1 // SPDX-License-Identifier: GPL-2.0-only
3 * Helpers for formatting and printing strings
5 * Copyright 31 August 2008 James Bottomley
6 * Copyright (C) 2013, Intel Corporation
9 #include <linux/kernel.h>
10 #include <linux/math64.h>
11 #include <linux/export.h>
12 #include <linux/ctype.h>
13 #include <linux/device.h>
14 #include <linux/errno.h>
16 #include <linux/limits.h>
18 #include <linux/slab.h>
19 #include <linux/string.h>
20 #include <linux/string_helpers.h>
21 #include <kunit/test.h>
22 #include <kunit/test-bug.h>
25 * string_get_size - get the size in the specified units
26 * @size: The size to be converted in blocks
27 * @blk_size: Size of the block (use 1 for size in bytes)
28 * @units: Units to use (powers of 1000 or 1024), whether to include space separator
29 * @buf: buffer to format to
30 * @len: length of buffer
32 * This function returns a string formatted to 3 significant figures
33 * giving the size in the required units. @buf should have room for
34 * at least 9 bytes and will always be zero terminated.
36 * Return value: number of characters of output that would have been written
37 * (which may be greater than len, if output was truncated).
39 int string_get_size(u64 size, u64 blk_size, const enum string_size_units units,
42 enum string_size_units units_base = units & STRING_UNITS_MASK;
43 static const char *const units_10[] = {
44 "", "k", "M", "G", "T", "P", "E", "Z", "Y",
46 static const char *const units_2[] = {
47 "", "Ki", "Mi", "Gi", "Ti", "Pi", "Ei", "Zi", "Yi",
49 static const char *const *const units_str[] = {
50 [STRING_UNITS_10] = units_10,
51 [STRING_UNITS_2] = units_2,
53 static const unsigned int divisor[] = {
54 [STRING_UNITS_10] = 1000,
55 [STRING_UNITS_2] = 1024,
57 static const unsigned int rounding[] = { 500, 50, 5 };
59 u32 remainder = 0, sf_cap;
70 /* This is Napier's algorithm. Reduce the original block size to
72 * coefficient * divisor[units_base]^i
74 * we do the reduction so both coefficients are just under 32 bits so
75 * that multiplying them together won't overflow 64 bits and we keep
76 * as much precision as possible in the numbers.
78 * Note: it's safe to throw away the remainders here because all the
79 * precision is in the coefficients.
81 while (blk_size >> 32) {
82 do_div(blk_size, divisor[units_base]);
87 do_div(size, divisor[units_base]);
91 /* now perform the actual multiplication keeping i as the sum of the
95 /* and logarithmically reduce it until it's just under the divisor */
96 while (size >= divisor[units_base]) {
97 remainder = do_div(size, divisor[units_base]);
101 /* work out in j how many digits of precision we need from the
104 for (j = 0; sf_cap*10 < 1000; j++)
107 if (units_base == STRING_UNITS_2) {
108 /* express the remainder as a decimal. It's currently the
109 * numerator of a fraction whose denominator is
110 * divisor[units_base], which is 1 << 10 for STRING_UNITS_2 */
115 /* add a 5 to the digit below what will be printed to ensure
116 * an arithmetical round up and carry it through to size */
117 remainder += rounding[j];
118 if (remainder >= 1000) {
124 snprintf(tmp, sizeof(tmp), ".%03u", remainder);
129 if (i >= ARRAY_SIZE(units_2))
132 unit = units_str[units_base][i];
134 return snprintf(buf, len, "%u%s%s%s%s", (u32)size, tmp,
135 (units & STRING_UNITS_NO_SPACE) ? "" : " ",
137 (units & STRING_UNITS_NO_BYTES) ? "" : "B");
139 EXPORT_SYMBOL(string_get_size);
142 * parse_int_array_user - Split string into a sequence of integers
143 * @from: The user space buffer to read from
144 * @count: The maximum number of bytes to read
145 * @array: Returned pointer to sequence of integers
147 * On success @array is allocated and initialized with a sequence of
148 * integers extracted from the @from plus an additional element that
149 * begins the sequence and specifies the integers count.
151 * Caller takes responsibility for freeing @array when it is no longer
154 int parse_int_array_user(const char __user *from, size_t count, int **array)
160 buf = memdup_user_nul(from, count);
164 get_options(buf, 0, &nints);
170 ints = kcalloc(nints + 1, sizeof(*ints), GFP_KERNEL);
176 get_options(buf, nints + 1, ints);
183 EXPORT_SYMBOL(parse_int_array_user);
185 static bool unescape_space(char **src, char **dst)
187 char *p = *dst, *q = *src;
213 static bool unescape_octal(char **src, char **dst)
215 char *p = *dst, *q = *src;
218 if (isodigit(*q) == 0)
222 while (num < 32 && isodigit(*q) && (q - *src < 3)) {
232 static bool unescape_hex(char **src, char **dst)
234 char *p = *dst, *q = *src;
241 num = digit = hex_to_bin(*q++);
245 digit = hex_to_bin(*q);
248 num = (num << 4) | digit;
256 static bool unescape_special(char **src, char **dst)
258 char *p = *dst, *q = *src;
282 * string_unescape - unquote characters in the given string
283 * @src: source buffer (escaped)
284 * @dst: destination buffer (unescaped)
285 * @size: size of the destination buffer (0 to unlimit)
286 * @flags: combination of the flags.
289 * The function unquotes characters in the given string.
291 * Because the size of the output will be the same as or less than the size of
292 * the input, the transformation may be performed in place.
294 * Caller must provide valid source and destination pointers. Be aware that
295 * destination buffer will always be NULL-terminated. Source string must be
296 * NULL-terminated as well. The supported flags are::
301 * '\r' - carriage return
302 * '\t' - horizontal tab
303 * '\v' - vertical tab
305 * '\NNN' - byte with octal value NNN (1 to 3 digits)
307 * '\xHH' - byte with hexadecimal value HH (1 to 2 digits)
309 * '\"' - double quote
314 * all previous together
317 * The amount of the characters processed to the destination buffer excluding
318 * trailing '\0' is returned.
320 int string_unescape(char *src, char *dst, size_t size, unsigned int flags)
324 while (*src && --size) {
325 if (src[0] == '\\' && src[1] != '\0' && size > 1) {
329 if (flags & UNESCAPE_SPACE &&
330 unescape_space(&src, &out))
333 if (flags & UNESCAPE_OCTAL &&
334 unescape_octal(&src, &out))
337 if (flags & UNESCAPE_HEX &&
338 unescape_hex(&src, &out))
341 if (flags & UNESCAPE_SPECIAL &&
342 unescape_special(&src, &out))
353 EXPORT_SYMBOL(string_unescape);
355 static bool escape_passthrough(unsigned char c, char **dst, char *end)
365 static bool escape_space(unsigned char c, char **dst, char *end)
401 static bool escape_special(unsigned char c, char **dst, char *end)
434 static bool escape_null(unsigned char c, char **dst, char *end)
452 static bool escape_octal(unsigned char c, char **dst, char *end)
460 *out = ((c >> 6) & 0x07) + '0';
463 *out = ((c >> 3) & 0x07) + '0';
466 *out = ((c >> 0) & 0x07) + '0';
473 static bool escape_hex(unsigned char c, char **dst, char *end)
484 *out = hex_asc_hi(c);
487 *out = hex_asc_lo(c);
495 * string_escape_mem - quote characters in the given memory buffer
496 * @src: source buffer (unescaped)
497 * @isz: source buffer size
498 * @dst: destination buffer (escaped)
499 * @osz: destination buffer size
500 * @flags: combination of the flags
501 * @only: NULL-terminated string containing characters used to limit
502 * the selected escape class. If characters are included in @only
503 * that would not normally be escaped by the classes selected
504 * in @flags, they will be copied to @dst unescaped.
507 * The process of escaping byte buffer includes several parts. They are applied
508 * in the following sequence.
510 * 1. The character is not matched to the one from @only string and thus
511 * must go as-is to the output.
512 * 2. The character is matched to the printable and ASCII classes, if asked,
513 * and in case of match it passes through to the output.
514 * 3. The character is matched to the printable or ASCII class, if asked,
515 * and in case of match it passes through to the output.
516 * 4. The character is checked if it falls into the class given by @flags.
517 * %ESCAPE_OCTAL and %ESCAPE_HEX are going last since they cover any
518 * character. Note that they actually can't go together, otherwise
519 * %ESCAPE_HEX will be ignored.
521 * Caller must provide valid source and destination pointers. Be aware that
522 * destination buffer will not be NULL-terminated, thus caller have to append
523 * it if needs. The supported flags are::
525 * %ESCAPE_SPACE: (special white space, not space itself)
528 * '\r' - carriage return
529 * '\t' - horizontal tab
530 * '\v' - vertical tab
532 * '\"' - double quote
539 * '\NNN' - byte with octal value NNN (3 digits)
541 * all previous together
543 * escape only non-printable characters, checked by isprint()
545 * all previous together
547 * '\xHH' - byte with hexadecimal value HH (2 digits)
549 * escape only non-ascii characters, checked by isascii()
551 * escape only non-printable or non-ascii characters
553 * append characters from @only to be escaped by the given classes
555 * %ESCAPE_APPEND would help to pass additional characters to the escaped, when
556 * one of %ESCAPE_NP, %ESCAPE_NA, or %ESCAPE_NAP is provided.
558 * One notable caveat, the %ESCAPE_NAP, %ESCAPE_NP and %ESCAPE_NA have the
559 * higher priority than the rest of the flags (%ESCAPE_NAP is the highest).
560 * It doesn't make much sense to use either of them without %ESCAPE_OCTAL
561 * or %ESCAPE_HEX, because they cover most of the other character classes.
562 * %ESCAPE_NAP can utilize %ESCAPE_SPACE or %ESCAPE_SPECIAL in addition to
566 * The total size of the escaped output that would be generated for
567 * the given input and flags. To check whether the output was
568 * truncated, compare the return value to osz. There is room left in
569 * dst for a '\0' terminator if and only if ret < osz.
571 int string_escape_mem(const char *src, size_t isz, char *dst, size_t osz,
572 unsigned int flags, const char *only)
576 bool is_dict = only && *only;
577 bool is_append = flags & ESCAPE_APPEND;
580 unsigned char c = *src++;
581 bool in_dict = is_dict && strchr(only, c);
584 * Apply rules in the following sequence:
585 * - the @only string is supplied and does not contain a
586 * character under question
587 * - the character is printable and ASCII, when @flags has
588 * %ESCAPE_NAP bit set
589 * - the character is printable, when @flags has
591 * - the character is ASCII, when @flags has
593 * - the character doesn't fall into a class of symbols
594 * defined by given @flags
595 * In these cases we just pass through a character to the
598 * When %ESCAPE_APPEND is passed, the characters from @only
599 * have been excluded from the %ESCAPE_NAP, %ESCAPE_NP, and
602 if (!(is_append || in_dict) && is_dict &&
603 escape_passthrough(c, &p, end))
606 if (!(is_append && in_dict) && isascii(c) && isprint(c) &&
607 flags & ESCAPE_NAP && escape_passthrough(c, &p, end))
610 if (!(is_append && in_dict) && isprint(c) &&
611 flags & ESCAPE_NP && escape_passthrough(c, &p, end))
614 if (!(is_append && in_dict) && isascii(c) &&
615 flags & ESCAPE_NA && escape_passthrough(c, &p, end))
618 if (flags & ESCAPE_SPACE && escape_space(c, &p, end))
621 if (flags & ESCAPE_SPECIAL && escape_special(c, &p, end))
624 if (flags & ESCAPE_NULL && escape_null(c, &p, end))
627 /* ESCAPE_OCTAL and ESCAPE_HEX always go last */
628 if (flags & ESCAPE_OCTAL && escape_octal(c, &p, end))
631 if (flags & ESCAPE_HEX && escape_hex(c, &p, end))
634 escape_passthrough(c, &p, end);
639 EXPORT_SYMBOL(string_escape_mem);
642 * Return an allocated string that has been escaped of special characters
643 * and double quotes, making it safe to log in quotes.
645 char *kstrdup_quotable(const char *src, gfp_t gfp)
649 const int flags = ESCAPE_HEX;
650 const char esc[] = "\f\n\r\t\v\a\e\\\"";
656 dlen = string_escape_mem(src, slen, NULL, 0, flags, esc);
657 dst = kmalloc(dlen + 1, gfp);
661 WARN_ON(string_escape_mem(src, slen, dst, dlen, flags, esc) != dlen);
666 EXPORT_SYMBOL_GPL(kstrdup_quotable);
669 * Returns allocated NULL-terminated string containing process
670 * command line, with inter-argument NULLs replaced with spaces,
671 * and other special characters escaped.
673 char *kstrdup_quotable_cmdline(struct task_struct *task, gfp_t gfp)
675 char *buffer, *quoted;
678 buffer = kmalloc(PAGE_SIZE, GFP_KERNEL);
682 res = get_cmdline(task, buffer, PAGE_SIZE - 1);
685 /* Collapse trailing NULLs, leave res pointing to last non-NULL. */
686 while (--res >= 0 && buffer[res] == '\0')
689 /* Replace inter-argument NULLs. */
690 for (i = 0; i <= res; i++)
691 if (buffer[i] == '\0')
694 /* Make sure result is printable. */
695 quoted = kstrdup_quotable(buffer, gfp);
699 EXPORT_SYMBOL_GPL(kstrdup_quotable_cmdline);
702 * Returns allocated NULL-terminated string containing pathname,
703 * with special characters escaped, able to be safely logged. If
704 * there is an error, the leading character will be "<".
706 char *kstrdup_quotable_file(struct file *file, gfp_t gfp)
708 char *temp, *pathname;
711 return kstrdup("<unknown>", gfp);
713 /* We add 11 spaces for ' (deleted)' to be appended */
714 temp = kmalloc(PATH_MAX + 11, GFP_KERNEL);
716 return kstrdup("<no_memory>", gfp);
718 pathname = file_path(file, temp, PATH_MAX + 11);
719 if (IS_ERR(pathname))
720 pathname = kstrdup("<too_long>", gfp);
722 pathname = kstrdup_quotable(pathname, gfp);
727 EXPORT_SYMBOL_GPL(kstrdup_quotable_file);
730 * Returns duplicate string in which the @old characters are replaced by @new.
732 char *kstrdup_and_replace(const char *src, char old, char new, gfp_t gfp)
736 dst = kstrdup(src, gfp);
740 return strreplace(dst, old, new);
742 EXPORT_SYMBOL_GPL(kstrdup_and_replace);
745 * kasprintf_strarray - allocate and fill array of sequential strings
746 * @gfp: flags for the slab allocator
747 * @prefix: prefix to be used
748 * @n: amount of lines to be allocated and filled
750 * Allocates and fills @n strings using pattern "%s-%zu", where prefix
751 * is provided by caller. The caller is responsible to free them with
752 * kfree_strarray() after use.
754 * Returns array of strings or NULL when memory can't be allocated.
756 char **kasprintf_strarray(gfp_t gfp, const char *prefix, size_t n)
761 names = kcalloc(n + 1, sizeof(char *), gfp);
765 for (i = 0; i < n; i++) {
766 names[i] = kasprintf(gfp, "%s-%zu", prefix, i);
768 kfree_strarray(names, i);
775 EXPORT_SYMBOL_GPL(kasprintf_strarray);
778 * kfree_strarray - free a number of dynamically allocated strings contained
779 * in an array and the array itself
781 * @array: Dynamically allocated array of strings to free.
782 * @n: Number of strings (starting from the beginning of the array) to free.
784 * Passing a non-NULL @array and @n == 0 as well as NULL @array are valid
785 * use-cases. If @array is NULL, the function does nothing.
787 void kfree_strarray(char **array, size_t n)
794 for (i = 0; i < n; i++)
798 EXPORT_SYMBOL_GPL(kfree_strarray);
805 static void devm_kfree_strarray(struct device *dev, void *res)
807 struct strarray *array = res;
809 kfree_strarray(array->array, array->n);
812 char **devm_kasprintf_strarray(struct device *dev, const char *prefix, size_t n)
814 struct strarray *ptr;
816 ptr = devres_alloc(devm_kfree_strarray, sizeof(*ptr), GFP_KERNEL);
818 return ERR_PTR(-ENOMEM);
820 ptr->array = kasprintf_strarray(GFP_KERNEL, prefix, n);
823 return ERR_PTR(-ENOMEM);
827 devres_add(dev, ptr);
831 EXPORT_SYMBOL_GPL(devm_kasprintf_strarray);
834 * skip_spaces - Removes leading whitespace from @str.
835 * @str: The string to be stripped.
837 * Returns a pointer to the first non-whitespace character in @str.
839 char *skip_spaces(const char *str)
841 while (isspace(*str))
845 EXPORT_SYMBOL(skip_spaces);
848 * strim - Removes leading and trailing whitespace from @s.
849 * @s: The string to be stripped.
851 * Note that the first trailing whitespace is replaced with a %NUL-terminator
852 * in the given string @s. Returns a pointer to the first non-whitespace
865 while (end >= s && isspace(*end))
869 return skip_spaces(s);
871 EXPORT_SYMBOL(strim);
874 * sysfs_streq - return true if strings are equal, modulo trailing newline
876 * @s2: another string
878 * This routine returns true iff two strings are equal, treating both
879 * NUL and newline-then-NUL as equivalent string terminations. It's
880 * geared for use with sysfs input strings, which generally terminate
881 * with newlines but are compared against values without newlines.
883 bool sysfs_streq(const char *s1, const char *s2)
885 while (*s1 && *s1 == *s2) {
892 if (!*s1 && *s2 == '\n' && !s2[1])
894 if (*s1 == '\n' && !s1[1] && !*s2)
898 EXPORT_SYMBOL(sysfs_streq);
901 * match_string - matches given string in an array
902 * @array: array of strings
903 * @n: number of strings in the array or -1 for NULL terminated arrays
904 * @string: string to match with
906 * This routine will look for a string in an array of strings up to the
907 * n-th element in the array or until the first NULL element.
909 * Historically the value of -1 for @n, was used to search in arrays that
910 * are NULL terminated. However, the function does not make a distinction
911 * when finishing the search: either @n elements have been compared OR
912 * the first NULL element was found.
915 * index of a @string in the @array if matches, or %-EINVAL otherwise.
917 int match_string(const char * const *array, size_t n, const char *string)
922 for (index = 0; index < n; index++) {
926 if (!strcmp(item, string))
932 EXPORT_SYMBOL(match_string);
935 * __sysfs_match_string - matches given string in an array
936 * @array: array of strings
937 * @n: number of strings in the array or -1 for NULL terminated arrays
938 * @str: string to match with
940 * Returns index of @str in the @array or -EINVAL, just like match_string().
941 * Uses sysfs_streq instead of strcmp for matching.
943 * This routine will look for a string in an array of strings up to the
944 * n-th element in the array or until the first NULL element.
946 * Historically the value of -1 for @n, was used to search in arrays that
947 * are NULL terminated. However, the function does not make a distinction
948 * when finishing the search: either @n elements have been compared OR
949 * the first NULL element was found.
951 int __sysfs_match_string(const char * const *array, size_t n, const char *str)
956 for (index = 0; index < n; index++) {
960 if (sysfs_streq(item, str))
966 EXPORT_SYMBOL(__sysfs_match_string);
969 * strreplace - Replace all occurrences of character in string.
970 * @str: The string to operate on.
971 * @old: The character being replaced.
972 * @new: The character @old is replaced with.
974 * Replaces the each @old character with a @new one in the given string @str.
976 * Return: pointer to the string @str itself.
978 char *strreplace(char *str, char old, char new)
987 EXPORT_SYMBOL(strreplace);
990 * memcpy_and_pad - Copy one buffer to another with padding
991 * @dest: Where to copy to
992 * @dest_len: The destination buffer size
993 * @src: Where to copy from
994 * @count: The number of bytes to copy
995 * @pad: Character to use for padding if space is left in destination.
997 void memcpy_and_pad(void *dest, size_t dest_len, const void *src, size_t count,
1000 if (dest_len > count) {
1001 memcpy(dest, src, count);
1002 memset(dest + count, pad, dest_len - count);
1004 memcpy(dest, src, dest_len);
1007 EXPORT_SYMBOL(memcpy_and_pad);
1009 #ifdef CONFIG_FORTIFY_SOURCE
1010 /* These are placeholders for fortify compile-time warnings. */
1011 void __read_overflow2_field(size_t avail, size_t wanted) { }
1012 EXPORT_SYMBOL(__read_overflow2_field);
1013 void __write_overflow_field(size_t avail, size_t wanted) { }
1014 EXPORT_SYMBOL(__write_overflow_field);
1016 static const char * const fortify_func_name[] = {
1017 #define MAKE_FORTIFY_FUNC_NAME(func) [MAKE_FORTIFY_FUNC(func)] = #func
1018 EACH_FORTIFY_FUNC(MAKE_FORTIFY_FUNC_NAME)
1019 #undef MAKE_FORTIFY_FUNC_NAME
1022 void __fortify_report(const u8 reason, const size_t avail, const size_t size)
1024 const u8 func = FORTIFY_REASON_FUNC(reason);
1025 const bool write = FORTIFY_REASON_DIR(reason);
1028 name = fortify_func_name[umin(func, FORTIFY_FUNC_UNKNOWN)];
1029 WARN(1, "%s: detected buffer overflow: %zu byte %s of buffer size %zu\n",
1030 name, size, str_read_write(!write), avail);
1032 EXPORT_SYMBOL(__fortify_report);
1034 void __fortify_panic(const u8 reason, const size_t avail, const size_t size)
1036 __fortify_report(reason, avail, size);
1039 EXPORT_SYMBOL(__fortify_panic);
1040 #endif /* CONFIG_FORTIFY_SOURCE */