Add assember CFI directives to millicode division and remainder routines.
[official-gcc.git] / gcc / gimple-ssa-sprintf.cc
blob18975708d2c24c5c51926b5711d8e2a2ab136e55
1 /* Copyright (C) 2016-2023 Free Software Foundation, Inc.
2 Contributed by Martin Sebor <msebor@redhat.com>.
4 This file is part of GCC.
6 GCC is free software; you can redistribute it and/or modify it under
7 the terms of the GNU General Public License as published by the Free
8 Software Foundation; either version 3, or (at your option) any later
9 version.
11 GCC is distributed in the hope that it will be useful, but WITHOUT ANY
12 WARRANTY; without even the implied warranty of MERCHANTABILITY or
13 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
14 for more details.
16 You should have received a copy of the GNU General Public License
17 along with GCC; see the file COPYING3. If not see
18 <http://www.gnu.org/licenses/>. */
20 /* This file implements the printf-return-value pass. The pass does
21 two things: 1) it analyzes calls to formatted output functions like
22 sprintf looking for possible buffer overflows and calls to bounded
23 functions like snprintf for early truncation (and under the control
24 of the -Wformat-length option issues warnings), and 2) under the
25 control of the -fprintf-return-value option it folds the return
26 value of safe calls into constants, making it possible to eliminate
27 code that depends on the value of those constants.
29 For all functions (bounded or not) the pass uses the size of the
30 destination object. That means that it will diagnose calls to
31 snprintf not on the basis of the size specified by the function's
32 second argument but rather on the basis of the size the first
33 argument points to (if possible). For bound-checking built-ins
34 like __builtin___snprintf_chk the pass uses the size typically
35 determined by __builtin_object_size and passed to the built-in
36 by the Glibc inline wrapper.
38 The pass handles all forms standard sprintf format directives,
39 including character, integer, floating point, pointer, and strings,
40 with the standard C flags, widths, and precisions. For integers
41 and strings it computes the length of output itself. For floating
42 point it uses MPFR to format known constants with up and down
43 rounding and uses the resulting range of output lengths. For
44 strings it uses the length of string literals and the sizes of
45 character arrays that a character pointer may point to as a bound
46 on the longest string. */
48 #include "config.h"
49 #include "system.h"
50 #include "coretypes.h"
51 #include "backend.h"
52 #include "tree.h"
53 #include "gimple.h"
54 #include "tree-pass.h"
55 #include "ssa.h"
56 #include "gimple-iterator.h"
57 #include "gimple-fold.h"
58 #include "gimple-pretty-print.h"
59 #include "diagnostic-core.h"
60 #include "fold-const.h"
61 #include "tree-ssa.h"
62 #include "tree-object-size.h"
63 #include "tree-cfg.h"
64 #include "tree-ssa-propagate.h"
65 #include "calls.h"
66 #include "cfgloop.h"
67 #include "tree-scalar-evolution.h"
68 #include "tree-ssa-loop.h"
69 #include "intl.h"
70 #include "langhooks.h"
72 #include "attribs.h"
73 #include "builtins.h"
74 #include "pointer-query.h"
75 #include "stor-layout.h"
77 #include "realmpfr.h"
78 #include "target.h"
80 #include "cpplib.h"
81 #include "input.h"
82 #include "toplev.h"
83 #include "substring-locations.h"
84 #include "diagnostic.h"
85 #include "domwalk.h"
86 #include "alloc-pool.h"
87 #include "vr-values.h"
88 #include "tree-ssa-strlen.h"
89 #include "tree-dfa.h"
91 /* The likely worst case value of MB_LEN_MAX for the target, large enough
92 for UTF-8. Ideally, this would be obtained by a target hook if it were
93 to be used for optimization but it's good enough as is for warnings. */
94 #define target_mb_len_max() 6
96 /* The maximum number of bytes a single non-string directive can result
97 in. This is the result of printf("%.*Lf", INT_MAX, -LDBL_MAX) for
98 LDBL_MAX_10_EXP of 4932. */
99 #define IEEE_MAX_10_EXP 4932
100 #define target_dir_max() (target_int_max () + IEEE_MAX_10_EXP + 2)
102 namespace {
104 /* Set to the warning level for the current function which is equal
105 either to warn_format_trunc for bounded functions or to
106 warn_format_overflow otherwise. */
108 static int warn_level;
110 /* The minimum, maximum, likely, and unlikely maximum number of bytes
111 of output either a formatting function or an individual directive
112 can result in. */
114 struct result_range
116 /* The absolute minimum number of bytes. The result of a successful
117 conversion is guaranteed to be no less than this. (An erroneous
118 conversion can be indicated by MIN > HOST_WIDE_INT_MAX.) */
119 unsigned HOST_WIDE_INT min;
120 /* The likely maximum result that is used in diagnostics. In most
121 cases MAX is the same as the worst case UNLIKELY result. */
122 unsigned HOST_WIDE_INT max;
123 /* The likely result used to trigger diagnostics. For conversions
124 that result in a range of bytes [MIN, MAX], LIKELY is somewhere
125 in that range. */
126 unsigned HOST_WIDE_INT likely;
127 /* In rare cases (e.g., for multibyte characters) UNLIKELY gives
128 the worst cases maximum result of a directive. In most cases
129 UNLIKELY == MAX. UNLIKELY is used to control the return value
130 optimization but not in diagnostics. */
131 unsigned HOST_WIDE_INT unlikely;
134 /* Return the value of INT_MIN for the target. */
136 static inline HOST_WIDE_INT
137 target_int_min ()
139 return tree_to_shwi (TYPE_MIN_VALUE (integer_type_node));
142 /* Return the value of INT_MAX for the target. */
144 static inline unsigned HOST_WIDE_INT
145 target_int_max ()
147 return tree_to_uhwi (TYPE_MAX_VALUE (integer_type_node));
150 /* Return the value of SIZE_MAX for the target. */
152 static inline unsigned HOST_WIDE_INT
153 target_size_max ()
155 return tree_to_uhwi (TYPE_MAX_VALUE (size_type_node));
158 /* A straightforward mapping from the execution character set to the host
159 character set indexed by execution character. */
161 static char target_to_host_charmap[256];
163 /* Initialize a mapping from the execution character set to the host
164 character set. */
166 static bool
167 init_target_to_host_charmap ()
169 /* If the percent sign is non-zero the mapping has already been
170 initialized. */
171 if (target_to_host_charmap['%'])
172 return true;
174 /* Initialize the target_percent character (done elsewhere). */
175 if (!init_target_chars ())
176 return false;
178 /* The subset of the source character set used by printf conversion
179 specifications (strictly speaking, not all letters are used but
180 they are included here for the sake of simplicity). The dollar
181 sign must be included even though it's not in the basic source
182 character set. */
183 const char srcset[] = " 0123456789!\"#%&'()*+,-./:;<=>?[\\]^_{|}~$"
184 "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
186 /* Set the mapping for all characters to some ordinary value (i,e.,
187 not none used in printf conversion specifications) and overwrite
188 those that are used by conversion specifications with their
189 corresponding values. */
190 memset (target_to_host_charmap + 1, '?', sizeof target_to_host_charmap - 1);
192 /* Are the two sets of characters the same? */
193 bool all_same_p = true;
195 for (const char *pc = srcset; *pc; ++pc)
197 /* Slice off the high end bits in case target characters are
198 signed. All values are expected to be non-nul, otherwise
199 there's a problem. */
200 if (unsigned char tc = lang_hooks.to_target_charset (*pc))
202 target_to_host_charmap[tc] = *pc;
203 if (tc != *pc)
204 all_same_p = false;
206 else
207 return false;
211 /* Set the first element to a non-zero value if the mapping
212 is 1-to-1, otherwise leave it clear (NUL is assumed to be
213 the same in both character sets). */
214 target_to_host_charmap[0] = all_same_p;
216 return true;
219 /* Return the host source character corresponding to the character
220 CH in the execution character set if one exists, or some innocuous
221 (non-special, non-nul) source character otherwise. */
223 static inline unsigned char
224 target_to_host (unsigned char ch)
226 return target_to_host_charmap[ch];
229 /* Convert an initial substring of the string TARGSTR consisting of
230 characters in the execution character set into a string in the
231 source character set on the host and store up to HOSTSZ characters
232 in the buffer pointed to by HOSTR. Return HOSTR. */
234 static const char*
235 target_to_host (char *hostr, size_t hostsz, const char *targstr)
237 /* Make sure the buffer is reasonably big. */
238 gcc_assert (hostsz > 4);
240 /* The interesting subset of source and execution characters are
241 the same so no conversion is necessary. However, truncate
242 overlong strings just like the translated strings are. */
243 if (target_to_host_charmap['\0'] == 1)
245 size_t len = strlen (targstr);
246 if (len >= hostsz)
248 memcpy (hostr, targstr, hostsz - 4);
249 strcpy (hostr + hostsz - 4, "...");
251 else
252 memcpy (hostr, targstr, len + 1);
253 return hostr;
256 /* Convert the initial substring of TARGSTR to the corresponding
257 characters in the host set, appending "..." if TARGSTR is too
258 long to fit. Using the static buffer assumes the function is
259 not called in between sequence points (which it isn't). */
260 for (char *ph = hostr; ; ++targstr)
262 *ph++ = target_to_host (*targstr);
263 if (!*targstr)
264 break;
266 if (size_t (ph - hostr) == hostsz)
268 strcpy (ph - 4, "...");
269 break;
273 return hostr;
276 /* Convert the sequence of decimal digits in the execution character
277 starting at *PS to a HOST_WIDE_INT, analogously to strtol. Return
278 the result and set *PS to one past the last converted character.
279 On range error set ERANGE to the digit that caused it. */
281 static inline HOST_WIDE_INT
282 target_strtowi (const char **ps, const char **erange)
284 unsigned HOST_WIDE_INT val = 0;
285 for ( ; ; ++*ps)
287 unsigned char c = target_to_host (**ps);
288 if (ISDIGIT (c))
290 c -= '0';
292 /* Check for overflow. */
293 if (val > ((unsigned HOST_WIDE_INT) HOST_WIDE_INT_MAX - c) / 10LU)
295 val = HOST_WIDE_INT_MAX;
296 *erange = *ps;
298 /* Skip the remaining digits. */
300 c = target_to_host (*++*ps);
301 while (ISDIGIT (c));
302 break;
304 else
305 val = val * 10 + c;
307 else
308 break;
311 return val;
314 /* Given FORMAT, set *PLOC to the source location of the format string
315 and return the format string if it is known or null otherwise. */
317 static const char*
318 get_format_string (tree format, location_t *ploc)
320 *ploc = EXPR_LOC_OR_LOC (format, input_location);
322 return c_getstr (format);
325 /* For convenience and brevity, shorter named entrypoints of
326 format_string_diagnostic_t::emit_warning_va and
327 format_string_diagnostic_t::emit_warning_n_va.
328 These have to be functions with the attribute so that exgettext
329 works properly. */
331 static bool
332 ATTRIBUTE_GCC_DIAG (5, 6)
333 fmtwarn (const substring_loc &fmt_loc, location_t param_loc,
334 const char *corrected_substring, opt_code opt,
335 const char *gmsgid, ...)
337 format_string_diagnostic_t diag (fmt_loc, NULL, param_loc, NULL,
338 corrected_substring);
339 va_list ap;
340 va_start (ap, gmsgid);
341 bool warned = diag.emit_warning_va (opt, gmsgid, &ap);
342 va_end (ap);
344 return warned;
347 static bool
348 ATTRIBUTE_GCC_DIAG (6, 8) ATTRIBUTE_GCC_DIAG (7, 8)
349 fmtwarn_n (const substring_loc &fmt_loc, location_t param_loc,
350 const char *corrected_substring, opt_code opt,
351 unsigned HOST_WIDE_INT n,
352 const char *singular_gmsgid, const char *plural_gmsgid, ...)
354 format_string_diagnostic_t diag (fmt_loc, NULL, param_loc, NULL,
355 corrected_substring);
356 va_list ap;
357 va_start (ap, plural_gmsgid);
358 bool warned = diag.emit_warning_n_va (opt, n, singular_gmsgid, plural_gmsgid,
359 &ap);
360 va_end (ap);
362 return warned;
365 /* Format length modifiers. */
367 enum format_lengths
369 FMT_LEN_none,
370 FMT_LEN_hh, // char argument
371 FMT_LEN_h, // short
372 FMT_LEN_l, // long
373 FMT_LEN_ll, // long long
374 FMT_LEN_L, // long double (and GNU long long)
375 FMT_LEN_z, // size_t
376 FMT_LEN_t, // ptrdiff_t
377 FMT_LEN_j // intmax_t
381 /* Description of the result of conversion either of a single directive
382 or the whole format string. */
384 class fmtresult
386 public:
387 /* Construct a FMTRESULT object with all counters initialized
388 to MIN. KNOWNRANGE is set when MIN is valid. */
389 fmtresult (unsigned HOST_WIDE_INT min = HOST_WIDE_INT_MAX)
390 : argmin (), argmax (), dst_offset (HOST_WIDE_INT_MIN), nonstr (),
391 knownrange (min < HOST_WIDE_INT_MAX),
392 mayfail (), nullp ()
394 range.min = min;
395 range.max = min;
396 range.likely = min;
397 range.unlikely = min;
400 /* Construct a FMTRESULT object with MIN, MAX, and LIKELY counters.
401 KNOWNRANGE is set when both MIN and MAX are valid. */
402 fmtresult (unsigned HOST_WIDE_INT min, unsigned HOST_WIDE_INT max,
403 unsigned HOST_WIDE_INT likely = HOST_WIDE_INT_MAX)
404 : argmin (), argmax (), dst_offset (HOST_WIDE_INT_MIN), nonstr (),
405 knownrange (min < HOST_WIDE_INT_MAX && max < HOST_WIDE_INT_MAX),
406 mayfail (), nullp ()
408 range.min = min;
409 range.max = max;
410 range.likely = max < likely ? min : likely;
411 range.unlikely = max;
414 /* Adjust result upward to reflect the RANGE of values the specified
415 width or precision is known to be in. */
416 fmtresult& adjust_for_width_or_precision (const HOST_WIDE_INT[2],
417 tree = NULL_TREE,
418 unsigned = 0, unsigned = 0);
420 /* Return the maximum number of decimal digits a value of TYPE
421 formats as on output. */
422 static unsigned type_max_digits (tree, int);
424 /* The range a directive's argument is in. */
425 tree argmin, argmax;
427 /* The starting offset into the destination of the formatted function
428 call of the %s argument that points into (aliases with) the same
429 destination array. */
430 HOST_WIDE_INT dst_offset;
432 /* The minimum and maximum number of bytes that a directive
433 results in on output for an argument in the range above. */
434 result_range range;
436 /* Non-nul when the argument of a string directive is not a nul
437 terminated string. */
438 tree nonstr;
440 /* True when the range above is obtained from a known value of
441 a directive's argument or its bounds and not the result of
442 heuristics that depend on warning levels. */
443 bool knownrange;
445 /* True for a directive that may fail (such as wide character
446 directives). */
447 bool mayfail;
449 /* True when the argument is a null pointer. */
450 bool nullp;
453 /* Adjust result upward to reflect the range ADJUST of values the
454 specified width or precision is known to be in. When non-null,
455 TYPE denotes the type of the directive whose result is being
456 adjusted, BASE gives the base of the directive (octal, decimal,
457 or hex), and ADJ denotes the additional adjustment to the LIKELY
458 counter that may need to be added when ADJUST is a range. */
460 fmtresult&
461 fmtresult::adjust_for_width_or_precision (const HOST_WIDE_INT adjust[2],
462 tree type /* = NULL_TREE */,
463 unsigned base /* = 0 */,
464 unsigned adj /* = 0 */)
466 bool minadjusted = false;
468 /* Adjust the minimum and likely counters. */
469 if (adjust[0] >= 0)
471 if (range.min < (unsigned HOST_WIDE_INT)adjust[0])
473 range.min = adjust[0];
474 minadjusted = true;
477 /* Adjust the likely counter. */
478 if (range.likely < range.min)
479 range.likely = range.min;
481 else if (adjust[0] == target_int_min ()
482 && (unsigned HOST_WIDE_INT)adjust[1] == target_int_max ())
483 knownrange = false;
485 /* Adjust the maximum counter. */
486 if (adjust[1] > 0)
488 if (range.max < (unsigned HOST_WIDE_INT)adjust[1])
490 range.max = adjust[1];
492 /* Set KNOWNRANGE if both the minimum and maximum have been
493 adjusted. Otherwise leave it at what it was before. */
494 knownrange = minadjusted;
498 if (warn_level > 1 && type)
500 /* For large non-constant width or precision whose range spans
501 the maximum number of digits produced by the directive for
502 any argument, set the likely number of bytes to be at most
503 the number digits plus other adjustment determined by the
504 caller (one for sign or two for the hexadecimal "0x"
505 prefix). */
506 unsigned dirdigs = type_max_digits (type, base);
507 if (adjust[0] < dirdigs && dirdigs < adjust[1]
508 && range.likely < dirdigs)
509 range.likely = dirdigs + adj;
511 else if (range.likely < (range.min ? range.min : 1))
513 /* Conservatively, set LIKELY to at least MIN but no less than
514 1 unless MAX is zero. */
515 range.likely = (range.min
516 ? range.min
517 : range.max && (range.max < HOST_WIDE_INT_MAX
518 || warn_level > 1) ? 1 : 0);
521 /* Finally adjust the unlikely counter to be at least as large as
522 the maximum. */
523 if (range.unlikely < range.max)
524 range.unlikely = range.max;
526 return *this;
529 /* Return the maximum number of digits a value of TYPE formats in
530 BASE on output, not counting base prefix . */
532 unsigned
533 fmtresult::type_max_digits (tree type, int base)
535 unsigned prec = TYPE_PRECISION (type);
536 switch (base)
538 case 2:
539 return prec;
540 case 8:
541 return (prec + 2) / 3;
542 case 10:
543 /* Decimal approximation: yields 3, 5, 10, and 20 for precision
544 of 8, 16, 32, and 64 bits. */
545 return prec * 301 / 1000 + 1;
546 case 16:
547 return prec / 4;
550 gcc_unreachable ();
553 static bool
554 get_int_range (tree, gimple *, HOST_WIDE_INT *, HOST_WIDE_INT *,
555 bool, HOST_WIDE_INT, range_query *);
557 struct call_info;
559 /* Description of a format directive. A directive is either a plain
560 string or a conversion specification that starts with '%'. */
562 struct directive
564 directive (const call_info *inf, unsigned dno)
565 : info (inf), dirno (dno), argno (), beg (), len (), flags (),
566 width (), prec (), modifier (), specifier (), arg (), fmtfunc ()
569 /* Reference to the info structure describing the call that this
570 directive is a part of. */
571 const call_info *info;
573 /* The 1-based directive number (for debugging). */
574 unsigned dirno;
576 /* The zero-based argument number of the directive's argument ARG in
577 the function's argument list. */
578 unsigned argno;
580 /* The first character of the directive and its length. */
581 const char *beg;
582 size_t len;
584 /* A bitmap of flags, one for each character. */
585 unsigned flags[256 / sizeof (int)];
587 /* The range of values of the specified width, or -1 if not specified. */
588 HOST_WIDE_INT width[2];
589 /* The range of values of the specified precision, or -1 if not
590 specified. */
591 HOST_WIDE_INT prec[2];
593 /* Length modifier. */
594 format_lengths modifier;
596 /* Format specifier character. */
597 char specifier;
599 /* The argument of the directive or null when the directive doesn't
600 take one or when none is available (such as for vararg functions). */
601 tree arg;
603 /* Format conversion function that given a directive and an argument
604 returns the formatting result. */
605 fmtresult (*fmtfunc) (const directive &, tree, pointer_query &);
607 /* Return True when the format flag CHR has been used. */
608 bool get_flag (char chr) const
610 unsigned char c = chr & 0xff;
611 return (flags[c / (CHAR_BIT * sizeof *flags)]
612 & (1U << (c % (CHAR_BIT * sizeof *flags))));
615 /* Make a record of the format flag CHR having been used. */
616 void set_flag (char chr)
618 unsigned char c = chr & 0xff;
619 flags[c / (CHAR_BIT * sizeof *flags)]
620 |= (1U << (c % (CHAR_BIT * sizeof *flags)));
623 /* Reset the format flag CHR. */
624 void clear_flag (char chr)
626 unsigned char c = chr & 0xff;
627 flags[c / (CHAR_BIT * sizeof *flags)]
628 &= ~(1U << (c % (CHAR_BIT * sizeof *flags)));
631 /* Set both bounds of the width range to VAL. */
632 void set_width (HOST_WIDE_INT val)
634 width[0] = width[1] = val;
637 /* Set the width range according to ARG, with both bounds being
638 no less than 0. For a constant ARG set both bounds to its value
639 or 0, whichever is greater. For a non-constant ARG in some range
640 set width to its range adjusting each bound to -1 if it's less.
641 For an indeterminate ARG set width to [0, INT_MAX]. */
642 void set_width (tree arg, range_query *);
644 /* Set both bounds of the precision range to VAL. */
645 void set_precision (HOST_WIDE_INT val)
647 prec[0] = prec[1] = val;
650 /* Set the precision range according to ARG, with both bounds being
651 no less than -1. For a constant ARG set both bounds to its value
652 or -1 whichever is greater. For a non-constant ARG in some range
653 set precision to its range adjusting each bound to -1 if it's less.
654 For an indeterminate ARG set precision to [-1, INT_MAX]. */
655 void set_precision (tree arg, range_query *query);
657 /* Return true if both width and precision are known to be
658 either constant or in some range, false otherwise. */
659 bool known_width_and_precision () const
661 return ((width[1] < 0
662 || (unsigned HOST_WIDE_INT)width[1] <= target_int_max ())
663 && (prec[1] < 0
664 || (unsigned HOST_WIDE_INT)prec[1] < target_int_max ()));
668 /* The result of a call to a formatted function. */
670 struct format_result
672 format_result ()
673 : range (), aliases (), alias_count (), knownrange (), posunder4k (),
674 floating (), warned () { /* No-op. */ }
676 ~format_result ()
678 XDELETEVEC (aliases);
681 /* Range of characters written by the formatted function.
682 Setting the minimum to HOST_WIDE_INT_MAX disables all
683 length tracking for the remainder of the format string. */
684 result_range range;
686 struct alias_info
688 directive dir; /* The directive that aliases the destination. */
689 HOST_WIDE_INT offset; /* The offset at which it aliases it. */
690 result_range range; /* The raw result of the directive. */
693 /* An array of directives whose pointer argument aliases a part
694 of the destination object of the formatted function. */
695 alias_info *aliases;
696 unsigned alias_count;
698 /* True when the range above is obtained from known values of
699 directive arguments, or bounds on the amount of output such
700 as width and precision, and not the result of heuristics that
701 depend on warning levels. It's used to issue stricter diagnostics
702 in cases where strings of unknown lengths are bounded by the arrays
703 they are determined to refer to. KNOWNRANGE must not be used for
704 the return value optimization. */
705 bool knownrange;
707 /* True if no individual directive could fail or result in more than
708 4095 bytes of output (the total NUMBER_CHARS_{MIN,MAX} might be
709 greater). Implementations are not required to handle directives
710 that produce more than 4K bytes (leading to undefined behavior)
711 and so when one is found it disables the return value optimization.
712 Similarly, directives that can fail (such as wide character
713 directives) disable the optimization. */
714 bool posunder4k;
716 /* True when a floating point directive has been seen in the format
717 string. */
718 bool floating;
720 /* True when an intermediate result has caused a warning. Used to
721 avoid issuing duplicate warnings while finishing the processing
722 of a call. WARNED also disables the return value optimization. */
723 bool warned;
725 /* Preincrement the number of output characters by 1. */
726 format_result& operator++ ()
728 return *this += 1;
731 /* Postincrement the number of output characters by 1. */
732 format_result operator++ (int)
734 format_result prev (*this);
735 *this += 1;
736 return prev;
739 /* Increment the number of output characters by N. */
740 format_result& operator+= (unsigned HOST_WIDE_INT);
742 /* Add a directive to the sequence of those with potentially aliasing
743 arguments. */
744 void append_alias (const directive &, HOST_WIDE_INT, const result_range &);
746 private:
747 /* Not copyable or assignable. */
748 format_result (format_result&);
749 void operator= (format_result&);
752 format_result&
753 format_result::operator+= (unsigned HOST_WIDE_INT n)
755 gcc_assert (n < HOST_WIDE_INT_MAX);
757 if (range.min < HOST_WIDE_INT_MAX)
758 range.min += n;
760 if (range.max < HOST_WIDE_INT_MAX)
761 range.max += n;
763 if (range.likely < HOST_WIDE_INT_MAX)
764 range.likely += n;
766 if (range.unlikely < HOST_WIDE_INT_MAX)
767 range.unlikely += n;
769 return *this;
772 void
773 format_result::append_alias (const directive &d, HOST_WIDE_INT off,
774 const result_range &resrng)
776 unsigned cnt = alias_count + 1;
777 alias_info *ar = XNEWVEC (alias_info, cnt);
779 for (unsigned i = 0; i != alias_count; ++i)
780 ar[i] = aliases[i];
782 ar[alias_count].dir = d;
783 ar[alias_count].offset = off;
784 ar[alias_count].range = resrng;
786 XDELETEVEC (aliases);
788 alias_count = cnt;
789 aliases = ar;
792 /* Return the logarithm of X in BASE. */
794 static int
795 ilog (unsigned HOST_WIDE_INT x, int base)
797 int res = 0;
800 ++res;
801 x /= base;
802 } while (x);
803 return res;
806 /* Return the number of bytes resulting from converting into a string
807 the INTEGER_CST tree node X in BASE with a minimum of PREC digits.
808 PLUS indicates whether 1 for a plus sign should be added for positive
809 numbers, and PREFIX whether the length of an octal ('0') or hexadecimal
810 ('0x') or binary ('0b') prefix should be added for nonzero numbers.
811 Return -1 if X cannot be represented. */
813 static HOST_WIDE_INT
814 tree_digits (tree x, int base, HOST_WIDE_INT prec, bool plus, bool prefix)
816 unsigned HOST_WIDE_INT absval;
818 HOST_WIDE_INT res;
820 if (TYPE_UNSIGNED (TREE_TYPE (x)))
822 if (tree_fits_uhwi_p (x))
824 absval = tree_to_uhwi (x);
825 res = plus;
827 else
828 return -1;
830 else
832 if (tree_fits_shwi_p (x))
834 HOST_WIDE_INT i = tree_to_shwi (x);
835 if (HOST_WIDE_INT_MIN == i)
837 /* Avoid undefined behavior due to negating a minimum. */
838 absval = HOST_WIDE_INT_MAX;
839 res = 1;
841 else if (i < 0)
843 absval = -i;
844 res = 1;
846 else
848 absval = i;
849 res = plus;
852 else
853 return -1;
856 int ndigs = ilog (absval, base);
858 res += prec < ndigs ? ndigs : prec;
860 /* Adjust a non-zero value for the base prefix, either hexadecimal,
861 or, unless precision has resulted in a leading zero, also octal. */
862 if (prefix && absval)
864 if (base == 8 && prec <= ndigs)
865 res += 1;
866 else if (base == 16 || base == 2) /* 0x...(0X...) or 0b...(0B...). */
867 res += 2;
870 return res;
873 /* Description of a call to a formatted function. */
875 struct call_info
877 /* Function call statement. */
878 gimple *callstmt;
880 /* Function called. */
881 tree func;
883 /* Called built-in function code. */
884 built_in_function fncode;
886 /* The "origin" of the destination pointer argument, which is either
887 the DECL of the destination buffer being written into or a pointer
888 that points to it, plus some offset. */
889 tree dst_origin;
891 /* For a destination pointing to a struct array member, the offset of
892 the member. */
893 HOST_WIDE_INT dst_field;
895 /* The offset into the destination buffer. */
896 HOST_WIDE_INT dst_offset;
898 /* Format argument and format string extracted from it. */
899 tree format;
900 const char *fmtstr;
902 /* The location of the format argument. */
903 location_t fmtloc;
905 /* The destination object size for __builtin___xxx_chk functions
906 typically determined by __builtin_object_size, or -1 if unknown. */
907 unsigned HOST_WIDE_INT objsize;
909 /* Number of the first variable argument. */
910 unsigned HOST_WIDE_INT argidx;
912 /* True for functions like snprintf that specify the size of
913 the destination, false for others like sprintf that don't. */
914 bool bounded;
916 /* True for bounded functions like snprintf that specify a zero-size
917 buffer as a request to compute the size of output without actually
918 writing any. NOWRITE is cleared in response to the %n directive
919 which has side-effects similar to writing output. */
920 bool nowrite;
922 /* Return true if the called function's return value is used. */
923 bool retval_used () const
925 return gimple_get_lhs (callstmt);
928 /* Return the warning option corresponding to the called function. */
929 opt_code warnopt () const
931 return bounded ? OPT_Wformat_truncation_ : OPT_Wformat_overflow_;
934 /* Return true for calls to file formatted functions. */
935 bool is_file_func () const
937 return (fncode == BUILT_IN_FPRINTF
938 || fncode == BUILT_IN_FPRINTF_CHK
939 || fncode == BUILT_IN_FPRINTF_UNLOCKED
940 || fncode == BUILT_IN_VFPRINTF
941 || fncode == BUILT_IN_VFPRINTF_CHK);
944 /* Return true for calls to string formatted functions. */
945 bool is_string_func () const
947 return (fncode == BUILT_IN_SPRINTF
948 || fncode == BUILT_IN_SPRINTF_CHK
949 || fncode == BUILT_IN_SNPRINTF
950 || fncode == BUILT_IN_SNPRINTF_CHK
951 || fncode == BUILT_IN_VSPRINTF
952 || fncode == BUILT_IN_VSPRINTF_CHK
953 || fncode == BUILT_IN_VSNPRINTF
954 || fncode == BUILT_IN_VSNPRINTF_CHK);
958 void
959 directive::set_width (tree arg, range_query *query)
961 get_int_range (arg, info->callstmt, width, width + 1, true, 0, query);
964 void
965 directive::set_precision (tree arg, range_query *query)
967 get_int_range (arg, info->callstmt, prec, prec + 1, false, -1, query);
970 /* Return the result of formatting a no-op directive (such as '%n'). */
972 static fmtresult
973 format_none (const directive &, tree, pointer_query &)
975 fmtresult res (0);
976 return res;
979 /* Return the result of formatting the '%%' directive. */
981 static fmtresult
982 format_percent (const directive &, tree, pointer_query &)
984 fmtresult res (1);
985 return res;
989 /* Compute intmax_type_node and uintmax_type_node similarly to how
990 tree.cc builds size_type_node. */
992 static void
993 build_intmax_type_nodes (tree *pintmax, tree *puintmax)
995 if (strcmp (UINTMAX_TYPE, "unsigned int") == 0)
997 *pintmax = integer_type_node;
998 *puintmax = unsigned_type_node;
1000 else if (strcmp (UINTMAX_TYPE, "long unsigned int") == 0)
1002 *pintmax = long_integer_type_node;
1003 *puintmax = long_unsigned_type_node;
1005 else if (strcmp (UINTMAX_TYPE, "long long unsigned int") == 0)
1007 *pintmax = long_long_integer_type_node;
1008 *puintmax = long_long_unsigned_type_node;
1010 else
1012 for (int i = 0; i < NUM_INT_N_ENTS; i++)
1013 if (int_n_enabled_p[i])
1015 char name[50], altname[50];
1016 sprintf (name, "__int%d unsigned", int_n_data[i].bitsize);
1017 sprintf (altname, "__int%d__ unsigned", int_n_data[i].bitsize);
1019 if (strcmp (name, UINTMAX_TYPE) == 0
1020 || strcmp (altname, UINTMAX_TYPE) == 0)
1022 *pintmax = int_n_trees[i].signed_type;
1023 *puintmax = int_n_trees[i].unsigned_type;
1024 return;
1027 gcc_unreachable ();
1031 /* Determine the range [*PMIN, *PMAX] that the expression ARG is
1032 in and that is representable in type int.
1033 Return true when the range is a subrange of that of int.
1034 When ARG is null it is as if it had the full range of int.
1035 When ABSOLUTE is true the range reflects the absolute value of
1036 the argument. When ABSOLUTE is false, negative bounds of
1037 the determined range are replaced with NEGBOUND. */
1039 static bool
1040 get_int_range (tree arg, gimple *stmt,
1041 HOST_WIDE_INT *pmin, HOST_WIDE_INT *pmax,
1042 bool absolute, HOST_WIDE_INT negbound,
1043 range_query *query)
1045 /* The type of the result. */
1046 const_tree type = integer_type_node;
1048 bool knownrange = false;
1050 if (!arg)
1052 *pmin = tree_to_shwi (TYPE_MIN_VALUE (type));
1053 *pmax = tree_to_shwi (TYPE_MAX_VALUE (type));
1055 else if (TREE_CODE (arg) == INTEGER_CST
1056 && TYPE_PRECISION (TREE_TYPE (arg)) <= TYPE_PRECISION (type))
1058 /* For a constant argument return its value adjusted as specified
1059 by NEGATIVE and NEGBOUND and return true to indicate that the
1060 result is known. */
1061 *pmin = tree_fits_shwi_p (arg) ? tree_to_shwi (arg) : tree_to_uhwi (arg);
1062 *pmax = *pmin;
1063 knownrange = true;
1065 else
1067 /* True if the argument's range cannot be determined. */
1068 bool unknown = true;
1070 tree argtype = TREE_TYPE (arg);
1072 /* Ignore invalid arguments with greater precision that that
1073 of the expected type (e.g., in sprintf("%*i", 12LL, i)).
1074 They will have been detected and diagnosed by -Wformat and
1075 so it's not important to complicate this code to try to deal
1076 with them again. */
1077 if (TREE_CODE (arg) == SSA_NAME
1078 && INTEGRAL_TYPE_P (argtype)
1079 && TYPE_PRECISION (argtype) <= TYPE_PRECISION (type))
1081 /* Try to determine the range of values of the integer argument. */
1082 value_range vr;
1083 query->range_of_expr (vr, arg, stmt);
1085 if (!vr.undefined_p () && !vr.varying_p ())
1087 HOST_WIDE_INT type_min
1088 = (TYPE_UNSIGNED (argtype)
1089 ? tree_to_uhwi (TYPE_MIN_VALUE (argtype))
1090 : tree_to_shwi (TYPE_MIN_VALUE (argtype)));
1092 HOST_WIDE_INT type_max = tree_to_uhwi (TYPE_MAX_VALUE (argtype));
1094 tree type = TREE_TYPE (arg);
1095 tree tmin = wide_int_to_tree (type, vr.lower_bound ());
1096 tree tmax = wide_int_to_tree (type, vr.upper_bound ());
1097 *pmin = TREE_INT_CST_LOW (tmin);
1098 *pmax = TREE_INT_CST_LOW (tmax);
1100 if (*pmin < *pmax)
1102 /* Return true if the adjusted range is a subrange of
1103 the full range of the argument's type. *PMAX may
1104 be less than *PMIN when the argument is unsigned
1105 and its upper bound is in excess of TYPE_MAX. In
1106 that (invalid) case disregard the range and use that
1107 of the expected type instead. */
1108 knownrange = type_min < *pmin || *pmax < type_max;
1110 unknown = false;
1115 /* Handle an argument with an unknown range as if none had been
1116 provided. */
1117 if (unknown)
1118 return get_int_range (NULL_TREE, NULL, pmin, pmax, absolute,
1119 negbound, query);
1122 /* Adjust each bound as specified by ABSOLUTE and NEGBOUND. */
1123 if (absolute)
1125 if (*pmin < 0)
1127 if (*pmin == *pmax)
1128 *pmin = *pmax = -*pmin;
1129 else
1131 /* Make sure signed overlow is avoided. */
1132 gcc_assert (*pmin != HOST_WIDE_INT_MIN);
1134 HOST_WIDE_INT tmp = -*pmin;
1135 *pmin = 0;
1136 if (*pmax < tmp)
1137 *pmax = tmp;
1141 else if (*pmin < negbound)
1142 *pmin = negbound;
1144 return knownrange;
1147 /* With the range [*ARGMIN, *ARGMAX] of an integer directive's actual
1148 argument, due to the conversion from either *ARGMIN or *ARGMAX to
1149 the type of the directive's formal argument it's possible for both
1150 to result in the same number of bytes or a range of bytes that's
1151 less than the number of bytes that would result from formatting
1152 some other value in the range [*ARGMIN, *ARGMAX]. This can be
1153 determined by checking for the actual argument being in the range
1154 of the type of the directive. If it isn't it must be assumed to
1155 take on the full range of the directive's type.
1156 Return true when the range has been adjusted to the full range
1157 of DIRTYPE, and false otherwise. */
1159 static bool
1160 adjust_range_for_overflow (tree dirtype, tree *argmin, tree *argmax)
1162 tree argtype = TREE_TYPE (*argmin);
1163 unsigned argprec = TYPE_PRECISION (argtype);
1164 unsigned dirprec = TYPE_PRECISION (dirtype);
1166 /* If the actual argument and the directive's argument have the same
1167 precision and sign there can be no overflow and so there is nothing
1168 to adjust. */
1169 if (argprec == dirprec && TYPE_SIGN (argtype) == TYPE_SIGN (dirtype))
1170 return false;
1172 /* The logic below was inspired/lifted from the CONVERT_EXPR_CODE_P
1173 branch in the extract_range_from_unary_expr function in tree-vrp.cc. */
1175 if (TREE_CODE (*argmin) == INTEGER_CST
1176 && TREE_CODE (*argmax) == INTEGER_CST
1177 && (dirprec >= argprec
1178 || integer_zerop (int_const_binop (RSHIFT_EXPR,
1179 int_const_binop (MINUS_EXPR,
1180 *argmax,
1181 *argmin),
1182 size_int (dirprec)))))
1184 *argmin = force_fit_type (dirtype, wi::to_widest (*argmin), 0, false);
1185 *argmax = force_fit_type (dirtype, wi::to_widest (*argmax), 0, false);
1187 /* If *ARGMIN is still less than *ARGMAX the conversion above
1188 is safe. Otherwise, it has overflowed and would be unsafe. */
1189 if (tree_int_cst_le (*argmin, *argmax))
1190 return false;
1193 *argmin = TYPE_MIN_VALUE (dirtype);
1194 *argmax = TYPE_MAX_VALUE (dirtype);
1195 return true;
1198 /* Return a range representing the minimum and maximum number of bytes
1199 that the format directive DIR will output for any argument given
1200 the WIDTH and PRECISION (extracted from DIR). This function is
1201 used when the directive argument or its value isn't known. */
1203 static fmtresult
1204 format_integer (const directive &dir, tree arg, pointer_query &ptr_qry)
1206 tree intmax_type_node;
1207 tree uintmax_type_node;
1209 /* Base to format the number in. */
1210 int base;
1212 /* True when a conversion is preceded by a prefix indicating the base
1213 of the argument (octal or hexadecimal). */
1214 const bool maybebase = dir.get_flag ('#');
1216 /* True when a signed conversion is preceded by a sign or space. */
1217 bool maybesign = false;
1219 /* True for signed conversions (i.e., 'd' and 'i'). */
1220 bool sign = false;
1222 switch (dir.specifier)
1224 case 'd':
1225 case 'i':
1226 /* Space and '+' are only meaningful for signed conversions. */
1227 maybesign = dir.get_flag (' ') | dir.get_flag ('+');
1228 sign = true;
1229 base = 10;
1230 break;
1231 case 'u':
1232 base = 10;
1233 break;
1234 case 'b':
1235 case 'B':
1236 base = 2;
1237 break;
1238 case 'o':
1239 base = 8;
1240 break;
1241 case 'X':
1242 case 'x':
1243 base = 16;
1244 break;
1245 default:
1246 gcc_unreachable ();
1249 const unsigned adj = (sign | maybebase) + (base == 2 || base == 16);
1251 /* The type of the "formal" argument expected by the directive. */
1252 tree dirtype = NULL_TREE;
1254 /* Determine the expected type of the argument from the length
1255 modifier. */
1256 switch (dir.modifier)
1258 case FMT_LEN_none:
1259 if (dir.specifier == 'p')
1260 dirtype = ptr_type_node;
1261 else
1262 dirtype = sign ? integer_type_node : unsigned_type_node;
1263 break;
1265 case FMT_LEN_h:
1266 dirtype = sign ? short_integer_type_node : short_unsigned_type_node;
1267 break;
1269 case FMT_LEN_hh:
1270 dirtype = sign ? signed_char_type_node : unsigned_char_type_node;
1271 break;
1273 case FMT_LEN_l:
1274 dirtype = sign ? long_integer_type_node : long_unsigned_type_node;
1275 break;
1277 case FMT_LEN_L:
1278 case FMT_LEN_ll:
1279 dirtype = (sign
1280 ? long_long_integer_type_node
1281 : long_long_unsigned_type_node);
1282 break;
1284 case FMT_LEN_z:
1285 dirtype = signed_or_unsigned_type_for (!sign, size_type_node);
1286 break;
1288 case FMT_LEN_t:
1289 dirtype = signed_or_unsigned_type_for (!sign, ptrdiff_type_node);
1290 break;
1292 case FMT_LEN_j:
1293 build_intmax_type_nodes (&intmax_type_node, &uintmax_type_node);
1294 dirtype = sign ? intmax_type_node : uintmax_type_node;
1295 break;
1297 default:
1298 return fmtresult ();
1301 /* The type of the argument to the directive, either deduced from
1302 the actual non-constant argument if one is known, or from
1303 the directive itself when none has been provided because it's
1304 a va_list. */
1305 tree argtype = NULL_TREE;
1307 if (!arg)
1309 /* When the argument has not been provided, use the type of
1310 the directive's argument as an approximation. This will
1311 result in false positives for directives like %i with
1312 arguments with smaller precision (such as short or char). */
1313 argtype = dirtype;
1315 else if (TREE_CODE (arg) == INTEGER_CST)
1317 /* When a constant argument has been provided use its value
1318 rather than type to determine the length of the output. */
1319 fmtresult res;
1321 if ((dir.prec[0] <= 0 && dir.prec[1] >= 0) && integer_zerop (arg))
1323 /* As a special case, a precision of zero with a zero argument
1324 results in zero bytes except in base 8 when the '#' flag is
1325 specified, and for signed conversions in base 8 and 10 when
1326 either the space or '+' flag has been specified and it results
1327 in just one byte (with width having the normal effect). This
1328 must extend to the case of a specified precision with
1329 an unknown value because it can be zero. */
1330 res.range.min = ((base == 8 && dir.get_flag ('#')) || maybesign);
1331 if (res.range.min == 0 && dir.prec[0] != dir.prec[1])
1333 res.range.max = 1;
1334 res.range.likely = 1;
1336 else
1338 res.range.max = res.range.min;
1339 res.range.likely = res.range.min;
1342 else
1344 /* Convert the argument to the type of the directive. */
1345 arg = fold_convert (dirtype, arg);
1347 res.range.min = tree_digits (arg, base, dir.prec[0],
1348 maybesign, maybebase);
1349 if (dir.prec[0] == dir.prec[1])
1350 res.range.max = res.range.min;
1351 else
1352 res.range.max = tree_digits (arg, base, dir.prec[1],
1353 maybesign, maybebase);
1354 res.range.likely = res.range.min;
1355 res.knownrange = true;
1358 res.range.unlikely = res.range.max;
1360 /* Bump up the counters if WIDTH is greater than LEN. */
1361 res.adjust_for_width_or_precision (dir.width, dirtype, base, adj);
1362 /* Bump up the counters again if PRECision is greater still. */
1363 res.adjust_for_width_or_precision (dir.prec, dirtype, base, adj);
1365 return res;
1367 else if (INTEGRAL_TYPE_P (TREE_TYPE (arg))
1368 || TREE_CODE (TREE_TYPE (arg)) == POINTER_TYPE)
1369 /* Determine the type of the provided non-constant argument. */
1370 argtype = TREE_TYPE (arg);
1371 else
1372 /* Don't bother with invalid arguments since they likely would
1373 have already been diagnosed, and disable any further checking
1374 of the format string by returning [-1, -1]. */
1375 return fmtresult ();
1377 fmtresult res;
1379 /* Using either the range the non-constant argument is in, or its
1380 type (either "formal" or actual), create a range of values that
1381 constrain the length of output given the warning level. */
1382 tree argmin = NULL_TREE;
1383 tree argmax = NULL_TREE;
1385 if (arg
1386 && TREE_CODE (arg) == SSA_NAME
1387 && INTEGRAL_TYPE_P (argtype))
1389 /* Try to determine the range of values of the integer argument
1390 (range information is not available for pointers). */
1391 value_range vr;
1392 ptr_qry.rvals->range_of_expr (vr, arg, dir.info->callstmt);
1394 if (!vr.varying_p () && !vr.undefined_p ())
1396 argmin = wide_int_to_tree (TREE_TYPE (arg), vr.lower_bound ());
1397 argmax = wide_int_to_tree (TREE_TYPE (arg), vr.upper_bound ());
1399 /* Set KNOWNRANGE if the argument is in a known subrange
1400 of the directive's type and neither width nor precision
1401 is unknown. (KNOWNRANGE may be reset below). */
1402 res.knownrange
1403 = ((!tree_int_cst_equal (TYPE_MIN_VALUE (dirtype), argmin)
1404 || !tree_int_cst_equal (TYPE_MAX_VALUE (dirtype), argmax))
1405 && dir.known_width_and_precision ());
1407 res.argmin = argmin;
1408 res.argmax = argmax;
1410 else
1412 /* The argument here may be the result of promoting the actual
1413 argument to int. Try to determine the type of the actual
1414 argument before promotion and narrow down its range that
1415 way. */
1416 gimple *def = SSA_NAME_DEF_STMT (arg);
1417 if (is_gimple_assign (def))
1419 tree_code code = gimple_assign_rhs_code (def);
1420 if (code == INTEGER_CST)
1422 arg = gimple_assign_rhs1 (def);
1423 return format_integer (dir, arg, ptr_qry);
1426 if (code == NOP_EXPR)
1428 tree type = TREE_TYPE (gimple_assign_rhs1 (def));
1429 if (INTEGRAL_TYPE_P (type)
1430 || TREE_CODE (type) == POINTER_TYPE)
1431 argtype = type;
1437 if (!argmin)
1439 if (TREE_CODE (argtype) == POINTER_TYPE)
1441 argmin = build_int_cst (pointer_sized_int_node, 0);
1442 argmax = build_all_ones_cst (pointer_sized_int_node);
1444 else
1446 argmin = TYPE_MIN_VALUE (argtype);
1447 argmax = TYPE_MAX_VALUE (argtype);
1451 /* Clear KNOWNRANGE if the range has been adjusted to the maximum
1452 of the directive. If it has been cleared then since ARGMIN and/or
1453 ARGMAX have been adjusted also adjust the corresponding ARGMIN and
1454 ARGMAX in the result to include in diagnostics. */
1455 if (adjust_range_for_overflow (dirtype, &argmin, &argmax))
1457 res.knownrange = false;
1458 res.argmin = argmin;
1459 res.argmax = argmax;
1462 /* Recursively compute the minimum and maximum from the known range. */
1463 if (TYPE_UNSIGNED (dirtype) || tree_int_cst_sgn (argmin) >= 0)
1465 /* For unsigned conversions/directives or signed when
1466 the minimum is positive, use the minimum and maximum to compute
1467 the shortest and longest output, respectively. */
1468 res.range.min = format_integer (dir, argmin, ptr_qry).range.min;
1469 res.range.max = format_integer (dir, argmax, ptr_qry).range.max;
1471 else if (tree_int_cst_sgn (argmax) < 0)
1473 /* For signed conversions/directives if maximum is negative,
1474 use the minimum as the longest output and maximum as the
1475 shortest output. */
1476 res.range.min = format_integer (dir, argmax, ptr_qry).range.min;
1477 res.range.max = format_integer (dir, argmin, ptr_qry).range.max;
1479 else
1481 /* Otherwise, 0 is inside of the range and minimum negative. Use 0
1482 as the shortest output and for the longest output compute the
1483 length of the output of both minimum and maximum and pick the
1484 longer. */
1485 unsigned HOST_WIDE_INT max1
1486 = format_integer (dir, argmin, ptr_qry).range.max;
1487 unsigned HOST_WIDE_INT max2
1488 = format_integer (dir, argmax, ptr_qry).range.max;
1489 res.range.min
1490 = format_integer (dir, integer_zero_node, ptr_qry).range.min;
1491 res.range.max = MAX (max1, max2);
1494 /* If the range is known, use the maximum as the likely length. */
1495 if (res.knownrange)
1496 res.range.likely = res.range.max;
1497 else
1499 /* Otherwise, use the minimum. Except for the case where for %#x or
1500 %#o the minimum is just for a single value in the range (0) and
1501 for all other values it is something longer, like 0x1 or 01.
1502 Use the length for value 1 in that case instead as the likely
1503 length. */
1504 res.range.likely = res.range.min;
1505 if (maybebase
1506 && base != 10
1507 && (tree_int_cst_sgn (argmin) < 0 || tree_int_cst_sgn (argmax) > 0))
1509 if (res.range.min == 1)
1510 res.range.likely += base == 8 ? 1 : 2;
1511 else if (res.range.min == 2
1512 && (base == 16 || base == 2)
1513 && (dir.width[0] == 2 || dir.prec[0] == 2))
1514 ++res.range.likely;
1518 res.range.unlikely = res.range.max;
1519 res.adjust_for_width_or_precision (dir.width, dirtype, base, adj);
1520 res.adjust_for_width_or_precision (dir.prec, dirtype, base, adj);
1522 return res;
1525 /* Return the number of bytes that a format directive consisting of FLAGS,
1526 PRECision, format SPECification, and MPFR rounding specifier RNDSPEC,
1527 would result for argument X under ideal conditions (i.e., if PREC
1528 weren't excessive). MPFR 3.1 allocates large amounts of memory for
1529 values of PREC with large magnitude and can fail (see MPFR bug #21056).
1530 This function works around those problems. */
1532 static unsigned HOST_WIDE_INT
1533 get_mpfr_format_length (mpfr_ptr x, const char *flags, HOST_WIDE_INT prec,
1534 char spec, char rndspec)
1536 char fmtstr[40];
1538 HOST_WIDE_INT len = strlen (flags);
1540 fmtstr[0] = '%';
1541 memcpy (fmtstr + 1, flags, len);
1542 memcpy (fmtstr + 1 + len, ".*R", 3);
1543 fmtstr[len + 4] = rndspec;
1544 fmtstr[len + 5] = spec;
1545 fmtstr[len + 6] = '\0';
1547 spec = TOUPPER (spec);
1548 if (spec == 'E' || spec == 'F')
1550 /* For %e, specify the precision explicitly since mpfr_sprintf
1551 does its own thing just to be different (see MPFR bug 21088). */
1552 if (prec < 0)
1553 prec = 6;
1555 else
1557 /* Avoid passing negative precisions with larger magnitude to MPFR
1558 to avoid exposing its bugs. (A negative precision is supposed
1559 to be ignored.) */
1560 if (prec < 0)
1561 prec = -1;
1564 HOST_WIDE_INT p = prec;
1566 if (spec == 'G' && !strchr (flags, '#'))
1568 /* For G/g without the pound flag, precision gives the maximum number
1569 of significant digits which is bounded by LDBL_MAX_10_EXP, or, for
1570 a 128 bit IEEE extended precision, 4932. Using twice as much here
1571 should be more than sufficient for any real format. */
1572 if ((IEEE_MAX_10_EXP * 2) < prec)
1573 prec = IEEE_MAX_10_EXP * 2;
1574 p = prec;
1576 else
1578 /* Cap precision arbitrarily at 1KB and add the difference
1579 (if any) to the MPFR result. */
1580 if (prec > 1024)
1581 p = 1024;
1584 len = mpfr_snprintf (NULL, 0, fmtstr, (int)p, x);
1586 /* Handle the unlikely (impossible?) error by returning more than
1587 the maximum dictated by the function's return type. */
1588 if (len < 0)
1589 return target_dir_max () + 1;
1591 /* Adjust the return value by the difference. */
1592 if (p < prec)
1593 len += prec - p;
1595 return len;
1598 /* Return the number of bytes to format using the format specifier
1599 SPEC and the precision PREC the largest value in the real floating
1600 TYPE. */
1602 static unsigned HOST_WIDE_INT
1603 format_floating_max (tree type, char spec, HOST_WIDE_INT prec)
1605 machine_mode mode = TYPE_MODE (type);
1607 /* IBM Extended mode. */
1608 if (MODE_COMPOSITE_P (mode))
1609 mode = DFmode;
1611 /* Get the real type format description for the target. */
1612 const real_format *rfmt = REAL_MODE_FORMAT (mode);
1613 REAL_VALUE_TYPE rv;
1615 real_maxval (&rv, 0, mode);
1617 /* Convert the GCC real value representation with the precision
1618 of the real type to the mpfr_t format with the GCC default
1619 round-to-nearest mode. */
1620 mpfr_t x;
1621 mpfr_init2 (x, rfmt->p);
1622 mpfr_from_real (x, &rv, MPFR_RNDN);
1624 /* Return a value one greater to account for the leading minus sign. */
1625 unsigned HOST_WIDE_INT r
1626 = 1 + get_mpfr_format_length (x, "", prec, spec, 'D');
1627 mpfr_clear (x);
1628 return r;
1631 /* Return a range representing the minimum and maximum number of bytes
1632 that the directive DIR will output for any argument. PREC gives
1633 the adjusted precision range to account for negative precisions
1634 meaning the default 6. This function is used when the directive
1635 argument or its value isn't known. */
1637 static fmtresult
1638 format_floating (const directive &dir, const HOST_WIDE_INT prec[2])
1640 tree type;
1642 switch (dir.modifier)
1644 case FMT_LEN_l:
1645 case FMT_LEN_none:
1646 type = double_type_node;
1647 break;
1649 case FMT_LEN_L:
1650 type = long_double_type_node;
1651 break;
1653 case FMT_LEN_ll:
1654 type = long_double_type_node;
1655 break;
1657 default:
1658 return fmtresult ();
1661 /* The minimum and maximum number of bytes produced by the directive. */
1662 fmtresult res;
1664 /* The minimum output as determined by flags. It's always at least 1.
1665 When plus or space are set the output is preceded by either a sign
1666 or a space. */
1667 unsigned flagmin = (1 /* for the first digit */
1668 + (dir.get_flag ('+') | dir.get_flag (' ')));
1670 /* The minimum is 3 for "inf" and "nan" for all specifiers, plus 1
1671 for the plus sign/space with the '+' and ' ' flags, respectively,
1672 unless reduced below. */
1673 res.range.min = 2 + flagmin;
1675 /* When the pound flag is set the decimal point is included in output
1676 regardless of precision. Whether or not a decimal point is included
1677 otherwise depends on the specification and precision. */
1678 bool radix = dir.get_flag ('#');
1680 switch (dir.specifier)
1682 case 'A':
1683 case 'a':
1685 HOST_WIDE_INT minprec = 6 + !radix /* decimal point */;
1686 if (dir.prec[0] <= 0)
1687 minprec = 0;
1688 else if (dir.prec[0] > 0)
1689 minprec = dir.prec[0] + !radix /* decimal point */;
1691 res.range.likely = (2 /* 0x */
1692 + flagmin
1693 + radix
1694 + minprec
1695 + 3 /* p+0 */);
1697 res.range.max = format_floating_max (type, 'a', prec[1]);
1699 /* The unlikely maximum accounts for the longest multibyte
1700 decimal point character. */
1701 res.range.unlikely = res.range.max;
1702 if (dir.prec[1] > 0)
1703 res.range.unlikely += target_mb_len_max () - 1;
1705 break;
1708 case 'E':
1709 case 'e':
1711 /* Minimum output attributable to precision and, when it's
1712 non-zero, decimal point. */
1713 HOST_WIDE_INT minprec = prec[0] ? prec[0] + !radix : 0;
1715 /* The likely minimum output is "[-+]1.234567e+00" regardless
1716 of the value of the actual argument. */
1717 res.range.likely = (flagmin
1718 + radix
1719 + minprec
1720 + 2 /* e+ */ + 2);
1722 res.range.max = format_floating_max (type, 'e', prec[1]);
1724 /* The unlikely maximum accounts for the longest multibyte
1725 decimal point character. */
1726 if (dir.prec[0] != dir.prec[1]
1727 || dir.prec[0] == -1 || dir.prec[0] > 0)
1728 res.range.unlikely = res.range.max + target_mb_len_max () -1;
1729 else
1730 res.range.unlikely = res.range.max;
1731 break;
1734 case 'F':
1735 case 'f':
1737 /* Minimum output attributable to precision and, when it's non-zero,
1738 decimal point. */
1739 HOST_WIDE_INT minprec = prec[0] ? prec[0] + !radix : 0;
1741 /* For finite numbers (i.e., not infinity or NaN) the lower bound
1742 when precision isn't specified is 8 bytes ("1.23456" since
1743 precision is taken to be 6). When precision is zero, the lower
1744 bound is 1 byte (e.g., "1"). Otherwise, when precision is greater
1745 than zero, then the lower bound is 2 plus precision (plus flags).
1746 But in all cases, the lower bound is no greater than 3. */
1747 unsigned HOST_WIDE_INT min = flagmin + radix + minprec;
1748 if (min < res.range.min)
1749 res.range.min = min;
1751 /* Compute the upper bound for -TYPE_MAX. */
1752 res.range.max = format_floating_max (type, 'f', prec[1]);
1754 /* The minimum output with unknown precision is a single byte
1755 (e.g., "0") but the more likely output is 3 bytes ("0.0"). */
1756 if (dir.prec[0] < 0 && dir.prec[1] > 0)
1757 res.range.likely = 3;
1758 else
1759 res.range.likely = min;
1761 /* The unlikely maximum accounts for the longest multibyte
1762 decimal point character. */
1763 if (dir.prec[0] != dir.prec[1]
1764 || dir.prec[0] == -1 || dir.prec[0] > 0)
1765 res.range.unlikely = res.range.max + target_mb_len_max () - 1;
1766 break;
1769 case 'G':
1770 case 'g':
1772 /* The %g output depends on precision and the exponent of
1773 the argument. Since the value of the argument isn't known
1774 the lower bound on the range of bytes (not counting flags
1775 or width) is 1 plus radix (i.e., either "0" or "0." for
1776 "%g" and "%#g", respectively, with a zero argument). */
1777 unsigned HOST_WIDE_INT min = flagmin + radix;
1778 if (min < res.range.min)
1779 res.range.min = min;
1781 char spec = 'g';
1782 HOST_WIDE_INT maxprec = dir.prec[1];
1783 if (radix && maxprec)
1785 /* When the pound flag (radix) is set, trailing zeros aren't
1786 trimmed and so the longest output is the same as for %e,
1787 except with precision minus 1 (as specified in C11). */
1788 spec = 'e';
1789 if (maxprec > 0)
1790 --maxprec;
1791 else if (maxprec < 0)
1792 maxprec = 5;
1794 else
1795 maxprec = prec[1];
1797 res.range.max = format_floating_max (type, spec, maxprec);
1799 /* The likely output is either the maximum computed above
1800 minus 1 (assuming the maximum is positive) when precision
1801 is known (or unspecified), or the same minimum as for %e
1802 (which is computed for a non-negative argument). Unlike
1803 for the other specifiers above the likely output isn't
1804 the minimum because for %g that's 1 which is unlikely. */
1805 if (dir.prec[1] < 0
1806 || (unsigned HOST_WIDE_INT)dir.prec[1] < target_int_max ())
1807 res.range.likely = res.range.max - 1;
1808 else
1810 HOST_WIDE_INT minprec = 6 + !radix /* decimal point */;
1811 res.range.likely = (flagmin
1812 + radix
1813 + minprec
1814 + 2 /* e+ */ + 2);
1817 /* The unlikely maximum accounts for the longest multibyte
1818 decimal point character. */
1819 res.range.unlikely = res.range.max + target_mb_len_max () - 1;
1820 break;
1823 default:
1824 return fmtresult ();
1827 /* Bump up the byte counters if WIDTH is greater. */
1828 res.adjust_for_width_or_precision (dir.width);
1829 return res;
1832 /* Return a range representing the minimum and maximum number of bytes
1833 that the directive DIR will write on output for the floating argument
1834 ARG. */
1836 static fmtresult
1837 format_floating (const directive &dir, tree arg, pointer_query &)
1839 HOST_WIDE_INT prec[] = { dir.prec[0], dir.prec[1] };
1840 tree type = (dir.modifier == FMT_LEN_L || dir.modifier == FMT_LEN_ll
1841 ? long_double_type_node : double_type_node);
1843 /* For an indeterminate precision the lower bound must be assumed
1844 to be zero. */
1845 if (TOUPPER (dir.specifier) == 'A')
1847 /* Get the number of fractional decimal digits needed to represent
1848 the argument without a loss of accuracy. */
1849 unsigned fmtprec
1850 = REAL_MODE_FORMAT (TYPE_MODE (type))->p;
1852 /* The precision of the IEEE 754 double format is 53.
1853 The precision of all other GCC binary double formats
1854 is 56 or less. */
1855 unsigned maxprec = fmtprec <= 56 ? 13 : 15;
1857 /* For %a, leave the minimum precision unspecified to let
1858 MFPR trim trailing zeros (as it and many other systems
1859 including Glibc happen to do) and set the maximum
1860 precision to reflect what it would be with trailing zeros
1861 present (as Solaris and derived systems do). */
1862 if (dir.prec[1] < 0)
1864 /* Both bounds are negative implies that precision has
1865 not been specified. */
1866 prec[0] = maxprec;
1867 prec[1] = -1;
1869 else if (dir.prec[0] < 0)
1871 /* With a negative lower bound and a non-negative upper
1872 bound set the minimum precision to zero and the maximum
1873 to the greater of the maximum precision (i.e., with
1874 trailing zeros present) and the specified upper bound. */
1875 prec[0] = 0;
1876 prec[1] = dir.prec[1] < maxprec ? maxprec : dir.prec[1];
1879 else if (dir.prec[0] < 0)
1881 if (dir.prec[1] < 0)
1883 /* A precision in a strictly negative range is ignored and
1884 the default of 6 is used instead. */
1885 prec[0] = prec[1] = 6;
1887 else
1889 /* For a precision in a partly negative range, the lower bound
1890 must be assumed to be zero and the new upper bound is the
1891 greater of 6 (the default precision used when the specified
1892 precision is negative) and the upper bound of the specified
1893 range. */
1894 prec[0] = 0;
1895 prec[1] = dir.prec[1] < 6 ? 6 : dir.prec[1];
1899 if (!arg
1900 || TREE_CODE (arg) != REAL_CST
1901 || !useless_type_conversion_p (type, TREE_TYPE (arg)))
1902 return format_floating (dir, prec);
1904 /* The minimum and maximum number of bytes produced by the directive. */
1905 fmtresult res;
1907 /* Get the real type format description for the target. */
1908 const REAL_VALUE_TYPE *rvp = TREE_REAL_CST_PTR (arg);
1909 const real_format *rfmt = REAL_MODE_FORMAT (TYPE_MODE (TREE_TYPE (arg)));
1911 if (!real_isfinite (rvp))
1913 /* The format for Infinity and NaN is "[-]inf"/"[-]infinity"
1914 and "[-]nan" with the choice being implementation-defined
1915 but not locale dependent. */
1916 bool sign = dir.get_flag ('+') || real_isneg (rvp);
1917 res.range.min = 3 + sign;
1919 res.range.likely = res.range.min;
1920 res.range.max = res.range.min;
1921 /* The unlikely maximum is "[-/+]infinity" or "[-/+][qs]nan".
1922 For NaN, the C/POSIX standards specify two formats:
1923 "[-/+]nan"
1925 "[-/+]nan(n-char-sequence)"
1926 No known printf implementation outputs the latter format but AIX
1927 outputs QNaN and SNaN for quiet and signalling NaN, respectively,
1928 so the unlikely maximum reflects that. */
1929 res.range.unlikely = sign + (real_isinf (rvp) ? 8 : 4);
1931 /* The range for infinity and NaN is known unless either width
1932 or precision is unknown. Width has the same effect regardless
1933 of whether the argument is finite. Precision is either ignored
1934 (e.g., Glibc) or can have an effect on the short vs long format
1935 such as inf/infinity (e.g., Solaris). */
1936 res.knownrange = dir.known_width_and_precision ();
1938 /* Adjust the range for width but ignore precision. */
1939 res.adjust_for_width_or_precision (dir.width);
1941 return res;
1944 char fmtstr [40];
1945 char *pfmt = fmtstr;
1947 /* Append flags. */
1948 for (const char *pf = "-+ #0"; *pf; ++pf)
1949 if (dir.get_flag (*pf))
1950 *pfmt++ = *pf;
1952 *pfmt = '\0';
1955 /* Set up an array to easily iterate over. */
1956 unsigned HOST_WIDE_INT* const minmax[] = {
1957 &res.range.min, &res.range.max
1960 for (int i = 0; i != ARRAY_SIZE (minmax); ++i)
1962 /* Convert the GCC real value representation with the precision
1963 of the real type to the mpfr_t format rounding down in the
1964 first iteration that computes the minimum and up in the second
1965 that computes the maximum. This order is arbitrary because
1966 rounding in either direction can result in longer output. */
1967 mpfr_t mpfrval;
1968 mpfr_init2 (mpfrval, rfmt->p);
1969 mpfr_from_real (mpfrval, rvp, i ? MPFR_RNDU : MPFR_RNDD);
1971 /* Use the MPFR rounding specifier to round down in the first
1972 iteration and then up. In most but not all cases this will
1973 result in the same number of bytes. */
1974 char rndspec = "DU"[i];
1976 /* Format it and store the result in the corresponding member
1977 of the result struct. */
1978 *minmax[i] = get_mpfr_format_length (mpfrval, fmtstr, prec[i],
1979 dir.specifier, rndspec);
1980 mpfr_clear (mpfrval);
1984 /* Make sure the minimum is less than the maximum (MPFR rounding
1985 in the call to mpfr_snprintf can result in the reverse. */
1986 if (res.range.max < res.range.min)
1988 unsigned HOST_WIDE_INT tmp = res.range.min;
1989 res.range.min = res.range.max;
1990 res.range.max = tmp;
1993 /* The range is known unless either width or precision is unknown. */
1994 res.knownrange = dir.known_width_and_precision ();
1996 /* For the same floating point constant, unless width or precision
1997 is unknown, use the longer output as the likely maximum since
1998 with round to nearest either is equally likely. Otherwise, when
1999 precision is unknown, use the greater of the minimum and 3 as
2000 the likely output (for "0.0" since zero precision is unlikely). */
2001 if (res.knownrange)
2002 res.range.likely = res.range.max;
2003 else if (res.range.min < 3
2004 && dir.prec[0] < 0
2005 && (unsigned HOST_WIDE_INT)dir.prec[1] == target_int_max ())
2006 res.range.likely = 3;
2007 else
2008 res.range.likely = res.range.min;
2010 res.range.unlikely = res.range.max;
2012 if (res.range.max > 2 && (prec[0] != 0 || prec[1] != 0))
2014 /* Unless the precision is zero output longer than 2 bytes may
2015 include the decimal point which must be a single character
2016 up to MB_LEN_MAX in length. This is overly conservative
2017 since in some conversions some constants result in no decimal
2018 point (e.g., in %g). */
2019 res.range.unlikely += target_mb_len_max () - 1;
2022 res.adjust_for_width_or_precision (dir.width);
2023 return res;
2026 /* Return a FMTRESULT struct set to the lengths of the shortest and longest
2027 strings referenced by the expression STR, or (-1, -1) when not known.
2028 Used by the format_string function below. */
2030 static fmtresult
2031 get_string_length (tree str, gimple *stmt, unsigned HOST_WIDE_INT max_size,
2032 unsigned eltsize, pointer_query &ptr_qry)
2034 if (!str)
2035 return fmtresult ();
2037 /* Try to determine the dynamic string length first.
2038 Set MAXBOUND to an arbitrary non-null non-integer node as a request
2039 to have it set to the length of the longest string in a PHI. */
2040 c_strlen_data lendata = { };
2041 lendata.maxbound = str;
2042 if (eltsize == 1)
2043 get_range_strlen_dynamic (str, stmt, &lendata, ptr_qry);
2044 else
2046 /* Determine the length of the shortest and longest string referenced
2047 by STR. Strings of unknown lengths are bounded by the sizes of
2048 arrays that subexpressions of STR may refer to. Pointers that
2049 aren't known to point any such arrays result in LENDATA.MAXLEN
2050 set to SIZE_MAX. */
2051 get_range_strlen (str, &lendata, eltsize);
2054 /* If LENDATA.MAXBOUND is not equal to .MINLEN it corresponds to the bound
2055 of the largest array STR refers to, if known, or it's set to SIZE_MAX
2056 otherwise. */
2058 /* Return the default result when nothing is known about the string. */
2059 if ((lendata.maxbound && !tree_fits_uhwi_p (lendata.maxbound))
2060 || !tree_fits_uhwi_p (lendata.maxlen))
2062 fmtresult res;
2063 res.nonstr = lendata.decl;
2064 return res;
2067 unsigned HOST_WIDE_INT lenmax = tree_to_uhwi (max_object_size ()) - 2;
2068 if (integer_zerop (lendata.minlen)
2069 && (!lendata.maxbound || lenmax <= tree_to_uhwi (lendata.maxbound))
2070 && lenmax <= tree_to_uhwi (lendata.maxlen))
2072 if (max_size > 0 && max_size < HOST_WIDE_INT_MAX)
2074 /* Adjust the conservative unknown/unbounded result if MAX_SIZE
2075 is valid. Set UNLIKELY to maximum in case MAX_SIZE refers
2076 to a subobject.
2077 TODO: This is overly conservative. Set UNLIKELY to the size
2078 of the outermost enclosing declared object. */
2079 fmtresult res (0, max_size - 1);
2080 res.nonstr = lendata.decl;
2081 res.range.likely = res.range.max;
2082 res.range.unlikely = HOST_WIDE_INT_MAX;
2083 return res;
2086 fmtresult res;
2087 res.nonstr = lendata.decl;
2088 return res;
2091 /* The minimum length of the string. */
2092 HOST_WIDE_INT min
2093 = (tree_fits_uhwi_p (lendata.minlen)
2094 ? tree_to_uhwi (lendata.minlen)
2095 : 0);
2097 /* The maximum length of the string; initially set to MAXBOUND which
2098 may be less than MAXLEN, but may be adjusted up below. */
2099 HOST_WIDE_INT max
2100 = (lendata.maxbound && tree_fits_uhwi_p (lendata.maxbound)
2101 ? tree_to_uhwi (lendata.maxbound)
2102 : HOST_WIDE_INT_M1U);
2104 /* True if either the maximum length is unknown or (conservatively)
2105 the array bound is less than the maximum length. That can happen
2106 when the length of the string is unknown but the array in which
2107 the string is stored is a member of a struct. The warning uses
2108 the size of the member as the upper bound but the optimization
2109 doesn't. The optimization could still use the size of
2110 enclosing object as the upper bound but that's not done here. */
2111 const bool unbounded
2112 = (integer_all_onesp (lendata.maxlen)
2113 || (lendata.maxbound
2114 && tree_int_cst_lt (lendata.maxbound, lendata.maxlen)));
2116 /* Set the max/likely counters to unbounded when a minimum is known
2117 but the maximum length isn't bounded. This implies that STR is
2118 a conditional expression involving a string of known length and
2119 an expression of unknown/unbounded length. */
2120 if (min
2121 && (unsigned HOST_WIDE_INT)min < HOST_WIDE_INT_M1U
2122 && unbounded)
2123 max = HOST_WIDE_INT_M1U;
2125 /* get_range_strlen() returns the target value of SIZE_MAX for
2126 strings of unknown length. Bump it up to HOST_WIDE_INT_M1U
2127 which may be bigger. */
2128 if ((unsigned HOST_WIDE_INT)min == target_size_max ())
2129 min = HOST_WIDE_INT_M1U;
2130 if ((unsigned HOST_WIDE_INT)max == target_size_max ())
2131 max = HOST_WIDE_INT_M1U;
2133 fmtresult res (min, max);
2134 res.nonstr = lendata.decl;
2136 /* Set RES.KNOWNRANGE to true if and only if all strings referenced
2137 by STR are known to be bounded (though not necessarily by their
2138 actual length but perhaps by their maximum possible length). */
2139 if (res.range.max < target_int_max ())
2141 res.knownrange = true;
2142 /* When the length of the longest string is known and not
2143 excessive use it as the likely length of the string(s). */
2144 res.range.likely = res.range.max;
2146 else
2148 /* When the upper bound is unknown (it can be zero or excessive)
2149 set the likely length to the greater of 1. If MAXBOUND is
2150 known, also reset the length of the lower bound to zero. */
2151 res.range.likely = res.range.min ? res.range.min : warn_level > 1;
2152 if (lendata.maxbound && !integer_all_onesp (lendata.maxbound))
2153 res.range.min = 0;
2156 res.range.unlikely = unbounded ? HOST_WIDE_INT_MAX : res.range.max;
2158 return res;
2161 /* Return the minimum and maximum number of characters formatted
2162 by the '%c' format directives and its wide character form for
2163 the argument ARG. ARG can be null (for functions such as
2164 vsprinf). */
2166 static fmtresult
2167 format_character (const directive &dir, tree arg, pointer_query &ptr_qry)
2169 fmtresult res;
2171 res.knownrange = true;
2173 if (dir.specifier == 'C'
2174 || dir.modifier == FMT_LEN_l)
2176 /* A wide character can result in as few as zero bytes. */
2177 res.range.min = 0;
2179 HOST_WIDE_INT min, max;
2180 if (get_int_range (arg, dir.info->callstmt, &min, &max, false, 0,
2181 ptr_qry.rvals))
2183 if (min == 0 && max == 0)
2185 /* The NUL wide character results in no bytes. */
2186 res.range.max = 0;
2187 res.range.likely = 0;
2188 res.range.unlikely = 0;
2190 else if (min >= 0 && min < 128)
2192 /* Be conservative if the target execution character set
2193 is not a 1-to-1 mapping to the source character set or
2194 if the source set is not ASCII. */
2195 bool one_2_one_ascii
2196 = (target_to_host_charmap[0] == 1 && target_to_host ('a') == 97);
2198 /* A wide character in the ASCII range most likely results
2199 in a single byte, and only unlikely in up to MB_LEN_MAX. */
2200 res.range.max = one_2_one_ascii ? 1 : target_mb_len_max ();;
2201 res.range.likely = 1;
2202 res.range.unlikely = target_mb_len_max ();
2203 res.mayfail = !one_2_one_ascii;
2205 else
2207 /* A wide character outside the ASCII range likely results
2208 in up to two bytes, and only unlikely in up to MB_LEN_MAX. */
2209 res.range.max = target_mb_len_max ();
2210 res.range.likely = 2;
2211 res.range.unlikely = res.range.max;
2212 /* Converting such a character may fail. */
2213 res.mayfail = true;
2216 else
2218 /* An unknown wide character is treated the same as a wide
2219 character outside the ASCII range. */
2220 res.range.max = target_mb_len_max ();
2221 res.range.likely = 2;
2222 res.range.unlikely = res.range.max;
2223 res.mayfail = true;
2226 else
2228 /* A plain '%c' directive. Its output is exactly 1. */
2229 res.range.min = res.range.max = 1;
2230 res.range.likely = res.range.unlikely = 1;
2231 res.knownrange = true;
2234 /* Bump up the byte counters if WIDTH is greater. */
2235 return res.adjust_for_width_or_precision (dir.width);
2238 /* If TYPE is an array or struct or union, increment *FLDOFF by the starting
2239 offset of the member that *OFF points into if one can be determined and
2240 set *FLDSIZE to its size in bytes and decrement *OFF by the same.
2241 Otherwise do nothing. */
2243 static void
2244 set_aggregate_size_and_offset (tree type, HOST_WIDE_INT *fldoff,
2245 HOST_WIDE_INT *fldsize, HOST_WIDE_INT *off)
2247 /* The byte offset of the most basic struct member the byte
2248 offset *OFF corresponds to, or for a (multidimensional)
2249 array member, the byte offset of the array element. */
2250 if (TREE_CODE (type) == ARRAY_TYPE
2251 && TREE_CODE (TREE_TYPE (type)) == ARRAY_TYPE)
2253 HOST_WIDE_INT index = 0, arrsize = 0;
2254 if (array_elt_at_offset (type, *off, &index, &arrsize))
2256 *fldoff += index;
2257 *fldsize = arrsize;
2259 /* Otherwise leave *FLDOFF et al. unchanged. */
2261 else if (RECORD_OR_UNION_TYPE_P (type))
2263 HOST_WIDE_INT index = 0;
2264 tree sub = field_at_offset (type, NULL_TREE, *off, &index);
2265 if (sub)
2267 tree subsize = DECL_SIZE_UNIT (sub);
2268 if (*fldsize < HOST_WIDE_INT_MAX
2269 && subsize
2270 && tree_fits_uhwi_p (subsize))
2271 *fldsize = tree_to_uhwi (subsize);
2272 else
2273 *fldsize = HOST_WIDE_INT_MAX;
2274 *fldoff += index;
2275 *off -= index;
2277 /* Otherwise leave *FLDOFF et al. unchanged. */
2281 /* For an expression X of pointer type, recursively try to find its origin
2282 (either object DECL or pointer such as PARM_DECL) Y and return such a Y.
2283 When X refers to an array element or struct member, set *FLDOFF to
2284 the offset of the element or member from the beginning of the "most
2285 derived" object and *FLDSIZE to its size. When nonnull, set *OFF to
2286 the overall offset from the beginning of the object so that
2287 *FLDOFF <= *OFF. */
2289 static tree
2290 get_origin_and_offset_r (tree x, HOST_WIDE_INT *fldoff, HOST_WIDE_INT *fldsize,
2291 HOST_WIDE_INT *off)
2293 HOST_WIDE_INT sizebuf = -1;
2294 if (!fldsize)
2295 fldsize = &sizebuf;
2297 if (DECL_P (x))
2299 /* Set the size if it hasn't been set yet. */
2300 if (tree size = DECL_SIZE_UNIT (x))
2301 if (*fldsize < 0 && tree_fits_shwi_p (size))
2302 *fldsize = tree_to_shwi (size);
2303 return x;
2306 switch (TREE_CODE (x))
2308 case ADDR_EXPR:
2309 x = TREE_OPERAND (x, 0);
2310 return get_origin_and_offset_r (x, fldoff, fldsize, off);
2312 case ARRAY_REF:
2314 tree sub = TREE_OPERAND (x, 1);
2315 unsigned HOST_WIDE_INT idx =
2316 tree_fits_uhwi_p (sub) ? tree_to_uhwi (sub) : HOST_WIDE_INT_MAX;
2318 tree elsz = array_ref_element_size (x);
2319 unsigned HOST_WIDE_INT elbytes =
2320 tree_fits_shwi_p (elsz) ? tree_to_shwi (elsz) : HOST_WIDE_INT_MAX;
2322 unsigned HOST_WIDE_INT byteoff = idx * elbytes;
2324 if (byteoff < HOST_WIDE_INT_MAX
2325 && elbytes < HOST_WIDE_INT_MAX
2326 && (elbytes == 0 || byteoff / elbytes == idx))
2328 /* For in-bounds constant offsets into constant-sized arrays
2329 bump up *OFF, and for what's likely arrays or structs of
2330 arrays, also *FLDOFF, as necessary. */
2331 if (off)
2332 *off += byteoff;
2333 if (elbytes > 1)
2334 *fldoff += byteoff;
2336 else
2337 *fldoff = HOST_WIDE_INT_MAX;
2339 x = TREE_OPERAND (x, 0);
2340 return get_origin_and_offset_r (x, fldoff, fldsize, off);
2343 case MEM_REF:
2344 if (off)
2346 tree offset = TREE_OPERAND (x, 1);
2347 *off = (tree_fits_uhwi_p (offset)
2348 ? tree_to_uhwi (offset) : HOST_WIDE_INT_MAX);
2351 x = TREE_OPERAND (x, 0);
2353 if (off)
2355 tree xtype
2356 = (TREE_CODE (x) == ADDR_EXPR
2357 ? TREE_TYPE (TREE_OPERAND (x, 0)) : TREE_TYPE (TREE_TYPE (x)));
2359 set_aggregate_size_and_offset (xtype, fldoff, fldsize, off);
2362 return get_origin_and_offset_r (x, fldoff, fldsize, nullptr);
2364 case COMPONENT_REF:
2366 tree foff = component_ref_field_offset (x);
2367 tree fld = TREE_OPERAND (x, 1);
2368 if (!tree_fits_shwi_p (foff)
2369 || !tree_fits_shwi_p (DECL_FIELD_BIT_OFFSET (fld)))
2370 return x;
2371 *fldoff += (tree_to_shwi (foff)
2372 + (tree_to_shwi (DECL_FIELD_BIT_OFFSET (fld))
2373 / BITS_PER_UNIT));
2375 get_origin_and_offset_r (fld, fldoff, fldsize, off);
2376 x = TREE_OPERAND (x, 0);
2377 return get_origin_and_offset_r (x, fldoff, nullptr, off);
2380 case SSA_NAME:
2382 gimple *def = SSA_NAME_DEF_STMT (x);
2383 if (is_gimple_assign (def))
2385 tree_code code = gimple_assign_rhs_code (def);
2386 if (code == ADDR_EXPR)
2388 x = gimple_assign_rhs1 (def);
2389 return get_origin_and_offset_r (x, fldoff, fldsize, off);
2392 if (code == POINTER_PLUS_EXPR)
2394 tree offset = gimple_assign_rhs2 (def);
2395 if (off && tree_fits_uhwi_p (offset))
2396 *off = tree_to_uhwi (offset);
2398 x = gimple_assign_rhs1 (def);
2399 x = get_origin_and_offset_r (x, fldoff, fldsize, off);
2400 if (off && !tree_fits_uhwi_p (offset))
2401 *off = HOST_WIDE_INT_MAX;
2402 if (off)
2404 tree xtype = TREE_TYPE (x);
2405 set_aggregate_size_and_offset (xtype, fldoff, fldsize, off);
2407 return x;
2409 else if (code == VAR_DECL)
2411 x = gimple_assign_rhs1 (def);
2412 return get_origin_and_offset_r (x, fldoff, fldsize, off);
2415 else if (gimple_nop_p (def) && SSA_NAME_VAR (x))
2416 x = SSA_NAME_VAR (x);
2418 tree xtype = TREE_TYPE (x);
2419 if (POINTER_TYPE_P (xtype))
2420 xtype = TREE_TYPE (xtype);
2422 if (off)
2423 set_aggregate_size_and_offset (xtype, fldoff, fldsize, off);
2426 default:
2427 break;
2430 return x;
2433 /* Nonrecursive version of the above.
2434 The function never returns null unless X is null to begin with. */
2436 static tree
2437 get_origin_and_offset (tree x, HOST_WIDE_INT *fldoff, HOST_WIDE_INT *off,
2438 HOST_WIDE_INT *fldsize = nullptr)
2440 if (!x)
2441 return NULL_TREE;
2443 HOST_WIDE_INT sizebuf;
2444 if (!fldsize)
2445 fldsize = &sizebuf;
2447 /* Invalidate *FLDSIZE. */
2448 *fldsize = -1;
2449 *fldoff = *off = 0;
2451 return get_origin_and_offset_r (x, fldoff, fldsize, off);
2454 /* If ARG refers to the same (sub)object or array element as described
2455 by DST and DST_FLD, return the byte offset into the struct member or
2456 array element referenced by ARG and set *ARG_SIZE to the size of
2457 the (sub)object. Otherwise return HOST_WIDE_INT_MIN to indicate
2458 that ARG and DST do not refer to the same object. */
2460 static HOST_WIDE_INT
2461 alias_offset (tree arg, HOST_WIDE_INT *arg_size,
2462 tree dst, HOST_WIDE_INT dst_fld)
2464 /* See if the argument refers to the same base object as the destination
2465 of the formatted function call, and if so, try to determine if they
2466 can alias. */
2467 if (!arg || !dst || !ptr_derefs_may_alias_p (arg, dst))
2468 return HOST_WIDE_INT_MIN;
2470 /* The two arguments may refer to the same object. If they both refer
2471 to a struct member, see if the members are one and the same. If so,
2472 return the offset into the member. */
2473 HOST_WIDE_INT arg_off = 0, arg_fld = 0;
2475 tree arg_orig = get_origin_and_offset (arg, &arg_fld, &arg_off, arg_size);
2477 if (arg_orig == dst && arg_fld == dst_fld)
2478 return arg_off;
2480 return HOST_WIDE_INT_MIN;
2483 /* Return the minimum and maximum number of characters formatted
2484 by the '%s' format directive and its wide character form for
2485 the argument ARG. ARG can be null (for functions such as
2486 vsprinf). */
2488 static fmtresult
2489 format_string (const directive &dir, tree arg, pointer_query &ptr_qry)
2491 fmtresult res;
2493 /* The size of the (sub)object ARG refers to. Used to adjust
2494 the conservative get_string_length() result. */
2495 HOST_WIDE_INT arg_size = 0;
2497 if (warn_restrict)
2499 /* See if ARG might alias the destination of the call with
2500 DST_ORIGIN and DST_FIELD. If so, store the starting offset
2501 so that the overlap can be determined for certain later,
2502 when the amount of output of the call (including subsequent
2503 directives) has been computed. Otherwise, store HWI_MIN. */
2504 res.dst_offset = alias_offset (arg, &arg_size, dir.info->dst_origin,
2505 dir.info->dst_field);
2506 if (res.dst_offset >= 0 && res.dst_offset <= arg_size)
2507 arg_size -= res.dst_offset;
2508 else
2509 arg_size = 0;
2512 /* Compute the range the argument's length can be in. */
2513 int count_by = 1;
2514 if (dir.specifier == 'S' || dir.modifier == FMT_LEN_l)
2516 /* Get a node for a C type that will be the same size
2517 as a wchar_t on the target. */
2518 tree node = get_typenode_from_name (MODIFIED_WCHAR_TYPE);
2520 /* Now that we have a suitable node, get the number of
2521 bytes it occupies. */
2522 count_by = int_size_in_bytes (node);
2523 gcc_checking_assert (count_by == 2 || count_by == 4);
2526 fmtresult slen =
2527 get_string_length (arg, dir.info->callstmt, arg_size, count_by, ptr_qry);
2528 if (slen.range.min == slen.range.max
2529 && slen.range.min < HOST_WIDE_INT_MAX)
2531 /* The argument is either a string constant or it refers
2532 to one of a number of strings of the same length. */
2534 /* A '%s' directive with a string argument with constant length. */
2535 res.range = slen.range;
2537 if (dir.specifier == 'S'
2538 || dir.modifier == FMT_LEN_l)
2540 /* In the worst case the length of output of a wide string S
2541 is bounded by MB_LEN_MAX * wcslen (S). */
2542 res.range.max *= target_mb_len_max ();
2543 res.range.unlikely = res.range.max;
2544 /* It's likely that the total length is not more that
2545 2 * wcslen (S).*/
2546 res.range.likely = res.range.min * 2;
2548 if (dir.prec[1] >= 0
2549 && (unsigned HOST_WIDE_INT)dir.prec[1] < res.range.max)
2551 res.range.max = dir.prec[1];
2552 res.range.likely = dir.prec[1];
2553 res.range.unlikely = dir.prec[1];
2556 if (dir.prec[0] < 0 && dir.prec[1] > -1)
2557 res.range.min = 0;
2558 else if (dir.prec[0] >= 0)
2559 res.range.likely = dir.prec[0];
2561 /* Even a non-empty wide character string need not convert into
2562 any bytes. */
2563 res.range.min = 0;
2565 /* A non-empty wide character conversion may fail. */
2566 if (slen.range.max > 0)
2567 res.mayfail = true;
2569 else
2571 res.knownrange = true;
2573 if (dir.prec[0] < 0 && dir.prec[1] > -1)
2574 res.range.min = 0;
2575 else if ((unsigned HOST_WIDE_INT)dir.prec[0] < res.range.min)
2576 res.range.min = dir.prec[0];
2578 if ((unsigned HOST_WIDE_INT)dir.prec[1] < res.range.max)
2580 res.range.max = dir.prec[1];
2581 res.range.likely = dir.prec[1];
2582 res.range.unlikely = dir.prec[1];
2586 else if (arg && integer_zerop (arg))
2588 /* Handle null pointer argument. */
2590 fmtresult res (0);
2591 res.nullp = true;
2592 return res;
2594 else
2596 /* For a '%s' and '%ls' directive with a non-constant string (either
2597 one of a number of strings of known length or an unknown string)
2598 the minimum number of characters is lesser of PRECISION[0] and
2599 the length of the shortest known string or zero, and the maximum
2600 is the lesser of the length of the longest known string or
2601 PTRDIFF_MAX and PRECISION[1]. The likely length is either
2602 the minimum at level 1 and the greater of the minimum and 1
2603 at level 2. This result is adjust upward for width (if it's
2604 specified). */
2606 if (dir.specifier == 'S'
2607 || dir.modifier == FMT_LEN_l)
2609 /* A wide character converts to as few as zero bytes. */
2610 slen.range.min = 0;
2611 if (slen.range.max < target_int_max ())
2612 slen.range.max *= target_mb_len_max ();
2614 if (slen.range.likely < target_int_max ())
2615 slen.range.likely *= 2;
2617 if (slen.range.likely < target_int_max ())
2618 slen.range.unlikely *= target_mb_len_max ();
2620 /* A non-empty wide character conversion may fail. */
2621 if (slen.range.max > 0)
2622 res.mayfail = true;
2625 res.range = slen.range;
2627 if (dir.prec[0] >= 0)
2629 /* Adjust the minimum to zero if the string length is unknown,
2630 or at most the lower bound of the precision otherwise. */
2631 if (slen.range.min >= target_int_max ())
2632 res.range.min = 0;
2633 else if ((unsigned HOST_WIDE_INT)dir.prec[0] < slen.range.min)
2634 res.range.min = dir.prec[0];
2636 /* Make both maxima no greater than the upper bound of precision. */
2637 if ((unsigned HOST_WIDE_INT)dir.prec[1] < slen.range.max
2638 || slen.range.max >= target_int_max ())
2640 res.range.max = dir.prec[1];
2641 res.range.unlikely = dir.prec[1];
2644 /* If precision is constant, set the likely counter to the lesser
2645 of it and the maximum string length. Otherwise, if the lower
2646 bound of precision is greater than zero, set the likely counter
2647 to the minimum. Otherwise set it to zero or one based on
2648 the warning level. */
2649 if (dir.prec[0] == dir.prec[1])
2650 res.range.likely
2651 = ((unsigned HOST_WIDE_INT)dir.prec[0] < slen.range.max
2652 ? dir.prec[0] : slen.range.max);
2653 else if (dir.prec[0] > 0)
2654 res.range.likely = res.range.min;
2655 else
2656 res.range.likely = warn_level > 1;
2658 else if (dir.prec[1] >= 0)
2660 res.range.min = 0;
2661 if ((unsigned HOST_WIDE_INT)dir.prec[1] < slen.range.max)
2662 res.range.max = dir.prec[1];
2663 res.range.likely = dir.prec[1] ? warn_level > 1 : 0;
2664 if ((unsigned HOST_WIDE_INT)dir.prec[1] < slen.range.unlikely)
2665 res.range.unlikely = dir.prec[1];
2667 else if (slen.range.min >= target_int_max ())
2669 res.range.min = 0;
2670 res.range.max = HOST_WIDE_INT_MAX;
2671 /* At level 1 strings of unknown length are assumed to be
2672 empty, while at level 1 they are assumed to be one byte
2673 long. */
2674 res.range.likely = warn_level > 1;
2675 res.range.unlikely = HOST_WIDE_INT_MAX;
2677 else
2679 /* A string of unknown length unconstrained by precision is
2680 assumed to be empty at level 1 and just one character long
2681 at higher levels. */
2682 if (res.range.likely >= target_int_max ())
2683 res.range.likely = warn_level > 1;
2687 /* If the argument isn't a nul-terminated string and the number
2688 of bytes on output isn't bounded by precision, set NONSTR. */
2689 if (slen.nonstr && slen.range.min < (unsigned HOST_WIDE_INT)dir.prec[0])
2690 res.nonstr = slen.nonstr;
2692 /* Bump up the byte counters if WIDTH is greater. */
2693 return res.adjust_for_width_or_precision (dir.width);
2696 /* Format plain string (part of the format string itself). */
2698 static fmtresult
2699 format_plain (const directive &dir, tree, pointer_query &)
2701 fmtresult res (dir.len);
2702 return res;
2705 /* Return true if the RESULT of a directive in a call describe by INFO
2706 should be diagnosed given the AVAILable space in the destination. */
2708 static bool
2709 should_warn_p (const call_info &info,
2710 const result_range &avail, const result_range &result)
2712 if (result.max <= avail.min)
2714 /* The least amount of space remaining in the destination is big
2715 enough for the longest output. */
2716 return false;
2719 if (info.bounded)
2721 if (warn_format_trunc == 1 && result.min <= avail.max
2722 && info.retval_used ())
2724 /* The likely amount of space remaining in the destination is big
2725 enough for the least output and the return value is used. */
2726 return false;
2729 if (warn_format_trunc == 1 && result.likely <= avail.likely
2730 && !info.retval_used ())
2732 /* The likely amount of space remaining in the destination is big
2733 enough for the likely output and the return value is unused. */
2734 return false;
2737 if (warn_format_trunc == 2
2738 && result.likely <= avail.min
2739 && (result.max <= avail.min
2740 || result.max > HOST_WIDE_INT_MAX))
2742 /* The minimum amount of space remaining in the destination is big
2743 enough for the longest output. */
2744 return false;
2747 else
2749 if (warn_level == 1 && result.likely <= avail.likely)
2751 /* The likely amount of space remaining in the destination is big
2752 enough for the likely output. */
2753 return false;
2756 if (warn_level == 2
2757 && result.likely <= avail.min
2758 && (result.max <= avail.min
2759 || result.max > HOST_WIDE_INT_MAX))
2761 /* The minimum amount of space remaining in the destination is big
2762 enough for the longest output. */
2763 return false;
2767 return true;
2770 /* At format string location describe by DIRLOC in a call described
2771 by INFO, issue a warning for a directive DIR whose output may be
2772 in excess of the available space AVAIL_RANGE in the destination
2773 given the formatting result FMTRES. This function does nothing
2774 except decide whether to issue a warning for a possible write
2775 past the end or truncation and, if so, format the warning.
2776 Return true if a warning has been issued. */
2778 static bool
2779 maybe_warn (substring_loc &dirloc, location_t argloc,
2780 const call_info &info,
2781 const result_range &avail_range, const result_range &res,
2782 const directive &dir)
2784 if (!should_warn_p (info, avail_range, res))
2785 return false;
2787 /* A warning will definitely be issued below. */
2789 /* The maximum byte count to reference in the warning. Larger counts
2790 imply that the upper bound is unknown (and could be anywhere between
2791 RES.MIN + 1 and SIZE_MAX / 2) are printed as "N or more bytes" rather
2792 than "between N and X" where X is some huge number. */
2793 unsigned HOST_WIDE_INT maxbytes = target_dir_max ();
2795 /* True when there is enough room in the destination for the least
2796 amount of a directive's output but not enough for its likely or
2797 maximum output. */
2798 bool maybe = (res.min <= avail_range.max
2799 && (avail_range.min < res.likely
2800 || (res.max < HOST_WIDE_INT_MAX
2801 && avail_range.min < res.max)));
2803 /* Buffer for the directive in the host character set (used when
2804 the source character set is different). */
2805 char hostdir[32];
2807 if (avail_range.min == avail_range.max)
2809 /* The size of the destination region is exact. */
2810 unsigned HOST_WIDE_INT navail = avail_range.max;
2812 if (target_to_host (*dir.beg) != '%')
2814 /* For plain character directives (i.e., the format string itself)
2815 but not others, point the caret at the first character that's
2816 past the end of the destination. */
2817 if (navail < dir.len)
2818 dirloc.set_caret_index (dirloc.get_caret_idx () + navail);
2821 if (*dir.beg == '\0')
2823 /* This is the terminating nul. */
2824 gcc_assert (res.min == 1 && res.min == res.max);
2826 return fmtwarn (dirloc, UNKNOWN_LOCATION, NULL, info.warnopt (),
2827 info.bounded
2828 ? (maybe
2829 ? G_("%qE output may be truncated before the "
2830 "last format character")
2831 : G_("%qE output truncated before the last "
2832 "format character"))
2833 : (maybe
2834 ? G_("%qE may write a terminating nul past the "
2835 "end of the destination")
2836 : G_("%qE writing a terminating nul past the "
2837 "end of the destination")),
2838 info.func);
2841 if (res.min == res.max)
2843 const char *d = target_to_host (hostdir, sizeof hostdir, dir.beg);
2844 if (!info.bounded)
2845 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2846 "%<%.*s%> directive writing %wu byte into a "
2847 "region of size %wu",
2848 "%<%.*s%> directive writing %wu bytes into a "
2849 "region of size %wu",
2850 (int) dir.len, d, res.min, navail);
2851 else if (maybe)
2852 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2853 "%<%.*s%> directive output may be truncated "
2854 "writing %wu byte into a region of size %wu",
2855 "%<%.*s%> directive output may be truncated "
2856 "writing %wu bytes into a region of size %wu",
2857 (int) dir.len, d, res.min, navail);
2858 else
2859 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2860 "%<%.*s%> directive output truncated writing "
2861 "%wu byte into a region of size %wu",
2862 "%<%.*s%> directive output truncated writing "
2863 "%wu bytes into a region of size %wu",
2864 (int) dir.len, d, res.min, navail);
2866 if (res.min == 0 && res.max < maxbytes)
2867 return fmtwarn (dirloc, argloc, NULL,
2868 info.warnopt (),
2869 info.bounded
2870 ? (maybe
2871 ? G_("%<%.*s%> directive output may be truncated "
2872 "writing up to %wu bytes into a region of "
2873 "size %wu")
2874 : G_("%<%.*s%> directive output truncated writing "
2875 "up to %wu bytes into a region of size %wu"))
2876 : G_("%<%.*s%> directive writing up to %wu bytes "
2877 "into a region of size %wu"), (int) dir.len,
2878 target_to_host (hostdir, sizeof hostdir, dir.beg),
2879 res.max, navail);
2881 if (res.min == 0 && maxbytes <= res.max)
2882 /* This is a special case to avoid issuing the potentially
2883 confusing warning:
2884 writing 0 or more bytes into a region of size 0. */
2885 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
2886 info.bounded
2887 ? (maybe
2888 ? G_("%<%.*s%> directive output may be truncated "
2889 "writing likely %wu or more bytes into a "
2890 "region of size %wu")
2891 : G_("%<%.*s%> directive output truncated writing "
2892 "likely %wu or more bytes into a region of "
2893 "size %wu"))
2894 : G_("%<%.*s%> directive writing likely %wu or more "
2895 "bytes into a region of size %wu"), (int) dir.len,
2896 target_to_host (hostdir, sizeof hostdir, dir.beg),
2897 res.likely, navail);
2899 if (res.max < maxbytes)
2900 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
2901 info.bounded
2902 ? (maybe
2903 ? G_("%<%.*s%> directive output may be truncated "
2904 "writing between %wu and %wu bytes into a "
2905 "region of size %wu")
2906 : G_("%<%.*s%> directive output truncated "
2907 "writing between %wu and %wu bytes into a "
2908 "region of size %wu"))
2909 : G_("%<%.*s%> directive writing between %wu and "
2910 "%wu bytes into a region of size %wu"),
2911 (int) dir.len,
2912 target_to_host (hostdir, sizeof hostdir, dir.beg),
2913 res.min, res.max, navail);
2915 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
2916 info.bounded
2917 ? (maybe
2918 ? G_("%<%.*s%> directive output may be truncated "
2919 "writing %wu or more bytes into a region of "
2920 "size %wu")
2921 : G_("%<%.*s%> directive output truncated writing "
2922 "%wu or more bytes into a region of size %wu"))
2923 : G_("%<%.*s%> directive writing %wu or more bytes "
2924 "into a region of size %wu"), (int) dir.len,
2925 target_to_host (hostdir, sizeof hostdir, dir.beg),
2926 res.min, navail);
2929 /* The size of the destination region is a range. */
2931 if (target_to_host (*dir.beg) != '%')
2933 unsigned HOST_WIDE_INT navail = avail_range.max;
2935 /* For plain character directives (i.e., the format string itself)
2936 but not others, point the caret at the first character that's
2937 past the end of the destination. */
2938 if (navail < dir.len)
2939 dirloc.set_caret_index (dirloc.get_caret_idx () + navail);
2942 if (*dir.beg == '\0')
2944 gcc_assert (res.min == 1 && res.min == res.max);
2946 return fmtwarn (dirloc, UNKNOWN_LOCATION, NULL, info.warnopt (),
2947 info.bounded
2948 ? (maybe
2949 ? G_("%qE output may be truncated before the last "
2950 "format character")
2951 : G_("%qE output truncated before the last format "
2952 "character"))
2953 : (maybe
2954 ? G_("%qE may write a terminating nul past the end "
2955 "of the destination")
2956 : G_("%qE writing a terminating nul past the end "
2957 "of the destination")), info.func);
2960 if (res.min == res.max)
2962 const char *d = target_to_host (hostdir, sizeof hostdir, dir.beg);
2963 if (!info.bounded)
2964 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2965 "%<%.*s%> directive writing %wu byte into a region "
2966 "of size between %wu and %wu",
2967 "%<%.*s%> directive writing %wu bytes into a region "
2968 "of size between %wu and %wu", (int) dir.len, d,
2969 res.min, avail_range.min, avail_range.max);
2970 else if (maybe)
2971 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2972 "%<%.*s%> directive output may be truncated writing "
2973 "%wu byte into a region of size between %wu and %wu",
2974 "%<%.*s%> directive output may be truncated writing "
2975 "%wu bytes into a region of size between %wu and "
2976 "%wu", (int) dir.len, d, res.min, avail_range.min,
2977 avail_range.max);
2978 else
2979 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2980 "%<%.*s%> directive output truncated writing %wu "
2981 "byte into a region of size between %wu and %wu",
2982 "%<%.*s%> directive output truncated writing %wu "
2983 "bytes into a region of size between %wu and %wu",
2984 (int) dir.len, d, res.min, avail_range.min,
2985 avail_range.max);
2988 if (res.min == 0 && res.max < maxbytes)
2989 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
2990 info.bounded
2991 ? (maybe
2992 ? G_("%<%.*s%> directive output may be truncated "
2993 "writing up to %wu bytes into a region of size "
2994 "between %wu and %wu")
2995 : G_("%<%.*s%> directive output truncated writing "
2996 "up to %wu bytes into a region of size between "
2997 "%wu and %wu"))
2998 : G_("%<%.*s%> directive writing up to %wu bytes "
2999 "into a region of size between %wu and %wu"),
3000 (int) dir.len,
3001 target_to_host (hostdir, sizeof hostdir, dir.beg),
3002 res.max, avail_range.min, avail_range.max);
3004 if (res.min == 0 && maxbytes <= res.max)
3005 /* This is a special case to avoid issuing the potentially confusing
3006 warning:
3007 writing 0 or more bytes into a region of size between 0 and N. */
3008 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3009 info.bounded
3010 ? (maybe
3011 ? G_("%<%.*s%> directive output may be truncated "
3012 "writing likely %wu or more bytes into a region "
3013 "of size between %wu and %wu")
3014 : G_("%<%.*s%> directive output truncated writing "
3015 "likely %wu or more bytes into a region of size "
3016 "between %wu and %wu"))
3017 : G_("%<%.*s%> directive writing likely %wu or more bytes "
3018 "into a region of size between %wu and %wu"),
3019 (int) dir.len,
3020 target_to_host (hostdir, sizeof hostdir, dir.beg),
3021 res.likely, avail_range.min, avail_range.max);
3023 if (res.max < maxbytes)
3024 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3025 info.bounded
3026 ? (maybe
3027 ? G_("%<%.*s%> directive output may be truncated "
3028 "writing between %wu and %wu bytes into a region "
3029 "of size between %wu and %wu")
3030 : G_("%<%.*s%> directive output truncated writing "
3031 "between %wu and %wu bytes into a region of size "
3032 "between %wu and %wu"))
3033 : G_("%<%.*s%> directive writing between %wu and "
3034 "%wu bytes into a region of size between %wu and "
3035 "%wu"), (int) dir.len,
3036 target_to_host (hostdir, sizeof hostdir, dir.beg),
3037 res.min, res.max, avail_range.min, avail_range.max);
3039 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3040 info.bounded
3041 ? (maybe
3042 ? G_("%<%.*s%> directive output may be truncated writing "
3043 "%wu or more bytes into a region of size between "
3044 "%wu and %wu")
3045 : G_("%<%.*s%> directive output truncated writing "
3046 "%wu or more bytes into a region of size between "
3047 "%wu and %wu"))
3048 : G_("%<%.*s%> directive writing %wu or more bytes "
3049 "into a region of size between %wu and %wu"),
3050 (int) dir.len,
3051 target_to_host (hostdir, sizeof hostdir, dir.beg),
3052 res.min, avail_range.min, avail_range.max);
3055 /* Given the formatting result described by RES and NAVAIL, the number
3056 of available bytes in the destination, return the range of bytes
3057 remaining in the destination. */
3059 static inline result_range
3060 bytes_remaining (unsigned HOST_WIDE_INT navail, const format_result &res)
3062 result_range range;
3064 if (HOST_WIDE_INT_MAX <= navail)
3066 range.min = range.max = range.likely = range.unlikely = navail;
3067 return range;
3070 /* The lower bound of the available range is the available size
3071 minus the maximum output size, and the upper bound is the size
3072 minus the minimum. */
3073 range.max = res.range.min < navail ? navail - res.range.min : 0;
3075 range.likely = res.range.likely < navail ? navail - res.range.likely : 0;
3077 if (res.range.max < HOST_WIDE_INT_MAX)
3078 range.min = res.range.max < navail ? navail - res.range.max : 0;
3079 else
3080 range.min = range.likely;
3082 range.unlikely = (res.range.unlikely < navail
3083 ? navail - res.range.unlikely : 0);
3085 return range;
3088 /* Compute the length of the output resulting from the directive DIR
3089 in a call described by INFO and update the overall result of the call
3090 in *RES. Return true if the directive has been handled. */
3092 static bool
3093 format_directive (const call_info &info,
3094 format_result *res, const directive &dir,
3095 pointer_query &ptr_qry)
3097 /* Offset of the beginning of the directive from the beginning
3098 of the format string. */
3099 size_t offset = dir.beg - info.fmtstr;
3100 size_t start = offset;
3101 size_t length = offset + dir.len - !!dir.len;
3103 /* Create a location for the whole directive from the % to the format
3104 specifier. */
3105 substring_loc dirloc (info.fmtloc, TREE_TYPE (info.format),
3106 offset, start, length);
3108 /* Also get the location of the argument if possible.
3109 This doesn't work for integer literals or function calls. */
3110 location_t argloc = UNKNOWN_LOCATION;
3111 if (dir.arg)
3112 argloc = EXPR_LOCATION (dir.arg);
3114 /* Bail when there is no function to compute the output length,
3115 or when minimum length checking has been disabled. */
3116 if (!dir.fmtfunc || res->range.min >= HOST_WIDE_INT_MAX)
3117 return false;
3119 /* Compute the range of lengths of the formatted output. */
3120 fmtresult fmtres = dir.fmtfunc (dir, dir.arg, ptr_qry);
3122 /* Record whether the output of all directives is known to be
3123 bounded by some maximum, implying that their arguments are
3124 either known exactly or determined to be in a known range
3125 or, for strings, limited by the upper bounds of the arrays
3126 they refer to. */
3127 res->knownrange &= fmtres.knownrange;
3129 if (!fmtres.knownrange)
3131 /* Only when the range is known, check it against the host value
3132 of INT_MAX + (the number of bytes of the "%.*Lf" directive with
3133 INT_MAX precision, which is the longest possible output of any
3134 single directive). That's the largest valid byte count (though
3135 not valid call to a printf-like function because it can never
3136 return such a count). Otherwise, the range doesn't correspond
3137 to known values of the argument. */
3138 if (fmtres.range.max > target_dir_max ())
3140 /* Normalize the MAX counter to avoid having to deal with it
3141 later. The counter can be less than HOST_WIDE_INT_M1U
3142 when compiling for an ILP32 target on an LP64 host. */
3143 fmtres.range.max = HOST_WIDE_INT_M1U;
3144 /* Disable exact and maximum length checking after a failure
3145 to determine the maximum number of characters (for example
3146 for wide characters or wide character strings) but continue
3147 tracking the minimum number of characters. */
3148 res->range.max = HOST_WIDE_INT_M1U;
3151 if (fmtres.range.min > target_dir_max ())
3153 /* Disable exact length checking after a failure to determine
3154 even the minimum number of characters (it shouldn't happen
3155 except in an error) but keep tracking the minimum and maximum
3156 number of characters. */
3157 return true;
3161 /* Buffer for the directive in the host character set (used when
3162 the source character set is different). */
3163 char hostdir[32];
3165 int dirlen = dir.len;
3167 if (fmtres.nullp)
3169 fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3170 "%<%.*s%> directive argument is null",
3171 dirlen, target_to_host (hostdir, sizeof hostdir, dir.beg));
3173 /* Don't bother processing the rest of the format string. */
3174 res->warned = true;
3175 res->range.min = HOST_WIDE_INT_M1U;
3176 res->range.max = HOST_WIDE_INT_M1U;
3177 return false;
3180 /* Compute the number of available bytes in the destination. There
3181 must always be at least one byte of space for the terminating
3182 NUL that's appended after the format string has been processed. */
3183 result_range avail_range = bytes_remaining (info.objsize, *res);
3185 /* If the argument aliases a part of the destination of the formatted
3186 call at offset FMTRES.DST_OFFSET append the directive and its result
3187 to the set of aliases for later processing. */
3188 if (fmtres.dst_offset != HOST_WIDE_INT_MIN)
3189 res->append_alias (dir, fmtres.dst_offset, fmtres.range);
3191 bool warned = res->warned;
3193 if (!warned)
3194 warned = maybe_warn (dirloc, argloc, info, avail_range,
3195 fmtres.range, dir);
3197 /* Bump up the total maximum if it isn't too big. */
3198 if (res->range.max < HOST_WIDE_INT_MAX
3199 && fmtres.range.max < HOST_WIDE_INT_MAX)
3200 res->range.max += fmtres.range.max;
3202 /* Raise the total unlikely maximum by the larger of the maximum
3203 and the unlikely maximum. */
3204 unsigned HOST_WIDE_INT save = res->range.unlikely;
3205 if (fmtres.range.max < fmtres.range.unlikely)
3206 res->range.unlikely += fmtres.range.unlikely;
3207 else
3208 res->range.unlikely += fmtres.range.max;
3210 if (res->range.unlikely < save)
3211 res->range.unlikely = HOST_WIDE_INT_M1U;
3213 res->range.min += fmtres.range.min;
3214 res->range.likely += fmtres.range.likely;
3216 /* Has the minimum directive output length exceeded the maximum
3217 of 4095 bytes required to be supported? */
3218 bool minunder4k = fmtres.range.min < 4096;
3219 bool maxunder4k = fmtres.range.max < 4096;
3220 /* Clear POSUNDER4K in the overall result if the maximum has exceeded
3221 the 4k (this is necessary to avoid the return value optimization
3222 that may not be safe in the maximum case). */
3223 if (!maxunder4k)
3224 res->posunder4k = false;
3225 /* Also clear POSUNDER4K if the directive may fail. */
3226 if (fmtres.mayfail)
3227 res->posunder4k = false;
3229 if (!warned
3230 /* Only warn at level 2. */
3231 && warn_level > 1
3232 /* Only warn for string functions. */
3233 && info.is_string_func ()
3234 && (!minunder4k
3235 || (!maxunder4k && fmtres.range.max < HOST_WIDE_INT_MAX)))
3237 /* The directive output may be longer than the maximum required
3238 to be handled by an implementation according to 7.21.6.1, p15
3239 of C11. Warn on this only at level 2 but remember this and
3240 prevent folding the return value when done. This allows for
3241 the possibility of the actual libc call failing due to ENOMEM
3242 (like Glibc does with very large precision or width).
3243 Issue the "may exceed" warning only for string functions and
3244 not for fprintf or printf. */
3246 if (fmtres.range.min == fmtres.range.max)
3247 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3248 "%<%.*s%> directive output of %wu bytes exceeds "
3249 "minimum required size of 4095", dirlen,
3250 target_to_host (hostdir, sizeof hostdir, dir.beg),
3251 fmtres.range.min);
3252 else if (!minunder4k)
3253 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3254 "%<%.*s%> directive output between %wu and %wu "
3255 "bytes exceeds minimum required size of 4095",
3256 dirlen,
3257 target_to_host (hostdir, sizeof hostdir, dir.beg),
3258 fmtres.range.min, fmtres.range.max);
3259 else if (!info.retval_used () && info.is_string_func ())
3260 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3261 "%<%.*s%> directive output between %wu and %wu "
3262 "bytes may exceed minimum required size of "
3263 "4095",
3264 dirlen,
3265 target_to_host (hostdir, sizeof hostdir, dir.beg),
3266 fmtres.range.min, fmtres.range.max);
3269 /* Has the likely and maximum directive output exceeded INT_MAX? */
3270 bool likelyximax = *dir.beg && res->range.likely > target_int_max ();
3271 /* Don't consider the maximum to be in excess when it's the result
3272 of a string of unknown length (i.e., whose maximum has been set
3273 to be greater than or equal to HOST_WIDE_INT_MAX. */
3274 bool maxximax = (*dir.beg
3275 && res->range.max > target_int_max ()
3276 && res->range.max < HOST_WIDE_INT_MAX);
3278 if (!warned
3279 /* Warn for the likely output size at level 1. */
3280 && (likelyximax
3281 /* But only warn for the maximum at level 2. */
3282 || (warn_level > 1
3283 && maxximax
3284 && fmtres.range.max < HOST_WIDE_INT_MAX)))
3286 if (fmtres.range.min > target_int_max ())
3288 /* The directive output exceeds INT_MAX bytes. */
3289 if (fmtres.range.min == fmtres.range.max)
3290 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3291 "%<%.*s%> directive output of %wu bytes exceeds "
3292 "%<INT_MAX%>", dirlen,
3293 target_to_host (hostdir, sizeof hostdir, dir.beg),
3294 fmtres.range.min);
3295 else
3296 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3297 "%<%.*s%> directive output between %wu and "
3298 "%wu bytes exceeds %<INT_MAX%>", dirlen,
3299 target_to_host (hostdir, sizeof hostdir, dir.beg),
3300 fmtres.range.min, fmtres.range.max);
3302 else if (res->range.min > target_int_max ())
3304 /* The directive output is under INT_MAX but causes the result
3305 to exceed INT_MAX bytes. */
3306 if (fmtres.range.min == fmtres.range.max)
3307 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3308 "%<%.*s%> directive output of %wu bytes causes "
3309 "result to exceed %<INT_MAX%>", dirlen,
3310 target_to_host (hostdir, sizeof hostdir, dir.beg),
3311 fmtres.range.min);
3312 else
3313 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3314 "%<%.*s%> directive output between %wu and "
3315 "%wu bytes causes result to exceed %<INT_MAX%>",
3316 dirlen,
3317 target_to_host (hostdir, sizeof hostdir, dir.beg),
3318 fmtres.range.min, fmtres.range.max);
3320 else if ((!info.retval_used () || !info.bounded)
3321 && (info.is_string_func ()))
3322 /* Warn for calls to string functions that either aren't bounded
3323 (sprintf) or whose return value isn't used. */
3324 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3325 "%<%.*s%> directive output between %wu and "
3326 "%wu bytes may cause result to exceed "
3327 "%<INT_MAX%>", dirlen,
3328 target_to_host (hostdir, sizeof hostdir, dir.beg),
3329 fmtres.range.min, fmtres.range.max);
3332 if (!warned && fmtres.nonstr)
3334 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3335 "%<%.*s%> directive argument is not a nul-terminated "
3336 "string",
3337 dirlen,
3338 target_to_host (hostdir, sizeof hostdir, dir.beg));
3339 if (warned && DECL_P (fmtres.nonstr))
3340 inform (DECL_SOURCE_LOCATION (fmtres.nonstr),
3341 "referenced argument declared here");
3342 return false;
3345 if (warned && fmtres.range.min < fmtres.range.likely
3346 && fmtres.range.likely < fmtres.range.max)
3347 inform_n (info.fmtloc, fmtres.range.likely,
3348 "assuming directive output of %wu byte",
3349 "assuming directive output of %wu bytes",
3350 fmtres.range.likely);
3352 if (warned && fmtres.argmin)
3354 if (fmtres.argmin == fmtres.argmax)
3355 inform (info.fmtloc, "directive argument %qE", fmtres.argmin);
3356 else if (fmtres.knownrange)
3357 inform (info.fmtloc, "directive argument in the range [%E, %E]",
3358 fmtres.argmin, fmtres.argmax);
3359 else
3360 inform (info.fmtloc,
3361 "using the range [%E, %E] for directive argument",
3362 fmtres.argmin, fmtres.argmax);
3365 res->warned |= warned;
3367 if (!dir.beg[0] && res->warned)
3369 location_t callloc = gimple_location (info.callstmt);
3371 unsigned HOST_WIDE_INT min = res->range.min;
3372 unsigned HOST_WIDE_INT max = res->range.max;
3374 if (info.objsize < HOST_WIDE_INT_MAX)
3376 /* If a warning has been issued for buffer overflow or truncation
3377 help the user figure out how big a buffer they need. */
3379 if (min == max)
3380 inform_n (callloc, min,
3381 "%qE output %wu byte into a destination of size %wu",
3382 "%qE output %wu bytes into a destination of size %wu",
3383 info.func, min, info.objsize);
3384 else if (max < HOST_WIDE_INT_MAX)
3385 inform (callloc,
3386 "%qE output between %wu and %wu bytes into "
3387 "a destination of size %wu",
3388 info.func, min, max, info.objsize);
3389 else if (min < res->range.likely && res->range.likely < max)
3390 inform (callloc,
3391 "%qE output %wu or more bytes (assuming %wu) into "
3392 "a destination of size %wu",
3393 info.func, min, res->range.likely, info.objsize);
3394 else
3395 inform (callloc,
3396 "%qE output %wu or more bytes into a destination of size "
3397 "%wu",
3398 info.func, min, info.objsize);
3400 else if (!info.is_string_func ())
3402 /* If the warning is for a file function like fprintf
3403 of printf with no destination size just print the computed
3404 result. */
3405 if (min == max)
3406 inform_n (callloc, min,
3407 "%qE output %wu byte", "%qE output %wu bytes",
3408 info.func, min);
3409 else if (max < HOST_WIDE_INT_MAX)
3410 inform (callloc,
3411 "%qE output between %wu and %wu bytes",
3412 info.func, min, max);
3413 else if (min < res->range.likely && res->range.likely < max)
3414 inform (callloc,
3415 "%qE output %wu or more bytes (assuming %wu)",
3416 info.func, min, res->range.likely);
3417 else
3418 inform (callloc,
3419 "%qE output %wu or more bytes",
3420 info.func, min);
3424 if (dump_file && *dir.beg)
3426 fprintf (dump_file,
3427 " Result: "
3428 HOST_WIDE_INT_PRINT_DEC ", " HOST_WIDE_INT_PRINT_DEC ", "
3429 HOST_WIDE_INT_PRINT_DEC ", " HOST_WIDE_INT_PRINT_DEC " ("
3430 HOST_WIDE_INT_PRINT_DEC ", " HOST_WIDE_INT_PRINT_DEC ", "
3431 HOST_WIDE_INT_PRINT_DEC ", " HOST_WIDE_INT_PRINT_DEC ")\n",
3432 fmtres.range.min, fmtres.range.likely,
3433 fmtres.range.max, fmtres.range.unlikely,
3434 res->range.min, res->range.likely,
3435 res->range.max, res->range.unlikely);
3438 return true;
3441 /* Parse a format directive in function call described by INFO starting
3442 at STR and populate DIR structure. Bump up *ARGNO by the number of
3443 arguments extracted for the directive. Return the length of
3444 the directive. */
3446 static size_t
3447 parse_directive (call_info &info,
3448 directive &dir, format_result *res,
3449 const char *str, unsigned *argno,
3450 range_query *query)
3452 const char *pcnt = strchr (str, target_percent);
3453 dir.beg = str;
3455 if (size_t len = pcnt ? pcnt - str : *str ? strlen (str) : 1)
3457 /* This directive is either a plain string or the terminating nul
3458 (which isn't really a directive but it simplifies things to
3459 handle it as if it were). */
3460 dir.len = len;
3461 dir.fmtfunc = format_plain;
3463 if (dump_file)
3465 fprintf (dump_file, " Directive %u at offset "
3466 HOST_WIDE_INT_PRINT_UNSIGNED ": \"%.*s\", "
3467 "length = " HOST_WIDE_INT_PRINT_UNSIGNED "\n",
3468 dir.dirno,
3469 (unsigned HOST_WIDE_INT)(size_t)(dir.beg - info.fmtstr),
3470 (int)dir.len, dir.beg, (unsigned HOST_WIDE_INT) dir.len);
3473 return len - !*str;
3476 /* Set the directive argument's number to correspond to its position
3477 in the formatted function call's argument list. */
3478 dir.argno = *argno;
3480 const char *pf = pcnt + 1;
3482 /* POSIX numbered argument index or zero when none. */
3483 HOST_WIDE_INT dollar = 0;
3485 /* With and precision. -1 when not specified, HOST_WIDE_INT_MIN
3486 when given by a va_list argument, and a non-negative value
3487 when specified in the format string itself. */
3488 HOST_WIDE_INT width = -1;
3489 HOST_WIDE_INT precision = -1;
3491 /* Pointers to the beginning of the width and precision decimal
3492 string (if any) within the directive. */
3493 const char *pwidth = 0;
3494 const char *pprec = 0;
3496 /* When the value of the decimal string that specifies width or
3497 precision is out of range, points to the digit that causes
3498 the value to exceed the limit. */
3499 const char *werange = NULL;
3500 const char *perange = NULL;
3502 /* Width specified via the asterisk. Need not be INTEGER_CST.
3503 For vararg functions set to void_node. */
3504 tree star_width = NULL_TREE;
3506 /* Width specified via the asterisk. Need not be INTEGER_CST.
3507 For vararg functions set to void_node. */
3508 tree star_precision = NULL_TREE;
3510 if (ISDIGIT (target_to_host (*pf)))
3512 /* This could be either a POSIX positional argument, the '0'
3513 flag, or a width, depending on what follows. Store it as
3514 width and sort it out later after the next character has
3515 been seen. */
3516 pwidth = pf;
3517 width = target_strtowi (&pf, &werange);
3519 else if (target_to_host (*pf) == '*')
3521 /* Similarly to the block above, this could be either a POSIX
3522 positional argument or a width, depending on what follows. */
3523 if (*argno < gimple_call_num_args (info.callstmt))
3524 star_width = gimple_call_arg (info.callstmt, (*argno)++);
3525 else
3526 star_width = void_node;
3527 ++pf;
3530 if (target_to_host (*pf) == '$')
3532 /* Handle the POSIX dollar sign which references the 1-based
3533 positional argument number. */
3534 if (width != -1)
3535 dollar = width + info.argidx;
3536 else if (star_width
3537 && TREE_CODE (star_width) == INTEGER_CST
3538 && (TYPE_PRECISION (TREE_TYPE (star_width))
3539 <= TYPE_PRECISION (integer_type_node)))
3540 dollar = width + tree_to_shwi (star_width);
3542 /* Bail when the numbered argument is out of range (it will
3543 have already been diagnosed by -Wformat). */
3544 if (dollar == 0
3545 || dollar == (int)info.argidx
3546 || dollar > gimple_call_num_args (info.callstmt))
3547 return false;
3549 --dollar;
3551 star_width = NULL_TREE;
3552 width = -1;
3553 ++pf;
3556 if (dollar || !star_width)
3558 if (width != -1)
3560 if (width == 0)
3562 /* The '0' that has been interpreted as a width above is
3563 actually a flag. Reset HAVE_WIDTH, set the '0' flag,
3564 and continue processing other flags. */
3565 width = -1;
3566 dir.set_flag ('0');
3568 else if (!dollar)
3570 /* (Non-zero) width has been seen. The next character
3571 is either a period or a digit. */
3572 goto start_precision;
3575 /* When either '$' has been seen, or width has not been seen,
3576 the next field is the optional flags followed by an optional
3577 width. */
3578 for ( ; ; ) {
3579 switch (target_to_host (*pf))
3581 case ' ':
3582 case '0':
3583 case '+':
3584 case '-':
3585 case '#':
3586 dir.set_flag (target_to_host (*pf++));
3587 break;
3589 default:
3590 goto start_width;
3594 start_width:
3595 if (ISDIGIT (target_to_host (*pf)))
3597 werange = 0;
3598 pwidth = pf;
3599 width = target_strtowi (&pf, &werange);
3601 else if (target_to_host (*pf) == '*')
3603 if (*argno < gimple_call_num_args (info.callstmt))
3604 star_width = gimple_call_arg (info.callstmt, (*argno)++);
3605 else
3607 /* This is (likely) a va_list. It could also be an invalid
3608 call with insufficient arguments. */
3609 star_width = void_node;
3611 ++pf;
3613 else if (target_to_host (*pf) == '\'')
3615 /* The POSIX apostrophe indicating a numeric grouping
3616 in the current locale. Even though it's possible to
3617 estimate the upper bound on the size of the output
3618 based on the number of digits it probably isn't worth
3619 continuing. */
3620 return 0;
3624 start_precision:
3625 if (target_to_host (*pf) == '.')
3627 ++pf;
3629 if (ISDIGIT (target_to_host (*pf)))
3631 pprec = pf;
3632 precision = target_strtowi (&pf, &perange);
3634 else if (target_to_host (*pf) == '*')
3636 if (*argno < gimple_call_num_args (info.callstmt))
3637 star_precision = gimple_call_arg (info.callstmt, (*argno)++);
3638 else
3640 /* This is (likely) a va_list. It could also be an invalid
3641 call with insufficient arguments. */
3642 star_precision = void_node;
3644 ++pf;
3646 else
3648 /* The decimal precision or the asterisk are optional.
3649 When neither is specified it's taken to be zero. */
3650 precision = 0;
3654 switch (target_to_host (*pf))
3656 case 'h':
3657 if (target_to_host (pf[1]) == 'h')
3659 ++pf;
3660 dir.modifier = FMT_LEN_hh;
3662 else
3663 dir.modifier = FMT_LEN_h;
3664 ++pf;
3665 break;
3667 case 'j':
3668 dir.modifier = FMT_LEN_j;
3669 ++pf;
3670 break;
3672 case 'L':
3673 dir.modifier = FMT_LEN_L;
3674 ++pf;
3675 break;
3677 case 'l':
3678 if (target_to_host (pf[1]) == 'l')
3680 ++pf;
3681 dir.modifier = FMT_LEN_ll;
3683 else
3684 dir.modifier = FMT_LEN_l;
3685 ++pf;
3686 break;
3688 case 't':
3689 dir.modifier = FMT_LEN_t;
3690 ++pf;
3691 break;
3693 case 'z':
3694 dir.modifier = FMT_LEN_z;
3695 ++pf;
3696 break;
3699 switch (target_to_host (*pf))
3701 /* Handle a sole '%' character the same as "%%" but since it's
3702 undefined prevent the result from being folded. */
3703 case '\0':
3704 --pf;
3705 res->range.min = res->range.max = HOST_WIDE_INT_M1U;
3706 /* FALLTHRU */
3707 case '%':
3708 dir.fmtfunc = format_percent;
3709 break;
3711 case 'a':
3712 case 'A':
3713 case 'e':
3714 case 'E':
3715 case 'f':
3716 case 'F':
3717 case 'g':
3718 case 'G':
3719 res->floating = true;
3720 dir.fmtfunc = format_floating;
3721 break;
3723 case 'd':
3724 case 'i':
3725 case 'o':
3726 case 'u':
3727 case 'x':
3728 case 'X':
3729 dir.fmtfunc = format_integer;
3730 break;
3732 case 'b':
3733 case 'B':
3734 dir.fmtfunc = format_integer;
3735 break;
3737 case 'p':
3738 /* The %p output is implementation-defined. It's possible
3739 to determine this format but due to extensions (especially
3740 those of the Linux kernel -- see bug 78512) the first %p
3741 in the format string disables any further processing. */
3742 return false;
3744 case 'n':
3745 /* %n has side-effects even when nothing is actually printed to
3746 any buffer. */
3747 info.nowrite = false;
3748 dir.fmtfunc = format_none;
3749 break;
3751 case 'C':
3752 case 'c':
3753 /* POSIX wide character and C/POSIX narrow character. */
3754 dir.fmtfunc = format_character;
3755 break;
3757 case 'S':
3758 case 's':
3759 /* POSIX wide string and C/POSIX narrow character string. */
3760 dir.fmtfunc = format_string;
3761 break;
3763 default:
3764 /* Unknown conversion specification. */
3765 return 0;
3768 dir.specifier = target_to_host (*pf++);
3770 /* Store the length of the format directive. */
3771 dir.len = pf - pcnt;
3773 /* Buffer for the directive in the host character set (used when
3774 the source character set is different). */
3775 char hostdir[32];
3777 if (star_width)
3779 if (INTEGRAL_TYPE_P (TREE_TYPE (star_width)))
3780 dir.set_width (star_width, query);
3781 else
3783 /* Width specified by a va_list takes on the range [0, -INT_MIN]
3784 (width is the absolute value of that specified). */
3785 dir.width[0] = 0;
3786 dir.width[1] = target_int_max () + 1;
3789 else
3791 if (width == HOST_WIDE_INT_MAX && werange)
3793 size_t begin = dir.beg - info.fmtstr + (pwidth - pcnt);
3794 size_t caret = begin + (werange - pcnt);
3795 size_t end = pf - info.fmtstr - 1;
3797 /* Create a location for the width part of the directive,
3798 pointing the caret at the first out-of-range digit. */
3799 substring_loc dirloc (info.fmtloc, TREE_TYPE (info.format),
3800 caret, begin, end);
3802 fmtwarn (dirloc, UNKNOWN_LOCATION, NULL, info.warnopt (),
3803 "%<%.*s%> directive width out of range", (int) dir.len,
3804 target_to_host (hostdir, sizeof hostdir, dir.beg));
3807 dir.set_width (width);
3810 if (star_precision)
3812 if (INTEGRAL_TYPE_P (TREE_TYPE (star_precision)))
3813 dir.set_precision (star_precision, query);
3814 else
3816 /* Precision specified by a va_list takes on the range [-1, INT_MAX]
3817 (unlike width, negative precision is ignored). */
3818 dir.prec[0] = -1;
3819 dir.prec[1] = target_int_max ();
3822 else
3824 if (precision == HOST_WIDE_INT_MAX && perange)
3826 size_t begin = dir.beg - info.fmtstr + (pprec - pcnt) - 1;
3827 size_t caret = dir.beg - info.fmtstr + (perange - pcnt) - 1;
3828 size_t end = pf - info.fmtstr - 2;
3830 /* Create a location for the precision part of the directive,
3831 including the leading period, pointing the caret at the first
3832 out-of-range digit . */
3833 substring_loc dirloc (info.fmtloc, TREE_TYPE (info.format),
3834 caret, begin, end);
3836 fmtwarn (dirloc, UNKNOWN_LOCATION, NULL, info.warnopt (),
3837 "%<%.*s%> directive precision out of range", (int) dir.len,
3838 target_to_host (hostdir, sizeof hostdir, dir.beg));
3841 dir.set_precision (precision);
3844 /* Extract the argument if the directive takes one and if it's
3845 available (e.g., the function doesn't take a va_list). Treat
3846 missing arguments the same as va_list, even though they will
3847 have likely already been diagnosed by -Wformat. */
3848 if (dir.specifier != '%'
3849 && *argno < gimple_call_num_args (info.callstmt))
3850 dir.arg = gimple_call_arg (info.callstmt, dollar ? dollar : (*argno)++);
3852 if (dump_file)
3854 fprintf (dump_file,
3855 " Directive %u at offset " HOST_WIDE_INT_PRINT_UNSIGNED
3856 ": \"%.*s\"",
3857 dir.dirno,
3858 (unsigned HOST_WIDE_INT)(size_t)(dir.beg - info.fmtstr),
3859 (int)dir.len, dir.beg);
3860 if (star_width)
3862 if (dir.width[0] == dir.width[1])
3863 fprintf (dump_file, ", width = " HOST_WIDE_INT_PRINT_DEC,
3864 dir.width[0]);
3865 else
3866 fprintf (dump_file,
3867 ", width in range [" HOST_WIDE_INT_PRINT_DEC
3868 ", " HOST_WIDE_INT_PRINT_DEC "]",
3869 dir.width[0], dir.width[1]);
3872 if (star_precision)
3874 if (dir.prec[0] == dir.prec[1])
3875 fprintf (dump_file, ", precision = " HOST_WIDE_INT_PRINT_DEC,
3876 dir.prec[0]);
3877 else
3878 fprintf (dump_file,
3879 ", precision in range [" HOST_WIDE_INT_PRINT_DEC
3880 HOST_WIDE_INT_PRINT_DEC "]",
3881 dir.prec[0], dir.prec[1]);
3883 fputc ('\n', dump_file);
3886 return dir.len;
3889 /* Diagnose overlap between destination and %s directive arguments. */
3891 static void
3892 maybe_warn_overlap (call_info &info, format_result *res)
3894 /* Two vectors of 1-based indices corresponding to either certainly
3895 or possibly aliasing arguments. */
3896 auto_vec<int, 16> aliasarg[2];
3898 /* Go through the array of potentially aliasing directives and collect
3899 argument numbers of those that do or may overlap the destination
3900 object given the full result. */
3901 for (unsigned i = 0; i != res->alias_count; ++i)
3903 const format_result::alias_info &alias = res->aliases[i];
3905 enum { possible = -1, none = 0, certain = 1 } overlap = none;
3907 /* If the precision is zero there is no overlap. (This only
3908 considers %s directives and ignores %n.) */
3909 if (alias.dir.prec[0] == 0 && alias.dir.prec[1] == 0)
3910 continue;
3912 if (alias.offset == HOST_WIDE_INT_MAX
3913 || info.dst_offset == HOST_WIDE_INT_MAX)
3914 overlap = possible;
3915 else if (alias.offset == info.dst_offset)
3916 overlap = alias.dir.prec[0] == 0 ? possible : certain;
3917 else
3919 /* Determine overlap from the range of output and offsets
3920 into the same destination as the source, and rule out
3921 impossible overlap. */
3922 unsigned HOST_WIDE_INT albeg = alias.offset;
3923 unsigned HOST_WIDE_INT dstbeg = info.dst_offset;
3925 unsigned HOST_WIDE_INT alend = albeg + alias.range.min;
3926 unsigned HOST_WIDE_INT dstend = dstbeg + res->range.min - 1;
3928 if ((albeg <= dstbeg && alend > dstbeg)
3929 || (albeg >= dstbeg && albeg < dstend))
3930 overlap = certain;
3931 else
3933 alend = albeg + alias.range.max;
3934 if (alend < albeg)
3935 alend = HOST_WIDE_INT_M1U;
3937 dstend = dstbeg + res->range.max - 1;
3938 if (dstend < dstbeg)
3939 dstend = HOST_WIDE_INT_M1U;
3941 if ((albeg >= dstbeg && albeg <= dstend)
3942 || (alend >= dstbeg && alend <= dstend))
3943 overlap = possible;
3947 if (overlap == none)
3948 continue;
3950 /* Append the 1-based argument number. */
3951 aliasarg[overlap != certain].safe_push (alias.dir.argno + 1);
3953 /* Disable any kind of optimization. */
3954 res->range.unlikely = HOST_WIDE_INT_M1U;
3957 tree arg0 = gimple_call_arg (info.callstmt, 0);
3958 location_t loc = gimple_location (info.callstmt);
3960 bool aliaswarn = false;
3962 unsigned ncertain = aliasarg[0].length ();
3963 unsigned npossible = aliasarg[1].length ();
3964 if (ncertain && npossible)
3966 /* If there are multiple arguments that overlap, some certainly
3967 and some possibly, handle both sets in a single diagnostic. */
3968 aliaswarn
3969 = warning_at (loc, OPT_Wrestrict,
3970 "%qE arguments %Z and maybe %Z overlap destination "
3971 "object %qE",
3972 info.func, aliasarg[0].address (), ncertain,
3973 aliasarg[1].address (), npossible,
3974 info.dst_origin);
3976 else if (ncertain)
3978 /* There is only one set of two or more arguments and they all
3979 certainly overlap the destination. */
3980 aliaswarn
3981 = warning_n (loc, OPT_Wrestrict, ncertain,
3982 "%qE argument %Z overlaps destination object %qE",
3983 "%qE arguments %Z overlap destination object %qE",
3984 info.func, aliasarg[0].address (), ncertain,
3985 info.dst_origin);
3987 else if (npossible)
3989 /* There is only one set of two or more arguments and they all
3990 may overlap (but need not). */
3991 aliaswarn
3992 = warning_n (loc, OPT_Wrestrict, npossible,
3993 "%qE argument %Z may overlap destination object %qE",
3994 "%qE arguments %Z may overlap destination object %qE",
3995 info.func, aliasarg[1].address (), npossible,
3996 info.dst_origin);
3999 if (aliaswarn)
4001 res->warned = true;
4003 if (info.dst_origin != arg0)
4005 /* If its location is different from the first argument of the call
4006 point either at the destination object itself or at the expression
4007 that was used to determine the overlap. */
4008 loc = (DECL_P (info.dst_origin)
4009 ? DECL_SOURCE_LOCATION (info.dst_origin)
4010 : EXPR_LOCATION (info.dst_origin));
4011 if (loc != UNKNOWN_LOCATION)
4012 inform (loc,
4013 "destination object referenced by %<restrict%>-qualified "
4014 "argument 1 was declared here");
4019 /* Compute the length of the output resulting from the call to a formatted
4020 output function described by INFO and store the result of the call in
4021 *RES. Issue warnings for detected past the end writes. Return true
4022 if the complete format string has been processed and *RES can be relied
4023 on, false otherwise (e.g., when a unknown or unhandled directive was seen
4024 that caused the processing to be terminated early). */
4026 static bool
4027 compute_format_length (call_info &info, format_result *res,
4028 pointer_query &ptr_qry)
4030 if (dump_file)
4032 location_t callloc = gimple_location (info.callstmt);
4033 fprintf (dump_file, "%s:%i: ",
4034 LOCATION_FILE (callloc), LOCATION_LINE (callloc));
4035 print_generic_expr (dump_file, info.func, dump_flags);
4037 fprintf (dump_file,
4038 ": objsize = " HOST_WIDE_INT_PRINT_UNSIGNED
4039 ", fmtstr = \"%s\"\n",
4040 info.objsize, info.fmtstr);
4043 /* Reset the minimum and maximum byte counters. */
4044 res->range.min = res->range.max = 0;
4046 /* No directive has been seen yet so the length of output is bounded
4047 by the known range [0, 0] (with no conversion resulting in a failure
4048 or producing more than 4K bytes) until determined otherwise. */
4049 res->knownrange = true;
4050 res->floating = false;
4051 res->warned = false;
4053 /* 1-based directive counter. */
4054 unsigned dirno = 1;
4056 /* The variadic argument counter. */
4057 unsigned argno = info.argidx;
4059 bool success = true;
4061 for (const char *pf = info.fmtstr; ; ++dirno)
4063 directive dir (&info, dirno);
4065 size_t n = parse_directive (info, dir, res, pf, &argno, ptr_qry.rvals);
4067 /* Return failure if the format function fails. */
4068 if (!format_directive (info, res, dir, ptr_qry))
4069 return false;
4071 /* Return success when the directive is zero bytes long and it's
4072 the last thing in the format string (i.e., it's the terminating
4073 nul, which isn't really a directive but handling it as one makes
4074 things simpler). */
4075 if (!n)
4077 success = *pf == '\0';
4078 break;
4081 pf += n;
4084 maybe_warn_overlap (info, res);
4086 /* The complete format string was processed (with or without warnings). */
4087 return success;
4090 /* Return the size of the object referenced by the expression DEST in
4091 statement STMT, if available, or the maximum possible size otherwise. */
4093 static unsigned HOST_WIDE_INT
4094 get_destination_size (tree dest, gimple *stmt, pointer_query &ptr_qry)
4096 /* When there is no destination return the maximum. */
4097 if (!dest)
4098 return HOST_WIDE_INT_MAX;
4100 /* Use compute_objsize to determine the size of the destination object. */
4101 access_ref aref;
4102 if (!ptr_qry.get_ref (dest, stmt, &aref))
4103 return HOST_WIDE_INT_MAX;
4105 offset_int remsize = aref.size_remaining ();
4106 if (!wi::fits_uhwi_p (remsize))
4107 return HOST_WIDE_INT_MAX;
4109 return remsize.to_uhwi ();
4112 /* Return true if the call described by INFO with result RES safe to
4113 optimize (i.e., no undefined behavior), and set RETVAL to the range
4114 of its return values. */
4116 static bool
4117 is_call_safe (const call_info &info,
4118 const format_result &res, bool under4k,
4119 unsigned HOST_WIDE_INT retval[2])
4121 if (under4k && !res.posunder4k)
4122 return false;
4124 /* The minimum return value. */
4125 retval[0] = res.range.min;
4127 /* The maximum return value is in most cases bounded by RES.RANGE.MAX
4128 but in cases involving multibyte characters could be as large as
4129 RES.RANGE.UNLIKELY. */
4130 retval[1]
4131 = res.range.unlikely < res.range.max ? res.range.max : res.range.unlikely;
4133 /* Adjust the number of bytes which includes the terminating nul
4134 to reflect the return value of the function which does not.
4135 Because the valid range of the function is [INT_MIN, INT_MAX],
4136 a valid range before the adjustment below is [0, INT_MAX + 1]
4137 (the functions only return negative values on error or undefined
4138 behavior). */
4139 if (retval[0] <= target_int_max () + 1)
4140 --retval[0];
4141 if (retval[1] <= target_int_max () + 1)
4142 --retval[1];
4144 /* Avoid the return value optimization when the behavior of the call
4145 is undefined either because any directive may have produced 4K or
4146 more of output, or the return value exceeds INT_MAX, or because
4147 the output overflows the destination object (but leave it enabled
4148 when the function is bounded because then the behavior is well-
4149 defined). */
4150 if (retval[0] == retval[1]
4151 && (info.bounded || retval[0] < info.objsize)
4152 && retval[0] <= target_int_max ())
4153 return true;
4155 if ((info.bounded || retval[1] < info.objsize)
4156 && (retval[0] < target_int_max ()
4157 && retval[1] < target_int_max ()))
4158 return true;
4160 if (!under4k && (info.bounded || retval[0] < info.objsize))
4161 return true;
4163 return false;
4166 /* Given a suitable result RES of a call to a formatted output function
4167 described by INFO, substitute the result for the return value of
4168 the call. The result is suitable if the number of bytes it represents
4169 is known and exact. A result that isn't suitable for substitution may
4170 have its range set to the range of return values, if that is known.
4171 Return true if the call is removed and gsi_next should not be performed
4172 in the caller. */
4174 static bool
4175 try_substitute_return_value (gimple_stmt_iterator *gsi,
4176 const call_info &info,
4177 const format_result &res)
4179 tree lhs = gimple_get_lhs (info.callstmt);
4181 /* Set to true when the entire call has been removed. */
4182 bool removed = false;
4184 /* The minimum and maximum return value. */
4185 unsigned HOST_WIDE_INT retval[2] = {0};
4186 bool safe = is_call_safe (info, res, true, retval);
4188 if (safe
4189 && retval[0] == retval[1]
4190 /* Not prepared to handle possibly throwing calls here; they shouldn't
4191 appear in non-artificial testcases, except when the __*_chk routines
4192 are badly declared. */
4193 && !stmt_ends_bb_p (info.callstmt))
4195 tree cst = build_int_cst (lhs ? TREE_TYPE (lhs) : integer_type_node,
4196 retval[0]);
4198 if (lhs == NULL_TREE && info.nowrite)
4200 /* Remove the call to the bounded function with a zero size
4201 (e.g., snprintf(0, 0, "%i", 123)) if there is no lhs. */
4202 unlink_stmt_vdef (info.callstmt);
4203 gsi_remove (gsi, true);
4204 removed = true;
4206 else if (info.nowrite)
4208 /* Replace the call to the bounded function with a zero size
4209 (e.g., snprintf(0, 0, "%i", 123) with the constant result
4210 of the function. */
4211 gimplify_and_update_call_from_tree (gsi, cst);
4212 gimple *callstmt = gsi_stmt (*gsi);
4213 update_stmt (callstmt);
4215 else if (lhs)
4217 /* Replace the left-hand side of the call with the constant
4218 result of the formatted function. */
4219 gimple_call_set_lhs (info.callstmt, NULL_TREE);
4220 gimple *g = gimple_build_assign (lhs, cst);
4221 gsi_insert_after (gsi, g, GSI_NEW_STMT);
4222 update_stmt (info.callstmt);
4225 if (dump_file)
4227 if (removed)
4228 fprintf (dump_file, " Removing call statement.");
4229 else
4231 fprintf (dump_file, " Substituting ");
4232 print_generic_expr (dump_file, cst, dump_flags);
4233 fprintf (dump_file, " for %s.\n",
4234 info.nowrite ? "statement" : "return value");
4238 else if (lhs && types_compatible_p (TREE_TYPE (lhs), integer_type_node))
4240 bool setrange = false;
4242 if (safe
4243 && (info.bounded || retval[1] < info.objsize)
4244 && (retval[0] < target_int_max ()
4245 && retval[1] < target_int_max ()))
4247 /* If the result is in a valid range bounded by the size of
4248 the destination set it so that it can be used for subsequent
4249 optimizations. */
4250 int prec = TYPE_PRECISION (integer_type_node);
4252 wide_int min = wi::shwi (retval[0], prec);
4253 wide_int max = wi::shwi (retval[1], prec);
4254 value_range r (TREE_TYPE (lhs), min, max);
4255 set_range_info (lhs, r);
4257 setrange = true;
4260 if (dump_file)
4262 const char *inbounds
4263 = (retval[0] < info.objsize
4264 ? (retval[1] < info.objsize
4265 ? "in" : "potentially out-of")
4266 : "out-of");
4268 const char *what = setrange ? "Setting" : "Discarding";
4269 if (retval[0] != retval[1])
4270 fprintf (dump_file,
4271 " %s %s-bounds return value range ["
4272 HOST_WIDE_INT_PRINT_UNSIGNED ", "
4273 HOST_WIDE_INT_PRINT_UNSIGNED "].\n",
4274 what, inbounds, retval[0], retval[1]);
4275 else
4276 fprintf (dump_file, " %s %s-bounds return value "
4277 HOST_WIDE_INT_PRINT_UNSIGNED ".\n",
4278 what, inbounds, retval[0]);
4282 if (dump_file)
4283 fputc ('\n', dump_file);
4285 return removed;
4288 /* Try to simplify a s{,n}printf call described by INFO with result
4289 RES by replacing it with a simpler and presumably more efficient
4290 call (such as strcpy). */
4292 static bool
4293 try_simplify_call (gimple_stmt_iterator *gsi,
4294 const call_info &info,
4295 const format_result &res)
4297 unsigned HOST_WIDE_INT dummy[2];
4298 if (!is_call_safe (info, res, info.retval_used (), dummy))
4299 return false;
4301 switch (info.fncode)
4303 case BUILT_IN_SNPRINTF:
4304 return gimple_fold_builtin_snprintf (gsi);
4306 case BUILT_IN_SPRINTF:
4307 return gimple_fold_builtin_sprintf (gsi);
4309 default:
4313 return false;
4316 /* Return the zero-based index of the format string argument of a printf
4317 like function and set *IDX_ARGS to the first format argument. When
4318 no such index exists return UINT_MAX. */
4320 static unsigned
4321 get_user_idx_format (tree fndecl, unsigned *idx_args)
4323 tree attrs = lookup_attribute ("format", DECL_ATTRIBUTES (fndecl));
4324 if (!attrs)
4325 attrs = lookup_attribute ("format", TYPE_ATTRIBUTES (TREE_TYPE (fndecl)));
4327 if (!attrs)
4328 return UINT_MAX;
4330 attrs = TREE_VALUE (attrs);
4332 tree archetype = TREE_VALUE (attrs);
4333 if (strcmp ("printf", IDENTIFIER_POINTER (archetype)))
4334 return UINT_MAX;
4336 attrs = TREE_CHAIN (attrs);
4337 tree fmtarg = TREE_VALUE (attrs);
4339 attrs = TREE_CHAIN (attrs);
4340 tree elliparg = TREE_VALUE (attrs);
4342 /* Attribute argument indices are 1-based but we use zero-based. */
4343 *idx_args = tree_to_uhwi (elliparg) - 1;
4344 return tree_to_uhwi (fmtarg) - 1;
4347 } /* Unnamed namespace. */
4349 /* Determine if a GIMPLE call at *GSI is to one of the sprintf-like built-in
4350 functions and if so, handle it. Return true if the call is removed and
4351 gsi_next should not be performed in the caller. */
4353 bool
4354 handle_printf_call (gimple_stmt_iterator *gsi, pointer_query &ptr_qry)
4356 init_target_to_host_charmap ();
4358 call_info info = call_info ();
4360 info.callstmt = gsi_stmt (*gsi);
4361 info.func = gimple_call_fndecl (info.callstmt);
4362 if (!info.func)
4363 return false;
4365 /* Format string argument number (valid for all functions). */
4366 unsigned idx_format = UINT_MAX;
4367 if (gimple_call_builtin_p (info.callstmt, BUILT_IN_NORMAL))
4368 info.fncode = DECL_FUNCTION_CODE (info.func);
4369 else
4371 unsigned idx_args;
4372 idx_format = get_user_idx_format (info.func, &idx_args);
4373 if (idx_format == UINT_MAX
4374 || idx_format >= gimple_call_num_args (info.callstmt)
4375 || idx_args > gimple_call_num_args (info.callstmt)
4376 || !POINTER_TYPE_P (TREE_TYPE (gimple_call_arg (info.callstmt,
4377 idx_format))))
4378 return false;
4379 info.fncode = BUILT_IN_NONE;
4380 info.argidx = idx_args;
4383 /* The size of the destination as in snprintf(dest, size, ...). */
4384 unsigned HOST_WIDE_INT dstsize = HOST_WIDE_INT_M1U;
4386 /* The size of the destination determined by __builtin_object_size. */
4387 unsigned HOST_WIDE_INT objsize = HOST_WIDE_INT_M1U;
4389 /* Zero-based buffer size argument number (snprintf and vsnprintf). */
4390 unsigned idx_dstsize = UINT_MAX;
4392 /* Object size argument number (snprintf_chk and vsnprintf_chk). */
4393 unsigned idx_objsize = UINT_MAX;
4395 /* Destinaton argument number (valid for sprintf functions only). */
4396 unsigned idx_dstptr = 0;
4398 switch (info.fncode)
4400 case BUILT_IN_NONE:
4401 // User-defined function with attribute format (printf).
4402 idx_dstptr = -1;
4403 break;
4405 case BUILT_IN_FPRINTF:
4406 // Signature:
4407 // __builtin_fprintf (FILE*, format, ...)
4408 idx_format = 1;
4409 info.argidx = 2;
4410 idx_dstptr = -1;
4411 break;
4413 case BUILT_IN_FPRINTF_CHK:
4414 // Signature:
4415 // __builtin_fprintf_chk (FILE*, ost, format, ...)
4416 idx_format = 2;
4417 info.argidx = 3;
4418 idx_dstptr = -1;
4419 break;
4421 case BUILT_IN_FPRINTF_UNLOCKED:
4422 // Signature:
4423 // __builtin_fprintf_unnlocked (FILE*, format, ...)
4424 idx_format = 1;
4425 info.argidx = 2;
4426 idx_dstptr = -1;
4427 break;
4429 case BUILT_IN_PRINTF:
4430 // Signature:
4431 // __builtin_printf (format, ...)
4432 idx_format = 0;
4433 info.argidx = 1;
4434 idx_dstptr = -1;
4435 break;
4437 case BUILT_IN_PRINTF_CHK:
4438 // Signature:
4439 // __builtin_printf_chk (ost, format, ...)
4440 idx_format = 1;
4441 info.argidx = 2;
4442 idx_dstptr = -1;
4443 break;
4445 case BUILT_IN_PRINTF_UNLOCKED:
4446 // Signature:
4447 // __builtin_printf (format, ...)
4448 idx_format = 0;
4449 info.argidx = 1;
4450 idx_dstptr = -1;
4451 break;
4453 case BUILT_IN_SPRINTF:
4454 // Signature:
4455 // __builtin_sprintf (dst, format, ...)
4456 idx_format = 1;
4457 info.argidx = 2;
4458 break;
4460 case BUILT_IN_SPRINTF_CHK:
4461 // Signature:
4462 // __builtin___sprintf_chk (dst, ost, objsize, format, ...)
4463 idx_objsize = 2;
4464 idx_format = 3;
4465 info.argidx = 4;
4466 break;
4468 case BUILT_IN_SNPRINTF:
4469 // Signature:
4470 // __builtin_snprintf (dst, size, format, ...)
4471 idx_dstsize = 1;
4472 idx_format = 2;
4473 info.argidx = 3;
4474 info.bounded = true;
4475 break;
4477 case BUILT_IN_SNPRINTF_CHK:
4478 // Signature:
4479 // __builtin___snprintf_chk (dst, size, ost, objsize, format, ...)
4480 idx_dstsize = 1;
4481 idx_objsize = 3;
4482 idx_format = 4;
4483 info.argidx = 5;
4484 info.bounded = true;
4485 break;
4487 case BUILT_IN_VFPRINTF:
4488 // Signature:
4489 // __builtin_vprintf (FILE*, format, va_list)
4490 idx_format = 1;
4491 info.argidx = -1;
4492 idx_dstptr = -1;
4493 break;
4495 case BUILT_IN_VFPRINTF_CHK:
4496 // Signature:
4497 // __builtin___vfprintf_chk (FILE*, ost, format, va_list)
4498 idx_format = 2;
4499 info.argidx = -1;
4500 idx_dstptr = -1;
4501 break;
4503 case BUILT_IN_VPRINTF:
4504 // Signature:
4505 // __builtin_vprintf (format, va_list)
4506 idx_format = 0;
4507 info.argidx = -1;
4508 idx_dstptr = -1;
4509 break;
4511 case BUILT_IN_VPRINTF_CHK:
4512 // Signature:
4513 // __builtin___vprintf_chk (ost, format, va_list)
4514 idx_format = 1;
4515 info.argidx = -1;
4516 idx_dstptr = -1;
4517 break;
4519 case BUILT_IN_VSNPRINTF:
4520 // Signature:
4521 // __builtin_vsprintf (dst, size, format, va)
4522 idx_dstsize = 1;
4523 idx_format = 2;
4524 info.argidx = -1;
4525 info.bounded = true;
4526 break;
4528 case BUILT_IN_VSNPRINTF_CHK:
4529 // Signature:
4530 // __builtin___vsnprintf_chk (dst, size, ost, objsize, format, va)
4531 idx_dstsize = 1;
4532 idx_objsize = 3;
4533 idx_format = 4;
4534 info.argidx = -1;
4535 info.bounded = true;
4536 break;
4538 case BUILT_IN_VSPRINTF:
4539 // Signature:
4540 // __builtin_vsprintf (dst, format, va)
4541 idx_format = 1;
4542 info.argidx = -1;
4543 break;
4545 case BUILT_IN_VSPRINTF_CHK:
4546 // Signature:
4547 // __builtin___vsprintf_chk (dst, ost, objsize, format, va)
4548 idx_format = 3;
4549 idx_objsize = 2;
4550 info.argidx = -1;
4551 break;
4553 default:
4554 return false;
4557 /* Set the global warning level for this function. */
4558 warn_level = info.bounded ? warn_format_trunc : warn_format_overflow;
4560 /* For all string functions the first argument is a pointer to
4561 the destination. */
4562 tree dstptr = (idx_dstptr < gimple_call_num_args (info.callstmt)
4563 ? gimple_call_arg (info.callstmt, 0) : NULL_TREE);
4565 info.format = gimple_call_arg (info.callstmt, idx_format);
4567 /* True when the destination size is constant as opposed to the lower
4568 or upper bound of a range. */
4569 bool dstsize_cst_p = true;
4570 bool posunder4k = true;
4572 if (idx_dstsize == UINT_MAX)
4574 /* For non-bounded functions like sprintf, determine the size
4575 of the destination from the object or pointer passed to it
4576 as the first argument. */
4577 dstsize = get_destination_size (dstptr, info.callstmt, ptr_qry);
4579 else if (tree size = gimple_call_arg (info.callstmt, idx_dstsize))
4581 /* For bounded functions try to get the size argument. */
4583 if (TREE_CODE (size) == INTEGER_CST)
4585 dstsize = tree_to_uhwi (size);
4586 /* No object can be larger than SIZE_MAX bytes (half the address
4587 space) on the target.
4588 The functions are defined only for output of at most INT_MAX
4589 bytes. Specifying a bound in excess of that limit effectively
4590 defeats the bounds checking (and on some implementations such
4591 as Solaris cause the function to fail with EINVAL). */
4592 if (dstsize > target_size_max () / 2)
4594 /* Avoid warning if -Wstringop-overflow is specified since
4595 it also warns for the same thing though only for the
4596 checking built-ins. */
4597 if ((idx_objsize == UINT_MAX
4598 || !warn_stringop_overflow))
4599 warning_at (gimple_location (info.callstmt), info.warnopt (),
4600 "specified bound %wu exceeds maximum object size "
4601 "%wu",
4602 dstsize, target_size_max () / 2);
4603 /* POSIX requires snprintf to fail if DSTSIZE is greater
4604 than INT_MAX. Even though not all POSIX implementations
4605 conform to the requirement, avoid folding in this case. */
4606 posunder4k = false;
4608 else if (dstsize > target_int_max ())
4610 warning_at (gimple_location (info.callstmt), info.warnopt (),
4611 "specified bound %wu exceeds %<INT_MAX%>",
4612 dstsize);
4613 /* POSIX requires snprintf to fail if DSTSIZE is greater
4614 than INT_MAX. Avoid folding in that case. */
4615 posunder4k = false;
4618 else if (TREE_CODE (size) == SSA_NAME)
4620 /* Try to determine the range of values of the argument
4621 and use the greater of the two at level 1 and the smaller
4622 of them at level 2. */
4623 value_range vr;
4624 ptr_qry.rvals->range_of_expr (vr, size, info.callstmt);
4626 if (!vr.undefined_p ())
4628 tree type = TREE_TYPE (size);
4629 tree tmin = wide_int_to_tree (type, vr.lower_bound ());
4630 tree tmax = wide_int_to_tree (type, vr.upper_bound ());
4631 unsigned HOST_WIDE_INT minsize = TREE_INT_CST_LOW (tmin);
4632 unsigned HOST_WIDE_INT maxsize = TREE_INT_CST_LOW (tmax);
4633 dstsize = warn_level < 2 ? maxsize : minsize;
4635 if (minsize > target_int_max ())
4636 warning_at (gimple_location (info.callstmt), info.warnopt (),
4637 "specified bound range [%wu, %wu] exceeds "
4638 "%<INT_MAX%>",
4639 minsize, maxsize);
4641 /* POSIX requires snprintf to fail if DSTSIZE is greater
4642 than INT_MAX. Avoid folding if that's possible. */
4643 if (maxsize > target_int_max ())
4644 posunder4k = false;
4647 /* The destination size is not constant. If the function is
4648 bounded (e.g., snprintf) a lower bound of zero doesn't
4649 necessarily imply it can be eliminated. */
4650 dstsize_cst_p = false;
4654 if (idx_objsize != UINT_MAX)
4655 if (tree size = gimple_call_arg (info.callstmt, idx_objsize))
4656 if (tree_fits_uhwi_p (size))
4657 objsize = tree_to_uhwi (size);
4659 if (info.bounded && !dstsize)
4661 /* As a special case, when the explicitly specified destination
4662 size argument (to a bounded function like snprintf) is zero
4663 it is a request to determine the number of bytes on output
4664 without actually producing any. Pretend the size is
4665 unlimited in this case. */
4666 info.objsize = HOST_WIDE_INT_MAX;
4667 info.nowrite = dstsize_cst_p;
4669 else
4671 /* For calls to non-bounded functions or to those of bounded
4672 functions with a non-zero size, warn if the destination
4673 pointer is null. */
4674 if (dstptr && integer_zerop (dstptr))
4676 /* This is diagnosed with -Wformat only when the null is a constant
4677 pointer. The warning here diagnoses instances where the pointer
4678 is not constant. */
4679 location_t loc = gimple_location (info.callstmt);
4680 warning_at (EXPR_LOC_OR_LOC (dstptr, loc),
4681 info.warnopt (), "null destination pointer");
4682 return false;
4685 /* Set the object size to the smaller of the two arguments
4686 of both have been specified and they're not equal. */
4687 info.objsize = dstsize < objsize ? dstsize : objsize;
4689 if (info.bounded
4690 && dstsize < target_size_max () / 2 && objsize < dstsize
4691 /* Avoid warning if -Wstringop-overflow is specified since
4692 it also warns for the same thing though only for the
4693 checking built-ins. */
4694 && (idx_objsize == UINT_MAX
4695 || !warn_stringop_overflow))
4697 warning_at (gimple_location (info.callstmt), info.warnopt (),
4698 "specified bound %wu exceeds the size %wu "
4699 "of the destination object", dstsize, objsize);
4703 /* Determine if the format argument may be null and warn if not
4704 and if the argument is null. */
4705 if (integer_zerop (info.format)
4706 && gimple_call_builtin_p (info.callstmt, BUILT_IN_NORMAL))
4708 location_t loc = gimple_location (info.callstmt);
4709 warning_at (EXPR_LOC_OR_LOC (info.format, loc),
4710 info.warnopt (), "null format string");
4711 return false;
4714 info.fmtstr = get_format_string (info.format, &info.fmtloc);
4715 if (!info.fmtstr)
4716 return false;
4718 if (warn_restrict)
4720 /* Compute the origin of the destination pointer and its offset
4721 from the base object/pointer if possible. */
4722 info.dst_offset = 0;
4723 info.dst_origin = get_origin_and_offset (dstptr, &info.dst_field,
4724 &info.dst_offset);
4727 /* The result is the number of bytes output by the formatted function,
4728 including the terminating NUL. */
4729 format_result res;
4731 /* I/O functions with no destination argument (i.e., all forms of fprintf
4732 and printf) may fail under any conditions. Others (i.e., all forms of
4733 sprintf) may only fail under specific conditions determined for each
4734 directive. Clear POSUNDER4K for the former set of functions and set
4735 it to true for the latter (it can only be cleared later, but it is
4736 never set to true again). */
4737 res.posunder4k = posunder4k && dstptr;
4739 bool success = compute_format_length (info, &res, ptr_qry);
4740 if (res.warned)
4741 suppress_warning (info.callstmt, info.warnopt ());
4743 /* When optimizing and the printf return value optimization is enabled,
4744 attempt to substitute the computed result for the return value of
4745 the call. Avoid this optimization when -frounding-math is in effect
4746 and the format string contains a floating point directive. */
4747 bool call_removed = false;
4748 if (success && optimize > 0)
4750 /* Save a copy of the iterator pointing at the call. The iterator
4751 may change to point past the call in try_substitute_return_value
4752 but the original value is needed in try_simplify_call. */
4753 gimple_stmt_iterator gsi_call = *gsi;
4755 if (flag_printf_return_value
4756 && (!flag_rounding_math || !res.floating))
4757 call_removed = try_substitute_return_value (gsi, info, res);
4759 if (!call_removed)
4760 try_simplify_call (&gsi_call, info, res);
4763 return call_removed;