xlogutils.c File Reference

#include "postgres.h"
#include <unistd.h>
#include "access/timeline.h"
#include "access/xlog.h"
#include "access/xlog_internal.h"
#include "access/xlogutils.h"
#include "miscadmin.h"
#include "pgstat.h"
#include "storage/smgr.h"
#include "utils/guc.h"
#include "utils/hsearch.h"
#include "utils/rel.h"

Include dependency graph for xlogutils.c:

Go to the source code of this file.

Data Structures
struct	xl_invalid_page_key
struct	xl_invalid_page
struct	FakeRelCacheEntryData

Typedefs
typedef struct xl_invalid_page_key	xl_invalid_page_key
typedef struct xl_invalid_page	xl_invalid_page
typedef FakeRelCacheEntryData *	FakeRelCacheEntry

Functions
static void	report_invalid_page (int elevel, RelFileNode node, ForkNumber forkno, BlockNumber blkno, bool present)
static void	log_invalid_page (RelFileNode node, ForkNumber forkno, BlockNumber blkno, bool present)
static void	forget_invalid_pages (RelFileNode node, ForkNumber forkno, BlockNumber minblkno)
static void	forget_invalid_pages_db (Oid dbid)
bool	XLogHaveInvalidPages (void)
void	XLogCheckInvalidPages (void)
XLogRedoAction	XLogReadBufferForRedo (XLogReaderState *record, uint8 block_id, Buffer *buf)
Buffer	XLogInitBufferForRedo (XLogReaderState *record, uint8 block_id)
XLogRedoAction	XLogReadBufferForRedoExtended (XLogReaderState *record, uint8 block_id, ReadBufferMode mode, bool get_cleanup_lock, Buffer *buf)
Buffer	XLogReadBufferExtended (RelFileNode rnode, ForkNumber forknum, BlockNumber blkno, ReadBufferMode mode, Buffer recent_buffer)
Relation	CreateFakeRelcacheEntry (RelFileNode rnode)
void	FreeFakeRelcacheEntry (Relation fakerel)
void	XLogDropRelation (RelFileNode rnode, ForkNumber forknum)
void	XLogDropDatabase (Oid dbid)
void	XLogTruncateRelation (RelFileNode rnode, ForkNumber forkNum, BlockNumber nblocks)
void	XLogReadDetermineTimeline (XLogReaderState *state, XLogRecPtr wantPage, uint32 wantLength)
void	wal_segment_open (XLogReaderState *state, XLogSegNo nextSegNo, TimeLineID *tli_p)
void	wal_segment_close (XLogReaderState *state)
bool	read_local_xlog_page (XLogReaderState *state)
void	WALReadRaiseError (WALReadError *errinfo)

Variables
bool	ignore_invalid_pages = false
static HTAB *	invalid_page_tab = NULL

Typedef Documentation

FakeRelCacheEntry

typedef FakeRelCacheEntryData* FakeRelCacheEntry

Definition at line 552 of file xlogutils.c.

xl_invalid_page

typedef struct xl_invalid_page xl_invalid_page

xl_invalid_page_key

typedef struct xl_invalid_page_key xl_invalid_page_key

Function Documentation

CreateFakeRelcacheEntry()

Relation CreateFakeRelcacheEntry

(

RelFileNode

rnode

)

Definition at line 569 of file xlogutils.c.

References LockRelId::dbId, RelFileNode::dbNode, InvalidBackendId, LockInfoData::lockRelId, palloc0(), FakeRelCacheEntryData::pgc, RelationData::rd_backend, RelationData::rd_lockInfo, RelationData::rd_node, RelationData::rd_rel, RelationData::rd_smgr, RelationGetRelationName, LockRelId::relId, RelFileNode::relNode, and sprintf.

Referenced by heap_xlog_delete(), heap_xlog_insert(), heap_xlog_lock(), heap_xlog_lock_updated(), heap_xlog_multi_insert(), heap_xlog_update(), heap_xlog_visible(), smgr_redo(), and smgrDoPendingSyncs().

{
    FakeRelCacheEntry fakeentry;
    Relation    rel;

    /* Allocate the Relation struct and all related space in one block. */
    fakeentry = palloc0(sizeof(FakeRelCacheEntryData));
    rel = (Relation) fakeentry;

    rel->rd_rel = &fakeentry->pgc;
    rel->rd_node = rnode;

    /*
     * We will never be working with temp rels during recovery or while
     * syncing WAL-skipped files.
     */
    rel->rd_backend = InvalidBackendId;

    /* It must be a permanent table here */
    rel->rd_rel->relpersistence = RELPERSISTENCE_PERMANENT;

    /* We don't know the name of the relation; use relfilenode instead */
    sprintf(RelationGetRelationName(rel), "%u", rnode.relNode);

    /*
     * We set up the lockRelId in case anything tries to lock the dummy
     * relation.  Note that this is fairly bogus since relNode may be
     * different from the relation's OID.  It shouldn't really matter though.
     * In recovery, we are running by ourselves and can't have any lock
     * conflicts.  While syncing, we already hold AccessExclusiveLock.
     */
    rel->rd_lockInfo.lockRelId.dbId = rnode.dbNode;
    rel->rd_lockInfo.lockRelId.relId = rnode.relNode;

    rel->rd_smgr = NULL;

    return rel;
}

forget_invalid_pages()

static void forget_invalid_pages ( RelFileNode  node, ForkNumber  forkno, BlockNumber  minblkno  )

static

Definition at line 145 of file xlogutils.c.

References xl_invalid_page_key::blkno, DEBUG2, elog, ERROR, xl_invalid_page_key::forkno, HASH_REMOVE, hash_search(), hash_seq_init(), hash_seq_search(), xl_invalid_page::key, message_level_is_interesting(), xl_invalid_page_key::node, pfree(), RelFileNodeEquals, relpathperm, and status().

Referenced by XLogDropRelation(), and XLogTruncateRelation().

{
    HASH_SEQ_STATUS status;
    xl_invalid_page *hentry;

    if (invalid_page_tab == NULL)
        return;                 /* nothing to do */

    hash_seq_init(&status, invalid_page_tab);

    while ((hentry = (xl_invalid_page *) hash_seq_search(&status)) != NULL)
    {
        if (RelFileNodeEquals(hentry->key.node, node) &&
            hentry->key.forkno == forkno &&
            hentry->key.blkno >= minblkno)
        {
            if (message_level_is_interesting(DEBUG2))
            {
                char       *path = relpathperm(hentry->key.node, forkno);

                elog(DEBUG2, "page %u of relation %s has been dropped",
                     hentry->key.blkno, path);
                pfree(path);
            }

            if (hash_search(invalid_page_tab,
                            (void *) &hentry->key,
                            HASH_REMOVE, NULL) == NULL)
                elog(ERROR, "hash table corrupted");
        }
    }
}

forget_invalid_pages_db()

static void forget_invalid_pages_db ( Oid  dbid )

static

Definition at line 180 of file xlogutils.c.

References xl_invalid_page_key::blkno, RelFileNode::dbNode, DEBUG2, elog, ERROR, xl_invalid_page_key::forkno, HASH_REMOVE, hash_search(), hash_seq_init(), hash_seq_search(), xl_invalid_page::key, message_level_is_interesting(), xl_invalid_page_key::node, pfree(), relpathperm, and status().

Referenced by XLogDropDatabase().

{
    HASH_SEQ_STATUS status;
    xl_invalid_page *hentry;

    if (invalid_page_tab == NULL)
        return;                 /* nothing to do */

    hash_seq_init(&status, invalid_page_tab);

    while ((hentry = (xl_invalid_page *) hash_seq_search(&status)) != NULL)
    {
        if (hentry->key.node.dbNode == dbid)
        {
            if (message_level_is_interesting(DEBUG2))
            {
                char       *path = relpathperm(hentry->key.node, hentry->key.forkno);

                elog(DEBUG2, "page %u of relation %s has been dropped",
                     hentry->key.blkno, path);
                pfree(path);
            }

            if (hash_search(invalid_page_tab,
                            (void *) &hentry->key,
                            HASH_REMOVE, NULL) == NULL)
                elog(ERROR, "hash table corrupted");
        }
    }
}

FreeFakeRelcacheEntry()

void FreeFakeRelcacheEntry

(

Relation

fakerel

)

Definition at line 612 of file xlogutils.c.

References pfree(), RelationData::rd_smgr, and smgrclearowner().

Referenced by heap_xlog_delete(), heap_xlog_insert(), heap_xlog_lock(), heap_xlog_lock_updated(), heap_xlog_multi_insert(), heap_xlog_update(), heap_xlog_visible(), smgr_redo(), and smgrDoPendingSyncs().

{
    /* make sure the fakerel is not referenced by the SmgrRelation anymore */
    if (fakerel->rd_smgr != NULL)
        smgrclearowner(&fakerel->rd_smgr, fakerel->rd_smgr);
    pfree(fakerel);
}

log_invalid_page()

static void log_invalid_page ( RelFileNode  node, ForkNumber  forkno, BlockNumber  blkno, bool  present  )

static

Definition at line 81 of file xlogutils.c.

References xl_invalid_page_key::blkno, DEBUG1, elog, HASHCTL::entrysize, xl_invalid_page_key::forkno, HASH_BLOBS, hash_create(), HASH_ELEM, HASH_ENTER, hash_search(), ignore_invalid_pages, sort-test::key, HASHCTL::keysize, message_level_is_interesting(), xl_invalid_page_key::node, PANIC, xl_invalid_page::present, reachedConsistency, report_invalid_page(), and WARNING.

Referenced by XLogReadBufferExtended().

{
    xl_invalid_page_key key;
    xl_invalid_page *hentry;
    bool        found;

    /*
     * Once recovery has reached a consistent state, the invalid-page table
     * should be empty and remain so. If a reference to an invalid page is
     * found after consistency is reached, PANIC immediately. This might seem
     * aggressive, but it's better than letting the invalid reference linger
     * in the hash table until the end of recovery and PANIC there, which
     * might come only much later if this is a standby server.
     */
    if (reachedConsistency)
    {
        report_invalid_page(WARNING, node, forkno, blkno, present);
        elog(ignore_invalid_pages ? WARNING : PANIC,
             "WAL contains references to invalid pages");
    }

    /*
     * Log references to invalid pages at DEBUG1 level.  This allows some
     * tracing of the cause (note the elog context mechanism will tell us
     * something about the XLOG record that generated the reference).
     */
    if (message_level_is_interesting(DEBUG1))
        report_invalid_page(DEBUG1, node, forkno, blkno, present);

    if (invalid_page_tab == NULL)
    {
        /* create hash table when first needed */
        HASHCTL     ctl;

        ctl.keysize = sizeof(xl_invalid_page_key);
        ctl.entrysize = sizeof(xl_invalid_page);

        invalid_page_tab = hash_create("XLOG invalid-page table",
                                       100,
                                       &ctl,
                                       HASH_ELEM | HASH_BLOBS);
    }

    /* we currently assume xl_invalid_page_key contains no padding */
    key.node = node;
    key.forkno = forkno;
    key.blkno = blkno;
    hentry = (xl_invalid_page *)
        hash_search(invalid_page_tab, (void *) &key, HASH_ENTER, &found);

    if (!found)
    {
        /* hash_search already filled in the key */
        hentry->present = present;
    }
    else
    {
        /* repeat reference ... leave "present" as it was */
    }
}

read_local_xlog_page()

bool read_local_xlog_page

(

XLogReaderState *

state

)

Definition at line 842 of file xlogutils.c.

References CHECK_FOR_INTERRUPTS, XLogReaderState::currTLI, XLogReaderState::currTLIValidUntil, GetFlushRecPtr(), GetXLogReplayRecPtr(), pg_usleep(), XLogReaderState::readBuf, XLogReaderState::readPagePtr, RecoveryInProgress(), XLogReaderState::reqLen, ThisTimeLineID, wal_segment_close(), wal_segment_open(), WALRead(), WALReadRaiseError(), XLogReadDetermineTimeline(), and XLogReaderSetInputData().

Referenced by create_logical_replication_slot(), pg_logical_replication_slot_advance(), pg_logical_slot_get_changes_guts(), and XlogReadTwoPhaseData().

{
    XLogRecPtr  targetPagePtr = state->readPagePtr;
    int         reqLen        = state->reqLen;
    char       *cur_page      = state->readBuf;
    XLogRecPtr  read_upto,
                loc;
    TimeLineID  tli;
    int         count;
    WALReadError errinfo;

    loc = targetPagePtr + reqLen;

    /* Loop waiting for xlog to be available if necessary */
    while (1)
    {
        /*
         * Determine the limit of xlog we can currently read to, and what the
         * most recent timeline is.
         *
         * RecoveryInProgress() will update ThisTimeLineID when it first
         * notices recovery finishes, so we only have to maintain it for the
         * local process until recovery ends.
         */
        if (!RecoveryInProgress())
            read_upto = GetFlushRecPtr();
        else
            read_upto = GetXLogReplayRecPtr(&ThisTimeLineID);
        tli = ThisTimeLineID;

        /*
         * Check which timeline to get the record from.
         *
         * We have to do it each time through the loop because if we're in
         * recovery as a cascading standby, the current timeline might've
         * become historical. We can't rely on RecoveryInProgress() because in
         * a standby configuration like
         *
         * A => B => C
         *
         * if we're a logical decoding session on C, and B gets promoted, our
         * timeline will change while we remain in recovery.
         *
         * We can't just keep reading from the old timeline as the last WAL
         * archive in the timeline will get renamed to .partial by
         * StartupXLOG().
         *
         * If that happens after our caller updated ThisTimeLineID but before
         * we actually read the xlog page, we might still try to read from the
         * old (now renamed) segment and fail. There's not much we can do
         * about this, but it can only happen when we're a leaf of a cascading
         * standby whose primary gets promoted while we're decoding, so a
         * one-off ERROR isn't too bad.
         */
        XLogReadDetermineTimeline(state, targetPagePtr, reqLen);

        if (state->currTLI == ThisTimeLineID)
        {

            if (loc <= read_upto)
                break;

            CHECK_FOR_INTERRUPTS();
            pg_usleep(1000L);
        }
        else
        {
            /*
             * We're on a historical timeline, so limit reading to the switch
             * point where we moved to the next timeline.
             *
             * We don't need to GetFlushRecPtr or GetXLogReplayRecPtr. We know
             * about the new timeline, so we must've received past the end of
             * it.
             */
            read_upto = state->currTLIValidUntil;

            /*
             * Setting tli to our wanted record's TLI is slightly wrong; the
             * page might begin on an older timeline if it contains a timeline
             * switch, since its xlog segment will have been copied from the
             * prior timeline. This is pretty harmless though, as nothing
             * cares so long as the timeline doesn't go backwards.  We should
             * read the page header instead; FIXME someday.
             */
            tli = state->currTLI;

            /* No need to wait on a historical timeline */
            break;
        }
    }

    if (targetPagePtr + XLOG_BLCKSZ <= read_upto)
    {
        /*
         * more than one block available; read only that block, have caller
         * come back if they need more.
         */
        count = XLOG_BLCKSZ;
    }
    else if (targetPagePtr + reqLen > read_upto)
    {
        /* not enough data there */
        XLogReaderSetInputData(state,  -1);
        return false;
    }
    else
    {
        /* enough bytes available to satisfy the request */
        count = read_upto - targetPagePtr;
    }

    /*
     * Even though we just determined how much of the page can be validly read
     * as 'count', read the whole page anyway. It's guaranteed to be
     * zero-padded up to the page boundary if it's incomplete.
     */
    if (!WALRead(state, wal_segment_open, wal_segment_close,
                 cur_page, targetPagePtr, XLOG_BLCKSZ, tli, &errinfo))
        WALReadRaiseError(&errinfo);

    /* number of valid bytes in the buffer */
    state->readPagePtr = targetPagePtr;
    XLogReaderSetInputData(state, count);
    return true;
}

report_invalid_page()

static void report_invalid_page ( int  elevel, RelFileNode  node, ForkNumber  forkno, BlockNumber  blkno, bool  present  )

static

Definition at line 65 of file xlogutils.c.

References elog, pfree(), and relpathperm.

Referenced by log_invalid_page(), and XLogCheckInvalidPages().

{
    char       *path = relpathperm(node, forkno);

    if (present)
        elog(elevel, "page %u of relation %s is uninitialized",
             blkno, path);
    else
        elog(elevel, "page %u of relation %s does not exist",
             blkno, path);
    pfree(path);
}

wal_segment_close()

void wal_segment_close

(

XLogReaderState *

state

)

Definition at line 823 of file xlogutils.c.

References close, XLogReaderState::seg, and WALOpenSegment::ws_file.

Referenced by create_logical_replication_slot(), CreateReplicationSlot(), logical_read_xlog_page(), pg_logical_replication_slot_advance(), pg_logical_slot_get_changes_guts(), read_local_xlog_page(), StartLogicalReplication(), StartReplication(), StartupXLOG(), WalSndErrorCleanup(), XlogReadTwoPhaseData(), and XLogSendPhysical().

{
    close(state->seg.ws_file);
    /* need to check errno? */
    state->seg.ws_file = -1;
}

wal_segment_open()

void wal_segment_open	(	XLogReaderState *	state,
		XLogSegNo	nextSegNo,
		TimeLineID *	tli_p
	)

Definition at line 798 of file xlogutils.c.

References BasicOpenFile(), ereport, errcode_for_file_access(), errmsg(), ERROR, MAXPGPATH, PG_BINARY, XLogReaderState::seg, XLogReaderState::segcxt, WALOpenSegment::ws_file, WALSegmentContext::ws_segsize, and XLogFilePath.

Referenced by read_local_xlog_page().

{
    TimeLineID  tli = *tli_p;
    char        path[MAXPGPATH];

    XLogFilePath(path, tli, nextSegNo, state->segcxt.ws_segsize);
    state->seg.ws_file = BasicOpenFile(path, O_RDONLY | PG_BINARY);
    if (state->seg.ws_file >= 0)
        return;

    if (errno == ENOENT)
        ereport(ERROR,
                (errcode_for_file_access(),
                 errmsg("requested WAL segment %s has already been removed",
                        path)));
    else
        ereport(ERROR,
                (errcode_for_file_access(),
                 errmsg("could not open file \"%s\": %m",
                        path)));
}

WALReadRaiseError()

void WALReadRaiseError

(

WALReadError *

errinfo

)

Definition at line 974 of file xlogutils.c.

References ereport, errcode(), ERRCODE_DATA_CORRUPTED, errcode_for_file_access(), errmsg(), ERROR, MAXFNAMELEN, wal_segment_size, WALReadError::wre_errno, WALReadError::wre_off, WALReadError::wre_read, WALReadError::wre_req, WALReadError::wre_seg, WALOpenSegment::ws_segno, WALOpenSegment::ws_tli, and XLogFileName.

Referenced by logical_read_xlog_page(), read_local_xlog_page(), and XLogSendPhysical().

{
    WALOpenSegment *seg = &errinfo->wre_seg;
    char        fname[MAXFNAMELEN];

    XLogFileName(fname, seg->ws_tli, seg->ws_segno, wal_segment_size);

    if (errinfo->wre_read < 0)
    {
        errno = errinfo->wre_errno;
        ereport(ERROR,
                (errcode_for_file_access(),
                 errmsg("could not read from log segment %s, offset %u: %m",
                        fname, errinfo->wre_off)));
    }
    else if (errinfo->wre_read == 0)
    {
        ereport(ERROR,
                (errcode(ERRCODE_DATA_CORRUPTED),
                 errmsg("could not read from log segment %s, offset %u: read %d of %zu",
                        fname, errinfo->wre_off, errinfo->wre_read,
                        (Size) errinfo->wre_req)));
    }
}

XLogCheckInvalidPages()

void XLogCheckInvalidPages

(

void

)

Definition at line 223 of file xlogutils.c.

References xl_invalid_page_key::blkno, elog, xl_invalid_page_key::forkno, hash_destroy(), hash_seq_init(), hash_seq_search(), ignore_invalid_pages, xl_invalid_page::key, xl_invalid_page_key::node, PANIC, xl_invalid_page::present, report_invalid_page(), status(), and WARNING.

Referenced by CheckRecoveryConsistency().

{
    HASH_SEQ_STATUS status;
    xl_invalid_page *hentry;
    bool        foundone = false;

    if (invalid_page_tab == NULL)
        return;                 /* nothing to do */

    hash_seq_init(&status, invalid_page_tab);

    /*
     * Our strategy is to emit WARNING messages for all remaining entries and
     * only PANIC after we've dumped all the available info.
     */
    while ((hentry = (xl_invalid_page *) hash_seq_search(&status)) != NULL)
    {
        report_invalid_page(WARNING, hentry->key.node, hentry->key.forkno,
                            hentry->key.blkno, hentry->present);
        foundone = true;
    }

    if (foundone)
        elog(ignore_invalid_pages ? WARNING : PANIC,
             "WAL contains references to invalid pages");

    hash_destroy(invalid_page_tab);
    invalid_page_tab = NULL;
}

XLogDropDatabase()

void XLogDropDatabase

(

Oid

dbid

)

Definition at line 638 of file xlogutils.c.

References forget_invalid_pages_db(), and smgrcloseall().

Referenced by dbase_redo().

{
    /*
     * This is unnecessarily heavy-handed, as it will close SMgrRelation
     * objects for other databases as well. DROP DATABASE occurs seldom enough
     * that it's not worth introducing a variant of smgrclose for just this
     * purpose. XXX: Or should we rather leave the smgr entries dangling?
     */
    smgrcloseall();

    forget_invalid_pages_db(dbid);
}

XLogDropRelation()

void XLogDropRelation	(	RelFileNode	rnode,
		ForkNumber	forknum
	)

Definition at line 627 of file xlogutils.c.

References forget_invalid_pages().

Referenced by DropRelationFiles().

{
    forget_invalid_pages(rnode, forknum, 0);
}

XLogHaveInvalidPages()

bool XLogHaveInvalidPages

(

void

)

Definition at line 213 of file xlogutils.c.

References hash_get_num_entries().

Referenced by RecoveryRestartPoint().

{
    if (invalid_page_tab != NULL &&
        hash_get_num_entries(invalid_page_tab) > 0)
        return true;
    return false;
}

XLogInitBufferForRedo()

Buffer XLogInitBufferForRedo	(	XLogReaderState *	record,
		uint8	block_id
	)

Definition at line 304 of file xlogutils.c.

References buf, RBM_ZERO_AND_LOCK, and XLogReadBufferForRedoExtended().

Referenced by _bt_restore_meta(), brin_xlog_createidx(), brin_xlog_insert_update(), brin_xlog_revmap_extend(), btree_xlog_mark_page_halfdead(), btree_xlog_newroot(), btree_xlog_split(), btree_xlog_unlink_page(), ginRedoCreatePTree(), ginRedoDeleteListPages(), ginRedoInsertListPage(), ginRedoUpdateMetapage(), gistRedoPageSplitRecord(), hash_xlog_add_ovfl_page(), hash_xlog_init_bitmap_page(), hash_xlog_init_meta_page(), hash_xlog_split_allocate_page(), heap_xlog_insert(), heap_xlog_multi_insert(), heap_xlog_update(), seq_redo(), spgRedoAddLeaf(), spgRedoAddNode(), spgRedoMoveLeafs(), spgRedoPickSplit(), and spgRedoSplitTuple().

{
    Buffer      buf;

    XLogReadBufferForRedoExtended(record, block_id, RBM_ZERO_AND_LOCK, false,
                                  &buf);
    return buf;
}

XLogReadBufferExtended()

Buffer XLogReadBufferExtended	(	RelFileNode	rnode,
		ForkNumber	forknum,
		BlockNumber	blkno,
		ReadBufferMode	mode,
		Buffer	recent_buffer
	)

Definition at line 443 of file xlogutils.c.

References Assert, BUFFER_LOCK_UNLOCK, BufferGetBlockNumber(), BufferGetPage, BufferIsValid, InRecovery, InvalidBackendId, InvalidBuffer, LockBuffer(), log_invalid_page(), P_NEW, PageIsNew, RBM_NORMAL, RBM_NORMAL_NO_LOG, RBM_ZERO_AND_CLEANUP_LOCK, RBM_ZERO_AND_LOCK, ReadBufferWithoutRelcache(), ReadRecentBuffer(), ReleaseBuffer(), smgrcreate(), smgrnblocks(), and smgropen().

Referenced by checkXLogConsistency(), XLogReadBufferForRedoExtended(), and XLogRecordPageWithFreeSpace().

{
    BlockNumber lastblock;
    Buffer      buffer;
    SMgrRelation smgr;

    Assert(blkno != P_NEW);

    /* Do we have a clue where the buffer might be already? */
    if (BufferIsValid(recent_buffer) &&
        mode == RBM_NORMAL &&
        ReadRecentBuffer(rnode, forknum, blkno, recent_buffer))
    {
        buffer = recent_buffer;
        goto recent_buffer_fast_path;
    }

    /* Open the relation at smgr level */
    smgr = smgropen(rnode, InvalidBackendId);

    /*
     * Create the target file if it doesn't already exist.  This lets us cope
     * if the replay sequence contains writes to a relation that is later
     * deleted.  (The original coding of this routine would instead suppress
     * the writes, but that seems like it risks losing valuable data if the
     * filesystem loses an inode during a crash.  Better to write the data
     * until we are actually told to delete the file.)
     */
    smgrcreate(smgr, forknum, true);

    lastblock = smgrnblocks(smgr, forknum);

    if (blkno < lastblock)
    {
        /* page exists in file */
        buffer = ReadBufferWithoutRelcache(rnode, forknum, blkno,
                                           mode, NULL);
    }
    else
    {
        /* hm, page doesn't exist in file */
        if (mode == RBM_NORMAL)
        {
            log_invalid_page(rnode, forknum, blkno, false);
            return InvalidBuffer;
        }
        if (mode == RBM_NORMAL_NO_LOG)
            return InvalidBuffer;
        /* OK to extend the file */
        /* we do this in recovery only - no rel-extension lock needed */
        Assert(InRecovery);
        buffer = InvalidBuffer;
        do
        {
            if (buffer != InvalidBuffer)
            {
                if (mode == RBM_ZERO_AND_LOCK || mode == RBM_ZERO_AND_CLEANUP_LOCK)
                    LockBuffer(buffer, BUFFER_LOCK_UNLOCK);
                ReleaseBuffer(buffer);
            }
            buffer = ReadBufferWithoutRelcache(rnode, forknum,
                                               P_NEW, mode, NULL);
        }
        while (BufferGetBlockNumber(buffer) < blkno);
        /* Handle the corner case that P_NEW returns non-consecutive pages */
        if (BufferGetBlockNumber(buffer) != blkno)
        {
            if (mode == RBM_ZERO_AND_LOCK || mode == RBM_ZERO_AND_CLEANUP_LOCK)
                LockBuffer(buffer, BUFFER_LOCK_UNLOCK);
            ReleaseBuffer(buffer);
            buffer = ReadBufferWithoutRelcache(rnode, forknum, blkno,
                                               mode, NULL);
        }
    }

recent_buffer_fast_path:
    if (mode == RBM_NORMAL)
    {
        /* check that page has been initialized */
        Page        page = (Page) BufferGetPage(buffer);

        /*
         * We assume that PageIsNew is safe without a lock. During recovery,
         * there should be no other backends that could modify the buffer at
         * the same time.
         */
        if (PageIsNew(page))
        {
            ReleaseBuffer(buffer);
            log_invalid_page(rnode, forknum, blkno, true);
            return InvalidBuffer;
        }
    }

    return buffer;
}

XLogReadBufferForRedo()

XLogRedoAction XLogReadBufferForRedo	(	XLogReaderState *	record,
		uint8	block_id,
		Buffer *	buf
	)

Definition at line 292 of file xlogutils.c.

References RBM_NORMAL, and XLogReadBufferForRedoExtended().

Referenced by _bt_clear_incomplete_split(), brin_xlog_desummarize_page(), brin_xlog_insert_update(), brin_xlog_revmap_extend(), brin_xlog_samepage_update(), brin_xlog_update(), btree_xlog_dedup(), btree_xlog_delete(), btree_xlog_insert(), btree_xlog_mark_page_halfdead(), btree_xlog_split(), btree_xlog_unlink_page(), generic_redo(), ginRedoClearIncompleteSplit(), ginRedoDeletePage(), ginRedoInsert(), ginRedoSplit(), ginRedoUpdateMetapage(), ginRedoVacuumDataLeafPage(), ginRedoVacuumPage(), gistRedoClearFollowRight(), gistRedoDeleteRecord(), gistRedoPageDelete(), gistRedoPageUpdateRecord(), hash_xlog_add_ovfl_page(), hash_xlog_delete(), hash_xlog_init_bitmap_page(), hash_xlog_insert(), hash_xlog_move_page_contents(), hash_xlog_split_allocate_page(), hash_xlog_split_cleanup(), hash_xlog_split_complete(), hash_xlog_split_page(), hash_xlog_squeeze_page(), hash_xlog_update_meta_page(), hash_xlog_vacuum_one_page(), heap_xlog_confirm(), heap_xlog_delete(), heap_xlog_freeze_page(), heap_xlog_inplace(), heap_xlog_insert(), heap_xlog_lock(), heap_xlog_lock_updated(), heap_xlog_multi_insert(), heap_xlog_update(), heap_xlog_visible(), spgRedoAddLeaf(), spgRedoAddNode(), spgRedoMoveLeafs(), spgRedoPickSplit(), spgRedoSplitTuple(), spgRedoVacuumLeaf(), spgRedoVacuumRedirect(), spgRedoVacuumRoot(), and xlog_redo().

{
    return XLogReadBufferForRedoExtended(record, block_id, RBM_NORMAL,
                                         false, buf);
}

XLogReadBufferForRedoExtended()

XLogRedoAction XLogReadBufferForRedoExtended	(	XLogReaderState *	record,
		uint8	block_id,
		ReadBufferMode	mode,
		bool	get_cleanup_lock,
		Buffer *	buf
	)

Definition at line 329 of file xlogutils.c.

References Assert, BKPBLOCK_WILL_INIT, BLK_DONE, BLK_NEEDS_REDO, BLK_NOTFOUND, BLK_RESTORED, xl_invalid_page_key::blkno, DecodedXLogRecord::blocks, BUFFER_LOCK_EXCLUSIVE, BufferGetPage, BufferIsValid, elog, XLogReaderState::EndRecPtr, ERROR, DecodedBkpBlock::flags, FlushOneBuffer(), INIT_FORKNUM, LockBuffer(), LockBufferForCleanup(), MarkBufferDirty(), PageGetLSN, PageIsNew, PageSetLSN, PANIC, RBM_ZERO_AND_CLEANUP_LOCK, RBM_ZERO_AND_LOCK, XLogReaderState::record, RestoreBlockImage(), XLogReadBufferExtended(), XLogRecBlockImageApply, XLogRecGetRecentBuffer(), and XLogRecHasBlockImage.

Referenced by btree_xlog_vacuum(), hash_xlog_delete(), hash_xlog_move_page_contents(), hash_xlog_split_allocate_page(), hash_xlog_squeeze_page(), hash_xlog_vacuum_one_page(), heap_xlog_prune(), heap_xlog_vacuum(), heap_xlog_visible(), XLogInitBufferForRedo(), and XLogReadBufferForRedo().

{
    XLogRecPtr  lsn = record->EndRecPtr;
    RelFileNode rnode;
    ForkNumber  forknum;
    BlockNumber blkno;
    Buffer      recent_buffer;
    Page        page;
    bool        zeromode;
    bool        willinit;

    if (!XLogRecGetRecentBuffer(record, block_id, &rnode, &forknum, &blkno,
                                &recent_buffer))
    {
        /* Caller specified a bogus block_id */
        elog(PANIC, "failed to locate backup block with ID %d", block_id);
    }

    /*
     * Make sure that if the block is marked with WILL_INIT, the caller is
     * going to initialize it. And vice versa.
     */
    zeromode = (mode == RBM_ZERO_AND_LOCK || mode == RBM_ZERO_AND_CLEANUP_LOCK);
    willinit = (record->record->blocks[block_id].flags & BKPBLOCK_WILL_INIT) != 0;
    if (willinit && !zeromode)
        elog(PANIC, "block with WILL_INIT flag in WAL record must be zeroed by redo routine");
    if (!willinit && zeromode)
        elog(PANIC, "block to be initialized in redo routine must be marked with WILL_INIT flag in the WAL record");

    /* If it has a full-page image and it should be restored, do it. */
    if (XLogRecBlockImageApply(record, block_id))
    {
        Assert(XLogRecHasBlockImage(record, block_id));
        *buf = XLogReadBufferExtended(rnode, forknum, blkno,
                                      get_cleanup_lock ? RBM_ZERO_AND_CLEANUP_LOCK : RBM_ZERO_AND_LOCK,
                                      recent_buffer);
        page = BufferGetPage(*buf);
        if (!RestoreBlockImage(record, block_id, page))
            elog(ERROR, "failed to restore block image");

        /*
         * The page may be uninitialized. If so, we can't set the LSN because
         * that would corrupt the page.
         */
        if (!PageIsNew(page))
        {
            PageSetLSN(page, lsn);
        }

        MarkBufferDirty(*buf);

        /*
         * At the end of crash recovery the init forks of unlogged relations
         * are copied, without going through shared buffers. So we need to
         * force the on-disk state of init forks to always be in sync with the
         * state in shared buffers.
         */
        if (forknum == INIT_FORKNUM)
            FlushOneBuffer(*buf);

        return BLK_RESTORED;
    }
    else
    {
        *buf = XLogReadBufferExtended(rnode, forknum, blkno, mode,
                                      recent_buffer);
        if (BufferIsValid(*buf))
        {
            if (mode != RBM_ZERO_AND_LOCK && mode != RBM_ZERO_AND_CLEANUP_LOCK)
            {
                if (get_cleanup_lock)
                    LockBufferForCleanup(*buf);
                else
                    LockBuffer(*buf, BUFFER_LOCK_EXCLUSIVE);
            }
            if (lsn <= PageGetLSN(BufferGetPage(*buf)))
                return BLK_DONE;
            else
                return BLK_NEEDS_REDO;
        }
        else
            return BLK_NOTFOUND;
    }
}

XLogReadDetermineTimeline()

void XLogReadDetermineTimeline	(	XLogReaderState *	state,
		XLogRecPtr	wantPage,
		uint32	wantLength
	)

Definition at line 702 of file xlogutils.c.

References Assert, XLogReaderState::currTLI, XLogReaderState::currTLIValidUntil, DEBUG3, elog, InvalidXLogRecPtr, list_free_deep(), LSN_FORMAT_ARGS, Min, XLogReaderState::nextTLI, XLogReaderState::page_verified, XLogReaderState::readLen, XLogReaderState::readPagePtr, readTimeLineHistory(), XLogReaderState::segcxt, ThisTimeLineID, tliOfPointInHistory(), tliSwitchPoint(), and WALSegmentContext::ws_segsize.

Referenced by logical_read_xlog_page(), and read_local_xlog_page().

{
    const XLogRecPtr lastReadPage = state->readPagePtr;

    Assert(wantPage != InvalidXLogRecPtr && wantPage % XLOG_BLCKSZ == 0);
    Assert(wantLength <= XLOG_BLCKSZ);
    Assert(state->readLen == 0 || state->readLen <= XLOG_BLCKSZ);

    /*
     * If the desired page is currently read in and valid, we have nothing to
     * do.
     *
     * The caller should've ensured that it didn't previously advance readOff
     * past the valid limit of this timeline, so it doesn't matter if the
     * current TLI has since become historical.
     */
    if (lastReadPage == wantPage &&
        state->page_verified &&
        lastReadPage + state->readLen >= wantPage + Min(wantLength, XLOG_BLCKSZ - 1))
        return;

    /*
     * If we're reading from the current timeline, it hasn't become historical
     * and the page we're reading is after the last page read, we can again
     * just carry on. (Seeking backwards requires a check to make sure the
     * older page isn't on a prior timeline).
     *
     * ThisTimeLineID might've become historical since we last looked, but the
     * caller is required not to read past the flush limit it saw at the time
     * it looked up the timeline. There's nothing we can do about it if
     * StartupXLOG() renames it to .partial concurrently.
     */
    if (state->currTLI == ThisTimeLineID && wantPage >= lastReadPage)
    {
        Assert(state->currTLIValidUntil == InvalidXLogRecPtr);
        return;
    }

    /*
     * If we're just reading pages from a previously validated historical
     * timeline and the timeline we're reading from is valid until the end of
     * the current segment we can just keep reading.
     */
    if (state->currTLIValidUntil != InvalidXLogRecPtr &&
        state->currTLI != ThisTimeLineID &&
        state->currTLI != 0 &&
        ((wantPage + wantLength) / state->segcxt.ws_segsize) <
        (state->currTLIValidUntil / state->segcxt.ws_segsize))
        return;

    /*
     * If we reach this point we're either looking up a page for random
     * access, the current timeline just became historical, or we're reading
     * from a new segment containing a timeline switch. In all cases we need
     * to determine the newest timeline on the segment.
     *
     * If it's the current timeline we can just keep reading from here unless
     * we detect a timeline switch that makes the current timeline historical.
     * If it's a historical timeline we can read all the segment on the newest
     * timeline because it contains all the old timelines' data too. So only
     * one switch check is required.
     */
    {
        /*
         * We need to re-read the timeline history in case it's been changed
         * by a promotion or replay from a cascaded replica.
         */
        List       *timelineHistory = readTimeLineHistory(ThisTimeLineID);
        XLogRecPtr  endOfSegment;

        endOfSegment = ((wantPage / state->segcxt.ws_segsize) + 1) *
            state->segcxt.ws_segsize - 1;
        Assert(wantPage / state->segcxt.ws_segsize ==
               endOfSegment / state->segcxt.ws_segsize);

        /*
         * Find the timeline of the last LSN on the segment containing
         * wantPage.
         */
        state->currTLI = tliOfPointInHistory(endOfSegment, timelineHistory);
        state->currTLIValidUntil = tliSwitchPoint(state->currTLI, timelineHistory,
                                                  &state->nextTLI);

        Assert(state->currTLIValidUntil == InvalidXLogRecPtr ||
               wantPage + wantLength < state->currTLIValidUntil);

        list_free_deep(timelineHistory);

        elog(DEBUG3, "switched to timeline %u valid until %X/%X",
             state->currTLI,
             LSN_FORMAT_ARGS(state->currTLIValidUntil));
    }
}

XLogTruncateRelation()

void XLogTruncateRelation	(	RelFileNode	rnode,
		ForkNumber	forkNum,
		BlockNumber	nblocks
	)

Definition at line 657 of file xlogutils.c.

References forget_invalid_pages().

Referenced by smgr_redo().

{
    forget_invalid_pages(rnode, forkNum, nblocks);
}

Variable Documentation

ignore_invalid_pages

bool ignore_invalid_pages = false

Definition at line 35 of file xlogutils.c.

Referenced by log_invalid_page(), and XLogCheckInvalidPages().

invalid_page_tab

HTAB* invalid_page_tab = NULL

static

Definition at line 60 of file xlogutils.c.