i386: move alignment defaults to processor_costs.
[official-gcc.git] / gcc / tree-ssa-strlen.c
blob3fa5ef56f38e726276d9146e542aff324d4c8c27
1 /* String length optimization
2 Copyright (C) 2011-2018 Free Software Foundation, Inc.
3 Contributed by Jakub Jelinek <jakub@redhat.com>
5 This file is part of GCC.
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.
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.
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/>. */
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-restrict.h"
34 #include "fold-const.h"
35 #include "stor-layout.h"
36 #include "gimple-fold.h"
37 #include "tree-eh.h"
38 #include "gimplify.h"
39 #include "gimple-iterator.h"
40 #include "gimplify-me.h"
41 #include "expr.h"
42 #include "tree-cfg.h"
43 #include "tree-dfa.h"
44 #include "domwalk.h"
45 #include "tree-ssa-alias.h"
46 #include "tree-ssa-propagate.h"
47 #include "tree-ssa-strlen.h"
48 #include "params.h"
49 #include "tree-hash-traits.h"
50 #include "tree-object-size.h"
51 #include "builtins.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"
59 /* A vector indexed by SSA_NAME_VERSION. 0 means unknown, positive value
60 is an index into strinfo vector, negative value stands for
61 string length of a string literal (~strlen). */
62 static vec<int> ssa_ver_to_stridx;
64 /* Number of currently active string indexes plus one. */
65 static int max_stridx;
67 /* String information record. */
68 struct strinfo
70 /* Number of leading characters that are known to be nonzero. This is
71 also the length of the string if FULL_STRING_P.
73 The values in a list of related string pointers must be consistent;
74 that is, if strinfo B comes X bytes after strinfo A, it must be
75 the case that A->nonzero_chars == X + B->nonzero_chars. */
76 tree nonzero_chars;
77 /* Any of the corresponding pointers for querying alias oracle. */
78 tree ptr;
79 /* This is used for two things:
81 - To record the statement that should be used for delayed length
82 computations. We maintain the invariant that all related strinfos
83 have delayed lengths or none do.
85 - To record the malloc or calloc call that produced this result. */
86 gimple *stmt;
87 /* Pointer to '\0' if known, if NULL, it can be computed as
88 ptr + length. */
89 tree endptr;
90 /* Reference count. Any changes to strinfo entry possibly shared
91 with dominating basic blocks need unshare_strinfo first, except
92 for dont_invalidate which affects only the immediately next
93 maybe_invalidate. */
94 int refcount;
95 /* Copy of index. get_strinfo (si->idx) should return si; */
96 int idx;
97 /* These 3 fields are for chaining related string pointers together.
98 E.g. for
99 bl = strlen (b); dl = strlen (d); strcpy (a, b); c = a + bl;
100 strcpy (c, d); e = c + dl;
101 strinfo(a) -> strinfo(c) -> strinfo(e)
102 All have ->first field equal to strinfo(a)->idx and are doubly
103 chained through prev/next fields. The later strinfos are required
104 to point into the same string with zero or more bytes after
105 the previous pointer and all bytes in between the two pointers
106 must be non-zero. Functions like strcpy or memcpy are supposed
107 to adjust all previous strinfo lengths, but not following strinfo
108 lengths (those are uncertain, usually invalidated during
109 maybe_invalidate, except when the alias oracle knows better).
110 Functions like strcat on the other side adjust the whole
111 related strinfo chain.
112 They are updated lazily, so to use the chain the same first fields
113 and si->prev->next == si->idx needs to be verified. */
114 int first;
115 int next;
116 int prev;
117 /* A flag whether the string is known to be written in the current
118 function. */
119 bool writable;
120 /* A flag for the next maybe_invalidate that this strinfo shouldn't
121 be invalidated. Always cleared by maybe_invalidate. */
122 bool dont_invalidate;
123 /* True if the string is known to be nul-terminated after NONZERO_CHARS
124 characters. False is useful when detecting strings that are built
125 up via successive memcpys. */
126 bool full_string_p;
129 /* Pool for allocating strinfo_struct entries. */
130 static object_allocator<strinfo> strinfo_pool ("strinfo pool");
132 /* Vector mapping positive string indexes to strinfo, for the
133 current basic block. The first pointer in the vector is special,
134 it is either NULL, meaning the vector isn't shared, or it is
135 a basic block pointer to the owner basic_block if shared.
136 If some other bb wants to modify the vector, the vector needs
137 to be unshared first, and only the owner bb is supposed to free it. */
138 static vec<strinfo *, va_heap, vl_embed> *stridx_to_strinfo;
140 /* One OFFSET->IDX mapping. */
141 struct stridxlist
143 struct stridxlist *next;
144 HOST_WIDE_INT offset;
145 int idx;
148 /* Hash table entry, mapping a DECL to a chain of OFFSET->IDX mappings. */
149 struct decl_stridxlist_map
151 struct tree_map_base base;
152 struct stridxlist list;
155 /* Hash table for mapping decls to a chained list of offset -> idx
156 mappings. */
157 static hash_map<tree_decl_hash, stridxlist> *decl_to_stridxlist_htab;
159 /* Hash table mapping strlen calls to stridx instances describing
160 the calls' arguments. Non-null only when warn_stringop_truncation
161 is non-zero. */
162 typedef std::pair<int, location_t> stridx_strlenloc;
163 static hash_map<tree, stridx_strlenloc> *strlen_to_stridx;
165 /* Obstack for struct stridxlist and struct decl_stridxlist_map. */
166 static struct obstack stridx_obstack;
168 /* Last memcpy statement if it could be adjusted if the trailing
169 '\0' written is immediately overwritten, or
170 *x = '\0' store that could be removed if it is immediately overwritten. */
171 struct laststmt_struct
173 gimple *stmt;
174 tree len;
175 int stridx;
176 } laststmt;
178 static int get_stridx_plus_constant (strinfo *, unsigned HOST_WIDE_INT, tree);
179 static void handle_builtin_stxncpy (built_in_function, gimple_stmt_iterator *);
181 /* Return:
183 - 1 if SI is known to start with more than OFF nonzero characters.
185 - 0 if SI is known to start with OFF nonzero characters,
186 but is not known to start with more.
188 - -1 if SI might not start with OFF nonzero characters. */
190 static inline int
191 compare_nonzero_chars (strinfo *si, unsigned HOST_WIDE_INT off)
193 if (si->nonzero_chars
194 && TREE_CODE (si->nonzero_chars) == INTEGER_CST)
195 return compare_tree_int (si->nonzero_chars, off);
196 else
197 return -1;
200 /* Return true if SI is known to be a zero-length string. */
202 static inline bool
203 zero_length_string_p (strinfo *si)
205 return si->full_string_p && integer_zerop (si->nonzero_chars);
208 /* Return strinfo vector entry IDX. */
210 static inline strinfo *
211 get_strinfo (int idx)
213 if (vec_safe_length (stridx_to_strinfo) <= (unsigned int) idx)
214 return NULL;
215 return (*stridx_to_strinfo)[idx];
218 /* Get the next strinfo in the chain after SI, or null if none. */
220 static inline strinfo *
221 get_next_strinfo (strinfo *si)
223 if (si->next == 0)
224 return NULL;
225 strinfo *nextsi = get_strinfo (si->next);
226 if (nextsi == NULL || nextsi->first != si->first || nextsi->prev != si->idx)
227 return NULL;
228 return nextsi;
231 /* Helper function for get_stridx. Return the strinfo index of the address
232 of EXP, which is available in PTR if nonnull. If OFFSET_OUT, it is
233 OK to return the index for some X <= &EXP and store &EXP - X in
234 *OFFSET_OUT. */
236 static int
237 get_addr_stridx (tree exp, tree ptr, unsigned HOST_WIDE_INT *offset_out)
239 HOST_WIDE_INT off;
240 struct stridxlist *list, *last = NULL;
241 tree base;
243 if (!decl_to_stridxlist_htab)
244 return 0;
246 poly_int64 poff;
247 base = get_addr_base_and_unit_offset (exp, &poff);
248 if (base == NULL || !DECL_P (base) || !poff.is_constant (&off))
249 return 0;
251 list = decl_to_stridxlist_htab->get (base);
252 if (list == NULL)
253 return 0;
257 if (list->offset == off)
259 if (offset_out)
260 *offset_out = 0;
261 return list->idx;
263 if (list->offset > off)
264 return 0;
265 last = list;
266 list = list->next;
268 while (list);
270 if ((offset_out || ptr) && last && last->idx > 0)
272 unsigned HOST_WIDE_INT rel_off
273 = (unsigned HOST_WIDE_INT) off - last->offset;
274 strinfo *si = get_strinfo (last->idx);
275 if (si && compare_nonzero_chars (si, rel_off) >= 0)
277 if (offset_out)
279 *offset_out = rel_off;
280 return last->idx;
282 else
283 return get_stridx_plus_constant (si, rel_off, ptr);
286 return 0;
289 /* Return string index for EXP. */
291 static int
292 get_stridx (tree exp)
294 if (TREE_CODE (exp) == SSA_NAME)
296 if (ssa_ver_to_stridx[SSA_NAME_VERSION (exp)])
297 return ssa_ver_to_stridx[SSA_NAME_VERSION (exp)];
298 int i;
299 tree e = exp;
300 HOST_WIDE_INT off = 0;
301 for (i = 0; i < 5; i++)
303 gimple *def_stmt = SSA_NAME_DEF_STMT (e);
304 if (!is_gimple_assign (def_stmt)
305 || gimple_assign_rhs_code (def_stmt) != POINTER_PLUS_EXPR)
306 return 0;
307 tree rhs1 = gimple_assign_rhs1 (def_stmt);
308 tree rhs2 = gimple_assign_rhs2 (def_stmt);
309 if (TREE_CODE (rhs1) != SSA_NAME
310 || !tree_fits_shwi_p (rhs2))
311 return 0;
312 HOST_WIDE_INT this_off = tree_to_shwi (rhs2);
313 if (this_off < 0)
314 return 0;
315 off = (unsigned HOST_WIDE_INT) off + this_off;
316 if (off < 0)
317 return 0;
318 if (ssa_ver_to_stridx[SSA_NAME_VERSION (rhs1)])
320 strinfo *si
321 = get_strinfo (ssa_ver_to_stridx[SSA_NAME_VERSION (rhs1)]);
322 if (si && compare_nonzero_chars (si, off) >= 0)
323 return get_stridx_plus_constant (si, off, exp);
325 e = rhs1;
327 return 0;
330 if (TREE_CODE (exp) == ADDR_EXPR)
332 int idx = get_addr_stridx (TREE_OPERAND (exp, 0), exp, NULL);
333 if (idx != 0)
334 return idx;
337 const char *p = c_getstr (exp);
338 if (p)
339 return ~(int) strlen (p);
341 return 0;
344 /* Return true if strinfo vector is shared with the immediate dominator. */
346 static inline bool
347 strinfo_shared (void)
349 return vec_safe_length (stridx_to_strinfo)
350 && (*stridx_to_strinfo)[0] != NULL;
353 /* Unshare strinfo vector that is shared with the immediate dominator. */
355 static void
356 unshare_strinfo_vec (void)
358 strinfo *si;
359 unsigned int i = 0;
361 gcc_assert (strinfo_shared ());
362 stridx_to_strinfo = vec_safe_copy (stridx_to_strinfo);
363 for (i = 1; vec_safe_iterate (stridx_to_strinfo, i, &si); ++i)
364 if (si != NULL)
365 si->refcount++;
366 (*stridx_to_strinfo)[0] = NULL;
369 /* Attempt to create a string index for exp, ADDR_EXPR's operand.
370 Return a pointer to the location where the string index can
371 be stored (if 0) or is stored, or NULL if this can't be tracked. */
373 static int *
374 addr_stridxptr (tree exp)
376 HOST_WIDE_INT off;
378 poly_int64 poff;
379 tree base = get_addr_base_and_unit_offset (exp, &poff);
380 if (base == NULL_TREE || !DECL_P (base) || !poff.is_constant (&off))
381 return NULL;
383 if (!decl_to_stridxlist_htab)
385 decl_to_stridxlist_htab
386 = new hash_map<tree_decl_hash, stridxlist> (64);
387 gcc_obstack_init (&stridx_obstack);
390 bool existed;
391 stridxlist *list = &decl_to_stridxlist_htab->get_or_insert (base, &existed);
392 if (existed)
394 int i;
395 stridxlist *before = NULL;
396 for (i = 0; i < 32; i++)
398 if (list->offset == off)
399 return &list->idx;
400 if (list->offset > off && before == NULL)
401 before = list;
402 if (list->next == NULL)
403 break;
404 list = list->next;
406 if (i == 32)
407 return NULL;
408 if (before)
410 list = before;
411 before = XOBNEW (&stridx_obstack, struct stridxlist);
412 *before = *list;
413 list->next = before;
414 list->offset = off;
415 list->idx = 0;
416 return &list->idx;
418 list->next = XOBNEW (&stridx_obstack, struct stridxlist);
419 list = list->next;
422 list->next = NULL;
423 list->offset = off;
424 list->idx = 0;
425 return &list->idx;
428 /* Create a new string index, or return 0 if reached limit. */
430 static int
431 new_stridx (tree exp)
433 int idx;
434 if (max_stridx >= PARAM_VALUE (PARAM_MAX_TRACKED_STRLENS))
435 return 0;
436 if (TREE_CODE (exp) == SSA_NAME)
438 if (SSA_NAME_OCCURS_IN_ABNORMAL_PHI (exp))
439 return 0;
440 idx = max_stridx++;
441 ssa_ver_to_stridx[SSA_NAME_VERSION (exp)] = idx;
442 return idx;
444 if (TREE_CODE (exp) == ADDR_EXPR)
446 int *pidx = addr_stridxptr (TREE_OPERAND (exp, 0));
447 if (pidx != NULL)
449 gcc_assert (*pidx == 0);
450 *pidx = max_stridx++;
451 return *pidx;
454 return 0;
457 /* Like new_stridx, but for ADDR_EXPR's operand instead. */
459 static int
460 new_addr_stridx (tree exp)
462 int *pidx;
463 if (max_stridx >= PARAM_VALUE (PARAM_MAX_TRACKED_STRLENS))
464 return 0;
465 pidx = addr_stridxptr (exp);
466 if (pidx != NULL)
468 gcc_assert (*pidx == 0);
469 *pidx = max_stridx++;
470 return *pidx;
472 return 0;
475 /* Create a new strinfo. */
477 static strinfo *
478 new_strinfo (tree ptr, int idx, tree nonzero_chars, bool full_string_p)
480 strinfo *si = strinfo_pool.allocate ();
481 si->nonzero_chars = nonzero_chars;
482 si->ptr = ptr;
483 si->stmt = NULL;
484 si->endptr = NULL_TREE;
485 si->refcount = 1;
486 si->idx = idx;
487 si->first = 0;
488 si->prev = 0;
489 si->next = 0;
490 si->writable = false;
491 si->dont_invalidate = false;
492 si->full_string_p = full_string_p;
493 return si;
496 /* Decrease strinfo refcount and free it if not referenced anymore. */
498 static inline void
499 free_strinfo (strinfo *si)
501 if (si && --si->refcount == 0)
502 strinfo_pool.remove (si);
505 /* Set strinfo in the vector entry IDX to SI. */
507 static inline void
508 set_strinfo (int idx, strinfo *si)
510 if (vec_safe_length (stridx_to_strinfo) && (*stridx_to_strinfo)[0])
511 unshare_strinfo_vec ();
512 if (vec_safe_length (stridx_to_strinfo) <= (unsigned int) idx)
513 vec_safe_grow_cleared (stridx_to_strinfo, idx + 1);
514 (*stridx_to_strinfo)[idx] = si;
517 /* Return the first strinfo in the related strinfo chain
518 if all strinfos in between belong to the chain, otherwise NULL. */
520 static strinfo *
521 verify_related_strinfos (strinfo *origsi)
523 strinfo *si = origsi, *psi;
525 if (origsi->first == 0)
526 return NULL;
527 for (; si->prev; si = psi)
529 if (si->first != origsi->first)
530 return NULL;
531 psi = get_strinfo (si->prev);
532 if (psi == NULL)
533 return NULL;
534 if (psi->next != si->idx)
535 return NULL;
537 if (si->idx != si->first)
538 return NULL;
539 return si;
542 /* Set SI's endptr to ENDPTR and compute its length based on SI->ptr.
543 Use LOC for folding. */
545 static void
546 set_endptr_and_length (location_t loc, strinfo *si, tree endptr)
548 si->endptr = endptr;
549 si->stmt = NULL;
550 tree start_as_size = fold_convert_loc (loc, size_type_node, si->ptr);
551 tree end_as_size = fold_convert_loc (loc, size_type_node, endptr);
552 si->nonzero_chars = fold_build2_loc (loc, MINUS_EXPR, size_type_node,
553 end_as_size, start_as_size);
554 si->full_string_p = true;
557 /* Return string length, or NULL if it can't be computed. */
559 static tree
560 get_string_length (strinfo *si)
562 if (si->nonzero_chars)
563 return si->full_string_p ? si->nonzero_chars : NULL;
565 if (si->stmt)
567 gimple *stmt = si->stmt, *lenstmt;
568 tree callee, lhs, fn, tem;
569 location_t loc;
570 gimple_stmt_iterator gsi;
572 gcc_assert (is_gimple_call (stmt));
573 callee = gimple_call_fndecl (stmt);
574 gcc_assert (callee && fndecl_built_in_p (callee, BUILT_IN_NORMAL));
575 lhs = gimple_call_lhs (stmt);
576 /* unshare_strinfo is intentionally not called here. The (delayed)
577 transformation of strcpy or strcat into stpcpy is done at the place
578 of the former strcpy/strcat call and so can affect all the strinfos
579 with the same stmt. If they were unshared before and transformation
580 has been already done, the handling of BUILT_IN_STPCPY{,_CHK} should
581 just compute the right length. */
582 switch (DECL_FUNCTION_CODE (callee))
584 case BUILT_IN_STRCAT:
585 case BUILT_IN_STRCAT_CHK:
586 gsi = gsi_for_stmt (stmt);
587 fn = builtin_decl_implicit (BUILT_IN_STRLEN);
588 gcc_assert (lhs == NULL_TREE);
589 tem = unshare_expr (gimple_call_arg (stmt, 0));
590 lenstmt = gimple_build_call (fn, 1, tem);
591 lhs = make_ssa_name (TREE_TYPE (TREE_TYPE (fn)), lenstmt);
592 gimple_call_set_lhs (lenstmt, lhs);
593 gimple_set_vuse (lenstmt, gimple_vuse (stmt));
594 gsi_insert_before (&gsi, lenstmt, GSI_SAME_STMT);
595 tem = gimple_call_arg (stmt, 0);
596 if (!ptrofftype_p (TREE_TYPE (lhs)))
598 lhs = convert_to_ptrofftype (lhs);
599 lhs = force_gimple_operand_gsi (&gsi, lhs, true, NULL_TREE,
600 true, GSI_SAME_STMT);
602 lenstmt = gimple_build_assign
603 (make_ssa_name (TREE_TYPE (gimple_call_arg (stmt, 0))),
604 POINTER_PLUS_EXPR,tem, lhs);
605 gsi_insert_before (&gsi, lenstmt, GSI_SAME_STMT);
606 gimple_call_set_arg (stmt, 0, gimple_assign_lhs (lenstmt));
607 lhs = NULL_TREE;
608 /* FALLTHRU */
609 case BUILT_IN_STRCPY:
610 case BUILT_IN_STRCPY_CHK:
611 gcc_assert (builtin_decl_implicit_p (BUILT_IN_STPCPY));
612 if (gimple_call_num_args (stmt) == 2)
613 fn = builtin_decl_implicit (BUILT_IN_STPCPY);
614 else
615 fn = builtin_decl_explicit (BUILT_IN_STPCPY_CHK);
616 gcc_assert (lhs == NULL_TREE);
617 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
619 fprintf (dump_file, "Optimizing: ");
620 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
622 gimple_call_set_fndecl (stmt, fn);
623 lhs = make_ssa_name (TREE_TYPE (TREE_TYPE (fn)), stmt);
624 gimple_call_set_lhs (stmt, lhs);
625 update_stmt (stmt);
626 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
628 fprintf (dump_file, "into: ");
629 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
631 /* FALLTHRU */
632 case BUILT_IN_STPCPY:
633 case BUILT_IN_STPCPY_CHK:
634 gcc_assert (lhs != NULL_TREE);
635 loc = gimple_location (stmt);
636 set_endptr_and_length (loc, si, lhs);
637 for (strinfo *chainsi = verify_related_strinfos (si);
638 chainsi != NULL;
639 chainsi = get_next_strinfo (chainsi))
640 if (chainsi->nonzero_chars == NULL)
641 set_endptr_and_length (loc, chainsi, lhs);
642 break;
643 case BUILT_IN_MALLOC:
644 break;
645 /* BUILT_IN_CALLOC always has si->nonzero_chars set. */
646 default:
647 gcc_unreachable ();
648 break;
652 return si->nonzero_chars;
655 /* Invalidate string length information for strings whose length
656 might change due to stores in stmt. */
658 static bool
659 maybe_invalidate (gimple *stmt)
661 strinfo *si;
662 unsigned int i;
663 bool nonempty = false;
665 for (i = 1; vec_safe_iterate (stridx_to_strinfo, i, &si); ++i)
666 if (si != NULL)
668 if (!si->dont_invalidate)
670 ao_ref r;
671 /* Do not use si->nonzero_chars. */
672 ao_ref_init_from_ptr_and_size (&r, si->ptr, NULL_TREE);
673 if (stmt_may_clobber_ref_p_1 (stmt, &r))
675 set_strinfo (i, NULL);
676 free_strinfo (si);
677 continue;
680 si->dont_invalidate = false;
681 nonempty = true;
683 return nonempty;
686 /* Unshare strinfo record SI, if it has refcount > 1 or
687 if stridx_to_strinfo vector is shared with some other
688 bbs. */
690 static strinfo *
691 unshare_strinfo (strinfo *si)
693 strinfo *nsi;
695 if (si->refcount == 1 && !strinfo_shared ())
696 return si;
698 nsi = new_strinfo (si->ptr, si->idx, si->nonzero_chars, si->full_string_p);
699 nsi->stmt = si->stmt;
700 nsi->endptr = si->endptr;
701 nsi->first = si->first;
702 nsi->prev = si->prev;
703 nsi->next = si->next;
704 nsi->writable = si->writable;
705 set_strinfo (si->idx, nsi);
706 free_strinfo (si);
707 return nsi;
710 /* Attempt to create a new strinfo for BASESI + OFF, or find existing
711 strinfo if there is any. Return it's idx, or 0 if no strinfo has
712 been created. */
714 static int
715 get_stridx_plus_constant (strinfo *basesi, unsigned HOST_WIDE_INT off,
716 tree ptr)
718 if (TREE_CODE (ptr) == SSA_NAME && SSA_NAME_OCCURS_IN_ABNORMAL_PHI (ptr))
719 return 0;
721 if (compare_nonzero_chars (basesi, off) < 0
722 || !tree_fits_uhwi_p (basesi->nonzero_chars))
723 return 0;
725 unsigned HOST_WIDE_INT nonzero_chars
726 = tree_to_uhwi (basesi->nonzero_chars) - off;
727 strinfo *si = basesi, *chainsi;
728 if (si->first || si->prev || si->next)
729 si = verify_related_strinfos (basesi);
730 if (si == NULL
731 || si->nonzero_chars == NULL_TREE
732 || TREE_CODE (si->nonzero_chars) != INTEGER_CST)
733 return 0;
735 if (TREE_CODE (ptr) == SSA_NAME
736 && ssa_ver_to_stridx.length () <= SSA_NAME_VERSION (ptr))
737 ssa_ver_to_stridx.safe_grow_cleared (num_ssa_names);
739 gcc_checking_assert (compare_tree_int (si->nonzero_chars, off) != -1);
740 for (chainsi = si; chainsi->next; chainsi = si)
742 si = get_next_strinfo (chainsi);
743 if (si == NULL
744 || si->nonzero_chars == NULL_TREE
745 || TREE_CODE (si->nonzero_chars) != INTEGER_CST)
746 break;
747 int r = compare_tree_int (si->nonzero_chars, nonzero_chars);
748 if (r != 1)
750 if (r == 0)
752 if (TREE_CODE (ptr) == SSA_NAME)
753 ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)] = si->idx;
754 else
756 int *pidx = addr_stridxptr (TREE_OPERAND (ptr, 0));
757 if (pidx != NULL && *pidx == 0)
758 *pidx = si->idx;
760 return si->idx;
762 break;
766 int idx = new_stridx (ptr);
767 if (idx == 0)
768 return 0;
769 si = new_strinfo (ptr, idx, build_int_cst (size_type_node, nonzero_chars),
770 basesi->full_string_p);
771 set_strinfo (idx, si);
772 if (strinfo *nextsi = get_strinfo (chainsi->next))
774 nextsi = unshare_strinfo (nextsi);
775 si->next = nextsi->idx;
776 nextsi->prev = idx;
778 chainsi = unshare_strinfo (chainsi);
779 if (chainsi->first == 0)
780 chainsi->first = chainsi->idx;
781 chainsi->next = idx;
782 if (chainsi->endptr == NULL_TREE && zero_length_string_p (si))
783 chainsi->endptr = ptr;
784 si->endptr = chainsi->endptr;
785 si->prev = chainsi->idx;
786 si->first = chainsi->first;
787 si->writable = chainsi->writable;
788 return si->idx;
791 /* Note that PTR, a pointer SSA_NAME initialized in the current stmt, points
792 to a zero-length string and if possible chain it to a related strinfo
793 chain whose part is or might be CHAINSI. */
795 static strinfo *
796 zero_length_string (tree ptr, strinfo *chainsi)
798 strinfo *si;
799 int idx;
800 if (ssa_ver_to_stridx.length () <= SSA_NAME_VERSION (ptr))
801 ssa_ver_to_stridx.safe_grow_cleared (num_ssa_names);
802 gcc_checking_assert (TREE_CODE (ptr) == SSA_NAME
803 && ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)] == 0);
805 if (SSA_NAME_OCCURS_IN_ABNORMAL_PHI (ptr))
806 return NULL;
807 if (chainsi != NULL)
809 si = verify_related_strinfos (chainsi);
810 if (si)
814 /* We shouldn't mix delayed and non-delayed lengths. */
815 gcc_assert (si->full_string_p);
816 if (si->endptr == NULL_TREE)
818 si = unshare_strinfo (si);
819 si->endptr = ptr;
821 chainsi = si;
822 si = get_next_strinfo (si);
824 while (si != NULL);
825 if (zero_length_string_p (chainsi))
827 if (chainsi->next)
829 chainsi = unshare_strinfo (chainsi);
830 chainsi->next = 0;
832 ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)] = chainsi->idx;
833 return chainsi;
836 else
838 /* We shouldn't mix delayed and non-delayed lengths. */
839 gcc_assert (chainsi->full_string_p);
840 if (chainsi->first || chainsi->prev || chainsi->next)
842 chainsi = unshare_strinfo (chainsi);
843 chainsi->first = 0;
844 chainsi->prev = 0;
845 chainsi->next = 0;
849 idx = new_stridx (ptr);
850 if (idx == 0)
851 return NULL;
852 si = new_strinfo (ptr, idx, build_int_cst (size_type_node, 0), true);
853 set_strinfo (idx, si);
854 si->endptr = ptr;
855 if (chainsi != NULL)
857 chainsi = unshare_strinfo (chainsi);
858 if (chainsi->first == 0)
859 chainsi->first = chainsi->idx;
860 chainsi->next = idx;
861 if (chainsi->endptr == NULL_TREE)
862 chainsi->endptr = ptr;
863 si->prev = chainsi->idx;
864 si->first = chainsi->first;
865 si->writable = chainsi->writable;
867 return si;
870 /* For strinfo ORIGSI whose length has been just updated, adjust other
871 related strinfos so that they match the new ORIGSI. This involves:
873 - adding ADJ to the nonzero_chars fields
874 - copying full_string_p from the new ORIGSI. */
876 static void
877 adjust_related_strinfos (location_t loc, strinfo *origsi, tree adj)
879 strinfo *si = verify_related_strinfos (origsi);
881 if (si == NULL)
882 return;
884 while (1)
886 strinfo *nsi;
888 if (si != origsi)
890 tree tem;
892 si = unshare_strinfo (si);
893 /* We shouldn't see delayed lengths here; the caller must have
894 calculated the old length in order to calculate the
895 adjustment. */
896 gcc_assert (si->nonzero_chars);
897 tem = fold_convert_loc (loc, TREE_TYPE (si->nonzero_chars), adj);
898 si->nonzero_chars = fold_build2_loc (loc, PLUS_EXPR,
899 TREE_TYPE (si->nonzero_chars),
900 si->nonzero_chars, tem);
901 si->full_string_p = origsi->full_string_p;
903 si->endptr = NULL_TREE;
904 si->dont_invalidate = true;
906 nsi = get_next_strinfo (si);
907 if (nsi == NULL)
908 return;
909 si = nsi;
913 /* Find if there are other SSA_NAME pointers equal to PTR
914 for which we don't track their string lengths yet. If so, use
915 IDX for them. */
917 static void
918 find_equal_ptrs (tree ptr, int idx)
920 if (TREE_CODE (ptr) != SSA_NAME)
921 return;
922 while (1)
924 gimple *stmt = SSA_NAME_DEF_STMT (ptr);
925 if (!is_gimple_assign (stmt))
926 return;
927 ptr = gimple_assign_rhs1 (stmt);
928 switch (gimple_assign_rhs_code (stmt))
930 case SSA_NAME:
931 break;
932 CASE_CONVERT:
933 if (!POINTER_TYPE_P (TREE_TYPE (ptr)))
934 return;
935 if (TREE_CODE (ptr) == SSA_NAME)
936 break;
937 if (TREE_CODE (ptr) != ADDR_EXPR)
938 return;
939 /* FALLTHRU */
940 case ADDR_EXPR:
942 int *pidx = addr_stridxptr (TREE_OPERAND (ptr, 0));
943 if (pidx != NULL && *pidx == 0)
944 *pidx = idx;
945 return;
947 default:
948 return;
951 /* We might find an endptr created in this pass. Grow the
952 vector in that case. */
953 if (ssa_ver_to_stridx.length () <= SSA_NAME_VERSION (ptr))
954 ssa_ver_to_stridx.safe_grow_cleared (num_ssa_names);
956 if (ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)] != 0)
957 return;
958 ssa_ver_to_stridx[SSA_NAME_VERSION (ptr)] = idx;
962 /* Return true if STMT is a call to a builtin function with the right
963 arguments and attributes that should be considered for optimization
964 by this pass. */
966 static bool
967 valid_builtin_call (gimple *stmt)
969 if (!gimple_call_builtin_p (stmt, BUILT_IN_NORMAL))
970 return false;
972 tree callee = gimple_call_fndecl (stmt);
973 switch (DECL_FUNCTION_CODE (callee))
975 case BUILT_IN_MEMCMP:
976 case BUILT_IN_MEMCMP_EQ:
977 case BUILT_IN_STRCHR:
978 case BUILT_IN_STRLEN:
979 /* The above functions should be pure. Punt if they aren't. */
980 if (gimple_vdef (stmt) || gimple_vuse (stmt) == NULL_TREE)
981 return false;
982 break;
984 case BUILT_IN_CALLOC:
985 case BUILT_IN_MALLOC:
986 case BUILT_IN_MEMCPY:
987 case BUILT_IN_MEMCPY_CHK:
988 case BUILT_IN_MEMPCPY:
989 case BUILT_IN_MEMPCPY_CHK:
990 case BUILT_IN_MEMSET:
991 case BUILT_IN_STPCPY:
992 case BUILT_IN_STPCPY_CHK:
993 case BUILT_IN_STRCAT:
994 case BUILT_IN_STRCAT_CHK:
995 case BUILT_IN_STRCPY:
996 case BUILT_IN_STRCPY_CHK:
997 /* The above functions should be neither const nor pure. Punt if they
998 aren't. */
999 if (gimple_vdef (stmt) == NULL_TREE || gimple_vuse (stmt) == NULL_TREE)
1000 return false;
1001 break;
1003 default:
1004 break;
1007 return true;
1010 /* If the last .MEM setter statement before STMT is
1011 memcpy (x, y, strlen (y) + 1), the only .MEM use of it is STMT
1012 and STMT is known to overwrite x[strlen (x)], adjust the last memcpy to
1013 just memcpy (x, y, strlen (y)). SI must be the zero length
1014 strinfo. */
1016 static void
1017 adjust_last_stmt (strinfo *si, gimple *stmt, bool is_strcat)
1019 tree vuse, callee, len;
1020 struct laststmt_struct last = laststmt;
1021 strinfo *lastsi, *firstsi;
1022 unsigned len_arg_no = 2;
1024 laststmt.stmt = NULL;
1025 laststmt.len = NULL_TREE;
1026 laststmt.stridx = 0;
1028 if (last.stmt == NULL)
1029 return;
1031 vuse = gimple_vuse (stmt);
1032 if (vuse == NULL_TREE
1033 || SSA_NAME_DEF_STMT (vuse) != last.stmt
1034 || !has_single_use (vuse))
1035 return;
1037 gcc_assert (last.stridx > 0);
1038 lastsi = get_strinfo (last.stridx);
1039 if (lastsi == NULL)
1040 return;
1042 if (lastsi != si)
1044 if (lastsi->first == 0 || lastsi->first != si->first)
1045 return;
1047 firstsi = verify_related_strinfos (si);
1048 if (firstsi == NULL)
1049 return;
1050 while (firstsi != lastsi)
1052 firstsi = get_next_strinfo (firstsi);
1053 if (firstsi == NULL)
1054 return;
1058 if (!is_strcat && !zero_length_string_p (si))
1059 return;
1061 if (is_gimple_assign (last.stmt))
1063 gimple_stmt_iterator gsi;
1065 if (!integer_zerop (gimple_assign_rhs1 (last.stmt)))
1066 return;
1067 if (stmt_could_throw_p (last.stmt))
1068 return;
1069 gsi = gsi_for_stmt (last.stmt);
1070 unlink_stmt_vdef (last.stmt);
1071 release_defs (last.stmt);
1072 gsi_remove (&gsi, true);
1073 return;
1076 if (!valid_builtin_call (last.stmt))
1077 return;
1079 callee = gimple_call_fndecl (last.stmt);
1080 switch (DECL_FUNCTION_CODE (callee))
1082 case BUILT_IN_MEMCPY:
1083 case BUILT_IN_MEMCPY_CHK:
1084 break;
1085 default:
1086 return;
1089 len = gimple_call_arg (last.stmt, len_arg_no);
1090 if (tree_fits_uhwi_p (len))
1092 if (!tree_fits_uhwi_p (last.len)
1093 || integer_zerop (len)
1094 || tree_to_uhwi (len) != tree_to_uhwi (last.len) + 1)
1095 return;
1096 /* Don't adjust the length if it is divisible by 4, it is more efficient
1097 to store the extra '\0' in that case. */
1098 if ((tree_to_uhwi (len) & 3) == 0)
1099 return;
1101 /* Don't fold away an out of bounds access, as this defeats proper
1102 warnings. */
1103 tree dst = gimple_call_arg (last.stmt, 0);
1104 tree size = compute_objsize (dst, 0);
1105 if (size && tree_int_cst_lt (size, len))
1106 return;
1108 else if (TREE_CODE (len) == SSA_NAME)
1110 gimple *def_stmt = SSA_NAME_DEF_STMT (len);
1111 if (!is_gimple_assign (def_stmt)
1112 || gimple_assign_rhs_code (def_stmt) != PLUS_EXPR
1113 || gimple_assign_rhs1 (def_stmt) != last.len
1114 || !integer_onep (gimple_assign_rhs2 (def_stmt)))
1115 return;
1117 else
1118 return;
1120 gimple_call_set_arg (last.stmt, len_arg_no, last.len);
1121 update_stmt (last.stmt);
1124 /* For an LHS that is an SSA_NAME and for strlen() or strnlen() argument
1125 SRC, set LHS range info to [0, min (N, BOUND)] if SRC refers to
1126 a character array A[N] with unknown length bounded by N, and for
1127 strnlen(), by min (N, BOUND). */
1129 static tree
1130 maybe_set_strlen_range (tree lhs, tree src, tree bound)
1132 if (TREE_CODE (lhs) != SSA_NAME
1133 || !INTEGRAL_TYPE_P (TREE_TYPE (lhs)))
1134 return NULL_TREE;
1136 if (TREE_CODE (src) == SSA_NAME)
1138 gimple *def = SSA_NAME_DEF_STMT (src);
1139 if (is_gimple_assign (def)
1140 && gimple_assign_rhs_code (def) == ADDR_EXPR)
1141 src = gimple_assign_rhs1 (def);
1144 wide_int max = wi::to_wide (TYPE_MAX_VALUE (ptrdiff_type_node));
1145 wide_int min = wi::zero (max.get_precision ());
1147 if (TREE_CODE (src) == ADDR_EXPR)
1149 /* The last array member of a struct can be bigger than its size
1150 suggests if it's treated as a poor-man's flexible array member. */
1151 src = TREE_OPERAND (src, 0);
1152 bool src_is_array = TREE_CODE (TREE_TYPE (src)) == ARRAY_TYPE;
1153 if (src_is_array
1154 && TREE_CODE (src) != MEM_REF
1155 && !array_at_struct_end_p (src))
1157 tree type = TREE_TYPE (src);
1158 if (tree size = TYPE_SIZE_UNIT (type))
1159 if (size && TREE_CODE (size) == INTEGER_CST)
1160 max = wi::to_wide (size);
1162 /* For strlen() the upper bound above is equal to
1163 the longest string that can be stored in the array
1164 (i.e., it accounts for the terminating nul. For
1165 strnlen() bump up the maximum by one since the array
1166 need not be nul-terminated. */
1167 if (!bound && max != 0)
1168 --max;
1170 else
1172 if (TREE_CODE (src) == COMPONENT_REF && !src_is_array)
1173 src = TREE_OPERAND (src, 1);
1174 if (DECL_P (src))
1176 /* Handle the unlikely case of strlen (&c) where c is some
1177 variable. */
1178 if (tree size = DECL_SIZE_UNIT (src))
1179 if (TREE_CODE (size) == INTEGER_CST)
1180 max = wi::to_wide (size);
1185 if (bound)
1187 /* For strnlen, adjust MIN and MAX as necessary. If the bound
1188 is less than the size of the array set MAX to it. It it's
1189 greater than MAX and MAX is non-zero bump MAX down to account
1190 for the necessary terminating nul. Otherwise leave it alone. */
1191 if (TREE_CODE (bound) == INTEGER_CST)
1193 wide_int wibnd = wi::to_wide (bound);
1194 int cmp = wi::cmpu (wibnd, max);
1195 if (cmp < 0)
1196 max = wibnd;
1197 else if (cmp && wi::ne_p (max, min))
1198 --max;
1200 else if (TREE_CODE (bound) == SSA_NAME)
1202 wide_int minbound, maxbound;
1203 value_range_type rng = get_range_info (bound, &minbound, &maxbound);
1204 if (rng == VR_RANGE)
1206 /* For a bound in a known range, adjust the range determined
1207 above as necessary. For a bound in some anti-range or
1208 in an unknown range, use the range determined above. */
1209 if (wi::ltu_p (minbound, min))
1210 min = minbound;
1211 if (wi::ltu_p (maxbound, max))
1212 max = maxbound;
1217 if (min == max)
1218 return wide_int_to_tree (size_type_node, min);
1220 set_range_info (lhs, VR_RANGE, min, max);
1221 return lhs;
1224 /* Handle a strlen call. If strlen of the argument is known, replace
1225 the strlen call with the known value, otherwise remember that strlen
1226 of the argument is stored in the lhs SSA_NAME. */
1228 static void
1229 handle_builtin_strlen (gimple_stmt_iterator *gsi)
1231 gimple *stmt = gsi_stmt (*gsi);
1232 tree lhs = gimple_call_lhs (stmt);
1234 if (lhs == NULL_TREE)
1235 return;
1237 location_t loc = gimple_location (stmt);
1238 tree callee = gimple_call_fndecl (stmt);
1239 tree src = gimple_call_arg (stmt, 0);
1240 tree bound = (DECL_FUNCTION_CODE (callee) == BUILT_IN_STRNLEN
1241 ? gimple_call_arg (stmt, 1) : NULL_TREE);
1242 int idx = get_stridx (src);
1243 if (idx)
1245 strinfo *si = NULL;
1246 tree rhs;
1248 if (idx < 0)
1249 rhs = build_int_cst (TREE_TYPE (lhs), ~idx);
1250 else
1252 rhs = NULL_TREE;
1253 si = get_strinfo (idx);
1254 if (si != NULL)
1255 rhs = get_string_length (si);
1257 if (rhs != NULL_TREE)
1259 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
1261 fprintf (dump_file, "Optimizing: ");
1262 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
1264 rhs = unshare_expr (rhs);
1265 if (!useless_type_conversion_p (TREE_TYPE (lhs), TREE_TYPE (rhs)))
1266 rhs = fold_convert_loc (loc, TREE_TYPE (lhs), rhs);
1268 /* Set for strnlen() calls with a non-constant bound. */
1269 bool noncst_bound = false;
1270 if (bound)
1272 tree new_rhs
1273 = fold_build2_loc (loc, MIN_EXPR, TREE_TYPE (rhs), rhs, bound);
1275 noncst_bound = (TREE_CODE (new_rhs) != INTEGER_CST
1276 || tree_int_cst_lt (new_rhs, rhs));
1278 rhs = new_rhs;
1281 if (!update_call_from_tree (gsi, rhs))
1282 gimplify_and_update_call_from_tree (gsi, rhs);
1283 stmt = gsi_stmt (*gsi);
1284 update_stmt (stmt);
1285 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
1287 fprintf (dump_file, "into: ");
1288 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
1291 /* Avoid storing the length for calls to strnlen() with
1292 a non-constant bound. */
1293 if (noncst_bound)
1294 return;
1296 if (si != NULL
1297 && TREE_CODE (si->nonzero_chars) != SSA_NAME
1298 && TREE_CODE (si->nonzero_chars) != INTEGER_CST
1299 && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (lhs))
1301 si = unshare_strinfo (si);
1302 si->nonzero_chars = lhs;
1303 gcc_assert (si->full_string_p);
1306 if (strlen_to_stridx)
1307 strlen_to_stridx->put (lhs, stridx_strlenloc (idx, loc));
1309 return;
1312 if (SSA_NAME_OCCURS_IN_ABNORMAL_PHI (lhs))
1313 return;
1315 if (idx == 0)
1316 idx = new_stridx (src);
1317 else
1319 strinfo *si = get_strinfo (idx);
1320 if (si != NULL)
1322 if (!si->full_string_p && !si->stmt)
1324 /* Until now we only had a lower bound on the string length.
1325 Install LHS as the actual length. */
1326 si = unshare_strinfo (si);
1327 tree old = si->nonzero_chars;
1328 si->nonzero_chars = lhs;
1329 si->full_string_p = true;
1330 if (TREE_CODE (old) == INTEGER_CST)
1332 old = fold_convert_loc (loc, TREE_TYPE (lhs), old);
1333 tree adj = fold_build2_loc (loc, MINUS_EXPR,
1334 TREE_TYPE (lhs), lhs, old);
1335 adjust_related_strinfos (loc, si, adj);
1337 else
1339 si->first = 0;
1340 si->prev = 0;
1341 si->next = 0;
1344 return;
1347 if (idx)
1349 if (!bound)
1351 /* Only store the new length information for calls to strlen(),
1352 not for those to strnlen(). */
1353 strinfo *si = new_strinfo (src, idx, lhs, true);
1354 set_strinfo (idx, si);
1355 find_equal_ptrs (src, idx);
1358 /* For SRC that is an array of N elements, set LHS's range
1359 to [0, min (N, BOUND)]. A constant return value means
1360 the range would have consisted of a single value. In
1361 that case, fold the result into the returned constant. */
1362 if (tree ret = maybe_set_strlen_range (lhs, src, bound))
1363 if (TREE_CODE (ret) == INTEGER_CST)
1365 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
1367 fprintf (dump_file, "Optimizing: ");
1368 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
1370 if (!useless_type_conversion_p (TREE_TYPE (lhs), TREE_TYPE (ret)))
1371 ret = fold_convert_loc (loc, TREE_TYPE (lhs), ret);
1372 if (!update_call_from_tree (gsi, ret))
1373 gimplify_and_update_call_from_tree (gsi, ret);
1374 stmt = gsi_stmt (*gsi);
1375 update_stmt (stmt);
1376 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
1378 fprintf (dump_file, "into: ");
1379 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
1383 if (strlen_to_stridx && !bound)
1384 strlen_to_stridx->put (lhs, stridx_strlenloc (idx, loc));
1388 /* Handle a strchr call. If strlen of the first argument is known, replace
1389 the strchr (x, 0) call with the endptr or x + strlen, otherwise remember
1390 that lhs of the call is endptr and strlen of the argument is endptr - x. */
1392 static void
1393 handle_builtin_strchr (gimple_stmt_iterator *gsi)
1395 int idx;
1396 tree src;
1397 gimple *stmt = gsi_stmt (*gsi);
1398 tree lhs = gimple_call_lhs (stmt);
1400 if (lhs == NULL_TREE)
1401 return;
1403 if (!integer_zerop (gimple_call_arg (stmt, 1)))
1404 return;
1406 src = gimple_call_arg (stmt, 0);
1407 idx = get_stridx (src);
1408 if (idx)
1410 strinfo *si = NULL;
1411 tree rhs;
1413 if (idx < 0)
1414 rhs = build_int_cst (size_type_node, ~idx);
1415 else
1417 rhs = NULL_TREE;
1418 si = get_strinfo (idx);
1419 if (si != NULL)
1420 rhs = get_string_length (si);
1422 if (rhs != NULL_TREE)
1424 location_t loc = gimple_location (stmt);
1426 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
1428 fprintf (dump_file, "Optimizing: ");
1429 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
1431 if (si != NULL && si->endptr != NULL_TREE)
1433 rhs = unshare_expr (si->endptr);
1434 if (!useless_type_conversion_p (TREE_TYPE (lhs),
1435 TREE_TYPE (rhs)))
1436 rhs = fold_convert_loc (loc, TREE_TYPE (lhs), rhs);
1438 else
1440 rhs = fold_convert_loc (loc, sizetype, unshare_expr (rhs));
1441 rhs = fold_build2_loc (loc, POINTER_PLUS_EXPR,
1442 TREE_TYPE (src), src, rhs);
1443 if (!useless_type_conversion_p (TREE_TYPE (lhs),
1444 TREE_TYPE (rhs)))
1445 rhs = fold_convert_loc (loc, TREE_TYPE (lhs), rhs);
1447 if (!update_call_from_tree (gsi, rhs))
1448 gimplify_and_update_call_from_tree (gsi, rhs);
1449 stmt = gsi_stmt (*gsi);
1450 update_stmt (stmt);
1451 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
1453 fprintf (dump_file, "into: ");
1454 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
1456 if (si != NULL
1457 && si->endptr == NULL_TREE
1458 && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (lhs))
1460 si = unshare_strinfo (si);
1461 si->endptr = lhs;
1463 zero_length_string (lhs, si);
1464 return;
1467 if (SSA_NAME_OCCURS_IN_ABNORMAL_PHI (lhs))
1468 return;
1469 if (TREE_CODE (src) != SSA_NAME || !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (src))
1471 if (idx == 0)
1472 idx = new_stridx (src);
1473 else if (get_strinfo (idx) != NULL)
1475 zero_length_string (lhs, NULL);
1476 return;
1478 if (idx)
1480 location_t loc = gimple_location (stmt);
1481 tree lhsu = fold_convert_loc (loc, size_type_node, lhs);
1482 tree srcu = fold_convert_loc (loc, size_type_node, src);
1483 tree length = fold_build2_loc (loc, MINUS_EXPR,
1484 size_type_node, lhsu, srcu);
1485 strinfo *si = new_strinfo (src, idx, length, true);
1486 si->endptr = lhs;
1487 set_strinfo (idx, si);
1488 find_equal_ptrs (src, idx);
1489 zero_length_string (lhs, si);
1492 else
1493 zero_length_string (lhs, NULL);
1496 /* Handle a strcpy-like ({st{r,p}cpy,__st{r,p}cpy_chk}) call.
1497 If strlen of the second argument is known, strlen of the first argument
1498 is the same after this call. Furthermore, attempt to convert it to
1499 memcpy. */
1501 static void
1502 handle_builtin_strcpy (enum built_in_function bcode, gimple_stmt_iterator *gsi)
1504 int idx, didx;
1505 tree src, dst, srclen, len, lhs, type, fn, oldlen;
1506 bool success;
1507 gimple *stmt = gsi_stmt (*gsi);
1508 strinfo *si, *dsi, *olddsi, *zsi;
1509 location_t loc;
1511 src = gimple_call_arg (stmt, 1);
1512 dst = gimple_call_arg (stmt, 0);
1513 lhs = gimple_call_lhs (stmt);
1514 idx = get_stridx (src);
1515 si = NULL;
1516 if (idx > 0)
1517 si = get_strinfo (idx);
1519 didx = get_stridx (dst);
1520 olddsi = NULL;
1521 oldlen = NULL_TREE;
1522 if (didx > 0)
1523 olddsi = get_strinfo (didx);
1524 else if (didx < 0)
1525 return;
1527 if (olddsi != NULL)
1528 adjust_last_stmt (olddsi, stmt, false);
1530 srclen = NULL_TREE;
1531 if (si != NULL)
1532 srclen = get_string_length (si);
1533 else if (idx < 0)
1534 srclen = build_int_cst (size_type_node, ~idx);
1536 loc = gimple_location (stmt);
1537 if (srclen == NULL_TREE)
1538 switch (bcode)
1540 case BUILT_IN_STRCPY:
1541 case BUILT_IN_STRCPY_CHK:
1542 if (lhs != NULL_TREE || !builtin_decl_implicit_p (BUILT_IN_STPCPY))
1543 return;
1544 break;
1545 case BUILT_IN_STPCPY:
1546 case BUILT_IN_STPCPY_CHK:
1547 if (lhs == NULL_TREE)
1548 return;
1549 else
1551 tree lhsuint = fold_convert_loc (loc, size_type_node, lhs);
1552 srclen = fold_convert_loc (loc, size_type_node, dst);
1553 srclen = fold_build2_loc (loc, MINUS_EXPR, size_type_node,
1554 lhsuint, srclen);
1556 break;
1557 default:
1558 gcc_unreachable ();
1561 if (didx == 0)
1563 didx = new_stridx (dst);
1564 if (didx == 0)
1565 return;
1567 if (olddsi != NULL)
1569 oldlen = olddsi->nonzero_chars;
1570 dsi = unshare_strinfo (olddsi);
1571 dsi->nonzero_chars = srclen;
1572 dsi->full_string_p = (srclen != NULL_TREE);
1573 /* Break the chain, so adjust_related_strinfo on later pointers in
1574 the chain won't adjust this one anymore. */
1575 dsi->next = 0;
1576 dsi->stmt = NULL;
1577 dsi->endptr = NULL_TREE;
1579 else
1581 dsi = new_strinfo (dst, didx, srclen, srclen != NULL_TREE);
1582 set_strinfo (didx, dsi);
1583 find_equal_ptrs (dst, didx);
1585 dsi->writable = true;
1586 dsi->dont_invalidate = true;
1588 if (dsi->nonzero_chars == NULL_TREE)
1590 strinfo *chainsi;
1592 /* If string length of src is unknown, use delayed length
1593 computation. If string lenth of dst will be needed, it
1594 can be computed by transforming this strcpy call into
1595 stpcpy and subtracting dst from the return value. */
1597 /* Look for earlier strings whose length could be determined if
1598 this strcpy is turned into an stpcpy. */
1600 if (dsi->prev != 0 && (chainsi = verify_related_strinfos (dsi)) != NULL)
1602 for (; chainsi && chainsi != dsi; chainsi = get_strinfo (chainsi->next))
1604 /* When setting a stmt for delayed length computation
1605 prevent all strinfos through dsi from being
1606 invalidated. */
1607 chainsi = unshare_strinfo (chainsi);
1608 chainsi->stmt = stmt;
1609 chainsi->nonzero_chars = NULL_TREE;
1610 chainsi->full_string_p = false;
1611 chainsi->endptr = NULL_TREE;
1612 chainsi->dont_invalidate = true;
1615 dsi->stmt = stmt;
1617 /* Try to detect overlap before returning. This catches cases
1618 like strcpy (d, d + n) where n is non-constant whose range
1619 is such that (n <= strlen (d) holds).
1621 OLDDSI->NONZERO_chars may have been reset by this point with
1622 oldlen holding it original value. */
1623 if (olddsi && oldlen)
1625 /* Add 1 for the terminating NUL. */
1626 tree type = TREE_TYPE (oldlen);
1627 oldlen = fold_build2 (PLUS_EXPR, type, oldlen,
1628 build_int_cst (type, 1));
1629 check_bounds_or_overlap (stmt, olddsi->ptr, src, oldlen, NULL_TREE);
1632 return;
1635 if (olddsi != NULL)
1637 tree adj = NULL_TREE;
1638 if (oldlen == NULL_TREE)
1640 else if (integer_zerop (oldlen))
1641 adj = srclen;
1642 else if (TREE_CODE (oldlen) == INTEGER_CST
1643 || TREE_CODE (srclen) == INTEGER_CST)
1644 adj = fold_build2_loc (loc, MINUS_EXPR,
1645 TREE_TYPE (srclen), srclen,
1646 fold_convert_loc (loc, TREE_TYPE (srclen),
1647 oldlen));
1648 if (adj != NULL_TREE)
1649 adjust_related_strinfos (loc, dsi, adj);
1650 else
1651 dsi->prev = 0;
1653 /* strcpy src may not overlap dst, so src doesn't need to be
1654 invalidated either. */
1655 if (si != NULL)
1656 si->dont_invalidate = true;
1658 fn = NULL_TREE;
1659 zsi = NULL;
1660 switch (bcode)
1662 case BUILT_IN_STRCPY:
1663 fn = builtin_decl_implicit (BUILT_IN_MEMCPY);
1664 if (lhs)
1665 ssa_ver_to_stridx[SSA_NAME_VERSION (lhs)] = didx;
1666 break;
1667 case BUILT_IN_STRCPY_CHK:
1668 fn = builtin_decl_explicit (BUILT_IN_MEMCPY_CHK);
1669 if (lhs)
1670 ssa_ver_to_stridx[SSA_NAME_VERSION (lhs)] = didx;
1671 break;
1672 case BUILT_IN_STPCPY:
1673 /* This would need adjustment of the lhs (subtract one),
1674 or detection that the trailing '\0' doesn't need to be
1675 written, if it will be immediately overwritten.
1676 fn = builtin_decl_explicit (BUILT_IN_MEMPCPY); */
1677 if (lhs)
1679 dsi->endptr = lhs;
1680 zsi = zero_length_string (lhs, dsi);
1682 break;
1683 case BUILT_IN_STPCPY_CHK:
1684 /* This would need adjustment of the lhs (subtract one),
1685 or detection that the trailing '\0' doesn't need to be
1686 written, if it will be immediately overwritten.
1687 fn = builtin_decl_explicit (BUILT_IN_MEMPCPY_CHK); */
1688 if (lhs)
1690 dsi->endptr = lhs;
1691 zsi = zero_length_string (lhs, dsi);
1693 break;
1694 default:
1695 gcc_unreachable ();
1697 if (zsi != NULL)
1698 zsi->dont_invalidate = true;
1700 if (fn)
1702 tree args = TYPE_ARG_TYPES (TREE_TYPE (fn));
1703 type = TREE_VALUE (TREE_CHAIN (TREE_CHAIN (args)));
1705 else
1706 type = size_type_node;
1708 len = fold_convert_loc (loc, type, unshare_expr (srclen));
1709 len = fold_build2_loc (loc, PLUS_EXPR, type, len, build_int_cst (type, 1));
1711 /* Set the no-warning bit on the transformed statement? */
1712 bool set_no_warning = false;
1714 if (const strinfo *chksi = olddsi ? olddsi : dsi)
1715 if (si
1716 && !check_bounds_or_overlap (stmt, chksi->ptr, si->ptr, NULL_TREE, len))
1718 gimple_set_no_warning (stmt, true);
1719 set_no_warning = true;
1722 if (fn == NULL_TREE)
1723 return;
1725 len = force_gimple_operand_gsi (gsi, len, true, NULL_TREE, true,
1726 GSI_SAME_STMT);
1727 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
1729 fprintf (dump_file, "Optimizing: ");
1730 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
1732 if (gimple_call_num_args (stmt) == 2)
1733 success = update_gimple_call (gsi, fn, 3, dst, src, len);
1734 else
1735 success = update_gimple_call (gsi, fn, 4, dst, src, len,
1736 gimple_call_arg (stmt, 2));
1737 if (success)
1739 stmt = gsi_stmt (*gsi);
1740 update_stmt (stmt);
1741 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
1743 fprintf (dump_file, "into: ");
1744 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
1746 /* Allow adjust_last_stmt to decrease this memcpy's size. */
1747 laststmt.stmt = stmt;
1748 laststmt.len = srclen;
1749 laststmt.stridx = dsi->idx;
1751 else if (dump_file && (dump_flags & TDF_DETAILS) != 0)
1752 fprintf (dump_file, "not possible.\n");
1754 if (set_no_warning)
1755 gimple_set_no_warning (stmt, true);
1758 /* Check the size argument to the built-in forms of stpncpy and strncpy
1759 for out-of-bounds offsets or overlapping access, and to see if the
1760 size argument is derived from a call to strlen() on the source argument,
1761 and if so, issue an appropriate warning. */
1763 static void
1764 handle_builtin_strncat (built_in_function bcode, gimple_stmt_iterator *gsi)
1766 /* Same as stxncpy(). */
1767 handle_builtin_stxncpy (bcode, gsi);
1770 /* Return true if LEN depends on a call to strlen(SRC) in an interesting
1771 way. LEN can either be an integer expression, or a pointer (to char).
1772 When it is the latter (such as in recursive calls to self) is is
1773 assumed to be the argument in some call to strlen() whose relationship
1774 to SRC is being ascertained. */
1776 bool
1777 is_strlen_related_p (tree src, tree len)
1779 if (TREE_CODE (TREE_TYPE (len)) == POINTER_TYPE
1780 && operand_equal_p (src, len, 0))
1781 return true;
1783 if (TREE_CODE (len) != SSA_NAME)
1784 return false;
1786 gimple *def_stmt = SSA_NAME_DEF_STMT (len);
1787 if (!def_stmt)
1788 return false;
1790 if (is_gimple_call (def_stmt))
1792 tree func = gimple_call_fndecl (def_stmt);
1793 if (!valid_builtin_call (def_stmt)
1794 || DECL_FUNCTION_CODE (func) != BUILT_IN_STRLEN)
1795 return false;
1797 tree arg = gimple_call_arg (def_stmt, 0);
1798 return is_strlen_related_p (src, arg);
1801 if (!is_gimple_assign (def_stmt))
1802 return false;
1804 tree_code code = gimple_assign_rhs_code (def_stmt);
1805 tree rhs1 = gimple_assign_rhs1 (def_stmt);
1806 tree rhstype = TREE_TYPE (rhs1);
1808 if ((TREE_CODE (rhstype) == POINTER_TYPE && code == POINTER_PLUS_EXPR)
1809 || (INTEGRAL_TYPE_P (rhstype)
1810 && (code == BIT_AND_EXPR
1811 || code == NOP_EXPR)))
1813 /* Pointer plus (an integer), and truncation are considered among
1814 the (potentially) related expressions to strlen. */
1815 return is_strlen_related_p (src, rhs1);
1818 if (tree rhs2 = gimple_assign_rhs2 (def_stmt))
1820 /* Integer subtraction is considered strlen-related when both
1821 arguments are integers and second one is strlen-related. */
1822 rhstype = TREE_TYPE (rhs2);
1823 if (INTEGRAL_TYPE_P (rhstype) && code == MINUS_EXPR)
1824 return is_strlen_related_p (src, rhs2);
1827 return false;
1830 /* Called by handle_builtin_stxncpy and by gimple_fold_builtin_strncpy
1831 in gimple-fold.c.
1832 Check to see if the specified bound is a) equal to the size of
1833 the destination DST and if so, b) if it's immediately followed by
1834 DST[CNT - 1] = '\0'. If a) holds and b) does not, warn. Otherwise,
1835 do nothing. Return true if diagnostic has been issued.
1837 The purpose is to diagnose calls to strncpy and stpncpy that do
1838 not nul-terminate the copy while allowing for the idiom where
1839 such a call is immediately followed by setting the last element
1840 to nul, as in:
1841 char a[32];
1842 strncpy (a, s, sizeof a);
1843 a[sizeof a - 1] = '\0';
1846 bool
1847 maybe_diag_stxncpy_trunc (gimple_stmt_iterator gsi, tree src, tree cnt)
1849 gimple *stmt = gsi_stmt (gsi);
1850 if (gimple_no_warning_p (stmt))
1851 return false;
1853 wide_int cntrange[2];
1855 if (TREE_CODE (cnt) == INTEGER_CST)
1856 cntrange[0] = cntrange[1] = wi::to_wide (cnt);
1857 else if (TREE_CODE (cnt) == SSA_NAME)
1859 enum value_range_type rng = get_range_info (cnt, cntrange, cntrange + 1);
1860 if (rng == VR_RANGE)
1862 else if (rng == VR_ANTI_RANGE)
1864 wide_int maxobjsize = wi::to_wide (TYPE_MAX_VALUE (ptrdiff_type_node));
1866 if (wi::ltu_p (cntrange[1], maxobjsize))
1868 cntrange[0] = cntrange[1] + 1;
1869 cntrange[1] = maxobjsize;
1871 else
1873 cntrange[1] = cntrange[0] - 1;
1874 cntrange[0] = wi::zero (TYPE_PRECISION (TREE_TYPE (cnt)));
1877 else
1878 return false;
1880 else
1881 return false;
1883 /* Negative value is the constant string length. If it's less than
1884 the lower bound there is no truncation. Avoid calling get_stridx()
1885 when ssa_ver_to_stridx is empty. That implies the caller isn't
1886 running under the control of this pass and ssa_ver_to_stridx hasn't
1887 been created yet. */
1888 int sidx = ssa_ver_to_stridx.length () ? get_stridx (src) : 0;
1889 if (sidx < 0 && wi::gtu_p (cntrange[0], ~sidx))
1890 return false;
1892 tree dst = gimple_call_arg (stmt, 0);
1893 tree dstdecl = dst;
1894 if (TREE_CODE (dstdecl) == ADDR_EXPR)
1895 dstdecl = TREE_OPERAND (dstdecl, 0);
1897 tree ref = NULL_TREE;
1899 if (!sidx)
1901 /* If the source is a non-string return early to avoid warning
1902 for possible truncation (if the truncation is certain SIDX
1903 is non-zero). */
1904 tree srcdecl = gimple_call_arg (stmt, 1);
1905 if (TREE_CODE (srcdecl) == ADDR_EXPR)
1906 srcdecl = TREE_OPERAND (srcdecl, 0);
1907 if (get_attr_nonstring_decl (srcdecl, &ref))
1908 return false;
1911 /* Likewise, if the destination refers to a an array/pointer declared
1912 nonstring return early. */
1913 if (get_attr_nonstring_decl (dstdecl, &ref))
1914 return false;
1916 /* Look for dst[i] = '\0'; after the stxncpy() call and if found
1917 avoid the truncation warning. */
1918 gsi_next_nondebug (&gsi);
1919 gimple *next_stmt = gsi_stmt (gsi);
1920 if (!next_stmt)
1922 /* When there is no statement in the same basic block check
1923 the immediate successor block. */
1924 if (basic_block bb = gimple_bb (stmt))
1926 if (single_succ_p (bb))
1928 /* For simplicity, ignore blocks with multiple outgoing
1929 edges for now and only consider successor blocks along
1930 normal edges. */
1931 edge e = EDGE_SUCC (bb, 0);
1932 if (!(e->flags & EDGE_ABNORMAL))
1934 gsi = gsi_start_bb (e->dest);
1935 next_stmt = gsi_stmt (gsi);
1936 if (next_stmt && is_gimple_debug (next_stmt))
1938 gsi_next_nondebug (&gsi);
1939 next_stmt = gsi_stmt (gsi);
1946 if (next_stmt && is_gimple_assign (next_stmt))
1948 tree lhs = gimple_assign_lhs (next_stmt);
1949 tree_code code = TREE_CODE (lhs);
1950 if (code == ARRAY_REF || code == MEM_REF)
1951 lhs = TREE_OPERAND (lhs, 0);
1953 tree func = gimple_call_fndecl (stmt);
1954 if (DECL_FUNCTION_CODE (func) == BUILT_IN_STPNCPY)
1956 tree ret = gimple_call_lhs (stmt);
1957 if (ret && operand_equal_p (ret, lhs, 0))
1958 return false;
1961 /* Determine the base address and offset of the reference,
1962 ignoring the innermost array index. */
1963 if (TREE_CODE (ref) == ARRAY_REF)
1964 ref = TREE_OPERAND (ref, 0);
1966 poly_int64 dstoff;
1967 tree dstbase = get_addr_base_and_unit_offset (ref, &dstoff);
1969 poly_int64 lhsoff;
1970 tree lhsbase = get_addr_base_and_unit_offset (lhs, &lhsoff);
1971 if (lhsbase
1972 && dstbase
1973 && known_eq (dstoff, lhsoff)
1974 && operand_equal_p (dstbase, lhsbase, 0))
1975 return false;
1978 int prec = TYPE_PRECISION (TREE_TYPE (cnt));
1979 wide_int lenrange[2];
1980 if (strinfo *sisrc = sidx > 0 ? get_strinfo (sidx) : NULL)
1982 lenrange[0] = (sisrc->nonzero_chars
1983 && TREE_CODE (sisrc->nonzero_chars) == INTEGER_CST
1984 ? wi::to_wide (sisrc->nonzero_chars)
1985 : wi::zero (prec));
1986 lenrange[1] = lenrange[0];
1988 else if (sidx < 0)
1989 lenrange[0] = lenrange[1] = wi::shwi (~sidx, prec);
1990 else
1992 tree range[2];
1993 get_range_strlen (src, range);
1994 if (range[0] != NULL_TREE
1995 && TREE_CODE (range[0]) == INTEGER_CST
1996 && range[1] != NULL_TREE
1997 && TREE_CODE (range[1]) == INTEGER_CST)
1999 lenrange[0] = wi::to_wide (range[0], prec);
2000 lenrange[1] = wi::to_wide (range[1], prec);
2002 else
2004 lenrange[0] = wi::shwi (0, prec);
2005 lenrange[1] = wi::shwi (-1, prec);
2009 location_t callloc = gimple_nonartificial_location (stmt);
2010 callloc = expansion_point_location_if_in_system_header (callloc);
2012 tree func = gimple_call_fndecl (stmt);
2014 if (lenrange[0] != 0 || !wi::neg_p (lenrange[1]))
2016 /* If the longest source string is shorter than the lower bound
2017 of the specified count the copy is definitely nul-terminated. */
2018 if (wi::ltu_p (lenrange[1], cntrange[0]))
2019 return false;
2021 if (wi::neg_p (lenrange[1]))
2023 /* The length of one of the strings is unknown but at least
2024 one has non-zero length and that length is stored in
2025 LENRANGE[1]. Swap the bounds to force a "may be truncated"
2026 warning below. */
2027 lenrange[1] = lenrange[0];
2028 lenrange[0] = wi::shwi (0, prec);
2031 /* Set to true for strncat whose bound is derived from the length
2032 of the destination (the expected usage pattern). */
2033 bool cat_dstlen_bounded = false;
2034 if (DECL_FUNCTION_CODE (func) == BUILT_IN_STRNCAT)
2035 cat_dstlen_bounded = is_strlen_related_p (dst, cnt);
2037 if (lenrange[0] == cntrange[1] && cntrange[0] == cntrange[1])
2038 return warning_n (callloc, OPT_Wstringop_truncation,
2039 cntrange[0].to_uhwi (),
2040 "%G%qD output truncated before terminating "
2041 "nul copying %E byte from a string of the "
2042 "same length",
2043 "%G%qD output truncated before terminating nul "
2044 "copying %E bytes from a string of the same "
2045 "length",
2046 stmt, func, cnt);
2047 else if (!cat_dstlen_bounded)
2049 if (wi::geu_p (lenrange[0], cntrange[1]))
2051 /* The shortest string is longer than the upper bound of
2052 the count so the truncation is certain. */
2053 if (cntrange[0] == cntrange[1])
2054 return warning_n (callloc, OPT_Wstringop_truncation,
2055 cntrange[0].to_uhwi (),
2056 "%G%qD output truncated copying %E byte "
2057 "from a string of length %wu",
2058 "%G%qD output truncated copying %E bytes "
2059 "from a string of length %wu",
2060 stmt, func, cnt, lenrange[0].to_uhwi ());
2062 return warning_at (callloc, OPT_Wstringop_truncation,
2063 "%G%qD output truncated copying between %wu "
2064 "and %wu bytes from a string of length %wu",
2065 stmt, func, cntrange[0].to_uhwi (),
2066 cntrange[1].to_uhwi (), lenrange[0].to_uhwi ());
2068 else if (wi::geu_p (lenrange[1], cntrange[1]))
2070 /* The longest string is longer than the upper bound of
2071 the count so the truncation is possible. */
2072 if (cntrange[0] == cntrange[1])
2073 return warning_n (callloc, OPT_Wstringop_truncation,
2074 cntrange[0].to_uhwi (),
2075 "%G%qD output may be truncated copying %E "
2076 "byte from a string of length %wu",
2077 "%G%qD output may be truncated copying %E "
2078 "bytes from a string of length %wu",
2079 stmt, func, cnt, lenrange[1].to_uhwi ());
2081 return warning_at (callloc, OPT_Wstringop_truncation,
2082 "%G%qD output may be truncated copying between "
2083 "%wu and %wu bytes from a string of length %wu",
2084 stmt, func, cntrange[0].to_uhwi (),
2085 cntrange[1].to_uhwi (), lenrange[1].to_uhwi ());
2089 if (!cat_dstlen_bounded
2090 && cntrange[0] != cntrange[1]
2091 && wi::leu_p (cntrange[0], lenrange[0])
2092 && wi::leu_p (cntrange[1], lenrange[0] + 1))
2094 /* If the source (including the terminating nul) is longer than
2095 the lower bound of the specified count but shorter than the
2096 upper bound the copy may (but need not) be truncated. */
2097 return warning_at (callloc, OPT_Wstringop_truncation,
2098 "%G%qD output may be truncated copying between "
2099 "%wu and %wu bytes from a string of length %wu",
2100 stmt, func, cntrange[0].to_uhwi (),
2101 cntrange[1].to_uhwi (), lenrange[0].to_uhwi ());
2105 if (tree dstsize = compute_objsize (dst, 1))
2107 /* The source length is uknown. Try to determine the destination
2108 size and see if it matches the specified bound. If not, bail.
2109 Otherwise go on to see if it should be diagnosed for possible
2110 truncation. */
2111 if (!dstsize)
2112 return false;
2114 if (wi::to_wide (dstsize) != cntrange[1])
2115 return false;
2117 if (cntrange[0] == cntrange[1])
2118 return warning_at (callloc, OPT_Wstringop_truncation,
2119 "%G%qD specified bound %E equals destination size",
2120 stmt, func, cnt);
2123 return false;
2126 /* Check the arguments to the built-in forms of stpncpy and strncpy for
2127 out-of-bounds offsets or overlapping access, and to see if the size
2128 is derived from calling strlen() on the source argument, and if so,
2129 issue the appropriate warning. */
2131 static void
2132 handle_builtin_stxncpy (built_in_function, gimple_stmt_iterator *gsi)
2134 if (!strlen_to_stridx)
2135 return;
2137 gimple *stmt = gsi_stmt (*gsi);
2139 tree dst = gimple_call_arg (stmt, 0);
2140 tree src = gimple_call_arg (stmt, 1);
2141 tree len = gimple_call_arg (stmt, 2);
2142 tree dstsize = NULL_TREE, srcsize = NULL_TREE;
2144 int didx = get_stridx (dst);
2145 if (strinfo *sidst = didx > 0 ? get_strinfo (didx) : NULL)
2147 /* Compute the size of the destination string including the NUL. */
2148 if (sidst->nonzero_chars)
2150 tree type = TREE_TYPE (sidst->nonzero_chars);
2151 dstsize = fold_build2 (PLUS_EXPR, type, sidst->nonzero_chars,
2152 build_int_cst (type, 1));
2154 dst = sidst->ptr;
2157 int sidx = get_stridx (src);
2158 strinfo *sisrc = sidx > 0 ? get_strinfo (sidx) : NULL;
2159 if (sisrc)
2161 /* strncat() and strncpy() can modify the source string by writing
2162 over the terminating nul so SISRC->DONT_INVALIDATE must be left
2163 clear. */
2165 /* Compute the size of the source string including the NUL. */
2166 if (sisrc->nonzero_chars)
2168 tree type = TREE_TYPE (sisrc->nonzero_chars);
2169 srcsize = fold_build2 (PLUS_EXPR, type, sisrc->nonzero_chars,
2170 build_int_cst (type, 1));
2173 src = sisrc->ptr;
2175 else
2176 srcsize = NULL_TREE;
2178 if (!check_bounds_or_overlap (stmt, dst, src, dstsize, srcsize))
2180 gimple_set_no_warning (stmt, true);
2181 return;
2184 /* If the length argument was computed from strlen(S) for some string
2185 S retrieve the strinfo index for the string (PSS->FIRST) alonng with
2186 the location of the strlen() call (PSS->SECOND). */
2187 stridx_strlenloc *pss = strlen_to_stridx->get (len);
2188 if (!pss || pss->first <= 0)
2190 if (maybe_diag_stxncpy_trunc (*gsi, src, len))
2191 gimple_set_no_warning (stmt, true);
2193 return;
2196 /* Retrieve the strinfo data for the string S that LEN was computed
2197 from as some function F of strlen (S) (i.e., LEN need not be equal
2198 to strlen(S)). */
2199 strinfo *silen = get_strinfo (pss->first);
2201 location_t callloc = gimple_nonartificial_location (stmt);
2202 callloc = expansion_point_location_if_in_system_header (callloc);
2204 tree func = gimple_call_fndecl (stmt);
2206 bool warned = false;
2208 /* When -Wstringop-truncation is set, try to determine truncation
2209 before diagnosing possible overflow. Truncation is implied by
2210 the LEN argument being equal to strlen(SRC), regardless of
2211 whether its value is known. Otherwise, issue the more generic
2212 -Wstringop-overflow which triggers for LEN arguments that in
2213 any meaningful way depend on strlen(SRC). */
2214 if (sisrc == silen
2215 && is_strlen_related_p (src, len)
2216 && warning_at (callloc, OPT_Wstringop_truncation,
2217 "%G%qD output truncated before terminating nul "
2218 "copying as many bytes from a string as its length",
2219 stmt, func))
2220 warned = true;
2221 else if (silen && is_strlen_related_p (src, silen->ptr))
2222 warned = warning_at (callloc, OPT_Wstringop_overflow_,
2223 "%G%qD specified bound depends on the length "
2224 "of the source argument",
2225 stmt, func);
2226 if (warned)
2228 location_t strlenloc = pss->second;
2229 if (strlenloc != UNKNOWN_LOCATION && strlenloc != callloc)
2230 inform (strlenloc, "length computed here");
2234 /* Handle a memcpy-like ({mem{,p}cpy,__mem{,p}cpy_chk}) call.
2235 If strlen of the second argument is known and length of the third argument
2236 is that plus one, strlen of the first argument is the same after this
2237 call. */
2239 static void
2240 handle_builtin_memcpy (enum built_in_function bcode, gimple_stmt_iterator *gsi)
2242 int idx, didx;
2243 tree src, dst, len, lhs, oldlen, newlen;
2244 gimple *stmt = gsi_stmt (*gsi);
2245 strinfo *si, *dsi, *olddsi;
2247 len = gimple_call_arg (stmt, 2);
2248 src = gimple_call_arg (stmt, 1);
2249 dst = gimple_call_arg (stmt, 0);
2250 idx = get_stridx (src);
2251 if (idx == 0)
2252 return;
2254 didx = get_stridx (dst);
2255 olddsi = NULL;
2256 if (didx > 0)
2257 olddsi = get_strinfo (didx);
2258 else if (didx < 0)
2259 return;
2261 if (olddsi != NULL
2262 && tree_fits_uhwi_p (len)
2263 && !integer_zerop (len))
2264 adjust_last_stmt (olddsi, stmt, false);
2266 bool full_string_p;
2267 if (idx > 0)
2269 gimple *def_stmt;
2271 /* Handle memcpy (x, y, l) where l's relationship with strlen (y)
2272 is known. */
2273 si = get_strinfo (idx);
2274 if (si == NULL || si->nonzero_chars == NULL_TREE)
2275 return;
2276 if (TREE_CODE (len) == INTEGER_CST
2277 && TREE_CODE (si->nonzero_chars) == INTEGER_CST)
2279 if (tree_int_cst_le (len, si->nonzero_chars))
2281 /* Copying LEN nonzero characters, where LEN is constant. */
2282 newlen = len;
2283 full_string_p = false;
2285 else
2287 /* Copying the whole of the analyzed part of SI. */
2288 newlen = si->nonzero_chars;
2289 full_string_p = si->full_string_p;
2292 else
2294 if (!si->full_string_p)
2295 return;
2296 if (TREE_CODE (len) != SSA_NAME)
2297 return;
2298 def_stmt = SSA_NAME_DEF_STMT (len);
2299 if (!is_gimple_assign (def_stmt)
2300 || gimple_assign_rhs_code (def_stmt) != PLUS_EXPR
2301 || gimple_assign_rhs1 (def_stmt) != si->nonzero_chars
2302 || !integer_onep (gimple_assign_rhs2 (def_stmt)))
2303 return;
2304 /* Copying variable-length string SI (and no more). */
2305 newlen = si->nonzero_chars;
2306 full_string_p = true;
2309 else
2311 si = NULL;
2312 /* Handle memcpy (x, "abcd", 5) or
2313 memcpy (x, "abc\0uvw", 7). */
2314 if (!tree_fits_uhwi_p (len))
2315 return;
2317 unsigned HOST_WIDE_INT clen = tree_to_uhwi (len);
2318 unsigned HOST_WIDE_INT nonzero_chars = ~idx;
2319 newlen = build_int_cst (size_type_node, MIN (nonzero_chars, clen));
2320 full_string_p = clen > nonzero_chars;
2323 if (!full_string_p
2324 && olddsi
2325 && olddsi->nonzero_chars
2326 && TREE_CODE (olddsi->nonzero_chars) == INTEGER_CST
2327 && tree_int_cst_le (newlen, olddsi->nonzero_chars))
2329 /* The SRC substring being written strictly overlaps
2330 a subsequence of the existing string OLDDSI. */
2331 newlen = olddsi->nonzero_chars;
2332 full_string_p = olddsi->full_string_p;
2335 if (olddsi != NULL && TREE_CODE (len) == SSA_NAME)
2336 adjust_last_stmt (olddsi, stmt, false);
2338 if (didx == 0)
2340 didx = new_stridx (dst);
2341 if (didx == 0)
2342 return;
2344 oldlen = NULL_TREE;
2345 if (olddsi != NULL)
2347 dsi = unshare_strinfo (olddsi);
2348 oldlen = olddsi->nonzero_chars;
2349 dsi->nonzero_chars = newlen;
2350 dsi->full_string_p = full_string_p;
2351 /* Break the chain, so adjust_related_strinfo on later pointers in
2352 the chain won't adjust this one anymore. */
2353 dsi->next = 0;
2354 dsi->stmt = NULL;
2355 dsi->endptr = NULL_TREE;
2357 else
2359 dsi = new_strinfo (dst, didx, newlen, full_string_p);
2360 set_strinfo (didx, dsi);
2361 find_equal_ptrs (dst, didx);
2363 dsi->writable = true;
2364 dsi->dont_invalidate = true;
2365 if (olddsi != NULL)
2367 tree adj = NULL_TREE;
2368 location_t loc = gimple_location (stmt);
2369 if (oldlen == NULL_TREE)
2371 else if (integer_zerop (oldlen))
2372 adj = newlen;
2373 else if (TREE_CODE (oldlen) == INTEGER_CST
2374 || TREE_CODE (newlen) == INTEGER_CST)
2375 adj = fold_build2_loc (loc, MINUS_EXPR, TREE_TYPE (newlen), newlen,
2376 fold_convert_loc (loc, TREE_TYPE (newlen),
2377 oldlen));
2378 if (adj != NULL_TREE)
2379 adjust_related_strinfos (loc, dsi, adj);
2380 else
2381 dsi->prev = 0;
2383 /* memcpy src may not overlap dst, so src doesn't need to be
2384 invalidated either. */
2385 if (si != NULL)
2386 si->dont_invalidate = true;
2388 if (full_string_p)
2390 lhs = gimple_call_lhs (stmt);
2391 switch (bcode)
2393 case BUILT_IN_MEMCPY:
2394 case BUILT_IN_MEMCPY_CHK:
2395 /* Allow adjust_last_stmt to decrease this memcpy's size. */
2396 laststmt.stmt = stmt;
2397 laststmt.len = dsi->nonzero_chars;
2398 laststmt.stridx = dsi->idx;
2399 if (lhs)
2400 ssa_ver_to_stridx[SSA_NAME_VERSION (lhs)] = didx;
2401 break;
2402 case BUILT_IN_MEMPCPY:
2403 case BUILT_IN_MEMPCPY_CHK:
2404 break;
2405 default:
2406 gcc_unreachable ();
2411 /* Handle a strcat-like ({strcat,__strcat_chk}) call.
2412 If strlen of the second argument is known, strlen of the first argument
2413 is increased by the length of the second argument. Furthermore, attempt
2414 to convert it to memcpy/strcpy if the length of the first argument
2415 is known. */
2417 static void
2418 handle_builtin_strcat (enum built_in_function bcode, gimple_stmt_iterator *gsi)
2420 int idx, didx;
2421 tree srclen, args, type, fn, objsz, endptr;
2422 bool success;
2423 gimple *stmt = gsi_stmt (*gsi);
2424 strinfo *si, *dsi;
2425 location_t loc = gimple_location (stmt);
2427 tree src = gimple_call_arg (stmt, 1);
2428 tree dst = gimple_call_arg (stmt, 0);
2430 /* Bail if the source is the same as destination. It will be diagnosed
2431 elsewhere. */
2432 if (operand_equal_p (src, dst, 0))
2433 return;
2435 tree lhs = gimple_call_lhs (stmt);
2437 didx = get_stridx (dst);
2438 if (didx < 0)
2439 return;
2441 dsi = NULL;
2442 if (didx > 0)
2443 dsi = get_strinfo (didx);
2445 srclen = NULL_TREE;
2446 si = NULL;
2447 idx = get_stridx (src);
2448 if (idx < 0)
2449 srclen = build_int_cst (size_type_node, ~idx);
2450 else if (idx > 0)
2452 si = get_strinfo (idx);
2453 if (si != NULL)
2454 srclen = get_string_length (si);
2457 /* Set the no-warning bit on the transformed statement? */
2458 bool set_no_warning = false;
2460 if (dsi == NULL || get_string_length (dsi) == NULL_TREE)
2463 /* The concatenation always involves copying at least one byte
2464 (the terminating nul), even if the source string is empty.
2465 If the source is unknown assume it's one character long and
2466 used that as both sizes. */
2467 tree slen = srclen;
2468 if (slen)
2470 tree type = TREE_TYPE (slen);
2471 slen = fold_build2 (PLUS_EXPR, type, slen, build_int_cst (type, 1));
2474 tree sptr = si && si->ptr ? si->ptr : src;
2476 if (!check_bounds_or_overlap (stmt, dst, sptr, NULL_TREE, slen))
2478 gimple_set_no_warning (stmt, true);
2479 set_no_warning = true;
2483 /* strcat (p, q) can be transformed into
2484 tmp = p + strlen (p); endptr = stpcpy (tmp, q);
2485 with length endptr - p if we need to compute the length
2486 later on. Don't do this transformation if we don't need
2487 it. */
2488 if (builtin_decl_implicit_p (BUILT_IN_STPCPY) && lhs == NULL_TREE)
2490 if (didx == 0)
2492 didx = new_stridx (dst);
2493 if (didx == 0)
2494 return;
2496 if (dsi == NULL)
2498 dsi = new_strinfo (dst, didx, NULL_TREE, false);
2499 set_strinfo (didx, dsi);
2500 find_equal_ptrs (dst, didx);
2502 else
2504 dsi = unshare_strinfo (dsi);
2505 dsi->nonzero_chars = NULL_TREE;
2506 dsi->full_string_p = false;
2507 dsi->next = 0;
2508 dsi->endptr = NULL_TREE;
2510 dsi->writable = true;
2511 dsi->stmt = stmt;
2512 dsi->dont_invalidate = true;
2514 return;
2517 tree dstlen = dsi->nonzero_chars;
2518 endptr = dsi->endptr;
2520 dsi = unshare_strinfo (dsi);
2521 dsi->endptr = NULL_TREE;
2522 dsi->stmt = NULL;
2523 dsi->writable = true;
2525 if (srclen != NULL_TREE)
2527 dsi->nonzero_chars = fold_build2_loc (loc, PLUS_EXPR,
2528 TREE_TYPE (dsi->nonzero_chars),
2529 dsi->nonzero_chars, srclen);
2530 gcc_assert (dsi->full_string_p);
2531 adjust_related_strinfos (loc, dsi, srclen);
2532 dsi->dont_invalidate = true;
2534 else
2536 dsi->nonzero_chars = NULL;
2537 dsi->full_string_p = false;
2538 if (lhs == NULL_TREE && builtin_decl_implicit_p (BUILT_IN_STPCPY))
2539 dsi->dont_invalidate = true;
2542 if (si != NULL)
2543 /* strcat src may not overlap dst, so src doesn't need to be
2544 invalidated either. */
2545 si->dont_invalidate = true;
2547 /* For now. Could remove the lhs from the call and add
2548 lhs = dst; afterwards. */
2549 if (lhs)
2550 return;
2552 fn = NULL_TREE;
2553 objsz = NULL_TREE;
2554 switch (bcode)
2556 case BUILT_IN_STRCAT:
2557 if (srclen != NULL_TREE)
2558 fn = builtin_decl_implicit (BUILT_IN_MEMCPY);
2559 else
2560 fn = builtin_decl_implicit (BUILT_IN_STRCPY);
2561 break;
2562 case BUILT_IN_STRCAT_CHK:
2563 if (srclen != NULL_TREE)
2564 fn = builtin_decl_explicit (BUILT_IN_MEMCPY_CHK);
2565 else
2566 fn = builtin_decl_explicit (BUILT_IN_STRCPY_CHK);
2567 objsz = gimple_call_arg (stmt, 2);
2568 break;
2569 default:
2570 gcc_unreachable ();
2573 if (fn == NULL_TREE)
2574 return;
2576 if (dsi && dstlen)
2578 tree type = TREE_TYPE (dstlen);
2580 /* Compute the size of the source sequence, including the nul. */
2581 tree srcsize = srclen ? srclen : size_zero_node;
2582 srcsize = fold_build2 (PLUS_EXPR, type, srcsize, build_int_cst (type, 1));
2584 tree sptr = si && si->ptr ? si->ptr : src;
2586 if (!check_bounds_or_overlap (stmt, dst, sptr, dstlen, srcsize))
2588 gimple_set_no_warning (stmt, true);
2589 set_no_warning = true;
2593 tree len = NULL_TREE;
2594 if (srclen != NULL_TREE)
2596 args = TYPE_ARG_TYPES (TREE_TYPE (fn));
2597 type = TREE_VALUE (TREE_CHAIN (TREE_CHAIN (args)));
2599 len = fold_convert_loc (loc, type, unshare_expr (srclen));
2600 len = fold_build2_loc (loc, PLUS_EXPR, type, len,
2601 build_int_cst (type, 1));
2602 len = force_gimple_operand_gsi (gsi, len, true, NULL_TREE, true,
2603 GSI_SAME_STMT);
2605 if (endptr)
2606 dst = fold_convert_loc (loc, TREE_TYPE (dst), unshare_expr (endptr));
2607 else
2608 dst = fold_build2_loc (loc, POINTER_PLUS_EXPR,
2609 TREE_TYPE (dst), unshare_expr (dst),
2610 fold_convert_loc (loc, sizetype,
2611 unshare_expr (dstlen)));
2612 dst = force_gimple_operand_gsi (gsi, dst, true, NULL_TREE, true,
2613 GSI_SAME_STMT);
2614 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2616 fprintf (dump_file, "Optimizing: ");
2617 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
2619 if (srclen != NULL_TREE)
2620 success = update_gimple_call (gsi, fn, 3 + (objsz != NULL_TREE),
2621 dst, src, len, objsz);
2622 else
2623 success = update_gimple_call (gsi, fn, 2 + (objsz != NULL_TREE),
2624 dst, src, objsz);
2625 if (success)
2627 stmt = gsi_stmt (*gsi);
2628 update_stmt (stmt);
2629 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2631 fprintf (dump_file, "into: ");
2632 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
2634 /* If srclen == NULL, note that current string length can be
2635 computed by transforming this strcpy into stpcpy. */
2636 if (srclen == NULL_TREE && dsi->dont_invalidate)
2637 dsi->stmt = stmt;
2638 adjust_last_stmt (dsi, stmt, true);
2639 if (srclen != NULL_TREE)
2641 laststmt.stmt = stmt;
2642 laststmt.len = srclen;
2643 laststmt.stridx = dsi->idx;
2646 else if (dump_file && (dump_flags & TDF_DETAILS) != 0)
2647 fprintf (dump_file, "not possible.\n");
2649 if (set_no_warning)
2650 gimple_set_no_warning (stmt, true);
2653 /* Handle a call to malloc or calloc. */
2655 static void
2656 handle_builtin_malloc (enum built_in_function bcode, gimple_stmt_iterator *gsi)
2658 gimple *stmt = gsi_stmt (*gsi);
2659 tree lhs = gimple_call_lhs (stmt);
2660 if (lhs == NULL_TREE)
2661 return;
2663 gcc_assert (get_stridx (lhs) == 0);
2664 int idx = new_stridx (lhs);
2665 tree length = NULL_TREE;
2666 if (bcode == BUILT_IN_CALLOC)
2667 length = build_int_cst (size_type_node, 0);
2668 strinfo *si = new_strinfo (lhs, idx, length, length != NULL_TREE);
2669 if (bcode == BUILT_IN_CALLOC)
2670 si->endptr = lhs;
2671 set_strinfo (idx, si);
2672 si->writable = true;
2673 si->stmt = stmt;
2674 si->dont_invalidate = true;
2677 /* Handle a call to memset.
2678 After a call to calloc, memset(,0,) is unnecessary.
2679 memset(malloc(n),0,n) is calloc(n,1).
2680 return true when the call is transfomred, false otherwise. */
2682 static bool
2683 handle_builtin_memset (gimple_stmt_iterator *gsi)
2685 gimple *stmt2 = gsi_stmt (*gsi);
2686 if (!integer_zerop (gimple_call_arg (stmt2, 1)))
2687 return false;
2688 tree ptr = gimple_call_arg (stmt2, 0);
2689 int idx1 = get_stridx (ptr);
2690 if (idx1 <= 0)
2691 return false;
2692 strinfo *si1 = get_strinfo (idx1);
2693 if (!si1)
2694 return false;
2695 gimple *stmt1 = si1->stmt;
2696 if (!stmt1 || !is_gimple_call (stmt1))
2697 return false;
2698 tree callee1 = gimple_call_fndecl (stmt1);
2699 if (!valid_builtin_call (stmt1))
2700 return false;
2701 enum built_in_function code1 = DECL_FUNCTION_CODE (callee1);
2702 tree size = gimple_call_arg (stmt2, 2);
2703 if (code1 == BUILT_IN_CALLOC)
2704 /* Not touching stmt1 */ ;
2705 else if (code1 == BUILT_IN_MALLOC
2706 && operand_equal_p (gimple_call_arg (stmt1, 0), size, 0))
2708 gimple_stmt_iterator gsi1 = gsi_for_stmt (stmt1);
2709 update_gimple_call (&gsi1, builtin_decl_implicit (BUILT_IN_CALLOC), 2,
2710 size, build_one_cst (size_type_node));
2711 si1->nonzero_chars = build_int_cst (size_type_node, 0);
2712 si1->full_string_p = true;
2713 si1->stmt = gsi_stmt (gsi1);
2715 else
2716 return false;
2717 tree lhs = gimple_call_lhs (stmt2);
2718 unlink_stmt_vdef (stmt2);
2719 if (lhs)
2721 gimple *assign = gimple_build_assign (lhs, ptr);
2722 gsi_replace (gsi, assign, false);
2724 else
2726 gsi_remove (gsi, true);
2727 release_defs (stmt2);
2730 return true;
2733 /* Handle a call to memcmp. We try to handle small comparisons by
2734 converting them to load and compare, and replacing the call to memcmp
2735 with a __builtin_memcmp_eq call where possible.
2736 return true when call is transformed, return false otherwise. */
2738 static bool
2739 handle_builtin_memcmp (gimple_stmt_iterator *gsi)
2741 gcall *stmt2 = as_a <gcall *> (gsi_stmt (*gsi));
2742 tree res = gimple_call_lhs (stmt2);
2743 tree arg1 = gimple_call_arg (stmt2, 0);
2744 tree arg2 = gimple_call_arg (stmt2, 1);
2745 tree len = gimple_call_arg (stmt2, 2);
2746 unsigned HOST_WIDE_INT leni;
2747 use_operand_p use_p;
2748 imm_use_iterator iter;
2750 if (!res)
2751 return false;
2753 FOR_EACH_IMM_USE_FAST (use_p, iter, res)
2755 gimple *ustmt = USE_STMT (use_p);
2757 if (is_gimple_debug (ustmt))
2758 continue;
2759 if (gimple_code (ustmt) == GIMPLE_ASSIGN)
2761 gassign *asgn = as_a <gassign *> (ustmt);
2762 tree_code code = gimple_assign_rhs_code (asgn);
2763 if ((code != EQ_EXPR && code != NE_EXPR)
2764 || !integer_zerop (gimple_assign_rhs2 (asgn)))
2765 return false;
2767 else if (gimple_code (ustmt) == GIMPLE_COND)
2769 tree_code code = gimple_cond_code (ustmt);
2770 if ((code != EQ_EXPR && code != NE_EXPR)
2771 || !integer_zerop (gimple_cond_rhs (ustmt)))
2772 return false;
2774 else
2775 return false;
2778 if (tree_fits_uhwi_p (len)
2779 && (leni = tree_to_uhwi (len)) <= GET_MODE_SIZE (word_mode)
2780 && pow2p_hwi (leni))
2782 leni *= CHAR_TYPE_SIZE;
2783 unsigned align1 = get_pointer_alignment (arg1);
2784 unsigned align2 = get_pointer_alignment (arg2);
2785 unsigned align = MIN (align1, align2);
2786 scalar_int_mode mode;
2787 if (int_mode_for_size (leni, 1).exists (&mode)
2788 && (align >= leni || !targetm.slow_unaligned_access (mode, align)))
2790 location_t loc = gimple_location (stmt2);
2791 tree type, off;
2792 type = build_nonstandard_integer_type (leni, 1);
2793 gcc_assert (known_eq (GET_MODE_BITSIZE (TYPE_MODE (type)), leni));
2794 tree ptrtype = build_pointer_type_for_mode (char_type_node,
2795 ptr_mode, true);
2796 off = build_int_cst (ptrtype, 0);
2797 arg1 = build2_loc (loc, MEM_REF, type, arg1, off);
2798 arg2 = build2_loc (loc, MEM_REF, type, arg2, off);
2799 tree tem1 = fold_const_aggregate_ref (arg1);
2800 if (tem1)
2801 arg1 = tem1;
2802 tree tem2 = fold_const_aggregate_ref (arg2);
2803 if (tem2)
2804 arg2 = tem2;
2805 res = fold_convert_loc (loc, TREE_TYPE (res),
2806 fold_build2_loc (loc, NE_EXPR,
2807 boolean_type_node,
2808 arg1, arg2));
2809 gimplify_and_update_call_from_tree (gsi, res);
2810 return true;
2814 gimple_call_set_fndecl (stmt2, builtin_decl_explicit (BUILT_IN_MEMCMP_EQ));
2815 return true;
2818 /* Given an index to the strinfo vector, compute the string length for the
2819 corresponding string. Return -1 when unknown. */
2821 static HOST_WIDE_INT
2822 compute_string_length (int idx)
2824 HOST_WIDE_INT string_leni = -1;
2825 gcc_assert (idx != 0);
2827 if (idx < 0)
2828 return ~idx;
2830 strinfo *si = get_strinfo (idx);
2831 if (si)
2833 tree const_string_len = get_string_length (si);
2834 if (const_string_len && tree_fits_shwi_p (const_string_len))
2835 string_leni = tree_to_shwi (const_string_len);
2838 if (string_leni < 0)
2839 return -1;
2841 return string_leni;
2844 /* Determine the minimum size of the object referenced by DEST expression which
2845 must have a pointer type.
2846 Return the minimum size of the object if successful or NULL when the size
2847 cannot be determined. */
2848 static tree
2849 determine_min_objsize (tree dest)
2851 unsigned HOST_WIDE_INT size = 0;
2853 if (compute_builtin_object_size (dest, 2, &size))
2854 return build_int_cst (sizetype, size);
2856 /* Try to determine the size of the object through the RHS of the
2857 assign statement. */
2858 if (TREE_CODE (dest) == SSA_NAME)
2860 gimple *stmt = SSA_NAME_DEF_STMT (dest);
2861 if (!is_gimple_assign (stmt))
2862 return NULL_TREE;
2864 if (!gimple_assign_single_p (stmt)
2865 && !gimple_assign_unary_nop_p (stmt))
2866 return NULL_TREE;
2868 dest = gimple_assign_rhs1 (stmt);
2869 return determine_min_objsize (dest);
2872 /* Try to determine the size of the object from its type. */
2873 if (TREE_CODE (dest) != ADDR_EXPR)
2874 return NULL_TREE;
2876 tree type = TREE_TYPE (dest);
2877 if (TREE_CODE (type) == POINTER_TYPE)
2878 type = TREE_TYPE (type);
2880 type = TYPE_MAIN_VARIANT (type);
2882 /* We cannot determine the size of the array if it's a flexible array,
2883 which is declared at the end of a structure. */
2884 if (TREE_CODE (type) == ARRAY_TYPE
2885 && !array_at_struct_end_p (dest))
2887 tree size_t = TYPE_SIZE_UNIT (type);
2888 if (size_t && TREE_CODE (size_t) == INTEGER_CST
2889 && !integer_zerop (size_t))
2890 return size_t;
2893 return NULL_TREE;
2896 /* Handle a call to strcmp or strncmp. When the result is ONLY used to do
2897 equality test against zero:
2899 A. When the lengths of both arguments are constant and it's a strcmp:
2900 * if the lengths are NOT equal, we can safely fold the call
2901 to a non-zero value.
2902 * otherwise, do nothing now.
2904 B. When the length of one argument is constant, try to replace the call with
2905 a __builtin_str(n)cmp_eq call where possible, i.e:
2907 strncmp (s, STR, C) (!)= 0 in which, s is a pointer to a string, STR is a
2908 string with constant length , C is a constant.
2909 if (C <= strlen(STR) && sizeof_array(s) > C)
2911 replace this call with
2912 strncmp_eq (s, STR, C) (!)= 0
2914 if (C > strlen(STR)
2916 it can be safely treated as a call to strcmp (s, STR) (!)= 0
2917 can handled by the following strcmp.
2920 strcmp (s, STR) (!)= 0 in which, s is a pointer to a string, STR is a
2921 string with constant length.
2922 if (sizeof_array(s) > strlen(STR))
2924 replace this call with
2925 strcmp_eq (s, STR, strlen(STR)+1) (!)= 0
2928 Return true when the call is transformed, return false otherwise.
2931 static bool
2932 handle_builtin_string_cmp (gimple_stmt_iterator *gsi)
2934 gcall *stmt = as_a <gcall *> (gsi_stmt (*gsi));
2935 tree res = gimple_call_lhs (stmt);
2936 use_operand_p use_p;
2937 imm_use_iterator iter;
2938 tree arg1 = gimple_call_arg (stmt, 0);
2939 tree arg2 = gimple_call_arg (stmt, 1);
2940 int idx1 = get_stridx (arg1);
2941 int idx2 = get_stridx (arg2);
2942 HOST_WIDE_INT length = -1;
2943 bool is_ncmp = false;
2945 if (!res)
2946 return false;
2948 /* When both arguments are unknown, do nothing. */
2949 if (idx1 == 0 && idx2 == 0)
2950 return false;
2952 /* Handle strncmp function. */
2953 if (gimple_call_num_args (stmt) == 3)
2955 tree len = gimple_call_arg (stmt, 2);
2956 if (tree_fits_shwi_p (len))
2957 length = tree_to_shwi (len);
2959 is_ncmp = true;
2962 /* For strncmp, if the length argument is NOT known, do nothing. */
2963 if (is_ncmp && length < 0)
2964 return false;
2966 /* When the result is ONLY used to do equality test against zero. */
2967 FOR_EACH_IMM_USE_FAST (use_p, iter, res)
2969 gimple *use_stmt = USE_STMT (use_p);
2971 if (is_gimple_debug (use_stmt))
2972 continue;
2973 if (gimple_code (use_stmt) == GIMPLE_ASSIGN)
2975 tree_code code = gimple_assign_rhs_code (use_stmt);
2976 if (code == COND_EXPR)
2978 tree cond_expr = gimple_assign_rhs1 (use_stmt);
2979 if ((TREE_CODE (cond_expr) != EQ_EXPR
2980 && (TREE_CODE (cond_expr) != NE_EXPR))
2981 || !integer_zerop (TREE_OPERAND (cond_expr, 1)))
2982 return false;
2984 else if (code == EQ_EXPR || code == NE_EXPR)
2986 if (!integer_zerop (gimple_assign_rhs2 (use_stmt)))
2987 return false;
2989 else
2990 return false;
2992 else if (gimple_code (use_stmt) == GIMPLE_COND)
2994 tree_code code = gimple_cond_code (use_stmt);
2995 if ((code != EQ_EXPR && code != NE_EXPR)
2996 || !integer_zerop (gimple_cond_rhs (use_stmt)))
2997 return false;
2999 else
3000 return false;
3003 /* When the lengths of both arguments are known, and they are unequal, we can
3004 safely fold the call to a non-zero value for strcmp;
3005 othewise, do nothing now. */
3006 if (idx1 != 0 && idx2 != 0)
3008 HOST_WIDE_INT const_string_leni1 = compute_string_length (idx1);
3009 HOST_WIDE_INT const_string_leni2 = compute_string_length (idx2);
3011 if (!is_ncmp
3012 && const_string_leni1 != -1
3013 && const_string_leni2 != -1
3014 && const_string_leni1 != const_string_leni2)
3016 replace_call_with_value (gsi, integer_one_node);
3017 return true;
3019 return false;
3022 /* When the length of one argument is constant. */
3023 tree var_string = NULL_TREE;
3024 HOST_WIDE_INT const_string_leni = -1;
3026 if (idx1)
3028 const_string_leni = compute_string_length (idx1);
3029 var_string = arg2;
3031 else
3033 gcc_checking_assert (idx2);
3034 const_string_leni = compute_string_length (idx2);
3035 var_string = arg1;
3038 if (const_string_leni < 0)
3039 return false;
3041 unsigned HOST_WIDE_INT var_sizei = 0;
3042 /* try to determine the minimum size of the object pointed by var_string. */
3043 tree size = determine_min_objsize (var_string);
3045 if (!size)
3046 return false;
3048 if (tree_fits_uhwi_p (size))
3049 var_sizei = tree_to_uhwi (size);
3051 if (var_sizei == 0)
3052 return false;
3054 /* For strncmp, if length > const_string_leni , this call can be safely
3055 transformed to a strcmp. */
3056 if (is_ncmp && length > const_string_leni)
3057 is_ncmp = false;
3059 unsigned HOST_WIDE_INT final_length
3060 = is_ncmp ? length : const_string_leni + 1;
3062 /* Replace strcmp or strncmp with the corresponding str(n)cmp_eq. */
3063 if (var_sizei > final_length)
3065 tree fn
3066 = (is_ncmp
3067 ? builtin_decl_implicit (BUILT_IN_STRNCMP_EQ)
3068 : builtin_decl_implicit (BUILT_IN_STRCMP_EQ));
3069 if (!fn)
3070 return false;
3071 tree const_string_len = build_int_cst (size_type_node, final_length);
3072 update_gimple_call (gsi, fn, 3, arg1, arg2, const_string_len);
3074 else
3075 return false;
3077 return true;
3080 /* Handle a POINTER_PLUS_EXPR statement.
3081 For p = "abcd" + 2; compute associated length, or if
3082 p = q + off is pointing to a '\0' character of a string, call
3083 zero_length_string on it. */
3085 static void
3086 handle_pointer_plus (gimple_stmt_iterator *gsi)
3088 gimple *stmt = gsi_stmt (*gsi);
3089 tree lhs = gimple_assign_lhs (stmt), off;
3090 int idx = get_stridx (gimple_assign_rhs1 (stmt));
3091 strinfo *si, *zsi;
3093 if (idx == 0)
3094 return;
3096 if (idx < 0)
3098 tree off = gimple_assign_rhs2 (stmt);
3099 if (tree_fits_uhwi_p (off)
3100 && tree_to_uhwi (off) <= (unsigned HOST_WIDE_INT) ~idx)
3101 ssa_ver_to_stridx[SSA_NAME_VERSION (lhs)]
3102 = ~(~idx - (int) tree_to_uhwi (off));
3103 return;
3106 si = get_strinfo (idx);
3107 if (si == NULL || si->nonzero_chars == NULL_TREE)
3108 return;
3110 off = gimple_assign_rhs2 (stmt);
3111 zsi = NULL;
3112 if (si->full_string_p && operand_equal_p (si->nonzero_chars, off, 0))
3113 zsi = zero_length_string (lhs, si);
3114 else if (TREE_CODE (off) == SSA_NAME)
3116 gimple *def_stmt = SSA_NAME_DEF_STMT (off);
3117 if (gimple_assign_single_p (def_stmt)
3118 && si->full_string_p
3119 && operand_equal_p (si->nonzero_chars,
3120 gimple_assign_rhs1 (def_stmt), 0))
3121 zsi = zero_length_string (lhs, si);
3123 if (zsi != NULL
3124 && si->endptr != NULL_TREE
3125 && si->endptr != lhs
3126 && TREE_CODE (si->endptr) == SSA_NAME)
3128 enum tree_code rhs_code
3129 = useless_type_conversion_p (TREE_TYPE (lhs), TREE_TYPE (si->endptr))
3130 ? SSA_NAME : NOP_EXPR;
3131 gimple_assign_set_rhs_with_ops (gsi, rhs_code, si->endptr);
3132 gcc_assert (gsi_stmt (*gsi) == stmt);
3133 update_stmt (stmt);
3137 /* If RHS, either directly or indirectly, refers to a string of constant
3138 length, return the length. Otherwise, if it refers to a character
3139 constant, return 1 if the constant is non-zero and 0 if it is nul.
3140 Otherwise, return a negative value. */
3142 static HOST_WIDE_INT
3143 get_min_string_length (tree rhs, bool *full_string_p)
3145 if (INTEGRAL_TYPE_P (TREE_TYPE (rhs)))
3147 if (tree_expr_nonzero_p (rhs))
3149 *full_string_p = false;
3150 return 1;
3153 *full_string_p = true;
3154 return 0;
3157 if (TREE_CODE (rhs) == MEM_REF
3158 && integer_zerop (TREE_OPERAND (rhs, 1)))
3160 rhs = TREE_OPERAND (rhs, 0);
3161 if (TREE_CODE (rhs) == ADDR_EXPR)
3163 tree rhs_addr = rhs;
3165 rhs = TREE_OPERAND (rhs, 0);
3166 if (TREE_CODE (rhs) != STRING_CST)
3168 int idx = get_stridx (rhs_addr);
3169 if (idx > 0)
3171 strinfo *si = get_strinfo (idx);
3172 if (si
3173 && tree_fits_shwi_p (si->nonzero_chars))
3175 *full_string_p = si->full_string_p;
3176 return tree_to_shwi (si->nonzero_chars);
3183 if (TREE_CODE (rhs) == VAR_DECL
3184 && TREE_READONLY (rhs))
3185 rhs = DECL_INITIAL (rhs);
3187 if (rhs && TREE_CODE (rhs) == STRING_CST)
3189 *full_string_p = true;
3190 return strlen (TREE_STRING_POINTER (rhs));
3193 return -1;
3196 /* Handle a single or multiple character store either by single
3197 character assignment or by multi-character assignment from
3198 MEM_REF. */
3200 static bool
3201 handle_char_store (gimple_stmt_iterator *gsi)
3203 int idx = -1;
3204 strinfo *si = NULL;
3205 gimple *stmt = gsi_stmt (*gsi);
3206 tree ssaname = NULL_TREE, lhs = gimple_assign_lhs (stmt);
3207 tree rhs = gimple_assign_rhs1 (stmt);
3208 unsigned HOST_WIDE_INT offset = 0;
3210 if (TREE_CODE (lhs) == MEM_REF
3211 && TREE_CODE (TREE_OPERAND (lhs, 0)) == SSA_NAME)
3213 tree mem_offset = TREE_OPERAND (lhs, 1);
3214 if (tree_fits_uhwi_p (mem_offset))
3216 /* Get the strinfo for the base, and use it if it starts with at
3217 least OFFSET nonzero characters. This is trivially true if
3218 OFFSET is zero. */
3219 offset = tree_to_uhwi (mem_offset);
3220 idx = get_stridx (TREE_OPERAND (lhs, 0));
3221 if (idx > 0)
3222 si = get_strinfo (idx);
3223 if (offset == 0)
3224 ssaname = TREE_OPERAND (lhs, 0);
3225 else if (si == NULL || compare_nonzero_chars (si, offset) < 0)
3226 return true;
3229 else
3231 idx = get_addr_stridx (lhs, NULL_TREE, &offset);
3232 if (idx > 0)
3233 si = get_strinfo (idx);
3236 /* STORING_NONZERO_P is true iff not all stored characters are zero.
3237 STORING_ALL_ZEROS_P is true iff all stored characters are zero.
3238 Both are false when it's impossible to determine which is true. */
3239 bool storing_nonzero_p;
3240 bool storing_all_zeros_p = initializer_zerop (rhs, &storing_nonzero_p);
3241 if (!storing_nonzero_p)
3242 storing_nonzero_p = tree_expr_nonzero_p (rhs);
3243 bool full_string_p = storing_all_zeros_p;
3245 /* Set to the length of the string being assigned if known. */
3246 HOST_WIDE_INT rhslen;
3248 if (si != NULL)
3250 int cmp = compare_nonzero_chars (si, offset);
3251 gcc_assert (offset == 0 || cmp >= 0);
3252 if (storing_all_zeros_p && cmp == 0 && si->full_string_p)
3254 /* When overwriting a '\0' with a '\0', the store can be removed
3255 if we know it has been stored in the current function. */
3256 if (!stmt_could_throw_p (stmt) && si->writable)
3258 unlink_stmt_vdef (stmt);
3259 release_defs (stmt);
3260 gsi_remove (gsi, true);
3261 return false;
3263 else
3265 si->writable = true;
3266 gsi_next (gsi);
3267 return false;
3270 /* If si->nonzero_chars > OFFSET, we aren't overwriting '\0',
3271 and if we aren't storing '\0', we know that the length of the
3272 string and any other zero terminated string in memory remains
3273 the same. In that case we move to the next gimple statement and
3274 return to signal the caller that it shouldn't invalidate anything.
3276 This is benefical for cases like:
3278 char p[20];
3279 void foo (char *q)
3281 strcpy (p, "foobar");
3282 size_t len = strlen (p); // This can be optimized into 6
3283 size_t len2 = strlen (q); // This has to be computed
3284 p[0] = 'X';
3285 size_t len3 = strlen (p); // This can be optimized into 6
3286 size_t len4 = strlen (q); // This can be optimized into len2
3287 bar (len, len2, len3, len4);
3290 else if (storing_nonzero_p && cmp > 0)
3292 gsi_next (gsi);
3293 return false;
3295 else if (storing_all_zeros_p || storing_nonzero_p || (offset != 0 && cmp > 0))
3297 /* When STORING_NONZERO_P, we know that the string will start
3298 with at least OFFSET + 1 nonzero characters. If storing
3299 a single character, set si->NONZERO_CHARS to the result.
3300 If storing multiple characters, try to determine the number
3301 of leading non-zero characters and set si->NONZERO_CHARS to
3302 the result instead.
3304 When STORING_ALL_ZEROS_P, we know that the string is now
3305 OFFSET characters long.
3307 Otherwise, we're storing an unknown value at offset OFFSET,
3308 so need to clip the nonzero_chars to OFFSET. */
3309 bool full_string_p = storing_all_zeros_p;
3310 HOST_WIDE_INT len = 1;
3311 if (storing_nonzero_p)
3313 /* Try to get the minimum length of the string (or
3314 individual character) being stored. If it fails,
3315 STORING_NONZERO_P guarantees it's at least 1. */
3316 len = get_min_string_length (rhs, &full_string_p);
3317 if (len < 0)
3318 len = 1;
3321 location_t loc = gimple_location (stmt);
3322 tree oldlen = si->nonzero_chars;
3323 if (cmp == 0 && si->full_string_p)
3324 /* We're overwriting the nul terminator with a nonzero or
3325 unknown character. If the previous stmt was a memcpy,
3326 its length may be decreased. */
3327 adjust_last_stmt (si, stmt, false);
3328 si = unshare_strinfo (si);
3329 if (storing_nonzero_p)
3331 gcc_assert (len >= 0);
3332 si->nonzero_chars = build_int_cst (size_type_node, offset + len);
3334 else
3335 si->nonzero_chars = build_int_cst (size_type_node, offset);
3336 si->full_string_p = full_string_p;
3337 if (storing_all_zeros_p
3338 && ssaname
3339 && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (ssaname))
3340 si->endptr = ssaname;
3341 else
3342 si->endptr = NULL;
3343 si->next = 0;
3344 si->stmt = NULL;
3345 si->writable = true;
3346 si->dont_invalidate = true;
3347 if (oldlen)
3349 tree adj = fold_build2_loc (loc, MINUS_EXPR, size_type_node,
3350 si->nonzero_chars, oldlen);
3351 adjust_related_strinfos (loc, si, adj);
3353 else
3354 si->prev = 0;
3357 else if (idx == 0 && (storing_all_zeros_p || storing_nonzero_p))
3359 if (ssaname)
3360 idx = new_stridx (ssaname);
3361 else
3362 idx = new_addr_stridx (lhs);
3363 if (idx != 0)
3365 tree ptr = (ssaname ? ssaname : build_fold_addr_expr (lhs));
3366 HOST_WIDE_INT slen = (storing_all_zeros_p
3368 : (storing_nonzero_p
3369 ? get_min_string_length (rhs, &full_string_p)
3370 : -1));
3371 tree len = (slen <= 0
3372 ? size_zero_node
3373 : build_int_cst (size_type_node, slen));
3374 si = new_strinfo (ptr, idx, len, slen >= 0 && full_string_p);
3375 set_strinfo (idx, si);
3376 if (storing_all_zeros_p
3377 && ssaname
3378 && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (ssaname))
3379 si->endptr = ssaname;
3380 si->dont_invalidate = true;
3381 si->writable = true;
3384 else if (idx == 0
3385 && (rhslen = get_min_string_length (rhs, &full_string_p)) >= 0
3386 && ssaname == NULL_TREE
3387 && TREE_CODE (TREE_TYPE (lhs)) == ARRAY_TYPE)
3389 HOST_WIDE_INT a = int_size_in_bytes (TREE_TYPE (lhs));
3390 if (a > 0 && (unsigned HOST_WIDE_INT) a > (unsigned HOST_WIDE_INT) rhslen)
3392 int idx = new_addr_stridx (lhs);
3393 if (idx != 0)
3395 si = new_strinfo (build_fold_addr_expr (lhs), idx,
3396 build_int_cst (size_type_node, rhslen),
3397 full_string_p);
3398 set_strinfo (idx, si);
3399 si->dont_invalidate = true;
3404 if (si != NULL && offset == 0 && storing_all_zeros_p)
3406 /* Allow adjust_last_stmt to remove it if the stored '\0'
3407 is immediately overwritten. */
3408 laststmt.stmt = stmt;
3409 laststmt.len = build_int_cst (size_type_node, 1);
3410 laststmt.stridx = si->idx;
3412 return true;
3415 /* Try to fold strstr (s, t) eq/ne s to strncmp (s, t, strlen (t)) eq/ne 0. */
3417 static void
3418 fold_strstr_to_strncmp (tree rhs1, tree rhs2, gimple *stmt)
3420 if (TREE_CODE (rhs1) != SSA_NAME
3421 || TREE_CODE (rhs2) != SSA_NAME)
3422 return;
3424 gimple *call_stmt = NULL;
3425 for (int pass = 0; pass < 2; pass++)
3427 gimple *g = SSA_NAME_DEF_STMT (rhs1);
3428 if (gimple_call_builtin_p (g, BUILT_IN_STRSTR)
3429 && has_single_use (rhs1)
3430 && gimple_call_arg (g, 0) == rhs2)
3432 call_stmt = g;
3433 break;
3435 std::swap (rhs1, rhs2);
3438 if (call_stmt)
3440 tree arg0 = gimple_call_arg (call_stmt, 0);
3442 if (arg0 == rhs2)
3444 tree arg1 = gimple_call_arg (call_stmt, 1);
3445 tree arg1_len = NULL_TREE;
3446 int idx = get_stridx (arg1);
3448 if (idx)
3450 if (idx < 0)
3451 arg1_len = build_int_cst (size_type_node, ~idx);
3452 else
3454 strinfo *si = get_strinfo (idx);
3455 if (si)
3456 arg1_len = get_string_length (si);
3460 if (arg1_len != NULL_TREE)
3462 gimple_stmt_iterator gsi = gsi_for_stmt (call_stmt);
3463 tree strncmp_decl = builtin_decl_explicit (BUILT_IN_STRNCMP);
3465 if (!is_gimple_val (arg1_len))
3467 tree arg1_len_tmp = make_ssa_name (TREE_TYPE (arg1_len));
3468 gassign *arg1_stmt = gimple_build_assign (arg1_len_tmp,
3469 arg1_len);
3470 gsi_insert_before (&gsi, arg1_stmt, GSI_SAME_STMT);
3471 arg1_len = arg1_len_tmp;
3474 gcall *strncmp_call = gimple_build_call (strncmp_decl, 3,
3475 arg0, arg1, arg1_len);
3476 tree strncmp_lhs = make_ssa_name (integer_type_node);
3477 gimple_set_vuse (strncmp_call, gimple_vuse (call_stmt));
3478 gimple_call_set_lhs (strncmp_call, strncmp_lhs);
3479 gsi_remove (&gsi, true);
3480 gsi_insert_before (&gsi, strncmp_call, GSI_SAME_STMT);
3481 tree zero = build_zero_cst (TREE_TYPE (strncmp_lhs));
3483 if (is_gimple_assign (stmt))
3485 if (gimple_assign_rhs_code (stmt) == COND_EXPR)
3487 tree cond = gimple_assign_rhs1 (stmt);
3488 TREE_OPERAND (cond, 0) = strncmp_lhs;
3489 TREE_OPERAND (cond, 1) = zero;
3491 else
3493 gimple_assign_set_rhs1 (stmt, strncmp_lhs);
3494 gimple_assign_set_rhs2 (stmt, zero);
3497 else
3499 gcond *cond = as_a<gcond *> (stmt);
3500 gimple_cond_set_lhs (cond, strncmp_lhs);
3501 gimple_cond_set_rhs (cond, zero);
3503 update_stmt (stmt);
3509 /* Attempt to check for validity of the performed access a single statement
3510 at *GSI using string length knowledge, and to optimize it.
3511 If the given basic block needs clean-up of EH, CLEANUP_EH is set to
3512 true. */
3514 static bool
3515 strlen_check_and_optimize_stmt (gimple_stmt_iterator *gsi, bool *cleanup_eh)
3517 gimple *stmt = gsi_stmt (*gsi);
3519 if (is_gimple_call (stmt))
3521 tree callee = gimple_call_fndecl (stmt);
3522 if (valid_builtin_call (stmt))
3523 switch (DECL_FUNCTION_CODE (callee))
3525 case BUILT_IN_STRLEN:
3526 case BUILT_IN_STRNLEN:
3527 handle_builtin_strlen (gsi);
3528 break;
3529 case BUILT_IN_STRCHR:
3530 handle_builtin_strchr (gsi);
3531 break;
3532 case BUILT_IN_STRCPY:
3533 case BUILT_IN_STRCPY_CHK:
3534 case BUILT_IN_STPCPY:
3535 case BUILT_IN_STPCPY_CHK:
3536 handle_builtin_strcpy (DECL_FUNCTION_CODE (callee), gsi);
3537 break;
3539 case BUILT_IN_STRNCAT:
3540 case BUILT_IN_STRNCAT_CHK:
3541 handle_builtin_strncat (DECL_FUNCTION_CODE (callee), gsi);
3542 break;
3544 case BUILT_IN_STPNCPY:
3545 case BUILT_IN_STPNCPY_CHK:
3546 case BUILT_IN_STRNCPY:
3547 case BUILT_IN_STRNCPY_CHK:
3548 handle_builtin_stxncpy (DECL_FUNCTION_CODE (callee), gsi);
3549 break;
3551 case BUILT_IN_MEMCPY:
3552 case BUILT_IN_MEMCPY_CHK:
3553 case BUILT_IN_MEMPCPY:
3554 case BUILT_IN_MEMPCPY_CHK:
3555 handle_builtin_memcpy (DECL_FUNCTION_CODE (callee), gsi);
3556 break;
3557 case BUILT_IN_STRCAT:
3558 case BUILT_IN_STRCAT_CHK:
3559 handle_builtin_strcat (DECL_FUNCTION_CODE (callee), gsi);
3560 break;
3561 case BUILT_IN_MALLOC:
3562 case BUILT_IN_CALLOC:
3563 handle_builtin_malloc (DECL_FUNCTION_CODE (callee), gsi);
3564 break;
3565 case BUILT_IN_MEMSET:
3566 if (handle_builtin_memset (gsi))
3567 return false;
3568 break;
3569 case BUILT_IN_MEMCMP:
3570 if (handle_builtin_memcmp (gsi))
3571 return false;
3572 break;
3573 case BUILT_IN_STRCMP:
3574 case BUILT_IN_STRNCMP:
3575 if (handle_builtin_string_cmp (gsi))
3576 return false;
3577 break;
3578 default:
3579 break;
3582 else if (is_gimple_assign (stmt) && !gimple_clobber_p (stmt))
3584 tree lhs = gimple_assign_lhs (stmt);
3586 if (TREE_CODE (lhs) == SSA_NAME && POINTER_TYPE_P (TREE_TYPE (lhs)))
3588 if (gimple_assign_single_p (stmt)
3589 || (gimple_assign_cast_p (stmt)
3590 && POINTER_TYPE_P (TREE_TYPE (gimple_assign_rhs1 (stmt)))))
3592 int idx = get_stridx (gimple_assign_rhs1 (stmt));
3593 ssa_ver_to_stridx[SSA_NAME_VERSION (lhs)] = idx;
3595 else if (gimple_assign_rhs_code (stmt) == POINTER_PLUS_EXPR)
3596 handle_pointer_plus (gsi);
3598 else if (TREE_CODE (lhs) == SSA_NAME && INTEGRAL_TYPE_P (TREE_TYPE (lhs)))
3600 enum tree_code code = gimple_assign_rhs_code (stmt);
3601 if (code == COND_EXPR)
3603 tree cond = gimple_assign_rhs1 (stmt);
3604 enum tree_code cond_code = TREE_CODE (cond);
3606 if (cond_code == EQ_EXPR || cond_code == NE_EXPR)
3607 fold_strstr_to_strncmp (TREE_OPERAND (cond, 0),
3608 TREE_OPERAND (cond, 1), stmt);
3610 else if (code == EQ_EXPR || code == NE_EXPR)
3611 fold_strstr_to_strncmp (gimple_assign_rhs1 (stmt),
3612 gimple_assign_rhs2 (stmt), stmt);
3613 else if (gimple_assign_load_p (stmt)
3614 && TREE_CODE (TREE_TYPE (lhs)) == INTEGER_TYPE
3615 && TYPE_MODE (TREE_TYPE (lhs)) == TYPE_MODE (char_type_node)
3616 && (TYPE_PRECISION (TREE_TYPE (lhs))
3617 == TYPE_PRECISION (char_type_node))
3618 && !gimple_has_volatile_ops (stmt))
3620 tree off = integer_zero_node;
3621 unsigned HOST_WIDE_INT coff = 0;
3622 int idx = 0;
3623 tree rhs1 = gimple_assign_rhs1 (stmt);
3624 if (code == MEM_REF)
3626 idx = get_stridx (TREE_OPERAND (rhs1, 0));
3627 if (idx > 0)
3629 strinfo *si = get_strinfo (idx);
3630 if (si
3631 && si->nonzero_chars
3632 && TREE_CODE (si->nonzero_chars) == INTEGER_CST
3633 && (wi::to_widest (si->nonzero_chars)
3634 >= wi::to_widest (off)))
3635 off = TREE_OPERAND (rhs1, 1);
3636 else
3637 /* This case is not useful. See if get_addr_stridx
3638 returns something usable. */
3639 idx = 0;
3642 if (idx <= 0)
3643 idx = get_addr_stridx (rhs1, NULL_TREE, &coff);
3644 if (idx > 0)
3646 strinfo *si = get_strinfo (idx);
3647 if (si
3648 && si->nonzero_chars
3649 && TREE_CODE (si->nonzero_chars) == INTEGER_CST)
3651 widest_int w1 = wi::to_widest (si->nonzero_chars);
3652 widest_int w2 = wi::to_widest (off) + coff;
3653 if (w1 == w2
3654 && si->full_string_p)
3656 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
3658 fprintf (dump_file, "Optimizing: ");
3659 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
3662 /* Reading the final '\0' character. */
3663 tree zero = build_int_cst (TREE_TYPE (lhs), 0);
3664 gimple_set_vuse (stmt, NULL_TREE);
3665 gimple_assign_set_rhs_from_tree (gsi, zero);
3666 *cleanup_eh
3667 |= maybe_clean_or_replace_eh_stmt (stmt,
3668 gsi_stmt (*gsi));
3669 stmt = gsi_stmt (*gsi);
3670 update_stmt (stmt);
3672 if (dump_file && (dump_flags & TDF_DETAILS) != 0)
3674 fprintf (dump_file, "into: ");
3675 print_gimple_stmt (dump_file, stmt, 0, TDF_SLIM);
3678 else if (w1 > w2)
3680 /* Reading a character before the final '\0'
3681 character. Just set the value range to ~[0, 0]
3682 if we don't have anything better. */
3683 wide_int min, max;
3684 tree type = TREE_TYPE (lhs);
3685 enum value_range_type vr
3686 = get_range_info (lhs, &min, &max);
3687 if (vr == VR_VARYING
3688 || (vr == VR_RANGE
3689 && min == wi::min_value (TYPE_PRECISION (type),
3690 TYPE_SIGN (type))
3691 && max == wi::max_value (TYPE_PRECISION (type),
3692 TYPE_SIGN (type))))
3693 set_range_info (lhs, VR_ANTI_RANGE,
3694 wi::zero (TYPE_PRECISION (type)),
3695 wi::zero (TYPE_PRECISION (type)));
3701 if (strlen_to_stridx)
3703 tree rhs1 = gimple_assign_rhs1 (stmt);
3704 if (stridx_strlenloc *ps = strlen_to_stridx->get (rhs1))
3705 strlen_to_stridx->put (lhs, stridx_strlenloc (*ps));
3708 else if (TREE_CODE (lhs) != SSA_NAME && !TREE_SIDE_EFFECTS (lhs))
3710 tree type = TREE_TYPE (lhs);
3711 if (TREE_CODE (type) == ARRAY_TYPE)
3712 type = TREE_TYPE (type);
3713 if (TREE_CODE (type) == INTEGER_TYPE
3714 && TYPE_MODE (type) == TYPE_MODE (char_type_node)
3715 && TYPE_PRECISION (type) == TYPE_PRECISION (char_type_node))
3717 if (! handle_char_store (gsi))
3718 return false;
3722 else if (gcond *cond = dyn_cast<gcond *> (stmt))
3724 enum tree_code code = gimple_cond_code (cond);
3725 if (code == EQ_EXPR || code == NE_EXPR)
3726 fold_strstr_to_strncmp (gimple_cond_lhs (stmt),
3727 gimple_cond_rhs (stmt), stmt);
3730 if (gimple_vdef (stmt))
3731 maybe_invalidate (stmt);
3732 return true;
3735 /* Recursively call maybe_invalidate on stmts that might be executed
3736 in between dombb and current bb and that contain a vdef. Stop when
3737 *count stmts are inspected, or if the whole strinfo vector has
3738 been invalidated. */
3740 static void
3741 do_invalidate (basic_block dombb, gimple *phi, bitmap visited, int *count)
3743 unsigned int i, n = gimple_phi_num_args (phi);
3745 for (i = 0; i < n; i++)
3747 tree vuse = gimple_phi_arg_def (phi, i);
3748 gimple *stmt = SSA_NAME_DEF_STMT (vuse);
3749 basic_block bb = gimple_bb (stmt);
3750 if (bb == NULL
3751 || bb == dombb
3752 || !bitmap_set_bit (visited, bb->index)
3753 || !dominated_by_p (CDI_DOMINATORS, bb, dombb))
3754 continue;
3755 while (1)
3757 if (gimple_code (stmt) == GIMPLE_PHI)
3759 do_invalidate (dombb, stmt, visited, count);
3760 if (*count == 0)
3761 return;
3762 break;
3764 if (--*count == 0)
3765 return;
3766 if (!maybe_invalidate (stmt))
3768 *count = 0;
3769 return;
3771 vuse = gimple_vuse (stmt);
3772 stmt = SSA_NAME_DEF_STMT (vuse);
3773 if (gimple_bb (stmt) != bb)
3775 bb = gimple_bb (stmt);
3776 if (bb == NULL
3777 || bb == dombb
3778 || !bitmap_set_bit (visited, bb->index)
3779 || !dominated_by_p (CDI_DOMINATORS, bb, dombb))
3780 break;
3786 class strlen_dom_walker : public dom_walker
3788 public:
3789 strlen_dom_walker (cdi_direction direction)
3790 : dom_walker (direction), m_cleanup_cfg (false)
3793 virtual edge before_dom_children (basic_block);
3794 virtual void after_dom_children (basic_block);
3796 /* Flag that will trigger TODO_cleanup_cfg to be returned in strlen
3797 execute function. */
3798 bool m_cleanup_cfg;
3801 /* Callback for walk_dominator_tree. Attempt to optimize various
3802 string ops by remembering string lengths pointed by pointer SSA_NAMEs. */
3804 edge
3805 strlen_dom_walker::before_dom_children (basic_block bb)
3807 basic_block dombb = get_immediate_dominator (CDI_DOMINATORS, bb);
3809 if (dombb == NULL)
3810 stridx_to_strinfo = NULL;
3811 else
3813 stridx_to_strinfo = ((vec<strinfo *, va_heap, vl_embed> *) dombb->aux);
3814 if (stridx_to_strinfo)
3816 for (gphi_iterator gsi = gsi_start_phis (bb); !gsi_end_p (gsi);
3817 gsi_next (&gsi))
3819 gphi *phi = gsi.phi ();
3820 if (virtual_operand_p (gimple_phi_result (phi)))
3822 bitmap visited = BITMAP_ALLOC (NULL);
3823 int count_vdef = 100;
3824 do_invalidate (dombb, phi, visited, &count_vdef);
3825 BITMAP_FREE (visited);
3826 if (count_vdef == 0)
3828 /* If there were too many vdefs in between immediate
3829 dominator and current bb, invalidate everything.
3830 If stridx_to_strinfo has been unshared, we need
3831 to free it, otherwise just set it to NULL. */
3832 if (!strinfo_shared ())
3834 unsigned int i;
3835 strinfo *si;
3837 for (i = 1;
3838 vec_safe_iterate (stridx_to_strinfo, i, &si);
3839 ++i)
3841 free_strinfo (si);
3842 (*stridx_to_strinfo)[i] = NULL;
3845 else
3846 stridx_to_strinfo = NULL;
3848 break;
3854 /* If all PHI arguments have the same string index, the PHI result
3855 has it as well. */
3856 for (gphi_iterator gsi = gsi_start_phis (bb); !gsi_end_p (gsi);
3857 gsi_next (&gsi))
3859 gphi *phi = gsi.phi ();
3860 tree result = gimple_phi_result (phi);
3861 if (!virtual_operand_p (result) && POINTER_TYPE_P (TREE_TYPE (result)))
3863 int idx = get_stridx (gimple_phi_arg_def (phi, 0));
3864 if (idx != 0)
3866 unsigned int i, n = gimple_phi_num_args (phi);
3867 for (i = 1; i < n; i++)
3868 if (idx != get_stridx (gimple_phi_arg_def (phi, i)))
3869 break;
3870 if (i == n)
3871 ssa_ver_to_stridx[SSA_NAME_VERSION (result)] = idx;
3876 bool cleanup_eh = false;
3878 /* Attempt to optimize individual statements. */
3879 for (gimple_stmt_iterator gsi = gsi_start_bb (bb); !gsi_end_p (gsi); )
3880 if (strlen_check_and_optimize_stmt (&gsi, &cleanup_eh))
3881 gsi_next (&gsi);
3883 if (cleanup_eh && gimple_purge_dead_eh_edges (bb))
3884 m_cleanup_cfg = true;
3886 bb->aux = stridx_to_strinfo;
3887 if (vec_safe_length (stridx_to_strinfo) && !strinfo_shared ())
3888 (*stridx_to_strinfo)[0] = (strinfo *) bb;
3889 return NULL;
3892 /* Callback for walk_dominator_tree. Free strinfo vector if it is
3893 owned by the current bb, clear bb->aux. */
3895 void
3896 strlen_dom_walker::after_dom_children (basic_block bb)
3898 if (bb->aux)
3900 stridx_to_strinfo = ((vec<strinfo *, va_heap, vl_embed> *) bb->aux);
3901 if (vec_safe_length (stridx_to_strinfo)
3902 && (*stridx_to_strinfo)[0] == (strinfo *) bb)
3904 unsigned int i;
3905 strinfo *si;
3907 for (i = 1; vec_safe_iterate (stridx_to_strinfo, i, &si); ++i)
3908 free_strinfo (si);
3909 vec_free (stridx_to_strinfo);
3911 bb->aux = NULL;
3915 /* Main entry point. */
3917 namespace {
3919 const pass_data pass_data_strlen =
3921 GIMPLE_PASS, /* type */
3922 "strlen", /* name */
3923 OPTGROUP_NONE, /* optinfo_flags */
3924 TV_TREE_STRLEN, /* tv_id */
3925 ( PROP_cfg | PROP_ssa ), /* properties_required */
3926 0, /* properties_provided */
3927 0, /* properties_destroyed */
3928 0, /* todo_flags_start */
3929 0, /* todo_flags_finish */
3932 class pass_strlen : public gimple_opt_pass
3934 public:
3935 pass_strlen (gcc::context *ctxt)
3936 : gimple_opt_pass (pass_data_strlen, ctxt)
3939 /* opt_pass methods: */
3940 virtual bool gate (function *) { return flag_optimize_strlen != 0; }
3941 virtual unsigned int execute (function *);
3943 }; // class pass_strlen
3945 unsigned int
3946 pass_strlen::execute (function *fun)
3948 gcc_assert (!strlen_to_stridx);
3949 if (warn_stringop_overflow || warn_stringop_truncation)
3950 strlen_to_stridx = new hash_map<tree, stridx_strlenloc> ();
3952 ssa_ver_to_stridx.safe_grow_cleared (num_ssa_names);
3953 max_stridx = 1;
3955 calculate_dominance_info (CDI_DOMINATORS);
3957 /* String length optimization is implemented as a walk of the dominator
3958 tree and a forward walk of statements within each block. */
3959 strlen_dom_walker walker (CDI_DOMINATORS);
3960 walker.walk (fun->cfg->x_entry_block_ptr);
3962 ssa_ver_to_stridx.release ();
3963 strinfo_pool.release ();
3964 if (decl_to_stridxlist_htab)
3966 obstack_free (&stridx_obstack, NULL);
3967 delete decl_to_stridxlist_htab;
3968 decl_to_stridxlist_htab = NULL;
3970 laststmt.stmt = NULL;
3971 laststmt.len = NULL_TREE;
3972 laststmt.stridx = 0;
3974 if (strlen_to_stridx)
3976 strlen_to_stridx->empty ();
3977 delete strlen_to_stridx;
3978 strlen_to_stridx = NULL;
3981 return walker.m_cleanup_cfg ? TODO_cleanup_cfg : 0;
3984 } // anon namespace
3986 gimple_opt_pass *
3987 make_pass_strlen (gcc::context *ctxt)
3989 return new pass_strlen (ctxt);