1 ------------------------------------------------------------------------------
3 -- GNAT COMPILER COMPONENTS --
9 -- Copyright (C) 1992-2019, Free Software Foundation, Inc. --
11 -- GNAT is free software; you can redistribute it and/or modify it under --
12 -- terms of the GNU General Public License as published by the Free Soft- --
13 -- ware Foundation; either version 3, or (at your option) any later ver- --
14 -- sion. GNAT is distributed in the hope that it will be useful, but WITH- --
15 -- OUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY --
16 -- or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License --
17 -- for more details. You should have received a copy of the GNU General --
18 -- Public License distributed with GNAT; see file COPYING3. If not, go to --
19 -- http://www.gnu.org/licenses for a complete copy of the license. --
21 -- GNAT was originally developed by the GNAT team at New York University. --
22 -- Extensive contributions were provided by Ada Core Technologies Inc. --
24 ------------------------------------------------------------------------------
26 -- Package containing routines used to deal with runtime checks. These
27 -- routines are used both by the semantics and by the expander. In some
28 -- cases, checks are enabled simply by setting flags for gigi, and in
29 -- other cases the code for the check is expanded.
31 -- The approach used for range and length checks, in regards to suppressed
32 -- checks, is to attempt to detect at compilation time that a constraint
33 -- error will occur. If this is detected a warning or error is issued and the
34 -- offending expression or statement replaced with a constraint error node.
35 -- This always occurs whether checks are suppressed or not. Dynamic range
36 -- checks are, of course, not inserted if checks are suppressed.
38 with Errout; use Errout;
39 with Namet; use Namet;
41 with Types; use Types;
42 with Uintp; use Uintp;
43 with Urealp; use Urealp;
48 -- Called for each new main source program, to initialize internal
49 -- variables used in the package body of the Checks unit.
51 function Access_Checks_Suppressed (E : Entity_Id) return Boolean;
52 function Accessibility_Checks_Suppressed (E : Entity_Id) return Boolean;
53 function Alignment_Checks_Suppressed (E : Entity_Id) return Boolean;
54 function Allocation_Checks_Suppressed (E : Entity_Id) return Boolean;
55 function Atomic_Synchronization_Disabled (E : Entity_Id) return Boolean;
56 function Discriminant_Checks_Suppressed (E : Entity_Id) return Boolean;
57 function Division_Checks_Suppressed (E : Entity_Id) return Boolean;
58 function Duplicated_Tag_Checks_Suppressed (E : Entity_Id) return Boolean;
59 function Elaboration_Checks_Suppressed (E : Entity_Id) return Boolean;
60 function Index_Checks_Suppressed (E : Entity_Id) return Boolean;
61 function Length_Checks_Suppressed (E : Entity_Id) return Boolean;
62 function Overflow_Checks_Suppressed (E : Entity_Id) return Boolean;
63 function Predicate_Checks_Suppressed (E : Entity_Id) return Boolean;
64 function Range_Checks_Suppressed (E : Entity_Id) return Boolean;
65 function Storage_Checks_Suppressed (E : Entity_Id) return Boolean;
66 function Tag_Checks_Suppressed (E : Entity_Id) return Boolean;
67 function Validity_Checks_Suppressed (E : Entity_Id) return Boolean;
68 -- These functions check to see if the named check is suppressed, either
69 -- by an active scope suppress setting, or because the check has been
70 -- specifically suppressed for the given entity. If no entity is relevant
71 -- for the current check, then Empty is used as an argument. Note: the
72 -- reason we insist on specifying Empty is to force the caller to think
73 -- about whether there is any relevant entity that should be checked.
75 function Is_Check_Suppressed (E : Entity_Id; C : Check_Id) return Boolean;
76 -- This function is called if Checks_May_Be_Suppressed (E) is True to
77 -- determine whether check C is suppressed either on the entity E or
78 -- as the result of a scope suppress pragma. If Checks_May_Be_Suppressed
79 -- is False, then the status of the check can be determined simply by
80 -- examining Scope_Suppress, so this routine is not called in that case.
82 function Overflow_Check_Mode return Overflow_Mode_Type;
83 -- Returns current overflow checking mode, taking into account whether
84 -- we are inside an assertion expression and the assertion policy.
86 -----------------------------------------
87 -- Control of Alignment Check Warnings --
88 -----------------------------------------
90 -- When we have address clauses, there is an issue of whether the address
91 -- specified is appropriate to the alignment. In the general case where the
92 -- address is dynamic, we generate a check and a possible warning (this
93 -- warning occurs for example if we have a restricted runtime with the
94 -- restriction No_Exception_Propagation). We also issue this warning in
95 -- the case where the address is static, but we don't know the alignment
96 -- at the time we process the address clause. In such a case, we issue the
97 -- warning, but we may be able to find out later (after the back end has
98 -- annotated the actual alignment chosen) that the warning was not needed.
100 -- To deal with deleting these potentially annoying warnings, we save the
101 -- warning information in a table, and then delete the warnings in the
102 -- post compilation validation stage if we can tell that the check would
103 -- never fail (in general the back end will also optimize away the check
106 -- Table used to record information
108 type Alignment_Warnings_Record is record
110 -- Entity whose alignment possibly warrants a warning
113 -- Compile time known value of address clause for which the alignment
114 -- is to be checked once we know the alignment.
117 -- Prefix of address clause when it is of the form X'Address
120 -- Id of warning message we might delete
123 package Alignment_Warnings is new Table.Table (
124 Table_Component_Type => Alignment_Warnings_Record,
125 Table_Index_Type => Int,
126 Table_Low_Bound => 0,
128 Table_Increment => 200,
129 Table_Name => "Alignment_Warnings");
131 procedure Validate_Alignment_Check_Warnings;
132 -- This routine is called after back annotation of type data to delete any
133 -- alignment warnings that turn out to be false alarms, based on knowing
134 -- the actual alignment, and a compile-time known alignment value.
136 -------------------------------------------
137 -- Procedures to Activate Checking Flags --
138 -------------------------------------------
140 procedure Activate_Division_Check (N : Node_Id);
141 pragma Inline (Activate_Division_Check);
142 -- Sets Do_Division_Check flag in node N, and handles possible local raise.
143 -- Always call this routine rather than calling Set_Do_Division_Check to
144 -- set an explicit value of True, to ensure handling the local raise case.
146 procedure Activate_Overflow_Check (N : Node_Id);
147 pragma Inline (Activate_Overflow_Check);
148 -- Sets Do_Overflow_Check flag in node N, and handles possible local raise.
149 -- Always call this routine rather than calling Set_Do_Overflow_Check to
150 -- set an explicit value of True, to ensure handling the local raise case.
151 -- Note that for discrete types, this call has no effect for MOD, REM, and
152 -- unary "+" for which overflow is never possible in any case.
154 -- Note: for the discrete-type case, it is legitimate to call this routine
155 -- on an unanalyzed node where the Etype field is not set. However, for the
156 -- floating-point case, Etype must be set (to a floating-point type).
158 -- For floating-point, we set the flag if we have automatic overflow checks
159 -- on the target, or if Check_Float_Overflow mode is set. For the floating-
160 -- point case, we ignore all the unary operators ("+", "-", and abs) since
161 -- none of these can result in overflow. If there are no overflow checks on
162 -- the target, and Check_Float_Overflow mode is not set, then the call has
163 -- no effect, since in such cases we want to generate NaN's and infinities.
165 procedure Activate_Range_Check (N : Node_Id);
166 pragma Inline (Activate_Range_Check);
167 -- Sets Do_Range_Check flag in node N, and handles possible local raise.
168 -- Always call this routine rather than calling Set_Do_Range_Check to
169 -- set an explicit value of True, to ensure handling the local raise case.
171 --------------------------------
172 -- Procedures to Apply Checks --
173 --------------------------------
175 -- General note on following checks. These checks are always active if
176 -- Expander_Active and not Inside_A_Generic. They are inactive and have
177 -- no effect Inside_A_Generic. In the case where not Expander_Active
178 -- and not Inside_A_Generic, most of them are inactive, but some of them
179 -- operate anyway since they may generate useful compile time warnings.
181 procedure Apply_Access_Check (N : Node_Id);
182 -- Determines whether an expression node requires a runtime access
183 -- check and if so inserts the appropriate run-time check.
185 procedure Apply_Accessibility_Check
188 Insert_Node : Node_Id);
189 -- Given a name N denoting an access parameter, emits a run-time
190 -- accessibility check (if necessary), checking that the level of
191 -- the object denoted by the access parameter is not deeper than the
192 -- level of the type Typ. Program_Error is raised if the check fails.
193 -- Insert_Node indicates the node where the check should be inserted.
195 procedure Apply_Address_Clause_Check (E : Entity_Id; N : Node_Id);
196 -- E is the entity for an object which has an address clause. If checks
197 -- are enabled, then this procedure generates a check that the specified
198 -- address has an alignment consistent with the alignment of the object,
199 -- raising PE if this is not the case. The resulting check (if one is
200 -- generated) is prepended to the Actions list of N_Freeze_Entity node N.
201 -- Note that the check references E'Alignment, so it cannot be emitted
202 -- before N (its freeze node), otherwise this would cause an illegal
203 -- access before elaboration error in GIGI. For the case of a clear overlay
204 -- situation, we also check that the size of the overlaying object is not
205 -- larger than the overlaid object.
207 procedure Apply_Arithmetic_Overflow_Check (N : Node_Id);
208 -- Handle overflow checking for an arithmetic operator. Also handles the
209 -- cases of ELIMINATED and MINIMIZED overflow checking mode. If the mode
210 -- is one of the latter two, then this routine can also be called with
211 -- an if or case expression node to make sure that we properly handle
212 -- overflow checking for dependent expressions. This routine handles
213 -- front end vs back end overflow checks (in the front end case it expands
214 -- the necessary check). Note that divide is handled separately using
215 -- Apply_Divide_Checks. Node N may or may not have Do_Overflow_Check.
216 -- In STRICT mode, there is nothing to do if this flag is off, but in
217 -- MINIMIZED/ELIMINATED mode we still have to deal with possible use
218 -- of doing operations in Long_Long_Integer or Bignum mode.
220 procedure Apply_Constraint_Check
223 No_Sliding : Boolean := False);
224 -- Top-level procedure, calls all the others depending on the class of
225 -- Typ. Checks that expression N satisfies the constraint of type Typ.
226 -- No_Sliding is only relevant for constrained array types, if set to
227 -- True, it checks that indexes are in range.
229 procedure Apply_Discriminant_Check
232 Lhs : Node_Id := Empty);
233 -- Given an expression N of a discriminated type, or of an access type
234 -- whose designated type is a discriminanted type, generates a check to
235 -- ensure that the expression can be converted to the subtype given as
236 -- the second parameter. Lhs is empty except in the case of assignments,
237 -- where the target object may be needed to determine the subtype to
238 -- check against (such as the cases of unconstrained formal parameters
239 -- and unconstrained aliased objects). For the case of unconstrained
240 -- formals, the check is performed only if the corresponding actual is
241 -- constrained, i.e., whether Lhs'Constrained is True.
243 procedure Apply_Divide_Checks (N : Node_Id);
244 -- The node kind is N_Op_Divide, N_Op_Mod, or N_Op_Rem if either of the
245 -- flags Do_Division_Check or Do_Overflow_Check is set, then this routine
246 -- ensures that the appropriate checks are made. Note that overflow can
247 -- occur in the signed case for the case of the largest negative number
248 -- divided by minus one. This procedure only applies to Integer types.
250 procedure Apply_Parameter_Aliasing_Checks
253 -- Given a subprogram call Call, add a check to verify that none of the
254 -- actuals overlap. Subp denotes the subprogram being called.
256 procedure Apply_Parameter_Validity_Checks (Subp : Entity_Id);
257 -- Given a subprogram Subp, add both a pre and post condition pragmas that
258 -- verify the proper initialization of scalars in parameters and function
261 procedure Apply_Predicate_Check
264 Fun : Entity_Id := Empty);
265 -- N is an expression to which a predicate check may need to be applied for
266 -- Typ, if Typ has a predicate function. When N is an actual in a call, Fun
267 -- is the function being called, which is used to generate a better warning
268 -- if the call leads to an infinite recursion.
270 procedure Apply_Type_Conversion_Checks (N : Node_Id);
271 -- N is an N_Type_Conversion node. A type conversion actually involves
272 -- two sorts of checks. The first check is the checks that ensures that
273 -- the operand in the type conversion fits onto the base type of the
274 -- subtype it is being converted to (see RM 4.6 (28)-(50)). The second
275 -- check is there to ensure that once the operand has been converted to
276 -- a value of the target type, this converted value meets the
277 -- constraints imposed by the target subtype (see RM 4.6 (51)).
279 procedure Apply_Universal_Integer_Attribute_Checks (N : Node_Id);
280 -- The argument N is an attribute reference node intended for processing
281 -- by gigi. The attribute is one that returns a universal integer, but
282 -- the attribute reference node is currently typed with the expected
283 -- result type. This routine deals with range and overflow checks needed
284 -- to make sure that the universal result is in range.
286 function Build_Discriminant_Checks
290 -- Subsidiary routine for Apply_Discriminant_Check. Builds the expression
291 -- that compares discriminants of the expression with discriminants of the
292 -- type. Also used directly for membership tests (see Exp_Ch4.Expand_N_In).
294 function Convert_From_Bignum (N : Node_Id) return Node_Id;
295 -- Returns result of converting node N from Bignum. The returned value is
296 -- not analyzed, the caller takes responsibility for this. Node N must be
297 -- a subexpression node of type Bignum. The result is Long_Long_Integer.
299 function Convert_To_Bignum (N : Node_Id) return Node_Id;
300 -- Returns result of converting node N to Bignum. The returned value is not
301 -- analyzed, the caller takes responsibility for this. Node N must be a
302 -- subexpression node of a signed integer type or Bignum type (if it is
303 -- already a Bignum, the returned value is Relocate_Node (N)).
305 procedure Determine_Range
310 Assume_Valid : Boolean := False);
311 -- N is a node for a subexpression. If N is of a discrete type with no
312 -- error indications, and no other peculiarities (e.g. missing Etype),
313 -- then OK is True on return, and Lo and Hi are set to a conservative
314 -- estimate of the possible range of values of N. Thus if OK is True on
315 -- return, the value of the subexpression N is known to lie in the range
316 -- Lo .. Hi (inclusive). For enumeration and character literals the values
317 -- returned are the Pos value in the relevant enumeration type. If the
318 -- expression is not of a discrete type, or some kind of error condition
319 -- is detected, then OK is False on exit, and Lo/Hi are set to No_Uint.
320 -- Thus the significance of OK being False on return is that no useful
321 -- information is available on the range of the expression. Assume_Valid
322 -- determines whether the processing is allowed to assume that values are
323 -- in range of their subtypes. If it is set to True, then this assumption
324 -- is valid, if False, then processing is done using base types to allow
327 procedure Determine_Range_R
332 Assume_Valid : Boolean := False);
333 -- Similar to Determine_Range, but for a node N of floating-point type. OK
334 -- is True on return only for IEEE floating-point types and only if we do
335 -- not have to worry about extended precision (i.e. on the x86, we must be
336 -- using -msse2 -mfpmath=sse). At the current time, this is used only in
337 -- GNATprove, though we could consider using it more generally in future.
338 -- For that to happen, the possibility of arguments of infinite or NaN
339 -- value should be taken into account, which is not the case currently.
341 procedure Install_Null_Excluding_Check (N : Node_Id);
342 -- Determines whether an access node requires a runtime access check and
343 -- if so inserts the appropriate run-time check.
345 procedure Install_Primitive_Elaboration_Check (Subp_Body : Node_Id);
346 -- Insert a check which ensures that subprogram body Subp_Body has been
347 -- properly elaborated. The check is installed only when Subp_Body is the
348 -- body of a nonabstract library-level primitive of a tagged type. Further
349 -- restrictions may apply, see the body for details.
351 function Make_Bignum_Block (Loc : Source_Ptr) return Node_Id;
352 -- This function is used by top level overflow checking routines to do a
353 -- mark/release operation on the secondary stack around bignum operations.
354 -- The block created looks like:
357 -- M : Mark_Id := SS_Mark;
362 -- The idea is that the caller will insert any needed extra declarations
363 -- after the declaration of M, and any needed statements (in particular
364 -- the bignum operations) before the call to SS_Release, and then do an
365 -- Insert_Action of the whole block (it is returned unanalyzed). The Loc
366 -- parameter is used to supply Sloc values for the constructed tree.
368 procedure Minimize_Eliminate_Overflows
372 Top_Level : Boolean);
373 -- This is the main routine for handling MINIMIZED and ELIMINATED overflow
374 -- processing. On entry N is a node whose result is a signed integer
375 -- subtype. The Do_Overflow_Check flag may or may not be set on N. If the
376 -- node is an arithmetic operation, then a range analysis is carried out,
377 -- and there are three possibilities:
379 -- The node is left unchanged (apart from expansion of an exponentiation
380 -- operation). This happens if the routine can determine that the result
381 -- is definitely in range. The Do_Overflow_Check flag is turned off in
384 -- The node is transformed into an arithmetic operation with a result
385 -- type of Long_Long_Integer.
387 -- The node is transformed into a function call that calls an appropriate
388 -- function in the System.Bignums package to compute a Bignum result.
390 -- In the first two cases, Lo and Hi are set to the bounds of the possible
391 -- range of results, computed as accurately as possible. In the third case
392 -- Lo and Hi are set to No_Uint (there are some cases where we could get an
393 -- advantage from keeping result ranges for Bignum values, but it could use
394 -- a lot of space and is very unlikely to be valuable).
396 -- If the node is not an arithmetic operation, then it is unchanged but
397 -- Lo and Hi are still set (to the bounds of the result subtype if nothing
398 -- better can be determined).
400 -- Note: this function is recursive, if called with an arithmetic operator,
401 -- recursive calls are made to process the operands using this procedure.
402 -- So we end up doing things top down. Nothing happens to an arithmetic
403 -- expression until this procedure is called on the top level node and
404 -- then the recursive calls process all the children. We have to do it
405 -- this way. If we try to do it bottom up in natural expansion order, then
406 -- there are two problems. First, where do we stash the bounds, and more
407 -- importantly, semantic processing will be messed up. Consider A+B+C where
408 -- A,B,C are all of type integer, if we processed A+B before doing semantic
409 -- analysis of the addition of this result to C, that addition could end up
410 -- with a Long_Long_Integer left operand and an Integer right operand, and
411 -- we would get a semantic error.
413 -- The routine is called in three situations if we are operating in either
414 -- MINIMIZED or ELIMINATED modes.
416 -- Overflow processing applied to the top node of an expression tree when
417 -- that node is an arithmetic operator. In this case the result is
418 -- converted to the appropriate result type (there is special processing
419 -- when the parent is a conversion, see body for details).
421 -- Overflow processing applied to the operands of a comparison operation.
422 -- In this case, the comparison is done on the result Long_Long_Integer
423 -- or Bignum values, without raising any exceptions.
425 -- Overflow processing applied to the left operand of a membership test.
426 -- In this case no exception is raised if a Long_Long_Integer or Bignum
427 -- result is outside the range of the type of that left operand (it is
428 -- just that the result of IN is false in that case).
430 -- Note that if Bignum values appear, the caller must take care of doing
431 -- the appropriate mark/release operations on the secondary stack.
433 -- Top_Level is used to avoid inefficient unnecessary transitions into the
434 -- Bignum domain. If Top_Level is True, it means that the caller will have
435 -- to convert any Bignum value back to Long_Long_Integer, possibly checking
436 -- that the value is in range. This is the normal case for a top level
437 -- operator in a subexpression. There is no point in going into Bignum mode
438 -- to avoid an overflow just so we can check for overflow the next moment.
439 -- For calls from comparisons and membership tests, and for all recursive
440 -- calls, we do want to transition into the Bignum domain if necessary.
441 -- Note that this setting is only relevant in ELIMINATED mode.
443 -------------------------------------------------------
444 -- Control and Optimization of Range/Overflow Checks --
445 -------------------------------------------------------
447 -- Range checks are controlled by the Do_Range_Check flag. The front end
448 -- is responsible for setting this flag in relevant nodes. Originally
449 -- the back end generated all corresponding range checks. But later on
450 -- we decided to generate many range checks in the front end. We are now
451 -- in the transitional phase where some of these checks are still done
452 -- by the back end, but many are done by the front end. It is possible
453 -- that in the future we might move all the checks to the front end. The
454 -- main remaining back end checks are for subscript checking.
456 -- Overflow checks are similarly controlled by the Do_Overflow_Check flag.
457 -- The difference here is that if back end overflow checks are inactive
458 -- (Backend_Overflow_Checks_On_Target set False), then the actual overflow
459 -- checks are generated by the front end, but if back end overflow checks
460 -- are active (Backend_Overflow_Checks_On_Target set True), then the back
461 -- end does generate the checks.
463 -- The following two routines are used to set these flags, they allow
464 -- for the possibility of eliminating checks. Checks can be eliminated
465 -- if an identical check has already been performed.
467 procedure Enable_Overflow_Check (N : Node_Id);
468 -- First this routine determines if an overflow check is needed by doing
469 -- an appropriate range check. If a check is not needed, then the call
470 -- has no effect. If a check is needed then this routine sets the flag
471 -- Do_Overflow_Check in node N to True, unless it can be determined that
472 -- the check is not needed. The only condition under which this is the
473 -- case is if there was an identical check earlier on.
475 procedure Enable_Range_Check (N : Node_Id);
476 -- Set Do_Range_Check flag in node N True, unless it can be determined
477 -- that the check is not needed. The only condition under which this is
478 -- the case is if there was an identical check earlier on. This routine
479 -- is not responsible for doing range analysis to determine whether or
480 -- not such a check is needed -- the caller is expected to do this. The
481 -- one other case in which the request to set the flag is ignored is
482 -- when Kill_Range_Check is set in an N_Unchecked_Conversion node.
484 -- The following routines are used to keep track of processing sequences
485 -- of statements (e.g. the THEN statements of an IF statement). A check
486 -- that appears within such a sequence can eliminate an identical check
487 -- within this sequence of statements. However, after the end of the
488 -- sequence of statements, such a check is no longer of interest, since
489 -- it may not have been executed.
491 procedure Conditional_Statements_Begin;
492 -- This call marks the start of processing of a sequence of statements.
493 -- Every call to this procedure must be followed by a matching call to
494 -- Conditional_Statements_End.
496 procedure Conditional_Statements_End;
497 -- This call removes from consideration all saved checks since the
498 -- corresponding call to Conditional_Statements_Begin. These two
499 -- procedures operate in a stack like manner.
501 -- The mechanism for optimizing checks works by remembering checks
502 -- that have already been made, but certain conditions, for example
503 -- an assignment to a variable involved in a check, may mean that the
504 -- remembered check is no longer valid, in the sense that if the same
505 -- expression appears again, another check is required because the
506 -- value may have changed.
508 -- The following routines are used to note conditions which may render
509 -- some or all of the stored and remembered checks to be invalidated.
511 procedure Kill_Checks (V : Entity_Id);
512 -- This procedure records an assignment or other condition that causes
513 -- the value of the variable to be changed, invalidating any stored
514 -- checks that reference the value. Note that all such checks must
515 -- be discarded, even if they are not in the current statement range.
517 procedure Kill_All_Checks;
518 -- This procedure kills all remembered checks
520 -----------------------------
521 -- Length and Range Checks --
522 -----------------------------
524 -- In the following procedures, there are three arguments which have
525 -- a common meaning as follows:
527 -- Expr The expression to be checked. If a check is required,
528 -- the appropriate flag will be placed on this node. Whether
529 -- this node is further examined depends on the setting of
530 -- the parameter Source_Typ, as described below.
532 -- ??? Apply_Length_Check and Apply_Range_Check do not have an Expr
535 -- ??? Apply_Length_Check and Apply_Range_Check have a Ck_Node formal
536 -- which is undocumented, is it the same as Expr?
538 -- Target_Typ The target type on which the check is to be based. For
539 -- example, if we have a scalar range check, then the check
540 -- is that we are in range of this type.
542 -- Source_Typ Normally Empty, but can be set to a type, in which case
543 -- this type is used for the check, see below.
545 -- The checks operate in one of two modes:
547 -- If Source_Typ is Empty, then the node Expr is examined, at the very
548 -- least to get the source subtype. In addition for some of the checks,
549 -- the actual form of the node may be examined. For example, a node of
550 -- type Integer whose actual form is an Integer conversion from a type
551 -- with range 0 .. 3 can be determined to have a value in range 0 .. 3.
553 -- If Source_Typ is given, then nothing can be assumed about the Expr,
554 -- and indeed its contents are not examined. In this case the check is
555 -- based on the assumption that Expr can be an arbitrary value of the
558 -- Currently, the only case in which a Source_Typ is explicitly supplied
559 -- is for the case of Out and In_Out parameters, where, for the conversion
560 -- on return (the Out direction), the types must be reversed. This is
561 -- handled by the caller.
563 procedure Apply_Length_Check
565 Target_Typ : Entity_Id;
566 Source_Typ : Entity_Id := Empty);
567 -- This procedure builds a sequence of declarations to do a length check
568 -- that checks if the lengths of the two arrays Target_Typ and source type
569 -- are the same. The resulting actions are inserted at Node using a call
570 -- to Insert_Actions.
572 -- For access types, the Directly_Designated_Type is retrieved and
573 -- processing continues as enumerated above, with a guard against null
576 -- Note: calls to Apply_Length_Check currently never supply an explicit
577 -- Source_Typ parameter, but Apply_Length_Check takes this parameter and
578 -- processes it as described above for consistency with the other routines
581 procedure Apply_Range_Check
583 Target_Typ : Entity_Id;
584 Source_Typ : Entity_Id := Empty);
585 -- For a Node of kind N_Range, constructs a range check action that tests
586 -- first that the range is not null and then that the range is contained in
587 -- the Target_Typ range.
589 -- For scalar types, constructs a range check action that first tests that
590 -- the expression is contained in the Target_Typ range. The difference
591 -- between this and Apply_Scalar_Range_Check is that the latter generates
592 -- the actual checking code against the Etype of the expression.
594 -- For constrained array types, construct series of range check actions
595 -- to check that each Expr range is properly contained in the range of
598 -- For a type conversion to an unconstrained array type, constructs a range
599 -- check action to check that the bounds of the source type are within the
600 -- constraints imposed by the Target_Typ.
602 -- For access types, the Directly_Designated_Type is retrieved and
603 -- processing continues as enumerated above, with a guard against null
606 -- The source type is used by type conversions to unconstrained array
607 -- types to retrieve the corresponding bounds.
609 procedure Apply_Static_Length_Check
611 Target_Typ : Entity_Id;
612 Source_Typ : Entity_Id := Empty);
613 -- Tries to determine statically whether the two array types source type
614 -- and Target_Typ have the same length. If it can be determined at compile
615 -- time that they do not, then an N_Raise_Constraint_Error node replaces
616 -- Expr, and a warning message is issued.
618 procedure Apply_Scalar_Range_Check
620 Target_Typ : Entity_Id;
621 Source_Typ : Entity_Id := Empty;
622 Fixed_Int : Boolean := False);
623 -- For scalar types, determines whether an expression node should be
624 -- flagged as needing a runtime range check. If the node requires such a
625 -- check, the Do_Range_Check flag is turned on. The Fixed_Int flag if set
626 -- causes any fixed-point values to be treated as though they were discrete
627 -- values (i.e. the underlying integer value is used).
629 type Check_Result is private;
630 -- Type used to return result of Get_Range_Checks call, for later use in
631 -- call to Insert_Range_Checks procedure.
633 function Get_Range_Checks
635 Target_Typ : Entity_Id;
636 Source_Typ : Entity_Id := Empty;
637 Warn_Node : Node_Id := Empty) return Check_Result;
638 -- Like Apply_Range_Check, except it does not modify anything. Instead
639 -- it returns an encapsulated result of the check operations for later
640 -- use in a call to Insert_Range_Checks. If Warn_Node is non-empty, its
641 -- Sloc is used, in the static case, for the generated warning or error.
642 -- Additionally, it is used rather than Expr (or Low/High_Bound of Expr)
643 -- in constructing the check.
645 procedure Append_Range_Checks
646 (Checks : Check_Result;
648 Suppress_Typ : Entity_Id;
649 Static_Sloc : Source_Ptr;
650 Flag_Node : Node_Id);
651 -- Called to append range checks as returned by a call to Get_Range_Checks.
652 -- Stmts is a list to which either the dynamic check is appended or the
653 -- raise Constraint_Error statement is appended (for static checks).
654 -- Static_Sloc is the Sloc at which the raise CE node points, Flag_Node is
655 -- used as the node at which to set the Has_Dynamic_Check flag. Checks_On
656 -- is a boolean value that says if range and index checking is on or not.
658 procedure Insert_Range_Checks
659 (Checks : Check_Result;
661 Suppress_Typ : Entity_Id;
662 Static_Sloc : Source_Ptr := No_Location;
663 Flag_Node : Node_Id := Empty;
664 Do_Before : Boolean := False);
665 -- Called to insert range checks as returned by a call to Get_Range_Checks.
666 -- Node is the node after which either the dynamic check is inserted or
667 -- the raise Constraint_Error statement is inserted (for static checks).
668 -- Suppress_Typ is the type to check to determine if checks are suppressed.
669 -- Static_Sloc, if passed, is the Sloc at which the raise CE node points,
670 -- otherwise Sloc (Node) is used. The Has_Dynamic_Check flag is normally
671 -- set at Node. If Flag_Node is present, then this is used instead as the
672 -- node at which to set the Has_Dynamic_Check flag. Normally the check is
673 -- inserted after, if Do_Before is True, the check is inserted before
676 -----------------------
677 -- Expander Routines --
678 -----------------------
680 -- Some of the earlier processing for checks results in temporarily setting
681 -- the Do_Range_Check flag rather than actually generating checks. Now we
682 -- are moving the generation of such checks into the front end for reasons
683 -- of efficiency and simplicity (there were difficulties in handling this
684 -- in the back end when side effects were present in the expressions being
687 -- Probably we could eliminate the Do_Range_Check flag entirely and
688 -- generate the checks earlier, but this is a delicate area and it
689 -- seemed safer to implement the following routines, which are called
690 -- late on in the expansion process. They check the Do_Range_Check flag
691 -- and if it is set, generate the actual checks and reset the flag.
693 procedure Generate_Range_Check
695 Target_Type : Entity_Id;
696 Reason : RT_Exception_Code);
697 -- This procedure is called to actually generate and insert a range check.
698 -- A check is generated to ensure that the value of N lies within the range
699 -- of the target type. Note that the base type of N may be different from
700 -- the base type of the target type. This happens in the conversion case.
701 -- The Reason parameter is the exception code to be used for the exception
704 -- Note: if the expander is not active, or if we are in GNATprove mode,
705 -- then we do not generate explicit range code. Instead we just turn the
706 -- Do_Range_Check flag on, since in these cases that's what we want to see
707 -- in the tree (GNATprove in particular depends on this flag being set). If
708 -- we generate the actual range check, then we make sure the flag is off,
709 -- since the code we generate takes complete care of the check.
711 -- Historical note: We used to just pass on the Do_Range_Check flag to the
712 -- back end to generate the check, but now in code-generation mode we never
713 -- have this flag set, since the front end takes care of the check. The
714 -- normal processing flow now is that the analyzer typically turns on the
715 -- Do_Range_Check flag, and if it is set, this routine is called, which
716 -- turns the flag off in code-generation mode.
718 procedure Generate_Index_Checks (N : Node_Id);
719 -- This procedure is called to generate index checks on the subscripts for
720 -- the indexed component node N. Each subscript expression is examined, and
721 -- if the Do_Range_Check flag is set, an appropriate index check is
722 -- generated and the flag is reset.
724 -- Similarly, we set the flag Do_Discriminant_Check in the semantic
725 -- analysis to indicate that a discriminant check is required for selected
726 -- component of a discriminated type. The following routine is called from
727 -- the expander to actually generate the call.
729 procedure Generate_Discriminant_Check (N : Node_Id);
730 -- N is a selected component for which a discriminant check is required to
731 -- make sure that the discriminants have appropriate values for the
732 -- selection. This is done by calling the appropriate discriminant checking
733 -- routine for the selector.
735 -----------------------
736 -- Validity Checking --
737 -----------------------
739 -- In (RM 13.9.1(9-11)) we have the following rules on invalid values
741 -- If the representation of a scalar object does not represent value of
742 -- the object's subtype (perhaps because the object was not initialized),
743 -- the object is said to have an invalid representation. It is a bounded
744 -- error to evaluate the value of such an object. If the error is
745 -- detected, either Constraint_Error or Program_Error is raised.
746 -- Otherwise, execution continues using the invalid representation. The
747 -- rules of the language outside this subclause assume that all objects
748 -- have valid representations. The semantics of operations on invalid
749 -- representations are as follows:
751 -- 10 If the representation of the object represents a value of the
752 -- object's type, the value of the type is used.
754 -- 11 If the representation of the object does not represent a value
755 -- of the object's type, the semantics of operations on such
756 -- representations is implementation-defined, but does not by
757 -- itself lead to erroneous or unpredictable execution, or to
758 -- other objects becoming abnormal.
760 -- We quote the rules in full here since they are quite delicate. Most
761 -- of the time, we can just compute away with wrong values, and get a
762 -- possibly wrong result, which is well within the range of allowed
763 -- implementation defined behavior. The two tricky cases are subscripted
764 -- array assignments, where we don't want to do wild stores, and case
765 -- statements where we don't want to do wild jumps.
767 -- In GNAT, we control validity checking with a switch -gnatV that can take
768 -- three parameters, n/d/f for None/Default/Full. These modes have the
769 -- following meanings:
771 -- None (no validity checking)
773 -- In this mode, there is no specific checking for invalid values
774 -- and the code generator assumes that all stored values are always
775 -- within the bounds of the object subtype. The consequences are as
778 -- For case statements, an out of range invalid value will cause
779 -- Constraint_Error to be raised, or an arbitrary one of the case
780 -- alternatives will be executed. Wild jumps cannot result even
781 -- in this mode, since we always do a range check
783 -- For subscripted array assignments, wild stores will result in
784 -- the expected manner when addresses are calculated using values
785 -- of subscripts that are out of range.
787 -- It could perhaps be argued that this mode is still conformant with
788 -- the letter of the RM, since implementation defined is a rather
789 -- broad category, but certainly it is not in the spirit of the
790 -- RM requirement, since wild stores certainly seem to be a case of
791 -- erroneous behavior.
793 -- Default (default standard RM-compatible validity checking)
795 -- In this mode, which is the default, minimal validity checking is
796 -- performed to ensure no erroneous behavior as follows:
798 -- For case statements, an out of range invalid value will cause
799 -- Constraint_Error to be raised.
801 -- For subscripted array assignments, invalid out of range
802 -- subscript values will cause Constraint_Error to be raised.
804 -- Full (Full validity checking)
806 -- In this mode, the protections guaranteed by the standard mode are
807 -- in place, and the following additional checks are made:
809 -- For every assignment, the right side is checked for validity
811 -- For every call, IN and IN OUT parameters are checked for validity
813 -- For every subscripted array reference, both for stores and loads,
814 -- all subscripts are checked for validity.
816 -- These checks are not required by the RM, but will in practice
817 -- improve the detection of uninitialized variables, particularly
818 -- if used in conjunction with pragma Normalize_Scalars.
820 -- In the above description, we talk about performing validity checks,
821 -- but we don't actually generate a check in a case where the compiler
822 -- can be sure that the value is valid. Note that this assurance must
823 -- be achieved without assuming that any uninitialized value lies within
824 -- the range of its type. The following are cases in which values are
825 -- known to be valid. The flag Is_Known_Valid is used to keep track of
826 -- some of these cases.
828 -- If all possible stored values are valid, then any uninitialized
829 -- value must be valid.
831 -- Literals, including enumeration literals, are clearly always valid
833 -- Constants are always assumed valid, with a validity check being
834 -- performed on the initializing value where necessary to ensure that
837 -- For variables, the status is set to known valid if there is an
838 -- initializing expression. Again a check is made on the initializing
839 -- value if necessary to ensure that this assumption is valid. The
840 -- status can change as a result of local assignments to a variable.
841 -- If a known valid value is unconditionally assigned, then we mark
842 -- the left side as known valid. If a value is assigned that is not
843 -- known to be valid, then we mark the left side as invalid. This
844 -- kind of processing does NOT apply to non-local variables since we
845 -- are not following the flow graph (more properly the flow of actual
846 -- processing only corresponds to the flow graph for local assignments).
847 -- For non-local variables, we preserve the current setting, i.e. a
848 -- validity check is performed when assigning to a knonwn valid global.
850 -- Note: no validity checking is required if range checks are suppressed
851 -- regardless of the setting of the validity checking mode.
853 -- The following procedures are used in handling validity checking
855 procedure Apply_Subscript_Validity_Checks (Expr : Node_Id);
856 -- Expr is the node for an indexed component. If validity checking and
857 -- range checking are enabled, all subscripts for this indexed component
858 -- are checked for validity.
860 procedure Check_Valid_Lvalue_Subscripts (Expr : Node_Id);
861 -- Expr is a lvalue, i.e. an expression representing the target of an
862 -- assignment. This procedure checks for this expression involving an
863 -- assignment to an array value. We have to be sure that all the subscripts
864 -- in such a case are valid, since according to the rules in (RM
865 -- 13.9.1(9-11)) such assignments are not permitted to result in erroneous
866 -- behavior in the case of invalid subscript values.
868 procedure Ensure_Valid
870 Holes_OK : Boolean := False;
871 Related_Id : Entity_Id := Empty;
872 Is_Low_Bound : Boolean := False;
873 Is_High_Bound : Boolean := False);
874 -- Ensure that Expr represents a valid value of its type. If this type
875 -- is not a scalar type, then the call has no effect, since validity
876 -- is only an issue for scalar types. The effect of this call is to
877 -- check if the value is known valid, if so, nothing needs to be done.
878 -- If this is not known, then either Expr is set to be range checked,
879 -- or specific checking code is inserted so that an exception is raised
880 -- if the value is not valid.
882 -- The optional argument Holes_OK indicates whether it is necessary to
883 -- worry about enumeration types with non-standard representations leading
884 -- to "holes" in the range of possible representations. If Holes_OK is
885 -- True, then such values are assumed valid (this is used when the caller
886 -- will make a separate check for this case anyway). If Holes_OK is False,
887 -- then this case is checked, and code is inserted to ensure that Expr is
888 -- valid, raising Constraint_Error if the value is not valid.
890 -- Related_Id denotes the entity of the context where Expr appears. Flags
891 -- Is_Low_Bound and Is_High_Bound specify whether the expression to check
892 -- is the low or the high bound of a range. These three optional arguments
893 -- signal Remove_Side_Effects to create an external symbol of the form
894 -- Chars (Related_Id)_FIRST/_LAST. For suggested use of these parameters
895 -- see the warning in the body of Sem_Ch3.Process_Range_Expr_In_Decl.
897 function Expr_Known_Valid (Expr : Node_Id) return Boolean;
898 -- This function tests it the value of Expr is known to be valid in the
899 -- sense of RM 13.9.1(9-11). In the case of GNAT, it is only discrete types
900 -- which are a concern, since for non-discrete types we simply continue
901 -- computation with invalid values, which does not lead to erroneous
902 -- behavior. Thus Expr_Known_Valid always returns True if the type of Expr
903 -- is non-discrete. For discrete types the value returned is True only if
904 -- it can be determined that the value is Valid. Otherwise False is
907 procedure Insert_Valid_Check
909 Related_Id : Entity_Id := Empty;
910 Is_Low_Bound : Boolean := False;
911 Is_High_Bound : Boolean := False);
912 -- Inserts code that will check for the value of Expr being valid, in the
913 -- sense of the 'Valid attribute returning True. Constraint_Error will be
914 -- raised if the value is not valid.
916 -- Related_Id denotes the entity of the context where Expr appears. Flags
917 -- Is_Low_Bound and Is_High_Bound specify whether the expression to check
918 -- is the low or the high bound of a range. These three optional arguments
919 -- signal Remove_Side_Effects to create an external symbol of the form
920 -- Chars (Related_Id)_FIRST/_LAST. For suggested use of these parameters
921 -- see the warning in the body of Sem_Ch3.Process_Range_Expr_In_Decl.
923 procedure Null_Exclusion_Static_Checks
925 Comp : Node_Id := Empty;
926 Array_Comp : Boolean := False);
927 -- Ada 2005 (AI-231): Test for and warn on null-excluding objects or
928 -- components that will raise an exception due to initialization by null.
930 -- When a value for Comp is supplied (as in the case of an uninitialized
931 -- null-excluding component within a composite object), a reported warning
932 -- will indicate the offending component instead of the object itself.
933 -- Array_Comp being True indicates an array object with null-excluding
934 -- components, and any reported warning will indicate that.
936 procedure Remove_Checks (Expr : Node_Id);
937 -- Remove all checks from Expr except those that are only executed
938 -- conditionally (on the right side of And Then/Or Else. This call
939 -- removes only embedded checks (Do_Range_Check, Do_Overflow_Check).
941 procedure Validity_Check_Range
943 Related_Id : Entity_Id := Empty);
944 -- If N is an N_Range node, then Ensure_Valid is called on its bounds, if
945 -- validity checking of operands is enabled. Related_Id denotes the entity
946 -- of the context where N appears.
948 -----------------------------
949 -- Handling of Check Names --
950 -----------------------------
952 -- The following table contains Name_Id's for recognized checks. The first
953 -- entries (corresponding to the values of the subtype Predefined_Check_Id)
954 -- contain the Name_Id values for the checks that are predefined, including
955 -- All_Checks (see Types). Remaining entries are those that are introduced
956 -- by pragma Check_Names.
958 package Check_Names is new Table.Table (
959 Table_Component_Type => Name_Id,
960 Table_Index_Type => Check_Id,
961 Table_Low_Bound => 1,
963 Table_Increment => 200,
964 Table_Name => "Name_Check_Names");
966 function Get_Check_Id (N : Name_Id) return Check_Id;
967 -- Function to search above table for matching name. If found returns the
968 -- corresponding Check_Id value in the range 1 .. Check_Name.Last. If not
969 -- found returns No_Check_Id.
973 type Check_Result is array (Positive range 1 .. 2) of Node_Id;
974 -- There are two cases for the result returned by Range_Check:
976 -- For the static case the result is one or two nodes that should cause
977 -- a Constraint_Error. Typically these will include Expr itself or the
978 -- direct descendants of Expr, such as Low/High_Bound (Expr)). It is the
979 -- responsibility of the caller to rewrite and substitute the nodes with
980 -- N_Raise_Constraint_Error nodes.
982 -- For the non-static case a single N_Raise_Constraint_Error node with a
983 -- non-empty Condition field is returned.
985 -- Unused entries in Check_Result, if any, are simply set to Empty For
986 -- external clients, the required processing on this result is achieved
987 -- using the Insert_Range_Checks routine.
989 pragma Inline (Apply_Length_Check);
990 pragma Inline (Apply_Range_Check);
991 pragma Inline (Apply_Static_Length_Check);
projects / thirdparty / gcc.git / blob