Skip gcc.dg/guality/example.c on hppa-linux.
[official-gcc.git] / gcc / gimple-ssa-sprintf.c
blobec78090b382ef8060fda9326c48b39f3f07680e3
1 /* Copyright (C) 2016-2021 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-fold.h"
57 #include "gimple-pretty-print.h"
58 #include "diagnostic-core.h"
59 #include "fold-const.h"
60 #include "gimple-iterator.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 8:
539 return (prec + 2) / 3;
540 case 10:
541 /* Decimal approximation: yields 3, 5, 10, and 20 for precision
542 of 8, 16, 32, and 64 bits. */
543 return prec * 301 / 1000 + 1;
544 case 16:
545 return prec / 4;
548 gcc_unreachable ();
551 static bool
552 get_int_range (tree, gimple *, HOST_WIDE_INT *, HOST_WIDE_INT *,
553 bool, HOST_WIDE_INT, range_query *);
555 struct call_info;
557 /* Description of a format directive. A directive is either a plain
558 string or a conversion specification that starts with '%'. */
560 struct directive
562 directive (const call_info *inf, unsigned dno)
563 : info (inf), dirno (dno), argno (), beg (), len (), flags (),
564 width (), prec (), modifier (), specifier (), arg (), fmtfunc ()
567 /* Reference to the info structure describing the call that this
568 directive is a part of. */
569 const call_info *info;
571 /* The 1-based directive number (for debugging). */
572 unsigned dirno;
574 /* The zero-based argument number of the directive's argument ARG in
575 the function's argument list. */
576 unsigned argno;
578 /* The first character of the directive and its length. */
579 const char *beg;
580 size_t len;
582 /* A bitmap of flags, one for each character. */
583 unsigned flags[256 / sizeof (int)];
585 /* The range of values of the specified width, or -1 if not specified. */
586 HOST_WIDE_INT width[2];
587 /* The range of values of the specified precision, or -1 if not
588 specified. */
589 HOST_WIDE_INT prec[2];
591 /* Length modifier. */
592 format_lengths modifier;
594 /* Format specifier character. */
595 char specifier;
597 /* The argument of the directive or null when the directive doesn't
598 take one or when none is available (such as for vararg functions). */
599 tree arg;
601 /* Format conversion function that given a directive and an argument
602 returns the formatting result. */
603 fmtresult (*fmtfunc) (const directive &, tree, range_query *);
605 /* Return True when the format flag CHR has been used. */
606 bool get_flag (char chr) const
608 unsigned char c = chr & 0xff;
609 return (flags[c / (CHAR_BIT * sizeof *flags)]
610 & (1U << (c % (CHAR_BIT * sizeof *flags))));
613 /* Make a record of the format flag CHR having been used. */
614 void set_flag (char chr)
616 unsigned char c = chr & 0xff;
617 flags[c / (CHAR_BIT * sizeof *flags)]
618 |= (1U << (c % (CHAR_BIT * sizeof *flags)));
621 /* Reset the format flag CHR. */
622 void clear_flag (char chr)
624 unsigned char c = chr & 0xff;
625 flags[c / (CHAR_BIT * sizeof *flags)]
626 &= ~(1U << (c % (CHAR_BIT * sizeof *flags)));
629 /* Set both bounds of the width range to VAL. */
630 void set_width (HOST_WIDE_INT val)
632 width[0] = width[1] = val;
635 /* Set the width range according to ARG, with both bounds being
636 no less than 0. For a constant ARG set both bounds to its value
637 or 0, whichever is greater. For a non-constant ARG in some range
638 set width to its range adjusting each bound to -1 if it's less.
639 For an indeterminate ARG set width to [0, INT_MAX]. */
640 void set_width (tree arg, range_query *);
642 /* Set both bounds of the precision range to VAL. */
643 void set_precision (HOST_WIDE_INT val)
645 prec[0] = prec[1] = val;
648 /* Set the precision range according to ARG, with both bounds being
649 no less than -1. For a constant ARG set both bounds to its value
650 or -1 whichever is greater. For a non-constant ARG in some range
651 set precision to its range adjusting each bound to -1 if it's less.
652 For an indeterminate ARG set precision to [-1, INT_MAX]. */
653 void set_precision (tree arg, range_query *query);
655 /* Return true if both width and precision are known to be
656 either constant or in some range, false otherwise. */
657 bool known_width_and_precision () const
659 return ((width[1] < 0
660 || (unsigned HOST_WIDE_INT)width[1] <= target_int_max ())
661 && (prec[1] < 0
662 || (unsigned HOST_WIDE_INT)prec[1] < target_int_max ()));
666 /* The result of a call to a formatted function. */
668 struct format_result
670 format_result ()
671 : range (), aliases (), alias_count (), knownrange (), posunder4k (),
672 floating (), warned () { /* No-op. */ }
674 ~format_result ()
676 XDELETEVEC (aliases);
679 /* Range of characters written by the formatted function.
680 Setting the minimum to HOST_WIDE_INT_MAX disables all
681 length tracking for the remainder of the format string. */
682 result_range range;
684 struct alias_info
686 directive dir; /* The directive that aliases the destination. */
687 HOST_WIDE_INT offset; /* The offset at which it aliases it. */
688 result_range range; /* The raw result of the directive. */
691 /* An array of directives whose pointer argument aliases a part
692 of the destination object of the formatted function. */
693 alias_info *aliases;
694 unsigned alias_count;
696 /* True when the range above is obtained from known values of
697 directive arguments, or bounds on the amount of output such
698 as width and precision, and not the result of heuristics that
699 depend on warning levels. It's used to issue stricter diagnostics
700 in cases where strings of unknown lengths are bounded by the arrays
701 they are determined to refer to. KNOWNRANGE must not be used for
702 the return value optimization. */
703 bool knownrange;
705 /* True if no individual directive could fail or result in more than
706 4095 bytes of output (the total NUMBER_CHARS_{MIN,MAX} might be
707 greater). Implementations are not required to handle directives
708 that produce more than 4K bytes (leading to undefined behavior)
709 and so when one is found it disables the return value optimization.
710 Similarly, directives that can fail (such as wide character
711 directives) disable the optimization. */
712 bool posunder4k;
714 /* True when a floating point directive has been seen in the format
715 string. */
716 bool floating;
718 /* True when an intermediate result has caused a warning. Used to
719 avoid issuing duplicate warnings while finishing the processing
720 of a call. WARNED also disables the return value optimization. */
721 bool warned;
723 /* Preincrement the number of output characters by 1. */
724 format_result& operator++ ()
726 return *this += 1;
729 /* Postincrement the number of output characters by 1. */
730 format_result operator++ (int)
732 format_result prev (*this);
733 *this += 1;
734 return prev;
737 /* Increment the number of output characters by N. */
738 format_result& operator+= (unsigned HOST_WIDE_INT);
740 /* Add a directive to the sequence of those with potentially aliasing
741 arguments. */
742 void append_alias (const directive &, HOST_WIDE_INT, const result_range &);
744 private:
745 /* Not copyable or assignable. */
746 format_result (format_result&);
747 void operator= (format_result&);
750 format_result&
751 format_result::operator+= (unsigned HOST_WIDE_INT n)
753 gcc_assert (n < HOST_WIDE_INT_MAX);
755 if (range.min < HOST_WIDE_INT_MAX)
756 range.min += n;
758 if (range.max < HOST_WIDE_INT_MAX)
759 range.max += n;
761 if (range.likely < HOST_WIDE_INT_MAX)
762 range.likely += n;
764 if (range.unlikely < HOST_WIDE_INT_MAX)
765 range.unlikely += n;
767 return *this;
770 void
771 format_result::append_alias (const directive &d, HOST_WIDE_INT off,
772 const result_range &resrng)
774 unsigned cnt = alias_count + 1;
775 alias_info *ar = XNEWVEC (alias_info, cnt);
777 for (unsigned i = 0; i != alias_count; ++i)
778 ar[i] = aliases[i];
780 ar[alias_count].dir = d;
781 ar[alias_count].offset = off;
782 ar[alias_count].range = resrng;
784 XDELETEVEC (aliases);
786 alias_count = cnt;
787 aliases = ar;
790 /* Return the logarithm of X in BASE. */
792 static int
793 ilog (unsigned HOST_WIDE_INT x, int base)
795 int res = 0;
798 ++res;
799 x /= base;
800 } while (x);
801 return res;
804 /* Return the number of bytes resulting from converting into a string
805 the INTEGER_CST tree node X in BASE with a minimum of PREC digits.
806 PLUS indicates whether 1 for a plus sign should be added for positive
807 numbers, and PREFIX whether the length of an octal ('O') or hexadecimal
808 ('0x') prefix should be added for nonzero numbers. Return -1 if X cannot
809 be represented. */
811 static HOST_WIDE_INT
812 tree_digits (tree x, int base, HOST_WIDE_INT prec, bool plus, bool prefix)
814 unsigned HOST_WIDE_INT absval;
816 HOST_WIDE_INT res;
818 if (TYPE_UNSIGNED (TREE_TYPE (x)))
820 if (tree_fits_uhwi_p (x))
822 absval = tree_to_uhwi (x);
823 res = plus;
825 else
826 return -1;
828 else
830 if (tree_fits_shwi_p (x))
832 HOST_WIDE_INT i = tree_to_shwi (x);
833 if (HOST_WIDE_INT_MIN == i)
835 /* Avoid undefined behavior due to negating a minimum. */
836 absval = HOST_WIDE_INT_MAX;
837 res = 1;
839 else if (i < 0)
841 absval = -i;
842 res = 1;
844 else
846 absval = i;
847 res = plus;
850 else
851 return -1;
854 int ndigs = ilog (absval, base);
856 res += prec < ndigs ? ndigs : prec;
858 /* Adjust a non-zero value for the base prefix, either hexadecimal,
859 or, unless precision has resulted in a leading zero, also octal. */
860 if (prefix && absval && (base == 16 || prec <= ndigs))
862 if (base == 8)
863 res += 1;
864 else if (base == 16)
865 res += 2;
868 return res;
871 /* Description of a call to a formatted function. */
873 struct call_info
875 /* Function call statement. */
876 gimple *callstmt;
878 /* Function called. */
879 tree func;
881 /* Called built-in function code. */
882 built_in_function fncode;
884 /* The "origin" of the destination pointer argument, which is either
885 the DECL of the destination buffer being written into or a pointer
886 that points to it, plus some offset. */
887 tree dst_origin;
889 /* For a destination pointing to a struct array member, the offset of
890 the member. */
891 HOST_WIDE_INT dst_field;
893 /* The offset into the destination buffer. */
894 HOST_WIDE_INT dst_offset;
896 /* Format argument and format string extracted from it. */
897 tree format;
898 const char *fmtstr;
900 /* The location of the format argument. */
901 location_t fmtloc;
903 /* The destination object size for __builtin___xxx_chk functions
904 typically determined by __builtin_object_size, or -1 if unknown. */
905 unsigned HOST_WIDE_INT objsize;
907 /* Number of the first variable argument. */
908 unsigned HOST_WIDE_INT argidx;
910 /* True for functions like snprintf that specify the size of
911 the destination, false for others like sprintf that don't. */
912 bool bounded;
914 /* True for bounded functions like snprintf that specify a zero-size
915 buffer as a request to compute the size of output without actually
916 writing any. NOWRITE is cleared in response to the %n directive
917 which has side-effects similar to writing output. */
918 bool nowrite;
920 /* Return true if the called function's return value is used. */
921 bool retval_used () const
923 return gimple_get_lhs (callstmt);
926 /* Return the warning option corresponding to the called function. */
927 opt_code warnopt () const
929 return bounded ? OPT_Wformat_truncation_ : OPT_Wformat_overflow_;
932 /* Return true for calls to file formatted functions. */
933 bool is_file_func () const
935 return (fncode == BUILT_IN_FPRINTF
936 || fncode == BUILT_IN_FPRINTF_CHK
937 || fncode == BUILT_IN_FPRINTF_UNLOCKED
938 || fncode == BUILT_IN_VFPRINTF
939 || fncode == BUILT_IN_VFPRINTF_CHK);
942 /* Return true for calls to string formatted functions. */
943 bool is_string_func () const
945 return (fncode == BUILT_IN_SPRINTF
946 || fncode == BUILT_IN_SPRINTF_CHK
947 || fncode == BUILT_IN_SNPRINTF
948 || fncode == BUILT_IN_SNPRINTF_CHK
949 || fncode == BUILT_IN_VSPRINTF
950 || fncode == BUILT_IN_VSPRINTF_CHK
951 || fncode == BUILT_IN_VSNPRINTF
952 || fncode == BUILT_IN_VSNPRINTF_CHK);
956 void
957 directive::set_width (tree arg, range_query *query)
959 get_int_range (arg, info->callstmt, width, width + 1, true, 0, query);
962 void
963 directive::set_precision (tree arg, range_query *query)
965 get_int_range (arg, info->callstmt, prec, prec + 1, false, -1, query);
968 /* Return the result of formatting a no-op directive (such as '%n'). */
970 static fmtresult
971 format_none (const directive &, tree, range_query *)
973 fmtresult res (0);
974 return res;
977 /* Return the result of formatting the '%%' directive. */
979 static fmtresult
980 format_percent (const directive &, tree, range_query *)
982 fmtresult res (1);
983 return res;
987 /* Compute intmax_type_node and uintmax_type_node similarly to how
988 tree.c builds size_type_node. */
990 static void
991 build_intmax_type_nodes (tree *pintmax, tree *puintmax)
993 if (strcmp (UINTMAX_TYPE, "unsigned int") == 0)
995 *pintmax = integer_type_node;
996 *puintmax = unsigned_type_node;
998 else if (strcmp (UINTMAX_TYPE, "long unsigned int") == 0)
1000 *pintmax = long_integer_type_node;
1001 *puintmax = long_unsigned_type_node;
1003 else if (strcmp (UINTMAX_TYPE, "long long unsigned int") == 0)
1005 *pintmax = long_long_integer_type_node;
1006 *puintmax = long_long_unsigned_type_node;
1008 else
1010 for (int i = 0; i < NUM_INT_N_ENTS; i++)
1011 if (int_n_enabled_p[i])
1013 char name[50], altname[50];
1014 sprintf (name, "__int%d unsigned", int_n_data[i].bitsize);
1015 sprintf (altname, "__int%d__ unsigned", int_n_data[i].bitsize);
1017 if (strcmp (name, UINTMAX_TYPE) == 0
1018 || strcmp (altname, UINTMAX_TYPE) == 0)
1020 *pintmax = int_n_trees[i].signed_type;
1021 *puintmax = int_n_trees[i].unsigned_type;
1022 return;
1025 gcc_unreachable ();
1029 /* Determine the range [*PMIN, *PMAX] that the expression ARG is
1030 in and that is representable in type int.
1031 Return true when the range is a subrange of that of int.
1032 When ARG is null it is as if it had the full range of int.
1033 When ABSOLUTE is true the range reflects the absolute value of
1034 the argument. When ABSOLUTE is false, negative bounds of
1035 the determined range are replaced with NEGBOUND. */
1037 static bool
1038 get_int_range (tree arg, gimple *stmt,
1039 HOST_WIDE_INT *pmin, HOST_WIDE_INT *pmax,
1040 bool absolute, HOST_WIDE_INT negbound,
1041 range_query *query)
1043 /* The type of the result. */
1044 const_tree type = integer_type_node;
1046 bool knownrange = false;
1048 if (!arg)
1050 *pmin = tree_to_shwi (TYPE_MIN_VALUE (type));
1051 *pmax = tree_to_shwi (TYPE_MAX_VALUE (type));
1053 else if (TREE_CODE (arg) == INTEGER_CST
1054 && TYPE_PRECISION (TREE_TYPE (arg)) <= TYPE_PRECISION (type))
1056 /* For a constant argument return its value adjusted as specified
1057 by NEGATIVE and NEGBOUND and return true to indicate that the
1058 result is known. */
1059 *pmin = tree_fits_shwi_p (arg) ? tree_to_shwi (arg) : tree_to_uhwi (arg);
1060 *pmax = *pmin;
1061 knownrange = true;
1063 else
1065 /* True if the argument's range cannot be determined. */
1066 bool unknown = true;
1068 tree argtype = TREE_TYPE (arg);
1070 /* Ignore invalid arguments with greater precision that that
1071 of the expected type (e.g., in sprintf("%*i", 12LL, i)).
1072 They will have been detected and diagnosed by -Wformat and
1073 so it's not important to complicate this code to try to deal
1074 with them again. */
1075 if (TREE_CODE (arg) == SSA_NAME
1076 && INTEGRAL_TYPE_P (argtype)
1077 && TYPE_PRECISION (argtype) <= TYPE_PRECISION (type))
1079 /* Try to determine the range of values of the integer argument. */
1080 value_range vr;
1081 query->range_of_expr (vr, arg, stmt);
1083 if (!vr.undefined_p () && !vr.varying_p ())
1085 HOST_WIDE_INT type_min
1086 = (TYPE_UNSIGNED (argtype)
1087 ? tree_to_uhwi (TYPE_MIN_VALUE (argtype))
1088 : tree_to_shwi (TYPE_MIN_VALUE (argtype)));
1090 HOST_WIDE_INT type_max = tree_to_uhwi (TYPE_MAX_VALUE (argtype));
1092 tree type = TREE_TYPE (arg);
1093 tree tmin = wide_int_to_tree (type, vr.lower_bound ());
1094 tree tmax = wide_int_to_tree (type, vr.upper_bound ());
1095 *pmin = TREE_INT_CST_LOW (tmin);
1096 *pmax = TREE_INT_CST_LOW (tmax);
1098 if (*pmin < *pmax)
1100 /* Return true if the adjusted range is a subrange of
1101 the full range of the argument's type. *PMAX may
1102 be less than *PMIN when the argument is unsigned
1103 and its upper bound is in excess of TYPE_MAX. In
1104 that (invalid) case disregard the range and use that
1105 of the expected type instead. */
1106 knownrange = type_min < *pmin || *pmax < type_max;
1108 unknown = false;
1113 /* Handle an argument with an unknown range as if none had been
1114 provided. */
1115 if (unknown)
1116 return get_int_range (NULL_TREE, NULL, pmin, pmax, absolute,
1117 negbound, query);
1120 /* Adjust each bound as specified by ABSOLUTE and NEGBOUND. */
1121 if (absolute)
1123 if (*pmin < 0)
1125 if (*pmin == *pmax)
1126 *pmin = *pmax = -*pmin;
1127 else
1129 /* Make sure signed overlow is avoided. */
1130 gcc_assert (*pmin != HOST_WIDE_INT_MIN);
1132 HOST_WIDE_INT tmp = -*pmin;
1133 *pmin = 0;
1134 if (*pmax < tmp)
1135 *pmax = tmp;
1139 else if (*pmin < negbound)
1140 *pmin = negbound;
1142 return knownrange;
1145 /* With the range [*ARGMIN, *ARGMAX] of an integer directive's actual
1146 argument, due to the conversion from either *ARGMIN or *ARGMAX to
1147 the type of the directive's formal argument it's possible for both
1148 to result in the same number of bytes or a range of bytes that's
1149 less than the number of bytes that would result from formatting
1150 some other value in the range [*ARGMIN, *ARGMAX]. This can be
1151 determined by checking for the actual argument being in the range
1152 of the type of the directive. If it isn't it must be assumed to
1153 take on the full range of the directive's type.
1154 Return true when the range has been adjusted to the full range
1155 of DIRTYPE, and false otherwise. */
1157 static bool
1158 adjust_range_for_overflow (tree dirtype, tree *argmin, tree *argmax)
1160 tree argtype = TREE_TYPE (*argmin);
1161 unsigned argprec = TYPE_PRECISION (argtype);
1162 unsigned dirprec = TYPE_PRECISION (dirtype);
1164 /* If the actual argument and the directive's argument have the same
1165 precision and sign there can be no overflow and so there is nothing
1166 to adjust. */
1167 if (argprec == dirprec && TYPE_SIGN (argtype) == TYPE_SIGN (dirtype))
1168 return false;
1170 /* The logic below was inspired/lifted from the CONVERT_EXPR_CODE_P
1171 branch in the extract_range_from_unary_expr function in tree-vrp.c. */
1173 if (TREE_CODE (*argmin) == INTEGER_CST
1174 && TREE_CODE (*argmax) == INTEGER_CST
1175 && (dirprec >= argprec
1176 || integer_zerop (int_const_binop (RSHIFT_EXPR,
1177 int_const_binop (MINUS_EXPR,
1178 *argmax,
1179 *argmin),
1180 size_int (dirprec)))))
1182 *argmin = force_fit_type (dirtype, wi::to_widest (*argmin), 0, false);
1183 *argmax = force_fit_type (dirtype, wi::to_widest (*argmax), 0, false);
1185 /* If *ARGMIN is still less than *ARGMAX the conversion above
1186 is safe. Otherwise, it has overflowed and would be unsafe. */
1187 if (tree_int_cst_le (*argmin, *argmax))
1188 return false;
1191 *argmin = TYPE_MIN_VALUE (dirtype);
1192 *argmax = TYPE_MAX_VALUE (dirtype);
1193 return true;
1196 /* Return a range representing the minimum and maximum number of bytes
1197 that the format directive DIR will output for any argument given
1198 the WIDTH and PRECISION (extracted from DIR). This function is
1199 used when the directive argument or its value isn't known. */
1201 static fmtresult
1202 format_integer (const directive &dir, tree arg, range_query *query)
1204 tree intmax_type_node;
1205 tree uintmax_type_node;
1207 /* Base to format the number in. */
1208 int base;
1210 /* True when a conversion is preceded by a prefix indicating the base
1211 of the argument (octal or hexadecimal). */
1212 bool maybebase = dir.get_flag ('#');
1214 /* True when a signed conversion is preceded by a sign or space. */
1215 bool maybesign = false;
1217 /* True for signed conversions (i.e., 'd' and 'i'). */
1218 bool sign = false;
1220 switch (dir.specifier)
1222 case 'd':
1223 case 'i':
1224 /* Space and '+' are only meaningful for signed conversions. */
1225 maybesign = dir.get_flag (' ') | dir.get_flag ('+');
1226 sign = true;
1227 base = 10;
1228 break;
1229 case 'u':
1230 base = 10;
1231 break;
1232 case 'o':
1233 base = 8;
1234 break;
1235 case 'X':
1236 case 'x':
1237 base = 16;
1238 break;
1239 default:
1240 gcc_unreachable ();
1243 /* The type of the "formal" argument expected by the directive. */
1244 tree dirtype = NULL_TREE;
1246 /* Determine the expected type of the argument from the length
1247 modifier. */
1248 switch (dir.modifier)
1250 case FMT_LEN_none:
1251 if (dir.specifier == 'p')
1252 dirtype = ptr_type_node;
1253 else
1254 dirtype = sign ? integer_type_node : unsigned_type_node;
1255 break;
1257 case FMT_LEN_h:
1258 dirtype = sign ? short_integer_type_node : short_unsigned_type_node;
1259 break;
1261 case FMT_LEN_hh:
1262 dirtype = sign ? signed_char_type_node : unsigned_char_type_node;
1263 break;
1265 case FMT_LEN_l:
1266 dirtype = sign ? long_integer_type_node : long_unsigned_type_node;
1267 break;
1269 case FMT_LEN_L:
1270 case FMT_LEN_ll:
1271 dirtype = (sign
1272 ? long_long_integer_type_node
1273 : long_long_unsigned_type_node);
1274 break;
1276 case FMT_LEN_z:
1277 dirtype = signed_or_unsigned_type_for (!sign, size_type_node);
1278 break;
1280 case FMT_LEN_t:
1281 dirtype = signed_or_unsigned_type_for (!sign, ptrdiff_type_node);
1282 break;
1284 case FMT_LEN_j:
1285 build_intmax_type_nodes (&intmax_type_node, &uintmax_type_node);
1286 dirtype = sign ? intmax_type_node : uintmax_type_node;
1287 break;
1289 default:
1290 return fmtresult ();
1293 /* The type of the argument to the directive, either deduced from
1294 the actual non-constant argument if one is known, or from
1295 the directive itself when none has been provided because it's
1296 a va_list. */
1297 tree argtype = NULL_TREE;
1299 if (!arg)
1301 /* When the argument has not been provided, use the type of
1302 the directive's argument as an approximation. This will
1303 result in false positives for directives like %i with
1304 arguments with smaller precision (such as short or char). */
1305 argtype = dirtype;
1307 else if (TREE_CODE (arg) == INTEGER_CST)
1309 /* When a constant argument has been provided use its value
1310 rather than type to determine the length of the output. */
1311 fmtresult res;
1313 if ((dir.prec[0] <= 0 && dir.prec[1] >= 0) && integer_zerop (arg))
1315 /* As a special case, a precision of zero with a zero argument
1316 results in zero bytes except in base 8 when the '#' flag is
1317 specified, and for signed conversions in base 8 and 10 when
1318 either the space or '+' flag has been specified and it results
1319 in just one byte (with width having the normal effect). This
1320 must extend to the case of a specified precision with
1321 an unknown value because it can be zero. */
1322 res.range.min = ((base == 8 && dir.get_flag ('#')) || maybesign);
1323 if (res.range.min == 0 && dir.prec[0] != dir.prec[1])
1325 res.range.max = 1;
1326 res.range.likely = 1;
1328 else
1330 res.range.max = res.range.min;
1331 res.range.likely = res.range.min;
1334 else
1336 /* Convert the argument to the type of the directive. */
1337 arg = fold_convert (dirtype, arg);
1339 res.range.min = tree_digits (arg, base, dir.prec[0],
1340 maybesign, maybebase);
1341 if (dir.prec[0] == dir.prec[1])
1342 res.range.max = res.range.min;
1343 else
1344 res.range.max = tree_digits (arg, base, dir.prec[1],
1345 maybesign, maybebase);
1346 res.range.likely = res.range.min;
1347 res.knownrange = true;
1350 res.range.unlikely = res.range.max;
1352 /* Bump up the counters if WIDTH is greater than LEN. */
1353 res.adjust_for_width_or_precision (dir.width, dirtype, base,
1354 (sign | maybebase) + (base == 16));
1355 /* Bump up the counters again if PRECision is greater still. */
1356 res.adjust_for_width_or_precision (dir.prec, dirtype, base,
1357 (sign | maybebase) + (base == 16));
1359 return res;
1361 else if (INTEGRAL_TYPE_P (TREE_TYPE (arg))
1362 || TREE_CODE (TREE_TYPE (arg)) == POINTER_TYPE)
1363 /* Determine the type of the provided non-constant argument. */
1364 argtype = TREE_TYPE (arg);
1365 else
1366 /* Don't bother with invalid arguments since they likely would
1367 have already been diagnosed, and disable any further checking
1368 of the format string by returning [-1, -1]. */
1369 return fmtresult ();
1371 fmtresult res;
1373 /* Using either the range the non-constant argument is in, or its
1374 type (either "formal" or actual), create a range of values that
1375 constrain the length of output given the warning level. */
1376 tree argmin = NULL_TREE;
1377 tree argmax = NULL_TREE;
1379 if (arg
1380 && TREE_CODE (arg) == SSA_NAME
1381 && INTEGRAL_TYPE_P (argtype))
1383 /* Try to determine the range of values of the integer argument
1384 (range information is not available for pointers). */
1385 value_range vr;
1386 query->range_of_expr (vr, arg, dir.info->callstmt);
1388 if (!vr.varying_p () && !vr.undefined_p ())
1390 argmin = wide_int_to_tree (TREE_TYPE (arg), vr.lower_bound ());
1391 argmax = wide_int_to_tree (TREE_TYPE (arg), vr.upper_bound ());
1393 /* Set KNOWNRANGE if the argument is in a known subrange
1394 of the directive's type and neither width nor precision
1395 is unknown. (KNOWNRANGE may be reset below). */
1396 res.knownrange
1397 = ((!tree_int_cst_equal (TYPE_MIN_VALUE (dirtype), argmin)
1398 || !tree_int_cst_equal (TYPE_MAX_VALUE (dirtype), argmax))
1399 && dir.known_width_and_precision ());
1401 res.argmin = argmin;
1402 res.argmax = argmax;
1404 else
1406 /* The argument here may be the result of promoting the actual
1407 argument to int. Try to determine the type of the actual
1408 argument before promotion and narrow down its range that
1409 way. */
1410 gimple *def = SSA_NAME_DEF_STMT (arg);
1411 if (is_gimple_assign (def))
1413 tree_code code = gimple_assign_rhs_code (def);
1414 if (code == INTEGER_CST)
1416 arg = gimple_assign_rhs1 (def);
1417 return format_integer (dir, arg, query);
1420 if (code == NOP_EXPR)
1422 tree type = TREE_TYPE (gimple_assign_rhs1 (def));
1423 if (INTEGRAL_TYPE_P (type)
1424 || TREE_CODE (type) == POINTER_TYPE)
1425 argtype = type;
1431 if (!argmin)
1433 if (TREE_CODE (argtype) == POINTER_TYPE)
1435 argmin = build_int_cst (pointer_sized_int_node, 0);
1436 argmax = build_all_ones_cst (pointer_sized_int_node);
1438 else
1440 argmin = TYPE_MIN_VALUE (argtype);
1441 argmax = TYPE_MAX_VALUE (argtype);
1445 /* Clear KNOWNRANGE if the range has been adjusted to the maximum
1446 of the directive. If it has been cleared then since ARGMIN and/or
1447 ARGMAX have been adjusted also adjust the corresponding ARGMIN and
1448 ARGMAX in the result to include in diagnostics. */
1449 if (adjust_range_for_overflow (dirtype, &argmin, &argmax))
1451 res.knownrange = false;
1452 res.argmin = argmin;
1453 res.argmax = argmax;
1456 /* Recursively compute the minimum and maximum from the known range. */
1457 if (TYPE_UNSIGNED (dirtype) || tree_int_cst_sgn (argmin) >= 0)
1459 /* For unsigned conversions/directives or signed when
1460 the minimum is positive, use the minimum and maximum to compute
1461 the shortest and longest output, respectively. */
1462 res.range.min = format_integer (dir, argmin, query).range.min;
1463 res.range.max = format_integer (dir, argmax, query).range.max;
1465 else if (tree_int_cst_sgn (argmax) < 0)
1467 /* For signed conversions/directives if maximum is negative,
1468 use the minimum as the longest output and maximum as the
1469 shortest output. */
1470 res.range.min = format_integer (dir, argmax, query).range.min;
1471 res.range.max = format_integer (dir, argmin, query).range.max;
1473 else
1475 /* Otherwise, 0 is inside of the range and minimum negative. Use 0
1476 as the shortest output and for the longest output compute the
1477 length of the output of both minimum and maximum and pick the
1478 longer. */
1479 unsigned HOST_WIDE_INT max1
1480 = format_integer (dir, argmin, query).range.max;
1481 unsigned HOST_WIDE_INT max2
1482 = format_integer (dir, argmax, query).range.max;
1483 res.range.min
1484 = format_integer (dir, integer_zero_node, query).range.min;
1485 res.range.max = MAX (max1, max2);
1488 /* If the range is known, use the maximum as the likely length. */
1489 if (res.knownrange)
1490 res.range.likely = res.range.max;
1491 else
1493 /* Otherwise, use the minimum. Except for the case where for %#x or
1494 %#o the minimum is just for a single value in the range (0) and
1495 for all other values it is something longer, like 0x1 or 01.
1496 Use the length for value 1 in that case instead as the likely
1497 length. */
1498 res.range.likely = res.range.min;
1499 if (maybebase
1500 && base != 10
1501 && (tree_int_cst_sgn (argmin) < 0 || tree_int_cst_sgn (argmax) > 0))
1503 if (res.range.min == 1)
1504 res.range.likely += base == 8 ? 1 : 2;
1505 else if (res.range.min == 2
1506 && base == 16
1507 && (dir.width[0] == 2 || dir.prec[0] == 2))
1508 ++res.range.likely;
1512 res.range.unlikely = res.range.max;
1513 res.adjust_for_width_or_precision (dir.width, dirtype, base,
1514 (sign | maybebase) + (base == 16));
1515 res.adjust_for_width_or_precision (dir.prec, dirtype, base,
1516 (sign | maybebase) + (base == 16));
1518 return res;
1521 /* Return the number of bytes that a format directive consisting of FLAGS,
1522 PRECision, format SPECification, and MPFR rounding specifier RNDSPEC,
1523 would result for argument X under ideal conditions (i.e., if PREC
1524 weren't excessive). MPFR 3.1 allocates large amounts of memory for
1525 values of PREC with large magnitude and can fail (see MPFR bug #21056).
1526 This function works around those problems. */
1528 static unsigned HOST_WIDE_INT
1529 get_mpfr_format_length (mpfr_ptr x, const char *flags, HOST_WIDE_INT prec,
1530 char spec, char rndspec)
1532 char fmtstr[40];
1534 HOST_WIDE_INT len = strlen (flags);
1536 fmtstr[0] = '%';
1537 memcpy (fmtstr + 1, flags, len);
1538 memcpy (fmtstr + 1 + len, ".*R", 3);
1539 fmtstr[len + 4] = rndspec;
1540 fmtstr[len + 5] = spec;
1541 fmtstr[len + 6] = '\0';
1543 spec = TOUPPER (spec);
1544 if (spec == 'E' || spec == 'F')
1546 /* For %e, specify the precision explicitly since mpfr_sprintf
1547 does its own thing just to be different (see MPFR bug 21088). */
1548 if (prec < 0)
1549 prec = 6;
1551 else
1553 /* Avoid passing negative precisions with larger magnitude to MPFR
1554 to avoid exposing its bugs. (A negative precision is supposed
1555 to be ignored.) */
1556 if (prec < 0)
1557 prec = -1;
1560 HOST_WIDE_INT p = prec;
1562 if (spec == 'G' && !strchr (flags, '#'))
1564 /* For G/g without the pound flag, precision gives the maximum number
1565 of significant digits which is bounded by LDBL_MAX_10_EXP, or, for
1566 a 128 bit IEEE extended precision, 4932. Using twice as much here
1567 should be more than sufficient for any real format. */
1568 if ((IEEE_MAX_10_EXP * 2) < prec)
1569 prec = IEEE_MAX_10_EXP * 2;
1570 p = prec;
1572 else
1574 /* Cap precision arbitrarily at 1KB and add the difference
1575 (if any) to the MPFR result. */
1576 if (prec > 1024)
1577 p = 1024;
1580 len = mpfr_snprintf (NULL, 0, fmtstr, (int)p, x);
1582 /* Handle the unlikely (impossible?) error by returning more than
1583 the maximum dictated by the function's return type. */
1584 if (len < 0)
1585 return target_dir_max () + 1;
1587 /* Adjust the return value by the difference. */
1588 if (p < prec)
1589 len += prec - p;
1591 return len;
1594 /* Return the number of bytes to format using the format specifier
1595 SPEC and the precision PREC the largest value in the real floating
1596 TYPE. */
1598 static unsigned HOST_WIDE_INT
1599 format_floating_max (tree type, char spec, HOST_WIDE_INT prec)
1601 machine_mode mode = TYPE_MODE (type);
1603 /* IBM Extended mode. */
1604 if (MODE_COMPOSITE_P (mode))
1605 mode = DFmode;
1607 /* Get the real type format description for the target. */
1608 const real_format *rfmt = REAL_MODE_FORMAT (mode);
1609 REAL_VALUE_TYPE rv;
1611 real_maxval (&rv, 0, mode);
1613 /* Convert the GCC real value representation with the precision
1614 of the real type to the mpfr_t format with the GCC default
1615 round-to-nearest mode. */
1616 mpfr_t x;
1617 mpfr_init2 (x, rfmt->p);
1618 mpfr_from_real (x, &rv, MPFR_RNDN);
1620 /* Return a value one greater to account for the leading minus sign. */
1621 unsigned HOST_WIDE_INT r
1622 = 1 + get_mpfr_format_length (x, "", prec, spec, 'D');
1623 mpfr_clear (x);
1624 return r;
1627 /* Return a range representing the minimum and maximum number of bytes
1628 that the directive DIR will output for any argument. PREC gives
1629 the adjusted precision range to account for negative precisions
1630 meaning the default 6. This function is used when the directive
1631 argument or its value isn't known. */
1633 static fmtresult
1634 format_floating (const directive &dir, const HOST_WIDE_INT prec[2])
1636 tree type;
1638 switch (dir.modifier)
1640 case FMT_LEN_l:
1641 case FMT_LEN_none:
1642 type = double_type_node;
1643 break;
1645 case FMT_LEN_L:
1646 type = long_double_type_node;
1647 break;
1649 case FMT_LEN_ll:
1650 type = long_double_type_node;
1651 break;
1653 default:
1654 return fmtresult ();
1657 /* The minimum and maximum number of bytes produced by the directive. */
1658 fmtresult res;
1660 /* The minimum output as determined by flags. It's always at least 1.
1661 When plus or space are set the output is preceded by either a sign
1662 or a space. */
1663 unsigned flagmin = (1 /* for the first digit */
1664 + (dir.get_flag ('+') | dir.get_flag (' ')));
1666 /* The minimum is 3 for "inf" and "nan" for all specifiers, plus 1
1667 for the plus sign/space with the '+' and ' ' flags, respectively,
1668 unless reduced below. */
1669 res.range.min = 2 + flagmin;
1671 /* When the pound flag is set the decimal point is included in output
1672 regardless of precision. Whether or not a decimal point is included
1673 otherwise depends on the specification and precision. */
1674 bool radix = dir.get_flag ('#');
1676 switch (dir.specifier)
1678 case 'A':
1679 case 'a':
1681 HOST_WIDE_INT minprec = 6 + !radix /* decimal point */;
1682 if (dir.prec[0] <= 0)
1683 minprec = 0;
1684 else if (dir.prec[0] > 0)
1685 minprec = dir.prec[0] + !radix /* decimal point */;
1687 res.range.likely = (2 /* 0x */
1688 + flagmin
1689 + radix
1690 + minprec
1691 + 3 /* p+0 */);
1693 res.range.max = format_floating_max (type, 'a', prec[1]);
1695 /* The unlikely maximum accounts for the longest multibyte
1696 decimal point character. */
1697 res.range.unlikely = res.range.max;
1698 if (dir.prec[1] > 0)
1699 res.range.unlikely += target_mb_len_max () - 1;
1701 break;
1704 case 'E':
1705 case 'e':
1707 /* Minimum output attributable to precision and, when it's
1708 non-zero, decimal point. */
1709 HOST_WIDE_INT minprec = prec[0] ? prec[0] + !radix : 0;
1711 /* The likely minimum output is "[-+]1.234567e+00" regardless
1712 of the value of the actual argument. */
1713 res.range.likely = (flagmin
1714 + radix
1715 + minprec
1716 + 2 /* e+ */ + 2);
1718 res.range.max = format_floating_max (type, 'e', prec[1]);
1720 /* The unlikely maximum accounts for the longest multibyte
1721 decimal point character. */
1722 if (dir.prec[0] != dir.prec[1]
1723 || dir.prec[0] == -1 || dir.prec[0] > 0)
1724 res.range.unlikely = res.range.max + target_mb_len_max () -1;
1725 else
1726 res.range.unlikely = res.range.max;
1727 break;
1730 case 'F':
1731 case 'f':
1733 /* Minimum output attributable to precision and, when it's non-zero,
1734 decimal point. */
1735 HOST_WIDE_INT minprec = prec[0] ? prec[0] + !radix : 0;
1737 /* For finite numbers (i.e., not infinity or NaN) the lower bound
1738 when precision isn't specified is 8 bytes ("1.23456" since
1739 precision is taken to be 6). When precision is zero, the lower
1740 bound is 1 byte (e.g., "1"). Otherwise, when precision is greater
1741 than zero, then the lower bound is 2 plus precision (plus flags).
1742 But in all cases, the lower bound is no greater than 3. */
1743 unsigned HOST_WIDE_INT min = flagmin + radix + minprec;
1744 if (min < res.range.min)
1745 res.range.min = min;
1747 /* Compute the upper bound for -TYPE_MAX. */
1748 res.range.max = format_floating_max (type, 'f', prec[1]);
1750 /* The minimum output with unknown precision is a single byte
1751 (e.g., "0") but the more likely output is 3 bytes ("0.0"). */
1752 if (dir.prec[0] < 0 && dir.prec[1] > 0)
1753 res.range.likely = 3;
1754 else
1755 res.range.likely = min;
1757 /* The unlikely maximum accounts for the longest multibyte
1758 decimal point character. */
1759 if (dir.prec[0] != dir.prec[1]
1760 || dir.prec[0] == -1 || dir.prec[0] > 0)
1761 res.range.unlikely = res.range.max + target_mb_len_max () - 1;
1762 break;
1765 case 'G':
1766 case 'g':
1768 /* The %g output depends on precision and the exponent of
1769 the argument. Since the value of the argument isn't known
1770 the lower bound on the range of bytes (not counting flags
1771 or width) is 1 plus radix (i.e., either "0" or "0." for
1772 "%g" and "%#g", respectively, with a zero argument). */
1773 unsigned HOST_WIDE_INT min = flagmin + radix;
1774 if (min < res.range.min)
1775 res.range.min = min;
1777 char spec = 'g';
1778 HOST_WIDE_INT maxprec = dir.prec[1];
1779 if (radix && maxprec)
1781 /* When the pound flag (radix) is set, trailing zeros aren't
1782 trimmed and so the longest output is the same as for %e,
1783 except with precision minus 1 (as specified in C11). */
1784 spec = 'e';
1785 if (maxprec > 0)
1786 --maxprec;
1787 else if (maxprec < 0)
1788 maxprec = 5;
1790 else
1791 maxprec = prec[1];
1793 res.range.max = format_floating_max (type, spec, maxprec);
1795 /* The likely output is either the maximum computed above
1796 minus 1 (assuming the maximum is positive) when precision
1797 is known (or unspecified), or the same minimum as for %e
1798 (which is computed for a non-negative argument). Unlike
1799 for the other specifiers above the likely output isn't
1800 the minimum because for %g that's 1 which is unlikely. */
1801 if (dir.prec[1] < 0
1802 || (unsigned HOST_WIDE_INT)dir.prec[1] < target_int_max ())
1803 res.range.likely = res.range.max - 1;
1804 else
1806 HOST_WIDE_INT minprec = 6 + !radix /* decimal point */;
1807 res.range.likely = (flagmin
1808 + radix
1809 + minprec
1810 + 2 /* e+ */ + 2);
1813 /* The unlikely maximum accounts for the longest multibyte
1814 decimal point character. */
1815 res.range.unlikely = res.range.max + target_mb_len_max () - 1;
1816 break;
1819 default:
1820 return fmtresult ();
1823 /* Bump up the byte counters if WIDTH is greater. */
1824 res.adjust_for_width_or_precision (dir.width);
1825 return res;
1828 /* Return a range representing the minimum and maximum number of bytes
1829 that the directive DIR will write on output for the floating argument
1830 ARG. */
1832 static fmtresult
1833 format_floating (const directive &dir, tree arg, range_query *)
1835 HOST_WIDE_INT prec[] = { dir.prec[0], dir.prec[1] };
1836 tree type = (dir.modifier == FMT_LEN_L || dir.modifier == FMT_LEN_ll
1837 ? long_double_type_node : double_type_node);
1839 /* For an indeterminate precision the lower bound must be assumed
1840 to be zero. */
1841 if (TOUPPER (dir.specifier) == 'A')
1843 /* Get the number of fractional decimal digits needed to represent
1844 the argument without a loss of accuracy. */
1845 unsigned fmtprec
1846 = REAL_MODE_FORMAT (TYPE_MODE (type))->p;
1848 /* The precision of the IEEE 754 double format is 53.
1849 The precision of all other GCC binary double formats
1850 is 56 or less. */
1851 unsigned maxprec = fmtprec <= 56 ? 13 : 15;
1853 /* For %a, leave the minimum precision unspecified to let
1854 MFPR trim trailing zeros (as it and many other systems
1855 including Glibc happen to do) and set the maximum
1856 precision to reflect what it would be with trailing zeros
1857 present (as Solaris and derived systems do). */
1858 if (dir.prec[1] < 0)
1860 /* Both bounds are negative implies that precision has
1861 not been specified. */
1862 prec[0] = maxprec;
1863 prec[1] = -1;
1865 else if (dir.prec[0] < 0)
1867 /* With a negative lower bound and a non-negative upper
1868 bound set the minimum precision to zero and the maximum
1869 to the greater of the maximum precision (i.e., with
1870 trailing zeros present) and the specified upper bound. */
1871 prec[0] = 0;
1872 prec[1] = dir.prec[1] < maxprec ? maxprec : dir.prec[1];
1875 else if (dir.prec[0] < 0)
1877 if (dir.prec[1] < 0)
1879 /* A precision in a strictly negative range is ignored and
1880 the default of 6 is used instead. */
1881 prec[0] = prec[1] = 6;
1883 else
1885 /* For a precision in a partly negative range, the lower bound
1886 must be assumed to be zero and the new upper bound is the
1887 greater of 6 (the default precision used when the specified
1888 precision is negative) and the upper bound of the specified
1889 range. */
1890 prec[0] = 0;
1891 prec[1] = dir.prec[1] < 6 ? 6 : dir.prec[1];
1895 if (!arg
1896 || TREE_CODE (arg) != REAL_CST
1897 || !useless_type_conversion_p (type, TREE_TYPE (arg)))
1898 return format_floating (dir, prec);
1900 /* The minimum and maximum number of bytes produced by the directive. */
1901 fmtresult res;
1903 /* Get the real type format description for the target. */
1904 const REAL_VALUE_TYPE *rvp = TREE_REAL_CST_PTR (arg);
1905 const real_format *rfmt = REAL_MODE_FORMAT (TYPE_MODE (TREE_TYPE (arg)));
1907 if (!real_isfinite (rvp))
1909 /* The format for Infinity and NaN is "[-]inf"/"[-]infinity"
1910 and "[-]nan" with the choice being implementation-defined
1911 but not locale dependent. */
1912 bool sign = dir.get_flag ('+') || real_isneg (rvp);
1913 res.range.min = 3 + sign;
1915 res.range.likely = res.range.min;
1916 res.range.max = res.range.min;
1917 /* The unlikely maximum is "[-/+]infinity" or "[-/+][qs]nan".
1918 For NaN, the C/POSIX standards specify two formats:
1919 "[-/+]nan"
1921 "[-/+]nan(n-char-sequence)"
1922 No known printf implementation outputs the latter format but AIX
1923 outputs QNaN and SNaN for quiet and signalling NaN, respectively,
1924 so the unlikely maximum reflects that. */
1925 res.range.unlikely = sign + (real_isinf (rvp) ? 8 : 4);
1927 /* The range for infinity and NaN is known unless either width
1928 or precision is unknown. Width has the same effect regardless
1929 of whether the argument is finite. Precision is either ignored
1930 (e.g., Glibc) or can have an effect on the short vs long format
1931 such as inf/infinity (e.g., Solaris). */
1932 res.knownrange = dir.known_width_and_precision ();
1934 /* Adjust the range for width but ignore precision. */
1935 res.adjust_for_width_or_precision (dir.width);
1937 return res;
1940 char fmtstr [40];
1941 char *pfmt = fmtstr;
1943 /* Append flags. */
1944 for (const char *pf = "-+ #0"; *pf; ++pf)
1945 if (dir.get_flag (*pf))
1946 *pfmt++ = *pf;
1948 *pfmt = '\0';
1951 /* Set up an array to easily iterate over. */
1952 unsigned HOST_WIDE_INT* const minmax[] = {
1953 &res.range.min, &res.range.max
1956 for (int i = 0; i != sizeof minmax / sizeof *minmax; ++i)
1958 /* Convert the GCC real value representation with the precision
1959 of the real type to the mpfr_t format rounding down in the
1960 first iteration that computes the minimum and up in the second
1961 that computes the maximum. This order is arbitrary because
1962 rounding in either direction can result in longer output. */
1963 mpfr_t mpfrval;
1964 mpfr_init2 (mpfrval, rfmt->p);
1965 mpfr_from_real (mpfrval, rvp, i ? MPFR_RNDU : MPFR_RNDD);
1967 /* Use the MPFR rounding specifier to round down in the first
1968 iteration and then up. In most but not all cases this will
1969 result in the same number of bytes. */
1970 char rndspec = "DU"[i];
1972 /* Format it and store the result in the corresponding member
1973 of the result struct. */
1974 *minmax[i] = get_mpfr_format_length (mpfrval, fmtstr, prec[i],
1975 dir.specifier, rndspec);
1976 mpfr_clear (mpfrval);
1980 /* Make sure the minimum is less than the maximum (MPFR rounding
1981 in the call to mpfr_snprintf can result in the reverse. */
1982 if (res.range.max < res.range.min)
1984 unsigned HOST_WIDE_INT tmp = res.range.min;
1985 res.range.min = res.range.max;
1986 res.range.max = tmp;
1989 /* The range is known unless either width or precision is unknown. */
1990 res.knownrange = dir.known_width_and_precision ();
1992 /* For the same floating point constant, unless width or precision
1993 is unknown, use the longer output as the likely maximum since
1994 with round to nearest either is equally likely. Otherwise, when
1995 precision is unknown, use the greater of the minimum and 3 as
1996 the likely output (for "0.0" since zero precision is unlikely). */
1997 if (res.knownrange)
1998 res.range.likely = res.range.max;
1999 else if (res.range.min < 3
2000 && dir.prec[0] < 0
2001 && (unsigned HOST_WIDE_INT)dir.prec[1] == target_int_max ())
2002 res.range.likely = 3;
2003 else
2004 res.range.likely = res.range.min;
2006 res.range.unlikely = res.range.max;
2008 if (res.range.max > 2 && (prec[0] != 0 || prec[1] != 0))
2010 /* Unless the precision is zero output longer than 2 bytes may
2011 include the decimal point which must be a single character
2012 up to MB_LEN_MAX in length. This is overly conservative
2013 since in some conversions some constants result in no decimal
2014 point (e.g., in %g). */
2015 res.range.unlikely += target_mb_len_max () - 1;
2018 res.adjust_for_width_or_precision (dir.width);
2019 return res;
2022 /* Return a FMTRESULT struct set to the lengths of the shortest and longest
2023 strings referenced by the expression STR, or (-1, -1) when not known.
2024 Used by the format_string function below. */
2026 static fmtresult
2027 get_string_length (tree str, gimple *stmt, unsigned HOST_WIDE_INT max_size,
2028 unsigned eltsize, range_query *query)
2030 if (!str)
2031 return fmtresult ();
2033 /* Try to determine the dynamic string length first.
2034 Set MAXBOUND to an arbitrary non-null non-integer node as a request
2035 to have it set to the length of the longest string in a PHI. */
2036 c_strlen_data lendata = { };
2037 lendata.maxbound = str;
2038 if (eltsize == 1)
2039 get_range_strlen_dynamic (str, stmt, &lendata, query);
2040 else
2042 /* Determine the length of the shortest and longest string referenced
2043 by STR. Strings of unknown lengths are bounded by the sizes of
2044 arrays that subexpressions of STR may refer to. Pointers that
2045 aren't known to point any such arrays result in LENDATA.MAXLEN
2046 set to SIZE_MAX. */
2047 get_range_strlen (str, &lendata, eltsize);
2050 /* If LENDATA.MAXBOUND is not equal to .MINLEN it corresponds to the bound
2051 of the largest array STR refers to, if known, or it's set to SIZE_MAX
2052 otherwise. */
2054 /* Return the default result when nothing is known about the string. */
2055 if ((lendata.maxbound && !tree_fits_uhwi_p (lendata.maxbound))
2056 || !tree_fits_uhwi_p (lendata.maxlen))
2058 fmtresult res;
2059 res.nonstr = lendata.decl;
2060 return res;
2063 unsigned HOST_WIDE_INT lenmax = tree_to_uhwi (max_object_size ()) - 2;
2064 if (integer_zerop (lendata.minlen)
2065 && (!lendata.maxbound || lenmax <= tree_to_uhwi (lendata.maxbound))
2066 && lenmax <= tree_to_uhwi (lendata.maxlen))
2068 if (max_size > 0 && max_size < HOST_WIDE_INT_MAX)
2070 /* Adjust the conservative unknown/unbounded result if MAX_SIZE
2071 is valid. Set UNLIKELY to maximum in case MAX_SIZE refers
2072 to a subobject.
2073 TODO: This is overly conservative. Set UNLIKELY to the size
2074 of the outermost enclosing declared object. */
2075 fmtresult res (0, max_size - 1);
2076 res.nonstr = lendata.decl;
2077 res.range.likely = res.range.max;
2078 res.range.unlikely = HOST_WIDE_INT_MAX;
2079 return res;
2082 fmtresult res;
2083 res.nonstr = lendata.decl;
2084 return res;
2087 HOST_WIDE_INT min
2088 = (tree_fits_uhwi_p (lendata.minlen)
2089 ? tree_to_uhwi (lendata.minlen)
2090 : 0);
2092 HOST_WIDE_INT max
2093 = (lendata.maxbound && tree_fits_uhwi_p (lendata.maxbound)
2094 ? tree_to_uhwi (lendata.maxbound)
2095 : HOST_WIDE_INT_M1U);
2097 const bool unbounded = integer_all_onesp (lendata.maxlen);
2099 /* Set the max/likely counters to unbounded when a minimum is known
2100 but the maximum length isn't bounded. This implies that STR is
2101 a conditional expression involving a string of known length and
2102 an expression of unknown/unbounded length. */
2103 if (min
2104 && (unsigned HOST_WIDE_INT)min < HOST_WIDE_INT_M1U
2105 && unbounded)
2106 max = HOST_WIDE_INT_M1U;
2108 /* get_range_strlen() returns the target value of SIZE_MAX for
2109 strings of unknown length. Bump it up to HOST_WIDE_INT_M1U
2110 which may be bigger. */
2111 if ((unsigned HOST_WIDE_INT)min == target_size_max ())
2112 min = HOST_WIDE_INT_M1U;
2113 if ((unsigned HOST_WIDE_INT)max == target_size_max ())
2114 max = HOST_WIDE_INT_M1U;
2116 fmtresult res (min, max);
2117 res.nonstr = lendata.decl;
2119 /* Set RES.KNOWNRANGE to true if and only if all strings referenced
2120 by STR are known to be bounded (though not necessarily by their
2121 actual length but perhaps by their maximum possible length). */
2122 if (res.range.max < target_int_max ())
2124 res.knownrange = true;
2125 /* When the length of the longest string is known and not
2126 excessive use it as the likely length of the string(s). */
2127 res.range.likely = res.range.max;
2129 else
2131 /* When the upper bound is unknown (it can be zero or excessive)
2132 set the likely length to the greater of 1. If MAXBOUND is
2133 known, also reset the length of the lower bound to zero. */
2134 res.range.likely = res.range.min ? res.range.min : warn_level > 1;
2135 if (lendata.maxbound && !integer_all_onesp (lendata.maxbound))
2136 res.range.min = 0;
2139 res.range.unlikely = unbounded ? HOST_WIDE_INT_MAX : res.range.max;
2141 return res;
2144 /* Return the minimum and maximum number of characters formatted
2145 by the '%c' format directives and its wide character form for
2146 the argument ARG. ARG can be null (for functions such as
2147 vsprinf). */
2149 static fmtresult
2150 format_character (const directive &dir, tree arg, range_query *query)
2152 fmtresult res;
2154 res.knownrange = true;
2156 if (dir.specifier == 'C'
2157 || dir.modifier == FMT_LEN_l)
2159 /* A wide character can result in as few as zero bytes. */
2160 res.range.min = 0;
2162 HOST_WIDE_INT min, max;
2163 if (get_int_range (arg, dir.info->callstmt, &min, &max, false, 0, query))
2165 if (min == 0 && max == 0)
2167 /* The NUL wide character results in no bytes. */
2168 res.range.max = 0;
2169 res.range.likely = 0;
2170 res.range.unlikely = 0;
2172 else if (min >= 0 && min < 128)
2174 /* Be conservative if the target execution character set
2175 is not a 1-to-1 mapping to the source character set or
2176 if the source set is not ASCII. */
2177 bool one_2_one_ascii
2178 = (target_to_host_charmap[0] == 1 && target_to_host ('a') == 97);
2180 /* A wide character in the ASCII range most likely results
2181 in a single byte, and only unlikely in up to MB_LEN_MAX. */
2182 res.range.max = one_2_one_ascii ? 1 : target_mb_len_max ();;
2183 res.range.likely = 1;
2184 res.range.unlikely = target_mb_len_max ();
2185 res.mayfail = !one_2_one_ascii;
2187 else
2189 /* A wide character outside the ASCII range likely results
2190 in up to two bytes, and only unlikely in up to MB_LEN_MAX. */
2191 res.range.max = target_mb_len_max ();
2192 res.range.likely = 2;
2193 res.range.unlikely = res.range.max;
2194 /* Converting such a character may fail. */
2195 res.mayfail = true;
2198 else
2200 /* An unknown wide character is treated the same as a wide
2201 character outside the ASCII range. */
2202 res.range.max = target_mb_len_max ();
2203 res.range.likely = 2;
2204 res.range.unlikely = res.range.max;
2205 res.mayfail = true;
2208 else
2210 /* A plain '%c' directive. Its output is exactly 1. */
2211 res.range.min = res.range.max = 1;
2212 res.range.likely = res.range.unlikely = 1;
2213 res.knownrange = true;
2216 /* Bump up the byte counters if WIDTH is greater. */
2217 return res.adjust_for_width_or_precision (dir.width);
2220 /* If TYPE is an array or struct or union, increment *FLDOFF by the starting
2221 offset of the member that *OFF point into and set *FLDSIZE to its size
2222 in bytes and decrement *OFF by the same. Otherwise do nothing. */
2224 static void
2225 set_aggregate_size_and_offset (tree type, HOST_WIDE_INT *fldoff,
2226 HOST_WIDE_INT *fldsize, HOST_WIDE_INT *off)
2228 /* The byte offset of the most basic struct member the byte
2229 offset *OFF corresponds to, or for a (multidimensional)
2230 array member, the byte offset of the array element. */
2231 if (TREE_CODE (type) == ARRAY_TYPE
2232 && TREE_CODE (TREE_TYPE (type)) == ARRAY_TYPE)
2234 HOST_WIDE_INT index = 0, arrsize = 0;
2235 if (array_elt_at_offset (type, *off, &index, &arrsize))
2237 *fldoff += index;
2238 *off -= index;
2239 *fldsize = arrsize;
2242 else if (RECORD_OR_UNION_TYPE_P (type))
2244 HOST_WIDE_INT index = 0;
2245 tree sub = field_at_offset (type, NULL_TREE, *off, &index);
2246 if (sub)
2248 tree subsize = DECL_SIZE_UNIT (sub);
2249 if (*fldsize < HOST_WIDE_INT_MAX
2250 && subsize
2251 && tree_fits_uhwi_p (subsize))
2252 *fldsize = tree_to_uhwi (subsize);
2253 else
2254 *fldsize = HOST_WIDE_INT_MAX;
2255 *fldoff += index;
2256 *off -= index;
2261 /* For an expression X of pointer type, recursively try to find the same
2262 origin (object or pointer) as Y it references and return such a Y.
2263 When X refers to an array element or struct member, set *FLDOFF to
2264 the offset of the element or member from the beginning of the "most
2265 derived" object and *FLDSIZE to its size. When nonnull, set *OFF to
2266 the overall offset from the beginning of the object so that
2267 *FLDOFF <= *OFF. */
2269 static tree
2270 get_origin_and_offset_r (tree x, HOST_WIDE_INT *fldoff, HOST_WIDE_INT *fldsize,
2271 HOST_WIDE_INT *off)
2273 if (!x)
2274 return NULL_TREE;
2276 HOST_WIDE_INT sizebuf = -1;
2277 if (!fldsize)
2278 fldsize = &sizebuf;
2280 if (DECL_P (x))
2282 /* Set the size if it hasn't been set yet. */
2283 if (tree size = DECL_SIZE_UNIT (x))
2284 if (*fldsize < 0 && tree_fits_shwi_p (size))
2285 *fldsize = tree_to_shwi (size);
2286 return x;
2289 switch (TREE_CODE (x))
2291 case ADDR_EXPR:
2292 x = TREE_OPERAND (x, 0);
2293 return get_origin_and_offset_r (x, fldoff, fldsize, off);
2295 case ARRAY_REF:
2297 tree offset = TREE_OPERAND (x, 1);
2298 HOST_WIDE_INT idx = (tree_fits_uhwi_p (offset)
2299 ? tree_to_uhwi (offset) : HOST_WIDE_INT_MAX);
2301 tree eltype = TREE_TYPE (x);
2302 if (TREE_CODE (eltype) == INTEGER_TYPE)
2304 if (off)
2305 *off = idx;
2307 else if (idx < HOST_WIDE_INT_MAX)
2308 *fldoff += idx * int_size_in_bytes (eltype);
2309 else
2310 *fldoff = idx;
2312 x = TREE_OPERAND (x, 0);
2313 return get_origin_and_offset_r (x, fldoff, fldsize, nullptr);
2316 case MEM_REF:
2317 if (off)
2319 tree offset = TREE_OPERAND (x, 1);
2320 *off = (tree_fits_uhwi_p (offset)
2321 ? tree_to_uhwi (offset) : HOST_WIDE_INT_MAX);
2324 x = TREE_OPERAND (x, 0);
2326 if (off)
2328 tree xtype
2329 = (TREE_CODE (x) == ADDR_EXPR
2330 ? TREE_TYPE (TREE_OPERAND (x, 0)) : TREE_TYPE (TREE_TYPE (x)));
2332 set_aggregate_size_and_offset (xtype, fldoff, fldsize, off);
2335 return get_origin_and_offset_r (x, fldoff, fldsize, nullptr);
2337 case COMPONENT_REF:
2339 tree fld = TREE_OPERAND (x, 1);
2340 *fldoff += int_byte_position (fld);
2342 get_origin_and_offset_r (fld, fldoff, fldsize, off);
2343 x = TREE_OPERAND (x, 0);
2344 return get_origin_and_offset_r (x, fldoff, nullptr, off);
2347 case SSA_NAME:
2349 gimple *def = SSA_NAME_DEF_STMT (x);
2350 if (is_gimple_assign (def))
2352 tree_code code = gimple_assign_rhs_code (def);
2353 if (code == ADDR_EXPR)
2355 x = gimple_assign_rhs1 (def);
2356 return get_origin_and_offset_r (x, fldoff, fldsize, off);
2359 if (code == POINTER_PLUS_EXPR)
2361 tree offset = gimple_assign_rhs2 (def);
2362 if (off && tree_fits_uhwi_p (offset))
2363 *off = tree_to_uhwi (offset);
2365 x = gimple_assign_rhs1 (def);
2366 x = get_origin_and_offset_r (x, fldoff, fldsize, off);
2367 if (off && !tree_fits_uhwi_p (offset))
2368 *off = HOST_WIDE_INT_MAX;
2369 if (off)
2371 tree xtype = TREE_TYPE (x);
2372 set_aggregate_size_and_offset (xtype, fldoff, fldsize, off);
2374 return x;
2376 else if (code == VAR_DECL)
2378 x = gimple_assign_rhs1 (def);
2379 return get_origin_and_offset_r (x, fldoff, fldsize, off);
2382 else if (gimple_nop_p (def) && SSA_NAME_VAR (x))
2383 x = SSA_NAME_VAR (x);
2385 tree xtype = TREE_TYPE (x);
2386 if (POINTER_TYPE_P (xtype))
2387 xtype = TREE_TYPE (xtype);
2389 if (off)
2390 set_aggregate_size_and_offset (xtype, fldoff, fldsize, off);
2393 default:
2394 break;
2397 return x;
2400 /* Nonrecursive version of the above. */
2402 static tree
2403 get_origin_and_offset (tree x, HOST_WIDE_INT *fldoff, HOST_WIDE_INT *off,
2404 HOST_WIDE_INT *fldsize = nullptr)
2406 HOST_WIDE_INT sizebuf;
2407 if (!fldsize)
2408 fldsize = &sizebuf;
2410 *fldsize = -1;
2412 *fldoff = *off = *fldsize = 0;
2413 tree orig = get_origin_and_offset_r (x, fldoff, fldsize, off);
2414 if (!orig)
2415 return NULL_TREE;
2417 if (!*fldoff && *off == *fldsize)
2419 *fldoff = *off;
2420 *off = 0;
2423 return orig;
2426 /* If ARG refers to the same (sub)object or array element as described
2427 by DST and DST_FLD, return the byte offset into the struct member or
2428 array element referenced by ARG and set *ARG_SIZE to the size of
2429 the (sub)object. Otherwise return HOST_WIDE_INT_MIN to indicate
2430 that ARG and DST do not refer to the same object. */
2432 static HOST_WIDE_INT
2433 alias_offset (tree arg, HOST_WIDE_INT *arg_size,
2434 tree dst, HOST_WIDE_INT dst_fld)
2436 /* See if the argument refers to the same base object as the destination
2437 of the formatted function call, and if so, try to determine if they
2438 can alias. */
2439 if (!arg || !dst || !ptr_derefs_may_alias_p (arg, dst))
2440 return HOST_WIDE_INT_MIN;
2442 /* The two arguments may refer to the same object. If they both refer
2443 to a struct member, see if the members are one and the same. */
2444 HOST_WIDE_INT arg_off = 0, arg_fld = 0;
2446 tree arg_orig = get_origin_and_offset (arg, &arg_fld, &arg_off, arg_size);
2448 if (arg_orig == dst && arg_fld == dst_fld)
2449 return arg_off;
2451 return HOST_WIDE_INT_MIN;
2454 /* Return the minimum and maximum number of characters formatted
2455 by the '%s' format directive and its wide character form for
2456 the argument ARG. ARG can be null (for functions such as
2457 vsprinf). */
2459 static fmtresult
2460 format_string (const directive &dir, tree arg, range_query *query)
2462 fmtresult res;
2464 /* The size of the (sub)object ARG refers to. Used to adjust
2465 the conservative get_string_length() result. */
2466 HOST_WIDE_INT arg_size = 0;
2468 if (warn_restrict)
2470 /* See if ARG might alias the destination of the call with
2471 DST_ORIGIN and DST_FIELD. If so, store the starting offset
2472 so that the overlap can be determined for certain later,
2473 when the amount of output of the call (including subsequent
2474 directives) has been computed. Otherwise, store HWI_MIN. */
2475 res.dst_offset = alias_offset (arg, &arg_size, dir.info->dst_origin,
2476 dir.info->dst_field);
2477 if (res.dst_offset >= 0 && res.dst_offset <= arg_size)
2478 arg_size -= res.dst_offset;
2479 else
2480 arg_size = 0;
2483 /* Compute the range the argument's length can be in. */
2484 int count_by = 1;
2485 if (dir.specifier == 'S' || dir.modifier == FMT_LEN_l)
2487 /* Get a node for a C type that will be the same size
2488 as a wchar_t on the target. */
2489 tree node = get_typenode_from_name (MODIFIED_WCHAR_TYPE);
2491 /* Now that we have a suitable node, get the number of
2492 bytes it occupies. */
2493 count_by = int_size_in_bytes (node);
2494 gcc_checking_assert (count_by == 2 || count_by == 4);
2497 fmtresult slen =
2498 get_string_length (arg, dir.info->callstmt, arg_size, count_by, query);
2499 if (slen.range.min == slen.range.max
2500 && slen.range.min < HOST_WIDE_INT_MAX)
2502 /* The argument is either a string constant or it refers
2503 to one of a number of strings of the same length. */
2505 /* A '%s' directive with a string argument with constant length. */
2506 res.range = slen.range;
2508 if (dir.specifier == 'S'
2509 || dir.modifier == FMT_LEN_l)
2511 /* In the worst case the length of output of a wide string S
2512 is bounded by MB_LEN_MAX * wcslen (S). */
2513 res.range.max *= target_mb_len_max ();
2514 res.range.unlikely = res.range.max;
2515 /* It's likely that the total length is not more that
2516 2 * wcslen (S).*/
2517 res.range.likely = res.range.min * 2;
2519 if (dir.prec[1] >= 0
2520 && (unsigned HOST_WIDE_INT)dir.prec[1] < res.range.max)
2522 res.range.max = dir.prec[1];
2523 res.range.likely = dir.prec[1];
2524 res.range.unlikely = dir.prec[1];
2527 if (dir.prec[0] < 0 && dir.prec[1] > -1)
2528 res.range.min = 0;
2529 else if (dir.prec[0] >= 0)
2530 res.range.likely = dir.prec[0];
2532 /* Even a non-empty wide character string need not convert into
2533 any bytes. */
2534 res.range.min = 0;
2536 /* A non-empty wide character conversion may fail. */
2537 if (slen.range.max > 0)
2538 res.mayfail = true;
2540 else
2542 res.knownrange = true;
2544 if (dir.prec[0] < 0 && dir.prec[1] > -1)
2545 res.range.min = 0;
2546 else if ((unsigned HOST_WIDE_INT)dir.prec[0] < res.range.min)
2547 res.range.min = dir.prec[0];
2549 if ((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];
2557 else if (arg && integer_zerop (arg))
2559 /* Handle null pointer argument. */
2561 fmtresult res (0);
2562 res.nullp = true;
2563 return res;
2565 else
2567 /* For a '%s' and '%ls' directive with a non-constant string (either
2568 one of a number of strings of known length or an unknown string)
2569 the minimum number of characters is lesser of PRECISION[0] and
2570 the length of the shortest known string or zero, and the maximum
2571 is the lesser of the length of the longest known string or
2572 PTRDIFF_MAX and PRECISION[1]. The likely length is either
2573 the minimum at level 1 and the greater of the minimum and 1
2574 at level 2. This result is adjust upward for width (if it's
2575 specified). */
2577 if (dir.specifier == 'S'
2578 || dir.modifier == FMT_LEN_l)
2580 /* A wide character converts to as few as zero bytes. */
2581 slen.range.min = 0;
2582 if (slen.range.max < target_int_max ())
2583 slen.range.max *= target_mb_len_max ();
2585 if (slen.range.likely < target_int_max ())
2586 slen.range.likely *= 2;
2588 if (slen.range.likely < target_int_max ())
2589 slen.range.unlikely *= target_mb_len_max ();
2591 /* A non-empty wide character conversion may fail. */
2592 if (slen.range.max > 0)
2593 res.mayfail = true;
2596 res.range = slen.range;
2598 if (dir.prec[0] >= 0)
2600 /* Adjust the minimum to zero if the string length is unknown,
2601 or at most the lower bound of the precision otherwise. */
2602 if (slen.range.min >= target_int_max ())
2603 res.range.min = 0;
2604 else if ((unsigned HOST_WIDE_INT)dir.prec[0] < slen.range.min)
2605 res.range.min = dir.prec[0];
2607 /* Make both maxima no greater than the upper bound of precision. */
2608 if ((unsigned HOST_WIDE_INT)dir.prec[1] < slen.range.max
2609 || slen.range.max >= target_int_max ())
2611 res.range.max = dir.prec[1];
2612 res.range.unlikely = dir.prec[1];
2615 /* If precision is constant, set the likely counter to the lesser
2616 of it and the maximum string length. Otherwise, if the lower
2617 bound of precision is greater than zero, set the likely counter
2618 to the minimum. Otherwise set it to zero or one based on
2619 the warning level. */
2620 if (dir.prec[0] == dir.prec[1])
2621 res.range.likely
2622 = ((unsigned HOST_WIDE_INT)dir.prec[0] < slen.range.max
2623 ? dir.prec[0] : slen.range.max);
2624 else if (dir.prec[0] > 0)
2625 res.range.likely = res.range.min;
2626 else
2627 res.range.likely = warn_level > 1;
2629 else if (dir.prec[1] >= 0)
2631 res.range.min = 0;
2632 if ((unsigned HOST_WIDE_INT)dir.prec[1] < slen.range.max)
2633 res.range.max = dir.prec[1];
2634 res.range.likely = dir.prec[1] ? warn_level > 1 : 0;
2635 if ((unsigned HOST_WIDE_INT)dir.prec[1] < slen.range.unlikely)
2636 res.range.unlikely = dir.prec[1];
2638 else if (slen.range.min >= target_int_max ())
2640 res.range.min = 0;
2641 res.range.max = HOST_WIDE_INT_MAX;
2642 /* At level 1 strings of unknown length are assumed to be
2643 empty, while at level 1 they are assumed to be one byte
2644 long. */
2645 res.range.likely = warn_level > 1;
2646 res.range.unlikely = HOST_WIDE_INT_MAX;
2648 else
2650 /* A string of unknown length unconstrained by precision is
2651 assumed to be empty at level 1 and just one character long
2652 at higher levels. */
2653 if (res.range.likely >= target_int_max ())
2654 res.range.likely = warn_level > 1;
2658 /* If the argument isn't a nul-terminated string and the number
2659 of bytes on output isn't bounded by precision, set NONSTR. */
2660 if (slen.nonstr && slen.range.min < (unsigned HOST_WIDE_INT)dir.prec[0])
2661 res.nonstr = slen.nonstr;
2663 /* Bump up the byte counters if WIDTH is greater. */
2664 return res.adjust_for_width_or_precision (dir.width);
2667 /* Format plain string (part of the format string itself). */
2669 static fmtresult
2670 format_plain (const directive &dir, tree, range_query *)
2672 fmtresult res (dir.len);
2673 return res;
2676 /* Return true if the RESULT of a directive in a call describe by INFO
2677 should be diagnosed given the AVAILable space in the destination. */
2679 static bool
2680 should_warn_p (const call_info &info,
2681 const result_range &avail, const result_range &result)
2683 if (result.max <= avail.min)
2685 /* The least amount of space remaining in the destination is big
2686 enough for the longest output. */
2687 return false;
2690 if (info.bounded)
2692 if (warn_format_trunc == 1 && result.min <= avail.max
2693 && info.retval_used ())
2695 /* The likely amount of space remaining in the destination is big
2696 enough for the least output and the return value is used. */
2697 return false;
2700 if (warn_format_trunc == 1 && result.likely <= avail.likely
2701 && !info.retval_used ())
2703 /* The likely amount of space remaining in the destination is big
2704 enough for the likely output and the return value is unused. */
2705 return false;
2708 if (warn_format_trunc == 2
2709 && result.likely <= avail.min
2710 && (result.max <= avail.min
2711 || result.max > HOST_WIDE_INT_MAX))
2713 /* The minimum amount of space remaining in the destination is big
2714 enough for the longest output. */
2715 return false;
2718 else
2720 if (warn_level == 1 && result.likely <= avail.likely)
2722 /* The likely amount of space remaining in the destination is big
2723 enough for the likely output. */
2724 return false;
2727 if (warn_level == 2
2728 && result.likely <= avail.min
2729 && (result.max <= avail.min
2730 || result.max > HOST_WIDE_INT_MAX))
2732 /* The minimum amount of space remaining in the destination is big
2733 enough for the longest output. */
2734 return false;
2738 return true;
2741 /* At format string location describe by DIRLOC in a call described
2742 by INFO, issue a warning for a directive DIR whose output may be
2743 in excess of the available space AVAIL_RANGE in the destination
2744 given the formatting result FMTRES. This function does nothing
2745 except decide whether to issue a warning for a possible write
2746 past the end or truncation and, if so, format the warning.
2747 Return true if a warning has been issued. */
2749 static bool
2750 maybe_warn (substring_loc &dirloc, location_t argloc,
2751 const call_info &info,
2752 const result_range &avail_range, const result_range &res,
2753 const directive &dir)
2755 if (!should_warn_p (info, avail_range, res))
2756 return false;
2758 /* A warning will definitely be issued below. */
2760 /* The maximum byte count to reference in the warning. Larger counts
2761 imply that the upper bound is unknown (and could be anywhere between
2762 RES.MIN + 1 and SIZE_MAX / 2) are printed as "N or more bytes" rather
2763 than "between N and X" where X is some huge number. */
2764 unsigned HOST_WIDE_INT maxbytes = target_dir_max ();
2766 /* True when there is enough room in the destination for the least
2767 amount of a directive's output but not enough for its likely or
2768 maximum output. */
2769 bool maybe = (res.min <= avail_range.max
2770 && (avail_range.min < res.likely
2771 || (res.max < HOST_WIDE_INT_MAX
2772 && avail_range.min < res.max)));
2774 /* Buffer for the directive in the host character set (used when
2775 the source character set is different). */
2776 char hostdir[32];
2778 if (avail_range.min == avail_range.max)
2780 /* The size of the destination region is exact. */
2781 unsigned HOST_WIDE_INT navail = avail_range.max;
2783 if (target_to_host (*dir.beg) != '%')
2785 /* For plain character directives (i.e., the format string itself)
2786 but not others, point the caret at the first character that's
2787 past the end of the destination. */
2788 if (navail < dir.len)
2789 dirloc.set_caret_index (dirloc.get_caret_idx () + navail);
2792 if (*dir.beg == '\0')
2794 /* This is the terminating nul. */
2795 gcc_assert (res.min == 1 && res.min == res.max);
2797 return fmtwarn (dirloc, UNKNOWN_LOCATION, NULL, info.warnopt (),
2798 info.bounded
2799 ? (maybe
2800 ? G_("%qE output may be truncated before the "
2801 "last format character")
2802 : G_("%qE output truncated before the last "
2803 "format character"))
2804 : (maybe
2805 ? G_("%qE may write a terminating nul past the "
2806 "end of the destination")
2807 : G_("%qE writing a terminating nul past the "
2808 "end of the destination")),
2809 info.func);
2812 if (res.min == res.max)
2814 const char *d = target_to_host (hostdir, sizeof hostdir, dir.beg);
2815 if (!info.bounded)
2816 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2817 "%<%.*s%> directive writing %wu byte into a "
2818 "region of size %wu",
2819 "%<%.*s%> directive writing %wu bytes into a "
2820 "region of size %wu",
2821 (int) dir.len, d, res.min, navail);
2822 else if (maybe)
2823 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2824 "%<%.*s%> directive output may be truncated "
2825 "writing %wu byte into a region of size %wu",
2826 "%<%.*s%> directive output may be truncated "
2827 "writing %wu bytes into a region of size %wu",
2828 (int) dir.len, d, res.min, navail);
2829 else
2830 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2831 "%<%.*s%> directive output truncated writing "
2832 "%wu byte into a region of size %wu",
2833 "%<%.*s%> directive output truncated writing "
2834 "%wu bytes into a region of size %wu",
2835 (int) dir.len, d, res.min, navail);
2837 if (res.min == 0 && res.max < maxbytes)
2838 return fmtwarn (dirloc, argloc, NULL,
2839 info.warnopt (),
2840 info.bounded
2841 ? (maybe
2842 ? G_("%<%.*s%> directive output may be truncated "
2843 "writing up to %wu bytes into a region of "
2844 "size %wu")
2845 : G_("%<%.*s%> directive output truncated writing "
2846 "up to %wu bytes into a region of size %wu"))
2847 : G_("%<%.*s%> directive writing up to %wu bytes "
2848 "into a region of size %wu"), (int) dir.len,
2849 target_to_host (hostdir, sizeof hostdir, dir.beg),
2850 res.max, navail);
2852 if (res.min == 0 && maxbytes <= res.max)
2853 /* This is a special case to avoid issuing the potentially
2854 confusing warning:
2855 writing 0 or more bytes into a region of size 0. */
2856 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
2857 info.bounded
2858 ? (maybe
2859 ? G_("%<%.*s%> directive output may be truncated "
2860 "writing likely %wu or more bytes into a "
2861 "region of size %wu")
2862 : G_("%<%.*s%> directive output truncated writing "
2863 "likely %wu or more bytes into a region of "
2864 "size %wu"))
2865 : G_("%<%.*s%> directive writing likely %wu or more "
2866 "bytes into a region of size %wu"), (int) dir.len,
2867 target_to_host (hostdir, sizeof hostdir, dir.beg),
2868 res.likely, navail);
2870 if (res.max < maxbytes)
2871 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
2872 info.bounded
2873 ? (maybe
2874 ? G_("%<%.*s%> directive output may be truncated "
2875 "writing between %wu and %wu bytes into a "
2876 "region of size %wu")
2877 : G_("%<%.*s%> directive output truncated "
2878 "writing between %wu and %wu bytes into a "
2879 "region of size %wu"))
2880 : G_("%<%.*s%> directive writing between %wu and "
2881 "%wu bytes into a region of size %wu"),
2882 (int) dir.len,
2883 target_to_host (hostdir, sizeof hostdir, dir.beg),
2884 res.min, res.max, navail);
2886 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
2887 info.bounded
2888 ? (maybe
2889 ? G_("%<%.*s%> directive output may be truncated "
2890 "writing %wu or more bytes into a region of "
2891 "size %wu")
2892 : G_("%<%.*s%> directive output truncated writing "
2893 "%wu or more bytes into a region of size %wu"))
2894 : G_("%<%.*s%> directive writing %wu or more bytes "
2895 "into a region of size %wu"), (int) dir.len,
2896 target_to_host (hostdir, sizeof hostdir, dir.beg),
2897 res.min, navail);
2900 /* The size of the destination region is a range. */
2902 if (target_to_host (*dir.beg) != '%')
2904 unsigned HOST_WIDE_INT navail = avail_range.max;
2906 /* For plain character directives (i.e., the format string itself)
2907 but not others, point the caret at the first character that's
2908 past the end of the destination. */
2909 if (navail < dir.len)
2910 dirloc.set_caret_index (dirloc.get_caret_idx () + navail);
2913 if (*dir.beg == '\0')
2915 gcc_assert (res.min == 1 && res.min == res.max);
2917 return fmtwarn (dirloc, UNKNOWN_LOCATION, NULL, info.warnopt (),
2918 info.bounded
2919 ? (maybe
2920 ? G_("%qE output may be truncated before the last "
2921 "format character")
2922 : G_("%qE output truncated before the last format "
2923 "character"))
2924 : (maybe
2925 ? G_("%qE may write a terminating nul past the end "
2926 "of the destination")
2927 : G_("%qE writing a terminating nul past the end "
2928 "of the destination")), info.func);
2931 if (res.min == res.max)
2933 const char *d = target_to_host (hostdir, sizeof hostdir, dir.beg);
2934 if (!info.bounded)
2935 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2936 "%<%.*s%> directive writing %wu byte into a region "
2937 "of size between %wu and %wu",
2938 "%<%.*s%> directive writing %wu bytes into a region "
2939 "of size between %wu and %wu", (int) dir.len, d,
2940 res.min, avail_range.min, avail_range.max);
2941 else if (maybe)
2942 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2943 "%<%.*s%> directive output may be truncated writing "
2944 "%wu byte into a region of size between %wu and %wu",
2945 "%<%.*s%> directive output may be truncated writing "
2946 "%wu bytes into a region of size between %wu and "
2947 "%wu", (int) dir.len, d, res.min, avail_range.min,
2948 avail_range.max);
2949 else
2950 return fmtwarn_n (dirloc, argloc, NULL, info.warnopt (), res.min,
2951 "%<%.*s%> directive output truncated writing %wu "
2952 "byte into a region of size between %wu and %wu",
2953 "%<%.*s%> directive output truncated writing %wu "
2954 "bytes into a region of size between %wu and %wu",
2955 (int) dir.len, d, res.min, avail_range.min,
2956 avail_range.max);
2959 if (res.min == 0 && res.max < maxbytes)
2960 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
2961 info.bounded
2962 ? (maybe
2963 ? G_("%<%.*s%> directive output may be truncated "
2964 "writing up to %wu bytes into a region of size "
2965 "between %wu and %wu")
2966 : G_("%<%.*s%> directive output truncated writing "
2967 "up to %wu bytes into a region of size between "
2968 "%wu and %wu"))
2969 : G_("%<%.*s%> directive writing up to %wu bytes "
2970 "into a region of size between %wu and %wu"),
2971 (int) dir.len,
2972 target_to_host (hostdir, sizeof hostdir, dir.beg),
2973 res.max, avail_range.min, avail_range.max);
2975 if (res.min == 0 && maxbytes <= res.max)
2976 /* This is a special case to avoid issuing the potentially confusing
2977 warning:
2978 writing 0 or more bytes into a region of size between 0 and N. */
2979 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
2980 info.bounded
2981 ? (maybe
2982 ? G_("%<%.*s%> directive output may be truncated "
2983 "writing likely %wu or more bytes into a region "
2984 "of size between %wu and %wu")
2985 : G_("%<%.*s%> directive output truncated writing "
2986 "likely %wu or more bytes into a region of size "
2987 "between %wu and %wu"))
2988 : G_("%<%.*s%> directive writing likely %wu or more bytes "
2989 "into a region of size between %wu and %wu"),
2990 (int) dir.len,
2991 target_to_host (hostdir, sizeof hostdir, dir.beg),
2992 res.likely, avail_range.min, avail_range.max);
2994 if (res.max < maxbytes)
2995 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
2996 info.bounded
2997 ? (maybe
2998 ? G_("%<%.*s%> directive output may be truncated "
2999 "writing between %wu and %wu bytes into a region "
3000 "of size between %wu and %wu")
3001 : G_("%<%.*s%> directive output truncated writing "
3002 "between %wu and %wu bytes into a region of size "
3003 "between %wu and %wu"))
3004 : G_("%<%.*s%> directive writing between %wu and "
3005 "%wu bytes into a region of size between %wu and "
3006 "%wu"), (int) dir.len,
3007 target_to_host (hostdir, sizeof hostdir, dir.beg),
3008 res.min, res.max, avail_range.min, avail_range.max);
3010 return fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3011 info.bounded
3012 ? (maybe
3013 ? G_("%<%.*s%> directive output may be truncated writing "
3014 "%wu or more bytes into a region of size between "
3015 "%wu and %wu")
3016 : G_("%<%.*s%> directive output truncated writing "
3017 "%wu or more bytes into a region of size between "
3018 "%wu and %wu"))
3019 : G_("%<%.*s%> directive writing %wu or more bytes "
3020 "into a region of size between %wu and %wu"),
3021 (int) dir.len,
3022 target_to_host (hostdir, sizeof hostdir, dir.beg),
3023 res.min, avail_range.min, avail_range.max);
3026 /* Given the formatting result described by RES and NAVAIL, the number
3027 of available bytes in the destination, return the range of bytes
3028 remaining in the destination. */
3030 static inline result_range
3031 bytes_remaining (unsigned HOST_WIDE_INT navail, const format_result &res)
3033 result_range range;
3035 if (HOST_WIDE_INT_MAX <= navail)
3037 range.min = range.max = range.likely = range.unlikely = navail;
3038 return range;
3041 /* The lower bound of the available range is the available size
3042 minus the maximum output size, and the upper bound is the size
3043 minus the minimum. */
3044 range.max = res.range.min < navail ? navail - res.range.min : 0;
3046 range.likely = res.range.likely < navail ? navail - res.range.likely : 0;
3048 if (res.range.max < HOST_WIDE_INT_MAX)
3049 range.min = res.range.max < navail ? navail - res.range.max : 0;
3050 else
3051 range.min = range.likely;
3053 range.unlikely = (res.range.unlikely < navail
3054 ? navail - res.range.unlikely : 0);
3056 return range;
3059 /* Compute the length of the output resulting from the directive DIR
3060 in a call described by INFO and update the overall result of the call
3061 in *RES. Return true if the directive has been handled. */
3063 static bool
3064 format_directive (const call_info &info,
3065 format_result *res, const directive &dir,
3066 range_query *query)
3068 /* Offset of the beginning of the directive from the beginning
3069 of the format string. */
3070 size_t offset = dir.beg - info.fmtstr;
3071 size_t start = offset;
3072 size_t length = offset + dir.len - !!dir.len;
3074 /* Create a location for the whole directive from the % to the format
3075 specifier. */
3076 substring_loc dirloc (info.fmtloc, TREE_TYPE (info.format),
3077 offset, start, length);
3079 /* Also get the location of the argument if possible.
3080 This doesn't work for integer literals or function calls. */
3081 location_t argloc = UNKNOWN_LOCATION;
3082 if (dir.arg)
3083 argloc = EXPR_LOCATION (dir.arg);
3085 /* Bail when there is no function to compute the output length,
3086 or when minimum length checking has been disabled. */
3087 if (!dir.fmtfunc || res->range.min >= HOST_WIDE_INT_MAX)
3088 return false;
3090 /* Compute the range of lengths of the formatted output. */
3091 fmtresult fmtres = dir.fmtfunc (dir, dir.arg, query);
3093 /* Record whether the output of all directives is known to be
3094 bounded by some maximum, implying that their arguments are
3095 either known exactly or determined to be in a known range
3096 or, for strings, limited by the upper bounds of the arrays
3097 they refer to. */
3098 res->knownrange &= fmtres.knownrange;
3100 if (!fmtres.knownrange)
3102 /* Only when the range is known, check it against the host value
3103 of INT_MAX + (the number of bytes of the "%.*Lf" directive with
3104 INT_MAX precision, which is the longest possible output of any
3105 single directive). That's the largest valid byte count (though
3106 not valid call to a printf-like function because it can never
3107 return such a count). Otherwise, the range doesn't correspond
3108 to known values of the argument. */
3109 if (fmtres.range.max > target_dir_max ())
3111 /* Normalize the MAX counter to avoid having to deal with it
3112 later. The counter can be less than HOST_WIDE_INT_M1U
3113 when compiling for an ILP32 target on an LP64 host. */
3114 fmtres.range.max = HOST_WIDE_INT_M1U;
3115 /* Disable exact and maximum length checking after a failure
3116 to determine the maximum number of characters (for example
3117 for wide characters or wide character strings) but continue
3118 tracking the minimum number of characters. */
3119 res->range.max = HOST_WIDE_INT_M1U;
3122 if (fmtres.range.min > target_dir_max ())
3124 /* Disable exact length checking after a failure to determine
3125 even the minimum number of characters (it shouldn't happen
3126 except in an error) but keep tracking the minimum and maximum
3127 number of characters. */
3128 return true;
3132 /* Buffer for the directive in the host character set (used when
3133 the source character set is different). */
3134 char hostdir[32];
3136 int dirlen = dir.len;
3138 if (fmtres.nullp)
3140 fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3141 "%<%.*s%> directive argument is null",
3142 dirlen, target_to_host (hostdir, sizeof hostdir, dir.beg));
3144 /* Don't bother processing the rest of the format string. */
3145 res->warned = true;
3146 res->range.min = HOST_WIDE_INT_M1U;
3147 res->range.max = HOST_WIDE_INT_M1U;
3148 return false;
3151 /* Compute the number of available bytes in the destination. There
3152 must always be at least one byte of space for the terminating
3153 NUL that's appended after the format string has been processed. */
3154 result_range avail_range = bytes_remaining (info.objsize, *res);
3156 /* If the argument aliases a part of the destination of the formatted
3157 call at offset FMTRES.DST_OFFSET append the directive and its result
3158 to the set of aliases for later processing. */
3159 if (fmtres.dst_offset != HOST_WIDE_INT_MIN)
3160 res->append_alias (dir, fmtres.dst_offset, fmtres.range);
3162 bool warned = res->warned;
3164 if (!warned)
3165 warned = maybe_warn (dirloc, argloc, info, avail_range,
3166 fmtres.range, dir);
3168 /* Bump up the total maximum if it isn't too big. */
3169 if (res->range.max < HOST_WIDE_INT_MAX
3170 && fmtres.range.max < HOST_WIDE_INT_MAX)
3171 res->range.max += fmtres.range.max;
3173 /* Raise the total unlikely maximum by the larger of the maximum
3174 and the unlikely maximum. */
3175 unsigned HOST_WIDE_INT save = res->range.unlikely;
3176 if (fmtres.range.max < fmtres.range.unlikely)
3177 res->range.unlikely += fmtres.range.unlikely;
3178 else
3179 res->range.unlikely += fmtres.range.max;
3181 if (res->range.unlikely < save)
3182 res->range.unlikely = HOST_WIDE_INT_M1U;
3184 res->range.min += fmtres.range.min;
3185 res->range.likely += fmtres.range.likely;
3187 /* Has the minimum directive output length exceeded the maximum
3188 of 4095 bytes required to be supported? */
3189 bool minunder4k = fmtres.range.min < 4096;
3190 bool maxunder4k = fmtres.range.max < 4096;
3191 /* Clear POSUNDER4K in the overall result if the maximum has exceeded
3192 the 4k (this is necessary to avoid the return value optimization
3193 that may not be safe in the maximum case). */
3194 if (!maxunder4k)
3195 res->posunder4k = false;
3196 /* Also clear POSUNDER4K if the directive may fail. */
3197 if (fmtres.mayfail)
3198 res->posunder4k = false;
3200 if (!warned
3201 /* Only warn at level 2. */
3202 && warn_level > 1
3203 /* Only warn for string functions. */
3204 && info.is_string_func ()
3205 && (!minunder4k
3206 || (!maxunder4k && fmtres.range.max < HOST_WIDE_INT_MAX)))
3208 /* The directive output may be longer than the maximum required
3209 to be handled by an implementation according to 7.21.6.1, p15
3210 of C11. Warn on this only at level 2 but remember this and
3211 prevent folding the return value when done. This allows for
3212 the possibility of the actual libc call failing due to ENOMEM
3213 (like Glibc does with very large precision or width).
3214 Issue the "may exceed" warning only for string functions and
3215 not for fprintf or printf. */
3217 if (fmtres.range.min == fmtres.range.max)
3218 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3219 "%<%.*s%> directive output of %wu bytes exceeds "
3220 "minimum required size of 4095", dirlen,
3221 target_to_host (hostdir, sizeof hostdir, dir.beg),
3222 fmtres.range.min);
3223 else if (!minunder4k)
3224 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3225 "%<%.*s%> directive output between %wu and %wu "
3226 "bytes exceeds minimum required size of 4095",
3227 dirlen,
3228 target_to_host (hostdir, sizeof hostdir, dir.beg),
3229 fmtres.range.min, fmtres.range.max);
3230 else if (!info.retval_used () && info.is_string_func ())
3231 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3232 "%<%.*s%> directive output between %wu and %wu "
3233 "bytes may exceed minimum required size of "
3234 "4095",
3235 dirlen,
3236 target_to_host (hostdir, sizeof hostdir, dir.beg),
3237 fmtres.range.min, fmtres.range.max);
3240 /* Has the likely and maximum directive output exceeded INT_MAX? */
3241 bool likelyximax = *dir.beg && res->range.likely > target_int_max ();
3242 /* Don't consider the maximum to be in excess when it's the result
3243 of a string of unknown length (i.e., whose maximum has been set
3244 to be greater than or equal to HOST_WIDE_INT_MAX. */
3245 bool maxximax = (*dir.beg
3246 && res->range.max > target_int_max ()
3247 && res->range.max < HOST_WIDE_INT_MAX);
3249 if (!warned
3250 /* Warn for the likely output size at level 1. */
3251 && (likelyximax
3252 /* But only warn for the maximum at level 2. */
3253 || (warn_level > 1
3254 && maxximax
3255 && fmtres.range.max < HOST_WIDE_INT_MAX)))
3257 if (fmtres.range.min > target_int_max ())
3259 /* The directive output exceeds INT_MAX bytes. */
3260 if (fmtres.range.min == fmtres.range.max)
3261 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3262 "%<%.*s%> directive output of %wu bytes exceeds "
3263 "%<INT_MAX%>", dirlen,
3264 target_to_host (hostdir, sizeof hostdir, dir.beg),
3265 fmtres.range.min);
3266 else
3267 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3268 "%<%.*s%> directive output between %wu and "
3269 "%wu bytes exceeds %<INT_MAX%>", dirlen,
3270 target_to_host (hostdir, sizeof hostdir, dir.beg),
3271 fmtres.range.min, fmtres.range.max);
3273 else if (res->range.min > target_int_max ())
3275 /* The directive output is under INT_MAX but causes the result
3276 to exceed INT_MAX bytes. */
3277 if (fmtres.range.min == fmtres.range.max)
3278 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3279 "%<%.*s%> directive output of %wu bytes causes "
3280 "result to exceed %<INT_MAX%>", dirlen,
3281 target_to_host (hostdir, sizeof hostdir, dir.beg),
3282 fmtres.range.min);
3283 else
3284 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3285 "%<%.*s%> directive output between %wu and "
3286 "%wu bytes causes result to exceed %<INT_MAX%>",
3287 dirlen,
3288 target_to_host (hostdir, sizeof hostdir, dir.beg),
3289 fmtres.range.min, fmtres.range.max);
3291 else if ((!info.retval_used () || !info.bounded)
3292 && (info.is_string_func ()))
3293 /* Warn for calls to string functions that either aren't bounded
3294 (sprintf) or whose return value isn't used. */
3295 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3296 "%<%.*s%> directive output between %wu and "
3297 "%wu bytes may cause result to exceed "
3298 "%<INT_MAX%>", dirlen,
3299 target_to_host (hostdir, sizeof hostdir, dir.beg),
3300 fmtres.range.min, fmtres.range.max);
3303 if (!warned && fmtres.nonstr)
3305 warned = fmtwarn (dirloc, argloc, NULL, info.warnopt (),
3306 "%<%.*s%> directive argument is not a nul-terminated "
3307 "string",
3308 dirlen,
3309 target_to_host (hostdir, sizeof hostdir, dir.beg));
3310 if (warned && DECL_P (fmtres.nonstr))
3311 inform (DECL_SOURCE_LOCATION (fmtres.nonstr),
3312 "referenced argument declared here");
3313 return false;
3316 if (warned && fmtres.range.min < fmtres.range.likely
3317 && fmtres.range.likely < fmtres.range.max)
3318 inform_n (info.fmtloc, fmtres.range.likely,
3319 "assuming directive output of %wu byte",
3320 "assuming directive output of %wu bytes",
3321 fmtres.range.likely);
3323 if (warned && fmtres.argmin)
3325 if (fmtres.argmin == fmtres.argmax)
3326 inform (info.fmtloc, "directive argument %qE", fmtres.argmin);
3327 else if (fmtres.knownrange)
3328 inform (info.fmtloc, "directive argument in the range [%E, %E]",
3329 fmtres.argmin, fmtres.argmax);
3330 else
3331 inform (info.fmtloc,
3332 "using the range [%E, %E] for directive argument",
3333 fmtres.argmin, fmtres.argmax);
3336 res->warned |= warned;
3338 if (!dir.beg[0] && res->warned)
3340 location_t callloc = gimple_location (info.callstmt);
3342 unsigned HOST_WIDE_INT min = res->range.min;
3343 unsigned HOST_WIDE_INT max = res->range.max;
3345 if (info.objsize < HOST_WIDE_INT_MAX)
3347 /* If a warning has been issued for buffer overflow or truncation
3348 help the user figure out how big a buffer they need. */
3350 if (min == max)
3351 inform_n (callloc, min,
3352 "%qE output %wu byte into a destination of size %wu",
3353 "%qE output %wu bytes into a destination of size %wu",
3354 info.func, min, info.objsize);
3355 else if (max < HOST_WIDE_INT_MAX)
3356 inform (callloc,
3357 "%qE output between %wu and %wu bytes into "
3358 "a destination of size %wu",
3359 info.func, min, max, info.objsize);
3360 else if (min < res->range.likely && res->range.likely < max)
3361 inform (callloc,
3362 "%qE output %wu or more bytes (assuming %wu) into "
3363 "a destination of size %wu",
3364 info.func, min, res->range.likely, info.objsize);
3365 else
3366 inform (callloc,
3367 "%qE output %wu or more bytes into a destination of size "
3368 "%wu",
3369 info.func, min, info.objsize);
3371 else if (!info.is_string_func ())
3373 /* If the warning is for a file function like fprintf
3374 of printf with no destination size just print the computed
3375 result. */
3376 if (min == max)
3377 inform_n (callloc, min,
3378 "%qE output %wu byte", "%qE output %wu bytes",
3379 info.func, min);
3380 else if (max < HOST_WIDE_INT_MAX)
3381 inform (callloc,
3382 "%qE output between %wu and %wu bytes",
3383 info.func, min, max);
3384 else if (min < res->range.likely && res->range.likely < max)
3385 inform (callloc,
3386 "%qE output %wu or more bytes (assuming %wu)",
3387 info.func, min, res->range.likely);
3388 else
3389 inform (callloc,
3390 "%qE output %wu or more bytes",
3391 info.func, min);
3395 if (dump_file && *dir.beg)
3397 fprintf (dump_file,
3398 " Result: "
3399 HOST_WIDE_INT_PRINT_DEC ", " HOST_WIDE_INT_PRINT_DEC ", "
3400 HOST_WIDE_INT_PRINT_DEC ", " HOST_WIDE_INT_PRINT_DEC " ("
3401 HOST_WIDE_INT_PRINT_DEC ", " HOST_WIDE_INT_PRINT_DEC ", "
3402 HOST_WIDE_INT_PRINT_DEC ", " HOST_WIDE_INT_PRINT_DEC ")\n",
3403 fmtres.range.min, fmtres.range.likely,
3404 fmtres.range.max, fmtres.range.unlikely,
3405 res->range.min, res->range.likely,
3406 res->range.max, res->range.unlikely);
3409 return true;
3412 /* Parse a format directive in function call described by INFO starting
3413 at STR and populate DIR structure. Bump up *ARGNO by the number of
3414 arguments extracted for the directive. Return the length of
3415 the directive. */
3417 static size_t
3418 parse_directive (call_info &info,
3419 directive &dir, format_result *res,
3420 const char *str, unsigned *argno,
3421 range_query *query)
3423 const char *pcnt = strchr (str, target_percent);
3424 dir.beg = str;
3426 if (size_t len = pcnt ? pcnt - str : *str ? strlen (str) : 1)
3428 /* This directive is either a plain string or the terminating nul
3429 (which isn't really a directive but it simplifies things to
3430 handle it as if it were). */
3431 dir.len = len;
3432 dir.fmtfunc = format_plain;
3434 if (dump_file)
3436 fprintf (dump_file, " Directive %u at offset "
3437 HOST_WIDE_INT_PRINT_UNSIGNED ": \"%.*s\", "
3438 "length = " HOST_WIDE_INT_PRINT_UNSIGNED "\n",
3439 dir.dirno,
3440 (unsigned HOST_WIDE_INT)(size_t)(dir.beg - info.fmtstr),
3441 (int)dir.len, dir.beg, (unsigned HOST_WIDE_INT) dir.len);
3444 return len - !*str;
3447 /* Set the directive argument's number to correspond to its position
3448 in the formatted function call's argument list. */
3449 dir.argno = *argno;
3451 const char *pf = pcnt + 1;
3453 /* POSIX numbered argument index or zero when none. */
3454 HOST_WIDE_INT dollar = 0;
3456 /* With and precision. -1 when not specified, HOST_WIDE_INT_MIN
3457 when given by a va_list argument, and a non-negative value
3458 when specified in the format string itself. */
3459 HOST_WIDE_INT width = -1;
3460 HOST_WIDE_INT precision = -1;
3462 /* Pointers to the beginning of the width and precision decimal
3463 string (if any) within the directive. */
3464 const char *pwidth = 0;
3465 const char *pprec = 0;
3467 /* When the value of the decimal string that specifies width or
3468 precision is out of range, points to the digit that causes
3469 the value to exceed the limit. */
3470 const char *werange = NULL;
3471 const char *perange = NULL;
3473 /* Width specified via the asterisk. Need not be INTEGER_CST.
3474 For vararg functions set to void_node. */
3475 tree star_width = NULL_TREE;
3477 /* Width specified via the asterisk. Need not be INTEGER_CST.
3478 For vararg functions set to void_node. */
3479 tree star_precision = NULL_TREE;
3481 if (ISDIGIT (target_to_host (*pf)))
3483 /* This could be either a POSIX positional argument, the '0'
3484 flag, or a width, depending on what follows. Store it as
3485 width and sort it out later after the next character has
3486 been seen. */
3487 pwidth = pf;
3488 width = target_strtowi (&pf, &werange);
3490 else if (target_to_host (*pf) == '*')
3492 /* Similarly to the block above, this could be either a POSIX
3493 positional argument or a width, depending on what follows. */
3494 if (*argno < gimple_call_num_args (info.callstmt))
3495 star_width = gimple_call_arg (info.callstmt, (*argno)++);
3496 else
3497 star_width = void_node;
3498 ++pf;
3501 if (target_to_host (*pf) == '$')
3503 /* Handle the POSIX dollar sign which references the 1-based
3504 positional argument number. */
3505 if (width != -1)
3506 dollar = width + info.argidx;
3507 else if (star_width
3508 && TREE_CODE (star_width) == INTEGER_CST
3509 && (TYPE_PRECISION (TREE_TYPE (star_width))
3510 <= TYPE_PRECISION (integer_type_node)))
3511 dollar = width + tree_to_shwi (star_width);
3513 /* Bail when the numbered argument is out of range (it will
3514 have already been diagnosed by -Wformat). */
3515 if (dollar == 0
3516 || dollar == (int)info.argidx
3517 || dollar > gimple_call_num_args (info.callstmt))
3518 return false;
3520 --dollar;
3522 star_width = NULL_TREE;
3523 width = -1;
3524 ++pf;
3527 if (dollar || !star_width)
3529 if (width != -1)
3531 if (width == 0)
3533 /* The '0' that has been interpreted as a width above is
3534 actually a flag. Reset HAVE_WIDTH, set the '0' flag,
3535 and continue processing other flags. */
3536 width = -1;
3537 dir.set_flag ('0');
3539 else if (!dollar)
3541 /* (Non-zero) width has been seen. The next character
3542 is either a period or a digit. */
3543 goto start_precision;
3546 /* When either '$' has been seen, or width has not been seen,
3547 the next field is the optional flags followed by an optional
3548 width. */
3549 for ( ; ; ) {
3550 switch (target_to_host (*pf))
3552 case ' ':
3553 case '0':
3554 case '+':
3555 case '-':
3556 case '#':
3557 dir.set_flag (target_to_host (*pf++));
3558 break;
3560 default:
3561 goto start_width;
3565 start_width:
3566 if (ISDIGIT (target_to_host (*pf)))
3568 werange = 0;
3569 pwidth = pf;
3570 width = target_strtowi (&pf, &werange);
3572 else if (target_to_host (*pf) == '*')
3574 if (*argno < gimple_call_num_args (info.callstmt))
3575 star_width = gimple_call_arg (info.callstmt, (*argno)++);
3576 else
3578 /* This is (likely) a va_list. It could also be an invalid
3579 call with insufficient arguments. */
3580 star_width = void_node;
3582 ++pf;
3584 else if (target_to_host (*pf) == '\'')
3586 /* The POSIX apostrophe indicating a numeric grouping
3587 in the current locale. Even though it's possible to
3588 estimate the upper bound on the size of the output
3589 based on the number of digits it probably isn't worth
3590 continuing. */
3591 return 0;
3595 start_precision:
3596 if (target_to_host (*pf) == '.')
3598 ++pf;
3600 if (ISDIGIT (target_to_host (*pf)))
3602 pprec = pf;
3603 precision = target_strtowi (&pf, &perange);
3605 else if (target_to_host (*pf) == '*')
3607 if (*argno < gimple_call_num_args (info.callstmt))
3608 star_precision = gimple_call_arg (info.callstmt, (*argno)++);
3609 else
3611 /* This is (likely) a va_list. It could also be an invalid
3612 call with insufficient arguments. */
3613 star_precision = void_node;
3615 ++pf;
3617 else
3619 /* The decimal precision or the asterisk are optional.
3620 When neither is specified it's taken to be zero. */
3621 precision = 0;
3625 switch (target_to_host (*pf))
3627 case 'h':
3628 if (target_to_host (pf[1]) == 'h')
3630 ++pf;
3631 dir.modifier = FMT_LEN_hh;
3633 else
3634 dir.modifier = FMT_LEN_h;
3635 ++pf;
3636 break;
3638 case 'j':
3639 dir.modifier = FMT_LEN_j;
3640 ++pf;
3641 break;
3643 case 'L':
3644 dir.modifier = FMT_LEN_L;
3645 ++pf;
3646 break;
3648 case 'l':
3649 if (target_to_host (pf[1]) == 'l')
3651 ++pf;
3652 dir.modifier = FMT_LEN_ll;
3654 else
3655 dir.modifier = FMT_LEN_l;
3656 ++pf;
3657 break;
3659 case 't':
3660 dir.modifier = FMT_LEN_t;
3661 ++pf;
3662 break;
3664 case 'z':
3665 dir.modifier = FMT_LEN_z;
3666 ++pf;
3667 break;
3670 switch (target_to_host (*pf))
3672 /* Handle a sole '%' character the same as "%%" but since it's
3673 undefined prevent the result from being folded. */
3674 case '\0':
3675 --pf;
3676 res->range.min = res->range.max = HOST_WIDE_INT_M1U;
3677 /* FALLTHRU */
3678 case '%':
3679 dir.fmtfunc = format_percent;
3680 break;
3682 case 'a':
3683 case 'A':
3684 case 'e':
3685 case 'E':
3686 case 'f':
3687 case 'F':
3688 case 'g':
3689 case 'G':
3690 res->floating = true;
3691 dir.fmtfunc = format_floating;
3692 break;
3694 case 'd':
3695 case 'i':
3696 case 'o':
3697 case 'u':
3698 case 'x':
3699 case 'X':
3700 dir.fmtfunc = format_integer;
3701 break;
3703 case 'p':
3704 /* The %p output is implementation-defined. It's possible
3705 to determine this format but due to extensions (especially
3706 those of the Linux kernel -- see bug 78512) the first %p
3707 in the format string disables any further processing. */
3708 return false;
3710 case 'n':
3711 /* %n has side-effects even when nothing is actually printed to
3712 any buffer. */
3713 info.nowrite = false;
3714 dir.fmtfunc = format_none;
3715 break;
3717 case 'C':
3718 case 'c':
3719 /* POSIX wide character and C/POSIX narrow character. */
3720 dir.fmtfunc = format_character;
3721 break;
3723 case 'S':
3724 case 's':
3725 /* POSIX wide string and C/POSIX narrow character string. */
3726 dir.fmtfunc = format_string;
3727 break;
3729 default:
3730 /* Unknown conversion specification. */
3731 return 0;
3734 dir.specifier = target_to_host (*pf++);
3736 /* Store the length of the format directive. */
3737 dir.len = pf - pcnt;
3739 /* Buffer for the directive in the host character set (used when
3740 the source character set is different). */
3741 char hostdir[32];
3743 if (star_width)
3745 if (INTEGRAL_TYPE_P (TREE_TYPE (star_width)))
3746 dir.set_width (star_width, query);
3747 else
3749 /* Width specified by a va_list takes on the range [0, -INT_MIN]
3750 (width is the absolute value of that specified). */
3751 dir.width[0] = 0;
3752 dir.width[1] = target_int_max () + 1;
3755 else
3757 if (width == HOST_WIDE_INT_MAX && werange)
3759 size_t begin = dir.beg - info.fmtstr + (pwidth - pcnt);
3760 size_t caret = begin + (werange - pcnt);
3761 size_t end = pf - info.fmtstr - 1;
3763 /* Create a location for the width part of the directive,
3764 pointing the caret at the first out-of-range digit. */
3765 substring_loc dirloc (info.fmtloc, TREE_TYPE (info.format),
3766 caret, begin, end);
3768 fmtwarn (dirloc, UNKNOWN_LOCATION, NULL, info.warnopt (),
3769 "%<%.*s%> directive width out of range", (int) dir.len,
3770 target_to_host (hostdir, sizeof hostdir, dir.beg));
3773 dir.set_width (width);
3776 if (star_precision)
3778 if (INTEGRAL_TYPE_P (TREE_TYPE (star_precision)))
3779 dir.set_precision (star_precision, query);
3780 else
3782 /* Precision specified by a va_list takes on the range [-1, INT_MAX]
3783 (unlike width, negative precision is ignored). */
3784 dir.prec[0] = -1;
3785 dir.prec[1] = target_int_max ();
3788 else
3790 if (precision == HOST_WIDE_INT_MAX && perange)
3792 size_t begin = dir.beg - info.fmtstr + (pprec - pcnt) - 1;
3793 size_t caret = dir.beg - info.fmtstr + (perange - pcnt) - 1;
3794 size_t end = pf - info.fmtstr - 2;
3796 /* Create a location for the precision part of the directive,
3797 including the leading period, pointing the caret at the first
3798 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 precision out of range", (int) dir.len,
3804 target_to_host (hostdir, sizeof hostdir, dir.beg));
3807 dir.set_precision (precision);
3810 /* Extract the argument if the directive takes one and if it's
3811 available (e.g., the function doesn't take a va_list). Treat
3812 missing arguments the same as va_list, even though they will
3813 have likely already been diagnosed by -Wformat. */
3814 if (dir.specifier != '%'
3815 && *argno < gimple_call_num_args (info.callstmt))
3816 dir.arg = gimple_call_arg (info.callstmt, dollar ? dollar : (*argno)++);
3818 if (dump_file)
3820 fprintf (dump_file,
3821 " Directive %u at offset " HOST_WIDE_INT_PRINT_UNSIGNED
3822 ": \"%.*s\"",
3823 dir.dirno,
3824 (unsigned HOST_WIDE_INT)(size_t)(dir.beg - info.fmtstr),
3825 (int)dir.len, dir.beg);
3826 if (star_width)
3828 if (dir.width[0] == dir.width[1])
3829 fprintf (dump_file, ", width = " HOST_WIDE_INT_PRINT_DEC,
3830 dir.width[0]);
3831 else
3832 fprintf (dump_file,
3833 ", width in range [" HOST_WIDE_INT_PRINT_DEC
3834 ", " HOST_WIDE_INT_PRINT_DEC "]",
3835 dir.width[0], dir.width[1]);
3838 if (star_precision)
3840 if (dir.prec[0] == dir.prec[1])
3841 fprintf (dump_file, ", precision = " HOST_WIDE_INT_PRINT_DEC,
3842 dir.prec[0]);
3843 else
3844 fprintf (dump_file,
3845 ", precision in range [" HOST_WIDE_INT_PRINT_DEC
3846 HOST_WIDE_INT_PRINT_DEC "]",
3847 dir.prec[0], dir.prec[1]);
3849 fputc ('\n', dump_file);
3852 return dir.len;
3855 /* Diagnose overlap between destination and %s directive arguments. */
3857 static void
3858 maybe_warn_overlap (call_info &info, format_result *res)
3860 /* Two vectors of 1-based indices corresponding to either certainly
3861 or possibly aliasing arguments. */
3862 auto_vec<int, 16> aliasarg[2];
3864 /* Go through the array of potentially aliasing directives and collect
3865 argument numbers of those that do or may overlap the destination
3866 object given the full result. */
3867 for (unsigned i = 0; i != res->alias_count; ++i)
3869 const format_result::alias_info &alias = res->aliases[i];
3871 enum { possible = -1, none = 0, certain = 1 } overlap = none;
3873 /* If the precision is zero there is no overlap. (This only
3874 considers %s directives and ignores %n.) */
3875 if (alias.dir.prec[0] == 0 && alias.dir.prec[1] == 0)
3876 continue;
3878 if (alias.offset == HOST_WIDE_INT_MAX
3879 || info.dst_offset == HOST_WIDE_INT_MAX)
3880 overlap = possible;
3881 else if (alias.offset == info.dst_offset)
3882 overlap = alias.dir.prec[0] == 0 ? possible : certain;
3883 else
3885 /* Determine overlap from the range of output and offsets
3886 into the same destination as the source, and rule out
3887 impossible overlap. */
3888 unsigned HOST_WIDE_INT albeg = alias.offset;
3889 unsigned HOST_WIDE_INT dstbeg = info.dst_offset;
3891 unsigned HOST_WIDE_INT alend = albeg + alias.range.min;
3892 unsigned HOST_WIDE_INT dstend = dstbeg + res->range.min - 1;
3894 if ((albeg <= dstbeg && alend > dstbeg)
3895 || (albeg >= dstbeg && albeg < dstend))
3896 overlap = certain;
3897 else
3899 alend = albeg + alias.range.max;
3900 if (alend < albeg)
3901 alend = HOST_WIDE_INT_M1U;
3903 dstend = dstbeg + res->range.max - 1;
3904 if (dstend < dstbeg)
3905 dstend = HOST_WIDE_INT_M1U;
3907 if ((albeg >= dstbeg && albeg <= dstend)
3908 || (alend >= dstbeg && alend <= dstend))
3909 overlap = possible;
3913 if (overlap == none)
3914 continue;
3916 /* Append the 1-based argument number. */
3917 aliasarg[overlap != certain].safe_push (alias.dir.argno + 1);
3919 /* Disable any kind of optimization. */
3920 res->range.unlikely = HOST_WIDE_INT_M1U;
3923 tree arg0 = gimple_call_arg (info.callstmt, 0);
3924 location_t loc = gimple_location (info.callstmt);
3926 bool aliaswarn = false;
3928 unsigned ncertain = aliasarg[0].length ();
3929 unsigned npossible = aliasarg[1].length ();
3930 if (ncertain && npossible)
3932 /* If there are multiple arguments that overlap, some certainly
3933 and some possibly, handle both sets in a single diagnostic. */
3934 aliaswarn
3935 = warning_at (loc, OPT_Wrestrict,
3936 "%qE arguments %Z and maybe %Z overlap destination "
3937 "object %qE",
3938 info.func, aliasarg[0].address (), ncertain,
3939 aliasarg[1].address (), npossible,
3940 info.dst_origin);
3942 else if (ncertain)
3944 /* There is only one set of two or more arguments and they all
3945 certainly overlap the destination. */
3946 aliaswarn
3947 = warning_n (loc, OPT_Wrestrict, ncertain,
3948 "%qE argument %Z overlaps destination object %qE",
3949 "%qE arguments %Z overlap destination object %qE",
3950 info.func, aliasarg[0].address (), ncertain,
3951 info.dst_origin);
3953 else if (npossible)
3955 /* There is only one set of two or more arguments and they all
3956 may overlap (but need not). */
3957 aliaswarn
3958 = warning_n (loc, OPT_Wrestrict, npossible,
3959 "%qE argument %Z may overlap destination object %qE",
3960 "%qE arguments %Z may overlap destination object %qE",
3961 info.func, aliasarg[1].address (), npossible,
3962 info.dst_origin);
3965 if (aliaswarn)
3967 res->warned = true;
3969 if (info.dst_origin != arg0)
3971 /* If its location is different from the first argument of the call
3972 point either at the destination object itself or at the expression
3973 that was used to determine the overlap. */
3974 loc = (DECL_P (info.dst_origin)
3975 ? DECL_SOURCE_LOCATION (info.dst_origin)
3976 : EXPR_LOCATION (info.dst_origin));
3977 if (loc != UNKNOWN_LOCATION)
3978 inform (loc,
3979 "destination object referenced by %<restrict%>-qualified "
3980 "argument 1 was declared here");
3985 /* Compute the length of the output resulting from the call to a formatted
3986 output function described by INFO and store the result of the call in
3987 *RES. Issue warnings for detected past the end writes. Return true
3988 if the complete format string has been processed and *RES can be relied
3989 on, false otherwise (e.g., when a unknown or unhandled directive was seen
3990 that caused the processing to be terminated early). */
3992 static bool
3993 compute_format_length (call_info &info, format_result *res, range_query *query)
3995 if (dump_file)
3997 location_t callloc = gimple_location (info.callstmt);
3998 fprintf (dump_file, "%s:%i: ",
3999 LOCATION_FILE (callloc), LOCATION_LINE (callloc));
4000 print_generic_expr (dump_file, info.func, dump_flags);
4002 fprintf (dump_file,
4003 ": objsize = " HOST_WIDE_INT_PRINT_UNSIGNED
4004 ", fmtstr = \"%s\"\n",
4005 info.objsize, info.fmtstr);
4008 /* Reset the minimum and maximum byte counters. */
4009 res->range.min = res->range.max = 0;
4011 /* No directive has been seen yet so the length of output is bounded
4012 by the known range [0, 0] (with no conversion resulting in a failure
4013 or producing more than 4K bytes) until determined otherwise. */
4014 res->knownrange = true;
4015 res->floating = false;
4016 res->warned = false;
4018 /* 1-based directive counter. */
4019 unsigned dirno = 1;
4021 /* The variadic argument counter. */
4022 unsigned argno = info.argidx;
4024 bool success = true;
4026 for (const char *pf = info.fmtstr; ; ++dirno)
4028 directive dir (&info, dirno);
4030 size_t n = parse_directive (info, dir, res, pf, &argno, query);
4032 /* Return failure if the format function fails. */
4033 if (!format_directive (info, res, dir, query))
4034 return false;
4036 /* Return success when the directive is zero bytes long and it's
4037 the last thing in the format string (i.e., it's the terminating
4038 nul, which isn't really a directive but handling it as one makes
4039 things simpler). */
4040 if (!n)
4042 success = *pf == '\0';
4043 break;
4046 pf += n;
4049 maybe_warn_overlap (info, res);
4051 /* The complete format string was processed (with or without warnings). */
4052 return success;
4055 /* Return the size of the object referenced by the expression DEST in
4056 statement STMT, if available, or the maximum possible size otherwise. */
4058 static unsigned HOST_WIDE_INT
4059 get_destination_size (tree dest, gimple *stmt, pointer_query &ptr_qry)
4061 /* When there is no destination return the maximum. */
4062 if (!dest)
4063 return HOST_WIDE_INT_MAX;
4065 /* Use compute_objsize to determine the size of the destination object. */
4066 access_ref aref;
4067 if (!ptr_qry.get_ref (dest, stmt, &aref))
4068 return HOST_WIDE_INT_MAX;
4070 offset_int remsize = aref.size_remaining ();
4071 if (!wi::fits_uhwi_p (remsize))
4072 return HOST_WIDE_INT_MAX;
4074 return remsize.to_uhwi ();
4077 /* Return true if the call described by INFO with result RES safe to
4078 optimize (i.e., no undefined behavior), and set RETVAL to the range
4079 of its return values. */
4081 static bool
4082 is_call_safe (const call_info &info,
4083 const format_result &res, bool under4k,
4084 unsigned HOST_WIDE_INT retval[2])
4086 if (under4k && !res.posunder4k)
4087 return false;
4089 /* The minimum return value. */
4090 retval[0] = res.range.min;
4092 /* The maximum return value is in most cases bounded by RES.RANGE.MAX
4093 but in cases involving multibyte characters could be as large as
4094 RES.RANGE.UNLIKELY. */
4095 retval[1]
4096 = res.range.unlikely < res.range.max ? res.range.max : res.range.unlikely;
4098 /* Adjust the number of bytes which includes the terminating nul
4099 to reflect the return value of the function which does not.
4100 Because the valid range of the function is [INT_MIN, INT_MAX],
4101 a valid range before the adjustment below is [0, INT_MAX + 1]
4102 (the functions only return negative values on error or undefined
4103 behavior). */
4104 if (retval[0] <= target_int_max () + 1)
4105 --retval[0];
4106 if (retval[1] <= target_int_max () + 1)
4107 --retval[1];
4109 /* Avoid the return value optimization when the behavior of the call
4110 is undefined either because any directive may have produced 4K or
4111 more of output, or the return value exceeds INT_MAX, or because
4112 the output overflows the destination object (but leave it enabled
4113 when the function is bounded because then the behavior is well-
4114 defined). */
4115 if (retval[0] == retval[1]
4116 && (info.bounded || retval[0] < info.objsize)
4117 && retval[0] <= target_int_max ())
4118 return true;
4120 if ((info.bounded || retval[1] < info.objsize)
4121 && (retval[0] < target_int_max ()
4122 && retval[1] < target_int_max ()))
4123 return true;
4125 if (!under4k && (info.bounded || retval[0] < info.objsize))
4126 return true;
4128 return false;
4131 /* Given a suitable result RES of a call to a formatted output function
4132 described by INFO, substitute the result for the return value of
4133 the call. The result is suitable if the number of bytes it represents
4134 is known and exact. A result that isn't suitable for substitution may
4135 have its range set to the range of return values, if that is known.
4136 Return true if the call is removed and gsi_next should not be performed
4137 in the caller. */
4139 static bool
4140 try_substitute_return_value (gimple_stmt_iterator *gsi,
4141 const call_info &info,
4142 const format_result &res)
4144 tree lhs = gimple_get_lhs (info.callstmt);
4146 /* Set to true when the entire call has been removed. */
4147 bool removed = false;
4149 /* The minimum and maximum return value. */
4150 unsigned HOST_WIDE_INT retval[2] = {0};
4151 bool safe = is_call_safe (info, res, true, retval);
4153 if (safe
4154 && retval[0] == retval[1]
4155 /* Not prepared to handle possibly throwing calls here; they shouldn't
4156 appear in non-artificial testcases, except when the __*_chk routines
4157 are badly declared. */
4158 && !stmt_ends_bb_p (info.callstmt))
4160 tree cst = build_int_cst (lhs ? TREE_TYPE (lhs) : integer_type_node,
4161 retval[0]);
4163 if (lhs == NULL_TREE && info.nowrite)
4165 /* Remove the call to the bounded function with a zero size
4166 (e.g., snprintf(0, 0, "%i", 123)) if there is no lhs. */
4167 unlink_stmt_vdef (info.callstmt);
4168 gsi_remove (gsi, true);
4169 removed = true;
4171 else if (info.nowrite)
4173 /* Replace the call to the bounded function with a zero size
4174 (e.g., snprintf(0, 0, "%i", 123) with the constant result
4175 of the function. */
4176 gimplify_and_update_call_from_tree (gsi, cst);
4177 gimple *callstmt = gsi_stmt (*gsi);
4178 update_stmt (callstmt);
4180 else if (lhs)
4182 /* Replace the left-hand side of the call with the constant
4183 result of the formatted function. */
4184 gimple_call_set_lhs (info.callstmt, NULL_TREE);
4185 gimple *g = gimple_build_assign (lhs, cst);
4186 gsi_insert_after (gsi, g, GSI_NEW_STMT);
4187 update_stmt (info.callstmt);
4190 if (dump_file)
4192 if (removed)
4193 fprintf (dump_file, " Removing call statement.");
4194 else
4196 fprintf (dump_file, " Substituting ");
4197 print_generic_expr (dump_file, cst, dump_flags);
4198 fprintf (dump_file, " for %s.\n",
4199 info.nowrite ? "statement" : "return value");
4203 else if (lhs && types_compatible_p (TREE_TYPE (lhs), integer_type_node))
4205 bool setrange = false;
4207 if (safe
4208 && (info.bounded || retval[1] < info.objsize)
4209 && (retval[0] < target_int_max ()
4210 && retval[1] < target_int_max ()))
4212 /* If the result is in a valid range bounded by the size of
4213 the destination set it so that it can be used for subsequent
4214 optimizations. */
4215 int prec = TYPE_PRECISION (integer_type_node);
4217 wide_int min = wi::shwi (retval[0], prec);
4218 wide_int max = wi::shwi (retval[1], prec);
4219 set_range_info (lhs, VR_RANGE, min, max);
4221 setrange = true;
4224 if (dump_file)
4226 const char *inbounds
4227 = (retval[0] < info.objsize
4228 ? (retval[1] < info.objsize
4229 ? "in" : "potentially out-of")
4230 : "out-of");
4232 const char *what = setrange ? "Setting" : "Discarding";
4233 if (retval[0] != retval[1])
4234 fprintf (dump_file,
4235 " %s %s-bounds return value range ["
4236 HOST_WIDE_INT_PRINT_UNSIGNED ", "
4237 HOST_WIDE_INT_PRINT_UNSIGNED "].\n",
4238 what, inbounds, retval[0], retval[1]);
4239 else
4240 fprintf (dump_file, " %s %s-bounds return value "
4241 HOST_WIDE_INT_PRINT_UNSIGNED ".\n",
4242 what, inbounds, retval[0]);
4246 if (dump_file)
4247 fputc ('\n', dump_file);
4249 return removed;
4252 /* Try to simplify a s{,n}printf call described by INFO with result
4253 RES by replacing it with a simpler and presumably more efficient
4254 call (such as strcpy). */
4256 static bool
4257 try_simplify_call (gimple_stmt_iterator *gsi,
4258 const call_info &info,
4259 const format_result &res)
4261 unsigned HOST_WIDE_INT dummy[2];
4262 if (!is_call_safe (info, res, info.retval_used (), dummy))
4263 return false;
4265 switch (info.fncode)
4267 case BUILT_IN_SNPRINTF:
4268 return gimple_fold_builtin_snprintf (gsi);
4270 case BUILT_IN_SPRINTF:
4271 return gimple_fold_builtin_sprintf (gsi);
4273 default:
4277 return false;
4280 /* Return the zero-based index of the format string argument of a printf
4281 like function and set *IDX_ARGS to the first format argument. When
4282 no such index exists return UINT_MAX. */
4284 static unsigned
4285 get_user_idx_format (tree fndecl, unsigned *idx_args)
4287 tree attrs = lookup_attribute ("format", DECL_ATTRIBUTES (fndecl));
4288 if (!attrs)
4289 attrs = lookup_attribute ("format", TYPE_ATTRIBUTES (TREE_TYPE (fndecl)));
4291 if (!attrs)
4292 return UINT_MAX;
4294 attrs = TREE_VALUE (attrs);
4296 tree archetype = TREE_VALUE (attrs);
4297 if (strcmp ("printf", IDENTIFIER_POINTER (archetype)))
4298 return UINT_MAX;
4300 attrs = TREE_CHAIN (attrs);
4301 tree fmtarg = TREE_VALUE (attrs);
4303 attrs = TREE_CHAIN (attrs);
4304 tree elliparg = TREE_VALUE (attrs);
4306 /* Attribute argument indices are 1-based but we use zero-based. */
4307 *idx_args = tree_to_uhwi (elliparg) - 1;
4308 return tree_to_uhwi (fmtarg) - 1;
4311 } /* Unnamed namespace. */
4313 /* Determine if a GIMPLE call at *GSI is to one of the sprintf-like built-in
4314 functions and if so, handle it. Return true if the call is removed and
4315 gsi_next should not be performed in the caller. */
4317 bool
4318 handle_printf_call (gimple_stmt_iterator *gsi, pointer_query &ptr_qry)
4320 init_target_to_host_charmap ();
4322 call_info info = call_info ();
4324 info.callstmt = gsi_stmt (*gsi);
4325 info.func = gimple_call_fndecl (info.callstmt);
4326 if (!info.func)
4327 return false;
4329 /* Format string argument number (valid for all functions). */
4330 unsigned idx_format = UINT_MAX;
4331 if (gimple_call_builtin_p (info.callstmt, BUILT_IN_NORMAL))
4332 info.fncode = DECL_FUNCTION_CODE (info.func);
4333 else
4335 unsigned idx_args;
4336 idx_format = get_user_idx_format (info.func, &idx_args);
4337 if (idx_format == UINT_MAX
4338 || idx_format >= gimple_call_num_args (info.callstmt)
4339 || idx_args > gimple_call_num_args (info.callstmt)
4340 || !POINTER_TYPE_P (TREE_TYPE (gimple_call_arg (info.callstmt,
4341 idx_format))))
4342 return false;
4343 info.fncode = BUILT_IN_NONE;
4344 info.argidx = idx_args;
4347 /* The size of the destination as in snprintf(dest, size, ...). */
4348 unsigned HOST_WIDE_INT dstsize = HOST_WIDE_INT_M1U;
4350 /* The size of the destination determined by __builtin_object_size. */
4351 unsigned HOST_WIDE_INT objsize = HOST_WIDE_INT_M1U;
4353 /* Zero-based buffer size argument number (snprintf and vsnprintf). */
4354 unsigned idx_dstsize = UINT_MAX;
4356 /* Object size argument number (snprintf_chk and vsnprintf_chk). */
4357 unsigned idx_objsize = UINT_MAX;
4359 /* Destinaton argument number (valid for sprintf functions only). */
4360 unsigned idx_dstptr = 0;
4362 switch (info.fncode)
4364 case BUILT_IN_NONE:
4365 // User-defined function with attribute format (printf).
4366 idx_dstptr = -1;
4367 break;
4369 case BUILT_IN_FPRINTF:
4370 // Signature:
4371 // __builtin_fprintf (FILE*, format, ...)
4372 idx_format = 1;
4373 info.argidx = 2;
4374 idx_dstptr = -1;
4375 break;
4377 case BUILT_IN_FPRINTF_CHK:
4378 // Signature:
4379 // __builtin_fprintf_chk (FILE*, ost, format, ...)
4380 idx_format = 2;
4381 info.argidx = 3;
4382 idx_dstptr = -1;
4383 break;
4385 case BUILT_IN_FPRINTF_UNLOCKED:
4386 // Signature:
4387 // __builtin_fprintf_unnlocked (FILE*, format, ...)
4388 idx_format = 1;
4389 info.argidx = 2;
4390 idx_dstptr = -1;
4391 break;
4393 case BUILT_IN_PRINTF:
4394 // Signature:
4395 // __builtin_printf (format, ...)
4396 idx_format = 0;
4397 info.argidx = 1;
4398 idx_dstptr = -1;
4399 break;
4401 case BUILT_IN_PRINTF_CHK:
4402 // Signature:
4403 // __builtin_printf_chk (ost, format, ...)
4404 idx_format = 1;
4405 info.argidx = 2;
4406 idx_dstptr = -1;
4407 break;
4409 case BUILT_IN_PRINTF_UNLOCKED:
4410 // Signature:
4411 // __builtin_printf (format, ...)
4412 idx_format = 0;
4413 info.argidx = 1;
4414 idx_dstptr = -1;
4415 break;
4417 case BUILT_IN_SPRINTF:
4418 // Signature:
4419 // __builtin_sprintf (dst, format, ...)
4420 idx_format = 1;
4421 info.argidx = 2;
4422 break;
4424 case BUILT_IN_SPRINTF_CHK:
4425 // Signature:
4426 // __builtin___sprintf_chk (dst, ost, objsize, format, ...)
4427 idx_objsize = 2;
4428 idx_format = 3;
4429 info.argidx = 4;
4430 break;
4432 case BUILT_IN_SNPRINTF:
4433 // Signature:
4434 // __builtin_snprintf (dst, size, format, ...)
4435 idx_dstsize = 1;
4436 idx_format = 2;
4437 info.argidx = 3;
4438 info.bounded = true;
4439 break;
4441 case BUILT_IN_SNPRINTF_CHK:
4442 // Signature:
4443 // __builtin___snprintf_chk (dst, size, ost, objsize, format, ...)
4444 idx_dstsize = 1;
4445 idx_objsize = 3;
4446 idx_format = 4;
4447 info.argidx = 5;
4448 info.bounded = true;
4449 break;
4451 case BUILT_IN_VFPRINTF:
4452 // Signature:
4453 // __builtin_vprintf (FILE*, format, va_list)
4454 idx_format = 1;
4455 info.argidx = -1;
4456 idx_dstptr = -1;
4457 break;
4459 case BUILT_IN_VFPRINTF_CHK:
4460 // Signature:
4461 // __builtin___vfprintf_chk (FILE*, ost, format, va_list)
4462 idx_format = 2;
4463 info.argidx = -1;
4464 idx_dstptr = -1;
4465 break;
4467 case BUILT_IN_VPRINTF:
4468 // Signature:
4469 // __builtin_vprintf (format, va_list)
4470 idx_format = 0;
4471 info.argidx = -1;
4472 idx_dstptr = -1;
4473 break;
4475 case BUILT_IN_VPRINTF_CHK:
4476 // Signature:
4477 // __builtin___vprintf_chk (ost, format, va_list)
4478 idx_format = 1;
4479 info.argidx = -1;
4480 idx_dstptr = -1;
4481 break;
4483 case BUILT_IN_VSNPRINTF:
4484 // Signature:
4485 // __builtin_vsprintf (dst, size, format, va)
4486 idx_dstsize = 1;
4487 idx_format = 2;
4488 info.argidx = -1;
4489 info.bounded = true;
4490 break;
4492 case BUILT_IN_VSNPRINTF_CHK:
4493 // Signature:
4494 // __builtin___vsnprintf_chk (dst, size, ost, objsize, format, va)
4495 idx_dstsize = 1;
4496 idx_objsize = 3;
4497 idx_format = 4;
4498 info.argidx = -1;
4499 info.bounded = true;
4500 break;
4502 case BUILT_IN_VSPRINTF:
4503 // Signature:
4504 // __builtin_vsprintf (dst, format, va)
4505 idx_format = 1;
4506 info.argidx = -1;
4507 break;
4509 case BUILT_IN_VSPRINTF_CHK:
4510 // Signature:
4511 // __builtin___vsprintf_chk (dst, ost, objsize, format, va)
4512 idx_format = 3;
4513 idx_objsize = 2;
4514 info.argidx = -1;
4515 break;
4517 default:
4518 return false;
4521 /* Set the global warning level for this function. */
4522 warn_level = info.bounded ? warn_format_trunc : warn_format_overflow;
4524 /* For all string functions the first argument is a pointer to
4525 the destination. */
4526 tree dstptr = (idx_dstptr < gimple_call_num_args (info.callstmt)
4527 ? gimple_call_arg (info.callstmt, 0) : NULL_TREE);
4529 info.format = gimple_call_arg (info.callstmt, idx_format);
4531 /* True when the destination size is constant as opposed to the lower
4532 or upper bound of a range. */
4533 bool dstsize_cst_p = true;
4534 bool posunder4k = true;
4536 if (idx_dstsize == UINT_MAX)
4538 /* For non-bounded functions like sprintf, determine the size
4539 of the destination from the object or pointer passed to it
4540 as the first argument. */
4541 dstsize = get_destination_size (dstptr, info.callstmt, ptr_qry);
4543 else if (tree size = gimple_call_arg (info.callstmt, idx_dstsize))
4545 /* For bounded functions try to get the size argument. */
4547 if (TREE_CODE (size) == INTEGER_CST)
4549 dstsize = tree_to_uhwi (size);
4550 /* No object can be larger than SIZE_MAX bytes (half the address
4551 space) on the target.
4552 The functions are defined only for output of at most INT_MAX
4553 bytes. Specifying a bound in excess of that limit effectively
4554 defeats the bounds checking (and on some implementations such
4555 as Solaris cause the function to fail with EINVAL). */
4556 if (dstsize > target_size_max () / 2)
4558 /* Avoid warning if -Wstringop-overflow is specified since
4559 it also warns for the same thing though only for the
4560 checking built-ins. */
4561 if ((idx_objsize == UINT_MAX
4562 || !warn_stringop_overflow))
4563 warning_at (gimple_location (info.callstmt), info.warnopt (),
4564 "specified bound %wu exceeds maximum object size "
4565 "%wu",
4566 dstsize, target_size_max () / 2);
4567 /* POSIX requires snprintf to fail if DSTSIZE is greater
4568 than INT_MAX. Even though not all POSIX implementations
4569 conform to the requirement, avoid folding in this case. */
4570 posunder4k = false;
4572 else if (dstsize > target_int_max ())
4574 warning_at (gimple_location (info.callstmt), info.warnopt (),
4575 "specified bound %wu exceeds %<INT_MAX%>",
4576 dstsize);
4577 /* POSIX requires snprintf to fail if DSTSIZE is greater
4578 than INT_MAX. Avoid folding in that case. */
4579 posunder4k = false;
4582 else if (TREE_CODE (size) == SSA_NAME)
4584 /* Try to determine the range of values of the argument
4585 and use the greater of the two at level 1 and the smaller
4586 of them at level 2. */
4587 value_range vr;
4588 ptr_qry.rvals->range_of_expr (vr, size, info.callstmt);
4590 if (!vr.undefined_p ())
4592 tree type = TREE_TYPE (size);
4593 tree tmin = wide_int_to_tree (type, vr.lower_bound ());
4594 tree tmax = wide_int_to_tree (type, vr.upper_bound ());
4595 unsigned HOST_WIDE_INT minsize = TREE_INT_CST_LOW (tmin);
4596 unsigned HOST_WIDE_INT maxsize = TREE_INT_CST_LOW (tmax);
4597 dstsize = warn_level < 2 ? maxsize : minsize;
4599 if (minsize > target_int_max ())
4600 warning_at (gimple_location (info.callstmt), info.warnopt (),
4601 "specified bound range [%wu, %wu] exceeds "
4602 "%<INT_MAX%>",
4603 minsize, maxsize);
4605 /* POSIX requires snprintf to fail if DSTSIZE is greater
4606 than INT_MAX. Avoid folding if that's possible. */
4607 if (maxsize > target_int_max ())
4608 posunder4k = false;
4611 /* The destination size is not constant. If the function is
4612 bounded (e.g., snprintf) a lower bound of zero doesn't
4613 necessarily imply it can be eliminated. */
4614 dstsize_cst_p = false;
4618 if (idx_objsize != UINT_MAX)
4619 if (tree size = gimple_call_arg (info.callstmt, idx_objsize))
4620 if (tree_fits_uhwi_p (size))
4621 objsize = tree_to_uhwi (size);
4623 if (info.bounded && !dstsize)
4625 /* As a special case, when the explicitly specified destination
4626 size argument (to a bounded function like snprintf) is zero
4627 it is a request to determine the number of bytes on output
4628 without actually producing any. Pretend the size is
4629 unlimited in this case. */
4630 info.objsize = HOST_WIDE_INT_MAX;
4631 info.nowrite = dstsize_cst_p;
4633 else
4635 /* For calls to non-bounded functions or to those of bounded
4636 functions with a non-zero size, warn if the destination
4637 pointer is null. */
4638 if (dstptr && integer_zerop (dstptr))
4640 /* This is diagnosed with -Wformat only when the null is a constant
4641 pointer. The warning here diagnoses instances where the pointer
4642 is not constant. */
4643 location_t loc = gimple_location (info.callstmt);
4644 warning_at (EXPR_LOC_OR_LOC (dstptr, loc),
4645 info.warnopt (), "null destination pointer");
4646 return false;
4649 /* Set the object size to the smaller of the two arguments
4650 of both have been specified and they're not equal. */
4651 info.objsize = dstsize < objsize ? dstsize : objsize;
4653 if (info.bounded
4654 && dstsize < target_size_max () / 2 && objsize < dstsize
4655 /* Avoid warning if -Wstringop-overflow is specified since
4656 it also warns for the same thing though only for the
4657 checking built-ins. */
4658 && (idx_objsize == UINT_MAX
4659 || !warn_stringop_overflow))
4661 warning_at (gimple_location (info.callstmt), info.warnopt (),
4662 "specified bound %wu exceeds the size %wu "
4663 "of the destination object", dstsize, objsize);
4667 /* Determine if the format argument may be null and warn if not
4668 and if the argument is null. */
4669 if (integer_zerop (info.format)
4670 && gimple_call_builtin_p (info.callstmt, BUILT_IN_NORMAL))
4672 location_t loc = gimple_location (info.callstmt);
4673 warning_at (EXPR_LOC_OR_LOC (info.format, loc),
4674 info.warnopt (), "null format string");
4675 return false;
4678 info.fmtstr = get_format_string (info.format, &info.fmtloc);
4679 if (!info.fmtstr)
4680 return false;
4682 if (warn_restrict)
4684 /* Compute the origin of the destination pointer and its offset
4685 from the base object/pointer if possible. */
4686 info.dst_offset = 0;
4687 info.dst_origin = get_origin_and_offset (dstptr, &info.dst_field,
4688 &info.dst_offset);
4691 /* The result is the number of bytes output by the formatted function,
4692 including the terminating NUL. */
4693 format_result res;
4695 /* I/O functions with no destination argument (i.e., all forms of fprintf
4696 and printf) may fail under any conditions. Others (i.e., all forms of
4697 sprintf) may only fail under specific conditions determined for each
4698 directive. Clear POSUNDER4K for the former set of functions and set
4699 it to true for the latter (it can only be cleared later, but it is
4700 never set to true again). */
4701 res.posunder4k = posunder4k && dstptr;
4703 bool success = compute_format_length (info, &res, ptr_qry.rvals);
4704 if (res.warned)
4705 suppress_warning (info.callstmt, info.warnopt ());
4707 /* When optimizing and the printf return value optimization is enabled,
4708 attempt to substitute the computed result for the return value of
4709 the call. Avoid this optimization when -frounding-math is in effect
4710 and the format string contains a floating point directive. */
4711 bool call_removed = false;
4712 if (success && optimize > 0)
4714 /* Save a copy of the iterator pointing at the call. The iterator
4715 may change to point past the call in try_substitute_return_value
4716 but the original value is needed in try_simplify_call. */
4717 gimple_stmt_iterator gsi_call = *gsi;
4719 if (flag_printf_return_value
4720 && (!flag_rounding_math || !res.floating))
4721 call_removed = try_substitute_return_value (gsi, info, res);
4723 if (!call_removed)
4724 try_simplify_call (&gsi_call, info, res);
4727 return call_removed;