Line data Source code
1 : /* String length optimization
2 : Copyright (C) 2011-2026 Free Software Foundation, Inc.
3 : Contributed by Jakub Jelinek <jakub@redhat.com>
4 :
5 : This file is part of GCC.
6 :
7 : GCC is free software; you can redistribute it and/or modify
8 : it under the terms of the GNU General Public License as published by
9 : the Free Software Foundation; either version 3, or (at your option)
10 : any later version.
11 :
12 : GCC is distributed in the hope that it will be useful,
13 : but WITHOUT ANY WARRANTY; without even the implied warranty of
14 : MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 : GNU General Public License for more details.
16 :
17 : You should have received a copy of the GNU General Public License
18 : along with GCC; see the file COPYING3. If not see
19 : <http://www.gnu.org/licenses/>. */
20 :
21 : #include "config.h"
22 : #include "system.h"
23 : #include "coretypes.h"
24 : #include "backend.h"
25 : #include "rtl.h"
26 : #include "tree.h"
27 : #include "gimple.h"
28 : #include "alloc-pool.h"
29 : #include "tree-pass.h"
30 : #include "ssa.h"
31 : #include "cgraph.h"
32 : #include "gimple-pretty-print.h"
33 : #include "gimple-ssa-warn-access.h"
34 : #include "gimple-ssa-warn-restrict.h"
35 : #include "fold-const.h"
36 : #include "stor-layout.h"
37 : #include "gimple-iterator.h"
38 : #include "gimple-fold.h"
39 : #include "tree-eh.h"
40 : #include "gimplify.h"
41 : #include "gimplify-me.h"
42 : #include "expr.h"
43 : #include "tree-cfg.h"
44 : #include "tree-dfa.h"
45 : #include "domwalk.h"
46 : #include "tree-ssa-alias.h"
47 : #include "tree-ssa-propagate.h"
48 : #include "tree-ssa-strlen.h"
49 : #include "tree-hash-traits.h"
50 : #include "builtins.h"
51 : #include "pointer-query.h"
52 : #include "target.h"
53 : #include "diagnostic-core.h"
54 : #include "diagnostic.h"
55 : #include "intl.h"
56 : #include "attribs.h"
57 : #include "calls.h"
58 : #include "cfgloop.h"
59 : #include "tree-ssa-loop.h"
60 : #include "tree-scalar-evolution.h"
61 : #include "vr-values.h"
62 : #include "gimple-range.h"
63 : #include "tree-ssa.h"
64 :
65 : /* A vector indexed by SSA_NAME_VERSION. 0 means unknown, positive value
66 : is an index into strinfo vector, negative value stands for
67 : string length of a string literal (~strlen). */
68 : static vec<int> ssa_ver_to_stridx;
69 :
70 : /* Number of currently active string indexes plus one. */
71 : static int max_stridx;
72 :
73 : /* Set to true to optimize, false when just checking. */
74 : static bool strlen_optimize;
75 :
76 : /* String information record. */
77 : struct strinfo
78 : {
79 : /* Number of leading characters that are known to be nonzero. This is
80 : also the length of the string if FULL_STRING_P.
81 :
82 : The values in a list of related string pointers must be consistent;
83 : that is, if strinfo B comes X bytes after strinfo A, it must be
84 : the case that A->nonzero_chars == X + B->nonzero_chars. */
85 : tree nonzero_chars;
86 : /* Any of the corresponding pointers for querying alias oracle. */
87 : tree ptr;
88 : /* STMT is used for two things:
89 :
90 : - To record the statement that should be used for delayed length
91 : computations. We maintain the invariant that all related strinfos
92 : have delayed lengths or none do.
93 :
94 : - To record the malloc or calloc call that produced this result
95 : to optimize away malloc/memset sequences. STMT is reset after
96 : a calloc-allocated object has been stored a non-zero value into. */
97 : gimple *stmt;
98 : /* Set to the dynamic allocation statement for the object (alloca,
99 : calloc, malloc, or VLA). Unlike STMT, once set for a strinfo
100 : object, ALLOC doesn't change. */
101 : gimple *alloc;
102 : /* Pointer to '\0' if known, if NULL, it can be computed as
103 : ptr + length. */
104 : tree endptr;
105 : /* Reference count. Any changes to strinfo entry possibly shared
106 : with dominating basic blocks need unshare_strinfo first, except
107 : for dont_invalidate which affects only the immediately next
108 : maybe_invalidate. */
109 : int refcount;
110 : /* Copy of index. get_strinfo (si->idx) should return si; */
111 : int idx;
112 : /* These 3 fields are for chaining related string pointers together.
113 : E.g. for
114 : bl = strlen (b); dl = strlen (d); strcpy (a, b); c = a + bl;
115 : strcpy (c, d); e = c + dl;
116 : strinfo(a) -> strinfo(c) -> strinfo(e)
117 : All have ->first field equal to strinfo(a)->idx and are doubly
118 : chained through prev/next fields. The later strinfos are required
119 : to point into the same string with zero or more bytes after
120 : the previous pointer and all bytes in between the two pointers
121 : must be non-zero. Functions like strcpy or memcpy are supposed
122 : to adjust all previous strinfo lengths, but not following strinfo
123 : lengths (those are uncertain, usually invalidated during
124 : maybe_invalidate, except when the alias oracle knows better).
125 : Functions like strcat on the other side adjust the whole
126 : related strinfo chain.
127 : They are updated lazily, so to use the chain the same first fields
128 : and si->prev->next == si->idx needs to be verified. */
129 : int first;
130 : int next;
131 : int prev;
132 : /* A flag whether the string is known to be written in the current
133 : function. */
134 : bool writable;
135 : /* A flag for the next maybe_invalidate that this strinfo shouldn't
136 : be invalidated. Always cleared by maybe_invalidate. */
137 : bool dont_invalidate;
138 : /* True if the string is known to be nul-terminated after NONZERO_CHARS
139 : characters. False is useful when detecting strings that are built
140 : up via successive memcpys. */
141 : bool full_string_p;
142 : };
143 :
144 : /* Pool for allocating strinfo_struct entries. */
145 : static object_allocator<strinfo> strinfo_pool ("strinfo pool");
146 :
147 : /* Vector mapping positive string indexes to strinfo, for the
148 : current basic block. The first pointer in the vector is special,
149 : it is either NULL, meaning the vector isn't shared, or it is
150 : a basic block pointer to the owner basic_block if shared.
151 : If some other bb wants to modify the vector, the vector needs
152 : to be unshared first, and only the owner bb is supposed to free it. */
153 : static vec<strinfo *, va_heap, vl_embed> *stridx_to_strinfo;
154 :
155 : /* One OFFSET->IDX mapping. */
156 : struct stridxlist
157 : {
158 : struct stridxlist *next;
159 : HOST_WIDE_INT offset;
160 : int idx;
161 : };
162 :
163 : /* Hash table entry, mapping a DECL to a chain of OFFSET->IDX mappings. */
164 : struct decl_stridxlist_map
165 : {
166 : struct tree_map_base base;
167 : struct stridxlist list;
168 : };
169 :
170 : /* Hash table for mapping decls to a chained list of offset -> idx
171 : mappings. */
172 : typedef hash_map<tree_decl_hash, stridxlist> decl_to_stridxlist_htab_t;
173 : static decl_to_stridxlist_htab_t *decl_to_stridxlist_htab;
174 :
175 : /* Hash table mapping strlen (or strnlen with constant bound and return
176 : smaller than bound) calls to stridx instances describing
177 : the calls' arguments. Non-null only when warn_stringop_truncation
178 : is non-zero. */
179 : typedef std::pair<int, location_t> stridx_strlenloc;
180 : static hash_map<tree, stridx_strlenloc> *strlen_to_stridx;
181 :
182 : /* Obstack for struct stridxlist and struct decl_stridxlist_map. */
183 : static struct obstack stridx_obstack;
184 :
185 : /* Last memcpy statement if it could be adjusted if the trailing
186 : '\0' written is immediately overwritten, or
187 : *x = '\0' store that could be removed if it is immediately overwritten. */
188 : struct laststmt_struct
189 : {
190 : gimple *stmt;
191 : tree len;
192 : int stridx;
193 : } laststmt;
194 :
195 : static int get_stridx_plus_constant (strinfo *, unsigned HOST_WIDE_INT, tree);
196 : static bool get_range_strlen_dynamic (tree, gimple *, c_strlen_data *,
197 : bitmap, pointer_query *, unsigned *);
198 :
199 : /* Sets MINMAX to either the constant value or the range VAL is in
200 : and returns either the constant value or VAL on success or null
201 : when the range couldn't be determined. Uses RVALS or CFUN for
202 : range info, whichever is nonnull. */
203 :
204 : tree
205 5174280 : get_range (tree val, gimple *stmt, wide_int minmax[2],
206 : range_query *rvals /* = NULL */)
207 : {
208 5174280 : if (!rvals)
209 : {
210 100643 : if (!cfun)
211 : /* When called from front ends for global initializers CFUN
212 : may be null. */
213 : return NULL_TREE;
214 :
215 100637 : rvals = get_range_query (cfun);
216 : }
217 :
218 5174274 : value_range vr (TREE_TYPE (val));
219 5174274 : if (!rvals->range_of_expr (vr, val, stmt))
220 : return NULL_TREE;
221 :
222 5174274 : tree vrmin, vrmax;
223 5174274 : value_range_kind rng = get_legacy_range (vr, vrmin, vrmax);
224 5174274 : if (rng == VR_RANGE)
225 : {
226 : /* Only handle straight ranges. */
227 4978598 : minmax[0] = wi::to_wide (vrmin);
228 4978598 : minmax[1] = wi::to_wide (vrmax);
229 4978598 : return val;
230 : }
231 :
232 : return NULL_TREE;
233 5174274 : }
234 :
235 : class strlen_pass : public dom_walker
236 : {
237 : public:
238 1124058 : strlen_pass (function *fun, cdi_direction direction)
239 : : dom_walker (direction),
240 1124058 : ptr_qry (get_range_query (fun)),
241 1124058 : m_cleanup_cfg (false)
242 : {
243 1124058 : }
244 :
245 : ~strlen_pass ();
246 :
247 : edge before_dom_children (basic_block) final override;
248 : void after_dom_children (basic_block) final override;
249 :
250 : bool check_and_optimize_stmt (bool *cleanup_eh);
251 : bool check_and_optimize_call (bool *zero_write);
252 : bool handle_assign (tree lhs, tree rhs, bool *zero_write);
253 : bool handle_store (bool *zero_write);
254 : void handle_pointer_plus ();
255 : void handle_builtin_strlen ();
256 : void handle_builtin_strchr ();
257 : void handle_builtin_strcpy (built_in_function);
258 : void handle_integral_assign (bool *cleanup_eh);
259 : void handle_builtin_stxncpy_strncat (bool append_p);
260 : void handle_builtin_memcpy (built_in_function bcode);
261 : void handle_builtin_strcat (built_in_function bcode);
262 : void handle_builtin_strncat (built_in_function);
263 : bool handle_builtin_memset (bool *zero_write);
264 : bool handle_builtin_memcmp ();
265 : bool handle_builtin_string_cmp ();
266 : void handle_alloc_call (built_in_function);
267 : void maybe_warn_overflow (gimple *stmt, bool call_lhs, tree len,
268 : strinfo *si = NULL, bool plus_one = false,
269 : bool rawmem = false);
270 : void maybe_warn_overflow (gimple *stmt, bool call_lhs,
271 : unsigned HOST_WIDE_INT len,
272 : strinfo *si = NULL,
273 : bool plus_one = false, bool rawmem = false);
274 : void adjust_last_stmt (strinfo *si, gimple *stmt, bool is_strcat);
275 : tree strxcmp_eqz_result (gimple *stmt, tree arg1, int idx1,
276 : tree arg2, int idx2,
277 : unsigned HOST_WIDE_INT bound,
278 : unsigned HOST_WIDE_INT len[2],
279 : unsigned HOST_WIDE_INT *psize);
280 : bool count_nonzero_bytes (tree expr_or_type,
281 : gimple *stmt,
282 : unsigned lenrange[3], bool *nulterm,
283 : bool *allnul, bool *allnonnul);
284 : bool count_nonzero_bytes (tree exp, tree vuse,
285 : gimple *stmt,
286 : unsigned HOST_WIDE_INT offset,
287 : unsigned HOST_WIDE_INT nbytes,
288 : unsigned lenrange[3], bool *nulterm,
289 : bool *allnul, bool *allnonnul,
290 : ssa_name_limit_t &snlim);
291 : bool count_nonzero_bytes_addr (tree exp, tree vuse,
292 : gimple *stmt,
293 : unsigned HOST_WIDE_INT offset,
294 : unsigned HOST_WIDE_INT nbytes,
295 : unsigned lenrange[3], bool *nulterm,
296 : bool *allnul, bool *allnonnul,
297 : ssa_name_limit_t &snlim);
298 : bool get_len_or_size (gimple *stmt, tree arg, int idx,
299 : unsigned HOST_WIDE_INT lenrng[2],
300 : unsigned HOST_WIDE_INT *size, bool *nulterm);
301 :
302 : /* A pointer_query object to store information about pointers and
303 : their targets in. */
304 : pointer_query ptr_qry;
305 :
306 : gimple_stmt_iterator m_gsi;
307 :
308 : /* Flag that will trigger TODO_cleanup_cfg to be returned in strlen
309 : execute function. */
310 : bool m_cleanup_cfg;
311 : };
312 :
313 : /* Return:
314 :
315 : * +1 if SI is known to start with more than OFF nonzero characters.
316 :
317 : * 0 if SI is known to start with exactly OFF nonzero characters.
318 :
319 : * -1 if SI either does not start with OFF nonzero characters
320 : or the relationship between the number of leading nonzero
321 : characters in SI and OFF is unknown. */
322 :
323 : static int
324 5840 : compare_nonzero_chars (strinfo *si, unsigned HOST_WIDE_INT off)
325 : {
326 5840 : if (si->nonzero_chars
327 5473 : && TREE_CODE (si->nonzero_chars) == INTEGER_CST)
328 5432 : return compare_tree_int (si->nonzero_chars, off);
329 : else
330 : return -1;
331 : }
332 :
333 : /* Same as above but suitable also for strings with non-constant lengths.
334 : Uses RVALS to determine length range. */
335 :
336 : static int
337 92957 : compare_nonzero_chars (strinfo *si, gimple *stmt,
338 : unsigned HOST_WIDE_INT off,
339 : range_query *rvals)
340 : {
341 92957 : if (!si->nonzero_chars)
342 : return -1;
343 :
344 86490 : if (TREE_CODE (si->nonzero_chars) == INTEGER_CST)
345 85996 : return compare_tree_int (si->nonzero_chars, off);
346 :
347 494 : if (!rvals || TREE_CODE (si->nonzero_chars) != SSA_NAME)
348 : return -1;
349 :
350 188 : int_range_max vr;
351 188 : if (!rvals->range_of_expr (vr, si->nonzero_chars, stmt)
352 188 : || vr.varying_p ()
353 376 : || vr.undefined_p ())
354 : return -1;
355 :
356 : /* If the offset is less than the minimum length or if the bounds
357 : of the length range are equal return the result of the comparison
358 : same as in the constant case. Otherwise return a conservative
359 : result. */
360 188 : signop sign = TYPE_SIGN (vr.type ());
361 188 : unsigned prec = TYPE_PRECISION (vr.type ());
362 188 : int cmpmin = wi::cmp (vr.lower_bound (), wi::uhwi (off, prec), sign);
363 188 : if (cmpmin > 0 || vr.singleton_p ())
364 135 : return cmpmin;
365 :
366 : return -1;
367 188 : }
368 :
369 : /* Return true if SI is known to be a zero-length string. */
370 :
371 : static inline bool
372 3575 : zero_length_string_p (strinfo *si)
373 : {
374 3575 : return si->full_string_p && integer_zerop (si->nonzero_chars);
375 : }
376 :
377 : /* Return strinfo vector entry IDX. */
378 :
379 : static inline strinfo *
380 349222 : get_strinfo (int idx)
381 : {
382 349222 : if (vec_safe_length (stridx_to_strinfo) <= (unsigned int) idx)
383 : return NULL;
384 288180 : return (*stridx_to_strinfo)[idx];
385 : }
386 :
387 : /* Get the next strinfo in the chain after SI, or null if none. */
388 :
389 : static inline strinfo *
390 8879 : get_next_strinfo (strinfo *si)
391 : {
392 8879 : if (si->next == 0)
393 : return NULL;
394 6864 : strinfo *nextsi = get_strinfo (si->next);
395 6864 : if (nextsi == NULL || nextsi->first != si->first || nextsi->prev != si->idx)
396 12 : return NULL;
397 : return nextsi;
398 : }
399 :
400 : /* Helper function for get_stridx. Return the strinfo index of the address
401 : of EXP, which is available in PTR if nonnull. If OFFSET_OUT, it is
402 : OK to return the index for some X <= &EXP and store &EXP - X in
403 : *OFFSET_OUT. When RVALS is nonnull uses it to determine range
404 : information. */
405 :
406 : static int
407 1935330 : get_addr_stridx (tree exp, gimple *stmt,
408 : tree ptr, unsigned HOST_WIDE_INT *offset_out,
409 : range_query *rvals = NULL)
410 : {
411 1935330 : HOST_WIDE_INT off;
412 1935330 : struct stridxlist *list, *last = NULL;
413 1935330 : tree base;
414 :
415 1935330 : if (!decl_to_stridxlist_htab)
416 : return 0;
417 :
418 651037 : poly_int64 poff;
419 651037 : base = get_addr_base_and_unit_offset (exp, &poff);
420 651037 : if (base == NULL || !DECL_P (base) || !poff.is_constant (&off))
421 : return 0;
422 :
423 382410 : list = decl_to_stridxlist_htab->get (base);
424 382410 : if (list == NULL)
425 : return 0;
426 :
427 318147 : do
428 : {
429 318147 : if (list->offset == off)
430 : {
431 91501 : if (offset_out)
432 39485 : *offset_out = 0;
433 91501 : return list->idx;
434 : }
435 226646 : if (list->offset > off)
436 : return 0;
437 178877 : last = list;
438 178877 : list = list->next;
439 : }
440 178877 : while (list);
441 :
442 59785 : if ((offset_out || ptr) && last && last->idx > 0)
443 : {
444 59785 : unsigned HOST_WIDE_INT rel_off
445 59785 : = (unsigned HOST_WIDE_INT) off - last->offset;
446 59785 : strinfo *si = get_strinfo (last->idx);
447 59785 : if (si && compare_nonzero_chars (si, stmt, rel_off, rvals) >= 0)
448 : {
449 15381 : if (offset_out)
450 : {
451 13856 : *offset_out = rel_off;
452 13856 : return last->idx;
453 : }
454 : else
455 1525 : return get_stridx_plus_constant (si, rel_off, ptr);
456 : }
457 : }
458 : return 0;
459 : }
460 :
461 : /* Returns string index for EXP. When EXP is an SSA_NAME that refers
462 : to a known strinfo with an offset and OFFRNG is non-null, sets
463 : both elements of the OFFRNG array to the range of the offset and
464 : returns the index of the known strinfo. In this case the result
465 : must not be used in for functions that modify the string.
466 : When nonnull, uses RVALS to determine range information. */
467 :
468 : static int
469 6753265 : get_stridx (tree exp, gimple *stmt,
470 : wide_int offrng[2] = NULL, range_query *rvals = NULL)
471 : {
472 6753265 : if (offrng)
473 30571 : offrng[0] = offrng[1] = wi::zero (TYPE_PRECISION (ptrdiff_type_node));
474 :
475 6753265 : if (TREE_CODE (exp) == SSA_NAME)
476 : {
477 2952825 : if (ssa_ver_to_stridx[SSA_NAME_VERSION (exp)])
478 : return ssa_ver_to_stridx[SSA_NAME_VERSION (exp)];
479 :
480 2773881 : tree e = exp;
481 2773881 : int last_idx = 0;
482 2773881 : HOST_WIDE_INT offset = 0;
483 : /* Follow a chain of at most 5 assignments. */
484 2987721 : for (int i = 0; i < 5; i++)
485 : {
486 2981853 : gimple *def_stmt = SSA_NAME_DEF_STMT (e);
487 2981853 : if (!is_gimple_assign (def_stmt))
488 : return last_idx;
489 :
490 1435082 : tree_code rhs_code = gimple_assign_rhs_code (def_stmt);
491 1435082 : tree ptr, off;
492 :
493 1435082 : if (rhs_code == ADDR_EXPR)
494 : {
495 : /* Handle indices/offsets into VLAs which are implemented
496 : as pointers to arrays. */
497 53036 : ptr = gimple_assign_rhs1 (def_stmt);
498 53036 : ptr = TREE_OPERAND (ptr, 0);
499 :
500 : /* Handle also VLAs of types larger than char. */
501 53036 : if (tree eltsize = TYPE_SIZE_UNIT (TREE_TYPE (ptr)))
502 : {
503 52849 : if (TREE_CODE (ptr) == ARRAY_REF)
504 : {
505 17015 : off = TREE_OPERAND (ptr, 1);
506 17015 : ptr = TREE_OPERAND (ptr, 0);
507 17015 : if (!integer_onep (eltsize))
508 : {
509 : /* Scale the array index by the size of the element
510 : type in the rare case that it's greater than
511 : the typical 1 for char, making sure both operands
512 : have the same type. */
513 9902 : eltsize = fold_convert (ssizetype, eltsize);
514 9902 : off = fold_convert (ssizetype, off);
515 9902 : off = fold_build2 (MULT_EXPR, ssizetype, off, eltsize);
516 : }
517 : }
518 : else
519 35834 : off = integer_zero_node;
520 : }
521 : else
522 : return 0;
523 :
524 52849 : if (TREE_CODE (ptr) != MEM_REF)
525 : return 0;
526 :
527 : /* Add the MEM_REF byte offset. */
528 13803 : tree mem_off = TREE_OPERAND (ptr, 1);
529 13803 : off = fold_build2 (PLUS_EXPR, TREE_TYPE (off), off, mem_off);
530 13803 : ptr = TREE_OPERAND (ptr, 0);
531 : }
532 1382046 : else if (rhs_code == POINTER_PLUS_EXPR)
533 : {
534 514568 : ptr = gimple_assign_rhs1 (def_stmt);
535 514568 : off = gimple_assign_rhs2 (def_stmt);
536 : }
537 : else
538 : return 0;
539 :
540 528371 : if (TREE_CODE (ptr) != SSA_NAME)
541 : return 0;
542 :
543 489347 : if (!tree_fits_shwi_p (off))
544 : {
545 273896 : if (int idx = ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)])
546 24187 : if (offrng)
547 : {
548 : /* Only when requested by setting OFFRNG to non-null,
549 : return the index corresponding to the SSA_NAME.
550 : Do this irrespective of the whether the offset
551 : is known. */
552 323 : if (get_range (off, def_stmt, offrng, rvals))
553 : {
554 : /* When the offset range is known, increment it
555 : it by the constant offset computed in prior
556 : iterations and store it in the OFFRNG array. */
557 210 : offrng[0] += offset;
558 210 : offrng[1] += offset;
559 : }
560 : else
561 : {
562 : /* When the offset range cannot be determined
563 : store [0, SIZE_MAX] and let the caller decide
564 : if the offset matters. */
565 113 : offrng[1] = wi::to_wide (TYPE_MAX_VALUE (sizetype));
566 113 : offrng[0] = wi::zero (offrng[1].get_precision ());
567 : }
568 : return idx;
569 : }
570 : return 0;
571 : }
572 :
573 215451 : HOST_WIDE_INT this_off = tree_to_shwi (off);
574 215451 : if (offrng)
575 : {
576 1814 : offrng[0] += wi::shwi (this_off, offrng->get_precision ());
577 1814 : offrng[1] += offrng[0];
578 : }
579 :
580 215451 : if (this_off < 0)
581 : return last_idx;
582 :
583 215451 : offset = (unsigned HOST_WIDE_INT) offset + this_off;
584 215451 : if (offset < 0)
585 : return last_idx;
586 :
587 215451 : if (int idx = ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)])
588 : {
589 13166 : strinfo *si = get_strinfo (idx);
590 13166 : if (si)
591 : {
592 2704 : if (compare_nonzero_chars (si, offset) >= 0)
593 1611 : return get_stridx_plus_constant (si, offset, exp);
594 :
595 1093 : if (offrng)
596 213840 : last_idx = idx;
597 : }
598 : }
599 213840 : e = ptr;
600 : }
601 :
602 : return last_idx;
603 : }
604 :
605 3800440 : if (TREE_CODE (exp) == ADDR_EXPR)
606 : {
607 1308011 : int idx = get_addr_stridx (TREE_OPERAND (exp, 0), stmt, exp, NULL);
608 1308011 : if (idx != 0)
609 : return idx;
610 : }
611 :
612 3746899 : const char *p = c_getstr (exp);
613 3746899 : if (p)
614 444608 : return ~(int) strlen (p);
615 :
616 : return 0;
617 : }
618 :
619 : /* Return true if strinfo vector is shared with the immediate dominator. */
620 :
621 : static inline bool
622 3659736 : strinfo_shared (void)
623 : {
624 3659736 : return vec_safe_length (stridx_to_strinfo)
625 3659736 : && (*stridx_to_strinfo)[0] != NULL;
626 : }
627 :
628 : /* Unshare strinfo vector that is shared with the immediate dominator. */
629 :
630 : static void
631 373160 : unshare_strinfo_vec (void)
632 : {
633 373160 : strinfo *si;
634 373160 : unsigned int i = 0;
635 :
636 373160 : gcc_assert (strinfo_shared ());
637 373160 : stridx_to_strinfo = vec_safe_copy (stridx_to_strinfo);
638 7187972 : for (i = 1; vec_safe_iterate (stridx_to_strinfo, i, &si); ++i)
639 6814812 : if (si != NULL)
640 1296131 : si->refcount++;
641 373160 : (*stridx_to_strinfo)[0] = NULL;
642 373160 : }
643 :
644 : /* Attempt to create a string index for exp, ADDR_EXPR's operand.
645 : Return a pointer to the location where the string index can
646 : be stored (if 0) or is stored, or NULL if this can't be tracked. */
647 :
648 : static int *
649 182014 : addr_stridxptr (tree exp)
650 : {
651 182014 : HOST_WIDE_INT off;
652 :
653 182014 : poly_int64 poff;
654 182014 : tree base = get_addr_base_and_unit_offset (exp, &poff);
655 182014 : if (base == NULL_TREE || !DECL_P (base) || !poff.is_constant (&off))
656 : return NULL;
657 :
658 156859 : if (!decl_to_stridxlist_htab)
659 : {
660 46973 : decl_to_stridxlist_htab
661 46973 : = new hash_map<tree_decl_hash, stridxlist> (64);
662 46973 : gcc_obstack_init (&stridx_obstack);
663 : }
664 :
665 156859 : bool existed;
666 156859 : stridxlist *list = &decl_to_stridxlist_htab->get_or_insert (base, &existed);
667 156859 : if (existed)
668 : {
669 : int i;
670 : stridxlist *before = NULL;
671 116674 : for (i = 0; i < 32; i++)
672 : {
673 115630 : if (list->offset == off)
674 0 : return &list->idx;
675 115630 : if (list->offset > off && before == NULL)
676 115630 : before = list;
677 115630 : if (list->next == NULL)
678 : break;
679 70878 : list = list->next;
680 : }
681 45796 : if (i == 32)
682 : return NULL;
683 44752 : if (before)
684 : {
685 6048 : list = before;
686 6048 : before = XOBNEW (&stridx_obstack, struct stridxlist);
687 6048 : *before = *list;
688 6048 : list->next = before;
689 6048 : list->offset = off;
690 6048 : list->idx = 0;
691 6048 : return &list->idx;
692 : }
693 38704 : list->next = XOBNEW (&stridx_obstack, struct stridxlist);
694 38704 : list = list->next;
695 : }
696 :
697 149767 : list->next = NULL;
698 149767 : list->offset = off;
699 149767 : list->idx = 0;
700 149767 : return &list->idx;
701 : }
702 :
703 : /* Create a new string index, or return 0 if reached limit. */
704 :
705 : static int
706 901672 : new_stridx (tree exp)
707 : {
708 901672 : int idx;
709 901672 : if (max_stridx >= param_max_tracked_strlens)
710 : return 0;
711 901672 : if (TREE_CODE (exp) == SSA_NAME)
712 : {
713 883516 : if (SSA_NAME_OCCURS_IN_ABNORMAL_PHI (exp))
714 : return 0;
715 883516 : idx = max_stridx++;
716 883516 : ssa_ver_to_stridx[SSA_NAME_VERSION (exp)] = idx;
717 883516 : return idx;
718 : }
719 18156 : if (TREE_CODE (exp) == ADDR_EXPR)
720 : {
721 18122 : int *pidx = addr_stridxptr (TREE_OPERAND (exp, 0));
722 18122 : if (pidx != NULL)
723 : {
724 17957 : gcc_assert (*pidx == 0);
725 17957 : *pidx = max_stridx++;
726 17957 : return *pidx;
727 : }
728 : }
729 : return 0;
730 : }
731 :
732 : /* Like new_stridx, but for ADDR_EXPR's operand instead. */
733 :
734 : static int
735 162797 : new_addr_stridx (tree exp)
736 : {
737 162797 : int *pidx;
738 162797 : if (max_stridx >= param_max_tracked_strlens)
739 : return 0;
740 162797 : pidx = addr_stridxptr (exp);
741 162797 : if (pidx != NULL)
742 : {
743 137858 : gcc_assert (*pidx == 0);
744 137858 : *pidx = max_stridx++;
745 137858 : return *pidx;
746 : }
747 : return 0;
748 : }
749 :
750 : /* Create a new strinfo. */
751 :
752 : static strinfo *
753 999761 : new_strinfo (tree ptr, int idx, tree nonzero_chars, bool full_string_p)
754 : {
755 999761 : strinfo *si = strinfo_pool.allocate ();
756 999761 : si->nonzero_chars = nonzero_chars;
757 999761 : STRIP_USELESS_TYPE_CONVERSION (ptr);
758 999761 : si->ptr = ptr;
759 999761 : si->stmt = NULL;
760 999761 : si->alloc = NULL;
761 999761 : si->endptr = NULL_TREE;
762 999761 : si->refcount = 1;
763 999761 : si->idx = idx;
764 999761 : si->first = 0;
765 999761 : si->prev = 0;
766 999761 : si->next = 0;
767 999761 : si->writable = false;
768 999761 : si->dont_invalidate = false;
769 999761 : si->full_string_p = full_string_p;
770 999761 : return si;
771 : }
772 :
773 : /* Decrease strinfo refcount and free it if not referenced anymore. */
774 :
775 : static inline void
776 14873174 : free_strinfo (strinfo *si)
777 : {
778 14873174 : if (si && --si->refcount == 0)
779 999761 : strinfo_pool.remove (si);
780 14873174 : }
781 :
782 : /* Set strinfo in the vector entry IDX to SI. */
783 :
784 : static inline void
785 1413637 : set_strinfo (int idx, strinfo *si)
786 : {
787 1413637 : if (vec_safe_length (stridx_to_strinfo) && (*stridx_to_strinfo)[0])
788 373160 : unshare_strinfo_vec ();
789 1413637 : if (vec_safe_length (stridx_to_strinfo) <= (unsigned int) idx)
790 986817 : vec_safe_grow_cleared (stridx_to_strinfo, idx + 1, true);
791 1413637 : (*stridx_to_strinfo)[idx] = si;
792 1413637 : }
793 :
794 : /* Return the first strinfo in the related strinfo chain
795 : if all strinfos in between belong to the chain, otherwise NULL. */
796 :
797 : static strinfo *
798 27304 : verify_related_strinfos (strinfo *origsi)
799 : {
800 27304 : strinfo *si = origsi, *psi;
801 :
802 27304 : if (origsi->first == 0)
803 : return NULL;
804 7823 : for (; si->prev; si = psi)
805 : {
806 4805 : if (si->first != origsi->first)
807 : return NULL;
808 4805 : psi = get_strinfo (si->prev);
809 4805 : if (psi == NULL)
810 : return NULL;
811 4805 : if (psi->next != si->idx)
812 : return NULL;
813 : }
814 3018 : if (si->idx != si->first)
815 0 : return NULL;
816 : return si;
817 : }
818 :
819 : /* Set SI's endptr to ENDPTR and compute its length based on SI->ptr.
820 : Use LOC for folding. */
821 :
822 : static void
823 31 : set_endptr_and_length (location_t loc, strinfo *si, tree endptr)
824 : {
825 31 : si->endptr = endptr;
826 31 : si->stmt = NULL;
827 31 : tree start_as_size = fold_convert_loc (loc, size_type_node, si->ptr);
828 31 : tree end_as_size = fold_convert_loc (loc, size_type_node, endptr);
829 31 : si->nonzero_chars = fold_build2_loc (loc, MINUS_EXPR, size_type_node,
830 : end_as_size, start_as_size);
831 31 : si->full_string_p = true;
832 31 : }
833 :
834 : /* Return the string length, or NULL if it can't be computed.
835 : The length may but need not be constant. Instead, it might be
836 : the result of a strlen() call. */
837 :
838 : static tree
839 6065 : get_string_length (strinfo *si)
840 : {
841 : /* If the length has already been computed return it if it's exact
842 : (i.e., the string is nul-terminated at NONZERO_CHARS), or return
843 : null if it isn't. */
844 6065 : if (si->nonzero_chars)
845 6034 : return si->full_string_p ? si->nonzero_chars : NULL;
846 :
847 : /* If the string is the result of one of the built-in calls below
848 : attempt to compute the length from the call statement. */
849 31 : if (si->stmt)
850 : {
851 30 : gimple *stmt = si->stmt, *lenstmt = NULL;
852 30 : tree callee, lhs, fn, tem;
853 30 : location_t loc;
854 30 : gimple_stmt_iterator gsi;
855 :
856 30 : gcc_assert (is_gimple_call (stmt));
857 30 : callee = gimple_call_fndecl (stmt);
858 30 : gcc_assert (callee && fndecl_built_in_p (callee, BUILT_IN_NORMAL));
859 30 : lhs = gimple_call_lhs (stmt);
860 : /* unshare_strinfo is intentionally not called here. The (delayed)
861 : transformation of strcpy or strcat into stpcpy is done at the place
862 : of the former strcpy/strcat call and so can affect all the strinfos
863 : with the same stmt. If they were unshared before and transformation
864 : has been already done, the handling of BUILT_IN_STPCPY{,_CHK} should
865 : just compute the right length. */
866 30 : switch (DECL_FUNCTION_CODE (callee))
867 : {
868 4 : case BUILT_IN_STRCAT:
869 4 : case BUILT_IN_STRCAT_CHK:
870 4 : gsi = gsi_for_stmt (stmt);
871 4 : fn = builtin_decl_implicit (BUILT_IN_STRLEN);
872 4 : gcc_assert (lhs == NULL_TREE);
873 4 : tem = unshare_expr (gimple_call_arg (stmt, 0));
874 4 : lenstmt = gimple_build_call (fn, 1, tem);
875 4 : lhs = make_ssa_name (TREE_TYPE (TREE_TYPE (fn)), lenstmt);
876 4 : gimple_call_set_lhs (lenstmt, lhs);
877 8 : gimple_set_vuse (lenstmt, gimple_vuse (stmt));
878 4 : gsi_insert_before (&gsi, lenstmt, GSI_SAME_STMT);
879 4 : tem = gimple_call_arg (stmt, 0);
880 4 : lhs = gimple_convert_to_ptrofftype (&gsi, true, GSI_SAME_STMT,
881 : gimple_location (stmt), lhs);
882 4 : tem = gimple_build (&gsi, true, GSI_SAME_STMT,
883 : gimple_location (stmt), POINTER_PLUS_EXPR,
884 4 : TREE_TYPE (gimple_call_arg (stmt, 0)),
885 : tem, lhs);
886 4 : gimple_call_set_arg (stmt, 0, tem);
887 4 : lhs = NULL_TREE;
888 : /* FALLTHRU */
889 28 : case BUILT_IN_STRCPY:
890 28 : case BUILT_IN_STRCPY_CHK:
891 28 : gcc_assert (builtin_decl_implicit_p (BUILT_IN_STPCPY));
892 28 : if (gimple_call_num_args (stmt) == 2)
893 : fn = builtin_decl_implicit (BUILT_IN_STPCPY);
894 : else
895 2 : fn = builtin_decl_explicit (BUILT_IN_STPCPY_CHK);
896 28 : gcc_assert (lhs == NULL_TREE);
897 28 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
898 : {
899 0 : fprintf (dump_file, "Optimizing: ");
900 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
901 : }
902 28 : gimple_call_set_fndecl (stmt, fn);
903 28 : lhs = make_ssa_name (TREE_TYPE (TREE_TYPE (fn)), stmt);
904 28 : gimple_call_set_lhs (stmt, lhs);
905 28 : if (DECL_FUNCTION_CODE (callee) == BUILT_IN_STRCAT_CHK)
906 : {
907 1 : tree objsz = gimple_call_lhs (lenstmt);
908 1 : gimple *g
909 1 : = gimple_build_assign (make_ssa_name (TREE_TYPE (objsz)),
910 : MINUS_EXPR, gimple_call_arg (stmt, 2),
911 : objsz);
912 1 : gimple_set_location (g, gimple_location (stmt));
913 1 : gsi_insert_before (&gsi, g, GSI_SAME_STMT);
914 1 : gimple_call_set_arg (stmt, 2, gimple_assign_lhs (g));
915 : }
916 28 : update_stmt (stmt);
917 28 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
918 : {
919 0 : fprintf (dump_file, "into: ");
920 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
921 : }
922 : /* FALLTHRU */
923 28 : case BUILT_IN_STPCPY:
924 28 : case BUILT_IN_STPCPY_CHK:
925 28 : gcc_assert (lhs != NULL_TREE);
926 28 : loc = gimple_location (stmt);
927 28 : set_endptr_and_length (loc, si, lhs);
928 28 : for (strinfo *chainsi = verify_related_strinfos (si);
929 39 : chainsi != NULL;
930 11 : chainsi = get_next_strinfo (chainsi))
931 11 : if (chainsi->nonzero_chars == NULL)
932 3 : set_endptr_and_length (loc, chainsi, lhs);
933 : break;
934 : case BUILT_IN_ALLOCA:
935 : case BUILT_IN_ALLOCA_WITH_ALIGN:
936 : case BUILT_IN_MALLOC:
937 : break;
938 : /* BUILT_IN_CALLOC always has si->nonzero_chars set. */
939 0 : default:
940 0 : gcc_unreachable ();
941 30 : break;
942 : }
943 : }
944 :
945 31 : return si->nonzero_chars;
946 : }
947 :
948 : /* Dump strlen data to FP for statement STMT. When non-null, RVALS
949 : points to the valuation engine used to calculate ranges, and is
950 : used to dump strlen range for non-constant results. */
951 :
952 : DEBUG_FUNCTION void
953 0 : dump_strlen_info (FILE *fp, gimple *stmt, range_query *rvals)
954 : {
955 0 : if (stmt)
956 : {
957 0 : fprintf (fp, "\nDumping strlen pass data after ");
958 0 : print_gimple_expr (fp, stmt, TDF_LINENO);
959 0 : fputc ('\n', fp);
960 : }
961 : else
962 0 : fprintf (fp, "\nDumping strlen pass data\n");
963 :
964 0 : fprintf (fp, "max_stridx = %i\n", max_stridx);
965 0 : fprintf (fp, "ssa_ver_to_stridx has %u elements\n",
966 : ssa_ver_to_stridx.length ());
967 0 : fprintf (fp, "stridx_to_strinfo");
968 0 : if (stridx_to_strinfo)
969 : {
970 0 : fprintf (fp, " has %u elements\n", stridx_to_strinfo->length ());
971 0 : for (unsigned i = 0; i != stridx_to_strinfo->length (); ++i)
972 : {
973 0 : if (strinfo *si = (*stridx_to_strinfo)[i])
974 : {
975 0 : if (!si->idx)
976 0 : continue;
977 0 : fprintf (fp, " idx = %i", si->idx);
978 0 : if (si->ptr)
979 : {
980 0 : fprintf (fp, ", ptr = ");
981 0 : print_generic_expr (fp, si->ptr);
982 : }
983 :
984 0 : if (si->nonzero_chars)
985 : {
986 0 : fprintf (fp, ", nonzero_chars = ");
987 0 : print_generic_expr (fp, si->nonzero_chars);
988 0 : if (TREE_CODE (si->nonzero_chars) == SSA_NAME)
989 : {
990 0 : int_range_max vr;
991 0 : if (rvals)
992 0 : rvals->range_of_expr (vr, si->nonzero_chars,
993 : si->stmt);
994 : else
995 0 : get_range_query (cfun)->range_of_expr (vr,
996 : si->nonzero_chars);
997 0 : vr.dump (fp);
998 0 : }
999 : }
1000 :
1001 0 : fprintf (fp, ", refcount = %i", si->refcount);
1002 0 : if (si->stmt)
1003 : {
1004 0 : fprintf (fp, ", stmt = ");
1005 0 : print_gimple_expr (fp, si->stmt, 0);
1006 : }
1007 0 : if (si->alloc)
1008 : {
1009 0 : fprintf (fp, ", alloc = ");
1010 0 : print_gimple_expr (fp, si->alloc, 0);
1011 : }
1012 0 : if (si->writable)
1013 0 : fprintf (fp, ", writable");
1014 0 : if (si->dont_invalidate)
1015 0 : fprintf (fp, ", dont_invalidate");
1016 0 : if (si->full_string_p)
1017 0 : fprintf (fp, ", full_string_p");
1018 0 : if (strinfo *next = get_next_strinfo (si))
1019 : {
1020 0 : fprintf (fp, ", {");
1021 0 : do
1022 0 : fprintf (fp, "%i%s", next->idx, next->first ? ", " : "");
1023 0 : while ((next = get_next_strinfo (next)));
1024 0 : fprintf (fp, "}");
1025 : }
1026 0 : fputs ("\n", fp);
1027 : }
1028 : }
1029 : }
1030 : else
1031 0 : fprintf (fp, " = null\n");
1032 :
1033 0 : fprintf (fp, "decl_to_stridxlist_htab");
1034 0 : if (decl_to_stridxlist_htab)
1035 : {
1036 0 : fputs ("\n", fp);
1037 0 : typedef decl_to_stridxlist_htab_t::iterator iter_t;
1038 0 : for (iter_t it = decl_to_stridxlist_htab->begin ();
1039 0 : it != decl_to_stridxlist_htab->end (); ++it)
1040 : {
1041 0 : tree decl = (*it).first;
1042 0 : stridxlist *list = &(*it).second;
1043 0 : fprintf (fp, " decl = ");
1044 0 : print_generic_expr (fp, decl);
1045 0 : if (list)
1046 : {
1047 0 : fprintf (fp, ", offsets = {");
1048 0 : for (; list; list = list->next)
1049 0 : fprintf (fp, "%lli%s", (long long) list->offset,
1050 0 : list->next ? ", " : "");
1051 0 : fputs ("}", fp);
1052 : }
1053 0 : fputs ("\n", fp);
1054 : }
1055 : }
1056 : else
1057 0 : fprintf (fp, " = null\n");
1058 :
1059 0 : if (laststmt.stmt)
1060 : {
1061 0 : fprintf (fp, "laststmt = ");
1062 0 : print_gimple_expr (fp, laststmt.stmt, 0);
1063 0 : fprintf (fp, ", len = ");
1064 0 : print_generic_expr (fp, laststmt.len);
1065 0 : fprintf (fp, ", stridx = %i\n", laststmt.stridx);
1066 : }
1067 0 : }
1068 :
1069 : /* Helper of get_range_strlen_dynamic(). See below. */
1070 :
1071 : static bool
1072 3008 : get_range_strlen_phi (tree src, gphi *phi,
1073 : c_strlen_data *pdata, bitmap visited,
1074 : pointer_query *ptr_qry, unsigned *pssa_def_max)
1075 : {
1076 3008 : if (!bitmap_set_bit (visited, SSA_NAME_VERSION (src)))
1077 : return true;
1078 :
1079 2884 : if (*pssa_def_max == 0)
1080 : return false;
1081 :
1082 2874 : --*pssa_def_max;
1083 :
1084 : /* Iterate over the PHI arguments and determine the minimum and maximum
1085 : length/size of each and incorporate them into the overall result. */
1086 12369 : for (unsigned i = 0; i != gimple_phi_num_args (phi); ++i)
1087 : {
1088 9495 : tree arg = gimple_phi_arg_def (phi, i);
1089 9495 : if (arg == gimple_phi_result (phi))
1090 3268 : continue;
1091 :
1092 9495 : c_strlen_data argdata = { };
1093 9495 : if (!get_range_strlen_dynamic (arg, phi, &argdata, visited, ptr_qry,
1094 : pssa_def_max))
1095 : {
1096 323 : pdata->maxlen = build_all_ones_cst (size_type_node);
1097 323 : continue;
1098 : }
1099 :
1100 : /* Set the DECL of an unterminated array this argument refers to
1101 : if one hasn't been found yet. */
1102 9172 : if (!pdata->decl && argdata.decl)
1103 9 : pdata->decl = argdata.decl;
1104 :
1105 12117 : if (!argdata.minlen
1106 9172 : || (integer_zerop (argdata.minlen)
1107 3305 : && (!argdata.maxbound
1108 867 : || integer_all_onesp (argdata.maxbound))
1109 2438 : && integer_all_onesp (argdata.maxlen)))
1110 : {
1111 : /* Set the upper bound of the length to unbounded. */
1112 2945 : pdata->maxlen = build_all_ones_cst (size_type_node);
1113 2945 : continue;
1114 : }
1115 :
1116 : /* Adjust the minimum and maximum length determined so far and
1117 : the upper bound on the array size. */
1118 6227 : if (TREE_CODE (argdata.minlen) == INTEGER_CST
1119 6227 : && (!pdata->minlen
1120 4730 : || tree_int_cst_lt (argdata.minlen, pdata->minlen)))
1121 2941 : pdata->minlen = argdata.minlen;
1122 :
1123 6227 : if (TREE_CODE (argdata.maxlen) == INTEGER_CST
1124 6227 : && (!pdata->maxlen
1125 : || (argdata.maxlen
1126 4856 : && tree_int_cst_lt (pdata->maxlen, argdata.maxlen))))
1127 2173 : pdata->maxlen = argdata.maxlen;
1128 :
1129 6227 : if (!pdata->maxbound
1130 6122 : || TREE_CODE (pdata->maxbound) != INTEGER_CST
1131 10943 : || (argdata.maxbound
1132 4714 : && tree_int_cst_lt (pdata->maxbound, argdata.maxbound)
1133 888 : && !integer_all_onesp (argdata.maxbound)))
1134 2399 : pdata->maxbound = argdata.maxbound;
1135 : }
1136 :
1137 : return true;
1138 : }
1139 :
1140 : /* Return the maximum possible length of the string PTR that's less
1141 : than MAXLEN given the size of the object of subobject it points
1142 : to at the given STMT. MAXLEN is the maximum length of the string
1143 : determined so far. Return null when no such maximum can be
1144 : determined. */
1145 :
1146 : static tree
1147 1834 : get_maxbound (tree ptr, gimple *stmt, offset_int maxlen,
1148 : pointer_query *ptr_qry)
1149 : {
1150 1834 : access_ref aref;
1151 1834 : if (!ptr_qry->get_ref (ptr, stmt, &aref))
1152 : return NULL_TREE;
1153 :
1154 1834 : offset_int sizrem = aref.size_remaining ();
1155 1834 : if (sizrem <= 0)
1156 : return NULL_TREE;
1157 :
1158 1834 : if (sizrem < maxlen)
1159 426 : maxlen = sizrem - 1;
1160 :
1161 : /* Try to determine the maximum from the subobject at the offset.
1162 : This handles MEM [&some-struct, member-offset] that's often
1163 : the result of folding COMPONENT_REF [some-struct, member]. */
1164 1834 : tree reftype = TREE_TYPE (aref.ref);
1165 1834 : if (!RECORD_OR_UNION_TYPE_P (reftype)
1166 9 : || aref.offrng[0] != aref.offrng[1]
1167 1843 : || !wi::fits_shwi_p (aref.offrng[0]))
1168 1825 : return wide_int_to_tree (size_type_node, maxlen);
1169 :
1170 9 : HOST_WIDE_INT off = aref.offrng[0].to_shwi ();
1171 9 : tree fld = field_at_offset (reftype, NULL_TREE, off);
1172 9 : if (!fld || !DECL_SIZE_UNIT (fld))
1173 1 : return wide_int_to_tree (size_type_node, maxlen);
1174 :
1175 8 : offset_int size = wi::to_offset (DECL_SIZE_UNIT (fld));
1176 8 : if (maxlen < size)
1177 0 : return wide_int_to_tree (size_type_node, maxlen);
1178 :
1179 8 : return wide_int_to_tree (size_type_node, size - 1);
1180 : }
1181 :
1182 : /* Attempt to determine the length of the string SRC. On success, store
1183 : the length in *PDATA and return true. Otherwise, return false.
1184 : VISITED is a bitmap of visited PHI nodes. RVALS points to the valuation
1185 : engine used to calculate ranges. PSSA_DEF_MAX to an SSA_NAME
1186 : assignment limit used to prevent runaway recursion. */
1187 :
1188 : static bool
1189 557714 : get_range_strlen_dynamic (tree src, gimple *stmt,
1190 : c_strlen_data *pdata, bitmap visited,
1191 : pointer_query *ptr_qry, unsigned *pssa_def_max)
1192 : {
1193 557714 : int idx = get_stridx (src, stmt);
1194 557714 : if (!idx)
1195 : {
1196 275459 : if (TREE_CODE (src) == SSA_NAME)
1197 : {
1198 269641 : gimple *def_stmt = SSA_NAME_DEF_STMT (src);
1199 269641 : if (gphi *phi = dyn_cast<gphi *>(def_stmt))
1200 3008 : return get_range_strlen_phi (src, phi, pdata, visited, ptr_qry,
1201 3008 : pssa_def_max);
1202 : }
1203 :
1204 : /* Return success regardless of the result and handle *PDATA
1205 : in the caller. */
1206 272451 : get_range_strlen (src, pdata, 1);
1207 272451 : return true;
1208 : }
1209 :
1210 282255 : if (idx < 0)
1211 : {
1212 : /* SRC is a string of constant length. */
1213 276332 : pdata->minlen = build_int_cst (size_type_node, ~idx);
1214 276332 : pdata->maxlen = pdata->minlen;
1215 276332 : pdata->maxbound = pdata->maxlen;
1216 276332 : return true;
1217 : }
1218 :
1219 5923 : if (strinfo *si = get_strinfo (idx))
1220 : {
1221 2064 : pdata->minlen = get_string_length (si);
1222 2064 : if (!pdata->minlen && si->nonzero_chars)
1223 : {
1224 39 : if (TREE_CODE (si->nonzero_chars) == INTEGER_CST)
1225 39 : pdata->minlen = si->nonzero_chars;
1226 0 : else if (TREE_CODE (si->nonzero_chars) == SSA_NAME)
1227 : {
1228 0 : int_range_max vr;
1229 0 : ptr_qry->rvals->range_of_expr (vr, si->nonzero_chars, si->stmt);
1230 0 : if (vr.undefined_p () || vr.varying_p ())
1231 0 : pdata->minlen = build_zero_cst (size_type_node);
1232 : else
1233 : {
1234 0 : tree type = vr.type ();
1235 0 : pdata->minlen = wide_int_to_tree (type, vr.lower_bound ());
1236 : }
1237 0 : }
1238 : else
1239 0 : pdata->minlen = build_zero_cst (size_type_node);
1240 :
1241 39 : tree base = si->ptr;
1242 39 : if (TREE_CODE (base) == ADDR_EXPR)
1243 39 : base = TREE_OPERAND (base, 0);
1244 :
1245 39 : HOST_WIDE_INT off;
1246 39 : poly_int64 poff;
1247 39 : base = get_addr_base_and_unit_offset (base, &poff);
1248 39 : if (base
1249 39 : && DECL_P (base)
1250 39 : && TREE_CODE (TREE_TYPE (base)) == ARRAY_TYPE
1251 37 : && TYPE_SIZE_UNIT (TREE_TYPE (base))
1252 39 : && poff.is_constant (&off))
1253 : {
1254 30 : tree basetype = TREE_TYPE (base);
1255 30 : tree size = TYPE_SIZE_UNIT (basetype);
1256 30 : if (TREE_CODE (size) == INTEGER_CST)
1257 : {
1258 30 : ++off; /* Increment for the terminating nul. */
1259 30 : tree toffset = build_int_cst (size_type_node, off);
1260 30 : pdata->maxlen = fold_build2 (MINUS_EXPR, size_type_node,
1261 : size, toffset);
1262 30 : if (tree_int_cst_lt (pdata->maxlen, pdata->minlen))
1263 : /* This can happen when triggering UB, when base is an
1264 : array which is known to be filled with at least size
1265 : non-zero bytes. E.g. for
1266 : char a[2]; memcpy (a, "12", sizeof a);
1267 : We don't want to create an invalid range [2, 1]
1268 : where 2 comes from the number of non-zero bytes and
1269 : 1 from longest valid zero-terminated string that can
1270 : be stored in such an array, so pick just one of
1271 : those, pdata->minlen. See PR110603. */
1272 3 : pdata->maxlen = build_all_ones_cst (size_type_node);
1273 : else
1274 27 : pdata->maxbound = pdata->maxlen;
1275 : }
1276 : else
1277 0 : pdata->maxlen = build_all_ones_cst (size_type_node);
1278 : }
1279 : else
1280 9 : pdata->maxlen = build_all_ones_cst (size_type_node);
1281 39 : }
1282 2025 : else if (pdata->minlen && TREE_CODE (pdata->minlen) == SSA_NAME)
1283 : {
1284 1834 : int_range_max vr;
1285 1834 : ptr_qry->rvals->range_of_expr (vr, si->nonzero_chars, stmt);
1286 1834 : if (vr.varying_p () || vr.undefined_p ())
1287 : {
1288 0 : pdata->minlen = build_zero_cst (size_type_node);
1289 0 : pdata->maxlen = build_all_ones_cst (size_type_node);
1290 : }
1291 : else
1292 : {
1293 1834 : tree type = vr.type ();
1294 1834 : pdata->minlen = wide_int_to_tree (type, vr.lower_bound ());
1295 1834 : pdata->maxlen = wide_int_to_tree (type, vr.upper_bound ());
1296 1834 : offset_int max = offset_int::from (vr.upper_bound (0), SIGNED);
1297 1834 : if (tree maxbound = get_maxbound (si->ptr, stmt, max, ptr_qry))
1298 : pdata->maxbound = maxbound;
1299 : else
1300 0 : pdata->maxbound = pdata->maxlen;
1301 : }
1302 1834 : }
1303 191 : else if (pdata->minlen && TREE_CODE (pdata->minlen) == INTEGER_CST)
1304 : {
1305 184 : pdata->maxlen = pdata->minlen;
1306 184 : pdata->maxbound = pdata->minlen;
1307 : }
1308 : else
1309 : {
1310 : /* For PDATA->MINLEN that's a non-constant expression such
1311 : as PLUS_EXPR whose value range is unknown, set the bounds
1312 : to zero and SIZE_MAX. */
1313 7 : pdata->minlen = build_zero_cst (size_type_node);
1314 7 : pdata->maxlen = build_all_ones_cst (size_type_node);
1315 : }
1316 :
1317 : return true;
1318 : }
1319 :
1320 : return false;
1321 : }
1322 :
1323 : /* Analogous to get_range_strlen but for dynamically created strings,
1324 : i.e., those created by calls to strcpy as opposed to just string
1325 : constants.
1326 : Try to obtain the range of the lengths of the string(s) referenced
1327 : by SRC, or the size of the largest array SRC refers to if the range
1328 : of lengths cannot be determined, and store all in *PDATA. RVALS
1329 : points to the valuation engine used to calculate ranges. */
1330 :
1331 : void
1332 548219 : get_range_strlen_dynamic (tree src, gimple *stmt, c_strlen_data *pdata,
1333 : pointer_query &ptr_qry)
1334 : {
1335 548219 : auto_bitmap visited;
1336 548219 : tree maxbound = pdata->maxbound;
1337 :
1338 548219 : unsigned limit = param_ssa_name_def_chain_limit;
1339 548219 : if (!get_range_strlen_dynamic (src, stmt, pdata, visited, &ptr_qry, &limit))
1340 : {
1341 : /* On failure extend the length range to an impossible maximum
1342 : (a valid MAXLEN must be less than PTRDIFF_MAX - 1). Other
1343 : members can stay unchanged regardless. */
1344 3546 : pdata->minlen = ssize_int (0);
1345 3546 : pdata->maxlen = build_all_ones_cst (size_type_node);
1346 : }
1347 544673 : else if (!pdata->minlen)
1348 990 : pdata->minlen = ssize_int (0);
1349 :
1350 : /* If it's unchanged from it initial non-null value, set the conservative
1351 : MAXBOUND to SIZE_MAX. Otherwise leave it null (if it is null). */
1352 548219 : if (maxbound && pdata->maxbound == maxbound)
1353 4552 : pdata->maxbound = build_all_ones_cst (size_type_node);
1354 548219 : }
1355 :
1356 : /* Invalidate string length information for strings whose length might
1357 : change due to stores in STMT, except those marked DONT_INVALIDATE.
1358 : For string-modifying statements, ZERO_WRITE is set when the statement
1359 : wrote only zeros.
1360 : Returns true if any STRIDX_TO_STRINFO entries were considered
1361 : for invalidation. */
1362 :
1363 : static bool
1364 12173267 : maybe_invalidate (gimple *stmt, bool zero_write = false)
1365 : {
1366 12173267 : if (dump_file && (dump_flags & TDF_DETAILS))
1367 : {
1368 0 : fprintf (dump_file, "%s called for ", __func__);
1369 0 : print_gimple_stmt (dump_file, stmt, TDF_LINENO);
1370 : }
1371 :
1372 12173267 : strinfo *si;
1373 12173267 : bool nonempty = false;
1374 :
1375 76305564 : for (unsigned i = 1; vec_safe_iterate (stridx_to_strinfo, i, &si); ++i)
1376 : {
1377 64132297 : if (si == NULL || !POINTER_TYPE_P (TREE_TYPE (si->ptr)))
1378 61276982 : continue;
1379 :
1380 3646039 : nonempty = true;
1381 :
1382 : /* Unconditionally reset DONT_INVALIDATE. */
1383 3646039 : bool dont_invalidate = si->dont_invalidate;
1384 3646039 : si->dont_invalidate = false;
1385 :
1386 3646039 : if (dont_invalidate)
1387 376848 : continue;
1388 :
1389 3269191 : ao_ref r;
1390 3269191 : tree size = si->nonzero_chars;
1391 3269191 : ao_ref_init_from_ptr_and_size (&r, si->ptr, size);
1392 : /* Include the terminating nul in the size of the string
1393 : to consider when determining possible clobber. But do not
1394 : add it to 'size' since we don't know whether it would
1395 : actually fit the allocated area. */
1396 3269191 : if (known_size_p (r.size))
1397 : {
1398 2814991 : if (known_le (r.size, HOST_WIDE_INT_MAX - BITS_PER_UNIT))
1399 2814991 : r.max_size += BITS_PER_UNIT;
1400 : else
1401 0 : r.max_size = -1;
1402 : }
1403 3269191 : if (stmt_may_clobber_ref_p_1 (stmt, &r))
1404 : {
1405 413876 : if (dump_file && (dump_flags & TDF_DETAILS))
1406 : {
1407 0 : fputs (" statement may clobber object ", dump_file);
1408 0 : print_generic_expr (dump_file, si->ptr);
1409 0 : if (size && tree_fits_uhwi_p (size))
1410 0 : fprintf (dump_file, " " HOST_WIDE_INT_PRINT_UNSIGNED
1411 : " bytes in size", tree_to_uhwi (size));
1412 0 : fputc ('\n', dump_file);
1413 : }
1414 :
1415 413876 : set_strinfo (i, NULL);
1416 413876 : free_strinfo (si);
1417 413876 : continue;
1418 : }
1419 :
1420 2855315 : if (size
1421 2855315 : && !zero_write
1422 2498797 : && si->stmt
1423 5529 : && is_gimple_call (si->stmt)
1424 2860844 : && (DECL_FUNCTION_CODE (gimple_call_fndecl (si->stmt))
1425 : == BUILT_IN_CALLOC))
1426 : {
1427 : /* If the clobber test above considered the length of
1428 : the string (including the nul), then for (potentially)
1429 : non-zero writes that might modify storage allocated by
1430 : calloc consider the whole object and if it might be
1431 : clobbered by the statement reset the statement. */
1432 5529 : ao_ref_init_from_ptr_and_size (&r, si->ptr, NULL_TREE);
1433 5529 : if (stmt_may_clobber_ref_p_1 (stmt, &r))
1434 63 : si->stmt = NULL;
1435 : }
1436 : }
1437 :
1438 12173267 : if (dump_file && (dump_flags & TDF_DETAILS))
1439 0 : fprintf (dump_file, "%s returns %i\n", __func__, nonempty);
1440 :
1441 12173267 : return nonempty;
1442 : }
1443 :
1444 : /* Unshare strinfo record SI, if it has refcount > 1 or
1445 : if stridx_to_strinfo vector is shared with some other
1446 : bbs. */
1447 :
1448 : static strinfo *
1449 42482 : unshare_strinfo (strinfo *si)
1450 : {
1451 42482 : strinfo *nsi;
1452 :
1453 42482 : if (si->refcount == 1 && !strinfo_shared ())
1454 : return si;
1455 :
1456 9156 : nsi = new_strinfo (si->ptr, si->idx, si->nonzero_chars, si->full_string_p);
1457 9156 : nsi->stmt = si->stmt;
1458 9156 : nsi->alloc = si->alloc;
1459 9156 : nsi->endptr = si->endptr;
1460 9156 : nsi->first = si->first;
1461 9156 : nsi->prev = si->prev;
1462 9156 : nsi->next = si->next;
1463 9156 : nsi->writable = si->writable;
1464 9156 : set_strinfo (si->idx, nsi);
1465 9156 : free_strinfo (si);
1466 9156 : return nsi;
1467 : }
1468 :
1469 : /* Attempt to create a new strinfo for BASESI + OFF, or find existing
1470 : strinfo if there is any. Return it's idx, or 0 if no strinfo has
1471 : been created. */
1472 :
1473 : static int
1474 3136 : get_stridx_plus_constant (strinfo *basesi, unsigned HOST_WIDE_INT off,
1475 : tree ptr)
1476 : {
1477 3136 : if (TREE_CODE (ptr) == SSA_NAME && SSA_NAME_OCCURS_IN_ABNORMAL_PHI (ptr))
1478 : return 0;
1479 :
1480 3136 : if (compare_nonzero_chars (basesi, off) < 0
1481 3136 : || !tree_fits_uhwi_p (basesi->nonzero_chars))
1482 : return 0;
1483 :
1484 3136 : unsigned HOST_WIDE_INT nonzero_chars
1485 3136 : = tree_to_uhwi (basesi->nonzero_chars) - off;
1486 3136 : strinfo *si = basesi, *chainsi;
1487 3136 : if (si->first || si->prev || si->next)
1488 670 : si = verify_related_strinfos (basesi);
1489 670 : if (si == NULL
1490 3136 : || si->nonzero_chars == NULL_TREE
1491 3136 : || TREE_CODE (si->nonzero_chars) != INTEGER_CST)
1492 : return 0;
1493 :
1494 3136 : if (TREE_CODE (ptr) == SSA_NAME
1495 4747 : && ssa_ver_to_stridx.length () <= SSA_NAME_VERSION (ptr))
1496 0 : ssa_ver_to_stridx.safe_grow_cleared (num_ssa_names, true);
1497 :
1498 3136 : gcc_checking_assert (compare_tree_int (si->nonzero_chars, off) != -1);
1499 5444 : for (chainsi = si; chainsi->next; chainsi = si)
1500 : {
1501 2308 : si = get_next_strinfo (chainsi);
1502 2308 : if (si == NULL
1503 2308 : || si->nonzero_chars == NULL_TREE
1504 2308 : || TREE_CODE (si->nonzero_chars) != INTEGER_CST)
1505 : break;
1506 2308 : int r = compare_tree_int (si->nonzero_chars, nonzero_chars);
1507 2308 : if (r != 1)
1508 : {
1509 0 : if (r == 0)
1510 : {
1511 0 : if (TREE_CODE (ptr) == SSA_NAME)
1512 0 : ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)] = si->idx;
1513 : else
1514 : {
1515 0 : int *pidx = addr_stridxptr (TREE_OPERAND (ptr, 0));
1516 0 : if (pidx != NULL && *pidx == 0)
1517 0 : *pidx = si->idx;
1518 : }
1519 0 : return si->idx;
1520 : }
1521 : break;
1522 : }
1523 : }
1524 :
1525 3136 : int idx = new_stridx (ptr);
1526 3136 : if (idx == 0)
1527 : return 0;
1528 3136 : si = new_strinfo (ptr, idx, build_int_cst (size_type_node, nonzero_chars),
1529 : basesi->full_string_p);
1530 3136 : set_strinfo (idx, si);
1531 3136 : if (strinfo *nextsi = get_strinfo (chainsi->next))
1532 : {
1533 0 : nextsi = unshare_strinfo (nextsi);
1534 0 : si->next = nextsi->idx;
1535 0 : nextsi->prev = idx;
1536 : }
1537 3136 : chainsi = unshare_strinfo (chainsi);
1538 3136 : if (chainsi->first == 0)
1539 2466 : chainsi->first = chainsi->idx;
1540 3136 : chainsi->next = idx;
1541 3136 : if (chainsi->endptr == NULL_TREE && zero_length_string_p (si))
1542 259 : chainsi->endptr = ptr;
1543 3136 : si->endptr = chainsi->endptr;
1544 3136 : si->prev = chainsi->idx;
1545 3136 : si->first = chainsi->first;
1546 3136 : si->writable = chainsi->writable;
1547 3136 : return si->idx;
1548 : }
1549 :
1550 : /* Note that PTR, a pointer SSA_NAME initialized in the current stmt, points
1551 : to a zero-length string and if possible chain it to a related strinfo
1552 : chain whose part is or might be CHAINSI. */
1553 :
1554 : static strinfo *
1555 2100 : zero_length_string (tree ptr, strinfo *chainsi)
1556 : {
1557 2100 : strinfo *si;
1558 2100 : int idx;
1559 4200 : if (ssa_ver_to_stridx.length () <= SSA_NAME_VERSION (ptr))
1560 0 : ssa_ver_to_stridx.safe_grow_cleared (num_ssa_names, true);
1561 2100 : gcc_checking_assert (TREE_CODE (ptr) == SSA_NAME
1562 : && ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)] == 0);
1563 :
1564 2100 : if (SSA_NAME_OCCURS_IN_ABNORMAL_PHI (ptr))
1565 : return NULL;
1566 2100 : if (chainsi != NULL)
1567 : {
1568 2100 : si = verify_related_strinfos (chainsi);
1569 2100 : if (si)
1570 : {
1571 267 : do
1572 : {
1573 : /* We shouldn't mix delayed and non-delayed lengths. */
1574 267 : gcc_assert (si->full_string_p);
1575 267 : if (si->endptr == NULL_TREE)
1576 : {
1577 221 : si = unshare_strinfo (si);
1578 221 : si->endptr = ptr;
1579 : }
1580 267 : chainsi = si;
1581 267 : si = get_next_strinfo (si);
1582 : }
1583 267 : while (si != NULL);
1584 109 : if (zero_length_string_p (chainsi))
1585 : {
1586 17 : if (chainsi->next)
1587 : {
1588 0 : chainsi = unshare_strinfo (chainsi);
1589 0 : chainsi->next = 0;
1590 : }
1591 17 : ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)] = chainsi->idx;
1592 17 : return chainsi;
1593 : }
1594 : }
1595 : else
1596 : {
1597 : /* We shouldn't mix delayed and non-delayed lengths. */
1598 1991 : gcc_assert (chainsi->full_string_p);
1599 1991 : if (chainsi->first || chainsi->prev || chainsi->next)
1600 : {
1601 0 : chainsi = unshare_strinfo (chainsi);
1602 0 : chainsi->first = 0;
1603 0 : chainsi->prev = 0;
1604 0 : chainsi->next = 0;
1605 : }
1606 : }
1607 : }
1608 2083 : idx = new_stridx (ptr);
1609 2083 : if (idx == 0)
1610 : return NULL;
1611 2083 : si = new_strinfo (ptr, idx, build_int_cst (size_type_node, 0), true);
1612 2083 : set_strinfo (idx, si);
1613 2083 : si->endptr = ptr;
1614 2083 : if (chainsi != NULL)
1615 : {
1616 2083 : chainsi = unshare_strinfo (chainsi);
1617 2083 : if (chainsi->first == 0)
1618 1991 : chainsi->first = chainsi->idx;
1619 2083 : chainsi->next = idx;
1620 2083 : if (chainsi->endptr == NULL_TREE)
1621 1565 : chainsi->endptr = ptr;
1622 2083 : si->prev = chainsi->idx;
1623 2083 : si->first = chainsi->first;
1624 2083 : si->writable = chainsi->writable;
1625 : }
1626 : return si;
1627 : }
1628 :
1629 : /* For strinfo ORIGSI whose length has been just updated, adjust other
1630 : related strinfos so that they match the new ORIGSI. This involves:
1631 :
1632 : - adding ADJ to the nonzero_chars fields
1633 : - copying full_string_p from the new ORIGSI. */
1634 :
1635 : static void
1636 24185 : adjust_related_strinfos (location_t loc, strinfo *origsi, tree adj)
1637 : {
1638 24185 : strinfo *si = verify_related_strinfos (origsi);
1639 :
1640 24185 : if (si == NULL)
1641 : return;
1642 :
1643 6117 : while (1)
1644 : {
1645 6117 : strinfo *nsi;
1646 :
1647 6117 : if (si != origsi)
1648 : {
1649 4207 : tree tem;
1650 :
1651 4207 : si = unshare_strinfo (si);
1652 : /* We shouldn't see delayed lengths here; the caller must
1653 : have calculated the old length in order to calculate
1654 : the adjustment. */
1655 4207 : gcc_assert (si->nonzero_chars);
1656 4207 : tem = fold_convert_loc (loc, TREE_TYPE (si->nonzero_chars), adj);
1657 4207 : si->nonzero_chars = fold_build2_loc (loc, PLUS_EXPR,
1658 4207 : TREE_TYPE (si->nonzero_chars),
1659 : si->nonzero_chars, tem);
1660 4207 : si->full_string_p = origsi->full_string_p;
1661 :
1662 4207 : si->endptr = NULL_TREE;
1663 4207 : si->dont_invalidate = true;
1664 : }
1665 6117 : nsi = get_next_strinfo (si);
1666 6117 : if (nsi == NULL)
1667 : return;
1668 : si = nsi;
1669 : }
1670 : }
1671 :
1672 : /* Find if there are other SSA_NAME pointers equal to PTR
1673 : for which we don't track their string lengths yet. If so, use
1674 : IDX for them. */
1675 :
1676 : static void
1677 38909 : find_equal_ptrs (tree ptr, int idx)
1678 : {
1679 38909 : if (TREE_CODE (ptr) != SSA_NAME)
1680 : return;
1681 19755 : while (1)
1682 : {
1683 19755 : gimple *stmt = SSA_NAME_DEF_STMT (ptr);
1684 19755 : if (!is_gimple_assign (stmt))
1685 : return;
1686 11569 : ptr = gimple_assign_rhs1 (stmt);
1687 11569 : switch (gimple_assign_rhs_code (stmt))
1688 : {
1689 : case SSA_NAME:
1690 : break;
1691 663 : CASE_CONVERT:
1692 663 : if (!POINTER_TYPE_P (TREE_TYPE (ptr)))
1693 : return;
1694 0 : if (TREE_CODE (ptr) == SSA_NAME)
1695 : break;
1696 0 : if (TREE_CODE (ptr) != ADDR_EXPR)
1697 : return;
1698 : /* FALLTHRU */
1699 1095 : case ADDR_EXPR:
1700 1095 : {
1701 1095 : int *pidx = addr_stridxptr (TREE_OPERAND (ptr, 0));
1702 1095 : if (pidx != NULL && *pidx == 0)
1703 0 : *pidx = idx;
1704 : return;
1705 : }
1706 : default:
1707 : return;
1708 : }
1709 :
1710 : /* We might find an endptr created in this pass. Grow the
1711 : vector in that case. */
1712 0 : if (ssa_ver_to_stridx.length () <= SSA_NAME_VERSION (ptr))
1713 0 : ssa_ver_to_stridx.safe_grow_cleared (num_ssa_names, true);
1714 :
1715 0 : if (ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)] != 0)
1716 : return;
1717 0 : ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)] = idx;
1718 0 : }
1719 : }
1720 :
1721 : /* Return true if STMT is a call to a builtin function with the right
1722 : arguments and attributes that should be considered for optimization
1723 : by this pass. */
1724 :
1725 : static bool
1726 4528337 : valid_builtin_call (gimple *stmt)
1727 : {
1728 4528337 : if (!gimple_call_builtin_p (stmt, BUILT_IN_NORMAL))
1729 : return false;
1730 :
1731 983612 : tree callee = gimple_call_fndecl (stmt);
1732 983612 : switch (DECL_FUNCTION_CODE (callee))
1733 : {
1734 209021 : case BUILT_IN_MEMCMP:
1735 209021 : case BUILT_IN_MEMCMP_EQ:
1736 209021 : case BUILT_IN_STRCMP:
1737 209021 : case BUILT_IN_STRNCMP:
1738 209021 : case BUILT_IN_STRCHR:
1739 209021 : case BUILT_IN_STRLEN:
1740 209021 : case BUILT_IN_STRNLEN:
1741 : /* The above functions should be pure. Punt if they aren't. */
1742 627063 : if (gimple_vdef (stmt) || gimple_vuse (stmt) == NULL_TREE)
1743 0 : return false;
1744 : break;
1745 :
1746 154335 : case BUILT_IN_ALLOCA:
1747 154335 : case BUILT_IN_ALLOCA_WITH_ALIGN:
1748 154335 : case BUILT_IN_CALLOC:
1749 154335 : case BUILT_IN_MALLOC:
1750 154335 : case BUILT_IN_MEMCPY:
1751 154335 : case BUILT_IN_MEMCPY_CHK:
1752 154335 : case BUILT_IN_MEMPCPY:
1753 154335 : case BUILT_IN_MEMPCPY_CHK:
1754 154335 : case BUILT_IN_MEMSET:
1755 154335 : case BUILT_IN_STPCPY:
1756 154335 : case BUILT_IN_STPCPY_CHK:
1757 154335 : case BUILT_IN_STPNCPY:
1758 154335 : case BUILT_IN_STPNCPY_CHK:
1759 154335 : case BUILT_IN_STRCAT:
1760 154335 : case BUILT_IN_STRCAT_CHK:
1761 154335 : case BUILT_IN_STRCPY:
1762 154335 : case BUILT_IN_STRCPY_CHK:
1763 154335 : case BUILT_IN_STRNCAT:
1764 154335 : case BUILT_IN_STRNCAT_CHK:
1765 154335 : case BUILT_IN_STRNCPY:
1766 154335 : case BUILT_IN_STRNCPY_CHK:
1767 : /* The above functions should be neither const nor pure. Punt if they
1768 : aren't. */
1769 308670 : if (gimple_vdef (stmt) == NULL_TREE || gimple_vuse (stmt) == NULL_TREE)
1770 7 : return false;
1771 : break;
1772 :
1773 : default:
1774 : break;
1775 : }
1776 :
1777 : return true;
1778 : }
1779 :
1780 : /* If the last .MEM setter statement before STMT is
1781 : memcpy (x, y, strlen (y) + 1), the only .MEM use of it is STMT
1782 : and STMT is known to overwrite x[strlen (x)], adjust the last memcpy to
1783 : just memcpy (x, y, strlen (y)). SI must be the zero length
1784 : strinfo. */
1785 :
1786 : void
1787 9818 : strlen_pass::adjust_last_stmt (strinfo *si, gimple *stmt, bool is_strcat)
1788 : {
1789 9818 : tree vuse, callee, len;
1790 9818 : struct laststmt_struct last = laststmt;
1791 9818 : strinfo *lastsi, *firstsi;
1792 9818 : unsigned len_arg_no = 2;
1793 :
1794 9818 : laststmt.stmt = NULL;
1795 9818 : laststmt.len = NULL_TREE;
1796 9818 : laststmt.stridx = 0;
1797 :
1798 9818 : if (last.stmt == NULL)
1799 9818 : return;
1800 :
1801 3679 : vuse = gimple_vuse (stmt);
1802 3679 : if (vuse == NULL_TREE
1803 3679 : || SSA_NAME_DEF_STMT (vuse) != last.stmt
1804 4549 : || !has_single_use (vuse))
1805 : return;
1806 :
1807 451 : gcc_assert (last.stridx > 0);
1808 451 : lastsi = get_strinfo (last.stridx);
1809 451 : if (lastsi == NULL)
1810 : return;
1811 :
1812 451 : if (lastsi != si)
1813 : {
1814 340 : if (lastsi->first == 0 || lastsi->first != si->first)
1815 : return;
1816 :
1817 319 : firstsi = verify_related_strinfos (si);
1818 319 : if (firstsi == NULL)
1819 : return;
1820 495 : while (firstsi != lastsi)
1821 : {
1822 176 : firstsi = get_next_strinfo (firstsi);
1823 176 : if (firstsi == NULL)
1824 : return;
1825 : }
1826 : }
1827 :
1828 430 : if (!is_strcat && !zero_length_string_p (si))
1829 : return;
1830 :
1831 349 : if (is_gimple_assign (last.stmt))
1832 : {
1833 35 : gimple_stmt_iterator gsi;
1834 :
1835 35 : if (!integer_zerop (gimple_assign_rhs1 (last.stmt)))
1836 : return;
1837 16 : if (stmt_could_throw_p (cfun, last.stmt))
1838 : return;
1839 16 : gsi = gsi_for_stmt (last.stmt);
1840 16 : unlink_stmt_vdef (last.stmt);
1841 16 : release_defs (last.stmt);
1842 16 : gsi_remove (&gsi, true);
1843 16 : return;
1844 : }
1845 :
1846 314 : if (!valid_builtin_call (last.stmt))
1847 : return;
1848 :
1849 314 : callee = gimple_call_fndecl (last.stmt);
1850 314 : switch (DECL_FUNCTION_CODE (callee))
1851 : {
1852 314 : case BUILT_IN_MEMCPY:
1853 314 : case BUILT_IN_MEMCPY_CHK:
1854 314 : break;
1855 : default:
1856 : return;
1857 : }
1858 :
1859 314 : len = gimple_call_arg (last.stmt, len_arg_no);
1860 314 : if (tree_fits_uhwi_p (len))
1861 : {
1862 237 : if (!tree_fits_uhwi_p (last.len)
1863 237 : || integer_zerop (len)
1864 474 : || tree_to_uhwi (len) != tree_to_uhwi (last.len) + 1)
1865 110 : return;
1866 : /* Don't adjust the length if it is divisible by 4, it is more efficient
1867 : to store the extra '\0' in that case. */
1868 237 : if ((tree_to_uhwi (len) & 3) == 0)
1869 : return;
1870 :
1871 : /* Don't fold away an out of bounds access, as this defeats proper
1872 : warnings. */
1873 138 : tree dst = gimple_call_arg (last.stmt, 0);
1874 :
1875 138 : access_ref aref;
1876 138 : tree size = compute_objsize (dst, stmt, 1, &aref, &ptr_qry);
1877 138 : if (size && tree_int_cst_lt (size, len))
1878 : return;
1879 : }
1880 77 : else if (TREE_CODE (len) == SSA_NAME)
1881 : {
1882 77 : gimple *def_stmt = SSA_NAME_DEF_STMT (len);
1883 77 : if (!is_gimple_assign (def_stmt)
1884 77 : || gimple_assign_rhs_code (def_stmt) != PLUS_EXPR
1885 77 : || gimple_assign_rhs1 (def_stmt) != last.len
1886 154 : || !integer_onep (gimple_assign_rhs2 (def_stmt)))
1887 : return;
1888 : }
1889 : else
1890 : return;
1891 :
1892 204 : gimple_call_set_arg (last.stmt, len_arg_no, last.len);
1893 204 : update_stmt (last.stmt);
1894 : }
1895 :
1896 : /* For an LHS that is an SSA_NAME that is the result of a strlen()
1897 : call, or when BOUND is non-null, of a strnlen() call, set LHS
1898 : range info to [0, min (MAX, BOUND)] when the range includes more
1899 : than one value and return LHS. Otherwise, when the range
1900 : [MIN, MAX] is such that MIN == MAX, return the tree representation
1901 : of (MIN). The latter allows callers to fold suitable strnlen() calls
1902 : to constants. */
1903 :
1904 : tree
1905 160576 : set_strlen_range (tree lhs, wide_int min, wide_int max,
1906 : tree bound /* = NULL_TREE */)
1907 : {
1908 160576 : if (TREE_CODE (lhs) != SSA_NAME
1909 160576 : || !INTEGRAL_TYPE_P (TREE_TYPE (lhs)))
1910 : return NULL_TREE;
1911 :
1912 156450 : if (bound)
1913 : {
1914 : /* For strnlen, adjust MIN and MAX as necessary. If the bound
1915 : is less than the size of the array set MAX to it. It it's
1916 : greater than MAX and MAX is non-zero bump MAX down to account
1917 : for the necessary terminating nul. Otherwise leave it alone. */
1918 518 : if (TREE_CODE (bound) == INTEGER_CST)
1919 : {
1920 325 : wide_int wibnd = wi::to_wide (bound);
1921 325 : int cmp = wi::cmpu (wibnd, max);
1922 325 : if (cmp < 0)
1923 246 : max = wibnd;
1924 133 : else if (cmp && wi::ne_p (max, min))
1925 54 : --max;
1926 325 : }
1927 193 : else if (TREE_CODE (bound) == SSA_NAME)
1928 : {
1929 193 : int_range_max r;
1930 386 : get_range_query (cfun)->range_of_expr (r, bound);
1931 193 : if (!r.undefined_p ())
1932 : {
1933 : /* For a bound in a known range, adjust the range determined
1934 : above as necessary. For a bound in some anti-range or
1935 : in an unknown range, use the range determined by callers. */
1936 193 : if (wi::ltu_p (r.lower_bound (), min))
1937 0 : min = r.lower_bound ();
1938 193 : if (wi::ltu_p (r.upper_bound (), max))
1939 33 : max = r.upper_bound ();
1940 : }
1941 193 : }
1942 : }
1943 :
1944 156450 : if (min == max)
1945 14 : return wide_int_to_tree (size_type_node, min);
1946 :
1947 156436 : int_range_max vr (TREE_TYPE (lhs), min, max);
1948 156436 : set_range_info (lhs, vr);
1949 156436 : return lhs;
1950 156436 : }
1951 :
1952 : /* For an LHS that is an SSA_NAME and for strlen() or strnlen() argument
1953 : SRC, set LHS range info to [0, min (N, BOUND)] if SRC refers to
1954 : a character array A[N] with unknown length bounded by N, and for
1955 : strnlen(), by min (N, BOUND). */
1956 :
1957 : static tree
1958 12737 : maybe_set_strlen_range (tree lhs, tree src, tree bound)
1959 : {
1960 12737 : if (TREE_CODE (lhs) != SSA_NAME
1961 12737 : || !INTEGRAL_TYPE_P (TREE_TYPE (lhs)))
1962 : return NULL_TREE;
1963 :
1964 12737 : if (TREE_CODE (src) == SSA_NAME)
1965 : {
1966 9499 : gimple *def = SSA_NAME_DEF_STMT (src);
1967 9499 : if (is_gimple_assign (def)
1968 9499 : && gimple_assign_rhs_code (def) == ADDR_EXPR)
1969 165 : src = gimple_assign_rhs1 (def);
1970 : }
1971 :
1972 : /* The longest string is PTRDIFF_MAX - 1 bytes including the final
1973 : NUL so that the difference between a pointer to just past it and
1974 : one to its beginning is positive. */
1975 12737 : wide_int max = wi::to_wide (TYPE_MAX_VALUE (ptrdiff_type_node)) - 2;
1976 :
1977 12737 : if (TREE_CODE (src) == ADDR_EXPR)
1978 : {
1979 : /* The last array member of a struct can be bigger than its size
1980 : suggests if it's treated as a poor-man's flexible array member. */
1981 3403 : src = TREE_OPERAND (src, 0);
1982 3403 : if (TREE_CODE (src) != MEM_REF
1983 3403 : && !array_ref_flexible_size_p (src))
1984 : {
1985 2351 : tree type = TREE_TYPE (src);
1986 2351 : tree size = TYPE_SIZE_UNIT (type);
1987 2351 : if (size
1988 2298 : && TREE_CODE (size) == INTEGER_CST
1989 4649 : && !integer_zerop (size))
1990 : {
1991 : /* Even though such uses of strlen would be undefined,
1992 : avoid relying on arrays of arrays in case some genius
1993 : decides to call strlen on an unterminated array element
1994 : that's followed by a terminated one. Likewise, avoid
1995 : assuming that a struct array member is necessarily
1996 : nul-terminated (the nul may be in the member that
1997 : follows). In those cases, assume that the length
1998 : of the string stored in such an array is bounded
1999 : by the size of the enclosing object if one can be
2000 : determined. */
2001 2292 : tree base = get_base_address (src);
2002 2292 : if (VAR_P (base))
2003 : {
2004 2203 : if (tree size = DECL_SIZE_UNIT (base))
2005 2203 : if (size
2006 2203 : && TREE_CODE (size) == INTEGER_CST
2007 2203 : && !POINTER_TYPE_P (TREE_TYPE (base)))
2008 2203 : max = wi::to_wide (size);
2009 : }
2010 : }
2011 :
2012 : /* For strlen() the upper bound above is equal to
2013 : the longest string that can be stored in the array
2014 : (i.e., it accounts for the terminating nul. For
2015 : strnlen() bump up the maximum by one since the array
2016 : need not be nul-terminated. */
2017 2351 : if (!bound && max != 0)
2018 2131 : --max;
2019 : }
2020 : }
2021 :
2022 12737 : wide_int min = wi::zero (max.get_precision ());
2023 12737 : return set_strlen_range (lhs, min, max, bound);
2024 12737 : }
2025 :
2026 : /* Diagnose buffer overflow by a STMT writing LEN + PLUS_ONE bytes,
2027 : either into a region allocated for the object SI when non-null,
2028 : or into an object designated by the LHS of STMT otherwise.
2029 : For a call STMT, when CALL_LHS is set use its left hand side
2030 : as the destination, otherwise use argument zero.
2031 : When nonnull uses RVALS to determine range information.
2032 : RAWMEM may be set by memcpy and other raw memory functions
2033 : to allow accesses across subobject boundaries. */
2034 :
2035 : void
2036 398595 : strlen_pass::maybe_warn_overflow (gimple *stmt, bool call_lhs, tree len,
2037 : strinfo *si, bool plus_one, bool rawmem)
2038 : {
2039 398595 : if (!len || warning_suppressed_p (stmt, OPT_Wstringop_overflow_))
2040 397745 : return;
2041 :
2042 : /* The DECL of the function performing the write if it is done
2043 : by one. */
2044 397307 : tree writefn = NULL_TREE;
2045 : /* The destination expression involved in the store or call STMT. */
2046 397307 : tree dest = NULL_TREE;
2047 :
2048 397307 : if (is_gimple_assign (stmt))
2049 343430 : dest = gimple_assign_lhs (stmt);
2050 53877 : else if (is_gimple_call (stmt))
2051 : {
2052 53877 : if (call_lhs)
2053 32855 : dest = gimple_call_lhs (stmt);
2054 : else
2055 : {
2056 21022 : gcc_assert (gimple_call_builtin_p (stmt, BUILT_IN_NORMAL));
2057 21022 : dest = gimple_call_arg (stmt, 0);
2058 : }
2059 :
2060 53877 : if (!dest)
2061 : return;
2062 53877 : writefn = gimple_call_fndecl (stmt);
2063 : }
2064 : else
2065 : return;
2066 :
2067 397307 : if (warning_suppressed_p (dest, OPT_Wstringop_overflow_))
2068 : return;
2069 :
2070 397135 : const int ostype = rawmem ? 0 : 1;
2071 :
2072 : /* Use maximum precision to avoid overflow in the addition below.
2073 : Make sure all operands have the same precision to keep wide_int
2074 : from ICE'ing. */
2075 :
2076 397135 : access_ref aref;
2077 : /* The size of the destination region (which is smaller than
2078 : the destination object for stores at a non-zero offset). */
2079 397135 : tree destsize = compute_objsize (dest, stmt, ostype, &aref, &ptr_qry);
2080 :
2081 397135 : if (!destsize)
2082 : {
2083 0 : aref.sizrng[0] = 0;
2084 0 : aref.sizrng[1] = wi::to_offset (max_object_size ());
2085 : }
2086 :
2087 : /* Return early if the DESTSIZE size expression is the same as LEN
2088 : and the offset into the destination is zero. This might happen
2089 : in the case of a pair of malloc and memset calls to allocate
2090 : an object and clear it as if by calloc. */
2091 397135 : if (destsize == len && !plus_one
2092 397397 : && aref.offrng[0] == 0 && aref.offrng[0] == aref.offrng[1])
2093 262 : return;
2094 :
2095 1984365 : wide_int rng[2];
2096 396873 : if (!get_range (len, stmt, rng, ptr_qry.rvals))
2097 1188069 : return;
2098 :
2099 395836 : widest_int lenrng[2] =
2100 1187508 : { widest_int::from (rng[0], SIGNED), widest_int::from (rng[1], SIGNED) };
2101 :
2102 395836 : if (plus_one)
2103 : {
2104 1633 : lenrng[0] += 1;
2105 1633 : lenrng[1] += 1;
2106 : }
2107 :
2108 : /* The size of the remaining space in the destination computed
2109 : as the size of the latter minus the offset into it. */
2110 1979180 : widest_int spcrng[2];
2111 395836 : {
2112 395836 : offset_int remrng[2];
2113 395836 : remrng[1] = aref.size_remaining (remrng);
2114 1181756 : spcrng[0] = remrng[0] == -1 ? 0 : widest_int::from (remrng[0], UNSIGNED);
2115 395836 : spcrng[1] = widest_int::from (remrng[1], UNSIGNED);
2116 : }
2117 :
2118 395836 : if (wi::leu_p (lenrng[0], spcrng[0])
2119 673353 : && wi::leu_p (lenrng[1], spcrng[1]))
2120 1184958 : return;
2121 :
2122 120763 : location_t loc = gimple_or_expr_nonartificial_location (stmt, dest);
2123 120763 : bool warned = false;
2124 120763 : if (wi::leu_p (lenrng[0], spcrng[1]))
2125 : {
2126 119592 : if (len != destsize
2127 119592 : && (!si || rawmem || !is_strlen_related_p (si->ptr, len)))
2128 : return;
2129 :
2130 12 : warned = (writefn
2131 6 : ? warning_at (loc, OPT_Wstringop_overflow_,
2132 : "%qD writing one too many bytes into a region "
2133 : "of a size that depends on %<strlen%>",
2134 : writefn)
2135 6 : : warning_at (loc, OPT_Wstringop_overflow_,
2136 : "writing one too many bytes into a region "
2137 : "of a size that depends on %<strlen%>"));
2138 : }
2139 1171 : else if (lenrng[0] == lenrng[1])
2140 : {
2141 1169 : if (spcrng[0] == spcrng[1])
2142 1075 : warned = (writefn
2143 1307 : ? warning_n (loc, OPT_Wstringop_overflow_,
2144 : lenrng[0].to_uhwi (),
2145 : "%qD writing %wu byte into a region "
2146 : "of size %wu",
2147 : "%qD writing %wu bytes into a region "
2148 : "of size %wu",
2149 : writefn, lenrng[0].to_uhwi (),
2150 : spcrng[0].to_uhwi ())
2151 1075 : : warning_n (loc, OPT_Wstringop_overflow_,
2152 : lenrng[0].to_uhwi (),
2153 : "writing %wu byte into a region "
2154 : "of size %wu",
2155 : "writing %wu bytes into a region "
2156 : "of size %wu",
2157 : lenrng[0].to_uhwi (),
2158 : spcrng[0].to_uhwi ()));
2159 : else
2160 94 : warned = (writefn
2161 185 : ? warning_n (loc, OPT_Wstringop_overflow_,
2162 : lenrng[0].to_uhwi (),
2163 : "%qD writing %wu byte into a region "
2164 : "of size between %wu and %wu",
2165 : "%qD writing %wu bytes into a region "
2166 : "of size between %wu and %wu",
2167 : writefn, lenrng[0].to_uhwi (),
2168 : spcrng[0].to_uhwi (), spcrng[1].to_uhwi ())
2169 94 : : warning_n (loc, OPT_Wstringop_overflow_,
2170 : lenrng[0].to_uhwi (),
2171 : "writing %wu byte into a region "
2172 : "of size between %wu and %wu",
2173 : "writing %wu bytes into a region "
2174 : "of size between %wu and %wu",
2175 : lenrng[0].to_uhwi (),
2176 : spcrng[0].to_uhwi (), spcrng[1].to_uhwi ()));
2177 : }
2178 2 : else if (spcrng[0] == spcrng[1])
2179 2 : warned = (writefn
2180 4 : ? warning_at (loc, OPT_Wstringop_overflow_,
2181 : "%qD writing between %wu and %wu bytes "
2182 : "into a region of size %wu",
2183 : writefn, lenrng[0].to_uhwi (),
2184 : lenrng[1].to_uhwi (),
2185 : spcrng[0].to_uhwi ())
2186 2 : : warning_at (loc, OPT_Wstringop_overflow_,
2187 : "writing between %wu and %wu bytes "
2188 : "into a region of size %wu",
2189 : lenrng[0].to_uhwi (),
2190 : lenrng[1].to_uhwi (),
2191 : spcrng[0].to_uhwi ()));
2192 : else
2193 0 : warned = (writefn
2194 0 : ? warning_at (loc, OPT_Wstringop_overflow_,
2195 : "%qD writing between %wu and %wu bytes "
2196 : "into a region of size between %wu and %wu",
2197 : writefn, lenrng[0].to_uhwi (),
2198 : lenrng[1].to_uhwi (),
2199 : spcrng[0].to_uhwi (), spcrng[1].to_uhwi ())
2200 0 : : warning_at (loc, OPT_Wstringop_overflow_,
2201 : "writing between %wu and %wu bytes "
2202 : "into a region of size between %wu and %wu",
2203 : lenrng[0].to_uhwi (),
2204 : lenrng[1].to_uhwi (),
2205 : spcrng[0].to_uhwi (), spcrng[1].to_uhwi ()));
2206 :
2207 1177 : if (!warned)
2208 : return;
2209 :
2210 850 : suppress_warning (stmt, OPT_Wstringop_overflow_);
2211 :
2212 850 : aref.inform_access (access_write_only);
2213 3565635 : }
2214 :
2215 : /* Convenience wrapper for the above. */
2216 :
2217 : void
2218 376395 : strlen_pass::maybe_warn_overflow (gimple *stmt, bool call_lhs,
2219 : unsigned HOST_WIDE_INT len,
2220 : strinfo *si, bool plus_one, bool rawmem)
2221 : {
2222 376395 : tree tlen = build_int_cst (size_type_node, len);
2223 376395 : maybe_warn_overflow (stmt, call_lhs, tlen, si, plus_one, rawmem);
2224 376395 : }
2225 :
2226 : /* Handle a strlen call. If strlen of the argument is known, replace
2227 : the strlen call with the known value, otherwise remember that strlen
2228 : of the argument is stored in the lhs SSA_NAME. */
2229 :
2230 : void
2231 14618 : strlen_pass::handle_builtin_strlen ()
2232 : {
2233 14618 : gimple *stmt = gsi_stmt (m_gsi);
2234 14618 : tree lhs = gimple_call_lhs (stmt);
2235 :
2236 14618 : if (lhs == NULL_TREE)
2237 1856 : return;
2238 :
2239 14618 : location_t loc = gimple_location (stmt);
2240 14618 : tree callee = gimple_call_fndecl (stmt);
2241 14618 : tree src = gimple_call_arg (stmt, 0);
2242 14618 : tree bound = (DECL_FUNCTION_CODE (callee) == BUILT_IN_STRNLEN
2243 14618 : ? gimple_call_arg (stmt, 1) : NULL_TREE);
2244 14618 : int idx = get_stridx (src, stmt);
2245 14618 : if (idx || (bound && integer_zerop (bound)))
2246 : {
2247 3847 : strinfo *si = NULL;
2248 3847 : tree rhs;
2249 :
2250 3847 : if (idx < 0)
2251 108 : rhs = build_int_cst (TREE_TYPE (lhs), ~idx);
2252 3739 : else if (idx == 0)
2253 : rhs = bound;
2254 : else
2255 : {
2256 3708 : rhs = NULL_TREE;
2257 3708 : si = get_strinfo (idx);
2258 3708 : if (si != NULL)
2259 : {
2260 1717 : rhs = get_string_length (si);
2261 : /* For strnlen, if bound is constant, even if si is not known
2262 : to be zero terminated, if we know at least bound bytes are
2263 : not zero, the return value will be bound. */
2264 1717 : if (rhs == NULL_TREE
2265 1717 : && bound != NULL_TREE
2266 18 : && TREE_CODE (bound) == INTEGER_CST
2267 16 : && si->nonzero_chars != NULL_TREE
2268 16 : && TREE_CODE (si->nonzero_chars) == INTEGER_CST
2269 1733 : && tree_int_cst_le (bound, si->nonzero_chars))
2270 : rhs = bound;
2271 : }
2272 : }
2273 1856 : if (rhs != NULL_TREE)
2274 : {
2275 1693 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2276 : {
2277 0 : fprintf (dump_file, "Optimizing: ");
2278 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
2279 : }
2280 1693 : rhs = unshare_expr (rhs);
2281 1693 : if (!useless_type_conversion_p (TREE_TYPE (lhs), TREE_TYPE (rhs)))
2282 0 : rhs = fold_convert_loc (loc, TREE_TYPE (lhs), rhs);
2283 :
2284 1693 : if (bound)
2285 154 : rhs = fold_build2_loc (loc, MIN_EXPR, TREE_TYPE (rhs), rhs, bound);
2286 :
2287 1693 : gimplify_and_update_call_from_tree (&m_gsi, rhs);
2288 1693 : stmt = gsi_stmt (m_gsi);
2289 1693 : update_stmt (stmt);
2290 1693 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2291 : {
2292 0 : fprintf (dump_file, "into: ");
2293 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
2294 : }
2295 :
2296 1693 : if (si != NULL
2297 : /* Don't update anything for strnlen. */
2298 1693 : && bound == NULL_TREE
2299 1524 : && TREE_CODE (si->nonzero_chars) != SSA_NAME
2300 905 : && TREE_CODE (si->nonzero_chars) != INTEGER_CST
2301 1741 : && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (lhs))
2302 : {
2303 48 : si = unshare_strinfo (si);
2304 48 : si->nonzero_chars = lhs;
2305 48 : gcc_assert (si->full_string_p);
2306 : }
2307 :
2308 1693 : if (strlen_to_stridx
2309 1693 : && (bound == NULL_TREE
2310 : /* For strnlen record this only if the call is proven
2311 : to return the same value as strlen would. */
2312 154 : || (TREE_CODE (bound) == INTEGER_CST
2313 58 : && TREE_CODE (rhs) == INTEGER_CST
2314 58 : && tree_int_cst_lt (rhs, bound))))
2315 1551 : strlen_to_stridx->put (lhs, stridx_strlenloc (idx, loc));
2316 :
2317 : return;
2318 : }
2319 : }
2320 12925 : if (SSA_NAME_OCCURS_IN_ABNORMAL_PHI (lhs))
2321 : return;
2322 :
2323 12925 : if (idx == 0)
2324 10771 : idx = new_stridx (src);
2325 : else
2326 : {
2327 2154 : strinfo *si = get_strinfo (idx);
2328 2154 : if (si != NULL)
2329 : {
2330 163 : if (!si->full_string_p && !si->stmt)
2331 : {
2332 : /* Until now we only had a lower bound on the string length.
2333 : Install LHS as the actual length. */
2334 161 : si = unshare_strinfo (si);
2335 161 : tree old = si->nonzero_chars;
2336 161 : si->nonzero_chars = lhs;
2337 161 : si->full_string_p = true;
2338 161 : if (old && TREE_CODE (old) == INTEGER_CST)
2339 : {
2340 158 : old = fold_convert_loc (loc, TREE_TYPE (lhs), old);
2341 158 : tree adj = fold_build2_loc (loc, MINUS_EXPR,
2342 158 : TREE_TYPE (lhs), lhs, old);
2343 158 : adjust_related_strinfos (loc, si, adj);
2344 : /* Use the constant minimum length as the lower bound
2345 : of the non-constant length. */
2346 158 : wide_int min = wi::to_wide (old);
2347 158 : wide_int max
2348 158 : = wi::to_wide (TYPE_MAX_VALUE (ptrdiff_type_node)) - 2;
2349 158 : if (wi::gtu_p (min, max))
2350 1 : max = wi::to_wide (TYPE_MAX_VALUE (TREE_TYPE (lhs)));
2351 158 : set_strlen_range (lhs, min, max);
2352 158 : }
2353 : else
2354 : {
2355 3 : si->first = 0;
2356 3 : si->prev = 0;
2357 3 : si->next = 0;
2358 : }
2359 : }
2360 : return;
2361 : }
2362 : }
2363 12762 : if (idx)
2364 : {
2365 12737 : if (!bound)
2366 : {
2367 : /* Only store the new length information for calls to strlen(),
2368 : not for those to strnlen(). */
2369 12219 : strinfo *si = new_strinfo (src, idx, lhs, true);
2370 12219 : set_strinfo (idx, si);
2371 12219 : find_equal_ptrs (src, idx);
2372 : }
2373 :
2374 : /* For SRC that is an array of N elements, set LHS's range
2375 : to [0, min (N, BOUND)]. A constant return value means
2376 : the range would have consisted of a single value. In
2377 : that case, fold the result into the returned constant. */
2378 12737 : if (tree ret = maybe_set_strlen_range (lhs, src, bound))
2379 12737 : if (TREE_CODE (ret) == INTEGER_CST)
2380 : {
2381 13 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2382 : {
2383 0 : fprintf (dump_file, "Optimizing: ");
2384 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
2385 : }
2386 13 : if (!useless_type_conversion_p (TREE_TYPE (lhs), TREE_TYPE (ret)))
2387 0 : ret = fold_convert_loc (loc, TREE_TYPE (lhs), ret);
2388 13 : gimplify_and_update_call_from_tree (&m_gsi, ret);
2389 13 : stmt = gsi_stmt (m_gsi);
2390 13 : update_stmt (stmt);
2391 13 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2392 : {
2393 0 : fprintf (dump_file, "into: ");
2394 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
2395 : }
2396 : }
2397 :
2398 12737 : if (strlen_to_stridx && !bound)
2399 12217 : strlen_to_stridx->put (lhs, stridx_strlenloc (idx, loc));
2400 : }
2401 : }
2402 :
2403 : /* Handle a strchr call. If strlen of the first argument is known, replace
2404 : the strchr (x, 0) call with the endptr or x + strlen, otherwise remember
2405 : that lhs of the call is endptr and strlen of the argument is endptr - x. */
2406 :
2407 : void
2408 492 : strlen_pass::handle_builtin_strchr ()
2409 : {
2410 492 : gimple *stmt = gsi_stmt (m_gsi);
2411 492 : tree lhs = gimple_call_lhs (stmt);
2412 :
2413 492 : if (lhs == NULL_TREE)
2414 : return;
2415 :
2416 492 : if (!integer_zerop (gimple_call_arg (stmt, 1)))
2417 : return;
2418 :
2419 0 : tree src = gimple_call_arg (stmt, 0);
2420 :
2421 : /* Avoid folding if the first argument is not a nul-terminated array.
2422 : Defer warning until later. */
2423 0 : if (!check_nul_terminated_array (NULL_TREE, src))
2424 : return;
2425 :
2426 0 : int idx = get_stridx (src, stmt);
2427 0 : if (idx)
2428 : {
2429 0 : strinfo *si = NULL;
2430 0 : tree rhs;
2431 :
2432 0 : if (idx < 0)
2433 0 : rhs = build_int_cst (size_type_node, ~idx);
2434 : else
2435 : {
2436 0 : rhs = NULL_TREE;
2437 0 : si = get_strinfo (idx);
2438 0 : if (si != NULL)
2439 0 : rhs = get_string_length (si);
2440 : }
2441 0 : if (rhs != NULL_TREE)
2442 : {
2443 0 : location_t loc = gimple_location (stmt);
2444 :
2445 0 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2446 : {
2447 0 : fprintf (dump_file, "Optimizing: ");
2448 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
2449 : }
2450 0 : if (si != NULL && si->endptr != NULL_TREE)
2451 : {
2452 0 : rhs = unshare_expr (si->endptr);
2453 0 : if (!useless_type_conversion_p (TREE_TYPE (lhs),
2454 0 : TREE_TYPE (rhs)))
2455 0 : rhs = fold_convert_loc (loc, TREE_TYPE (lhs), rhs);
2456 : }
2457 : else
2458 : {
2459 0 : rhs = fold_convert_loc (loc, sizetype, unshare_expr (rhs));
2460 0 : rhs = fold_build2_loc (loc, POINTER_PLUS_EXPR,
2461 0 : TREE_TYPE (src), src, rhs);
2462 0 : if (!useless_type_conversion_p (TREE_TYPE (lhs),
2463 0 : TREE_TYPE (rhs)))
2464 0 : rhs = fold_convert_loc (loc, TREE_TYPE (lhs), rhs);
2465 : }
2466 0 : gimplify_and_update_call_from_tree (&m_gsi, rhs);
2467 0 : stmt = gsi_stmt (m_gsi);
2468 0 : update_stmt (stmt);
2469 0 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2470 : {
2471 0 : fprintf (dump_file, "into: ");
2472 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
2473 : }
2474 0 : if (si != NULL
2475 0 : && si->endptr == NULL_TREE
2476 0 : && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (lhs))
2477 : {
2478 0 : si = unshare_strinfo (si);
2479 0 : si->endptr = lhs;
2480 : }
2481 0 : zero_length_string (lhs, si);
2482 0 : return;
2483 : }
2484 : }
2485 0 : if (SSA_NAME_OCCURS_IN_ABNORMAL_PHI (lhs))
2486 : return;
2487 0 : if (TREE_CODE (src) != SSA_NAME || !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (src))
2488 : {
2489 0 : if (idx == 0)
2490 0 : idx = new_stridx (src);
2491 0 : else if (get_strinfo (idx) != NULL)
2492 : {
2493 0 : zero_length_string (lhs, NULL);
2494 0 : return;
2495 : }
2496 0 : if (idx)
2497 : {
2498 0 : location_t loc = gimple_location (stmt);
2499 0 : tree lhsu = fold_convert_loc (loc, size_type_node, lhs);
2500 0 : tree srcu = fold_convert_loc (loc, size_type_node, src);
2501 0 : tree length = fold_build2_loc (loc, MINUS_EXPR,
2502 : size_type_node, lhsu, srcu);
2503 0 : strinfo *si = new_strinfo (src, idx, length, true);
2504 0 : si->endptr = lhs;
2505 0 : set_strinfo (idx, si);
2506 0 : find_equal_ptrs (src, idx);
2507 0 : zero_length_string (lhs, si);
2508 : }
2509 : }
2510 : else
2511 0 : zero_length_string (lhs, NULL);
2512 : }
2513 :
2514 : /* Handle a strcpy-like ({st{r,p}cpy,__st{r,p}cpy_chk}) call.
2515 : If strlen of the second argument is known, strlen of the first argument
2516 : is the same after this call. Furthermore, attempt to convert it to
2517 : memcpy. Uses RVALS to determine range information. */
2518 :
2519 : void
2520 2811 : strlen_pass::handle_builtin_strcpy (built_in_function bcode)
2521 : {
2522 2811 : int idx, didx;
2523 2811 : tree src, dst, srclen, len, lhs, type, fn, oldlen;
2524 2811 : bool success;
2525 2811 : gimple *stmt = gsi_stmt (m_gsi);
2526 2811 : strinfo *si, *dsi, *olddsi, *zsi;
2527 2811 : location_t loc;
2528 :
2529 2811 : src = gimple_call_arg (stmt, 1);
2530 2811 : dst = gimple_call_arg (stmt, 0);
2531 2811 : lhs = gimple_call_lhs (stmt);
2532 2811 : idx = get_stridx (src, stmt);
2533 2811 : didx = get_stridx (dst, stmt);
2534 2811 : si = NULL;
2535 2811 : if (idx > 0)
2536 1726 : si = get_strinfo (idx);
2537 :
2538 2811 : olddsi = NULL;
2539 2811 : oldlen = NULL_TREE;
2540 2811 : if (didx > 0)
2541 1319 : olddsi = get_strinfo (didx);
2542 1492 : else if (didx < 0)
2543 : return;
2544 :
2545 1319 : if (olddsi != NULL)
2546 483 : adjust_last_stmt (olddsi, stmt, false);
2547 :
2548 2811 : srclen = NULL_TREE;
2549 2811 : if (si != NULL)
2550 1624 : srclen = get_string_length (si);
2551 1187 : else if (idx < 0)
2552 9 : srclen = build_int_cst (size_type_node, ~idx);
2553 :
2554 2811 : maybe_warn_overflow (stmt, false, srclen, olddsi, true);
2555 :
2556 2811 : if (olddsi != NULL)
2557 483 : adjust_last_stmt (olddsi, stmt, false);
2558 :
2559 2811 : loc = gimple_location (stmt);
2560 2811 : if (srclen == NULL_TREE)
2561 1178 : switch (bcode)
2562 : {
2563 865 : case BUILT_IN_STRCPY:
2564 865 : case BUILT_IN_STRCPY_CHK:
2565 865 : if (lhs != NULL_TREE || !builtin_decl_implicit_p (BUILT_IN_STPCPY))
2566 : return;
2567 : break;
2568 313 : case BUILT_IN_STPCPY:
2569 313 : case BUILT_IN_STPCPY_CHK:
2570 313 : if (lhs == NULL_TREE)
2571 : return;
2572 : else
2573 : {
2574 313 : tree lhsuint = fold_convert_loc (loc, size_type_node, lhs);
2575 313 : srclen = fold_convert_loc (loc, size_type_node, dst);
2576 313 : srclen = fold_build2_loc (loc, MINUS_EXPR, size_type_node,
2577 : lhsuint, srclen);
2578 : }
2579 313 : break;
2580 0 : default:
2581 0 : gcc_unreachable ();
2582 : }
2583 :
2584 2258 : if (didx == 0)
2585 : {
2586 1125 : didx = new_stridx (dst);
2587 1125 : if (didx == 0)
2588 : return;
2589 : }
2590 2256 : if (olddsi != NULL)
2591 : {
2592 467 : oldlen = olddsi->nonzero_chars;
2593 467 : dsi = unshare_strinfo (olddsi);
2594 467 : dsi->nonzero_chars = srclen;
2595 467 : dsi->full_string_p = (srclen != NULL_TREE);
2596 : /* Break the chain, so adjust_related_strinfo on later pointers in
2597 : the chain won't adjust this one anymore. */
2598 467 : dsi->next = 0;
2599 467 : dsi->stmt = NULL;
2600 467 : dsi->endptr = NULL_TREE;
2601 : }
2602 : else
2603 : {
2604 1789 : dsi = new_strinfo (dst, didx, srclen, srclen != NULL_TREE);
2605 1789 : set_strinfo (didx, dsi);
2606 1789 : find_equal_ptrs (dst, didx);
2607 : }
2608 2256 : dsi->writable = true;
2609 2256 : dsi->dont_invalidate = true;
2610 :
2611 2256 : if (dsi->nonzero_chars == NULL_TREE)
2612 : {
2613 312 : strinfo *chainsi;
2614 :
2615 : /* If string length of src is unknown, use delayed length
2616 : computation. If string length of dst will be needed, it
2617 : can be computed by transforming this strcpy call into
2618 : stpcpy and subtracting dst from the return value. */
2619 :
2620 : /* Look for earlier strings whose length could be determined if
2621 : this strcpy is turned into an stpcpy. */
2622 :
2623 312 : if (dsi->prev != 0 && (chainsi = verify_related_strinfos (dsi)) != NULL)
2624 : {
2625 5 : for (; chainsi && chainsi != dsi; chainsi = get_strinfo (chainsi->next))
2626 : {
2627 : /* When setting a stmt for delayed length computation
2628 : prevent all strinfos through dsi from being
2629 : invalidated. */
2630 3 : chainsi = unshare_strinfo (chainsi);
2631 3 : chainsi->stmt = stmt;
2632 3 : chainsi->nonzero_chars = NULL_TREE;
2633 3 : chainsi->full_string_p = false;
2634 3 : chainsi->endptr = NULL_TREE;
2635 3 : chainsi->dont_invalidate = true;
2636 : }
2637 : }
2638 312 : dsi->stmt = stmt;
2639 :
2640 : /* Try to detect overlap before returning. This catches cases
2641 : like strcpy (d, d + n) where n is non-constant whose range
2642 : is such that (n <= strlen (d) holds).
2643 :
2644 : OLDDSI->NONZERO_chars may have been reset by this point with
2645 : oldlen holding it original value. */
2646 312 : if (olddsi && oldlen)
2647 : {
2648 : /* Add 1 for the terminating NUL. */
2649 111 : tree type = TREE_TYPE (oldlen);
2650 111 : oldlen = fold_build2 (PLUS_EXPR, type, oldlen,
2651 : build_int_cst (type, 1));
2652 111 : check_bounds_or_overlap (stmt, olddsi->ptr, src, oldlen, NULL_TREE);
2653 : }
2654 :
2655 : return;
2656 : }
2657 :
2658 1944 : if (olddsi != NULL)
2659 : {
2660 356 : tree adj = NULL_TREE;
2661 356 : if (oldlen == NULL_TREE)
2662 : ;
2663 74 : else if (integer_zerop (oldlen))
2664 : adj = srclen;
2665 61 : else if (TREE_CODE (oldlen) == INTEGER_CST
2666 3 : || TREE_CODE (srclen) == INTEGER_CST)
2667 116 : adj = fold_build2_loc (loc, MINUS_EXPR,
2668 58 : TREE_TYPE (srclen), srclen,
2669 58 : fold_convert_loc (loc, TREE_TYPE (srclen),
2670 : oldlen));
2671 71 : if (adj != NULL_TREE)
2672 71 : adjust_related_strinfos (loc, dsi, adj);
2673 : else
2674 285 : dsi->prev = 0;
2675 : }
2676 : /* strcpy src may not overlap dst, so src doesn't need to be
2677 : invalidated either. */
2678 1944 : if (si != NULL)
2679 1624 : si->dont_invalidate = true;
2680 :
2681 1944 : fn = NULL_TREE;
2682 1944 : zsi = NULL;
2683 1944 : switch (bcode)
2684 : {
2685 1528 : case BUILT_IN_STRCPY:
2686 1528 : fn = builtin_decl_implicit (BUILT_IN_MEMCPY);
2687 1528 : if (lhs)
2688 94 : ssa_ver_to_stridx[SSA_NAME_VERSION (lhs)] = didx;
2689 : break;
2690 1 : case BUILT_IN_STRCPY_CHK:
2691 1 : fn = builtin_decl_explicit (BUILT_IN_MEMCPY_CHK);
2692 1 : if (lhs)
2693 0 : ssa_ver_to_stridx[SSA_NAME_VERSION (lhs)] = didx;
2694 : break;
2695 343 : case BUILT_IN_STPCPY:
2696 : /* This would need adjustment of the lhs (subtract one),
2697 : or detection that the trailing '\0' doesn't need to be
2698 : written, if it will be immediately overwritten.
2699 : fn = builtin_decl_explicit (BUILT_IN_MEMPCPY); */
2700 343 : if (lhs)
2701 : {
2702 343 : dsi->endptr = lhs;
2703 343 : zsi = zero_length_string (lhs, dsi);
2704 : }
2705 : break;
2706 72 : case BUILT_IN_STPCPY_CHK:
2707 : /* This would need adjustment of the lhs (subtract one),
2708 : or detection that the trailing '\0' doesn't need to be
2709 : written, if it will be immediately overwritten.
2710 : fn = builtin_decl_explicit (BUILT_IN_MEMPCPY_CHK); */
2711 72 : if (lhs)
2712 : {
2713 72 : dsi->endptr = lhs;
2714 72 : zsi = zero_length_string (lhs, dsi);
2715 : }
2716 : break;
2717 0 : default:
2718 0 : gcc_unreachable ();
2719 : }
2720 509 : if (zsi != NULL)
2721 415 : zsi->dont_invalidate = true;
2722 :
2723 1944 : if (fn)
2724 : {
2725 1529 : tree args = TYPE_ARG_TYPES (TREE_TYPE (fn));
2726 1529 : type = TREE_VALUE (TREE_CHAIN (TREE_CHAIN (args)));
2727 : }
2728 : else
2729 415 : type = size_type_node;
2730 :
2731 1944 : len = fold_convert_loc (loc, type, unshare_expr (srclen));
2732 1944 : len = fold_build2_loc (loc, PLUS_EXPR, type, len, build_int_cst (type, 1));
2733 :
2734 : /* Disable warning for the transformed statement? */
2735 1944 : opt_code no_warning_opt = no_warning;
2736 :
2737 1944 : if (const strinfo *chksi = si ? olddsi ? olddsi : dsi : NULL)
2738 : {
2739 1624 : no_warning_opt = check_bounds_or_overlap (stmt, chksi->ptr, si->ptr,
2740 : NULL_TREE, len);
2741 1624 : if (no_warning_opt)
2742 247 : suppress_warning (stmt, no_warning_opt);
2743 : }
2744 :
2745 1944 : if (fn == NULL_TREE)
2746 : return;
2747 :
2748 1529 : len = force_gimple_operand_gsi (&m_gsi, len, true, NULL_TREE, true,
2749 : GSI_SAME_STMT);
2750 1529 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2751 : {
2752 0 : fprintf (dump_file, "Optimizing: ");
2753 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
2754 : }
2755 1529 : if (gimple_call_num_args (stmt) == 2)
2756 1528 : success = update_gimple_call (&m_gsi, fn, 3, dst, src, len);
2757 : else
2758 1 : success = update_gimple_call (&m_gsi, fn, 4, dst, src, len,
2759 : gimple_call_arg (stmt, 2));
2760 1529 : if (success)
2761 : {
2762 1529 : stmt = gsi_stmt (m_gsi);
2763 1529 : update_stmt (stmt);
2764 1529 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2765 : {
2766 0 : fprintf (dump_file, "into: ");
2767 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
2768 : }
2769 : /* Allow adjust_last_stmt to decrease this memcpy's size. */
2770 1529 : laststmt.stmt = stmt;
2771 1529 : laststmt.len = srclen;
2772 1529 : laststmt.stridx = dsi->idx;
2773 : }
2774 0 : else if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2775 0 : fprintf (dump_file, "not possible.\n");
2776 :
2777 1529 : if (no_warning_opt)
2778 247 : suppress_warning (stmt, no_warning_opt);
2779 : }
2780 :
2781 : /* Check the size argument to the built-in forms of stpncpy and strncpy
2782 : for out-of-bounds offsets or overlapping access, and to see if the
2783 : size argument is derived from a call to strlen() on the source argument,
2784 : and if so, issue an appropriate warning. */
2785 :
2786 : void
2787 805 : strlen_pass::handle_builtin_strncat (built_in_function)
2788 : {
2789 : /* Same as stxncpy(). */
2790 805 : handle_builtin_stxncpy_strncat (true);
2791 805 : }
2792 :
2793 : /* Return true if LEN depends on a call to strlen(SRC) in an interesting
2794 : way. LEN can either be an integer expression, or a pointer (to char).
2795 : When it is the latter (such as in recursive calls to self) it is
2796 : assumed to be the argument in some call to strlen() whose relationship
2797 : to SRC is being ascertained. */
2798 :
2799 : bool
2800 875 : is_strlen_related_p (tree src, tree len)
2801 : {
2802 2523 : if (POINTER_TYPE_P (TREE_TYPE (len))
2803 1333 : && operand_equal_p (src, len, 0))
2804 : return true;
2805 :
2806 1196 : if (TREE_CODE (len) != SSA_NAME)
2807 : return false;
2808 :
2809 826 : if (TREE_CODE (src) == SSA_NAME)
2810 : {
2811 578 : gimple *srcdef = SSA_NAME_DEF_STMT (src);
2812 578 : if (is_gimple_assign (srcdef))
2813 : {
2814 : /* Handle bitwise AND used in conversions from wider size_t
2815 : to narrower unsigned types. */
2816 244 : tree_code code = gimple_assign_rhs_code (srcdef);
2817 244 : if (code == BIT_AND_EXPR
2818 244 : || code == NOP_EXPR)
2819 151 : return is_strlen_related_p (gimple_assign_rhs1 (srcdef), len);
2820 :
2821 : return false;
2822 : }
2823 :
2824 334 : if (gimple_call_builtin_p (srcdef, BUILT_IN_NORMAL))
2825 : {
2826 : /* If SRC is the result of a call to an allocation function
2827 : or strlen, use the function's argument instead. */
2828 97 : tree func = gimple_call_fndecl (srcdef);
2829 97 : built_in_function code = DECL_FUNCTION_CODE (func);
2830 97 : if (code == BUILT_IN_ALLOCA
2831 97 : || code == BUILT_IN_ALLOCA_WITH_ALIGN
2832 16 : || code == BUILT_IN_MALLOC
2833 16 : || code == BUILT_IN_STRLEN)
2834 95 : return is_strlen_related_p (gimple_call_arg (srcdef, 0), len);
2835 :
2836 : /* FIXME: Handle other functions with attribute alloc_size. */
2837 : return false;
2838 : }
2839 : }
2840 :
2841 485 : gimple *lendef = SSA_NAME_DEF_STMT (len);
2842 485 : if (!lendef)
2843 : return false;
2844 :
2845 485 : if (is_gimple_call (lendef))
2846 : {
2847 68 : tree func = gimple_call_fndecl (lendef);
2848 68 : if (!valid_builtin_call (lendef)
2849 68 : || DECL_FUNCTION_CODE (func) != BUILT_IN_STRLEN)
2850 : return false;
2851 :
2852 68 : tree arg = gimple_call_arg (lendef, 0);
2853 68 : return is_strlen_related_p (src, arg);
2854 : }
2855 :
2856 417 : if (!is_gimple_assign (lendef))
2857 : return false;
2858 :
2859 213 : tree_code code = gimple_assign_rhs_code (lendef);
2860 213 : tree rhs1 = gimple_assign_rhs1 (lendef);
2861 213 : tree rhstype = TREE_TYPE (rhs1);
2862 :
2863 213 : if ((POINTER_TYPE_P (rhstype) && code == POINTER_PLUS_EXPR)
2864 213 : || (INTEGRAL_TYPE_P (rhstype)
2865 213 : && (code == BIT_AND_EXPR
2866 213 : || code == NOP_EXPR)))
2867 : {
2868 : /* Pointer plus (an integer), and truncation are considered among
2869 : the (potentially) related expressions to strlen. */
2870 : return is_strlen_related_p (src, rhs1);
2871 : }
2872 :
2873 81 : if (tree rhs2 = gimple_assign_rhs2 (lendef))
2874 : {
2875 : /* Integer subtraction is considered strlen-related when both
2876 : arguments are integers and second one is strlen-related. */
2877 78 : rhstype = TREE_TYPE (rhs2);
2878 78 : if (INTEGRAL_TYPE_P (rhstype) && code == MINUS_EXPR)
2879 : return is_strlen_related_p (src, rhs2);
2880 : }
2881 :
2882 : return false;
2883 : }
2884 :
2885 : /* Called by handle_builtin_stxncpy_strncat and by
2886 : gimple_fold_builtin_strncpy in gimple-fold.cc.
2887 : Check to see if the specified bound is a) equal to the size of
2888 : the destination DST and if so, b) if it's immediately followed by
2889 : DST[CNT - 1] = '\0'. If a) holds and b) does not, warn. Otherwise,
2890 : do nothing. Return true if diagnostic has been issued.
2891 :
2892 : The purpose is to diagnose calls to strncpy and stpncpy that do
2893 : not nul-terminate the copy while allowing for the idiom where
2894 : such a call is immediately followed by setting the last element
2895 : to nul, as in:
2896 : char a[32];
2897 : strncpy (a, s, sizeof a);
2898 : a[sizeof a - 1] = '\0';
2899 : */
2900 :
2901 : bool
2902 3637 : maybe_diag_stxncpy_trunc (gimple_stmt_iterator gsi, tree src, tree cnt,
2903 : pointer_query *ptr_qry /* = NULL */)
2904 : {
2905 3637 : gimple *stmt = gsi_stmt (gsi);
2906 3637 : if (warning_suppressed_p (stmt, OPT_Wstringop_truncation))
2907 : return false;
2908 :
2909 18095 : wide_int cntrange[2];
2910 3619 : int_range_max r;
2911 7238 : if (!get_range_query (cfun)->range_of_expr (r, cnt)
2912 3619 : || r.varying_p ()
2913 6794 : || r.undefined_p ())
2914 : return false;
2915 :
2916 3175 : tree min, max;
2917 3175 : value_range_kind kind = get_legacy_range (r, min, max);
2918 3175 : cntrange[0] = wi::to_wide (min);
2919 3175 : cntrange[1] = wi::to_wide (max);
2920 3175 : if (kind == VR_ANTI_RANGE)
2921 : {
2922 178 : wide_int maxobjsize = wi::to_wide (TYPE_MAX_VALUE (ptrdiff_type_node));
2923 :
2924 178 : if (wi::ltu_p (cntrange[1], maxobjsize))
2925 : {
2926 0 : cntrange[0] = cntrange[1] + 1;
2927 0 : cntrange[1] = maxobjsize;
2928 : }
2929 : else
2930 : {
2931 178 : cntrange[1] = cntrange[0] - 1;
2932 178 : cntrange[0] = wi::zero (TYPE_PRECISION (TREE_TYPE (cnt)));
2933 : }
2934 178 : }
2935 :
2936 : /* Negative value is the constant string length. If it's less than
2937 : the lower bound there is no truncation. Avoid calling get_stridx()
2938 : when ssa_ver_to_stridx is empty. That implies the caller isn't
2939 : running under the control of this pass and ssa_ver_to_stridx hasn't
2940 : been created yet. */
2941 3175 : int sidx = ssa_ver_to_stridx.length () ? get_stridx (src, stmt) : 0;
2942 2480 : if (sidx < 0 && wi::gtu_p (cntrange[0], ~sidx))
2943 199 : return false;
2944 :
2945 2976 : tree dst = gimple_call_arg (stmt, 0);
2946 2976 : tree dstdecl = dst;
2947 2976 : if (TREE_CODE (dstdecl) == ADDR_EXPR)
2948 1115 : dstdecl = TREE_OPERAND (dstdecl, 0);
2949 :
2950 2976 : tree ref = NULL_TREE;
2951 :
2952 2976 : if (!sidx)
2953 : {
2954 : /* If the source is a non-string return early to avoid warning
2955 : for possible truncation (if the truncation is certain SIDX
2956 : is non-zero). */
2957 2401 : tree srcdecl = gimple_call_arg (stmt, 1);
2958 2401 : if (TREE_CODE (srcdecl) == ADDR_EXPR)
2959 1231 : srcdecl = TREE_OPERAND (srcdecl, 0);
2960 2401 : if (get_attr_nonstring_decl (srcdecl, &ref))
2961 : return false;
2962 : }
2963 :
2964 : /* Likewise, if the destination refers to an array/pointer declared
2965 : nonstring return early. */
2966 2566 : if (get_attr_nonstring_decl (dstdecl, &ref))
2967 : return false;
2968 :
2969 : /* Look for dst[i] = '\0'; after the stxncpy() call and if found
2970 : avoid the truncation warning. */
2971 1956 : gsi_next_nondebug (&gsi);
2972 1956 : gimple *next_stmt = gsi_stmt (gsi);
2973 1956 : if (!next_stmt)
2974 : {
2975 : /* When there is no statement in the same basic block check
2976 : the immediate successor block. */
2977 56 : if (basic_block bb = gimple_bb (stmt))
2978 : {
2979 31 : if (single_succ_p (bb))
2980 : {
2981 : /* For simplicity, ignore blocks with multiple outgoing
2982 : edges for now and only consider successor blocks along
2983 : normal edges. */
2984 31 : edge e = EDGE_SUCC (bb, 0);
2985 31 : if (!(e->flags & EDGE_ABNORMAL))
2986 : {
2987 31 : gsi = gsi_start_bb (e->dest);
2988 31 : next_stmt = gsi_stmt (gsi);
2989 31 : if (next_stmt && is_gimple_debug (next_stmt))
2990 : {
2991 22 : gsi_next_nondebug (&gsi);
2992 22 : next_stmt = gsi_stmt (gsi);
2993 : }
2994 : }
2995 : }
2996 : }
2997 : }
2998 :
2999 1956 : if (next_stmt && is_gimple_assign (next_stmt))
3000 : {
3001 473 : tree lhs = gimple_assign_lhs (next_stmt);
3002 473 : tree_code code = TREE_CODE (lhs);
3003 473 : if (code == ARRAY_REF || code == MEM_REF)
3004 125 : lhs = TREE_OPERAND (lhs, 0);
3005 :
3006 473 : tree func = gimple_call_fndecl (stmt);
3007 473 : if (DECL_FUNCTION_CODE (func) == BUILT_IN_STPNCPY)
3008 : {
3009 65 : tree ret = gimple_call_lhs (stmt);
3010 65 : if (ret && operand_equal_p (ret, lhs, 0))
3011 113 : return false;
3012 : }
3013 :
3014 : /* Determine the base address and offset of the reference,
3015 : ignoring the innermost array index. */
3016 457 : if (TREE_CODE (ref) == ARRAY_REF)
3017 61 : ref = TREE_OPERAND (ref, 0);
3018 :
3019 457 : poly_int64 dstoff;
3020 457 : tree dstbase = get_addr_base_and_unit_offset (ref, &dstoff);
3021 :
3022 457 : poly_int64 lhsoff;
3023 457 : tree lhsbase = get_addr_base_and_unit_offset (lhs, &lhsoff);
3024 457 : if (lhsbase
3025 457 : && dstbase
3026 454 : && known_eq (dstoff, lhsoff)
3027 831 : && operand_equal_p (dstbase, lhsbase, 0))
3028 : return false;
3029 : }
3030 :
3031 1843 : int prec = TYPE_PRECISION (TREE_TYPE (cnt));
3032 9215 : wide_int lenrange[2];
3033 1843 : if (strinfo *sisrc = sidx > 0 ? get_strinfo (sidx) : NULL)
3034 : {
3035 109 : lenrange[0] = (sisrc->nonzero_chars
3036 109 : && TREE_CODE (sisrc->nonzero_chars) == INTEGER_CST
3037 218 : ? wi::to_wide (sisrc->nonzero_chars)
3038 9 : : wi::zero (prec));
3039 109 : lenrange[1] = lenrange[0];
3040 : }
3041 1734 : else if (sidx < 0)
3042 375 : lenrange[0] = lenrange[1] = wi::shwi (~sidx, prec);
3043 : else
3044 : {
3045 1359 : c_strlen_data lendata = { };
3046 : /* Set MAXBOUND to an arbitrary non-null non-integer node as a request
3047 : to have it set to the length of the longest string in a PHI. */
3048 1359 : lendata.maxbound = src;
3049 1359 : get_range_strlen (src, &lendata, /* eltsize = */1);
3050 1359 : if (TREE_CODE (lendata.minlen) == INTEGER_CST
3051 1359 : && TREE_CODE (lendata.maxbound) == INTEGER_CST)
3052 : {
3053 : /* When LENDATA.MAXLEN is unknown, reset LENDATA.MINLEN
3054 : which stores the length of the shortest known string. */
3055 1359 : if (integer_all_onesp (lendata.maxlen))
3056 613 : lenrange[0] = wi::shwi (0, prec);
3057 : else
3058 746 : lenrange[0] = wi::to_wide (lendata.minlen, prec);
3059 1359 : lenrange[1] = wi::to_wide (lendata.maxbound, prec);
3060 : }
3061 : else
3062 : {
3063 0 : lenrange[0] = wi::shwi (0, prec);
3064 0 : lenrange[1] = wi::shwi (-1, prec);
3065 : }
3066 : }
3067 :
3068 1843 : location_t callloc = gimple_or_expr_nonartificial_location (stmt, dst);
3069 1843 : tree func = gimple_call_fndecl (stmt);
3070 :
3071 1843 : if (lenrange[0] != 0 || !wi::neg_p (lenrange[1]))
3072 : {
3073 : /* If the longest source string is shorter than the lower bound
3074 : of the specified count the copy is definitely nul-terminated. */
3075 1261 : if (wi::ltu_p (lenrange[1], cntrange[0]))
3076 : return false;
3077 :
3078 916 : if (wi::neg_p (lenrange[1]))
3079 : {
3080 : /* The length of one of the strings is unknown but at least
3081 : one has non-zero length and that length is stored in
3082 : LENRANGE[1]. Swap the bounds to force a "may be truncated"
3083 : warning below. */
3084 0 : lenrange[1] = lenrange[0];
3085 0 : lenrange[0] = wi::shwi (0, prec);
3086 : }
3087 :
3088 : /* Set to true for strncat whose bound is derived from the length
3089 : of the destination (the expected usage pattern). */
3090 916 : bool cat_dstlen_bounded = false;
3091 916 : if (DECL_FUNCTION_CODE (func) == BUILT_IN_STRNCAT)
3092 160 : cat_dstlen_bounded = is_strlen_related_p (dst, cnt);
3093 :
3094 1128 : if (lenrange[0] == cntrange[1] && cntrange[0] == cntrange[1])
3095 203 : return warning_n (callloc, OPT_Wstringop_truncation,
3096 : cntrange[0].to_uhwi (),
3097 : "%qD output truncated before terminating "
3098 : "nul copying %E byte from a string of the "
3099 : "same length",
3100 : "%qD output truncated before terminating nul "
3101 : "copying %E bytes from a string of the same "
3102 : "length",
3103 : func, cnt);
3104 713 : else if (!cat_dstlen_bounded)
3105 : {
3106 709 : if (wi::geu_p (lenrange[0], cntrange[1]))
3107 : {
3108 : /* The shortest string is longer than the upper bound of
3109 : the count so the truncation is certain. */
3110 345 : if (cntrange[0] == cntrange[1])
3111 304 : return warning_n (callloc, OPT_Wstringop_truncation,
3112 : cntrange[0].to_uhwi (),
3113 : "%qD output truncated copying %E byte "
3114 : "from a string of length %wu",
3115 : "%qD output truncated copying %E bytes "
3116 : "from a string of length %wu",
3117 : func, cnt, lenrange[0].to_uhwi ());
3118 :
3119 41 : return warning_at (callloc, OPT_Wstringop_truncation,
3120 : "%qD output truncated copying between %wu "
3121 : "and %wu bytes from a string of length %wu",
3122 : func, cntrange[0].to_uhwi (),
3123 : cntrange[1].to_uhwi (), lenrange[0].to_uhwi ());
3124 : }
3125 364 : else if (wi::geu_p (lenrange[1], cntrange[1]))
3126 : {
3127 : /* The longest string is longer than the upper bound of
3128 : the count so the truncation is possible. */
3129 80 : if (cntrange[0] == cntrange[1])
3130 60 : return warning_n (callloc, OPT_Wstringop_truncation,
3131 : cntrange[0].to_uhwi (),
3132 : "%qD output may be truncated copying %E "
3133 : "byte from a string of length %wu",
3134 : "%qD output may be truncated copying %E "
3135 : "bytes from a string of length %wu",
3136 : func, cnt, lenrange[1].to_uhwi ());
3137 :
3138 20 : return warning_at (callloc, OPT_Wstringop_truncation,
3139 : "%qD output may be truncated copying between "
3140 : "%wu and %wu bytes from a string of length %wu",
3141 : func, cntrange[0].to_uhwi (),
3142 : cntrange[1].to_uhwi (), lenrange[1].to_uhwi ());
3143 : }
3144 : }
3145 :
3146 284 : if (!cat_dstlen_bounded
3147 284 : && cntrange[0] != cntrange[1]
3148 284 : && wi::leu_p (cntrange[0], lenrange[0])
3149 555 : && wi::leu_p (cntrange[1], lenrange[0] + 1))
3150 : {
3151 : /* If the source (including the terminating nul) is longer than
3152 : the lower bound of the specified count but shorter than the
3153 : upper bound the copy may (but need not) be truncated. */
3154 20 : return warning_at (callloc, OPT_Wstringop_truncation,
3155 : "%qD output may be truncated copying between "
3156 : "%wu and %wu bytes from a string of length %wu",
3157 : func, cntrange[0].to_uhwi (),
3158 : cntrange[1].to_uhwi (), lenrange[0].to_uhwi ());
3159 : }
3160 : }
3161 :
3162 850 : access_ref aref;
3163 850 : if (tree dstsize = compute_objsize (dst, stmt, 1, &aref, ptr_qry))
3164 : {
3165 : /* The source length is unknown. Try to determine the destination
3166 : size and see if it matches the specified bound. If not, bail.
3167 : Otherwise go on to see if it should be diagnosed for possible
3168 : truncation. */
3169 850 : if (!dstsize)
3170 : return false;
3171 :
3172 850 : if (wi::to_wide (dstsize) != cntrange[1])
3173 : return false;
3174 :
3175 : /* Avoid warning for strncpy(a, b, N) calls where the following
3176 : equalities hold:
3177 : N == sizeof a && N == sizeof b */
3178 264 : if (tree srcsize = compute_objsize (src, stmt, 1, &aref, ptr_qry))
3179 264 : if (wi::to_wide (srcsize) == cntrange[1])
3180 : return false;
3181 :
3182 250 : if (cntrange[0] == cntrange[1])
3183 193 : return warning_at (callloc, OPT_Wstringop_truncation,
3184 : "%qD specified bound %E equals destination size",
3185 : func, cnt);
3186 : }
3187 :
3188 : return false;
3189 16386 : }
3190 :
3191 : /* Check the arguments to the built-in forms of stpncpy, strncpy, and
3192 : strncat, for out-of-bounds offsets or overlapping access, and to see
3193 : if the size is derived from calling strlen() on the source argument,
3194 : and if so, issue the appropriate warning.
3195 : APPEND_P is true for strncat. */
3196 :
3197 : void
3198 3300 : strlen_pass::handle_builtin_stxncpy_strncat (bool append_p)
3199 : {
3200 3300 : if (!strlen_to_stridx)
3201 3196 : return;
3202 :
3203 3257 : gimple *stmt = gsi_stmt (m_gsi);
3204 :
3205 3257 : tree dst = gimple_call_arg (stmt, 0);
3206 3257 : tree src = gimple_call_arg (stmt, 1);
3207 3257 : tree len = gimple_call_arg (stmt, 2);
3208 : /* An upper bound of the size of the destination. */
3209 3257 : tree dstsize = NULL_TREE;
3210 : /* The length of the destination and source strings (plus 1 for those
3211 : whose FULL_STRING_P is set, i.e., whose length is exact rather than
3212 : a lower bound). */
3213 3257 : tree dstlenp1 = NULL_TREE, srclenp1 = NULL_TREE;;
3214 :
3215 3257 : int didx = get_stridx (dst, stmt);
3216 3257 : if (strinfo *sidst = didx > 0 ? get_strinfo (didx) : NULL)
3217 : {
3218 : /* Compute the size of the destination string including the nul
3219 : if it is known to be nul-terminated. */
3220 310 : if (sidst->nonzero_chars)
3221 : {
3222 268 : if (sidst->full_string_p)
3223 : {
3224 : /* String is known to be nul-terminated. */
3225 250 : tree type = TREE_TYPE (sidst->nonzero_chars);
3226 250 : dstlenp1 = fold_build2 (PLUS_EXPR, type, sidst->nonzero_chars,
3227 : build_int_cst (type, 1));
3228 : }
3229 : else
3230 : dstlenp1 = sidst->nonzero_chars;
3231 : }
3232 42 : else if (TREE_CODE (sidst->ptr) == SSA_NAME)
3233 : {
3234 42 : gimple *def_stmt = SSA_NAME_DEF_STMT (sidst->ptr);
3235 42 : dstsize = gimple_call_alloc_size (def_stmt);
3236 : }
3237 :
3238 310 : dst = sidst->ptr;
3239 : }
3240 :
3241 3257 : int sidx = get_stridx (src, stmt);
3242 3257 : strinfo *sisrc = sidx > 0 ? get_strinfo (sidx) : NULL;
3243 373 : if (sisrc)
3244 : {
3245 : /* strncat() and strncpy() can modify the source string by writing
3246 : over the terminating nul so SISRC->DONT_INVALIDATE must be left
3247 : clear. */
3248 :
3249 : /* Compute the size of the source string including the terminating
3250 : nul if its known to be nul-terminated. */
3251 342 : if (sisrc->nonzero_chars)
3252 : {
3253 342 : if (sisrc->full_string_p)
3254 : {
3255 334 : tree type = TREE_TYPE (sisrc->nonzero_chars);
3256 334 : srclenp1 = fold_build2 (PLUS_EXPR, type, sisrc->nonzero_chars,
3257 : build_int_cst (type, 1));
3258 : }
3259 : else
3260 : srclenp1 = sisrc->nonzero_chars;
3261 : }
3262 :
3263 342 : src = sisrc->ptr;
3264 : }
3265 : else
3266 : srclenp1 = NULL_TREE;
3267 :
3268 3257 : opt_code opt = check_bounds_or_overlap (stmt, dst, src, dstlenp1, srclenp1);
3269 3257 : if (opt != no_warning)
3270 : {
3271 211 : suppress_warning (stmt, opt);
3272 211 : return;
3273 : }
3274 :
3275 : /* If the length argument was computed from strlen(S) for some string
3276 : S retrieve the strinfo index for the string (PSS->FIRST) along with
3277 : the location of the strlen() call (PSS->SECOND). */
3278 3046 : stridx_strlenloc *pss = strlen_to_stridx->get (len);
3279 3046 : if (!pss || pss->first <= 0)
3280 : {
3281 2942 : if (maybe_diag_stxncpy_trunc (m_gsi, src, len))
3282 241 : suppress_warning (stmt, OPT_Wstringop_truncation);
3283 :
3284 : return;
3285 : }
3286 :
3287 : /* Retrieve the strinfo data for the string S that LEN was computed
3288 : from as some function F of strlen (S) (i.e., LEN need not be equal
3289 : to strlen(S)). */
3290 104 : strinfo *silen = get_strinfo (pss->first);
3291 :
3292 104 : location_t callloc = gimple_or_expr_nonartificial_location (stmt, dst);
3293 :
3294 104 : tree func = gimple_call_fndecl (stmt);
3295 :
3296 104 : bool warned = false;
3297 :
3298 : /* When -Wstringop-truncation is set, try to determine truncation
3299 : before diagnosing possible overflow. Truncation is implied by
3300 : the LEN argument being equal to strlen(SRC), regardless of
3301 : whether its value is known. Otherwise, when appending, or
3302 : when copying into a destination of known size, issue the more
3303 : generic -Wstringop-overflow which triggers for LEN arguments
3304 : that in any meaningful way depend on strlen(SRC). */
3305 104 : if (!append_p
3306 104 : && sisrc == silen
3307 60 : && is_strlen_related_p (src, len)
3308 148 : && warning_at (callloc, OPT_Wstringop_truncation,
3309 : "%qD output truncated before terminating nul "
3310 : "copying as many bytes from a string as its length",
3311 : func))
3312 : warned = true;
3313 1 : else if ((append_p || !dstsize || len == dstlenp1)
3314 76 : && silen && is_strlen_related_p (src, silen->ptr))
3315 : {
3316 : /* Issue -Wstringop-overflow when appending or when writing into
3317 : a destination of a known size. Otherwise, when copying into
3318 : a destination of an unknown size, it's truncation. */
3319 31 : opt_code opt = (append_p || dstsize
3320 75 : ? OPT_Wstringop_overflow_ : OPT_Wstringop_truncation);
3321 75 : warned = warning_at (callloc, opt,
3322 : "%qD specified bound depends on the length "
3323 : "of the source argument",
3324 : func);
3325 : }
3326 103 : if (warned)
3327 : {
3328 83 : location_t strlenloc = pss->second;
3329 83 : if (strlenloc != UNKNOWN_LOCATION && strlenloc != callloc)
3330 83 : inform (strlenloc, "length computed here");
3331 : }
3332 : }
3333 :
3334 : /* Handle a memcpy-like ({mem{,p}cpy,__mem{,p}cpy_chk}) call.
3335 : If strlen of the second argument is known and length of the third argument
3336 : is that plus one, strlen of the first argument is the same after this
3337 : call. Uses RVALS to determine range information. */
3338 :
3339 : void
3340 81636 : strlen_pass::handle_builtin_memcpy (built_in_function bcode)
3341 : {
3342 81636 : tree lhs, oldlen, newlen;
3343 81636 : gimple *stmt = gsi_stmt (m_gsi);
3344 81636 : strinfo *si, *dsi;
3345 :
3346 81636 : tree len = gimple_call_arg (stmt, 2);
3347 81636 : tree src = gimple_call_arg (stmt, 1);
3348 81636 : tree dst = gimple_call_arg (stmt, 0);
3349 :
3350 81636 : int didx = get_stridx (dst, stmt);
3351 81636 : if (didx < 0)
3352 : return;
3353 81636 : int idx = get_stridx (src, stmt);
3354 81636 : strinfo *olddsi = NULL;
3355 81636 : if (didx > 0)
3356 26293 : olddsi = get_strinfo (didx);
3357 :
3358 26293 : if (olddsi != NULL
3359 26293 : && !integer_zerop (len))
3360 : {
3361 18647 : maybe_warn_overflow (stmt, false, len, olddsi, false, true);
3362 18647 : if (tree_fits_uhwi_p (len))
3363 6745 : adjust_last_stmt (olddsi, stmt, false);
3364 : }
3365 :
3366 81636 : if (idx == 0)
3367 : return;
3368 :
3369 26004 : bool full_string_p;
3370 26004 : if (idx > 0)
3371 : {
3372 9237 : gimple *def_stmt;
3373 :
3374 : /* Handle memcpy (x, y, l) where l's relationship with strlen (y)
3375 : is known. */
3376 9237 : si = get_strinfo (idx);
3377 9237 : if (si == NULL || si->nonzero_chars == NULL_TREE)
3378 : return;
3379 3983 : if (TREE_CODE (len) == INTEGER_CST
3380 1564 : && TREE_CODE (si->nonzero_chars) == INTEGER_CST)
3381 : {
3382 1552 : if (tree_int_cst_le (len, si->nonzero_chars))
3383 : {
3384 : /* Copying LEN nonzero characters, where LEN is constant. */
3385 : newlen = len;
3386 : full_string_p = false;
3387 : }
3388 : else
3389 : {
3390 : /* Copying the whole of the analyzed part of SI. */
3391 225 : newlen = si->nonzero_chars;
3392 225 : full_string_p = si->full_string_p;
3393 : }
3394 : }
3395 : else
3396 : {
3397 2431 : if (!si->full_string_p)
3398 : return;
3399 2395 : if (TREE_CODE (len) != SSA_NAME)
3400 : return;
3401 2383 : def_stmt = SSA_NAME_DEF_STMT (len);
3402 2383 : if (!is_gimple_assign (def_stmt)
3403 231 : || gimple_assign_rhs_code (def_stmt) != PLUS_EXPR
3404 114 : || gimple_assign_rhs1 (def_stmt) != si->nonzero_chars
3405 2487 : || !integer_onep (gimple_assign_rhs2 (def_stmt)))
3406 : return;
3407 : /* Copying variable-length string SI (and no more). */
3408 104 : newlen = si->nonzero_chars;
3409 104 : full_string_p = true;
3410 : }
3411 : }
3412 : else
3413 : {
3414 16767 : si = NULL;
3415 : /* Handle memcpy (x, "abcd", 5) or
3416 : memcpy (x, "abc\0uvw", 7). */
3417 16767 : if (!tree_fits_uhwi_p (len))
3418 : return;
3419 :
3420 16687 : unsigned HOST_WIDE_INT clen = tree_to_uhwi (len);
3421 16687 : unsigned HOST_WIDE_INT nonzero_chars = ~idx;
3422 16687 : newlen = build_int_cst (size_type_node, MIN (nonzero_chars, clen));
3423 16687 : full_string_p = clen > nonzero_chars;
3424 : }
3425 :
3426 18343 : if (!full_string_p
3427 18343 : && olddsi
3428 3413 : && olddsi->nonzero_chars
3429 391 : && TREE_CODE (olddsi->nonzero_chars) == INTEGER_CST
3430 18734 : && tree_int_cst_le (newlen, olddsi->nonzero_chars))
3431 : {
3432 : /* The SRC substring being written strictly overlaps
3433 : a subsequence of the existing string OLDDSI. */
3434 140 : newlen = olddsi->nonzero_chars;
3435 140 : full_string_p = olddsi->full_string_p;
3436 : }
3437 :
3438 18343 : if (olddsi != NULL && TREE_CODE (len) == SSA_NAME)
3439 75 : adjust_last_stmt (olddsi, stmt, false);
3440 :
3441 18343 : if (didx == 0)
3442 : {
3443 11309 : didx = new_stridx (dst);
3444 11309 : if (didx == 0)
3445 : return;
3446 : }
3447 18340 : oldlen = NULL_TREE;
3448 18340 : if (olddsi != NULL)
3449 : {
3450 4910 : dsi = unshare_strinfo (olddsi);
3451 4910 : oldlen = olddsi->nonzero_chars;
3452 4910 : dsi->nonzero_chars = newlen;
3453 4910 : dsi->full_string_p = full_string_p;
3454 : /* Break the chain, so adjust_related_strinfo on later pointers in
3455 : the chain won't adjust this one anymore. */
3456 4910 : dsi->next = 0;
3457 4910 : dsi->stmt = NULL;
3458 4910 : dsi->endptr = NULL_TREE;
3459 : }
3460 : else
3461 : {
3462 13430 : dsi = new_strinfo (dst, didx, newlen, full_string_p);
3463 13430 : set_strinfo (didx, dsi);
3464 13430 : find_equal_ptrs (dst, didx);
3465 : }
3466 18340 : dsi->writable = true;
3467 18340 : dsi->dont_invalidate = true;
3468 18340 : if (olddsi != NULL)
3469 : {
3470 4910 : tree adj = NULL_TREE;
3471 4910 : location_t loc = gimple_location (stmt);
3472 4910 : if (oldlen == NULL_TREE)
3473 : ;
3474 1275 : else if (integer_zerop (oldlen))
3475 : adj = newlen;
3476 562 : else if (TREE_CODE (oldlen) == INTEGER_CST
3477 6 : || TREE_CODE (newlen) == INTEGER_CST)
3478 562 : adj = fold_build2_loc (loc, MINUS_EXPR, TREE_TYPE (newlen), newlen,
3479 562 : fold_convert_loc (loc, TREE_TYPE (newlen),
3480 : oldlen));
3481 1275 : if (adj != NULL_TREE)
3482 1275 : adjust_related_strinfos (loc, dsi, adj);
3483 : else
3484 3635 : dsi->prev = 0;
3485 : }
3486 : /* memcpy src may not overlap dst, so src doesn't need to be
3487 : invalidated either. */
3488 18340 : if (si != NULL)
3489 1656 : si->dont_invalidate = true;
3490 :
3491 18340 : if (full_string_p)
3492 : {
3493 3401 : lhs = gimple_call_lhs (stmt);
3494 3401 : switch (bcode)
3495 : {
3496 3345 : case BUILT_IN_MEMCPY:
3497 3345 : case BUILT_IN_MEMCPY_CHK:
3498 : /* Allow adjust_last_stmt to decrease this memcpy's size. */
3499 3345 : laststmt.stmt = stmt;
3500 3345 : laststmt.len = dsi->nonzero_chars;
3501 3345 : laststmt.stridx = dsi->idx;
3502 3345 : if (lhs)
3503 176 : ssa_ver_to_stridx[SSA_NAME_VERSION (lhs)] = didx;
3504 : break;
3505 : case BUILT_IN_MEMPCPY:
3506 : case BUILT_IN_MEMPCPY_CHK:
3507 : break;
3508 0 : default:
3509 0 : gcc_unreachable ();
3510 : }
3511 : }
3512 : }
3513 :
3514 : /* Handle a strcat-like ({strcat,__strcat_chk}) call.
3515 : If strlen of the second argument is known, strlen of the first argument
3516 : is increased by the length of the second argument. Furthermore, attempt
3517 : to convert it to memcpy/strcpy if the length of the first argument
3518 : is known. */
3519 :
3520 : void
3521 809 : strlen_pass::handle_builtin_strcat (built_in_function bcode)
3522 : {
3523 809 : int idx, didx;
3524 809 : tree srclen, args, type, fn, objsz, endptr;
3525 809 : bool success;
3526 809 : gimple *stmt = gsi_stmt (m_gsi);
3527 809 : strinfo *si, *dsi;
3528 809 : location_t loc = gimple_location (stmt);
3529 :
3530 809 : tree src = gimple_call_arg (stmt, 1);
3531 809 : tree dst = gimple_call_arg (stmt, 0);
3532 :
3533 : /* Bail if the source is the same as destination. It will be diagnosed
3534 : elsewhere. */
3535 809 : if (operand_equal_p (src, dst, 0))
3536 : return;
3537 :
3538 770 : tree lhs = gimple_call_lhs (stmt);
3539 :
3540 770 : didx = get_stridx (dst, stmt);
3541 770 : if (didx < 0)
3542 : return;
3543 770 : idx = get_stridx (src, stmt);
3544 :
3545 770 : dsi = NULL;
3546 770 : if (didx > 0)
3547 467 : dsi = get_strinfo (didx);
3548 :
3549 770 : srclen = NULL_TREE;
3550 770 : si = NULL;
3551 770 : if (idx < 0)
3552 66 : srclen = build_int_cst (size_type_node, ~idx);
3553 704 : else if (idx > 0)
3554 : {
3555 410 : si = get_strinfo (idx);
3556 410 : if (si != NULL)
3557 353 : srclen = get_string_length (si);
3558 : }
3559 :
3560 : /* Disable warning for the transformed statement? */
3561 770 : opt_code no_warning_opt = no_warning;
3562 :
3563 770 : if (dsi == NULL || get_string_length (dsi) == NULL_TREE)
3564 : {
3565 464 : {
3566 : /* The concatenation always involves copying at least one byte
3567 : (the terminating nul), even if the source string is empty.
3568 : If the source is unknown assume it's one character long and
3569 : used that as both sizes. */
3570 464 : tree slen = srclen;
3571 464 : if (slen)
3572 : {
3573 197 : tree type = TREE_TYPE (slen);
3574 197 : slen = fold_build2 (PLUS_EXPR, type, slen, build_int_cst (type, 1));
3575 : }
3576 :
3577 464 : tree sptr = si && si->ptr ? si->ptr : src;
3578 464 : no_warning_opt = check_bounds_or_overlap (stmt, dst, sptr, NULL_TREE,
3579 : slen);
3580 464 : if (no_warning_opt)
3581 49 : suppress_warning (stmt, no_warning_opt);
3582 : }
3583 :
3584 : /* strcat (p, q) can be transformed into
3585 : tmp = p + strlen (p); endptr = stpcpy (tmp, q);
3586 : with length endptr - p if we need to compute the length
3587 : later on. Don't do this transformation if we don't need
3588 : it. */
3589 602 : if (builtin_decl_implicit_p (BUILT_IN_STPCPY) && lhs == NULL_TREE)
3590 : {
3591 102 : if (didx == 0)
3592 : {
3593 58 : didx = new_stridx (dst);
3594 58 : if (didx == 0)
3595 : return;
3596 : }
3597 102 : if (dsi == NULL)
3598 : {
3599 102 : dsi = new_strinfo (dst, didx, NULL_TREE, false);
3600 102 : set_strinfo (didx, dsi);
3601 102 : find_equal_ptrs (dst, didx);
3602 : }
3603 : else
3604 : {
3605 0 : dsi = unshare_strinfo (dsi);
3606 0 : dsi->nonzero_chars = NULL_TREE;
3607 0 : dsi->full_string_p = false;
3608 0 : dsi->next = 0;
3609 0 : dsi->endptr = NULL_TREE;
3610 : }
3611 102 : dsi->writable = true;
3612 102 : dsi->stmt = stmt;
3613 102 : dsi->dont_invalidate = true;
3614 : }
3615 : return;
3616 : }
3617 :
3618 306 : tree dstlen = dsi->nonzero_chars;
3619 306 : endptr = dsi->endptr;
3620 :
3621 306 : dsi = unshare_strinfo (dsi);
3622 306 : dsi->endptr = NULL_TREE;
3623 306 : dsi->stmt = NULL;
3624 306 : dsi->writable = true;
3625 :
3626 306 : if (srclen != NULL_TREE)
3627 : {
3628 222 : dsi->nonzero_chars = fold_build2_loc (loc, PLUS_EXPR,
3629 222 : TREE_TYPE (dsi->nonzero_chars),
3630 : dsi->nonzero_chars, srclen);
3631 222 : gcc_assert (dsi->full_string_p);
3632 222 : adjust_related_strinfos (loc, dsi, srclen);
3633 222 : dsi->dont_invalidate = true;
3634 : }
3635 : else
3636 : {
3637 84 : dsi->nonzero_chars = NULL;
3638 84 : dsi->full_string_p = false;
3639 84 : if (lhs == NULL_TREE && builtin_decl_implicit_p (BUILT_IN_STPCPY))
3640 64 : dsi->dont_invalidate = true;
3641 : }
3642 :
3643 306 : if (si != NULL)
3644 : /* strcat src may not overlap dst, so src doesn't need to be
3645 : invalidated either. */
3646 191 : si->dont_invalidate = true;
3647 :
3648 : /* For now. Could remove the lhs from the call and add
3649 : lhs = dst; afterwards. */
3650 306 : if (lhs)
3651 : return;
3652 :
3653 208 : fn = NULL_TREE;
3654 208 : objsz = NULL_TREE;
3655 208 : switch (bcode)
3656 : {
3657 174 : case BUILT_IN_STRCAT:
3658 174 : if (srclen != NULL_TREE)
3659 94 : fn = builtin_decl_implicit (BUILT_IN_MEMCPY);
3660 : else
3661 80 : fn = builtin_decl_implicit (BUILT_IN_STRCPY);
3662 : break;
3663 34 : case BUILT_IN_STRCAT_CHK:
3664 34 : if (srclen != NULL_TREE)
3665 32 : fn = builtin_decl_explicit (BUILT_IN_MEMCPY_CHK);
3666 : else
3667 2 : fn = builtin_decl_explicit (BUILT_IN_STRCPY_CHK);
3668 34 : objsz = gimple_call_arg (stmt, 2);
3669 34 : break;
3670 0 : default:
3671 0 : gcc_unreachable ();
3672 : }
3673 :
3674 208 : if (fn == NULL_TREE)
3675 : return;
3676 :
3677 208 : if (dsi && dstlen)
3678 : {
3679 208 : tree type = TREE_TYPE (dstlen);
3680 :
3681 : /* Compute the size of the source sequence, including the nul. */
3682 208 : tree srcsize = srclen ? srclen : size_zero_node;
3683 208 : tree one = build_int_cst (type, 1);
3684 208 : srcsize = fold_build2 (PLUS_EXPR, type, srcsize, one);
3685 208 : tree dstsize = fold_build2 (PLUS_EXPR, type, dstlen, one);
3686 208 : tree sptr = si && si->ptr ? si->ptr : src;
3687 :
3688 208 : no_warning_opt = check_bounds_or_overlap (stmt, dst, sptr, dstsize,
3689 : srcsize);
3690 208 : if (no_warning_opt)
3691 71 : suppress_warning (stmt, no_warning_opt);
3692 : }
3693 :
3694 208 : tree len = NULL_TREE;
3695 208 : if (srclen != NULL_TREE)
3696 : {
3697 126 : args = TYPE_ARG_TYPES (TREE_TYPE (fn));
3698 126 : type = TREE_VALUE (TREE_CHAIN (TREE_CHAIN (args)));
3699 :
3700 126 : len = fold_convert_loc (loc, type, unshare_expr (srclen));
3701 126 : len = fold_build2_loc (loc, PLUS_EXPR, type, len,
3702 : build_int_cst (type, 1));
3703 126 : len = force_gimple_operand_gsi (&m_gsi, len, true, NULL_TREE, true,
3704 : GSI_SAME_STMT);
3705 : }
3706 208 : if (endptr)
3707 24 : dst = fold_convert_loc (loc, TREE_TYPE (dst), unshare_expr (endptr));
3708 : else
3709 184 : dst = fold_build2_loc (loc, POINTER_PLUS_EXPR, TREE_TYPE (dst), dst,
3710 : fold_convert_loc (loc, sizetype,
3711 : unshare_expr (dstlen)));
3712 208 : dst = force_gimple_operand_gsi (&m_gsi, dst, true, NULL_TREE, true,
3713 : GSI_SAME_STMT);
3714 208 : if (objsz)
3715 : {
3716 68 : objsz = fold_build2_loc (loc, MINUS_EXPR, TREE_TYPE (objsz), objsz,
3717 34 : fold_convert_loc (loc, TREE_TYPE (objsz),
3718 : unshare_expr (dstlen)));
3719 34 : objsz = force_gimple_operand_gsi (&m_gsi, objsz, true, NULL_TREE, true,
3720 : GSI_SAME_STMT);
3721 : }
3722 208 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
3723 : {
3724 0 : fprintf (dump_file, "Optimizing: ");
3725 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
3726 : }
3727 208 : if (srclen != NULL_TREE)
3728 220 : success = update_gimple_call (&m_gsi, fn, 3 + (objsz != NULL_TREE),
3729 : dst, src, len, objsz);
3730 : else
3731 162 : success = update_gimple_call (&m_gsi, fn, 2 + (objsz != NULL_TREE),
3732 : dst, src, objsz);
3733 208 : if (success)
3734 : {
3735 208 : stmt = gsi_stmt (m_gsi);
3736 208 : update_stmt (stmt);
3737 208 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
3738 : {
3739 0 : fprintf (dump_file, "into: ");
3740 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
3741 : }
3742 : /* If srclen == NULL, note that current string length can be
3743 : computed by transforming this strcpy into stpcpy. */
3744 208 : if (srclen == NULL_TREE && dsi->dont_invalidate)
3745 64 : dsi->stmt = stmt;
3746 208 : adjust_last_stmt (dsi, stmt, true);
3747 208 : if (srclen != NULL_TREE)
3748 : {
3749 126 : laststmt.stmt = stmt;
3750 126 : laststmt.len = srclen;
3751 126 : laststmt.stridx = dsi->idx;
3752 : }
3753 : }
3754 0 : else if (dump_file && (dump_flags & TDF_DETAILS) != 0)
3755 0 : fprintf (dump_file, "not possible.\n");
3756 :
3757 208 : if (no_warning_opt)
3758 71 : suppress_warning (stmt, no_warning_opt);
3759 : }
3760 :
3761 : /* Handle a call to an allocation function like alloca, malloc or calloc,
3762 : or an ordinary allocation function declared with attribute alloc_size. */
3763 :
3764 : void
3765 98896 : strlen_pass::handle_alloc_call (built_in_function bcode)
3766 : {
3767 98896 : gimple *stmt = gsi_stmt (m_gsi);
3768 98896 : tree lhs = gimple_call_lhs (stmt);
3769 98896 : if (lhs == NULL_TREE)
3770 : return;
3771 :
3772 98804 : gcc_assert (get_stridx (lhs, stmt) == 0);
3773 98804 : int idx = new_stridx (lhs);
3774 98804 : tree length = NULL_TREE;
3775 98804 : if (bcode == BUILT_IN_CALLOC)
3776 505 : length = build_int_cst (size_type_node, 0);
3777 98804 : strinfo *si = new_strinfo (lhs, idx, length, length != NULL_TREE);
3778 98804 : if (bcode == BUILT_IN_CALLOC)
3779 : {
3780 : /* Only set STMT for calloc and malloc. */
3781 505 : si->stmt = stmt;
3782 : /* Only set ENDPTR for calloc. */
3783 505 : si->endptr = lhs;
3784 : }
3785 98299 : else if (bcode == BUILT_IN_MALLOC)
3786 28487 : si->stmt = stmt;
3787 :
3788 : /* Set ALLOC is set for all allocation functions. */
3789 98804 : si->alloc = stmt;
3790 98804 : set_strinfo (idx, si);
3791 98804 : si->writable = true;
3792 98804 : si->dont_invalidate = true;
3793 : }
3794 :
3795 : /* Returns true of the last statement of the bb is a conditional
3796 : that checks ptr for null-ness. */
3797 : static bool
3798 236 : last_stmt_ptr_check (tree ptr, basic_block bb)
3799 : {
3800 236 : gimple_stmt_iterator gsi = gsi_last_nondebug_bb (bb);
3801 236 : gcond *cstmt = dyn_cast <gcond *>(gsi_stmt (gsi));
3802 235 : if (!cstmt)
3803 : return false;
3804 235 : if (gimple_cond_code (cstmt) != EQ_EXPR && gimple_cond_code (cstmt) != NE_EXPR)
3805 : return false;
3806 205 : if (!integer_zerop (gimple_cond_rhs (cstmt)))
3807 : return false;
3808 205 : if (!operand_equal_p (gimple_cond_lhs (cstmt), ptr))
3809 : return false;
3810 : return true;
3811 : }
3812 :
3813 : /* Check if doing a malloc+memset to calloc is a good idea. PTR is the
3814 : return value of the malloc/where the memset happens. MALLOC_BB is
3815 : the basic block of the malloc. MEMSET_BB is basic block of the memset. */
3816 :
3817 : static bool
3818 391 : allow_memset_malloc_to_calloc (tree ptr, basic_block malloc_bb,
3819 : basic_block memset_bb)
3820 : {
3821 : /* If the malloc and memset are in the same block, then always
3822 : allow the transformation. Don't need post dominator calculation. */
3823 391 : if (malloc_bb == memset_bb)
3824 : return true;
3825 :
3826 235 : if (!dom_info_available_p (cfun, CDI_POST_DOMINATORS))
3827 186 : calculate_dominance_info (CDI_POST_DOMINATORS);
3828 :
3829 : /* If the memset is always executed after the malloc, then allow
3830 : to optimize to calloc. */
3831 235 : if (dominated_by_p (CDI_POST_DOMINATORS, malloc_bb, memset_bb))
3832 : return true;
3833 :
3834 : /* If the malloc bb ends in a ptr check, then we need to check if
3835 : either successor is post dominated by the memset bb. */
3836 222 : if (last_stmt_ptr_check (ptr, malloc_bb))
3837 : {
3838 199 : if (dominated_by_p (CDI_POST_DOMINATORS, EDGE_SUCC (malloc_bb, 0)->dest, memset_bb))
3839 : return true;
3840 189 : if (dominated_by_p (CDI_POST_DOMINATORS, EDGE_SUCC (malloc_bb, 1)->dest, memset_bb))
3841 : return true;
3842 : }
3843 :
3844 : /* At this point we want to only handle:
3845 : malloc();
3846 : ...
3847 : if (ptr) goto memset_bb; */
3848 65 : if (!single_pred_p (memset_bb))
3849 : return false;
3850 :
3851 : /* If the predecessor of the memset bb is not post dominated by malloc, then the memset is
3852 : conditionalized by something more than just the checking if ptr is non-null. */
3853 57 : if (!dominated_by_p (CDI_POST_DOMINATORS, malloc_bb, single_pred_edge (memset_bb)->src))
3854 : return false;
3855 :
3856 14 : return last_stmt_ptr_check (ptr, single_pred_edge (memset_bb)->src);
3857 : }
3858 :
3859 : /* Handle a call to memset.
3860 : After a call to calloc, memset(,0,) is unnecessary.
3861 : memset(malloc(n),0,n) is calloc(n,1).
3862 : return true when the call is transformed, false otherwise.
3863 : When nonnull uses RVALS to determine range information. */
3864 :
3865 : bool
3866 26940 : strlen_pass::handle_builtin_memset (bool *zero_write)
3867 : {
3868 26940 : gimple *memset_stmt = gsi_stmt (m_gsi);
3869 26940 : tree ptr = gimple_call_arg (memset_stmt, 0);
3870 26940 : tree memset_val = gimple_call_arg (memset_stmt, 1);
3871 26940 : tree memset_size = gimple_call_arg (memset_stmt, 2);
3872 :
3873 : /* Set to the non-constant offset added to PTR. */
3874 134700 : wide_int offrng[2];
3875 26940 : int idx1 = get_stridx (ptr, memset_stmt, offrng, ptr_qry.rvals);
3876 26940 : if (idx1 == 0
3877 17205 : && TREE_CODE (memset_val) == INTEGER_CST
3878 43183 : && ((TREE_CODE (memset_size) == INTEGER_CST
3879 8460 : && !integer_zerop (memset_size))
3880 7783 : || TREE_CODE (memset_size) == SSA_NAME))
3881 : {
3882 16243 : unsigned HOST_WIDE_INT mask = (HOST_WIDE_INT_1U << CHAR_TYPE_SIZE) - 1;
3883 16243 : bool full_string_p = (wi::to_wide (memset_val) & mask) == 0;
3884 :
3885 : /* We only handle symbolic lengths when writing non-zero values. */
3886 16243 : if (full_string_p && TREE_CODE (memset_size) != INTEGER_CST)
3887 : return false;
3888 :
3889 11538 : idx1 = new_stridx (ptr);
3890 11538 : if (idx1 == 0)
3891 : return false;
3892 11369 : tree newlen;
3893 11369 : if (full_string_p)
3894 6172 : newlen = build_int_cst (size_type_node, 0);
3895 5197 : else if (TREE_CODE (memset_size) == INTEGER_CST)
3896 2119 : newlen = fold_convert (size_type_node, memset_size);
3897 : else
3898 : newlen = memset_size;
3899 :
3900 11369 : strinfo *dsi = new_strinfo (ptr, idx1, newlen, full_string_p);
3901 11369 : set_strinfo (idx1, dsi);
3902 11369 : find_equal_ptrs (ptr, idx1);
3903 11369 : dsi->dont_invalidate = true;
3904 11369 : dsi->writable = true;
3905 11369 : return false;
3906 : }
3907 :
3908 10697 : if (idx1 <= 0)
3909 : return false;
3910 9735 : strinfo *si1 = get_strinfo (idx1);
3911 9735 : if (!si1)
3912 : return false;
3913 3973 : gimple *alloc_stmt = si1->alloc;
3914 3973 : if (!alloc_stmt || !is_gimple_call (alloc_stmt))
3915 : return false;
3916 2498 : tree callee1 = gimple_call_fndecl (alloc_stmt);
3917 2498 : if (!valid_builtin_call (alloc_stmt))
3918 : return false;
3919 742 : tree alloc_size = gimple_call_arg (alloc_stmt, 0);
3920 :
3921 : /* Check for overflow. */
3922 742 : maybe_warn_overflow (memset_stmt, false, memset_size, NULL, false, true);
3923 :
3924 : /* Bail when there is no statement associated with the destination
3925 : (the statement may be null even when SI1->ALLOC is not). */
3926 742 : if (!si1->stmt)
3927 : return false;
3928 :
3929 : /* Avoid optimizing if store is at a variable offset from the beginning
3930 : of the allocated object. */
3931 1023 : if (offrng[0] != 0 || offrng[0] != offrng[1])
3932 : return false;
3933 :
3934 : /* Bail when the call writes a non-zero value. */
3935 473 : if (!integer_zerop (memset_val))
3936 : return false;
3937 :
3938 : /* Let the caller know the memset call cleared the destination. */
3939 394 : *zero_write = true;
3940 :
3941 394 : enum built_in_function code1 = DECL_FUNCTION_CODE (callee1);
3942 394 : if (code1 == BUILT_IN_CALLOC)
3943 : /* Not touching alloc_stmt */ ;
3944 386 : else if (!allow_memset_malloc_to_calloc (ptr, gimple_bb (si1->stmt), gimple_bb (memset_stmt)))
3945 : return false;
3946 322 : else if (code1 == BUILT_IN_MALLOC
3947 322 : && operand_equal_p (memset_size, alloc_size, 0))
3948 : {
3949 : /* Replace the malloc + memset calls with calloc. */
3950 281 : gimple_stmt_iterator gsi1 = gsi_for_stmt (si1->stmt);
3951 562 : update_gimple_call (&gsi1, builtin_decl_implicit (BUILT_IN_CALLOC), 2,
3952 : alloc_size, build_one_cst (size_type_node));
3953 281 : si1->nonzero_chars = build_int_cst (size_type_node, 0);
3954 281 : si1->full_string_p = true;
3955 281 : si1->stmt = gsi_stmt (gsi1);
3956 : }
3957 : else
3958 : return false;
3959 289 : tree lhs = gimple_call_lhs (memset_stmt);
3960 289 : unlink_stmt_vdef (memset_stmt);
3961 289 : if (lhs)
3962 : {
3963 1 : gimple *assign = gimple_build_assign (lhs, ptr);
3964 1 : gsi_replace (&m_gsi, assign, false);
3965 : }
3966 : else
3967 : {
3968 288 : gsi_remove (&m_gsi, true);
3969 288 : release_defs (memset_stmt);
3970 : }
3971 :
3972 : return true;
3973 80820 : }
3974 :
3975 : /* Return first such statement if RES is used in statements testing its
3976 : equality to zero, and null otherwise. If EXCLUSIVE is true, return
3977 : nonnull if and only RES is used in such expressions exclusively and
3978 : in none other. */
3979 :
3980 : gimple *
3981 501469 : use_in_zero_equality (tree res, bool exclusive)
3982 : {
3983 501469 : gimple *first_use = NULL;
3984 :
3985 501469 : use_operand_p use_p;
3986 501469 : imm_use_iterator iter;
3987 :
3988 977788 : FOR_EACH_IMM_USE_FAST (use_p, iter, res)
3989 : {
3990 518681 : gimple *use_stmt = USE_STMT (use_p);
3991 :
3992 518681 : if (is_gimple_debug (use_stmt))
3993 15944 : continue;
3994 :
3995 502737 : if (gimple_code (use_stmt) == GIMPLE_ASSIGN)
3996 : {
3997 74392 : tree_code code = gimple_assign_rhs_code (use_stmt);
3998 74392 : if (code == EQ_EXPR || code == NE_EXPR)
3999 : {
4000 53551 : if (!integer_zerop (gimple_assign_rhs2 (use_stmt)))
4001 : {
4002 51 : if (exclusive)
4003 : return NULL;
4004 1 : continue;
4005 : }
4006 : }
4007 20841 : else if (exclusive)
4008 : return NULL;
4009 : else
4010 18 : continue;
4011 : }
4012 428345 : else if (gimple_code (use_stmt) == GIMPLE_COND)
4013 : {
4014 412019 : tree_code code = gimple_cond_code (use_stmt);
4015 412021 : if ((code != EQ_EXPR && code != NE_EXPR)
4016 412019 : || !integer_zerop (gimple_cond_rhs (use_stmt)))
4017 : {
4018 5266 : if (exclusive)
4019 : return NULL;
4020 2 : continue;
4021 : }
4022 : }
4023 16326 : else if (exclusive)
4024 : return NULL;
4025 : else
4026 101 : continue;
4027 :
4028 460253 : if (!first_use)
4029 476319 : first_use = use_stmt;
4030 42362 : }
4031 :
4032 459107 : return first_use;
4033 : }
4034 :
4035 : /* Given strinfo IDX for ARG, sets LENRNG[] to the range of lengths
4036 : of the string(s) referenced by ARG if it can be determined.
4037 : If the length cannot be determined, sets *SIZE to the size of
4038 : the array the string is stored in, if any. If no such array is
4039 : known, sets *SIZE to -1. When the strings are nul-terminated sets
4040 : *NULTERM to true, otherwise to false. When nonnull uses RVALS to
4041 : determine range information. Returns true on success. */
4042 :
4043 : bool
4044 513255 : strlen_pass::get_len_or_size (gimple *stmt, tree arg, int idx,
4045 : unsigned HOST_WIDE_INT lenrng[2],
4046 : unsigned HOST_WIDE_INT *size, bool *nulterm)
4047 : {
4048 : /* Invalidate. */
4049 513255 : *size = HOST_WIDE_INT_M1U;
4050 :
4051 513255 : if (idx < 0)
4052 : {
4053 : /* IDX is the inverted constant string length. */
4054 253699 : lenrng[0] = ~idx;
4055 253699 : lenrng[1] = lenrng[0];
4056 253699 : *nulterm = true;
4057 253699 : return true;
4058 : }
4059 :
4060 : /* Set so that both LEN and ~LEN are invalid lengths, i.e., maximum
4061 : possible length + 1. */
4062 259556 : lenrng[0] = lenrng[1] = HOST_WIDE_INT_MAX;
4063 :
4064 259556 : if (strinfo *si = idx ? get_strinfo (idx) : NULL)
4065 : {
4066 : /* FIXME: Handle all this in_range_strlen_dynamic. */
4067 1068 : if (!si->nonzero_chars)
4068 : ;
4069 1065 : else if (tree_fits_uhwi_p (si->nonzero_chars))
4070 : {
4071 948 : lenrng[0] = tree_to_uhwi (si->nonzero_chars);
4072 948 : *nulterm = si->full_string_p;
4073 : /* Set the upper bound only if the string is known to be
4074 : nul-terminated, otherwise leave it at maximum + 1. */
4075 948 : if (*nulterm)
4076 592 : lenrng[1] = lenrng[0];
4077 : }
4078 117 : else if (TREE_CODE (si->nonzero_chars) == SSA_NAME)
4079 : {
4080 114 : int_range_max r;
4081 228 : if (get_range_query (cfun)->range_of_expr (r, si->nonzero_chars)
4082 114 : && !r.undefined_p ()
4083 228 : && !r.varying_p ())
4084 : {
4085 114 : lenrng[0] = r.lower_bound ().to_uhwi ();
4086 114 : lenrng[1] = r.upper_bound ().to_uhwi ();
4087 114 : *nulterm = si->full_string_p;
4088 : }
4089 114 : }
4090 : }
4091 :
4092 259556 : if (lenrng[0] != HOST_WIDE_INT_MAX)
4093 : return true;
4094 :
4095 : /* Compute the minimum and maximum real or possible lengths. */
4096 258494 : c_strlen_data lendata = { };
4097 : /* Set MAXBOUND to an arbitrary non-null non-integer node as a request
4098 : to have it set to the length of the longest string in a PHI. */
4099 258494 : lendata.maxbound = arg;
4100 258494 : get_range_strlen_dynamic (arg, stmt, &lendata, ptr_qry);
4101 :
4102 258494 : unsigned HOST_WIDE_INT maxbound = HOST_WIDE_INT_M1U;
4103 258494 : if (tree_fits_uhwi_p (lendata.maxbound)
4104 258494 : && !integer_all_onesp (lendata.maxbound))
4105 2126 : maxbound = tree_to_uhwi (lendata.maxbound);
4106 :
4107 258494 : if (tree_fits_uhwi_p (lendata.minlen) && tree_fits_uhwi_p (lendata.maxlen))
4108 : {
4109 258494 : unsigned HOST_WIDE_INT minlen = tree_to_uhwi (lendata.minlen);
4110 258494 : unsigned HOST_WIDE_INT maxlen = tree_to_uhwi (lendata.maxlen);
4111 :
4112 : /* The longest string in this data model. */
4113 258494 : const unsigned HOST_WIDE_INT lenmax
4114 258494 : = tree_to_uhwi (max_object_size ()) - 2;
4115 :
4116 258494 : if (maxbound == HOST_WIDE_INT_M1U)
4117 : {
4118 256368 : lenrng[0] = minlen;
4119 256368 : lenrng[1] = maxlen;
4120 256368 : *nulterm = minlen == maxlen;
4121 : }
4122 2126 : else if (maxlen < lenmax)
4123 : {
4124 1847 : *size = maxbound + 1;
4125 1847 : *nulterm = false;
4126 : }
4127 : else
4128 : return false;
4129 :
4130 : return true;
4131 : }
4132 :
4133 0 : if (maxbound != HOST_WIDE_INT_M1U
4134 0 : && lendata.maxlen
4135 0 : && !integer_all_onesp (lendata.maxlen))
4136 : {
4137 : /* Set *SIZE to LENDATA.MAXBOUND which is a conservative estimate
4138 : of the longest string based on the sizes of the arrays referenced
4139 : by ARG. */
4140 0 : *size = maxbound + 1;
4141 0 : *nulterm = false;
4142 0 : return true;
4143 : }
4144 :
4145 : return false;
4146 : }
4147 :
4148 : /* If IDX1 and IDX2 refer to strings A and B of unequal lengths, return
4149 : the result of 0 == strncmp (A, B, BOUND) (which is the same as strcmp
4150 : for a sufficiently large BOUND). If the result is based on the length
4151 : of one string being greater than the longest string that would fit in
4152 : the array pointer to by the argument, set *PLEN and *PSIZE to
4153 : the corresponding length (or its complement when the string is known
4154 : to be at least as long and need not be nul-terminated) and size.
4155 : Otherwise return null. */
4156 :
4157 : tree
4158 129051 : strlen_pass::strxcmp_eqz_result (gimple *stmt, tree arg1, int idx1,
4159 : tree arg2, int idx2,
4160 : unsigned HOST_WIDE_INT bound,
4161 : unsigned HOST_WIDE_INT len[2],
4162 : unsigned HOST_WIDE_INT *psize)
4163 : {
4164 : /* Determine the range the length of each string is in and whether it's
4165 : known to be nul-terminated, or the size of the array it's stored in. */
4166 129051 : bool nul1, nul2;
4167 129051 : unsigned HOST_WIDE_INT siz1, siz2;
4168 129051 : unsigned HOST_WIDE_INT len1rng[2], len2rng[2];
4169 129051 : if (!get_len_or_size (stmt, arg1, idx1, len1rng, &siz1, &nul1)
4170 129051 : || !get_len_or_size (stmt, arg2, idx2, len2rng, &siz2, &nul2))
4171 : return NULL_TREE;
4172 :
4173 : /* BOUND is set to HWI_M1U for strcmp and less to strncmp, and LENiRNG
4174 : to HWI_MAX when invalid. Adjust the length of each string to consider
4175 : to be no more than BOUND. */
4176 128853 : if (len1rng[0] < HOST_WIDE_INT_MAX && len1rng[0] > bound)
4177 33 : len1rng[0] = bound;
4178 128853 : if (len1rng[1] < HOST_WIDE_INT_MAX && len1rng[1] > bound)
4179 71 : len1rng[1] = bound;
4180 128853 : if (len2rng[0] < HOST_WIDE_INT_MAX && len2rng[0] > bound)
4181 109 : len2rng[0] = bound;
4182 128853 : if (len2rng[1] < HOST_WIDE_INT_MAX && len2rng[1] > bound)
4183 109 : len2rng[1] = bound;
4184 :
4185 : /* Two empty strings are equal. */
4186 128853 : if (len1rng[1] == 0 && len2rng[1] == 0)
4187 3 : return integer_one_node;
4188 :
4189 : /* The strings are definitely unequal when the lower bound of the length
4190 : of one of them is greater than the length of the longest string that
4191 : would fit into the other array. */
4192 128850 : if (len1rng[0] == HOST_WIDE_INT_MAX
4193 806 : && len2rng[0] != HOST_WIDE_INT_MAX
4194 699 : && ((len2rng[0] < bound && len2rng[0] >= siz1)
4195 626 : || len2rng[0] > siz1))
4196 : {
4197 99 : *psize = siz1;
4198 99 : len[0] = len1rng[0];
4199 : /* Set LEN[0] to the lower bound of ARG1's length when it's
4200 : nul-terminated or to the complement of its minimum length
4201 : otherwise, */
4202 99 : len[1] = nul2 ? len2rng[0] : ~len2rng[0];
4203 99 : return integer_zero_node;
4204 : }
4205 :
4206 128751 : if (len2rng[0] == HOST_WIDE_INT_MAX
4207 353 : && len1rng[0] != HOST_WIDE_INT_MAX
4208 246 : && ((len1rng[0] < bound && len1rng[0] >= siz2)
4209 186 : || len1rng[0] > siz2))
4210 : {
4211 82 : *psize = siz2;
4212 82 : len[0] = nul1 ? len1rng[0] : ~len1rng[0];
4213 82 : len[1] = len2rng[0];
4214 82 : return integer_zero_node;
4215 : }
4216 :
4217 : /* The strings are also definitely unequal when their lengths are unequal
4218 : and at least one is nul-terminated. */
4219 128669 : if (len1rng[0] != HOST_WIDE_INT_MAX
4220 127962 : && len2rng[0] != HOST_WIDE_INT_MAX
4221 127798 : && ((len1rng[1] < len2rng[0] && nul1)
4222 127771 : || (len2rng[1] < len1rng[0] && nul2)))
4223 : {
4224 36 : if (bound <= len1rng[0] || bound <= len2rng[0])
4225 6 : *psize = bound;
4226 : else
4227 30 : *psize = HOST_WIDE_INT_M1U;
4228 :
4229 36 : len[0] = len1rng[0];
4230 36 : len[1] = len2rng[0];
4231 36 : return integer_zero_node;
4232 : }
4233 :
4234 : /* The string lengths may be equal or unequal. Even when equal and
4235 : both strings nul-terminated, without the string contents there's
4236 : no way to determine whether they are equal. */
4237 : return NULL_TREE;
4238 : }
4239 :
4240 : /* Diagnose pointless calls to strcmp or strncmp STMT with string
4241 : arguments of lengths LEN or size SIZ and (for strncmp) BOUND,
4242 : whose result is used in equality expressions that evaluate to
4243 : a constant due to one argument being longer than the size of
4244 : the other. */
4245 :
4246 : static void
4247 217 : maybe_warn_pointless_strcmp (gimple *stmt, HOST_WIDE_INT bound,
4248 : unsigned HOST_WIDE_INT len[2],
4249 : unsigned HOST_WIDE_INT siz)
4250 : {
4251 217 : tree lhs = gimple_call_lhs (stmt);
4252 217 : gimple *use = use_in_zero_equality (lhs, /* exclusive = */ false);
4253 217 : if (!use)
4254 : return;
4255 :
4256 110 : bool at_least = false;
4257 :
4258 : /* Excessive LEN[i] indicates a lower bound. */
4259 110 : if (len[0] > HOST_WIDE_INT_MAX)
4260 : {
4261 3 : at_least = true;
4262 3 : len[0] = ~len[0];
4263 : }
4264 :
4265 110 : if (len[1] > HOST_WIDE_INT_MAX)
4266 : {
4267 0 : at_least = true;
4268 0 : len[1] = ~len[1];
4269 : }
4270 :
4271 110 : unsigned HOST_WIDE_INT minlen = MIN (len[0], len[1]);
4272 :
4273 : /* FIXME: Include a note pointing to the declaration of the smaller
4274 : array. */
4275 110 : location_t stmt_loc = gimple_or_expr_nonartificial_location (stmt, lhs);
4276 :
4277 110 : tree callee = gimple_call_fndecl (stmt);
4278 110 : bool warned = false;
4279 110 : if (siz <= minlen && bound == -1)
4280 119 : warned = warning_at (stmt_loc, OPT_Wstring_compare,
4281 : (at_least
4282 : ? G_("%qD of a string of length %wu or more and "
4283 : "an array of size %wu evaluates to nonzero")
4284 : : G_("%qD of a string of length %wu and an array "
4285 : "of size %wu evaluates to nonzero")),
4286 : callee, minlen, siz);
4287 49 : else if (!at_least && siz <= HOST_WIDE_INT_MAX)
4288 : {
4289 21 : if (len[0] != HOST_WIDE_INT_MAX && len[1] != HOST_WIDE_INT_MAX)
4290 6 : warned = warning_at (stmt_loc, OPT_Wstring_compare,
4291 : "%qD of strings of length %wu and %wu "
4292 : "and bound of %wu evaluates to nonzero",
4293 : callee, len[0], len[1], bound);
4294 : else
4295 15 : warned = warning_at (stmt_loc, OPT_Wstring_compare,
4296 : "%qD of a string of length %wu, an array "
4297 : "of size %wu and bound of %wu evaluates to "
4298 : "nonzero",
4299 : callee, minlen, siz, bound);
4300 : }
4301 :
4302 82 : if (!warned)
4303 : return;
4304 :
4305 58 : location_t use_loc = gimple_location (use);
4306 58 : if (LOCATION_LINE (stmt_loc) != LOCATION_LINE (use_loc))
4307 12 : inform (use_loc, "in this expression");
4308 : }
4309 :
4310 :
4311 : /* Optimize a call to strcmp or strncmp either by folding it to a constant
4312 : when possible or by transforming the latter to the former. Warn about
4313 : calls where the length of one argument is greater than the size of
4314 : the array to which the other argument points if the latter's length
4315 : is not known. Return true when the call has been transformed into
4316 : another and false otherwise. */
4317 :
4318 : bool
4319 129565 : strlen_pass::handle_builtin_string_cmp ()
4320 : {
4321 129565 : gcall *stmt = as_a <gcall *> (gsi_stmt (m_gsi));
4322 129565 : tree lhs = gimple_call_lhs (stmt);
4323 :
4324 129565 : if (!lhs)
4325 : return false;
4326 :
4327 129565 : tree arg1 = gimple_call_arg (stmt, 0);
4328 129565 : tree arg2 = gimple_call_arg (stmt, 1);
4329 129565 : int idx1 = get_stridx (arg1, stmt);
4330 129565 : int idx2 = get_stridx (arg2, stmt);
4331 :
4332 : /* For strncmp set to the value of the third argument if known. */
4333 129565 : HOST_WIDE_INT bound = -1;
4334 129565 : tree len = NULL_TREE;
4335 : /* Extract the strncmp bound. */
4336 129565 : if (gimple_call_num_args (stmt) == 3)
4337 : {
4338 1784 : len = gimple_call_arg (stmt, 2);
4339 1784 : if (tree_fits_shwi_p (len))
4340 1277 : bound = tree_to_shwi (len);
4341 :
4342 : /* If the bound argument is NOT known, do nothing. */
4343 1277 : if (bound < 0)
4344 : return false;
4345 : }
4346 :
4347 : /* Avoid folding if either argument is not a nul-terminated array.
4348 : Defer warning until later. */
4349 129058 : if (!check_nul_terminated_array (NULL_TREE, arg1, len)
4350 129058 : || !check_nul_terminated_array (NULL_TREE, arg2, len))
4351 : return false;
4352 :
4353 129051 : {
4354 : /* Set to the length of one argument (or its complement if it's
4355 : the lower bound of a range) and the size of the array storing
4356 : the other if the result is based on the former being equal to
4357 : or greater than the latter. */
4358 129051 : unsigned HOST_WIDE_INT len[2] = { HOST_WIDE_INT_MAX, HOST_WIDE_INT_MAX };
4359 129051 : unsigned HOST_WIDE_INT siz = HOST_WIDE_INT_M1U;
4360 :
4361 : /* Try to determine if the two strings are either definitely equal
4362 : or definitely unequal and if so, either fold the result to zero
4363 : (when equal) or set the range of the result to ~[0, 0] otherwise. */
4364 129051 : if (tree eqz = strxcmp_eqz_result (stmt, arg1, idx1, arg2, idx2, bound,
4365 : len, &siz))
4366 : {
4367 220 : if (integer_zerop (eqz))
4368 : {
4369 217 : maybe_warn_pointless_strcmp (stmt, bound, len, siz);
4370 :
4371 : /* When the lengths of the first two string arguments are
4372 : known to be unequal set the range of the result to non-zero.
4373 : This allows the call to be eliminated if its result is only
4374 : used in tests for equality to zero. */
4375 217 : int_range_max nz;
4376 217 : nz.set_nonzero (TREE_TYPE (lhs));
4377 217 : set_range_info (lhs, nz);
4378 217 : return false;
4379 217 : }
4380 : /* When the two strings are definitely equal (such as when they
4381 : are both empty) fold the call to the constant result. */
4382 3 : replace_call_with_value (&m_gsi, integer_zero_node);
4383 3 : return true;
4384 : }
4385 : }
4386 :
4387 : /* Return if nothing is known about the strings pointed to by ARG1
4388 : and ARG2. */
4389 128831 : if (idx1 == 0 && idx2 == 0)
4390 : return false;
4391 :
4392 : /* Determine either the length or the size of each of the strings,
4393 : whichever is available. */
4394 127691 : HOST_WIDE_INT cstlen1 = -1, cstlen2 = -1;
4395 127691 : HOST_WIDE_INT arysiz1 = -1, arysiz2 = -1;
4396 :
4397 127691 : {
4398 127691 : unsigned HOST_WIDE_INT len1rng[2], len2rng[2];
4399 127691 : unsigned HOST_WIDE_INT arsz1, arsz2;
4400 127691 : bool nulterm[2];
4401 :
4402 127691 : if (!get_len_or_size (stmt, arg1, idx1, len1rng, &arsz1, nulterm)
4403 127691 : || !get_len_or_size (stmt, arg2, idx2, len2rng, &arsz2, nulterm + 1))
4404 81 : return false;
4405 :
4406 127610 : if (len1rng[0] == len1rng[1] && len1rng[0] < HOST_WIDE_INT_MAX)
4407 496 : cstlen1 = len1rng[0];
4408 127114 : else if (arsz1 < HOST_WIDE_INT_M1U)
4409 550 : arysiz1 = arsz1;
4410 :
4411 127610 : if (len2rng[0] == len2rng[1] && len2rng[0] < HOST_WIDE_INT_MAX)
4412 126519 : cstlen2 = len2rng[0];
4413 1091 : else if (arsz2 < HOST_WIDE_INT_M1U)
4414 130 : arysiz2 = arsz2;
4415 : }
4416 :
4417 : /* Bail if neither the string length nor the size of the array
4418 : it is stored in can be determined. */
4419 127610 : if ((cstlen1 < 0 && arysiz1 < 0)
4420 1046 : || (cstlen2 < 0 && arysiz2 < 0)
4421 574 : || (cstlen1 < 0 && cstlen2 < 0))
4422 : return false;
4423 :
4424 574 : if (cstlen1 >= 0)
4425 349 : ++cstlen1;
4426 574 : if (cstlen2 >= 0)
4427 456 : ++cstlen2;
4428 :
4429 : /* The exact number of characters to compare. */
4430 574 : HOST_WIDE_INT cmpsiz;
4431 574 : if (cstlen1 >= 0 && cstlen2 >= 0)
4432 231 : cmpsiz = MIN (cstlen1, cstlen2);
4433 343 : else if (cstlen1 >= 0)
4434 : cmpsiz = cstlen1;
4435 : else
4436 225 : cmpsiz = cstlen2;
4437 574 : if (bound >= 0)
4438 108 : cmpsiz = MIN (cmpsiz, bound);
4439 : /* The size of the array in which the unknown string is stored. */
4440 574 : HOST_WIDE_INT varsiz = arysiz1 < 0 ? arysiz2 : arysiz1;
4441 :
4442 574 : if ((varsiz < 0 || cmpsiz < varsiz) && use_in_zero_equality (lhs))
4443 : {
4444 : /* If the known length is less than the size of the other array
4445 : and the strcmp result is only used to test equality to zero,
4446 : transform the call to the equivalent _eq call. */
4447 444 : if (tree fn = builtin_decl_implicit (bound < 0 ? BUILT_IN_STRCMP_EQ
4448 : : BUILT_IN_STRNCMP_EQ))
4449 : {
4450 406 : tree n = build_int_cst (size_type_node, cmpsiz);
4451 406 : update_gimple_call (&m_gsi, fn, 3, arg1, arg2, n);
4452 406 : return true;
4453 : }
4454 : }
4455 :
4456 : return false;
4457 : }
4458 :
4459 : /* Handle a POINTER_PLUS_EXPR statement.
4460 : For p = "abcd" + 2; compute associated length, or if
4461 : p = q + off is pointing to a '\0' character of a string, call
4462 : zero_length_string on it. */
4463 :
4464 : void
4465 764874 : strlen_pass::handle_pointer_plus ()
4466 : {
4467 764874 : gimple *stmt = gsi_stmt (m_gsi);
4468 764874 : tree lhs = gimple_assign_lhs (stmt), off;
4469 764874 : int idx = get_stridx (gimple_assign_rhs1 (stmt), stmt);
4470 764874 : strinfo *si, *zsi;
4471 :
4472 764874 : if (idx == 0)
4473 : return;
4474 :
4475 82225 : if (idx < 0)
4476 : {
4477 3038 : tree off = gimple_assign_rhs2 (stmt);
4478 3038 : if (tree_fits_uhwi_p (off)
4479 64 : && tree_to_uhwi (off) <= (unsigned HOST_WIDE_INT) ~idx)
4480 59 : ssa_ver_to_stridx[SSA_NAME_VERSION (lhs)]
4481 118 : = ~(~idx - (int) tree_to_uhwi (off));
4482 : return;
4483 : }
4484 :
4485 79187 : si = get_strinfo (idx);
4486 79187 : if (si == NULL || si->nonzero_chars == NULL_TREE)
4487 : return;
4488 :
4489 7886 : off = gimple_assign_rhs2 (stmt);
4490 7886 : zsi = NULL;
4491 7886 : if (si->full_string_p && operand_equal_p (si->nonzero_chars, off, 0))
4492 1248 : zsi = zero_length_string (lhs, si);
4493 6638 : else if (TREE_CODE (off) == SSA_NAME)
4494 : {
4495 2842 : gimple *def_stmt = SSA_NAME_DEF_STMT (off);
4496 2842 : if (gimple_assign_single_p (def_stmt)
4497 476 : && si->full_string_p
4498 3315 : && operand_equal_p (si->nonzero_chars,
4499 473 : gimple_assign_rhs1 (def_stmt), 0))
4500 437 : zsi = zero_length_string (lhs, si);
4501 : }
4502 1685 : if (zsi != NULL
4503 1685 : && si->endptr != NULL_TREE
4504 1442 : && si->endptr != lhs
4505 37 : && TREE_CODE (si->endptr) == SSA_NAME)
4506 : {
4507 37 : enum tree_code rhs_code
4508 37 : = useless_type_conversion_p (TREE_TYPE (lhs), TREE_TYPE (si->endptr))
4509 37 : ? SSA_NAME : NOP_EXPR;
4510 37 : gimple_assign_set_rhs_with_ops (&m_gsi, rhs_code, si->endptr);
4511 37 : gcc_assert (gsi_stmt (m_gsi) == stmt);
4512 37 : update_stmt (stmt);
4513 : }
4514 : }
4515 :
4516 : /* Set LENRANGE to the number of nonzero bytes for a store of TYPE and
4517 : clear all flags. Return true on success and false on failure. */
4518 :
4519 : static bool
4520 82613 : nonzero_bytes_for_type (tree type, unsigned lenrange[3],
4521 : bool *nulterm, bool *allnul, bool *allnonnul)
4522 : {
4523 : /* Use the size of the type of the expression as the size of the store,
4524 : and set the upper bound of the length range to that of the size.
4525 : Nothing is known about the contents so clear all flags. */
4526 82613 : tree typesize = TYPE_SIZE_UNIT (type);
4527 82613 : if (!type)
4528 : return false;
4529 :
4530 82613 : if (!tree_fits_uhwi_p (typesize))
4531 : return false;
4532 :
4533 82599 : unsigned HOST_WIDE_INT sz = tree_to_uhwi (typesize);
4534 82599 : if (sz > UINT_MAX)
4535 : return false;
4536 :
4537 82599 : lenrange[2] = sz;
4538 82599 : lenrange[1] = lenrange[2] ? lenrange[2] - 1 : 0;
4539 82599 : lenrange[0] = 0;
4540 82599 : *nulterm = false;
4541 82599 : *allnul = false;
4542 82599 : *allnonnul = false;
4543 82599 : return true;
4544 : }
4545 :
4546 : /* Recursively determine the minimum and maximum number of leading nonzero
4547 : bytes in the representation of EXP at memory state VUSE and set
4548 : LENRANGE[0] and LENRANGE[1] to each.
4549 : Sets LENRANGE[2] to the total size of the access (which may be less
4550 : than LENRANGE[1] when what's being referenced by EXP is a pointer
4551 : rather than an array).
4552 : Sets *NULTERM if the representation contains a zero byte, sets *ALLNUL
4553 : if all the bytes are zero, and *ALLNONNUL is all are nonzero.
4554 : OFFSET and NBYTES are the offset into the representation and
4555 : the size of the access to it determined from an ADDR_EXPR (i.e.,
4556 : a pointer) or MEM_REF or zero for other expressions.
4557 : Uses RVALS to determine range information.
4558 : Avoids recursing deeper than the limits in SNLIM allow.
4559 : Returns true on success and false otherwise. */
4560 :
4561 : bool
4562 1503298 : strlen_pass::count_nonzero_bytes (tree exp, tree vuse, gimple *stmt,
4563 : unsigned HOST_WIDE_INT offset,
4564 : unsigned HOST_WIDE_INT nbytes,
4565 : unsigned lenrange[3], bool *nulterm,
4566 : bool *allnul, bool *allnonnul,
4567 : ssa_name_limit_t &snlim)
4568 : {
4569 1515985 : if (TREE_CODE (exp) == SSA_NAME)
4570 : {
4571 : /* Handle non-zero single-character stores specially. */
4572 171696 : tree type = TREE_TYPE (exp);
4573 171696 : if (TREE_CODE (type) == INTEGER_TYPE
4574 164397 : && TYPE_MODE (type) == TYPE_MODE (char_type_node)
4575 157669 : && TYPE_PRECISION (type) == TYPE_PRECISION (char_type_node)
4576 329365 : && tree_expr_nonzero_p (exp))
4577 : {
4578 : /* If the character EXP is known to be non-zero (even if its
4579 : exact value is not known) recurse once to set the range
4580 : for an arbitrary constant. */
4581 12687 : exp = build_int_cst (type, 1);
4582 12687 : return count_nonzero_bytes (exp, vuse, stmt,
4583 : offset, 1, lenrange,
4584 12687 : nulterm, allnul, allnonnul, snlim);
4585 : }
4586 :
4587 159009 : gimple *g = SSA_NAME_DEF_STMT (exp);
4588 159009 : if (gimple_assign_single_p (g))
4589 : {
4590 77756 : exp = gimple_assign_rhs1 (g);
4591 77756 : if (!DECL_P (exp)
4592 77483 : && TREE_CODE (exp) != CONSTRUCTOR
4593 76801 : && TREE_CODE (exp) != MEM_REF)
4594 : return false;
4595 : /* Handle DECLs, CONSTRUCTOR and MEM_REF below. */
4596 : stmt = g;
4597 : }
4598 81253 : else if (gimple_code (g) == GIMPLE_PHI)
4599 : {
4600 : /* Avoid processing an SSA_NAME that has already been visited
4601 : or if an SSA_NAME limit has been reached. Indicate success
4602 : if the former and failure if the latter. */
4603 33496 : if (int res = snlim.next_phi (exp))
4604 12960 : return res > 0;
4605 :
4606 : /* Determine the minimum and maximum from the PHI arguments. */
4607 20536 : unsigned int n = gimple_phi_num_args (g);
4608 81628 : for (unsigned i = 0; i != n; i++)
4609 : {
4610 68560 : tree def = gimple_phi_arg_def (g, i);
4611 68560 : if (!count_nonzero_bytes (def, vuse, g,
4612 : offset, nbytes, lenrange, nulterm,
4613 : allnul, allnonnul, snlim))
4614 : return false;
4615 : }
4616 :
4617 : return true;
4618 : }
4619 : }
4620 :
4621 1441119 : if (TREE_CODE (exp) == CONSTRUCTOR)
4622 : {
4623 46515 : if (nbytes)
4624 : /* If NBYTES has already been determined by an outer MEM_REF
4625 : fail rather than overwriting it (this shouldn't happen). */
4626 : return false;
4627 :
4628 46515 : tree type = TREE_TYPE (exp);
4629 46515 : tree size = TYPE_SIZE_UNIT (type);
4630 46515 : if (!size || !tree_fits_uhwi_p (size))
4631 : return false;
4632 :
4633 46515 : unsigned HOST_WIDE_INT byte_size = tree_to_uhwi (size);
4634 46515 : if (byte_size <= offset)
4635 : return false;
4636 :
4637 46515 : nbytes = byte_size - offset;
4638 : }
4639 :
4640 1441119 : if (TREE_CODE (exp) == MEM_REF)
4641 : {
4642 750091 : if (nbytes)
4643 : return false;
4644 :
4645 749860 : tree arg = TREE_OPERAND (exp, 0);
4646 749860 : tree off = TREE_OPERAND (exp, 1);
4647 :
4648 749860 : if (TREE_CODE (off) != INTEGER_CST || !tree_fits_uhwi_p (off))
4649 : return false;
4650 :
4651 749860 : unsigned HOST_WIDE_INT wioff = tree_to_uhwi (off);
4652 749860 : if (INT_MAX < wioff)
4653 : return false;
4654 :
4655 703810 : offset += wioff;
4656 703810 : if (INT_MAX < offset)
4657 : return false;
4658 :
4659 : /* The size of the MEM_REF access determines the number of bytes. */
4660 703810 : tree type = TREE_TYPE (exp);
4661 703810 : tree typesize = TYPE_SIZE_UNIT (type);
4662 703810 : if (!typesize || !tree_fits_uhwi_p (typesize))
4663 : return false;
4664 703810 : nbytes = tree_to_uhwi (typesize);
4665 703810 : if (!nbytes)
4666 : return false;
4667 :
4668 : /* Handle MEM_REF = SSA_NAME types of assignments. */
4669 703810 : return count_nonzero_bytes_addr (arg, vuse, stmt,
4670 : offset, nbytes, lenrange, nulterm,
4671 703810 : allnul, allnonnul, snlim);
4672 : }
4673 :
4674 691028 : if (VAR_P (exp) || TREE_CODE (exp) == CONST_DECL)
4675 : {
4676 : /* If EXP can be folded into a constant use the result. Otherwise
4677 : proceed to use EXP to determine a range of the result. */
4678 350040 : if (tree fold_exp = ctor_for_folding (exp))
4679 350038 : if (fold_exp != error_mark_node)
4680 691028 : exp = fold_exp;
4681 : }
4682 :
4683 691028 : const char *prep = NULL;
4684 691028 : if (TREE_CODE (exp) == STRING_CST)
4685 : {
4686 15189 : unsigned nchars = TREE_STRING_LENGTH (exp);
4687 15189 : if (nchars <= offset)
4688 : return false;
4689 :
4690 15183 : if (!nbytes)
4691 : /* If NBYTES hasn't been determined earlier, either from ADDR_EXPR
4692 : (i.e., it's the size of a pointer), or from MEM_REF (as the size
4693 : of the access), set it here to the size of the string, including
4694 : all internal and trailing nuls if the string has any. */
4695 6708 : nbytes = nchars - offset;
4696 8475 : else if (nchars - offset < nbytes)
4697 : return false;
4698 :
4699 15178 : prep = TREE_STRING_POINTER (exp) + offset;
4700 : }
4701 :
4702 691017 : unsigned char buf[256];
4703 15178 : if (!prep)
4704 : {
4705 675839 : if (CHAR_BIT != 8 || BITS_PER_UNIT != 8 || offset > INT_MAX)
4706 : return false;
4707 : /* If the pointer to representation hasn't been set above
4708 : for STRING_CST point it at the buffer. */
4709 675839 : prep = reinterpret_cast <char *>(buf);
4710 : /* Try to extract the representation of the constant object
4711 : or expression starting from the offset. */
4712 675839 : unsigned repsize = native_encode_expr (exp, buf, sizeof buf, offset);
4713 675839 : if (repsize < nbytes)
4714 : {
4715 : /* Handle vector { 0x12345678, 0x23003412, x_1(D), y_2(D) }
4716 : and similar cases. Even when not all the elements are constant,
4717 : we can perhaps figure out something from the constant ones
4718 : and assume the others can be anything. */
4719 362594 : if (TREE_CODE (exp) == CONSTRUCTOR
4720 4461 : && CONSTRUCTOR_NELTS (exp)
4721 3072 : && VECTOR_TYPE_P (TREE_TYPE (exp))
4722 363276 : && nbytes <= sizeof buf)
4723 : {
4724 682 : tree v0 = CONSTRUCTOR_ELT (exp, 0)->value;
4725 682 : unsigned HOST_WIDE_INT elt_sz
4726 682 : = int_size_in_bytes (TREE_TYPE (v0));
4727 682 : unsigned int i, s = 0;
4728 682 : tree v, idx;
4729 716 : FOR_EACH_CONSTRUCTOR_ELT (CONSTRUCTOR_ELTS (exp), i, idx, v)
4730 : {
4731 716 : if (idx
4732 716 : && (VECTOR_TYPE_P (TREE_TYPE (v0))
4733 15 : || !tree_fits_uhwi_p (idx)
4734 15 : || tree_to_uhwi (idx) != i))
4735 : {
4736 : s = 0;
4737 : break;
4738 : }
4739 716 : if ((i + 1) * elt_sz <= offset)
4740 0 : continue;
4741 716 : unsigned int o = 0;
4742 716 : if (i * elt_sz < offset)
4743 0 : o = offset % elt_sz;
4744 1432 : repsize = native_encode_expr (v, buf + s,
4745 716 : sizeof (buf) - s, o);
4746 716 : if (repsize != elt_sz - o)
4747 : break;
4748 34 : s += repsize;
4749 : }
4750 682 : if (s != 0 && s < nbytes)
4751 : {
4752 20 : unsigned HOST_WIDE_INT n = strnlen (prep, s);
4753 20 : if (n < lenrange[0])
4754 20 : lenrange[0] = n;
4755 20 : if (lenrange[1] < n && n != s)
4756 2 : lenrange[1] = n;
4757 20 : if (lenrange[2] < nbytes)
4758 20 : lenrange[2] = nbytes;
4759 : /* We haven't processed all bytes, the rest are unknown.
4760 : So, clear NULTERM if none of the initial bytes are
4761 : zero, and clear ALLNUL and ALLNONNULL because we don't
4762 : know about the remaining bytes. */
4763 20 : if (n == s)
4764 6 : *nulterm = false;
4765 20 : *allnul = false;
4766 20 : *allnonnul = false;
4767 20 : return true;
4768 : }
4769 662 : else if (s != nbytes)
4770 : {
4771 : /* See below. */
4772 662 : lenrange[0] = 0;
4773 662 : prep = NULL;
4774 : }
4775 : }
4776 : else
4777 : {
4778 : /* This should only happen when REPSIZE is zero because EXP
4779 : doesn't denote an object with a known initializer, except
4780 : perhaps when the reference reads past its end. */
4781 361912 : lenrange[0] = 0;
4782 361912 : prep = NULL;
4783 : }
4784 : }
4785 313245 : else if (!nbytes)
4786 : nbytes = repsize;
4787 57138 : else if (nbytes < repsize)
4788 : return false;
4789 : }
4790 :
4791 690997 : if (!nbytes)
4792 49744 : return nonzero_bytes_for_type (TREE_TYPE (exp), lenrange,
4793 49744 : nulterm, allnul, allnonnul);
4794 :
4795 : /* Compute the number of leading nonzero bytes in the representation
4796 : and update the minimum and maximum. */
4797 641253 : unsigned HOST_WIDE_INT n = prep ? strnlen (prep, nbytes) : nbytes;
4798 :
4799 641253 : if (n < lenrange[0])
4800 253514 : lenrange[0] = n;
4801 641253 : if (lenrange[1] < n)
4802 482425 : lenrange[1] = n;
4803 :
4804 : /* Set the size of the representation. */
4805 641253 : if (lenrange[2] < nbytes)
4806 612209 : lenrange[2] = nbytes;
4807 :
4808 : /* Clear NULTERM if none of the bytes is zero. */
4809 641253 : if (n == nbytes)
4810 490850 : *nulterm = false;
4811 :
4812 641253 : if (n)
4813 : {
4814 : /* When the initial number of non-zero bytes N is non-zero, reset
4815 : *ALLNUL; if N is less than that the size of the representation
4816 : also clear *ALLNONNUL. */
4817 501139 : *allnul = false;
4818 501139 : if (n < nbytes)
4819 10289 : *allnonnul = false;
4820 : }
4821 140114 : else if (*allnul || *allnonnul)
4822 : {
4823 134464 : *allnonnul = false;
4824 :
4825 134464 : if (*allnul)
4826 : {
4827 : /* When either ALLNUL is set and N is zero, also determine
4828 : whether all subsequent bytes after the first one (which
4829 : is nul) are zero or nonzero and clear ALLNUL if not. */
4830 1016985 : for (const char *p = prep; p != prep + nbytes; ++p)
4831 890862 : if (*p)
4832 : {
4833 2580 : *allnul = false;
4834 2580 : break;
4835 : }
4836 : }
4837 : }
4838 :
4839 : return true;
4840 : }
4841 :
4842 : /* Like count_nonzero_bytes, but instead of counting bytes in EXP, count
4843 : bytes that are pointed to by EXP, which should be a pointer. */
4844 :
4845 : bool
4846 886782 : strlen_pass::count_nonzero_bytes_addr (tree exp, tree vuse, gimple *stmt,
4847 : unsigned HOST_WIDE_INT offset,
4848 : unsigned HOST_WIDE_INT nbytes,
4849 : unsigned lenrange[3], bool *nulterm,
4850 : bool *allnul, bool *allnonnul,
4851 : ssa_name_limit_t &snlim)
4852 : {
4853 886782 : int idx = get_stridx (exp, stmt);
4854 886782 : if (idx > 0)
4855 : {
4856 : /* get_strinfo reflects string lengths before the current statement,
4857 : where the current statement is the outermost count_nonzero_bytes
4858 : stmt. If there are any stores in between stmt and that
4859 : current statement, the string length information might describe
4860 : something significantly different. */
4861 27382 : if (gimple_vuse (stmt) != vuse)
4862 : return false;
4863 :
4864 11465 : strinfo *si = get_strinfo (idx);
4865 11465 : if (!si)
4866 : return false;
4867 :
4868 : /* Handle both constant lengths as well non-constant lengths
4869 : in some range. */
4870 4413 : unsigned HOST_WIDE_INT minlen, maxlen;
4871 4413 : if (tree_fits_shwi_p (si->nonzero_chars))
4872 1078 : minlen = maxlen = tree_to_shwi (si->nonzero_chars);
4873 3335 : else if (si->nonzero_chars
4874 3272 : && TREE_CODE (si->nonzero_chars) == SSA_NAME)
4875 : {
4876 3272 : int_range_max vr;
4877 3272 : if (!ptr_qry.rvals->range_of_expr (vr, si->nonzero_chars, stmt)
4878 3272 : || vr.undefined_p ()
4879 6535 : || vr.varying_p ())
4880 9 : return false;
4881 :
4882 3263 : minlen = vr.lower_bound ().to_uhwi ();
4883 3263 : maxlen = vr.upper_bound ().to_uhwi ();
4884 3272 : }
4885 : else
4886 : return false;
4887 :
4888 4341 : if (maxlen < offset)
4889 : return false;
4890 :
4891 4138 : minlen = minlen < offset ? 0 : minlen - offset;
4892 4138 : maxlen -= offset;
4893 4138 : if (maxlen + 1 < nbytes)
4894 : return false;
4895 :
4896 3676 : if (nbytes <= minlen || !si->full_string_p)
4897 3302 : *nulterm = false;
4898 :
4899 3676 : if (nbytes < minlen)
4900 : {
4901 159 : minlen = nbytes;
4902 159 : if (nbytes < maxlen)
4903 : maxlen = nbytes;
4904 : }
4905 :
4906 3676 : if (!si->full_string_p)
4907 350 : maxlen = nbytes;
4908 :
4909 3676 : if (minlen < lenrange[0])
4910 3676 : lenrange[0] = minlen;
4911 3676 : if (lenrange[1] < maxlen)
4912 3676 : lenrange[1] = maxlen;
4913 :
4914 3676 : if (lenrange[2] < nbytes)
4915 3676 : lenrange[2] = nbytes;
4916 :
4917 : /* Since only the length of the string are known and not its contents,
4918 : clear ALLNUL and ALLNONNUL purely on the basis of the length. */
4919 3676 : *allnul = false;
4920 3676 : if (minlen < nbytes)
4921 393 : *allnonnul = false;
4922 :
4923 : return true;
4924 : }
4925 :
4926 873091 : if (TREE_CODE (exp) == ADDR_EXPR)
4927 369242 : return count_nonzero_bytes (TREE_OPERAND (exp, 0), vuse, stmt,
4928 : offset, nbytes,
4929 369242 : lenrange, nulterm, allnul, allnonnul, snlim);
4930 :
4931 503849 : if (TREE_CODE (exp) == SSA_NAME)
4932 : {
4933 501734 : gimple *g = SSA_NAME_DEF_STMT (exp);
4934 501734 : if (gimple_code (g) == GIMPLE_PHI)
4935 : {
4936 : /* Avoid processing an SSA_NAME that has already been visited
4937 : or if an SSA_NAME limit has been reached. Indicate success
4938 : if the former and failure if the latter. */
4939 106111 : if (int res = snlim.next_phi (exp))
4940 26160 : return res > 0;
4941 :
4942 : /* Determine the minimum and maximum from the PHI arguments. */
4943 79951 : unsigned int n = gimple_phi_num_args (g);
4944 259556 : for (unsigned i = 0; i != n; i++)
4945 : {
4946 182972 : tree def = gimple_phi_arg_def (g, i);
4947 182972 : if (!count_nonzero_bytes_addr (def, vuse, g,
4948 : offset, nbytes, lenrange,
4949 : nulterm, allnul, allnonnul,
4950 : snlim))
4951 : return false;
4952 : }
4953 :
4954 : return true;
4955 : }
4956 : }
4957 :
4958 : /* Otherwise we don't know anything. */
4959 397738 : lenrange[0] = 0;
4960 397738 : if (lenrange[1] < nbytes)
4961 325052 : lenrange[1] = nbytes;
4962 397738 : if (lenrange[2] < nbytes)
4963 323624 : lenrange[2] = nbytes;
4964 397738 : *nulterm = false;
4965 397738 : *allnul = false;
4966 397738 : *allnonnul = false;
4967 397738 : return true;
4968 : }
4969 :
4970 : /* Same as above except with an implicit SSA_NAME limit. When EXPR_OR_TYPE
4971 : is a type rather than an expression use its size to compute the range.
4972 : RVALS is used to determine ranges of dynamically computed string lengths
4973 : (the results of strlen). */
4974 :
4975 : bool
4976 1098365 : strlen_pass::count_nonzero_bytes (tree expr_or_type, gimple *stmt,
4977 : unsigned lenrange[3], bool *nulterm,
4978 : bool *allnul, bool *allnonnul)
4979 : {
4980 1098365 : if (TYPE_P (expr_or_type))
4981 32869 : return nonzero_bytes_for_type (expr_or_type, lenrange,
4982 32869 : nulterm, allnul, allnonnul);
4983 :
4984 : /* Set to optimistic values so the caller doesn't have to worry about
4985 : initializing these and to what. On success, the function will clear
4986 : these if it determines their values are different but being recursive
4987 : it never sets either to true. On failure, their values are
4988 : unspecified. */
4989 1065496 : *nulterm = true;
4990 1065496 : *allnul = true;
4991 1065496 : *allnonnul = true;
4992 :
4993 1065496 : ssa_name_limit_t snlim;
4994 1065496 : tree expr = expr_or_type;
4995 2130992 : return count_nonzero_bytes (expr, gimple_vuse (stmt), stmt,
4996 : 0, 0, lenrange, nulterm, allnul, allnonnul,
4997 : snlim);
4998 1065496 : }
4999 :
5000 : /* Handle a single or multibyte store other than by a built-in function,
5001 : either via a single character assignment or by multi-byte assignment
5002 : either via MEM_REF or via a type other than char (such as in
5003 : '*(int*)a = 12345'). Return true to let the caller advance *GSI to
5004 : the next statement in the basic block and false otherwise. */
5005 :
5006 : bool
5007 408366 : strlen_pass::handle_store (bool *zero_write)
5008 : {
5009 408366 : gimple *stmt = gsi_stmt (m_gsi);
5010 : /* The LHS and RHS of the store. The RHS is null if STMT is a function
5011 : call. STORETYPE is the type of the store (determined from either
5012 : the RHS of the assignment statement or the LHS of a function call. */
5013 408366 : tree lhs, rhs, storetype;
5014 408366 : if (is_gimple_assign (stmt))
5015 : {
5016 375497 : lhs = gimple_assign_lhs (stmt);
5017 375497 : rhs = gimple_assign_rhs1 (stmt);
5018 375497 : storetype = TREE_TYPE (rhs);
5019 : }
5020 32869 : else if (is_gimple_call (stmt))
5021 : {
5022 32869 : lhs = gimple_call_lhs (stmt);
5023 32869 : rhs = NULL_TREE;
5024 32869 : storetype = TREE_TYPE (lhs);
5025 : }
5026 : else
5027 : return true;
5028 :
5029 408366 : tree ssaname = NULL_TREE;
5030 408366 : strinfo *si = NULL;
5031 408366 : int idx = -1;
5032 :
5033 408366 : range_query *const rvals = ptr_qry.rvals;
5034 :
5035 : /* The offset of the first byte in LHS modified by the store. */
5036 408366 : unsigned HOST_WIDE_INT offset = 0;
5037 :
5038 408366 : if (TREE_CODE (lhs) == MEM_REF
5039 408366 : && TREE_CODE (TREE_OPERAND (lhs, 0)) == SSA_NAME)
5040 : {
5041 121388 : tree mem_offset = TREE_OPERAND (lhs, 1);
5042 121388 : if (tree_fits_uhwi_p (mem_offset))
5043 : {
5044 : /* Get the strinfo for the base, and use it if it starts with at
5045 : least OFFSET nonzero characters. This is trivially true if
5046 : OFFSET is zero. */
5047 121388 : offset = tree_to_uhwi (mem_offset);
5048 121388 : idx = get_stridx (TREE_OPERAND (lhs, 0), stmt);
5049 121388 : if (idx > 0)
5050 14570 : si = get_strinfo (idx);
5051 121388 : if (offset == 0)
5052 99787 : ssaname = TREE_OPERAND (lhs, 0);
5053 21601 : else if (si == NULL
5054 21601 : || compare_nonzero_chars (si, stmt, offset, rvals) < 0)
5055 : {
5056 18675 : *zero_write = rhs ? initializer_zerop (rhs) : false;
5057 :
5058 18675 : bool dummy;
5059 18675 : unsigned lenrange[] = { UINT_MAX, 0, 0 };
5060 18675 : if (count_nonzero_bytes (rhs ? rhs : storetype, stmt, lenrange,
5061 : &dummy, &dummy, &dummy))
5062 16749 : maybe_warn_overflow (stmt, true, lenrange[2]);
5063 :
5064 18675 : return true;
5065 : }
5066 : }
5067 : }
5068 : else
5069 : {
5070 286978 : idx = get_addr_stridx (lhs, stmt, NULL_TREE, &offset, rvals);
5071 286978 : if (idx > 0)
5072 46903 : si = get_strinfo (idx);
5073 : }
5074 :
5075 : /* Minimum and maximum leading non-zero bytes and the size of the store. */
5076 389691 : unsigned lenrange[] = { UINT_MAX, 0, 0 };
5077 :
5078 : /* Set to the minimum length of the string being assigned if known. */
5079 389691 : unsigned HOST_WIDE_INT rhs_minlen;
5080 :
5081 : /* STORING_NONZERO_P is true iff not all stored characters are zero.
5082 : STORING_ALL_NONZERO_P is true if all stored characters are zero.
5083 : STORING_ALL_ZEROS_P is true iff all stored characters are zero.
5084 : Both are false when it's impossible to determine which is true. */
5085 389691 : bool storing_nonzero_p;
5086 389691 : bool storing_all_nonzero_p;
5087 389691 : bool storing_all_zeros_p;
5088 : /* FULL_STRING_P is set when the stored sequence of characters form
5089 : a nul-terminated string. */
5090 389691 : bool full_string_p;
5091 :
5092 389691 : const bool ranges_valid
5093 422560 : = count_nonzero_bytes (rhs ? rhs : storetype, stmt,
5094 : lenrange, &full_string_p,
5095 : &storing_all_zeros_p, &storing_all_nonzero_p);
5096 :
5097 389691 : if (ranges_valid)
5098 : {
5099 359646 : rhs_minlen = lenrange[0];
5100 359646 : storing_nonzero_p = lenrange[1] > 0;
5101 359646 : *zero_write = storing_all_zeros_p;
5102 :
5103 359646 : maybe_warn_overflow (stmt, true, lenrange[2]);
5104 : }
5105 : else
5106 : {
5107 30045 : rhs_minlen = HOST_WIDE_INT_M1U;
5108 30045 : full_string_p = false;
5109 30045 : storing_nonzero_p = false;
5110 30045 : storing_all_zeros_p = false;
5111 30045 : storing_all_nonzero_p = false;
5112 : }
5113 :
5114 389691 : if (si != NULL)
5115 : {
5116 : /* The count_nonzero_bytes call above might have unshared si.
5117 : Fetch it again from the vector. */
5118 28771 : si = get_strinfo (idx);
5119 : /* The corresponding element is set to 1 if the first and last
5120 : element, respectively, of the sequence of characters being
5121 : written over the string described by SI ends before
5122 : the terminating nul (if it has one), to zero if the nul is
5123 : being overwritten but not beyond, or negative otherwise. */
5124 28771 : int store_before_nul[2];
5125 28771 : if (ranges_valid)
5126 : {
5127 : /* The offset of the last stored byte. */
5128 27789 : unsigned HOST_WIDE_INT endoff = offset + lenrange[2] - 1;
5129 27789 : store_before_nul[0]
5130 27789 : = compare_nonzero_chars (si, stmt, offset, rvals);
5131 27789 : if (endoff == offset)
5132 : store_before_nul[1] = store_before_nul[0];
5133 : else
5134 9526 : store_before_nul[1]
5135 9526 : = compare_nonzero_chars (si, stmt, endoff, rvals);
5136 : }
5137 : else
5138 : {
5139 982 : store_before_nul[0]
5140 982 : = compare_nonzero_chars (si, stmt, offset, rvals);
5141 982 : store_before_nul[1] = store_before_nul[0];
5142 982 : gcc_assert (offset == 0 || store_before_nul[0] >= 0);
5143 : }
5144 :
5145 28771 : if (storing_all_zeros_p
5146 13671 : && store_before_nul[0] == 0
5147 12669 : && store_before_nul[1] == 0
5148 10227 : && si->full_string_p)
5149 : {
5150 : /* When overwriting a '\0' with a '\0', the store can be removed
5151 : if we know it has been stored in the current function. */
5152 434 : if (!stmt_could_throw_p (cfun, stmt) && si->writable)
5153 : {
5154 432 : unlink_stmt_vdef (stmt);
5155 432 : release_defs (stmt);
5156 432 : gsi_remove (&m_gsi, true);
5157 432 : return false;
5158 : }
5159 : else
5160 : {
5161 2 : si->writable = true;
5162 2 : gsi_next (&m_gsi);
5163 2 : return false;
5164 : }
5165 : }
5166 :
5167 28337 : if (store_before_nul[1] > 0
5168 1031 : && storing_nonzero_p
5169 573 : && lenrange[0] == lenrange[1]
5170 566 : && lenrange[0] == lenrange[2]
5171 562 : && TREE_CODE (storetype) == INTEGER_TYPE)
5172 : {
5173 : /* Handle a store of one or more non-nul characters that ends
5174 : before the terminating nul of the destination and so does
5175 : not affect its length
5176 : If si->nonzero_chars > OFFSET, we aren't overwriting '\0',
5177 : and if we aren't storing '\0', we know that the length of
5178 : the string and any other zero terminated string in memory
5179 : remains the same. In that case we move to the next gimple
5180 : statement and return to signal the caller that it shouldn't
5181 : invalidate anything.
5182 :
5183 : This is beneficial for cases like:
5184 :
5185 : char p[20];
5186 : void foo (char *q)
5187 : {
5188 : strcpy (p, "foobar");
5189 : size_t len = strlen (p); // can be folded to 6
5190 : size_t len2 = strlen (q); // has to be computed
5191 : p[0] = 'X';
5192 : size_t len3 = strlen (p); // can be folded to 6
5193 : size_t len4 = strlen (q); // can be folded to len2
5194 : bar (len, len2, len3, len4);
5195 : } */
5196 266 : gsi_next (&m_gsi);
5197 266 : return false;
5198 : }
5199 :
5200 27613 : if (storing_nonzero_p
5201 14674 : || storing_all_zeros_p
5202 1437 : || (full_string_p && lenrange[1] == 0)
5203 1136 : || (offset != 0 && store_before_nul[1] > 0))
5204 : {
5205 : /* When STORING_NONZERO_P, we know that the string will start
5206 : with at least OFFSET + 1 nonzero characters. If storing
5207 : a single character, set si->NONZERO_CHARS to the result.
5208 : If storing multiple characters, try to determine the number
5209 : of leading non-zero characters and set si->NONZERO_CHARS to
5210 : the result instead.
5211 :
5212 : When STORING_ALL_ZEROS_P, or the first byte written is zero,
5213 : i.e. FULL_STRING_P && LENRANGE[1] == 0, we know that the
5214 : string is now OFFSET characters long.
5215 :
5216 : Otherwise, we're storing an unknown value at offset OFFSET,
5217 : so need to clip the nonzero_chars to OFFSET.
5218 : Use the minimum length of the string (or individual character)
5219 : being stored if it's known. Otherwise, STORING_NONZERO_P
5220 : guarantees it's at least 1. */
5221 40337 : HOST_WIDE_INT len
5222 26940 : = storing_nonzero_p && ranges_valid ? lenrange[0] : 1;
5223 26940 : location_t loc = gimple_location (stmt);
5224 26940 : tree oldlen = si->nonzero_chars;
5225 26940 : if (store_before_nul[1] == 0 && si->full_string_p)
5226 : /* We're overwriting the nul terminator with a nonzero or
5227 : unknown character. If the previous stmt was a memcpy,
5228 : its length may be decreased. */
5229 1824 : adjust_last_stmt (si, stmt, false);
5230 26940 : si = unshare_strinfo (si);
5231 26940 : if (storing_nonzero_p)
5232 : {
5233 13397 : gcc_assert (len >= 0);
5234 13397 : si->nonzero_chars = build_int_cst (size_type_node, offset + len);
5235 : }
5236 : else
5237 13543 : si->nonzero_chars = build_int_cst (size_type_node, offset);
5238 :
5239 : /* Set FULL_STRING_P only if the length of the strings being
5240 : written is the same, and clear it if the strings have
5241 : different lengths. In the latter case the length stored
5242 : in si->NONZERO_CHARS becomes the lower bound.
5243 : FIXME: Handle the upper bound of the length if possible. */
5244 26940 : si->full_string_p = full_string_p && lenrange[0] == lenrange[1];
5245 :
5246 26940 : if (storing_all_zeros_p
5247 13237 : && ssaname
5248 27530 : && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (ssaname))
5249 590 : si->endptr = ssaname;
5250 : else
5251 26350 : si->endptr = NULL;
5252 26940 : si->next = 0;
5253 26940 : si->stmt = NULL;
5254 26940 : si->writable = true;
5255 26940 : si->dont_invalidate = true;
5256 26940 : if (oldlen)
5257 : {
5258 22459 : tree adj = fold_build2_loc (loc, MINUS_EXPR, size_type_node,
5259 : si->nonzero_chars, oldlen);
5260 22459 : adjust_related_strinfos (loc, si, adj);
5261 : }
5262 : else
5263 4481 : si->prev = 0;
5264 : }
5265 : }
5266 360920 : else if (idx == 0 && (storing_all_zeros_p || storing_nonzero_p))
5267 : {
5268 235151 : if (ssaname)
5269 72849 : idx = new_stridx (ssaname);
5270 : else
5271 162302 : idx = new_addr_stridx (lhs);
5272 235151 : if (idx != 0)
5273 : {
5274 210212 : tree ptr = (ssaname ? ssaname : build_fold_addr_expr (lhs));
5275 :
5276 210212 : HOST_WIDE_INT slen;
5277 210212 : if (storing_all_zeros_p)
5278 : slen = 0;
5279 115809 : else if (storing_nonzero_p && ranges_valid)
5280 : {
5281 : /* FIXME: Handle the upper bound of the length when
5282 : LENRANGE[0] != LENRANGE[1]. */
5283 115809 : slen = lenrange[0];
5284 115809 : if (lenrange[0] != lenrange[1])
5285 : /* Set the minimum length but ignore the maximum
5286 : for now. */
5287 38944 : full_string_p = false;
5288 : }
5289 : else
5290 : slen = -1;
5291 :
5292 115809 : tree len = (slen <= 0
5293 115809 : ? size_zero_node
5294 210212 : : build_int_cst (size_type_node, slen));
5295 210212 : si = new_strinfo (ptr, idx, len, slen >= 0 && full_string_p);
5296 210212 : set_strinfo (idx, si);
5297 210212 : if (storing_all_zeros_p
5298 94403 : && ssaname
5299 242492 : && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (ssaname))
5300 32280 : si->endptr = ssaname;
5301 210212 : si->dont_invalidate = true;
5302 210212 : si->writable = true;
5303 : }
5304 : }
5305 95074 : else if (idx == 0
5306 95074 : && rhs_minlen < HOST_WIDE_INT_M1U
5307 68073 : && ssaname == NULL_TREE
5308 59265 : && TREE_CODE (TREE_TYPE (lhs)) == ARRAY_TYPE)
5309 : {
5310 495 : HOST_WIDE_INT a = int_size_in_bytes (TREE_TYPE (lhs));
5311 495 : if (a > 0 && (unsigned HOST_WIDE_INT) a > rhs_minlen)
5312 : {
5313 495 : int idx = new_addr_stridx (lhs);
5314 495 : if (idx != 0)
5315 : {
5316 495 : si = new_strinfo (build_fold_addr_expr (lhs), idx,
5317 495 : build_int_cst (size_type_node, rhs_minlen),
5318 : full_string_p);
5319 495 : set_strinfo (idx, si);
5320 495 : si->dont_invalidate = true;
5321 : }
5322 : }
5323 : }
5324 :
5325 388991 : if (si != NULL && offset == 0 && storing_all_zeros_p && lenrange[2] == 1)
5326 : {
5327 : /* For single-byte stores only, allow adjust_last_stmt to remove
5328 : the statement if the stored '\0' is immediately overwritten. */
5329 55200 : laststmt.stmt = stmt;
5330 55200 : laststmt.len = build_int_cst (size_type_node, 1);
5331 55200 : laststmt.stridx = si->idx;
5332 : }
5333 : return true;
5334 : }
5335 :
5336 : /* Try to fold strstr (s, t) eq/ne s to strncmp (s, t, strlen (t)) eq/ne 0. */
5337 :
5338 : static void
5339 3508404 : fold_strstr_to_strncmp (tree rhs1, tree rhs2, gimple *stmt)
5340 : {
5341 3508404 : if (TREE_CODE (rhs1) != SSA_NAME
5342 3507869 : || TREE_CODE (rhs2) != SSA_NAME)
5343 : return;
5344 :
5345 2087309 : gimple *call_stmt = NULL;
5346 2087309 : for (int pass = 0; pass < 2; pass++)
5347 : {
5348 1391544 : gimple *g = SSA_NAME_DEF_STMT (rhs1);
5349 1391544 : if (gimple_call_builtin_p (g, BUILT_IN_STRSTR)
5350 26 : && has_single_use (rhs1)
5351 1391555 : && gimple_call_arg (g, 0) == rhs2)
5352 : {
5353 : call_stmt = g;
5354 : break;
5355 : }
5356 1391533 : std::swap (rhs1, rhs2);
5357 : }
5358 :
5359 695776 : if (call_stmt)
5360 : {
5361 11 : tree arg0 = gimple_call_arg (call_stmt, 0);
5362 :
5363 11 : if (arg0 == rhs2)
5364 : {
5365 11 : tree arg1 = gimple_call_arg (call_stmt, 1);
5366 11 : tree arg1_len = NULL_TREE;
5367 11 : int idx = get_stridx (arg1, call_stmt);
5368 :
5369 11 : if (idx)
5370 : {
5371 10 : if (idx < 0)
5372 9 : arg1_len = build_int_cst (size_type_node, ~idx);
5373 : else
5374 : {
5375 1 : strinfo *si = get_strinfo (idx);
5376 1 : if (si)
5377 1 : arg1_len = get_string_length (si);
5378 : }
5379 : }
5380 :
5381 10 : if (arg1_len != NULL_TREE)
5382 : {
5383 10 : gimple_stmt_iterator gsi = gsi_for_stmt (call_stmt);
5384 10 : tree strncmp_decl = builtin_decl_explicit (BUILT_IN_STRNCMP);
5385 :
5386 10 : if (!is_gimple_val (arg1_len))
5387 : {
5388 1 : tree arg1_len_tmp = make_ssa_name (TREE_TYPE (arg1_len));
5389 1 : gassign *arg1_stmt = gimple_build_assign (arg1_len_tmp,
5390 : arg1_len);
5391 1 : gsi_insert_before (&gsi, arg1_stmt, GSI_SAME_STMT);
5392 1 : arg1_len = arg1_len_tmp;
5393 : }
5394 :
5395 10 : gcall *strncmp_call = gimple_build_call (strncmp_decl, 3,
5396 : arg0, arg1, arg1_len);
5397 10 : tree strncmp_lhs = make_ssa_name (integer_type_node);
5398 20 : gimple_set_vuse (strncmp_call, gimple_vuse (call_stmt));
5399 10 : gimple_call_set_lhs (strncmp_call, strncmp_lhs);
5400 10 : gsi_remove (&gsi, true);
5401 10 : gsi_insert_before (&gsi, strncmp_call, GSI_SAME_STMT);
5402 10 : tree zero = build_zero_cst (TREE_TYPE (strncmp_lhs));
5403 :
5404 10 : if (is_gimple_assign (stmt))
5405 : {
5406 8 : if (gimple_assign_rhs_code (stmt) == COND_EXPR)
5407 : {
5408 0 : tree cond = gimple_assign_rhs1 (stmt);
5409 0 : TREE_OPERAND (cond, 0) = strncmp_lhs;
5410 0 : TREE_OPERAND (cond, 1) = zero;
5411 : }
5412 : else
5413 : {
5414 8 : gimple_assign_set_rhs1 (stmt, strncmp_lhs);
5415 8 : gimple_assign_set_rhs2 (stmt, zero);
5416 : }
5417 : }
5418 : else
5419 : {
5420 2 : gcond *cond = as_a<gcond *> (stmt);
5421 2 : gimple_cond_set_lhs (cond, strncmp_lhs);
5422 2 : gimple_cond_set_rhs (cond, zero);
5423 : }
5424 10 : update_stmt (stmt);
5425 : }
5426 : }
5427 : }
5428 : }
5429 :
5430 : /* Return true if TYPE corresponds to a narrow character type. */
5431 :
5432 : static bool
5433 9516104 : is_char_type (tree type)
5434 : {
5435 9516104 : return (TREE_CODE (type) == INTEGER_TYPE
5436 3196739 : && TYPE_MODE (type) == TYPE_MODE (char_type_node)
5437 9956564 : && TYPE_PRECISION (type) == TYPE_PRECISION (char_type_node));
5438 : }
5439 :
5440 : /* Check the built-in call at GSI for validity and optimize it.
5441 : Uses RVALS to determine range information.
5442 : Return true to let the caller advance *GSI to the next statement
5443 : in the basic block and false otherwise. */
5444 :
5445 : bool
5446 5541823 : strlen_pass::check_and_optimize_call (bool *zero_write)
5447 : {
5448 5541823 : gimple *stmt = gsi_stmt (m_gsi);
5449 :
5450 5541823 : if (!gimple_call_builtin_p (stmt, BUILT_IN_NORMAL))
5451 : {
5452 4284423 : tree fntype = gimple_call_fntype (stmt);
5453 4284423 : if (!fntype)
5454 : return true;
5455 :
5456 4125966 : if (lookup_attribute ("alloc_size", TYPE_ATTRIBUTES (fntype)))
5457 : {
5458 61125 : handle_alloc_call (BUILT_IN_NONE);
5459 61125 : return true;
5460 : }
5461 :
5462 4064841 : if (tree lhs = gimple_call_lhs (stmt))
5463 1395475 : handle_assign (lhs, NULL_TREE, zero_write);
5464 :
5465 : /* Proceed to handle user-defined formatting functions. */
5466 : }
5467 :
5468 : /* When not optimizing we must be checking printf calls which
5469 : we do even for user-defined functions when they are declared
5470 : with attribute format. */
5471 5322241 : if (!flag_optimize_strlen
5472 4525452 : || !strlen_optimize
5473 9847693 : || !valid_builtin_call (stmt))
5474 4339765 : return !handle_printf_call (&m_gsi, ptr_qry);
5475 :
5476 982476 : tree callee = gimple_call_fndecl (stmt);
5477 982476 : switch (DECL_FUNCTION_CODE (callee))
5478 : {
5479 14618 : case BUILT_IN_STRLEN:
5480 14618 : case BUILT_IN_STRNLEN:
5481 14618 : handle_builtin_strlen ();
5482 14618 : break;
5483 492 : case BUILT_IN_STRCHR:
5484 492 : handle_builtin_strchr ();
5485 492 : break;
5486 2811 : case BUILT_IN_STRCPY:
5487 2811 : case BUILT_IN_STRCPY_CHK:
5488 2811 : case BUILT_IN_STPCPY:
5489 2811 : case BUILT_IN_STPCPY_CHK:
5490 2811 : handle_builtin_strcpy (DECL_FUNCTION_CODE (callee));
5491 2811 : break;
5492 :
5493 805 : case BUILT_IN_STRNCAT:
5494 805 : case BUILT_IN_STRNCAT_CHK:
5495 805 : handle_builtin_strncat (DECL_FUNCTION_CODE (callee));
5496 805 : break;
5497 :
5498 2495 : case BUILT_IN_STPNCPY:
5499 2495 : case BUILT_IN_STPNCPY_CHK:
5500 2495 : case BUILT_IN_STRNCPY:
5501 2495 : case BUILT_IN_STRNCPY_CHK:
5502 2495 : handle_builtin_stxncpy_strncat (false);
5503 2495 : break;
5504 :
5505 81636 : case BUILT_IN_MEMCPY:
5506 81636 : case BUILT_IN_MEMCPY_CHK:
5507 81636 : case BUILT_IN_MEMPCPY:
5508 81636 : case BUILT_IN_MEMPCPY_CHK:
5509 81636 : handle_builtin_memcpy (DECL_FUNCTION_CODE (callee));
5510 81636 : break;
5511 809 : case BUILT_IN_STRCAT:
5512 809 : case BUILT_IN_STRCAT_CHK:
5513 809 : handle_builtin_strcat (DECL_FUNCTION_CODE (callee));
5514 809 : break;
5515 37771 : case BUILT_IN_ALLOCA:
5516 37771 : case BUILT_IN_ALLOCA_WITH_ALIGN:
5517 37771 : case BUILT_IN_MALLOC:
5518 37771 : case BUILT_IN_CALLOC:
5519 37771 : handle_alloc_call (DECL_FUNCTION_CODE (callee));
5520 37771 : break;
5521 26940 : case BUILT_IN_MEMSET:
5522 26940 : if (handle_builtin_memset (zero_write))
5523 289 : return false;
5524 : break;
5525 129565 : case BUILT_IN_STRCMP:
5526 129565 : case BUILT_IN_STRNCMP:
5527 129565 : if (handle_builtin_string_cmp ())
5528 409 : return false;
5529 : break;
5530 684534 : default:
5531 684534 : if (handle_printf_call (&m_gsi, ptr_qry))
5532 10 : return false;
5533 : break;
5534 : }
5535 :
5536 : return true;
5537 : }
5538 :
5539 : /* Handle an assignment statement at *GSI to a LHS of integral type.
5540 : If GSI's basic block needs clean-up of EH, set *CLEANUP_EH to true. */
5541 :
5542 : void
5543 9611499 : strlen_pass::handle_integral_assign (bool *cleanup_eh)
5544 : {
5545 9611499 : gimple *stmt = gsi_stmt (m_gsi);
5546 9611499 : tree lhs = gimple_assign_lhs (stmt);
5547 9611499 : tree lhs_type = TREE_TYPE (lhs);
5548 :
5549 9611499 : enum tree_code code = gimple_assign_rhs_code (stmt);
5550 9611499 : if (code == COND_EXPR)
5551 : {
5552 10187 : tree cond = gimple_assign_rhs1 (stmt);
5553 10187 : enum tree_code cond_code = TREE_CODE (cond);
5554 :
5555 10187 : if (cond_code == EQ_EXPR || cond_code == NE_EXPR)
5556 0 : fold_strstr_to_strncmp (TREE_OPERAND (cond, 0),
5557 0 : TREE_OPERAND (cond, 1), stmt);
5558 : }
5559 9601312 : else if (code == EQ_EXPR || code == NE_EXPR)
5560 245343 : fold_strstr_to_strncmp (gimple_assign_rhs1 (stmt),
5561 : gimple_assign_rhs2 (stmt), stmt);
5562 9355969 : else if (gimple_assign_load_p (stmt)
5563 2656844 : && TREE_CODE (lhs_type) == INTEGER_TYPE
5564 2498634 : && TYPE_MODE (lhs_type) == TYPE_MODE (char_type_node)
5565 359159 : && (TYPE_PRECISION (lhs_type)
5566 359159 : == TYPE_PRECISION (char_type_node))
5567 9707939 : && !gimple_has_volatile_ops (stmt))
5568 : {
5569 340473 : tree off = integer_zero_node;
5570 340473 : unsigned HOST_WIDE_INT coff = 0;
5571 340473 : int idx = 0;
5572 340473 : tree rhs1 = gimple_assign_rhs1 (stmt);
5573 340473 : if (code == MEM_REF)
5574 : {
5575 164869 : idx = get_stridx (TREE_OPERAND (rhs1, 0), stmt);
5576 164869 : if (idx > 0)
5577 : {
5578 6964 : strinfo *si = get_strinfo (idx);
5579 6964 : if (si
5580 3448 : && si->nonzero_chars
5581 3438 : && TREE_CODE (si->nonzero_chars) == INTEGER_CST
5582 7096 : && (wi::to_widest (si->nonzero_chars)
5583 264 : >= wi::to_widest (off)))
5584 132 : off = TREE_OPERAND (rhs1, 1);
5585 : else
5586 : /* This case is not useful. See if get_addr_stridx
5587 : returns something usable. */
5588 : idx = 0;
5589 : }
5590 : }
5591 6964 : if (idx <= 0)
5592 340341 : idx = get_addr_stridx (rhs1, stmt, NULL_TREE, &coff);
5593 340341 : if (idx > 0)
5594 : {
5595 6570 : strinfo *si = get_strinfo (idx);
5596 6570 : if (si
5597 1169 : && si->nonzero_chars
5598 1169 : && TREE_CODE (si->nonzero_chars) == INTEGER_CST)
5599 : {
5600 169 : widest_int w1 = wi::to_widest (si->nonzero_chars);
5601 169 : widest_int w2 = wi::to_widest (off) + coff;
5602 169 : if (w1 == w2
5603 169 : && si->full_string_p)
5604 : {
5605 31 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
5606 : {
5607 0 : fprintf (dump_file, "Optimizing: ");
5608 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
5609 : }
5610 :
5611 : /* Reading the final '\0' character. */
5612 31 : tree zero = build_int_cst (lhs_type, 0);
5613 31 : gimple_set_vuse (stmt, NULL_TREE);
5614 31 : gimple_assign_set_rhs_from_tree (&m_gsi, zero);
5615 31 : *cleanup_eh
5616 31 : |= maybe_clean_or_replace_eh_stmt (stmt,
5617 : gsi_stmt (m_gsi));
5618 31 : stmt = gsi_stmt (m_gsi);
5619 31 : update_stmt (stmt);
5620 :
5621 31 : if (dump_file && (dump_flags & TDF_DETAILS) != 0)
5622 : {
5623 0 : fprintf (dump_file, "into: ");
5624 0 : print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
5625 : }
5626 : }
5627 138 : else if (w1 > w2)
5628 : {
5629 : /* Reading a character before the final '\0'
5630 : character. Just set the value range to ~[0, 0]
5631 : if we don't have anything better. */
5632 7 : int_range_max r;
5633 14 : if (!get_range_query (cfun)->range_of_expr (r, lhs)
5634 7 : || r.varying_p ())
5635 : {
5636 7 : r.set_nonzero (lhs_type);
5637 7 : set_range_info (lhs, r);
5638 : }
5639 7 : }
5640 169 : }
5641 : }
5642 : }
5643 9015496 : else if (code == MEM_REF && TREE_CODE (lhs) == SSA_NAME)
5644 : {
5645 689999 : if (int idx = new_stridx (lhs))
5646 : {
5647 : /* Record multi-byte assignments from MEM_REFs. */
5648 689999 : bool storing_all_nonzero_p;
5649 689999 : bool storing_all_zeros_p;
5650 689999 : bool full_string_p;
5651 689999 : unsigned lenrange[] = { UINT_MAX, 0, 0 };
5652 689999 : tree rhs = gimple_assign_rhs1 (stmt);
5653 689999 : const bool ranges_valid
5654 689999 : = count_nonzero_bytes (rhs, stmt,
5655 : lenrange, &full_string_p,
5656 : &storing_all_zeros_p,
5657 : &storing_all_nonzero_p);
5658 689999 : if (ranges_valid)
5659 : {
5660 636966 : tree length = build_int_cst (sizetype, lenrange[0]);
5661 636966 : strinfo *si = new_strinfo (lhs, idx, length, full_string_p);
5662 636966 : set_strinfo (idx, si);
5663 636966 : si->writable = true;
5664 636966 : si->dont_invalidate = true;
5665 : }
5666 : }
5667 : }
5668 :
5669 9611499 : if (strlen_to_stridx)
5670 : {
5671 9608952 : tree rhs1 = gimple_assign_rhs1 (stmt);
5672 9608952 : if (stridx_strlenloc *ps = strlen_to_stridx->get (rhs1))
5673 10763 : strlen_to_stridx->put (lhs, stridx_strlenloc (*ps));
5674 : }
5675 9611499 : }
5676 :
5677 : /* Handle assignment statement at *GSI to LHS. Set *ZERO_WRITE if
5678 : the assignment stores all zero bytes. RHS is the rhs of the
5679 : statement if not a call. */
5680 :
5681 : bool
5682 6527793 : strlen_pass::handle_assign (tree lhs, tree rhs, bool *zero_write)
5683 : {
5684 6527793 : tree type = TREE_TYPE (lhs);
5685 6527793 : if (TREE_CODE (type) == ARRAY_TYPE)
5686 117325 : type = TREE_TYPE (type);
5687 :
5688 5132318 : if (rhs && TREE_CODE (rhs) == CONSTRUCTOR
5689 89856 : && TREE_CODE (lhs) == MEM_REF
5690 46995 : && TREE_CODE (TREE_OPERAND (lhs, 0)) == SSA_NAME
5691 6531684 : && integer_zerop (TREE_OPERAND (lhs, 1)))
5692 : {
5693 : /* Set to the non-constant offset added to PTR. */
5694 18155 : wide_int offrng[2];
5695 3631 : gcc_assert (CONSTRUCTOR_NELTS (rhs) == 0);
5696 3631 : tree ptr = TREE_OPERAND (lhs, 0);
5697 3631 : tree len = TYPE_SIZE_UNIT (TREE_TYPE (lhs));
5698 3631 : int idx1 = get_stridx (ptr, gsi_stmt (m_gsi), offrng, ptr_qry.rvals);
5699 3631 : if (idx1 > 0)
5700 : {
5701 129 : strinfo *si1 = get_strinfo (idx1);
5702 10 : if (si1 && si1->stmt
5703 5 : && si1->alloc && is_gimple_call (si1->alloc)
5704 5 : && valid_builtin_call (si1->stmt)
5705 139 : && offrng[0] == 0 && offrng[1] == 0)
5706 : {
5707 5 : gimple *malloc_stmt = si1->stmt;
5708 5 : basic_block malloc_bb = gimple_bb (malloc_stmt);
5709 5 : if ((DECL_FUNCTION_CODE (gimple_call_fndecl (malloc_stmt))
5710 : == BUILT_IN_MALLOC)
5711 5 : && operand_equal_p (len, gimple_call_arg (malloc_stmt, 0), 0)
5712 10 : && allow_memset_malloc_to_calloc (ptr, malloc_bb,
5713 : gsi_bb (m_gsi)))
5714 : {
5715 5 : tree alloc_size = gimple_call_arg (malloc_stmt, 0);
5716 5 : gimple_stmt_iterator gsi1 = gsi_for_stmt (malloc_stmt);
5717 5 : tree calloc_decl = builtin_decl_implicit (BUILT_IN_CALLOC);
5718 5 : update_gimple_call (&gsi1, calloc_decl, 2, alloc_size,
5719 : build_one_cst (size_type_node));
5720 5 : si1->nonzero_chars = build_int_cst (size_type_node, 0);
5721 5 : si1->full_string_p = true;
5722 5 : si1->stmt = gsi_stmt (gsi1);
5723 5 : gimple *stmt = gsi_stmt (m_gsi);
5724 5 : unlink_stmt_vdef (stmt);
5725 5 : gsi_remove (&m_gsi, true);
5726 5 : release_defs (stmt);
5727 5 : return false;
5728 : }
5729 : }
5730 : }
5731 10893 : }
5732 :
5733 6527788 : bool is_char_store = is_char_type (type);
5734 6527788 : if (!is_char_store && TREE_CODE (lhs) == MEM_REF)
5735 : {
5736 : /* To consider stores into char objects via integer types other
5737 : than char but not those to non-character objects, determine
5738 : the type of the destination rather than just the type of
5739 : the access. */
5740 4445933 : for (int i = 0; i != 2; ++i)
5741 : {
5742 2988316 : tree ref = TREE_OPERAND (lhs, i);
5743 2988316 : type = TREE_TYPE (ref);
5744 2988316 : if (POINTER_TYPE_P (type))
5745 2988316 : type = TREE_TYPE (type);
5746 2988316 : if (TREE_CODE (type) == ARRAY_TYPE)
5747 166447 : type = TREE_TYPE (type);
5748 2988316 : if (is_char_type (type))
5749 : {
5750 : is_char_store = true;
5751 : break;
5752 : }
5753 : }
5754 : }
5755 :
5756 : /* Handle a single or multibyte assignment. */
5757 6527788 : if (is_char_store && !handle_store (zero_write))
5758 : return false;
5759 :
5760 : return true;
5761 : }
5762 :
5763 :
5764 : /* Attempt to check for validity of the performed access a single statement
5765 : at *GSI using string length knowledge, and to optimize it.
5766 : If the given basic block needs clean-up of EH, CLEANUP_EH is set to
5767 : true. Return true to let the caller advance *GSI to the next statement
5768 : in the basic block and false otherwise. */
5769 :
5770 : bool
5771 95427104 : strlen_pass::check_and_optimize_stmt (bool *cleanup_eh)
5772 : {
5773 95427104 : gimple *stmt = gsi_stmt (m_gsi);
5774 :
5775 : /* For statements that modify a string, set to true if the write
5776 : is only zeros. */
5777 95427104 : bool zero_write = false;
5778 :
5779 95427104 : if (is_gimple_call (stmt))
5780 : {
5781 5541823 : if (!check_and_optimize_call (&zero_write))
5782 : return false;
5783 : }
5784 89885281 : else if (!flag_optimize_strlen || !strlen_optimize)
5785 : return true;
5786 85104327 : else if (is_gimple_assign (stmt) && !gimple_clobber_p (stmt))
5787 : {
5788 : /* Handle non-clobbering assignment. */
5789 20647489 : tree lhs = gimple_assign_lhs (stmt);
5790 20647489 : tree lhs_type = TREE_TYPE (lhs);
5791 :
5792 20647489 : if (TREE_CODE (lhs) == SSA_NAME && POINTER_TYPE_P (lhs_type))
5793 : {
5794 3833792 : if (gimple_assign_single_p (stmt)
5795 3833792 : || (gimple_assign_cast_p (stmt)
5796 327079 : && POINTER_TYPE_P (TREE_TYPE (gimple_assign_rhs1 (stmt)))))
5797 : {
5798 2764427 : int idx = get_stridx (gimple_assign_rhs1 (stmt), stmt);
5799 2764427 : ssa_ver_to_stridx[SSA_NAME_VERSION (lhs)] = idx;
5800 : }
5801 1069365 : else if (gimple_assign_rhs_code (stmt) == POINTER_PLUS_EXPR)
5802 764874 : handle_pointer_plus ();
5803 : }
5804 16813697 : else if (TREE_CODE (lhs) == SSA_NAME && INTEGRAL_TYPE_P (lhs_type))
5805 : /* Handle assignment to a character. */
5806 9611499 : handle_integral_assign (cleanup_eh);
5807 7202198 : else if (TREE_CODE (lhs) != SSA_NAME && !TREE_SIDE_EFFECTS (lhs))
5808 5132318 : if (!handle_assign (lhs, gimple_assign_rhs1 (stmt), &zero_write))
5809 : return false;
5810 : }
5811 64456838 : else if (gcond *cond = dyn_cast<gcond *> (stmt))
5812 : {
5813 4190309 : enum tree_code code = gimple_cond_code (cond);
5814 4190309 : if (code == EQ_EXPR || code == NE_EXPR)
5815 3263061 : fold_strstr_to_strncmp (gimple_cond_lhs (stmt),
5816 : gimple_cond_rhs (stmt), stmt);
5817 : }
5818 :
5819 118994635 : if (gimple_vdef (stmt))
5820 11285840 : maybe_invalidate (stmt, zero_write);
5821 : return true;
5822 : }
5823 :
5824 : /* Recursively call maybe_invalidate on stmts that might be executed
5825 : in between dombb and current bb and that contain a vdef. Stop when
5826 : *count stmts are inspected, or if the whole strinfo vector has
5827 : been invalidated. */
5828 :
5829 : static void
5830 722907 : do_invalidate (basic_block dombb, gimple *phi, bitmap visited, int *count)
5831 : {
5832 722907 : unsigned int i, n = gimple_phi_num_args (phi);
5833 :
5834 1250131 : for (i = 0; i < n; i++)
5835 : {
5836 1076765 : tree vuse = gimple_phi_arg_def (phi, i);
5837 1076765 : gimple *stmt = SSA_NAME_DEF_STMT (vuse);
5838 1076765 : basic_block bb = gimple_bb (stmt);
5839 1394290 : if (bb == NULL
5840 1076765 : || bb == dombb
5841 934404 : || !bitmap_set_bit (visited, bb->index)
5842 1891075 : || !dominated_by_p (CDI_DOMINATORS, bb, dombb))
5843 317525 : continue;
5844 1087160 : while (1)
5845 : {
5846 1087160 : if (gimple_code (stmt) == GIMPLE_PHI)
5847 : {
5848 199166 : do_invalidate (dombb, stmt, visited, count);
5849 199166 : if (*count == 0)
5850 : return;
5851 : break;
5852 : }
5853 887994 : if (--*count == 0)
5854 : return;
5855 887427 : if (!maybe_invalidate (stmt))
5856 : {
5857 454311 : *count = 0;
5858 454311 : return;
5859 : }
5860 433116 : vuse = gimple_vuse (stmt);
5861 433116 : stmt = SSA_NAME_DEF_STMT (vuse);
5862 433116 : if (gimple_bb (stmt) != bb)
5863 : {
5864 164596 : bb = gimple_bb (stmt);
5865 164596 : if (bb == NULL
5866 164596 : || bb == dombb
5867 107554 : || !bitmap_set_bit (visited, bb->index)
5868 235306 : || !dominated_by_p (CDI_DOMINATORS, bb, dombb))
5869 : break;
5870 : }
5871 : }
5872 : }
5873 : }
5874 :
5875 : /* Release pointer_query cache. */
5876 :
5877 1124058 : strlen_pass::~strlen_pass ()
5878 : {
5879 1124058 : ptr_qry.flush_cache ();
5880 2248116 : }
5881 :
5882 : /* Callback for walk_dominator_tree. Attempt to optimize various
5883 : string ops by remembering string lengths pointed by pointer SSA_NAMEs. */
5884 :
5885 : edge
5886 12208797 : strlen_pass::before_dom_children (basic_block bb)
5887 : {
5888 12208797 : basic_block dombb = get_immediate_dominator (CDI_DOMINATORS, bb);
5889 :
5890 12208797 : if (dombb == NULL)
5891 1124058 : stridx_to_strinfo = NULL;
5892 : else
5893 : {
5894 11084739 : stridx_to_strinfo = ((vec<strinfo *, va_heap, vl_embed> *) dombb->aux);
5895 11084739 : if (stridx_to_strinfo)
5896 : {
5897 3177787 : for (gphi_iterator gsi = gsi_start_phis (bb); !gsi_end_p (gsi);
5898 526811 : gsi_next (&gsi))
5899 : {
5900 1050552 : gphi *phi = gsi.phi ();
5901 2101104 : if (virtual_operand_p (gimple_phi_result (phi)))
5902 : {
5903 523741 : bitmap visited = BITMAP_ALLOC (NULL);
5904 523741 : int count_vdef = 100;
5905 523741 : do_invalidate (dombb, phi, visited, &count_vdef);
5906 523741 : BITMAP_FREE (visited);
5907 523741 : if (count_vdef == 0)
5908 : {
5909 : /* If there were too many vdefs in between immediate
5910 : dominator and current bb, invalidate everything.
5911 : If stridx_to_strinfo has been unshared, we need
5912 : to free it, otherwise just set it to NULL. */
5913 454878 : if (!strinfo_shared ())
5914 : {
5915 : unsigned int i;
5916 : strinfo *si;
5917 :
5918 712474 : for (i = 1;
5919 770212 : vec_safe_iterate (stridx_to_strinfo, i, &si);
5920 : ++i)
5921 : {
5922 712474 : free_strinfo (si);
5923 712474 : (*stridx_to_strinfo)[i] = NULL;
5924 : }
5925 : }
5926 : else
5927 397140 : stridx_to_strinfo = NULL;
5928 : }
5929 523741 : break;
5930 : }
5931 : }
5932 : }
5933 : }
5934 :
5935 : /* If all PHI arguments have the same string index, the PHI result
5936 : has it as well. */
5937 17197948 : for (gphi_iterator gsi = gsi_start_phis (bb); !gsi_end_p (gsi);
5938 4989151 : gsi_next (&gsi))
5939 : {
5940 4989151 : gphi *phi = gsi.phi ();
5941 4989151 : tree result = gimple_phi_result (phi);
5942 9978302 : if (!virtual_operand_p (result) && POINTER_TYPE_P (TREE_TYPE (result)))
5943 : {
5944 869078 : int idx = get_stridx (gimple_phi_arg_def (phi, 0), phi);
5945 869078 : if (idx != 0)
5946 : {
5947 41406 : unsigned int i, n = gimple_phi_num_args (phi);
5948 52268 : for (i = 1; i < n; i++)
5949 41571 : if (idx != get_stridx (gimple_phi_arg_def (phi, i), phi))
5950 : break;
5951 41406 : if (i == n)
5952 10697 : ssa_ver_to_stridx[SSA_NAME_VERSION (result)] = idx;
5953 : }
5954 : }
5955 : }
5956 :
5957 12208797 : bool cleanup_eh = false;
5958 :
5959 : /* Attempt to optimize individual statements. */
5960 132053495 : for (m_gsi = gsi_start_bb (bb); !gsi_end_p (m_gsi); )
5961 : {
5962 : /* Reset search depth performance counter. */
5963 95427104 : ptr_qry.depth = 0;
5964 :
5965 95427104 : if (check_and_optimize_stmt (&cleanup_eh))
5966 95425688 : gsi_next (&m_gsi);
5967 : }
5968 :
5969 12208797 : if (cleanup_eh && gimple_purge_dead_eh_edges (bb))
5970 0 : m_cleanup_cfg = true;
5971 :
5972 12208797 : bb->aux = stridx_to_strinfo;
5973 12208797 : if (vec_safe_length (stridx_to_strinfo) && !strinfo_shared ())
5974 910164 : (*stridx_to_strinfo)[0] = (strinfo *) bb;
5975 12208797 : return NULL;
5976 : }
5977 :
5978 : /* Callback for walk_dominator_tree. Free strinfo vector if it is
5979 : owned by the current bb, clear bb->aux. */
5980 :
5981 : void
5982 12208797 : strlen_pass::after_dom_children (basic_block bb)
5983 : {
5984 12208797 : if (bb->aux)
5985 : {
5986 2790840 : stridx_to_strinfo = ((vec<strinfo *, va_heap, vl_embed> *) bb->aux);
5987 5581680 : if (vec_safe_length (stridx_to_strinfo)
5988 2790840 : && (*stridx_to_strinfo)[0] == (strinfo *) bb)
5989 : {
5990 : unsigned int i;
5991 : strinfo *si;
5992 :
5993 14647832 : for (i = 1; vec_safe_iterate (stridx_to_strinfo, i, &si); ++i)
5994 13737668 : free_strinfo (si);
5995 910164 : vec_free (stridx_to_strinfo);
5996 : }
5997 2790840 : bb->aux = NULL;
5998 : }
5999 12208797 : }
6000 :
6001 : namespace {
6002 :
6003 : static unsigned int
6004 1124058 : printf_strlen_execute (function *fun, bool warn_only)
6005 : {
6006 1124058 : strlen_optimize = !warn_only;
6007 :
6008 1124058 : calculate_dominance_info (CDI_DOMINATORS);
6009 1124058 : loop_optimizer_init (LOOPS_NORMAL);
6010 1124058 : scev_initialize ();
6011 :
6012 1124058 : gcc_assert (!strlen_to_stridx);
6013 1124058 : if (warn_stringop_overflow || warn_stringop_truncation)
6014 1123818 : strlen_to_stridx = new hash_map<tree, stridx_strlenloc> ();
6015 :
6016 : /* This has to happen after initializing the loop optimizer
6017 : and initializing SCEV as they create new SSA_NAMEs. */
6018 2248116 : ssa_ver_to_stridx.safe_grow_cleared (num_ssa_names, true);
6019 1124058 : max_stridx = 1;
6020 :
6021 1124058 : enable_ranger (fun);
6022 : /* String length optimization is implemented as a walk of the dominator
6023 : tree and a forward walk of statements within each block. */
6024 1124058 : strlen_pass walker (fun, CDI_DOMINATORS);
6025 1124058 : walker.walk (ENTRY_BLOCK_PTR_FOR_FN (fun));
6026 :
6027 1124058 : if (dump_file && (dump_flags & TDF_DETAILS))
6028 0 : walker.ptr_qry.dump (dump_file, true);
6029 :
6030 1124058 : ssa_ver_to_stridx.release ();
6031 1124058 : strinfo_pool.release ();
6032 1124058 : if (decl_to_stridxlist_htab)
6033 : {
6034 46973 : obstack_free (&stridx_obstack, NULL);
6035 93946 : delete decl_to_stridxlist_htab;
6036 46973 : decl_to_stridxlist_htab = NULL;
6037 : }
6038 1124058 : laststmt.stmt = NULL;
6039 1124058 : laststmt.len = NULL_TREE;
6040 1124058 : laststmt.stridx = 0;
6041 :
6042 1124058 : if (strlen_to_stridx)
6043 : {
6044 1123818 : strlen_to_stridx->empty ();
6045 2247636 : delete strlen_to_stridx;
6046 1123818 : strlen_to_stridx = NULL;
6047 : }
6048 :
6049 1124058 : disable_ranger (fun);
6050 1124058 : scev_finalize ();
6051 1124058 : loop_optimizer_finalize ();
6052 1124058 : free_dominance_info (CDI_POST_DOMINATORS);
6053 :
6054 1124058 : return walker.m_cleanup_cfg ? TODO_cleanup_cfg : 0;
6055 1124058 : }
6056 :
6057 : /* This file defines two passes: one for warnings that runs only when
6058 : optimization is disabled, and another that implements optimizations
6059 : and also issues warnings. */
6060 :
6061 : const pass_data pass_data_warn_printf =
6062 : {
6063 : GIMPLE_PASS, /* type */
6064 : "warn-printf", /* name */
6065 : OPTGROUP_NONE, /* optinfo_flags */
6066 : TV_NONE, /* tv_id */
6067 : /* Normally an optimization pass would require PROP_ssa but because
6068 : this pass runs early, with no optimization, to do sprintf format
6069 : checking, it only requires PROP_cfg. */
6070 : PROP_cfg, /* properties_required */
6071 : 0, /* properties_provided */
6072 : 0, /* properties_destroyed */
6073 : 0, /* todo_flags_start */
6074 : 0, /* todo_flags_finish */
6075 : };
6076 :
6077 : class pass_warn_printf : public gimple_opt_pass
6078 : {
6079 : public:
6080 294196 : pass_warn_printf (gcc::context *ctxt)
6081 588392 : : gimple_opt_pass (pass_data_warn_printf, ctxt)
6082 : {}
6083 :
6084 : bool gate (function *) final override;
6085 61299 : unsigned int execute (function *fun) final override
6086 : {
6087 61299 : return printf_strlen_execute (fun, true);
6088 : }
6089 : };
6090 :
6091 :
6092 : /* Return true to run the warning pass only when not optimizing and
6093 : iff either -Wformat-overflow or -Wformat-truncation is specified. */
6094 :
6095 : bool
6096 3004029 : pass_warn_printf::gate (function *)
6097 : {
6098 3004029 : return !optimize && (warn_format_overflow > 0 || warn_format_trunc > 0);
6099 : }
6100 :
6101 : const pass_data pass_data_strlen =
6102 : {
6103 : GIMPLE_PASS, /* type */
6104 : "strlen", /* name */
6105 : OPTGROUP_NONE, /* optinfo_flags */
6106 : TV_TREE_STRLEN, /* tv_id */
6107 : PROP_cfg | PROP_ssa, /* properties_required */
6108 : 0, /* properties_provided */
6109 : 0, /* properties_destroyed */
6110 : 0, /* todo_flags_start */
6111 : 0, /* todo_flags_finish */
6112 : };
6113 :
6114 : class pass_strlen : public gimple_opt_pass
6115 : {
6116 : public:
6117 588392 : pass_strlen (gcc::context *ctxt)
6118 1176784 : : gimple_opt_pass (pass_data_strlen, ctxt)
6119 : {}
6120 :
6121 294196 : opt_pass * clone () final override { return new pass_strlen (m_ctxt); }
6122 :
6123 : bool gate (function *) final override;
6124 1062759 : unsigned int execute (function *fun) final override
6125 : {
6126 1062759 : return printf_strlen_execute (fun, false);
6127 : }
6128 : };
6129 :
6130 : /* Return true to run the pass only when the sprintf and/or strlen
6131 : optimizations are enabled and -Wformat-overflow or -Wformat-truncation
6132 : are specified. */
6133 :
6134 : bool
6135 1062796 : pass_strlen::gate (function *)
6136 : {
6137 1062796 : return ((warn_format_overflow > 0
6138 949772 : || warn_format_trunc > 0
6139 949767 : || warn_restrict > 0
6140 949580 : || flag_optimize_strlen > 0
6141 124620 : || flag_printf_return_value)
6142 2012556 : && optimize > 0);
6143 : }
6144 :
6145 : } // anon namespace
6146 :
6147 : gimple_opt_pass *
6148 294196 : make_pass_warn_printf (gcc::context *ctxt)
6149 : {
6150 294196 : return new pass_warn_printf (ctxt);
6151 : }
6152 :
6153 : gimple_opt_pass *
6154 294196 : make_pass_strlen (gcc::context *ctxt)
6155 : {
6156 294196 : return new pass_strlen (ctxt);
6157 : }
|
