1
0
mirror of https://github.com/postgres/postgres.git synced 2025-05-18 17:41:14 +03:00
postgres/src/backend/executor/execCurrent.c
Tom Lane 188cd4f440 Remove faulty support for MergeAppend plan with WHERE CURRENT OF.
Somebody extended search_plan_tree() to treat MergeAppend exactly
like Append, which is 100% wrong, because unlike Append we can't
assume that only one input node is actively returning tuples.
Hence a cursor using a MergeAppend across a UNION ALL or inheritance
tree could falsely match a WHERE CURRENT OF query at a row that
isn't actually the cursor's current output row, but coincidentally
has the same TID (in a different table) as the current output row.

Delete the faulty code; this means that such a case will now return
an error like 'cursor "foo" is not a simply updatable scan of table
"bar"', instead of silently misbehaving.  Users should not find that
surprising though, as the same cursor query could have failed that way
already depending on the chosen plan.  (It would fail like that if the
sort were done with an explicit Sort node instead of MergeAppend.)

Expand the clearly-inadequate commentary to be more explicit about
what this code is doing, in hopes of forestalling future mistakes.

It's been like this for awhile, so back-patch to all supported
branches.

Discussion: https://postgr.es/m/482865.1611075182@sss.pgh.pa.us
2021-01-19 13:25:40 -05:00

426 lines
13 KiB
C

/*-------------------------------------------------------------------------
*
* execCurrent.c
* executor support for WHERE CURRENT OF cursor
*
* Portions Copyright (c) 1996-2020, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
* src/backend/executor/execCurrent.c
*
*-------------------------------------------------------------------------
*/
#include "postgres.h"
#include "access/genam.h"
#include "access/relscan.h"
#include "access/sysattr.h"
#include "catalog/pg_type.h"
#include "executor/executor.h"
#include "utils/builtins.h"
#include "utils/lsyscache.h"
#include "utils/portal.h"
#include "utils/rel.h"
static char *fetch_cursor_param_value(ExprContext *econtext, int paramId);
static ScanState *search_plan_tree(PlanState *node, Oid table_oid,
bool *pending_rescan);
/*
* execCurrentOf
*
* Given a CURRENT OF expression and the OID of a table, determine which row
* of the table is currently being scanned by the cursor named by CURRENT OF,
* and return the row's TID into *current_tid.
*
* Returns true if a row was identified. Returns false if the cursor is valid
* for the table but is not currently scanning a row of the table (this is a
* legal situation in inheritance cases). Raises error if cursor is not a
* valid updatable scan of the specified table.
*/
bool
execCurrentOf(CurrentOfExpr *cexpr,
ExprContext *econtext,
Oid table_oid,
ItemPointer current_tid)
{
char *cursor_name;
char *table_name;
Portal portal;
QueryDesc *queryDesc;
/* Get the cursor name --- may have to look up a parameter reference */
if (cexpr->cursor_name)
cursor_name = cexpr->cursor_name;
else
cursor_name = fetch_cursor_param_value(econtext, cexpr->cursor_param);
/* Fetch table name for possible use in error messages */
table_name = get_rel_name(table_oid);
if (table_name == NULL)
elog(ERROR, "cache lookup failed for relation %u", table_oid);
/* Find the cursor's portal */
portal = GetPortalByName(cursor_name);
if (!PortalIsValid(portal))
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_CURSOR),
errmsg("cursor \"%s\" does not exist", cursor_name)));
/*
* We have to watch out for non-SELECT queries as well as held cursors,
* both of which may have null queryDesc.
*/
if (portal->strategy != PORTAL_ONE_SELECT)
ereport(ERROR,
(errcode(ERRCODE_INVALID_CURSOR_STATE),
errmsg("cursor \"%s\" is not a SELECT query",
cursor_name)));
queryDesc = portal->queryDesc;
if (queryDesc == NULL || queryDesc->estate == NULL)
ereport(ERROR,
(errcode(ERRCODE_INVALID_CURSOR_STATE),
errmsg("cursor \"%s\" is held from a previous transaction",
cursor_name)));
/*
* We have two different strategies depending on whether the cursor uses
* FOR UPDATE/SHARE or not. The reason for supporting both is that the
* FOR UPDATE code is able to identify a target table in many cases where
* the other code can't, while the non-FOR-UPDATE case allows use of WHERE
* CURRENT OF with an insensitive cursor.
*/
if (queryDesc->estate->es_rowmarks)
{
ExecRowMark *erm;
Index i;
/*
* Here, the query must have exactly one FOR UPDATE/SHARE reference to
* the target table, and we dig the ctid info out of that.
*/
erm = NULL;
for (i = 0; i < queryDesc->estate->es_range_table_size; i++)
{
ExecRowMark *thiserm = queryDesc->estate->es_rowmarks[i];
if (thiserm == NULL ||
!RowMarkRequiresRowShareLock(thiserm->markType))
continue; /* ignore non-FOR UPDATE/SHARE items */
if (thiserm->relid == table_oid)
{
if (erm)
ereport(ERROR,
(errcode(ERRCODE_INVALID_CURSOR_STATE),
errmsg("cursor \"%s\" has multiple FOR UPDATE/SHARE references to table \"%s\"",
cursor_name, table_name)));
erm = thiserm;
}
}
if (erm == NULL)
ereport(ERROR,
(errcode(ERRCODE_INVALID_CURSOR_STATE),
errmsg("cursor \"%s\" does not have a FOR UPDATE/SHARE reference to table \"%s\"",
cursor_name, table_name)));
/*
* The cursor must have a current result row: per the SQL spec, it's
* an error if not.
*/
if (portal->atStart || portal->atEnd)
ereport(ERROR,
(errcode(ERRCODE_INVALID_CURSOR_STATE),
errmsg("cursor \"%s\" is not positioned on a row",
cursor_name)));
/* Return the currently scanned TID, if there is one */
if (ItemPointerIsValid(&(erm->curCtid)))
{
*current_tid = erm->curCtid;
return true;
}
/*
* This table didn't produce the cursor's current row; some other
* inheritance child of the same parent must have. Signal caller to
* do nothing on this table.
*/
return false;
}
else
{
/*
* Without FOR UPDATE, we dig through the cursor's plan to find the
* scan node. Fail if it's not there or buried underneath
* aggregation.
*/
ScanState *scanstate;
bool pending_rescan = false;
scanstate = search_plan_tree(queryDesc->planstate, table_oid,
&pending_rescan);
if (!scanstate)
ereport(ERROR,
(errcode(ERRCODE_INVALID_CURSOR_STATE),
errmsg("cursor \"%s\" is not a simply updatable scan of table \"%s\"",
cursor_name, table_name)));
/*
* The cursor must have a current result row: per the SQL spec, it's
* an error if not. We test this at the top level, rather than at the
* scan node level, because in inheritance cases any one table scan
* could easily not be on a row. We want to return false, not raise
* error, if the passed-in table OID is for one of the inactive scans.
*/
if (portal->atStart || portal->atEnd)
ereport(ERROR,
(errcode(ERRCODE_INVALID_CURSOR_STATE),
errmsg("cursor \"%s\" is not positioned on a row",
cursor_name)));
/*
* Now OK to return false if we found an inactive scan. It is
* inactive either if it's not positioned on a row, or there's a
* rescan pending for it.
*/
if (TupIsNull(scanstate->ss_ScanTupleSlot) || pending_rescan)
return false;
/*
* Extract TID of the scan's current row. The mechanism for this is
* in principle scan-type-dependent, but for most scan types, we can
* just dig the TID out of the physical scan tuple.
*/
if (IsA(scanstate, IndexOnlyScanState))
{
/*
* For IndexOnlyScan, the tuple stored in ss_ScanTupleSlot may be
* a virtual tuple that does not have the ctid column, so we have
* to get the TID from xs_ctup.t_self.
*/
IndexScanDesc scan = ((IndexOnlyScanState *) scanstate)->ioss_ScanDesc;
*current_tid = scan->xs_heaptid;
}
else
{
/*
* Default case: try to fetch TID from the scan node's current
* tuple. As an extra cross-check, verify tableoid in the current
* tuple. If the scan hasn't provided a physical tuple, we have
* to fail.
*/
Datum ldatum;
bool lisnull;
ItemPointer tuple_tid;
#ifdef USE_ASSERT_CHECKING
ldatum = slot_getsysattr(scanstate->ss_ScanTupleSlot,
TableOidAttributeNumber,
&lisnull);
if (lisnull)
ereport(ERROR,
(errcode(ERRCODE_INVALID_CURSOR_STATE),
errmsg("cursor \"%s\" is not a simply updatable scan of table \"%s\"",
cursor_name, table_name)));
Assert(DatumGetObjectId(ldatum) == table_oid);
#endif
ldatum = slot_getsysattr(scanstate->ss_ScanTupleSlot,
SelfItemPointerAttributeNumber,
&lisnull);
if (lisnull)
ereport(ERROR,
(errcode(ERRCODE_INVALID_CURSOR_STATE),
errmsg("cursor \"%s\" is not a simply updatable scan of table \"%s\"",
cursor_name, table_name)));
tuple_tid = (ItemPointer) DatumGetPointer(ldatum);
*current_tid = *tuple_tid;
}
Assert(ItemPointerIsValid(current_tid));
return true;
}
}
/*
* fetch_cursor_param_value
*
* Fetch the string value of a param, verifying it is of type REFCURSOR.
*/
static char *
fetch_cursor_param_value(ExprContext *econtext, int paramId)
{
ParamListInfo paramInfo = econtext->ecxt_param_list_info;
if (paramInfo &&
paramId > 0 && paramId <= paramInfo->numParams)
{
ParamExternData *prm;
ParamExternData prmdata;
/* give hook a chance in case parameter is dynamic */
if (paramInfo->paramFetch != NULL)
prm = paramInfo->paramFetch(paramInfo, paramId, false, &prmdata);
else
prm = &paramInfo->params[paramId - 1];
if (OidIsValid(prm->ptype) && !prm->isnull)
{
/* safety check in case hook did something unexpected */
if (prm->ptype != REFCURSOROID)
ereport(ERROR,
(errcode(ERRCODE_DATATYPE_MISMATCH),
errmsg("type of parameter %d (%s) does not match that when preparing the plan (%s)",
paramId,
format_type_be(prm->ptype),
format_type_be(REFCURSOROID))));
/* We know that refcursor uses text's I/O routines */
return TextDatumGetCString(prm->value);
}
}
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("no value found for parameter %d", paramId)));
return NULL;
}
/*
* search_plan_tree
*
* Search through a PlanState tree for a scan node on the specified table.
* Return NULL if not found or multiple candidates.
*
* CAUTION: this function is not charged simply with finding some candidate
* scan, but with ensuring that that scan returned the plan tree's current
* output row. That's why we must reject multiple-match cases.
*
* If a candidate is found, set *pending_rescan to true if that candidate
* or any node above it has a pending rescan action, i.e. chgParam != NULL.
* That indicates that we shouldn't consider the node to be positioned on a
* valid tuple, even if its own state would indicate that it is. (Caller
* must initialize *pending_rescan to false, and should not trust its state
* if multiple candidates are found.)
*/
static ScanState *
search_plan_tree(PlanState *node, Oid table_oid,
bool *pending_rescan)
{
ScanState *result = NULL;
if (node == NULL)
return NULL;
switch (nodeTag(node))
{
/*
* Relation scan nodes can all be treated alike: check to see if
* they are scanning the specified table.
*
* ForeignScan and CustomScan might not have a currentRelation, in
* which case we just ignore them. (We dare not descend to any
* child plan nodes they might have, since we do not know the
* relationship of such a node's current output tuple to the
* children's current outputs.)
*/
case T_SeqScanState:
case T_SampleScanState:
case T_IndexScanState:
case T_IndexOnlyScanState:
case T_BitmapHeapScanState:
case T_TidScanState:
case T_ForeignScanState:
case T_CustomScanState:
{
ScanState *sstate = (ScanState *) node;
if (sstate->ss_currentRelation &&
RelationGetRelid(sstate->ss_currentRelation) == table_oid)
result = sstate;
break;
}
/*
* For Append, we can check each input node. It is safe to
* descend to the inputs because only the input that resulted in
* the Append's current output node could be positioned on a tuple
* at all; the other inputs are either at EOF or not yet started.
* Hence, if the desired table is scanned by some
* currently-inactive input node, we will find that node but then
* our caller will realize that it didn't emit the tuple of
* interest.
*
* We do need to watch out for multiple matches (possible if
* Append was from UNION ALL rather than an inheritance tree).
*
* Note: we can NOT descend through MergeAppend similarly, since
* its inputs are likely all active, and we don't know which one
* returned the current output tuple. (Perhaps that could be
* fixed if we were to let this code know more about MergeAppend's
* internal state, but it does not seem worth the trouble. Users
* should not expect plans for ORDER BY queries to be considered
* simply-updatable, since they won't be if the sorting is
* implemented by a Sort node.)
*/
case T_AppendState:
{
AppendState *astate = (AppendState *) node;
int i;
for (i = 0; i < astate->as_nplans; i++)
{
ScanState *elem = search_plan_tree(astate->appendplans[i],
table_oid,
pending_rescan);
if (!elem)
continue;
if (result)
return NULL; /* multiple matches */
result = elem;
}
break;
}
/*
* Result and Limit can be descended through (these are safe
* because they always return their input's current row)
*/
case T_ResultState:
case T_LimitState:
result = search_plan_tree(node->lefttree,
table_oid,
pending_rescan);
break;
/*
* SubqueryScan too, but it keeps the child in a different place
*/
case T_SubqueryScanState:
result = search_plan_tree(((SubqueryScanState *) node)->subplan,
table_oid,
pending_rescan);
break;
default:
/* Otherwise, assume we can't descend through it */
break;
}
/*
* If we found a candidate at or below this node, then this node's
* chgParam indicates a pending rescan that will affect the candidate.
*/
if (result && node->chgParam != NULL)
*pending_rescan = true;
return result;
}