From 5d991456d26e9b4f35b3653c0049625eac006900 Mon Sep 17 00:00:00 2001 From: Thomas Munro Date: Mon, 15 Jul 2019 21:25:08 +1200 Subject: [PATCH 1/3] Add SmgrId to smgropen() and BufferTag. To use bufmgr.c for new kinds of data in addition to plain old relations, add an SMGR argument to places that identify blocks and the files that hold them (smgropen(), block references in the WAL, BufferTag). To avoid making BufferTag wider, take some space away from the fork number for this new member, since there are just a few values possible (a suggestion from Anton Shyrabokau). Create a new callback for smgropen() calls so that some md.c- specific stuff can move out of smgropen(), and future implementations can also run their own initialization code. Author: Thomas Munro Reviewed-by: Shawn Debnath Discussion: https://postgr.es/m/CA%2BhUKG%2BOZqOiOuDm5tC5DyQZtJ3FH4%2BFSVMqtdC4P1atpJ%2Bqhg%40mail.gmail.com Discussion: https://postgr.es/m/CA%2BhUKG%2BDE0mmiBZMtZyvwWtgv1sZCniSVhXYsXkvJ_Wo%2B83vvw%40mail.gmail.com --- contrib/bloom/blinsert.c | 2 +- src/backend/access/brin/brin_xlog.c | 2 +- src/backend/access/gin/ginxlog.c | 3 +- src/backend/access/gist/gistxlog.c | 6 +-- src/backend/access/hash/hash_xlog.c | 12 +++--- src/backend/access/hash/hashpage.c | 6 ++- src/backend/access/heap/heapam.c | 20 +++++----- src/backend/access/heap/heapam_handler.c | 2 +- src/backend/access/heap/rewriteheap.c | 6 ++- src/backend/access/nbtree/nbtree.c | 2 +- src/backend/access/nbtree/nbtsort.c | 3 +- src/backend/access/nbtree/nbtxlog.c | 8 ++-- src/backend/access/spgist/spginsert.c | 6 +-- src/backend/access/spgist/spgxlog.c | 12 +++--- src/backend/access/transam/xlog.c | 6 ++- src/backend/access/transam/xloginsert.c | 39 ++++++++++++------- src/backend/access/transam/xlogreader.c | 11 +++++- src/backend/access/transam/xlogutils.c | 17 ++++---- src/backend/catalog/storage.c | 11 +++--- src/backend/commands/tablecmds.c | 2 +- src/backend/replication/logical/decode.c | 10 ++--- .../replication/logical/reorderbuffer.c | 3 +- src/backend/storage/buffer/bufmgr.c | 29 +++++++++----- src/backend/storage/buffer/localbuf.c | 8 ++-- src/backend/storage/freespace/freespace.c | 3 +- src/backend/storage/freespace/fsmpage.c | 3 +- src/backend/storage/smgr/md.c | 29 +++++++++----- src/backend/storage/smgr/smgr.c | 13 +++---- src/bin/pg_rewind/parsexlog.c | 8 +++- src/bin/pg_waldump/pg_waldump.c | 16 +++++--- src/include/access/xloginsert.h | 13 ++++--- src/include/access/xlogreader.h | 6 ++- src/include/access/xlogutils.h | 4 +- src/include/storage/buf_internals.h | 11 +++++- src/include/storage/bufmgr.h | 9 +++-- src/include/storage/md.h | 1 + src/include/storage/smgr.h | 8 +++- src/include/utils/rel.h | 6 ++- 38 files changed, 223 insertions(+), 133 deletions(-) diff --git a/contrib/bloom/blinsert.c b/contrib/bloom/blinsert.c index 4b2186b8dda..e39d21df1f6 100644 --- a/contrib/bloom/blinsert.c +++ b/contrib/bloom/blinsert.c @@ -181,7 +181,7 @@ blbuildempty(Relation index) PageSetChecksumInplace(metapage, BLOOM_METAPAGE_BLKNO); smgrwrite(index->rd_smgr, INIT_FORKNUM, BLOOM_METAPAGE_BLKNO, (char *) metapage, true); - log_newpage(&index->rd_smgr->smgr_rnode.node, INIT_FORKNUM, + log_newpage(SMGR_MD, &index->rd_smgr->smgr_rnode.node, INIT_FORKNUM, BLOOM_METAPAGE_BLKNO, metapage, true); /* diff --git a/src/backend/access/brin/brin_xlog.c b/src/backend/access/brin/brin_xlog.c index db1f47ca218..a13b3cd2575 100644 --- a/src/backend/access/brin/brin_xlog.c +++ b/src/backend/access/brin/brin_xlog.c @@ -217,7 +217,7 @@ brin_xlog_revmap_extend(XLogReaderState *record) xlrec = (xl_brin_revmap_extend *) XLogRecGetData(record); - XLogRecGetBlockTag(record, 1, NULL, NULL, &targetBlk); + XLogRecGetBlockTag(record, 1, NULL, NULL, NULL, &targetBlk); Assert(xlrec->targetBlk == targetBlk); /* Update the metapage */ diff --git a/src/backend/access/gin/ginxlog.c b/src/backend/access/gin/ginxlog.c index c945b282721..261881c4184 100644 --- a/src/backend/access/gin/ginxlog.c +++ b/src/backend/access/gin/ginxlog.c @@ -95,11 +95,12 @@ ginRedoInsertEntry(Buffer buffer, bool isLeaf, BlockNumber rightblkno, void *rda if (PageAddItem(page, (Item) itup, IndexTupleSize(itup), offset, false, false) == InvalidOffsetNumber) { + SmgrId smgrid; RelFileNode node; ForkNumber forknum; BlockNumber blknum; - BufferGetTag(buffer, &node, &forknum, &blknum); + BufferGetTag(buffer, &smgrid, &node, &forknum, &blknum); elog(ERROR, "failed to add item to index page in %u/%u/%u", node.spcNode, node.dbNode, node.relNode); } diff --git a/src/backend/access/gist/gistxlog.c b/src/backend/access/gist/gistxlog.c index 503db34d863..bf945b9fb50 100644 --- a/src/backend/access/gist/gistxlog.c +++ b/src/backend/access/gist/gistxlog.c @@ -193,7 +193,7 @@ gistRedoDeleteRecord(XLogReaderState *record) { RelFileNode rnode; - XLogRecGetBlockTag(record, 0, &rnode, NULL, NULL); + XLogRecGetBlockTag(record, 0, NULL, &rnode, NULL, NULL); ResolveRecoveryConflictWithSnapshot(xldata->latestRemovedXid, rnode); } @@ -278,7 +278,7 @@ gistRedoPageSplitRecord(XLogReaderState *record) BlockNumber blkno; IndexTuple *tuples; - XLogRecGetBlockTag(record, i + 1, NULL, NULL, &blkno); + XLogRecGetBlockTag(record, i + 1, NULL, NULL, NULL, &blkno); if (blkno == GIST_ROOT_BLKNO) { Assert(i == 0); @@ -313,7 +313,7 @@ gistRedoPageSplitRecord(XLogReaderState *record) { BlockNumber nextblkno; - XLogRecGetBlockTag(record, i + 2, NULL, NULL, &nextblkno); + XLogRecGetBlockTag(record, i + 2, NULL, NULL, NULL, &nextblkno); GistPageGetOpaque(page)->rightlink = nextblkno; } else diff --git a/src/backend/access/hash/hash_xlog.c b/src/backend/access/hash/hash_xlog.c index d7b70981101..ec604a7d428 100644 --- a/src/backend/access/hash/hash_xlog.c +++ b/src/backend/access/hash/hash_xlog.c @@ -51,7 +51,7 @@ hash_xlog_init_meta_page(XLogReaderState *record) * special handling for init forks as create index operations don't log a * full page image of the metapage. */ - XLogRecGetBlockTag(record, 0, NULL, &forknum, NULL); + XLogRecGetBlockTag(record, 0, NULL, NULL, &forknum, NULL); if (forknum == INIT_FORKNUM) FlushOneBuffer(metabuf); @@ -89,7 +89,7 @@ hash_xlog_init_bitmap_page(XLogReaderState *record) * special handling for init forks as create index operations don't log a * full page image of the metapage. */ - XLogRecGetBlockTag(record, 0, NULL, &forknum, NULL); + XLogRecGetBlockTag(record, 0, NULL, NULL, &forknum, NULL); if (forknum == INIT_FORKNUM) FlushOneBuffer(bitmapbuf); UnlockReleaseBuffer(bitmapbuf); @@ -113,7 +113,7 @@ hash_xlog_init_bitmap_page(XLogReaderState *record) PageSetLSN(page, lsn); MarkBufferDirty(metabuf); - XLogRecGetBlockTag(record, 1, NULL, &forknum, NULL); + XLogRecGetBlockTag(record, 1, NULL, NULL, &forknum, NULL); if (forknum == INIT_FORKNUM) FlushOneBuffer(metabuf); } @@ -190,8 +190,8 @@ hash_xlog_add_ovfl_page(XLogReaderState *record) Size datalen PG_USED_FOR_ASSERTS_ONLY; bool new_bmpage = false; - XLogRecGetBlockTag(record, 0, NULL, NULL, &rightblk); - XLogRecGetBlockTag(record, 1, NULL, NULL, &leftblk); + XLogRecGetBlockTag(record, 0, NULL, NULL, NULL, &rightblk); + XLogRecGetBlockTag(record, 1, NULL, NULL, NULL, &leftblk); ovflbuf = XLogInitBufferForRedo(record, 0); Assert(BufferIsValid(ovflbuf)); @@ -1001,7 +1001,7 @@ hash_xlog_vacuum_one_page(XLogReaderState *record) { RelFileNode rnode; - XLogRecGetBlockTag(record, 0, &rnode, NULL, NULL); + XLogRecGetBlockTag(record, 0, NULL, &rnode, NULL, NULL); ResolveRecoveryConflictWithSnapshot(xldata->latestRemovedXid, rnode); } diff --git a/src/backend/access/hash/hashpage.c b/src/backend/access/hash/hashpage.c index 376ee2a63b5..f6042e42f91 100644 --- a/src/backend/access/hash/hashpage.c +++ b/src/backend/access/hash/hashpage.c @@ -427,7 +427,8 @@ _hash_init(Relation rel, double num_tuples, ForkNumber forkNum) MarkBufferDirty(buf); if (use_wal) - log_newpage(&rel->rd_node, + log_newpage(SMGR_MD, + &rel->rd_node, forkNum, blkno, BufferGetPage(buf), @@ -1021,7 +1022,8 @@ _hash_alloc_buckets(Relation rel, BlockNumber firstblock, uint32 nblocks) ovflopaque->hasho_page_id = HASHO_PAGE_ID; if (RelationNeedsWAL(rel)) - log_newpage(&rel->rd_node, + log_newpage(SMGR_MD, + &rel->rd_node, MAIN_FORKNUM, lastblock, zerobuf.data, diff --git a/src/backend/access/heap/heapam.c b/src/backend/access/heap/heapam.c index d768b9b061c..96e732cd12f 100644 --- a/src/backend/access/heap/heapam.c +++ b/src/backend/access/heap/heapam.c @@ -7725,7 +7725,7 @@ heap_xlog_clean(XLogReaderState *record) BlockNumber blkno; XLogRedoAction action; - XLogRecGetBlockTag(record, 0, &rnode, NULL, &blkno); + XLogRecGetBlockTag(record, 0, NULL, &rnode, NULL, &blkno); /* * We're about to remove tuples. In Hot Standby mode, ensure that there's @@ -7820,7 +7820,7 @@ heap_xlog_visible(XLogReaderState *record) BlockNumber blkno; XLogRedoAction action; - XLogRecGetBlockTag(record, 1, &rnode, NULL, &blkno); + XLogRecGetBlockTag(record, 1, NULL, &rnode, NULL, &blkno); /* * If there are any Hot Standby transactions running that have an xmin @@ -7968,7 +7968,7 @@ heap_xlog_freeze_page(XLogReaderState *record) TransactionIdRetreat(latestRemovedXid); - XLogRecGetBlockTag(record, 0, &rnode, NULL, NULL); + XLogRecGetBlockTag(record, 0, NULL, &rnode, NULL, NULL); ResolveRecoveryConflictWithSnapshot(latestRemovedXid, rnode); } @@ -8040,7 +8040,7 @@ heap_xlog_delete(XLogReaderState *record) RelFileNode target_node; ItemPointerData target_tid; - XLogRecGetBlockTag(record, 0, &target_node, NULL, &blkno); + XLogRecGetBlockTag(record, 0, NULL, &target_node, NULL, &blkno); ItemPointerSetBlockNumber(&target_tid, blkno); ItemPointerSetOffsetNumber(&target_tid, xlrec->offnum); @@ -8121,7 +8121,7 @@ heap_xlog_insert(XLogReaderState *record) ItemPointerData target_tid; XLogRedoAction action; - XLogRecGetBlockTag(record, 0, &target_node, NULL, &blkno); + XLogRecGetBlockTag(record, 0, NULL, &target_node, NULL, &blkno); ItemPointerSetBlockNumber(&target_tid, blkno); ItemPointerSetOffsetNumber(&target_tid, xlrec->offnum); @@ -8243,7 +8243,7 @@ heap_xlog_multi_insert(XLogReaderState *record) */ xlrec = (xl_heap_multi_insert *) XLogRecGetData(record); - XLogRecGetBlockTag(record, 0, &rnode, NULL, &blkno); + XLogRecGetBlockTag(record, 0, NULL, &rnode, NULL, &blkno); /* * The visibility map may need to be fixed even if the heap page is @@ -8389,8 +8389,8 @@ heap_xlog_update(XLogReaderState *record, bool hot_update) oldtup.t_data = NULL; oldtup.t_len = 0; - XLogRecGetBlockTag(record, 0, &rnode, NULL, &newblk); - if (XLogRecGetBlockTag(record, 1, NULL, NULL, &oldblk)) + XLogRecGetBlockTag(record, 0, NULL, &rnode, NULL, &newblk); + if (XLogRecGetBlockTag(record, 1, NULL, NULL, NULL, &oldblk)) { /* HOT updates are never done across pages */ Assert(!hot_update); @@ -8685,7 +8685,7 @@ heap_xlog_lock(XLogReaderState *record) BlockNumber block; Relation reln; - XLogRecGetBlockTag(record, 0, &rnode, NULL, &block); + XLogRecGetBlockTag(record, 0, NULL, &rnode, NULL, &block); reln = CreateFakeRelcacheEntry(rnode); visibilitymap_pin(reln, block, &vmbuffer); @@ -8758,7 +8758,7 @@ heap_xlog_lock_updated(XLogReaderState *record) BlockNumber block; Relation reln; - XLogRecGetBlockTag(record, 0, &rnode, NULL, &block); + XLogRecGetBlockTag(record, 0, NULL, &rnode, NULL, &block); reln = CreateFakeRelcacheEntry(rnode); visibilitymap_pin(reln, block, &vmbuffer); diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c index 09bc6fe98a7..39afef1d3cf 100644 --- a/src/backend/access/heap/heapam_handler.c +++ b/src/backend/access/heap/heapam_handler.c @@ -634,7 +634,7 @@ heapam_relation_copy_data(Relation rel, const RelFileNode *newrnode) { SMgrRelation dstrel; - dstrel = smgropen(*newrnode, rel->rd_backend); + dstrel = smgropen(SMGR_MD, *newrnode, rel->rd_backend); RelationOpenSmgr(rel); /* diff --git a/src/backend/access/heap/rewriteheap.c b/src/backend/access/heap/rewriteheap.c index 72a448ad316..f47df39bd28 100644 --- a/src/backend/access/heap/rewriteheap.c +++ b/src/backend/access/heap/rewriteheap.c @@ -331,7 +331,8 @@ end_heap_rewrite(RewriteState state) if (state->rs_buffer_valid) { if (state->rs_use_wal) - log_newpage(&state->rs_new_rel->rd_node, + log_newpage(SMGR_MD, + &state->rs_new_rel->rd_node, MAIN_FORKNUM, state->rs_blockno, state->rs_buffer, @@ -696,7 +697,8 @@ raw_heap_insert(RewriteState state, HeapTuple tup) /* XLOG stuff */ if (state->rs_use_wal) - log_newpage(&state->rs_new_rel->rd_node, + log_newpage(SMGR_MD, + &state->rs_new_rel->rd_node, MAIN_FORKNUM, state->rs_blockno, page, diff --git a/src/backend/access/nbtree/nbtree.c b/src/backend/access/nbtree/nbtree.c index 4cfd5289ad7..f2ce02f07c6 100644 --- a/src/backend/access/nbtree/nbtree.c +++ b/src/backend/access/nbtree/nbtree.c @@ -172,7 +172,7 @@ btbuildempty(Relation index) PageSetChecksumInplace(metapage, BTREE_METAPAGE); smgrwrite(index->rd_smgr, INIT_FORKNUM, BTREE_METAPAGE, (char *) metapage, true); - log_newpage(&index->rd_smgr->smgr_rnode.node, INIT_FORKNUM, + log_newpage(SMGR_MD, &index->rd_smgr->smgr_rnode.node, INIT_FORKNUM, BTREE_METAPAGE, metapage, true); /* diff --git a/src/backend/access/nbtree/nbtsort.c b/src/backend/access/nbtree/nbtsort.c index d0b9013caf4..931d6cd32b2 100644 --- a/src/backend/access/nbtree/nbtsort.c +++ b/src/backend/access/nbtree/nbtsort.c @@ -658,7 +658,8 @@ _bt_blwritepage(BTWriteState *wstate, Page page, BlockNumber blkno) if (wstate->btws_use_wal) { /* We use the heap NEWPAGE record type for this */ - log_newpage(&wstate->index->rd_node, MAIN_FORKNUM, blkno, page, true); + log_newpage(SMGR_MD, &wstate->index->rd_node, MAIN_FORKNUM, blkno, + page, true); } /* diff --git a/src/backend/access/nbtree/nbtxlog.c b/src/backend/access/nbtree/nbtxlog.c index 3147ea47268..1c074f77ac3 100644 --- a/src/backend/access/nbtree/nbtxlog.c +++ b/src/backend/access/nbtree/nbtxlog.c @@ -216,9 +216,9 @@ btree_xlog_split(bool onleft, XLogReaderState *record) BlockNumber rightsib; BlockNumber rnext; - XLogRecGetBlockTag(record, 0, NULL, NULL, &leftsib); - XLogRecGetBlockTag(record, 1, NULL, NULL, &rightsib); - if (!XLogRecGetBlockTag(record, 2, NULL, NULL, &rnext)) + XLogRecGetBlockTag(record, 0, NULL, NULL, NULL, &leftsib); + XLogRecGetBlockTag(record, 1, NULL, NULL, NULL, &rightsib); + if (!XLogRecGetBlockTag(record, 2, NULL, NULL, NULL, &rnext)) rnext = P_NONE; /* @@ -524,7 +524,7 @@ btree_xlog_delete(XLogReaderState *record) { RelFileNode rnode; - XLogRecGetBlockTag(record, 0, &rnode, NULL, NULL); + XLogRecGetBlockTag(record, 0, NULL, &rnode, NULL, NULL); ResolveRecoveryConflictWithSnapshot(xlrec->latestRemovedXid, rnode); } diff --git a/src/backend/access/spgist/spginsert.c b/src/backend/access/spgist/spginsert.c index b40bd440cf0..8019f6839de 100644 --- a/src/backend/access/spgist/spginsert.c +++ b/src/backend/access/spgist/spginsert.c @@ -171,7 +171,7 @@ spgbuildempty(Relation index) PageSetChecksumInplace(page, SPGIST_METAPAGE_BLKNO); smgrwrite(index->rd_smgr, INIT_FORKNUM, SPGIST_METAPAGE_BLKNO, (char *) page, true); - log_newpage(&index->rd_smgr->smgr_rnode.node, INIT_FORKNUM, + log_newpage(SMGR_MD, &index->rd_smgr->smgr_rnode.node, INIT_FORKNUM, SPGIST_METAPAGE_BLKNO, page, true); /* Likewise for the root page. */ @@ -180,7 +180,7 @@ spgbuildempty(Relation index) PageSetChecksumInplace(page, SPGIST_ROOT_BLKNO); smgrwrite(index->rd_smgr, INIT_FORKNUM, SPGIST_ROOT_BLKNO, (char *) page, true); - log_newpage(&index->rd_smgr->smgr_rnode.node, INIT_FORKNUM, + log_newpage(SMGR_MD, &index->rd_smgr->smgr_rnode.node, INIT_FORKNUM, SPGIST_ROOT_BLKNO, page, true); /* Likewise for the null-tuples root page. */ @@ -189,7 +189,7 @@ spgbuildempty(Relation index) PageSetChecksumInplace(page, SPGIST_NULL_BLKNO); smgrwrite(index->rd_smgr, INIT_FORKNUM, SPGIST_NULL_BLKNO, (char *) page, true); - log_newpage(&index->rd_smgr->smgr_rnode.node, INIT_FORKNUM, + log_newpage(SMGR_MD, &index->rd_smgr->smgr_rnode.node, INIT_FORKNUM, SPGIST_NULL_BLKNO, page, true); /* diff --git a/src/backend/access/spgist/spgxlog.c b/src/backend/access/spgist/spgxlog.c index ebe6ae8715b..3ce35feee69 100644 --- a/src/backend/access/spgist/spgxlog.c +++ b/src/backend/access/spgist/spgxlog.c @@ -151,7 +151,7 @@ spgRedoAddLeaf(XLogReaderState *record) SpGistInnerTuple tuple; BlockNumber blknoLeaf; - XLogRecGetBlockTag(record, 0, NULL, NULL, &blknoLeaf); + XLogRecGetBlockTag(record, 0, NULL, NULL, NULL, &blknoLeaf); page = BufferGetPage(buffer); @@ -184,7 +184,7 @@ spgRedoMoveLeafs(XLogReaderState *record) XLogRedoAction action; BlockNumber blknoDst; - XLogRecGetBlockTag(record, 1, NULL, NULL, &blknoDst); + XLogRecGetBlockTag(record, 1, NULL, NULL, NULL, &blknoDst); fillFakeState(&state, xldata->stateSrc); @@ -328,8 +328,8 @@ spgRedoAddNode(XLogReaderState *record) BlockNumber blkno; BlockNumber blknoNew; - XLogRecGetBlockTag(record, 0, NULL, NULL, &blkno); - XLogRecGetBlockTag(record, 1, NULL, NULL, &blknoNew); + XLogRecGetBlockTag(record, 0, NULL, NULL, NULL, &blkno); + XLogRecGetBlockTag(record, 1, NULL, NULL, NULL, &blknoNew); /* * In normal operation we would have all three pages (source, dest, @@ -549,7 +549,7 @@ spgRedoPickSplit(XLogReaderState *record) BlockNumber blknoInner; XLogRedoAction action; - XLogRecGetBlockTag(record, 2, NULL, NULL, &blknoInner); + XLogRecGetBlockTag(record, 2, NULL, NULL, NULL, &blknoInner); fillFakeState(&state, xldata->stateSrc); @@ -879,7 +879,7 @@ spgRedoVacuumRedirect(XLogReaderState *record) { RelFileNode node; - XLogRecGetBlockTag(record, 0, &node, NULL, NULL); + XLogRecGetBlockTag(record, 0, NULL, &node, NULL, NULL); ResolveRecoveryConflictWithSnapshot(xldata->newestRedirectXid, node); } diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c index b6c9353cbd2..edcbbf0bc94 100644 --- a/src/backend/access/transam/xlog.c +++ b/src/backend/access/transam/xlog.c @@ -1372,6 +1372,7 @@ checkXLogConsistency(XLogReaderState *record) ForkNumber forknum; BlockNumber blkno; int block_id; + SmgrId smgrid; /* Records with no backup blocks have no need for consistency checks. */ if (!XLogRecHasAnyBlockRefs(record)) @@ -1384,7 +1385,8 @@ checkXLogConsistency(XLogReaderState *record) Buffer buf; Page page; - if (!XLogRecGetBlockTag(record, block_id, &rnode, &forknum, &blkno)) + if (!XLogRecGetBlockTag(record, block_id, &smgrid, &rnode, &forknum, + &blkno)) { /* * WAL record doesn't contain a block reference with the given id. @@ -1409,7 +1411,7 @@ checkXLogConsistency(XLogReaderState *record) * Read the contents from the current buffer and store it in a * temporary page. */ - buf = XLogReadBufferExtended(rnode, forknum, blkno, + buf = XLogReadBufferExtended(smgrid, rnode, forknum, blkno, RBM_NORMAL_NO_LOG); if (!BufferIsValid(buf)) continue; diff --git a/src/backend/access/transam/xloginsert.c b/src/backend/access/transam/xloginsert.c index 3ec67d468b5..1697797bc08 100644 --- a/src/backend/access/transam/xloginsert.c +++ b/src/backend/access/transam/xloginsert.c @@ -43,7 +43,8 @@ typedef struct { bool in_use; /* is this slot in use? */ uint8 flags; /* REGBUF_* flags */ - RelFileNode rnode; /* identifies the relation and block */ + SmgrId smgrid; /* identifies the SGMR, relation and block */ + RelFileNode rnode; ForkNumber forkno; BlockNumber block; Page page; /* page content */ @@ -227,7 +228,8 @@ XLogRegisterBuffer(uint8 block_id, Buffer buffer, uint8 flags) regbuf = ®istered_buffers[block_id]; - BufferGetTag(buffer, ®buf->rnode, ®buf->forkno, ®buf->block); + BufferGetTag(buffer, ®buf->smgrid, ®buf->rnode, ®buf->forkno, + ®buf->block); regbuf->page = BufferGetPage(buffer); regbuf->flags = flags; regbuf->rdata_tail = (XLogRecData *) ®buf->rdata_head; @@ -248,7 +250,8 @@ XLogRegisterBuffer(uint8 block_id, Buffer buffer, uint8 flags) if (i == block_id || !regbuf_old->in_use) continue; - Assert(!RelFileNodeEquals(regbuf_old->rnode, regbuf->rnode) || + Assert(regbuf_old->smgrid != regbuf->smgrid || + !RelFileNodeEquals(regbuf_old->rnode, regbuf->rnode) || regbuf_old->forkno != regbuf->forkno || regbuf_old->block != regbuf->block); } @@ -263,8 +266,9 @@ XLogRegisterBuffer(uint8 block_id, Buffer buffer, uint8 flags) * shared buffer pool (i.e. when you don't have a Buffer for it). */ void -XLogRegisterBlock(uint8 block_id, RelFileNode *rnode, ForkNumber forknum, - BlockNumber blknum, Page page, uint8 flags) +XLogRegisterBlock(uint8 block_id, SmgrId smgrid, RelFileNode *rnode, + ForkNumber forknum, BlockNumber blknum, Page page, + uint8 flags) { registered_buffer *regbuf; @@ -280,6 +284,7 @@ XLogRegisterBlock(uint8 block_id, RelFileNode *rnode, ForkNumber forknum, regbuf = ®istered_buffers[block_id]; + regbuf->smgrid = smgrid; regbuf->rnode = *rnode; regbuf->forkno = forknum; regbuf->block = blknum; @@ -303,7 +308,8 @@ XLogRegisterBlock(uint8 block_id, RelFileNode *rnode, ForkNumber forknum, if (i == block_id || !regbuf_old->in_use) continue; - Assert(!RelFileNodeEquals(regbuf_old->rnode, regbuf->rnode) || + Assert(regbuf_old->smgrid != regbuf->smgrid || + !RelFileNodeEquals(regbuf_old->rnode, regbuf->rnode) || regbuf_old->forkno != regbuf->forkno || regbuf_old->block != regbuf->block); } @@ -702,7 +708,8 @@ XLogRecordAssemble(RmgrId rmid, uint8 info, rdt_datas_last = regbuf->rdata_tail; } - if (prev_regbuf && RelFileNodeEquals(regbuf->rnode, prev_regbuf->rnode)) + if (prev_regbuf && regbuf->smgrid == prev_regbuf->smgrid && + RelFileNodeEquals(regbuf->rnode, prev_regbuf->rnode)) { samerel = true; bkpb.fork_flags |= BKPBLOCK_SAME_REL; @@ -727,6 +734,8 @@ XLogRecordAssemble(RmgrId rmid, uint8 info, } if (!samerel) { + memcpy(scratch, ®buf->smgrid, sizeof(SmgrId)); + scratch += sizeof(SmgrId); memcpy(scratch, ®buf->rnode, sizeof(RelFileNode)); scratch += sizeof(RelFileNode); } @@ -919,6 +928,7 @@ XLogSaveBufferForHint(Buffer buffer, bool buffer_std) int flags; PGAlignedBlock copied_buffer; char *origdata = (char *) BufferGetBlock(buffer); + SmgrId smgrid; RelFileNode rnode; ForkNumber forkno; BlockNumber blkno; @@ -947,8 +957,8 @@ XLogSaveBufferForHint(Buffer buffer, bool buffer_std) if (buffer_std) flags |= REGBUF_STANDARD; - BufferGetTag(buffer, &rnode, &forkno, &blkno); - XLogRegisterBlock(0, &rnode, forkno, blkno, copied_buffer.data, flags); + BufferGetTag(buffer, &smgrid, &rnode, &forkno, &blkno); + XLogRegisterBlock(0, smgrid, &rnode, forkno, blkno, copied_buffer.data, flags); recptr = XLogInsert(RM_XLOG_ID, XLOG_FPI_FOR_HINT); } @@ -969,8 +979,8 @@ XLogSaveBufferForHint(Buffer buffer, bool buffer_std) * the unused space to be left out from the WAL record, making it smaller. */ XLogRecPtr -log_newpage(RelFileNode *rnode, ForkNumber forkNum, BlockNumber blkno, - Page page, bool page_std) +log_newpage(SmgrId smgrid, RelFileNode *rnode, ForkNumber forkNum, + BlockNumber blkno, Page page, bool page_std) { int flags; XLogRecPtr recptr; @@ -980,7 +990,7 @@ log_newpage(RelFileNode *rnode, ForkNumber forkNum, BlockNumber blkno, flags |= REGBUF_STANDARD; XLogBeginInsert(); - XLogRegisterBlock(0, rnode, forkNum, blkno, page, flags); + XLogRegisterBlock(0, smgrid, rnode, forkNum, blkno, page, flags); recptr = XLogInsert(RM_XLOG_ID, XLOG_FPI); /* @@ -1009,6 +1019,7 @@ XLogRecPtr log_newpage_buffer(Buffer buffer, bool page_std) { Page page = BufferGetPage(buffer); + SmgrId smgrid; RelFileNode rnode; ForkNumber forkNum; BlockNumber blkno; @@ -1016,9 +1027,9 @@ log_newpage_buffer(Buffer buffer, bool page_std) /* Shared buffers should be modified in a critical section. */ Assert(CritSectionCount > 0); - BufferGetTag(buffer, &rnode, &forkNum, &blkno); + BufferGetTag(buffer, &smgrid, &rnode, &forkNum, &blkno); - return log_newpage(&rnode, forkNum, blkno, page, page_std); + return log_newpage(smgrid, &rnode, forkNum, blkno, page, page_std); } /* diff --git a/src/backend/access/transam/xlogreader.c b/src/backend/access/transam/xlogreader.c index 33ccfc15531..aeeb7790b7d 100644 --- a/src/backend/access/transam/xlogreader.c +++ b/src/backend/access/transam/xlogreader.c @@ -1047,6 +1047,7 @@ DecodeXLogRecord(XLogReaderState *state, XLogRecord *record, char **errormsg) uint32 remaining; uint32 datatotal; RelFileNode *rnode = NULL; + SmgrId smgrid = -1; uint8 block_id; ResetDecoder(state); @@ -1220,8 +1221,10 @@ DecodeXLogRecord(XLogReaderState *state, XLogRecord *record, char **errormsg) } if (!(fork_flags & BKPBLOCK_SAME_REL)) { + COPY_HEADER_FIELD(&blk->smgrid, sizeof(SmgrId)); COPY_HEADER_FIELD(&blk->rnode, sizeof(RelFileNode)); rnode = &blk->rnode; + smgrid = blk->smgrid; } else { @@ -1233,6 +1236,7 @@ DecodeXLogRecord(XLogReaderState *state, XLogRecord *record, char **errormsg) goto err; } + blk->smgrid = smgrid; blk->rnode = *rnode; } COPY_HEADER_FIELD(&blk->blkno, sizeof(BlockNumber)); @@ -1340,12 +1344,13 @@ err: /* * Returns information about the block that a block reference refers to. * - * If the WAL record contains a block reference with the given ID, *rnode, - * *forknum, and *blknum are filled in (if not NULL), and returns true. + * If the WAL record contains a block reference with the given ID, *smgrid, + * *rnode, *forknum and *blknum are filled in (if not NULL), and returns true. * Otherwise returns false. */ bool XLogRecGetBlockTag(XLogReaderState *record, uint8 block_id, + SmgrId *smgrid, RelFileNode *rnode, ForkNumber *forknum, BlockNumber *blknum) { DecodedBkpBlock *bkpb; @@ -1354,6 +1359,8 @@ XLogRecGetBlockTag(XLogReaderState *record, uint8 block_id, return false; bkpb = &record->blocks[block_id]; + if (smgrid) + *smgrid = bkpb->smgrid; if (rnode) *rnode = bkpb->rnode; if (forknum) diff --git a/src/backend/access/transam/xlogutils.c b/src/backend/access/transam/xlogutils.c index 10a663bae62..c5f27fb0e17 100644 --- a/src/backend/access/transam/xlogutils.c +++ b/src/backend/access/transam/xlogutils.c @@ -335,8 +335,9 @@ XLogReadBufferForRedoExtended(XLogReaderState *record, Page page; bool zeromode; bool willinit; + SmgrId smgrid; - if (!XLogRecGetBlockTag(record, block_id, &rnode, &forknum, &blkno)) + if (!XLogRecGetBlockTag(record, block_id, &smgrid, &rnode, &forknum, &blkno)) { /* Caller specified a bogus block_id */ elog(PANIC, "failed to locate backup block with ID %d", block_id); @@ -357,7 +358,7 @@ XLogReadBufferForRedoExtended(XLogReaderState *record, if (XLogRecBlockImageApply(record, block_id)) { Assert(XLogRecHasBlockImage(record, block_id)); - *buf = XLogReadBufferExtended(rnode, forknum, blkno, + *buf = XLogReadBufferExtended(smgrid, rnode, forknum, blkno, get_cleanup_lock ? RBM_ZERO_AND_CLEANUP_LOCK : RBM_ZERO_AND_LOCK); page = BufferGetPage(*buf); if (!RestoreBlockImage(record, block_id, page)) @@ -387,7 +388,7 @@ XLogReadBufferForRedoExtended(XLogReaderState *record, } else { - *buf = XLogReadBufferExtended(rnode, forknum, blkno, mode); + *buf = XLogReadBufferExtended(smgrid, rnode, forknum, blkno, mode); if (BufferIsValid(*buf)) { if (mode != RBM_ZERO_AND_LOCK && mode != RBM_ZERO_AND_CLEANUP_LOCK) @@ -434,7 +435,7 @@ XLogReadBufferForRedoExtended(XLogReaderState *record, * modified. */ Buffer -XLogReadBufferExtended(RelFileNode rnode, ForkNumber forknum, +XLogReadBufferExtended(SmgrId smgrid, RelFileNode rnode, ForkNumber forknum, BlockNumber blkno, ReadBufferMode mode) { BlockNumber lastblock; @@ -444,7 +445,7 @@ XLogReadBufferExtended(RelFileNode rnode, ForkNumber forknum, Assert(blkno != P_NEW); /* Open the relation at smgr level */ - smgr = smgropen(rnode, InvalidBackendId); + smgr = smgropen(smgrid, rnode, InvalidBackendId); /* * Create the target file if it doesn't already exist. This lets us cope @@ -461,7 +462,7 @@ XLogReadBufferExtended(RelFileNode rnode, ForkNumber forknum, if (blkno < lastblock) { /* page exists in file */ - buffer = ReadBufferWithoutRelcache(rnode, forknum, blkno, + buffer = ReadBufferWithoutRelcache(smgrid, rnode, forknum, blkno, mode, NULL); } else @@ -486,7 +487,7 @@ XLogReadBufferExtended(RelFileNode rnode, ForkNumber forknum, LockBuffer(buffer, BUFFER_LOCK_UNLOCK); ReleaseBuffer(buffer); } - buffer = ReadBufferWithoutRelcache(rnode, forknum, + buffer = ReadBufferWithoutRelcache(smgrid, rnode, forknum, P_NEW, mode, NULL); } while (BufferGetBlockNumber(buffer) < blkno); @@ -496,7 +497,7 @@ XLogReadBufferExtended(RelFileNode rnode, ForkNumber forknum, if (mode == RBM_ZERO_AND_LOCK || mode == RBM_ZERO_AND_CLEANUP_LOCK) LockBuffer(buffer, BUFFER_LOCK_UNLOCK); ReleaseBuffer(buffer); - buffer = ReadBufferWithoutRelcache(rnode, forknum, blkno, + buffer = ReadBufferWithoutRelcache(smgrid, rnode, forknum, blkno, mode, NULL); } } diff --git a/src/backend/catalog/storage.c b/src/backend/catalog/storage.c index 3cc886f7fe2..9509c197786 100644 --- a/src/backend/catalog/storage.c +++ b/src/backend/catalog/storage.c @@ -102,7 +102,7 @@ RelationCreateStorage(RelFileNode rnode, char relpersistence) return NULL; /* placate compiler */ } - srel = smgropen(rnode, backend); + srel = smgropen(SMGR_MD, rnode, backend); smgrcreate(srel, MAIN_FORKNUM, false); if (needs_wal) @@ -353,7 +353,8 @@ RelationCopyStorage(SMgrRelation src, SMgrRelation dst, * space. */ if (use_wal) - log_newpage(&dst->smgr_rnode.node, forkNum, blkno, page, false); + log_newpage(SMGR_MD, &dst->smgr_rnode.node, forkNum, blkno, page, + false); PageSetChecksumInplace(page, blkno); @@ -428,7 +429,7 @@ smgrDoPendingDeletes(bool isCommit) { SMgrRelation srel; - srel = smgropen(pending->relnode, pending->backend); + srel = smgropen(SMGR_MD, pending->relnode, pending->backend); /* allocate the initial array, or extend it, if needed */ if (maxrels == 0) @@ -580,7 +581,7 @@ smgr_redo(XLogReaderState *record) xl_smgr_create *xlrec = (xl_smgr_create *) XLogRecGetData(record); SMgrRelation reln; - reln = smgropen(xlrec->rnode, InvalidBackendId); + reln = smgropen(SMGR_MD, xlrec->rnode, InvalidBackendId); smgrcreate(reln, xlrec->forkNum, true); } else if (info == XLOG_SMGR_TRUNCATE) @@ -589,7 +590,7 @@ smgr_redo(XLogReaderState *record) SMgrRelation reln; Relation rel; - reln = smgropen(xlrec->rnode, InvalidBackendId); + reln = smgropen(SMGR_MD, xlrec->rnode, InvalidBackendId); /* * Forcibly create relation if it doesn't exist (which suggests that diff --git a/src/backend/commands/tablecmds.c b/src/backend/commands/tablecmds.c index 0f1a9f0e548..2896df0d2ba 100644 --- a/src/backend/commands/tablecmds.c +++ b/src/backend/commands/tablecmds.c @@ -12623,7 +12623,7 @@ index_copy_data(Relation rel, RelFileNode newrnode) { SMgrRelation dstrel; - dstrel = smgropen(newrnode, rel->rd_backend); + dstrel = smgropen(SMGR_MD, newrnode, rel->rd_backend); RelationOpenSmgr(rel); /* diff --git a/src/backend/replication/logical/decode.c b/src/backend/replication/logical/decode.c index 151c3ef8825..3e96d2a0752 100644 --- a/src/backend/replication/logical/decode.c +++ b/src/backend/replication/logical/decode.c @@ -683,7 +683,7 @@ DecodeInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf) return; /* only interested in our database */ - XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL); + XLogRecGetBlockTag(r, 0, NULL, &target_node, NULL, NULL); if (target_node.dbNode != ctx->slot->data.database) return; @@ -731,7 +731,7 @@ DecodeUpdate(LogicalDecodingContext *ctx, XLogRecordBuffer *buf) xlrec = (xl_heap_update *) XLogRecGetData(r); /* only interested in our database */ - XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL); + XLogRecGetBlockTag(r, 0, NULL, &target_node, NULL, NULL); if (target_node.dbNode != ctx->slot->data.database) return; @@ -796,7 +796,7 @@ DecodeDelete(LogicalDecodingContext *ctx, XLogRecordBuffer *buf) xlrec = (xl_heap_delete *) XLogRecGetData(r); /* only interested in our database */ - XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL); + XLogRecGetBlockTag(r, 0, NULL, &target_node, NULL, NULL); if (target_node.dbNode != ctx->slot->data.database) return; @@ -892,7 +892,7 @@ DecodeMultiInsert(LogicalDecodingContext *ctx, XLogRecordBuffer *buf) xlrec = (xl_heap_multi_insert *) XLogRecGetData(r); /* only interested in our database */ - XLogRecGetBlockTag(r, 0, &rnode, NULL, NULL); + XLogRecGetBlockTag(r, 0, NULL, &rnode, NULL, NULL); if (rnode.dbNode != ctx->slot->data.database) return; @@ -991,7 +991,7 @@ DecodeSpecConfirm(LogicalDecodingContext *ctx, XLogRecordBuffer *buf) RelFileNode target_node; /* only interested in our database */ - XLogRecGetBlockTag(r, 0, &target_node, NULL, NULL); + XLogRecGetBlockTag(r, 0, NULL, &target_node, NULL, NULL); if (target_node.dbNode != ctx->slot->data.database) return; diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c index 591377d2cd7..5cd44970110 100644 --- a/src/backend/replication/logical/reorderbuffer.c +++ b/src/backend/replication/logical/reorderbuffer.c @@ -3496,6 +3496,7 @@ ResolveCminCmaxDuringDecoding(HTAB *tuplecid_data, ReorderBufferTupleCidEnt *ent; ForkNumber forkno; BlockNumber blockno; + SmgrId smgrid; bool updated_mapping = false; /* be careful about padding */ @@ -3507,7 +3508,7 @@ ResolveCminCmaxDuringDecoding(HTAB *tuplecid_data, * get relfilenode from the buffer, no convenient way to access it other * than that. */ - BufferGetTag(buffer, &key.relnode, &forkno, &blockno); + BufferGetTag(buffer, &smgrid, &key.relnode, &forkno, &blockno); /* tuples can only be in the main fork */ Assert(forkno == MAIN_FORKNUM); diff --git a/src/backend/storage/buffer/bufmgr.c b/src/backend/storage/buffer/bufmgr.c index 7332e6b5903..8046334c253 100644 --- a/src/backend/storage/buffer/bufmgr.c +++ b/src/backend/storage/buffer/bufmgr.c @@ -555,7 +555,8 @@ PrefetchBuffer(Relation reln, ForkNumber forkNum, BlockNumber blockNum) int buf_id; /* create a tag so we can lookup the buffer */ - INIT_BUFFERTAG(newTag, reln->rd_smgr->smgr_rnode.node, + INIT_BUFFERTAG(newTag, reln->rd_smgr->smgr_which, + reln->rd_smgr->smgr_rnode.node, forkNum, blockNum); /* determine its hash code and partition lock ID */ @@ -680,13 +681,13 @@ ReadBufferExtended(Relation reln, ForkNumber forkNum, BlockNumber blockNum, * parameters. */ Buffer -ReadBufferWithoutRelcache(RelFileNode rnode, ForkNumber forkNum, +ReadBufferWithoutRelcache(SmgrId smgrid, RelFileNode rnode, ForkNumber forkNum, BlockNumber blockNum, ReadBufferMode mode, BufferAccessStrategy strategy) { bool hit; - SMgrRelation smgr = smgropen(rnode, InvalidBackendId); + SMgrRelation smgr = smgropen(smgrid, rnode, InvalidBackendId); Assert(InRecovery); @@ -1009,7 +1010,8 @@ BufferAlloc(SMgrRelation smgr, char relpersistence, ForkNumber forkNum, uint32 buf_state; /* create a tag so we can lookup the buffer */ - INIT_BUFFERTAG(newTag, smgr->smgr_rnode.node, forkNum, blockNum); + INIT_BUFFERTAG(newTag, smgr->smgr_which, + smgr->smgr_rnode.node, forkNum, blockNum); /* determine its hash code and partition lock ID */ newHash = BufTableHashCode(&newTag); @@ -1843,6 +1845,7 @@ BufferSync(int flags) buf_state |= BM_CHECKPOINT_NEEDED; item = &CkptBufferIds[num_to_scan++]; + item->smgrid = bufHdr->tag.smgrid; item->buf_id = buf_id; item->tsId = bufHdr->tag.rnode.spcNode; item->relNode = bufHdr->tag.rnode.relNode; @@ -2626,12 +2629,12 @@ BufferGetBlockNumber(Buffer buffer) /* * BufferGetTag - * Returns the relfilenode, fork number and block number associated with - * a buffer. + * Returns the SMGR ID, relfilenode, fork number and block number + * associated with a buffer. */ void -BufferGetTag(Buffer buffer, RelFileNode *rnode, ForkNumber *forknum, - BlockNumber *blknum) +BufferGetTag(Buffer buffer, SmgrId *smgrid, RelFileNode *rnode, + ForkNumber *forknum, BlockNumber *blknum) { BufferDesc *bufHdr; @@ -2644,6 +2647,7 @@ BufferGetTag(Buffer buffer, RelFileNode *rnode, ForkNumber *forknum, bufHdr = GetBufferDescriptor(buffer - 1); /* pinned, so OK to read tag without spinlock */ + *smgrid = bufHdr->tag.smgrid; *rnode = bufHdr->tag.rnode; *forknum = bufHdr->tag.forkNum; *blknum = bufHdr->tag.blockNum; @@ -2695,7 +2699,7 @@ FlushBuffer(BufferDesc *buf, SMgrRelation reln) /* Find smgr relation for buffer */ if (reln == NULL) - reln = smgropen(buf->tag.rnode, InvalidBackendId); + reln = smgropen(buf->tag.smgrid, buf->tag.rnode, InvalidBackendId); TRACE_POSTGRESQL_BUFFER_FLUSH_START(buf->tag.forkNum, buf->tag.blockNum, @@ -4220,6 +4224,11 @@ ckpt_buforder_comparator(const void *pa, const void *pb) const CkptSortItem *a = (const CkptSortItem *) pa; const CkptSortItem *b = (const CkptSortItem *) pb; + /* compare smgr */ + if (a->smgrid < b->smgrid) + return -1; + else if (a->smgrid > b->smgrid) + return 1; /* compare tablespace */ if (a->tsId < b->tsId) return -1; @@ -4377,7 +4386,7 @@ IssuePendingWritebacks(WritebackContext *context) i += ahead; /* and finally tell the kernel to write the data to storage */ - reln = smgropen(tag.rnode, InvalidBackendId); + reln = smgropen(tag.smgrid, tag.rnode, InvalidBackendId); smgrwriteback(reln, tag.forkNum, tag.blockNum, nblocks); } diff --git a/src/backend/storage/buffer/localbuf.c b/src/backend/storage/buffer/localbuf.c index f5f6a29222b..79a185e034f 100644 --- a/src/backend/storage/buffer/localbuf.c +++ b/src/backend/storage/buffer/localbuf.c @@ -68,7 +68,8 @@ LocalPrefetchBuffer(SMgrRelation smgr, ForkNumber forkNum, BufferTag newTag; /* identity of requested block */ LocalBufferLookupEnt *hresult; - INIT_BUFFERTAG(newTag, smgr->smgr_rnode.node, forkNum, blockNum); + INIT_BUFFERTAG(newTag, smgr->smgr_which, + smgr->smgr_rnode.node, forkNum, blockNum); /* Initialize local buffers if first request in this session */ if (LocalBufHash == NULL) @@ -111,7 +112,8 @@ LocalBufferAlloc(SMgrRelation smgr, ForkNumber forkNum, BlockNumber blockNum, bool found; uint32 buf_state; - INIT_BUFFERTAG(newTag, smgr->smgr_rnode.node, forkNum, blockNum); + INIT_BUFFERTAG(newTag, smgr->smgr_which, + smgr->smgr_rnode.node, forkNum, blockNum); /* Initialize local buffers if first request in this session */ if (LocalBufHash == NULL) @@ -209,7 +211,7 @@ LocalBufferAlloc(SMgrRelation smgr, ForkNumber forkNum, BlockNumber blockNum, Page localpage = (char *) LocalBufHdrGetBlock(bufHdr); /* Find smgr relation for buffer */ - oreln = smgropen(bufHdr->tag.rnode, MyBackendId); + oreln = smgropen(bufHdr->tag.smgrid, bufHdr->tag.rnode, MyBackendId); PageSetChecksumInplace(localpage, bufHdr->tag.blockNum); diff --git a/src/backend/storage/freespace/freespace.c b/src/backend/storage/freespace/freespace.c index c17b3f49dd0..78a2274e55c 100644 --- a/src/backend/storage/freespace/freespace.c +++ b/src/backend/storage/freespace/freespace.c @@ -210,7 +210,8 @@ XLogRecordPageWithFreeSpace(RelFileNode rnode, BlockNumber heapBlk, blkno = fsm_logical_to_physical(addr); /* If the page doesn't exist already, extend */ - buf = XLogReadBufferExtended(rnode, FSM_FORKNUM, blkno, RBM_ZERO_ON_ERROR); + buf = XLogReadBufferExtended(SMGR_MD, rnode, FSM_FORKNUM, blkno, + RBM_ZERO_ON_ERROR); LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE); page = BufferGetPage(buf); diff --git a/src/backend/storage/freespace/fsmpage.c b/src/backend/storage/freespace/fsmpage.c index cf7f03f12dd..da3b286ca6c 100644 --- a/src/backend/storage/freespace/fsmpage.c +++ b/src/backend/storage/freespace/fsmpage.c @@ -268,11 +268,12 @@ restart: * * Fix the corruption and restart. */ + SmgrId smgrid; RelFileNode rnode; ForkNumber forknum; BlockNumber blknum; - BufferGetTag(buf, &rnode, &forknum, &blknum); + BufferGetTag(buf, &smgrid, &rnode, &forknum, &blknum); elog(DEBUG1, "fixing corrupt FSM block %u, relation %u/%u/%u", blknum, rnode.spcNode, rnode.dbNode, rnode.relNode); diff --git a/src/backend/storage/smgr/md.c b/src/backend/storage/smgr/md.c index 58c94e9257a..b2c42cf8f0a 100644 --- a/src/backend/storage/smgr/md.c +++ b/src/backend/storage/smgr/md.c @@ -120,7 +120,7 @@ static MemoryContext MdCxt; /* context for all MdfdVec objects */ /* local routines */ static void mdunlinkfork(RelFileNodeBackend rnode, ForkNumber forkNum, bool isRedo); -static MdfdVec *mdopen(SMgrRelation reln, ForkNumber forknum, int behavior); +static MdfdVec *mdopenfork(SMgrRelation reln, ForkNumber forknum, int behavior); static void register_dirty_segment(SMgrRelation reln, ForkNumber forknum, MdfdVec *seg); static void register_unlink_segment(RelFileNodeBackend rnode, ForkNumber forknum, @@ -151,6 +151,17 @@ mdinit(void) ALLOCSET_DEFAULT_SIZES); } +/* + * mdopen() -- Initialize a newly-opened relation. + */ +void +mdopen(SMgrRelation reln) +{ + /* mark it not open */ + for (int forknum = 0; forknum <= MAX_FORKNUM; forknum++) + reln->md_num_open_segs[forknum] = 0; +} + /* * mdexists() -- Does the physical file exist? * @@ -165,7 +176,7 @@ mdexists(SMgrRelation reln, ForkNumber forkNum) */ mdclose(reln, forkNum); - return (mdopen(reln, forkNum, EXTENSION_RETURN_NULL) != NULL); + return (mdopenfork(reln, forkNum, EXTENSION_RETURN_NULL) != NULL); } /* @@ -425,7 +436,7 @@ mdextend(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum, } /* - * mdopen() -- Open the specified relation. + * mdopenfork() -- Open the specified relation. * * Note we only open the first segment, when there are multiple segments. * @@ -435,7 +446,7 @@ mdextend(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum, * invent one out of whole cloth. */ static MdfdVec * -mdopen(SMgrRelation reln, ForkNumber forknum, int behavior) +mdopenfork(SMgrRelation reln, ForkNumber forknum, int behavior) { MdfdVec *mdfd; char *path; @@ -713,11 +724,11 @@ mdwrite(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum, BlockNumber mdnblocks(SMgrRelation reln, ForkNumber forknum) { - MdfdVec *v = mdopen(reln, forknum, EXTENSION_FAIL); + MdfdVec *v = mdopenfork(reln, forknum, EXTENSION_FAIL); BlockNumber nblocks; BlockNumber segno = 0; - /* mdopen has opened the first segment */ + /* mdopenfork has opened the first segment */ Assert(reln->md_num_open_segs[forknum] > 0); /* @@ -981,7 +992,7 @@ DropRelationFiles(RelFileNode *delrels, int ndelrels, bool isRedo) srels = palloc(sizeof(SMgrRelation) * ndelrels); for (i = 0; i < ndelrels; i++) { - SMgrRelation srel = smgropen(delrels[i], InvalidBackendId); + SMgrRelation srel = smgropen(SMGR_MD, delrels[i], InvalidBackendId); if (isRedo) { @@ -1137,7 +1148,7 @@ _mdfd_getseg(SMgrRelation reln, ForkNumber forknum, BlockNumber blkno, v = &reln->md_seg_fds[forknum][reln->md_num_open_segs[forknum] - 1]; else { - v = mdopen(reln, forknum, behavior); + v = mdopenfork(reln, forknum, behavior); if (!v) return NULL; /* if behavior & EXTENSION_RETURN_NULL */ } @@ -1254,7 +1265,7 @@ _mdnblocks(SMgrRelation reln, ForkNumber forknum, MdfdVec *seg) int mdsyncfiletag(const FileTag *ftag, char *path) { - SMgrRelation reln = smgropen(ftag->rnode, InvalidBackendId); + SMgrRelation reln = smgropen(SMGR_MD, ftag->rnode, InvalidBackendId); MdfdVec *v; char *p; diff --git a/src/backend/storage/smgr/smgr.c b/src/backend/storage/smgr/smgr.c index dba8c397feb..26281fab51d 100644 --- a/src/backend/storage/smgr/smgr.c +++ b/src/backend/storage/smgr/smgr.c @@ -41,6 +41,7 @@ typedef struct f_smgr { void (*smgr_init) (void); /* may be NULL */ void (*smgr_shutdown) (void); /* may be NULL */ + void (*smgr_open) (SMgrRelation reln); void (*smgr_close) (SMgrRelation reln, ForkNumber forknum); void (*smgr_create) (SMgrRelation reln, ForkNumber forknum, bool isRedo); @@ -68,6 +69,7 @@ static const f_smgr smgrsw[] = { { .smgr_init = mdinit, .smgr_shutdown = NULL, + .smgr_open = mdopen, .smgr_close = mdclose, .smgr_create = mdcreate, .smgr_exists = mdexists, @@ -141,7 +143,7 @@ smgrshutdown(int code, Datum arg) * This does not attempt to actually open the underlying file. */ SMgrRelation -smgropen(RelFileNode rnode, BackendId backend) +smgropen(SmgrId smgrid, RelFileNode rnode, BackendId backend) { RelFileNodeBackend brnode; SMgrRelation reln; @@ -170,18 +172,15 @@ smgropen(RelFileNode rnode, BackendId backend) /* Initialize it if not present before */ if (!found) { - int forknum; - /* hash_search already filled in the lookup key */ reln->smgr_owner = NULL; reln->smgr_targblock = InvalidBlockNumber; reln->smgr_fsm_nblocks = InvalidBlockNumber; reln->smgr_vm_nblocks = InvalidBlockNumber; - reln->smgr_which = 0; /* we only have md.c at present */ + reln->smgr_which = smgrid; - /* mark it not open */ - for (forknum = 0; forknum <= MAX_FORKNUM; forknum++) - reln->md_num_open_segs[forknum] = 0; + /* implementation-specific initialization */ + smgrsw[reln->smgr_which].smgr_open(reln); /* it has no owner yet */ dlist_push_tail(&unowned_relns, &reln->node); diff --git a/src/bin/pg_rewind/parsexlog.c b/src/bin/pg_rewind/parsexlog.c index 287af60c4e7..89c183abc61 100644 --- a/src/bin/pg_rewind/parsexlog.c +++ b/src/bin/pg_rewind/parsexlog.c @@ -394,8 +394,14 @@ extractPageInfo(XLogReaderState *record) RelFileNode rnode; ForkNumber forknum; BlockNumber blkno; + SmgrId smgrid; - if (!XLogRecGetBlockTag(record, block_id, &rnode, &forknum, &blkno)) + if (!XLogRecGetBlockTag(record, block_id, &smgrid, &rnode, &forknum, + &blkno)) + continue; + + /* We only care about SMGR_MD for now */ + if (smgrid != SMGR_MD) continue; /* We only care about the main fork; others are copied in toto */ diff --git a/src/bin/pg_waldump/pg_waldump.c b/src/bin/pg_waldump/pg_waldump.c index b95d467805a..9b9b4502b3b 100644 --- a/src/bin/pg_waldump/pg_waldump.c +++ b/src/bin/pg_waldump/pg_waldump.c @@ -524,6 +524,7 @@ XLogDumpDisplayRecord(XLogDumpConfig *config, XLogReaderState *record) const RmgrDescData *desc = &RmgrDescTable[XLogRecGetRmid(record)]; uint32 rec_len; uint32 fpi_len; + SmgrId smgrid; RelFileNode rnode; ForkNumber forknum; BlockNumber blk; @@ -556,16 +557,19 @@ XLogDumpDisplayRecord(XLogDumpConfig *config, XLogReaderState *record) if (!XLogRecHasBlockRef(record, block_id)) continue; - XLogRecGetBlockTag(record, block_id, &rnode, &forknum, &blk); + XLogRecGetBlockTag(record, block_id, &smgrid, &rnode, &forknum, + &blk); if (forknum != MAIN_FORKNUM) - printf(", blkref #%u: rel %u/%u/%u fork %s blk %u", + printf(", blkref #%u: smgr %d rel %u/%u/%u fork %s blk %u", block_id, + smgrid, rnode.spcNode, rnode.dbNode, rnode.relNode, forkNames[forknum], blk); else - printf(", blkref #%u: rel %u/%u/%u blk %u", + printf(", blkref #%u: smgr %d rel %u/%u/%u blk %u", block_id, + smgrid, rnode.spcNode, rnode.dbNode, rnode.relNode, blk); if (XLogRecHasBlockImage(record, block_id)) @@ -587,9 +591,11 @@ XLogDumpDisplayRecord(XLogDumpConfig *config, XLogReaderState *record) if (!XLogRecHasBlockRef(record, block_id)) continue; - XLogRecGetBlockTag(record, block_id, &rnode, &forknum, &blk); - printf("\tblkref #%u: rel %u/%u/%u fork %s blk %u", + XLogRecGetBlockTag(record, block_id, &smgrid, &rnode, &forknum, + &blk); + printf("\tblkref #%u: smgr %d rel %u/%u/%u fork %s blk %u", block_id, + smgrid, rnode.spcNode, rnode.dbNode, rnode.relNode, forkNames[forknum], blk); diff --git a/src/include/access/xloginsert.h b/src/include/access/xloginsert.h index df24089ea45..3ae9d225bbf 100644 --- a/src/include/access/xloginsert.h +++ b/src/include/access/xloginsert.h @@ -16,6 +16,7 @@ #include "storage/block.h" #include "storage/buf.h" #include "storage/relfilenode.h" +#include "storage/smgr.h" #include "utils/relcache.h" /* @@ -45,15 +46,17 @@ extern XLogRecPtr XLogInsert(RmgrId rmid, uint8 info); extern void XLogEnsureRecordSpace(int nbuffers, int ndatas); extern void XLogRegisterData(char *data, int len); extern void XLogRegisterBuffer(uint8 block_id, Buffer buffer, uint8 flags); -extern void XLogRegisterBlock(uint8 block_id, RelFileNode *rnode, - ForkNumber forknum, BlockNumber blknum, char *page, - uint8 flags); +extern void XLogRegisterBlock(uint8 block_id, SmgrId smgrid, + RelFileNode *rnode, ForkNumber forknum, + BlockNumber blknum, + char *page, uint8 flags); extern void XLogRegisterBufData(uint8 block_id, char *data, int len); extern void XLogResetInsertion(void); extern bool XLogCheckBufferNeedsBackup(Buffer buffer); -extern XLogRecPtr log_newpage(RelFileNode *rnode, ForkNumber forkNum, - BlockNumber blk, char *page, bool page_std); +extern XLogRecPtr log_newpage(SmgrId smgrid, RelFileNode *rnode, + ForkNumber forkNum, BlockNumber blk, + char *page, bool page_std); extern XLogRecPtr log_newpage_buffer(Buffer buffer, bool page_std); extern void log_newpage_range(Relation rel, ForkNumber forkNum, BlockNumber startblk, BlockNumber endblk, bool page_std); diff --git a/src/include/access/xlogreader.h b/src/include/access/xlogreader.h index a12c94cba67..6ec50b4151e 100644 --- a/src/include/access/xlogreader.h +++ b/src/include/access/xlogreader.h @@ -30,6 +30,7 @@ #endif #include "access/xlogrecord.h" +#include "storage/smgr.h" typedef struct XLogReaderState XLogReaderState; @@ -47,6 +48,7 @@ typedef struct bool in_use; /* Identify the block this refers to */ + SmgrId smgrid; RelFileNode rnode; ForkNumber forknum; BlockNumber blkno; @@ -251,7 +253,7 @@ extern FullTransactionId XLogRecGetFullXid(XLogReaderState *record); extern bool RestoreBlockImage(XLogReaderState *recoder, uint8 block_id, char *dst); extern char *XLogRecGetBlockData(XLogReaderState *record, uint8 block_id, Size *len); extern bool XLogRecGetBlockTag(XLogReaderState *record, uint8 block_id, - RelFileNode *rnode, ForkNumber *forknum, - BlockNumber *blknum); + SmgrId *smgrid, RelFileNode *rnode, + ForkNumber *forknum, BlockNumber *blknum); #endif /* XLOGREADER_H */ diff --git a/src/include/access/xlogutils.h b/src/include/access/xlogutils.h index 4105b59904b..366b8d32812 100644 --- a/src/include/access/xlogutils.h +++ b/src/include/access/xlogutils.h @@ -13,6 +13,7 @@ #include "access/xlogreader.h" #include "storage/bufmgr.h" +#include "storage/smgr.h" extern bool XLogHaveInvalidPages(void); @@ -41,7 +42,8 @@ extern XLogRedoAction XLogReadBufferForRedoExtended(XLogReaderState *record, ReadBufferMode mode, bool get_cleanup_lock, Buffer *buf); -extern Buffer XLogReadBufferExtended(RelFileNode rnode, ForkNumber forknum, +extern Buffer XLogReadBufferExtended(SmgrId smgrid, RelFileNode rnode, + ForkNumber forknum, BlockNumber blkno, ReadBufferMode mode); extern Relation CreateFakeRelcacheEntry(RelFileNode rnode); diff --git a/src/include/storage/buf_internals.h b/src/include/storage/buf_internals.h index df2dda7e7e7..e0d7e8f9fa3 100644 --- a/src/include/storage/buf_internals.h +++ b/src/include/storage/buf_internals.h @@ -87,16 +87,20 @@ * * Note: if there's any pad bytes in the struct, INIT_BUFFERTAG will have * to be fixed to zero them, since this struct is used as a hash key. + * Conceptually the SmgrId should go first, but we put it next to the + * ForkNumber so that it packs better with typical alignment rules. */ typedef struct buftag { RelFileNode rnode; /* physical relation identifier */ - ForkNumber forkNum; + int16 smgrid; /* SmgrId */ + int16 forkNum; /* ForkNumber */ BlockNumber blockNum; /* blknum relative to begin of reln */ } BufferTag; #define CLEAR_BUFFERTAG(a) \ ( \ + (a).smgrid = SMGR_INVALID, \ (a).rnode.spcNode = InvalidOid, \ (a).rnode.dbNode = InvalidOid, \ (a).rnode.relNode = InvalidOid, \ @@ -104,8 +108,9 @@ typedef struct buftag (a).blockNum = InvalidBlockNumber \ ) -#define INIT_BUFFERTAG(a,xx_rnode,xx_forkNum,xx_blockNum) \ +#define INIT_BUFFERTAG(a,xx_smgrid,xx_rnode,xx_forkNum,xx_blockNum) \ ( \ + (a).smgrid = (xx_smgrid), \ (a).rnode = (xx_rnode), \ (a).forkNum = (xx_forkNum), \ (a).blockNum = (xx_blockNum) \ @@ -113,6 +118,7 @@ typedef struct buftag #define BUFFERTAGS_EQUAL(a,b) \ ( \ + (a).smgrid == (b).smgrid && \ RelFileNodeEquals((a).rnode, (b).rnode) && \ (a).blockNum == (b).blockNum && \ (a).forkNum == (b).forkNum \ @@ -288,6 +294,7 @@ extern BufferDesc *LocalBufferDescriptors; */ typedef struct CkptSortItem { + SmgrId smgrid; Oid tsId; Oid relNode; ForkNumber forkNum; diff --git a/src/include/storage/bufmgr.h b/src/include/storage/bufmgr.h index 509f4b7ef1c..013a7de338e 100644 --- a/src/include/storage/bufmgr.h +++ b/src/include/storage/bufmgr.h @@ -18,6 +18,7 @@ #include "storage/buf.h" #include "storage/bufpage.h" #include "storage/relfilenode.h" +#include "storage/smgr.h" #include "utils/relcache.h" #include "utils/snapmgr.h" @@ -168,9 +169,9 @@ extern Buffer ReadBuffer(Relation reln, BlockNumber blockNum); extern Buffer ReadBufferExtended(Relation reln, ForkNumber forkNum, BlockNumber blockNum, ReadBufferMode mode, BufferAccessStrategy strategy); -extern Buffer ReadBufferWithoutRelcache(RelFileNode rnode, - ForkNumber forkNum, BlockNumber blockNum, - ReadBufferMode mode, BufferAccessStrategy strategy); +extern Buffer ReadBufferWithoutRelcache(SmgrId smgrid, RelFileNode rnode, + ForkNumber forkNum, BlockNumber blockNum, + ReadBufferMode mode, BufferAccessStrategy strategy); extern void ReleaseBuffer(Buffer buffer); extern void UnlockReleaseBuffer(Buffer buffer); extern void MarkBufferDirty(Buffer buffer); @@ -205,7 +206,7 @@ extern XLogRecPtr BufferGetLSNAtomic(Buffer buffer); extern void PrintPinnedBufs(void); #endif extern Size BufferShmemSize(void); -extern void BufferGetTag(Buffer buffer, RelFileNode *rnode, +extern void BufferGetTag(Buffer buffer, SmgrId *smgrid, RelFileNode *rnode, ForkNumber *forknum, BlockNumber *blknum); extern void MarkBufferDirtyHint(Buffer buffer, bool buffer_std); diff --git a/src/include/storage/md.h b/src/include/storage/md.h index df24b931613..c0f05e23ff9 100644 --- a/src/include/storage/md.h +++ b/src/include/storage/md.h @@ -21,6 +21,7 @@ /* md storage manager functionality */ extern void mdinit(void); +extern void mdopen(SMgrRelation reln); extern void mdclose(SMgrRelation reln, ForkNumber forknum); extern void mdcreate(SMgrRelation reln, ForkNumber forknum, bool isRedo); extern bool mdexists(SMgrRelation reln, ForkNumber forknum); diff --git a/src/include/storage/smgr.h b/src/include/storage/smgr.h index d286c8c7b11..243efc6cdb1 100644 --- a/src/include/storage/smgr.h +++ b/src/include/storage/smgr.h @@ -79,8 +79,14 @@ typedef SMgrRelationData *SMgrRelation; #define SmgrIsTemp(smgr) \ RelFileNodeBackendIsTemp((smgr)->smgr_rnode) +typedef enum SmgrId +{ + SMGR_INVALID = -1, + SMGR_MD = 0, /* md.c */ +} SmgrId; + extern void smgrinit(void); -extern SMgrRelation smgropen(RelFileNode rnode, BackendId backend); +extern SMgrRelation smgropen(SmgrId which, RelFileNode rnode, BackendId backend); extern bool smgrexists(SMgrRelation reln, ForkNumber forknum); extern void smgrsetowner(SMgrRelation *owner, SMgrRelation reln); extern void smgrclearowner(SMgrRelation *owner, SMgrRelation reln); diff --git a/src/include/utils/rel.h b/src/include/utils/rel.h index d35b4a5061c..2b19683eb52 100644 --- a/src/include/utils/rel.h +++ b/src/include/utils/rel.h @@ -52,6 +52,7 @@ typedef LockInfoData *LockInfo; typedef struct RelationData { + SmgrId rd_smgrid; /* relation storage manager */ RelFileNode rd_node; /* relation physical identifier */ /* use "struct" here to avoid needing to include smgr.h: */ struct SMgrRelationData *rd_smgr; /* cached file handle, or NULL */ @@ -471,7 +472,10 @@ typedef struct ViewOptions #define RelationOpenSmgr(relation) \ do { \ if ((relation)->rd_smgr == NULL) \ - smgrsetowner(&((relation)->rd_smgr), smgropen((relation)->rd_node, (relation)->rd_backend)); \ + smgrsetowner(&((relation)->rd_smgr), \ + smgropen((relation)->rd_smgrid, \ + (relation)->rd_node, \ + (relation)->rd_backend)); \ } while (0) /* -- 2.21.0