#include "access/xlogreader.h"
#include "access/xlogrecord.h"
+#include "access/xlog_internal.h"
#include "access/transam.h"
#include "common/fe_memutils.h"
#include "getopt_long.h"
* Store per-rmgr and per-record statistics for a given record.
*/
static void
-XLogDumpCountRecord(XLogDumpConfig *config, XLogDumpStats *stats, XLogRecPtr ReadRecPtr, XLogRecord *record)
+XLogDumpCountRecord(XLogDumpConfig *config, XLogDumpStats *stats,
+ XLogReaderState *record)
{
RmgrId rmid;
uint8 recid;
+ uint32 rec_len;
+ uint32 fpi_len;
stats->count++;
/* Update per-rmgr statistics */
- rmid = record->xl_rmid;
+ rmid = XLogRecGetRmid(record);
+ rec_len = XLogRecGetDataLen(record) + SizeOfXLogRecord;
+ fpi_len = record->decoded_record->xl_tot_len - rec_len;
stats->rmgr_stats[rmid].count++;
- stats->rmgr_stats[rmid].rec_len +=
- record->xl_len + SizeOfXLogRecord;
- stats->rmgr_stats[rmid].fpi_len +=
- record->xl_tot_len - (record->xl_len + SizeOfXLogRecord);
+ stats->rmgr_stats[rmid].rec_len += rec_len;
+ stats->rmgr_stats[rmid].fpi_len += fpi_len;
/*
* Update per-record statistics, where the record is identified by a
- * combination of the RmgrId and the four bits of the xl_info field
- * that are the rmgr's domain (resulting in sixteen possible entries
- * per RmgrId).
+ * combination of the RmgrId and the four bits of the xl_info field that
+ * are the rmgr's domain (resulting in sixteen possible entries per
+ * RmgrId).
*/
- recid = record->xl_info >> 4;
+ recid = XLogRecGetInfo(record) >> 4;
stats->record_stats[rmid][recid].count++;
- stats->record_stats[rmid][recid].rec_len +=
- record->xl_len + SizeOfXLogRecord;
- stats->record_stats[rmid][recid].fpi_len +=
- record->xl_tot_len - (record->xl_len + SizeOfXLogRecord);
+ stats->record_stats[rmid][recid].rec_len += rec_len;
+ stats->record_stats[rmid][recid].fpi_len += fpi_len;
}
/*
* Print a record to stdout
*/
static void
-XLogDumpDisplayRecord(XLogDumpConfig *config, XLogRecPtr ReadRecPtr, XLogRecord *record)
+XLogDumpDisplayRecord(XLogDumpConfig *config, XLogReaderState *record)
{
- const char *id;
- const RmgrDescData *desc = &RmgrDescTable[record->xl_rmid];
-
- id = desc->rm_identify(record->xl_info);
+ const char *id;
+ const RmgrDescData *desc = &RmgrDescTable[XLogRecGetRmid(record)];
+ RelFileNode rnode;
+ ForkNumber forknum;
+ BlockNumber blk;
+ int block_id;
+ uint8 info = XLogRecGetInfo(record);
+ XLogRecPtr xl_prev = XLogRecGetPrev(record);
+
+ id = desc->rm_identify(info);
if (id == NULL)
- id = psprintf("UNKNOWN (%x)", record->xl_info & ~XLR_INFO_MASK);
+ id = psprintf("UNKNOWN (%x)", info & ~XLR_INFO_MASK);
- printf("rmgr: %-11s len (rec/tot): %6u/%6u, tx: %10u, lsn: %X/%08X, prev %X/%08X, bkp: %u%u%u%u, desc: %s ",
+ printf("rmgr: %-11s len (rec/tot): %6u/%6u, tx: %10u, lsn: %X/%08X, prev %X/%08X, ",
desc->rm_name,
- record->xl_len, record->xl_tot_len,
- record->xl_xid,
- (uint32) (ReadRecPtr >> 32), (uint32) ReadRecPtr,
- (uint32) (record->xl_prev >> 32), (uint32) record->xl_prev,
- !!(XLR_BKP_BLOCK(0) & record->xl_info),
- !!(XLR_BKP_BLOCK(1) & record->xl_info),
- !!(XLR_BKP_BLOCK(2) & record->xl_info),
- !!(XLR_BKP_BLOCK(3) & record->xl_info),
- id);
+ XLogRecGetDataLen(record), XLogRecGetTotalLen(record),
+ XLogRecGetXid(record),
+ (uint32) (record->ReadRecPtr >> 32), (uint32) record->ReadRecPtr,
+ (uint32) (xl_prev >> 32), (uint32) xl_prev);
+ printf("desc: %s ", id);
/* the desc routine will printf the description directly to stdout */
desc->rm_desc(NULL, record);
- putchar('\n');
-
- if (config->bkp_details)
+ if (!config->bkp_details)
{
- int bkpnum;
- char *blk = (char *) XLogRecGetData(record) + record->xl_len;
-
- for (bkpnum = 0; bkpnum < XLR_MAX_BKP_BLOCKS; bkpnum++)
+ /* print block references (short format) */
+ for (block_id = 0; block_id <= record->max_block_id; block_id++)
{
- BkpBlock bkpb;
-
- if (!(XLR_BKP_BLOCK(bkpnum) & record->xl_info))
+ if (!XLogRecHasBlockRef(record, block_id))
continue;
- memcpy(&bkpb, blk, sizeof(BkpBlock));
- blk += sizeof(BkpBlock);
- blk += BLCKSZ - bkpb.hole_length;
+ XLogRecGetBlockTag(record, block_id, &rnode, &forknum, &blk);
+ if (forknum != MAIN_FORKNUM)
+ printf(", blkref #%u: rel %u/%u/%u fork %s blk %u",
+ block_id,
+ rnode.spcNode, rnode.dbNode, rnode.relNode,
+ forkNames[forknum],
+ blk);
+ else
+ printf(", blkref #%u: rel %u/%u/%u blk %u",
+ block_id,
+ rnode.spcNode, rnode.dbNode, rnode.relNode,
+ blk);
+ if (XLogRecHasBlockImage(record, block_id))
+ printf(" FPW");
+ }
+ putchar('\n');
+ }
+ else
+ {
+ /* print block references (detailed format) */
+ putchar('\n');
+ for (block_id = 0; block_id <= record->max_block_id; block_id++)
+ {
+ if (!XLogRecHasBlockRef(record, block_id))
+ continue;
- printf("\tbackup bkp #%u; rel %u/%u/%u; fork: %s; block: %u; hole: offset: %u, length: %u\n",
- bkpnum,
- bkpb.node.spcNode, bkpb.node.dbNode, bkpb.node.relNode,
- forkNames[bkpb.fork],
- bkpb.block, bkpb.hole_offset, bkpb.hole_length);
+ XLogRecGetBlockTag(record, block_id, &rnode, &forknum, &blk);
+ printf("\tblkref #%u: rel %u/%u/%u fork %s blk %u",
+ block_id,
+ rnode.spcNode, rnode.dbNode, rnode.relNode,
+ forkNames[forknum],
+ blk);
+ if (XLogRecHasBlockImage(record, block_id))
+ {
+ printf(" (FPW); hole: offset: %u, length: %u\n",
+ record->blocks[block_id].hole_offset,
+ record->blocks[block_id].hole_length);
+ }
+ putchar('\n');
}
}
}
/* process the record */
if (config.stats == true)
- XLogDumpCountRecord(&config, &stats, xlogreader_state->ReadRecPtr, record);
+ XLogDumpCountRecord(&config, &stats, xlogreader_state);
else
- XLogDumpDisplayRecord(&config, xlogreader_state->ReadRecPtr, record);
+ XLogDumpDisplayRecord(&config, xlogreader_state);
/* check whether we printed enough */
config.already_displayed_records++;
typedef struct RmgrDescData
{
const char *rm_name;
- void (*rm_desc) (StringInfo buf, XLogRecord *record);
+ void (*rm_desc) (StringInfo buf, XLogReaderState *record);
const char *(*rm_identify) (uint8 info);
} RmgrDescData;
{
xl_brin_createidx xlrec;
XLogRecPtr recptr;
- XLogRecData rdata;
Page page;
- xlrec.node = index->rd_node;
xlrec.version = BRIN_CURRENT_VERSION;
xlrec.pagesPerRange = BrinGetPagesPerRange(index);
- rdata.buffer = InvalidBuffer;
- rdata.data = (char *) &xlrec;
- rdata.len = SizeOfBrinCreateIdx;
- rdata.next = NULL;
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfBrinCreateIdx);
+ XLogRegisterBuffer(0, meta, REGBUF_WILL_INIT);
- recptr = XLogInsert(RM_BRIN_ID, XLOG_BRIN_CREATE_INDEX, &rdata);
+ recptr = XLogInsert(RM_BRIN_ID, XLOG_BRIN_CREATE_INDEX);
page = BufferGetPage(meta);
PageSetLSN(page, recptr);
/* XLOG stuff */
if (RelationNeedsWAL(idxrel))
{
- BlockNumber blk = BufferGetBlockNumber(oldbuf);
xl_brin_samepage_update xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[2];
uint8 info = XLOG_BRIN_SAMEPAGE_UPDATE;
- xlrec.node = idxrel->rd_node;
- ItemPointerSetBlockNumber(&xlrec.tid, blk);
- ItemPointerSetOffsetNumber(&xlrec.tid, oldoff);
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfBrinSamepageUpdate;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
+ xlrec.offnum = oldoff;
- rdata[1].data = (char *) newtup;
- rdata[1].len = newsz;
- rdata[1].buffer = oldbuf;
- rdata[1].buffer_std = true;
- rdata[1].next = NULL;
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfBrinSamepageUpdate);
- recptr = XLogInsert(RM_BRIN_ID, info, rdata);
+ XLogRegisterBuffer(0, oldbuf, REGBUF_STANDARD);
+ XLogRegisterBufData(0, (char *) newtup, newsz);
+
+ recptr = XLogInsert(RM_BRIN_ID, info);
PageSetLSN(oldpage, recptr);
}
{
xl_brin_update xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[4];
uint8 info;
info = XLOG_BRIN_UPDATE | (extended ? XLOG_BRIN_INIT_PAGE : 0);
- xlrec.insert.node = idxrel->rd_node;
- ItemPointerSet(&xlrec.insert.tid, BufferGetBlockNumber(newbuf), newoff);
+ xlrec.insert.offnum = newoff;
xlrec.insert.heapBlk = heapBlk;
- xlrec.insert.tuplen = newsz;
- xlrec.insert.revmapBlk = BufferGetBlockNumber(revmapbuf);
xlrec.insert.pagesPerRange = pagesPerRange;
- ItemPointerSet(&xlrec.oldtid, BufferGetBlockNumber(oldbuf), oldoff);
+ xlrec.oldOffnum = oldoff;
+
+ XLogBeginInsert();
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfBrinUpdate;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
+ /* new page */
+ XLogRegisterData((char *) &xlrec, SizeOfBrinUpdate);
- rdata[1].data = (char *) newtup;
- rdata[1].len = newsz;
- rdata[1].buffer = extended ? InvalidBuffer : newbuf;
- rdata[1].buffer_std = true;
- rdata[1].next = &(rdata[2]);
+ XLogRegisterBuffer(0, newbuf, REGBUF_STANDARD | (extended ? REGBUF_WILL_INIT : 0));
+ XLogRegisterBufData(0, (char *) newtup, newsz);
- rdata[2].data = (char *) NULL;
- rdata[2].len = 0;
- rdata[2].buffer = revmapbuf;
- rdata[2].buffer_std = true;
- rdata[2].next = &(rdata[3]);
+ /* revmap page */
+ XLogRegisterBuffer(1, revmapbuf, REGBUF_STANDARD);
- rdata[3].data = (char *) NULL;
- rdata[3].len = 0;
- rdata[3].buffer = oldbuf;
- rdata[3].buffer_std = true;
- rdata[3].next = NULL;
+ /* old page */
+ XLogRegisterBuffer(2, oldbuf, REGBUF_STANDARD);
- recptr = XLogInsert(RM_BRIN_ID, info, rdata);
+ recptr = XLogInsert(RM_BRIN_ID, info);
PageSetLSN(oldpage, recptr);
PageSetLSN(newpage, recptr);
{
xl_brin_insert xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[3];
uint8 info;
info = XLOG_BRIN_INSERT | (extended ? XLOG_BRIN_INIT_PAGE : 0);
- xlrec.node = idxrel->rd_node;
xlrec.heapBlk = heapBlk;
xlrec.pagesPerRange = pagesPerRange;
- xlrec.revmapBlk = BufferGetBlockNumber(revmapbuf);
- xlrec.tuplen = itemsz;
- ItemPointerSet(&xlrec.tid, blk, off);
-
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfBrinInsert;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].buffer_std = false;
- rdata[0].next = &(rdata[1]);
-
- rdata[1].data = (char *) tup;
- rdata[1].len = itemsz;
- rdata[1].buffer = extended ? InvalidBuffer : *buffer;
- rdata[1].buffer_std = true;
- rdata[1].next = &(rdata[2]);
-
- rdata[2].data = (char *) NULL;
- rdata[2].len = 0;
- rdata[2].buffer = revmapbuf;
- rdata[2].buffer_std = false;
- rdata[2].next = NULL;
-
- recptr = XLogInsert(RM_BRIN_ID, info, rdata);
+ 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);
{
xl_brin_revmap_extend xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[2];
- xlrec.node = revmap->rm_irel->rd_node;
xlrec.targetBlk = mapBlk;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfBrinRevmapExtend;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].buffer_std = false;
- rdata[0].next = &(rdata[1]);
-
- rdata[1].data = (char *) NULL;
- rdata[1].len = 0;
- rdata[1].buffer = revmap->rm_metaBuf;
- rdata[1].buffer_std = false;
- rdata[1].next = NULL;
-
- recptr = XLogInsert(RM_BRIN_ID, XLOG_BRIN_REVMAP_EXTEND, rdata);
+
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfBrinRevmapExtend);
+ XLogRegisterBuffer(0, revmap->rm_metaBuf, 0);
+
+ XLogRegisterBuffer(1, buf, REGBUF_WILL_INIT);
+
+ recptr = XLogInsert(RM_BRIN_ID, XLOG_BRIN_REVMAP_EXTEND);
PageSetLSN(metapage, recptr);
PageSetLSN(page, recptr);
}
* xlog replay routines
*/
static void
-brin_xlog_createidx(XLogRecPtr lsn, XLogRecord *record)
+brin_xlog_createidx(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_brin_createidx *xlrec = (xl_brin_createidx *) XLogRecGetData(record);
Buffer buf;
Page page;
- /* Backup blocks are not used in create_index records */
- Assert(!(record->xl_info & XLR_BKP_BLOCK_MASK));
-
/* create the index' metapage */
- buf = XLogReadBuffer(xlrec->node, BRIN_METAPAGE_BLKNO, true);
+ buf = XLogInitBufferForRedo(record, 0);
Assert(BufferIsValid(buf));
page = (Page) BufferGetPage(buf);
brin_metapage_init(page, xlrec->pagesPerRange, xlrec->version);
* revmap.
*/
static void
-brin_xlog_insert_update(XLogRecPtr lsn, XLogRecord *record,
- xl_brin_insert *xlrec, BrinTuple *tuple)
+brin_xlog_insert_update(XLogReaderState *record,
+ xl_brin_insert *xlrec)
{
- BlockNumber blkno;
+ XLogRecPtr lsn = record->EndRecPtr;
Buffer buffer;
Page page;
XLogRedoAction action;
- blkno = ItemPointerGetBlockNumber(&xlrec->tid);
-
/*
* If we inserted the first and only tuple on the page, re-initialize the
* page from scratch.
*/
- if (record->xl_info & XLOG_BRIN_INIT_PAGE)
+ if (XLogRecGetInfo(record) & XLOG_BRIN_INIT_PAGE)
{
- /*
- * No full-page image here. Don't try to read it, because there
- * might be one for the revmap buffer, below.
- */
- buffer = XLogReadBuffer(xlrec->node, blkno, true);
+ buffer = XLogInitBufferForRedo(record, 0);
page = BufferGetPage(buffer);
brin_page_init(page, BRIN_PAGETYPE_REGULAR);
action = BLK_NEEDS_REDO;
}
else
{
- action = XLogReadBufferForRedo(lsn, record, 0,
- xlrec->node, blkno, &buffer);
+ action = XLogReadBufferForRedo(record, 0, &buffer);
}
/* insert the index item into the page */
if (action == BLK_NEEDS_REDO)
{
OffsetNumber offnum;
+ BrinTuple *tuple;
+ Size tuplen;
+
+ tuple = (BrinTuple *) XLogRecGetBlockData(record, 0, &tuplen);
Assert(tuple->bt_blkno == xlrec->heapBlk);
page = (Page) BufferGetPage(buffer);
- offnum = ItemPointerGetOffsetNumber(&(xlrec->tid));
+ offnum = xlrec->offnum;
if (PageGetMaxOffsetNumber(page) + 1 < offnum)
elog(PANIC, "brin_xlog_insert_update: invalid max offset number");
- offnum = PageAddItem(page, (Item) tuple, xlrec->tuplen, offnum, true,
- false);
+ offnum = PageAddItem(page, (Item) tuple, tuplen, offnum, true, false);
if (offnum == InvalidOffsetNumber)
elog(PANIC, "brin_xlog_insert_update: failed to add tuple");
UnlockReleaseBuffer(buffer);
/* update the revmap */
- action = XLogReadBufferForRedo(lsn, record,
- record->xl_info & XLOG_BRIN_INIT_PAGE ? 0 : 1,
- xlrec->node,
- xlrec->revmapBlk, &buffer);
+ action = XLogReadBufferForRedo(record, 1, &buffer);
if (action == BLK_NEEDS_REDO)
{
+ ItemPointerData tid;
+ BlockNumber blkno = BufferGetBlockNumber(buffer);
+
+ ItemPointerSet(&tid, blkno, xlrec->offnum);
page = (Page) BufferGetPage(buffer);
brinSetHeapBlockItemptr(buffer, xlrec->pagesPerRange, xlrec->heapBlk,
- xlrec->tid);
+ tid);
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
* replay a BRIN index insertion
*/
static void
-brin_xlog_insert(XLogRecPtr lsn, XLogRecord *record)
+brin_xlog_insert(XLogReaderState *record)
{
xl_brin_insert *xlrec = (xl_brin_insert *) XLogRecGetData(record);
- BrinTuple *newtup;
- newtup = (BrinTuple *) ((char *) xlrec + SizeOfBrinInsert);
-
- brin_xlog_insert_update(lsn, record, xlrec, newtup);
+ brin_xlog_insert_update(record, xlrec);
}
/*
* replay a BRIN index update
*/
static void
-brin_xlog_update(XLogRecPtr lsn, XLogRecord *record)
+brin_xlog_update(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_brin_update *xlrec = (xl_brin_update *) XLogRecGetData(record);
- BlockNumber blkno;
Buffer buffer;
- BrinTuple *newtup;
XLogRedoAction action;
- newtup = (BrinTuple *) ((char *) xlrec + SizeOfBrinUpdate);
-
/* First remove the old tuple */
- blkno = ItemPointerGetBlockNumber(&(xlrec->oldtid));
- action = XLogReadBufferForRedo(lsn, record, 2, xlrec->insert.node,
- blkno, &buffer);
+ action = XLogReadBufferForRedo(record, 2, &buffer);
if (action == BLK_NEEDS_REDO)
{
Page page;
page = (Page) BufferGetPage(buffer);
- offnum = ItemPointerGetOffsetNumber(&(xlrec->oldtid));
+ offnum = xlrec->oldOffnum;
if (PageGetMaxOffsetNumber(page) + 1 < offnum)
elog(PANIC, "brin_xlog_update: invalid max offset number");
}
/* Then insert the new tuple and update revmap, like in an insertion. */
- brin_xlog_insert_update(lsn, record, &xlrec->insert, newtup);
+ brin_xlog_insert_update(record, &xlrec->insert);
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
* Update a tuple on a single page.
*/
static void
-brin_xlog_samepage_update(XLogRecPtr lsn, XLogRecord *record)
+brin_xlog_samepage_update(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_brin_samepage_update *xlrec;
- BlockNumber blkno;
Buffer buffer;
XLogRedoAction action;
xlrec = (xl_brin_samepage_update *) XLogRecGetData(record);
- blkno = ItemPointerGetBlockNumber(&(xlrec->tid));
- action = XLogReadBufferForRedo(lsn, record, 0, xlrec->node, blkno,
- &buffer);
+ action = XLogReadBufferForRedo(record, 0, &buffer);
if (action == BLK_NEEDS_REDO)
{
- int tuplen;
+ Size tuplen;
BrinTuple *mmtuple;
Page page;
OffsetNumber offnum;
- tuplen = record->xl_len - SizeOfBrinSamepageUpdate;
- mmtuple = (BrinTuple *) ((char *) xlrec + SizeOfBrinSamepageUpdate);
+ mmtuple = (BrinTuple *) XLogRecGetBlockData(record, 0, &tuplen);
page = (Page) BufferGetPage(buffer);
- offnum = ItemPointerGetOffsetNumber(&(xlrec->tid));
+ offnum = xlrec->offnum;
if (PageGetMaxOffsetNumber(page) + 1 < offnum)
elog(PANIC, "brin_xlog_samepage_update: invalid max offset number");
* Replay a revmap page extension
*/
static void
-brin_xlog_revmap_extend(XLogRecPtr lsn, XLogRecord *record)
+brin_xlog_revmap_extend(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_brin_revmap_extend *xlrec;
Buffer metabuf;
Buffer buf;
Page page;
+ BlockNumber targetBlk;
XLogRedoAction action;
xlrec = (xl_brin_revmap_extend *) XLogRecGetData(record);
+
+ XLogRecGetBlockTag(record, 1, NULL, NULL, &targetBlk);
+ Assert(xlrec->targetBlk == targetBlk);
+
/* Update the metapage */
- action = XLogReadBufferForRedo(lsn, record, 0, xlrec->node,
- BRIN_METAPAGE_BLKNO, &metabuf);
+ action = XLogReadBufferForRedo(record, 0, &metabuf);
if (action == BLK_NEEDS_REDO)
{
Page metapg;
* image here.
*/
- buf = XLogReadBuffer(xlrec->node, xlrec->targetBlk, true);
+ buf = XLogInitBufferForRedo(record, 1);
page = (Page) BufferGetPage(buf);
brin_page_init(page, BRIN_PAGETYPE_REVMAP);
}
void
-brin_redo(XLogRecPtr lsn, XLogRecord *record)
+brin_redo(XLogReaderState *record)
{
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
switch (info & XLOG_BRIN_OPMASK)
{
case XLOG_BRIN_CREATE_INDEX:
- brin_xlog_createidx(lsn, record);
+ brin_xlog_createidx(record);
break;
case XLOG_BRIN_INSERT:
- brin_xlog_insert(lsn, record);
+ brin_xlog_insert(record);
break;
case XLOG_BRIN_UPDATE:
- brin_xlog_update(lsn, record);
+ brin_xlog_update(record);
break;
case XLOG_BRIN_SAMEPAGE_UPDATE:
- brin_xlog_samepage_update(lsn, record);
+ brin_xlog_samepage_update(record);
break;
case XLOG_BRIN_REVMAP_EXTEND:
- brin_xlog_revmap_extend(lsn, record);
+ brin_xlog_revmap_extend(record);
break;
default:
elog(PANIC, "brin_redo: unknown op code %u", info);
Buffer childbuf, GinStatsData *buildStats)
{
Page page = BufferGetPage(stack->buffer);
- XLogRecData *payloadrdata;
GinPlaceToPageRC rc;
uint16 xlflags = 0;
Page childpage = NULL;
/*
* Try to put the incoming tuple on the page. placeToPage will decide if
* the page needs to be split.
+ *
+ * WAL-logging this operation is a bit funny:
+ *
+ * We're responsible for calling XLogBeginInsert() and XLogInsert().
+ * XLogBeginInsert() must be called before placeToPage, because
+ * placeToPage can register some data to the WAL record.
+ *
+ * If placeToPage returns INSERTED, placeToPage has already called
+ * START_CRIT_SECTION(), and we're responsible for calling
+ * END_CRIT_SECTION. When it returns INSERTED, it is also responsible for
+ * registering any data required to replay the operation with
+ * XLogRegisterData(0, ...). It may only add data to block index 0; the
+ * main data of the WAL record is reserved for this function.
+ *
+ * If placeToPage returns SPLIT, we're wholly responsible for WAL logging.
+ * Splits happen infrequently, so we just make a full-page image of all
+ * the pages involved.
*/
+
+ if (RelationNeedsWAL(btree->index))
+ XLogBeginInsert();
+
rc = btree->placeToPage(btree, stack->buffer, stack,
insertdata, updateblkno,
- &payloadrdata, &newlpage, &newrpage);
+ &newlpage, &newrpage);
if (rc == UNMODIFIED)
+ {
+ XLogResetInsertion();
return true;
+ }
else if (rc == INSERTED)
{
/* placeToPage did START_CRIT_SECTION() */
if (RelationNeedsWAL(btree->index))
{
XLogRecPtr recptr;
- XLogRecData rdata[3];
ginxlogInsert xlrec;
BlockIdData childblknos[2];
- xlrec.node = btree->index->rd_node;
- xlrec.blkno = BufferGetBlockNumber(stack->buffer);
+ /*
+ * placetopage already registered stack->buffer as block 0.
+ */
xlrec.flags = xlflags;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = sizeof(ginxlogInsert);
+ if (childbuf != InvalidBuffer)
+ XLogRegisterBuffer(1, childbuf, REGBUF_STANDARD);
+
+ XLogRegisterData((char *) &xlrec, sizeof(ginxlogInsert));
/*
* Log information about child if this was an insertion of a
*/
if (childbuf != InvalidBuffer)
{
- rdata[0].next = &rdata[1];
-
BlockIdSet(&childblknos[0], BufferGetBlockNumber(childbuf));
BlockIdSet(&childblknos[1], GinPageGetOpaque(childpage)->rightlink);
-
- rdata[1].buffer = InvalidBuffer;
- rdata[1].data = (char *) childblknos;
- rdata[1].len = sizeof(BlockIdData) * 2;
- rdata[1].next = &rdata[2];
-
- rdata[2].buffer = childbuf;
- rdata[2].buffer_std = false;
- rdata[2].data = NULL;
- rdata[2].len = 0;
- rdata[2].next = payloadrdata;
+ XLogRegisterData((char *) childblknos,
+ sizeof(BlockIdData) * 2);
}
- else
- rdata[0].next = payloadrdata;
- recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_INSERT, rdata);
+ recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_INSERT);
PageSetLSN(page, recptr);
if (childbuf != InvalidBuffer)
PageSetLSN(childpage, recptr);
}
else if (rc == SPLIT)
{
- /* Didn't fit, have to split */
+ /* Didn't fit, had to split */
Buffer rbuffer;
BlockNumber savedRightLink;
- XLogRecData rdata[2];
ginxlogSplit data;
Buffer lbuffer = InvalidBuffer;
Page newrootpg = NULL;
*/
data.node = btree->index->rd_node;
- data.rblkno = BufferGetBlockNumber(rbuffer);
data.flags = xlflags;
if (childbuf != InvalidBuffer)
{
else
data.leftChildBlkno = data.rightChildBlkno = InvalidBlockNumber;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].data = (char *) &data;
- rdata[0].len = sizeof(ginxlogSplit);
-
- if (childbuf != InvalidBuffer)
- {
- rdata[0].next = &rdata[1];
-
- rdata[1].buffer = childbuf;
- rdata[1].buffer_std = false;
- rdata[1].data = NULL;
- rdata[1].len = 0;
- rdata[1].next = payloadrdata;
- }
- else
- rdata[0].next = payloadrdata;
-
if (stack->parent == NULL)
{
/*
buildStats->nEntryPages++;
}
- /*
- * root never has a right-link, so we borrow the rrlink field to
- * store the root block number.
- */
- data.rrlink = BufferGetBlockNumber(stack->buffer);
- data.lblkno = BufferGetBlockNumber(lbuffer);
+ data.rrlink = InvalidBlockNumber;
data.flags |= GIN_SPLIT_ROOT;
GinPageGetOpaque(newrpage)->rightlink = InvalidBlockNumber;
{
/* split non-root page */
data.rrlink = savedRightLink;
- data.lblkno = BufferGetBlockNumber(stack->buffer);
GinPageGetOpaque(newrpage)->rightlink = savedRightLink;
GinPageGetOpaque(newlpage)->flags |= GIN_INCOMPLETE_SPLIT;
{
XLogRecPtr recptr;
- recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_SPLIT, rdata);
+ /*
+ * We just take full page images of all the split pages. Splits
+ * are uncommon enough that it's not worth complicating the code
+ * to be more efficient.
+ */
+ if (stack->parent == NULL)
+ {
+ XLogRegisterBuffer(0, lbuffer, REGBUF_FORCE_IMAGE | REGBUF_STANDARD);
+ XLogRegisterBuffer(1, rbuffer, REGBUF_FORCE_IMAGE | REGBUF_STANDARD);
+ XLogRegisterBuffer(2, stack->buffer, REGBUF_FORCE_IMAGE | REGBUF_STANDARD);
+ }
+ else
+ {
+ XLogRegisterBuffer(0, stack->buffer, REGBUF_FORCE_IMAGE | REGBUF_STANDARD);
+ XLogRegisterBuffer(1, rbuffer, REGBUF_FORCE_IMAGE | REGBUF_STANDARD);
+ }
+ if (BufferIsValid(childbuf))
+ XLogRegisterBuffer(3, childbuf, 0);
+
+ XLogRegisterData((char *) &data, sizeof(ginxlogSplit));
+
+ recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_SPLIT);
PageSetLSN(BufferGetPage(stack->buffer), recptr);
PageSetLSN(BufferGetPage(rbuffer), recptr);
if (stack->parent == NULL)
static void dataSplitPageInternal(GinBtree btree, Buffer origbuf,
GinBtreeStack *stack,
void *insertdata, BlockNumber updateblkno,
- XLogRecData **prdata, Page *newlpage, Page *newrpage);
+ Page *newlpage, Page *newrpage);
static disassembledLeaf *disassembleLeaf(Page page);
static bool leafRepackItems(disassembledLeaf *leaf, ItemPointer remaining);
static bool addItemsToLeaf(disassembledLeaf *leaf, ItemPointer newItems,
int nNewItems);
-static XLogRecData *constructLeafRecompressWALData(Buffer buf,
- disassembledLeaf *leaf);
+static void registerLeafRecompressWALData(Buffer buf, disassembledLeaf *leaf);
static void dataPlaceToPageLeafRecompress(Buffer buf, disassembledLeaf *leaf);
static void dataPlaceToPageLeafSplit(Buffer buf,
disassembledLeaf *leaf,
ItemPointerData lbound, ItemPointerData rbound,
- XLogRecData **prdata, Page lpage, Page rpage);
+ Page lpage, Page rpage);
/*
* Read TIDs from leaf data page to single uncompressed array. The TIDs are
*/
static GinPlaceToPageRC
dataPlaceToPageLeaf(GinBtree btree, Buffer buf, GinBtreeStack *stack,
- void *insertdata, XLogRecData **prdata,
- Page *newlpage, Page *newrpage)
+ void *insertdata, Page *newlpage, Page *newrpage)
{
GinBtreeDataLeafInsertData *items = insertdata;
ItemPointer newItems = &items->items[items->curitem];
*/
MemoryContextSwitchTo(oldCxt);
if (RelationNeedsWAL(btree->index))
- *prdata = constructLeafRecompressWALData(buf, leaf);
- else
- *prdata = NULL;
+ registerLeafRecompressWALData(buf, leaf);
START_CRIT_SECTION();
dataPlaceToPageLeafRecompress(buf, leaf);
*newrpage = MemoryContextAlloc(oldCxt, BLCKSZ);
dataPlaceToPageLeafSplit(buf, leaf, lbound, rbound,
- prdata, *newlpage, *newrpage);
+ *newlpage, *newrpage);
Assert(GinPageRightMost(page) ||
ginCompareItemPointers(GinDataPageGetRightBound(*newlpage),
*/
if (removedsomething)
{
- XLogRecData *payloadrdata = NULL;
bool modified;
/*
}
if (RelationNeedsWAL(indexrel))
- payloadrdata = constructLeafRecompressWALData(buffer, leaf);
+ {
+ XLogBeginInsert();
+ registerLeafRecompressWALData(buffer, leaf);
+ }
START_CRIT_SECTION();
dataPlaceToPageLeafRecompress(buffer, leaf);
if (RelationNeedsWAL(indexrel))
{
XLogRecPtr recptr;
- XLogRecData rdata;
- ginxlogVacuumDataLeafPage xlrec;
- xlrec.node = indexrel->rd_node;
- xlrec.blkno = BufferGetBlockNumber(buffer);
-
- rdata.buffer = InvalidBuffer;
- rdata.data = (char *) &xlrec;
- rdata.len = offsetof(ginxlogVacuumDataLeafPage, data);
- rdata.next = payloadrdata;
-
- recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_VACUUM_DATA_LEAF_PAGE, &rdata);
+ recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_VACUUM_DATA_LEAF_PAGE);
PageSetLSN(page, recptr);
}
* Construct a ginxlogRecompressDataLeaf record representing the changes
* in *leaf.
*/
-static XLogRecData *
-constructLeafRecompressWALData(Buffer buf, disassembledLeaf *leaf)
+static void
+registerLeafRecompressWALData(Buffer buf, disassembledLeaf *leaf)
{
int nmodified = 0;
char *walbufbegin;
char *walbufend;
- XLogRecData *rdata;
dlist_iter iter;
int segno;
ginxlogRecompressDataLeaf *recompress_xlog;
nmodified++;
}
- walbufbegin = palloc(
- sizeof(ginxlogRecompressDataLeaf) +
- BLCKSZ + /* max size needed to hold the segment
- * data */
- nmodified * 2 + /* (segno + action) per action */
- sizeof(XLogRecData));
+ walbufbegin =
+ palloc(sizeof(ginxlogRecompressDataLeaf) +
+ BLCKSZ + /* max size needed to hold the segment data */
+ nmodified * 2 /* (segno + action) per action */
+ );
walbufend = walbufbegin;
recompress_xlog = (ginxlogRecompressDataLeaf *) walbufend;
segno++;
}
- rdata = (XLogRecData *) MAXALIGN(walbufend);
- rdata->buffer = buf;
- rdata->buffer_std = TRUE;
- rdata->data = walbufbegin;
- rdata->len = walbufend - walbufbegin;
- rdata->next = NULL;
- return rdata;
+ XLogRegisterBuffer(0, buf, REGBUF_STANDARD);
+ XLogRegisterBufData(0, walbufbegin, walbufend - walbufbegin);
+
}
/*
static void
dataPlaceToPageLeafSplit(Buffer buf, disassembledLeaf *leaf,
ItemPointerData lbound, ItemPointerData rbound,
- XLogRecData **prdata, Page lpage, Page rpage)
+ Page lpage, Page rpage)
{
char *ptr;
int segsize;
dlist_node *firstright;
leafSegmentInfo *seginfo;
- /* these must be static so they can be returned to caller */
- static ginxlogSplitDataLeaf split_xlog;
- static XLogRecData rdata[3];
-
/* Initialize temporary pages to hold the new left and right pages */
GinInitPage(lpage, GIN_DATA | GIN_LEAF | GIN_COMPRESSED, BLCKSZ);
GinInitPage(rpage, GIN_DATA | GIN_LEAF | GIN_COMPRESSED, BLCKSZ);
Assert(rsize == leaf->rsize);
GinDataPageSetDataSize(rpage, rsize);
*GinDataPageGetRightBound(rpage) = rbound;
-
- /* Create WAL record */
- split_xlog.lsize = lsize;
- split_xlog.rsize = rsize;
- split_xlog.lrightbound = lbound;
- split_xlog.rrightbound = rbound;
-
- rdata[0].buffer = InvalidBuffer;
- rdata[0].data = (char *) &split_xlog;
- rdata[0].len = sizeof(ginxlogSplitDataLeaf);
- rdata[0].next = &rdata[1];
-
- rdata[1].buffer = InvalidBuffer;
- rdata[1].data = (char *) GinDataLeafPageGetPostingList(lpage);
- rdata[1].len = lsize;
- rdata[1].next = &rdata[2];
-
- rdata[2].buffer = InvalidBuffer;
- rdata[2].data = (char *) GinDataLeafPageGetPostingList(rpage);
- rdata[2].len = rsize;
- rdata[2].next = NULL;
-
- *prdata = rdata;
}
/*
*
* 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
dataPlaceToPageInternal(GinBtree btree, Buffer buf, GinBtreeStack *stack,
void *insertdata, BlockNumber updateblkno,
- XLogRecData **prdata, Page *newlpage, Page *newrpage)
+ Page *newlpage, Page *newrpage)
{
Page page = BufferGetPage(buf);
OffsetNumber off = stack->off;
PostingItem *pitem;
- /* these must be static so they can be returned to caller */
- static XLogRecData rdata;
+ /* this must be static so it can be returned to caller */
static ginxlogInsertDataInternal data;
/* split if we have to */
if (GinNonLeafDataPageGetFreeSpace(page) < sizeof(PostingItem))
{
dataSplitPageInternal(btree, buf, stack, insertdata, updateblkno,
- prdata, newlpage, newrpage);
+ newlpage, newrpage);
return SPLIT;
}
- *prdata = &rdata;
Assert(GinPageIsData(page));
START_CRIT_SECTION();
pitem = (PostingItem *) insertdata;
GinDataPageAddPostingItem(page, pitem, off);
- data.offset = off;
- data.newitem = *pitem;
+ if (RelationNeedsWAL(btree->index))
+ {
+ data.offset = off;
+ data.newitem = *pitem;
- rdata.buffer = buf;
- rdata.buffer_std = TRUE;
- rdata.data = (char *) &data;
- rdata.len = sizeof(ginxlogInsertDataInternal);
- rdata.next = NULL;
+ XLogRegisterBuffer(0, buf, REGBUF_STANDARD);
+ XLogRegisterBufData(0, (char *) &data,
+ sizeof(ginxlogInsertDataInternal));
+ }
return INSERTED;
}
static GinPlaceToPageRC
dataPlaceToPage(GinBtree btree, Buffer buf, GinBtreeStack *stack,
void *insertdata, BlockNumber updateblkno,
- XLogRecData **prdata,
Page *newlpage, Page *newrpage)
{
Page page = BufferGetPage(buf);
if (GinPageIsLeaf(page))
return dataPlaceToPageLeaf(btree, buf, stack, insertdata,
- prdata, newlpage, newrpage);
+ newlpage, newrpage);
else
return dataPlaceToPageInternal(btree, buf, stack,
insertdata, updateblkno,
- prdata, newlpage, newrpage);
+ newlpage, newrpage);
}
/*
dataSplitPageInternal(GinBtree btree, Buffer origbuf,
GinBtreeStack *stack,
void *insertdata, BlockNumber updateblkno,
- XLogRecData **prdata, Page *newlpage, Page *newrpage)
+ Page *newlpage, Page *newrpage)
{
Page oldpage = BufferGetPage(origbuf);
OffsetNumber off = stack->off;
Page lpage;
Page rpage;
OffsetNumber separator;
-
- /* these must be static so they can be returned to caller */
- static ginxlogSplitDataInternal data;
- static XLogRecData rdata[4];
- static PostingItem allitems[(BLCKSZ / sizeof(PostingItem)) + 1];
+ PostingItem allitems[(BLCKSZ / sizeof(PostingItem)) + 1];
lpage = PageGetTempPage(oldpage);
rpage = PageGetTempPage(oldpage);
GinInitPage(lpage, GinPageGetOpaque(oldpage)->flags, pageSize);
GinInitPage(rpage, GinPageGetOpaque(oldpage)->flags, pageSize);
- *prdata = rdata;
-
/*
* First construct a new list of PostingItems, which includes all the old
* items, and the new item.
/* set up right bound for right page */
*GinDataPageGetRightBound(rpage) = oldbound;
- data.separator = separator;
- data.nitem = nitems;
- data.rightbound = oldbound;
-
- rdata[0].buffer = InvalidBuffer;
- rdata[0].data = (char *) &data;
- rdata[0].len = sizeof(ginxlogSplitDataInternal);
- rdata[0].next = &rdata[1];
-
- rdata[1].buffer = InvalidBuffer;
- rdata[1].data = (char *) allitems;
- rdata[1].len = nitems * sizeof(PostingItem);
- rdata[1].next = NULL;
-
*newlpage = lpage;
*newrpage = rpage;
}
if (RelationNeedsWAL(index))
{
XLogRecPtr recptr;
- XLogRecData rdata[2];
ginxlogCreatePostingTree data;
- data.node = index->rd_node;
- data.blkno = blkno;
data.size = rootsize;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].data = (char *) &data;
- rdata[0].len = sizeof(ginxlogCreatePostingTree);
- rdata[0].next = &rdata[1];
+ XLogBeginInsert();
+ XLogRegisterData((char *) &data, sizeof(ginxlogCreatePostingTree));
- rdata[1].buffer = InvalidBuffer;
- rdata[1].data = (char *) GinDataLeafPageGetPostingList(page);
- rdata[1].len = rootsize;
- rdata[1].next = NULL;
+ XLogRegisterData((char *) GinDataLeafPageGetPostingList(page),
+ rootsize);
+ XLogRegisterBuffer(0, buffer, REGBUF_WILL_INIT);
- recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_CREATE_PTREE, rdata);
+ recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_CREATE_PTREE);
PageSetLSN(page, recptr);
}
static void entrySplitPage(GinBtree btree, Buffer origbuf,
GinBtreeStack *stack,
void *insertPayload,
- BlockNumber updateblkno, XLogRecData **prdata,
+ BlockNumber updateblkno,
Page *newlpage, Page *newrpage);
/*
* 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,
- XLogRecData **prdata, Page *newlpage, Page *newrpage)
+ 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];
+ /* 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,
- prdata, newlpage, newrpage);
+ newlpage, newrpage);
return SPLIT;
}
START_CRIT_SECTION();
- *prdata = rdata;
entryPreparePage(btree, page, off, insertData, updateblkno);
placed = PageAddItem(page,
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;
+ if (RelationNeedsWAL(btree->index))
+ {
+ data.isDelete = insertData->isDelete;
+ data.offset = off;
+
+ XLogRegisterBuffer(0, buf, REGBUF_STANDARD);
+ XLogRegisterBufData(0, (char *) &data,
+ offsetof(ginxlogInsertEntry, tuple));
+ XLogRegisterBufData(0, (char *) insertData->entry,
+ IndexTupleSize(insertData->entry));
+ }
return INSERTED;
}
entrySplitPage(GinBtree btree, Buffer origbuf,
GinBtreeStack *stack,
void *insertPayload,
- BlockNumber updateblkno, XLogRecData **prdata,
+ BlockNumber updateblkno,
Page *newlpage, Page *newrpage)
{
GinBtreeEntryInsertData *insertData = insertPayload;
maxoff,
separator = InvalidOffsetNumber;
Size totalsize = 0;
- Size tupstoresize;
Size lsize = 0,
size;
char *ptr;
Page lpage = PageGetTempPageCopy(BufferGetPage(origbuf));
Page rpage = PageGetTempPageCopy(BufferGetPage(origbuf));
Size pageSize = PageGetPageSize(lpage);
+ char tupstore[2 * BLCKSZ];
- /* these must be static so they can be returned to caller */
- static XLogRecData rdata[2];
- static ginxlogSplitEntry data;
- static char tupstore[2 * BLCKSZ];
-
- *prdata = rdata;
entryPreparePage(btree, lpage, off, insertData, updateblkno);
/*
ptr += size;
totalsize += size + sizeof(ItemIdData);
}
- tupstoresize = ptr - tupstore;
/*
* Initialize the left and right pages, and copy all the tuples back to
ptr += MAXALIGN(IndexTupleSize(itup));
}
- data.separator = separator;
- data.nitem = maxoff;
-
- rdata[0].buffer = InvalidBuffer;
- rdata[0].data = (char *) &data;
- rdata[0].len = sizeof(ginxlogSplitEntry);
- rdata[0].next = &rdata[1];
-
- rdata[1].buffer = InvalidBuffer;
- rdata[1].data = tupstore;
- rdata[1].len = tupstoresize;
- rdata[1].next = NULL;
-
*newlpage = lpage;
*newrpage = rpage;
}
if (RelationNeedsWAL(index))
{
- XLogRecData rdata[2];
ginxlogInsertListPage data;
XLogRecPtr recptr;
- data.node = index->rd_node;
- data.blkno = BufferGetBlockNumber(buffer);
data.rightlink = rightlink;
data.ntuples = ntuples;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].data = (char *) &data;
- rdata[0].len = sizeof(ginxlogInsertListPage);
- rdata[0].next = rdata + 1;
+ XLogBeginInsert();
+ XLogRegisterData((char *) &data, sizeof(ginxlogInsertListPage));
- rdata[1].buffer = InvalidBuffer;
- rdata[1].data = workspace;
- rdata[1].len = size;
- rdata[1].next = NULL;
+ XLogRegisterBuffer(0, buffer, REGBUF_WILL_INIT);
+ XLogRegisterBufData(0, workspace, size);
- recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_INSERT_LISTPAGE, rdata);
+ recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_INSERT_LISTPAGE);
PageSetLSN(page, recptr);
}
Buffer metabuffer;
Page metapage;
GinMetaPageData *metadata = NULL;
- XLogRecData rdata[2];
Buffer buffer = InvalidBuffer;
Page page = NULL;
ginxlogUpdateMeta data;
bool separateList = false;
bool needCleanup = false;
int cleanupSize;
+ bool needWal;
if (collector->ntuples == 0)
return;
+ needWal = RelationNeedsWAL(index);
+
data.node = index->rd_node;
data.ntuples = 0;
data.newRightlink = data.prevTail = InvalidBlockNumber;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].data = (char *) &data;
- rdata[0].len = sizeof(ginxlogUpdateMeta);
- rdata[0].next = NULL;
-
metabuffer = ReadBuffer(index, GIN_METAPAGE_BLKNO);
metapage = BufferGetPage(metabuffer);
memset(&sublist, 0, sizeof(GinMetaPageData));
makeSublist(index, collector->tuples, collector->ntuples, &sublist);
+ if (needWal)
+ XLogBeginInsert();
+
/*
* metapage was unlocked, see above
*/
LockBuffer(buffer, GIN_EXCLUSIVE);
page = BufferGetPage(buffer);
- rdata[0].next = rdata + 1;
-
- rdata[1].buffer = buffer;
- rdata[1].buffer_std = true;
- rdata[1].data = NULL;
- rdata[1].len = 0;
- rdata[1].next = NULL;
-
Assert(GinPageGetOpaque(page)->rightlink == InvalidBlockNumber);
START_CRIT_SECTION();
metadata->nPendingPages += sublist.nPendingPages;
metadata->nPendingHeapTuples += sublist.nPendingHeapTuples;
+
+ if (needWal)
+ XLogRegisterBuffer(1, buffer, REGBUF_STANDARD);
}
}
else
int i,
tupsize;
char *ptr;
+ char *collectordata;
buffer = ReadBuffer(index, metadata->tail);
LockBuffer(buffer, GIN_EXCLUSIVE);
off = (PageIsEmpty(page)) ? FirstOffsetNumber :
OffsetNumberNext(PageGetMaxOffsetNumber(page));
- rdata[0].next = rdata + 1;
-
- rdata[1].buffer = buffer;
- rdata[1].buffer_std = true;
- ptr = rdata[1].data = (char *) palloc(collector->sumsize);
- rdata[1].len = collector->sumsize;
- rdata[1].next = NULL;
+ collectordata = ptr = (char *) palloc(collector->sumsize);
data.ntuples = collector->ntuples;
+ if (needWal)
+ XLogBeginInsert();
+
START_CRIT_SECTION();
/*
off++;
}
- Assert((ptr - rdata[1].data) <= collector->sumsize);
+ Assert((ptr - collectordata) <= collector->sumsize);
+ if (needWal)
+ {
+ XLogRegisterBuffer(1, buffer, REGBUF_STANDARD);
+ XLogRegisterBufData(1, collectordata, collector->sumsize);
+ }
metadata->tailFreeSize = PageGetExactFreeSpace(page);
*/
MarkBufferDirty(metabuffer);
- if (RelationNeedsWAL(index))
+ if (needWal)
{
XLogRecPtr recptr;
memcpy(&data.metadata, metadata, sizeof(GinMetaPageData));
- recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_UPDATE_META_PAGE, rdata);
+ XLogRegisterBuffer(0, metabuffer, REGBUF_WILL_INIT);
+ XLogRegisterData((char *) &data, sizeof(ginxlogUpdateMeta));
+
+ recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_UPDATE_META_PAGE);
PageSetLSN(metapage, recptr);
if (buffer != InvalidBuffer)
int i;
int64 nDeletedHeapTuples = 0;
ginxlogDeleteListPages data;
- XLogRecData rdata[1];
Buffer buffers[GIN_NDELETE_AT_ONCE];
- data.node = index->rd_node;
-
- rdata[0].buffer = InvalidBuffer;
- rdata[0].data = (char *) &data;
- rdata[0].len = sizeof(ginxlogDeleteListPages);
- rdata[0].next = NULL;
-
data.ndeleted = 0;
while (data.ndeleted < GIN_NDELETE_AT_ONCE && blknoToDelete != newHead)
{
- data.toDelete[data.ndeleted] = blknoToDelete;
buffers[data.ndeleted] = ReadBuffer(index, blknoToDelete);
LockBuffer(buffers[data.ndeleted], GIN_EXCLUSIVE);
page = BufferGetPage(buffers[data.ndeleted]);
if (stats)
stats->pages_deleted += data.ndeleted;
+ /*
+ * This operation touches an unusually large number of pages, so
+ * prepare the XLogInsert machinery for that before entering the
+ * critical section.
+ */
+ XLogEnsureRecordSpace(data.ndeleted, 0);
+
START_CRIT_SECTION();
metadata->head = blknoToDelete;
{
XLogRecPtr recptr;
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, metabuffer, REGBUF_WILL_INIT);
+ for (i = 0; i < data.ndeleted; i++)
+ XLogRegisterBuffer(i + 1, buffers[i], REGBUF_WILL_INIT);
+
memcpy(&data.metadata, metadata, sizeof(GinMetaPageData));
- recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_DELETE_LISTPAGE, rdata);
+ XLogRegisterData((char *) &data,
+ sizeof(ginxlogDeleteListPages));
+
+ recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_DELETE_LISTPAGE);
PageSetLSN(metapage, recptr);
for (i = 0; i < data.ndeleted; i++)
if (RelationNeedsWAL(index))
{
XLogRecPtr recptr;
- XLogRecData rdata;
Page page;
- rdata.buffer = InvalidBuffer;
- rdata.data = (char *) &(index->rd_node);
- rdata.len = sizeof(RelFileNode);
- rdata.next = NULL;
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, MetaBuffer, REGBUF_WILL_INIT);
+ XLogRegisterBuffer(1, RootBuffer, REGBUF_WILL_INIT);
- recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_CREATE_INDEX, &rdata);
+ recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_CREATE_INDEX);
page = BufferGetPage(RootBuffer);
PageSetLSN(page, recptr);
{
XLogRecPtr recptr;
ginxlogUpdateMeta data;
- XLogRecData rdata;
data.node = index->rd_node;
data.ntuples = 0;
data.newRightlink = data.prevTail = InvalidBlockNumber;
memcpy(&data.metadata, metadata, sizeof(GinMetaPageData));
- rdata.buffer = InvalidBuffer;
- rdata.data = (char *) &data;
- rdata.len = sizeof(ginxlogUpdateMeta);
- rdata.next = NULL;
+ XLogBeginInsert();
+ XLogRegisterData((char *) &data, sizeof(ginxlogUpdateMeta));
+ XLogRegisterBuffer(0, metabuffer, REGBUF_WILL_INIT);
- recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_UPDATE_META_PAGE, &rdata);
+ recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_UPDATE_META_PAGE);
PageSetLSN(metapage, recptr);
}
{
Page page = BufferGetPage(buffer);
XLogRecPtr recptr;
- XLogRecData rdata[3];
- ginxlogVacuumPage xlrec;
- uint16 lower;
- uint16 upper;
/* This is only used for entry tree leaf pages. */
Assert(!GinPageIsData(page));
if (!RelationNeedsWAL(index))
return;
- xlrec.node = index->rd_node;
- xlrec.blkno = BufferGetBlockNumber(buffer);
-
- /* Assume we can omit data between pd_lower and pd_upper */
- lower = ((PageHeader) page)->pd_lower;
- upper = ((PageHeader) page)->pd_upper;
-
- Assert(lower < BLCKSZ);
- Assert(upper < BLCKSZ);
-
- if (lower >= SizeOfPageHeaderData &&
- upper > lower &&
- upper <= BLCKSZ)
- {
- xlrec.hole_offset = lower;
- xlrec.hole_length = upper - lower;
- }
- else
- {
- /* No "hole" to compress out */
- xlrec.hole_offset = 0;
- xlrec.hole_length = 0;
- }
-
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = sizeof(ginxlogVacuumPage);
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &rdata[1];
-
- if (xlrec.hole_length == 0)
- {
- rdata[1].data = (char *) page;
- rdata[1].len = BLCKSZ;
- rdata[1].buffer = InvalidBuffer;
- rdata[1].next = NULL;
- }
- else
- {
- /* must skip the hole */
- rdata[1].data = (char *) page;
- rdata[1].len = xlrec.hole_offset;
- rdata[1].buffer = InvalidBuffer;
- rdata[1].next = &rdata[2];
-
- rdata[2].data = (char *) page + (xlrec.hole_offset + xlrec.hole_length);
- rdata[2].len = BLCKSZ - (xlrec.hole_offset + xlrec.hole_length);
- rdata[2].buffer = InvalidBuffer;
- rdata[2].next = NULL;
- }
+ /*
+ * Always create a full image, we don't track the changes on the page at
+ * any more fine-grained level. This could obviously be improved...
+ */
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, buffer, REGBUF_FORCE_IMAGE | REGBUF_STANDARD);
- recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_VACUUM_PAGE, rdata);
+ recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_VACUUM_PAGE);
PageSetLSN(page, recptr);
}
if (RelationNeedsWAL(gvs->index))
{
XLogRecPtr recptr;
- XLogRecData rdata[4];
ginxlogDeletePage data;
- data.node = gvs->index->rd_node;
- data.blkno = deleteBlkno;
- data.parentBlkno = parentBlkno;
+ /*
+ * We can't pass REGBUF_STANDARD for the deleted page, because we
+ * didn't set pd_lower on pre-9.4 versions. The page might've been
+ * binary-upgraded from an older version, and hence not have pd_lower
+ * set correctly. Ditto for the left page, but removing the item from
+ * the parent updated its pd_lower, so we know that's OK at this
+ * point.
+ */
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, dBuffer, 0);
+ XLogRegisterBuffer(1, pBuffer, REGBUF_STANDARD);
+ XLogRegisterBuffer(2, lBuffer, 0);
+
data.parentOffset = myoff;
- data.leftBlkno = leftBlkno;
data.rightLink = GinPageGetOpaque(page)->rightlink;
- /*
- * We can't pass buffer_std = TRUE, because we didn't set pd_lower on
- * pre-9.4 versions. The page might've been binary-upgraded from an
- * older version, and hence not have pd_lower set correctly. Ditto for
- * the left page, but removing the item from the parent updated its
- * pd_lower, so we know that's OK at this point.
- */
- rdata[0].buffer = dBuffer;
- rdata[0].buffer_std = FALSE;
- rdata[0].data = NULL;
- rdata[0].len = 0;
- rdata[0].next = rdata + 1;
-
- rdata[1].buffer = pBuffer;
- rdata[1].buffer_std = TRUE;
- rdata[1].data = NULL;
- rdata[1].len = 0;
- rdata[1].next = rdata + 2;
-
- rdata[2].buffer = lBuffer;
- rdata[2].buffer_std = FALSE;
- rdata[2].data = NULL;
- rdata[2].len = 0;
- rdata[2].next = rdata + 3;
-
- rdata[3].buffer = InvalidBuffer;
- rdata[3].buffer_std = FALSE;
- rdata[3].len = sizeof(ginxlogDeletePage);
- rdata[3].data = (char *) &data;
- rdata[3].next = NULL;
-
- recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_DELETE_PAGE, rdata);
+ XLogRegisterData((char *) &data, sizeof(ginxlogDeletePage));
+
+ recptr = XLogInsert(RM_GIN_ID, XLOG_GIN_DELETE_PAGE);
PageSetLSN(page, recptr);
PageSetLSN(parentPage, recptr);
PageSetLSN(BufferGetPage(lBuffer), recptr);
static MemoryContext opCtx; /* working memory for operations */
static void
-ginRedoClearIncompleteSplit(XLogRecPtr lsn, XLogRecord *record,
- int block_index,
- RelFileNode node, BlockNumber blkno)
+ginRedoClearIncompleteSplit(XLogReaderState *record, uint8 block_id)
{
+ XLogRecPtr lsn = record->EndRecPtr;
Buffer buffer;
Page page;
- if (XLogReadBufferForRedo(lsn, record, block_index, node, blkno, &buffer)
- == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, block_id, &buffer) == BLK_NEEDS_REDO)
{
page = (Page) BufferGetPage(buffer);
-
GinPageGetOpaque(page)->flags &= ~GIN_INCOMPLETE_SPLIT;
PageSetLSN(page, lsn);
}
static void
-ginRedoCreateIndex(XLogRecPtr lsn, XLogRecord *record)
+ginRedoCreateIndex(XLogReaderState *record)
{
- RelFileNode *node = (RelFileNode *) XLogRecGetData(record);
+ XLogRecPtr lsn = record->EndRecPtr;
Buffer RootBuffer,
MetaBuffer;
Page page;
- /* Backup blocks are not used in create_index records */
- Assert(!(record->xl_info & XLR_BKP_BLOCK_MASK));
-
- MetaBuffer = XLogReadBuffer(*node, GIN_METAPAGE_BLKNO, true);
- Assert(BufferIsValid(MetaBuffer));
+ MetaBuffer = XLogInitBufferForRedo(record, 0);
+ Assert(BufferGetBlockNumber(MetaBuffer) == GIN_METAPAGE_BLKNO);
page = (Page) BufferGetPage(MetaBuffer);
GinInitMetabuffer(MetaBuffer);
PageSetLSN(page, lsn);
MarkBufferDirty(MetaBuffer);
- RootBuffer = XLogReadBuffer(*node, GIN_ROOT_BLKNO, true);
- Assert(BufferIsValid(RootBuffer));
+ RootBuffer = XLogInitBufferForRedo(record, 1);
+ Assert(BufferGetBlockNumber(RootBuffer) == GIN_ROOT_BLKNO);
page = (Page) BufferGetPage(RootBuffer);
GinInitBuffer(RootBuffer, GIN_LEAF);
}
static void
-ginRedoCreatePTree(XLogRecPtr lsn, XLogRecord *record)
+ginRedoCreatePTree(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
ginxlogCreatePostingTree *data = (ginxlogCreatePostingTree *) XLogRecGetData(record);
char *ptr;
Buffer buffer;
Page page;
- /* Backup blocks are not used in create_ptree records */
- Assert(!(record->xl_info & XLR_BKP_BLOCK_MASK));
-
- buffer = XLogReadBuffer(data->node, data->blkno, true);
- Assert(BufferIsValid(buffer));
+ buffer = XLogInitBufferForRedo(record, 0);
page = (Page) BufferGetPage(buffer);
GinInitBuffer(buffer, GIN_DATA | GIN_LEAF | GIN_COMPRESSED);
}
static void
-ginRedoInsert(XLogRecPtr lsn, XLogRecord *record)
+ginRedoInsert(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
ginxlogInsert *data = (ginxlogInsert *) XLogRecGetData(record);
Buffer buffer;
- char *payload;
+#ifdef NOT_USED
BlockNumber leftChildBlkno = InvalidBlockNumber;
+#endif
BlockNumber rightChildBlkno = InvalidBlockNumber;
bool isLeaf = (data->flags & GIN_INSERT_ISLEAF) != 0;
- payload = XLogRecGetData(record) + sizeof(ginxlogInsert);
-
/*
* First clear incomplete-split flag on child page if this finishes a
* split.
*/
if (!isLeaf)
{
+ char *payload = XLogRecGetData(record) + sizeof(ginxlogInsert);
+
+#ifdef NOT_USED
leftChildBlkno = BlockIdGetBlockNumber((BlockId) payload);
+#endif
payload += sizeof(BlockIdData);
rightChildBlkno = BlockIdGetBlockNumber((BlockId) payload);
payload += sizeof(BlockIdData);
- ginRedoClearIncompleteSplit(lsn, record, 0, data->node, leftChildBlkno);
+ ginRedoClearIncompleteSplit(record, 1);
}
- if (XLogReadBufferForRedo(lsn, record, isLeaf ? 0 : 1, data->node,
- data->blkno, &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
Page page = BufferGetPage(buffer);
+ Size len;
+ char *payload = XLogRecGetBlockData(record, 0, &len);
/* How to insert the payload is tree-type specific */
if (data->flags & GIN_INSERT_ISDATA)
}
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
-ginRedoSplitData(Page lpage, Page rpage, void *rdata)
-{
- bool isleaf = GinPageIsLeaf(lpage);
-
- if (isleaf)
- {
- ginxlogSplitDataLeaf *data = (ginxlogSplitDataLeaf *) rdata;
- Pointer lptr = (Pointer) rdata + sizeof(ginxlogSplitDataLeaf);
- Pointer rptr = lptr + data->lsize;
-
- Assert(data->lsize > 0 && data->lsize <= GinDataPageMaxDataSize);
- Assert(data->rsize > 0 && data->rsize <= GinDataPageMaxDataSize);
-
- memcpy(GinDataLeafPageGetPostingList(lpage), lptr, data->lsize);
- memcpy(GinDataLeafPageGetPostingList(rpage), rptr, data->rsize);
-
- GinDataPageSetDataSize(lpage, data->lsize);
- GinDataPageSetDataSize(rpage, data->rsize);
- *GinDataPageGetRightBound(lpage) = data->lrightbound;
- *GinDataPageGetRightBound(rpage) = data->rrightbound;
- }
- else
- {
- ginxlogSplitDataInternal *data = (ginxlogSplitDataInternal *) rdata;
- PostingItem *items = (PostingItem *) ((char *) rdata + sizeof(ginxlogSplitDataInternal));
- OffsetNumber i;
- OffsetNumber maxoff;
-
- for (i = 0; i < data->separator; i++)
- GinDataPageAddPostingItem(lpage, &items[i], InvalidOffsetNumber);
- for (i = data->separator; i < data->nitem; i++)
- GinDataPageAddPostingItem(rpage, &items[i], InvalidOffsetNumber);
-
- /* set up right key */
- maxoff = GinPageGetOpaque(lpage)->maxoff;
- *GinDataPageGetRightBound(lpage) = GinDataPageGetPostingItem(lpage, maxoff)->key;
- *GinDataPageGetRightBound(rpage) = data->rightbound;
- }
-}
-
-static void
-ginRedoSplit(XLogRecPtr lsn, XLogRecord *record)
+ginRedoSplit(XLogReaderState *record)
{
ginxlogSplit *data = (ginxlogSplit *) XLogRecGetData(record);
Buffer lbuffer,
- rbuffer;
- Page lpage,
- rpage;
- uint32 flags;
- uint32 lflags,
- rflags;
- char *payload;
+ rbuffer,
+ rootbuf;
bool isLeaf = (data->flags & GIN_INSERT_ISLEAF) != 0;
- bool isData = (data->flags & GIN_INSERT_ISDATA) != 0;
bool isRoot = (data->flags & GIN_SPLIT_ROOT) != 0;
- payload = XLogRecGetData(record) + sizeof(ginxlogSplit);
-
/*
* First clear incomplete-split flag on child page if this finishes a
* split
*/
if (!isLeaf)
- ginRedoClearIncompleteSplit(lsn, record, 0, data->node, data->leftChildBlkno);
-
- flags = 0;
- if (isLeaf)
- flags |= GIN_LEAF;
- if (isData)
- flags |= GIN_DATA;
- if (isLeaf && isData)
- flags |= GIN_COMPRESSED;
-
- lflags = rflags = flags;
- if (!isRoot)
- lflags |= GIN_INCOMPLETE_SPLIT;
-
- lbuffer = XLogReadBuffer(data->node, data->lblkno, true);
- Assert(BufferIsValid(lbuffer));
- lpage = (Page) BufferGetPage(lbuffer);
- GinInitBuffer(lbuffer, lflags);
-
- rbuffer = XLogReadBuffer(data->node, data->rblkno, true);
- Assert(BufferIsValid(rbuffer));
- rpage = (Page) BufferGetPage(rbuffer);
- GinInitBuffer(rbuffer, rflags);
-
- GinPageGetOpaque(lpage)->rightlink = BufferGetBlockNumber(rbuffer);
- GinPageGetOpaque(rpage)->rightlink = isRoot ? InvalidBlockNumber : data->rrlink;
-
- /* Do the tree-type specific portion to restore the page contents */
- if (isData)
- ginRedoSplitData(lpage, rpage, payload);
- else
- ginRedoSplitEntry(lpage, rpage, payload);
+ ginRedoClearIncompleteSplit(record, 3);
- PageSetLSN(rpage, lsn);
- MarkBufferDirty(rbuffer);
+ if (XLogReadBufferForRedo(record, 0, &lbuffer) != BLK_RESTORED)
+ elog(ERROR, "GIN split record did not contain a full-page image of left page");
- PageSetLSN(lpage, lsn);
- MarkBufferDirty(lbuffer);
+ if (XLogReadBufferForRedo(record, 1, &rbuffer) != BLK_RESTORED)
+ elog(ERROR, "GIN split record did not contain a full-page image of right page");
if (isRoot)
{
- BlockNumber rootBlkno = data->rrlink;
- Buffer rootBuf = XLogReadBuffer(data->node, rootBlkno, true);
- Page rootPage = BufferGetPage(rootBuf);
-
- GinInitBuffer(rootBuf, flags & ~GIN_LEAF & ~GIN_COMPRESSED);
-
- if (isData)
- {
- Assert(rootBlkno != GIN_ROOT_BLKNO);
- ginDataFillRoot(NULL, BufferGetPage(rootBuf),
- BufferGetBlockNumber(lbuffer),
- BufferGetPage(lbuffer),
- BufferGetBlockNumber(rbuffer),
- BufferGetPage(rbuffer));
- }
- else
- {
- Assert(rootBlkno == GIN_ROOT_BLKNO);
- ginEntryFillRoot(NULL, BufferGetPage(rootBuf),
- BufferGetBlockNumber(lbuffer),
- BufferGetPage(lbuffer),
- BufferGetBlockNumber(rbuffer),
- BufferGetPage(rbuffer));
- }
-
- PageSetLSN(rootPage, lsn);
-
- MarkBufferDirty(rootBuf);
- UnlockReleaseBuffer(rootBuf);
+ if (XLogReadBufferForRedo(record, 2, &rootbuf) != BLK_RESTORED)
+ elog(ERROR, "GIN split record did not contain a full-page image of root page");
+ UnlockReleaseBuffer(rootbuf);
}
UnlockReleaseBuffer(rbuffer);
* a XLOG_FPI record.
*/
static void
-ginRedoVacuumPage(XLogRecPtr lsn, XLogRecord *record)
+ginRedoVacuumPage(XLogReaderState *record)
{
- ginxlogVacuumPage *xlrec = (ginxlogVacuumPage *) XLogRecGetData(record);
- char *blk = ((char *) xlrec) + sizeof(ginxlogVacuumPage);
Buffer buffer;
- Page page;
-
- Assert(xlrec->hole_offset < BLCKSZ);
- Assert(xlrec->hole_length < BLCKSZ);
-
- /* Backup blocks are not used, we'll re-initialize the page always. */
- Assert(!(record->xl_info & XLR_BKP_BLOCK_MASK));
- buffer = XLogReadBuffer(xlrec->node, xlrec->blkno, true);
- if (!BufferIsValid(buffer))
- return;
- page = (Page) BufferGetPage(buffer);
-
- if (xlrec->hole_length == 0)
+ if (XLogReadBufferForRedo(record, 0, &buffer) != BLK_RESTORED)
{
- memcpy((char *) page, blk, BLCKSZ);
+ elog(ERROR, "replay of gin entry tree page vacuum did not restore the page");
}
- else
- {
- memcpy((char *) page, blk, xlrec->hole_offset);
- /* must zero-fill the hole */
- MemSet((char *) page + xlrec->hole_offset, 0, xlrec->hole_length);
- memcpy((char *) page + (xlrec->hole_offset + xlrec->hole_length),
- blk + xlrec->hole_offset,
- BLCKSZ - (xlrec->hole_offset + xlrec->hole_length));
- }
-
- PageSetLSN(page, lsn);
-
- MarkBufferDirty(buffer);
UnlockReleaseBuffer(buffer);
}
static void
-ginRedoVacuumDataLeafPage(XLogRecPtr lsn, XLogRecord *record)
+ginRedoVacuumDataLeafPage(XLogReaderState *record)
{
- ginxlogVacuumDataLeafPage *xlrec = (ginxlogVacuumDataLeafPage *) XLogRecGetData(record);
+ XLogRecPtr lsn = record->EndRecPtr;
Buffer buffer;
- if (XLogReadBufferForRedo(lsn, record, 0, xlrec->node, xlrec->blkno,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
Page page = BufferGetPage(buffer);
+ Size len;
+ ginxlogVacuumDataLeafPage *xlrec;
+
+ xlrec = (ginxlogVacuumDataLeafPage *) XLogRecGetBlockData(record, 0, &len);
Assert(GinPageIsLeaf(page));
Assert(GinPageIsData(page));
}
static void
-ginRedoDeletePage(XLogRecPtr lsn, XLogRecord *record)
+ginRedoDeletePage(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
ginxlogDeletePage *data = (ginxlogDeletePage *) XLogRecGetData(record);
Buffer dbuffer;
Buffer pbuffer;
Buffer lbuffer;
Page page;
- if (XLogReadBufferForRedo(lsn, record, 0, data->node, data->blkno, &dbuffer)
- == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &dbuffer) == BLK_NEEDS_REDO)
{
page = BufferGetPage(dbuffer);
-
Assert(GinPageIsData(page));
GinPageGetOpaque(page)->flags = GIN_DELETED;
PageSetLSN(page, lsn);
MarkBufferDirty(dbuffer);
}
- if (XLogReadBufferForRedo(lsn, record, 1, data->node, data->parentBlkno,
- &pbuffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 1, &pbuffer) == BLK_NEEDS_REDO)
{
page = BufferGetPage(pbuffer);
-
Assert(GinPageIsData(page));
Assert(!GinPageIsLeaf(page));
GinPageDeletePostingItem(page, data->parentOffset);
MarkBufferDirty(pbuffer);
}
- if (XLogReadBufferForRedo(lsn, record, 2, data->node, data->leftBlkno,
- &lbuffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 2, &lbuffer) == BLK_NEEDS_REDO)
{
page = BufferGetPage(lbuffer);
-
Assert(GinPageIsData(page));
GinPageGetOpaque(page)->rightlink = data->rightLink;
PageSetLSN(page, lsn);
}
static void
-ginRedoUpdateMetapage(XLogRecPtr lsn, XLogRecord *record)
+ginRedoUpdateMetapage(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
ginxlogUpdateMeta *data = (ginxlogUpdateMeta *) XLogRecGetData(record);
Buffer metabuffer;
Page metapage;
* image, so restore the metapage unconditionally without looking at the
* LSN, to avoid torn page hazards.
*/
- metabuffer = XLogReadBuffer(data->node, GIN_METAPAGE_BLKNO, false);
- if (!BufferIsValid(metabuffer))
- return; /* assume index was deleted, nothing to do */
+ metabuffer = XLogInitBufferForRedo(record, 0);
+ Assert(BufferGetBlockNumber(metabuffer) == GIN_METAPAGE_BLKNO);
metapage = BufferGetPage(metabuffer);
memcpy(GinPageGetMeta(metapage), &data->metadata, sizeof(GinMetaPageData));
/*
* insert into tail page
*/
- if (XLogReadBufferForRedo(lsn, record, 0, data->node,
- data->metadata.tail, &buffer)
- == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 1, &buffer) == BLK_NEEDS_REDO)
{
Page page = BufferGetPage(buffer);
OffsetNumber off;
int i;
Size tupsize;
+ char *payload;
IndexTuple tuples;
+ Size totaltupsize;
- tuples = (IndexTuple) (XLogRecGetData(record) + sizeof(ginxlogUpdateMeta));
+ payload = XLogRecGetBlockData(record, 1, &totaltupsize);
+ tuples = (IndexTuple) payload;
if (PageIsEmpty(page))
off = FirstOffsetNumber;
off++;
}
+ Assert(payload + totaltupsize == (char *) tuples);
/*
* Increase counter of heap tuples
/*
* New tail
*/
- if (XLogReadBufferForRedo(lsn, record, 0, data->node, data->prevTail,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 1, &buffer) == BLK_NEEDS_REDO)
{
Page page = BufferGetPage(buffer);
}
static void
-ginRedoInsertListPage(XLogRecPtr lsn, XLogRecord *record)
+ginRedoInsertListPage(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
ginxlogInsertListPage *data = (ginxlogInsertListPage *) XLogRecGetData(record);
Buffer buffer;
Page page;
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));
+ char *payload;
+ IndexTuple tuples;
+ Size totaltupsize;
- buffer = XLogReadBuffer(data->node, data->blkno, true);
- Assert(BufferIsValid(buffer));
+ /* We always re-initialize the page. */
+ buffer = XLogInitBufferForRedo(record, 0);
page = BufferGetPage(buffer);
GinInitBuffer(buffer, GIN_LIST);
GinPageGetOpaque(page)->maxoff = 0;
}
+ payload = XLogRecGetBlockData(record, 0, &totaltupsize);
+
+ tuples = (IndexTuple) payload;
for (i = 0; i < data->ntuples; i++)
{
tupsize = IndexTupleSize(tuples);
tuples = (IndexTuple) (((char *) tuples) + tupsize);
off++;
}
+ Assert((char *) tuples == payload + totaltupsize);
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
static void
-ginRedoDeleteListPages(XLogRecPtr lsn, XLogRecord *record)
+ginRedoDeleteListPages(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
ginxlogDeleteListPages *data = (ginxlogDeleteListPages *) XLogRecGetData(record);
Buffer metabuffer;
Page metapage;
int i;
- /* Backup blocks are not used in delete_listpage records */
- Assert(!(record->xl_info & XLR_BKP_BLOCK_MASK));
-
- metabuffer = XLogReadBuffer(data->node, GIN_METAPAGE_BLKNO, false);
- if (!BufferIsValid(metabuffer))
- return; /* assume index was deleted, nothing to do */
+ metabuffer = XLogInitBufferForRedo(record, 0);
+ Assert(BufferGetBlockNumber(metabuffer) == GIN_METAPAGE_BLKNO);
metapage = BufferGetPage(metabuffer);
+ GinInitPage(metapage, GIN_META, BufferGetPageSize(metabuffer));
+
memcpy(GinPageGetMeta(metapage), &data->metadata, sizeof(GinMetaPageData));
PageSetLSN(metapage, lsn);
MarkBufferDirty(metabuffer);
Buffer buffer;
Page page;
- buffer = XLogReadBuffer(data->node, data->toDelete[i], true);
+ buffer = XLogInitBufferForRedo(record, i + 1);
page = BufferGetPage(buffer);
GinInitBuffer(buffer, GIN_DELETED);
}
void
-gin_redo(XLogRecPtr lsn, XLogRecord *record)
+gin_redo(XLogReaderState *record)
{
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
MemoryContext oldCtx;
/*
switch (info)
{
case XLOG_GIN_CREATE_INDEX:
- ginRedoCreateIndex(lsn, record);
+ ginRedoCreateIndex(record);
break;
case XLOG_GIN_CREATE_PTREE:
- ginRedoCreatePTree(lsn, record);
+ ginRedoCreatePTree(record);
break;
case XLOG_GIN_INSERT:
- ginRedoInsert(lsn, record);
+ ginRedoInsert(record);
break;
case XLOG_GIN_SPLIT:
- ginRedoSplit(lsn, record);
+ ginRedoSplit(record);
break;
case XLOG_GIN_VACUUM_PAGE:
- ginRedoVacuumPage(lsn, record);
+ ginRedoVacuumPage(record);
break;
case XLOG_GIN_VACUUM_DATA_LEAF_PAGE:
- ginRedoVacuumDataLeafPage(lsn, record);
+ ginRedoVacuumDataLeafPage(record);
break;
case XLOG_GIN_DELETE_PAGE:
- ginRedoDeletePage(lsn, record);
+ ginRedoDeletePage(record);
break;
case XLOG_GIN_UPDATE_META_PAGE:
- ginRedoUpdateMetapage(lsn, record);
+ ginRedoUpdateMetapage(record);
break;
case XLOG_GIN_INSERT_LISTPAGE:
- ginRedoInsertListPage(lsn, record);
+ ginRedoInsertListPage(record);
break;
case XLOG_GIN_DELETE_LISTPAGE:
- ginRedoDeleteListPages(lsn, record);
+ ginRedoDeleteListPages(record);
break;
default:
elog(PANIC, "gin_redo: unknown op code %u", info);
#include "access/genam.h"
#include "access/gist_private.h"
+#include "access/xloginsert.h"
#include "catalog/index.h"
#include "catalog/pg_collation.h"
#include "miscadmin.h"
GistPageSetNSN(ptr->page, oldnsn);
}
+ /*
+ * gistXLogSplit() needs to WAL log a lot of pages, prepare WAL
+ * insertion for that. NB: The number of pages and data segments
+ * specified here must match the calculations in gistXLogSplit()!
+ */
+ if (RelationNeedsWAL(rel))
+ XLogEnsureRecordSpace(npage, 1 + npage * 2);
+
START_CRIT_SECTION();
/*
if (RelationNeedsWAL(index))
{
XLogRecPtr recptr;
- XLogRecData rdata;
- rdata.data = (char *) &(index->rd_node);
- rdata.len = sizeof(RelFileNode);
- rdata.buffer = InvalidBuffer;
- rdata.next = NULL;
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, buffer, REGBUF_WILL_INIT);
- recptr = XLogInsert(RM_GIST_ID, XLOG_GIST_CREATE_INDEX, &rdata);
+ recptr = XLogInsert(RM_GIST_ID, XLOG_GIST_CREATE_INDEX);
PageSetLSN(page, recptr);
}
else
#include "access/xlogutils.h"
#include "utils/memutils.h"
-typedef struct
-{
- gistxlogPage *header;
- IndexTuple *itup;
-} NewPage;
-
-typedef struct
-{
- gistxlogPageSplit *data;
- NewPage *page;
-} PageSplitRecord;
-
static MemoryContext opCtx; /* working memory for operations */
/*
* action.)
*/
static void
-gistRedoClearFollowRight(XLogRecPtr lsn, XLogRecord *record, int block_index,
- RelFileNode node, BlockNumber childblkno)
+gistRedoClearFollowRight(XLogReaderState *record, uint8 block_id)
{
+ XLogRecPtr lsn = record->EndRecPtr;
Buffer buffer;
Page page;
XLogRedoAction action;
* Note that we still update the page even if it was restored from a full
* page image, because the updated NSN is not included in the image.
*/
- action = XLogReadBufferForRedo(lsn, record, block_index, node, childblkno,
- &buffer);
+ action = XLogReadBufferForRedo(record, block_id, &buffer);
if (action == BLK_NEEDS_REDO || action == BLK_RESTORED)
{
page = BufferGetPage(buffer);
* redo any page update (except page split)
*/
static void
-gistRedoPageUpdateRecord(XLogRecPtr lsn, XLogRecord *record)
+gistRedoPageUpdateRecord(XLogReaderState *record)
{
- char *begin = XLogRecGetData(record);
- gistxlogPageUpdate *xldata = (gistxlogPageUpdate *) begin;
+ XLogRecPtr lsn = record->EndRecPtr;
+ gistxlogPageUpdate *xldata = (gistxlogPageUpdate *) XLogRecGetData(record);
Buffer buffer;
Page page;
- char *data;
- if (XLogReadBufferForRedo(lsn, record, 0, xldata->node, xldata->blkno,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
- page = (Page) BufferGetPage(buffer);
+ char *begin;
+ char *data;
+ Size datalen;
+ int ninserted = 0;
- data = begin + sizeof(gistxlogPageUpdate);
+ data = begin = XLogRecGetBlockData(record, 0, &datalen);
+
+ page = (Page) BufferGetPage(buffer);
/* Delete old tuples */
if (xldata->ntodelete > 0)
}
/* add tuples */
- if (data - begin < record->xl_len)
+ if (data - begin < datalen)
{
OffsetNumber off = (PageIsEmpty(page)) ? FirstOffsetNumber :
OffsetNumberNext(PageGetMaxOffsetNumber(page));
- while (data - begin < record->xl_len)
+ while (data - begin < datalen)
{
IndexTuple itup = (IndexTuple) data;
Size sz = IndexTupleSize(itup);
elog(ERROR, "failed to add item to GiST index page, size %d bytes",
(int) sz);
off++;
+ ninserted++;
}
}
+ Assert(ninserted == xldata->ntoinsert);
+
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
* that even if the target page no longer exists, we still attempt to
* replay the change on the child page.
*/
- if (BlockNumberIsValid(xldata->leftchild))
- gistRedoClearFollowRight(lsn, record, 1,
- xldata->node, xldata->leftchild);
+ if (XLogRecHasBlockRef(record, 1))
+ gistRedoClearFollowRight(record, 1);
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
}
-static void
-decodePageSplitRecord(PageSplitRecord *decoded, XLogRecord *record)
+/*
+ * Returns an array of index pointers.
+ */
+static IndexTuple *
+decodePageSplitRecord(char *begin, int len, int *n)
{
- char *begin = XLogRecGetData(record),
- *ptr;
- int j,
- i = 0;
+ char *ptr;
+ int i = 0;
+ IndexTuple *tuples;
+
+ /* extract the number of tuples */
+ memcpy(n, begin, sizeof(int));
+ ptr = begin + sizeof(int);
- decoded->data = (gistxlogPageSplit *) begin;
- decoded->page = (NewPage *) palloc(sizeof(NewPage) * decoded->data->npage);
+ tuples = palloc(*n * sizeof(IndexTuple));
- ptr = begin + sizeof(gistxlogPageSplit);
- for (i = 0; i < decoded->data->npage; i++)
+ for (i = 0; i < *n; 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++;
- }
+ Assert(ptr - begin < len);
+ tuples[i] = (IndexTuple) ptr;
+ ptr += IndexTupleSize((IndexTuple) ptr);
}
+ Assert(ptr - begin == len);
+
+ return tuples;
}
static void
-gistRedoPageSplitRecord(XLogRecPtr lsn, XLogRecord *record)
+gistRedoPageSplitRecord(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
gistxlogPageSplit *xldata = (gistxlogPageSplit *) XLogRecGetData(record);
- PageSplitRecord xlrec;
Buffer firstbuffer = InvalidBuffer;
Buffer buffer;
Page page;
int i;
bool isrootsplit = false;
- decodePageSplitRecord(&xlrec, record);
-
/*
* We must hold lock on the first-listed page throughout the action,
* including while updating the left child page (if any). We can unlock
*/
/* loop around all pages */
- for (i = 0; i < xlrec.data->npage; i++)
+ for (i = 0; i < xldata->npage; i++)
{
- NewPage *newpage = xlrec.page + i;
int flags;
-
- if (newpage->header->blkno == GIST_ROOT_BLKNO)
+ char *data;
+ Size datalen;
+ int num;
+ BlockNumber blkno;
+ IndexTuple *tuples;
+
+ XLogRecGetBlockTag(record, i + 1, NULL, NULL, &blkno);
+ if (blkno == GIST_ROOT_BLKNO)
{
Assert(i == 0);
isrootsplit = true;
}
- buffer = XLogReadBuffer(xlrec.data->node, newpage->header->blkno, true);
- Assert(BufferIsValid(buffer));
+ buffer = XLogInitBufferForRedo(record, i + 1);
page = (Page) BufferGetPage(buffer);
+ data = XLogRecGetBlockData(record, i + 1, &datalen);
+
+ tuples = decodePageSplitRecord(data, datalen, &num);
/* ok, clear buffer */
- if (xlrec.data->origleaf && newpage->header->blkno != GIST_ROOT_BLKNO)
+ if (xldata->origleaf && blkno != GIST_ROOT_BLKNO)
flags = F_LEAF;
else
flags = 0;
GISTInitBuffer(buffer, flags);
/* and fill it */
- gistfillbuffer(page, newpage->itup, newpage->header->num, FirstOffsetNumber);
+ gistfillbuffer(page, tuples, num, FirstOffsetNumber);
- if (newpage->header->blkno == GIST_ROOT_BLKNO)
+ if (blkno == GIST_ROOT_BLKNO)
{
GistPageGetOpaque(page)->rightlink = InvalidBlockNumber;
GistPageSetNSN(page, xldata->orignsn);
}
else
{
- if (i < xlrec.data->npage - 1)
- GistPageGetOpaque(page)->rightlink = xlrec.page[i + 1].header->blkno;
+ if (i < xldata->npage - 1)
+ {
+ BlockNumber nextblkno;
+
+ XLogRecGetBlockTag(record, i + 2, NULL, NULL, &nextblkno);
+ GistPageGetOpaque(page)->rightlink = nextblkno;
+ }
else
GistPageGetOpaque(page)->rightlink = xldata->origrlink;
GistPageSetNSN(page, xldata->orignsn);
- if (i < xlrec.data->npage - 1 && !isrootsplit &&
+ if (i < xldata->npage - 1 && !isrootsplit &&
xldata->markfollowright)
GistMarkFollowRight(page);
else
}
/* Fix follow-right data on left child page, if any */
- if (BlockNumberIsValid(xldata->leftchild))
- gistRedoClearFollowRight(lsn, record, 0,
- xldata->node, xldata->leftchild);
+ if (XLogRecHasBlockRef(record, 0))
+ gistRedoClearFollowRight(record, 0);
/* Finally, release lock on the first page */
UnlockReleaseBuffer(firstbuffer);
}
static void
-gistRedoCreateIndex(XLogRecPtr lsn, XLogRecord *record)
+gistRedoCreateIndex(XLogReaderState *record)
{
- RelFileNode *node = (RelFileNode *) XLogRecGetData(record);
+ XLogRecPtr lsn = record->EndRecPtr;
Buffer buffer;
Page page;
- /* Backup blocks are not used in create_index records */
- Assert(!(record->xl_info & XLR_BKP_BLOCK_MASK));
-
- buffer = XLogReadBuffer(*node, GIST_ROOT_BLKNO, true);
- Assert(BufferIsValid(buffer));
+ buffer = XLogInitBufferForRedo(record, 0);
+ Assert(BufferGetBlockNumber(buffer) == GIST_ROOT_BLKNO);
page = (Page) BufferGetPage(buffer);
GISTInitBuffer(buffer, F_LEAF);
}
void
-gist_redo(XLogRecPtr lsn, XLogRecord *record)
+gist_redo(XLogReaderState *record)
{
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
MemoryContext oldCxt;
/*
switch (info)
{
case XLOG_GIST_PAGE_UPDATE:
- gistRedoPageUpdateRecord(lsn, record);
+ gistRedoPageUpdateRecord(record);
break;
case XLOG_GIST_PAGE_SPLIT:
- gistRedoPageSplitRecord(lsn, record);
+ gistRedoPageSplitRecord(record);
break;
case XLOG_GIST_CREATE_INDEX:
- gistRedoCreateIndex(lsn, record);
+ gistRedoCreateIndex(record);
break;
default:
elog(PANIC, "gist_redo: unknown op code %u", info);
BlockNumber origrlink, GistNSN orignsn,
Buffer leftchildbuf, bool markfollowright)
{
- XLogRecData rdata[GIST_MAX_SPLIT_PAGES * 2 + 2];
gistxlogPageSplit xlrec;
SplitedPageLayout *ptr;
- int npage = 0,
- cur;
+ int npage = 0;
XLogRecPtr recptr;
+ int i;
for (ptr = dist; ptr; ptr = ptr->next)
npage++;
- /*
- * the caller should've checked this already, but doesn't hurt to check
- * again.
- */
- if (npage > GIST_MAX_SPLIT_PAGES)
- elog(ERROR, "GiST page split into too many halves");
-
- xlrec.node = node;
- xlrec.origblkno = blkno;
xlrec.origrlink = origrlink;
xlrec.orignsn = orignsn;
xlrec.origleaf = page_is_leaf;
xlrec.npage = (uint16) npage;
- xlrec.leftchild =
- BufferIsValid(leftchildbuf) ? BufferGetBlockNumber(leftchildbuf) : InvalidBlockNumber;
xlrec.markfollowright = markfollowright;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = sizeof(gistxlogPageSplit);
- rdata[0].buffer = InvalidBuffer;
-
- cur = 1;
+ XLogBeginInsert();
/*
* Include a full page image of the child buf. (only necessary if a
* checkpoint happened since the child page was split)
*/
if (BufferIsValid(leftchildbuf))
- {
- rdata[cur - 1].next = &(rdata[cur]);
- rdata[cur].data = NULL;
- rdata[cur].len = 0;
- rdata[cur].buffer = leftchildbuf;
- rdata[cur].buffer_std = true;
- cur++;
- }
+ XLogRegisterBuffer(0, leftchildbuf, REGBUF_STANDARD);
+ /*
+ * NOTE: We register a lot of data. The caller must've called
+ * XLogEnsureRecordSpace() to prepare for that. We cannot do it here,
+ * because we're already in a critical section. If you change the number
+ * of buffer or data registrations here, make sure you modify the
+ * XLogEnsureRecordSpace() calls accordingly!
+ */
+ XLogRegisterData((char *) &xlrec, sizeof(gistxlogPageSplit));
+
+ i = 1;
for (ptr = dist; ptr; ptr = ptr->next)
{
- rdata[cur - 1].next = &(rdata[cur]);
- rdata[cur].buffer = InvalidBuffer;
- rdata[cur].data = (char *) &(ptr->block);
- rdata[cur].len = sizeof(gistxlogPage);
- cur++;
-
- rdata[cur - 1].next = &(rdata[cur]);
- rdata[cur].buffer = InvalidBuffer;
- rdata[cur].data = (char *) (ptr->list);
- rdata[cur].len = ptr->lenlist;
- cur++;
+ XLogRegisterBuffer(i, ptr->buffer, REGBUF_WILL_INIT);
+ XLogRegisterBufData(i, (char *) &(ptr->block.num), sizeof(int));
+ XLogRegisterBufData(i, (char *) ptr->list, ptr->lenlist);
+ i++;
}
- rdata[cur - 1].next = NULL;
- recptr = XLogInsert(RM_GIST_ID, XLOG_GIST_PAGE_SPLIT, rdata);
+ recptr = XLogInsert(RM_GIST_ID, XLOG_GIST_PAGE_SPLIT);
return recptr;
}
*
* 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. Also, we take care that there's
- * at least one rdata item referencing the buffer, even when ntodelete and
- * ituplen are both zero; this ensures that XLogInsert knows about the buffer.
+ * to log the whole buffer contents instead.
*/
XLogRecPtr
gistXLogUpdate(RelFileNode node, Buffer buffer,
IndexTuple *itup, int ituplen,
Buffer leftchildbuf)
{
- XLogRecData rdata[MaxIndexTuplesPerPage + 3];
gistxlogPageUpdate xlrec;
- int cur,
- i;
+ int i;
XLogRecPtr recptr;
- xlrec.node = node;
- xlrec.blkno = BufferGetBlockNumber(buffer);
xlrec.ntodelete = ntodelete;
- xlrec.leftchild =
- BufferIsValid(leftchildbuf) ? BufferGetBlockNumber(leftchildbuf) : InvalidBlockNumber;
-
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = sizeof(gistxlogPageUpdate);
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
+ xlrec.ntoinsert = ituplen;
- rdata[1].data = (char *) todelete;
- rdata[1].len = sizeof(OffsetNumber) * ntodelete;
- rdata[1].buffer = buffer;
- rdata[1].buffer_std = true;
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, sizeof(gistxlogPageUpdate));
- cur = 2;
+ XLogRegisterBuffer(0, buffer, REGBUF_STANDARD);
+ XLogRegisterBufData(0, (char *) todelete, sizeof(OffsetNumber) * ntodelete);
/* new tuples */
for (i = 0; i < ituplen; i++)
- {
- rdata[cur - 1].next = &(rdata[cur]);
- rdata[cur].data = (char *) (itup[i]);
- rdata[cur].len = IndexTupleSize(itup[i]);
- rdata[cur].buffer = buffer;
- rdata[cur].buffer_std = true;
- cur++;
- }
+ 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))
- {
- rdata[cur - 1].next = &(rdata[cur]);
- rdata[cur].data = NULL;
- rdata[cur].len = 0;
- rdata[cur].buffer = leftchildbuf;
- rdata[cur].buffer_std = true;
- cur++;
- }
- rdata[cur - 1].next = NULL;
+ XLogRegisterBuffer(1, leftchildbuf, REGBUF_STANDARD);
- recptr = XLogInsert(RM_GIST_ID, XLOG_GIST_PAGE_UPDATE, rdata);
+ recptr = XLogInsert(RM_GIST_ID, XLOG_GIST_PAGE_UPDATE);
return recptr;
}
void
-hash_redo(XLogRecPtr lsn, XLogRecord *record)
+hash_redo(XLogReaderState *record)
{
elog(PANIC, "hash_redo: unimplemented");
}
xl_heap_insert xlrec;
xl_heap_header xlhdr;
XLogRecPtr recptr;
- XLogRecData rdata[4];
Page page = BufferGetPage(buffer);
uint8 info = XLOG_HEAP_INSERT;
- bool need_tuple_data;
+ int bufflags = 0;
/*
- * For logical decoding, we need the tuple even if we're doing a full
- * page write, so make sure to log it separately. (XXX We could
- * alternatively store a pointer into the FPW).
- *
- * Also, if this is a catalog, we need to transmit combocids to
- * properly decode, so log that as well.
+ * If this is a catalog, we need to transmit combocids to properly
+ * decode, so log that as well.
*/
- need_tuple_data = RelationIsLogicallyLogged(relation);
if (RelationIsAccessibleInLogicalDecoding(relation))
log_heap_new_cid(relation, heaptup);
- xlrec.flags = all_visible_cleared ? XLOG_HEAP_ALL_VISIBLE_CLEARED : 0;
- xlrec.target.node = relation->rd_node;
- xlrec.target.tid = heaptup->t_self;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfHeapInsert;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
-
- xlhdr.t_infomask2 = heaptup->t_data->t_infomask2;
- xlhdr.t_infomask = heaptup->t_data->t_infomask;
- xlhdr.t_hoff = heaptup->t_data->t_hoff;
-
/*
- * note we mark rdata[1] as belonging to buffer; if XLogInsert decides
- * to write the whole page to the xlog, we don't need to store
- * xl_heap_header in the xlog.
+ * If this is the single and first tuple on page, we can reinit the
+ * page instead of restoring the whole thing. Set flag, and hide
+ * buffer references from XLogInsert.
*/
- rdata[1].data = (char *) &xlhdr;
- rdata[1].len = SizeOfHeapHeader;
- rdata[1].buffer = need_tuple_data ? InvalidBuffer : buffer;
- rdata[1].buffer_std = true;
- rdata[1].next = &(rdata[2]);
+ if (ItemPointerGetOffsetNumber(&(heaptup->t_self)) == FirstOffsetNumber &&
+ PageGetMaxOffsetNumber(page) == FirstOffsetNumber)
+ {
+ info |= XLOG_HEAP_INIT_PAGE;
+ bufflags |= REGBUF_WILL_INIT;
+ }
- /* PG73FORMAT: write bitmap [+ padding] [+ oid] + data */
- rdata[2].data = (char *) heaptup->t_data + offsetof(HeapTupleHeaderData, t_bits);
- rdata[2].len = heaptup->t_len - offsetof(HeapTupleHeaderData, t_bits);
- rdata[2].buffer = need_tuple_data ? InvalidBuffer : buffer;
- rdata[2].buffer_std = true;
- rdata[2].next = NULL;
+ xlrec.offnum = ItemPointerGetOffsetNumber(&heaptup->t_self);
+ xlrec.flags = all_visible_cleared ? XLOG_HEAP_ALL_VISIBLE_CLEARED : 0;
+ Assert(ItemPointerGetBlockNumber(&heaptup->t_self) == BufferGetBlockNumber(buffer));
/*
- * Make a separate rdata entry for the tuple's buffer if we're doing
- * logical decoding, so that an eventual FPW doesn't remove the
- * tuple's data.
+ * For logical decoding, we need the tuple even if we're doing a full
+ * page write, so make sure it's included even if we take a full-page
+ * image. (XXX We could alternatively store a pointer into the FPW).
*/
- if (need_tuple_data)
+ if (RelationIsLogicallyLogged(relation))
{
- rdata[2].next = &(rdata[3]);
-
- rdata[3].data = NULL;
- rdata[3].len = 0;
- rdata[3].buffer = buffer;
- rdata[3].buffer_std = true;
- rdata[3].next = NULL;
-
xlrec.flags |= XLOG_HEAP_CONTAINS_NEW_TUPLE;
+ bufflags |= REGBUF_KEEP_DATA;
}
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfHeapInsert);
+
+ xlhdr.t_infomask2 = heaptup->t_data->t_infomask2;
+ xlhdr.t_infomask = heaptup->t_data->t_infomask;
+ xlhdr.t_hoff = heaptup->t_data->t_hoff;
+
/*
- * If this is the single and first tuple on page, we can reinit the
- * page instead of restoring the whole thing. Set flag, and hide
- * buffer references from XLogInsert.
+ * note we mark xlhdr as belonging to buffer; if XLogInsert decides to
+ * write the whole page to the xlog, we don't need to store
+ * xl_heap_header in the xlog.
*/
- if (ItemPointerGetOffsetNumber(&(heaptup->t_self)) == FirstOffsetNumber &&
- PageGetMaxOffsetNumber(page) == FirstOffsetNumber)
- {
- info |= XLOG_HEAP_INIT_PAGE;
- rdata[1].buffer = rdata[2].buffer = rdata[3].buffer = InvalidBuffer;
- }
+ XLogRegisterBuffer(0, buffer, REGBUF_STANDARD | bufflags);
+ XLogRegisterBufData(0, (char *) &xlhdr, SizeOfHeapHeader);
+ /* PG73FORMAT: write bitmap [+ padding] [+ oid] + data */
+ XLogRegisterBufData(0,
+ (char *) heaptup->t_data + offsetof(HeapTupleHeaderData, t_bits),
+ heaptup->t_len - offsetof(HeapTupleHeaderData, t_bits));
- recptr = XLogInsert(RM_HEAP_ID, info, rdata);
+ recptr = XLogInsert(RM_HEAP_ID, info);
PageSetLSN(page, recptr);
}
break;
RelationPutHeapTuple(relation, buffer, heaptup);
+
+ /*
+ * We don't use heap_multi_insert for catalog tuples yet, but
+ * better be prepared...
+ */
+ if (needwal && need_cids)
+ log_heap_new_cid(relation, heaptup);
}
if (PageIsAllVisible(page))
{
XLogRecPtr recptr;
xl_heap_multi_insert *xlrec;
- XLogRecData rdata[3];
uint8 info = XLOG_HEAP2_MULTI_INSERT;
char *tupledata;
int totaldatalen;
char *scratchptr = scratch;
bool init;
+ int bufflags = 0;
/*
* If the page was previously empty, we can reinit the page
tupledata = scratchptr;
xlrec->flags = all_visible_cleared ? XLOG_HEAP_ALL_VISIBLE_CLEARED : 0;
- xlrec->node = relation->rd_node;
- xlrec->blkno = BufferGetBlockNumber(buffer);
xlrec->ntuples = nthispage;
/*
datalen);
tuphdr->datalen = datalen;
scratchptr += datalen;
-
- /*
- * We don't use heap_multi_insert for catalog tuples yet, but
- * better be prepared...
- */
- if (need_cids)
- log_heap_new_cid(relation, heaptup);
}
totaldatalen = scratchptr - tupledata;
Assert((scratchptr - scratch) < BLCKSZ);
- rdata[0].data = (char *) xlrec;
- rdata[0].len = tupledata - scratch;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &rdata[1];
-
- rdata[1].data = tupledata;
- rdata[1].len = totaldatalen;
- rdata[1].buffer = need_tuple_data ? InvalidBuffer : buffer;
- rdata[1].buffer_std = true;
- rdata[1].next = NULL;
-
- /*
- * Make a separate rdata entry for the tuple's buffer if we're
- * doing logical decoding, so that an eventual FPW doesn't remove
- * the tuple's data.
- */
if (need_tuple_data)
- {
- rdata[1].next = &(rdata[2]);
-
- rdata[2].data = NULL;
- rdata[2].len = 0;
- rdata[2].buffer = buffer;
- rdata[2].buffer_std = true;
- rdata[2].next = NULL;
xlrec->flags |= XLOG_HEAP_CONTAINS_NEW_TUPLE;
- }
/*
- * If we're going to reinitialize the whole page using the WAL
- * record, hide buffer reference from XLogInsert.
+ * Signal that this is the last xl_heap_multi_insert record
+ * emitted by this call to heap_multi_insert(). Needed for logical
+ * decoding so it knows when to cleanup temporary data.
*/
+ if (ndone + nthispage == ntuples)
+ xlrec->flags |= XLOG_HEAP_LAST_MULTI_INSERT;
+
if (init)
{
- rdata[1].buffer = rdata[2].buffer = InvalidBuffer;
info |= XLOG_HEAP_INIT_PAGE;
+ bufflags |= REGBUF_WILL_INIT;
}
/*
- * Signal that this is the last xl_heap_multi_insert record
- * emitted by this call to heap_multi_insert(). Needed for logical
- * decoding so it knows when to cleanup temporary data.
+ * If we're doing logical decoding, include the new tuple data
+ * even if we take a full-page image of the page.
*/
- if (ndone + nthispage == ntuples)
- xlrec->flags |= XLOG_HEAP_LAST_MULTI_INSERT;
+ if (need_tuple_data)
+ bufflags |= REGBUF_KEEP_DATA;
+
+ XLogBeginInsert();
+ XLogRegisterData((char *) xlrec, tupledata - scratch);
+ XLogRegisterBuffer(0, buffer, REGBUF_STANDARD | bufflags);
- recptr = XLogInsert(RM_HEAP2_ID, info, rdata);
+ XLogRegisterBufData(0, tupledata, totaldatalen);
+ recptr = XLogInsert(RM_HEAP2_ID, info);
PageSetLSN(page, recptr);
}
{
xl_heap_delete xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[4];
/* For logical decode we need combocids to properly decode the catalog */
if (RelationIsAccessibleInLogicalDecoding(relation))
xlrec.flags = all_visible_cleared ? XLOG_HEAP_ALL_VISIBLE_CLEARED : 0;
xlrec.infobits_set = compute_infobits(tp.t_data->t_infomask,
tp.t_data->t_infomask2);
- xlrec.target.node = relation->rd_node;
- xlrec.target.tid = tp.t_self;
+ xlrec.offnum = ItemPointerGetOffsetNumber(&tp.t_self);
xlrec.xmax = new_xmax;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfHeapDelete;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
- rdata[1].data = NULL;
- rdata[1].len = 0;
- rdata[1].buffer = buffer;
- rdata[1].buffer_std = true;
- rdata[1].next = NULL;
+ if (old_key_tuple != NULL)
+ {
+ if (relation->rd_rel->relreplident == REPLICA_IDENTITY_FULL)
+ xlrec.flags |= XLOG_HEAP_CONTAINS_OLD_TUPLE;
+ else
+ xlrec.flags |= XLOG_HEAP_CONTAINS_OLD_KEY;
+ }
+
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfHeapDelete);
+
+ XLogRegisterBuffer(0, buffer, REGBUF_STANDARD);
/*
* Log replica identity of the deleted tuple if there is one
xlhdr.t_infomask = old_key_tuple->t_data->t_infomask;
xlhdr.t_hoff = old_key_tuple->t_data->t_hoff;
- rdata[1].next = &(rdata[2]);
- rdata[2].data = (char *) &xlhdr;
- rdata[2].len = SizeOfHeapHeader;
- rdata[2].buffer = InvalidBuffer;
- rdata[2].next = NULL;
-
- rdata[2].next = &(rdata[3]);
- rdata[3].data = (char *) old_key_tuple->t_data
- + offsetof(HeapTupleHeaderData, t_bits);
- rdata[3].len = old_key_tuple->t_len
- - offsetof(HeapTupleHeaderData, t_bits);
- rdata[3].buffer = InvalidBuffer;
- rdata[3].next = NULL;
-
- if (relation->rd_rel->relreplident == REPLICA_IDENTITY_FULL)
- xlrec.flags |= XLOG_HEAP_CONTAINS_OLD_TUPLE;
- else
- xlrec.flags |= XLOG_HEAP_CONTAINS_OLD_KEY;
+ XLogRegisterData((char *) &xlhdr, SizeOfHeapHeader);
+ XLogRegisterData((char *) old_key_tuple->t_data
+ + offsetof(HeapTupleHeaderData, t_bits),
+ old_key_tuple->t_len
+ - offsetof(HeapTupleHeaderData, t_bits));
}
- recptr = XLogInsert(RM_HEAP_ID, XLOG_HEAP_DELETE, rdata);
+ recptr = XLogInsert(RM_HEAP_ID, XLOG_HEAP_DELETE);
PageSetLSN(page, recptr);
}
{
xl_heap_lock xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[2];
- xlrec.target.node = relation->rd_node;
- xlrec.target.tid = tuple->t_self;
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, *buffer, REGBUF_STANDARD);
+
+ xlrec.offnum = ItemPointerGetOffsetNumber(&tuple->t_self);
xlrec.locking_xid = xid;
xlrec.infobits_set = compute_infobits(new_infomask,
tuple->t_data->t_infomask2);
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfHeapLock;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
-
- rdata[1].data = NULL;
- rdata[1].len = 0;
- rdata[1].buffer = *buffer;
- rdata[1].buffer_std = true;
- rdata[1].next = NULL;
+ XLogRegisterData((char *) &xlrec, SizeOfHeapLock);
- recptr = XLogInsert(RM_HEAP_ID, XLOG_HEAP_LOCK, rdata);
+ recptr = XLogInsert(RM_HEAP_ID, XLOG_HEAP_LOCK);
PageSetLSN(page, recptr);
}
{
xl_heap_lock_updated xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[2];
Page page = BufferGetPage(buf);
- xlrec.target.node = rel->rd_node;
- xlrec.target.tid = mytup.t_self;
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, buf, REGBUF_STANDARD);
+
+ xlrec.offnum = ItemPointerGetOffsetNumber(&mytup.t_self);
xlrec.xmax = new_xmax;
xlrec.infobits_set = compute_infobits(new_infomask, new_infomask2);
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfHeapLockUpdated;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
+ XLogRegisterData((char *) &xlrec, SizeOfHeapLockUpdated);
- rdata[1].data = NULL;
- rdata[1].len = 0;
- rdata[1].buffer = buf;
- rdata[1].buffer_std = true;
- rdata[1].next = NULL;
-
- recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_LOCK_UPDATED, rdata);
+ recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_LOCK_UPDATED);
PageSetLSN(page, recptr);
}
{
xl_heap_inplace xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[2];
- xlrec.target.node = relation->rd_node;
- xlrec.target.tid = tuple->t_self;
+ xlrec.offnum = ItemPointerGetOffsetNumber(&tuple->t_self);
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfHeapInplace;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfHeapInplace);
- rdata[1].data = (char *) htup + htup->t_hoff;
- rdata[1].len = newlen;
- rdata[1].buffer = buffer;
- rdata[1].buffer_std = true;
- rdata[1].next = NULL;
+ XLogRegisterBuffer(0, buffer, REGBUF_STANDARD);
+ XLogRegisterBufData(0, (char *) htup + htup->t_hoff, newlen);
- recptr = XLogInsert(RM_HEAP_ID, XLOG_HEAP_INPLACE, rdata);
+ recptr = XLogInsert(RM_HEAP_ID, XLOG_HEAP_INPLACE);
PageSetLSN(page, recptr);
}
{
xl_heap_cleanup_info xlrec;
XLogRecPtr recptr;
- XLogRecData rdata;
xlrec.node = rnode;
xlrec.latestRemovedXid = latestRemovedXid;
- rdata.data = (char *) &xlrec;
- rdata.len = SizeOfHeapCleanupInfo;
- rdata.buffer = InvalidBuffer;
- rdata.next = NULL;
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfHeapCleanupInfo);
- recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_CLEANUP_INFO, &rdata);
+ recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_CLEANUP_INFO);
return recptr;
}
TransactionId latestRemovedXid)
{
xl_heap_clean xlrec;
- uint8 info;
XLogRecPtr recptr;
- XLogRecData rdata[4];
/* Caller should not call me on a non-WAL-logged relation */
Assert(RelationNeedsWAL(reln));
- xlrec.node = reln->rd_node;
- xlrec.block = BufferGetBlockNumber(buffer);
xlrec.latestRemovedXid = latestRemovedXid;
xlrec.nredirected = nredirected;
xlrec.ndead = ndead;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfHeapClean;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfHeapClean);
+
+ XLogRegisterBuffer(0, buffer, REGBUF_STANDARD);
/*
* The OffsetNumber arrays are not actually in the buffer, but we pretend
* even if no item pointers changed state.
*/
if (nredirected > 0)
- {
- rdata[1].data = (char *) redirected;
- rdata[1].len = nredirected * sizeof(OffsetNumber) * 2;
- }
- else
- {
- rdata[1].data = NULL;
- rdata[1].len = 0;
- }
- rdata[1].buffer = buffer;
- rdata[1].buffer_std = true;
- rdata[1].next = &(rdata[2]);
+ XLogRegisterBufData(0, (char *) redirected,
+ nredirected * sizeof(OffsetNumber) * 2);
if (ndead > 0)
- {
- rdata[2].data = (char *) nowdead;
- rdata[2].len = ndead * sizeof(OffsetNumber);
- }
- else
- {
- rdata[2].data = NULL;
- rdata[2].len = 0;
- }
- rdata[2].buffer = buffer;
- rdata[2].buffer_std = true;
- rdata[2].next = &(rdata[3]);
+ XLogRegisterBufData(0, (char *) nowdead,
+ ndead * sizeof(OffsetNumber));
if (nunused > 0)
- {
- rdata[3].data = (char *) nowunused;
- rdata[3].len = nunused * sizeof(OffsetNumber);
- }
- else
- {
- rdata[3].data = NULL;
- rdata[3].len = 0;
- }
- rdata[3].buffer = buffer;
- rdata[3].buffer_std = true;
- rdata[3].next = NULL;
+ XLogRegisterBufData(0, (char *) nowunused,
+ nunused * sizeof(OffsetNumber));
- info = XLOG_HEAP2_CLEAN;
- recptr = XLogInsert(RM_HEAP2_ID, info, rdata);
+ recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_CLEAN);
return recptr;
}
{
xl_heap_freeze_page xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[2];
/* Caller should not call me on a non-WAL-logged relation */
Assert(RelationNeedsWAL(reln));
/* nor when there are no tuples to freeze */
Assert(ntuples > 0);
- xlrec.node = reln->rd_node;
- xlrec.block = BufferGetBlockNumber(buffer);
xlrec.cutoff_xid = cutoff_xid;
xlrec.ntuples = ntuples;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfHeapFreezePage;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfHeapFreezePage);
/*
* The freeze plan array is not actually in the buffer, but pretend that
* it is. When XLogInsert stores the whole buffer, the freeze plan need
* not be stored too.
*/
- rdata[1].data = (char *) tuples;
- rdata[1].len = ntuples * sizeof(xl_heap_freeze_tuple);
- rdata[1].buffer = buffer;
- rdata[1].buffer_std = true;
- rdata[1].next = NULL;
+ XLogRegisterBuffer(0, buffer, REGBUF_STANDARD);
+ XLogRegisterBufData(0, (char *) tuples,
+ ntuples * sizeof(xl_heap_freeze_tuple));
- recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_FREEZE_PAGE, rdata);
+ recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_FREEZE_PAGE);
return recptr;
}
* corresponding visibility map block. Both should have already been modified
* and dirtied.
*
- * If checksums are enabled, we also add the heap_buffer to the chain to
- * protect it from being torn.
+ * If checksums are enabled, we also generate a full-page image of
+ * heap_buffer, if necessary.
*/
XLogRecPtr
log_heap_visible(RelFileNode rnode, Buffer heap_buffer, Buffer vm_buffer,
{
xl_heap_visible xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[3];
+ uint8 flags;
Assert(BufferIsValid(heap_buffer));
Assert(BufferIsValid(vm_buffer));
- xlrec.node = rnode;
- xlrec.block = BufferGetBlockNumber(heap_buffer);
xlrec.cutoff_xid = cutoff_xid;
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfHeapVisible);
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfHeapVisible;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
+ XLogRegisterBuffer(0, vm_buffer, 0);
- rdata[1].data = NULL;
- rdata[1].len = 0;
- rdata[1].buffer = vm_buffer;
- rdata[1].buffer_std = false;
- rdata[1].next = NULL;
+ flags = REGBUF_STANDARD;
+ if (!XLogHintBitIsNeeded())
+ flags |= REGBUF_NO_IMAGE;
+ XLogRegisterBuffer(1, heap_buffer, flags);
- if (XLogHintBitIsNeeded())
- {
- rdata[1].next = &(rdata[2]);
-
- rdata[2].data = NULL;
- rdata[2].len = 0;
- rdata[2].buffer = heap_buffer;
- rdata[2].buffer_std = true;
- rdata[2].next = NULL;
- }
-
- recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_VISIBLE, rdata);
+ recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_VISIBLE);
return recptr;
}
bool all_visible_cleared, bool new_all_visible_cleared)
{
xl_heap_update xlrec;
- xl_heap_header_len xlhdr;
- xl_heap_header_len xlhdr_idx;
+ xl_heap_header xlhdr;
+ xl_heap_header xlhdr_idx;
uint8 info;
uint16 prefix_suffix[2];
uint16 prefixlen = 0,
suffixlen = 0;
XLogRecPtr recptr;
- XLogRecData rdata[9];
Page page = BufferGetPage(newbuf);
bool need_tuple_data = RelationIsLogicallyLogged(reln);
- int nr;
- Buffer newbufref;
+ bool init;
+ int bufflags;
/* Caller should not call me on a non-WAL-logged relation */
Assert(RelationNeedsWAL(reln));
+ XLogBeginInsert();
+
if (HeapTupleIsHeapOnly(newtup))
info = XLOG_HEAP_HOT_UPDATE;
else
suffixlen = 0;
}
- xlrec.target.node = reln->rd_node;
- xlrec.target.tid = oldtup->t_self;
- xlrec.old_xmax = HeapTupleHeaderGetRawXmax(oldtup->t_data);
- xlrec.old_infobits_set = compute_infobits(oldtup->t_data->t_infomask,
- oldtup->t_data->t_infomask2);
- xlrec.new_xmax = HeapTupleHeaderGetRawXmax(newtup->t_data);
+ /* Prepare main WAL data chain */
xlrec.flags = 0;
if (all_visible_cleared)
xlrec.flags |= XLOG_HEAP_ALL_VISIBLE_CLEARED;
- xlrec.newtid = newtup->t_self;
if (new_all_visible_cleared)
xlrec.flags |= XLOG_HEAP_NEW_ALL_VISIBLE_CLEARED;
if (prefixlen > 0)
xlrec.flags |= XLOG_HEAP_PREFIX_FROM_OLD;
if (suffixlen > 0)
xlrec.flags |= XLOG_HEAP_SUFFIX_FROM_OLD;
+ if (need_tuple_data)
+ {
+ xlrec.flags |= XLOG_HEAP_CONTAINS_NEW_TUPLE;
+ if (old_key_tuple)
+ {
+ if (reln->rd_rel->relreplident == REPLICA_IDENTITY_FULL)
+ xlrec.flags |= XLOG_HEAP_CONTAINS_OLD_TUPLE;
+ else
+ xlrec.flags |= XLOG_HEAP_CONTAINS_OLD_KEY;
+ }
+ }
/* If new tuple is the single and first tuple on page... */
if (ItemPointerGetOffsetNumber(&(newtup->t_self)) == FirstOffsetNumber &&
PageGetMaxOffsetNumber(page) == FirstOffsetNumber)
{
info |= XLOG_HEAP_INIT_PAGE;
- newbufref = InvalidBuffer;
+ init = true;
}
else
- newbufref = newbuf;
+ init = false;
- rdata[0].data = NULL;
- rdata[0].len = 0;
- rdata[0].buffer = oldbuf;
- rdata[0].buffer_std = true;
- rdata[0].next = &(rdata[1]);
+ /* Prepare WAL data for the old page */
+ xlrec.old_offnum = ItemPointerGetOffsetNumber(&oldtup->t_self);
+ xlrec.old_xmax = HeapTupleHeaderGetRawXmax(oldtup->t_data);
+ xlrec.old_infobits_set = compute_infobits(oldtup->t_data->t_infomask,
+ oldtup->t_data->t_infomask2);
+
+ /* Prepare WAL data for the new page */
+ xlrec.new_offnum = ItemPointerGetOffsetNumber(&newtup->t_self);
+ xlrec.new_xmax = HeapTupleHeaderGetRawXmax(newtup->t_data);
+
+ bufflags = REGBUF_STANDARD;
+ if (init)
+ bufflags |= REGBUF_WILL_INIT;
+ if (need_tuple_data)
+ bufflags |= REGBUF_KEEP_DATA;
- rdata[1].data = (char *) &xlrec;
- rdata[1].len = SizeOfHeapUpdate;
- rdata[1].buffer = InvalidBuffer;
- rdata[1].next = &(rdata[2]);
+ XLogRegisterBuffer(0, newbuf, bufflags);
+ if (oldbuf != newbuf)
+ XLogRegisterBuffer(1, oldbuf, REGBUF_STANDARD);
- /* prefix and/or suffix length fields */
+ XLogRegisterData((char *) &xlrec, SizeOfHeapUpdate);
+
+ /*
+ * Prepare WAL data for the new tuple.
+ */
if (prefixlen > 0 || suffixlen > 0)
{
if (prefixlen > 0 && suffixlen > 0)
{
prefix_suffix[0] = prefixlen;
prefix_suffix[1] = suffixlen;
- rdata[2].data = (char *) &prefix_suffix;
- rdata[2].len = 2 * sizeof(uint16);
+ XLogRegisterBufData(0, (char *) &prefix_suffix, sizeof(uint16) * 2);
}
else if (prefixlen > 0)
{
- rdata[2].data = (char *) &prefixlen;
- rdata[2].len = sizeof(uint16);
+ XLogRegisterBufData(0, (char *) &prefixlen, sizeof(uint16));
}
else
{
- rdata[2].data = (char *) &suffixlen;
- rdata[2].len = sizeof(uint16);
+ XLogRegisterBufData(0, (char *) &suffixlen, sizeof(uint16));
}
- rdata[2].buffer = newbufref;
- rdata[2].buffer_std = true;
- rdata[2].next = &(rdata[3]);
- nr = 3;
}
- else
- nr = 2;
-
- xlhdr.header.t_infomask2 = newtup->t_data->t_infomask2;
- xlhdr.header.t_infomask = newtup->t_data->t_infomask;
- xlhdr.header.t_hoff = newtup->t_data->t_hoff;
- Assert(offsetof(HeapTupleHeaderData, t_bits) +prefixlen + suffixlen <= newtup->t_len);
- xlhdr.t_len = newtup->t_len - offsetof(HeapTupleHeaderData, t_bits) -prefixlen - suffixlen;
- /*
- * As with insert records, we need not store this rdata segment if we
- * decide to store the whole buffer instead, unless we're doing logical
- * decoding.
- */
- rdata[nr].data = (char *) &xlhdr;
- rdata[nr].len = SizeOfHeapHeaderLen;
- rdata[nr].buffer = need_tuple_data ? InvalidBuffer : newbufref;
- rdata[nr].buffer_std = true;
- rdata[nr].next = &(rdata[nr + 1]);
- nr++;
+ xlhdr.t_infomask2 = newtup->t_data->t_infomask2;
+ xlhdr.t_infomask = newtup->t_data->t_infomask;
+ xlhdr.t_hoff = newtup->t_data->t_hoff;
+ Assert(offsetof(HeapTupleHeaderData, t_bits) + prefixlen + suffixlen <= newtup->t_len);
/*
* PG73FORMAT: write bitmap [+ padding] [+ oid] + data
*
* The 'data' doesn't include the common prefix or suffix.
*/
+ XLogRegisterBufData(0, (char *) &xlhdr, SizeOfHeapHeader);
if (prefixlen == 0)
{
- rdata[nr].data = ((char *) newtup->t_data) + offsetof(HeapTupleHeaderData, t_bits);
- rdata[nr].len = newtup->t_len - offsetof(HeapTupleHeaderData, t_bits) -suffixlen;
- rdata[nr].buffer = need_tuple_data ? InvalidBuffer : newbufref;
- rdata[nr].buffer_std = true;
- rdata[nr].next = NULL;
- nr++;
+ XLogRegisterBufData(0,
+ ((char *) newtup->t_data) + offsetof(HeapTupleHeaderData, t_bits),
+ newtup->t_len - offsetof(HeapTupleHeaderData, t_bits) -suffixlen);
}
else
{
/* bitmap [+ padding] [+ oid] */
if (newtup->t_data->t_hoff - offsetof(HeapTupleHeaderData, t_bits) >0)
{
- rdata[nr - 1].next = &(rdata[nr]);
- rdata[nr].data = ((char *) newtup->t_data) + offsetof(HeapTupleHeaderData, t_bits);
- rdata[nr].len = newtup->t_data->t_hoff - offsetof(HeapTupleHeaderData, t_bits);
- rdata[nr].buffer = need_tuple_data ? InvalidBuffer : newbufref;
- rdata[nr].buffer_std = true;
- rdata[nr].next = NULL;
- nr++;
+ XLogRegisterBufData(0,
+ ((char *) newtup->t_data) + offsetof(HeapTupleHeaderData, t_bits),
+ newtup->t_data->t_hoff - offsetof(HeapTupleHeaderData, t_bits));
}
/* data after common prefix */
- rdata[nr - 1].next = &(rdata[nr]);
- rdata[nr].data = ((char *) newtup->t_data) + newtup->t_data->t_hoff + prefixlen;
- rdata[nr].len = newtup->t_len - newtup->t_data->t_hoff - prefixlen - suffixlen;
- rdata[nr].buffer = need_tuple_data ? InvalidBuffer : newbufref;
- rdata[nr].buffer_std = true;
- rdata[nr].next = NULL;
- nr++;
+ XLogRegisterBufData(0,
+ ((char *) newtup->t_data) + newtup->t_data->t_hoff + prefixlen,
+ newtup->t_len - newtup->t_data->t_hoff - prefixlen - suffixlen);
}
- /*
- * Separate storage for the FPW buffer reference of the new page in the
- * wal_level >= logical case.
- */
- if (need_tuple_data)
+ /* We need to log a tuple identity */
+ if (need_tuple_data && old_key_tuple)
{
- rdata[nr - 1].next = &(rdata[nr]);
-
- rdata[nr].data = NULL,
- rdata[nr].len = 0;
- rdata[nr].buffer = newbufref;
- rdata[nr].buffer_std = true;
- rdata[nr].next = NULL;
- nr++;
-
- xlrec.flags |= XLOG_HEAP_CONTAINS_NEW_TUPLE;
+ /* don't really need this, but its more comfy to decode */
+ xlhdr_idx.t_infomask2 = old_key_tuple->t_data->t_infomask2;
+ xlhdr_idx.t_infomask = old_key_tuple->t_data->t_infomask;
+ xlhdr_idx.t_hoff = old_key_tuple->t_data->t_hoff;
- /* We need to log a tuple identity */
- if (old_key_tuple)
- {
- /* don't really need this, but its more comfy to decode */
- xlhdr_idx.header.t_infomask2 = old_key_tuple->t_data->t_infomask2;
- xlhdr_idx.header.t_infomask = old_key_tuple->t_data->t_infomask;
- xlhdr_idx.header.t_hoff = old_key_tuple->t_data->t_hoff;
- xlhdr_idx.t_len = old_key_tuple->t_len;
-
- rdata[nr - 1].next = &(rdata[nr]);
- rdata[nr].data = (char *) &xlhdr_idx;
- rdata[nr].len = SizeOfHeapHeaderLen;
- rdata[nr].buffer = InvalidBuffer;
- rdata[nr].next = &(rdata[nr + 1]);
- nr++;
-
- /* PG73FORMAT: write bitmap [+ padding] [+ oid] + data */
- rdata[nr].data = (char *) old_key_tuple->t_data
- + offsetof(HeapTupleHeaderData, t_bits);
- rdata[nr].len = old_key_tuple->t_len
- - offsetof(HeapTupleHeaderData, t_bits);
- rdata[nr].buffer = InvalidBuffer;
- rdata[nr].next = NULL;
- nr++;
+ XLogRegisterData((char *) &xlhdr_idx, SizeOfHeapHeader);
- if (reln->rd_rel->relreplident == REPLICA_IDENTITY_FULL)
- xlrec.flags |= XLOG_HEAP_CONTAINS_OLD_TUPLE;
- else
- xlrec.flags |= XLOG_HEAP_CONTAINS_OLD_KEY;
- }
+ /* PG73FORMAT: write bitmap [+ padding] [+ oid] + data */
+ XLogRegisterData((char *) old_key_tuple->t_data + offsetof(HeapTupleHeaderData, t_bits),
+ old_key_tuple->t_len - offsetof(HeapTupleHeaderData, t_bits));
}
- recptr = XLogInsert(RM_HEAP_ID, info, rdata);
+ recptr = XLogInsert(RM_HEAP_ID, info);
return recptr;
}
xl_heap_new_cid xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[1];
HeapTupleHeader hdr = tup->t_data;
Assert(ItemPointerIsValid(&tup->t_self));
Assert(tup->t_tableOid != InvalidOid);
xlrec.top_xid = GetTopTransactionId();
- xlrec.target.node = relation->rd_node;
- xlrec.target.tid = tup->t_self;
+ xlrec.target_node = relation->rd_node;
+ xlrec.target_tid = tup->t_self;
/*
* If the tuple got inserted & deleted in the same TX we definitely have a
xlrec.combocid = InvalidCommandId;
}
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfHeapNewCid;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = NULL;
+ /*
+ * Note that we don't need to register the buffer here, because this
+ * operation does not modify the page. The insert/update/delete that
+ * called us certainly did, but that's WAL-logged separately.
+ */
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfHeapNewCid);
- recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_NEW_CID, rdata);
+ recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_NEW_CID);
return recptr;
}
* Handles CLEANUP_INFO
*/
static void
-heap_xlog_cleanup_info(XLogRecPtr lsn, XLogRecord *record)
+heap_xlog_cleanup_info(XLogReaderState *record)
{
xl_heap_cleanup_info *xlrec = (xl_heap_cleanup_info *) XLogRecGetData(record);
*/
/* Backup blocks are not used in cleanup_info records */
- Assert(!(record->xl_info & XLR_BKP_BLOCK_MASK));
+ Assert(!XLogRecHasAnyBlockRefs(record));
}
/*
* Handles HEAP2_CLEAN record type
*/
static void
-heap_xlog_clean(XLogRecPtr lsn, XLogRecord *record)
+heap_xlog_clean(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_heap_clean *xlrec = (xl_heap_clean *) XLogRecGetData(record);
Buffer buffer;
Size freespace = 0;
BlockNumber blkno;
XLogRedoAction action;
- rnode = xlrec->node;
- blkno = xlrec->block;
+ XLogRecGetBlockTag(record, 0, &rnode, NULL, &blkno);
/*
* We're about to remove tuples. In Hot Standby mode, ensure that there's
* If we have a full-page image, restore it (using a cleanup lock) and
* we're done.
*/
- action = XLogReadBufferForRedoExtended(lsn, record, 0,
- rnode, MAIN_FORKNUM, blkno,
- RBM_NORMAL, true, &buffer);
+ action = XLogReadBufferForRedoExtended(record, 0, RBM_NORMAL, true,
+ &buffer);
if (action == BLK_NEEDS_REDO)
{
Page page = (Page) BufferGetPage(buffer);
int nredirected;
int ndead;
int nunused;
+ Size datalen;
+
+ redirected = (OffsetNumber *) XLogRecGetBlockData(record, 0, &datalen);
nredirected = xlrec->nredirected;
ndead = xlrec->ndead;
- end = (OffsetNumber *) ((char *) xlrec + record->xl_len);
- redirected = (OffsetNumber *) ((char *) xlrec + SizeOfHeapClean);
+ end = (OffsetNumber *) ((char *) redirected + datalen);
nowdead = redirected + (nredirected * 2);
nowunused = nowdead + ndead;
nunused = (end - nowunused);
* totally accurate anyway.
*/
if (action == BLK_NEEDS_REDO)
- XLogRecordPageWithFreeSpace(xlrec->node, xlrec->block, freespace);
+ XLogRecordPageWithFreeSpace(rnode, blkno, freespace);
}
/*
* page modification would fail to clear the visibility map bit.
*/
static void
-heap_xlog_visible(XLogRecPtr lsn, XLogRecord *record)
+heap_xlog_visible(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_heap_visible *xlrec = (xl_heap_visible *) XLogRecGetData(record);
+ Buffer vmbuffer = InvalidBuffer;
Buffer buffer;
Page page;
RelFileNode rnode;
BlockNumber blkno;
XLogRedoAction action;
- rnode = xlrec->node;
- blkno = xlrec->block;
+ XLogRecGetBlockTag(record, 1, &rnode, NULL, &blkno);
/*
* If there are any Hot Standby transactions running that have an xmin
* truncated later in recovery, we don't need to update the page, but we'd
* better still update the visibility map.
*/
- action = XLogReadBufferForRedo(lsn, record, 1, rnode, blkno, &buffer);
+ action = XLogReadBufferForRedo(record, 1, &buffer);
if (action == BLK_NEEDS_REDO)
{
/*
* the visibility map bit does so before checking the page LSN, so any
* bits that need to be cleared will still be cleared.
*/
- if (record->xl_info & XLR_BKP_BLOCK(0))
- (void) RestoreBackupBlock(lsn, record, 0, false, false);
- else
+ if (XLogReadBufferForRedoExtended(record, 0, RBM_ZERO_ON_ERROR, false,
+ &vmbuffer) == BLK_NEEDS_REDO)
{
+ Page vmpage = BufferGetPage(vmbuffer);
Relation reln;
- Buffer vmbuffer = InvalidBuffer;
+
+ /* initialize the page if it was read as zeros */
+ if (PageIsNew(vmpage))
+ PageInit(vmpage, BLCKSZ, 0);
+
+ /*
+ * XLogReplayBufferExtended locked the buffer. But visibilitymap_set
+ * will handle locking itself.
+ */
+ LockBuffer(vmbuffer, BUFFER_LOCK_UNLOCK);
reln = CreateFakeRelcacheEntry(rnode);
visibilitymap_pin(reln, blkno, &vmbuffer);
* we did for the heap page. If this results in a dropped bit, no
* real harm is done; and the next VACUUM will fix it.
*/
- if (lsn > PageGetLSN(BufferGetPage(vmbuffer)))
+ if (lsn > PageGetLSN(vmpage))
visibilitymap_set(reln, blkno, InvalidBuffer, lsn, vmbuffer,
xlrec->cutoff_xid);
ReleaseBuffer(vmbuffer);
FreeFakeRelcacheEntry(reln);
}
+ else if (BufferIsValid(vmbuffer))
+ UnlockReleaseBuffer(vmbuffer);
}
/*
* Replay XLOG_HEAP2_FREEZE_PAGE records
*/
static void
-heap_xlog_freeze_page(XLogRecPtr lsn, XLogRecord *record)
+heap_xlog_freeze_page(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_heap_freeze_page *xlrec = (xl_heap_freeze_page *) XLogRecGetData(record);
TransactionId cutoff_xid = xlrec->cutoff_xid;
Buffer buffer;
- Page page;
int ntup;
/*
* consider the frozen xids as running.
*/
if (InHotStandby)
- ResolveRecoveryConflictWithSnapshot(cutoff_xid, xlrec->node);
+ {
+ RelFileNode rnode;
- if (XLogReadBufferForRedo(lsn, record, 0, xlrec->node, xlrec->block,
- &buffer) == BLK_NEEDS_REDO)
+ XLogRecGetBlockTag(record, 0, &rnode, NULL, NULL);
+ ResolveRecoveryConflictWithSnapshot(cutoff_xid, rnode);
+ }
+
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
- page = BufferGetPage(buffer);
+ Page page = BufferGetPage(buffer);
+ xl_heap_freeze_tuple *tuples;
+
+ tuples = (xl_heap_freeze_tuple *) XLogRecGetBlockData(record, 0, NULL);
/* now execute freeze plan for each frozen tuple */
for (ntup = 0; ntup < xlrec->ntuples; ntup++)
ItemId lp;
HeapTupleHeader tuple;
- xlrec_tp = &xlrec->tuples[ntup];
+ xlrec_tp = &tuples[ntup];
lp = PageGetItemId(page, xlrec_tp->offset); /* offsets are one-based */
tuple = (HeapTupleHeader) PageGetItem(page, lp);
}
static void
-heap_xlog_delete(XLogRecPtr lsn, XLogRecord *record)
+heap_xlog_delete(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_heap_delete *xlrec = (xl_heap_delete *) XLogRecGetData(record);
Buffer buffer;
Page page;
- OffsetNumber offnum;
ItemId lp = NULL;
HeapTupleHeader htup;
BlockNumber blkno;
RelFileNode target_node;
+ ItemPointerData target_tid;
- blkno = ItemPointerGetBlockNumber(&(xlrec->target.tid));
- target_node = xlrec->target.node;
+ XLogRecGetBlockTag(record, 0, &target_node, NULL, &blkno);
+ ItemPointerSetBlockNumber(&target_tid, blkno);
+ ItemPointerSetOffsetNumber(&target_tid, xlrec->offnum);
/*
* The visibility map may need to be fixed even if the heap page is
FreeFakeRelcacheEntry(reln);
}
- if (XLogReadBufferForRedo(lsn, record, 0, target_node, blkno, &buffer)
- == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
- page = (Page) BufferGetPage(buffer);
+ page = BufferGetPage(buffer);
- offnum = ItemPointerGetOffsetNumber(&(xlrec->target.tid));
- if (PageGetMaxOffsetNumber(page) >= offnum)
- lp = PageGetItemId(page, offnum);
+ if (PageGetMaxOffsetNumber(page) >= xlrec->offnum)
+ lp = PageGetItemId(page, xlrec->offnum);
- if (PageGetMaxOffsetNumber(page) < offnum || !ItemIdIsNormal(lp))
+ if (PageGetMaxOffsetNumber(page) < xlrec->offnum || !ItemIdIsNormal(lp))
elog(PANIC, "heap_delete_redo: invalid lp");
htup = (HeapTupleHeader) PageGetItem(page, lp);
HeapTupleHeaderSetCmax(htup, FirstCommandId, false);
/* Mark the page as a candidate for pruning */
- PageSetPrunable(page, record->xl_xid);
+ PageSetPrunable(page, XLogRecGetXid(record));
if (xlrec->flags & XLOG_HEAP_ALL_VISIBLE_CLEARED)
PageClearAllVisible(page);
/* Make sure there is no forward chain link in t_ctid */
- htup->t_ctid = xlrec->target.tid;
+ htup->t_ctid = target_tid;
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
}
static void
-heap_xlog_insert(XLogRecPtr lsn, XLogRecord *record)
+heap_xlog_insert(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_heap_insert *xlrec = (xl_heap_insert *) XLogRecGetData(record);
Buffer buffer;
Page page;
- OffsetNumber offnum;
struct
{
HeapTupleHeaderData hdr;
Size freespace = 0;
RelFileNode target_node;
BlockNumber blkno;
+ ItemPointerData target_tid;
XLogRedoAction action;
- target_node = xlrec->target.node;
- blkno = ItemPointerGetBlockNumber(&(xlrec->target.tid));
+ XLogRecGetBlockTag(record, 0, &target_node, NULL, &blkno);
+ ItemPointerSetBlockNumber(&target_tid, blkno);
+ ItemPointerSetOffsetNumber(&target_tid, xlrec->offnum);
/*
* The visibility map may need to be fixed even if the heap page is
}
/*
- * If we inserted the first and only tuple on the page, re-initialize
- * the page from scratch.
+ * If we inserted the first and only tuple on the page, re-initialize the
+ * page from scratch.
*/
- if (record->xl_info & XLOG_HEAP_INIT_PAGE)
+ if (XLogRecGetInfo(record) & XLOG_HEAP_INIT_PAGE)
{
- XLogReadBufferForRedoExtended(lsn, record, 0,
- target_node, MAIN_FORKNUM, blkno,
- RBM_ZERO_AND_LOCK, false, &buffer);
+ buffer = XLogInitBufferForRedo(record, 0);
page = BufferGetPage(buffer);
PageInit(page, BufferGetPageSize(buffer), 0);
action = BLK_NEEDS_REDO;
}
else
- action = XLogReadBufferForRedo(lsn, record, 0, target_node, blkno,
- &buffer);
-
+ action = XLogReadBufferForRedo(record, 0, &buffer);
if (action == BLK_NEEDS_REDO)
{
+ Size datalen;
+ char *data;
+
page = BufferGetPage(buffer);
- offnum = ItemPointerGetOffsetNumber(&(xlrec->target.tid));
- if (PageGetMaxOffsetNumber(page) + 1 < offnum)
+ if (PageGetMaxOffsetNumber(page) + 1 < xlrec->offnum)
elog(PANIC, "heap_insert_redo: invalid max offset number");
- newlen = record->xl_len - SizeOfHeapInsert - SizeOfHeapHeader;
- Assert(newlen <= MaxHeapTupleSize);
- memcpy((char *) &xlhdr,
- (char *) xlrec + SizeOfHeapInsert,
- SizeOfHeapHeader);
+ data = XLogRecGetBlockData(record, 0, &datalen);
+
+ newlen = datalen - SizeOfHeapHeader;
+ Assert(datalen > SizeOfHeapHeader && newlen <= MaxHeapTupleSize);
+ memcpy((char *) &xlhdr, data, SizeOfHeapHeader);
+ data += SizeOfHeapHeader;
+
htup = &tbuf.hdr;
MemSet((char *) htup, 0, sizeof(HeapTupleHeaderData));
/* PG73FORMAT: get bitmap [+ padding] [+ oid] + data */
memcpy((char *) htup + offsetof(HeapTupleHeaderData, t_bits),
- (char *) xlrec + SizeOfHeapInsert + SizeOfHeapHeader,
+ data,
newlen);
newlen += offsetof(HeapTupleHeaderData, t_bits);
htup->t_infomask2 = xlhdr.t_infomask2;
htup->t_infomask = xlhdr.t_infomask;
htup->t_hoff = xlhdr.t_hoff;
- HeapTupleHeaderSetXmin(htup, record->xl_xid);
+ HeapTupleHeaderSetXmin(htup, XLogRecGetXid(record));
HeapTupleHeaderSetCmin(htup, FirstCommandId);
- htup->t_ctid = xlrec->target.tid;
+ htup->t_ctid = target_tid;
- offnum = PageAddItem(page, (Item) htup, newlen, offnum, true, true);
- if (offnum == InvalidOffsetNumber)
+ if (PageAddItem(page, (Item) htup, newlen, xlrec->offnum,
+ true, true) == InvalidOffsetNumber)
elog(PANIC, "heap_insert_redo: failed to add tuple");
freespace = PageGetHeapFreeSpace(page); /* needed to update FSM below */
* totally accurate anyway.
*/
if (action == BLK_NEEDS_REDO && freespace < BLCKSZ / 5)
- XLogRecordPageWithFreeSpace(xlrec->target.node, blkno, freespace);
+ XLogRecordPageWithFreeSpace(target_node, blkno, freespace);
}
/*
* Handles MULTI_INSERT record type.
*/
static void
-heap_xlog_multi_insert(XLogRecPtr lsn, XLogRecord *record)
+heap_xlog_multi_insert(XLogReaderState *record)
{
- char *recdata = XLogRecGetData(record);
+ XLogRecPtr lsn = record->EndRecPtr;
xl_heap_multi_insert *xlrec;
RelFileNode rnode;
BlockNumber blkno;
uint32 newlen;
Size freespace = 0;
int i;
- bool isinit = (record->xl_info & XLOG_HEAP_INIT_PAGE) != 0;
+ bool isinit = (XLogRecGetInfo(record) & XLOG_HEAP_INIT_PAGE) != 0;
XLogRedoAction action;
/*
* Insertion doesn't overwrite MVCC data, so no conflict processing is
* required.
*/
+ xlrec = (xl_heap_multi_insert *) XLogRecGetData(record);
- xlrec = (xl_heap_multi_insert *) recdata;
- recdata += SizeOfHeapMultiInsert;
-
- rnode = xlrec->node;
- blkno = xlrec->blkno;
-
- /*
- * If we're reinitializing the page, the tuples are stored in order from
- * FirstOffsetNumber. Otherwise there's an array of offsets in the WAL
- * record.
- */
- if (!isinit)
- recdata += sizeof(OffsetNumber) * xlrec->ntuples;
+ XLogRecGetBlockTag(record, 0, &rnode, NULL, &blkno);
/*
* The visibility map may need to be fixed even if the heap page is
if (isinit)
{
- XLogReadBufferForRedoExtended(lsn, record, 0,
- rnode, MAIN_FORKNUM, blkno,
- RBM_ZERO_AND_LOCK, false, &buffer);
+ buffer = XLogInitBufferForRedo(record, 0);
page = BufferGetPage(buffer);
PageInit(page, BufferGetPageSize(buffer), 0);
action = BLK_NEEDS_REDO;
}
else
- action = XLogReadBufferForRedo(lsn, record, 0, rnode, blkno, &buffer);
-
+ action = XLogReadBufferForRedo(record, 0, &buffer);
if (action == BLK_NEEDS_REDO)
{
- page = BufferGetPage(buffer);
+ char *tupdata;
+ char *endptr;
+ Size len;
+
+ /* Tuples are stored as block data */
+ tupdata = XLogRecGetBlockData(record, 0, &len);
+ endptr = tupdata + len;
+
+ page = (Page) BufferGetPage(buffer);
+
for (i = 0; i < xlrec->ntuples; i++)
{
OffsetNumber offnum;
xl_multi_insert_tuple *xlhdr;
+ /*
+ * If we're reinitializing the page, the tuples are stored in
+ * order from FirstOffsetNumber. Otherwise there's an array of
+ * offsets in the WAL record, and the tuples come after that.
+ */
if (isinit)
offnum = FirstOffsetNumber + i;
else
if (PageGetMaxOffsetNumber(page) + 1 < offnum)
elog(PANIC, "heap_multi_insert_redo: invalid max offset number");
- xlhdr = (xl_multi_insert_tuple *) SHORTALIGN(recdata);
- recdata = ((char *) xlhdr) + SizeOfMultiInsertTuple;
+ xlhdr = (xl_multi_insert_tuple *) SHORTALIGN(tupdata);
+ tupdata = ((char *) xlhdr) + SizeOfMultiInsertTuple;
newlen = xlhdr->datalen;
Assert(newlen <= MaxHeapTupleSize);
MemSet((char *) htup, 0, sizeof(HeapTupleHeaderData));
/* PG73FORMAT: get bitmap [+ padding] [+ oid] + data */
memcpy((char *) htup + offsetof(HeapTupleHeaderData, t_bits),
- (char *) recdata,
+ (char *) tupdata,
newlen);
- recdata += newlen;
+ tupdata += newlen;
newlen += offsetof(HeapTupleHeaderData, t_bits);
htup->t_infomask2 = xlhdr->t_infomask2;
htup->t_infomask = xlhdr->t_infomask;
htup->t_hoff = xlhdr->t_hoff;
- HeapTupleHeaderSetXmin(htup, record->xl_xid);
+ HeapTupleHeaderSetXmin(htup, XLogRecGetXid(record));
HeapTupleHeaderSetCmin(htup, FirstCommandId);
ItemPointerSetBlockNumber(&htup->t_ctid, blkno);
ItemPointerSetOffsetNumber(&htup->t_ctid, offnum);
if (offnum == InvalidOffsetNumber)
elog(PANIC, "heap_multi_insert_redo: failed to add tuple");
}
+ if (tupdata != endptr)
+ elog(PANIC, "heap_multi_insert_redo: total tuple length mismatch");
freespace = PageGetHeapFreeSpace(page); /* needed to update FSM below */
* totally accurate anyway.
*/
if (action == BLK_NEEDS_REDO && freespace < BLCKSZ / 5)
- XLogRecordPageWithFreeSpace(xlrec->node, blkno, freespace);
+ XLogRecordPageWithFreeSpace(rnode, blkno, freespace);
}
/*
* Handles UPDATE and HOT_UPDATE
*/
static void
-heap_xlog_update(XLogRecPtr lsn, XLogRecord *record, bool hot_update)
+heap_xlog_update(XLogReaderState *record, bool hot_update)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_heap_update *xlrec = (xl_heap_update *) XLogRecGetData(record);
RelFileNode rnode;
BlockNumber oldblk;
BlockNumber newblk;
+ ItemPointerData newtid;
Buffer obuffer,
nbuffer;
Page page;
ItemId lp = NULL;
HeapTupleData oldtup;
HeapTupleHeader htup;
- char *recdata;
uint16 prefixlen = 0,
suffixlen = 0;
char *newp;
HeapTupleHeaderData hdr;
char data[MaxHeapTupleSize];
} tbuf;
- xl_heap_header_len xlhdr;
+ xl_heap_header xlhdr;
uint32 newlen;
Size freespace = 0;
XLogRedoAction oldaction;
oldtup.t_data = NULL;
oldtup.t_len = 0;
- rnode = xlrec->target.node;
- newblk = ItemPointerGetBlockNumber(&xlrec->newtid);
- oldblk = ItemPointerGetBlockNumber(&xlrec->target.tid);
+ XLogRecGetBlockTag(record, 0, &rnode, NULL, &newblk);
+ if (XLogRecGetBlockTag(record, 1, NULL, NULL, &oldblk))
+ {
+ /* HOT updates are never done across pages */
+ Assert(!hot_update);
+ }
+ else
+ oldblk = newblk;
+
+ ItemPointerSet(&newtid, newblk, xlrec->new_offnum);
/*
* The visibility map may need to be fixed even if the heap page is
*/
/* Deal with old tuple version */
- oldaction = XLogReadBufferForRedo(lsn, record, 0, rnode, oldblk, &obuffer);
+ oldaction = XLogReadBufferForRedo(record, (oldblk == newblk) ? 0 : 1,
+ &obuffer);
if (oldaction == BLK_NEEDS_REDO)
{
- page = (Page) BufferGetPage(obuffer);
-
- offnum = ItemPointerGetOffsetNumber(&(xlrec->target.tid));
+ page = BufferGetPage(obuffer);
+ offnum = xlrec->old_offnum;
if (PageGetMaxOffsetNumber(page) >= offnum)
lp = PageGetItemId(page, offnum);
HeapTupleHeaderSetXmax(htup, xlrec->old_xmax);
HeapTupleHeaderSetCmax(htup, FirstCommandId, false);
/* Set forward chain link in t_ctid */
- htup->t_ctid = xlrec->newtid;
+ htup->t_ctid = newtid;
/* Mark the page as a candidate for pruning */
- PageSetPrunable(page, record->xl_xid);
+ PageSetPrunable(page, XLogRecGetXid(record));
if (xlrec->flags & XLOG_HEAP_ALL_VISIBLE_CLEARED)
PageClearAllVisible(page);
nbuffer = obuffer;
newaction = oldaction;
}
- else if (record->xl_info & XLOG_HEAP_INIT_PAGE)
+ else if (XLogRecGetInfo(record) & XLOG_HEAP_INIT_PAGE)
{
- XLogReadBufferForRedoExtended(lsn, record, 1,
- rnode, MAIN_FORKNUM, newblk,
- RBM_ZERO_AND_LOCK, false, &nbuffer);
+ nbuffer = XLogInitBufferForRedo(record, 0);
page = (Page) BufferGetPage(nbuffer);
PageInit(page, BufferGetPageSize(nbuffer), 0);
newaction = BLK_NEEDS_REDO;
}
else
- newaction = XLogReadBufferForRedo(lsn, record, 1, rnode, newblk,
- &nbuffer);
+ newaction = XLogReadBufferForRedo(record, 0, &nbuffer);
/*
* The visibility map may need to be fixed even if the heap page is
*/
if (xlrec->flags & XLOG_HEAP_NEW_ALL_VISIBLE_CLEARED)
{
- Relation reln = CreateFakeRelcacheEntry(xlrec->target.node);
+ Relation reln = CreateFakeRelcacheEntry(rnode);
Buffer vmbuffer = InvalidBuffer;
visibilitymap_pin(reln, newblk, &vmbuffer);
/* Deal with new tuple */
if (newaction == BLK_NEEDS_REDO)
{
- page = (Page) BufferGetPage(nbuffer);
+ char *recdata;
+ char *recdata_end;
+ Size datalen;
+ Size tuplen;
+
+ recdata = XLogRecGetBlockData(record, 0, &datalen);
+ recdata_end = recdata + datalen;
- offnum = ItemPointerGetOffsetNumber(&(xlrec->newtid));
+ page = BufferGetPage(nbuffer);
+
+ offnum = xlrec->new_offnum;
if (PageGetMaxOffsetNumber(page) + 1 < offnum)
elog(PANIC, "heap_update_redo: invalid max offset number");
- recdata = (char *) xlrec + SizeOfHeapUpdate;
-
if (xlrec->flags & XLOG_HEAP_PREFIX_FROM_OLD)
{
Assert(newblk == oldblk);
recdata += sizeof(uint16);
}
- memcpy((char *) &xlhdr, recdata, SizeOfHeapHeaderLen);
- recdata += SizeOfHeapHeaderLen;
+ memcpy((char *) &xlhdr, recdata, SizeOfHeapHeader);
+ recdata += SizeOfHeapHeader;
+
+ tuplen = recdata_end - recdata;
+ Assert(tuplen <= MaxHeapTupleSize);
- Assert(xlhdr.t_len + prefixlen + suffixlen <= MaxHeapTupleSize);
htup = &tbuf.hdr;
MemSet((char *) htup, 0, sizeof(HeapTupleHeaderData));
int len;
/* copy bitmap [+ padding] [+ oid] from WAL record */
- len = xlhdr.header.t_hoff - offsetof(HeapTupleHeaderData, t_bits);
+ len = xlhdr.t_hoff - offsetof(HeapTupleHeaderData, t_bits);
memcpy(newp, recdata, len);
recdata += len;
newp += len;
newp += prefixlen;
/* copy new tuple data from WAL record */
- len = xlhdr.t_len - (xlhdr.header.t_hoff - offsetof(HeapTupleHeaderData, t_bits));
+ len = tuplen - (xlhdr.t_hoff - offsetof(HeapTupleHeaderData, t_bits));
memcpy(newp, recdata, len);
recdata += len;
newp += len;
* copy bitmap [+ padding] [+ oid] + data from record, all in one
* go
*/
- memcpy(newp, recdata, xlhdr.t_len);
- recdata += xlhdr.t_len;
- newp += xlhdr.t_len;
+ memcpy(newp, recdata, tuplen);
+ recdata += tuplen;
+ newp += tuplen;
}
+ Assert(recdata == recdata_end);
+
/* copy suffix from old tuple */
if (suffixlen > 0)
memcpy(newp, (char *) oldtup.t_data + oldtup.t_len - suffixlen, suffixlen);
- newlen = offsetof(HeapTupleHeaderData, t_bits) + xlhdr.t_len + prefixlen + suffixlen;
- htup->t_infomask2 = xlhdr.header.t_infomask2;
- htup->t_infomask = xlhdr.header.t_infomask;
- htup->t_hoff = xlhdr.header.t_hoff;
+ newlen = offsetof(HeapTupleHeaderData, t_bits) + tuplen + prefixlen + suffixlen;
+ htup->t_infomask2 = xlhdr.t_infomask2;
+ htup->t_infomask = xlhdr.t_infomask;
+ htup->t_hoff = xlhdr.t_hoff;
- HeapTupleHeaderSetXmin(htup, record->xl_xid);
+ HeapTupleHeaderSetXmin(htup, XLogRecGetXid(record));
HeapTupleHeaderSetCmin(htup, FirstCommandId);
HeapTupleHeaderSetXmax(htup, xlrec->new_xmax);
/* Make sure there is no forward chain link in t_ctid */
- htup->t_ctid = xlrec->newtid;
+ htup->t_ctid = newtid;
offnum = PageAddItem(page, (Item) htup, newlen, offnum, true, true);
if (offnum == InvalidOffsetNumber)
PageSetLSN(page, lsn);
MarkBufferDirty(nbuffer);
}
+
if (BufferIsValid(nbuffer) && nbuffer != obuffer)
UnlockReleaseBuffer(nbuffer);
if (BufferIsValid(obuffer))
* totally accurate anyway.
*/
if (newaction == BLK_NEEDS_REDO && !hot_update && freespace < BLCKSZ / 5)
- XLogRecordPageWithFreeSpace(xlrec->target.node,
- ItemPointerGetBlockNumber(&(xlrec->newtid)),
- freespace);
+ XLogRecordPageWithFreeSpace(rnode, newblk, freespace);
}
static void
-heap_xlog_lock(XLogRecPtr lsn, XLogRecord *record)
+heap_xlog_lock(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_heap_lock *xlrec = (xl_heap_lock *) XLogRecGetData(record);
Buffer buffer;
Page page;
ItemId lp = NULL;
HeapTupleHeader htup;
- if (XLogReadBufferForRedo(lsn, record, 0, xlrec->target.node,
- ItemPointerGetBlockNumber(&xlrec->target.tid),
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
page = (Page) BufferGetPage(buffer);
- offnum = ItemPointerGetOffsetNumber(&(xlrec->target.tid));
+ offnum = xlrec->offnum;
if (PageGetMaxOffsetNumber(page) >= offnum)
lp = PageGetItemId(page, offnum);
{
HeapTupleHeaderClearHotUpdated(htup);
/* Make sure there is no forward chain link in t_ctid */
- htup->t_ctid = xlrec->target.tid;
+ ItemPointerSet(&htup->t_ctid,
+ BufferGetBlockNumber(buffer),
+ offnum);
}
HeapTupleHeaderSetXmax(htup, xlrec->locking_xid);
HeapTupleHeaderSetCmax(htup, FirstCommandId, false);
}
static void
-heap_xlog_lock_updated(XLogRecPtr lsn, XLogRecord *record)
+heap_xlog_lock_updated(XLogReaderState *record)
{
- xl_heap_lock_updated *xlrec =
- (xl_heap_lock_updated *) XLogRecGetData(record);
+ XLogRecPtr lsn = record->EndRecPtr;
+ xl_heap_lock_updated *xlrec;
Buffer buffer;
Page page;
OffsetNumber offnum;
ItemId lp = NULL;
HeapTupleHeader htup;
- if (XLogReadBufferForRedo(lsn, record, 0, xlrec->target.node,
- ItemPointerGetBlockNumber(&(xlrec->target.tid)),
- &buffer) == BLK_NEEDS_REDO)
+ xlrec = (xl_heap_lock_updated *) XLogRecGetData(record);
+
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
page = BufferGetPage(buffer);
- offnum = ItemPointerGetOffsetNumber(&(xlrec->target.tid));
+
+ offnum = xlrec->offnum;
if (PageGetMaxOffsetNumber(page) >= offnum)
lp = PageGetItemId(page, offnum);
}
static void
-heap_xlog_inplace(XLogRecPtr lsn, XLogRecord *record)
+heap_xlog_inplace(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_heap_inplace *xlrec = (xl_heap_inplace *) XLogRecGetData(record);
Buffer buffer;
Page page;
ItemId lp = NULL;
HeapTupleHeader htup;
uint32 oldlen;
- uint32 newlen;
+ Size newlen;
- if (XLogReadBufferForRedo(lsn, record, 0, xlrec->target.node,
- ItemPointerGetBlockNumber(&(xlrec->target.tid)),
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
+ char *newtup = XLogRecGetBlockData(record, 0, &newlen);
+
page = BufferGetPage(buffer);
- offnum = ItemPointerGetOffsetNumber(&(xlrec->target.tid));
+ offnum = xlrec->offnum;
if (PageGetMaxOffsetNumber(page) >= offnum)
lp = PageGetItemId(page, offnum);
htup = (HeapTupleHeader) PageGetItem(page, lp);
oldlen = ItemIdGetLength(lp) - htup->t_hoff;
- newlen = record->xl_len - SizeOfHeapInplace;
if (oldlen != newlen)
elog(PANIC, "heap_inplace_redo: wrong tuple length");
- memcpy((char *) htup + htup->t_hoff,
- (char *) xlrec + SizeOfHeapInplace,
- newlen);
+ memcpy((char *) htup + htup->t_hoff, newtup, newlen);
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
void
-heap_redo(XLogRecPtr lsn, XLogRecord *record)
+heap_redo(XLogReaderState *record)
{
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
/*
* These operations don't overwrite MVCC data so no conflict processing is
switch (info & XLOG_HEAP_OPMASK)
{
case XLOG_HEAP_INSERT:
- heap_xlog_insert(lsn, record);
+ heap_xlog_insert(record);
break;
case XLOG_HEAP_DELETE:
- heap_xlog_delete(lsn, record);
+ heap_xlog_delete(record);
break;
case XLOG_HEAP_UPDATE:
- heap_xlog_update(lsn, record, false);
+ heap_xlog_update(record, false);
break;
case XLOG_HEAP_HOT_UPDATE:
- heap_xlog_update(lsn, record, true);
+ heap_xlog_update(record, true);
break;
case XLOG_HEAP_LOCK:
- heap_xlog_lock(lsn, record);
+ heap_xlog_lock(record);
break;
case XLOG_HEAP_INPLACE:
- heap_xlog_inplace(lsn, record);
+ heap_xlog_inplace(record);
break;
default:
elog(PANIC, "heap_redo: unknown op code %u", info);
}
void
-heap2_redo(XLogRecPtr lsn, XLogRecord *record)
+heap2_redo(XLogReaderState *record)
{
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
switch (info & XLOG_HEAP_OPMASK)
{
case XLOG_HEAP2_CLEAN:
- heap_xlog_clean(lsn, record);
+ heap_xlog_clean(record);
break;
case XLOG_HEAP2_FREEZE_PAGE:
- heap_xlog_freeze_page(lsn, record);
+ heap_xlog_freeze_page(record);
break;
case XLOG_HEAP2_CLEANUP_INFO:
- heap_xlog_cleanup_info(lsn, record);
+ heap_xlog_cleanup_info(record);
break;
case XLOG_HEAP2_VISIBLE:
- heap_xlog_visible(lsn, record);
+ heap_xlog_visible(record);
break;
case XLOG_HEAP2_MULTI_INSERT:
- heap_xlog_multi_insert(lsn, record);
+ heap_xlog_multi_insert(record);
break;
case XLOG_HEAP2_LOCK_UPDATED:
- heap_xlog_lock_updated(lsn, record);
+ heap_xlog_lock_updated(record);
break;
case XLOG_HEAP2_NEW_CID:
*/
break;
case XLOG_HEAP2_REWRITE:
- heap_xlog_logical_rewrite(lsn, record);
+ heap_xlog_logical_rewrite(record);
break;
default:
elog(PANIC, "heap2_redo: unknown op code %u", info);
hash_seq_init(&seq_status, state->rs_logical_mappings);
while ((src = (RewriteMappingFile *) hash_seq_search(&seq_status)) != NULL)
{
- XLogRecData rdata[2];
char *waldata;
char *waldata_start;
xl_heap_rewrite_mapping xlrec;
xlrec.offset = src->off;
xlrec.start_lsn = state->rs_begin_lsn;
- rdata[0].data = (char *) (&xlrec);
- rdata[0].len = sizeof(xlrec);
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
-
/* write all mappings consecutively */
len = src->num_mappings * sizeof(LogicalRewriteMappingData);
waldata_start = waldata = palloc(len);
written, len)));
src->off += len;
- rdata[1].data = waldata_start;
- rdata[1].len = len;
- rdata[1].buffer = InvalidBuffer;
- rdata[1].next = NULL;
+ XLogBeginInsert();
+ XLogRegisterData((char *) (&xlrec), sizeof(xlrec));
+ XLogRegisterData(waldata_start, len);
/* write xlog record */
- XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_REWRITE, rdata);
+ XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_REWRITE);
pfree(waldata_start);
}
* Replay XLOG_HEAP2_REWRITE records
*/
void
-heap_xlog_logical_rewrite(XLogRecPtr lsn, XLogRecord *r)
+heap_xlog_logical_rewrite(XLogReaderState *r)
{
char path[MAXPGPATH];
int fd;
xlrec->mapped_db, xlrec->mapped_rel,
(uint32) (xlrec->start_lsn >> 32),
(uint32) xlrec->start_lsn,
- xlrec->mapped_xid, r->xl_xid);
+ xlrec->mapped_xid, XLogRecGetXid(r));
fd = OpenTransientFile(path,
O_CREAT | O_WRONLY | PG_BINARY,
if (RelationNeedsWAL(rel))
{
xl_btree_insert xlrec;
- BlockNumber xlleftchild;
xl_btree_metadata xlmeta;
uint8 xlinfo;
XLogRecPtr recptr;
- XLogRecData rdata[4];
- XLogRecData *nextrdata;
IndexTupleData trunctuple;
- xlrec.target.node = rel->rd_node;
- ItemPointerSet(&(xlrec.target.tid), itup_blkno, itup_off);
+ xlrec.offnum = itup_off;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfBtreeInsert;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = nextrdata = &(rdata[1]);
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfBtreeInsert);
if (P_ISLEAF(lpageop))
xlinfo = XLOG_BTREE_INSERT_LEAF;
else
{
/*
- * Include the block number of the left child, whose
- * INCOMPLETE_SPLIT flag was cleared.
+ * Register the left child whose INCOMPLETE_SPLIT flag was
+ * cleared.
*/
- xlleftchild = BufferGetBlockNumber(cbuf);
- nextrdata->data = (char *) &xlleftchild;
- nextrdata->len = sizeof(BlockNumber);
- nextrdata->buffer = cbuf;
- nextrdata->buffer_std = true;
- nextrdata->next = nextrdata + 1;
- nextrdata++;
+ XLogRegisterBuffer(1, cbuf, REGBUF_STANDARD);
xlinfo = XLOG_BTREE_INSERT_UPPER;
}
xlmeta.fastroot = metad->btm_fastroot;
xlmeta.fastlevel = metad->btm_fastlevel;
- nextrdata->data = (char *) &xlmeta;
- nextrdata->len = sizeof(xl_btree_metadata);
- nextrdata->buffer = InvalidBuffer;
- nextrdata->next = nextrdata + 1;
- nextrdata++;
+ XLogRegisterBuffer(2, metabuf, REGBUF_WILL_INIT);
+ XLogRegisterBufData(2, (char *) &xlmeta, sizeof(xl_btree_metadata));
xlinfo = XLOG_BTREE_INSERT_META;
}
/* Read comments in _bt_pgaddtup */
+ XLogRegisterBuffer(0, buf, REGBUF_STANDARD);
if (!P_ISLEAF(lpageop) && newitemoff == P_FIRSTDATAKEY(lpageop))
{
trunctuple = *itup;
trunctuple.t_info = sizeof(IndexTupleData);
- nextrdata->data = (char *) &trunctuple;
- nextrdata->len = sizeof(IndexTupleData);
+ XLogRegisterBufData(0, (char *) &trunctuple,
+ sizeof(IndexTupleData));
}
else
- {
- nextrdata->data = (char *) itup;
- nextrdata->len = IndexTupleDSize(*itup);
- }
- nextrdata->buffer = buf;
- nextrdata->buffer_std = true;
- nextrdata->next = NULL;
+ XLogRegisterBufData(0, (char *) itup, IndexTupleDSize(*itup));
- recptr = XLogInsert(RM_BTREE_ID, xlinfo, rdata);
+ recptr = XLogInsert(RM_BTREE_ID, xlinfo);
if (BufferIsValid(metabuf))
{
xl_btree_split xlrec;
uint8 xlinfo;
XLogRecPtr recptr;
- XLogRecData rdata[7];
- XLogRecData *lastrdata;
- BlockNumber cblkno;
-
- xlrec.node = rel->rd_node;
- xlrec.leftsib = origpagenumber;
- xlrec.rightsib = rightpagenumber;
- xlrec.rnext = ropaque->btpo_next;
+
xlrec.level = ropaque->btpo.level;
xlrec.firstright = firstright;
+ xlrec.newitemoff = newitemoff;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfBtreeSplit;
- rdata[0].buffer = InvalidBuffer;
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfBtreeSplit);
- lastrdata = &rdata[0];
+ XLogRegisterBuffer(0, buf, REGBUF_STANDARD);
+ XLogRegisterBuffer(1, rbuf, REGBUF_WILL_INIT);
+ /* Log the right sibling, because we've changed its prev-pointer. */
+ if (!P_RIGHTMOST(ropaque))
+ XLogRegisterBuffer(2, sbuf, REGBUF_STANDARD);
+ if (BufferIsValid(cbuf))
+ XLogRegisterBuffer(3, cbuf, REGBUF_STANDARD);
/*
- * Log the new item and its offset, if it was inserted on the left
- * page. (If it was put on the right page, we don't need to explicitly
- * WAL log it because it's included with all the other items on the
- * right page.) Show the new item as belonging to the left page
- * buffer, so that it is not stored if XLogInsert decides it needs a
- * full-page image of the left page. We store the offset anyway,
- * though, to support archive compression of these records.
+ * Log the new item, if it was inserted on the left page. (If it was
+ * put on the right page, we don't need to explicitly WAL log it
+ * because it's included with all the other items on the right page.)
+ * Show the new item as belonging to the left page buffer, so that it
+ * is not stored if XLogInsert decides it needs a full-page image of
+ * the left page. We store the offset anyway, though, to support
+ * archive compression of these records.
*/
if (newitemonleft)
- {
- lastrdata->next = lastrdata + 1;
- lastrdata++;
-
- lastrdata->data = (char *) &newitemoff;
- lastrdata->len = sizeof(OffsetNumber);
- lastrdata->buffer = InvalidBuffer;
-
- lastrdata->next = lastrdata + 1;
- lastrdata++;
-
- lastrdata->data = (char *) newitem;
- lastrdata->len = MAXALIGN(newitemsz);
- lastrdata->buffer = buf; /* backup block 0 */
- lastrdata->buffer_std = true;
- }
+ XLogRegisterBufData(0, (char *) newitem, MAXALIGN(newitemsz));
/* Log left page */
if (!isleaf)
{
- lastrdata->next = lastrdata + 1;
- lastrdata++;
-
/*
* We must also log the left page's high key, because the right
* page's leftmost key is suppressed on non-leaf levels. Show it
*/
itemid = PageGetItemId(origpage, P_HIKEY);
item = (IndexTuple) PageGetItem(origpage, itemid);
- lastrdata->data = (char *) item;
- lastrdata->len = MAXALIGN(IndexTupleSize(item));
- lastrdata->buffer = buf; /* backup block 0 */
- lastrdata->buffer_std = true;
- }
-
- if (isleaf && !newitemonleft)
- {
- lastrdata->next = lastrdata + 1;
- lastrdata++;
-
- /*
- * Although we don't need to WAL-log anything on the left page, we
- * still need XLogInsert to consider storing a full-page image of
- * the left page, so make an empty entry referencing that buffer.
- * This also ensures that the left page is always backup block 0.
- */
- lastrdata->data = NULL;
- lastrdata->len = 0;
- lastrdata->buffer = buf; /* backup block 0 */
- lastrdata->buffer_std = true;
- }
-
- /*
- * Log block number of left child, whose INCOMPLETE_SPLIT flag this
- * insertion clears.
- */
- if (!isleaf)
- {
- lastrdata->next = lastrdata + 1;
- lastrdata++;
-
- cblkno = BufferGetBlockNumber(cbuf);
- lastrdata->data = (char *) &cblkno;
- lastrdata->len = sizeof(BlockNumber);
- lastrdata->buffer = cbuf; /* backup block 1 */
- lastrdata->buffer_std = true;
+ XLogRegisterBufData(0, (char *) item, MAXALIGN(IndexTupleSize(item)));
}
/*
* and so the item pointers can be reconstructed. See comments for
* _bt_restore_page().
*/
- lastrdata->next = lastrdata + 1;
- lastrdata++;
-
- lastrdata->data = (char *) rightpage +
- ((PageHeader) rightpage)->pd_upper;
- lastrdata->len = ((PageHeader) rightpage)->pd_special -
- ((PageHeader) rightpage)->pd_upper;
- lastrdata->buffer = InvalidBuffer;
-
- /* Log the right sibling, because we've changed its' prev-pointer. */
- if (!P_RIGHTMOST(ropaque))
- {
- lastrdata->next = lastrdata + 1;
- lastrdata++;
-
- lastrdata->data = NULL;
- lastrdata->len = 0;
- lastrdata->buffer = sbuf; /* bkp block 1 (leaf) or 2 (non-leaf) */
- lastrdata->buffer_std = true;
- }
-
- lastrdata->next = NULL;
+ XLogRegisterBufData(1,
+ (char *) rightpage + ((PageHeader) rightpage)->pd_upper,
+ ((PageHeader) rightpage)->pd_special - ((PageHeader) rightpage)->pd_upper);
if (isroot)
xlinfo = newitemonleft ? XLOG_BTREE_SPLIT_L_ROOT : XLOG_BTREE_SPLIT_R_ROOT;
else
xlinfo = newitemonleft ? XLOG_BTREE_SPLIT_L : XLOG_BTREE_SPLIT_R;
- recptr = XLogInsert(RM_BTREE_ID, xlinfo, rdata);
+ recptr = XLogInsert(RM_BTREE_ID, xlinfo);
PageSetLSN(origpage, recptr);
PageSetLSN(rightpage, recptr);
{
xl_btree_newroot xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[3];
+ xl_btree_metadata md;
- xlrec.node = rel->rd_node;
xlrec.rootblk = rootblknum;
xlrec.level = metad->btm_level;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfBtreeNewroot;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfBtreeNewroot);
+
+ XLogRegisterBuffer(0, rootbuf, REGBUF_WILL_INIT);
+ XLogRegisterBuffer(1, lbuf, REGBUF_STANDARD);
+ XLogRegisterBuffer(2, metabuf, REGBUF_WILL_INIT);
+
+ md.root = rootblknum;
+ md.level = metad->btm_level;
+ md.fastroot = rootblknum;
+ md.fastlevel = metad->btm_level;
+
+ XLogRegisterBufData(2, (char *) &md, sizeof(xl_btree_metadata));
/*
* Direct access to page is not good but faster - we should implement
* some new func in page API.
*/
- rdata[1].data = (char *) rootpage + ((PageHeader) rootpage)->pd_upper;
- rdata[1].len = ((PageHeader) rootpage)->pd_special -
- ((PageHeader) rootpage)->pd_upper;
- rdata[1].buffer = InvalidBuffer;
- rdata[1].next = &(rdata[2]);
-
- /* Make a full-page image of the left child if needed */
- rdata[2].data = NULL;
- rdata[2].len = 0;
- rdata[2].buffer = lbuf;
- rdata[2].next = NULL;
-
- recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_NEWROOT, rdata);
+ XLogRegisterBufData(0,
+ (char *) rootpage + ((PageHeader) rootpage)->pd_upper,
+ ((PageHeader) rootpage)->pd_special -
+ ((PageHeader) rootpage)->pd_upper);
+
+ recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_NEWROOT);
PageSetLSN(lpage, recptr);
PageSetLSN(rootpage, recptr);
{
xl_btree_newroot xlrec;
XLogRecPtr recptr;
- XLogRecData rdata;
+ xl_btree_metadata md;
+
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, rootbuf, REGBUF_WILL_INIT);
+ XLogRegisterBuffer(2, metabuf, REGBUF_WILL_INIT);
+
+ md.root = rootblkno;
+ md.level = 0;
+ md.fastroot = rootblkno;
+ md.fastlevel = 0;
+
+ XLogRegisterBufData(2, (char *) &md, sizeof(xl_btree_metadata));
- xlrec.node = rel->rd_node;
xlrec.rootblk = rootblkno;
xlrec.level = 0;
- rdata.data = (char *) &xlrec;
- rdata.len = SizeOfBtreeNewroot;
- rdata.buffer = InvalidBuffer;
- rdata.next = NULL;
+ XLogRegisterData((char *) &xlrec, SizeOfBtreeNewroot);
- recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_NEWROOT, &rdata);
+ recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_NEWROOT);
PageSetLSN(rootpage, recptr);
PageSetLSN(metapg, recptr);
static void
_bt_log_reuse_page(Relation rel, BlockNumber blkno, TransactionId latestRemovedXid)
{
- if (!RelationNeedsWAL(rel))
- return;
-
- /* No ereport(ERROR) until changes are logged */
- START_CRIT_SECTION();
+ xl_btree_reuse_page xlrec_reuse;
/*
- * We don't do MarkBufferDirty here because we're about to initialise the
- * page, and nobody else can see it yet.
+ * Note that we don't register the buffer with the record, because this
+ * operation doesn't modify the page. This record only exists to provide a
+ * conflict point for Hot Standby.
*/
/* XLOG stuff */
- {
- XLogRecData rdata[1];
- xl_btree_reuse_page xlrec_reuse;
+ xlrec_reuse.node = rel->rd_node;
+ xlrec_reuse.block = blkno;
+ xlrec_reuse.latestRemovedXid = latestRemovedXid;
- xlrec_reuse.node = rel->rd_node;
- xlrec_reuse.block = blkno;
- xlrec_reuse.latestRemovedXid = latestRemovedXid;
- rdata[0].data = (char *) &xlrec_reuse;
- rdata[0].len = SizeOfBtreeReusePage;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = NULL;
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec_reuse, SizeOfBtreeReusePage);
- XLogInsert(RM_BTREE_ID, XLOG_BTREE_REUSE_PAGE, rdata);
-
- /*
- * We don't do PageSetLSN here because we're about to initialise the
- * page, so no need.
- */
- }
-
- END_CRIT_SECTION();
+ XLogInsert(RM_BTREE_ID, XLOG_BTREE_REUSE_PAGE);
}
/*
* WAL record that will allow us to conflict with queries
* running on standby.
*/
- if (XLogStandbyInfoActive())
+ if (XLogStandbyInfoActive() && RelationNeedsWAL(rel))
{
BTPageOpaque opaque = (BTPageOpaque) PageGetSpecialPointer(page);
if (RelationNeedsWAL(rel))
{
XLogRecPtr recptr;
- XLogRecData rdata[2];
xl_btree_vacuum xlrec_vacuum;
- xlrec_vacuum.node = rel->rd_node;
- xlrec_vacuum.block = BufferGetBlockNumber(buf);
-
xlrec_vacuum.lastBlockVacuumed = lastBlockVacuumed;
- rdata[0].data = (char *) &xlrec_vacuum;
- rdata[0].len = SizeOfBtreeVacuum;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
+
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, buf, REGBUF_STANDARD);
+ XLogRegisterData((char *) &xlrec_vacuum, SizeOfBtreeVacuum);
/*
* The target-offsets array is not in the buffer, but pretend that it
* need not be stored too.
*/
if (nitems > 0)
- {
- rdata[1].data = (char *) itemnos;
- rdata[1].len = nitems * sizeof(OffsetNumber);
- }
- else
- {
- rdata[1].data = NULL;
- rdata[1].len = 0;
- }
- rdata[1].buffer = buf;
- rdata[1].buffer_std = true;
- rdata[1].next = NULL;
+ XLogRegisterBufData(0, (char *) itemnos, nitems * sizeof(OffsetNumber));
- recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_VACUUM, rdata);
+ recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_VACUUM);
PageSetLSN(page, recptr);
}
if (RelationNeedsWAL(rel))
{
XLogRecPtr recptr;
- XLogRecData rdata[3];
xl_btree_delete xlrec_delete;
- xlrec_delete.node = rel->rd_node;
xlrec_delete.hnode = heapRel->rd_node;
- xlrec_delete.block = BufferGetBlockNumber(buf);
xlrec_delete.nitems = nitems;
- rdata[0].data = (char *) &xlrec_delete;
- rdata[0].len = SizeOfBtreeDelete;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, buf, REGBUF_STANDARD);
+ XLogRegisterData((char *) &xlrec_delete, SizeOfBtreeDelete);
/*
* We need the target-offsets array whether or not we store the whole
* buffer, to allow us to find the latestRemovedXid on a standby
* server.
*/
- rdata[1].data = (char *) itemnos;
- rdata[1].len = nitems * sizeof(OffsetNumber);
- rdata[1].buffer = InvalidBuffer;
- rdata[1].next = &(rdata[2]);
-
- rdata[2].data = NULL;
- rdata[2].len = 0;
- rdata[2].buffer = buf;
- rdata[2].buffer_std = true;
- rdata[2].next = NULL;
+ XLogRegisterData((char *) itemnos, nitems * sizeof(OffsetNumber));
- recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_DELETE, rdata);
+ recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_DELETE);
PageSetLSN(page, recptr);
}
{
xl_btree_mark_page_halfdead xlrec;
XLogRecPtr recptr;
- XLogRecData rdata[2];
- xlrec.target.node = rel->rd_node;
- ItemPointerSet(&(xlrec.target.tid), BufferGetBlockNumber(topparent), topoff);
+ xlrec.poffset = topoff;
xlrec.leafblk = leafblkno;
if (target != leafblkno)
xlrec.topparent = target;
else
xlrec.topparent = InvalidBlockNumber;
+ XLogBeginInsert();
+ XLogRegisterBuffer(0, leafbuf, REGBUF_WILL_INIT);
+ XLogRegisterBuffer(1, topparent, REGBUF_STANDARD);
+
page = BufferGetPage(leafbuf);
opaque = (BTPageOpaque) PageGetSpecialPointer(page);
xlrec.leftblk = opaque->btpo_prev;
xlrec.rightblk = opaque->btpo_next;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfBtreeMarkPageHalfDead;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = &(rdata[1]);
-
- rdata[1].data = NULL;
- rdata[1].len = 0;
- rdata[1].buffer = topparent;
- rdata[1].buffer_std = true;
- rdata[1].next = NULL;
+ XLogRegisterData((char *) &xlrec, SizeOfBtreeMarkPageHalfDead);
- recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_MARK_PAGE_HALFDEAD, rdata);
+ recptr = XLogInsert(RM_BTREE_ID, XLOG_BTREE_MARK_PAGE_HALFDEAD);
page = BufferGetPage(topparent);
PageSetLSN(page, recptr);
xl_btree_metadata xlmeta;
uint8 xlinfo;
XLogRecPtr recptr;
- XLogRecData rdata[4];
- XLogRecData *nextrdata;
- xlrec.node = rel->rd_node;
+ XLogBeginInsert();
+
+ XLogRegisterBuffer(0, buf, REGBUF_WILL_INIT);
+ if (BufferIsValid(lbuf))
+ XLogRegisterBuffer(1, lbuf, REGBUF_STANDARD);
+ XLogRegisterBuffer(2, rbuf, REGBUF_STANDARD);
+ if (target != leafblkno)
+ XLogRegisterBuffer(3, leafbuf, REGBUF_WILL_INIT);
/* information on the unlinked block */
- xlrec.deadblk = target;
xlrec.leftsib = leftsib;
xlrec.rightsib = rightsib;
xlrec.btpo_xact = opaque->btpo.xact;
/* information needed to recreate the leaf block (if not the target) */
- xlrec.leafblk = leafblkno;
xlrec.leafleftsib = leafleftsib;
xlrec.leafrightsib = leafrightsib;
xlrec.topparent = nextchild;
- rdata[0].data = (char *) &xlrec;
- rdata[0].len = SizeOfBtreeUnlinkPage;
- rdata[0].buffer = InvalidBuffer;
- rdata[0].next = nextrdata = &(rdata[1]);
+ XLogRegisterData((char *) &xlrec, SizeOfBtreeUnlinkPage);
if (BufferIsValid(metabuf))
{
+ XLogRegisterBuffer(4, metabuf, REGBUF_WILL_INIT);
+
xlmeta.root = metad->btm_root;
xlmeta.level = metad->btm_level;
xlmeta.fastroot = metad->btm_fastroot;
xlmeta.fastlevel = metad->btm_fastlevel;
- nextrdata->data = (char *) &xlmeta;
- nextrdata->len = sizeof(xl_btree_metadata);
- nextrdata->buffer = InvalidBuffer;
- nextrdata->next = nextrdata + 1;
- nextrdata++;
+ XLogRegisterBufData(4, (char *) &xlmeta, sizeof(xl_btree_metadata));
xlinfo = XLOG_BTREE_UNLINK_PAGE_META;
}
else
xlinfo = XLOG_BTREE_UNLINK_PAGE;
- nextrdata->data = NULL;
- nextrdata->len = 0;
- nextrdata->buffer = rbuf;
- nextrdata->buffer_std = true;
- nextrdata->next = NULL;
-
- if (BufferIsValid(lbuf))
- {
- nextrdata->next = nextrdata + 1;
- nextrdata++;
- nextrdata->data = NULL;
- nextrdata->len = 0;
- nextrdata->buffer = lbuf;
- nextrdata->buffer_std = true;
- nextrdata->next = NULL;
- }
-
- recptr = XLogInsert(RM_BTREE_ID, xlinfo, rdata);
+ recptr = XLogInsert(RM_BTREE_ID, xlinfo);
if (BufferIsValid(metabuf))
{
}
static void
-_bt_restore_meta(RelFileNode rnode, XLogRecPtr lsn,
- BlockNumber root, uint32 level,
- BlockNumber fastroot, uint32 fastlevel)
+_bt_restore_meta(XLogReaderState *record, uint8 block_id)
{
+ XLogRecPtr lsn = record->EndRecPtr;
Buffer metabuf;
Page metapg;
BTMetaPageData *md;
BTPageOpaque pageop;
+ xl_btree_metadata *xlrec;
+ char *ptr;
+ Size len;
- metabuf = XLogReadBuffer(rnode, BTREE_METAPAGE, true);
- Assert(BufferIsValid(metabuf));
+ metabuf = XLogInitBufferForRedo(record, block_id);
+ ptr = XLogRecGetBlockData(record, block_id, &len);
+
+ Assert(len == sizeof(xl_btree_metadata));
+ Assert(BufferGetBlockNumber(metabuf) == BTREE_METAPAGE);
+ xlrec = (xl_btree_metadata *) ptr;
metapg = BufferGetPage(metabuf);
_bt_pageinit(metapg, BufferGetPageSize(metabuf));
md = BTPageGetMeta(metapg);
md->btm_magic = BTREE_MAGIC;
md->btm_version = BTREE_VERSION;
- md->btm_root = root;
- md->btm_level = level;
- md->btm_fastroot = fastroot;
- md->btm_fastlevel = fastlevel;
+ md->btm_root = xlrec->root;
+ md->btm_level = xlrec->level;
+ md->btm_fastroot = xlrec->fastroot;
+ md->btm_fastlevel = xlrec->fastlevel;
pageop = (BTPageOpaque) PageGetSpecialPointer(metapg);
pageop->btpo_flags = BTP_META;
* types that can insert a downlink: insert, split, and newroot.
*/
static void
-_bt_clear_incomplete_split(XLogRecPtr lsn, XLogRecord *record,
- int block_index,
- RelFileNode rnode, BlockNumber cblock)
+_bt_clear_incomplete_split(XLogReaderState *record, uint8 block_id)
{
+ XLogRecPtr lsn = record->EndRecPtr;
Buffer buf;
- if (XLogReadBufferForRedo(lsn, record, block_index, rnode, cblock, &buf)
- == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, block_id, &buf) == BLK_NEEDS_REDO)
{
Page page = (Page) BufferGetPage(buf);
BTPageOpaque pageop = (BTPageOpaque) PageGetSpecialPointer(page);
}
static void
-btree_xlog_insert(bool isleaf, bool ismeta,
- XLogRecPtr lsn, XLogRecord *record)
+btree_xlog_insert(bool isleaf, bool ismeta, XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_btree_insert *xlrec = (xl_btree_insert *) XLogRecGetData(record);
Buffer buffer;
Page page;
- char *datapos;
- int datalen;
- xl_btree_metadata md;
- BlockNumber cblkno = 0;
- int main_blk_index;
-
- datapos = (char *) xlrec + SizeOfBtreeInsert;
- datalen = record->xl_len - SizeOfBtreeInsert;
-
- /*
- * if this insert finishes a split at lower level, extract the block
- * number of the (left) child.
- */
- if (!isleaf && (record->xl_info & XLR_BKP_BLOCK(0)) == 0)
- {
- memcpy(&cblkno, datapos, sizeof(BlockNumber));
- Assert(cblkno != 0);
- datapos += sizeof(BlockNumber);
- datalen -= sizeof(BlockNumber);
- }
- if (ismeta)
- {
- memcpy(&md, datapos, sizeof(xl_btree_metadata));
- datapos += sizeof(xl_btree_metadata);
- datalen -= sizeof(xl_btree_metadata);
- }
/*
* Insertion to an internal page finishes an incomplete split at the child
* cannot be updates happening.
*/
if (!isleaf)
+ _bt_clear_incomplete_split(record, 1);
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
- _bt_clear_incomplete_split(lsn, record, 0, xlrec->target.node, cblkno);
- main_blk_index = 1;
- }
- else
- main_blk_index = 0;
+ Size datalen;
+ char *datapos = XLogRecGetBlockData(record, 0, &datalen);
- if (XLogReadBufferForRedo(lsn, record, main_blk_index, xlrec->target.node,
- ItemPointerGetBlockNumber(&(xlrec->target.tid)),
- &buffer) == BLK_NEEDS_REDO)
- {
page = BufferGetPage(buffer);
- if (PageAddItem(page, (Item) datapos, datalen,
- ItemPointerGetOffsetNumber(&(xlrec->target.tid)),
+ if (PageAddItem(page, (Item) datapos, datalen, xlrec->offnum,
false, false) == InvalidOffsetNumber)
elog(PANIC, "btree_insert_redo: failed to add item");
* obsolete link from the metapage.
*/
if (ismeta)
- _bt_restore_meta(xlrec->target.node, lsn,
- md.root, md.level,
- md.fastroot, md.fastlevel);
+ _bt_restore_meta(record, 2);
}
static void
-btree_xlog_split(bool onleft, bool isroot,
- XLogRecPtr lsn, XLogRecord *record)
+btree_xlog_split(bool onleft, bool isroot, XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_btree_split *xlrec = (xl_btree_split *) XLogRecGetData(record);
bool isleaf = (xlrec->level == 0);
Buffer lbuf;
Page rpage;
BTPageOpaque ropaque;
char *datapos;
- int datalen;
- OffsetNumber newitemoff = 0;
- Item newitem = NULL;
- Size newitemsz = 0;
+ Size datalen;
Item left_hikey = NULL;
Size left_hikeysz = 0;
- BlockNumber cblkno = InvalidBlockNumber;
-
- datapos = (char *) xlrec + SizeOfBtreeSplit;
- datalen = record->xl_len - SizeOfBtreeSplit;
-
- /* Extract newitemoff and newitem, if present */
- if (onleft)
- {
- memcpy(&newitemoff, datapos, sizeof(OffsetNumber));
- datapos += sizeof(OffsetNumber);
- datalen -= sizeof(OffsetNumber);
- }
- if (onleft && !(record->xl_info & XLR_BKP_BLOCK(0)))
- {
- /*
- * We assume that 16-bit alignment is enough to apply IndexTupleSize
- * (since it's fetching from a uint16 field) and also enough for
- * PageAddItem to insert the tuple.
- */
- newitem = (Item) datapos;
- newitemsz = MAXALIGN(IndexTupleSize(newitem));
- datapos += newitemsz;
- datalen -= newitemsz;
- }
-
- /* Extract left hikey and its size (still assuming 16-bit alignment) */
- if (!isleaf && !(record->xl_info & XLR_BKP_BLOCK(0)))
- {
- left_hikey = (Item) datapos;
- left_hikeysz = MAXALIGN(IndexTupleSize(left_hikey));
- datapos += left_hikeysz;
- datalen -= left_hikeysz;
- }
+ BlockNumber leftsib;
+ BlockNumber rightsib;
+ BlockNumber rnext;
- /*
- * If this insertion finishes an incomplete split, get the block number of
- * the child.
- */
- if (!isleaf && !(record->xl_info & XLR_BKP_BLOCK(1)))
- {
- memcpy(&cblkno, datapos, sizeof(BlockNumber));
- datapos += sizeof(BlockNumber);
- datalen -= sizeof(BlockNumber);
- }
+ XLogRecGetBlockTag(record, 0, NULL, NULL, &leftsib);
+ XLogRecGetBlockTag(record, 1, NULL, NULL, &rightsib);
+ if (!XLogRecGetBlockTag(record, 2, NULL, NULL, &rnext))
+ rnext = P_NONE;
/*
* Clear the incomplete split flag on the left sibling of the child page
* before locking the other pages)
*/
if (!isleaf)
- _bt_clear_incomplete_split(lsn, record, 1, xlrec->node, cblkno);
+ _bt_clear_incomplete_split(record, 3);
/* Reconstruct right (new) sibling page from scratch */
- rbuf = XLogReadBuffer(xlrec->node, xlrec->rightsib, true);
- Assert(BufferIsValid(rbuf));
+ rbuf = XLogInitBufferForRedo(record, 1);
+ datapos = XLogRecGetBlockData(record, 1, &datalen);
rpage = (Page) BufferGetPage(rbuf);
_bt_pageinit(rpage, BufferGetPageSize(rbuf));
ropaque = (BTPageOpaque) PageGetSpecialPointer(rpage);
- ropaque->btpo_prev = xlrec->leftsib;
- ropaque->btpo_next = xlrec->rnext;
+ ropaque->btpo_prev = leftsib;
+ ropaque->btpo_next = rnext;
ropaque->btpo.level = xlrec->level;
ropaque->btpo_flags = isleaf ? BTP_LEAF : 0;
ropaque->btpo_cycleid = 0;
/* don't release the buffer yet; we touch right page's first item below */
/* Now reconstruct left (original) sibling page */
- if (XLogReadBufferForRedo(lsn, record, 0, xlrec->node, xlrec->leftsib,
- &lbuf) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &lbuf) == BLK_NEEDS_REDO)
{
/*
* To retain the same physical order of the tuples that they had, we
Page lpage = (Page) BufferGetPage(lbuf);
BTPageOpaque lopaque = (BTPageOpaque) PageGetSpecialPointer(lpage);
OffsetNumber off;
+ Item newitem;
+ Size newitemsz = 0;
Page newlpage;
OffsetNumber leftoff;
+ datapos = XLogRecGetBlockData(record, 0, &datalen);
+
+ if (onleft)
+ {
+ newitem = (Item) datapos;
+ newitemsz = MAXALIGN(IndexTupleSize(newitem));
+ datapos += newitemsz;
+ datalen -= newitemsz;
+ }
+
+ /* Extract left hikey and its size (assuming 16-bit alignment) */
+ if (!isleaf)
+ {
+ left_hikey = (Item) datapos;
+ left_hikeysz = MAXALIGN(IndexTupleSize(left_hikey));
+ datapos += left_hikeysz;
+ datalen -= left_hikeysz;
+ }
+ Assert(datalen == 0);
+
newlpage = PageGetTempPageCopySpecial(lpage);
/* Set high key */
Item item;
/* add the new item if it was inserted on left page */
- if (onleft && off == newitemoff)
+ if (onleft && off == xlrec->newitemoff)
{
if (PageAddItem(newlpage, newitem, newitemsz, leftoff,
false, false) == InvalidOffsetNumber)
}
/* cope with possibility that newitem goes at the end */
- if (onleft && off == newitemoff)
+ if (onleft && off == xlrec->newitemoff)
{
if (PageAddItem(newlpage, newitem, newitemsz, leftoff,
false, false) == InvalidOffsetNumber)
lopaque->btpo_flags = BTP_INCOMPLETE_SPLIT;
if (isleaf)
lopaque->btpo_flags |= BTP_LEAF;
- lopaque->btpo_next = xlrec->rightsib;
+ lopaque->btpo_next = rightsib;
lopaque->btpo_cycleid = 0;
PageSetLSN(lpage, lsn);
* replay, because no other index update can be in progress, and readers
* will cope properly when following an obsolete left-link.
*/
- if (xlrec->rnext != P_NONE)
+ if (rnext != P_NONE)
{
- /*
- * the backup block containing right sibling is 1 or 2, depending
- * whether this was a leaf or internal page.
- */
- int rnext_index = isleaf ? 1 : 2;
Buffer buffer;
- if (XLogReadBufferForRedo(lsn, record, rnext_index, xlrec->node,
- xlrec->rnext, &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 2, &buffer) == BLK_NEEDS_REDO)
{
Page page = (Page) BufferGetPage(buffer);
BTPageOpaque pageop = (BTPageOpaque) PageGetSpecialPointer(page);
- pageop->btpo_prev = xlrec->rightsib;
+ pageop->btpo_prev = rightsib;
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
static void
-btree_xlog_vacuum(XLogRecPtr lsn, XLogRecord *record)
+btree_xlog_vacuum(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_btree_vacuum *xlrec = (xl_btree_vacuum *) XLogRecGetData(record);
Buffer buffer;
Page page;
*/
if (HotStandbyActiveInReplay())
{
+ RelFileNode thisrnode;
+ BlockNumber thisblkno;
BlockNumber blkno;
- for (blkno = xlrec->lastBlockVacuumed + 1; blkno < xlrec->block; blkno++)
+ XLogRecGetBlockTag(record, 0, &thisrnode, NULL, &thisblkno);
+
+ for (blkno = xlrec->lastBlockVacuumed + 1; blkno < thisblkno; blkno++)
{
/*
* We use RBM_NORMAL_NO_LOG mode because it's not an error
* buffer manager we could optimise this so that if the block is
* not in shared_buffers we confirm it as unpinned.
*/
- buffer = XLogReadBufferExtended(xlrec->node, MAIN_FORKNUM, blkno,
+ buffer = XLogReadBufferExtended(thisrnode, MAIN_FORKNUM, blkno,
RBM_NORMAL_NO_LOG);
if (BufferIsValid(buffer))
{
* Like in btvacuumpage(), we need to take a cleanup lock on every leaf
* page. See nbtree/README for details.
*/
- if (XLogReadBufferForRedoExtended(lsn, record, 0,
- xlrec->node, MAIN_FORKNUM, xlrec->block,
- RBM_NORMAL, true, &buffer)
+ if (XLogReadBufferForRedoExtended(record, 0, RBM_NORMAL, true, &buffer)
== BLK_NEEDS_REDO)
{
+ char *ptr;
+ Size len;
+
+ ptr = XLogRecGetBlockData(record, 0, &len);
+
page = (Page) BufferGetPage(buffer);
- if (record->xl_len > SizeOfBtreeVacuum)
+ if (len > 0)
{
OffsetNumber *unused;
OffsetNumber *unend;
- unused = (OffsetNumber *) ((char *) xlrec + SizeOfBtreeVacuum);
- unend = (OffsetNumber *) ((char *) xlrec + record->xl_len);
+ unused = (OffsetNumber *) ptr;
+ unend = (OffsetNumber *) ((char *) ptr + len);
if ((unend - unused) > 0)
PageIndexMultiDelete(page, unused, unend - unused);
* XXX optimise later with something like XLogPrefetchBuffer()
*/
static TransactionId
-btree_xlog_delete_get_latestRemovedXid(xl_btree_delete *xlrec)
+btree_xlog_delete_get_latestRemovedXid(XLogReaderState *record)
{
+ xl_btree_delete *xlrec = (xl_btree_delete *) XLogRecGetData(record);
OffsetNumber *unused;
Buffer ibuffer,
hbuffer;
Page ipage,
hpage;
+ RelFileNode rnode;
+ BlockNumber blkno;
ItemId iitemid,
hitemid;
IndexTuple itup;
* InvalidTransactionId to cancel all HS transactions. That's probably
* overkill, but it's safe, and certainly better than panicking here.
*/
- ibuffer = XLogReadBuffer(xlrec->node, xlrec->block, false);
+ XLogRecGetBlockTag(record, 0, &rnode, NULL, &blkno);
+ ibuffer = XLogReadBufferExtended(rnode, MAIN_FORKNUM, blkno, RBM_NORMAL);
if (!BufferIsValid(ibuffer))
return InvalidTransactionId;
+ LockBuffer(ibuffer, BT_READ);
ipage = (Page) BufferGetPage(ibuffer);
/*
* Locate the heap page that the index tuple points at
*/
hblkno = ItemPointerGetBlockNumber(&(itup->t_tid));
- hbuffer = XLogReadBuffer(xlrec->hnode, hblkno, false);
+ hbuffer = XLogReadBufferExtended(xlrec->hnode, MAIN_FORKNUM, hblkno, RBM_NORMAL);
if (!BufferIsValid(hbuffer))
{
UnlockReleaseBuffer(ibuffer);
return InvalidTransactionId;
}
+ LockBuffer(hbuffer, BUFFER_LOCK_SHARE);
hpage = (Page) BufferGetPage(hbuffer);
/*
}
static void
-btree_xlog_delete(XLogRecPtr lsn, XLogRecord *record)
+btree_xlog_delete(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_btree_delete *xlrec = (xl_btree_delete *) XLogRecGetData(record);
Buffer buffer;
Page page;
*/
if (InHotStandby)
{
- TransactionId latestRemovedXid = btree_xlog_delete_get_latestRemovedXid(xlrec);
+ TransactionId latestRemovedXid = btree_xlog_delete_get_latestRemovedXid(record);
+ RelFileNode rnode;
- ResolveRecoveryConflictWithSnapshot(latestRemovedXid, xlrec->node);
+ XLogRecGetBlockTag(record, 0, &rnode, NULL, NULL);
+
+ ResolveRecoveryConflictWithSnapshot(latestRemovedXid, rnode);
}
/*
* We don't need to take a cleanup lock to apply these changes. See
* nbtree/README for details.
*/
- if (XLogReadBufferForRedo(lsn, record, 0, xlrec->node, xlrec->block,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
page = (Page) BufferGetPage(buffer);
- if (record->xl_len > SizeOfBtreeDelete)
+ if (XLogRecGetDataLen(record) > SizeOfBtreeDelete)
{
OffsetNumber *unused;
}
static void
-btree_xlog_mark_page_halfdead(uint8 info, XLogRecPtr lsn, XLogRecord *record)
+btree_xlog_mark_page_halfdead(uint8 info, XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_btree_mark_page_halfdead *xlrec = (xl_btree_mark_page_halfdead *) XLogRecGetData(record);
- BlockNumber parent;
Buffer buffer;
Page page;
BTPageOpaque pageop;
IndexTupleData trunctuple;
- parent = ItemPointerGetBlockNumber(&(xlrec->target.tid));
-
/*
* In normal operation, we would lock all the pages this WAL record
* touches before changing any of them. In WAL replay, it should be okay
*/
/* parent page */
- if (XLogReadBufferForRedo(lsn, record, 0, xlrec->target.node, parent,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 1, &buffer) == BLK_NEEDS_REDO)
{
OffsetNumber poffset;
ItemId itemid;
page = (Page) BufferGetPage(buffer);
pageop = (BTPageOpaque) PageGetSpecialPointer(page);
- poffset = ItemPointerGetOffsetNumber(&(xlrec->target.tid));
+ poffset = xlrec->poffset;
nextoffset = OffsetNumberNext(poffset);
itemid = PageGetItemId(page, nextoffset);
UnlockReleaseBuffer(buffer);
/* Rewrite the leaf page as a halfdead page */
- buffer = XLogReadBuffer(xlrec->target.node, xlrec->leafblk, true);
- Assert(BufferIsValid(buffer));
+ buffer = XLogInitBufferForRedo(record, 0);
page = (Page) BufferGetPage(buffer);
_bt_pageinit(page, BufferGetPageSize(buffer));
static void
-btree_xlog_unlink_page(uint8 info, XLogRecPtr lsn, XLogRecord *record)
+btree_xlog_unlink_page(uint8 info, XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_btree_unlink_page *xlrec = (xl_btree_unlink_page *) XLogRecGetData(record);
- BlockNumber target;
BlockNumber leftsib;
BlockNumber rightsib;
Buffer buffer;
Page page;
BTPageOpaque pageop;
- target = xlrec->deadblk;
leftsib = xlrec->leftsib;
rightsib = xlrec->rightsib;
*/
/* Fix left-link of right sibling */
- if (XLogReadBufferForRedo(lsn, record, 0, xlrec->node, rightsib, &buffer)
- == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 2, &buffer) == BLK_NEEDS_REDO)
{
page = (Page) BufferGetPage(buffer);
pageop = (BTPageOpaque) PageGetSpecialPointer(page);
/* Fix right-link of left sibling, if any */
if (leftsib != P_NONE)
{
- if (XLogReadBufferForRedo(lsn, record, 1, xlrec->node, leftsib, &buffer)
- == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 1, &buffer) == BLK_NEEDS_REDO)
{
page = (Page) BufferGetPage(buffer);
pageop = (BTPageOpaque) PageGetSpecialPointer(page);
}
/* Rewrite target page as empty deleted page */
- buffer = XLogReadBuffer(xlrec->node, target, true);
- Assert(BufferIsValid(buffer));
+ buffer = XLogInitBufferForRedo(record, 0);
page = (Page) BufferGetPage(buffer);
_bt_pageinit(page, BufferGetPageSize(buffer));
* itself, update the leaf to point to the next remaining child in the
* branch.
*/
- if (target != xlrec->leafblk)
+ if (XLogRecHasBlockRef(record, 3))
{
/*
* There is no real data on the page, so we just re-create it from
*/
IndexTupleData trunctuple;
- buffer = XLogReadBuffer(xlrec->node, xlrec->leafblk, true);
- Assert(BufferIsValid(buffer));
+ buffer = XLogInitBufferForRedo(record, 3);
page = (Page) BufferGetPage(buffer);
pageop = (BTPageOpaque) PageGetSpecialPointer(page);
/* Update metapage if needed */
if (info == XLOG_BTREE_UNLINK_PAGE_META)
- {
- xl_btree_metadata md;
-
- memcpy(&md, (char *) xlrec + SizeOfBtreeUnlinkPage,
- sizeof(xl_btree_metadata));
- _bt_restore_meta(xlrec->node, lsn,
- md.root, md.level,
- md.fastroot, md.fastlevel);
- }
+ _bt_restore_meta(record, 4);
}
static void
-btree_xlog_newroot(XLogRecPtr lsn, XLogRecord *record)
+btree_xlog_newroot(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
xl_btree_newroot *xlrec = (xl_btree_newroot *) XLogRecGetData(record);
Buffer buffer;
Page page;
BTPageOpaque pageop;
+ char *ptr;
+ Size len;
- buffer = XLogReadBuffer(xlrec->node, xlrec->rootblk, true);
- Assert(BufferIsValid(buffer));
+ buffer = XLogInitBufferForRedo(record, 0);
page = (Page) BufferGetPage(buffer);
_bt_pageinit(page, BufferGetPageSize(buffer));
pageop->btpo_flags |= BTP_LEAF;
pageop->btpo_cycleid = 0;
- if (record->xl_len > SizeOfBtreeNewroot)
+ if (xlrec->level > 0)
{
- IndexTuple itup;
- BlockNumber cblkno;
-
- _bt_restore_page(page,
- (char *) xlrec + SizeOfBtreeNewroot,
- record->xl_len - SizeOfBtreeNewroot);
- /* extract block number of the left-hand split page */
- itup = (IndexTuple) PageGetItem(page, PageGetItemId(page, P_HIKEY));
- cblkno = ItemPointerGetBlockNumber(&(itup->t_tid));
- Assert(ItemPointerGetOffsetNumber(&(itup->t_tid)) == P_HIKEY);
+ ptr = XLogRecGetBlockData(record, 0, &len);
+ _bt_restore_page(page, ptr, len);
/* Clear the incomplete-split flag in left child */
- _bt_clear_incomplete_split(lsn, record, 0, xlrec->node, cblkno);
+ _bt_clear_incomplete_split(record, 1);
}
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
UnlockReleaseBuffer(buffer);
- _bt_restore_meta(xlrec->node, lsn,
- xlrec->rootblk, xlrec->level,
- xlrec->rootblk, xlrec->level);
+ _bt_restore_meta(record, 2);
}
static void
-btree_xlog_reuse_page(XLogRecPtr lsn, XLogRecord *record)
+btree_xlog_reuse_page(XLogReaderState *record)
{
xl_btree_reuse_page *xlrec = (xl_btree_reuse_page *) XLogRecGetData(record);
ResolveRecoveryConflictWithSnapshot(xlrec->latestRemovedXid,
xlrec->node);
}
-
- /* Backup blocks are not used in reuse_page records */
- Assert(!(record->xl_info & XLR_BKP_BLOCK_MASK));
}
void
-btree_redo(XLogRecPtr lsn, XLogRecord *record)
+btree_redo(XLogReaderState *record)
{
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
switch (info)
{
case XLOG_BTREE_INSERT_LEAF:
- btree_xlog_insert(true, false, lsn, record);
+ btree_xlog_insert(true, false, record);
break;
case XLOG_BTREE_INSERT_UPPER:
- btree_xlog_insert(false, false, lsn, record);
+ btree_xlog_insert(false, false, record);
break;
case XLOG_BTREE_INSERT_META:
- btree_xlog_insert(false, true, lsn, record);
+ btree_xlog_insert(false, true, record);
break;
case XLOG_BTREE_SPLIT_L:
- btree_xlog_split(true, false, lsn, record);
+ btree_xlog_split(true, false, record);
break;
case XLOG_BTREE_SPLIT_R:
- btree_xlog_split(false, false, lsn, record);
+ btree_xlog_split(false, false, record);
break;
case XLOG_BTREE_SPLIT_L_ROOT:
- btree_xlog_split(true, true, lsn, record);
+ btree_xlog_split(true, true, record);
break;
case XLOG_BTREE_SPLIT_R_ROOT:
- btree_xlog_split(false, true, lsn, record);
+ btree_xlog_split(false, true, record);
break;
case XLOG_BTREE_VACUUM:
- btree_xlog_vacuum(lsn, record);
+ btree_xlog_vacuum(record);
break;
case XLOG_BTREE_DELETE:
- btree_xlog_delete(lsn, record);
+ btree_xlog_delete(record);
break;
case XLOG_BTREE_MARK_PAGE_HALFDEAD:
- btree_xlog_mark_page_halfdead(info, lsn, record);
+ btree_xlog_mark_page_halfdead(info, record);
break;
case XLOG_BTREE_UNLINK_PAGE:
case XLOG_BTREE_UNLINK_PAGE_META:
- btree_xlog_unlink_page(info, lsn, record);
+ btree_xlog_unlink_page(info, record);
break;
case XLOG_BTREE_NEWROOT:
- btree_xlog_newroot(lsn, record);
+ btree_xlog_newroot(record);
break;
case XLOG_BTREE_REUSE_PAGE:
- btree_xlog_reuse_page(lsn, record);
+ btree_xlog_reuse_page(record);
break;
default:
elog(PANIC, "btree_redo: unknown op code %u", info);
#include "access/brin_xlog.h"
void
-brin_desc(StringInfo buf, XLogRecord *record)
+brin_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
info &= XLOG_BRIN_OPMASK;
if (info == XLOG_BRIN_CREATE_INDEX)
{
xl_brin_createidx *xlrec = (xl_brin_createidx *) rec;
- appendStringInfo(buf, "v%d pagesPerRange %u rel %u/%u/%u",
- xlrec->version, xlrec->pagesPerRange,
- xlrec->node.spcNode, xlrec->node.dbNode,
- xlrec->node.relNode);
+ appendStringInfo(buf, "v%d pagesPerRange %u",
+ xlrec->version, xlrec->pagesPerRange);
}
else if (info == XLOG_BRIN_INSERT)
{
xl_brin_insert *xlrec = (xl_brin_insert *) rec;
- appendStringInfo(buf, "rel %u/%u/%u heapBlk %u revmapBlk %u pagesPerRange %u TID (%u,%u)",
- xlrec->node.spcNode, xlrec->node.dbNode,
- xlrec->node.relNode,
- xlrec->heapBlk, xlrec->revmapBlk,
+ appendStringInfo(buf, "heapBlk %u pagesPerRange %u offnum %u",
+ xlrec->heapBlk,
xlrec->pagesPerRange,
- ItemPointerGetBlockNumber(&xlrec->tid),
- ItemPointerGetOffsetNumber(&xlrec->tid));
+ xlrec->offnum);
}
else if (info == XLOG_BRIN_UPDATE)
{
xl_brin_update *xlrec = (xl_brin_update *) rec;
- appendStringInfo(buf, "rel %u/%u/%u heapBlk %u revmapBlk %u pagesPerRange %u old TID (%u,%u) TID (%u,%u)",
- xlrec->insert.node.spcNode, xlrec->insert.node.dbNode,
- xlrec->insert.node.relNode,
- xlrec->insert.heapBlk, xlrec->insert.revmapBlk,
+ appendStringInfo(buf, "heapBlk %u pagesPerRange %u old offnum %u, new offnum %u",
+ xlrec->insert.heapBlk,
xlrec->insert.pagesPerRange,
- ItemPointerGetBlockNumber(&xlrec->oldtid),
- ItemPointerGetOffsetNumber(&xlrec->oldtid),
- ItemPointerGetBlockNumber(&xlrec->insert.tid),
- ItemPointerGetOffsetNumber(&xlrec->insert.tid));
+ xlrec->oldOffnum,
+ xlrec->insert.offnum);
}
else if (info == XLOG_BRIN_SAMEPAGE_UPDATE)
{
xl_brin_samepage_update *xlrec = (xl_brin_samepage_update *) rec;
- appendStringInfo(buf, "rel %u/%u/%u TID (%u,%u)",
- xlrec->node.spcNode, xlrec->node.dbNode,
- xlrec->node.relNode,
- ItemPointerGetBlockNumber(&xlrec->tid),
- ItemPointerGetOffsetNumber(&xlrec->tid));
+ appendStringInfo(buf, "offnum %u", xlrec->offnum);
}
else if (info == XLOG_BRIN_REVMAP_EXTEND)
{
xl_brin_revmap_extend *xlrec = (xl_brin_revmap_extend *) rec;
- appendStringInfo(buf, "rel %u/%u/%u targetBlk %u",
- xlrec->node.spcNode, xlrec->node.dbNode,
- xlrec->node.relNode, xlrec->targetBlk);
+ appendStringInfo(buf, "targetBlk %u", xlrec->targetBlk);
}
}
void
-clog_desc(StringInfo buf, XLogRecord *record)
+clog_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
if (info == CLOG_ZEROPAGE || info == CLOG_TRUNCATE)
{
void
-dbase_desc(StringInfo buf, XLogRecord *record)
+dbase_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
if (info == XLOG_DBASE_CREATE)
{
#include "postgres.h"
#include "access/gin_private.h"
+#include "access/xlogutils.h"
#include "lib/stringinfo.h"
#include "storage/relfilenode.h"
-static void
-desc_node(StringInfo buf, RelFileNode node, BlockNumber blkno)
-{
- appendStringInfo(buf, "node: %u/%u/%u blkno: %u",
- node.spcNode, node.dbNode, node.relNode, blkno);
-}
-
static void
desc_recompress_leaf(StringInfo buf, ginxlogRecompressDataLeaf *insertData)
{
}
void
-gin_desc(StringInfo buf, XLogRecord *record)
+gin_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
switch (info)
{
case XLOG_GIN_CREATE_INDEX:
- desc_node(buf, *(RelFileNode *) rec, GIN_ROOT_BLKNO);
+ /* no further information */
break;
case XLOG_GIN_CREATE_PTREE:
- desc_node(buf, ((ginxlogCreatePostingTree *) rec)->node, ((ginxlogCreatePostingTree *) rec)->blkno);
+ /* no further information */
break;
case XLOG_GIN_INSERT:
{
ginxlogInsert *xlrec = (ginxlogInsert *) rec;
char *payload = rec + sizeof(ginxlogInsert);
- desc_node(buf, xlrec->node, xlrec->blkno);
- appendStringInfo(buf, " isdata: %c isleaf: %c",
+ appendStringInfo(buf, "isdata: %c isleaf: %c",
(xlrec->flags & GIN_INSERT_ISDATA) ? 'T' : 'F',
(xlrec->flags & GIN_INSERT_ISLEAF) ? 'T' : 'F');
if (!(xlrec->flags & GIN_INSERT_ISLEAF))
ginxlogRecompressDataLeaf *insertData =
(ginxlogRecompressDataLeaf *) payload;
- if (record->xl_info & XLR_BKP_BLOCK(0))
+ if (XLogRecHasBlockImage(record, 0))
appendStringInfo(buf, " (full page image)");
else
desc_recompress_leaf(buf, insertData);
{
ginxlogSplit *xlrec = (ginxlogSplit *) rec;
- desc_node(buf, ((ginxlogSplit *) rec)->node, ((ginxlogSplit *) rec)->lblkno);
- appendStringInfo(buf, " isrootsplit: %c", (((ginxlogSplit *) rec)->flags & GIN_SPLIT_ROOT) ? 'T' : 'F');
+ appendStringInfo(buf, "isrootsplit: %c",
+ (((ginxlogSplit *) rec)->flags & GIN_SPLIT_ROOT) ? 'T' : 'F');
appendStringInfo(buf, " isdata: %c isleaf: %c",
(xlrec->flags & GIN_INSERT_ISDATA) ? 'T' : 'F',
(xlrec->flags & GIN_INSERT_ISLEAF) ? 'T' : 'F');
}
break;
case XLOG_GIN_VACUUM_PAGE:
- desc_node(buf, ((ginxlogVacuumPage *) rec)->node, ((ginxlogVacuumPage *) rec)->blkno);
+ /* no further information */
break;
case XLOG_GIN_VACUUM_DATA_LEAF_PAGE:
{
ginxlogVacuumDataLeafPage *xlrec = (ginxlogVacuumDataLeafPage *) rec;
- desc_node(buf, xlrec->node, xlrec->blkno);
- if (record->xl_info & XLR_BKP_BLOCK(0))
+ if (XLogRecHasBlockImage(record, 0))
appendStringInfo(buf, " (full page image)");
else
desc_recompress_leaf(buf, &xlrec->data);
}
break;
case XLOG_GIN_DELETE_PAGE:
- desc_node(buf, ((ginxlogDeletePage *) rec)->node, ((ginxlogDeletePage *) rec)->blkno);
+ /* no further information */
break;
case XLOG_GIN_UPDATE_META_PAGE:
- desc_node(buf, ((ginxlogUpdateMeta *) rec)->node, GIN_METAPAGE_BLKNO);
+ /* no further information */
break;
case XLOG_GIN_INSERT_LISTPAGE:
- desc_node(buf, ((ginxlogInsertListPage *) rec)->node, ((ginxlogInsertListPage *) rec)->blkno);
+ /* no further information */
break;
case XLOG_GIN_DELETE_LISTPAGE:
- appendStringInfo(buf, "%d pages, ", ((ginxlogDeleteListPages *) rec)->ndeleted);
- desc_node(buf, ((ginxlogDeleteListPages *) rec)->node, GIN_METAPAGE_BLKNO);
+ appendStringInfo(buf, "ndeleted: %d",
+ ((ginxlogDeleteListPages *) rec)->ndeleted);
break;
}
}
#include "lib/stringinfo.h"
#include "storage/relfilenode.h"
-static void
-out_target(StringInfo buf, RelFileNode node)
-{
- appendStringInfo(buf, "rel %u/%u/%u",
- node.spcNode, node.dbNode, node.relNode);
-}
-
static void
out_gistxlogPageUpdate(StringInfo buf, gistxlogPageUpdate *xlrec)
{
- out_target(buf, xlrec->node);
- appendStringInfo(buf, "; block number %u", xlrec->blkno);
}
static void
out_gistxlogPageSplit(StringInfo buf, gistxlogPageSplit *xlrec)
{
- appendStringInfoString(buf, "page_split: ");
- out_target(buf, xlrec->node);
- appendStringInfo(buf, "; block number %u splits to %d pages",
- xlrec->origblkno, xlrec->npage);
+ appendStringInfo(buf, "page_split: splits to %d pages",
+ xlrec->npage);
}
void
-gist_desc(StringInfo buf, XLogRecord *record)
+gist_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
switch (info)
{
out_gistxlogPageSplit(buf, (gistxlogPageSplit *) rec);
break;
case XLOG_GIST_CREATE_INDEX:
- appendStringInfo(buf, "rel %u/%u/%u",
- ((RelFileNode *) rec)->spcNode,
- ((RelFileNode *) rec)->dbNode,
- ((RelFileNode *) rec)->relNode);
break;
}
}
#include "access/hash.h"
void
-hash_desc(StringInfo buf, XLogRecord *record)
+hash_desc(StringInfo buf, XLogReaderState *record)
{
}
#include "access/heapam_xlog.h"
-static void
-out_target(StringInfo buf, xl_heaptid *target)
-{
- appendStringInfo(buf, "rel %u/%u/%u; tid %u/%u",
- target->node.spcNode, target->node.dbNode, target->node.relNode,
- ItemPointerGetBlockNumber(&(target->tid)),
- ItemPointerGetOffsetNumber(&(target->tid)));
-}
-
static void
out_infobits(StringInfo buf, uint8 infobits)
{
}
void
-heap_desc(StringInfo buf, XLogRecord *record)
+heap_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
info &= XLOG_HEAP_OPMASK;
if (info == XLOG_HEAP_INSERT)
{
xl_heap_insert *xlrec = (xl_heap_insert *) rec;
- out_target(buf, &(xlrec->target));
+ appendStringInfo(buf, "off %u", xlrec->offnum);
}
else if (info == XLOG_HEAP_DELETE)
{
xl_heap_delete *xlrec = (xl_heap_delete *) rec;
- out_target(buf, &(xlrec->target));
+ appendStringInfo(buf, "off %u", xlrec->offnum);
appendStringInfoChar(buf, ' ');
out_infobits(buf, xlrec->infobits_set);
}
{
xl_heap_update *xlrec = (xl_heap_update *) rec;
- out_target(buf, &(xlrec->target));
- appendStringInfo(buf, " xmax %u ", xlrec->old_xmax);
+ appendStringInfo(buf, "off %u xmax %u",
+ xlrec->old_offnum,
+ xlrec->old_xmax);
out_infobits(buf, xlrec->old_infobits_set);
- appendStringInfo(buf, "; new tid %u/%u xmax %u",
- ItemPointerGetBlockNumber(&(xlrec->newtid)),
- ItemPointerGetOffsetNumber(&(xlrec->newtid)),
+ appendStringInfo(buf, "; new off %u xmax %u",
+ xlrec->new_offnum,
xlrec->new_xmax);
}
else if (info == XLOG_HEAP_HOT_UPDATE)
{
xl_heap_update *xlrec = (xl_heap_update *) rec;
- out_target(buf, &(xlrec->target));
- appendStringInfo(buf, " xmax %u ", xlrec->old_xmax);
+ appendStringInfo(buf, "off %u xmax %u",
+ xlrec->old_offnum,
+ xlrec->old_xmax);
out_infobits(buf, xlrec->old_infobits_set);
- appendStringInfo(buf, "; new tid %u/%u xmax %u",
- ItemPointerGetBlockNumber(&(xlrec->newtid)),
- ItemPointerGetOffsetNumber(&(xlrec->newtid)),
+ appendStringInfo(buf, "; new off %u xmax %u",
+ xlrec->new_offnum,
xlrec->new_xmax);
}
else if (info == XLOG_HEAP_LOCK)
xl_heap_lock *xlrec = (xl_heap_lock *) rec;
appendStringInfo(buf, "xid %u: ", xlrec->locking_xid);
- out_target(buf, &(xlrec->target));
- appendStringInfoChar(buf, ' ');
+ appendStringInfo(buf, "off %u ", xlrec->offnum);
out_infobits(buf, xlrec->infobits_set);
}
else if (info == XLOG_HEAP_INPLACE)
{
xl_heap_inplace *xlrec = (xl_heap_inplace *) rec;
- out_target(buf, &(xlrec->target));
+ appendStringInfo(buf, "off %u", xlrec->offnum);
}
}
void
-heap2_desc(StringInfo buf, XLogRecord *record)
+heap2_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
info &= XLOG_HEAP_OPMASK;
if (info == XLOG_HEAP2_CLEAN)
{
xl_heap_clean *xlrec = (xl_heap_clean *) rec;
- appendStringInfo(buf, "rel %u/%u/%u; blk %u remxid %u",
- xlrec->node.spcNode, xlrec->node.dbNode,
- xlrec->node.relNode, xlrec->block,
- xlrec->latestRemovedXid);
+ appendStringInfo(buf, "remxid %u", xlrec->latestRemovedXid);
}
else if (info == XLOG_HEAP2_FREEZE_PAGE)
{
xl_heap_freeze_page *xlrec = (xl_heap_freeze_page *) rec;
- appendStringInfo(buf, "rel %u/%u/%u; blk %u; cutoff xid %u ntuples %u",
- xlrec->node.spcNode, xlrec->node.dbNode,
- xlrec->node.relNode, xlrec->block,
+ appendStringInfo(buf, "cutoff xid %u ntuples %u",
xlrec->cutoff_xid, xlrec->ntuples);
}
else if (info == XLOG_HEAP2_CLEANUP_INFO)
{
xl_heap_visible *xlrec = (xl_heap_visible *) rec;
- appendStringInfo(buf, "rel %u/%u/%u; blk %u",
- xlrec->node.spcNode, xlrec->node.dbNode,
- xlrec->node.relNode, xlrec->block);
+ appendStringInfo(buf, "cutoff xid %u", xlrec->cutoff_xid);
}
else if (info == XLOG_HEAP2_MULTI_INSERT)
{
xl_heap_multi_insert *xlrec = (xl_heap_multi_insert *) rec;
- appendStringInfo(buf, "rel %u/%u/%u; blk %u; %d tuples",
- xlrec->node.spcNode, xlrec->node.dbNode, xlrec->node.relNode,
- xlrec->blkno, xlrec->ntuples);
+ appendStringInfo(buf, "%d tuples", xlrec->ntuples);
}
else if (info == XLOG_HEAP2_LOCK_UPDATED)
{
appendStringInfo(buf, "xmax %u msk %04x; ", xlrec->xmax,
xlrec->infobits_set);
- out_target(buf, &(xlrec->target));
+ appendStringInfo(buf, "off %u", xlrec->offnum);
}
else if (info == XLOG_HEAP2_NEW_CID)
{
xl_heap_new_cid *xlrec = (xl_heap_new_cid *) rec;
- out_target(buf, &(xlrec->target));
+ appendStringInfo(buf, "rel %u/%u/%u; tid %u/%u",
+ xlrec->target_node.spcNode,
+ xlrec->target_node.dbNode,
+ xlrec->target_node.relNode,
+ ItemPointerGetBlockNumber(&(xlrec->target_tid)),
+ ItemPointerGetOffsetNumber(&(xlrec->target_tid)));
appendStringInfo(buf, "; cmin: %u, cmax: %u, combo: %u",
xlrec->cmin, xlrec->cmax, xlrec->combocid);
}
}
void
-multixact_desc(StringInfo buf, XLogRecord *record)
+multixact_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
if (info == XLOG_MULTIXACT_ZERO_OFF_PAGE ||
info == XLOG_MULTIXACT_ZERO_MEM_PAGE)
#include "access/nbtree.h"
-static void
-out_target(StringInfo buf, xl_btreetid *target)
-{
- appendStringInfo(buf, "rel %u/%u/%u; tid %u/%u",
- target->node.spcNode, target->node.dbNode, target->node.relNode,
- ItemPointerGetBlockNumber(&(target->tid)),
- ItemPointerGetOffsetNumber(&(target->tid)));
-}
-
void
-btree_desc(StringInfo buf, XLogRecord *record)
+btree_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
switch (info)
{
{
xl_btree_insert *xlrec = (xl_btree_insert *) rec;
- out_target(buf, &(xlrec->target));
+ appendStringInfo(buf, "off %u", xlrec->offnum);
break;
}
case XLOG_BTREE_SPLIT_L:
{
xl_btree_split *xlrec = (xl_btree_split *) rec;
- appendStringInfo(buf, "rel %u/%u/%u ",
- xlrec->node.spcNode, xlrec->node.dbNode,
- xlrec->node.relNode);
- appendStringInfo(buf, "left %u, right %u, next %u, level %u, firstright %d",
- xlrec->leftsib, xlrec->rightsib, xlrec->rnext,
+ appendStringInfo(buf, "level %u, firstright %d",
xlrec->level, xlrec->firstright);
break;
}
{
xl_btree_vacuum *xlrec = (xl_btree_vacuum *) rec;
- appendStringInfo(buf, "rel %u/%u/%u; blk %u, lastBlockVacuumed %u",
- xlrec->node.spcNode, xlrec->node.dbNode,
- xlrec->node.relNode, xlrec->block,
+ appendStringInfo(buf, "lastBlockVacuumed %u",
xlrec->lastBlockVacuumed);
break;
}
{
xl_btree_delete *xlrec = (xl_btree_delete *) rec;
- appendStringInfo(buf, "index %u/%u/%u; iblk %u, heap %u/%u/%u;",
- xlrec->node.spcNode, xlrec->node.dbNode, xlrec->node.relNode,
- xlrec->block,
- xlrec->hnode.spcNode, xlrec->hnode.dbNode, xlrec->hnode.relNode);
+ appendStringInfo(buf, "%d items", xlrec->nitems);
break;
}
case XLOG_BTREE_MARK_PAGE_HALFDEAD:
{
xl_btree_mark_page_halfdead *xlrec = (xl_btree_mark_page_halfdead *) rec;
- out_target(buf, &(xlrec->target));
- appendStringInfo(buf, "; topparent %u; leaf %u; left %u; right %u",
+ appendStringInfo(buf, "topparent %u; leaf %u; left %u; right %u",
xlrec->topparent, xlrec->leafblk, xlrec->leftblk, xlrec->rightblk);
break;
}
{
xl_btree_unlink_page *xlrec = (xl_btree_unlink_page *) rec;
- appendStringInfo(buf, "rel %u/%u/%u; ",
- xlrec->node.spcNode, xlrec->node.dbNode, xlrec->node.relNode);
- appendStringInfo(buf, "dead %u; left %u; right %u; btpo_xact %u; ",
- xlrec->deadblk, xlrec->leftsib, xlrec->rightsib, xlrec->btpo_xact);
- appendStringInfo(buf, "leaf %u; leafleft %u; leafright %u; topparent %u",
- xlrec->leafblk, xlrec->leafleftsib, xlrec->leafrightsib, xlrec->topparent);
+ appendStringInfo(buf, "left %u; right %u; btpo_xact %u; ",
+ xlrec->leftsib, xlrec->rightsib,
+ xlrec->btpo_xact);
+ appendStringInfo(buf, "leafleft %u; leafright %u; topparent %u",
+ xlrec->leafleftsib, xlrec->leafrightsib,
+ xlrec->topparent);
break;
}
case XLOG_BTREE_NEWROOT:
{
xl_btree_newroot *xlrec = (xl_btree_newroot *) rec;
- appendStringInfo(buf, "rel %u/%u/%u; root %u lev %u",
- xlrec->node.spcNode, xlrec->node.dbNode,
- xlrec->node.relNode,
- xlrec->rootblk, xlrec->level);
+ appendStringInfo(buf, "lev %u", xlrec->level);
break;
}
case XLOG_BTREE_REUSE_PAGE:
appendStringInfo(buf, "rel %u/%u/%u; latestRemovedXid %u",
xlrec->node.spcNode, xlrec->node.dbNode,
- xlrec->node.relNode, xlrec->latestRemovedXid);
+ xlrec->node.relNode, xlrec->latestRemovedXid);
break;
}
}
#include "utils/relmapper.h"
void
-relmap_desc(StringInfo buf, XLogRecord *record)
+relmap_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
if (info == XLOG_RELMAP_UPDATE)
{
void
-seq_desc(StringInfo buf, XLogRecord *record)
+seq_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
xl_seq_rec *xlrec = (xl_seq_rec *) rec;
if (info == XLOG_SEQ_LOG)
void
-smgr_desc(StringInfo buf, XLogRecord *record)
+smgr_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
if (info == XLOG_SMGR_CREATE)
{
#include "access/spgist_private.h"
-static void
-out_target(StringInfo buf, RelFileNode node)
-{
- appendStringInfo(buf, "rel %u/%u/%u ",
- node.spcNode, node.dbNode, node.relNode);
-}
-
void
-spg_desc(StringInfo buf, XLogRecord *record)
+spg_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
switch (info)
{
case XLOG_SPGIST_CREATE_INDEX:
- appendStringInfo(buf, "rel %u/%u/%u",
- ((RelFileNode *) rec)->spcNode,
- ((RelFileNode *) rec)->dbNode,
- ((RelFileNode *) rec)->relNode);
break;
case XLOG_SPGIST_ADD_LEAF:
- out_target(buf, ((spgxlogAddLeaf *) rec)->node);
- appendStringInfo(buf, "%u",
- ((spgxlogAddLeaf *) rec)->blknoLeaf);
+ {
+ spgxlogAddLeaf *xlrec = (spgxlogAddLeaf *) rec;
+
+ appendStringInfo(buf, "add leaf to page");
+ appendStringInfo(buf, "; off %u; headoff %u; parentoff %u",
+ xlrec->offnumLeaf, xlrec->offnumHeadLeaf,
+ xlrec->offnumParent);
+ if (xlrec->newPage)
+ appendStringInfo(buf, " (newpage)");
+ if (xlrec->storesNulls)
+ appendStringInfo(buf, " (nulls)");
+ }
break;
case XLOG_SPGIST_MOVE_LEAFS:
- out_target(buf, ((spgxlogMoveLeafs *) rec)->node);
- appendStringInfo(buf, "%u leafs from page %u to page %u",
- ((spgxlogMoveLeafs *) rec)->nMoves,
- ((spgxlogMoveLeafs *) rec)->blknoSrc,
- ((spgxlogMoveLeafs *) rec)->blknoDst);
+ appendStringInfo(buf, "%u leafs",
+ ((spgxlogMoveLeafs *) rec)->nMoves);
break;
case XLOG_SPGIST_ADD_NODE:
- out_target(buf, ((spgxlogAddNode *) rec)->node);
- appendStringInfo(buf, "%u:%u",
- ((spgxlogAddNode *) rec)->blkno,
+ appendStringInfo(buf, "off %u",
((spgxlogAddNode *) rec)->offnum);
break;
case XLOG_SPGIST_SPLIT_TUPLE:
- out_target(buf, ((spgxlogSplitTuple *) rec)->node);
- appendStringInfo(buf, "%u:%u to %u:%u",
- ((spgxlogSplitTuple *) rec)->blknoPrefix,
+ appendStringInfo(buf, "prefix off: %u, postfix off: %u (same %d, new %d)",
((spgxlogSplitTuple *) rec)->offnumPrefix,
- ((spgxlogSplitTuple *) rec)->blknoPostfix,
- ((spgxlogSplitTuple *) rec)->offnumPostfix);
+ ((spgxlogSplitTuple *) rec)->offnumPostfix,
+ ((spgxlogSplitTuple *) rec)->postfixBlkSame,
+ ((spgxlogSplitTuple *) rec)->newPage
+ );
break;
case XLOG_SPGIST_PICKSPLIT:
- out_target(buf, ((spgxlogPickSplit *) rec)->node);
+ {
+ spgxlogPickSplit *xlrec = (spgxlogPickSplit *) rec;
+
+ appendStringInfo(buf, "ndel %u; nins %u",
+ xlrec->nDelete, xlrec->nInsert);
+ if (xlrec->innerIsParent)
+ appendStringInfo(buf, " (innerIsParent)");
+ if (xlrec->isRootSplit)
+ appendStringInfo(buf, " (isRootSplit)");
+ }
break;
case XLOG_SPGIST_VACUUM_LEAF:
- out_target(buf, ((spgxlogVacuumLeaf *) rec)->node);
- appendStringInfo(buf, "page %u",
- ((spgxlogVacuumLeaf *) rec)->blkno);
+ /* no further information */
break;
case XLOG_SPGIST_VACUUM_ROOT:
- out_target(buf, ((spgxlogVacuumRoot *) rec)->node);
- appendStringInfo(buf, "page %u",
- ((spgxlogVacuumRoot *) rec)->blkno);
+ /* no further information */
break;
case XLOG_SPGIST_VACUUM_REDIRECT:
- out_target(buf, ((spgxlogVacuumRedirect *) rec)->node);
- appendStringInfo(buf, "page %u, newest XID %u",
- ((spgxlogVacuumRedirect *) rec)->blkno,
+ appendStringInfo(buf, "newest XID %u",
((spgxlogVacuumRedirect *) rec)->newestRedirectXid);
break;
}
}
void
-standby_desc(StringInfo buf, XLogRecord *record)
+standby_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
if (info == XLOG_STANDBY_LOCK)
{
void
-tblspc_desc(StringInfo buf, XLogRecord *record)
+tblspc_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
if (info == XLOG_TBLSPC_CREATE)
{
}
void
-xact_desc(StringInfo buf, XLogRecord *record)
+xact_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
if (info == XLOG_XACT_COMMIT_COMPACT)
{
};
void
-xlog_desc(StringInfo buf, XLogRecord *record)
+xlog_desc(StringInfo buf, XLogReaderState *record)
{
char *rec = XLogRecGetData(record);
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
if (info == XLOG_CHECKPOINT_SHUTDOWN ||
info == XLOG_CHECKPOINT_ONLINE)
}
else if (info == XLOG_FPI)
{
- BkpBlock *bkp = (BkpBlock *) rec;
-
- appendStringInfo(buf, "%s block %u",
- relpathperm(bkp->node, bkp->fork),
- bkp->block);
+ /* no further information to print */
}
else if (info == XLOG_BACKUP_END)
{
#include "postgres.h"
#include "access/genam.h"
-#include "access/xloginsert.h"
#include "access/spgist_private.h"
+#include "access/xloginsert.h"
#include "miscadmin.h"
#include "storage/bufmgr.h"
#include "utils/rel.h"
addLeafTuple(Relation index, SpGistState *state, SpGistLeafTuple leafTuple,
SPPageDesc *current, SPPageDesc *parent, bool isNulls, bool isNew)
{
- XLogRecData rdata[4];
spgxlogAddLeaf xlrec;
- xlrec.node = index->rd_node;
- xlrec.blknoLeaf = current->blkno;
xlrec.newPage = isNew;
xlrec.storesNulls = isNulls;
/* these will be filled below as needed */
xlrec.offnumLeaf = InvalidOffsetNumber;
xlrec.offnumHeadLeaf = InvalidOffsetNumber;
- xlrec.blknoParent = InvalidBlockNumber;
xlrec.offnumParent = InvalidOffsetNumber;
xlrec.nodeI = 0;
- ACCEPT_RDATA_DATA(&xlrec, sizeof(xlrec), 0);
- ACCEPT_RDATA_DATA(leafTuple, leafTuple->size, 1);
- ACCEPT_RDATA_BUFFER(current->buffer, 2);
-
START_CRIT_SECTION();
if (current->offnum == InvalidOffsetNumber ||
/* Must update parent's downlink if any */
if (parent->buffer != InvalidBuffer)
{
- xlrec.blknoParent = parent->blkno;
xlrec.offnumParent = parent->offnum;
xlrec.nodeI = parent->node;
saveNodeLink(index, parent, current->blkno, current->offnum);
-
- ACCEPT_RDATA_BUFFER(parent->buffer, 3);
}
}
else
{
XLogRecPtr recptr;
- recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_ADD_LEAF, rdata);
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, sizeof(xlrec));
+ XLogRegisterData((char *) leafTuple, leafTuple->size);
+
+ XLogRegisterBuffer(0, current->buffer, REGBUF_STANDARD);
+ if (xlrec.offnumParent != InvalidOffsetNumber)
+ XLogRegisterBuffer(1, parent->buffer, REGBUF_STANDARD);
+
+ recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_ADD_LEAF);
PageSetLSN(current->page, recptr);
/* update parent only if we actually changed it */
- if (xlrec.blknoParent != InvalidBlockNumber)
+ if (xlrec.offnumParent != InvalidOffsetNumber)
{
PageSetLSN(parent->page, recptr);
}
OffsetNumber *toDelete;
OffsetNumber *toInsert;
BlockNumber nblkno;
- XLogRecData rdata[7];
spgxlogMoveLeafs xlrec;
char *leafdata,
*leafptr;
nblkno = BufferGetBlockNumber(nbuf);
Assert(nblkno != current->blkno);
- /* prepare WAL info */
- xlrec.node = index->rd_node;
- STORE_STATE(state, xlrec.stateSrc);
-
- xlrec.blknoSrc = current->blkno;
- xlrec.blknoDst = nblkno;
- xlrec.nMoves = nDelete;
- xlrec.replaceDead = replaceDead;
- xlrec.storesNulls = isNulls;
-
- xlrec.blknoParent = parent->blkno;
- xlrec.offnumParent = parent->offnum;
- xlrec.nodeI = parent->node;
-
leafdata = leafptr = palloc(size);
START_CRIT_SECTION();
{
XLogRecPtr recptr;
- ACCEPT_RDATA_DATA(&xlrec, SizeOfSpgxlogMoveLeafs, 0);
- ACCEPT_RDATA_DATA(toDelete, sizeof(OffsetNumber) * nDelete, 1);
- ACCEPT_RDATA_DATA(toInsert, sizeof(OffsetNumber) * nInsert, 2);
- ACCEPT_RDATA_DATA(leafdata, leafptr - leafdata, 3);
- ACCEPT_RDATA_BUFFER(current->buffer, 4);
- ACCEPT_RDATA_BUFFER(nbuf, 5);
- ACCEPT_RDATA_BUFFER(parent->buffer, 6);
+ /* prepare WAL info */
+ STORE_STATE(state, xlrec.stateSrc);
- recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_MOVE_LEAFS, rdata);
+ xlrec.nMoves = nDelete;
+ xlrec.replaceDead = replaceDead;
+ xlrec.storesNulls = isNulls;
+
+ xlrec.offnumParent = parent->offnum;
+ xlrec.nodeI = parent->node;
+
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, SizeOfSpgxlogMoveLeafs);
+ XLogRegisterData((char *) toDelete,
+ sizeof(OffsetNumber) * nDelete);
+ XLogRegisterData((char *) toInsert,
+ sizeof(OffsetNumber) * nInsert);
+ XLogRegisterData((char *) leafdata, leafptr - leafdata);
+
+ XLogRegisterBuffer(0, current->buffer, REGBUF_STANDARD);
+ XLogRegisterBuffer(1, nbuf, REGBUF_STANDARD | (xlrec.newPage ? REGBUF_WILL_INIT : 0));
+ XLogRegisterBuffer(2, parent->buffer, REGBUF_STANDARD);
+
+ recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_MOVE_LEAFS);
PageSetLSN(current->page, recptr);
PageSetLSN(npage, recptr);
int currentFreeSpace;
int totalLeafSizes;
bool allTheSame;
- XLogRecData rdata[10];
- int nRdata;
spgxlogPickSplit xlrec;
char *leafdata,
*leafptr;
newLeafs = (SpGistLeafTuple *) palloc(sizeof(SpGistLeafTuple) * n);
leafPageSelect = (uint8 *) palloc(sizeof(uint8) * n);
- xlrec.node = index->rd_node;
STORE_STATE(state, xlrec.stateSrc);
/*
}
/*
- * Because a WAL record can't involve more than four buffers, we can only
- * afford to deal with two leaf pages in each picksplit action, ie the
- * current page and at most one other.
- *
* The new leaf tuples converted from the existing ones should require the
* same or less space, and therefore should all fit onto one page
* (although that's not necessarily the current page, since we can't
}
/* Start preparing WAL record */
- xlrec.blknoSrc = current->blkno;
- xlrec.blknoDest = InvalidBlockNumber;
xlrec.nDelete = 0;
xlrec.initSrc = isNew;
xlrec.storesNulls = isNulls;
+ xlrec.isRootSplit = SpGistBlockIsRoot(current->blkno);
leafdata = leafptr = (char *) palloc(totalLeafSizes);
- ACCEPT_RDATA_DATA(&xlrec, SizeOfSpgxlogPickSplit, 0);
- nRdata = 1;
-
/* Here we begin making the changes to the target pages */
START_CRIT_SECTION();
else
{
xlrec.nDelete = nToDelete;
- ACCEPT_RDATA_DATA(toDelete,
- sizeof(OffsetNumber) * nToDelete,
- nRdata);
- nRdata++;
- ACCEPT_RDATA_BUFFER(current->buffer, nRdata);
- nRdata++;
if (!state->isBuild)
{
if (newLeafBuffer != InvalidBuffer)
{
MarkBufferDirty(newLeafBuffer);
- /* also save block number for WAL */
- xlrec.blknoDest = BufferGetBlockNumber(newLeafBuffer);
- if (!xlrec.initDest)
- {
- ACCEPT_RDATA_BUFFER(newLeafBuffer, nRdata);
- nRdata++;
- }
}
- xlrec.nInsert = nToInsert;
- ACCEPT_RDATA_DATA(toInsert, sizeof(OffsetNumber) * nToInsert, nRdata);
- nRdata++;
- ACCEPT_RDATA_DATA(leafPageSelect, sizeof(uint8) * nToInsert, nRdata);
- nRdata++;
- ACCEPT_RDATA_DATA(innerTuple, innerTuple->size, nRdata);
- nRdata++;
- ACCEPT_RDATA_DATA(leafdata, leafptr - leafdata, nRdata);
- nRdata++;
-
/* Remember current buffer, since we're about to change "current" */
saveCurrent = *current;
current->blkno = parent->blkno;
current->buffer = parent->buffer;
current->page = parent->page;
- xlrec.blknoInner = current->blkno;
xlrec.offnumInner = current->offnum =
SpGistPageAddNewItem(state, current->page,
(Item) innerTuple, innerTuple->size,
/*
* Update parent node link and mark parent page dirty
*/
- xlrec.blknoParent = parent->blkno;
+ xlrec.innerIsParent = true;
xlrec.offnumParent = parent->offnum;
xlrec.nodeI = parent->node;
saveNodeLink(index, parent, current->blkno, current->offnum);
- ACCEPT_RDATA_BUFFER(parent->buffer, nRdata);
- nRdata++;
-
/*
* Update redirection link (in old current buffer)
*/
current->buffer = newInnerBuffer;
current->blkno = BufferGetBlockNumber(current->buffer);
current->page = BufferGetPage(current->buffer);
- xlrec.blknoInner = current->blkno;
xlrec.offnumInner = current->offnum =
SpGistPageAddNewItem(state, current->page,
(Item) innerTuple, innerTuple->size,
/*
* Update parent node link and mark parent page dirty
*/
- xlrec.blknoParent = parent->blkno;
+ xlrec.innerIsParent = (parent->buffer == current->buffer);
xlrec.offnumParent = parent->offnum;
xlrec.nodeI = parent->node;
saveNodeLink(index, parent, current->blkno, current->offnum);
- ACCEPT_RDATA_BUFFER(current->buffer, nRdata);
- nRdata++;
- ACCEPT_RDATA_BUFFER(parent->buffer, nRdata);
- nRdata++;
-
/*
* Update redirection link (in old current buffer)
*/
SpGistInitBuffer(current->buffer, (isNulls ? SPGIST_NULLS : 0));
xlrec.initInner = true;
+ xlrec.innerIsParent = false;
- xlrec.blknoInner = current->blkno;
xlrec.offnumInner = current->offnum =
PageAddItem(current->page, (Item) innerTuple, innerTuple->size,
InvalidOffsetNumber, false, false);
innerTuple->size);
/* No parent link to update, nor redirection to do */
- xlrec.blknoParent = InvalidBlockNumber;
xlrec.offnumParent = InvalidOffsetNumber;
xlrec.nodeI = 0;
if (RelationNeedsWAL(index))
{
XLogRecPtr recptr;
+ int flags;
+
+ XLogBeginInsert();
+
+ xlrec.nInsert = nToInsert;
+ XLogRegisterData((char *) &xlrec, SizeOfSpgxlogPickSplit);
+
+ XLogRegisterData((char *) toDelete,
+ sizeof(OffsetNumber) * xlrec.nDelete);
+ XLogRegisterData((char *) toInsert,
+ sizeof(OffsetNumber) * xlrec.nInsert);
+ XLogRegisterData((char *) leafPageSelect,
+ sizeof(uint8) * xlrec.nInsert);
+ XLogRegisterData((char *) innerTuple, innerTuple->size);
+ XLogRegisterData(leafdata, leafptr - leafdata);
+
+ flags = REGBUF_STANDARD;
+ if (xlrec.initSrc)
+ flags |= REGBUF_WILL_INIT;
+ if (BufferIsValid(saveCurrent.buffer))
+ XLogRegisterBuffer(0, saveCurrent.buffer, flags);
+
+ if (BufferIsValid(newLeafBuffer))
+ {
+ flags = REGBUF_STANDARD;
+ if (xlrec.initDest)
+ flags |= REGBUF_WILL_INIT;
+ XLogRegisterBuffer(1, newLeafBuffer, flags);
+ }
+ XLogRegisterBuffer(2, current->buffer, REGBUF_STANDARD);
+ if (parent->buffer != InvalidBuffer)
+ {
+ if (parent->buffer != current->buffer)
+ XLogRegisterBuffer(3, parent->buffer, REGBUF_STANDARD);
+ else
+ Assert(xlrec.innerIsParent);
+ }
/* Issue the WAL record */
- recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_PICKSPLIT, rdata);
+ recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_PICKSPLIT);
/* Update page LSNs on all affected pages */
if (newLeafBuffer != InvalidBuffer)
int nodeN, Datum nodeLabel)
{
SpGistInnerTuple newInnerTuple;
- XLogRecData rdata[5];
spgxlogAddNode xlrec;
/* Should not be applied to nulls */
newInnerTuple = addNode(state, innerTuple, nodeLabel, nodeN);
/* Prepare WAL record */
- xlrec.node = index->rd_node;
STORE_STATE(state, xlrec.stateSrc);
- xlrec.blkno = current->blkno;
xlrec.offnum = current->offnum;
/* we don't fill these unless we need to change the parent downlink */
- xlrec.blknoParent = InvalidBlockNumber;
+ xlrec.parentBlk = -1;
xlrec.offnumParent = InvalidOffsetNumber;
xlrec.nodeI = 0;
/* we don't fill these unless tuple has to be moved */
- xlrec.blknoNew = InvalidBlockNumber;
xlrec.offnumNew = InvalidOffsetNumber;
xlrec.newPage = false;
- ACCEPT_RDATA_DATA(&xlrec, sizeof(xlrec), 0);
- ACCEPT_RDATA_DATA(newInnerTuple, newInnerTuple->size, 1);
- ACCEPT_RDATA_BUFFER(current->buffer, 2);
-
if (PageGetExactFreeSpace(current->page) >=
newInnerTuple->size - innerTuple->size)
{
{
XLogRecPtr recptr;
- recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_ADD_NODE, rdata);
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, sizeof(xlrec));
+ XLogRegisterData((char *) newInnerTuple, newInnerTuple->size);
+
+ XLogRegisterBuffer(0, current->buffer, REGBUF_STANDARD);
+
+ recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_ADD_NODE);
PageSetLSN(current->page, recptr);
}
saveCurrent = *current;
- xlrec.blknoParent = parent->blkno;
xlrec.offnumParent = parent->offnum;
xlrec.nodeI = parent->node;
current->blkno = BufferGetBlockNumber(current->buffer);
current->page = BufferGetPage(current->buffer);
- xlrec.blknoNew = current->blkno;
-
/*
* Let's just make real sure new current isn't same as old. Right now
* that's impossible, but if SpGistGetBuffer ever got smart enough to
* replay would be subtly wrong, so I think a mere assert isn't enough
* here.
*/
- if (xlrec.blknoNew == xlrec.blkno)
+ if (current->blkno == saveCurrent.blkno)
elog(ERROR, "SPGiST new buffer shouldn't be same as old buffer");
/*
* New current and parent buffer will both be modified; but note that
* parent buffer could be same as either new or old current.
*/
- ACCEPT_RDATA_BUFFER(current->buffer, 3);
- if (parent->buffer != current->buffer &&
- parent->buffer != saveCurrent.buffer)
- ACCEPT_RDATA_BUFFER(parent->buffer, 4);
+ if (parent->buffer == saveCurrent.buffer)
+ xlrec.parentBlk = 0;
+ else if (parent->buffer == current->buffer)
+ xlrec.parentBlk = 1;
+ else
+ xlrec.parentBlk = 2;
START_CRIT_SECTION();
{
XLogRecPtr recptr;
- recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_ADD_NODE, rdata);
+ XLogBeginInsert();
+
+ /* orig page */
+ XLogRegisterBuffer(0, saveCurrent.buffer, REGBUF_STANDARD);
+ /* new page */
+ XLogRegisterBuffer(1, current->buffer, REGBUF_STANDARD);
+ /* parent page (if different from orig and new) */
+ if (xlrec.parentBlk == 2)
+ XLogRegisterBuffer(2, parent->buffer, REGBUF_STANDARD);
+
+ XLogRegisterData((char *) &xlrec, sizeof(xlrec));
+ XLogRegisterData((char *) newInnerTuple, newInnerTuple->size);
+
+ recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_ADD_NODE);
/* we don't bother to check if any of these are redundant */
PageSetLSN(current->page, recptr);
BlockNumber postfixBlkno;
OffsetNumber postfixOffset;
int i;
- XLogRecData rdata[5];
spgxlogSplitTuple xlrec;
Buffer newBuffer = InvalidBuffer;
postfixTuple->allTheSame = innerTuple->allTheSame;
/* prep data for WAL record */
- xlrec.node = index->rd_node;
xlrec.newPage = false;
- ACCEPT_RDATA_DATA(&xlrec, sizeof(xlrec), 0);
- ACCEPT_RDATA_DATA(prefixTuple, prefixTuple->size, 1);
- ACCEPT_RDATA_DATA(postfixTuple, postfixTuple->size, 2);
- ACCEPT_RDATA_BUFFER(current->buffer, 3);
-
/*
* If we can't fit both tuples on the current page, get a new page for the
* postfix tuple. In particular, can't split to the root page.
GBUF_INNER_PARITY(current->blkno + 1),
postfixTuple->size + sizeof(ItemIdData),
&xlrec.newPage);
- ACCEPT_RDATA_BUFFER(newBuffer, 4);
}
START_CRIT_SECTION();
if (xlrec.offnumPrefix != current->offnum)
elog(ERROR, "failed to add item of size %u to SPGiST index page",
prefixTuple->size);
- xlrec.blknoPrefix = current->blkno;
/*
* put postfix tuple into appropriate page
*/
if (newBuffer == InvalidBuffer)
{
- xlrec.blknoPostfix = postfixBlkno = current->blkno;
+ postfixBlkno = current->blkno;
xlrec.offnumPostfix = postfixOffset =
SpGistPageAddNewItem(state, current->page,
(Item) postfixTuple, postfixTuple->size,
NULL, false);
+ xlrec.postfixBlkSame = true;
}
else
{
- xlrec.blknoPostfix = postfixBlkno = BufferGetBlockNumber(newBuffer);
+ postfixBlkno = BufferGetBlockNumber(newBuffer);
xlrec.offnumPostfix = postfixOffset =
SpGistPageAddNewItem(state, BufferGetPage(newBuffer),
(Item) postfixTuple, postfixTuple->size,
NULL, false);
MarkBufferDirty(newBuffer);
+ xlrec.postfixBlkSame = false;
}
/*
{
XLogRecPtr recptr;
- recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_SPLIT_TUPLE, rdata);
+ XLogBeginInsert();
+ XLogRegisterData((char *) &xlrec, sizeof(xlrec));
+ XLogRegisterData((char *) prefixTuple, prefixTuple->size);
+ XLogRegisterData((char *) postfixTuple, postfixTuple->size);
+
+ XLogRegisterBuffer(0, current->buffer, REGBUF_STANDARD);
+ if (newBuffer != InvalidBuffer)
+ {
+ int flags;
+
+ flags = REGBUF_STANDARD;
+ if (xlrec.newPage)
+ flags |= REGBUF_WILL_INIT;
+ XLogRegisterBuffer(1, newBuffer, flags);
+ }
+
+ recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_SPLIT_TUPLE);
PageSetLSN(current->page, recptr);
if (RelationNeedsWAL(index))
{
XLogRecPtr recptr;
- XLogRecData rdata;
- /* WAL data is just the relfilenode */
- rdata.data = (char *) &(index->rd_node);
- rdata.len = sizeof(RelFileNode);
- rdata.buffer = InvalidBuffer;
- rdata.next = NULL;
+ XLogBeginInsert();
- recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_CREATE_INDEX, &rdata);
+ /*
+ * Replay will re-initialize the pages, so don't take full pages
+ * images. No other data to log.
+ */
+ XLogRegisterBuffer(0, metabuffer, REGBUF_WILL_INIT);
+ XLogRegisterBuffer(1, rootbuffer, REGBUF_WILL_INIT | REGBUF_STANDARD);
+ XLogRegisterBuffer(2, nullbuffer, REGBUF_WILL_INIT | REGBUF_STANDARD);
+
+ recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_CREATE_INDEX);
PageSetLSN(BufferGetPage(metabuffer), recptr);
PageSetLSN(BufferGetPage(rootbuffer), recptr);
{
Page page = BufferGetPage(buffer);
spgxlogVacuumLeaf xlrec;
- XLogRecData rdata[8];
OffsetNumber toDead[MaxIndexTuplesPerPage];
OffsetNumber toPlaceholder[MaxIndexTuplesPerPage];
OffsetNumber moveSrc[MaxIndexTuplesPerPage];
if (nDeletable != xlrec.nDead + xlrec.nPlaceholder + xlrec.nMove)
elog(ERROR, "inconsistent counts of deletable tuples");
- /* Prepare WAL record */
- xlrec.node = index->rd_node;
- xlrec.blkno = BufferGetBlockNumber(buffer);
- STORE_STATE(&bds->spgstate, xlrec.stateSrc);
-
- ACCEPT_RDATA_DATA(&xlrec, SizeOfSpgxlogVacuumLeaf, 0);
- ACCEPT_RDATA_DATA(toDead, sizeof(OffsetNumber) * xlrec.nDead, 1);
- ACCEPT_RDATA_DATA(toPlaceholder, sizeof(OffsetNumber) * xlrec.nPlaceholder, 2);
- ACCEPT_RDATA_DATA(moveSrc, sizeof(OffsetNumber) * xlrec.nMove, 3);
- ACCEPT_RDATA_DATA(moveDest, sizeof(OffsetNumber) * xlrec.nMove, 4);
- ACCEPT_RDATA_DATA(chainSrc, sizeof(OffsetNumber) * xlrec.nChain, 5);
- ACCEPT_RDATA_DATA(chainDest, sizeof(OffsetNumber) * xlrec.nChain, 6);
- ACCEPT_RDATA_BUFFER(buffer, 7);
-
/* Do the updates */
START_CRIT_SECTION();
{
XLogRecPtr recptr;
- recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_VACUUM_LEAF, rdata);
+ XLogBeginInsert();
+
+ STORE_STATE(&bds->spgstate, xlrec.stateSrc);
+
+ XLogRegisterData((char *) &xlrec, SizeOfSpgxlogVacuumLeaf);
+ /* sizeof(xlrec) should be a multiple of sizeof(OffsetNumber) */
+ XLogRegisterData((char *) toDead, sizeof(OffsetNumber) * xlrec.nDead);
+ XLogRegisterData((char *) toPlaceholder, sizeof(OffsetNumber) * xlrec.nPlaceholder);
+ XLogRegisterData((char *) moveSrc, sizeof(OffsetNumber) * xlrec.nMove);
+ XLogRegisterData((char *) moveDest, sizeof(OffsetNumber) * xlrec.nMove);
+ XLogRegisterData((char *) chainSrc, sizeof(OffsetNumber) * xlrec.nChain);
+ XLogRegisterData((char *) chainDest, sizeof(OffsetNumber) * xlrec.nChain);
+
+ XLogRegisterBuffer(0, buffer, REGBUF_STANDARD);
+
+ recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_VACUUM_LEAF);
PageSetLSN(page, recptr);
}
{
Page page = BufferGetPage(buffer);
spgxlogVacuumRoot xlrec;
- XLogRecData rdata[3];
OffsetNumber toDelete[MaxIndexTuplesPerPage];
OffsetNumber i,
max = PageGetMaxOffsetNumber(page);
- xlrec.blkno = BufferGetBlockNumber(buffer);
xlrec.nDelete = 0;
/* Scan page, identify tuples to delete, accumulate stats */
if (xlrec.nDelete == 0)
return; /* nothing more to do */
- /* Prepare WAL record */
- xlrec.node = index->rd_node;
- STORE_STATE(&bds->spgstate, xlrec.stateSrc);
-
- ACCEPT_RDATA_DATA(&xlrec, SizeOfSpgxlogVacuumRoot, 0);
- /* sizeof(xlrec) should be a multiple of sizeof(OffsetNumber) */
- ACCEPT_RDATA_DATA(toDelete, sizeof(OffsetNumber) * xlrec.nDelete, 1);
- ACCEPT_RDATA_BUFFER(buffer, 2);
-
/* Do the update */
START_CRIT_SECTION();
{
XLogRecPtr recptr;
- recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_VACUUM_ROOT, rdata);
+ XLogBeginInsert();
+
+ /* Prepare WAL record */
+ STORE_STATE(&bds->spgstate, xlrec.stateSrc);
+
+ XLogRegisterData((char *) &xlrec, SizeOfSpgxlogVacuumRoot);
+ /* sizeof(xlrec) should be a multiple of sizeof(OffsetNumber) */
+ XLogRegisterData((char *) toDelete,
+ sizeof(OffsetNumber) * xlrec.nDelete);
+
+ XLogRegisterBuffer(0, buffer, REGBUF_STANDARD);
+
+ recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_VACUUM_ROOT);
PageSetLSN(page, recptr);
}
OffsetNumber itemToPlaceholder[MaxIndexTuplesPerPage];
OffsetNumber itemnos[MaxIndexTuplesPerPage];
spgxlogVacuumRedirect xlrec;
- XLogRecData rdata[3];
- xlrec.node = index->rd_node;
- xlrec.blkno = BufferGetBlockNumber(buffer);
xlrec.nToPlaceholder = 0;
xlrec.newestRedirectXid = InvalidTransactionId;
{
XLogRecPtr recptr;
- ACCEPT_RDATA_DATA(&xlrec, SizeOfSpgxlogVacuumRedirect, 0);
- ACCEPT_RDATA_DATA(itemToPlaceholder, sizeof(OffsetNumber) * xlrec.nToPlaceholder, 1);
- ACCEPT_RDATA_BUFFER(buffer, 2);
+ XLogBeginInsert();
+
+ XLogRegisterData((char *) &xlrec, SizeOfSpgxlogVacuumRedirect);
+ XLogRegisterData((char *) itemToPlaceholder,
+ sizeof(OffsetNumber) * xlrec.nToPlaceholder);
+
+ XLogRegisterBuffer(0, buffer, REGBUF_STANDARD);
- recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_VACUUM_REDIRECT, rdata);
+ recptr = XLogInsert(RM_SPGIST_ID, XLOG_SPGIST_VACUUM_REDIRECT);
PageSetLSN(page, recptr);
}
}
static void
-spgRedoCreateIndex(XLogRecPtr lsn, XLogRecord *record)
+spgRedoCreateIndex(XLogReaderState *record)
{
- RelFileNode *node = (RelFileNode *) XLogRecGetData(record);
+ XLogRecPtr lsn = record->EndRecPtr;
Buffer buffer;
Page page;
- /* Backup blocks are not used in create_index records */
- Assert(!(record->xl_info & XLR_BKP_BLOCK_MASK));
-
- buffer = XLogReadBuffer(*node, SPGIST_METAPAGE_BLKNO, true);
- Assert(BufferIsValid(buffer));
+ buffer = XLogInitBufferForRedo(record, 0);
+ Assert(BufferGetBlockNumber(buffer) == SPGIST_METAPAGE_BLKNO);
page = (Page) BufferGetPage(buffer);
SpGistInitMetapage(page);
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
UnlockReleaseBuffer(buffer);
- buffer = XLogReadBuffer(*node, SPGIST_ROOT_BLKNO, true);
- Assert(BufferIsValid(buffer));
+ buffer = XLogInitBufferForRedo(record, 1);
+ Assert(BufferGetBlockNumber(buffer) == SPGIST_ROOT_BLKNO);
SpGistInitBuffer(buffer, SPGIST_LEAF);
page = (Page) BufferGetPage(buffer);
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
UnlockReleaseBuffer(buffer);
- buffer = XLogReadBuffer(*node, SPGIST_NULL_BLKNO, true);
- Assert(BufferIsValid(buffer));
+ buffer = XLogInitBufferForRedo(record, 2);
+ Assert(BufferGetBlockNumber(buffer) == SPGIST_NULL_BLKNO);
SpGistInitBuffer(buffer, SPGIST_LEAF | SPGIST_NULLS);
page = (Page) BufferGetPage(buffer);
PageSetLSN(page, lsn);
}
static void
-spgRedoAddLeaf(XLogRecPtr lsn, XLogRecord *record)
+spgRedoAddLeaf(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
char *ptr = XLogRecGetData(record);
spgxlogAddLeaf *xldata = (spgxlogAddLeaf *) ptr;
char *leafTuple;
*/
if (xldata->newPage)
{
- buffer = XLogReadBuffer(xldata->node, xldata->blknoLeaf, true);
+ buffer = XLogInitBufferForRedo(record, 0);
SpGistInitBuffer(buffer,
SPGIST_LEAF | (xldata->storesNulls ? SPGIST_NULLS : 0));
action = BLK_NEEDS_REDO;
}
else
- action = XLogReadBufferForRedo(lsn, record, 0,
- xldata->node, xldata->blknoLeaf,
- &buffer);
+ action = XLogReadBufferForRedo(record, 0, &buffer);
if (action == BLK_NEEDS_REDO)
{
{
/* replacing a DEAD tuple */
PageIndexTupleDelete(page, xldata->offnumLeaf);
- if (PageAddItem(page, (Item) leafTuple, leafTupleHdr.size,
+ if (PageAddItem(page,
+ (Item) leafTuple, leafTupleHdr.size,
xldata->offnumLeaf, false, false) != xldata->offnumLeaf)
elog(ERROR, "failed to add item of size %u to SPGiST index page",
leafTupleHdr.size);
UnlockReleaseBuffer(buffer);
/* update parent downlink if necessary */
- if (xldata->blknoParent != InvalidBlockNumber)
+ if (xldata->offnumParent != InvalidOffsetNumber)
{
- if (XLogReadBufferForRedo(lsn, record, 1,
- xldata->node, xldata->blknoParent,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 1, &buffer) == BLK_NEEDS_REDO)
{
SpGistInnerTuple tuple;
+ BlockNumber blknoLeaf;
+
+ XLogRecGetBlockTag(record, 0, NULL, NULL, &blknoLeaf);
page = BufferGetPage(buffer);
PageGetItemId(page, xldata->offnumParent));
spgUpdateNodeLink(tuple, xldata->nodeI,
- xldata->blknoLeaf, xldata->offnumLeaf);
+ blknoLeaf, xldata->offnumLeaf);
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
static void
-spgRedoMoveLeafs(XLogRecPtr lsn, XLogRecord *record)
+spgRedoMoveLeafs(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
char *ptr = XLogRecGetData(record);
spgxlogMoveLeafs *xldata = (spgxlogMoveLeafs *) ptr;
SpGistState state;
Buffer buffer;
Page page;
XLogRedoAction action;
+ BlockNumber blknoDst;
+
+ XLogRecGetBlockTag(record, 1, NULL, NULL, &blknoDst);
fillFakeState(&state, xldata->stateSrc);
/* Insert tuples on the dest page (do first, so redirect is valid) */
if (xldata->newPage)
{
- buffer = XLogReadBuffer(xldata->node, xldata->blknoDst, true);
+ buffer = XLogInitBufferForRedo(record, 1);
SpGistInitBuffer(buffer,
SPGIST_LEAF | (xldata->storesNulls ? SPGIST_NULLS : 0));
action = BLK_NEEDS_REDO;
}
else
- action = XLogReadBufferForRedo(lsn, record, 1,
- xldata->node, xldata->blknoDst,
- &buffer);
+ action = XLogReadBufferForRedo(record, 1, &buffer);
+
if (action == BLK_NEEDS_REDO)
{
int i;
* field.
*/
leafTuple = ptr;
- memcpy(&leafTupleHdr, leafTuple, sizeof(SpGistLeafTupleData));
+ memcpy(&leafTupleHdr, leafTuple,
+ sizeof(SpGistLeafTupleData));
addOrReplaceTuple(page, (Item) leafTuple,
leafTupleHdr.size, toInsert[i]);
UnlockReleaseBuffer(buffer);
/* Delete tuples from the source page, inserting a redirection pointer */
- if (XLogReadBufferForRedo(lsn, record, 0, xldata->node, xldata->blknoSrc,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
page = BufferGetPage(buffer);
+
spgPageIndexMultiDelete(&state, page, toDelete, xldata->nMoves,
state.isBuild ? SPGIST_PLACEHOLDER : SPGIST_REDIRECT,
SPGIST_PLACEHOLDER,
- xldata->blknoDst,
+ blknoDst,
toInsert[nInsert - 1]);
PageSetLSN(page, lsn);
UnlockReleaseBuffer(buffer);
/* And update the parent downlink */
- if (XLogReadBufferForRedo(lsn, record, 2, xldata->node, xldata->blknoParent,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 2, &buffer) == BLK_NEEDS_REDO)
{
SpGistInnerTuple tuple;
PageGetItemId(page, xldata->offnumParent));
spgUpdateNodeLink(tuple, xldata->nodeI,
- xldata->blknoDst, toInsert[nInsert - 1]);
+ blknoDst, toInsert[nInsert - 1]);
PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
static void
-spgRedoAddNode(XLogRecPtr lsn, XLogRecord *record)
+spgRedoAddNode(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
char *ptr = XLogRecGetData(record);
spgxlogAddNode *xldata = (spgxlogAddNode *) ptr;
char *innerTuple;
SpGistState state;
Buffer buffer;
Page page;
- int bbi;
XLogRedoAction action;
ptr += sizeof(spgxlogAddNode);
fillFakeState(&state, xldata->stateSrc);
- if (xldata->blknoNew == InvalidBlockNumber)
+ if (!XLogRecHasBlockRef(record, 1))
{
/* update in place */
- Assert(xldata->blknoParent == InvalidBlockNumber);
- if (XLogReadBufferForRedo(lsn, record, 0, xldata->node, xldata->blkno,
- &buffer) == BLK_NEEDS_REDO)
+ Assert(xldata->parentBlk == -1);
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
page = BufferGetPage(buffer);
+
PageIndexTupleDelete(page, xldata->offnum);
if (PageAddItem(page, (Item) innerTuple, innerTupleHdr.size,
- xldata->offnum, false, false) != xldata->offnum)
+ xldata->offnum,
+ false, false) != xldata->offnum)
elog(ERROR, "failed to add item of size %u to SPGiST index page",
innerTupleHdr.size);
}
else
{
+ BlockNumber blkno;
+ BlockNumber blknoNew;
+
+ XLogRecGetBlockTag(record, 0, NULL, NULL, &blkno);
+ XLogRecGetBlockTag(record, 1, NULL, NULL, &blknoNew);
+
/*
* In normal operation we would have all three pages (source, dest,
* and parent) locked simultaneously; but in WAL replay it should be
* safe to update them one at a time, as long as we do it in the right
- * order.
- *
- * The logic here depends on the assumption that blkno != blknoNew,
- * else we can't tell which BKP bit goes with which page, and the LSN
- * checks could go wrong too.
+ * order. We must insert the new tuple before replacing the old tuple
+ * with the redirect tuple.
*/
- Assert(xldata->blkno != xldata->blknoNew);
/* Install new tuple first so redirect is valid */
if (xldata->newPage)
{
- buffer = XLogReadBuffer(xldata->node, xldata->blknoNew, true);
/* AddNode is not used for nulls pages */
+ buffer = XLogInitBufferForRedo(record, 1);
SpGistInitBuffer(buffer, 0);
action = BLK_NEEDS_REDO;
}
else
- action = XLogReadBufferForRedo(lsn, record, 1,
- xldata->node, xldata->blknoNew,
- &buffer);
+ action = XLogReadBufferForRedo(record, 1, &buffer);
if (action == BLK_NEEDS_REDO)
{
page = BufferGetPage(buffer);
innerTupleHdr.size, xldata->offnumNew);
/*
- * If parent is in this same page, don't advance LSN; doing so
- * would fool us into not applying the parent downlink update
- * below. We'll update the LSN when we fix the parent downlink.
+ * If parent is in this same page, update it now.
*/
- if (xldata->blknoParent != xldata->blknoNew)
+ if (xldata->parentBlk == 1)
{
- PageSetLSN(page, lsn);
+ SpGistInnerTuple parentTuple;
+
+ parentTuple = (SpGistInnerTuple) PageGetItem(page,
+ PageGetItemId(page, xldata->offnumParent));
+
+ spgUpdateNodeLink(parentTuple, xldata->nodeI,
+ blknoNew, xldata->offnumNew);
}
+ PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
/* Delete old tuple, replacing it with redirect or placeholder tuple */
- if (XLogReadBufferForRedo(lsn, record, 0, xldata->node, xldata->blkno,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
SpGistDeadTuple dt;
InvalidOffsetNumber);
else
dt = spgFormDeadTuple(&state, SPGIST_REDIRECT,
- xldata->blknoNew,
+ blknoNew,
xldata->offnumNew);
PageIndexTupleDelete(page, xldata->offnum);
- if (PageAddItem(page, (Item) dt, dt->size, xldata->offnum,
+ if (PageAddItem(page, (Item) dt, dt->size,
+ xldata->offnum,
false, false) != xldata->offnum)
elog(ERROR, "failed to add item of size %u to SPGiST index page",
dt->size);
SpGistPageGetOpaque(page)->nRedirection++;
/*
- * If parent is in this same page, don't advance LSN; doing so
- * would fool us into not applying the parent downlink update
- * below. We'll update the LSN when we fix the parent downlink.
+ * If parent is in this same page, update it now.
*/
- if (xldata->blknoParent != xldata->blkno)
+ if (xldata->parentBlk == 0)
{
- PageSetLSN(page, lsn);
+ SpGistInnerTuple parentTuple;
+
+ parentTuple = (SpGistInnerTuple) PageGetItem(page,
+ PageGetItemId(page, xldata->offnumParent));
+
+ spgUpdateNodeLink(parentTuple, xldata->nodeI,
+ blknoNew, xldata->offnumNew);
}
+ PageSetLSN(page, lsn);
MarkBufferDirty(buffer);
}
if (BufferIsValid(buffer))
UnlockReleaseBuffer(buffer);
/*
- * Update parent downlink. Since parent could be in either of the
- * previous two buffers, it's a bit tricky to determine which BKP bit
- * applies.
+ * Update parent downlink (if we didn't do it as part of the source or
+ * destination page update already).
*/
- if (xldata->blknoParent == xldata->blkno)
- bbi = 0;
- else if (xldata->blknoParent == xldata->blknoNew)
- bbi = 1;
- else
- bbi = 2;
-
- if (record->xl_info & XLR_BKP_BLOCK(bbi))
+ if (xldata->parentBlk == 2)
{
- if (bbi == 2) /* else we already did it */
- (void) RestoreBackupBlock(lsn, record, bbi, false, false);
- action = BLK_RESTORED;
- buffer = InvalidBuffer;
- }
- else
- {
- action = XLogReadBufferForRedo(lsn, record, bbi, xldata->node,
- xldata->blknoParent, &buffer);
- Assert(action != BLK_RESTORED);
- }
- if (action == BLK_NEEDS_REDO)
- {
- SpGistInnerTuple innerTuple;
+ if (XLogReadBufferForRedo(record, 2, &buffer) == BLK_NEEDS_REDO)
+ {
+ SpGistInnerTuple parentTuple;
- page = BufferGetPage(buffer);
+ page = BufferGetPage(buffer);
- innerTuple = (SpGistInnerTuple) PageGetItem(page,
+ parentTuple = (SpGistInnerTuple) PageGetItem(page,
PageGetItemId(page, xldata->offnumParent));
- spgUpdateNodeLink(innerTuple, xldata->nodeI,
- xldata->blknoNew, xldata->offnumNew);
+ spgUpdateNodeLink(parentTuple, xldata->nodeI,
+ blknoNew, xldata->offnumNew);
- PageSetLSN(page, lsn);
- MarkBufferDirty(buffer);
+ PageSetLSN(page, lsn);
+ MarkBufferDirty(buffer);
+ }
+ if (BufferIsValid(buffer))
+ UnlockReleaseBuffer(buffer);
}
- if (BufferIsValid(buffer))
- UnlockReleaseBuffer(buffer);
}
}
static void
-spgRedoSplitTuple(XLogRecPtr lsn, XLogRecord *record)
+spgRedoSplitTuple(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
char *ptr = XLogRecGetData(record);
spgxlogSplitTuple *xldata = (spgxlogSplitTuple *) ptr;
char *prefixTuple;
SpGistInnerTupleData postfixTupleHdr;
Buffer buffer;
Page page;
+ XLogRedoAction action;
ptr += sizeof(spgxlogSplitTuple);
prefixTuple = ptr;
*/
/* insert postfix tuple first to avoid dangling link */
- if (xldata->blknoPostfix != xldata->blknoPrefix)
+ if (!xldata->postfixBlkSame)
{
- XLogRedoAction action;
-
if (xldata->newPage)
{
- buffer = XLogReadBuffer(xldata->node, xldata->blknoPostfix, true);
+ buffer = XLogInitBufferForRedo(record, 1);
/* SplitTuple is not used for nulls pages */
SpGistInitBuffer(buffer, 0);
action = BLK_NEEDS_REDO;
}
else
- action = XLogReadBufferForRedo(lsn, record, 1,
- xldata->node, xldata->blknoPostfix,
- &buffer);
-
+ action = XLogReadBufferForRedo(record, 1, &buffer);
if (action == BLK_NEEDS_REDO)
{
page = BufferGetPage(buffer);
}
/* now handle the original page */
- if (XLogReadBufferForRedo(lsn, record, 0, xldata->node, xldata->blknoPrefix,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
page = BufferGetPage(buffer);
+
PageIndexTupleDelete(page, xldata->offnumPrefix);
if (PageAddItem(page, (Item) prefixTuple, prefixTupleHdr.size,
xldata->offnumPrefix, false, false) != xldata->offnumPrefix)
elog(ERROR, "failed to add item of size %u to SPGiST index page",
prefixTupleHdr.size);
- if (xldata->blknoPostfix == xldata->blknoPrefix)
- addOrReplaceTuple(page, (Item) postfixTuple, postfixTupleHdr.size,
+ if (xldata->postfixBlkSame)
+ addOrReplaceTuple(page, (Item) postfixTuple,
+ postfixTupleHdr.size,
xldata->offnumPostfix);
PageSetLSN(page, lsn);
}
static void
-spgRedoPickSplit(XLogRecPtr lsn, XLogRecord *record)
+spgRedoPickSplit(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
char *ptr = XLogRecGetData(record);
spgxlogPickSplit *xldata = (spgxlogPickSplit *) ptr;
char *innerTuple;
uint8 *leafPageSelect;
Buffer srcBuffer;
Buffer destBuffer;
+ Buffer innerBuffer;
Page srcPage;
Page destPage;
- Buffer innerBuffer;
Page page;
- int bbi;
int i;
+ BlockNumber blknoInner;
XLogRedoAction action;
+ XLogRecGetBlockTag(record, 2, NULL, NULL, &blknoInner);
+
fillFakeState(&state, xldata->stateSrc);
ptr += SizeOfSpgxlogPickSplit;
/* now ptr points to the list of leaf tuples */
- /*
- * It's a bit tricky to identify which pages have been handled as
- * full-page images, so we explicitly count each referenced buffer.
- */
- bbi = 0;
-
- if (SpGistBlockIsRoot(xldata->blknoSrc))
+ if (xldata->isRootSplit)
{
/* when splitting root, we touch it only in the guise of new inner */
srcBuffer = InvalidBuffer;
else if (xldata->initSrc)
{
/* just re-init the source page */
- srcBuffer = XLogReadBuffer(xldata->node, xldata->blknoSrc, true);
- Assert(BufferIsValid(srcBuffer));
+ srcBuffer = XLogInitBufferForRedo(record, 0);
srcPage = (Page) BufferGetPage(srcBuffer);
SpGistInitBuffer(srcBuffer,
* inserting leaf tuples and the new inner tuple, else the added
* redirect tuple will be a dangling link.)
*/
- if (XLogReadBufferForRedo(lsn, record, bbi,
- xldata->node, xldata->blknoSrc,
- &srcBuffer) == BLK_NEEDS_REDO)
+ srcPage = NULL;
+ if (XLogReadBufferForRedo(record, 0, &srcBuffer) == BLK_NEEDS_REDO)
{
srcPage = BufferGetPage(srcBuffer);
toDelete, xldata->nDelete,
SPGIST_REDIRECT,
SPGIST_PLACEHOLDER,
- xldata->blknoInner,
+ blknoInner,
xldata->offnumInner);
else
spgPageIndexMultiDelete(&state, srcPage,
/* don't update LSN etc till we're done with it */
}
- else
- {
- srcPage = NULL; /* don't do any page updates */
- }
- bbi++;
}
/* try to access dest page if any */
- if (xldata->blknoDest == InvalidBlockNumber)
+ if (!XLogRecHasBlockRef(record, 1))
{
destBuffer = InvalidBuffer;
destPage = NULL;
else if (xldata->initDest)
{
/* just re-init the dest page */
- destBuffer = XLogReadBuffer(xldata->node, xldata->blknoDest, true);
- Assert(BufferIsValid(destBuffer));
+ destBuffer = XLogInitBufferForRedo(record, 1);
destPage = (Page) BufferGetPage(destBuffer);
SpGistInitBuffer(destBuffer,
* We could probably release the page lock immediately in the
* full-page-image case, but for safety let's hold it till later.
*/
- if (XLogReadBufferForRedo(lsn, record, bbi,
- xldata->node, xldata->blknoDest,
- &destBuffer) == BLK_NEEDS_REDO)
- {
+ if (XLogReadBufferForRedo(record, 1, &destBuffer) == BLK_NEEDS_REDO)
destPage = (Page) BufferGetPage(destBuffer);
- }
else
- {
destPage = NULL; /* don't do any page updates */
- }
- bbi++;
}
/* restore leaf tuples to src and/or dest page */
/* restore new inner tuple */
if (xldata->initInner)
{
- innerBuffer = XLogReadBuffer(xldata->node, xldata->blknoInner, true);
- SpGistInitBuffer(innerBuffer,
- (xldata->storesNulls ? SPGIST_NULLS : 0));
+ innerBuffer = XLogInitBufferForRedo(record, 2);
+ SpGistInitBuffer(innerBuffer, (xldata->storesNulls ? SPGIST_NULLS : 0));
action = BLK_NEEDS_REDO;
}
else
- action = XLogReadBufferForRedo(lsn, record, bbi, xldata->node,
- xldata->blknoInner, &innerBuffer);
+ action = XLogReadBufferForRedo(record, 2, &innerBuffer);
if (action == BLK_NEEDS_REDO)
{
xldata->offnumInner);
/* if inner is also parent, update link while we're here */
- if (xldata->blknoInner == xldata->blknoParent)
+ if (xldata->innerIsParent)
{
SpGistInnerTuple parent;
parent = (SpGistInnerTuple) PageGetItem(page,
PageGetItemId(page, xldata->offnumParent));
spgUpdateNodeLink(parent, xldata->nodeI,
- xldata->blknoInner, xldata->offnumInner);
+ blknoInner, xldata->offnumInner);
}
PageSetLSN(page, lsn);
}
if (BufferIsValid(innerBuffer))
UnlockReleaseBuffer(innerBuffer);
- bbi++;
/*
* Now we can release the leaf-page locks. It's okay to do this before
UnlockReleaseBuffer(destBuffer);
/* update parent downlink, unless we did it above */
- if (xldata->blknoParent == InvalidBlockNumber)
- {
- /* no parent cause we split the root */
- Assert(SpGistBlockIsRoot(xldata->blknoInner));
- }
- else if (xldata->blknoInner != xldata->blknoParent)
+ if (XLogRecHasBlockRef(record, 3))
{
Buffer parentBuffer;
- if (XLogReadBufferForRedo(lsn, record, bbi,
- xldata->node, xldata->blknoParent,
- &parentBuffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 3, &parentBuffer) == BLK_NEEDS_REDO)
{
SpGistInnerTuple parent;
parent = (SpGistInnerTuple) PageGetItem(page,
PageGetItemId(page, xldata->offnumParent));
spgUpdateNodeLink(parent, xldata->nodeI,
- xldata->blknoInner, xldata->offnumInner);
+ blknoInner, xldata->offnumInner);
PageSetLSN(page, lsn);
MarkBufferDirty(parentBuffer);
if (BufferIsValid(parentBuffer))
UnlockReleaseBuffer(parentBuffer);
}
+ else
+ Assert(xldata->innerIsParent || xldata->isRootSplit);
}
static void
-spgRedoVacuumLeaf(XLogRecPtr lsn, XLogRecord *record)
+spgRedoVacuumLeaf(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
char *ptr = XLogRecGetData(record);
spgxlogVacuumLeaf *xldata = (spgxlogVacuumLeaf *) ptr;
OffsetNumber *toDead;
ptr += sizeof(OffsetNumber) * xldata->nChain;
chainDest = (OffsetNumber *) ptr;
- if (XLogReadBufferForRedo(lsn, record, 0, xldata->node, xldata->blkno,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
page = BufferGetPage(buffer);
}
static void
-spgRedoVacuumRoot(XLogRecPtr lsn, XLogRecord *record)
+spgRedoVacuumRoot(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
char *ptr = XLogRecGetData(record);
spgxlogVacuumRoot *xldata = (spgxlogVacuumRoot *) ptr;
OffsetNumber *toDelete;
toDelete = xldata->offsets;
- if (XLogReadBufferForRedo(lsn, record, 0, xldata->node, xldata->blkno,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
page = BufferGetPage(buffer);
}
static void
-spgRedoVacuumRedirect(XLogRecPtr lsn, XLogRecord *record)
+spgRedoVacuumRedirect(XLogReaderState *record)
{
+ XLogRecPtr lsn = record->EndRecPtr;
char *ptr = XLogRecGetData(record);
spgxlogVacuumRedirect *xldata = (spgxlogVacuumRedirect *) ptr;
OffsetNumber *itemToPlaceholder;
if (InHotStandby)
{
if (TransactionIdIsValid(xldata->newestRedirectXid))
+ {
+ RelFileNode node;
+
+ XLogRecGetBlockTag(record, 0, &node, NULL, NULL);
ResolveRecoveryConflictWithSnapshot(xldata->newestRedirectXid,
- xldata->node);
+ node);
+ }
}
- if (XLogReadBufferForRedo(lsn, record, 0, xldata->node, xldata->blkno,
- &buffer) == BLK_NEEDS_REDO)
+ if (XLogReadBufferForRedo(record, 0, &buffer) == BLK_NEEDS_REDO)
{
Page page = BufferGetPage(buffer);
SpGistPageOpaque opaque = SpGistPageGetOpaque(page);
}
void
-spg_redo(XLogRecPtr lsn, XLogRecord *record)
+spg_redo(XLogReaderState *record)
{
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
MemoryContext oldCxt;
oldCxt = MemoryContextSwitchTo(opCtx);
switch (info)
{
case XLOG_SPGIST_CREATE_INDEX:
- spgRedoCreateIndex(lsn, record);
+ spgRedoCreateIndex(record);
break;
case XLOG_SPGIST_ADD_LEAF:
- spgRedoAddLeaf(lsn, record);
+ spgRedoAddLeaf(record);
break;
case XLOG_SPGIST_MOVE_LEAFS:
- spgRedoMoveLeafs(lsn, record);
+ spgRedoMoveLeafs(record);
break;
case XLOG_SPGIST_ADD_NODE:
- spgRedoAddNode(lsn, record);
+ spgRedoAddNode(record);
break;
case XLOG_SPGIST_SPLIT_TUPLE:
- spgRedoSplitTuple(lsn, record);
+ spgRedoSplitTuple(record);
break;
case XLOG_SPGIST_PICKSPLIT:
- spgRedoPickSplit(lsn, record);
+ spgRedoPickSplit(record);
break;
case XLOG_SPGIST_VACUUM_LEAF:
- spgRedoVacuumLeaf(lsn, record);
+ spgRedoVacuumLeaf(record);
break;
case XLOG_SPGIST_VACUUM_ROOT:
- spgRedoVacuumRoot(lsn, record);
+ spgRedoVacuumRoot(record);
break;
case XLOG_SPGIST_VACUUM_REDIRECT:
- spgRedoVacuumRedirect(lsn, record);
+ spgRedoVacuumRedirect(record);
break;
default:
elog(PANIC, "spg_redo: unknown op code %u", info);
Note that marking a buffer dirty with MarkBufferDirty() should only
happen iff you write a WAL record; see Writing Hints below.
-5. If the relation requires WAL-logging, build a WAL log record and pass it
-to XLogInsert(); then update the page's LSN using the returned XLOG
-location. For instance,
+5. If the relation requires WAL-logging, build a WAL record using
+XLogBeginInsert and XLogRegister* functions, and insert it. (See
+"Constructing a WAL record" below). Then update the page's LSN using the
+returned XLOG location. For instance,
- recptr = XLogInsert(rmgr_id, info, rdata);
+ XLogBeginInsert();
+ XLogRegisterBuffer(...)
+ XLogRegisterData(...)
+ recptr = XLogInsert(rmgr_id, info);
PageSetLSN(dp, recptr);
- // Note that we no longer do PageSetTLI() from 9.3 onwards
- // since that field on a page has now changed its meaning.
6. END_CRIT_SECTION()
7. Unlock and unpin the buffer(s).
-XLogInsert's "rdata" argument is an array of pointer/size items identifying
-chunks of data to be written in the XLOG record, plus optional shared-buffer
-IDs for chunks that are in shared buffers rather than temporary variables.
-The "rdata" array must mention (at least once) each of the shared buffers
-being modified, unless the action is such that the WAL replay routine can
-reconstruct the entire page contents. XLogInsert includes the logic that
-tests to see whether a shared buffer has been modified since the last
-checkpoint. If not, the entire page contents are logged rather than just the
-portion(s) pointed to by "rdata".
-
-Because XLogInsert drops the rdata components associated with buffers it
-chooses to log in full, the WAL replay routines normally need to test to see
-which buffers were handled that way --- otherwise they may be misled about
-what the XLOG record actually contains. XLOG records that describe multi-page
-changes therefore require some care to design: you must be certain that you
-know what data is indicated by each "BKP" bit. An example of the trickiness
-is that in a HEAP_UPDATE record, BKP(0) normally is associated with the source
-page and BKP(1) is associated with the destination page --- but if these are
-the same page, only BKP(0) would have been set.
-
-For this reason as well as the risk of deadlocking on buffer locks, it's best
-to design WAL records so that they reflect small atomic actions involving just
-one or a few pages. The current XLOG infrastructure cannot handle WAL records
-involving references to more than four shared buffers, anyway.
-
-In the case where the WAL record contains enough information to re-generate
-the entire contents of a page, do *not* show that page's buffer ID in the
-rdata array, even if some of the rdata items point into the buffer. This is
-because you don't want XLogInsert to log the whole page contents. The
-standard replay-routine pattern for this case is
-
- buffer = XLogReadBuffer(rnode, blkno, true);
- Assert(BufferIsValid(buffer));
- page = (Page) BufferGetPage(buffer);
-
- ... initialize the page ...
-
- PageSetLSN(page, lsn);
- MarkBufferDirty(buffer);
- UnlockReleaseBuffer(buffer);
-
-In the case where the WAL record provides only enough information to
-incrementally update the page, the rdata array *must* mention the buffer
-ID at least once; otherwise there is no defense against torn-page problems.
-The standard replay-routine pattern for this case is
-
- if (XLogReadBufferForRedo(lsn, record, N, rnode, blkno, &buffer) == BLK_NEEDS_REDO)
- {
- page = (Page) BufferGetPage(buffer);
-
- ... apply the change ...
-
- PageSetLSN(page, lsn);
- MarkBufferDirty(buffer);
- }
- if (BufferIsValid(buffer))
- UnlockReleaseBuffer(buffer);
-
-XLogReadBufferForRedo reads the page from disk, and checks what action needs to
-be taken to the page. If the XLR_BKP_BLOCK(N) flag is set, it restores the
-full page image and returns BLK_RESTORED. If there is no full page image, but
-page cannot be found or if the change has already been replayed (i.e. the
-page's LSN >= the record we're replaying), it returns BLK_NOTFOUND or BLK_DONE,
-respectively. Usually, the redo routine only needs to pay attention to the
-BLK_NEEDS_REDO return code, which means that the routine should apply the
-incremental change. In any case, the caller is responsible for unlocking and
-releasing the buffer. Note that XLogReadBufferForRedo returns the buffer
-locked even if no redo is required, unless the page does not exist.
-
-As noted above, for a multi-page update you need to be able to determine
-which XLR_BKP_BLOCK(N) flag applies to each page. If a WAL record reflects
-a combination of fully-rewritable and incremental updates, then the rewritable
-pages don't count for the XLR_BKP_BLOCK(N) numbering. (XLR_BKP_BLOCK(N) is
-associated with the N'th distinct buffer ID seen in the "rdata" array, and
-per the above discussion, fully-rewritable buffers shouldn't be mentioned in
-"rdata".)
+Complex changes (such as a multilevel index insertion) normally need to be
+described by a series of atomic-action WAL records. The intermediate states
+must be self-consistent, so that if the replay is interrupted between any
+two actions, the system is fully functional. In btree indexes, for example,
+a page split requires a new page to be allocated, and an insertion of a new
+key in the parent btree level, but for locking reasons this has to be
+reflected by two separate WAL records. Replaying the first record, to
+allocate the new page and move tuples to it, sets a flag on the page to
+indicate that the key has not been inserted to the parent yet. Replaying the
+second record clears the flag. This intermediate state is never seen by
+other backends during normal operation, because the lock on the child page
+is held across the two actions, but will be seen if the operation is
+interrupted before writing the second WAL record. The search algorithm works
+with the intermediate state as normal, but if an insertion encounters a page
+with the incomplete-split flag set, it will finish the interrupted split by
+inserting the key to the parent, before proceeding.
+
+
+Constructing a WAL record
+-------------------------
+
+A WAL record consists of a header common to all WAL record types,
+record-specific data, and information about the data blocks modified. Each
+modified data block is identified by an ID number, and can optionally have
+more record-specific data associated with the block. If XLogInsert decides
+that a full-page image of a block needs to be taken, the data associated
+with that block is not included.
+
+The API for constructing a WAL record consists of five functions:
+XLogBeginInsert, XLogRegisterBuffer, XLogRegisterData, XLogRegisterBufData,
+and XLogInsert. First, call XLogBeginInsert(). Then register all the buffers
+modified, and data needed to replay the changes, using XLogRegister*
+functions. Finally, insert the constructed record to the WAL by calling
+XLogInsert().
+
+ XLogBeginInsert();
+
+ /* register buffers modified as part of this WAL-logged action */
+ XLogRegisterBuffer(0, lbuffer, REGBUF_STANDARD);
+ XLogRegisterBuffer(1, rbuffer, REGBUF_STANDARD);
+
+ /* register data that is always included in the WAL record */
+ XLogRegisterData(&xlrec, SizeOfFictionalAction);
+
+ /*
+ * register data associated with a buffer. This will not be included
+ * in the record if a full-page image is taken.
+ */
+ XLogRegisterBufData(0, tuple->data, tuple->len);
+
+ /* more data associated with the buffer */
+ XLogRegisterBufData(0, data2, len2);
+
+ /*
+ * Ok, all the data and buffers to include in the WAL record have
+ * been registered. Insert the record.
+ */
+ recptr = XLogInsert(RM_FOO_ID, XLOG_FOOBAR_DO_STUFF);
+
+Details of the API functions:
+
+void XLogBeginInsert(void)
+
+ Must be called before XLogRegisterBuffer and XLogRegisterData.
+
+void XLogResetInsertion(void)
+
+ Clear any currently registered data and buffers from the WAL record
+ construction workspace. This is only needed if you have already called
+ XLogBeginInsert(), but decide to not insert the record after all.
+
+void XLogEnsureRecordSpace(int max_block_id, int nrdatas)
+
+ Normally, the WAL record construction buffers have the following limits:
+
+ * highest block ID that can be used is 4 (allowing five block references)
+ * Max 20 chunks of registered data
+
+ These default limits are enough for most record types that change some
+ on-disk structures. For the odd case that requires more data, or needs to
+ modify more buffers, these limits can be raised by calling
+ XLogEnsureRecordSpace(). XLogEnsureRecordSpace() must be called before
+ XLogBeginInsert(), and outside a critical section.
+
+void XLogRegisterBuffer(uint8 block_id, Buffer buf, uint8 flags);
+
+ XLogRegisterBuffer adds information about a data block to the WAL record.
+ block_id is an arbitrary number used to identify this page reference in
+ the redo routine. The information needed to re-find the page at redo -
+ relfilenode, fork, and block number - are included in the WAL record.
+
+ XLogInsert will automatically include a full copy of the page contents, if
+ this is the first modification of the buffer since the last checkpoint.
+ It is important to register every buffer modified by the action with
+ XLogRegisterBuffer, to avoid torn-page hazards.
+
+ The flags control when and how the buffer contents are included in the
+ WAL record. Normally, a full-page image is taken only if the page has not
+ been modified since the last checkpoint, and only if full_page_writes=on
+ or an online backup is in progress. The REGBUF_FORCE_IMAGE flag can be
+ used to force a full-page image to always be included; that is useful
+ e.g. for an operation that rewrites most of the page, so that tracking the
+ details is not worth it. For the rare case where it is not necessary to
+ protect from torn pages, REGBUF_NO_IMAGE flag can be used to suppress
+ full page image from being taken. REGBUF_WILL_INIT also suppresses a full
+ page image, but the redo routine must re-generate the page from scratch,
+ without looking at the old page contents. Re-initializing the page
+ protects from torn page hazards like a full page image does.
+
+ The REGBUF_STANDARD flag can be specified together with the other flags to
+ indicate that the page follows the standard page layout. It causes the
+ area between pd_lower and pd_upper to be left out from the image, reducing
+ WAL volume.
+
+ If the REGBUF_KEEP_DATA flag is given, any per-buffer data registered with
+ XLogRegisterBufData() is included in the WAL record even if a full-page
+ image is taken.
+
+void XLogRegisterData(char *data, int len);
+
+ XLogRegisterData is used to include arbitrary data in the WAL record. If
+ XLogRegisterData() is called multiple times, the data are appended, and
+ will be made available to the redo routine as one contiguous chunk.
+
+void XLogRegisterBufData(uint8 block_id, char *data, int len);
+
+ XLogRegisterBufData is used to include data associated with a particular
+ buffer that was registered earlier with XLogRegisterBuffer(). If
+ XLogRegisterBufData() is called multiple times with the same block ID, the
+ data are appended, and will be made available to the redo routine as one
+ contiguous chunk.
+
+ If a full-page image of the buffer is taken at insertion, the data is not
+ included in the WAL record, unless the REGBUF_KEEP_DATA flag is used.
+
+
+Writing a REDO routine
+----------------------
+
+A REDO routine uses the data and page references included in the WAL record
+to reconstruct the new state of the page. The record decoding functions
+and macros in xlogreader.c/h can be used to extract the data from the record.
When replaying a WAL record that describes changes on multiple pages, you
must be careful to lock the pages properly to prevent concurrent Hot Standby
or be writing the data block directly rather than through shared buffers
while holding AccessExclusiveLock on the relation.
-Due to all these constraints, complex changes (such as a multilevel index
-insertion) normally need to be described by a series of atomic-action WAL
-records. The intermediate states must be self-consistent, so that if the
-replay is interrupted between any two actions, the system is fully
-functional. In btree indexes, for example, a page split requires a new page
-to be allocated, and an insertion of a new key in the parent btree level,
-but for locking reasons this has to be reflected by two separate WAL
-records. Replaying the first record, to allocate the new page and move
-tuples to it, sets a flag on the page to indicate that the key has not been
-inserted to the parent yet. Replaying the second record clears the flag.
-This intermediate state is never seen by other backends during normal
-operation, because the lock on the child page is held across the two
-actions, but will be seen if the operation is interrupted before writing
-the second WAL record. The search algorithm works with the intermediate
-state as normal, but if an insertion encounters a page with the
-incomplete-split flag set, it will finish the interrupted split by
-inserting the key to the parent, before proceeding.
Writing Hints
-------------
static void
WriteZeroPageXlogRec(int pageno)
{
- XLogRecData rdata;
-
- rdata.data = (char *) (&pageno);
- rdata.len = sizeof(int);
- rdata.buffer = InvalidBuffer;
- rdata.next = NULL;
- (void) XLogInsert(RM_CLOG_ID, CLOG_ZEROPAGE, &rdata);
+ XLogBeginInsert();
+ XLogRegisterData((char *) (&pageno), sizeof(int));
+ (void) XLogInsert(RM_CLOG_ID, CLOG_ZEROPAGE);
}
/*
static void
WriteTruncateXlogRec(int pageno)
{
- XLogRecData rdata;
XLogRecPtr recptr;
- rdata.data = (char *) (&pageno);
- rdata.len = sizeof(int);
- rdata.buffer = InvalidBuffer;
- rdata.next = NULL;
- recptr = XLogInsert(RM_CLOG_ID, CLOG_TRUNCATE, &rdata);
+ XLogBeginInsert();
+ XLogRegisterData((char *) (&pageno), sizeof(int));
+ recptr = XLogInsert(RM_CLOG_ID, CLOG_TRUNCATE);
XLogFlush(recptr);
}
* CLOG resource manager's routines
*/
void
-clog_redo(XLogRecPtr lsn, XLogRecord *record)
+clog_redo(XLogReaderState *record)
{
- uint8 info = record->xl_info & ~XLR_INFO_MASK;
+ uint8 info = XLogRecGetInfo(record) & ~XLR_INFO_MASK;
/* Backup blocks are not used in clog records */
- Assert(!(record->xl_info & XLR_BKP_BLOCK_MASK));
+ Assert(!XLogRecHasAnyBlockRefs(record));
if (info == CLOG_ZEROPAGE)
{
projects / users / bernd / postgres.git / commitdiff