1/* Scalar evolution detector.
2 Copyright (C) 2003-2024 Free Software Foundation, Inc.
3 Contributed by Sebastian Pop <s.pop@laposte.net>
4
5This file is part of GCC.
6
7GCC is free software; you can redistribute it and/or modify it under
8the terms of the GNU General Public License as published by the Free
9Software Foundation; either version 3, or (at your option) any later
10version.
11
12GCC is distributed in the hope that it will be useful, but WITHOUT ANY
13WARRANTY; without even the implied warranty of MERCHANTABILITY or
14FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
15for more details.
16
17You should have received a copy of the GNU General Public License
18along with GCC; see the file COPYING3. If not see
19<http://www.gnu.org/licenses/>. */
20
21/*
22 Description:
23
24 This pass analyzes the evolution of scalar variables in loop
25 structures. The algorithm is based on the SSA representation,
26 and on the loop hierarchy tree. This algorithm is not based on
27 the notion of versions of a variable, as it was the case for the
28 previous implementations of the scalar evolution algorithm, but
29 it assumes that each defined name is unique.
30
31 The notation used in this file is called "chains of recurrences",
32 and has been proposed by Eugene Zima, Robert Van Engelen, and
33 others for describing induction variables in programs. For example
34 "b -> {0, +, 2}_1" means that the scalar variable "b" is equal to 0
35 when entering in the loop_1 and has a step 2 in this loop, in other
36 words "for (b = 0; b < N; b+=2);". Note that the coefficients of
37 this chain of recurrence (or chrec [shrek]) can contain the name of
38 other variables, in which case they are called parametric chrecs.
39 For example, "b -> {a, +, 2}_1" means that the initial value of "b"
40 is the value of "a". In most of the cases these parametric chrecs
41 are fully instantiated before their use because symbolic names can
42 hide some difficult cases such as self-references described later
43 (see the Fibonacci example).
44
45 A short sketch of the algorithm is:
46
47 Given a scalar variable to be analyzed, follow the SSA edge to
48 its definition:
49
50 - When the definition is a GIMPLE_ASSIGN: if the right hand side
51 (RHS) of the definition cannot be statically analyzed, the answer
52 of the analyzer is: "don't know".
53 Otherwise, for all the variables that are not yet analyzed in the
54 RHS, try to determine their evolution, and finally try to
55 evaluate the operation of the RHS that gives the evolution
56 function of the analyzed variable.
57
58 - When the definition is a condition-phi-node: determine the
59 evolution function for all the branches of the phi node, and
60 finally merge these evolutions (see chrec_merge).
61
62 - When the definition is a loop-phi-node: determine its initial
63 condition, that is the SSA edge defined in an outer loop, and
64 keep it symbolic. Then determine the SSA edges that are defined
65 in the body of the loop. Follow the inner edges until ending on
66 another loop-phi-node of the same analyzed loop. If the reached
67 loop-phi-node is not the starting loop-phi-node, then we keep
68 this definition under a symbolic form. If the reached
69 loop-phi-node is the same as the starting one, then we compute a
70 symbolic stride on the return path. The result is then the
71 symbolic chrec {initial_condition, +, symbolic_stride}_loop.
72
73 Examples:
74
75 Example 1: Illustration of the basic algorithm.
76
77 | a = 3
78 | loop_1
79 | b = phi (a, c)
80 | c = b + 1
81 | if (c > 10) exit_loop
82 | endloop
83
84 Suppose that we want to know the number of iterations of the
85 loop_1. The exit_loop is controlled by a COND_EXPR (c > 10). We
86 ask the scalar evolution analyzer two questions: what's the
87 scalar evolution (scev) of "c", and what's the scev of "10". For
88 "10" the answer is "10" since it is a scalar constant. For the
89 scalar variable "c", it follows the SSA edge to its definition,
90 "c = b + 1", and then asks again what's the scev of "b".
91 Following the SSA edge, we end on a loop-phi-node "b = phi (a,
92 c)", where the initial condition is "a", and the inner loop edge
93 is "c". The initial condition is kept under a symbolic form (it
94 may be the case that the copy constant propagation has done its
95 work and we end with the constant "3" as one of the edges of the
96 loop-phi-node). The update edge is followed to the end of the
97 loop, and until reaching again the starting loop-phi-node: b -> c
98 -> b. At this point we have drawn a path from "b" to "b" from
99 which we compute the stride in the loop: in this example it is
100 "+1". The resulting scev for "b" is "b -> {a, +, 1}_1". Now
101 that the scev for "b" is known, it is possible to compute the
102 scev for "c", that is "c -> {a + 1, +, 1}_1". In order to
103 determine the number of iterations in the loop_1, we have to
104 instantiate_parameters (loop_1, {a + 1, +, 1}_1), that gives after some
105 more analysis the scev {4, +, 1}_1, or in other words, this is
106 the function "f (x) = x + 4", where x is the iteration count of
107 the loop_1. Now we have to solve the inequality "x + 4 > 10",
108 and take the smallest iteration number for which the loop is
109 exited: x = 7. This loop runs from x = 0 to x = 7, and in total
110 there are 8 iterations. In terms of loop normalization, we have
111 created a variable that is implicitly defined, "x" or just "_1",
112 and all the other analyzed scalars of the loop are defined in
113 function of this variable:
114
115 a -> 3
116 b -> {3, +, 1}_1
117 c -> {4, +, 1}_1
118
119 or in terms of a C program:
120
121 | a = 3
122 | for (x = 0; x <= 7; x++)
123 | {
124 | b = x + 3
125 | c = x + 4
126 | }
127
128 Example 2a: Illustration of the algorithm on nested loops.
129
130 | loop_1
131 | a = phi (1, b)
132 | c = a + 2
133 | loop_2 10 times
134 | b = phi (c, d)
135 | d = b + 3
136 | endloop
137 | endloop
138
139 For analyzing the scalar evolution of "a", the algorithm follows
140 the SSA edge into the loop's body: "a -> b". "b" is an inner
141 loop-phi-node, and its analysis as in Example 1, gives:
142
143 b -> {c, +, 3}_2
144 d -> {c + 3, +, 3}_2
145
146 Following the SSA edge for the initial condition, we end on "c = a
147 + 2", and then on the starting loop-phi-node "a". From this point,
148 the loop stride is computed: back on "c = a + 2" we get a "+2" in
149 the loop_1, then on the loop-phi-node "b" we compute the overall
150 effect of the inner loop that is "b = c + 30", and we get a "+30"
151 in the loop_1. That means that the overall stride in loop_1 is
152 equal to "+32", and the result is:
153
154 a -> {1, +, 32}_1
155 c -> {3, +, 32}_1
156
157 Example 2b: Multivariate chains of recurrences.
158
159 | loop_1
160 | k = phi (0, k + 1)
161 | loop_2 4 times
162 | j = phi (0, j + 1)
163 | loop_3 4 times
164 | i = phi (0, i + 1)
165 | A[j + k] = ...
166 | endloop
167 | endloop
168 | endloop
169
170 Analyzing the access function of array A with
171 instantiate_parameters (loop_1, "j + k"), we obtain the
172 instantiation and the analysis of the scalar variables "j" and "k"
173 in loop_1. This leads to the scalar evolution {4, +, 1}_1: the end
174 value of loop_2 for "j" is 4, and the evolution of "k" in loop_1 is
175 {0, +, 1}_1. To obtain the evolution function in loop_3 and
176 instantiate the scalar variables up to loop_1, one has to use:
177 instantiate_scev (block_before_loop (loop_1), loop_3, "j + k").
178 The result of this call is {{0, +, 1}_1, +, 1}_2.
179
180 Example 3: Higher degree polynomials.
181
182 | loop_1
183 | a = phi (2, b)
184 | c = phi (5, d)
185 | b = a + 1
186 | d = c + a
187 | endloop
188
189 a -> {2, +, 1}_1
190 b -> {3, +, 1}_1
191 c -> {5, +, a}_1
192 d -> {5 + a, +, a}_1
193
194 instantiate_parameters (loop_1, {5, +, a}_1) -> {5, +, 2, +, 1}_1
195 instantiate_parameters (loop_1, {5 + a, +, a}_1) -> {7, +, 3, +, 1}_1
196
197 Example 4: Lucas, Fibonacci, or mixers in general.
198
199 | loop_1
200 | a = phi (1, b)
201 | c = phi (3, d)
202 | b = c
203 | d = c + a
204 | endloop
205
206 a -> (1, c)_1
207 c -> {3, +, a}_1
208
209 The syntax "(1, c)_1" stands for a PEELED_CHREC that has the
210 following semantics: during the first iteration of the loop_1, the
211 variable contains the value 1, and then it contains the value "c".
212 Note that this syntax is close to the syntax of the loop-phi-node:
213 "a -> (1, c)_1" vs. "a = phi (1, c)".
214
215 The symbolic chrec representation contains all the semantics of the
216 original code. What is more difficult is to use this information.
217
218 Example 5: Flip-flops, or exchangers.
219
220 | loop_1
221 | a = phi (1, b)
222 | c = phi (3, d)
223 | b = c
224 | d = a
225 | endloop
226
227 a -> (1, c)_1
228 c -> (3, a)_1
229
230 Based on these symbolic chrecs, it is possible to refine this
231 information into the more precise PERIODIC_CHRECs:
232
233 a -> |1, 3|_1
234 c -> |3, 1|_1
235
236 This transformation is not yet implemented.
237
238 Further readings:
239
240 You can find a more detailed description of the algorithm in:
241 http://icps.u-strasbg.fr/~pop/DEA_03_Pop.pdf
242 http://icps.u-strasbg.fr/~pop/DEA_03_Pop.ps.gz. But note that
243 this is a preliminary report and some of the details of the
244 algorithm have changed. I'm working on a research report that
245 updates the description of the algorithms to reflect the design
246 choices used in this implementation.
247
248 A set of slides show a high level overview of the algorithm and run
249 an example through the scalar evolution analyzer:
250 http://cri.ensmp.fr/~pop/gcc/mar04/slides.pdf
251
252 The slides that I have presented at the GCC Summit'04 are available
253 at: http://cri.ensmp.fr/~pop/gcc/20040604/gccsummit-lno-spop.pdf
254*/
255
256#include "config.h"
257#include "system.h"
258#include "coretypes.h"
259#include "backend.h"
260#include "target.h"
261#include "rtl.h"
262#include "optabs-query.h"
263#include "tree.h"
264#include "gimple.h"
265#include "ssa.h"
266#include "gimple-pretty-print.h"
267#include "fold-const.h"
268#include "gimplify.h"
269#include "gimple-iterator.h"
270#include "gimplify-me.h"
271#include "tree-cfg.h"
272#include "tree-ssa-loop-ivopts.h"
273#include "tree-ssa-loop-manip.h"
274#include "tree-ssa-loop-niter.h"
275#include "tree-ssa-loop.h"
276#include "tree-ssa.h"
277#include "cfgloop.h"
278#include "tree-chrec.h"
279#include "tree-affine.h"
280#include "tree-scalar-evolution.h"
281#include "dumpfile.h"
282#include "tree-ssa-propagate.h"
283#include "gimple-fold.h"
284#include "tree-into-ssa.h"
285#include "builtins.h"
286#include "case-cfn-macros.h"
287
288static tree analyze_scalar_evolution_1 (class loop *, tree);
289static tree analyze_scalar_evolution_for_address_of (class loop *loop,
290 tree var);
291
292/* The cached information about an SSA name with version NAME_VERSION,
293 claiming that below basic block with index INSTANTIATED_BELOW, the
294 value of the SSA name can be expressed as CHREC. */
295
296struct GTY((for_user)) scev_info_str {
297 unsigned int name_version;
298 int instantiated_below;
299 tree chrec;
300};
301
302/* Counters for the scev database. */
303static unsigned nb_set_scev = 0;
304static unsigned nb_get_scev = 0;
305
306struct scev_info_hasher : ggc_ptr_hash<scev_info_str>
307{
308 static hashval_t hash (scev_info_str *i);
309 static bool equal (const scev_info_str *a, const scev_info_str *b);
310};
311
312static GTY (()) hash_table<scev_info_hasher> *scalar_evolution_info;
313
314
315/* Constructs a new SCEV_INFO_STR structure for VAR and INSTANTIATED_BELOW. */
316
317static inline struct scev_info_str *
318new_scev_info_str (basic_block instantiated_below, tree var)
319{
320 struct scev_info_str *res;
321
322 res = ggc_alloc<scev_info_str> ();
323 res->name_version = SSA_NAME_VERSION (var);
324 res->chrec = chrec_not_analyzed_yet;
325 res->instantiated_below = instantiated_below->index;
326
327 return res;
328}
329
330/* Computes a hash function for database element ELT. */
331
332hashval_t
333scev_info_hasher::hash (scev_info_str *elt)
334{
335 return elt->name_version ^ elt->instantiated_below;
336}
337
338/* Compares database elements E1 and E2. */
339
340bool
341scev_info_hasher::equal (const scev_info_str *elt1, const scev_info_str *elt2)
342{
343 return (elt1->name_version == elt2->name_version
344 && elt1->instantiated_below == elt2->instantiated_below);
345}
346
347/* Get the scalar evolution of VAR for INSTANTIATED_BELOW basic block.
348 A first query on VAR returns chrec_not_analyzed_yet. */
349
350static tree *
351find_var_scev_info (basic_block instantiated_below, tree var)
352{
353 struct scev_info_str *res;
354 struct scev_info_str tmp;
355
356 tmp.name_version = SSA_NAME_VERSION (var);
357 tmp.instantiated_below = instantiated_below->index;
358 scev_info_str **slot = scalar_evolution_info->find_slot (value: &tmp, insert: INSERT);
359
360 if (!*slot)
361 *slot = new_scev_info_str (instantiated_below, var);
362 res = *slot;
363
364 return &res->chrec;
365}
366
367
368/* Hashtable helpers for a temporary hash-table used when
369 analyzing a scalar evolution, instantiating a CHREC or
370 resolving mixers. */
371
372class instantiate_cache_type
373{
374public:
375 htab_t map;
376 vec<scev_info_str> entries;
377
378 instantiate_cache_type () : map (NULL), entries (vNULL) {}
379 ~instantiate_cache_type ();
380 tree get (unsigned slot) { return entries[slot].chrec; }
381 void set (unsigned slot, tree chrec) { entries[slot].chrec = chrec; }
382};
383
384instantiate_cache_type::~instantiate_cache_type ()
385{
386 if (map != NULL)
387 {
388 htab_delete (map);
389 entries.release ();
390 }
391}
392
393/* Cache to avoid infinite recursion when instantiating an SSA name.
394 Live during the outermost analyze_scalar_evolution, instantiate_scev
395 or resolve_mixers call. */
396static instantiate_cache_type *global_cache;
397
398
399/* Return true when PHI is a loop-phi-node. */
400
401static bool
402loop_phi_node_p (gimple *phi)
403{
404 /* The implementation of this function is based on the following
405 property: "all the loop-phi-nodes of a loop are contained in the
406 loop's header basic block". */
407
408 return loop_containing_stmt (stmt: phi)->header == gimple_bb (g: phi);
409}
410
411/* Compute the scalar evolution for EVOLUTION_FN after crossing LOOP.
412 In general, in the case of multivariate evolutions we want to get
413 the evolution in different loops. LOOP specifies the level for
414 which to get the evolution.
415
416 Example:
417
418 | for (j = 0; j < 100; j++)
419 | {
420 | for (k = 0; k < 100; k++)
421 | {
422 | i = k + j; - Here the value of i is a function of j, k.
423 | }
424 | ... = i - Here the value of i is a function of j.
425 | }
426 | ... = i - Here the value of i is a scalar.
427
428 Example:
429
430 | i_0 = ...
431 | loop_1 10 times
432 | i_1 = phi (i_0, i_2)
433 | i_2 = i_1 + 2
434 | endloop
435
436 This loop has the same effect as:
437 LOOP_1 has the same effect as:
438
439 | i_1 = i_0 + 20
440
441 The overall effect of the loop, "i_0 + 20" in the previous example,
442 is obtained by passing in the parameters: LOOP = 1,
443 EVOLUTION_FN = {i_0, +, 2}_1.
444*/
445
446tree
447compute_overall_effect_of_inner_loop (class loop *loop, tree evolution_fn)
448{
449 bool val = false;
450
451 if (evolution_fn == chrec_dont_know)
452 return chrec_dont_know;
453
454 else if (TREE_CODE (evolution_fn) == POLYNOMIAL_CHREC)
455 {
456 class loop *inner_loop = get_chrec_loop (chrec: evolution_fn);
457
458 if (inner_loop == loop
459 || flow_loop_nested_p (loop, inner_loop))
460 {
461 tree nb_iter = number_of_latch_executions (inner_loop);
462
463 if (nb_iter == chrec_dont_know)
464 return chrec_dont_know;
465 else
466 {
467 tree res;
468
469 /* evolution_fn is the evolution function in LOOP. Get
470 its value in the nb_iter-th iteration. */
471 res = chrec_apply (inner_loop->num, evolution_fn, nb_iter);
472
473 if (chrec_contains_symbols_defined_in_loop (res, loop->num))
474 res = instantiate_parameters (loop, chrec: res);
475
476 /* Continue the computation until ending on a parent of LOOP. */
477 return compute_overall_effect_of_inner_loop (loop, evolution_fn: res);
478 }
479 }
480 else
481 return evolution_fn;
482 }
483
484 /* If the evolution function is an invariant, there is nothing to do. */
485 else if (no_evolution_in_loop_p (chrec: evolution_fn, loop_num: loop->num, res: &val) && val)
486 return evolution_fn;
487
488 else
489 return chrec_dont_know;
490}
491
492/* Associate CHREC to SCALAR. */
493
494static void
495set_scalar_evolution (basic_block instantiated_below, tree scalar, tree chrec)
496{
497 tree *scalar_info;
498
499 if (TREE_CODE (scalar) != SSA_NAME)
500 return;
501
502 scalar_info = find_var_scev_info (instantiated_below, var: scalar);
503
504 if (dump_file)
505 {
506 if (dump_flags & TDF_SCEV)
507 {
508 fprintf (stream: dump_file, format: "(set_scalar_evolution \n");
509 fprintf (stream: dump_file, format: " instantiated_below = %d \n",
510 instantiated_below->index);
511 fprintf (stream: dump_file, format: " (scalar = ");
512 print_generic_expr (dump_file, scalar);
513 fprintf (stream: dump_file, format: ")\n (scalar_evolution = ");
514 print_generic_expr (dump_file, chrec);
515 fprintf (stream: dump_file, format: "))\n");
516 }
517 if (dump_flags & TDF_STATS)
518 nb_set_scev++;
519 }
520
521 *scalar_info = chrec;
522}
523
524/* Retrieve the chrec associated to SCALAR instantiated below
525 INSTANTIATED_BELOW block. */
526
527static tree
528get_scalar_evolution (basic_block instantiated_below, tree scalar)
529{
530 tree res;
531
532 if (dump_file)
533 {
534 if (dump_flags & TDF_SCEV)
535 {
536 fprintf (stream: dump_file, format: "(get_scalar_evolution \n");
537 fprintf (stream: dump_file, format: " (scalar = ");
538 print_generic_expr (dump_file, scalar);
539 fprintf (stream: dump_file, format: ")\n");
540 }
541 if (dump_flags & TDF_STATS)
542 nb_get_scev++;
543 }
544
545 if (VECTOR_TYPE_P (TREE_TYPE (scalar))
546 || TREE_CODE (TREE_TYPE (scalar)) == COMPLEX_TYPE)
547 /* For chrec_dont_know we keep the symbolic form. */
548 res = scalar;
549 else
550 switch (TREE_CODE (scalar))
551 {
552 case SSA_NAME:
553 if (SSA_NAME_IS_DEFAULT_DEF (scalar))
554 res = scalar;
555 else
556 res = *find_var_scev_info (instantiated_below, var: scalar);
557 break;
558
559 case REAL_CST:
560 case FIXED_CST:
561 case INTEGER_CST:
562 res = scalar;
563 break;
564
565 default:
566 res = chrec_not_analyzed_yet;
567 break;
568 }
569
570 if (dump_file && (dump_flags & TDF_SCEV))
571 {
572 fprintf (stream: dump_file, format: " (scalar_evolution = ");
573 print_generic_expr (dump_file, res);
574 fprintf (stream: dump_file, format: "))\n");
575 }
576
577 return res;
578}
579
580
581/* Depth first search algorithm. */
582
583enum t_bool {
584 t_false,
585 t_true,
586 t_dont_know
587};
588
589class scev_dfs
590{
591public:
592 scev_dfs (class loop *loop_, gphi *phi_, tree init_cond_)
593 : loop (loop_), loop_phi_node (phi_), init_cond (init_cond_) {}
594 t_bool get_ev (tree *, tree);
595
596private:
597 t_bool follow_ssa_edge_expr (gimple *, tree, tree *, int);
598 t_bool follow_ssa_edge_binary (gimple *at_stmt,
599 tree type, tree rhs0, enum tree_code code,
600 tree rhs1, tree *evolution_of_loop, int limit);
601 t_bool follow_ssa_edge_in_condition_phi_branch (int i,
602 gphi *condition_phi,
603 tree *evolution_of_branch,
604 tree init_cond, int limit);
605 t_bool follow_ssa_edge_in_condition_phi (gphi *condition_phi,
606 tree *evolution_of_loop, int limit);
607 t_bool follow_ssa_edge_inner_loop_phi (gphi *loop_phi_node,
608 tree *evolution_of_loop, int limit);
609 tree add_to_evolution (tree chrec_before, enum tree_code code,
610 tree to_add, gimple *at_stmt);
611 tree add_to_evolution_1 (tree chrec_before, tree to_add, gimple *at_stmt);
612
613 class loop *loop;
614 gphi *loop_phi_node;
615 tree init_cond;
616};
617
618t_bool
619scev_dfs::get_ev (tree *ev_fn, tree arg)
620{
621 *ev_fn = chrec_dont_know;
622 return follow_ssa_edge_expr (loop_phi_node, arg, ev_fn, 0);
623}
624
625/* Helper function for add_to_evolution. Returns the evolution
626 function for an assignment of the form "a = b + c", where "a" and
627 "b" are on the strongly connected component. CHREC_BEFORE is the
628 information that we already have collected up to this point.
629 TO_ADD is the evolution of "c".
630
631 When CHREC_BEFORE has an evolution part in LOOP_NB, add to this
632 evolution the expression TO_ADD, otherwise construct an evolution
633 part for this loop. */
634
635tree
636scev_dfs::add_to_evolution_1 (tree chrec_before, tree to_add, gimple *at_stmt)
637{
638 tree type, left, right;
639 unsigned loop_nb = loop->num;
640 class loop *chloop;
641
642 switch (TREE_CODE (chrec_before))
643 {
644 case POLYNOMIAL_CHREC:
645 chloop = get_chrec_loop (chrec: chrec_before);
646 if (chloop == loop
647 || flow_loop_nested_p (chloop, loop))
648 {
649 unsigned var;
650
651 type = chrec_type (chrec: chrec_before);
652
653 /* When there is no evolution part in this loop, build it. */
654 if (chloop != loop)
655 {
656 var = loop_nb;
657 left = chrec_before;
658 right = SCALAR_FLOAT_TYPE_P (type)
659 ? build_real (type, dconst0)
660 : build_int_cst (type, 0);
661 }
662 else
663 {
664 var = CHREC_VARIABLE (chrec_before);
665 left = CHREC_LEFT (chrec_before);
666 right = CHREC_RIGHT (chrec_before);
667 }
668
669 to_add = chrec_convert (type, to_add, at_stmt);
670 right = chrec_convert_rhs (type, right, at_stmt);
671 right = chrec_fold_plus (chrec_type (chrec: right), right, to_add);
672 return build_polynomial_chrec (loop_num: var, left, right);
673 }
674 else
675 {
676 gcc_assert (flow_loop_nested_p (loop, chloop));
677
678 /* Search the evolution in LOOP_NB. */
679 left = add_to_evolution_1 (CHREC_LEFT (chrec_before),
680 to_add, at_stmt);
681 right = CHREC_RIGHT (chrec_before);
682 right = chrec_convert_rhs (chrec_type (chrec: left), right, at_stmt);
683 return build_polynomial_chrec (CHREC_VARIABLE (chrec_before),
684 left, right);
685 }
686
687 default:
688 /* These nodes do not depend on a loop. */
689 if (chrec_before == chrec_dont_know)
690 return chrec_dont_know;
691
692 left = chrec_before;
693 right = chrec_convert_rhs (chrec_type (chrec: left), to_add, at_stmt);
694 /* When we add the first evolution we need to replace the symbolic
695 evolution we've put in when the DFS reached the loop PHI node
696 with the initial value. There's only a limited cases of
697 extra operations ontop of that symbol allowed, namely
698 sign-conversions we can look through. For other cases we leave
699 the symbolic initial condition which causes build_polynomial_chrec
700 to return chrec_dont_know. See PR42512, PR66375 and PR107176 for
701 cases we mishandled before. */
702 STRIP_NOPS (chrec_before);
703 if (chrec_before == gimple_phi_result (gs: loop_phi_node))
704 left = fold_convert (TREE_TYPE (left), init_cond);
705 return build_polynomial_chrec (loop_num: loop_nb, left, right);
706 }
707}
708
709/* Add TO_ADD to the evolution part of CHREC_BEFORE in the dimension
710 of LOOP_NB.
711
712 Description (provided for completeness, for those who read code in
713 a plane, and for my poor 62 bytes brain that would have forgotten
714 all this in the next two or three months):
715
716 The algorithm of translation of programs from the SSA representation
717 into the chrecs syntax is based on a pattern matching. After having
718 reconstructed the overall tree expression for a loop, there are only
719 two cases that can arise:
720
721 1. a = loop-phi (init, a + expr)
722 2. a = loop-phi (init, expr)
723
724 where EXPR is either a scalar constant with respect to the analyzed
725 loop (this is a degree 0 polynomial), or an expression containing
726 other loop-phi definitions (these are higher degree polynomials).
727
728 Examples:
729
730 1.
731 | init = ...
732 | loop_1
733 | a = phi (init, a + 5)
734 | endloop
735
736 2.
737 | inita = ...
738 | initb = ...
739 | loop_1
740 | a = phi (inita, 2 * b + 3)
741 | b = phi (initb, b + 1)
742 | endloop
743
744 For the first case, the semantics of the SSA representation is:
745
746 | a (x) = init + \sum_{j = 0}^{x - 1} expr (j)
747
748 that is, there is a loop index "x" that determines the scalar value
749 of the variable during the loop execution. During the first
750 iteration, the value is that of the initial condition INIT, while
751 during the subsequent iterations, it is the sum of the initial
752 condition with the sum of all the values of EXPR from the initial
753 iteration to the before last considered iteration.
754
755 For the second case, the semantics of the SSA program is:
756
757 | a (x) = init, if x = 0;
758 | expr (x - 1), otherwise.
759
760 The second case corresponds to the PEELED_CHREC, whose syntax is
761 close to the syntax of a loop-phi-node:
762
763 | phi (init, expr) vs. (init, expr)_x
764
765 The proof of the translation algorithm for the first case is a
766 proof by structural induction based on the degree of EXPR.
767
768 Degree 0:
769 When EXPR is a constant with respect to the analyzed loop, or in
770 other words when EXPR is a polynomial of degree 0, the evolution of
771 the variable A in the loop is an affine function with an initial
772 condition INIT, and a step EXPR. In order to show this, we start
773 from the semantics of the SSA representation:
774
775 f (x) = init + \sum_{j = 0}^{x - 1} expr (j)
776
777 and since "expr (j)" is a constant with respect to "j",
778
779 f (x) = init + x * expr
780
781 Finally, based on the semantics of the pure sum chrecs, by
782 identification we get the corresponding chrecs syntax:
783
784 f (x) = init * \binom{x}{0} + expr * \binom{x}{1}
785 f (x) -> {init, +, expr}_x
786
787 Higher degree:
788 Suppose that EXPR is a polynomial of degree N with respect to the
789 analyzed loop_x for which we have already determined that it is
790 written under the chrecs syntax:
791
792 | expr (x) -> {b_0, +, b_1, +, ..., +, b_{n-1}} (x)
793
794 We start from the semantics of the SSA program:
795
796 | f (x) = init + \sum_{j = 0}^{x - 1} expr (j)
797 |
798 | f (x) = init + \sum_{j = 0}^{x - 1}
799 | (b_0 * \binom{j}{0} + ... + b_{n-1} * \binom{j}{n-1})
800 |
801 | f (x) = init + \sum_{j = 0}^{x - 1}
802 | \sum_{k = 0}^{n - 1} (b_k * \binom{j}{k})
803 |
804 | f (x) = init + \sum_{k = 0}^{n - 1}
805 | (b_k * \sum_{j = 0}^{x - 1} \binom{j}{k})
806 |
807 | f (x) = init + \sum_{k = 0}^{n - 1}
808 | (b_k * \binom{x}{k + 1})
809 |
810 | f (x) = init + b_0 * \binom{x}{1} + ...
811 | + b_{n-1} * \binom{x}{n}
812 |
813 | f (x) = init * \binom{x}{0} + b_0 * \binom{x}{1} + ...
814 | + b_{n-1} * \binom{x}{n}
815 |
816
817 And finally from the definition of the chrecs syntax, we identify:
818 | f (x) -> {init, +, b_0, +, ..., +, b_{n-1}}_x
819
820 This shows the mechanism that stands behind the add_to_evolution
821 function. An important point is that the use of symbolic
822 parameters avoids the need of an analysis schedule.
823
824 Example:
825
826 | inita = ...
827 | initb = ...
828 | loop_1
829 | a = phi (inita, a + 2 + b)
830 | b = phi (initb, b + 1)
831 | endloop
832
833 When analyzing "a", the algorithm keeps "b" symbolically:
834
835 | a -> {inita, +, 2 + b}_1
836
837 Then, after instantiation, the analyzer ends on the evolution:
838
839 | a -> {inita, +, 2 + initb, +, 1}_1
840
841*/
842
843tree
844scev_dfs::add_to_evolution (tree chrec_before, enum tree_code code,
845 tree to_add, gimple *at_stmt)
846{
847 tree type = chrec_type (chrec: to_add);
848 tree res = NULL_TREE;
849
850 if (to_add == NULL_TREE)
851 return chrec_before;
852
853 /* TO_ADD is either a scalar, or a parameter. TO_ADD is not
854 instantiated at this point. */
855 if (TREE_CODE (to_add) == POLYNOMIAL_CHREC)
856 /* This should not happen. */
857 return chrec_dont_know;
858
859 if (dump_file && (dump_flags & TDF_SCEV))
860 {
861 fprintf (stream: dump_file, format: "(add_to_evolution \n");
862 fprintf (stream: dump_file, format: " (loop_nb = %d)\n", loop->num);
863 fprintf (stream: dump_file, format: " (chrec_before = ");
864 print_generic_expr (dump_file, chrec_before);
865 fprintf (stream: dump_file, format: ")\n (to_add = ");
866 print_generic_expr (dump_file, to_add);
867 fprintf (stream: dump_file, format: ")\n");
868 }
869
870 if (code == MINUS_EXPR)
871 to_add = chrec_fold_multiply (type, to_add, SCALAR_FLOAT_TYPE_P (type)
872 ? build_real (type, dconstm1)
873 : build_int_cst_type (type, -1));
874
875 res = add_to_evolution_1 (chrec_before, to_add, at_stmt);
876
877 if (dump_file && (dump_flags & TDF_SCEV))
878 {
879 fprintf (stream: dump_file, format: " (res = ");
880 print_generic_expr (dump_file, res);
881 fprintf (stream: dump_file, format: "))\n");
882 }
883
884 return res;
885}
886
887
888/* Follow the ssa edge into the binary expression RHS0 CODE RHS1.
889 Return true if the strongly connected component has been found. */
890
891t_bool
892scev_dfs::follow_ssa_edge_binary (gimple *at_stmt, tree type, tree rhs0,
893 enum tree_code code, tree rhs1,
894 tree *evolution_of_loop, int limit)
895{
896 t_bool res = t_false;
897 tree evol;
898
899 switch (code)
900 {
901 case POINTER_PLUS_EXPR:
902 case PLUS_EXPR:
903 if (TREE_CODE (rhs0) == SSA_NAME)
904 {
905 if (TREE_CODE (rhs1) == SSA_NAME)
906 {
907 /* Match an assignment under the form:
908 "a = b + c". */
909
910 /* We want only assignments of form "name + name" contribute to
911 LIMIT, as the other cases do not necessarily contribute to
912 the complexity of the expression. */
913 limit++;
914
915 evol = *evolution_of_loop;
916 res = follow_ssa_edge_expr (at_stmt, rhs0, &evol, limit);
917 if (res == t_true)
918 *evolution_of_loop = add_to_evolution
919 (chrec_before: chrec_convert (type, evol, at_stmt), code, to_add: rhs1, at_stmt);
920 else if (res == t_false)
921 {
922 res = follow_ssa_edge_expr
923 (at_stmt, rhs1, evolution_of_loop, limit);
924 if (res == t_true)
925 *evolution_of_loop = add_to_evolution
926 (chrec_before: chrec_convert (type, *evolution_of_loop, at_stmt),
927 code, to_add: rhs0, at_stmt);
928 }
929 }
930
931 else
932 gcc_unreachable (); /* Handled in caller. */
933 }
934
935 else if (TREE_CODE (rhs1) == SSA_NAME)
936 {
937 /* Match an assignment under the form:
938 "a = ... + c". */
939 res = follow_ssa_edge_expr (at_stmt, rhs1, evolution_of_loop, limit);
940 if (res == t_true)
941 *evolution_of_loop = add_to_evolution
942 (chrec_before: chrec_convert (type, *evolution_of_loop, at_stmt),
943 code, to_add: rhs0, at_stmt);
944 }
945
946 else
947 /* Otherwise, match an assignment under the form:
948 "a = ... + ...". */
949 /* And there is nothing to do. */
950 res = t_false;
951 break;
952
953 case MINUS_EXPR:
954 /* This case is under the form "opnd0 = rhs0 - rhs1". */
955 if (TREE_CODE (rhs0) == SSA_NAME)
956 gcc_unreachable (); /* Handled in caller. */
957 else
958 /* Otherwise, match an assignment under the form:
959 "a = ... - ...". */
960 /* And there is nothing to do. */
961 res = t_false;
962 break;
963
964 default:
965 res = t_false;
966 }
967
968 return res;
969}
970
971/* Checks whether the I-th argument of a PHI comes from a backedge. */
972
973static bool
974backedge_phi_arg_p (gphi *phi, int i)
975{
976 const_edge e = gimple_phi_arg_edge (phi, i);
977
978 /* We would in fact like to test EDGE_DFS_BACK here, but we do not care
979 about updating it anywhere, and this should work as well most of the
980 time. */
981 if (e->flags & EDGE_IRREDUCIBLE_LOOP)
982 return true;
983
984 return false;
985}
986
987/* Helper function for one branch of the condition-phi-node. Return
988 true if the strongly connected component has been found following
989 this path. */
990
991t_bool
992scev_dfs::follow_ssa_edge_in_condition_phi_branch (int i,
993 gphi *condition_phi,
994 tree *evolution_of_branch,
995 tree init_cond, int limit)
996{
997 tree branch = PHI_ARG_DEF (condition_phi, i);
998 *evolution_of_branch = chrec_dont_know;
999
1000 /* Do not follow back edges (they must belong to an irreducible loop, which
1001 we really do not want to worry about). */
1002 if (backedge_phi_arg_p (phi: condition_phi, i))
1003 return t_false;
1004
1005 if (TREE_CODE (branch) == SSA_NAME)
1006 {
1007 *evolution_of_branch = init_cond;
1008 return follow_ssa_edge_expr (condition_phi, branch,
1009 evolution_of_branch, limit);
1010 }
1011
1012 /* This case occurs when one of the condition branches sets
1013 the variable to a constant: i.e. a phi-node like
1014 "a_2 = PHI <a_7(5), 2(6)>;".
1015
1016 FIXME: This case have to be refined correctly:
1017 in some cases it is possible to say something better than
1018 chrec_dont_know, for example using a wrap-around notation. */
1019 return t_false;
1020}
1021
1022/* This function merges the branches of a condition-phi-node in a
1023 loop. */
1024
1025t_bool
1026scev_dfs::follow_ssa_edge_in_condition_phi (gphi *condition_phi,
1027 tree *evolution_of_loop, int limit)
1028{
1029 int i, n;
1030 tree init = *evolution_of_loop;
1031 tree evolution_of_branch;
1032 t_bool res = follow_ssa_edge_in_condition_phi_branch (i: 0, condition_phi,
1033 evolution_of_branch: &evolution_of_branch,
1034 init_cond: init, limit);
1035 if (res == t_false || res == t_dont_know)
1036 return res;
1037
1038 *evolution_of_loop = evolution_of_branch;
1039
1040 n = gimple_phi_num_args (gs: condition_phi);
1041 for (i = 1; i < n; i++)
1042 {
1043 /* Quickly give up when the evolution of one of the branches is
1044 not known. */
1045 if (*evolution_of_loop == chrec_dont_know)
1046 return t_true;
1047
1048 /* Increase the limit by the PHI argument number to avoid exponential
1049 time and memory complexity. */
1050 res = follow_ssa_edge_in_condition_phi_branch (i, condition_phi,
1051 evolution_of_branch: &evolution_of_branch,
1052 init_cond: init, limit: limit + i);
1053 if (res == t_false || res == t_dont_know)
1054 return res;
1055
1056 *evolution_of_loop = chrec_merge (*evolution_of_loop,
1057 evolution_of_branch);
1058 }
1059
1060 return t_true;
1061}
1062
1063/* Follow an SSA edge in an inner loop. It computes the overall
1064 effect of the loop, and following the symbolic initial conditions,
1065 it follows the edges in the parent loop. The inner loop is
1066 considered as a single statement. */
1067
1068t_bool
1069scev_dfs::follow_ssa_edge_inner_loop_phi (gphi *loop_phi_node,
1070 tree *evolution_of_loop, int limit)
1071{
1072 class loop *loop = loop_containing_stmt (stmt: loop_phi_node);
1073 tree ev = analyze_scalar_evolution (loop, PHI_RESULT (loop_phi_node));
1074
1075 /* Sometimes, the inner loop is too difficult to analyze, and the
1076 result of the analysis is a symbolic parameter. */
1077 if (ev == PHI_RESULT (loop_phi_node))
1078 {
1079 t_bool res = t_false;
1080 int i, n = gimple_phi_num_args (gs: loop_phi_node);
1081
1082 for (i = 0; i < n; i++)
1083 {
1084 tree arg = PHI_ARG_DEF (loop_phi_node, i);
1085 basic_block bb;
1086
1087 /* Follow the edges that exit the inner loop. */
1088 bb = gimple_phi_arg_edge (phi: loop_phi_node, i)->src;
1089 if (!flow_bb_inside_loop_p (loop, bb))
1090 res = follow_ssa_edge_expr (loop_phi_node,
1091 arg, evolution_of_loop, limit);
1092 if (res == t_true)
1093 break;
1094 }
1095
1096 /* If the path crosses this loop-phi, give up. */
1097 if (res == t_true)
1098 *evolution_of_loop = chrec_dont_know;
1099
1100 return res;
1101 }
1102
1103 /* Otherwise, compute the overall effect of the inner loop. */
1104 ev = compute_overall_effect_of_inner_loop (loop, evolution_fn: ev);
1105 return follow_ssa_edge_expr (loop_phi_node, ev, evolution_of_loop, limit);
1106}
1107
1108/* Follow the ssa edge into the expression EXPR.
1109 Return true if the strongly connected component has been found. */
1110
1111t_bool
1112scev_dfs::follow_ssa_edge_expr (gimple *at_stmt, tree expr,
1113 tree *evolution_of_loop, int limit)
1114{
1115 gphi *halting_phi = loop_phi_node;
1116 enum tree_code code;
1117 tree type, rhs0, rhs1 = NULL_TREE;
1118
1119 /* The EXPR is one of the following cases:
1120 - an SSA_NAME,
1121 - an INTEGER_CST,
1122 - a PLUS_EXPR,
1123 - a POINTER_PLUS_EXPR,
1124 - a MINUS_EXPR,
1125 - other cases are not yet handled. */
1126
1127 /* For SSA_NAME look at the definition statement, handling
1128 PHI nodes and otherwise expand appropriately for the expression
1129 handling below. */
1130 if (TREE_CODE (expr) == SSA_NAME)
1131 {
1132 gimple *def = SSA_NAME_DEF_STMT (expr);
1133
1134 if (gimple_nop_p (g: def))
1135 return t_false;
1136
1137 /* Give up if the path is longer than the MAX that we allow. */
1138 if (limit > param_scev_max_expr_complexity)
1139 {
1140 *evolution_of_loop = chrec_dont_know;
1141 return t_dont_know;
1142 }
1143
1144 if (gphi *phi = dyn_cast <gphi *>(p: def))
1145 {
1146 if (!loop_phi_node_p (phi))
1147 /* DEF is a condition-phi-node. Follow the branches, and
1148 record their evolutions. Finally, merge the collected
1149 information and set the approximation to the main
1150 variable. */
1151 return follow_ssa_edge_in_condition_phi (condition_phi: phi, evolution_of_loop,
1152 limit);
1153
1154 /* When the analyzed phi is the halting_phi, the
1155 depth-first search is over: we have found a path from
1156 the halting_phi to itself in the loop. */
1157 if (phi == halting_phi)
1158 {
1159 *evolution_of_loop = expr;
1160 return t_true;
1161 }
1162
1163 /* Otherwise, the evolution of the HALTING_PHI depends
1164 on the evolution of another loop-phi-node, i.e. the
1165 evolution function is a higher degree polynomial. */
1166 class loop *def_loop = loop_containing_stmt (stmt: def);
1167 if (def_loop == loop)
1168 return t_false;
1169
1170 /* Inner loop. */
1171 if (flow_loop_nested_p (loop, def_loop))
1172 return follow_ssa_edge_inner_loop_phi (loop_phi_node: phi, evolution_of_loop,
1173 limit: limit + 1);
1174
1175 /* Outer loop. */
1176 return t_false;
1177 }
1178
1179 /* At this level of abstraction, the program is just a set
1180 of GIMPLE_ASSIGNs and PHI_NODEs. In principle there is no
1181 other def to be handled. */
1182 if (!is_gimple_assign (gs: def))
1183 return t_false;
1184
1185 code = gimple_assign_rhs_code (gs: def);
1186 switch (get_gimple_rhs_class (code))
1187 {
1188 case GIMPLE_BINARY_RHS:
1189 rhs0 = gimple_assign_rhs1 (gs: def);
1190 rhs1 = gimple_assign_rhs2 (gs: def);
1191 break;
1192 case GIMPLE_UNARY_RHS:
1193 case GIMPLE_SINGLE_RHS:
1194 rhs0 = gimple_assign_rhs1 (gs: def);
1195 break;
1196 default:
1197 return t_false;
1198 }
1199 type = TREE_TYPE (gimple_assign_lhs (def));
1200 at_stmt = def;
1201 }
1202 else
1203 {
1204 code = TREE_CODE (expr);
1205 type = TREE_TYPE (expr);
1206 /* Via follow_ssa_edge_inner_loop_phi we arrive here with the
1207 GENERIC scalar evolution of the inner loop. */
1208 switch (code)
1209 {
1210 CASE_CONVERT:
1211 rhs0 = TREE_OPERAND (expr, 0);
1212 break;
1213 case POINTER_PLUS_EXPR:
1214 case PLUS_EXPR:
1215 case MINUS_EXPR:
1216 rhs0 = TREE_OPERAND (expr, 0);
1217 rhs1 = TREE_OPERAND (expr, 1);
1218 STRIP_USELESS_TYPE_CONVERSION (rhs0);
1219 STRIP_USELESS_TYPE_CONVERSION (rhs1);
1220 break;
1221 default:
1222 rhs0 = expr;
1223 }
1224 }
1225
1226 switch (code)
1227 {
1228 CASE_CONVERT:
1229 {
1230 /* This assignment is under the form "a_1 = (cast) rhs. We cannot
1231 validate any precision altering conversion during the SCC
1232 analysis, so don't even try. */
1233 if (!tree_nop_conversion_p (type, TREE_TYPE (rhs0)))
1234 return t_false;
1235 t_bool res = follow_ssa_edge_expr (at_stmt, expr: rhs0,
1236 evolution_of_loop, limit);
1237 if (res == t_true)
1238 *evolution_of_loop = chrec_convert (type, *evolution_of_loop,
1239 at_stmt);
1240 return res;
1241 }
1242
1243 case INTEGER_CST:
1244 /* This assignment is under the form "a_1 = 7". */
1245 return t_false;
1246
1247 case ADDR_EXPR:
1248 {
1249 /* Handle &MEM[ptr + CST] which is equivalent to POINTER_PLUS_EXPR. */
1250 if (TREE_CODE (TREE_OPERAND (rhs0, 0)) != MEM_REF)
1251 return t_false;
1252 tree mem = TREE_OPERAND (rhs0, 0);
1253 rhs0 = TREE_OPERAND (mem, 0);
1254 rhs1 = TREE_OPERAND (mem, 1);
1255 code = POINTER_PLUS_EXPR;
1256 }
1257 /* Fallthru. */
1258 case POINTER_PLUS_EXPR:
1259 case PLUS_EXPR:
1260 case MINUS_EXPR:
1261 /* This case is under the form "rhs0 +- rhs1". */
1262 if (TREE_CODE (rhs0) == SSA_NAME
1263 && (TREE_CODE (rhs1) != SSA_NAME || code == MINUS_EXPR))
1264 {
1265 /* Match an assignment under the form:
1266 "a = b +- ...". */
1267 t_bool res = follow_ssa_edge_expr (at_stmt, expr: rhs0,
1268 evolution_of_loop, limit);
1269 if (res == t_true)
1270 *evolution_of_loop = add_to_evolution
1271 (chrec_before: chrec_convert (type, *evolution_of_loop, at_stmt),
1272 code, to_add: rhs1, at_stmt);
1273 return res;
1274 }
1275 /* Else search for the SCC in both rhs0 and rhs1. */
1276 return follow_ssa_edge_binary (at_stmt, type, rhs0, code, rhs1,
1277 evolution_of_loop, limit);
1278
1279 default:
1280 return t_false;
1281 }
1282}
1283
1284
1285/* This section selects the loops that will be good candidates for the
1286 scalar evolution analysis. For the moment, greedily select all the
1287 loop nests we could analyze. */
1288
1289/* For a loop with a single exit edge, return the COND_EXPR that
1290 guards the exit edge. If the expression is too difficult to
1291 analyze, then give up. */
1292
1293gcond *
1294get_loop_exit_condition (const class loop *loop)
1295{
1296 return get_loop_exit_condition (single_exit (loop));
1297}
1298
1299/* If the statement just before the EXIT_EDGE contains a condition then
1300 return the condition, otherwise NULL. */
1301
1302gcond *
1303get_loop_exit_condition (const_edge exit_edge)
1304{
1305 gcond *res = NULL;
1306
1307 if (dump_file && (dump_flags & TDF_SCEV))
1308 fprintf (stream: dump_file, format: "(get_loop_exit_condition \n ");
1309
1310 if (exit_edge)
1311 res = safe_dyn_cast <gcond *> (p: *gsi_last_bb (bb: exit_edge->src));
1312
1313 if (dump_file && (dump_flags & TDF_SCEV))
1314 {
1315 print_gimple_stmt (dump_file, res, 0);
1316 fprintf (stream: dump_file, format: ")\n");
1317 }
1318
1319 return res;
1320}
1321
1322
1323/* Simplify PEELED_CHREC represented by (init_cond, arg) in LOOP.
1324 Handle below case and return the corresponding POLYNOMIAL_CHREC:
1325
1326 # i_17 = PHI <i_13(5), 0(3)>
1327 # _20 = PHI <_5(5), start_4(D)(3)>
1328 ...
1329 i_13 = i_17 + 1;
1330 _5 = start_4(D) + i_13;
1331
1332 Though variable _20 appears as a PEELED_CHREC in the form of
1333 (start_4, _5)_LOOP, it's a POLYNOMIAL_CHREC like {start_4, 1}_LOOP.
1334
1335 See PR41488. */
1336
1337static tree
1338simplify_peeled_chrec (class loop *loop, tree arg, tree init_cond)
1339{
1340 aff_tree aff1, aff2;
1341 tree ev, left, right, type, step_val;
1342 hash_map<tree, name_expansion *> *peeled_chrec_map = NULL;
1343
1344 ev = instantiate_parameters (loop, chrec: analyze_scalar_evolution (loop, arg));
1345 if (ev == NULL_TREE || TREE_CODE (ev) != POLYNOMIAL_CHREC)
1346 return chrec_dont_know;
1347
1348 left = CHREC_LEFT (ev);
1349 right = CHREC_RIGHT (ev);
1350 type = TREE_TYPE (left);
1351 step_val = chrec_fold_plus (type, init_cond, right);
1352
1353 /* Transform (init, {left, right}_LOOP)_LOOP to {init, right}_LOOP
1354 if "left" equals to "init + right". */
1355 if (operand_equal_p (left, step_val, flags: 0))
1356 {
1357 if (dump_file && (dump_flags & TDF_SCEV))
1358 fprintf (stream: dump_file, format: "Simplify PEELED_CHREC into POLYNOMIAL_CHREC.\n");
1359
1360 return build_polynomial_chrec (loop_num: loop->num, left: init_cond, right);
1361 }
1362
1363 /* The affine code only deals with pointer and integer types. */
1364 if (!POINTER_TYPE_P (type)
1365 && !INTEGRAL_TYPE_P (type))
1366 return chrec_dont_know;
1367
1368 /* Try harder to check if they are equal. */
1369 tree_to_aff_combination_expand (left, type, &aff1, &peeled_chrec_map);
1370 tree_to_aff_combination_expand (step_val, type, &aff2, &peeled_chrec_map);
1371 free_affine_expand_cache (&peeled_chrec_map);
1372 aff_combination_scale (&aff2, -1);
1373 aff_combination_add (&aff1, &aff2);
1374
1375 /* Transform (init, {left, right}_LOOP)_LOOP to {init, right}_LOOP
1376 if "left" equals to "init + right". */
1377 if (aff_combination_zero_p (aff: &aff1))
1378 {
1379 if (dump_file && (dump_flags & TDF_SCEV))
1380 fprintf (stream: dump_file, format: "Simplify PEELED_CHREC into POLYNOMIAL_CHREC.\n");
1381
1382 return build_polynomial_chrec (loop_num: loop->num, left: init_cond, right);
1383 }
1384 return chrec_dont_know;
1385}
1386
1387/* Given a LOOP_PHI_NODE, this function determines the evolution
1388 function from LOOP_PHI_NODE to LOOP_PHI_NODE in the loop. */
1389
1390static tree
1391analyze_evolution_in_loop (gphi *loop_phi_node,
1392 tree init_cond)
1393{
1394 int i, n = gimple_phi_num_args (gs: loop_phi_node);
1395 tree evolution_function = chrec_not_analyzed_yet;
1396 class loop *loop = loop_containing_stmt (stmt: loop_phi_node);
1397 basic_block bb;
1398 static bool simplify_peeled_chrec_p = true;
1399
1400 if (dump_file && (dump_flags & TDF_SCEV))
1401 {
1402 fprintf (stream: dump_file, format: "(analyze_evolution_in_loop \n");
1403 fprintf (stream: dump_file, format: " (loop_phi_node = ");
1404 print_gimple_stmt (dump_file, loop_phi_node, 0);
1405 fprintf (stream: dump_file, format: ")\n");
1406 }
1407
1408 for (i = 0; i < n; i++)
1409 {
1410 tree arg = PHI_ARG_DEF (loop_phi_node, i);
1411 tree ev_fn = chrec_dont_know;
1412 t_bool res;
1413
1414 /* Select the edges that enter the loop body. */
1415 bb = gimple_phi_arg_edge (phi: loop_phi_node, i)->src;
1416 if (!flow_bb_inside_loop_p (loop, bb))
1417 continue;
1418
1419 if (TREE_CODE (arg) == SSA_NAME)
1420 {
1421 bool val = false;
1422
1423 /* Pass in the initial condition to the follow edge function. */
1424 scev_dfs dfs (loop, loop_phi_node, init_cond);
1425 res = dfs.get_ev (ev_fn: &ev_fn, arg);
1426
1427 /* If ev_fn has no evolution in the inner loop, and the
1428 init_cond is not equal to ev_fn, then we have an
1429 ambiguity between two possible values, as we cannot know
1430 the number of iterations at this point. */
1431 if (TREE_CODE (ev_fn) != POLYNOMIAL_CHREC
1432 && no_evolution_in_loop_p (chrec: ev_fn, loop_num: loop->num, res: &val) && val
1433 && !operand_equal_p (init_cond, ev_fn, flags: 0))
1434 ev_fn = chrec_dont_know;
1435 }
1436 else
1437 res = t_false;
1438
1439 /* When it is impossible to go back on the same
1440 loop_phi_node by following the ssa edges, the
1441 evolution is represented by a peeled chrec, i.e. the
1442 first iteration, EV_FN has the value INIT_COND, then
1443 all the other iterations it has the value of ARG.
1444 For the moment, PEELED_CHREC nodes are not built. */
1445 if (res != t_true)
1446 {
1447 ev_fn = chrec_dont_know;
1448 /* Try to recognize POLYNOMIAL_CHREC which appears in
1449 the form of PEELED_CHREC, but guard the process with
1450 a bool variable to keep the analyzer from infinite
1451 recurrence for real PEELED_RECs. */
1452 if (simplify_peeled_chrec_p && TREE_CODE (arg) == SSA_NAME)
1453 {
1454 simplify_peeled_chrec_p = false;
1455 ev_fn = simplify_peeled_chrec (loop, arg, init_cond);
1456 simplify_peeled_chrec_p = true;
1457 }
1458 }
1459
1460 /* When there are multiple back edges of the loop (which in fact never
1461 happens currently, but nevertheless), merge their evolutions. */
1462 evolution_function = chrec_merge (evolution_function, ev_fn);
1463
1464 if (evolution_function == chrec_dont_know)
1465 break;
1466 }
1467
1468 if (dump_file && (dump_flags & TDF_SCEV))
1469 {
1470 fprintf (stream: dump_file, format: " (evolution_function = ");
1471 print_generic_expr (dump_file, evolution_function);
1472 fprintf (stream: dump_file, format: "))\n");
1473 }
1474
1475 return evolution_function;
1476}
1477
1478/* Looks to see if VAR is a copy of a constant (via straightforward assignments
1479 or degenerate phi's). If so, returns the constant; else, returns VAR. */
1480
1481static tree
1482follow_copies_to_constant (tree var)
1483{
1484 tree res = var;
1485 while (TREE_CODE (res) == SSA_NAME
1486 /* We face not updated SSA form in multiple places and this walk
1487 may end up in sibling loops so we have to guard it. */
1488 && !name_registered_for_update_p (res))
1489 {
1490 gimple *def = SSA_NAME_DEF_STMT (res);
1491 if (gphi *phi = dyn_cast <gphi *> (p: def))
1492 {
1493 if (tree rhs = degenerate_phi_result (phi))
1494 res = rhs;
1495 else
1496 break;
1497 }
1498 else if (gimple_assign_single_p (gs: def))
1499 /* Will exit loop if not an SSA_NAME. */
1500 res = gimple_assign_rhs1 (gs: def);
1501 else
1502 break;
1503 }
1504 if (CONSTANT_CLASS_P (res))
1505 return res;
1506 return var;
1507}
1508
1509/* Given a loop-phi-node, return the initial conditions of the
1510 variable on entry of the loop. When the CCP has propagated
1511 constants into the loop-phi-node, the initial condition is
1512 instantiated, otherwise the initial condition is kept symbolic.
1513 This analyzer does not analyze the evolution outside the current
1514 loop, and leaves this task to the on-demand tree reconstructor. */
1515
1516static tree
1517analyze_initial_condition (gphi *loop_phi_node)
1518{
1519 int i, n;
1520 tree init_cond = chrec_not_analyzed_yet;
1521 class loop *loop = loop_containing_stmt (stmt: loop_phi_node);
1522
1523 if (dump_file && (dump_flags & TDF_SCEV))
1524 {
1525 fprintf (stream: dump_file, format: "(analyze_initial_condition \n");
1526 fprintf (stream: dump_file, format: " (loop_phi_node = \n");
1527 print_gimple_stmt (dump_file, loop_phi_node, 0);
1528 fprintf (stream: dump_file, format: ")\n");
1529 }
1530
1531 n = gimple_phi_num_args (gs: loop_phi_node);
1532 for (i = 0; i < n; i++)
1533 {
1534 tree branch = PHI_ARG_DEF (loop_phi_node, i);
1535 basic_block bb = gimple_phi_arg_edge (phi: loop_phi_node, i)->src;
1536
1537 /* When the branch is oriented to the loop's body, it does
1538 not contribute to the initial condition. */
1539 if (flow_bb_inside_loop_p (loop, bb))
1540 continue;
1541
1542 if (init_cond == chrec_not_analyzed_yet)
1543 {
1544 init_cond = branch;
1545 continue;
1546 }
1547
1548 if (TREE_CODE (branch) == SSA_NAME)
1549 {
1550 init_cond = chrec_dont_know;
1551 break;
1552 }
1553
1554 init_cond = chrec_merge (init_cond, branch);
1555 }
1556
1557 /* Ooops -- a loop without an entry??? */
1558 if (init_cond == chrec_not_analyzed_yet)
1559 init_cond = chrec_dont_know;
1560
1561 /* We may not have fully constant propagated IL. Handle degenerate PHIs here
1562 to not miss important early loop unrollings. */
1563 init_cond = follow_copies_to_constant (var: init_cond);
1564
1565 if (dump_file && (dump_flags & TDF_SCEV))
1566 {
1567 fprintf (stream: dump_file, format: " (init_cond = ");
1568 print_generic_expr (dump_file, init_cond);
1569 fprintf (stream: dump_file, format: "))\n");
1570 }
1571
1572 return init_cond;
1573}
1574
1575/* Analyze the scalar evolution for LOOP_PHI_NODE. */
1576
1577static tree
1578interpret_loop_phi (class loop *loop, gphi *loop_phi_node)
1579{
1580 class loop *phi_loop = loop_containing_stmt (stmt: loop_phi_node);
1581 tree init_cond;
1582
1583 gcc_assert (phi_loop == loop);
1584
1585 /* Otherwise really interpret the loop phi. */
1586 init_cond = analyze_initial_condition (loop_phi_node);
1587 return analyze_evolution_in_loop (loop_phi_node, init_cond);
1588}
1589
1590/* This function merges the branches of a condition-phi-node,
1591 contained in the outermost loop, and whose arguments are already
1592 analyzed. */
1593
1594static tree
1595interpret_condition_phi (class loop *loop, gphi *condition_phi)
1596{
1597 int i, n = gimple_phi_num_args (gs: condition_phi);
1598 tree res = chrec_not_analyzed_yet;
1599
1600 for (i = 0; i < n; i++)
1601 {
1602 tree branch_chrec;
1603
1604 if (backedge_phi_arg_p (phi: condition_phi, i))
1605 {
1606 res = chrec_dont_know;
1607 break;
1608 }
1609
1610 branch_chrec = analyze_scalar_evolution
1611 (loop, PHI_ARG_DEF (condition_phi, i));
1612
1613 res = chrec_merge (res, branch_chrec);
1614 if (res == chrec_dont_know)
1615 break;
1616 }
1617
1618 return res;
1619}
1620
1621/* Interpret the operation RHS1 OP RHS2. If we didn't
1622 analyze this node before, follow the definitions until ending
1623 either on an analyzed GIMPLE_ASSIGN, or on a loop-phi-node. On the
1624 return path, this function propagates evolutions (ala constant copy
1625 propagation). OPND1 is not a GIMPLE expression because we could
1626 analyze the effect of an inner loop: see interpret_loop_phi. */
1627
1628static tree
1629interpret_rhs_expr (class loop *loop, gimple *at_stmt,
1630 tree type, tree rhs1, enum tree_code code, tree rhs2)
1631{
1632 tree res, chrec1, chrec2, ctype;
1633 gimple *def;
1634
1635 if (get_gimple_rhs_class (code) == GIMPLE_SINGLE_RHS)
1636 {
1637 if (is_gimple_min_invariant (rhs1))
1638 return chrec_convert (type, rhs1, at_stmt);
1639
1640 if (code == SSA_NAME)
1641 return chrec_convert (type, analyze_scalar_evolution (loop, rhs1),
1642 at_stmt);
1643 }
1644
1645 switch (code)
1646 {
1647 case ADDR_EXPR:
1648 if (TREE_CODE (TREE_OPERAND (rhs1, 0)) == MEM_REF
1649 || handled_component_p (TREE_OPERAND (rhs1, 0)))
1650 {
1651 machine_mode mode;
1652 poly_int64 bitsize, bitpos;
1653 int unsignedp, reversep;
1654 int volatilep = 0;
1655 tree base, offset;
1656 tree chrec3;
1657 tree unitpos;
1658
1659 base = get_inner_reference (TREE_OPERAND (rhs1, 0),
1660 &bitsize, &bitpos, &offset, &mode,
1661 &unsignedp, &reversep, &volatilep);
1662
1663 if (TREE_CODE (base) == MEM_REF)
1664 {
1665 rhs2 = TREE_OPERAND (base, 1);
1666 rhs1 = TREE_OPERAND (base, 0);
1667
1668 chrec1 = analyze_scalar_evolution (loop, rhs1);
1669 chrec2 = analyze_scalar_evolution (loop, rhs2);
1670 chrec1 = chrec_convert (type, chrec1, at_stmt);
1671 chrec2 = chrec_convert (TREE_TYPE (rhs2), chrec2, at_stmt);
1672 chrec1 = instantiate_parameters (loop, chrec: chrec1);
1673 chrec2 = instantiate_parameters (loop, chrec: chrec2);
1674 res = chrec_fold_plus (type, chrec1, chrec2);
1675 }
1676 else
1677 {
1678 chrec1 = analyze_scalar_evolution_for_address_of (loop, var: base);
1679 chrec1 = chrec_convert (type, chrec1, at_stmt);
1680 res = chrec1;
1681 }
1682
1683 if (offset != NULL_TREE)
1684 {
1685 chrec2 = analyze_scalar_evolution (loop, offset);
1686 chrec2 = chrec_convert (TREE_TYPE (offset), chrec2, at_stmt);
1687 chrec2 = instantiate_parameters (loop, chrec: chrec2);
1688 res = chrec_fold_plus (type, res, chrec2);
1689 }
1690
1691 if (maybe_ne (a: bitpos, b: 0))
1692 {
1693 unitpos = size_int (exact_div (bitpos, BITS_PER_UNIT));
1694 chrec3 = analyze_scalar_evolution (loop, unitpos);
1695 chrec3 = chrec_convert (TREE_TYPE (unitpos), chrec3, at_stmt);
1696 chrec3 = instantiate_parameters (loop, chrec: chrec3);
1697 res = chrec_fold_plus (type, res, chrec3);
1698 }
1699 }
1700 else
1701 res = chrec_dont_know;
1702 break;
1703
1704 case POINTER_PLUS_EXPR:
1705 chrec1 = analyze_scalar_evolution (loop, rhs1);
1706 chrec2 = analyze_scalar_evolution (loop, rhs2);
1707 chrec1 = chrec_convert (type, chrec1, at_stmt);
1708 chrec2 = chrec_convert (TREE_TYPE (rhs2), chrec2, at_stmt);
1709 chrec1 = instantiate_parameters (loop, chrec: chrec1);
1710 chrec2 = instantiate_parameters (loop, chrec: chrec2);
1711 res = chrec_fold_plus (type, chrec1, chrec2);
1712 break;
1713
1714 case PLUS_EXPR:
1715 chrec1 = analyze_scalar_evolution (loop, rhs1);
1716 chrec2 = analyze_scalar_evolution (loop, rhs2);
1717 ctype = type;
1718 /* When the stmt is conditionally executed re-write the CHREC
1719 into a form that has well-defined behavior on overflow. */
1720 if (at_stmt
1721 && INTEGRAL_TYPE_P (type)
1722 && ! TYPE_OVERFLOW_WRAPS (type)
1723 && ! dominated_by_p (CDI_DOMINATORS, loop->latch,
1724 gimple_bb (g: at_stmt)))
1725 ctype = unsigned_type_for (type);
1726 chrec1 = chrec_convert (ctype, chrec1, at_stmt);
1727 chrec2 = chrec_convert (ctype, chrec2, at_stmt);
1728 chrec1 = instantiate_parameters (loop, chrec: chrec1);
1729 chrec2 = instantiate_parameters (loop, chrec: chrec2);
1730 res = chrec_fold_plus (ctype, chrec1, chrec2);
1731 if (type != ctype)
1732 res = chrec_convert (type, res, at_stmt);
1733 break;
1734
1735 case MINUS_EXPR:
1736 chrec1 = analyze_scalar_evolution (loop, rhs1);
1737 chrec2 = analyze_scalar_evolution (loop, rhs2);
1738 ctype = type;
1739 /* When the stmt is conditionally executed re-write the CHREC
1740 into a form that has well-defined behavior on overflow. */
1741 if (at_stmt
1742 && INTEGRAL_TYPE_P (type)
1743 && ! TYPE_OVERFLOW_WRAPS (type)
1744 && ! dominated_by_p (CDI_DOMINATORS,
1745 loop->latch, gimple_bb (g: at_stmt)))
1746 ctype = unsigned_type_for (type);
1747 chrec1 = chrec_convert (ctype, chrec1, at_stmt);
1748 chrec2 = chrec_convert (ctype, chrec2, at_stmt);
1749 chrec1 = instantiate_parameters (loop, chrec: chrec1);
1750 chrec2 = instantiate_parameters (loop, chrec: chrec2);
1751 res = chrec_fold_minus (ctype, chrec1, chrec2);
1752 if (type != ctype)
1753 res = chrec_convert (type, res, at_stmt);
1754 break;
1755
1756 case NEGATE_EXPR:
1757 chrec1 = analyze_scalar_evolution (loop, rhs1);
1758 ctype = type;
1759 /* When the stmt is conditionally executed re-write the CHREC
1760 into a form that has well-defined behavior on overflow. */
1761 if (at_stmt
1762 && INTEGRAL_TYPE_P (type)
1763 && ! TYPE_OVERFLOW_WRAPS (type)
1764 && ! dominated_by_p (CDI_DOMINATORS,
1765 loop->latch, gimple_bb (g: at_stmt)))
1766 ctype = unsigned_type_for (type);
1767 chrec1 = chrec_convert (ctype, chrec1, at_stmt);
1768 /* TYPE may be integer, real or complex, so use fold_convert. */
1769 chrec1 = instantiate_parameters (loop, chrec: chrec1);
1770 res = chrec_fold_multiply (ctype, chrec1,
1771 fold_convert (ctype, integer_minus_one_node));
1772 if (type != ctype)
1773 res = chrec_convert (type, res, at_stmt);
1774 break;
1775
1776 case BIT_NOT_EXPR:
1777 /* Handle ~X as -1 - X. */
1778 chrec1 = analyze_scalar_evolution (loop, rhs1);
1779 chrec1 = chrec_convert (type, chrec1, at_stmt);
1780 chrec1 = instantiate_parameters (loop, chrec: chrec1);
1781 res = chrec_fold_minus (type,
1782 fold_convert (type, integer_minus_one_node),
1783 chrec1);
1784 break;
1785
1786 case MULT_EXPR:
1787 chrec1 = analyze_scalar_evolution (loop, rhs1);
1788 chrec2 = analyze_scalar_evolution (loop, rhs2);
1789 ctype = type;
1790 /* When the stmt is conditionally executed re-write the CHREC
1791 into a form that has well-defined behavior on overflow. */
1792 if (at_stmt
1793 && INTEGRAL_TYPE_P (type)
1794 && ! TYPE_OVERFLOW_WRAPS (type)
1795 && ! dominated_by_p (CDI_DOMINATORS,
1796 loop->latch, gimple_bb (g: at_stmt)))
1797 ctype = unsigned_type_for (type);
1798 chrec1 = chrec_convert (ctype, chrec1, at_stmt);
1799 chrec2 = chrec_convert (ctype, chrec2, at_stmt);
1800 chrec1 = instantiate_parameters (loop, chrec: chrec1);
1801 chrec2 = instantiate_parameters (loop, chrec: chrec2);
1802 res = chrec_fold_multiply (ctype, chrec1, chrec2);
1803 if (type != ctype)
1804 res = chrec_convert (type, res, at_stmt);
1805 break;
1806
1807 case LSHIFT_EXPR:
1808 {
1809 /* Handle A<<B as A * (1<<B). */
1810 tree uns = unsigned_type_for (type);
1811 chrec1 = analyze_scalar_evolution (loop, rhs1);
1812 chrec2 = analyze_scalar_evolution (loop, rhs2);
1813 chrec1 = chrec_convert (uns, chrec1, at_stmt);
1814 chrec1 = instantiate_parameters (loop, chrec: chrec1);
1815 chrec2 = instantiate_parameters (loop, chrec: chrec2);
1816
1817 tree one = build_int_cst (uns, 1);
1818 chrec2 = fold_build2 (LSHIFT_EXPR, uns, one, chrec2);
1819 res = chrec_fold_multiply (uns, chrec1, chrec2);
1820 res = chrec_convert (type, res, at_stmt);
1821 }
1822 break;
1823
1824 CASE_CONVERT:
1825 /* In case we have a truncation of a widened operation that in
1826 the truncated type has undefined overflow behavior analyze
1827 the operation done in an unsigned type of the same precision
1828 as the final truncation. We cannot derive a scalar evolution
1829 for the widened operation but for the truncated result. */
1830 if (TREE_CODE (type) == INTEGER_TYPE
1831 && TREE_CODE (TREE_TYPE (rhs1)) == INTEGER_TYPE
1832 && TYPE_PRECISION (type) < TYPE_PRECISION (TREE_TYPE (rhs1))
1833 && TYPE_OVERFLOW_UNDEFINED (type)
1834 && TREE_CODE (rhs1) == SSA_NAME
1835 && (def = SSA_NAME_DEF_STMT (rhs1))
1836 && is_gimple_assign (gs: def)
1837 && TREE_CODE_CLASS (gimple_assign_rhs_code (def)) == tcc_binary
1838 && TREE_CODE (gimple_assign_rhs2 (def)) == INTEGER_CST)
1839 {
1840 tree utype = unsigned_type_for (type);
1841 chrec1 = interpret_rhs_expr (loop, at_stmt, type: utype,
1842 rhs1: gimple_assign_rhs1 (gs: def),
1843 code: gimple_assign_rhs_code (gs: def),
1844 rhs2: gimple_assign_rhs2 (gs: def));
1845 }
1846 else
1847 chrec1 = analyze_scalar_evolution (loop, rhs1);
1848 res = chrec_convert (type, chrec1, at_stmt, true, rhs1);
1849 break;
1850
1851 case BIT_AND_EXPR:
1852 /* Given int variable A, handle A&0xffff as (int)(unsigned short)A.
1853 If A is SCEV and its value is in the range of representable set
1854 of type unsigned short, the result expression is a (no-overflow)
1855 SCEV. */
1856 res = chrec_dont_know;
1857 if (tree_fits_uhwi_p (rhs2))
1858 {
1859 int precision;
1860 unsigned HOST_WIDE_INT val = tree_to_uhwi (rhs2);
1861
1862 val ++;
1863 /* Skip if value of rhs2 wraps in unsigned HOST_WIDE_INT or
1864 it's not the maximum value of a smaller type than rhs1. */
1865 if (val != 0
1866 && (precision = exact_log2 (x: val)) > 0
1867 && (unsigned) precision < TYPE_PRECISION (TREE_TYPE (rhs1)))
1868 {
1869 tree utype = build_nonstandard_integer_type (precision, 1);
1870
1871 if (TYPE_PRECISION (utype) < TYPE_PRECISION (TREE_TYPE (rhs1)))
1872 {
1873 chrec1 = analyze_scalar_evolution (loop, rhs1);
1874 chrec1 = chrec_convert (utype, chrec1, at_stmt);
1875 res = chrec_convert (TREE_TYPE (rhs1), chrec1, at_stmt);
1876 }
1877 }
1878 }
1879 break;
1880
1881 default:
1882 res = chrec_dont_know;
1883 break;
1884 }
1885
1886 return res;
1887}
1888
1889/* Interpret the expression EXPR. */
1890
1891static tree
1892interpret_expr (class loop *loop, gimple *at_stmt, tree expr)
1893{
1894 enum tree_code code;
1895 tree type = TREE_TYPE (expr), op0, op1;
1896
1897 if (automatically_generated_chrec_p (chrec: expr))
1898 return expr;
1899
1900 if (TREE_CODE (expr) == POLYNOMIAL_CHREC
1901 || TREE_CODE (expr) == CALL_EXPR
1902 || get_gimple_rhs_class (TREE_CODE (expr)) == GIMPLE_TERNARY_RHS)
1903 return chrec_dont_know;
1904
1905 extract_ops_from_tree (expr, code: &code, op0: &op0, op1: &op1);
1906
1907 return interpret_rhs_expr (loop, at_stmt, type,
1908 rhs1: op0, code, rhs2: op1);
1909}
1910
1911/* Interpret the rhs of the assignment STMT. */
1912
1913static tree
1914interpret_gimple_assign (class loop *loop, gimple *stmt)
1915{
1916 tree type = TREE_TYPE (gimple_assign_lhs (stmt));
1917 enum tree_code code = gimple_assign_rhs_code (gs: stmt);
1918
1919 return interpret_rhs_expr (loop, at_stmt: stmt, type,
1920 rhs1: gimple_assign_rhs1 (gs: stmt), code,
1921 rhs2: gimple_assign_rhs2 (gs: stmt));
1922}
1923
1924
1925
1926/* This section contains all the entry points:
1927 - number_of_iterations_in_loop,
1928 - analyze_scalar_evolution,
1929 - instantiate_parameters.
1930*/
1931
1932/* Helper recursive function. */
1933
1934static tree
1935analyze_scalar_evolution_1 (class loop *loop, tree var)
1936{
1937 gimple *def;
1938 basic_block bb;
1939 class loop *def_loop;
1940 tree res;
1941
1942 if (TREE_CODE (var) != SSA_NAME)
1943 return interpret_expr (loop, NULL, expr: var);
1944
1945 def = SSA_NAME_DEF_STMT (var);
1946 bb = gimple_bb (g: def);
1947 def_loop = bb->loop_father;
1948
1949 if (!flow_bb_inside_loop_p (loop, bb))
1950 {
1951 /* Keep symbolic form, but look through obvious copies for constants. */
1952 res = follow_copies_to_constant (var);
1953 goto set_and_end;
1954 }
1955
1956 if (loop != def_loop)
1957 {
1958 res = analyze_scalar_evolution_1 (loop: def_loop, var);
1959 class loop *loop_to_skip = superloop_at_depth (def_loop,
1960 loop_depth (loop) + 1);
1961 res = compute_overall_effect_of_inner_loop (loop: loop_to_skip, evolution_fn: res);
1962 if (chrec_contains_symbols_defined_in_loop (res, loop->num))
1963 res = analyze_scalar_evolution_1 (loop, var: res);
1964 goto set_and_end;
1965 }
1966
1967 switch (gimple_code (g: def))
1968 {
1969 case GIMPLE_ASSIGN:
1970 res = interpret_gimple_assign (loop, stmt: def);
1971 break;
1972
1973 case GIMPLE_PHI:
1974 if (loop_phi_node_p (phi: def))
1975 res = interpret_loop_phi (loop, loop_phi_node: as_a <gphi *> (p: def));
1976 else
1977 res = interpret_condition_phi (loop, condition_phi: as_a <gphi *> (p: def));
1978 break;
1979
1980 default:
1981 res = chrec_dont_know;
1982 break;
1983 }
1984
1985 set_and_end:
1986
1987 /* Keep the symbolic form. */
1988 if (res == chrec_dont_know)
1989 res = var;
1990
1991 if (loop == def_loop)
1992 set_scalar_evolution (instantiated_below: block_before_loop (loop), scalar: var, chrec: res);
1993
1994 return res;
1995}
1996
1997/* Analyzes and returns the scalar evolution of the ssa_name VAR in
1998 LOOP. LOOP is the loop in which the variable is used.
1999
2000 Example of use: having a pointer VAR to a SSA_NAME node, STMT a
2001 pointer to the statement that uses this variable, in order to
2002 determine the evolution function of the variable, use the following
2003 calls:
2004
2005 loop_p loop = loop_containing_stmt (stmt);
2006 tree chrec_with_symbols = analyze_scalar_evolution (loop, var);
2007 tree chrec_instantiated = instantiate_parameters (loop, chrec_with_symbols);
2008*/
2009
2010tree
2011analyze_scalar_evolution (class loop *loop, tree var)
2012{
2013 tree res;
2014
2015 /* ??? Fix callers. */
2016 if (! loop)
2017 return var;
2018
2019 if (dump_file && (dump_flags & TDF_SCEV))
2020 {
2021 fprintf (stream: dump_file, format: "(analyze_scalar_evolution \n");
2022 fprintf (stream: dump_file, format: " (loop_nb = %d)\n", loop->num);
2023 fprintf (stream: dump_file, format: " (scalar = ");
2024 print_generic_expr (dump_file, var);
2025 fprintf (stream: dump_file, format: ")\n");
2026 }
2027
2028 res = get_scalar_evolution (instantiated_below: block_before_loop (loop), scalar: var);
2029 if (res == chrec_not_analyzed_yet)
2030 {
2031 /* We'll recurse into instantiate_scev, avoid tearing down the
2032 instantiate cache repeatedly and keep it live from here. */
2033 bool destr = false;
2034 if (!global_cache)
2035 {
2036 global_cache = new instantiate_cache_type;
2037 destr = true;
2038 }
2039 res = analyze_scalar_evolution_1 (loop, var);
2040 if (destr)
2041 {
2042 delete global_cache;
2043 global_cache = NULL;
2044 }
2045 }
2046
2047 if (dump_file && (dump_flags & TDF_SCEV))
2048 fprintf (stream: dump_file, format: ")\n");
2049
2050 return res;
2051}
2052
2053/* If CHREC doesn't overflow, set the nonwrapping flag. */
2054
2055void record_nonwrapping_chrec (tree chrec)
2056{
2057 CHREC_NOWRAP(chrec) = 1;
2058
2059 if (dump_file && (dump_flags & TDF_SCEV))
2060 {
2061 fprintf (stream: dump_file, format: "(record_nonwrapping_chrec: ");
2062 print_generic_expr (dump_file, chrec);
2063 fprintf (stream: dump_file, format: ")\n");
2064 }
2065}
2066
2067/* Return true if CHREC's nonwrapping flag is set. */
2068
2069bool nonwrapping_chrec_p (tree chrec)
2070{
2071 if (!chrec || TREE_CODE(chrec) != POLYNOMIAL_CHREC)
2072 return false;
2073
2074 return CHREC_NOWRAP(chrec);
2075}
2076
2077/* Analyzes and returns the scalar evolution of VAR address in LOOP. */
2078
2079static tree
2080analyze_scalar_evolution_for_address_of (class loop *loop, tree var)
2081{
2082 return analyze_scalar_evolution (loop, build_fold_addr_expr (var));
2083}
2084
2085/* Analyze scalar evolution of use of VERSION in USE_LOOP with respect to
2086 WRTO_LOOP (which should be a superloop of USE_LOOP)
2087
2088 FOLDED_CASTS is set to true if resolve_mixers used
2089 chrec_convert_aggressive (TODO -- not really, we are way too conservative
2090 at the moment in order to keep things simple).
2091
2092 To illustrate the meaning of USE_LOOP and WRTO_LOOP, consider the following
2093 example:
2094
2095 for (i = 0; i < 100; i++) -- loop 1
2096 {
2097 for (j = 0; j < 100; j++) -- loop 2
2098 {
2099 k1 = i;
2100 k2 = j;
2101
2102 use2 (k1, k2);
2103
2104 for (t = 0; t < 100; t++) -- loop 3
2105 use3 (k1, k2);
2106
2107 }
2108 use1 (k1, k2);
2109 }
2110
2111 Both k1 and k2 are invariants in loop3, thus
2112 analyze_scalar_evolution_in_loop (loop3, loop3, k1) = k1
2113 analyze_scalar_evolution_in_loop (loop3, loop3, k2) = k2
2114
2115 As they are invariant, it does not matter whether we consider their
2116 usage in loop 3 or loop 2, hence
2117 analyze_scalar_evolution_in_loop (loop2, loop3, k1) =
2118 analyze_scalar_evolution_in_loop (loop2, loop2, k1) = i
2119 analyze_scalar_evolution_in_loop (loop2, loop3, k2) =
2120 analyze_scalar_evolution_in_loop (loop2, loop2, k2) = [0,+,1]_2
2121
2122 Similarly for their evolutions with respect to loop 1. The values of K2
2123 in the use in loop 2 vary independently on loop 1, thus we cannot express
2124 the evolution with respect to loop 1:
2125 analyze_scalar_evolution_in_loop (loop1, loop3, k1) =
2126 analyze_scalar_evolution_in_loop (loop1, loop2, k1) = [0,+,1]_1
2127 analyze_scalar_evolution_in_loop (loop1, loop3, k2) =
2128 analyze_scalar_evolution_in_loop (loop1, loop2, k2) = dont_know
2129
2130 The value of k2 in the use in loop 1 is known, though:
2131 analyze_scalar_evolution_in_loop (loop1, loop1, k1) = [0,+,1]_1
2132 analyze_scalar_evolution_in_loop (loop1, loop1, k2) = 100
2133 */
2134
2135static tree
2136analyze_scalar_evolution_in_loop (class loop *wrto_loop, class loop *use_loop,
2137 tree version, bool *folded_casts)
2138{
2139 bool val = false;
2140 tree ev = version, tmp;
2141
2142 /* We cannot just do
2143
2144 tmp = analyze_scalar_evolution (use_loop, version);
2145 ev = resolve_mixers (wrto_loop, tmp, folded_casts);
2146
2147 as resolve_mixers would query the scalar evolution with respect to
2148 wrto_loop. For example, in the situation described in the function
2149 comment, suppose that wrto_loop = loop1, use_loop = loop3 and
2150 version = k2. Then
2151
2152 analyze_scalar_evolution (use_loop, version) = k2
2153
2154 and resolve_mixers (loop1, k2, folded_casts) finds that the value of
2155 k2 in loop 1 is 100, which is a wrong result, since we are interested
2156 in the value in loop 3.
2157
2158 Instead, we need to proceed from use_loop to wrto_loop loop by loop,
2159 each time checking that there is no evolution in the inner loop. */
2160
2161 if (folded_casts)
2162 *folded_casts = false;
2163 while (1)
2164 {
2165 tmp = analyze_scalar_evolution (loop: use_loop, var: ev);
2166 ev = resolve_mixers (use_loop, tmp, folded_casts);
2167
2168 if (use_loop == wrto_loop)
2169 return ev;
2170
2171 /* If the value of the use changes in the inner loop, we cannot express
2172 its value in the outer loop (we might try to return interval chrec,
2173 but we do not have a user for it anyway) */
2174 if (!no_evolution_in_loop_p (chrec: ev, loop_num: use_loop->num, res: &val)
2175 || !val)
2176 return chrec_dont_know;
2177
2178 use_loop = loop_outer (loop: use_loop);
2179 }
2180}
2181
2182
2183/* Computes a hash function for database element ELT. */
2184
2185static inline hashval_t
2186hash_idx_scev_info (const void *elt_)
2187{
2188 unsigned idx = ((size_t) elt_) - 2;
2189 return scev_info_hasher::hash (elt: &global_cache->entries[idx]);
2190}
2191
2192/* Compares database elements E1 and E2. */
2193
2194static inline int
2195eq_idx_scev_info (const void *e1, const void *e2)
2196{
2197 unsigned idx1 = ((size_t) e1) - 2;
2198 return scev_info_hasher::equal (elt1: &global_cache->entries[idx1],
2199 elt2: (const scev_info_str *) e2);
2200}
2201
2202/* Returns from CACHE the slot number of the cached chrec for NAME. */
2203
2204static unsigned
2205get_instantiated_value_entry (instantiate_cache_type &cache,
2206 tree name, edge instantiate_below)
2207{
2208 if (!cache.map)
2209 {
2210 cache.map = htab_create (10, hash_idx_scev_info, eq_idx_scev_info, NULL);
2211 cache.entries.create (nelems: 10);
2212 }
2213
2214 scev_info_str e;
2215 e.name_version = SSA_NAME_VERSION (name);
2216 e.instantiated_below = instantiate_below->dest->index;
2217 void **slot = htab_find_slot_with_hash (cache.map, &e,
2218 scev_info_hasher::hash (elt: &e), INSERT);
2219 if (!*slot)
2220 {
2221 e.chrec = chrec_not_analyzed_yet;
2222 *slot = (void *)(size_t)(cache.entries.length () + 2);
2223 cache.entries.safe_push (obj: e);
2224 }
2225
2226 return ((size_t)*slot) - 2;
2227}
2228
2229
2230/* Return the closed_loop_phi node for VAR. If there is none, return
2231 NULL_TREE. */
2232
2233static tree
2234loop_closed_phi_def (tree var)
2235{
2236 class loop *loop;
2237 edge exit;
2238 gphi *phi;
2239 gphi_iterator psi;
2240
2241 if (var == NULL_TREE
2242 || TREE_CODE (var) != SSA_NAME)
2243 return NULL_TREE;
2244
2245 loop = loop_containing_stmt (SSA_NAME_DEF_STMT (var));
2246 exit = single_exit (loop);
2247 if (!exit)
2248 return NULL_TREE;
2249
2250 for (psi = gsi_start_phis (exit->dest); !gsi_end_p (i: psi); gsi_next (i: &psi))
2251 {
2252 phi = psi.phi ();
2253 if (PHI_ARG_DEF_FROM_EDGE (phi, exit) == var)
2254 return PHI_RESULT (phi);
2255 }
2256
2257 return NULL_TREE;
2258}
2259
2260static tree instantiate_scev_r (edge, class loop *, class loop *,
2261 tree, bool *, int);
2262
2263/* Analyze all the parameters of the chrec, between INSTANTIATE_BELOW
2264 and EVOLUTION_LOOP, that were left under a symbolic form.
2265
2266 CHREC is an SSA_NAME to be instantiated.
2267
2268 CACHE is the cache of already instantiated values.
2269
2270 Variable pointed by FOLD_CONVERSIONS is set to TRUE when the
2271 conversions that may wrap in signed/pointer type are folded, as long
2272 as the value of the chrec is preserved. If FOLD_CONVERSIONS is NULL
2273 then we don't do such fold.
2274
2275 SIZE_EXPR is used for computing the size of the expression to be
2276 instantiated, and to stop if it exceeds some limit. */
2277
2278static tree
2279instantiate_scev_name (edge instantiate_below,
2280 class loop *evolution_loop, class loop *inner_loop,
2281 tree chrec,
2282 bool *fold_conversions,
2283 int size_expr)
2284{
2285 tree res;
2286 class loop *def_loop;
2287 basic_block def_bb = gimple_bb (SSA_NAME_DEF_STMT (chrec));
2288
2289 /* A parameter, nothing to do. */
2290 if (!def_bb
2291 || !dominated_by_p (CDI_DOMINATORS, def_bb, instantiate_below->dest))
2292 return chrec;
2293
2294 /* We cache the value of instantiated variable to avoid exponential
2295 time complexity due to reevaluations. We also store the convenient
2296 value in the cache in order to prevent infinite recursion -- we do
2297 not want to instantiate the SSA_NAME if it is in a mixer
2298 structure. This is used for avoiding the instantiation of
2299 recursively defined functions, such as:
2300
2301 | a_2 -> {0, +, 1, +, a_2}_1 */
2302
2303 unsigned si = get_instantiated_value_entry (cache&: *global_cache,
2304 name: chrec, instantiate_below);
2305 if (global_cache->get (slot: si) != chrec_not_analyzed_yet)
2306 return global_cache->get (slot: si);
2307
2308 /* On recursion return chrec_dont_know. */
2309 global_cache->set (slot: si, chrec_dont_know);
2310
2311 def_loop = find_common_loop (evolution_loop, def_bb->loop_father);
2312
2313 if (! dominated_by_p (CDI_DOMINATORS,
2314 def_loop->header, instantiate_below->dest))
2315 {
2316 gimple *def = SSA_NAME_DEF_STMT (chrec);
2317 if (gassign *ass = dyn_cast <gassign *> (p: def))
2318 {
2319 switch (gimple_assign_rhs_class (gs: ass))
2320 {
2321 case GIMPLE_UNARY_RHS:
2322 {
2323 tree op0 = instantiate_scev_r (instantiate_below, evolution_loop,
2324 inner_loop, gimple_assign_rhs1 (gs: ass),
2325 fold_conversions, size_expr);
2326 if (op0 == chrec_dont_know)
2327 return chrec_dont_know;
2328 res = fold_build1 (gimple_assign_rhs_code (ass),
2329 TREE_TYPE (chrec), op0);
2330 break;
2331 }
2332 case GIMPLE_BINARY_RHS:
2333 {
2334 tree op0 = instantiate_scev_r (instantiate_below, evolution_loop,
2335 inner_loop, gimple_assign_rhs1 (gs: ass),
2336 fold_conversions, size_expr);
2337 if (op0 == chrec_dont_know)
2338 return chrec_dont_know;
2339 tree op1 = instantiate_scev_r (instantiate_below, evolution_loop,
2340 inner_loop, gimple_assign_rhs2 (gs: ass),
2341 fold_conversions, size_expr);
2342 if (op1 == chrec_dont_know)
2343 return chrec_dont_know;
2344 res = fold_build2 (gimple_assign_rhs_code (ass),
2345 TREE_TYPE (chrec), op0, op1);
2346 break;
2347 }
2348 default:
2349 res = chrec_dont_know;
2350 }
2351 }
2352 else
2353 res = chrec_dont_know;
2354 global_cache->set (slot: si, chrec: res);
2355 return res;
2356 }
2357
2358 /* If the analysis yields a parametric chrec, instantiate the
2359 result again. */
2360 res = analyze_scalar_evolution (loop: def_loop, var: chrec);
2361
2362 /* Don't instantiate default definitions. */
2363 if (TREE_CODE (res) == SSA_NAME
2364 && SSA_NAME_IS_DEFAULT_DEF (res))
2365 ;
2366
2367 /* Don't instantiate loop-closed-ssa phi nodes. */
2368 else if (TREE_CODE (res) == SSA_NAME
2369 && loop_depth (loop: loop_containing_stmt (SSA_NAME_DEF_STMT (res)))
2370 > loop_depth (loop: def_loop))
2371 {
2372 if (res == chrec)
2373 res = loop_closed_phi_def (var: chrec);
2374 else
2375 res = chrec;
2376
2377 /* When there is no loop_closed_phi_def, it means that the
2378 variable is not used after the loop: try to still compute the
2379 value of the variable when exiting the loop. */
2380 if (res == NULL_TREE)
2381 {
2382 loop_p loop = loop_containing_stmt (SSA_NAME_DEF_STMT (chrec));
2383 res = analyze_scalar_evolution (loop, var: chrec);
2384 res = compute_overall_effect_of_inner_loop (loop, evolution_fn: res);
2385 res = instantiate_scev_r (instantiate_below, evolution_loop,
2386 inner_loop, res,
2387 fold_conversions, size_expr);
2388 }
2389 else if (dominated_by_p (CDI_DOMINATORS,
2390 gimple_bb (SSA_NAME_DEF_STMT (res)),
2391 instantiate_below->dest))
2392 res = chrec_dont_know;
2393 }
2394
2395 else if (res != chrec_dont_know)
2396 {
2397 if (inner_loop
2398 && def_bb->loop_father != inner_loop
2399 && !flow_loop_nested_p (def_bb->loop_father, inner_loop))
2400 /* ??? We could try to compute the overall effect of the loop here. */
2401 res = chrec_dont_know;
2402 else
2403 res = instantiate_scev_r (instantiate_below, evolution_loop,
2404 inner_loop, res,
2405 fold_conversions, size_expr);
2406 }
2407
2408 /* Store the correct value to the cache. */
2409 global_cache->set (slot: si, chrec: res);
2410 return res;
2411}
2412
2413/* Analyze all the parameters of the chrec, between INSTANTIATE_BELOW
2414 and EVOLUTION_LOOP, that were left under a symbolic form.
2415
2416 CHREC is a polynomial chain of recurrence to be instantiated.
2417
2418 CACHE is the cache of already instantiated values.
2419
2420 Variable pointed by FOLD_CONVERSIONS is set to TRUE when the
2421 conversions that may wrap in signed/pointer type are folded, as long
2422 as the value of the chrec is preserved. If FOLD_CONVERSIONS is NULL
2423 then we don't do such fold.
2424
2425 SIZE_EXPR is used for computing the size of the expression to be
2426 instantiated, and to stop if it exceeds some limit. */
2427
2428static tree
2429instantiate_scev_poly (edge instantiate_below,
2430 class loop *evolution_loop, class loop *,
2431 tree chrec, bool *fold_conversions, int size_expr)
2432{
2433 tree op1;
2434 tree op0 = instantiate_scev_r (instantiate_below, evolution_loop,
2435 get_chrec_loop (chrec),
2436 CHREC_LEFT (chrec), fold_conversions,
2437 size_expr);
2438 if (op0 == chrec_dont_know)
2439 return chrec_dont_know;
2440
2441 op1 = instantiate_scev_r (instantiate_below, evolution_loop,
2442 get_chrec_loop (chrec),
2443 CHREC_RIGHT (chrec), fold_conversions,
2444 size_expr);
2445 if (op1 == chrec_dont_know)
2446 return chrec_dont_know;
2447
2448 if (CHREC_LEFT (chrec) != op0
2449 || CHREC_RIGHT (chrec) != op1)
2450 {
2451 op1 = chrec_convert_rhs (chrec_type (chrec: op0), op1, NULL);
2452 chrec = build_polynomial_chrec (CHREC_VARIABLE (chrec), left: op0, right: op1);
2453 }
2454
2455 return chrec;
2456}
2457
2458/* Analyze all the parameters of the chrec, between INSTANTIATE_BELOW
2459 and EVOLUTION_LOOP, that were left under a symbolic form.
2460
2461 "C0 CODE C1" is a binary expression of type TYPE to be instantiated.
2462
2463 CACHE is the cache of already instantiated values.
2464
2465 Variable pointed by FOLD_CONVERSIONS is set to TRUE when the
2466 conversions that may wrap in signed/pointer type are folded, as long
2467 as the value of the chrec is preserved. If FOLD_CONVERSIONS is NULL
2468 then we don't do such fold.
2469
2470 SIZE_EXPR is used for computing the size of the expression to be
2471 instantiated, and to stop if it exceeds some limit. */
2472
2473static tree
2474instantiate_scev_binary (edge instantiate_below,
2475 class loop *evolution_loop, class loop *inner_loop,
2476 tree chrec, enum tree_code code,
2477 tree type, tree c0, tree c1,
2478 bool *fold_conversions, int size_expr)
2479{
2480 tree op1;
2481 tree op0 = instantiate_scev_r (instantiate_below, evolution_loop, inner_loop,
2482 c0, fold_conversions, size_expr);
2483 if (op0 == chrec_dont_know)
2484 return chrec_dont_know;
2485
2486 /* While we eventually compute the same op1 if c0 == c1 the process
2487 of doing this is expensive so the following short-cut prevents
2488 exponential compile-time behavior. */
2489 if (c0 != c1)
2490 {
2491 op1 = instantiate_scev_r (instantiate_below, evolution_loop, inner_loop,
2492 c1, fold_conversions, size_expr);
2493 if (op1 == chrec_dont_know)
2494 return chrec_dont_know;
2495 }
2496 else
2497 op1 = op0;
2498
2499 if (c0 != op0
2500 || c1 != op1)
2501 {
2502 op0 = chrec_convert (type, op0, NULL);
2503 op1 = chrec_convert_rhs (type, op1, NULL);
2504
2505 switch (code)
2506 {
2507 case POINTER_PLUS_EXPR:
2508 case PLUS_EXPR:
2509 return chrec_fold_plus (type, op0, op1);
2510
2511 case MINUS_EXPR:
2512 return chrec_fold_minus (type, op0, op1);
2513
2514 case MULT_EXPR:
2515 return chrec_fold_multiply (type, op0, op1);
2516
2517 default:
2518 gcc_unreachable ();
2519 }
2520 }
2521
2522 return chrec ? chrec : fold_build2 (code, type, c0, c1);
2523}
2524
2525/* Analyze all the parameters of the chrec, between INSTANTIATE_BELOW
2526 and EVOLUTION_LOOP, that were left under a symbolic form.
2527
2528 "CHREC" that stands for a convert expression "(TYPE) OP" is to be
2529 instantiated.
2530
2531 CACHE is the cache of already instantiated values.
2532
2533 Variable pointed by FOLD_CONVERSIONS is set to TRUE when the
2534 conversions that may wrap in signed/pointer type are folded, as long
2535 as the value of the chrec is preserved. If FOLD_CONVERSIONS is NULL
2536 then we don't do such fold.
2537
2538 SIZE_EXPR is used for computing the size of the expression to be
2539 instantiated, and to stop if it exceeds some limit. */
2540
2541static tree
2542instantiate_scev_convert (edge instantiate_below,
2543 class loop *evolution_loop, class loop *inner_loop,
2544 tree chrec, tree type, tree op,
2545 bool *fold_conversions, int size_expr)
2546{
2547 tree op0 = instantiate_scev_r (instantiate_below, evolution_loop,
2548 inner_loop, op,
2549 fold_conversions, size_expr);
2550
2551 if (op0 == chrec_dont_know)
2552 return chrec_dont_know;
2553
2554 if (fold_conversions)
2555 {
2556 tree tmp = chrec_convert_aggressive (type, op0, fold_conversions);
2557 if (tmp)
2558 return tmp;
2559
2560 /* If we used chrec_convert_aggressive, we can no longer assume that
2561 signed chrecs do not overflow, as chrec_convert does, so avoid
2562 calling it in that case. */
2563 if (*fold_conversions)
2564 {
2565 if (chrec && op0 == op)
2566 return chrec;
2567
2568 return fold_convert (type, op0);
2569 }
2570 }
2571
2572 return chrec_convert (type, op0, NULL);
2573}
2574
2575/* Analyze all the parameters of the chrec, between INSTANTIATE_BELOW
2576 and EVOLUTION_LOOP, that were left under a symbolic form.
2577
2578 CHREC is a BIT_NOT_EXPR or a NEGATE_EXPR expression to be instantiated.
2579 Handle ~X as -1 - X.
2580 Handle -X as -1 * X.
2581
2582 CACHE is the cache of already instantiated values.
2583
2584 Variable pointed by FOLD_CONVERSIONS is set to TRUE when the
2585 conversions that may wrap in signed/pointer type are folded, as long
2586 as the value of the chrec is preserved. If FOLD_CONVERSIONS is NULL
2587 then we don't do such fold.
2588
2589 SIZE_EXPR is used for computing the size of the expression to be
2590 instantiated, and to stop if it exceeds some limit. */
2591
2592static tree
2593instantiate_scev_not (edge instantiate_below,
2594 class loop *evolution_loop, class loop *inner_loop,
2595 tree chrec,
2596 enum tree_code code, tree type, tree op,
2597 bool *fold_conversions, int size_expr)
2598{
2599 tree op0 = instantiate_scev_r (instantiate_below, evolution_loop,
2600 inner_loop, op,
2601 fold_conversions, size_expr);
2602
2603 if (op0 == chrec_dont_know)
2604 return chrec_dont_know;
2605
2606 if (op != op0)
2607 {
2608 op0 = chrec_convert (type, op0, NULL);
2609
2610 switch (code)
2611 {
2612 case BIT_NOT_EXPR:
2613 return chrec_fold_minus
2614 (type, fold_convert (type, integer_minus_one_node), op0);
2615
2616 case NEGATE_EXPR:
2617 return chrec_fold_multiply
2618 (type, fold_convert (type, integer_minus_one_node), op0);
2619
2620 default:
2621 gcc_unreachable ();
2622 }
2623 }
2624
2625 return chrec ? chrec : fold_build1 (code, type, op0);
2626}
2627
2628/* Analyze all the parameters of the chrec, between INSTANTIATE_BELOW
2629 and EVOLUTION_LOOP, that were left under a symbolic form.
2630
2631 CHREC is the scalar evolution to instantiate.
2632
2633 CACHE is the cache of already instantiated values.
2634
2635 Variable pointed by FOLD_CONVERSIONS is set to TRUE when the
2636 conversions that may wrap in signed/pointer type are folded, as long
2637 as the value of the chrec is preserved. If FOLD_CONVERSIONS is NULL
2638 then we don't do such fold.
2639
2640 SIZE_EXPR is used for computing the size of the expression to be
2641 instantiated, and to stop if it exceeds some limit. */
2642
2643static tree
2644instantiate_scev_r (edge instantiate_below,
2645 class loop *evolution_loop, class loop *inner_loop,
2646 tree chrec,
2647 bool *fold_conversions, int size_expr)
2648{
2649 /* Give up if the expression is larger than the MAX that we allow. */
2650 if (size_expr++ > param_scev_max_expr_size)
2651 return chrec_dont_know;
2652
2653 if (chrec == NULL_TREE
2654 || automatically_generated_chrec_p (chrec)
2655 || is_gimple_min_invariant (chrec))
2656 return chrec;
2657
2658 switch (TREE_CODE (chrec))
2659 {
2660 case SSA_NAME:
2661 return instantiate_scev_name (instantiate_below, evolution_loop,
2662 inner_loop, chrec,
2663 fold_conversions, size_expr);
2664
2665 case POLYNOMIAL_CHREC:
2666 return instantiate_scev_poly (instantiate_below, evolution_loop,
2667 inner_loop, chrec,
2668 fold_conversions, size_expr);
2669
2670 case POINTER_PLUS_EXPR:
2671 case PLUS_EXPR:
2672 case MINUS_EXPR:
2673 case MULT_EXPR:
2674 return instantiate_scev_binary (instantiate_below, evolution_loop,
2675 inner_loop, chrec,
2676 TREE_CODE (chrec), type: chrec_type (chrec),
2677 TREE_OPERAND (chrec, 0),
2678 TREE_OPERAND (chrec, 1),
2679 fold_conversions, size_expr);
2680
2681 CASE_CONVERT:
2682 return instantiate_scev_convert (instantiate_below, evolution_loop,
2683 inner_loop, chrec,
2684 TREE_TYPE (chrec), TREE_OPERAND (chrec, 0),
2685 fold_conversions, size_expr);
2686
2687 case NEGATE_EXPR:
2688 case BIT_NOT_EXPR:
2689 return instantiate_scev_not (instantiate_below, evolution_loop,
2690 inner_loop, chrec,
2691 TREE_CODE (chrec), TREE_TYPE (chrec),
2692 TREE_OPERAND (chrec, 0),
2693 fold_conversions, size_expr);
2694
2695 case ADDR_EXPR:
2696 if (is_gimple_min_invariant (chrec))
2697 return chrec;
2698 /* Fallthru. */
2699 case SCEV_NOT_KNOWN:
2700 return chrec_dont_know;
2701
2702 case SCEV_KNOWN:
2703 return chrec_known;
2704
2705 default:
2706 if (CONSTANT_CLASS_P (chrec))
2707 return chrec;
2708 return chrec_dont_know;
2709 }
2710}
2711
2712/* Analyze all the parameters of the chrec that were left under a
2713 symbolic form. INSTANTIATE_BELOW is the basic block that stops the
2714 recursive instantiation of parameters: a parameter is a variable
2715 that is defined in a basic block that dominates INSTANTIATE_BELOW or
2716 a function parameter. */
2717
2718tree
2719instantiate_scev (edge instantiate_below, class loop *evolution_loop,
2720 tree chrec)
2721{
2722 tree res;
2723
2724 if (dump_file && (dump_flags & TDF_SCEV))
2725 {
2726 fprintf (stream: dump_file, format: "(instantiate_scev \n");
2727 fprintf (stream: dump_file, format: " (instantiate_below = %d -> %d)\n",
2728 instantiate_below->src->index, instantiate_below->dest->index);
2729 if (evolution_loop)
2730 fprintf (stream: dump_file, format: " (evolution_loop = %d)\n", evolution_loop->num);
2731 fprintf (stream: dump_file, format: " (chrec = ");
2732 print_generic_expr (dump_file, chrec);
2733 fprintf (stream: dump_file, format: ")\n");
2734 }
2735
2736 bool destr = false;
2737 if (!global_cache)
2738 {
2739 global_cache = new instantiate_cache_type;
2740 destr = true;
2741 }
2742
2743 res = instantiate_scev_r (instantiate_below, evolution_loop,
2744 NULL, chrec, NULL, size_expr: 0);
2745
2746 if (destr)
2747 {
2748 delete global_cache;
2749 global_cache = NULL;
2750 }
2751
2752 if (dump_file && (dump_flags & TDF_SCEV))
2753 {
2754 fprintf (stream: dump_file, format: " (res = ");
2755 print_generic_expr (dump_file, res);
2756 fprintf (stream: dump_file, format: "))\n");
2757 }
2758
2759 return res;
2760}
2761
2762/* Similar to instantiate_parameters, but does not introduce the
2763 evolutions in outer loops for LOOP invariants in CHREC, and does not
2764 care about causing overflows, as long as they do not affect value
2765 of an expression. */
2766
2767tree
2768resolve_mixers (class loop *loop, tree chrec, bool *folded_casts)
2769{
2770 bool destr = false;
2771 bool fold_conversions = false;
2772 if (!global_cache)
2773 {
2774 global_cache = new instantiate_cache_type;
2775 destr = true;
2776 }
2777
2778 tree ret = instantiate_scev_r (instantiate_below: loop_preheader_edge (loop), evolution_loop: loop, NULL,
2779 chrec, fold_conversions: &fold_conversions, size_expr: 0);
2780
2781 if (folded_casts && !*folded_casts)
2782 *folded_casts = fold_conversions;
2783
2784 if (destr)
2785 {
2786 delete global_cache;
2787 global_cache = NULL;
2788 }
2789
2790 return ret;
2791}
2792
2793/* Entry point for the analysis of the number of iterations pass.
2794 This function tries to safely approximate the number of iterations
2795 the loop will run. When this property is not decidable at compile
2796 time, the result is chrec_dont_know. Otherwise the result is a
2797 scalar or a symbolic parameter. When the number of iterations may
2798 be equal to zero and the property cannot be determined at compile
2799 time, the result is a COND_EXPR that represents in a symbolic form
2800 the conditions under which the number of iterations is not zero.
2801
2802 Example of analysis: suppose that the loop has an exit condition:
2803
2804 "if (b > 49) goto end_loop;"
2805
2806 and that in a previous analysis we have determined that the
2807 variable 'b' has an evolution function:
2808
2809 "EF = {23, +, 5}_2".
2810
2811 When we evaluate the function at the point 5, i.e. the value of the
2812 variable 'b' after 5 iterations in the loop, we have EF (5) = 48,
2813 and EF (6) = 53. In this case the value of 'b' on exit is '53' and
2814 the loop body has been executed 6 times. */
2815
2816tree
2817number_of_latch_executions (class loop *loop)
2818{
2819 edge exit;
2820 class tree_niter_desc niter_desc;
2821 tree may_be_zero;
2822 tree res;
2823
2824 /* Determine whether the number of iterations in loop has already
2825 been computed. */
2826 res = loop->nb_iterations;
2827 if (res)
2828 return res;
2829
2830 may_be_zero = NULL_TREE;
2831
2832 if (dump_file && (dump_flags & TDF_SCEV))
2833 fprintf (stream: dump_file, format: "(number_of_iterations_in_loop = \n");
2834
2835 res = chrec_dont_know;
2836 exit = single_exit (loop);
2837
2838 if (exit && number_of_iterations_exit (loop, exit, niter: &niter_desc, false))
2839 {
2840 may_be_zero = niter_desc.may_be_zero;
2841 res = niter_desc.niter;
2842 }
2843
2844 if (res == chrec_dont_know
2845 || !may_be_zero
2846 || integer_zerop (may_be_zero))
2847 ;
2848 else if (integer_nonzerop (may_be_zero))
2849 res = build_int_cst (TREE_TYPE (res), 0);
2850
2851 else if (COMPARISON_CLASS_P (may_be_zero))
2852 res = fold_build3 (COND_EXPR, TREE_TYPE (res), may_be_zero,
2853 build_int_cst (TREE_TYPE (res), 0), res);
2854 else
2855 res = chrec_dont_know;
2856
2857 if (dump_file && (dump_flags & TDF_SCEV))
2858 {
2859 fprintf (stream: dump_file, format: " (set_nb_iterations_in_loop = ");
2860 print_generic_expr (dump_file, res);
2861 fprintf (stream: dump_file, format: "))\n");
2862 }
2863
2864 loop->nb_iterations = res;
2865 return res;
2866}
2867
2868
2869/* Counters for the stats. */
2870
2871struct chrec_stats
2872{
2873 unsigned nb_chrecs;
2874 unsigned nb_affine;
2875 unsigned nb_affine_multivar;
2876 unsigned nb_higher_poly;
2877 unsigned nb_chrec_dont_know;
2878 unsigned nb_undetermined;
2879};
2880
2881/* Reset the counters. */
2882
2883static inline void
2884reset_chrecs_counters (struct chrec_stats *stats)
2885{
2886 stats->nb_chrecs = 0;
2887 stats->nb_affine = 0;
2888 stats->nb_affine_multivar = 0;
2889 stats->nb_higher_poly = 0;
2890 stats->nb_chrec_dont_know = 0;
2891 stats->nb_undetermined = 0;
2892}
2893
2894/* Dump the contents of a CHREC_STATS structure. */
2895
2896static void
2897dump_chrecs_stats (FILE *file, struct chrec_stats *stats)
2898{
2899 fprintf (stream: file, format: "\n(\n");
2900 fprintf (stream: file, format: "-----------------------------------------\n");
2901 fprintf (stream: file, format: "%d\taffine univariate chrecs\n", stats->nb_affine);
2902 fprintf (stream: file, format: "%d\taffine multivariate chrecs\n", stats->nb_affine_multivar);
2903 fprintf (stream: file, format: "%d\tdegree greater than 2 polynomials\n",
2904 stats->nb_higher_poly);
2905 fprintf (stream: file, format: "%d\tchrec_dont_know chrecs\n", stats->nb_chrec_dont_know);
2906 fprintf (stream: file, format: "-----------------------------------------\n");
2907 fprintf (stream: file, format: "%d\ttotal chrecs\n", stats->nb_chrecs);
2908 fprintf (stream: file, format: "%d\twith undetermined coefficients\n",
2909 stats->nb_undetermined);
2910 fprintf (stream: file, format: "-----------------------------------------\n");
2911 fprintf (stream: file, format: "%d\tchrecs in the scev database\n",
2912 (int) scalar_evolution_info->elements ());
2913 fprintf (stream: file, format: "%d\tsets in the scev database\n", nb_set_scev);
2914 fprintf (stream: file, format: "%d\tgets in the scev database\n", nb_get_scev);
2915 fprintf (stream: file, format: "-----------------------------------------\n");
2916 fprintf (stream: file, format: ")\n\n");
2917}
2918
2919/* Gather statistics about CHREC. */
2920
2921static void
2922gather_chrec_stats (tree chrec, struct chrec_stats *stats)
2923{
2924 if (dump_file && (dump_flags & TDF_STATS))
2925 {
2926 fprintf (stream: dump_file, format: "(classify_chrec ");
2927 print_generic_expr (dump_file, chrec);
2928 fprintf (stream: dump_file, format: "\n");
2929 }
2930
2931 stats->nb_chrecs++;
2932
2933 if (chrec == NULL_TREE)
2934 {
2935 stats->nb_undetermined++;
2936 return;
2937 }
2938
2939 switch (TREE_CODE (chrec))
2940 {
2941 case POLYNOMIAL_CHREC:
2942 if (evolution_function_is_affine_p (chrec))
2943 {
2944 if (dump_file && (dump_flags & TDF_STATS))
2945 fprintf (stream: dump_file, format: " affine_univariate\n");
2946 stats->nb_affine++;
2947 }
2948 else if (evolution_function_is_affine_multivariate_p (chrec, 0))
2949 {
2950 if (dump_file && (dump_flags & TDF_STATS))
2951 fprintf (stream: dump_file, format: " affine_multivariate\n");
2952 stats->nb_affine_multivar++;
2953 }
2954 else
2955 {
2956 if (dump_file && (dump_flags & TDF_STATS))
2957 fprintf (stream: dump_file, format: " higher_degree_polynomial\n");
2958 stats->nb_higher_poly++;
2959 }
2960
2961 break;
2962
2963 default:
2964 break;
2965 }
2966
2967 if (chrec_contains_undetermined (chrec))
2968 {
2969 if (dump_file && (dump_flags & TDF_STATS))
2970 fprintf (stream: dump_file, format: " undetermined\n");
2971 stats->nb_undetermined++;
2972 }
2973
2974 if (dump_file && (dump_flags & TDF_STATS))
2975 fprintf (stream: dump_file, format: ")\n");
2976}
2977
2978/* Classify the chrecs of the whole database. */
2979
2980void
2981gather_stats_on_scev_database (void)
2982{
2983 struct chrec_stats stats;
2984
2985 if (!dump_file)
2986 return;
2987
2988 reset_chrecs_counters (stats: &stats);
2989
2990 hash_table<scev_info_hasher>::iterator iter;
2991 scev_info_str *elt;
2992 FOR_EACH_HASH_TABLE_ELEMENT (*scalar_evolution_info, elt, scev_info_str *,
2993 iter)
2994 gather_chrec_stats (chrec: elt->chrec, stats: &stats);
2995
2996 dump_chrecs_stats (file: dump_file, stats: &stats);
2997}
2998
2999
3000/* Initialize the analysis of scalar evolutions for LOOPS. */
3001
3002void
3003scev_initialize (void)
3004{
3005 gcc_assert (! scev_initialized_p ()
3006 && loops_state_satisfies_p (cfun, LOOPS_NORMAL));
3007
3008 scalar_evolution_info = hash_table<scev_info_hasher>::create_ggc (n: 100);
3009
3010 for (auto loop : loops_list (cfun, 0))
3011 loop->nb_iterations = NULL_TREE;
3012}
3013
3014/* Return true if SCEV is initialized. */
3015
3016bool
3017scev_initialized_p (void)
3018{
3019 return scalar_evolution_info != NULL;
3020}
3021
3022/* Cleans up the information cached by the scalar evolutions analysis
3023 in the hash table. */
3024
3025void
3026scev_reset_htab (void)
3027{
3028 if (!scalar_evolution_info)
3029 return;
3030
3031 scalar_evolution_info->empty ();
3032}
3033
3034/* Cleans up the information cached by the scalar evolutions analysis
3035 in the hash table and in the loop->nb_iterations. */
3036
3037void
3038scev_reset (void)
3039{
3040 scev_reset_htab ();
3041
3042 for (auto loop : loops_list (cfun, 0))
3043 loop->nb_iterations = NULL_TREE;
3044}
3045
3046/* Return true if the IV calculation in TYPE can overflow based on the knowledge
3047 of the upper bound on the number of iterations of LOOP, the BASE and STEP
3048 of IV.
3049
3050 We do not use information whether TYPE can overflow so it is safe to
3051 use this test even for derived IVs not computed every iteration or
3052 hypotetical IVs to be inserted into code. */
3053
3054bool
3055iv_can_overflow_p (class loop *loop, tree type, tree base, tree step)
3056{
3057 widest_int nit;
3058 wide_int base_min, base_max, step_min, step_max, type_min, type_max;
3059 signop sgn = TYPE_SIGN (type);
3060 value_range r;
3061
3062 if (integer_zerop (step))
3063 return false;
3064
3065 if (!INTEGRAL_TYPE_P (TREE_TYPE (base))
3066 || !get_range_query (cfun)->range_of_expr (r, expr: base)
3067 || r.varying_p ()
3068 || r.undefined_p ())
3069 return true;
3070
3071 base_min = r.lower_bound ();
3072 base_max = r.upper_bound ();
3073
3074 if (!INTEGRAL_TYPE_P (TREE_TYPE (step))
3075 || !get_range_query (cfun)->range_of_expr (r, expr: step)
3076 || r.varying_p ()
3077 || r.undefined_p ())
3078 return true;
3079
3080 step_min = r.lower_bound ();
3081 step_max = r.upper_bound ();
3082
3083 if (!get_max_loop_iterations (loop, nit: &nit))
3084 return true;
3085
3086 type_min = wi::min_value (type);
3087 type_max = wi::max_value (type);
3088
3089 /* Just sanity check that we don't see values out of the range of the type.
3090 In this case the arithmetics bellow would overflow. */
3091 gcc_checking_assert (wi::ge_p (base_min, type_min, sgn)
3092 && wi::le_p (base_max, type_max, sgn));
3093
3094 /* Account the possible increment in the last ieration. */
3095 wi::overflow_type overflow = wi::OVF_NONE;
3096 nit = wi::add (x: nit, y: 1, sgn: SIGNED, overflow: &overflow);
3097 if (overflow)
3098 return true;
3099
3100 /* NIT is typeless and can exceed the precision of the type. In this case
3101 overflow is always possible, because we know STEP is non-zero. */
3102 if (wi::min_precision (x: nit, sgn: UNSIGNED) > TYPE_PRECISION (type))
3103 return true;
3104 wide_int nit2 = wide_int::from (x: nit, TYPE_PRECISION (type), sgn: UNSIGNED);
3105
3106 /* If step can be positive, check that nit*step <= type_max-base.
3107 This can be done by unsigned arithmetic and we only need to watch overflow
3108 in the multiplication. The right hand side can always be represented in
3109 the type. */
3110 if (sgn == UNSIGNED || !wi::neg_p (x: step_max))
3111 {
3112 wi::overflow_type overflow = wi::OVF_NONE;
3113 if (wi::gtu_p (x: wi::mul (x: step_max, y: nit2, sgn: UNSIGNED, overflow: &overflow),
3114 y: type_max - base_max)
3115 || overflow)
3116 return true;
3117 }
3118 /* If step can be negative, check that nit*(-step) <= base_min-type_min. */
3119 if (sgn == SIGNED && wi::neg_p (x: step_min))
3120 {
3121 wi::overflow_type overflow, overflow2;
3122 overflow = overflow2 = wi::OVF_NONE;
3123 if (wi::gtu_p (x: wi::mul (x: wi::neg (x: step_min, overflow: &overflow2),
3124 y: nit2, sgn: UNSIGNED, overflow: &overflow),
3125 y: base_min - type_min)
3126 || overflow || overflow2)
3127 return true;
3128 }
3129
3130 return false;
3131}
3132
3133/* Given EV with form of "(type) {inner_base, inner_step}_loop", this
3134 function tries to derive condition under which it can be simplified
3135 into "{(type)inner_base, (type)inner_step}_loop". The condition is
3136 the maximum number that inner iv can iterate. */
3137
3138static tree
3139derive_simple_iv_with_niters (tree ev, tree *niters)
3140{
3141 if (!CONVERT_EXPR_P (ev))
3142 return ev;
3143
3144 tree inner_ev = TREE_OPERAND (ev, 0);
3145 if (TREE_CODE (inner_ev) != POLYNOMIAL_CHREC)
3146 return ev;
3147
3148 tree init = CHREC_LEFT (inner_ev);
3149 tree step = CHREC_RIGHT (inner_ev);
3150 if (TREE_CODE (init) != INTEGER_CST
3151 || TREE_CODE (step) != INTEGER_CST || integer_zerop (step))
3152 return ev;
3153
3154 tree type = TREE_TYPE (ev);
3155 tree inner_type = TREE_TYPE (inner_ev);
3156 if (TYPE_PRECISION (inner_type) >= TYPE_PRECISION (type))
3157 return ev;
3158
3159 /* Type conversion in "(type) {inner_base, inner_step}_loop" can be
3160 folded only if inner iv won't overflow. We compute the maximum
3161 number the inner iv can iterate before overflowing and return the
3162 simplified affine iv. */
3163 tree delta;
3164 init = fold_convert (type, init);
3165 step = fold_convert (type, step);
3166 ev = build_polynomial_chrec (CHREC_VARIABLE (inner_ev), left: init, right: step);
3167 if (tree_int_cst_sign_bit (step))
3168 {
3169 tree bound = lower_bound_in_type (inner_type, inner_type);
3170 delta = fold_build2 (MINUS_EXPR, type, init, fold_convert (type, bound));
3171 step = fold_build1 (NEGATE_EXPR, type, step);
3172 }
3173 else
3174 {
3175 tree bound = upper_bound_in_type (inner_type, inner_type);
3176 delta = fold_build2 (MINUS_EXPR, type, fold_convert (type, bound), init);
3177 }
3178 *niters = fold_build2 (FLOOR_DIV_EXPR, type, delta, step);
3179 return ev;
3180}
3181
3182/* Checks whether use of OP in USE_LOOP behaves as a simple affine iv with
3183 respect to WRTO_LOOP and returns its base and step in IV if possible
3184 (see analyze_scalar_evolution_in_loop for more details on USE_LOOP
3185 and WRTO_LOOP). If ALLOW_NONCONSTANT_STEP is true, we want step to be
3186 invariant in LOOP. Otherwise we require it to be an integer constant.
3187
3188 IV->no_overflow is set to true if we are sure the iv cannot overflow (e.g.
3189 because it is computed in signed arithmetics). Consequently, adding an
3190 induction variable
3191
3192 for (i = IV->base; ; i += IV->step)
3193
3194 is only safe if IV->no_overflow is false, or TYPE_OVERFLOW_UNDEFINED is
3195 false for the type of the induction variable, or you can prove that i does
3196 not wrap by some other argument. Otherwise, this might introduce undefined
3197 behavior, and
3198
3199 i = iv->base;
3200 for (; ; i = (type) ((unsigned type) i + (unsigned type) iv->step))
3201
3202 must be used instead.
3203
3204 When IV_NITERS is not NULL, this function also checks case in which OP
3205 is a conversion of an inner simple iv of below form:
3206
3207 (outer_type){inner_base, inner_step}_loop.
3208
3209 If type of inner iv has smaller precision than outer_type, it can't be
3210 folded into {(outer_type)inner_base, (outer_type)inner_step}_loop because
3211 the inner iv could overflow/wrap. In this case, we derive a condition
3212 under which the inner iv won't overflow/wrap and do the simplification.
3213 The derived condition normally is the maximum number the inner iv can
3214 iterate, and will be stored in IV_NITERS. This is useful in loop niter
3215 analysis, to derive break conditions when a loop must terminate, when is
3216 infinite. */
3217
3218bool
3219simple_iv_with_niters (class loop *wrto_loop, class loop *use_loop,
3220 tree op, affine_iv *iv, tree *iv_niters,
3221 bool allow_nonconstant_step)
3222{
3223 enum tree_code code;
3224 tree type, ev, base, e;
3225 wide_int extreme;
3226 bool folded_casts;
3227
3228 iv->base = NULL_TREE;
3229 iv->step = NULL_TREE;
3230 iv->no_overflow = false;
3231
3232 type = TREE_TYPE (op);
3233 if (!POINTER_TYPE_P (type)
3234 && !INTEGRAL_TYPE_P (type))
3235 return false;
3236
3237 ev = analyze_scalar_evolution_in_loop (wrto_loop, use_loop, version: op,
3238 folded_casts: &folded_casts);
3239 if (chrec_contains_undetermined (ev)
3240 || chrec_contains_symbols_defined_in_loop (ev, wrto_loop->num))
3241 return false;
3242
3243 if (tree_does_not_contain_chrecs (expr: ev))
3244 {
3245 iv->base = ev;
3246 iv->step = build_int_cst (TREE_TYPE (ev), 0);
3247 iv->no_overflow = true;
3248 return true;
3249 }
3250
3251 /* If we can derive valid scalar evolution with assumptions. */
3252 if (iv_niters && TREE_CODE (ev) != POLYNOMIAL_CHREC)
3253 ev = derive_simple_iv_with_niters (ev, niters: iv_niters);
3254
3255 if (TREE_CODE (ev) != POLYNOMIAL_CHREC)
3256 return false;
3257
3258 if (CHREC_VARIABLE (ev) != (unsigned) wrto_loop->num)
3259 return false;
3260
3261 iv->step = CHREC_RIGHT (ev);
3262 if ((!allow_nonconstant_step && TREE_CODE (iv->step) != INTEGER_CST)
3263 || tree_contains_chrecs (iv->step, NULL))
3264 return false;
3265
3266 iv->base = CHREC_LEFT (ev);
3267 if (tree_contains_chrecs (iv->base, NULL))
3268 return false;
3269
3270 iv->no_overflow = !folded_casts && nowrap_type_p (type);
3271
3272 if (!iv->no_overflow
3273 && !iv_can_overflow_p (loop: wrto_loop, type, base: iv->base, step: iv->step))
3274 iv->no_overflow = true;
3275
3276 /* Try to simplify iv base:
3277
3278 (signed T) ((unsigned T)base + step) ;; TREE_TYPE (base) == signed T
3279 == (signed T)(unsigned T)base + step
3280 == base + step
3281
3282 If we can prove operation (base + step) doesn't overflow or underflow.
3283 Specifically, we try to prove below conditions are satisfied:
3284
3285 base <= UPPER_BOUND (type) - step ;;step > 0
3286 base >= LOWER_BOUND (type) - step ;;step < 0
3287
3288 This is done by proving the reverse conditions are false using loop's
3289 initial conditions.
3290
3291 The is necessary to make loop niter, or iv overflow analysis easier
3292 for below example:
3293
3294 int foo (int *a, signed char s, signed char l)
3295 {
3296 signed char i;
3297 for (i = s; i < l; i++)
3298 a[i] = 0;
3299 return 0;
3300 }
3301
3302 Note variable I is firstly converted to type unsigned char, incremented,
3303 then converted back to type signed char. */
3304
3305 if (wrto_loop->num != use_loop->num)
3306 return true;
3307
3308 if (!CONVERT_EXPR_P (iv->base) || TREE_CODE (iv->step) != INTEGER_CST)
3309 return true;
3310
3311 type = TREE_TYPE (iv->base);
3312 e = TREE_OPERAND (iv->base, 0);
3313 if (!tree_nop_conversion_p (type, TREE_TYPE (e))
3314 || TREE_CODE (e) != PLUS_EXPR
3315 || TREE_CODE (TREE_OPERAND (e, 1)) != INTEGER_CST
3316 || !tree_int_cst_equal (iv->step,
3317 fold_convert (type, TREE_OPERAND (e, 1))))
3318 return true;
3319 e = TREE_OPERAND (e, 0);
3320 if (!CONVERT_EXPR_P (e))
3321 return true;
3322 base = TREE_OPERAND (e, 0);
3323 if (!useless_type_conversion_p (type, TREE_TYPE (base)))
3324 return true;
3325
3326 if (tree_int_cst_sign_bit (iv->step))
3327 {
3328 code = LT_EXPR;
3329 extreme = wi::min_value (type);
3330 }
3331 else
3332 {
3333 code = GT_EXPR;
3334 extreme = wi::max_value (type);
3335 }
3336 wi::overflow_type overflow = wi::OVF_NONE;
3337 extreme = wi::sub (x: extreme, y: wi::to_wide (t: iv->step),
3338 TYPE_SIGN (type), overflow: &overflow);
3339 if (overflow)
3340 return true;
3341 e = fold_build2 (code, boolean_type_node, base,
3342 wide_int_to_tree (type, extreme));
3343 e = simplify_using_initial_conditions (use_loop, e);
3344 if (!integer_zerop (e))
3345 return true;
3346
3347 if (POINTER_TYPE_P (TREE_TYPE (base)))
3348 code = POINTER_PLUS_EXPR;
3349 else
3350 code = PLUS_EXPR;
3351
3352 iv->base = fold_build2 (code, TREE_TYPE (base), base, iv->step);
3353 return true;
3354}
3355
3356/* Like simple_iv_with_niters, but return TRUE when OP behaves as a simple
3357 affine iv unconditionally. */
3358
3359bool
3360simple_iv (class loop *wrto_loop, class loop *use_loop, tree op,
3361 affine_iv *iv, bool allow_nonconstant_step)
3362{
3363 return simple_iv_with_niters (wrto_loop, use_loop, op, iv,
3364 NULL, allow_nonconstant_step);
3365}
3366
3367/* Finalize the scalar evolution analysis. */
3368
3369void
3370scev_finalize (void)
3371{
3372 if (!scalar_evolution_info)
3373 return;
3374 scalar_evolution_info->empty ();
3375 scalar_evolution_info = NULL;
3376 free_numbers_of_iterations_estimates (cfun);
3377}
3378
3379/* Returns true if the expression EXPR is considered to be too expensive
3380 for scev_const_prop. Sets *COND_OVERFLOW_P to true when the
3381 expression might contain a sub-expression that is subject to undefined
3382 overflow behavior and conditionally evaluated. */
3383
3384static bool
3385expression_expensive_p (tree expr, bool *cond_overflow_p,
3386 hash_map<tree, uint64_t> &cache, uint64_t &cost)
3387{
3388 enum tree_code code;
3389
3390 if (is_gimple_val (expr))
3391 return false;
3392
3393 code = TREE_CODE (expr);
3394 if (code == TRUNC_DIV_EXPR
3395 || code == CEIL_DIV_EXPR
3396 || code == FLOOR_DIV_EXPR
3397 || code == ROUND_DIV_EXPR
3398 || code == TRUNC_MOD_EXPR
3399 || code == CEIL_MOD_EXPR
3400 || code == FLOOR_MOD_EXPR
3401 || code == ROUND_MOD_EXPR
3402 || code == EXACT_DIV_EXPR)
3403 {
3404 /* Division by power of two is usually cheap, so we allow it.
3405 Forbid anything else. */
3406 if (!integer_pow2p (TREE_OPERAND (expr, 1)))
3407 return true;
3408 }
3409
3410 bool visited_p;
3411 uint64_t &local_cost = cache.get_or_insert (k: expr, existed: &visited_p);
3412 if (visited_p)
3413 {
3414 uint64_t tem = cost + local_cost;
3415 if (tem < cost)
3416 return true;
3417 cost = tem;
3418 return false;
3419 }
3420 local_cost = 1;
3421
3422 uint64_t op_cost = 0;
3423 if (code == CALL_EXPR)
3424 {
3425 tree arg;
3426 call_expr_arg_iterator iter;
3427 /* Even though is_inexpensive_builtin might say true, we will get a
3428 library call for popcount when backend does not have an instruction
3429 to do so. We consider this to be expensive and generate
3430 __builtin_popcount only when backend defines it. */
3431 optab optab;
3432 combined_fn cfn = get_call_combined_fn (expr);
3433 switch (cfn)
3434 {
3435 CASE_CFN_POPCOUNT:
3436 optab = popcount_optab;
3437 goto bitcount_call;
3438 CASE_CFN_CLZ:
3439 optab = clz_optab;
3440 goto bitcount_call;
3441 CASE_CFN_CTZ:
3442 optab = ctz_optab;
3443bitcount_call:
3444 /* Check if opcode for popcount is available in the mode required. */
3445 if (optab_handler (op: optab,
3446 TYPE_MODE (TREE_TYPE (CALL_EXPR_ARG (expr, 0))))
3447 == CODE_FOR_nothing)
3448 {
3449 machine_mode mode;
3450 mode = TYPE_MODE (TREE_TYPE (CALL_EXPR_ARG (expr, 0)));
3451 scalar_int_mode int_mode;
3452
3453 /* If the mode is of 2 * UNITS_PER_WORD size, we can handle
3454 double-word popcount by emitting two single-word popcount
3455 instructions. */
3456 if (is_a <scalar_int_mode> (m: mode, result: &int_mode)
3457 && GET_MODE_SIZE (mode: int_mode) == 2 * UNITS_PER_WORD
3458 && (optab_handler (op: optab, mode: word_mode)
3459 != CODE_FOR_nothing))
3460 break;
3461 return true;
3462 }
3463 break;
3464
3465 default:
3466 if (cfn == CFN_LAST
3467 || !is_inexpensive_builtin (get_callee_fndecl (expr)))
3468 return true;
3469 break;
3470 }
3471
3472 FOR_EACH_CALL_EXPR_ARG (arg, iter, expr)
3473 if (expression_expensive_p (expr: arg, cond_overflow_p, cache, cost&: op_cost))
3474 return true;
3475 *cache.get (k: expr) += op_cost;
3476 cost += op_cost + 1;
3477 return false;
3478 }
3479
3480 if (code == COND_EXPR)
3481 {
3482 if (expression_expensive_p (TREE_OPERAND (expr, 0), cond_overflow_p,
3483 cache, cost&: op_cost)
3484 || (EXPR_P (TREE_OPERAND (expr, 1))
3485 && EXPR_P (TREE_OPERAND (expr, 2)))
3486 /* If either branch has side effects or could trap. */
3487 || TREE_SIDE_EFFECTS (TREE_OPERAND (expr, 1))
3488 || generic_expr_could_trap_p (TREE_OPERAND (expr, 1))
3489 || TREE_SIDE_EFFECTS (TREE_OPERAND (expr, 0))
3490 || generic_expr_could_trap_p (TREE_OPERAND (expr, 0))
3491 || expression_expensive_p (TREE_OPERAND (expr, 1), cond_overflow_p,
3492 cache, cost&: op_cost)
3493 || expression_expensive_p (TREE_OPERAND (expr, 2), cond_overflow_p,
3494 cache, cost&: op_cost))
3495 return true;
3496 /* Conservatively assume there's overflow for now. */
3497 *cond_overflow_p = true;
3498 *cache.get (k: expr) += op_cost;
3499 cost += op_cost + 1;
3500 return false;
3501 }
3502
3503 switch (TREE_CODE_CLASS (code))
3504 {
3505 case tcc_binary:
3506 case tcc_comparison:
3507 if (expression_expensive_p (TREE_OPERAND (expr, 1), cond_overflow_p,
3508 cache, cost&: op_cost))
3509 return true;
3510
3511 /* Fallthru. */
3512 case tcc_unary:
3513 if (expression_expensive_p (TREE_OPERAND (expr, 0), cond_overflow_p,
3514 cache, cost&: op_cost))
3515 return true;
3516 *cache.get (k: expr) += op_cost;
3517 cost += op_cost + 1;
3518 return false;
3519
3520 default:
3521 return true;
3522 }
3523}
3524
3525bool
3526expression_expensive_p (tree expr, bool *cond_overflow_p)
3527{
3528 hash_map<tree, uint64_t> cache;
3529 uint64_t expanded_size = 0;
3530 *cond_overflow_p = false;
3531 return (expression_expensive_p (expr, cond_overflow_p, cache, cost&: expanded_size)
3532 /* ??? Both the explicit unsharing and gimplification of expr will
3533 expand shared trees to multiple copies.
3534 Guard against exponential growth by counting the visits and
3535 comparing againt the number of original nodes. Allow a tiny
3536 bit of duplication to catch some additional optimizations. */
3537 || expanded_size > (cache.elements () + 1));
3538}
3539
3540/* Match.pd function to match bitwise inductive expression.
3541 .i.e.
3542 _2 = 1 << _1;
3543 _3 = ~_2;
3544 tmp_9 = _3 & tmp_12; */
3545extern bool gimple_bitwise_induction_p (tree, tree *, tree (*)(tree));
3546
3547/* Return the inductive expression of bitwise operation if possible,
3548 otherwise returns DEF. */
3549static tree
3550analyze_and_compute_bitwise_induction_effect (class loop* loop,
3551 tree phidef,
3552 unsigned HOST_WIDE_INT niter)
3553{
3554 tree match_op[3],inv, bitwise_scev;
3555 tree type = TREE_TYPE (phidef);
3556 gphi* header_phi = NULL;
3557
3558 /* Match things like op2(MATCH_OP[2]), op1(MATCH_OP[1]), phidef(PHIDEF)
3559
3560 op2 = PHI <phidef, inv>
3561 _1 = (int) bit_17;
3562 _3 = 1 << _1;
3563 op1 = ~_3;
3564 phidef = op1 & op2; */
3565 if (!gimple_bitwise_induction_p (phidef, &match_op[0], NULL)
3566 || TREE_CODE (match_op[2]) != SSA_NAME
3567 || !(header_phi = dyn_cast <gphi *> (SSA_NAME_DEF_STMT (match_op[2])))
3568 || gimple_bb (g: header_phi) != loop->header
3569 || gimple_phi_num_args (gs: header_phi) != 2)
3570 return NULL_TREE;
3571
3572 if (PHI_ARG_DEF_FROM_EDGE (header_phi, loop_latch_edge (loop)) != phidef)
3573 return NULL_TREE;
3574
3575 bitwise_scev = analyze_scalar_evolution (loop, var: match_op[1]);
3576 bitwise_scev = instantiate_parameters (loop, chrec: bitwise_scev);
3577
3578 /* Make sure bits is in range of type precision. */
3579 if (TREE_CODE (bitwise_scev) != POLYNOMIAL_CHREC
3580 || !INTEGRAL_TYPE_P (TREE_TYPE (bitwise_scev))
3581 || !tree_fits_uhwi_p (CHREC_LEFT (bitwise_scev))
3582 || tree_to_uhwi (CHREC_LEFT (bitwise_scev)) >= TYPE_PRECISION (type)
3583 || !tree_fits_shwi_p (CHREC_RIGHT (bitwise_scev)))
3584 return NULL_TREE;
3585
3586enum bit_op_kind
3587 {
3588 INDUCTION_BIT_CLEAR,
3589 INDUCTION_BIT_IOR,
3590 INDUCTION_BIT_XOR,
3591 INDUCTION_BIT_RESET,
3592 INDUCTION_ZERO,
3593 INDUCTION_ALL
3594 };
3595
3596 enum bit_op_kind induction_kind;
3597 enum tree_code code1
3598 = gimple_assign_rhs_code (SSA_NAME_DEF_STMT (phidef));
3599 enum tree_code code2
3600 = gimple_assign_rhs_code (SSA_NAME_DEF_STMT (match_op[0]));
3601
3602 /* BIT_CLEAR: A &= ~(1 << bit)
3603 BIT_RESET: A ^= (1 << bit).
3604 BIT_IOR: A |= (1 << bit)
3605 BIT_ZERO: A &= (1 << bit)
3606 BIT_ALL: A |= ~(1 << bit)
3607 BIT_XOR: A ^= ~(1 << bit).
3608 bit is induction variable. */
3609 switch (code1)
3610 {
3611 case BIT_AND_EXPR:
3612 induction_kind = code2 == BIT_NOT_EXPR
3613 ? INDUCTION_BIT_CLEAR
3614 : INDUCTION_ZERO;
3615 break;
3616 case BIT_IOR_EXPR:
3617 induction_kind = code2 == BIT_NOT_EXPR
3618 ? INDUCTION_ALL
3619 : INDUCTION_BIT_IOR;
3620 break;
3621 case BIT_XOR_EXPR:
3622 induction_kind = code2 == BIT_NOT_EXPR
3623 ? INDUCTION_BIT_XOR
3624 : INDUCTION_BIT_RESET;
3625 break;
3626 /* A ^ ~(1 << bit) is equal to ~(A ^ (1 << bit)). */
3627 case BIT_NOT_EXPR:
3628 gcc_assert (code2 == BIT_XOR_EXPR);
3629 induction_kind = INDUCTION_BIT_XOR;
3630 break;
3631 default:
3632 gcc_unreachable ();
3633 }
3634
3635 if (induction_kind == INDUCTION_ZERO)
3636 return build_zero_cst (type);
3637 if (induction_kind == INDUCTION_ALL)
3638 return build_all_ones_cst (type);
3639
3640 wide_int bits = wi::zero (TYPE_PRECISION (type));
3641 HOST_WIDE_INT bit_start = tree_to_shwi (CHREC_LEFT (bitwise_scev));
3642 HOST_WIDE_INT step = tree_to_shwi (CHREC_RIGHT (bitwise_scev));
3643 HOST_WIDE_INT bit_final = bit_start + step * niter;
3644
3645 /* bit_start, bit_final in range of [0,TYPE_PRECISION)
3646 implies all bits are set in range. */
3647 if (bit_final >= TYPE_PRECISION (type)
3648 || bit_final < 0)
3649 return NULL_TREE;
3650
3651 /* Loop tripcount should be niter + 1. */
3652 for (unsigned i = 0; i != niter + 1; i++)
3653 {
3654 bits = wi::set_bit (x: bits, bit: bit_start);
3655 bit_start += step;
3656 }
3657
3658 bool inverted = false;
3659 switch (induction_kind)
3660 {
3661 case INDUCTION_BIT_CLEAR:
3662 code1 = BIT_AND_EXPR;
3663 inverted = true;
3664 break;
3665 case INDUCTION_BIT_IOR:
3666 code1 = BIT_IOR_EXPR;
3667 break;
3668 case INDUCTION_BIT_RESET:
3669 code1 = BIT_XOR_EXPR;
3670 break;
3671 /* A ^= ~(1 << bit) is special, when loop tripcount is even,
3672 it's equal to A ^= bits, else A ^= ~bits. */
3673 case INDUCTION_BIT_XOR:
3674 code1 = BIT_XOR_EXPR;
3675 if (niter % 2 == 0)
3676 inverted = true;
3677 break;
3678 default:
3679 gcc_unreachable ();
3680 }
3681
3682 if (inverted)
3683 bits = wi::bit_not (x: bits);
3684
3685 inv = PHI_ARG_DEF_FROM_EDGE (header_phi, loop_preheader_edge (loop));
3686 return fold_build2 (code1, type, inv, wide_int_to_tree (type, bits));
3687}
3688
3689/* Match.pd function to match bitop with invariant expression
3690 .i.e.
3691 tmp_7 = _0 & _1; */
3692extern bool gimple_bitop_with_inv_p (tree, tree *, tree (*)(tree));
3693
3694/* Return the inductive expression of bitop with invariant if possible,
3695 otherwise returns DEF. */
3696static tree
3697analyze_and_compute_bitop_with_inv_effect (class loop* loop, tree phidef,
3698 tree niter)
3699{
3700 tree match_op[2],inv;
3701 tree type = TREE_TYPE (phidef);
3702 gphi* header_phi = NULL;
3703 enum tree_code code;
3704 /* match thing like op0 (match[0]), op1 (match[1]), phidef (PHIDEF)
3705
3706 op1 = PHI <phidef, inv>
3707 phidef = op0 & op1
3708 if op0 is an invariant, it could change to
3709 phidef = op0 & inv. */
3710 gimple *def;
3711 def = SSA_NAME_DEF_STMT (phidef);
3712 if (!(is_gimple_assign (gs: def)
3713 && ((code = gimple_assign_rhs_code (gs: def)), true)
3714 && (code == BIT_AND_EXPR || code == BIT_IOR_EXPR
3715 || code == BIT_XOR_EXPR)))
3716 return NULL_TREE;
3717
3718 match_op[0] = gimple_assign_rhs1 (gs: def);
3719 match_op[1] = gimple_assign_rhs2 (gs: def);
3720
3721 if (expr_invariant_in_loop_p (loop, match_op[1]))
3722 std::swap (a&: match_op[0], b&: match_op[1]);
3723
3724 if (TREE_CODE (match_op[1]) != SSA_NAME
3725 || !expr_invariant_in_loop_p (loop, match_op[0])
3726 || !(header_phi = dyn_cast <gphi *> (SSA_NAME_DEF_STMT (match_op[1])))
3727 || gimple_bb (g: header_phi) != loop->header
3728 || gimple_phi_num_args (gs: header_phi) != 2)
3729 return NULL_TREE;
3730
3731 if (PHI_ARG_DEF_FROM_EDGE (header_phi, loop_latch_edge (loop)) != phidef)
3732 return NULL_TREE;
3733
3734 enum tree_code code1
3735 = gimple_assign_rhs_code (gs: def);
3736
3737 if (code1 == BIT_XOR_EXPR)
3738 {
3739 if (!tree_fits_uhwi_p (niter))
3740 return NULL_TREE;
3741 unsigned HOST_WIDE_INT niter_num;
3742 niter_num = tree_to_uhwi (niter);
3743 if (niter_num % 2 != 0)
3744 match_op[0] = build_zero_cst (type);
3745 }
3746
3747 inv = PHI_ARG_DEF_FROM_EDGE (header_phi, loop_preheader_edge (loop));
3748 return fold_build2 (code1, type, inv, match_op[0]);
3749}
3750
3751/* Do final value replacement for LOOP, return true if we did anything. */
3752
3753bool
3754final_value_replacement_loop (class loop *loop)
3755{
3756 /* If we do not know exact number of iterations of the loop, we cannot
3757 replace the final value. */
3758 edge exit = single_exit (loop);
3759 if (!exit)
3760 return false;
3761
3762 tree niter = number_of_latch_executions (loop);
3763 if (niter == chrec_dont_know)
3764 return false;
3765
3766 /* Ensure that it is possible to insert new statements somewhere. */
3767 if (!single_pred_p (bb: exit->dest))
3768 split_loop_exit_edge (exit);
3769
3770 /* Set stmt insertion pointer. All stmts are inserted before this point. */
3771
3772 class loop *ex_loop
3773 = superloop_at_depth (loop,
3774 loop_depth (loop: exit->dest->loop_father) + 1);
3775
3776 bool any = false;
3777 gphi_iterator psi;
3778 for (psi = gsi_start_phis (exit->dest); !gsi_end_p (i: psi); )
3779 {
3780 gphi *phi = psi.phi ();
3781 tree rslt = PHI_RESULT (phi);
3782 tree phidef = PHI_ARG_DEF_FROM_EDGE (phi, exit);
3783 tree def = phidef;
3784 if (virtual_operand_p (op: def))
3785 {
3786 gsi_next (i: &psi);
3787 continue;
3788 }
3789
3790 if (!POINTER_TYPE_P (TREE_TYPE (def))
3791 && !INTEGRAL_TYPE_P (TREE_TYPE (def)))
3792 {
3793 gsi_next (i: &psi);
3794 continue;
3795 }
3796
3797 bool folded_casts;
3798 def = analyze_scalar_evolution_in_loop (wrto_loop: ex_loop, use_loop: loop, version: def,
3799 folded_casts: &folded_casts);
3800
3801 tree bitinv_def, bit_def;
3802 unsigned HOST_WIDE_INT niter_num;
3803
3804 if (def != chrec_dont_know)
3805 def = compute_overall_effect_of_inner_loop (loop: ex_loop, evolution_fn: def);
3806
3807 /* Handle bitop with invariant induction expression.
3808
3809 .i.e
3810 for (int i =0 ;i < 32; i++)
3811 tmp &= bit2;
3812 if bit2 is an invariant in loop which could simple to
3813 tmp &= bit2. */
3814 else if ((bitinv_def
3815 = analyze_and_compute_bitop_with_inv_effect (loop,
3816 phidef, niter)))
3817 def = bitinv_def;
3818
3819 /* Handle bitwise induction expression.
3820
3821 .i.e.
3822 for (int i = 0; i != 64; i+=3)
3823 res &= ~(1UL << i);
3824
3825 RES can't be analyzed out by SCEV because it is not polynomially
3826 expressible, but in fact final value of RES can be replaced by
3827 RES & CONSTANT where CONSTANT all ones with bit {0,3,6,9,... ,63}
3828 being cleared, similar for BIT_IOR_EXPR/BIT_XOR_EXPR. */
3829 else if (tree_fits_uhwi_p (niter)
3830 && (niter_num = tree_to_uhwi (niter)) != 0
3831 && niter_num < TYPE_PRECISION (TREE_TYPE (phidef))
3832 && (bit_def
3833 = analyze_and_compute_bitwise_induction_effect (loop,
3834 phidef,
3835 niter: niter_num)))
3836 def = bit_def;
3837
3838 bool cond_overflow_p;
3839 if (!tree_does_not_contain_chrecs (expr: def)
3840 || chrec_contains_symbols_defined_in_loop (def, ex_loop->num)
3841 /* Moving the computation from the loop may prolong life range
3842 of some ssa names, which may cause problems if they appear
3843 on abnormal edges. */
3844 || contains_abnormal_ssa_name_p (def)
3845 /* Do not emit expensive expressions. The rationale is that
3846 when someone writes a code like
3847
3848 while (n > 45) n -= 45;
3849
3850 he probably knows that n is not large, and does not want it
3851 to be turned into n %= 45. */
3852 || expression_expensive_p (expr: def, cond_overflow_p: &cond_overflow_p))
3853 {
3854 if (dump_file && (dump_flags & TDF_DETAILS))
3855 {
3856 fprintf (stream: dump_file, format: "not replacing:\n ");
3857 print_gimple_stmt (dump_file, phi, 0);
3858 fprintf (stream: dump_file, format: "\n");
3859 }
3860 gsi_next (i: &psi);
3861 continue;
3862 }
3863
3864 /* Eliminate the PHI node and replace it by a computation outside
3865 the loop. */
3866 if (dump_file)
3867 {
3868 fprintf (stream: dump_file, format: "\nfinal value replacement:\n ");
3869 print_gimple_stmt (dump_file, phi, 0);
3870 fprintf (stream: dump_file, format: " with expr: ");
3871 print_generic_expr (dump_file, def);
3872 fprintf (stream: dump_file, format: "\n");
3873 }
3874 any = true;
3875 /* ??? Here we'd like to have a unshare_expr that would assign
3876 shared sub-trees to new temporary variables either gimplified
3877 to a GIMPLE sequence or to a statement list (keeping this a
3878 GENERIC interface). */
3879 def = unshare_expr (def);
3880 auto loc = gimple_phi_arg_location (phi, i: exit->dest_idx);
3881 remove_phi_node (&psi, false);
3882
3883 /* Propagate constants immediately, but leave an unused initialization
3884 around to avoid invalidating the SCEV cache. */
3885 if (CONSTANT_CLASS_P (def) && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (rslt))
3886 replace_uses_by (rslt, def);
3887
3888 /* Create the replacement statements. */
3889 gimple_seq stmts;
3890 def = force_gimple_operand (def, &stmts, false, NULL_TREE);
3891 gassign *ass = gimple_build_assign (rslt, def);
3892 gimple_set_location (g: ass, location: loc);
3893 gimple_seq_add_stmt (&stmts, ass);
3894
3895 /* If def's type has undefined overflow and there were folded
3896 casts, rewrite all stmts added for def into arithmetics
3897 with defined overflow behavior. */
3898 if ((folded_casts
3899 && ANY_INTEGRAL_TYPE_P (TREE_TYPE (def))
3900 && TYPE_OVERFLOW_UNDEFINED (TREE_TYPE (def)))
3901 || cond_overflow_p)
3902 {
3903 gimple_stmt_iterator gsi2;
3904 gsi2 = gsi_start (seq&: stmts);
3905 while (!gsi_end_p (i: gsi2))
3906 {
3907 gimple *stmt = gsi_stmt (i: gsi2);
3908 if (is_gimple_assign (gs: stmt)
3909 && arith_code_with_undefined_signed_overflow
3910 (gimple_assign_rhs_code (gs: stmt)))
3911 rewrite_to_defined_overflow (&gsi2);
3912 gsi_next (i: &gsi2);
3913 }
3914 }
3915 gimple_stmt_iterator gsi = gsi_after_labels (bb: exit->dest);
3916 gsi_insert_seq_before (&gsi, stmts, GSI_SAME_STMT);
3917 if (dump_file)
3918 {
3919 fprintf (stream: dump_file, format: " final stmt:\n ");
3920 print_gimple_stmt (dump_file, SSA_NAME_DEF_STMT (rslt), 0);
3921 fprintf (stream: dump_file, format: "\n");
3922 }
3923 }
3924
3925 return any;
3926}
3927
3928#include "gt-tree-scalar-evolution.h"
3929

source code of gcc/tree-scalar-evolution.cc