verify_heapam.c File Reference

#include "postgres.h"
#include "access/detoast.h"
#include "access/genam.h"
#include "access/heaptoast.h"
#include "access/multixact.h"
#include "access/relation.h"
#include "access/table.h"
#include "access/toast_internals.h"
#include "access/visibilitymap.h"
#include "access/xact.h"
#include "catalog/pg_am.h"
#include "catalog/pg_class.h"
#include "funcapi.h"
#include "miscadmin.h"
#include "storage/bufmgr.h"
#include "storage/procarray.h"
#include "storage/read_stream.h"
#include "utils/builtins.h"
#include "utils/fmgroids.h"
#include "utils/rel.h"

Include dependency graph for verify_heapam.c:

Go to the source code of this file.

Data Structures
struct	ToastedAttribute
struct	HeapCheckContext
struct	HeapCheckReadStreamData

Macros
#define	HEAPCHECK_RELATION_COLS   4
#define	VARLENA_SIZE_LIMIT   0x3FFFFFFF

Typedefs
typedef enum XidBoundsViolation	XidBoundsViolation
typedef enum XidCommitStatus	XidCommitStatus
typedef enum SkipPages	SkipPages
typedef struct ToastedAttribute	ToastedAttribute
typedef struct HeapCheckContext	HeapCheckContext
typedef struct HeapCheckReadStreamData	HeapCheckReadStreamData

Enumerations
enum	XidBoundsViolation {   XID_INVALID , XID_IN_FUTURE , XID_PRECEDES_CLUSTERMIN , XID_PRECEDES_RELMIN ,   XID_BOUNDS_OK }
enum	XidCommitStatus { XID_COMMITTED , XID_IS_CURRENT_XID , XID_IN_PROGRESS , XID_ABORTED }
enum	SkipPages { SKIP_PAGES_ALL_FROZEN , SKIP_PAGES_ALL_VISIBLE , SKIP_PAGES_NONE }

Functions
	PG_FUNCTION_INFO_V1 (verify_heapam)
static BlockNumber	heapcheck_read_stream_next_unskippable (ReadStream *stream, void *callback_private_data, void *per_buffer_data)
static void	check_tuple (HeapCheckContext *ctx, bool *xmin_commit_status_ok, XidCommitStatus *xmin_commit_status)
static void	check_toast_tuple (HeapTuple toasttup, HeapCheckContext *ctx, ToastedAttribute *ta, int32 *expected_chunk_seq, uint32 extsize)
static bool	check_tuple_attribute (HeapCheckContext *ctx)
static void	check_toasted_attribute (HeapCheckContext *ctx, ToastedAttribute *ta)
static bool	check_tuple_header (HeapCheckContext *ctx)
static bool	check_tuple_visibility (HeapCheckContext *ctx, bool *xmin_commit_status_ok, XidCommitStatus *xmin_commit_status)
static void	report_corruption (HeapCheckContext *ctx, char *msg)
static void	report_toast_corruption (HeapCheckContext *ctx, ToastedAttribute *ta, char *msg)
static FullTransactionId	FullTransactionIdFromXidAndCtx (TransactionId xid, const HeapCheckContext *ctx)
static void	update_cached_xid_range (HeapCheckContext *ctx)
static void	update_cached_mxid_range (HeapCheckContext *ctx)
static XidBoundsViolation	check_mxid_in_range (MultiXactId mxid, HeapCheckContext *ctx)
static XidBoundsViolation	check_mxid_valid_in_rel (MultiXactId mxid, HeapCheckContext *ctx)
static XidBoundsViolation	get_xid_status (TransactionId xid, HeapCheckContext *ctx, XidCommitStatus *status)
Datum	verify_heapam (PG_FUNCTION_ARGS)
static void	report_corruption_internal (Tuplestorestate *tupstore, TupleDesc tupdesc, BlockNumber blkno, OffsetNumber offnum, AttrNumber attnum, char *msg)
static bool	fxid_in_cached_range (FullTransactionId fxid, const HeapCheckContext *ctx)

Macro Definition Documentation

HEAPCHECK_RELATION_COLS

#define HEAPCHECK_RELATION_COLS   4

Definition at line 36 of file verify_heapam.c.

VARLENA_SIZE_LIMIT

#define VARLENA_SIZE_LIMIT   0x3FFFFFFF

Definition at line 39 of file verify_heapam.c.

Typedef Documentation

HeapCheckContext

typedef struct HeapCheckContext HeapCheckContext

HeapCheckReadStreamData

typedef struct HeapCheckReadStreamData HeapCheckReadStreamData

SkipPages

typedef enum SkipPages SkipPages

ToastedAttribute

typedef struct ToastedAttribute ToastedAttribute

XidBoundsViolation

typedef enum XidBoundsViolation XidBoundsViolation

XidCommitStatus

typedef enum XidCommitStatus XidCommitStatus

Enumeration Type Documentation

SkipPages

enum SkipPages

Enumerator
SKIP_PAGES_ALL_FROZEN
SKIP_PAGES_ALL_VISIBLE
SKIP_PAGES_NONE

Definition at line 62 of file verify_heapam.c.

{
    SKIP_PAGES_ALL_FROZEN,
    SKIP_PAGES_ALL_VISIBLE,
    SKIP_PAGES_NONE,
} SkipPages;

XidBoundsViolation

enum XidBoundsViolation

Enumerator
XID_INVALID
XID_IN_FUTURE
XID_PRECEDES_CLUSTERMIN
XID_PRECEDES_RELMIN
XID_BOUNDS_OK

Definition at line 45 of file verify_heapam.c.

{
    XID_INVALID,
    XID_IN_FUTURE,
    XID_PRECEDES_CLUSTERMIN,
    XID_PRECEDES_RELMIN,
    XID_BOUNDS_OK,
} XidBoundsViolation;

XidCommitStatus

enum XidCommitStatus

Enumerator
XID_COMMITTED
XID_IS_CURRENT_XID
XID_IN_PROGRESS
XID_ABORTED

Definition at line 54 of file verify_heapam.c.

{
    XID_COMMITTED,
    XID_IS_CURRENT_XID,
    XID_IN_PROGRESS,
    XID_ABORTED,
} XidCommitStatus;

Function Documentation

check_mxid_in_range()

static XidBoundsViolation check_mxid_in_range ( MultiXactId  mxid, HeapCheckContext *  ctx  )

static

Definition at line 2057 of file verify_heapam.c.

{
    if (!TransactionIdIsValid(mxid))
        return XID_INVALID;
    if (MultiXactIdPrecedes(mxid, ctx->relminmxid))
        return XID_PRECEDES_RELMIN;
    if (MultiXactIdPrecedes(mxid, ctx->oldest_mxact))
        return XID_PRECEDES_CLUSTERMIN;
    if (MultiXactIdPrecedesOrEquals(ctx->next_mxact, mxid))
        return XID_IN_FUTURE;
    return XID_BOUNDS_OK;
}

References MultiXactIdPrecedes(), MultiXactIdPrecedesOrEquals(), HeapCheckContext::next_mxact, HeapCheckContext::oldest_mxact, HeapCheckContext::relminmxid, TransactionIdIsValid, XID_BOUNDS_OK, XID_IN_FUTURE, XID_INVALID, XID_PRECEDES_CLUSTERMIN, and XID_PRECEDES_RELMIN.

Referenced by check_mxid_valid_in_rel().

check_mxid_valid_in_rel()

static XidBoundsViolation check_mxid_valid_in_rel ( MultiXactId  mxid, HeapCheckContext *  ctx  )

static

Definition at line 2079 of file verify_heapam.c.

{
    XidBoundsViolation result;
 
    result = check_mxid_in_range(mxid, ctx);
    if (result == XID_BOUNDS_OK)
        return XID_BOUNDS_OK;
 
    /* The range may have advanced.  Recheck. */
    update_cached_mxid_range(ctx);
    return check_mxid_in_range(mxid, ctx);
}

References check_mxid_in_range(), update_cached_mxid_range(), and XID_BOUNDS_OK.

Referenced by check_tuple_visibility().

check_toast_tuple()

static void check_toast_tuple ( HeapTuple  toasttup, HeapCheckContext *  ctx, ToastedAttribute *  ta, int32 *  expected_chunk_seq, uint32  extsize  )

static

Definition at line 1554 of file verify_heapam.c.

{
    int32       chunk_seq;
    int32       last_chunk_seq = (extsize - 1) / TOAST_MAX_CHUNK_SIZE;
    Pointer     chunk;
    bool        isnull;
    int32       chunksize;
    int32       expected_size;
 
    /* Sanity-check the sequence number. */
    chunk_seq = DatumGetInt32(fastgetattr(toasttup, 2,
                                          ctx->toast_rel->rd_att, &isnull));
    if (isnull)
    {
        report_toast_corruption(ctx, ta,
                                psprintf("toast value %u has toast chunk with null sequence number",
                                         ta->toast_pointer.va_valueid));
        return;
    }
    if (chunk_seq != *expected_chunk_seq)
    {
        /* Either the TOAST index is corrupt, or we don't have all chunks. */
        report_toast_corruption(ctx, ta,
                                psprintf("toast value %u index scan returned chunk %d when expecting chunk %d",
                                         ta->toast_pointer.va_valueid,
                                         chunk_seq, *expected_chunk_seq));
    }
    *expected_chunk_seq = chunk_seq + 1;
 
    /* Sanity-check the chunk data. */
    chunk = DatumGetPointer(fastgetattr(toasttup, 3,
                                        ctx->toast_rel->rd_att, &isnull));
    if (isnull)
    {
        report_toast_corruption(ctx, ta,
                                psprintf("toast value %u chunk %d has null data",
                                         ta->toast_pointer.va_valueid,
                                         chunk_seq));
        return;
    }
    if (!VARATT_IS_EXTENDED(chunk))
        chunksize = VARSIZE(chunk) - VARHDRSZ;
    else if (VARATT_IS_SHORT(chunk))
    {
        /*
         * could happen due to heap_form_tuple doing its thing
         */
        chunksize = VARSIZE_SHORT(chunk) - VARHDRSZ_SHORT;
    }
    else
    {
        /* should never happen */
        uint32      header = ((varattrib_4b *) chunk)->va_4byte.va_header;
 
        report_toast_corruption(ctx, ta,
                                psprintf("toast value %u chunk %d has invalid varlena header %0x",
                                         ta->toast_pointer.va_valueid,
                                         chunk_seq, header));
        return;
    }
 
    /*
     * Some checks on the data we've found
     */
    if (chunk_seq > last_chunk_seq)
    {
        report_toast_corruption(ctx, ta,
                                psprintf("toast value %u chunk %d follows last expected chunk %d",
                                         ta->toast_pointer.va_valueid,
                                         chunk_seq, last_chunk_seq));
        return;
    }
 
    expected_size = chunk_seq < last_chunk_seq ? TOAST_MAX_CHUNK_SIZE
        : extsize - (last_chunk_seq * TOAST_MAX_CHUNK_SIZE);
 
    if (chunksize != expected_size)
        report_toast_corruption(ctx, ta,
                                psprintf("toast value %u chunk %d has size %u, but expected size %u",
                                         ta->toast_pointer.va_valueid,
                                         chunk_seq, chunksize, expected_size));
}

References DatumGetInt32(), DatumGetPointer(), fastgetattr(), psprintf(), RelationData::rd_att, report_toast_corruption(), TOAST_MAX_CHUNK_SIZE, ToastedAttribute::toast_pointer, HeapCheckContext::toast_rel, varatt_external::va_valueid, VARATT_IS_EXTENDED, VARATT_IS_SHORT, VARHDRSZ, VARHDRSZ_SHORT, VARSIZE, and VARSIZE_SHORT.

Referenced by check_toasted_attribute().

check_toasted_attribute()

static void check_toasted_attribute ( HeapCheckContext *  ctx, ToastedAttribute *  ta  )

static

Definition at line 1860 of file verify_heapam.c.

{
    ScanKeyData toastkey;
    SysScanDesc toastscan;
    bool        found_toasttup;
    HeapTuple   toasttup;
    uint32      extsize;
    int32       expected_chunk_seq = 0;
    int32       last_chunk_seq;
 
    extsize = VARATT_EXTERNAL_GET_EXTSIZE(ta->toast_pointer);
    last_chunk_seq = (extsize - 1) / TOAST_MAX_CHUNK_SIZE;
 
    /*
     * Setup a scan key to find chunks in toast table with matching va_valueid
     */
    ScanKeyInit(&toastkey,
                (AttrNumber) 1,
                BTEqualStrategyNumber, F_OIDEQ,
                ObjectIdGetDatum(ta->toast_pointer.va_valueid));
 
    /*
     * Check if any chunks for this toasted object exist in the toast table,
     * accessible via the index.
     */
    toastscan = systable_beginscan_ordered(ctx->toast_rel,
                                           ctx->valid_toast_index,
                                           get_toast_snapshot(), 1,
                                           &toastkey);
    found_toasttup = false;
    while ((toasttup =
            systable_getnext_ordered(toastscan,
                                     ForwardScanDirection)) != NULL)
    {
        found_toasttup = true;
        check_toast_tuple(toasttup, ctx, ta, &expected_chunk_seq, extsize);
    }
    systable_endscan_ordered(toastscan);
 
    if (!found_toasttup)
        report_toast_corruption(ctx, ta,
                                psprintf("toast value %u not found in toast table",
                                         ta->toast_pointer.va_valueid));
    else if (expected_chunk_seq <= last_chunk_seq)
        report_toast_corruption(ctx, ta,
                                psprintf("toast value %u was expected to end at chunk %d, but ended while expecting chunk %d",
                                         ta->toast_pointer.va_valueid,
                                         last_chunk_seq, expected_chunk_seq));
}

References BTEqualStrategyNumber, check_toast_tuple(), ForwardScanDirection, get_toast_snapshot(), ObjectIdGetDatum(), psprintf(), report_toast_corruption(), ScanKeyInit(), systable_beginscan_ordered(), systable_endscan_ordered(), systable_getnext_ordered(), TOAST_MAX_CHUNK_SIZE, ToastedAttribute::toast_pointer, HeapCheckContext::toast_rel, varatt_external::va_valueid, HeapCheckContext::valid_toast_index, and VARATT_EXTERNAL_GET_EXTSIZE.

Referenced by verify_heapam().

check_tuple()

static void check_tuple ( HeapCheckContext *  ctx, bool *  xmin_commit_status_ok, XidCommitStatus *  xmin_commit_status  )

static

Definition at line 1918 of file verify_heapam.c.

{
    /*
     * Check various forms of tuple header corruption, and if the header is
     * too corrupt, do not continue with other checks.
     */
    if (!check_tuple_header(ctx))
        return;
 
    /*
     * Check tuple visibility.  If the inserting transaction aborted, we
     * cannot assume our relation description matches the tuple structure, and
     * therefore cannot check it.
     */
    if (!check_tuple_visibility(ctx, xmin_commit_status_ok,
                                xmin_commit_status))
        return;
 
    /*
     * The tuple is visible, so it must be compatible with the current version
     * of the relation descriptor. It might have fewer columns than are
     * present in the relation descriptor, but it cannot have more.
     */
    if (RelationGetDescr(ctx->rel)->natts < ctx->natts)
    {
        report_corruption(ctx,
                          psprintf("number of attributes %u exceeds maximum expected for table %u",
                                   ctx->natts,
                                   RelationGetDescr(ctx->rel)->natts));
        return;
    }
 
    /*
     * Check each attribute unless we hit corruption that confuses what to do
     * next, at which point we abort further attribute checks for this tuple.
     * Note that we don't abort for all types of corruption, only for those
     * types where we don't know how to continue.  We also don't abort the
     * checking of toasted attributes collected from the tuple prior to
     * aborting.  Those will still be checked later along with other toasted
     * attributes collected from the page.
     */
    ctx->offset = 0;
    for (ctx->attnum = 0; ctx->attnum < ctx->natts; ctx->attnum++)
        if (!check_tuple_attribute(ctx))
            break;              /* cannot continue */
 
    /* revert attnum to -1 until we again examine individual attributes */
    ctx->attnum = -1;
}

References HeapCheckContext::attnum, check_tuple_attribute(), check_tuple_header(), check_tuple_visibility(), HeapCheckContext::natts, HeapCheckContext::offset, psprintf(), HeapCheckContext::rel, RelationGetDescr, and report_corruption().

Referenced by verify_heapam().

check_tuple_attribute()

static bool check_tuple_attribute ( HeapCheckContext *  ctx )

static

Definition at line 1660 of file verify_heapam.c.

{
    Datum       attdatum;
    struct varlena *attr;
    char       *tp;             /* pointer to the tuple data */
    uint16      infomask;
    CompactAttribute *thisatt;
    struct varatt_external toast_pointer;
 
    infomask = ctx->tuphdr->t_infomask;
    thisatt = TupleDescCompactAttr(RelationGetDescr(ctx->rel), ctx->attnum);
 
    tp = (char *) ctx->tuphdr + ctx->tuphdr->t_hoff;
 
    if (ctx->tuphdr->t_hoff + ctx->offset > ctx->lp_len)
    {
        report_corruption(ctx,
                          psprintf("attribute with length %u starts at offset %u beyond total tuple length %u",
                                   thisatt->attlen,
                                   ctx->tuphdr->t_hoff + ctx->offset,
                                   ctx->lp_len));
        return false;
    }
 
    /* Skip null values */
    if (infomask & HEAP_HASNULL && att_isnull(ctx->attnum, ctx->tuphdr->t_bits))
        return true;
 
    /* Skip non-varlena values, but update offset first */
    if (thisatt->attlen != -1)
    {
        ctx->offset = att_nominal_alignby(ctx->offset, thisatt->attalignby);
        ctx->offset = att_addlength_pointer(ctx->offset, thisatt->attlen,
                                            tp + ctx->offset);
        if (ctx->tuphdr->t_hoff + ctx->offset > ctx->lp_len)
        {
            report_corruption(ctx,
                              psprintf("attribute with length %u ends at offset %u beyond total tuple length %u",
                                       thisatt->attlen,
                                       ctx->tuphdr->t_hoff + ctx->offset,
                                       ctx->lp_len));
            return false;
        }
        return true;
    }
 
    /* Ok, we're looking at a varlena attribute. */
    ctx->offset = att_pointer_alignby(ctx->offset, thisatt->attalignby, -1,
                                      tp + ctx->offset);
 
    /* Get the (possibly corrupt) varlena datum */
    attdatum = fetchatt(thisatt, tp + ctx->offset);
 
    /*
     * We have the datum, but we cannot decode it carelessly, as it may still
     * be corrupt.
     */
 
    /*
     * Check that VARTAG_SIZE won't hit an Assert on a corrupt va_tag before
     * risking a call into att_addlength_pointer
     */
    if (VARATT_IS_EXTERNAL(tp + ctx->offset))
    {
        uint8       va_tag = VARTAG_EXTERNAL(tp + ctx->offset);
 
        if (va_tag != VARTAG_ONDISK)
        {
            report_corruption(ctx,
                              psprintf("toasted attribute has unexpected TOAST tag %u",
                                       va_tag));
            /* We can't know where the next attribute begins */
            return false;
        }
    }
 
    /* Ok, should be safe now */
    ctx->offset = att_addlength_pointer(ctx->offset, thisatt->attlen,
                                        tp + ctx->offset);
 
    if (ctx->tuphdr->t_hoff + ctx->offset > ctx->lp_len)
    {
        report_corruption(ctx,
                          psprintf("attribute with length %u ends at offset %u beyond total tuple length %u",
                                   thisatt->attlen,
                                   ctx->tuphdr->t_hoff + ctx->offset,
                                   ctx->lp_len));
 
        return false;
    }
 
    /*
     * heap_deform_tuple would be done with this attribute at this point,
     * having stored it in values[], and would continue to the next attribute.
     * We go further, because we need to check if the toast datum is corrupt.
     */
 
    attr = (struct varlena *) DatumGetPointer(attdatum);
 
    /*
     * Now we follow the logic of detoast_external_attr(), with the same
     * caveats about being paranoid about corruption.
     */
 
    /* Skip values that are not external */
    if (!VARATT_IS_EXTERNAL(attr))
        return true;
 
    /* It is external, and we're looking at a page on disk */
 
    /*
     * Must copy attr into toast_pointer for alignment considerations
     */
    VARATT_EXTERNAL_GET_POINTER(toast_pointer, attr);
 
    /* Toasted attributes too large to be untoasted should never be stored */
    if (toast_pointer.va_rawsize > VARLENA_SIZE_LIMIT)
        report_corruption(ctx,
                          psprintf("toast value %u rawsize %d exceeds limit %d",
                                   toast_pointer.va_valueid,
                                   toast_pointer.va_rawsize,
                                   VARLENA_SIZE_LIMIT));
 
    if (VARATT_EXTERNAL_IS_COMPRESSED(toast_pointer))
    {
        ToastCompressionId cmid;
        bool        valid = false;
 
        /* Compressed attributes should have a valid compression method */
        cmid = TOAST_COMPRESS_METHOD(&toast_pointer);
        switch (cmid)
        {
                /* List of all valid compression method IDs */
            case TOAST_PGLZ_COMPRESSION_ID:
            case TOAST_LZ4_COMPRESSION_ID:
                valid = true;
                break;
 
                /* Recognized but invalid compression method ID */
            case TOAST_INVALID_COMPRESSION_ID:
                break;
 
                /* Intentionally no default here */
        }
        if (!valid)
            report_corruption(ctx,
                              psprintf("toast value %u has invalid compression method id %d",
                                       toast_pointer.va_valueid, cmid));
    }
 
    /* The tuple header better claim to contain toasted values */
    if (!(infomask & HEAP_HASEXTERNAL))
    {
        report_corruption(ctx,
                          psprintf("toast value %u is external but tuple header flag HEAP_HASEXTERNAL not set",
                                   toast_pointer.va_valueid));
        return true;
    }
 
    /* The relation better have a toast table */
    if (!ctx->rel->rd_rel->reltoastrelid)
    {
        report_corruption(ctx,
                          psprintf("toast value %u is external but relation has no toast relation",
                                   toast_pointer.va_valueid));
        return true;
    }
 
    /* If we were told to skip toast checking, then we're done. */
    if (ctx->toast_rel == NULL)
        return true;
 
    /*
     * If this tuple is eligible to be pruned, we cannot check the toast.
     * Otherwise, we push a copy of the toast tuple so we can check it after
     * releasing the main table buffer lock.
     */
    if (!ctx->tuple_could_be_pruned)
    {
        ToastedAttribute *ta;
 
        ta = (ToastedAttribute *) palloc0(sizeof(ToastedAttribute));
 
        VARATT_EXTERNAL_GET_POINTER(ta->toast_pointer, attr);
        ta->blkno = ctx->blkno;
        ta->offnum = ctx->offnum;
        ta->attnum = ctx->attnum;
        ctx->toasted_attributes = lappend(ctx->toasted_attributes, ta);
    }
 
    return true;
}

References att_addlength_pointer, att_isnull(), att_nominal_alignby, att_pointer_alignby, CompactAttribute::attalignby, CompactAttribute::attlen, ToastedAttribute::attnum, HeapCheckContext::attnum, ToastedAttribute::blkno, HeapCheckContext::blkno, DatumGetPointer(), fetchatt, HEAP_HASEXTERNAL, HEAP_HASNULL, if(), lappend(), HeapCheckContext::lp_len, ToastedAttribute::offnum, HeapCheckContext::offnum, HeapCheckContext::offset, palloc0(), psprintf(), RelationData::rd_rel, HeapCheckContext::rel, RelationGetDescr, report_corruption(), HeapTupleHeaderData::t_bits, HeapTupleHeaderData::t_hoff, HeapTupleHeaderData::t_infomask, TOAST_COMPRESS_METHOD, TOAST_INVALID_COMPRESSION_ID, TOAST_LZ4_COMPRESSION_ID, TOAST_PGLZ_COMPRESSION_ID, ToastedAttribute::toast_pointer, HeapCheckContext::toast_rel, HeapCheckContext::toasted_attributes, HeapCheckContext::tuphdr, HeapCheckContext::tuple_could_be_pruned, TupleDescCompactAttr(), varatt_external::va_rawsize, varatt_external::va_valueid, VARATT_EXTERNAL_GET_POINTER, VARATT_EXTERNAL_IS_COMPRESSED, VARATT_IS_EXTERNAL, VARLENA_SIZE_LIMIT, VARTAG_EXTERNAL, and VARTAG_ONDISK.

Referenced by check_tuple().

check_tuple_header()

static bool check_tuple_header ( HeapCheckContext *  ctx )

static

Definition at line 999 of file verify_heapam.c.

{
    HeapTupleHeader tuphdr = ctx->tuphdr;
    uint16      infomask = tuphdr->t_infomask;
    TransactionId curr_xmax = HeapTupleHeaderGetUpdateXid(tuphdr);
    bool        result = true;
    unsigned    expected_hoff;
 
    if (ctx->tuphdr->t_hoff > ctx->lp_len)
    {
        report_corruption(ctx,
                          psprintf("data begins at offset %u beyond the tuple length %u",
                                   ctx->tuphdr->t_hoff, ctx->lp_len));
        result = false;
    }
 
    if ((ctx->tuphdr->t_infomask & HEAP_XMAX_COMMITTED) &&
        (ctx->tuphdr->t_infomask & HEAP_XMAX_IS_MULTI))
    {
        report_corruption(ctx,
                          pstrdup("multixact should not be marked committed"));
 
        /*
         * This condition is clearly wrong, but it's not enough to justify
         * skipping further checks, because we don't rely on this to determine
         * whether the tuple is visible or to interpret other relevant header
         * fields.
         */
    }
 
    if (!TransactionIdIsValid(curr_xmax) &&
        HeapTupleHeaderIsHotUpdated(tuphdr))
    {
        report_corruption(ctx,
                          psprintf("tuple has been HOT updated, but xmax is 0"));
 
        /*
         * As above, even though this shouldn't happen, it's not sufficient
         * justification for skipping further checks, we should still be able
         * to perform sensibly.
         */
    }
 
    if (HeapTupleHeaderIsHeapOnly(tuphdr) &&
        ((tuphdr->t_infomask & HEAP_UPDATED) == 0))
    {
        report_corruption(ctx,
                          psprintf("tuple is heap only, but not the result of an update"));
 
        /* Here again, we can still perform further checks. */
    }
 
    if (infomask & HEAP_HASNULL)
        expected_hoff = MAXALIGN(SizeofHeapTupleHeader + BITMAPLEN(ctx->natts));
    else
        expected_hoff = MAXALIGN(SizeofHeapTupleHeader);
    if (ctx->tuphdr->t_hoff != expected_hoff)
    {
        if ((infomask & HEAP_HASNULL) && ctx->natts == 1)
            report_corruption(ctx,
                              psprintf("tuple data should begin at byte %u, but actually begins at byte %u (1 attribute, has nulls)",
                                       expected_hoff, ctx->tuphdr->t_hoff));
        else if ((infomask & HEAP_HASNULL))
            report_corruption(ctx,
                              psprintf("tuple data should begin at byte %u, but actually begins at byte %u (%u attributes, has nulls)",
                                       expected_hoff, ctx->tuphdr->t_hoff, ctx->natts));
        else if (ctx->natts == 1)
            report_corruption(ctx,
                              psprintf("tuple data should begin at byte %u, but actually begins at byte %u (1 attribute, no nulls)",
                                       expected_hoff, ctx->tuphdr->t_hoff));
        else
            report_corruption(ctx,
                              psprintf("tuple data should begin at byte %u, but actually begins at byte %u (%u attributes, no nulls)",
                                       expected_hoff, ctx->tuphdr->t_hoff, ctx->natts));
        result = false;
    }
 
    return result;
}

References BITMAPLEN(), HEAP_HASNULL, HEAP_UPDATED, HEAP_XMAX_COMMITTED, HEAP_XMAX_IS_MULTI, HeapTupleHeaderGetUpdateXid(), HeapTupleHeaderIsHeapOnly(), HeapTupleHeaderIsHotUpdated(), HeapCheckContext::lp_len, MAXALIGN, HeapCheckContext::natts, psprintf(), pstrdup(), report_corruption(), SizeofHeapTupleHeader, HeapTupleHeaderData::t_hoff, HeapTupleHeaderData::t_infomask, TransactionIdIsValid, and HeapCheckContext::tuphdr.

Referenced by check_tuple().

check_tuple_visibility()

static bool check_tuple_visibility ( HeapCheckContext *  ctx, bool *  xmin_commit_status_ok, XidCommitStatus *  xmin_commit_status  )

static

Definition at line 1112 of file verify_heapam.c.

{
    TransactionId xmin;
    TransactionId xvac;
    TransactionId xmax;
    XidCommitStatus xmin_status;
    XidCommitStatus xvac_status;
    XidCommitStatus xmax_status;
    HeapTupleHeader tuphdr = ctx->tuphdr;
 
    ctx->tuple_could_be_pruned = true;  /* have not yet proven otherwise */
    *xmin_commit_status_ok = false; /* have not yet proven otherwise */
 
    /* If xmin is normal, it should be within valid range */
    xmin = HeapTupleHeaderGetXmin(tuphdr);
    switch (get_xid_status(xmin, ctx, &xmin_status))
    {
        case XID_INVALID:
            /* Could be the result of a speculative insertion that aborted. */
            return false;
        case XID_BOUNDS_OK:
            *xmin_commit_status_ok = true;
            *xmin_commit_status = xmin_status;
            break;
        case XID_IN_FUTURE:
            report_corruption(ctx,
                              psprintf("xmin %u equals or exceeds next valid transaction ID %u:%u",
                                       xmin,
                                       EpochFromFullTransactionId(ctx->next_fxid),
                                       XidFromFullTransactionId(ctx->next_fxid)));
            return false;
        case XID_PRECEDES_CLUSTERMIN:
            report_corruption(ctx,
                              psprintf("xmin %u precedes oldest valid transaction ID %u:%u",
                                       xmin,
                                       EpochFromFullTransactionId(ctx->oldest_fxid),
                                       XidFromFullTransactionId(ctx->oldest_fxid)));
            return false;
        case XID_PRECEDES_RELMIN:
            report_corruption(ctx,
                              psprintf("xmin %u precedes relation freeze threshold %u:%u",
                                       xmin,
                                       EpochFromFullTransactionId(ctx->relfrozenfxid),
                                       XidFromFullTransactionId(ctx->relfrozenfxid)));
            return false;
    }
 
    /*
     * Has inserting transaction committed?
     */
    if (!HeapTupleHeaderXminCommitted(tuphdr))
    {
        if (HeapTupleHeaderXminInvalid(tuphdr))
            return false;       /* inserter aborted, don't check */
        /* Used by pre-9.0 binary upgrades */
        else if (tuphdr->t_infomask & HEAP_MOVED_OFF)
        {
            xvac = HeapTupleHeaderGetXvac(tuphdr);
 
            switch (get_xid_status(xvac, ctx, &xvac_status))
            {
                case XID_INVALID:
                    report_corruption(ctx,
                                      pstrdup("old-style VACUUM FULL transaction ID for moved off tuple is invalid"));
                    return false;
                case XID_IN_FUTURE:
                    report_corruption(ctx,
                                      psprintf("old-style VACUUM FULL transaction ID %u for moved off tuple equals or exceeds next valid transaction ID %u:%u",
                                               xvac,
                                               EpochFromFullTransactionId(ctx->next_fxid),
                                               XidFromFullTransactionId(ctx->next_fxid)));
                    return false;
                case XID_PRECEDES_RELMIN:
                    report_corruption(ctx,
                                      psprintf("old-style VACUUM FULL transaction ID %u for moved off tuple precedes relation freeze threshold %u:%u",
                                               xvac,
                                               EpochFromFullTransactionId(ctx->relfrozenfxid),
                                               XidFromFullTransactionId(ctx->relfrozenfxid)));
                    return false;
                case XID_PRECEDES_CLUSTERMIN:
                    report_corruption(ctx,
                                      psprintf("old-style VACUUM FULL transaction ID %u for moved off tuple precedes oldest valid transaction ID %u:%u",
                                               xvac,
                                               EpochFromFullTransactionId(ctx->oldest_fxid),
                                               XidFromFullTransactionId(ctx->oldest_fxid)));
                    return false;
                case XID_BOUNDS_OK:
                    break;
            }
 
            switch (xvac_status)
            {
                case XID_IS_CURRENT_XID:
                    report_corruption(ctx,
                                      psprintf("old-style VACUUM FULL transaction ID %u for moved off tuple matches our current transaction ID",
                                               xvac));
                    return false;
                case XID_IN_PROGRESS:
                    report_corruption(ctx,
                                      psprintf("old-style VACUUM FULL transaction ID %u for moved off tuple appears to be in progress",
                                               xvac));
                    return false;
 
                case XID_COMMITTED:
 
                    /*
                     * The tuple is dead, because the xvac transaction moved
                     * it off and committed. It's checkable, but also
                     * prunable.
                     */
                    return true;
 
                case XID_ABORTED:
 
                    /*
                     * The original xmin must have committed, because the xvac
                     * transaction tried to move it later. Since xvac is
                     * aborted, whether it's still alive now depends on the
                     * status of xmax.
                     */
                    break;
            }
        }
        /* Used by pre-9.0 binary upgrades */
        else if (tuphdr->t_infomask & HEAP_MOVED_IN)
        {
            xvac = HeapTupleHeaderGetXvac(tuphdr);
 
            switch (get_xid_status(xvac, ctx, &xvac_status))
            {
                case XID_INVALID:
                    report_corruption(ctx,
                                      pstrdup("old-style VACUUM FULL transaction ID for moved in tuple is invalid"));
                    return false;
                case XID_IN_FUTURE:
                    report_corruption(ctx,
                                      psprintf("old-style VACUUM FULL transaction ID %u for moved in tuple equals or exceeds next valid transaction ID %u:%u",
                                               xvac,
                                               EpochFromFullTransactionId(ctx->next_fxid),
                                               XidFromFullTransactionId(ctx->next_fxid)));
                    return false;
                case XID_PRECEDES_RELMIN:
                    report_corruption(ctx,
                                      psprintf("old-style VACUUM FULL transaction ID %u for moved in tuple precedes relation freeze threshold %u:%u",
                                               xvac,
                                               EpochFromFullTransactionId(ctx->relfrozenfxid),
                                               XidFromFullTransactionId(ctx->relfrozenfxid)));
                    return false;
                case XID_PRECEDES_CLUSTERMIN:
                    report_corruption(ctx,
                                      psprintf("old-style VACUUM FULL transaction ID %u for moved in tuple precedes oldest valid transaction ID %u:%u",
                                               xvac,
                                               EpochFromFullTransactionId(ctx->oldest_fxid),
                                               XidFromFullTransactionId(ctx->oldest_fxid)));
                    return false;
                case XID_BOUNDS_OK:
                    break;
            }
 
            switch (xvac_status)
            {
                case XID_IS_CURRENT_XID:
                    report_corruption(ctx,
                                      psprintf("old-style VACUUM FULL transaction ID %u for moved in tuple matches our current transaction ID",
                                               xvac));
                    return false;
                case XID_IN_PROGRESS:
                    report_corruption(ctx,
                                      psprintf("old-style VACUUM FULL transaction ID %u for moved in tuple appears to be in progress",
                                               xvac));
                    return false;
 
                case XID_COMMITTED:
 
                    /*
                     * The original xmin must have committed, because the xvac
                     * transaction moved it later. Whether it's still alive
                     * now depends on the status of xmax.
                     */
                    break;
 
                case XID_ABORTED:
 
                    /*
                     * The tuple is dead, because the xvac transaction moved
                     * it off and committed. It's checkable, but also
                     * prunable.
                     */
                    return true;
            }
        }
        else if (xmin_status != XID_COMMITTED)
        {
            /*
             * Inserting transaction is not in progress, and not committed, so
             * it might have changed the TupleDesc in ways we don't know
             * about. Thus, don't try to check the tuple structure.
             *
             * If xmin_status happens to be XID_IS_CURRENT_XID, then in theory
             * any such DDL changes ought to be visible to us, so perhaps we
             * could check anyway in that case. But, for now, let's be
             * conservative and treat this like any other uncommitted insert.
             */
            return false;
        }
    }
 
    /*
     * Okay, the inserter committed, so it was good at some point.  Now what
     * about the deleting transaction?
     */
 
    if (tuphdr->t_infomask & HEAP_XMAX_IS_MULTI)
    {
        /*
         * xmax is a multixact, so sanity-check the MXID. Note that we do this
         * prior to checking for HEAP_XMAX_INVALID or
         * HEAP_XMAX_IS_LOCKED_ONLY. This might therefore complain about
         * things that wouldn't actually be a problem during a normal scan,
         * but eventually we're going to have to freeze, and that process will
         * ignore hint bits.
         *
         * Even if the MXID is out of range, we still know that the original
         * insert committed, so we can check the tuple itself. However, we
         * can't rule out the possibility that this tuple is dead, so don't
         * clear ctx->tuple_could_be_pruned. Possibly we should go ahead and
         * clear that flag anyway if HEAP_XMAX_INVALID is set or if
         * HEAP_XMAX_IS_LOCKED_ONLY is true, but for now we err on the side of
         * avoiding possibly-bogus complaints about missing TOAST entries.
         */
        xmax = HeapTupleHeaderGetRawXmax(tuphdr);
        switch (check_mxid_valid_in_rel(xmax, ctx))
        {
            case XID_INVALID:
                report_corruption(ctx,
                                  pstrdup("multitransaction ID is invalid"));
                return true;
            case XID_PRECEDES_RELMIN:
                report_corruption(ctx,
                                  psprintf("multitransaction ID %u precedes relation minimum multitransaction ID threshold %u",
                                           xmax, ctx->relminmxid));
                return true;
            case XID_PRECEDES_CLUSTERMIN:
                report_corruption(ctx,
                                  psprintf("multitransaction ID %u precedes oldest valid multitransaction ID threshold %u",
                                           xmax, ctx->oldest_mxact));
                return true;
            case XID_IN_FUTURE:
                report_corruption(ctx,
                                  psprintf("multitransaction ID %u equals or exceeds next valid multitransaction ID %u",
                                           xmax,
                                           ctx->next_mxact));
                return true;
            case XID_BOUNDS_OK:
                break;
        }
    }
 
    if (tuphdr->t_infomask & HEAP_XMAX_INVALID)
    {
        /*
         * This tuple is live.  A concurrently running transaction could
         * delete it before we get around to checking the toast, but any such
         * running transaction is surely not less than our safe_xmin, so the
         * toast cannot be vacuumed out from under us.
         */
        ctx->tuple_could_be_pruned = false;
        return true;
    }
 
    if (HEAP_XMAX_IS_LOCKED_ONLY(tuphdr->t_infomask))
    {
        /*
         * "Deleting" xact really only locked it, so the tuple is live in any
         * case.  As above, a concurrently running transaction could delete
         * it, but it cannot be vacuumed out from under us.
         */
        ctx->tuple_could_be_pruned = false;
        return true;
    }
 
    if (tuphdr->t_infomask & HEAP_XMAX_IS_MULTI)
    {
        /*
         * We already checked above that this multixact is within limits for
         * this table.  Now check the update xid from this multixact.
         */
        xmax = HeapTupleGetUpdateXid(tuphdr);
        switch (get_xid_status(xmax, ctx, &xmax_status))
        {
            case XID_INVALID:
                /* not LOCKED_ONLY, so it has to have an xmax */
                report_corruption(ctx,
                                  pstrdup("update xid is invalid"));
                return true;
            case XID_IN_FUTURE:
                report_corruption(ctx,
                                  psprintf("update xid %u equals or exceeds next valid transaction ID %u:%u",
                                           xmax,
                                           EpochFromFullTransactionId(ctx->next_fxid),
                                           XidFromFullTransactionId(ctx->next_fxid)));
                return true;
            case XID_PRECEDES_RELMIN:
                report_corruption(ctx,
                                  psprintf("update xid %u precedes relation freeze threshold %u:%u",
                                           xmax,
                                           EpochFromFullTransactionId(ctx->relfrozenfxid),
                                           XidFromFullTransactionId(ctx->relfrozenfxid)));
                return true;
            case XID_PRECEDES_CLUSTERMIN:
                report_corruption(ctx,
                                  psprintf("update xid %u precedes oldest valid transaction ID %u:%u",
                                           xmax,
                                           EpochFromFullTransactionId(ctx->oldest_fxid),
                                           XidFromFullTransactionId(ctx->oldest_fxid)));
                return true;
            case XID_BOUNDS_OK:
                break;
        }
 
        switch (xmax_status)
        {
            case XID_IS_CURRENT_XID:
            case XID_IN_PROGRESS:
 
                /*
                 * The delete is in progress, so it cannot be visible to our
                 * snapshot.
                 */
                ctx->tuple_could_be_pruned = false;
                break;
            case XID_COMMITTED:
 
                /*
                 * The delete committed.  Whether the toast can be vacuumed
                 * away depends on how old the deleting transaction is.
                 */
                ctx->tuple_could_be_pruned = TransactionIdPrecedes(xmax,
                                                                   ctx->safe_xmin);
                break;
            case XID_ABORTED:
 
                /*
                 * The delete aborted or crashed.  The tuple is still live.
                 */
                ctx->tuple_could_be_pruned = false;
                break;
        }
 
        /* Tuple itself is checkable even if it's dead. */
        return true;
    }
 
    /* xmax is an XID, not a MXID. Sanity check it. */
    xmax = HeapTupleHeaderGetRawXmax(tuphdr);
    switch (get_xid_status(xmax, ctx, &xmax_status))
    {
        case XID_INVALID:
            ctx->tuple_could_be_pruned = false;
            return true;
        case XID_IN_FUTURE:
            report_corruption(ctx,
                              psprintf("xmax %u equals or exceeds next valid transaction ID %u:%u",
                                       xmax,
                                       EpochFromFullTransactionId(ctx->next_fxid),
                                       XidFromFullTransactionId(ctx->next_fxid)));
            return false;       /* corrupt */
        case XID_PRECEDES_RELMIN:
            report_corruption(ctx,
                              psprintf("xmax %u precedes relation freeze threshold %u:%u",
                                       xmax,
                                       EpochFromFullTransactionId(ctx->relfrozenfxid),
                                       XidFromFullTransactionId(ctx->relfrozenfxid)));
            return false;       /* corrupt */
        case XID_PRECEDES_CLUSTERMIN:
            report_corruption(ctx,
                              psprintf("xmax %u precedes oldest valid transaction ID %u:%u",
                                       xmax,
                                       EpochFromFullTransactionId(ctx->oldest_fxid),
                                       XidFromFullTransactionId(ctx->oldest_fxid)));
            return false;       /* corrupt */
        case XID_BOUNDS_OK:
            break;
    }
 
    /*
     * Whether the toast can be vacuumed away depends on how old the deleting
     * transaction is.
     */
    switch (xmax_status)
    {
        case XID_IS_CURRENT_XID:
        case XID_IN_PROGRESS:
 
            /*
             * The delete is in progress, so it cannot be visible to our
             * snapshot.
             */
            ctx->tuple_could_be_pruned = false;
            break;
 
        case XID_COMMITTED:
 
            /*
             * The delete committed.  Whether the toast can be vacuumed away
             * depends on how old the deleting transaction is.
             */
            ctx->tuple_could_be_pruned = TransactionIdPrecedes(xmax,
                                                               ctx->safe_xmin);
            break;
 
        case XID_ABORTED:
 
            /*
             * The delete aborted or crashed.  The tuple is still live.
             */
            ctx->tuple_could_be_pruned = false;
            break;
    }
 
    /* Tuple itself is checkable even if it's dead. */
    return true;
}

References check_mxid_valid_in_rel(), EpochFromFullTransactionId, get_xid_status(), HEAP_MOVED_IN, HEAP_MOVED_OFF, HEAP_XMAX_INVALID, HEAP_XMAX_IS_LOCKED_ONLY(), HEAP_XMAX_IS_MULTI, HeapTupleGetUpdateXid(), HeapTupleHeaderGetRawXmax(), HeapTupleHeaderGetXmin(), HeapTupleHeaderGetXvac(), HeapTupleHeaderXminCommitted(), HeapTupleHeaderXminInvalid(), HeapCheckContext::next_fxid, HeapCheckContext::next_mxact, HeapCheckContext::oldest_fxid, HeapCheckContext::oldest_mxact, psprintf(), pstrdup(), HeapCheckContext::relfrozenfxid, HeapCheckContext::relminmxid, report_corruption(), HeapCheckContext::safe_xmin, HeapTupleHeaderData::t_infomask, TransactionIdPrecedes(), HeapCheckContext::tuphdr, HeapCheckContext::tuple_could_be_pruned, XID_ABORTED, XID_BOUNDS_OK, XID_COMMITTED, XID_IN_FUTURE, XID_IN_PROGRESS, XID_INVALID, XID_IS_CURRENT_XID, XID_PRECEDES_CLUSTERMIN, XID_PRECEDES_RELMIN, and XidFromFullTransactionId.

Referenced by check_tuple().

FullTransactionIdFromXidAndCtx()

static FullTransactionId FullTransactionIdFromXidAndCtx ( TransactionId  xid, const HeapCheckContext *  ctx  )

static

Definition at line 1977 of file verify_heapam.c.

{
    uint64      nextfxid_i;
    int32       diff;
    FullTransactionId fxid;
 
    Assert(TransactionIdIsNormal(ctx->next_xid));
    Assert(FullTransactionIdIsNormal(ctx->next_fxid));
    Assert(XidFromFullTransactionId(ctx->next_fxid) == ctx->next_xid);
 
    if (!TransactionIdIsNormal(xid))
        return FullTransactionIdFromEpochAndXid(0, xid);
 
    nextfxid_i = U64FromFullTransactionId(ctx->next_fxid);
 
    /* compute the 32bit modulo difference */
    diff = (int32) (ctx->next_xid - xid);
 
    /*
     * In cases of corruption we might see a 32bit xid that is before epoch 0.
     * We can't represent that as a 64bit xid, due to 64bit xids being
     * unsigned integers, without the modulo arithmetic of 32bit xid. There's
     * no really nice way to deal with that, but it works ok enough to use
     * FirstNormalFullTransactionId in that case, as a freshly initdb'd
     * cluster already has a newer horizon.
     */
    if (diff > 0 && (nextfxid_i - FirstNormalTransactionId) < (int64) diff)
    {
        Assert(EpochFromFullTransactionId(ctx->next_fxid) == 0);
        fxid = FirstNormalFullTransactionId;
    }
    else
        fxid = FullTransactionIdFromU64(nextfxid_i - diff);
 
    Assert(FullTransactionIdIsNormal(fxid));
    return fxid;
}

References Assert(), EpochFromFullTransactionId, FirstNormalFullTransactionId, FirstNormalTransactionId, FullTransactionIdFromEpochAndXid(), FullTransactionIdFromU64(), FullTransactionIdIsNormal, HeapCheckContext::next_fxid, HeapCheckContext::next_xid, TransactionIdIsNormal, U64FromFullTransactionId, and XidFromFullTransactionId.

Referenced by get_xid_status(), update_cached_xid_range(), and verify_heapam().

fxid_in_cached_range()

static bool fxid_in_cached_range ( FullTransactionId  fxid, const HeapCheckContext *  ctx  )

inlinestatic

Definition at line 2046 of file verify_heapam.c.

{
    return (FullTransactionIdPrecedesOrEquals(ctx->oldest_fxid, fxid) &&
            FullTransactionIdPrecedes(fxid, ctx->next_fxid));
}

References FullTransactionIdPrecedes, FullTransactionIdPrecedesOrEquals, HeapCheckContext::next_fxid, and HeapCheckContext::oldest_fxid.

Referenced by get_xid_status().

get_xid_status()

static XidBoundsViolation get_xid_status ( TransactionId  xid, HeapCheckContext *  ctx, XidCommitStatus *  status  )

static

Definition at line 2111 of file verify_heapam.c.

{
    FullTransactionId fxid;
    FullTransactionId clog_horizon;
 
    /* Quick check for special xids */
    if (!TransactionIdIsValid(xid))
        return XID_INVALID;
    else if (xid == BootstrapTransactionId || xid == FrozenTransactionId)
    {
        if (status != NULL)
            *status = XID_COMMITTED;
        return XID_BOUNDS_OK;
    }
 
    /* Check if the xid is within bounds */
    fxid = FullTransactionIdFromXidAndCtx(xid, ctx);
    if (!fxid_in_cached_range(fxid, ctx))
    {
        /*
         * We may have been checking against stale values.  Update the cached
         * range to be sure, and since we relied on the cached range when we
         * performed the full xid conversion, reconvert.
         */
        update_cached_xid_range(ctx);
        fxid = FullTransactionIdFromXidAndCtx(xid, ctx);
    }
 
    if (FullTransactionIdPrecedesOrEquals(ctx->next_fxid, fxid))
        return XID_IN_FUTURE;
    if (FullTransactionIdPrecedes(fxid, ctx->oldest_fxid))
        return XID_PRECEDES_CLUSTERMIN;
    if (FullTransactionIdPrecedes(fxid, ctx->relfrozenfxid))
        return XID_PRECEDES_RELMIN;
 
    /* Early return if the caller does not request clog checking */
    if (status == NULL)
        return XID_BOUNDS_OK;
 
    /* Early return if we just checked this xid in a prior call */
    if (xid == ctx->cached_xid)
    {
        *status = ctx->cached_status;
        return XID_BOUNDS_OK;
    }
 
    *status = XID_COMMITTED;
    LWLockAcquire(XactTruncationLock, LW_SHARED);
    clog_horizon =
        FullTransactionIdFromXidAndCtx(TransamVariables->oldestClogXid,
                                       ctx);
    if (FullTransactionIdPrecedesOrEquals(clog_horizon, fxid))
    {
        if (TransactionIdIsCurrentTransactionId(xid))
            *status = XID_IS_CURRENT_XID;
        else if (TransactionIdIsInProgress(xid))
            *status = XID_IN_PROGRESS;
        else if (TransactionIdDidCommit(xid))
            *status = XID_COMMITTED;
        else
            *status = XID_ABORTED;
    }
    LWLockRelease(XactTruncationLock);
    ctx->cached_xid = xid;
    ctx->cached_status = *status;
    return XID_BOUNDS_OK;
}

References BootstrapTransactionId, HeapCheckContext::cached_status, HeapCheckContext::cached_xid, FrozenTransactionId, FullTransactionIdFromXidAndCtx(), FullTransactionIdPrecedes, FullTransactionIdPrecedesOrEquals, fxid_in_cached_range(), LW_SHARED, LWLockAcquire(), LWLockRelease(), HeapCheckContext::next_fxid, HeapCheckContext::oldest_fxid, TransamVariablesData::oldestClogXid, HeapCheckContext::relfrozenfxid, TransactionIdDidCommit(), TransactionIdIsCurrentTransactionId(), TransactionIdIsInProgress(), TransactionIdIsValid, TransamVariables, update_cached_xid_range(), XID_ABORTED, XID_BOUNDS_OK, XID_COMMITTED, XID_IN_FUTURE, XID_IN_PROGRESS, XID_INVALID, XID_IS_CURRENT_XID, XID_PRECEDES_CLUSTERMIN, and XID_PRECEDES_RELMIN.

Referenced by check_tuple_visibility().

heapcheck_read_stream_next_unskippable()

static BlockNumber heapcheck_read_stream_next_unskippable ( ReadStream *  stream, void *  callback_private_data, void *  per_buffer_data  )

static

Definition at line 881 of file verify_heapam.c.

{
    HeapCheckReadStreamData *p = callback_private_data;
 
    /* Loops over [current_blocknum, last_exclusive) blocks */
    for (BlockNumber i; (i = p->range.current_blocknum++) < p->range.last_exclusive;)
    {
        uint8       mapbits = visibilitymap_get_status(p->rel, i, p->vmbuffer);
 
        if (p->skip_option == SKIP_PAGES_ALL_FROZEN)
        {
            if ((mapbits & VISIBILITYMAP_ALL_FROZEN) != 0)
                continue;
        }
 
        if (p->skip_option == SKIP_PAGES_ALL_VISIBLE)
        {
            if ((mapbits & VISIBILITYMAP_ALL_VISIBLE) != 0)
                continue;
        }
 
        return i;
    }
 
    return InvalidBlockNumber;
}

References BlockRangeReadStreamPrivate::current_blocknum, i, InvalidBlockNumber, BlockRangeReadStreamPrivate::last_exclusive, HeapCheckReadStreamData::range, HeapCheckReadStreamData::rel, HeapCheckReadStreamData::skip_option, SKIP_PAGES_ALL_FROZEN, SKIP_PAGES_ALL_VISIBLE, VISIBILITYMAP_ALL_FROZEN, VISIBILITYMAP_ALL_VISIBLE, visibilitymap_get_status(), and HeapCheckReadStreamData::vmbuffer.

Referenced by verify_heapam().

PG_FUNCTION_INFO_V1()

PG_FUNCTION_INFO_V1

(

verify_heapam

)

report_corruption()

static void report_corruption ( HeapCheckContext *  ctx, char *  msg  )

static

Definition at line 951 of file verify_heapam.c.

{
    report_corruption_internal(ctx->tupstore, ctx->tupdesc, ctx->blkno,
                               ctx->offnum, ctx->attnum, msg);
    ctx->is_corrupt = true;
}

References HeapCheckContext::attnum, HeapCheckContext::blkno, HeapCheckContext::is_corrupt, HeapCheckContext::offnum, report_corruption_internal(), HeapCheckContext::tupdesc, and HeapCheckContext::tupstore.

Referenced by check_tuple(), check_tuple_attribute(), check_tuple_header(), check_tuple_visibility(), and verify_heapam().

report_corruption_internal()

static void report_corruption_internal ( Tuplestorestate *  tupstore, TupleDesc  tupdesc, BlockNumber  blkno, OffsetNumber  offnum, AttrNumber  attnum, char *  msg  )

static

Definition at line 915 of file verify_heapam.c.

{
    Datum       values[HEAPCHECK_RELATION_COLS] = {0};
    bool        nulls[HEAPCHECK_RELATION_COLS] = {0};
    HeapTuple   tuple;
 
    values[0] = Int64GetDatum(blkno);
    values[1] = Int32GetDatum(offnum);
    values[2] = Int32GetDatum(attnum);
    nulls[2] = (attnum < 0);
    values[3] = CStringGetTextDatum(msg);
 
    /*
     * In principle, there is nothing to prevent a scan over a large, highly
     * corrupted table from using work_mem worth of memory building up the
     * tuplestore.  That's ok, but if we also leak the msg argument memory
     * until the end of the query, we could exceed work_mem by more than a
     * trivial amount.  Therefore, free the msg argument each time we are
     * called rather than waiting for our current memory context to be freed.
     */
    pfree(msg);
 
    tuple = heap_form_tuple(tupdesc, values, nulls);
    tuplestore_puttuple(tupstore, tuple);
}

References attnum, CStringGetTextDatum, heap_form_tuple(), HEAPCHECK_RELATION_COLS, Int32GetDatum(), Int64GetDatum(), pfree(), tuplestore_puttuple(), and values.

Referenced by report_corruption(), and report_toast_corruption().

report_toast_corruption()

static void report_toast_corruption ( HeapCheckContext *  ctx, ToastedAttribute *  ta, char *  msg  )

static

Definition at line 967 of file verify_heapam.c.

{
    report_corruption_internal(ctx->tupstore, ctx->tupdesc, ta->blkno,
                               ta->offnum, ta->attnum, msg);
    ctx->is_corrupt = true;
}

References ToastedAttribute::attnum, ToastedAttribute::blkno, HeapCheckContext::is_corrupt, ToastedAttribute::offnum, report_corruption_internal(), HeapCheckContext::tupdesc, and HeapCheckContext::tupstore.

Referenced by check_toast_tuple(), and check_toasted_attribute().

update_cached_mxid_range()

static void update_cached_mxid_range ( HeapCheckContext *  ctx )

static

Definition at line 2036 of file verify_heapam.c.

{
    ReadMultiXactIdRange(&ctx->oldest_mxact, &ctx->next_mxact);
}

References HeapCheckContext::next_mxact, HeapCheckContext::oldest_mxact, and ReadMultiXactIdRange().

Referenced by check_mxid_valid_in_rel(), and verify_heapam().

update_cached_xid_range()

static void update_cached_xid_range ( HeapCheckContext *  ctx )

static

Definition at line 2019 of file verify_heapam.c.

{
    /* Make cached copies */
    LWLockAcquire(XidGenLock, LW_SHARED);
    ctx->next_fxid = TransamVariables->nextXid;
    ctx->oldest_xid = TransamVariables->oldestXid;
    LWLockRelease(XidGenLock);
 
    /* And compute alternate versions of the same */
    ctx->next_xid = XidFromFullTransactionId(ctx->next_fxid);
    ctx->oldest_fxid = FullTransactionIdFromXidAndCtx(ctx->oldest_xid, ctx);
}

References FullTransactionIdFromXidAndCtx(), LW_SHARED, LWLockAcquire(), LWLockRelease(), HeapCheckContext::next_fxid, HeapCheckContext::next_xid, TransamVariablesData::nextXid, HeapCheckContext::oldest_fxid, HeapCheckContext::oldest_xid, TransamVariablesData::oldestXid, TransamVariables, and XidFromFullTransactionId.

Referenced by get_xid_status(), and verify_heapam().

verify_heapam()

Datum verify_heapam

(

PG_FUNCTION_ARGS

)

Definition at line 250 of file verify_heapam.c.

{
    ReturnSetInfo *rsinfo = (ReturnSetInfo *) fcinfo->resultinfo;
    HeapCheckContext ctx;
    Buffer      vmbuffer = InvalidBuffer;
    Oid         relid;
    bool        on_error_stop;
    bool        check_toast;
    SkipPages   skip_option = SKIP_PAGES_NONE;
    BlockNumber first_block;
    BlockNumber last_block;
    BlockNumber nblocks;
    const char *skip;
    ReadStream *stream;
    int         stream_flags;
    ReadStreamBlockNumberCB stream_cb;
    void       *stream_data;
    HeapCheckReadStreamData stream_skip_data;
 
    /* Check supplied arguments */
    if (PG_ARGISNULL(0))
        ereport(ERROR,
                (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
                 errmsg("relation cannot be null")));
    relid = PG_GETARG_OID(0);
 
    if (PG_ARGISNULL(1))
        ereport(ERROR,
                (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
                 errmsg("on_error_stop cannot be null")));
    on_error_stop = PG_GETARG_BOOL(1);
 
    if (PG_ARGISNULL(2))
        ereport(ERROR,
                (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
                 errmsg("check_toast cannot be null")));
    check_toast = PG_GETARG_BOOL(2);
 
    if (PG_ARGISNULL(3))
        ereport(ERROR,
                (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
                 errmsg("skip cannot be null")));
    skip = text_to_cstring(PG_GETARG_TEXT_PP(3));
    if (pg_strcasecmp(skip, "all-visible") == 0)
        skip_option = SKIP_PAGES_ALL_VISIBLE;
    else if (pg_strcasecmp(skip, "all-frozen") == 0)
        skip_option = SKIP_PAGES_ALL_FROZEN;
    else if (pg_strcasecmp(skip, "none") == 0)
        skip_option = SKIP_PAGES_NONE;
    else
        ereport(ERROR,
                (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
                 errmsg("invalid skip option"),
                 errhint("Valid skip options are \"all-visible\", \"all-frozen\", and \"none\".")));
 
    memset(&ctx, 0, sizeof(HeapCheckContext));
    ctx.cached_xid = InvalidTransactionId;
    ctx.toasted_attributes = NIL;
 
    /*
     * Any xmin newer than the xmin of our snapshot can't become all-visible
     * while we're running.
     */
    ctx.safe_xmin = GetTransactionSnapshot()->xmin;
 
    /*
     * If we report corruption when not examining some individual attribute,
     * we need attnum to be reported as NULL.  Set that up before any
     * corruption reporting might happen.
     */
    ctx.attnum = -1;
 
    /* Construct the tuplestore and tuple descriptor */
    InitMaterializedSRF(fcinfo, 0);
    ctx.tupdesc = rsinfo->setDesc;
    ctx.tupstore = rsinfo->setResult;
 
    /* Open relation, check relkind and access method */
    ctx.rel = relation_open(relid, AccessShareLock);
 
    /*
     * Check that a relation's relkind and access method are both supported.
     */
    if (!RELKIND_HAS_TABLE_AM(ctx.rel->rd_rel->relkind) &&
        ctx.rel->rd_rel->relkind != RELKIND_SEQUENCE)
        ereport(ERROR,
                (errcode(ERRCODE_WRONG_OBJECT_TYPE),
                 errmsg("cannot check relation \"%s\"",
                        RelationGetRelationName(ctx.rel)),
                 errdetail_relkind_not_supported(ctx.rel->rd_rel->relkind)));
 
    /*
     * Sequences always use heap AM, but they don't show that in the catalogs.
     * Other relkinds might be using a different AM, so check.
     */
    if (ctx.rel->rd_rel->relkind != RELKIND_SEQUENCE &&
        ctx.rel->rd_rel->relam != HEAP_TABLE_AM_OID)
        ereport(ERROR,
                (errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
                 errmsg("only heap AM is supported")));
 
    /*
     * Early exit for unlogged relations during recovery.  These will have no
     * relation fork, so there won't be anything to check.  We behave as if
     * the relation is empty.
     */
    if (ctx.rel->rd_rel->relpersistence == RELPERSISTENCE_UNLOGGED &&
        RecoveryInProgress())
    {
        ereport(DEBUG1,
                (errcode(ERRCODE_READ_ONLY_SQL_TRANSACTION),
                 errmsg("cannot verify unlogged relation \"%s\" during recovery, skipping",
                        RelationGetRelationName(ctx.rel))));
        relation_close(ctx.rel, AccessShareLock);
        PG_RETURN_NULL();
    }
 
    /* Early exit if the relation is empty */
    nblocks = RelationGetNumberOfBlocks(ctx.rel);
    if (!nblocks)
    {
        relation_close(ctx.rel, AccessShareLock);
        PG_RETURN_NULL();
    }
 
    ctx.bstrategy = GetAccessStrategy(BAS_BULKREAD);
    ctx.buffer = InvalidBuffer;
    ctx.page = NULL;
 
    /* Validate block numbers, or handle nulls. */
    if (PG_ARGISNULL(4))
        first_block = 0;
    else
    {
        int64       fb = PG_GETARG_INT64(4);
 
        if (fb < 0 || fb >= nblocks)
            ereport(ERROR,
                    (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
                     errmsg("starting block number must be between 0 and %u",
                            nblocks - 1)));
        first_block = (BlockNumber) fb;
    }
    if (PG_ARGISNULL(5))
        last_block = nblocks - 1;
    else
    {
        int64       lb = PG_GETARG_INT64(5);
 
        if (lb < 0 || lb >= nblocks)
            ereport(ERROR,
                    (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
                     errmsg("ending block number must be between 0 and %u",
                            nblocks - 1)));
        last_block = (BlockNumber) lb;
    }
 
    /* Optionally open the toast relation, if any. */
    if (ctx.rel->rd_rel->reltoastrelid && check_toast)
    {
        int         offset;
 
        /* Main relation has associated toast relation */
        ctx.toast_rel = table_open(ctx.rel->rd_rel->reltoastrelid,
                                   AccessShareLock);
        offset = toast_open_indexes(ctx.toast_rel,
                                    AccessShareLock,
                                    &(ctx.toast_indexes),
                                    &(ctx.num_toast_indexes));
        ctx.valid_toast_index = ctx.toast_indexes[offset];
    }
    else
    {
        /*
         * Main relation has no associated toast relation, or we're
         * intentionally skipping it.
         */
        ctx.toast_rel = NULL;
        ctx.toast_indexes = NULL;
        ctx.num_toast_indexes = 0;
    }
 
    update_cached_xid_range(&ctx);
    update_cached_mxid_range(&ctx);
    ctx.relfrozenxid = ctx.rel->rd_rel->relfrozenxid;
    ctx.relfrozenfxid = FullTransactionIdFromXidAndCtx(ctx.relfrozenxid, &ctx);
    ctx.relminmxid = ctx.rel->rd_rel->relminmxid;
 
    if (TransactionIdIsNormal(ctx.relfrozenxid))
        ctx.oldest_xid = ctx.relfrozenxid;
 
    /* Now that `ctx` is set up, set up the read stream */
    stream_skip_data.range.current_blocknum = first_block;
    stream_skip_data.range.last_exclusive = last_block + 1;
    stream_skip_data.skip_option = skip_option;
    stream_skip_data.rel = ctx.rel;
    stream_skip_data.vmbuffer = &vmbuffer;
 
    if (skip_option == SKIP_PAGES_NONE)
    {
        /*
         * It is safe to use batchmode as block_range_read_stream_cb takes no
         * locks.
         */
        stream_cb = block_range_read_stream_cb;
        stream_flags = READ_STREAM_SEQUENTIAL |
            READ_STREAM_FULL |
            READ_STREAM_USE_BATCHING;
        stream_data = &stream_skip_data.range;
    }
    else
    {
        /*
         * It would not be safe to naively use batchmode, as
         * heapcheck_read_stream_next_unskippable takes locks. It shouldn't be
         * too hard to convert though.
         */
        stream_cb = heapcheck_read_stream_next_unskippable;
        stream_flags = READ_STREAM_DEFAULT;
        stream_data = &stream_skip_data;
    }
 
    stream = read_stream_begin_relation(stream_flags,
                                        ctx.bstrategy,
                                        ctx.rel,
                                        MAIN_FORKNUM,
                                        stream_cb,
                                        stream_data,
                                        0);
 
    while ((ctx.buffer = read_stream_next_buffer(stream, NULL)) != InvalidBuffer)
    {
        OffsetNumber maxoff;
        OffsetNumber predecessor[MaxOffsetNumber];
        OffsetNumber successor[MaxOffsetNumber];
        bool        lp_valid[MaxOffsetNumber];
        bool        xmin_commit_status_ok[MaxOffsetNumber];
        XidCommitStatus xmin_commit_status[MaxOffsetNumber];
 
        CHECK_FOR_INTERRUPTS();
 
        memset(predecessor, 0, sizeof(OffsetNumber) * MaxOffsetNumber);
 
        /* Lock the next page. */
        Assert(BufferIsValid(ctx.buffer));
        LockBuffer(ctx.buffer, BUFFER_LOCK_SHARE);
 
        ctx.blkno = BufferGetBlockNumber(ctx.buffer);
        ctx.page = BufferGetPage(ctx.buffer);
 
        /* Perform tuple checks */
        maxoff = PageGetMaxOffsetNumber(ctx.page);
        for (ctx.offnum = FirstOffsetNumber; ctx.offnum <= maxoff;
             ctx.offnum = OffsetNumberNext(ctx.offnum))
        {
            BlockNumber nextblkno;
            OffsetNumber nextoffnum;
 
            successor[ctx.offnum] = InvalidOffsetNumber;
            lp_valid[ctx.offnum] = false;
            xmin_commit_status_ok[ctx.offnum] = false;
            ctx.itemid = PageGetItemId(ctx.page, ctx.offnum);
 
            /* Skip over unused/dead line pointers */
            if (!ItemIdIsUsed(ctx.itemid) || ItemIdIsDead(ctx.itemid))
                continue;
 
            /*
             * If this line pointer has been redirected, check that it
             * redirects to a valid offset within the line pointer array
             */
            if (ItemIdIsRedirected(ctx.itemid))
            {
                OffsetNumber rdoffnum = ItemIdGetRedirect(ctx.itemid);
                ItemId      rditem;
 
                if (rdoffnum < FirstOffsetNumber)
                {
                    report_corruption(&ctx,
                                      psprintf("line pointer redirection to item at offset %u precedes minimum offset %u",
                                               (unsigned) rdoffnum,
                                               (unsigned) FirstOffsetNumber));
                    continue;
                }
                if (rdoffnum > maxoff)
                {
                    report_corruption(&ctx,
                                      psprintf("line pointer redirection to item at offset %u exceeds maximum offset %u",
                                               (unsigned) rdoffnum,
                                               (unsigned) maxoff));
                    continue;
                }
 
                /*
                 * Since we've checked that this redirect points to a line
                 * pointer between FirstOffsetNumber and maxoff, it should now
                 * be safe to fetch the referenced line pointer. We expect it
                 * to be LP_NORMAL; if not, that's corruption.
                 */
                rditem = PageGetItemId(ctx.page, rdoffnum);
                if (!ItemIdIsUsed(rditem))
                {
                    report_corruption(&ctx,
                                      psprintf("redirected line pointer points to an unused item at offset %u",
                                               (unsigned) rdoffnum));
                    continue;
                }
                else if (ItemIdIsDead(rditem))
                {
                    report_corruption(&ctx,
                                      psprintf("redirected line pointer points to a dead item at offset %u",
                                               (unsigned) rdoffnum));
                    continue;
                }
                else if (ItemIdIsRedirected(rditem))
                {
                    report_corruption(&ctx,
                                      psprintf("redirected line pointer points to another redirected line pointer at offset %u",
                                               (unsigned) rdoffnum));
                    continue;
                }
 
                /*
                 * Record the fact that this line pointer has passed basic
                 * sanity checking, and also the offset number to which it
                 * points.
                 */
                lp_valid[ctx.offnum] = true;
                successor[ctx.offnum] = rdoffnum;
                continue;
            }
 
            /* Sanity-check the line pointer's offset and length values */
            ctx.lp_len = ItemIdGetLength(ctx.itemid);
            ctx.lp_off = ItemIdGetOffset(ctx.itemid);
 
            if (ctx.lp_off != MAXALIGN(ctx.lp_off))
            {
                report_corruption(&ctx,
                                  psprintf("line pointer to page offset %u is not maximally aligned",
                                           ctx.lp_off));
                continue;
            }
            if (ctx.lp_len < MAXALIGN(SizeofHeapTupleHeader))
            {
                report_corruption(&ctx,
                                  psprintf("line pointer length %u is less than the minimum tuple header size %u",
                                           ctx.lp_len,
                                           (unsigned) MAXALIGN(SizeofHeapTupleHeader)));
                continue;
            }
            if (ctx.lp_off + ctx.lp_len > BLCKSZ)
            {
                report_corruption(&ctx,
                                  psprintf("line pointer to page offset %u with length %u ends beyond maximum page offset %u",
                                           ctx.lp_off,
                                           ctx.lp_len,
                                           (unsigned) BLCKSZ));
                continue;
            }
 
            /* It should be safe to examine the tuple's header, at least */
            lp_valid[ctx.offnum] = true;
            ctx.tuphdr = (HeapTupleHeader) PageGetItem(ctx.page, ctx.itemid);
            ctx.natts = HeapTupleHeaderGetNatts(ctx.tuphdr);
 
            /* Ok, ready to check this next tuple */
            check_tuple(&ctx,
                        &xmin_commit_status_ok[ctx.offnum],
                        &xmin_commit_status[ctx.offnum]);
 
            /*
             * If the CTID field of this tuple seems to point to another tuple
             * on the same page, record that tuple as the successor of this
             * one.
             */
            nextblkno = ItemPointerGetBlockNumber(&(ctx.tuphdr)->t_ctid);
            nextoffnum = ItemPointerGetOffsetNumber(&(ctx.tuphdr)->t_ctid);
            if (nextblkno == ctx.blkno && nextoffnum != ctx.offnum &&
                nextoffnum >= FirstOffsetNumber && nextoffnum <= maxoff)
                successor[ctx.offnum] = nextoffnum;
        }
 
        /*
         * Update chain validation. Check each line pointer that's got a valid
         * successor against that successor.
         */
        ctx.attnum = -1;
        for (ctx.offnum = FirstOffsetNumber; ctx.offnum <= maxoff;
             ctx.offnum = OffsetNumberNext(ctx.offnum))
        {
            ItemId      curr_lp;
            ItemId      next_lp;
            HeapTupleHeader curr_htup;
            HeapTupleHeader next_htup;
            TransactionId curr_xmin;
            TransactionId curr_xmax;
            TransactionId next_xmin;
            OffsetNumber nextoffnum = successor[ctx.offnum];
 
            /*
             * The current line pointer may not have a successor, either
             * because it's not valid or because it didn't point to anything.
             * In either case, we have to give up.
             *
             * If the current line pointer does point to something, it's
             * possible that the target line pointer isn't valid. We have to
             * give up in that case, too.
             */
            if (nextoffnum == InvalidOffsetNumber || !lp_valid[nextoffnum])
                continue;
 
            /* We have two valid line pointers that we can examine. */
            curr_lp = PageGetItemId(ctx.page, ctx.offnum);
            next_lp = PageGetItemId(ctx.page, nextoffnum);
 
            /* Handle the cases where the current line pointer is a redirect. */
            if (ItemIdIsRedirected(curr_lp))
            {
                /*
                 * We should not have set successor[ctx.offnum] to a value
                 * other than InvalidOffsetNumber unless that line pointer is
                 * LP_NORMAL.
                 */
                Assert(ItemIdIsNormal(next_lp));
 
                /* Can only redirect to a HOT tuple. */
                next_htup = (HeapTupleHeader) PageGetItem(ctx.page, next_lp);
                if (!HeapTupleHeaderIsHeapOnly(next_htup))
                {
                    report_corruption(&ctx,
                                      psprintf("redirected line pointer points to a non-heap-only tuple at offset %u",
                                               (unsigned) nextoffnum));
                }
 
                /* HOT chains should not intersect. */
                if (predecessor[nextoffnum] != InvalidOffsetNumber)
                {
                    report_corruption(&ctx,
                                      psprintf("redirect line pointer points to offset %u, but offset %u also points there",
                                               (unsigned) nextoffnum, (unsigned) predecessor[nextoffnum]));
                    continue;
                }
 
                /*
                 * This redirect and the tuple to which it points seem to be
                 * part of an update chain.
                 */
                predecessor[nextoffnum] = ctx.offnum;
                continue;
            }
 
            /*
             * If the next line pointer is a redirect, or if it's a tuple but
             * the XMAX of this tuple doesn't match the XMIN of the next
             * tuple, then the two aren't part of the same update chain and
             * there is nothing more to do.
             */
            if (ItemIdIsRedirected(next_lp))
                continue;
            curr_htup = (HeapTupleHeader) PageGetItem(ctx.page, curr_lp);
            curr_xmax = HeapTupleHeaderGetUpdateXid(curr_htup);
            next_htup = (HeapTupleHeader) PageGetItem(ctx.page, next_lp);
            next_xmin = HeapTupleHeaderGetXmin(next_htup);
            if (!TransactionIdIsValid(curr_xmax) ||
                !TransactionIdEquals(curr_xmax, next_xmin))
                continue;
 
            /* HOT chains should not intersect. */
            if (predecessor[nextoffnum] != InvalidOffsetNumber)
            {
                report_corruption(&ctx,
                                  psprintf("tuple points to new version at offset %u, but offset %u also points there",
                                           (unsigned) nextoffnum, (unsigned) predecessor[nextoffnum]));
                continue;
            }
 
            /*
             * This tuple and the tuple to which it points seem to be part of
             * an update chain.
             */
            predecessor[nextoffnum] = ctx.offnum;
 
            /*
             * If the current tuple is marked as HOT-updated, then the next
             * tuple should be marked as a heap-only tuple. Conversely, if the
             * current tuple isn't marked as HOT-updated, then the next tuple
             * shouldn't be marked as a heap-only tuple.
             *
             * NB: Can't use HeapTupleHeaderIsHotUpdated() as it checks if
             * hint bits indicate xmin/xmax aborted.
             */
            if (!(curr_htup->t_infomask2 & HEAP_HOT_UPDATED) &&
                HeapTupleHeaderIsHeapOnly(next_htup))
            {
                report_corruption(&ctx,
                                  psprintf("non-heap-only update produced a heap-only tuple at offset %u",
                                           (unsigned) nextoffnum));
            }
            if ((curr_htup->t_infomask2 & HEAP_HOT_UPDATED) &&
                !HeapTupleHeaderIsHeapOnly(next_htup))
            {
                report_corruption(&ctx,
                                  psprintf("heap-only update produced a non-heap only tuple at offset %u",
                                           (unsigned) nextoffnum));
            }
 
            /*
             * If the current tuple's xmin is still in progress but the
             * successor tuple's xmin is committed, that's corruption.
             *
             * NB: We recheck the commit status of the current tuple's xmin
             * here, because it might have committed after we checked it and
             * before we checked the commit status of the successor tuple's
             * xmin. This should be safe because the xmin itself can't have
             * changed, only its commit status.
             */
            curr_xmin = HeapTupleHeaderGetXmin(curr_htup);
            if (xmin_commit_status_ok[ctx.offnum] &&
                xmin_commit_status[ctx.offnum] == XID_IN_PROGRESS &&
                xmin_commit_status_ok[nextoffnum] &&
                xmin_commit_status[nextoffnum] == XID_COMMITTED &&
                TransactionIdIsInProgress(curr_xmin))
            {
                report_corruption(&ctx,
                                  psprintf("tuple with in-progress xmin %u was updated to produce a tuple at offset %u with committed xmin %u",
                                           (unsigned) curr_xmin,
                                           (unsigned) ctx.offnum,
                                           (unsigned) next_xmin));
            }
 
            /*
             * If the current tuple's xmin is aborted but the successor
             * tuple's xmin is in-progress or committed, that's corruption.
             */
            if (xmin_commit_status_ok[ctx.offnum] &&
                xmin_commit_status[ctx.offnum] == XID_ABORTED &&
                xmin_commit_status_ok[nextoffnum])
            {
                if (xmin_commit_status[nextoffnum] == XID_IN_PROGRESS)
                    report_corruption(&ctx,
                                      psprintf("tuple with aborted xmin %u was updated to produce a tuple at offset %u with in-progress xmin %u",
                                               (unsigned) curr_xmin,
                                               (unsigned) ctx.offnum,
                                               (unsigned) next_xmin));
                else if (xmin_commit_status[nextoffnum] == XID_COMMITTED)
                    report_corruption(&ctx,
                                      psprintf("tuple with aborted xmin %u was updated to produce a tuple at offset %u with committed xmin %u",
                                               (unsigned) curr_xmin,
                                               (unsigned) ctx.offnum,
                                               (unsigned) next_xmin));
            }
        }
 
        /*
         * An update chain can start either with a non-heap-only tuple or with
         * a redirect line pointer, but not with a heap-only tuple.
         *
         * (This check is in a separate loop because we need the predecessor
         * array to be fully populated before we can perform it.)
         */
        for (ctx.offnum = FirstOffsetNumber;
             ctx.offnum <= maxoff;
             ctx.offnum = OffsetNumberNext(ctx.offnum))
        {
            if (xmin_commit_status_ok[ctx.offnum] &&
                (xmin_commit_status[ctx.offnum] == XID_COMMITTED ||
                 xmin_commit_status[ctx.offnum] == XID_IN_PROGRESS) &&
                predecessor[ctx.offnum] == InvalidOffsetNumber)
            {
                ItemId      curr_lp;
 
                curr_lp = PageGetItemId(ctx.page, ctx.offnum);
                if (!ItemIdIsRedirected(curr_lp))
                {
                    HeapTupleHeader curr_htup;
 
                    curr_htup = (HeapTupleHeader)
                        PageGetItem(ctx.page, curr_lp);
                    if (HeapTupleHeaderIsHeapOnly(curr_htup))
                        report_corruption(&ctx,
                                          psprintf("tuple is root of chain but is marked as heap-only tuple"));
                }
            }
        }
 
        /* clean up */
        UnlockReleaseBuffer(ctx.buffer);
 
        /*
         * Check any toast pointers from the page whose lock we just released
         */
        if (ctx.toasted_attributes != NIL)
        {
            ListCell   *cell;
 
            foreach(cell, ctx.toasted_attributes)
                check_toasted_attribute(&ctx, lfirst(cell));
            list_free_deep(ctx.toasted_attributes);
            ctx.toasted_attributes = NIL;
        }
 
        if (on_error_stop && ctx.is_corrupt)
            break;
    }
 
    read_stream_end(stream);
 
    if (vmbuffer != InvalidBuffer)
        ReleaseBuffer(vmbuffer);
 
    /* Close the associated toast table and indexes, if any. */
    if (ctx.toast_indexes)
        toast_close_indexes(ctx.toast_indexes, ctx.num_toast_indexes,
                            AccessShareLock);
    if (ctx.toast_rel)
        table_close(ctx.toast_rel, AccessShareLock);
 
    /* Close the main relation */
    relation_close(ctx.rel, AccessShareLock);
 
    PG_RETURN_NULL();
}

References AccessShareLock, Assert(), BAS_BULKREAD, block_range_read_stream_cb(), BUFFER_LOCK_SHARE, BufferGetBlockNumber(), BufferGetPage(), BufferIsValid(), CHECK_FOR_INTERRUPTS, check_toasted_attribute(), check_tuple(), DEBUG1, ereport, errcode(), errdetail_relkind_not_supported(), errhint(), errmsg(), ERROR, fb(), FirstOffsetNumber, FullTransactionIdFromXidAndCtx(), GetAccessStrategy(), GetTransactionSnapshot(), HEAP_HOT_UPDATED, heapcheck_read_stream_next_unskippable(), HeapTupleHeaderGetNatts, HeapTupleHeaderGetUpdateXid(), HeapTupleHeaderGetXmin(), HeapTupleHeaderIsHeapOnly(), if(), InitMaterializedSRF(), InvalidBuffer, InvalidOffsetNumber, InvalidTransactionId, ItemIdGetLength, ItemIdGetOffset, ItemIdGetRedirect, ItemIdIsDead, ItemIdIsNormal, ItemIdIsRedirected, ItemIdIsUsed, ItemPointerGetBlockNumber(), ItemPointerGetOffsetNumber(), lfirst, list_free_deep(), LockBuffer(), MAIN_FORKNUM, MAXALIGN, MaxOffsetNumber, NIL, OffsetNumberNext, PageGetItem(), PageGetItemId(), PageGetMaxOffsetNumber(), PG_ARGISNULL, PG_GETARG_BOOL, PG_GETARG_INT64, PG_GETARG_OID, PG_GETARG_TEXT_PP, PG_RETURN_NULL, pg_strcasecmp(), psprintf(), read_stream_begin_relation(), READ_STREAM_DEFAULT, read_stream_end(), READ_STREAM_FULL, read_stream_next_buffer(), READ_STREAM_SEQUENTIAL, READ_STREAM_USE_BATCHING, RecoveryInProgress(), relation_close(), relation_open(), RelationGetNumberOfBlocks, RelationGetRelationName, ReleaseBuffer(), report_corruption(), ReturnSetInfo::setDesc, ReturnSetInfo::setResult, SizeofHeapTupleHeader, skip, SKIP_PAGES_ALL_FROZEN, SKIP_PAGES_ALL_VISIBLE, SKIP_PAGES_NONE, HeapTupleHeaderData::t_infomask2, table_close(), table_open(), text_to_cstring(), toast_close_indexes(), toast_open_indexes(), TransactionIdEquals, TransactionIdIsInProgress(), TransactionIdIsNormal, TransactionIdIsValid, UnlockReleaseBuffer(), update_cached_mxid_range(), update_cached_xid_range(), XID_ABORTED, XID_COMMITTED, XID_IN_PROGRESS, and SnapshotData::xmin.