loliObj* loliSym::eval(loliObj* e){
this->type = typeSYM;
if(this->name == "nil" || this == nil){
return nil;
}
if(this->name == "t" || this == t){
return t;
}
if(e->nilp() || e == NULL){
loliObj* r = lookup_top_env(this);
if(r->nilp()){
loli_err("Symbol: " + this->toString() + " is unbound.");
return nil;
}else{
return lcons(r)->head();
}
}else{
loliObj* r = lookup_env(this, e);
if(r->nilp()){
loli_err("Symbol: " + this->toString() + " is unbound in its environment.");
return nil;
}else{
return lcons(r)->head();
}
}
}
static void validate_structure_layout(size_t slots, lref_t layout)
{
if (!CONSP(layout))
vmerror_wrong_type_n(2, layout);
size_t len = (size_t) get_c_long(llength(layout));
if (len != 2)
vmerror_arg_out_of_range(layout, _T("bad structure layout, length<>2"));
lref_t slot_layout = CAR(CDR(layout));
if (get_c_long(llength(slot_layout)) != (long) slots)
vmerror_arg_out_of_range(lcons(slot_layout, fixcons(slots)),
_T("bad structure layout, wrong number of slots"));
for (; CONSP(slot_layout); slot_layout = CDR(slot_layout))
{
if (!CONSP(CAR(slot_layout)))
vmerror_arg_out_of_range(lcons(slot_layout, layout),
_T("bad structure layout, bad slot layout"));
if (!SYMBOLP(CAR(CAR(slot_layout))))
vmerror_arg_out_of_range(layout,
_T("bad structure layout, missing slot name"));
}
}
static void accept_command_line_arguments(int argc, _TCHAR * argv[])
{
lref_t arg_list = NIL;
lref_t arg_list_bud = NIL;
for (int ii = 0; ii < argc; ii++)
{
if (is_vm_argument(argv[ii]))
continue;
lref_t new_cell = lcons(strconsbuf(argv[ii]), NIL);
if (NULLP(arg_list_bud))
{
arg_list = arg_list_bud = new_cell;
}
else
{
SET_CDR(arg_list_bud, new_cell);
arg_list_bud = new_cell;
}
}
interp.startup_args = arg_list;
}
lref_t lenvlookup(lref_t var, lref_t env)
{
lref_t frame;
for (frame = env; CONSP(frame); frame = CDR(frame))
{
lref_t tmp = CAR(frame);
if (!CONSP(tmp))
panic("damaged frame");
lref_t al, fl;
for (fl = CAR(tmp), al = CDR(tmp);
CONSP(fl);
fl = CDR(fl), al = CDR(al))
{
if (!CONSP(al))
vmerror_arg_out_of_range(NIL, _T("too few arguments"));
if (EQ(CAR(fl), var))
return al;
}
if (SYMBOLP(fl) && EQ(fl, var))
return lcons(al, NIL);
}
if (!NULLP(frame))
panic("damaged env");
return NIL;
}
/*
* Finish an incomplete split by inserting/updating the downlinks in parent
* page. 'splitinfo' contains all the child pages involved in the split,
* from left-to-right.
*
* On entry, the caller must hold a lock on stack->buffer and all the child
* pages in 'splitinfo'. If 'unlockbuf' is true, the lock on stack->buffer is
* released on return. The child pages are always unlocked and unpinned.
*/
static void
gistfinishsplit(GISTInsertState *state, GISTInsertStack *stack,
GISTSTATE *giststate, List *splitinfo, bool unlockbuf)
{
ListCell *lc;
List *reversed;
GISTPageSplitInfo *right;
GISTPageSplitInfo *left;
IndexTuple tuples[2];
/* A split always contains at least two halves */
Assert(list_length(splitinfo) >= 2);
/*
* We need to insert downlinks for each new page, and update the downlink
* for the original (leftmost) page in the split. Begin at the rightmost
* page, inserting one downlink at a time until there's only two pages
* left. Finally insert the downlink for the last new page and update the
* downlink for the original page as one operation.
*/
/* for convenience, create a copy of the list in reverse order */
reversed = NIL;
foreach(lc, splitinfo)
{
reversed = lcons(lfirst(lc), reversed);
}
/*
* listCopy--
* this copy function only copies the "lcons-cells" of the list but not
* its contents. (good for list of pointers as well as list of integers).
*/
List *
listCopy(List *list)
{
List *newlist=NIL;
List *l, *nl;
foreach(l, list) {
if (newlist==NIL) {
newlist = nl = lcons(lfirst(l),NIL);
}else {
lnext(nl) = lcons(lfirst(l),NIL);
nl = lnext(nl);
}
}
return newlist;
}
/*
* Async_Notify
*
* This is executed by the SQL notify command.
*
* Adds the relation to the list of pending notifies.
* Actual notification happens during transaction commit.
* ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
*/
void
Async_Notify(const char *relname)
{
if (Trace_notify)
elog(DEBUG1, "Async_Notify(%s)", relname);
/* no point in making duplicate entries in the list ... */
if (!AsyncExistsPendingNotify(relname))
{
/*
* The name list needs to live until end of transaction, so store it
* in the transaction context.
*/
MemoryContext oldcontext;
oldcontext = MemoryContextSwitchTo(CurTransactionContext);
/*
* Ordering of the list isn't important. We choose to put new entries
* on the front, as this might make duplicate-elimination a tad faster
* when the same condition is signaled many times in a row.
*/
pendingNotifies = lcons(pstrdup(relname), pendingNotifies);
MemoryContextSwitchTo(oldcontext);
}
}
static lref_t arg_list_from_buffer(size_t argc, lref_t argv[])
{
lref_t result = NIL;
for (size_t ii = argc; ii > 0; ii--)
result = lcons(argv[ii - 1], result);
return result;
}
/*
* CreateAndPushRestrictionContext creates a new restriction context, inserts it to the
* beginning of the context list, and returns the newly created context.
*/
RelationRestrictionContext *
CreateAndPushRestrictionContext(void)
{
RelationRestrictionContext *restrictionContext =
palloc0(sizeof(RelationRestrictionContext));
relationRestrictionContextList = lcons(restrictionContext,
relationRestrictionContextList);
return restrictionContext;
}
/* ----------------
* CreateExprContext
*
* Create a context for expression evaluation within an EState.
*
* An executor run may require multiple ExprContexts (we usually make one
* for each Plan node, and a separate one for per-output-tuple processing
* such as constraint checking). Each ExprContext has its own "per-tuple"
* memory context.
*
* Note we make no assumption about the caller's memory context.
* ----------------
*/
ExprContext *
CreateExprContext(EState *estate)
{
ExprContext *econtext;
MemoryContext oldcontext;
/* Create the ExprContext node within the per-query memory context */
oldcontext = MemoryContextSwitchTo(estate->es_query_cxt);
econtext = makeNode(ExprContext);
/* Initialize fields of ExprContext */
econtext->ecxt_scantuple = NULL;
econtext->ecxt_innertuple = NULL;
econtext->ecxt_outertuple = NULL;
econtext->ecxt_per_query_memory = estate->es_query_cxt;
/*
* Create working memory for expression evaluation in this context.
*/
econtext->ecxt_per_tuple_memory =
AllocSetContextCreate(estate->es_query_cxt,
"ExprContext",
ALLOCSET_DEFAULT_MINSIZE,
ALLOCSET_DEFAULT_INITSIZE,
ALLOCSET_DEFAULT_MAXSIZE);
econtext->ecxt_param_exec_vals = estate->es_param_exec_vals;
econtext->ecxt_param_list_info = estate->es_param_list_info;
econtext->ecxt_aggvalues = NULL;
econtext->ecxt_aggnulls = NULL;
econtext->caseValue_datum = (Datum) 0;
econtext->caseValue_isNull = true;
econtext->domainValue_datum = (Datum) 0;
econtext->domainValue_isNull = true;
econtext->ecxt_estate = estate;
econtext->ecxt_callbacks = NULL;
/*
* Link the ExprContext into the EState to ensure it is shut down when the
* EState is freed. Because we use lcons(), shutdowns will occur in
* reverse order of creation, which may not be essential but can't hurt.
*/
estate->es_exprcontexts = lcons(econtext, estate->es_exprcontexts);
MemoryContextSwitchTo(oldcontext);
return econtext;
}
lref_t lstress_lisp_heap(lref_t c)
{
if (!FIXNUMP(c))
vmerror_wrong_type_n(1, c);
fixnum_t count = FIXNM(c);
for (fixnum_t i = 0; i < count; i++)
lcons(NIL, NIL);
return NIL;
}
static lref_t extend_env(lref_t actuals, lref_t formals, lref_t env)
{
if (SYMBOLP(formals))
return lcons(lcons(lcons(formals, NIL), lcons(actuals, NIL)), env);
else
return lcons(lcons(formals, actuals), env);
}
loliObj* loliCons::eval(loliObj* env){
this->type = typeCONS;
// std::cout<<this->type->toString()<<std::endl;
if(this->head() == SYM("if")){
loliObj* cond = lcons(this->tail())->head();
if(this->tail()->nilp()){
loli_err("Need at least one expression for if");
return nil;
}
loliObj* wt = lcons(lcons(this->tail())->tail())->head();
loliObj* wf = lcons(lcons(lcons(this->tail())->tail())->tail())->head();
if(cond->eval(env)==boolt){
return wt->eval(top_env);
}else if(cond->eval(env)==boolf){
if(wf){
return wf->eval(top_env);
}else{
return nil;
}
}else{
loli_err("Condition error");
return nil;
}
}
return eval_list(this, env);
}
static void
pushIncompleteInsert(RelFileNode node, XLogRecPtr lsn, ItemPointerData key,
BlockNumber *blkno, int lenblk,
PageSplitRecord *xlinfo /* to extract blkno info */ )
{
MemoryContext oldCxt;
gistIncompleteInsert *ninsert;
if (!ItemPointerIsValid(&key))
/*
* if key is null then we should not store insertion as incomplete,
* because it's a vacuum operation..
*/
return;
oldCxt = MemoryContextSwitchTo(insertCtx);
ninsert = (gistIncompleteInsert *) palloc(sizeof(gistIncompleteInsert));
ninsert->node = node;
ninsert->key = key;
ninsert->lsn = lsn;
if (lenblk && blkno)
{
ninsert->lenblk = lenblk;
ninsert->blkno = (BlockNumber *) palloc(sizeof(BlockNumber) * ninsert->lenblk);
memcpy(ninsert->blkno, blkno, sizeof(BlockNumber) * ninsert->lenblk);
ninsert->origblkno = *blkno;
}
else
{
int i;
Assert(xlinfo);
ninsert->lenblk = xlinfo->data->npage;
ninsert->blkno = (BlockNumber *) palloc(sizeof(BlockNumber) * ninsert->lenblk);
for (i = 0; i < ninsert->lenblk; i++)
ninsert->blkno[i] = xlinfo->page[i].header->blkno;
ninsert->origblkno = xlinfo->data->origblkno;
}
Assert(ninsert->lenblk > 0);
/*
* Stick the new incomplete insert onto the front of the list, not the
* back. This is so that gist_xlog_cleanup will process incompletions in
* last-in-first-out order.
*/
incomplete_inserts = lcons(ninsert, incomplete_inserts);
MemoryContextSwitchTo(oldCxt);
}
/*
* Creates a new gang by logging on a session to each segDB involved.
*
* elog ERROR or return a non-NULL gang.
*/
Gang *
AllocateGang(CdbDispatcherState *ds, GangType type, List *segments)
{
MemoryContext oldContext;
SegmentType segmentType;
Gang *newGang = NULL;
int i;
ELOG_DISPATCHER_DEBUG("AllocateGang begin.");
if (Gp_role != GP_ROLE_DISPATCH)
{
elog(FATAL, "dispatch process called with role %d", Gp_role);
}
if (segments == NIL)
return NULL;
Assert(DispatcherContext);
oldContext = MemoryContextSwitchTo(DispatcherContext);
if (type == GANGTYPE_PRIMARY_WRITER)
segmentType = SEGMENTTYPE_EXPLICT_WRITER;
/* for extended query like cursor, must specify a reader */
else if (ds->isExtendedQuery)
segmentType = SEGMENTTYPE_EXPLICT_READER;
else
segmentType = SEGMENTTYPE_ANY;
newGang = cdbgang_createGang(segments, segmentType);
newGang->allocated = true;
newGang->type = type;
ds->allocatedGangs = lcons(newGang, ds->allocatedGangs);
ds->largestGangSize = Max(ds->largestGangSize, newGang->size);
ELOG_DISPATCHER_DEBUG("AllocateGang end.");
if (type == GANGTYPE_PRIMARY_WRITER)
{
/*
* set "whoami" for utility statement. non-utility statement will
* overwrite it in function getCdbProcessList.
*/
for (i = 0; i < newGang->size; i++)
cdbconn_setQEIdentifier(newGang->db_descriptors[i], -1);
}
MemoryContextSwitchTo(oldContext);
return newGang;
}
loliObj* eval_list(loliObj* lst, loliObj* env){
loliObj* car = lcons(lst)->head()->eval(env);
loliObj* cdr = lcons(lst)->tail();
std::cout<<"HEAD: "<<car->toString()<<"\tTAIL: "<<cdr->toString()<<std::endl;
// std::cout<<cdr->type->toString()<<std::endl;
if(lfunc(car)->rtype){
if(lcons(lcons(lcons(cdr)->tail())->head())->hd == NULL){
std::cout<<lcons(lcons(lcons(cdr)->tail())->head())->toString()<<std::endl;
return c_apply(car, cdr->eval(env), env);
}
return c_apply(car, cdr, env);
}
else{
loli_err(car->toString() + " is not a function!");
}
return nil;
}
/*
* AtSubStart_Notify() --- Take care of subtransaction start.
*
* Push empty state for the new subtransaction.
*/
void
AtSubStart_Notify(void)
{
MemoryContext old_cxt;
/* Keep the list-of-lists in TopTransactionContext for simplicity */
old_cxt = MemoryContextSwitchTo(TopTransactionContext);
upperPendingNotifies = lcons(pendingNotifies, upperPendingNotifies);
Assert(list_length(upperPendingNotifies) ==
GetCurrentTransactionNestLevel() - 1);
pendingNotifies = NIL;
MemoryContextSwitchTo(old_cxt);
}
/* ----------------
* create_expr_ctx
*
* Create a context for expression evaluation within an EState.
*
* An executor run may require multiple ExprContexts (we usually make one
* for each plan_n node, and a separate one for per-output-tuple processing
* such as constraint checking). Each expr_ctx_n has its own "per-tuple"
* memory context.
*
* Note we make no assumption about the caller's memory context.
* ----------------
*/
expr_ctx_n*
create_expr_ctx(exec_state_n *estate)
{
expr_ctx_n *econtext;
struct mctx * oldcontext;
/* Create the expr_ctx_n node within the per-query memory context */
oldcontext = mctx_switch(estate->es_query_cxt);
econtext = MK_N(ExprContext,expr_ctx_n);
/* Initialize fields of expr_ctx_n */
econtext->ecxt_scantuple = NULL;
econtext->ecxt_innertuple = NULL;
econtext->ecxt_outertuple = NULL;
econtext->ecxt_per_query_memory = estate->es_query_cxt;
/*
* Create working memory for expression evaluation in this context.
*/
econtext->ecxt_per_tuple_memory = aset_create_normal(estate->es_query_cxt, "ExprContext");
econtext->ecxt_param_exec_vals = estate->es_param_exec_vals;
econtext->ecxt_param_list_info = estate->es_param_list_info;
econtext->ecxt_aggvalues = NULL;
econtext->ecxt_aggnulls = NULL;
econtext->caseValue_datum = (datum_t) 0;
econtext->caseValue_isNull = true;
econtext->domainValue_datum = (datum_t) 0;
econtext->domainValue_isNull = true;
econtext->ecxt_estate = estate;
econtext->ecxt_callbacks = NULL;
/*
* Link the expr_ctx_n into the exec_state_n to ensure it is shut down when the
* exec_state_n is freed. Because we use lcons(), shutdowns will occur in
* reverse order of creation, which may not be essential but can't hurt.
*/
estate->es_exprcontexts = lcons(econtext, estate->es_exprcontexts);
mctx_switch(oldcontext);
return econtext;
}
/*
* subxact.__enter__() or subxact.enter()
*
* Start an explicit subtransaction. SPI calls within an explicit
* subtransaction will not start another one, so you can atomically
* execute many SPI calls and still get a controllable exception if
* one of them fails.
*/
static PyObject *
PLy_subtransaction_enter(PyObject *self, PyObject *unused)
{
PLySubtransactionData *subxactdata;
MemoryContext oldcontext;
PLySubtransactionObject *subxact = (PLySubtransactionObject *) self;
if (subxact->started)
{
PLy_exception_set(PyExc_ValueError, "this subtransaction has already been entered");
return NULL;
}
if (subxact->exited)
{
PLy_exception_set(PyExc_ValueError, "this subtransaction has already been exited");
return NULL;
}
subxact->started = true;
oldcontext = CurrentMemoryContext;
subxactdata = (PLySubtransactionData *)
MemoryContextAlloc(TopTransactionContext,
sizeof(PLySubtransactionData));
subxactdata->oldcontext = oldcontext;
subxactdata->oldowner = CurrentResourceOwner;
BeginInternalSubTransaction(NULL);
/* Be sure that cells of explicit_subtransactions list are long-lived */
MemoryContextSwitchTo(TopTransactionContext);
explicit_subtransactions = lcons(subxactdata, explicit_subtransactions);
/* Caller wants to stay in original memory context */
MemoryContextSwitchTo(oldcontext);
Py_INCREF(self);
return self;
}
/*
* infer_tupledesc
*
* Given a stream, infer a TupleDesc based on the supertype of all
* the casted types for each of the stream's columns
*/
static void
infer_tupledesc(StreamTargetsEntry *stream)
{
HASH_SEQ_STATUS status;
StreamColumnsEntry *entry;
List *names = NIL;
List *types = NIL;
List *mods = NIL;
List *collations = NIL;
Const *preferred = makeConst(NUMERIC_OID, -1, 0, -1, 0, false, false);
hash_seq_init(&status, stream->colstotypes);
while ((entry = (StreamColumnsEntry *) hash_seq_search(&status)) != NULL)
{
char err[128];
Oid supertype;
char category;
bool typispreferred;
Oid t = exprType(linitial(entry->types));
/*
* If there are any numeric types in our target types, we prepend a float8
* to the list of types to select from, as that is our preferred type when
* there is any ambiguity about how to interpret numeric types.
*/
get_type_category_preferred(t, &category, &typispreferred);
if (category == TYPCATEGORY_NUMERIC)
entry->types = lcons(preferred, entry->types);
sprintf(err, "type conflict with stream \"%s\":", get_rel_name(stream->relid));
supertype = select_common_type(NULL, entry->types, err, NULL);
names = lappend(names, makeString(entry->name));
types = lappend_int(types, supertype);
mods = lappend_int(mods, -1);
collations = lappend_int(collations, 0);
}
stream->desc = BuildDescFromLists(names, types, mods, collations);
}
static void fast_read_list(lref_t reader, bool read_listd, lref_t * list)
{
*list = NIL;
lref_t list_bud = NIL;
lref_t next_list_cell = NIL;
lref_t list_length;
fast_read(reader, &list_length, false);
if (!FIXNUMP(list_length))
vmerror_fast_read("expected fixnum for list length", reader, list_length);
*list = NIL;
for (fixnum_t ii = 0; ii < FIXNM(list_length); ii++)
{
next_list_cell = lcons(NIL, NIL);
if (NULLP(*list))
*list = next_list_cell;
else
SET_CDR(list_bud, next_list_cell);
list_bud = next_list_cell;
fast_read(reader, &(next_list_cell->as.cons.car), false);
if (EOFP(CAR(next_list_cell)))
vmerror_fast_read("incomplete list definition", reader, NIL);
}
if (read_listd)
{
fast_read(reader, &(list_bud->as.cons.cdr), false);
if (EOFP(CDR(list_bud)))
vmerror_fast_read("incomplete list defintion, missing cdr", reader, NIL);
}
}
/*
*--------------------------------------------------------------
* Async_Notify
*
* This is executed by the SQL notify command.
*
* Adds the relation to the list of pending notifies.
* Actual notification happens during transaction commit.
* ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
*
*--------------------------------------------------------------
*/
void
Async_Notify(const char *relname)
{
if (Trace_notify)
elog(DEBUG1, "Async_Notify(%s)", relname);
/* no point in making duplicate entries in the list ... */
if (!AsyncExistsPendingNotify(relname))
{
/*
* The name list needs to live until end of transaction, so store it
* in the transaction context.
*/
MemoryContext oldcontext;
oldcontext = MemoryContextSwitchTo(CurTransactionContext);
pendingNotifies = lcons(pstrdup(relname), pendingNotifies);
MemoryContextSwitchTo(oldcontext);
}
}
/*
* Returns a List* of WorkloadElements.
*/
List *
ExtractWorkload(void){
//for the beginning just get all the childNodes of the ModelGraph, simulating a single reference of every Node
//TODO: extract the relevant nodes from the workload, a separate structure will be needed: WorkloadElement{Node, relativeCount, Horizon}
List *init = GetAllChildren(NULL, true),
*res = NIL;
ListCell *lc;
Horizon *h;
foreach(lc, init)
{
WorkloadElement *w = palloc0(sizeof(WorkloadElement));
w->mgin = (ModelGraphIndexNode *)lfirst(lc);
h = palloc0(sizeof(Horizon));
h->error = 0.0;
h->horizon = 1;
h->frequency = 1.0/list_length(init);
h->stillCheck = true;
w->horizons = lcons(h, w->horizons);
// res = lcons(w, res);
res = lappend(res, w);
}
/*
* get_relation_info -
* Retrieves catalog information for a given relation.
*
* Given the Oid of the relation, return the following info into fields
* of the RelOptInfo struct:
*
* min_attr lowest valid AttrNumber
* max_attr highest valid AttrNumber
* indexlist list of IndexOptInfos for relation's indexes
* pages number of pages
* tuples number of tuples
*
* Also, initialize the attr_needed[] and attr_widths[] arrays. In most
* cases these are left as zeroes, but sometimes we need to compute attr
* widths here, and we may as well cache the results for costsize.c.
*
* If inhparent is true, all we need to do is set up the attr arrays:
* the RelOptInfo actually represents the appendrel formed by an inheritance
* tree, and so the parent rel's physical size and index information isn't
* important for it.
*/
void
get_relation_info(PlannerInfo *root, Oid relationObjectId, bool inhparent,
RelOptInfo *rel)
{
Index varno = rel->relid;
Relation relation;
bool hasindex;
List *indexinfos = NIL;
/*
* We need not lock the relation since it was already locked, either by
* the rewriter or when expand_inherited_rtentry() added it to the query's
* rangetable.
*/
relation = heap_open(relationObjectId, NoLock);
rel->min_attr = FirstLowInvalidHeapAttributeNumber + 1;
rel->max_attr = RelationGetNumberOfAttributes(relation);
rel->reltablespace = RelationGetForm(relation)->reltablespace;
Assert(rel->max_attr >= rel->min_attr);
rel->attr_needed = (Relids *)
palloc0((rel->max_attr - rel->min_attr + 1) * sizeof(Relids));
rel->attr_widths = (int32 *)
palloc0((rel->max_attr - rel->min_attr + 1) * sizeof(int32));
/*
* Estimate relation size --- unless it's an inheritance parent, in which
* case the size will be computed later in set_append_rel_pathlist, and we
* must leave it zero for now to avoid bollixing the total_table_pages
* calculation.
*/
if (!inhparent)
estimate_rel_size(relation, rel->attr_widths - rel->min_attr,
&rel->pages, &rel->tuples);
/*
* Make list of indexes. Ignore indexes on system catalogs if told to.
* Don't bother with indexes for an inheritance parent, either.
*/
if (inhparent ||
(IgnoreSystemIndexes && IsSystemClass(relation->rd_rel)))
hasindex = false;
else
hasindex = relation->rd_rel->relhasindex;
if (hasindex)
{
List *indexoidlist;
ListCell *l;
LOCKMODE lmode;
indexoidlist = RelationGetIndexList(relation);
/*
* For each index, we get the same type of lock that the executor will
* need, and do not release it. This saves a couple of trips to the
* shared lock manager while not creating any real loss of
* concurrency, because no schema changes could be happening on the
* index while we hold lock on the parent rel, and neither lock type
* blocks any other kind of index operation.
*/
if (rel->relid == root->parse->resultRelation)
lmode = RowExclusiveLock;
else
lmode = AccessShareLock;
foreach(l, indexoidlist)
{
Oid indexoid = lfirst_oid(l);
Relation indexRelation;
Form_pg_index index;
IndexOptInfo *info;
int ncolumns;
int i;
/*
* Extract info from the relation descriptor for the index.
*/
indexRelation = index_open(indexoid, lmode);
index = indexRelation->rd_index;
/*
* Ignore invalid indexes, since they can't safely be used for
* queries. Note that this is OK because the data structure we
* are constructing is only used by the planner --- the executor
* still needs to insert into "invalid" indexes!
*/
if (!index->indisvalid)
{
index_close(indexRelation, NoLock);
continue;
}
/*
* If the index is valid, but cannot yet be used, ignore it; but
* mark the plan we are generating as transient. See
* src/backend/access/heap/README.HOT for discussion.
*/
if (index->indcheckxmin &&
!TransactionIdPrecedes(HeapTupleHeaderGetXmin(indexRelation->rd_indextuple->t_data),
TransactionXmin))
{
root->glob->transientPlan = true;
index_close(indexRelation, NoLock);
continue;
}
info = makeNode(IndexOptInfo);
info->indexoid = index->indexrelid;
info->reltablespace =
RelationGetForm(indexRelation)->reltablespace;
info->rel = rel;
info->ncolumns = ncolumns = index->indnatts;
/*
* Allocate per-column info arrays. To save a few palloc cycles
* we allocate all the Oid-type arrays in one request. Note that
* the opfamily array needs an extra, terminating zero at the end.
* We pre-zero the ordering info in case the index is unordered.
*/
info->indexkeys = (int *) palloc(sizeof(int) * ncolumns);
info->opfamily = (Oid *) palloc0(sizeof(Oid) * (4 * ncolumns + 1));
info->opcintype = info->opfamily + (ncolumns + 1);
info->fwdsortop = info->opcintype + ncolumns;
info->revsortop = info->fwdsortop + ncolumns;
info->nulls_first = (bool *) palloc0(sizeof(bool) * ncolumns);
for (i = 0; i < ncolumns; i++)
{
info->indexkeys[i] = index->indkey.values[i];
info->opfamily[i] = indexRelation->rd_opfamily[i];
info->opcintype[i] = indexRelation->rd_opcintype[i];
}
info->relam = indexRelation->rd_rel->relam;
info->amcostestimate = indexRelation->rd_am->amcostestimate;
info->amoptionalkey = indexRelation->rd_am->amoptionalkey;
info->amsearchnulls = indexRelation->rd_am->amsearchnulls;
info->amhasgettuple = OidIsValid(indexRelation->rd_am->amgettuple);
info->amhasgetbitmap = OidIsValid(indexRelation->rd_am->amgetbitmap);
/*
* Fetch the ordering operators associated with the index, if any.
* We expect that all ordering-capable indexes use btree's
* strategy numbers for the ordering operators.
*/
if (indexRelation->rd_am->amcanorder)
{
int nstrat = indexRelation->rd_am->amstrategies;
for (i = 0; i < ncolumns; i++)
{
int16 opt = indexRelation->rd_indoption[i];
int fwdstrat;
int revstrat;
if (opt & INDOPTION_DESC)
{
fwdstrat = BTGreaterStrategyNumber;
revstrat = BTLessStrategyNumber;
}
else
{
fwdstrat = BTLessStrategyNumber;
revstrat = BTGreaterStrategyNumber;
}
/*
* Index AM must have a fixed set of strategies for it to
* make sense to specify amcanorder, so we need not allow
* the case amstrategies == 0.
*/
if (fwdstrat > 0)
{
Assert(fwdstrat <= nstrat);
info->fwdsortop[i] = indexRelation->rd_operator[i * nstrat + fwdstrat - 1];
}
if (revstrat > 0)
{
Assert(revstrat <= nstrat);
info->revsortop[i] = indexRelation->rd_operator[i * nstrat + revstrat - 1];
}
info->nulls_first[i] = (opt & INDOPTION_NULLS_FIRST) != 0;
}
}
/*
* Fetch the index expressions and predicate, if any. We must
* modify the copies we obtain from the relcache to have the
* correct varno for the parent relation, so that they match up
* correctly against qual clauses.
*/
info->indexprs = RelationGetIndexExpressions(indexRelation);
info->indpred = RelationGetIndexPredicate(indexRelation);
if (info->indexprs && varno != 1)
ChangeVarNodes((Node *) info->indexprs, 1, varno, 0);
if (info->indpred && varno != 1)
ChangeVarNodes((Node *) info->indpred, 1, varno, 0);
info->predOK = false; /* set later in indxpath.c */
info->unique = index->indisunique;
/*
* Estimate the index size. If it's not a partial index, we lock
* the number-of-tuples estimate to equal the parent table; if it
* is partial then we have to use the same methods as we would for
* a table, except we can be sure that the index is not larger
* than the table.
*/
if (info->indpred == NIL)
{
info->pages = RelationGetNumberOfBlocks(indexRelation);
info->tuples = rel->tuples;
}
else
{
estimate_rel_size(indexRelation, NULL,
&info->pages, &info->tuples);
if (info->tuples > rel->tuples)
info->tuples = rel->tuples;
}
index_close(indexRelation, NoLock);
indexinfos = lcons(info, indexinfos);
}
list_free(indexoidlist);
}
/*
* get_relation_info -
* Retrieves catalog information for a given relation.
*
* Given the Oid of the relation, return the following info into fields
* of the RelOptInfo struct:
*
* min_attr lowest valid AttrNumber
* max_attr highest valid AttrNumber
* indexlist list of IndexOptInfos for relation's indexes
* pages number of pages
* tuples number of tuples
*
* Also, initialize the attr_needed[] and attr_widths[] arrays. In most
* cases these are left as zeroes, but sometimes we need to compute attr
* widths here, and we may as well cache the results for costsize.c.
*
* If inhparent is true, all we need to do is set up the attr arrays:
* the RelOptInfo actually represents the appendrel formed by an inheritance
* tree, and so the parent rel's physical size and index information isn't
* important for it.
*/
void
get_relation_info(PlannerInfo *root, Oid relationObjectId, bool inhparent,
RelOptInfo *rel)
{
Index varno = rel->relid;
Relation relation;
bool hasindex;
List *indexinfos = NIL;
/*
* We need not lock the relation since it was already locked, either by
* the rewriter or when expand_inherited_rtentry() added it to the query's
* rangetable.
*/
relation = heap_open(relationObjectId, NoLock);
/* Temporary and unlogged relations are inaccessible during recovery. */
if (!RelationNeedsWAL(relation) && RecoveryInProgress())
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("cannot access temporary or unlogged relations during recovery")));
rel->min_attr = FirstLowInvalidHeapAttributeNumber + 1;
rel->max_attr = RelationGetNumberOfAttributes(relation);
rel->reltablespace = RelationGetForm(relation)->reltablespace;
Assert(rel->max_attr >= rel->min_attr);
rel->attr_needed = (Relids *)
palloc0((rel->max_attr - rel->min_attr + 1) * sizeof(Relids));
rel->attr_widths = (int32 *)
palloc0((rel->max_attr - rel->min_attr + 1) * sizeof(int32));
/*
* Estimate relation size --- unless it's an inheritance parent, in which
* case the size will be computed later in set_append_rel_pathlist, and we
* must leave it zero for now to avoid bollixing the total_table_pages
* calculation.
*/
if (!inhparent)
estimate_rel_size(relation, rel->attr_widths - rel->min_attr,
&rel->pages, &rel->tuples, &rel->allvisfrac);
/*
* Make list of indexes. Ignore indexes on system catalogs if told to.
* Don't bother with indexes for an inheritance parent, either.
*/
if (inhparent ||
(IgnoreSystemIndexes && IsSystemClass(relation->rd_rel)))
hasindex = false;
else
hasindex = relation->rd_rel->relhasindex;
if (hasindex)
{
List *indexoidlist;
ListCell *l;
LOCKMODE lmode;
indexoidlist = RelationGetIndexList(relation);
/*
* For each index, we get the same type of lock that the executor will
* need, and do not release it. This saves a couple of trips to the
* shared lock manager while not creating any real loss of
* concurrency, because no schema changes could be happening on the
* index while we hold lock on the parent rel, and neither lock type
* blocks any other kind of index operation.
*/
if (rel->relid == root->parse->resultRelation)
lmode = RowExclusiveLock;
else
lmode = AccessShareLock;
foreach(l, indexoidlist)
{
Oid indexoid = lfirst_oid(l);
Relation indexRelation;
Form_pg_index index;
IndexOptInfo *info;
int ncolumns;
int i;
/*
* Extract info from the relation descriptor for the index.
*/
indexRelation = index_open(indexoid, lmode);
index = indexRelation->rd_index;
/*
* Ignore invalid indexes, since they can't safely be used for
* queries. Note that this is OK because the data structure we
* are constructing is only used by the planner --- the executor
* still needs to insert into "invalid" indexes!
*/
if (!index->indisvalid)
{
index_close(indexRelation, NoLock);
continue;
}
/*
* If the index is valid, but cannot yet be used, ignore it; but
* mark the plan we are generating as transient. See
* src/backend/access/heap/README.HOT for discussion.
*/
if (index->indcheckxmin &&
!TransactionIdPrecedes(HeapTupleHeaderGetXmin(indexRelation->rd_indextuple->t_data),
TransactionXmin))
{
root->glob->transientPlan = true;
index_close(indexRelation, NoLock);
continue;
}
info = makeNode(IndexOptInfo);
info->indexoid = index->indexrelid;
info->reltablespace =
RelationGetForm(indexRelation)->reltablespace;
info->rel = rel;
info->ncolumns = ncolumns = index->indnatts;
info->indexkeys = (int *) palloc(sizeof(int) * ncolumns);
info->indexcollations = (Oid *) palloc(sizeof(Oid) * ncolumns);
info->opfamily = (Oid *) palloc(sizeof(Oid) * ncolumns);
info->opcintype = (Oid *) palloc(sizeof(Oid) * ncolumns);
for (i = 0; i < ncolumns; i++)
{
info->indexkeys[i] = index->indkey.values[i];
info->indexcollations[i] = indexRelation->rd_indcollation[i];
info->opfamily[i] = indexRelation->rd_opfamily[i];
info->opcintype[i] = indexRelation->rd_opcintype[i];
}
info->relam = indexRelation->rd_rel->relam;
info->amcostestimate = indexRelation->rd_am->amcostestimate;
info->canreturn = index_can_return(indexRelation);
info->amcanorderbyop = indexRelation->rd_am->amcanorderbyop;
info->amoptionalkey = indexRelation->rd_am->amoptionalkey;
info->amsearcharray = indexRelation->rd_am->amsearcharray;
info->amsearchnulls = indexRelation->rd_am->amsearchnulls;
info->amhasgettuple = OidIsValid(indexRelation->rd_am->amgettuple);
info->amhasgetbitmap = OidIsValid(indexRelation->rd_am->amgetbitmap);
/*
* Fetch the ordering information for the index, if any.
*/
if (info->relam == BTREE_AM_OID)
{
/*
* If it's a btree index, we can use its opfamily OIDs
* directly as the sort ordering opfamily OIDs.
*/
Assert(indexRelation->rd_am->amcanorder);
info->sortopfamily = info->opfamily;
info->reverse_sort = (bool *) palloc(sizeof(bool) * ncolumns);
info->nulls_first = (bool *) palloc(sizeof(bool) * ncolumns);
for (i = 0; i < ncolumns; i++)
{
int16 opt = indexRelation->rd_indoption[i];
info->reverse_sort[i] = (opt & INDOPTION_DESC) != 0;
info->nulls_first[i] = (opt & INDOPTION_NULLS_FIRST) != 0;
}
}
else if (indexRelation->rd_am->amcanorder)
{
/*
* Otherwise, identify the corresponding btree opfamilies by
* trying to map this index's "<" operators into btree. Since
* "<" uniquely defines the behavior of a sort order, this is
* a sufficient test.
*
* XXX This method is rather slow and also requires the
* undesirable assumption that the other index AM numbers its
* strategies the same as btree. It'd be better to have a way
* to explicitly declare the corresponding btree opfamily for
* each opfamily of the other index type. But given the lack
* of current or foreseeable amcanorder index types, it's not
* worth expending more effort on now.
*/
info->sortopfamily = (Oid *) palloc(sizeof(Oid) * ncolumns);
info->reverse_sort = (bool *) palloc(sizeof(bool) * ncolumns);
info->nulls_first = (bool *) palloc(sizeof(bool) * ncolumns);
for (i = 0; i < ncolumns; i++)
{
int16 opt = indexRelation->rd_indoption[i];
Oid ltopr;
Oid btopfamily;
Oid btopcintype;
int16 btstrategy;
info->reverse_sort[i] = (opt & INDOPTION_DESC) != 0;
info->nulls_first[i] = (opt & INDOPTION_NULLS_FIRST) != 0;
ltopr = get_opfamily_member(info->opfamily[i],
info->opcintype[i],
info->opcintype[i],
BTLessStrategyNumber);
if (OidIsValid(ltopr) &&
get_ordering_op_properties(ltopr,
&btopfamily,
&btopcintype,
&btstrategy) &&
btopcintype == info->opcintype[i] &&
btstrategy == BTLessStrategyNumber)
{
/* Successful mapping */
info->sortopfamily[i] = btopfamily;
}
else
{
/* Fail ... quietly treat index as unordered */
info->sortopfamily = NULL;
info->reverse_sort = NULL;
info->nulls_first = NULL;
break;
}
}
}
else
{
info->sortopfamily = NULL;
info->reverse_sort = NULL;
info->nulls_first = NULL;
}
/*
* Fetch the index expressions and predicate, if any. We must
* modify the copies we obtain from the relcache to have the
* correct varno for the parent relation, so that they match up
* correctly against qual clauses.
*/
info->indexprs = RelationGetIndexExpressions(indexRelation);
info->indpred = RelationGetIndexPredicate(indexRelation);
if (info->indexprs && varno != 1)
ChangeVarNodes((Node *) info->indexprs, 1, varno, 0);
if (info->indpred && varno != 1)
ChangeVarNodes((Node *) info->indpred, 1, varno, 0);
/* Build targetlist using the completed indexprs data */
info->indextlist = build_index_tlist(root, info, relation);
info->predOK = false; /* set later in indxpath.c */
info->unique = index->indisunique;
info->immediate = index->indimmediate;
info->hypothetical = false;
/*
* Estimate the index size. If it's not a partial index, we lock
* the number-of-tuples estimate to equal the parent table; if it
* is partial then we have to use the same methods as we would for
* a table, except we can be sure that the index is not larger
* than the table.
*/
if (info->indpred == NIL)
{
info->pages = RelationGetNumberOfBlocks(indexRelation);
info->tuples = rel->tuples;
}
else
{
double allvisfrac; /* dummy */
estimate_rel_size(indexRelation, NULL,
&info->pages, &info->tuples, &allvisfrac);
if (info->tuples > rel->tuples)
info->tuples = rel->tuples;
}
index_close(indexRelation, NoLock);
indexinfos = lcons(info, indexinfos);
}
list_free(indexoidlist);
}
/*
* Get any row security quals and check quals that should be applied to the
* specified RTE.
*
* In addition, hasRowSecurity is set to true if row level security is enabled
* (even if this RTE doesn't have any row security quals), and hasSubLinks is
* set to true if any of the quals returned contain sublinks.
*/
void
get_row_security_policies(Query *root, CmdType commandType, RangeTblEntry *rte,
int rt_index, List **securityQuals,
List **withCheckOptions, bool *hasRowSecurity,
bool *hasSubLinks)
{
Expr *rowsec_expr = NULL;
Expr *rowsec_with_check_expr = NULL;
Expr *hook_expr_restrictive = NULL;
Expr *hook_with_check_expr_restrictive = NULL;
Expr *hook_expr_permissive = NULL;
Expr *hook_with_check_expr_permissive = NULL;
List *rowsec_policies;
List *hook_policies_restrictive = NIL;
List *hook_policies_permissive = NIL;
Relation rel;
Oid user_id;
int rls_status;
bool defaultDeny = false;
/* Defaults for the return values */
*securityQuals = NIL;
*withCheckOptions = NIL;
*hasRowSecurity = false;
*hasSubLinks = false;
/* If this is not a normal relation, just return immediately */
if (rte->relkind != RELKIND_RELATION)
return;
/* Switch to checkAsUser if it's set */
user_id = rte->checkAsUser ? rte->checkAsUser : GetUserId();
/* Determine the state of RLS for this, pass checkAsUser explicitly */
rls_status = check_enable_rls(rte->relid, rte->checkAsUser, false);
/* If there is no RLS on this table at all, nothing to do */
if (rls_status == RLS_NONE)
return;
/*
* RLS_NONE_ENV means we are not doing any RLS now, but that may change
* with changes to the environment, so we mark it as hasRowSecurity to
* force a re-plan when the environment changes.
*/
if (rls_status == RLS_NONE_ENV)
{
/*
* Indicate that this query may involve RLS and must therefore be
* replanned if the environment changes (GUCs, role), but we are not
* adding anything here.
*/
*hasRowSecurity = true;
return;
}
/* Grab the built-in policies which should be applied to this relation. */
rel = heap_open(rte->relid, NoLock);
rowsec_policies = pull_row_security_policies(commandType, rel,
user_id);
/*
* Check if this is only the default-deny policy.
*
* Normally, if the table has row security enabled but there are no
* policies, we use a default-deny policy and not allow anything. However,
* when an extension uses the hook to add their own policies, we don't
* want to include the default deny policy or there won't be any way for a
* user to use an extension exclusively for the policies to be used.
*/
if (((RowSecurityPolicy *) linitial(rowsec_policies))->policy_id
== InvalidOid)
defaultDeny = true;
/* Now that we have our policies, build the expressions from them. */
process_policies(root, rowsec_policies, rt_index, &rowsec_expr,
&rowsec_with_check_expr, hasSubLinks, OR_EXPR);
/*
* Also, allow extensions to add their own policies.
*
* extensions can add either permissive or restrictive policies.
*
* Note that, as with the internal policies, if multiple policies are
* returned then they will be combined into a single expression with all
* of them OR'd (for permissive) or AND'd (for restrictive) together.
*
* If only a USING policy is returned by the extension then it will be
* used for WITH CHECK as well, similar to how internal policies are
* handled.
*
* The only caveat to this is that if there are NO internal policies
* defined, there ARE policies returned by the extension, and RLS is
* enabled on the table, then we will ignore the internally-generated
* default-deny policy and use only the policies returned by the
* extension.
*/
if (row_security_policy_hook_restrictive)
{
hook_policies_restrictive = (*row_security_policy_hook_restrictive) (commandType, rel);
/* Build the expression from any policies returned. */
if (hook_policies_restrictive != NIL)
process_policies(root, hook_policies_restrictive, rt_index,
&hook_expr_restrictive,
&hook_with_check_expr_restrictive,
hasSubLinks,
AND_EXPR);
}
if (row_security_policy_hook_permissive)
{
hook_policies_permissive = (*row_security_policy_hook_permissive) (commandType, rel);
/* Build the expression from any policies returned. */
if (hook_policies_permissive != NIL)
process_policies(root, hook_policies_permissive, rt_index,
&hook_expr_permissive,
&hook_with_check_expr_permissive, hasSubLinks,
OR_EXPR);
}
/*
* If the only built-in policy is the default-deny one, and hook policies
* exist, then use the hook policies only and do not apply the
* default-deny policy. Otherwise, we will apply both sets below.
*/
if (defaultDeny &&
(hook_policies_restrictive != NIL || hook_policies_permissive != NIL))
{
rowsec_expr = NULL;
rowsec_with_check_expr = NULL;
}
/*
* For INSERT or UPDATE, we need to add the WITH CHECK quals to Query's
* withCheckOptions to verify that any new records pass the WITH CHECK
* policy (this will be a copy of the USING policy, if no explicit WITH
* CHECK policy exists).
*/
if (commandType == CMD_INSERT || commandType == CMD_UPDATE)
{
/*
* WITH CHECK OPTIONS wants a WCO node which wraps each Expr, so
* create them as necessary.
*/
/*
* Handle any restrictive policies first.
*
* They can simply be added.
*/
if (hook_with_check_expr_restrictive)
{
WithCheckOption *wco;
wco = (WithCheckOption *) makeNode(WithCheckOption);
wco->kind = commandType == CMD_INSERT ? WCO_RLS_INSERT_CHECK :
WCO_RLS_UPDATE_CHECK;
wco->relname = pstrdup(RelationGetRelationName(rel));
wco->qual = (Node *) hook_with_check_expr_restrictive;
wco->cascaded = false;
*withCheckOptions = lappend(*withCheckOptions, wco);
}
/*
* Handle built-in policies, if there are no permissive policies from
* the hook.
*/
if (rowsec_with_check_expr && !hook_with_check_expr_permissive)
{
WithCheckOption *wco;
wco = (WithCheckOption *) makeNode(WithCheckOption);
wco->kind = commandType == CMD_INSERT ? WCO_RLS_INSERT_CHECK :
WCO_RLS_UPDATE_CHECK;
wco->relname = pstrdup(RelationGetRelationName(rel));
wco->qual = (Node *) rowsec_with_check_expr;
wco->cascaded = false;
*withCheckOptions = lappend(*withCheckOptions, wco);
}
/* Handle the hook policies, if there are no built-in ones. */
else if (!rowsec_with_check_expr && hook_with_check_expr_permissive)
{
WithCheckOption *wco;
wco = (WithCheckOption *) makeNode(WithCheckOption);
wco->kind = commandType == CMD_INSERT ? WCO_RLS_INSERT_CHECK :
WCO_RLS_UPDATE_CHECK;
wco->relname = pstrdup(RelationGetRelationName(rel));
wco->qual = (Node *) hook_with_check_expr_permissive;
wco->cascaded = false;
*withCheckOptions = lappend(*withCheckOptions, wco);
}
/* Handle the case where there are both. */
else if (rowsec_with_check_expr && hook_with_check_expr_permissive)
{
WithCheckOption *wco;
List *combined_quals = NIL;
Expr *combined_qual_eval;
combined_quals = lcons(copyObject(rowsec_with_check_expr),
combined_quals);
combined_quals = lcons(copyObject(hook_with_check_expr_permissive),
combined_quals);
combined_qual_eval = makeBoolExpr(OR_EXPR, combined_quals, -1);
wco = (WithCheckOption *) makeNode(WithCheckOption);
wco->kind = commandType == CMD_INSERT ? WCO_RLS_INSERT_CHECK :
WCO_RLS_UPDATE_CHECK;
wco->relname = pstrdup(RelationGetRelationName(rel));
wco->qual = (Node *) combined_qual_eval;
wco->cascaded = false;
*withCheckOptions = lappend(*withCheckOptions, wco);
}
/*
* ON CONFLICT DO UPDATE has an RTE that is subject to both INSERT and
* UPDATE RLS enforcement. Those are enforced (as a special, distinct
* kind of WCO) on the target tuple.
*
* Make a second, recursive pass over the RTE for this, gathering
* UPDATE-applicable RLS checks/WCOs, and gathering and converting
* UPDATE-applicable security quals into WCO_RLS_CONFLICT_CHECK RLS
* checks/WCOs. Finally, these distinct kinds of RLS checks/WCOs are
* concatenated with our own INSERT-applicable list.
*/
if (root->onConflict && root->onConflict->action == ONCONFLICT_UPDATE &&
commandType == CMD_INSERT)
{
List *conflictSecurityQuals = NIL;
List *conflictWCOs = NIL;
ListCell *item;
bool conflictHasRowSecurity = false;
bool conflictHasSublinks = false;
/* Assume that RTE is target resultRelation */
get_row_security_policies(root, CMD_UPDATE, rte, rt_index,
&conflictSecurityQuals, &conflictWCOs,
&conflictHasRowSecurity,
&conflictHasSublinks);
if (conflictHasRowSecurity)
*hasRowSecurity = true;
if (conflictHasSublinks)
*hasSubLinks = true;
/*
* Append WITH CHECK OPTIONs/RLS checks, which should not conflict
* between this INSERT and the auxiliary UPDATE
*/
*withCheckOptions = list_concat(*withCheckOptions,
conflictWCOs);
foreach(item, conflictSecurityQuals)
{
Expr *conflict_rowsec_expr = (Expr *) lfirst(item);
WithCheckOption *wco;
wco = (WithCheckOption *) makeNode(WithCheckOption);
wco->kind = WCO_RLS_CONFLICT_CHECK;
wco->relname = pstrdup(RelationGetRelationName(rel));
wco->qual = (Node *) copyObject(conflict_rowsec_expr);
wco->cascaded = false;
*withCheckOptions = lappend(*withCheckOptions, wco);
}
}
/*
* Traverse the tree to find path from root page to specified "child" block.
*
* returns a new insertion stack, starting from the parent of "child", up
* to the root. *downlinkoffnum is set to the offset of the downlink in the
* direct parent of child.
*
* To prevent deadlocks, this should lock only one page at a time.
*/
static GISTInsertStack *
gistFindPath(Relation r, BlockNumber child, OffsetNumber *downlinkoffnum)
{
Page page;
Buffer buffer;
OffsetNumber i,
maxoff;
ItemId iid;
IndexTuple idxtuple;
List *fifo;
GISTInsertStack *top,
*ptr;
BlockNumber blkno;
top = (GISTInsertStack *) palloc0(sizeof(GISTInsertStack));
top->blkno = GIST_ROOT_BLKNO;
top->downlinkoffnum = InvalidOffsetNumber;
fifo = list_make1(top);
while (fifo != NIL)
{
/* Get next page to visit */
top = linitial(fifo);
fifo = list_delete_first(fifo);
buffer = ReadBuffer(r, top->blkno);
LockBuffer(buffer, GIST_SHARE);
gistcheckpage(r, buffer);
page = (Page) BufferGetPage(buffer);
if (GistPageIsLeaf(page))
{
/*
* Because we scan the index top-down, all the rest of the pages
* in the queue must be leaf pages as well.
*/
UnlockReleaseBuffer(buffer);
break;
}
top->lsn = PageGetLSN(page);
/*
* If F_FOLLOW_RIGHT is set, the page to the right doesn't have a
* downlink. This should not normally happen..
*/
if (GistFollowRight(page))
elog(ERROR, "concurrent GiST page split was incomplete");
if (top->parent && top->parent->lsn < GistPageGetNSN(page) &&
GistPageGetOpaque(page)->rightlink != InvalidBlockNumber /* sanity check */ )
{
/*
* Page was split while we looked elsewhere. We didn't see the
* downlink to the right page when we scanned the parent, so add
* it to the queue now.
*
* Put the right page ahead of the queue, so that we visit it
* next. That's important, because if this is the lowest internal
* level, just above leaves, we might already have queued up some
* leaf pages, and we assume that there can't be any non-leaf
* pages behind leaf pages.
*/
ptr = (GISTInsertStack *) palloc0(sizeof(GISTInsertStack));
ptr->blkno = GistPageGetOpaque(page)->rightlink;
ptr->downlinkoffnum = InvalidOffsetNumber;
ptr->parent = top->parent;
fifo = lcons(ptr, fifo);
}
maxoff = PageGetMaxOffsetNumber(page);
for (i = FirstOffsetNumber; i <= maxoff; i = OffsetNumberNext(i))
{
iid = PageGetItemId(page, i);
idxtuple = (IndexTuple) PageGetItem(page, iid);
blkno = ItemPointerGetBlockNumber(&(idxtuple->t_tid));
if (blkno == child)
{
/* Found it! */
UnlockReleaseBuffer(buffer);
*downlinkoffnum = i;
return top;
}
else
{
/* Append this child to the list of pages to visit later */
ptr = (GISTInsertStack *) palloc0(sizeof(GISTInsertStack));
ptr->blkno = blkno;
ptr->downlinkoffnum = i;
ptr->parent = top;
fifo = lappend(fifo, ptr);
}
}
UnlockReleaseBuffer(buffer);
}
elog(ERROR, "failed to re-find parent of a page in index \"%s\", block %u",
RelationGetRelationName(r), child);
return NULL; /* keep compiler quiet */
}
static lref_t execute_fast_op(lref_t fop, lref_t env)
{
lref_t retval = NIL;
lref_t sym;
lref_t binding;
lref_t fn;
lref_t args;
size_t argc;
lref_t argv[ARG_BUF_LEN];
lref_t after;
lref_t tag;
lref_t cell;
lref_t escape_retval;
jmp_buf *jmpbuf;
STACK_CHECK(&fop);
_process_interrupts();
fstack_enter_eval_frame(&fop, fop, env);
while(!NULLP(fop)) {
switch(fop->header.opcode)
{
case FOP_LITERAL:
retval = fop->as.fast_op.arg1;
fop = fop->as.fast_op.next;
break;
case FOP_GLOBAL_REF:
sym = fop->as.fast_op.arg1;
binding = SYMBOL_VCELL(sym);
if (UNBOUND_MARKER_P(binding))
vmerror_unbound(sym);
retval = binding;
fop = fop->as.fast_op.next;
break;
case FOP_GLOBAL_SET:
sym = fop->as.fast_op.arg1;
binding = SYMBOL_VCELL(sym);
if (UNBOUND_MARKER_P(binding))
vmerror_unbound(sym);
SET_SYMBOL_VCELL(sym, retval);
fop = fop->as.fast_op.next;
break;
case FOP_APPLY_GLOBAL:
sym = fop->as.fast_op.arg1;
fn = SYMBOL_VCELL(sym);
if (UNBOUND_MARKER_P(fn))
vmerror_unbound(sym);
argc = 0;
args = fop->as.fast_op.arg2;
while (CONSP(args)) {
if (argc >= ARG_BUF_LEN) {
vmerror_unsupported(_T("too many actual arguments"));
break;
}
argv[argc] = execute_fast_op(CAR(args), env);
args = CDR(args);
argc++;
}
if (!NULLP(args))
vmerror_arg_out_of_range(fop->as.fast_op.arg2,
_T("bad formal argument list"));
fop = apply(fn, argc, argv, &env, &retval);
break;
case FOP_APPLY:
argc = 0;
fn = execute_fast_op(fop->as.fast_op.arg1, env);
args = fop->as.fast_op.arg2;
while (CONSP(args)) {
if (argc >= ARG_BUF_LEN) {
vmerror_unsupported(_T("too many actual arguments"));
break;
}
argv[argc] = execute_fast_op(CAR(args), env);
args = CDR(args);
argc++;
}
if (!NULLP(args))
vmerror_arg_out_of_range(fop->as.fast_op.arg2,
_T("bad formal argument list"));
fop = apply(fn, argc, argv, &env, &retval);
break;
case FOP_IF_TRUE:
if (TRUEP(retval))
fop = fop->as.fast_op.arg1;
else
fop = fop->as.fast_op.arg2;
break;
case FOP_RETVAL:
fop = fop->as.fast_op.next;
break;
case FOP_SEQUENCE:
retval = execute_fast_op(fop->as.fast_op.arg1, env);
fop = fop->as.fast_op.arg2;
break;
case FOP_THROW:
tag = execute_fast_op(fop->as.fast_op.arg1, env);
escape_retval = execute_fast_op(fop->as.fast_op.arg2, env);
dscwritef(DF_SHOW_THROWS, (_T("; DEBUG: throw ~a, retval = ~a\n"), tag, escape_retval));
CURRENT_TIB()->escape_frame = find_matching_escape(CURRENT_TIB()->frame, tag);
CURRENT_TIB()->escape_value = escape_retval;
if (CURRENT_TIB()->escape_frame == NULL) {
/* If we don't find a matching catch for the throw, we have a
* problem and need to invoke a trap. */
vmtrap(TRAP_UNCAUGHT_THROW,
(enum vmt_options_t)(VMT_MANDATORY_TRAP | VMT_HANDLER_MUST_ESCAPE),
2, tag, escape_retval);
}
unwind_stack_for_throw();
fop = fop->as.fast_op.next;
break;
case FOP_CATCH:
tag = execute_fast_op(fop->as.fast_op.arg1, env);
jmpbuf = fstack_enter_catch_frame(tag, CURRENT_TIB()->frame);
dscwritef(DF_SHOW_THROWS, (_T("; DEBUG: setjmp tag: ~a, frame: ~c&, jmpbuf: ~c&\n"), tag, CURRENT_TIB()->frame, jmpbuf));
if (setjmp(*jmpbuf) == 0) {
retval = execute_fast_op(fop->as.fast_op.arg2, env);
} else {
dscwritef(DF_SHOW_THROWS, (_T("; DEBUG: catch, retval = ~a\n"), CURRENT_TIB()->escape_value));
retval = CURRENT_TIB()->escape_value;
CURRENT_TIB()->escape_value = NIL;
}
fstack_leave_frame();
fop = fop->as.fast_op.next;
break;
case FOP_WITH_UNWIND_FN:
fstack_enter_unwind_frame(execute_fast_op(fop->as.fast_op.arg1, env));
retval = execute_fast_op(fop->as.fast_op.arg2, env);
after = CURRENT_TIB()->frame[FOFS_UNWIND_AFTER];
fstack_leave_frame();
apply1(after, 0, NULL);
fop = fop->as.fast_op.next;
break;
case FOP_CLOSURE:
retval = lclosurecons(env,
lcons(lcar(fop->as.fast_op.arg1),
fop->as.fast_op.arg2),
lcdr(fop->as.fast_op.arg1));
fop = fop->as.fast_op.next;
break;
case FOP_CAR:
retval = lcar(retval);
fop = fop->as.fast_op.next;
break;
case FOP_CDR:
retval = lcdr(retval);
fop = fop->as.fast_op.next;
break;
case FOP_NOT:
retval = boolcons(!TRUEP(retval));
fop = fop->as.fast_op.next;
break;
case FOP_NULLP:
retval = boolcons(NULLP(retval));
fop = fop->as.fast_op.next;
break;
case FOP_EQP:
retval = boolcons(EQ(execute_fast_op(fop->as.fast_op.arg1, env),
execute_fast_op(fop->as.fast_op.arg2, env)));
fop = fop->as.fast_op.next;
break;
case FOP_GET_ENV:
retval = env;
fop = fop->as.fast_op.next;
break;
case FOP_GLOBAL_DEF: // three args, third was genv, but currently unused
retval = lidefine_global(fop->as.fast_op.arg1, fop->as.fast_op.arg2);
fop = fop->as.fast_op.next;
break;
case FOP_GET_FSP:
retval = fixcons((fixnum_t)CURRENT_TIB()->fsp);
fop = fop->as.fast_op.next;
break;
case FOP_GET_FRAME:
retval = fixcons((fixnum_t)CURRENT_TIB()->frame);
fop = fop->as.fast_op.next;
break;
case FOP_GET_HFRAMES:
retval = CURRENT_TIB()->handler_frames;
fop = fop->as.fast_op.next;
break;
case FOP_SET_HFRAMES:
CURRENT_TIB()->handler_frames = execute_fast_op(fop->as.fast_op.arg1, env);
fop = fop->as.fast_op.next;
break;
case FOP_GLOBAL_PRESERVE_FRAME:
sym = fop->as.fast_op.arg1;
binding = SYMBOL_VCELL(sym);
if (UNBOUND_MARKER_P(binding))
vmerror_unbound(sym);
SET_SYMBOL_VCELL(sym, fixcons((fixnum_t)CURRENT_TIB()->frame));
retval = execute_fast_op(fop->as.fast_op.arg2, env);
fop = fop->as.fast_op.next;
break;
case FOP_STACK_BOUNDARY:
sym = execute_fast_op(fop->as.fast_op.arg1, env);
fstack_enter_boundary_frame(sym);
retval = execute_fast_op(fop->as.fast_op.arg2, env);
fstack_leave_frame();
fop = fop->as.fast_op.next;
break;
case FOP_FAST_ENQUEUE_CELL:
retval = execute_fast_op(fop->as.fast_op.arg2, env);
cell = execute_fast_op(fop->as.fast_op.arg1, env);
SET_CDR(CAR(retval), cell);
SET_CAR(retval, cell);
fop = fop->as.fast_op.next;
break;
case FOP_WHILE_TRUE:
while(TRUEP(execute_fast_op(fop->as.fast_op.arg1, env))) {
retval = execute_fast_op(fop->as.fast_op.arg2, env);
}
fop = fop->as.fast_op.next;
break;
case FOP_LOCAL_REF_BY_INDEX:
retval = lenvlookup_by_index(FIXNM(fop->as.fast_op.arg1),
FIXNM(fop->as.fast_op.arg2),
env);
fop = fop->as.fast_op.next;
break;
case FOP_LOCAL_REF_RESTARG:
retval = lenvlookup_restarg_by_index(FIXNM(fop->as.fast_op.arg1),
FIXNM(fop->as.fast_op.arg2),
env);
fop = fop->as.fast_op.next;
break;
case FOP_LOCAL_SET_BY_INDEX:
lenvlookup_set_by_index(FIXNM(fop->as.fast_op.arg1),
FIXNM(fop->as.fast_op.arg2),
env,
retval);
fop = fop->as.fast_op.next;
break;
default:
panic("Unsupported fast-op");
}
}
fstack_leave_frame();
return retval;
}
/*
* sort_inner_and_outer
* Create mergejoin join paths by explicitly sorting both the outer and
* inner join relations on each available merge ordering.
*
* 'joinrel' is the join relation
* 'outerrel' is the outer join relation
* 'innerrel' is the inner join relation
* 'restrictlist' contains all of the RestrictInfo nodes for restriction
* clauses that apply to this join
* 'mergeclause_list' is a list of RestrictInfo nodes for available
* mergejoin clauses in this join
* 'jointype' is the type of join to do
*/
static void
sort_inner_and_outer(PlannerInfo *root,
RelOptInfo *joinrel,
RelOptInfo *outerrel,
RelOptInfo *innerrel,
List *restrictlist,
List *mergeclause_list,
JoinType jointype)
{
bool useallclauses;
Path *outer_path;
Path *inner_path;
List *all_pathkeys;
ListCell *l;
/*
* If we are doing a right or full join, we must use *all* the
* mergeclauses as join clauses, else we will not have a valid plan.
*/
switch (jointype)
{
case JOIN_INNER:
case JOIN_LEFT:
case JOIN_IN:
case JOIN_UNIQUE_OUTER:
case JOIN_UNIQUE_INNER:
useallclauses = false;
break;
case JOIN_RIGHT:
case JOIN_FULL:
useallclauses = true;
break;
default:
elog(ERROR, "unrecognized join type: %d",
(int) jointype);
useallclauses = false; /* keep compiler quiet */
break;
}
/*
* We only consider the cheapest-total-cost input paths, since we are
* assuming here that a sort is required. We will consider
* cheapest-startup-cost input paths later, and only if they don't need a
* sort.
*
* If unique-ification is requested, do it and then handle as a plain
* inner join.
*/
outer_path = outerrel->cheapest_total_path;
inner_path = innerrel->cheapest_total_path;
if (jointype == JOIN_UNIQUE_OUTER)
{
outer_path = (Path *) create_unique_path(root, outerrel, outer_path);
jointype = JOIN_INNER;
}
else if (jointype == JOIN_UNIQUE_INNER)
{
inner_path = (Path *) create_unique_path(root, innerrel, inner_path);
jointype = JOIN_INNER;
}
/*
* Each possible ordering of the available mergejoin clauses will generate
* a differently-sorted result path at essentially the same cost. We have
* no basis for choosing one over another at this level of joining, but
* some sort orders may be more useful than others for higher-level
* mergejoins, so it's worth considering multiple orderings.
*
* Actually, it's not quite true that every mergeclause ordering will
* generate a different path order, because some of the clauses may be
* partially redundant (refer to the same EquivalenceClasses). Therefore,
* what we do is convert the mergeclause list to a list of canonical
* pathkeys, and then consider different orderings of the pathkeys.
*
* Generating a path for *every* permutation of the pathkeys doesn't seem
* like a winning strategy; the cost in planning time is too high. For
* now, we generate one path for each pathkey, listing that pathkey first
* and the rest in random order. This should allow at least a one-clause
* mergejoin without re-sorting against any other possible mergejoin
* partner path. But if we've not guessed the right ordering of secondary
* keys, we may end up evaluating clauses as qpquals when they could have
* been done as mergeclauses. (In practice, it's rare that there's more
* than two or three mergeclauses, so expending a huge amount of thought
* on that is probably not worth it.)
*
* The pathkey order returned by select_outer_pathkeys_for_merge() has
* some heuristics behind it (see that function), so be sure to try it
* exactly as-is as well as making variants.
*/
all_pathkeys = select_outer_pathkeys_for_merge(root,
mergeclause_list,
joinrel);
foreach(l, all_pathkeys)
{
List *front_pathkey = (List *) lfirst(l);
List *cur_mergeclauses;
List *outerkeys;
List *innerkeys;
List *merge_pathkeys;
/* Make a pathkey list with this guy first */
if (l != list_head(all_pathkeys))
outerkeys = lcons(front_pathkey,
list_delete_ptr(list_copy(all_pathkeys),
front_pathkey));
else
outerkeys = all_pathkeys; /* no work at first one... */
/* Sort the mergeclauses into the corresponding ordering */
cur_mergeclauses = find_mergeclauses_for_pathkeys(root,
outerkeys,
true,
mergeclause_list);
/* Should have used them all... */
Assert(list_length(cur_mergeclauses) == list_length(mergeclause_list));
/* Build sort pathkeys for the inner side */
innerkeys = make_inner_pathkeys_for_merge(root,
cur_mergeclauses,
outerkeys);
/* Build pathkeys representing output sort order */
merge_pathkeys = build_join_pathkeys(root, joinrel, jointype,
outerkeys);
/*
* And now we can make the path.
*
* Note: it's possible that the cheapest paths will already be sorted
* properly. create_mergejoin_path will detect that case and suppress
* an explicit sort step, so we needn't do so here.
*/
add_path(joinrel, (Path *)
create_mergejoin_path(root,
joinrel,
jointype,
outer_path,
inner_path,
restrictlist,
merge_pathkeys,
cur_mergeclauses,
outerkeys,
innerkeys));
}
/*
* sepgsql_avc_compute
*
* A fallback path, when cache mishit. It asks SELinux its access control
* decision for the supplied pair of security context and object class.
*/
static avc_cache *
sepgsql_avc_compute(const char *scontext, const char *tcontext, uint16 tclass)
{
char *ucontext = NULL;
char *ncontext = NULL;
MemoryContext oldctx;
avc_cache *cache;
uint32 hash;
int index;
struct av_decision avd;
hash = sepgsql_avc_hash(scontext, tcontext, tclass);
index = hash % AVC_NUM_SLOTS;
/*
* Validation check of the supplied security context. Because it always
* invoke system-call, frequent check should be avoided. Unless security
* policy is reloaded, validation status shall be kept, so we also cache
* whether the supplied security context was valid, or not.
*/
if (security_check_context_raw((security_context_t) tcontext) != 0)
ucontext = sepgsql_avc_unlabeled();
/*
* Ask SELinux its access control decision
*/
if (!ucontext)
sepgsql_compute_avd(scontext, tcontext, tclass, &avd);
else
sepgsql_compute_avd(scontext, ucontext, tclass, &avd);
/*
* It also caches a security label to be switched when a client labeled as
* 'scontext' executes a procedure labeled as 'tcontext', not only access
* control decision on the procedure. The security label to be switched
* shall be computed uniquely on a pair of 'scontext' and 'tcontext',
* thus, it is reasonable to cache the new label on avc, and enables to
* reduce unnecessary system calls. It shall be referenced at
* sepgsql_needs_fmgr_hook to check whether the supplied function is a
* trusted procedure, or not.
*/
if (tclass == SEPG_CLASS_DB_PROCEDURE)
{
if (!ucontext)
ncontext = sepgsql_compute_create(scontext, tcontext,
SEPG_CLASS_PROCESS, NULL);
else
ncontext = sepgsql_compute_create(scontext, ucontext,
SEPG_CLASS_PROCESS, NULL);
if (strcmp(scontext, ncontext) == 0)
{
pfree(ncontext);
ncontext = NULL;
}
}
/*
* Set up an avc_cache object
*/
oldctx = MemoryContextSwitchTo(avc_mem_cxt);
cache = palloc0(sizeof(avc_cache));
cache->hash = hash;
cache->scontext = pstrdup(scontext);
cache->tcontext = pstrdup(tcontext);
cache->tclass = tclass;
cache->allowed = avd.allowed;
cache->auditallow = avd.auditallow;
cache->auditdeny = avd.auditdeny;
cache->hot_cache = true;
if (avd.flags & SELINUX_AVD_FLAGS_PERMISSIVE)
cache->permissive = true;
if (!ucontext)
cache->tcontext_is_valid = true;
if (ncontext)
cache->ncontext = pstrdup(ncontext);
avc_num_caches++;
if (avc_num_caches > avc_threshold)
sepgsql_avc_reclaim();
avc_slots[index] = lcons(cache, avc_slots[index]);
MemoryContextSwitchTo(oldctx);
return cache;
}