PageAddItem(newpage, (Item) newtup, newsz,
InvalidOffsetNumber, false, false);
if (newoff == InvalidOffsetNumber)
elog(ERROR, "failed to add BRIN tuple to new page");
MarkBufferDirty(oldbuf);
MarkBufferDirty(newbuf);
/* needed to update FSM below */
if (extended)
{
newblk = BufferGetBlockNumber(newbuf);
freespace = br_page_get_freespace(newpage);
}
ItemPointerSet(&newtid, BufferGetBlockNumber(newbuf), newoff);
brinSetHeapBlockItemptr(revmapbuf, pagesPerRange, heapBlk, newtid);
MarkBufferDirty(revmapbuf);
/* XLOG stuff */
if (RelationNeedsWAL(idxrel))
{
xl_brin_update xlrec;
XLogRecPtr recptr;
uint8 info;
info = XLOG_BRIN_UPDATE | (extended ? XLOG_BRIN_INIT_PAGE : 0);
xlrec.insert.offnum = newoff;
xlrec.insert.heapBlk = heapBlk;
xlrec.insert.pagesPerRange = pagesPerRange;
xlrec.oldOffnum = oldoff;
XLogBeginInsert();
/* new page */
XLogRegisterData((char *) &xlrec, SizeOfBrinUpdate);
XLogRegisterBuffer(0, newbuf, REGBUF_STANDARD | (extended ? REGBUF_WILL_INIT : 0));
XLogRegisterBufData(0, (char *) newtup, newsz);
/* revmap page */
XLogRegisterBuffer(1, revmapbuf, REGBUF_STANDARD);
/* old page */
XLogRegisterBuffer(2, oldbuf, REGBUF_STANDARD);
recptr = XLogInsert(RM_BRIN_ID, info);
PageSetLSN(oldpage, recptr);
PageSetLSN(newpage, recptr);
PageSetLSN(BufferGetPage(revmapbuf), recptr);
}
END_CRIT_SECTION();
LockBuffer(revmapbuf, BUFFER_LOCK_UNLOCK);
LockBuffer(oldbuf, BUFFER_LOCK_UNLOCK);
UnlockReleaseBuffer(newbuf);
if (extended)
{
Assert(BlockNumberIsValid(newblk));
RecordPageWithFreeSpace(idxrel, newblk, freespace);
FreeSpaceMapVacuum(idxrel);
}
return true;
}
}
/*
* PageIndexTupleDelete
*
* This routine does the work of removing a tuple from an index page.
*
* Unlike heap pages, we compact out the line pointer for the removed tuple.
*/
void
PageIndexTupleDelete(Page page, OffsetNumber offnum)
{
PageHeader phdr = (PageHeader) page;
char *addr;
ItemId tup;
Size size;
unsigned offset;
int nbytes;
int offidx;
int nline;
/*
* As with PageRepairFragmentation, paranoia seems justified.
*/
if (phdr->pd_lower < SizeOfPageHeaderData ||
phdr->pd_lower > phdr->pd_upper ||
phdr->pd_upper > phdr->pd_special ||
phdr->pd_special > BLCKSZ)
ereport(ERROR,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted page pointers: lower = %u, upper = %u, special = %u",
phdr->pd_lower, phdr->pd_upper, phdr->pd_special)));
nline = PageGetMaxOffsetNumber(page);
if ((int) offnum <= 0 || (int) offnum > nline)
elog(ERROR, "invalid index offnum: %u", offnum);
/* change offset number to offset index */
offidx = offnum - 1;
tup = PageGetItemId(page, offnum);
Assert(ItemIdHasStorage(tup));
size = ItemIdGetLength(tup);
offset = ItemIdGetOffset(tup);
if (offset < phdr->pd_upper || (offset + size) > phdr->pd_special ||
offset != MAXALIGN(offset) || size != MAXALIGN(size))
ereport(ERROR,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted item pointer: offset = %u, size = %u",
offset, (unsigned int) size)));
/*
* First, we want to get rid of the pd_linp entry for the index tuple. We
* copy all subsequent linp's back one slot in the array. We don't use
* PageGetItemId, because we are manipulating the _array_, not individual
* linp's.
*/
nbytes = phdr->pd_lower -
((char *) &phdr->pd_linp[offidx + 1] - (char *) phdr);
if (nbytes > 0)
memmove((char *) &(phdr->pd_linp[offidx]),
(char *) &(phdr->pd_linp[offidx + 1]),
nbytes);
/*
* Now move everything between the old upper bound (beginning of tuple
* space) and the beginning of the deleted tuple forward, so that space in
* the middle of the page is left free. If we've just deleted the tuple
* at the beginning of tuple space, then there's no need to do the copy
* (and bcopy on some architectures SEGV's if asked to move zero bytes).
*/
/* beginning of tuple space */
addr = (char *) page + phdr->pd_upper;
if (offset > phdr->pd_upper)
memmove(addr + size, addr, (int) (offset - phdr->pd_upper));
/* adjust free space boundary pointers */
phdr->pd_upper += size;
phdr->pd_lower -= sizeof(ItemIdData);
/*
* Finally, we need to adjust the linp entries that remain.
*
* Anything that used to be before the deleted tuple's data was moved
* forward by the size of the deleted tuple.
*/
if (!PageIsEmpty(page))
{
int i;
nline--; /* there's one less than when we started */
for (i = 1; i <= nline; i++)
{
ItemId ii = PageGetItemId(phdr, i);
Assert(ItemIdHasStorage(ii));
if (ItemIdGetOffset(ii) <= offset)
ii->lp_off += size;
}
}
}
/*
* PageIndexMultiDelete
*
* This routine handles the case of deleting multiple tuples from an
* index page at once. It is considerably faster than a loop around
* PageIndexTupleDelete ... however, the caller *must* supply the array
* of item numbers to be deleted in item number order!
*/
void
PageIndexMultiDelete(Page page, OffsetNumber *itemnos, int nitems)
{
PageHeader phdr = (PageHeader) page;
Offset pd_lower = phdr->pd_lower;
Offset pd_upper = phdr->pd_upper;
Offset pd_special = phdr->pd_special;
itemIdSortData itemidbase[MaxIndexTuplesPerPage];
ItemIdData newitemids[MaxIndexTuplesPerPage];
itemIdSort itemidptr;
ItemId lp;
int nline,
nused;
Size totallen;
Size size;
unsigned offset;
int nextitm;
OffsetNumber offnum;
Assert(nitems <= MaxIndexTuplesPerPage);
/*
* If there aren't very many items to delete, then retail
* PageIndexTupleDelete is the best way. Delete the items in reverse
* order so we don't have to think about adjusting item numbers for
* previous deletions.
*
* TODO: tune the magic number here
*/
if (nitems <= 2)
{
while (--nitems >= 0)
PageIndexTupleDelete(page, itemnos[nitems]);
return;
}
/*
* As with PageRepairFragmentation, paranoia seems justified.
*/
if (pd_lower < SizeOfPageHeaderData ||
pd_lower > pd_upper ||
pd_upper > pd_special ||
pd_special > BLCKSZ ||
pd_special != MAXALIGN(pd_special))
ereport(ERROR,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted page pointers: lower = %u, upper = %u, special = %u",
pd_lower, pd_upper, pd_special)));
/*
* Scan the item pointer array and build a list of just the ones we are
* going to keep. Notice we do not modify the page yet, since we are
* still validity-checking.
*/
nline = PageGetMaxOffsetNumber(page);
itemidptr = itemidbase;
totallen = 0;
nused = 0;
nextitm = 0;
for (offnum = FirstOffsetNumber; offnum <= nline; offnum = OffsetNumberNext(offnum))
{
lp = PageGetItemId(page, offnum);
Assert(ItemIdHasStorage(lp));
size = ItemIdGetLength(lp);
offset = ItemIdGetOffset(lp);
if (offset < pd_upper ||
(offset + size) > pd_special ||
offset != MAXALIGN(offset))
ereport(ERROR,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted item pointer: offset = %u, length = %u",
offset, (unsigned int) size)));
if (nextitm < nitems && offnum == itemnos[nextitm])
{
/* skip item to be deleted */
nextitm++;
}
else
{
itemidptr->offsetindex = nused; /* where it will go */
itemidptr->itemoff = offset;
itemidptr->alignedlen = MAXALIGN(size);
totallen += itemidptr->alignedlen;
newitemids[nused] = *lp;
itemidptr++;
nused++;
}
}
/* this will catch invalid or out-of-order itemnos[] */
if (nextitm != nitems)
elog(ERROR, "incorrect index offsets supplied");
if (totallen > (Size) (pd_special - pd_lower))
ereport(ERROR,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted item lengths: total %u, available space %u",
(unsigned int) totallen, pd_special - pd_lower)));
/*
* Looks good. Overwrite the line pointers with the copy, from which we've
* removed all the unused items.
*/
memcpy(phdr->pd_linp, newitemids, nused * sizeof(ItemIdData));
phdr->pd_lower = SizeOfPageHeaderData + nused * sizeof(ItemIdData);
/* and compactify the tuple data */
compactify_tuples(itemidbase, nused, page);
}
/*
* _hash_init() -- 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_init(Relation rel, double num_tuples, ForkNumber forkNum)
{
Buffer metabuf;
Buffer buf;
Buffer bitmapbuf;
Page pg;
HashMetaPage metap;
RegProcedure procid;
int32 data_width;
int32 item_width;
int32 ffactor;
uint32 num_buckets;
uint32 i;
bool use_wal;
/* safety check */
if (RelationGetNumberOfBlocksInFork(rel, forkNum) != 0)
elog(ERROR, "cannot initialize non-empty hash index \"%s\"",
RelationGetRelationName(rel));
/*
* WAL log creation of pages if the relation is persistent, or this is the
* init fork. Init forks for unlogged relations always need to be WAL
* logged.
*/
use_wal = RelationNeedsWAL(rel) || forkNum == INIT_FORKNUM;
/*
* 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;
procid = index_getprocid(rel, 1, HASHSTANDARD_PROC);
/*
* 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.
*
* Critical section not required, because on error the creation of the
* whole relation will be rolled back.
*/
metabuf = _hash_getnewbuf(rel, HASH_METAPAGE, forkNum);
_hash_init_metabuffer(metabuf, num_tuples, procid, ffactor, false);
MarkBufferDirty(metabuf);
pg = BufferGetPage(metabuf);
metap = HashPageGetMeta(pg);
/* XLOG stuff */
if (use_wal)
{
xl_hash_init_meta_page xlrec;
XLogRecPtr recptr;
xlrec.num_tuples = num_tuples;
xlrec.procid = metap->hashm_procid;
xlrec.ffactor = metap->hashm_ffactor;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHashInitMetaPage);
XLogRegisterBuffer(0, metabuf, REGBUF_WILL_INIT | REGBUF_STANDARD);
recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_INIT_META_PAGE);
PageSetLSN(BufferGetPage(metabuf), recptr);
}
num_buckets = metap->hashm_maxbucket + 1;
/*
* 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.
*/
LockBuffer(metabuf, BUFFER_LOCK_UNLOCK);
/*
* Initialize and WAL Log the first N buckets
*/
for (i = 0; i < num_buckets; i++)
{
BlockNumber blkno;
/* Allow interrupts, in case N is huge */
CHECK_FOR_INTERRUPTS();
blkno = BUCKET_TO_BLKNO(metap, i);
buf = _hash_getnewbuf(rel, blkno, forkNum);
_hash_initbuf(buf, metap->hashm_maxbucket, i, LH_BUCKET_PAGE, false);
MarkBufferDirty(buf);
if (use_wal)
log_newpage(&rel->rd_node,
forkNum,
blkno,
BufferGetPage(buf),
true);
_hash_relbuf(rel, buf);
}
/* Now reacquire buffer lock on metapage */
LockBuffer(metabuf, BUFFER_LOCK_EXCLUSIVE);
/*
* Initialize bitmap page
*/
bitmapbuf = _hash_getnewbuf(rel, num_buckets + 1, forkNum);
_hash_initbitmapbuffer(bitmapbuf, metap->hashm_bmsize, false);
MarkBufferDirty(bitmapbuf);
/* add the new bitmap page to the metapage's list of bitmaps */
/* metapage already has a write lock */
if (metap->hashm_nmaps >= HASH_MAX_BITMAPS)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("out of overflow pages in hash index \"%s\"",
RelationGetRelationName(rel))));
metap->hashm_mapp[metap->hashm_nmaps] = num_buckets + 1;
metap->hashm_nmaps++;
MarkBufferDirty(metabuf);
/* XLOG stuff */
if (use_wal)
{
xl_hash_init_bitmap_page xlrec;
XLogRecPtr recptr;
xlrec.bmsize = metap->hashm_bmsize;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHashInitBitmapPage);
XLogRegisterBuffer(0, bitmapbuf, REGBUF_WILL_INIT);
/*
* This is safe only because nobody else can be modifying the index at
* this stage; it's only visible to the transaction that is creating
* it.
*/
XLogRegisterBuffer(1, metabuf, REGBUF_STANDARD);
recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_INIT_BITMAP_PAGE);
PageSetLSN(BufferGetPage(bitmapbuf), recptr);
PageSetLSN(BufferGetPage(metabuf), recptr);
}
/* all done */
_hash_relbuf(rel, bitmapbuf);
_hash_relbuf(rel, metabuf);
return num_buckets;
}
static void
makeSublist(Relation index, IndexTuple *tuples, int32 ntuples,
GinMetaPageData *res)
{
Buffer curBuffer = InvalidBuffer;
Buffer prevBuffer = InvalidBuffer;
int i,
size = 0,
tupsize;
int startTuple = 0;
Assert(ntuples > 0);
/*
* Split tuples into pages
*/
for (i = 0; i < ntuples; i++)
{
if (curBuffer == InvalidBuffer)
{
curBuffer = GinNewBuffer(index);
if (prevBuffer != InvalidBuffer)
{
res->nPendingPages++;
writeListPage(index, prevBuffer,
tuples + startTuple,
i - startTuple,
BufferGetBlockNumber(curBuffer));
}
else
{
res->head = BufferGetBlockNumber(curBuffer);
}
prevBuffer = curBuffer;
startTuple = i;
size = 0;
}
tupsize = MAXALIGN(IndexTupleSize(tuples[i])) + sizeof(ItemIdData);
if (size + tupsize > GinListPageSize)
{
/* won't fit, force a new page and reprocess */
i--;
curBuffer = InvalidBuffer;
}
else
{
size += tupsize;
}
}
/*
* Write last page
*/
res->tail = BufferGetBlockNumber(curBuffer);
res->tailFreeSize = writeListPage(index, curBuffer,
tuples + startTuple,
ntuples - startTuple,
InvalidBlockNumber);
res->nPendingPages++;
/* that was only one heap tuple */
res->nPendingHeapTuples = 1;
}
/*
* Write bytes into a shared message queue.
*/
static shm_mq_result
shm_mq_send_bytes(shm_mq_handle *mqh, Size nbytes, const void *data,
bool nowait, Size *bytes_written)
{
shm_mq *mq = mqh->mqh_queue;
Size sent = 0;
uint64 used;
Size ringsize = mq->mq_ring_size;
Size available;
while (sent < nbytes)
{
bool detached;
uint64 rb;
/* Compute number of ring buffer bytes used and available. */
rb = shm_mq_get_bytes_read(mq, &detached);
Assert(mq->mq_bytes_written >= rb);
used = mq->mq_bytes_written - rb;
Assert(used <= ringsize);
available = Min(ringsize - used, nbytes - sent);
/* Bail out if the queue has been detached. */
if (detached)
{
*bytes_written = sent;
return SHM_MQ_DETACHED;
}
if (available == 0 && !mqh->mqh_counterparty_attached)
{
/*
* The queue is full, so if the receiver isn't yet known to be
* attached, we must wait for that to happen.
*/
if (nowait)
{
if (shm_mq_get_receiver(mq) == NULL)
{
*bytes_written = sent;
return SHM_MQ_WOULD_BLOCK;
}
}
else if (!shm_mq_wait_internal(mq, &mq->mq_receiver,
mqh->mqh_handle))
{
mq->mq_detached = true;
*bytes_written = sent;
return SHM_MQ_DETACHED;
}
mqh->mqh_counterparty_attached = true;
/*
* The receiver may have read some data after attaching, so we
* must not wait without rechecking the queue state.
*/
}
else if (available == 0)
{
shm_mq_result res;
/* Let the receiver know that we need them to read some data. */
res = shm_mq_notify_receiver(mq);
if (res != SHM_MQ_SUCCESS)
{
*bytes_written = sent;
return res;
}
/* Skip manipulation of our latch if nowait = true. */
if (nowait)
{
*bytes_written = sent;
return SHM_MQ_WOULD_BLOCK;
}
/*
* Wait for our latch to be set. It might already be set for some
* unrelated reason, but that'll just result in one extra trip
* through the loop. It's worth it to avoid resetting the latch
* at top of loop, because setting an already-set latch is much
* cheaper than setting one that has been reset.
*/
WaitLatch(MyLatch, WL_LATCH_SET, 0);
/* An interrupt may have occurred while we were waiting. */
CHECK_FOR_INTERRUPTS();
/* Reset the latch so we don't spin. */
ResetLatch(MyLatch);
}
else
{
Size offset = mq->mq_bytes_written % (uint64) ringsize;
Size sendnow = Min(available, ringsize - offset);
/* Write as much data as we can via a single memcpy(). */
memcpy(&mq->mq_ring[mq->mq_ring_offset + offset],
(char *) data + sent, sendnow);
sent += sendnow;
/*
* Update count of bytes written, with alignment padding. Note
* that this will never actually insert any padding except at the
* end of a run of bytes, because the buffer size is a multiple of
* MAXIMUM_ALIGNOF, and each read is as well.
*/
Assert(sent == nbytes || sendnow == MAXALIGN(sendnow));
shm_mq_inc_bytes_written(mq, MAXALIGN(sendnow));
/*
* For efficiency, we don't set the reader's latch here. We'll do
* that only when the buffer fills up or after writing an entire
* message.
*/
}
}
*bytes_written = sent;
return SHM_MQ_SUCCESS;
}
/*
* _hash_splitbucket -- split 'obucket' into 'obucket' and 'nbucket'
*
* This routine is used to partition the tuples between old and new bucket and
* is used to finish the incomplete split operations. To finish the previously
* interrupted split operation, the caller needs to fill htab. If htab is set,
* then we skip the movement of tuples that exists in htab, otherwise NULL
* value of htab indicates movement of all the tuples that belong to the new
* bucket.
*
* We are splitting a bucket that consists of a base bucket page and zero
* or more overflow (bucket chain) pages. We must relocate tuples that
* belong in the new bucket.
*
* The caller must hold cleanup locks on both buckets to ensure that
* no one else is trying to access them (see README).
*
* The caller must hold a pin, but no lock, on the metapage buffer.
* The buffer is returned in the same state. (The metapage is only
* touched if it becomes necessary to add or remove overflow pages.)
*
* Split needs to retain pin on primary bucket pages of both old and new
* buckets till end of operation. This is to prevent vacuum from starting
* while a split is in progress.
*
* In addition, the caller must have created the new bucket's base page,
* which is passed in buffer nbuf, pinned and write-locked. The lock will be
* released here and pin must be released by the caller. (The API is set up
* this way because we must do _hash_getnewbuf() before releasing the metapage
* write lock. So instead of passing the new bucket's start block number, we
* pass an actual buffer.)
*/
static void
_hash_splitbucket(Relation rel,
Buffer metabuf,
Bucket obucket,
Bucket nbucket,
Buffer obuf,
Buffer nbuf,
HTAB *htab,
uint32 maxbucket,
uint32 highmask,
uint32 lowmask)
{
Buffer bucket_obuf;
Buffer bucket_nbuf;
Page opage;
Page npage;
HashPageOpaque oopaque;
HashPageOpaque nopaque;
OffsetNumber itup_offsets[MaxIndexTuplesPerPage];
IndexTuple itups[MaxIndexTuplesPerPage];
Size all_tups_size = 0;
int i;
uint16 nitups = 0;
bucket_obuf = obuf;
opage = BufferGetPage(obuf);
oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);
bucket_nbuf = nbuf;
npage = BufferGetPage(nbuf);
nopaque = (HashPageOpaque) PageGetSpecialPointer(npage);
/* Copy the predicate locks from old bucket to new bucket. */
PredicateLockPageSplit(rel,
BufferGetBlockNumber(bucket_obuf),
BufferGetBlockNumber(bucket_nbuf));
/*
* Partition the tuples in the old bucket between the old bucket and the
* new bucket, advancing along the old bucket's overflow bucket chain and
* adding overflow pages to the new bucket as needed. Outer loop iterates
* once per page in old bucket.
*/
for (;;)
{
BlockNumber oblkno;
OffsetNumber ooffnum;
OffsetNumber omaxoffnum;
/* Scan each tuple in old page */
omaxoffnum = PageGetMaxOffsetNumber(opage);
for (ooffnum = FirstOffsetNumber;
ooffnum <= omaxoffnum;
ooffnum = OffsetNumberNext(ooffnum))
{
IndexTuple itup;
Size itemsz;
Bucket bucket;
bool found = false;
/* skip dead tuples */
if (ItemIdIsDead(PageGetItemId(opage, ooffnum)))
continue;
/*
* Before inserting a tuple, probe the hash table containing TIDs
* of tuples belonging to new bucket, if we find a match, then
* skip that tuple, else fetch the item's hash key (conveniently
* stored in the item) and determine which bucket it now belongs
* in.
*/
itup = (IndexTuple) PageGetItem(opage,
PageGetItemId(opage, ooffnum));
if (htab)
(void) hash_search(htab, &itup->t_tid, HASH_FIND, &found);
if (found)
continue;
bucket = _hash_hashkey2bucket(_hash_get_indextuple_hashkey(itup),
maxbucket, highmask, lowmask);
if (bucket == nbucket)
{
IndexTuple new_itup;
/*
* make a copy of index tuple as we have to scribble on it.
*/
new_itup = CopyIndexTuple(itup);
/*
* mark the index tuple as moved by split, such tuples are
* skipped by scan if there is split in progress for a bucket.
*/
new_itup->t_info |= INDEX_MOVED_BY_SPLIT_MASK;
/*
* insert the tuple into the new bucket. if it doesn't fit on
* the current page in the new bucket, we must allocate a new
* overflow page and place the tuple on that page instead.
*/
itemsz = IndexTupleSize(new_itup);
itemsz = MAXALIGN(itemsz);
if (PageGetFreeSpaceForMultipleTuples(npage, nitups + 1) < (all_tups_size + itemsz))
{
/*
* Change the shared buffer state in critical section,
* otherwise any error could make it unrecoverable.
*/
START_CRIT_SECTION();
_hash_pgaddmultitup(rel, nbuf, itups, itup_offsets, nitups);
MarkBufferDirty(nbuf);
/* log the split operation before releasing the lock */
log_split_page(rel, nbuf);
END_CRIT_SECTION();
/* drop lock, but keep pin */
LockBuffer(nbuf, BUFFER_LOCK_UNLOCK);
/* be tidy */
for (i = 0; i < nitups; i++)
pfree(itups[i]);
nitups = 0;
all_tups_size = 0;
/* chain to a new overflow page */
nbuf = _hash_addovflpage(rel, metabuf, nbuf, (nbuf == bucket_nbuf) ? true : false);
npage = BufferGetPage(nbuf);
nopaque = (HashPageOpaque) PageGetSpecialPointer(npage);
}
itups[nitups++] = new_itup;
all_tups_size += itemsz;
}
else
{
/*
* the tuple stays on this page, so nothing to do.
*/
Assert(bucket == obucket);
}
}
oblkno = oopaque->hasho_nextblkno;
/* retain the pin on the old primary bucket */
if (obuf == bucket_obuf)
LockBuffer(obuf, BUFFER_LOCK_UNLOCK);
else
_hash_relbuf(rel, obuf);
/* Exit loop if no more overflow pages in old bucket */
if (!BlockNumberIsValid(oblkno))
{
/*
* Change the shared buffer state in critical section, otherwise
* any error could make it unrecoverable.
*/
START_CRIT_SECTION();
_hash_pgaddmultitup(rel, nbuf, itups, itup_offsets, nitups);
MarkBufferDirty(nbuf);
/* log the split operation before releasing the lock */
log_split_page(rel, nbuf);
END_CRIT_SECTION();
if (nbuf == bucket_nbuf)
LockBuffer(nbuf, BUFFER_LOCK_UNLOCK);
else
_hash_relbuf(rel, nbuf);
/* be tidy */
for (i = 0; i < nitups; i++)
pfree(itups[i]);
break;
}
/* Else, advance to next old page */
obuf = _hash_getbuf(rel, oblkno, HASH_READ, LH_OVERFLOW_PAGE);
opage = BufferGetPage(obuf);
oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);
}
/*
* We're at the end of the old bucket chain, so we're done partitioning
* the tuples. Mark the old and new buckets to indicate split is
* finished.
*
* To avoid deadlocks due to locking order of buckets, first lock the old
* bucket and then the new bucket.
*/
LockBuffer(bucket_obuf, BUFFER_LOCK_EXCLUSIVE);
opage = BufferGetPage(bucket_obuf);
oopaque = (HashPageOpaque) PageGetSpecialPointer(opage);
LockBuffer(bucket_nbuf, BUFFER_LOCK_EXCLUSIVE);
npage = BufferGetPage(bucket_nbuf);
nopaque = (HashPageOpaque) PageGetSpecialPointer(npage);
START_CRIT_SECTION();
oopaque->hasho_flag &= ~LH_BUCKET_BEING_SPLIT;
nopaque->hasho_flag &= ~LH_BUCKET_BEING_POPULATED;
/*
* After the split is finished, mark the old bucket to indicate that it
* contains deletable tuples. We will clear split-cleanup flag after
* deleting such tuples either at the end of split or at the next split
* from old bucket or at the time of vacuum.
*/
oopaque->hasho_flag |= LH_BUCKET_NEEDS_SPLIT_CLEANUP;
/*
* now write the buffers, here we don't release the locks as caller is
* responsible to release locks.
*/
MarkBufferDirty(bucket_obuf);
MarkBufferDirty(bucket_nbuf);
if (RelationNeedsWAL(rel))
{
XLogRecPtr recptr;
xl_hash_split_complete xlrec;
xlrec.old_bucket_flag = oopaque->hasho_flag;
xlrec.new_bucket_flag = nopaque->hasho_flag;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHashSplitComplete);
XLogRegisterBuffer(0, bucket_obuf, REGBUF_STANDARD);
XLogRegisterBuffer(1, bucket_nbuf, REGBUF_STANDARD);
recptr = XLogInsert(RM_HASH_ID, XLOG_HASH_SPLIT_COMPLETE);
PageSetLSN(BufferGetPage(bucket_obuf), recptr);
PageSetLSN(BufferGetPage(bucket_nbuf), recptr);
}
END_CRIT_SECTION();
/*
* If possible, clean up the old bucket. We might not be able to do this
* if someone else has a pin on it, but if not then we can go ahead. This
* isn't absolutely necessary, but it reduces bloat; if we don't do it
* now, VACUUM will do it eventually, but maybe not until new overflow
* pages have been allocated. Note that there's no need to clean up the
* new bucket.
*/
if (IsBufferCleanupOK(bucket_obuf))
{
LockBuffer(bucket_nbuf, BUFFER_LOCK_UNLOCK);
hashbucketcleanup(rel, obucket, bucket_obuf,
BufferGetBlockNumber(bucket_obuf), NULL,
maxbucket, highmask, lowmask, NULL, NULL, true,
NULL, NULL);
}
else
{
LockBuffer(bucket_nbuf, BUFFER_LOCK_UNLOCK);
LockBuffer(bucket_obuf, BUFFER_LOCK_UNLOCK);
}
}
/*
* Receive a message from a shared message queue.
*
* We set *nbytes to the message length and *data to point to the message
* payload. If the entire message exists in the queue as a single,
* contiguous chunk, *data will point directly into shared memory; otherwise,
* it will point to a temporary buffer. This mostly avoids data copying in
* the hoped-for case where messages are short compared to the buffer size,
* while still allowing longer messages. In either case, the return value
* remains valid until the next receive operation is perfomed on the queue.
*
* When nowait = false, we'll wait on our process latch when the ring buffer
* is empty and we have not yet received a full message. The sender will
* set our process latch after more data has been written, and we'll resume
* processing. Each call will therefore return a complete message
* (unless the sender detaches the queue).
*
* When nowait = true, we do not manipulate the state of the process latch;
* instead, whenever the buffer is empty and we need to read from it, we
* return SHM_MQ_WOULD_BLOCK. In this case, the caller should call this
* function again after the process latch has been set.
*/
shm_mq_result
shm_mq_receive(shm_mq_handle *mqh, Size *nbytesp, void **datap, bool nowait)
{
shm_mq *mq = mqh->mqh_queue;
shm_mq_result res;
Size rb = 0;
Size nbytes;
void *rawdata;
Assert(mq->mq_receiver == MyProc);
/* We can't receive data until the sender has attached. */
if (!mqh->mqh_counterparty_attached)
{
if (nowait)
{
if (shm_mq_get_sender(mq) == NULL)
return SHM_MQ_WOULD_BLOCK;
}
else if (!shm_mq_wait_internal(mq, &mq->mq_sender, mqh->mqh_handle)
&& shm_mq_get_sender(mq) == NULL)
{
mq->mq_detached = true;
return SHM_MQ_DETACHED;
}
mqh->mqh_counterparty_attached = true;
}
/* Consume any zero-copy data from previous receive operation. */
if (mqh->mqh_consume_pending > 0)
{
shm_mq_inc_bytes_read(mq, mqh->mqh_consume_pending);
mqh->mqh_consume_pending = 0;
}
/* Try to read, or finish reading, the length word from the buffer. */
while (!mqh->mqh_length_word_complete)
{
/* Try to receive the message length word. */
Assert(mqh->mqh_partial_bytes < sizeof(Size));
res = shm_mq_receive_bytes(mq, sizeof(Size) - mqh->mqh_partial_bytes,
nowait, &rb, &rawdata);
if (res != SHM_MQ_SUCCESS)
return res;
/*
* Hopefully, we'll receive the entire message length word at once.
* But if sizeof(Size) > MAXIMUM_ALIGNOF, then it might be split over
* multiple reads.
*/
if (mqh->mqh_partial_bytes == 0 && rb >= sizeof(Size))
{
Size needed;
nbytes = *(Size *) rawdata;
/* If we've already got the whole message, we're done. */
needed = MAXALIGN(sizeof(Size)) + MAXALIGN(nbytes);
if (rb >= needed)
{
/*
* Technically, we could consume the message length
* information at this point, but the extra write to shared
* memory wouldn't be free and in most cases we would reap no
* benefit.
*/
mqh->mqh_consume_pending = needed;
*nbytesp = nbytes;
*datap = ((char *) rawdata) + MAXALIGN(sizeof(Size));
return SHM_MQ_SUCCESS;
}
/*
* We don't have the whole message, but we at least have the whole
* length word.
*/
mqh->mqh_expected_bytes = nbytes;
mqh->mqh_length_word_complete = true;
shm_mq_inc_bytes_read(mq, MAXALIGN(sizeof(Size)));
rb -= MAXALIGN(sizeof(Size));
}
else
{
Size lengthbytes;
/* Can't be split unless bigger than required alignment. */
Assert(sizeof(Size) > MAXIMUM_ALIGNOF);
/* Message word is split; need buffer to reassemble. */
if (mqh->mqh_buffer == NULL)
{
mqh->mqh_buffer = MemoryContextAlloc(mqh->mqh_context,
MQH_INITIAL_BUFSIZE);
mqh->mqh_buflen = MQH_INITIAL_BUFSIZE;
}
Assert(mqh->mqh_buflen >= sizeof(Size));
/* Copy and consume partial length word. */
if (mqh->mqh_partial_bytes + rb > sizeof(Size))
lengthbytes = sizeof(Size) - mqh->mqh_partial_bytes;
else
lengthbytes = rb;
memcpy(&mqh->mqh_buffer[mqh->mqh_partial_bytes], rawdata,
lengthbytes);
mqh->mqh_partial_bytes += lengthbytes;
shm_mq_inc_bytes_read(mq, MAXALIGN(lengthbytes));
rb -= lengthbytes;
/* If we now have the whole word, we're ready to read payload. */
if (mqh->mqh_partial_bytes >= sizeof(Size))
{
Assert(mqh->mqh_partial_bytes == sizeof(Size));
mqh->mqh_expected_bytes = *(Size *) mqh->mqh_buffer;
mqh->mqh_length_word_complete = true;
mqh->mqh_partial_bytes = 0;
}
}
}
nbytes = mqh->mqh_expected_bytes;
if (mqh->mqh_partial_bytes == 0)
{
/*
* Try to obtain the whole message in a single chunk. If this works,
* we need not copy the data and can return a pointer directly into
* shared memory.
*/
res = shm_mq_receive_bytes(mq, nbytes, nowait, &rb, &rawdata);
if (res != SHM_MQ_SUCCESS)
return res;
if (rb >= nbytes)
{
mqh->mqh_length_word_complete = false;
mqh->mqh_consume_pending = MAXALIGN(nbytes);
*nbytesp = nbytes;
*datap = rawdata;
return SHM_MQ_SUCCESS;
}
/*
* The message has wrapped the buffer. We'll need to copy it in order
* to return it to the client in one chunk. First, make sure we have
* a large enough buffer available.
*/
if (mqh->mqh_buflen < nbytes)
{
Size newbuflen = Max(mqh->mqh_buflen, MQH_INITIAL_BUFSIZE);
while (newbuflen < nbytes)
newbuflen *= 2;
if (mqh->mqh_buffer != NULL)
{
pfree(mqh->mqh_buffer);
mqh->mqh_buffer = NULL;
mqh->mqh_buflen = 0;
}
mqh->mqh_buffer = MemoryContextAlloc(mqh->mqh_context, newbuflen);
mqh->mqh_buflen = newbuflen;
}
}
/* Loop until we've copied the entire message. */
for (;;)
{
Size still_needed;
/* Copy as much as we can. */
Assert(mqh->mqh_partial_bytes + rb <= nbytes);
memcpy(&mqh->mqh_buffer[mqh->mqh_partial_bytes], rawdata, rb);
mqh->mqh_partial_bytes += rb;
/*
* Update count of bytes read, with alignment padding. Note that this
* will never actually insert any padding except at the end of a
* message, because the buffer size is a multiple of MAXIMUM_ALIGNOF,
* and each read and write is as well.
*/
Assert(mqh->mqh_partial_bytes == nbytes || rb == MAXALIGN(rb));
shm_mq_inc_bytes_read(mq, MAXALIGN(rb));
/* If we got all the data, exit the loop. */
if (mqh->mqh_partial_bytes >= nbytes)
break;
/* Wait for some more data. */
still_needed = nbytes - mqh->mqh_partial_bytes;
res = shm_mq_receive_bytes(mq, still_needed, nowait, &rb, &rawdata);
if (res != SHM_MQ_SUCCESS)
return res;
if (rb > still_needed)
rb = still_needed;
}
/* Return the complete message, and reset for next message. */
*nbytesp = nbytes;
*datap = mqh->mqh_buffer;
mqh->mqh_length_word_complete = false;
mqh->mqh_partial_bytes = 0;
return SHM_MQ_SUCCESS;
}
static shm_mq_result shm_mq_send_bytes(shm_mq_handle *mq, Size nbytes,
const void *data, bool nowait, Size *bytes_written);
static shm_mq_result shm_mq_receive_bytes(shm_mq *mq, Size bytes_needed,
bool nowait, Size *nbytesp, void **datap);
static bool shm_mq_wait_internal(volatile shm_mq *mq, PGPROC *volatile * ptr,
BackgroundWorkerHandle *handle);
static uint64 shm_mq_get_bytes_read(volatile shm_mq *mq, bool *detached);
static void shm_mq_inc_bytes_read(volatile shm_mq *mq, Size n);
static uint64 shm_mq_get_bytes_written(volatile shm_mq *mq, bool *detached);
static void shm_mq_inc_bytes_written(volatile shm_mq *mq, Size n);
static shm_mq_result shm_mq_notify_receiver(volatile shm_mq *mq);
static void shm_mq_detach_callback(dsm_segment *seg, Datum arg);
/* Minimum queue size is enough for header and at least one chunk of data. */
const Size shm_mq_minimum_size =
MAXALIGN(offsetof(shm_mq, mq_ring)) + MAXIMUM_ALIGNOF;
#define MQH_INITIAL_BUFSIZE 8192
/*
* Initialize a new shared message queue.
*/
shm_mq *
shm_mq_create(void *address, Size size)
{
shm_mq *mq = address;
Size data_offset = MAXALIGN(offsetof(shm_mq, mq_ring));
/* If the size isn't MAXALIGN'd, just discard the odd bytes. */
size = MAXALIGN_DOWN(size);
/*----------
* Add an item to a disk page from the sort output.
*
* We must be careful to observe the page layout conventions of nbtsearch.c:
* - rightmost pages start data items at P_HIKEY instead of at P_FIRSTKEY.
* - on non-leaf pages, the key portion of the first item need not be
* stored, we should store only the link.
*
* A leaf page being built looks like:
*
* +----------------+---------------------------------+
* | PageHeaderData | linp0 linp1 linp2 ... |
* +-----------+----+---------------------------------+
* | ... linpN | |
* +-----------+--------------------------------------+
* | ^ last |
* | |
* +-------------+------------------------------------+
* | | itemN ... |
* +-------------+------------------+-----------------+
* | ... item3 item2 item1 | "special space" |
* +--------------------------------+-----------------+
*
* Contrast this with the diagram in bufpage.h; note the mismatch
* between linps and items. This is because we reserve linp0 as a
* placeholder for the pointer to the "high key" item; when we have
* filled up the page, we will set linp0 to point to itemN and clear
* linpN. On the other hand, if we find this is the last (rightmost)
* page, we leave the items alone and slide the linp array over.
*
* 'last' pointer indicates the last offset added to the page.
*----------
*/
static void
_bt_buildadd(BTWriteState *wstate, BTPageState *state, IndexTuple itup)
{
Page npage;
BlockNumber nblkno;
OffsetNumber last_off;
Size pgspc;
Size itupsz;
/*
* This is a handy place to check for cancel interrupts during the btree
* load phase of index creation.
*/
CHECK_FOR_INTERRUPTS();
npage = state->btps_page;
nblkno = state->btps_blkno;
last_off = state->btps_lastoff;
pgspc = PageGetFreeSpace(npage);
itupsz = IndexTupleDSize(*itup);
itupsz = MAXALIGN(itupsz);
/*
* Check whether the item can fit on a btree page at all. (Eventually, we
* ought to try to apply TOAST methods if not.) We actually need to be
* able to fit three items on every page, so restrict any one item to 1/3
* the per-page available space. Note that at this point, itupsz doesn't
* include the ItemId.
*
* NOTE: similar code appears in _bt_insertonpg() to defend against
* oversize items being inserted into an already-existing index. But
* during creation of an index, we don't go through there.
*/
if (itupsz > BTMaxItemSize(npage))
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("index row size %zu exceeds maximum %zu for index \"%s\"",
itupsz, BTMaxItemSize(npage),
RelationGetRelationName(wstate->index)),
errhint("Values larger than 1/3 of a buffer page cannot be indexed.\n"
"Consider a function index of an MD5 hash of the value, "
"or use full text indexing."),
errtableconstraint(wstate->heap,
RelationGetRelationName(wstate->index))));
/*
* Check to see if page is "full". It's definitely full if the item won't
* fit. Otherwise, compare to the target freespace derived from the
* fillfactor. However, we must put at least two items on each page, so
* disregard fillfactor if we don't have that many.
*/
if (pgspc < itupsz || (pgspc < state->btps_full && last_off > P_FIRSTKEY))
{
/*
* Finish off the page and write it out.
*/
Page opage = npage;
BlockNumber oblkno = nblkno;
ItemId ii;
ItemId hii;
IndexTuple oitup;
/* Create new page of same level */
npage = _bt_blnewpage(state->btps_level);
/* and assign it a page position */
nblkno = wstate->btws_pages_alloced++;
/*
* We copy the last item on the page into the new page, and then
* rearrange the old page so that the 'last item' becomes its high key
* rather than a true data item. There had better be at least two
* items on the page already, else the page would be empty of useful
* data.
*/
Assert(last_off > P_FIRSTKEY);
ii = PageGetItemId(opage, last_off);
oitup = (IndexTuple) PageGetItem(opage, ii);
_bt_sortaddtup(npage, ItemIdGetLength(ii), oitup, P_FIRSTKEY);
/*
* Move 'last' into the high key position on opage
*/
hii = PageGetItemId(opage, P_HIKEY);
*hii = *ii;
ItemIdSetUnused(ii); /* redundant */
((PageHeader) opage)->pd_lower -= sizeof(ItemIdData);
/*
* Link the old page into its parent, using its minimum key. If we
* don't have a parent, we have to create one; this adds a new btree
* level.
*/
if (state->btps_next == NULL)
state->btps_next = _bt_pagestate(wstate, state->btps_level + 1);
Assert(state->btps_minkey != NULL);
ItemPointerSet(&(state->btps_minkey->t_tid), oblkno, P_HIKEY);
_bt_buildadd(wstate, state->btps_next, state->btps_minkey);
pfree(state->btps_minkey);
/*
* Save a copy of the minimum key for the new page. We have to copy
* it off the old page, not the new one, in case we are not at leaf
* level.
*/
state->btps_minkey = CopyIndexTuple(oitup);
/*
* Set the sibling links for both pages.
*/
{
BTPageOpaque oopaque = (BTPageOpaque) PageGetSpecialPointer(opage);
BTPageOpaque nopaque = (BTPageOpaque) PageGetSpecialPointer(npage);
oopaque->btpo_next = nblkno;
nopaque->btpo_prev = oblkno;
nopaque->btpo_next = P_NONE; /* redundant */
}
/*
* Write out the old page. We never need to touch it again, so we can
* free the opage workspace too.
*/
_bt_blwritepage(wstate, opage, oblkno);
/*
* Reset last_off to point to new page
*/
last_off = P_FIRSTKEY;
}
/*
* If the new item is the first for its page, stash a copy for later. Note
* this will only happen for the first item on a level; on later pages,
* the first item for a page is copied from the prior page in the code
* above.
*/
if (last_off == P_HIKEY)
{
Assert(state->btps_minkey == NULL);
state->btps_minkey = CopyIndexTuple(itup);
}
/*
* Add the new item into the current page.
*/
last_off = OffsetNumberNext(last_off);
_bt_sortaddtup(npage, itupsz, itup, last_off);
state->btps_page = npage;
state->btps_blkno = nblkno;
state->btps_lastoff = last_off;
}
/*
* PageAddItemExtended
*
* Add an item to a page. Return value is the offset at which it was
* inserted, or InvalidOffsetNumber if the item is not inserted for any
* reason. A WARNING is issued indicating the reason for the refusal.
*
* If flag PAI_OVERWRITE is set, we just store the item at the specified
* offsetNumber (which must be either a currently-unused item pointer,
* or one past the last existing item). Otherwise,
* if offsetNumber is valid and <= current max offset in the page,
* insert item into the array at that position by shuffling ItemId's
* down to make room.
* If offsetNumber is not valid, then assign one by finding the first
* one that is both unused and deallocated.
*
* If flag PAI_IS_HEAP is set, we enforce that there can't be more than
* MaxHeapTuplesPerPage line pointers on the page.
*
* If flag PAI_ALLOW_FAR_OFFSET is not set, we disallow placing items
* beyond one past the last existing item.
*
* !!! EREPORT(ERROR) IS DISALLOWED HERE !!!
*/
OffsetNumber
PageAddItemExtended(Page page,
Item item,
Size size,
OffsetNumber offsetNumber,
int flags)
{
PageHeader phdr = (PageHeader) page;
Size alignedSize;
int lower;
int upper;
ItemId itemId;
OffsetNumber limit;
bool needshuffle = false;
/*
* Be wary about corrupted page pointers
*/
if (phdr->pd_lower < SizeOfPageHeaderData ||
phdr->pd_lower > phdr->pd_upper ||
phdr->pd_upper > phdr->pd_special ||
phdr->pd_special > BLCKSZ)
ereport(PANIC,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted page pointers: lower = %u, upper = %u, special = %u",
phdr->pd_lower, phdr->pd_upper, phdr->pd_special)));
/*
* Select offsetNumber to place the new item at
*/
limit = OffsetNumberNext(PageGetMaxOffsetNumber(page));
/* was offsetNumber passed in? */
if (OffsetNumberIsValid(offsetNumber))
{
/* yes, check it */
if ((flags & PAI_OVERWRITE) != 0)
{
if (offsetNumber < limit)
{
itemId = PageGetItemId(phdr, offsetNumber);
if (ItemIdIsUsed(itemId) || ItemIdHasStorage(itemId))
{
elog(WARNING, "will not overwrite a used ItemId");
return InvalidOffsetNumber;
}
}
}
else
{
if (offsetNumber < limit)
needshuffle = true; /* need to move existing linp's */
}
}
else
{
/* offsetNumber was not passed in, so find a free slot */
/* if no free slot, we'll put it at limit (1st open slot) */
if (PageHasFreeLinePointers(phdr))
{
/*
* Look for "recyclable" (unused) ItemId. We check for no storage
* as well, just to be paranoid --- unused items should never have
* storage.
*/
for (offsetNumber = 1; offsetNumber < limit; offsetNumber++)
{
itemId = PageGetItemId(phdr, offsetNumber);
if (!ItemIdIsUsed(itemId) && !ItemIdHasStorage(itemId))
break;
}
if (offsetNumber >= limit)
{
/* the hint is wrong, so reset it */
PageClearHasFreeLinePointers(phdr);
}
}
else
{
/* don't bother searching if hint says there's no free slot */
offsetNumber = limit;
}
}
/*
* Reject placing items beyond the first unused line pointer, unless
* caller asked for that behavior specifically.
*/
if ((flags & PAI_ALLOW_FAR_OFFSET) == 0 && offsetNumber > limit)
{
elog(WARNING, "specified item offset is too large");
return InvalidOffsetNumber;
}
/* Reject placing items beyond heap boundary, if heap */
if ((flags & PAI_IS_HEAP) != 0 && offsetNumber > MaxHeapTuplesPerPage)
{
elog(WARNING, "can't put more than MaxHeapTuplesPerPage items in a heap page");
return InvalidOffsetNumber;
}
/*
* Compute new lower and upper pointers for page, see if it'll fit.
*
* Note: do arithmetic as signed ints, to avoid mistakes if, say,
* alignedSize > pd_upper.
*/
if ((flags & PAI_ALLOW_FAR_OFFSET) != 0)
lower = Max(phdr->pd_lower,
SizeOfPageHeaderData + sizeof(ItemIdData) * offsetNumber);
else if (offsetNumber == limit || needshuffle)
lower = phdr->pd_lower + sizeof(ItemIdData);
else
lower = phdr->pd_lower;
alignedSize = MAXALIGN(size);
upper = (int) phdr->pd_upper - (int) alignedSize;
if (lower > upper)
return InvalidOffsetNumber;
/*
* OK to insert the item. First, shuffle the existing pointers if needed.
*/
itemId = PageGetItemId(phdr, offsetNumber);
if (needshuffle)
memmove(itemId + 1, itemId,
(limit - offsetNumber) * sizeof(ItemIdData));
/* set the item pointer */
ItemIdSetNormal(itemId, upper, size);
/*
* Items normally contain no uninitialized bytes. Core bufpage consumers
* conform, but this is not a necessary coding rule; a new index AM could
* opt to depart from it. However, data type input functions and other
* C-language functions that synthesize datums should initialize all
* bytes; datumIsEqual() relies on this. Testing here, along with the
* similar check in printtup(), helps to catch such mistakes.
*
* Values of the "name" type retrieved via index-only scans may contain
* uninitialized bytes; see comment in btrescan(). Valgrind will report
* this as an error, but it is safe to ignore.
*/
VALGRIND_CHECK_MEM_IS_DEFINED(item, size);
/* copy the item's data onto the page */
memcpy((char *) page + upper, item, size);
/* adjust page header */
phdr->pd_lower = (LocationIndex) lower;
phdr->pd_upper = (LocationIndex) upper;
return offsetNumber;
}
/*
* PGSharedMemoryCreate
*
* Create a shared memory segment of the given size and initialize its
* standard header. Also, register an on_shmem_exit callback to release
* the storage.
*
* Dead Postgres segments are recycled if found, but we do not fail upon
* collision with non-Postgres shmem segments. The idea here is to detect and
* re-use keys that may have been assigned by a crashed postmaster or backend.
*
* makePrivate means to always create a new segment, rather than attach to
* or recycle any existing segment.
*
* The port number is passed for possible use as a key (for SysV, we use
* it to generate the starting shmem key). In a standalone backend,
* zero will be passed.
*/
PGShmemHeader *
PGSharedMemoryCreate(Size size, bool makePrivate, int port,
PGShmemHeader **shim)
{
IpcMemoryKey NextShmemSegID;
void *memAddress;
PGShmemHeader *hdr;
IpcMemoryId shmid;
struct stat statbuf;
Size sysvsize;
/* Complain if hugepages demanded but we can't possibly support them */
#if !defined(USE_ANONYMOUS_SHMEM) || !defined(MAP_HUGETLB)
if (huge_pages == HUGE_PAGES_ON)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("huge pages not supported on this platform")));
#endif
/* Room for a header? */
Assert(size > MAXALIGN(sizeof(PGShmemHeader)));
#ifdef USE_ANONYMOUS_SHMEM
AnonymousShmem = CreateAnonymousSegment(&size);
AnonymousShmemSize = size;
/* Register on-exit routine to unmap the anonymous segment */
on_shmem_exit(AnonymousShmemDetach, (Datum) 0);
/* Now we need only allocate a minimal-sized SysV shmem block. */
sysvsize = sizeof(PGShmemHeader);
#else
sysvsize = size;
#endif
/* Make sure PGSharedMemoryAttach doesn't fail without need */
UsedShmemSegAddr = NULL;
/* Loop till we find a free IPC key */
NextShmemSegID = port * 1000;
for (NextShmemSegID++;; NextShmemSegID++)
{
/* Try to create new segment */
memAddress = InternalIpcMemoryCreate(NextShmemSegID, sysvsize);
if (memAddress)
break; /* successful create and attach */
/* Check shared memory and possibly remove and recreate */
if (makePrivate) /* a standalone backend shouldn't do this */
continue;
if ((memAddress = PGSharedMemoryAttach(NextShmemSegID, &shmid)) == NULL)
continue; /* can't attach, not one of mine */
/*
* If I am not the creator and it belongs to an extant process,
* continue.
*/
hdr = (PGShmemHeader *) memAddress;
if (hdr->creatorPID != getpid())
{
if (kill(hdr->creatorPID, 0) == 0 || errno != ESRCH)
{
shmdt(memAddress);
continue; /* segment belongs to a live process */
}
}
/*
* The segment appears to be from a dead Postgres process, or from a
* previous cycle of life in this same process. Zap it, if possible,
* and any associated dynamic shared memory segments, as well. This
* probably shouldn't fail, but if it does, assume the segment belongs
* to someone else after all, and continue quietly.
*/
if (hdr->dsm_control != 0)
dsm_cleanup_using_control_segment(hdr->dsm_control);
shmdt(memAddress);
if (shmctl(shmid, IPC_RMID, NULL) < 0)
continue;
/*
* Now try again to create the segment.
*/
memAddress = InternalIpcMemoryCreate(NextShmemSegID, sysvsize);
if (memAddress)
break; /* successful create and attach */
/*
* Can only get here if some other process managed to create the same
* shmem key before we did. Let him have that one, loop around to try
* next key.
*/
}
/*
* OK, we created a new segment. Mark it as created by this process. The
* order of assignments here is critical so that another Postgres process
* can't see the header as valid but belonging to an invalid PID!
*/
hdr = (PGShmemHeader *) memAddress;
hdr->creatorPID = getpid();
hdr->magic = PGShmemMagic;
hdr->dsm_control = 0;
/* Fill in the data directory ID info, too */
if (stat(DataDir, &statbuf) < 0)
ereport(FATAL,
(errcode_for_file_access(),
errmsg("could not stat data directory \"%s\": %m",
DataDir)));
hdr->device = statbuf.st_dev;
hdr->inode = statbuf.st_ino;
/*
* Initialize space allocation status for segment.
*/
hdr->totalsize = size;
hdr->freeoffset = MAXALIGN(sizeof(PGShmemHeader));
*shim = hdr;
/* Save info for possible future use */
UsedShmemSegAddr = memAddress;
UsedShmemSegID = (unsigned long) NextShmemSegID;
/*
* If AnonymousShmem is NULL here, then we're not using anonymous shared
* memory, and should return a pointer to the System V shared memory
* block. Otherwise, the System V shared memory block is only a shim, and
* we must return a pointer to the real block.
*/
#ifdef USE_ANONYMOUS_SHMEM
if (AnonymousShmem == NULL)
return hdr;
memcpy(AnonymousShmem, hdr, sizeof(PGShmemHeader));
return (PGShmemHeader *) AnonymousShmem;
#else
return hdr;
#endif
}
/*
* PageRepairFragmentation
*
* Frees fragmented space on a page.
* It doesn't remove unused line pointers! Please don't change this.
*
* This routine is usable for heap pages only, but see PageIndexMultiDelete.
*
* As a side effect, the page's PD_HAS_FREE_LINES hint bit is updated.
*/
void
PageRepairFragmentation(Page page)
{
Offset pd_lower = ((PageHeader) page)->pd_lower;
Offset pd_upper = ((PageHeader) page)->pd_upper;
Offset pd_special = ((PageHeader) page)->pd_special;
ItemId lp;
int nline,
nstorage,
nunused;
int i;
Size totallen;
/*
* It's worth the trouble to be more paranoid here than in most places,
* because we are about to reshuffle data in (what is usually) a shared
* disk buffer. If we aren't careful then corrupted pointers, lengths,
* etc could cause us to clobber adjacent disk buffers, spreading the data
* loss further. So, check everything.
*/
if (pd_lower < SizeOfPageHeaderData ||
pd_lower > pd_upper ||
pd_upper > pd_special ||
pd_special > BLCKSZ ||
pd_special != MAXALIGN(pd_special))
ereport(ERROR,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted page pointers: lower = %u, upper = %u, special = %u",
pd_lower, pd_upper, pd_special)));
nline = PageGetMaxOffsetNumber(page);
nunused = nstorage = 0;
for (i = FirstOffsetNumber; i <= nline; i++)
{
lp = PageGetItemId(page, i);
if (ItemIdIsUsed(lp))
{
if (ItemIdHasStorage(lp))
nstorage++;
}
else
{
/* Unused entries should have lp_len = 0, but make sure */
ItemIdSetUnused(lp);
nunused++;
}
}
if (nstorage == 0)
{
/* Page is completely empty, so just reset it quickly */
((PageHeader) page)->pd_upper = pd_special;
}
else
{
/* Need to compact the page the hard way */
itemIdSortData itemidbase[MaxHeapTuplesPerPage];
itemIdSort itemidptr = itemidbase;
totallen = 0;
for (i = 0; i < nline; i++)
{
lp = PageGetItemId(page, i + 1);
if (ItemIdHasStorage(lp))
{
itemidptr->offsetindex = i;
itemidptr->itemoff = ItemIdGetOffset(lp);
if (itemidptr->itemoff < (int) pd_upper ||
itemidptr->itemoff >= (int) pd_special)
ereport(ERROR,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted item pointer: %u",
itemidptr->itemoff)));
itemidptr->alignedlen = MAXALIGN(ItemIdGetLength(lp));
totallen += itemidptr->alignedlen;
itemidptr++;
}
}
if (totallen > (Size) (pd_special - pd_lower))
ereport(ERROR,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted item lengths: total %u, available space %u",
(unsigned int) totallen, pd_special - pd_lower)));
compactify_tuples(itemidbase, nstorage, page);
}
/* Set hint bit for PageAddItem */
if (nunused > 0)
PageSetHasFreeLinePointers(page);
else
PageClearHasFreeLinePointers(page);
}
void
SimpleLruInit(SlruCtl ctl, const char *name, int nslots, int nlsns,
LWLock *ctllock, const char *subdir)
{
SlruShared shared;
bool found;
shared = (SlruShared) ShmemInitStruct(name,
SimpleLruShmemSize(nslots, nlsns),
&found);
if (!IsUnderPostmaster)
{
/* Initialize locks and shared memory area */
char *ptr;
Size offset;
int slotno;
Assert(!found);
memset(shared, 0, sizeof(SlruSharedData));
shared->ControlLock = ctllock;
shared->num_slots = nslots;
shared->lsn_groups_per_page = nlsns;
shared->cur_lru_count = 0;
/* shared->latest_page_number will be set later */
ptr = (char *) shared;
offset = MAXALIGN(sizeof(SlruSharedData));
shared->page_buffer = (char **) (ptr + offset);
offset += MAXALIGN(nslots * sizeof(char *));
shared->page_status = (SlruPageStatus *) (ptr + offset);
offset += MAXALIGN(nslots * sizeof(SlruPageStatus));
shared->page_dirty = (bool *) (ptr + offset);
offset += MAXALIGN(nslots * sizeof(bool));
shared->page_number = (int *) (ptr + offset);
offset += MAXALIGN(nslots * sizeof(int));
shared->page_lru_count = (int *) (ptr + offset);
offset += MAXALIGN(nslots * sizeof(int));
shared->buffer_locks = (LWLock **) (ptr + offset);
offset += MAXALIGN(nslots * sizeof(LWLock *));
if (nlsns > 0)
{
shared->group_lsn = (XLogRecPtr *) (ptr + offset);
offset += MAXALIGN(nslots * nlsns * sizeof(XLogRecPtr));
}
ptr += BUFFERALIGN(offset);
for (slotno = 0; slotno < nslots; slotno++)
{
shared->page_buffer[slotno] = ptr;
shared->page_status[slotno] = SLRU_PAGE_EMPTY;
shared->page_dirty[slotno] = false;
shared->page_lru_count[slotno] = 0;
shared->buffer_locks[slotno] = LWLockAssign();
ptr += BLCKSZ;
}
}
else
Assert(found);
/*
* Initialize the unshared control struct, including directory path. We
* assume caller set PagePrecedes.
*/
ctl->shared = shared;
ctl->do_fsync = true; /* default behavior */
StrNCpy(ctl->Dir, subdir, sizeof(ctl->Dir));
}
/*
* PageIsVerified
* Check that the page header and checksum (if any) appear valid.
*
* This is called when a page has just been read in from disk. The idea is
* to cheaply detect trashed pages before we go nuts following bogus item
* pointers, testing invalid transaction identifiers, etc.
*
* It turns out to be necessary to allow zeroed pages here too. Even though
* this routine is *not* called when deliberately adding a page to a relation,
* there are scenarios in which a zeroed page might be found in a table.
* (Example: a backend extends a relation, then crashes before it can write
* any WAL entry about the new page. The kernel will already have the
* zeroed page in the file, and it will stay that way after restart.) So we
* allow zeroed pages here, and are careful that the page access macros
* treat such a page as empty and without free space. Eventually, VACUUM
* will clean up such a page and make it usable.
*/
bool
PageIsVerified(Page page, BlockNumber blkno)
{
PageHeader p = (PageHeader) page;
char *pagebytes;
int i;
bool checksum_failure = false;
bool header_sane = false;
bool all_zeroes = false;
uint16 checksum = 0;
/*
* Don't verify page data unless the page passes basic non-zero test
*/
if (!PageIsNew(page))
{
if (DataChecksumsEnabled())
{
checksum = pg_checksum_page((char *) page, blkno);
if (checksum != p->pd_checksum)
checksum_failure = true;
}
/*
* The following checks don't prove the header is correct, only that
* it looks sane enough to allow into the buffer pool. Later usage of
* the block can still reveal problems, which is why we offer the
* checksum option.
*/
if ((p->pd_flags & ~PD_VALID_FLAG_BITS) == 0 &&
p->pd_lower <= p->pd_upper &&
p->pd_upper <= p->pd_special &&
p->pd_special <= BLCKSZ &&
p->pd_special == MAXALIGN(p->pd_special))
header_sane = true;
if (header_sane && !checksum_failure)
return true;
}
/* Check all-zeroes case */
all_zeroes = true;
pagebytes = (char *) page;
for (i = 0; i < BLCKSZ; i++)
{
if (pagebytes[i] != 0)
{
all_zeroes = false;
break;
}
}
if (all_zeroes)
return true;
/*
* Throw a WARNING if the checksum fails, but only after we've checked for
* the all-zeroes case.
*/
if (checksum_failure)
{
ereport(WARNING,
(ERRCODE_DATA_CORRUPTED,
errmsg("page verification failed, calculated checksum %u but expected %u",
checksum, p->pd_checksum)));
if (header_sane && ignore_checksum_failure)
return true;
}
return false;
}
/*
* Place tuple and split page, original buffer(lbuf) leaves untouched,
* returns shadow page of lbuf filled new data.
* Tuples are distributed between pages by equal size on its, not
* an equal number!
*/
static Page
entrySplitPage(GinBtree btree, Buffer lbuf, Buffer rbuf, OffsetNumber off, XLogRecData **prdata)
{
static XLogRecData rdata[2];
OffsetNumber i,
maxoff,
separator = InvalidOffsetNumber;
Size totalsize = 0;
Size lsize = 0,
size;
static char tupstore[2 * BLCKSZ];
char *ptr;
IndexTuple itup,
leftrightmost = NULL;
static ginxlogSplit data;
Page page;
Page lpage = PageGetTempPageCopy(BufferGetPage(lbuf));
Page rpage = BufferGetPage(rbuf);
Size pageSize = PageGetPageSize(lpage);
*prdata = rdata;
data.leftChildBlkno = (GinPageIsLeaf(lpage)) ?
InvalidOffsetNumber : GinItemPointerGetBlockNumber(&(btree->entry->t_tid));
data.updateBlkno = entryPreparePage(btree, lpage, off);
maxoff = PageGetMaxOffsetNumber(lpage);
ptr = tupstore;
for (i = FirstOffsetNumber; i <= maxoff; i++)
{
if (i == off)
{
size = MAXALIGN(IndexTupleSize(btree->entry));
memcpy(ptr, btree->entry, size);
ptr += size;
totalsize += size + sizeof(ItemIdData);
}
itup = (IndexTuple) PageGetItem(lpage, PageGetItemId(lpage, i));
size = MAXALIGN(IndexTupleSize(itup));
memcpy(ptr, itup, size);
ptr += size;
totalsize += size + sizeof(ItemIdData);
}
if (off == maxoff + 1)
{
size = MAXALIGN(IndexTupleSize(btree->entry));
memcpy(ptr, btree->entry, size);
ptr += size;
totalsize += size + sizeof(ItemIdData);
}
GinInitPage(rpage, GinPageGetOpaque(lpage)->flags, pageSize);
GinInitPage(lpage, GinPageGetOpaque(rpage)->flags, pageSize);
ptr = tupstore;
maxoff++;
lsize = 0;
page = lpage;
for (i = FirstOffsetNumber; i <= maxoff; i++)
{
itup = (IndexTuple) ptr;
if (lsize > totalsize / 2)
{
if (separator == InvalidOffsetNumber)
separator = i - 1;
page = rpage;
}
else
{
leftrightmost = itup;
lsize += MAXALIGN(IndexTupleSize(itup)) + sizeof(ItemIdData);
}
if (PageAddItem(page, (Item) itup, IndexTupleSize(itup), InvalidOffsetNumber, false, false) == InvalidOffsetNumber)
elog(ERROR, "failed to add item to index page in \"%s\"",
RelationGetRelationName(btree->index));
ptr += MAXALIGN(IndexTupleSize(itup));
}
btree->entry = copyIndexTuple(leftrightmost, lpage);
ItemPointerSet(&(btree->entry)->t_tid, BufferGetBlockNumber(lbuf), InvalidOffsetNumber);
btree->rightblkno = BufferGetBlockNumber(rbuf);
data.node = btree->index->rd_node;
data.rootBlkno = InvalidBlockNumber;
data.lblkno = BufferGetBlockNumber(lbuf);
data.rblkno = BufferGetBlockNumber(rbuf);
data.separator = separator;
data.nitem = maxoff;
data.isData = FALSE;
data.isLeaf = GinPageIsLeaf(lpage) ? TRUE : FALSE;
data.isRootSplit = FALSE;
rdata[0].buffer = InvalidBuffer;
rdata[0].data = (char *) &data;
rdata[0].len = sizeof(ginxlogSplit);
rdata[0].next = &rdata[1];
rdata[1].buffer = InvalidBuffer;
rdata[1].data = tupstore;
rdata[1].len = MAXALIGN(totalsize);
rdata[1].next = NULL;
return lpage;
}
/*
* PageIndexDeleteNoCompact
* Delete the given items for an index page, and defragment the resulting
* free space, but do not compact the item pointers array.
*
* itemnos is the array of tuples to delete; nitems is its size. maxIdxTuples
* is the maximum number of tuples that can exist in a page.
*
* Unused items at the end of the array are removed.
*
* This is used for index AMs that require that existing TIDs of live tuples
* remain unchanged.
*/
void
PageIndexDeleteNoCompact(Page page, OffsetNumber *itemnos, int nitems)
{
PageHeader phdr = (PageHeader) page;
LocationIndex pd_lower = phdr->pd_lower;
LocationIndex pd_upper = phdr->pd_upper;
LocationIndex pd_special = phdr->pd_special;
int nline;
bool empty;
OffsetNumber offnum;
int nextitm;
/*
* As with PageRepairFragmentation, paranoia seems justified.
*/
if (pd_lower < SizeOfPageHeaderData ||
pd_lower > pd_upper ||
pd_upper > pd_special ||
pd_special > BLCKSZ ||
pd_special != MAXALIGN(pd_special))
ereport(ERROR,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted page pointers: lower = %u, upper = %u, special = %u",
pd_lower, pd_upper, pd_special)));
/*
* Scan the existing item pointer array and mark as unused those that are
* in our kill-list; make sure any non-interesting ones are marked unused
* as well.
*/
nline = PageGetMaxOffsetNumber(page);
empty = true;
nextitm = 0;
for (offnum = FirstOffsetNumber; offnum <= nline; offnum = OffsetNumberNext(offnum))
{
ItemId lp;
ItemLength itemlen;
ItemOffset offset;
lp = PageGetItemId(page, offnum);
itemlen = ItemIdGetLength(lp);
offset = ItemIdGetOffset(lp);
if (ItemIdIsUsed(lp))
{
if (offset < pd_upper ||
(offset + itemlen) > pd_special ||
offset != MAXALIGN(offset))
ereport(ERROR,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted item pointer: offset = %u, length = %u",
offset, (unsigned int) itemlen)));
if (nextitm < nitems && offnum == itemnos[nextitm])
{
/* this one is on our list to delete, so mark it unused */
ItemIdSetUnused(lp);
nextitm++;
}
else if (ItemIdHasStorage(lp))
{
/* This one's live -- must do the compaction dance */
empty = false;
}
else
{
/* get rid of this one too */
ItemIdSetUnused(lp);
}
}
}
/* this will catch invalid or out-of-order itemnos[] */
if (nextitm != nitems)
elog(ERROR, "incorrect index offsets supplied");
if (empty)
{
/* Page is completely empty, so just reset it quickly */
phdr->pd_lower = SizeOfPageHeaderData;
phdr->pd_upper = pd_special;
}
else
{
/* There are live items: need to compact the page the hard way */
itemIdSortData itemidbase[MaxOffsetNumber];
itemIdSort itemidptr;
int i;
Size totallen;
/*
* Scan the page taking note of each item that we need to preserve.
* This includes both live items (those that contain data) and
* interspersed unused ones. It's critical to preserve these unused
* items, because otherwise the offset numbers for later live items
* would change, which is not acceptable. Unused items might get used
* again later; that is fine.
*/
itemidptr = itemidbase;
totallen = 0;
PageClearHasFreeLinePointers(page);
for (i = 0; i < nline; i++)
{
ItemId lp;
itemidptr->offsetindex = i;
lp = PageGetItemId(page, i + 1);
if (ItemIdHasStorage(lp))
{
itemidptr->itemoff = ItemIdGetOffset(lp);
itemidptr->alignedlen = MAXALIGN(ItemIdGetLength(lp));
totallen += itemidptr->alignedlen;
itemidptr++;
}
else
{
PageSetHasFreeLinePointers(page);
ItemIdSetUnused(lp);
}
}
nline = itemidptr - itemidbase;
/* By here, there are exactly nline elements in itemidbase array */
if (totallen > (Size) (pd_special - pd_lower))
ereport(ERROR,
(errcode(ERRCODE_DATA_CORRUPTED),
errmsg("corrupted item lengths: total %u, available space %u",
(unsigned int) totallen, pd_special - pd_lower)));
/*
* Defragment the data areas of each tuple, being careful to preserve
* each item's position in the linp array.
*/
compactify_tuples(itemidbase, nline, page);
}
}
/*
* Form a tuple for entry tree.
*
* If the tuple would be too big to be stored, function throws a suitable
* error if errorTooBig is TRUE, or returns NULL if errorTooBig is FALSE.
*
* 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(Relation index, GinState *ginstate,
OffsetNumber attnum, Datum key,
ItemPointerData *ipd, uint32 nipd, bool errorTooBig)
{
bool isnull[2] = {FALSE, FALSE};
IndexTuple itup;
uint32 newsize;
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)
{
newsize = MAXALIGN(SHORTALIGN(IndexTupleSize(itup)) + sizeof(ItemPointerData) * nipd);
if (newsize > Min(INDEX_SIZE_MASK, GinMaxItemSize))
{
if (errorTooBig)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("index row size %lu exceeds maximum %lu for index \"%s\"",
(unsigned long) newsize,
(unsigned long) Min(INDEX_SIZE_MASK,
GinMaxItemSize),
RelationGetRelationName(index))));
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
{
/*
* Gin tuple without any ItemPointers should be large enough to keep
* one ItemPointer, to prevent inconsistency between
* ginHeapTupleFastCollect and ginEntryInsert called by
* ginHeapTupleInsert. ginHeapTupleFastCollect forms tuple without
* extra pointer to heap, but ginEntryInsert (called for pending list
* cleanup during vacuum) will form the same tuple with one
* ItemPointer.
*/
newsize = MAXALIGN(SHORTALIGN(IndexTupleSize(itup)) + sizeof(ItemPointerData));
if (newsize > Min(INDEX_SIZE_MASK, GinMaxItemSize))
{
if (errorTooBig)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("index row size %lu exceeds maximum %lu for index \"%s\"",
(unsigned long) newsize,
(unsigned long) Min(INDEX_SIZE_MASK,
GinMaxItemSize),
RelationGetRelationName(index))));
return NULL;
}
GinSetNPosting(itup, 0);
}
return itup;
}
static Size
PersistentFilespace_SharedDataSize(void)
{
return MAXALIGN(sizeof(PersistentFilespaceSharedData));
}
/*
* Insert an index tuple into the index relation. The revmap is updated to
* mark the range containing the given page as pointing to the inserted entry.
* A WAL record is written.
*
* The buffer, if valid, is first checked for free space to insert the new
* entry; if there isn't enough, a new buffer is obtained and pinned. No
* buffer lock must be held on entry, no buffer lock is held on exit.
*
* Return value is the offset number where the tuple was inserted.
*/
OffsetNumber
brin_doinsert(Relation idxrel, BlockNumber pagesPerRange,
BrinRevmap *revmap, Buffer *buffer, BlockNumber heapBlk,
BrinTuple *tup, Size itemsz)
{
Page page;
BlockNumber blk;
OffsetNumber off;
Buffer revmapbuf;
ItemPointerData tid;
bool extended;
Assert(itemsz == MAXALIGN(itemsz));
/* If the item is oversized, don't even bother. */
if (itemsz > BrinMaxItemSize)
{
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("index row size %lu exceeds maximum %lu for index \"%s\"",
(unsigned long) itemsz,
(unsigned long) BrinMaxItemSize,
RelationGetRelationName(idxrel))));
return InvalidOffsetNumber; /* keep compiler quiet */
}
/* Make sure the revmap is long enough to contain the entry we need */
brinRevmapExtend(revmap, heapBlk);
/*
* Acquire lock on buffer supplied by caller, if any. If it doesn't have
* enough space, unpin it to obtain a new one below.
*/
if (BufferIsValid(*buffer))
{
/*
* It's possible that another backend (or ourselves!) extended the
* revmap over the page we held a pin on, so we cannot assume that
* it's still a regular page.
*/
LockBuffer(*buffer, BUFFER_LOCK_EXCLUSIVE);
if (br_page_get_freespace(BufferGetPage(*buffer)) < itemsz)
{
UnlockReleaseBuffer(*buffer);
*buffer = InvalidBuffer;
}
}
/*
* If we still don't have a usable buffer, have brin_getinsertbuffer
* obtain one for us.
*/
if (!BufferIsValid(*buffer))
{
do
*buffer = brin_getinsertbuffer(idxrel, InvalidBuffer, itemsz, &extended);
while (!BufferIsValid(*buffer));
}
else
extended = false;
/* Now obtain lock on revmap buffer */
revmapbuf = brinLockRevmapPageForUpdate(revmap, heapBlk);
page = BufferGetPage(*buffer);
blk = BufferGetBlockNumber(*buffer);
/* Execute the actual insertion */
START_CRIT_SECTION();
if (extended)
brin_page_init(BufferGetPage(*buffer), BRIN_PAGETYPE_REGULAR);
off = PageAddItem(page, (Item) tup, itemsz, InvalidOffsetNumber,
false, false);
if (off == InvalidOffsetNumber)
elog(ERROR, "could not insert new index tuple to page");
MarkBufferDirty(*buffer);
BRIN_elog((DEBUG2, "inserted tuple (%u,%u) for range starting at %u",
blk, off, heapBlk));
ItemPointerSet(&tid, blk, off);
brinSetHeapBlockItemptr(revmapbuf, pagesPerRange, heapBlk, tid);
MarkBufferDirty(revmapbuf);
/* XLOG stuff */
if (RelationNeedsWAL(idxrel))
{
xl_brin_insert xlrec;
XLogRecPtr recptr;
uint8 info;
info = XLOG_BRIN_INSERT | (extended ? XLOG_BRIN_INIT_PAGE : 0);
xlrec.heapBlk = heapBlk;
xlrec.pagesPerRange = pagesPerRange;
xlrec.offnum = off;
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfBrinInsert);
XLogRegisterBuffer(0, *buffer, REGBUF_STANDARD | (extended ? REGBUF_WILL_INIT : 0));
XLogRegisterBufData(0, (char *) tup, itemsz);
XLogRegisterBuffer(1, revmapbuf, 0);
recptr = XLogInsert(RM_BRIN_ID, info);
PageSetLSN(page, recptr);
PageSetLSN(BufferGetPage(revmapbuf), recptr);
}
END_CRIT_SECTION();
/* Tuple is firmly on buffer; we can release our locks */
LockBuffer(*buffer, BUFFER_LOCK_UNLOCK);
LockBuffer(revmapbuf, BUFFER_LOCK_UNLOCK);
if (extended)
FreeSpaceMapVacuum(idxrel);
return off;
}
