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>
23 * string_get_size - get the size in the specified units
24 * @size: The size to be converted in blocks
25 * @blk_size: Size of the block (use 1 for size in bytes)
26 * @units: units to use (powers of 1000 or 1024)
27 * @buf: buffer to format to
28 * @len: length of buffer
30 * This function returns a string formatted to 3 significant figures
31 * giving the size in the required units. @buf should have room for
32 * at least 9 bytes and will always be zero terminated.
34 * Return value: number of characters of output that would have been written
35 * (which may be greater than len, if output was truncated).
37 int string_get_size(u64 size, u64 blk_size, const enum string_size_units units,
40 static const char *const units_10[] = {
41 "B", "kB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB"
43 static const char *const units_2[] = {
44 "B", "KiB", "MiB", "GiB", "TiB", "PiB", "EiB", "ZiB", "YiB"
46 static const char *const *const units_str[] = {
47 [STRING_UNITS_10] = units_10,
48 [STRING_UNITS_2] = units_2,
50 static const unsigned int divisor[] = {
51 [STRING_UNITS_10] = 1000,
52 [STRING_UNITS_2] = 1024,
54 static const unsigned int rounding[] = { 500, 50, 5 };
56 u32 remainder = 0, sf_cap;
67 /* This is Napier's algorithm. Reduce the original block size to
69 * coefficient * divisor[units]^i
71 * we do the reduction so both coefficients are just under 32 bits so
72 * that multiplying them together won't overflow 64 bits and we keep
73 * as much precision as possible in the numbers.
75 * Note: it's safe to throw away the remainders here because all the
76 * precision is in the coefficients.
78 while (blk_size >> 32) {
79 do_div(blk_size, divisor[units]);
84 do_div(size, divisor[units]);
88 /* now perform the actual multiplication keeping i as the sum of the
92 /* and logarithmically reduce it until it's just under the divisor */
93 while (size >= divisor[units]) {
94 remainder = do_div(size, divisor[units]);
98 /* work out in j how many digits of precision we need from the
101 for (j = 0; sf_cap*10 < 1000; j++)
104 if (units == STRING_UNITS_2) {
105 /* express the remainder as a decimal. It's currently the
106 * numerator of a fraction whose denominator is
107 * divisor[units], which is 1 << 10 for STRING_UNITS_2 */
112 /* add a 5 to the digit below what will be printed to ensure
113 * an arithmetical round up and carry it through to size */
114 remainder += rounding[j];
115 if (remainder >= 1000) {
121 snprintf(tmp, sizeof(tmp), ".%03u", remainder);
126 if (i >= ARRAY_SIZE(units_2))
129 unit = units_str[units][i];
131 return snprintf(buf, len, "%u%s %s", (u32)size,
134 EXPORT_SYMBOL(string_get_size);
137 * parse_int_array_user - Split string into a sequence of integers
138 * @from: The user space buffer to read from
139 * @count: The maximum number of bytes to read
140 * @array: Returned pointer to sequence of integers
142 * On success @array is allocated and initialized with a sequence of
143 * integers extracted from the @from plus an additional element that
144 * begins the sequence and specifies the integers count.
146 * Caller takes responsibility for freeing @array when it is no longer
149 int parse_int_array_user(const char __user *from, size_t count, int **array)
155 buf = memdup_user_nul(from, count);
159 get_options(buf, 0, &nints);
165 ints = kcalloc(nints + 1, sizeof(*ints), GFP_KERNEL);
171 get_options(buf, nints + 1, ints);
178 EXPORT_SYMBOL(parse_int_array_user);
180 static bool unescape_space(char **src, char **dst)
182 char *p = *dst, *q = *src;
208 static bool unescape_octal(char **src, char **dst)
210 char *p = *dst, *q = *src;
213 if (isodigit(*q) == 0)
217 while (num < 32 && isodigit(*q) && (q - *src < 3)) {
227 static bool unescape_hex(char **src, char **dst)
229 char *p = *dst, *q = *src;
236 num = digit = hex_to_bin(*q++);
240 digit = hex_to_bin(*q);
243 num = (num << 4) | digit;
251 static bool unescape_special(char **src, char **dst)
253 char *p = *dst, *q = *src;
277 * string_unescape - unquote characters in the given string
278 * @src: source buffer (escaped)
279 * @dst: destination buffer (unescaped)
280 * @size: size of the destination buffer (0 to unlimit)
281 * @flags: combination of the flags.
284 * The function unquotes characters in the given string.
286 * Because the size of the output will be the same as or less than the size of
287 * the input, the transformation may be performed in place.
289 * Caller must provide valid source and destination pointers. Be aware that
290 * destination buffer will always be NULL-terminated. Source string must be
291 * NULL-terminated as well. The supported flags are::
296 * '\r' - carriage return
297 * '\t' - horizontal tab
298 * '\v' - vertical tab
300 * '\NNN' - byte with octal value NNN (1 to 3 digits)
302 * '\xHH' - byte with hexadecimal value HH (1 to 2 digits)
304 * '\"' - double quote
309 * all previous together
312 * The amount of the characters processed to the destination buffer excluding
313 * trailing '\0' is returned.
315 int string_unescape(char *src, char *dst, size_t size, unsigned int flags)
319 while (*src && --size) {
320 if (src[0] == '\\' && src[1] != '\0' && size > 1) {
324 if (flags & UNESCAPE_SPACE &&
325 unescape_space(&src, &out))
328 if (flags & UNESCAPE_OCTAL &&
329 unescape_octal(&src, &out))
332 if (flags & UNESCAPE_HEX &&
333 unescape_hex(&src, &out))
336 if (flags & UNESCAPE_SPECIAL &&
337 unescape_special(&src, &out))
348 EXPORT_SYMBOL(string_unescape);
350 static bool escape_passthrough(unsigned char c, char **dst, char *end)
360 static bool escape_space(unsigned char c, char **dst, char *end)
396 static bool escape_special(unsigned char c, char **dst, char *end)
429 static bool escape_null(unsigned char c, char **dst, char *end)
447 static bool escape_octal(unsigned char c, char **dst, char *end)
455 *out = ((c >> 6) & 0x07) + '0';
458 *out = ((c >> 3) & 0x07) + '0';
461 *out = ((c >> 0) & 0x07) + '0';
468 static bool escape_hex(unsigned char c, char **dst, char *end)
479 *out = hex_asc_hi(c);
482 *out = hex_asc_lo(c);
490 * string_escape_mem - quote characters in the given memory buffer
491 * @src: source buffer (unescaped)
492 * @isz: source buffer size
493 * @dst: destination buffer (escaped)
494 * @osz: destination buffer size
495 * @flags: combination of the flags
496 * @only: NULL-terminated string containing characters used to limit
497 * the selected escape class. If characters are included in @only
498 * that would not normally be escaped by the classes selected
499 * in @flags, they will be copied to @dst unescaped.
502 * The process of escaping byte buffer includes several parts. They are applied
503 * in the following sequence.
505 * 1. The character is not matched to the one from @only string and thus
506 * must go as-is to the output.
507 * 2. The character is matched to the printable and ASCII classes, if asked,
508 * and in case of match it passes through to the output.
509 * 3. The character is matched to the printable or ASCII class, if asked,
510 * and in case of match it passes through to the output.
511 * 4. The character is checked if it falls into the class given by @flags.
512 * %ESCAPE_OCTAL and %ESCAPE_HEX are going last since they cover any
513 * character. Note that they actually can't go together, otherwise
514 * %ESCAPE_HEX will be ignored.
516 * Caller must provide valid source and destination pointers. Be aware that
517 * destination buffer will not be NULL-terminated, thus caller have to append
518 * it if needs. The supported flags are::
520 * %ESCAPE_SPACE: (special white space, not space itself)
523 * '\r' - carriage return
524 * '\t' - horizontal tab
525 * '\v' - vertical tab
527 * '\"' - double quote
534 * '\NNN' - byte with octal value NNN (3 digits)
536 * all previous together
538 * escape only non-printable characters, checked by isprint()
540 * all previous together
542 * '\xHH' - byte with hexadecimal value HH (2 digits)
544 * escape only non-ascii characters, checked by isascii()
546 * escape only non-printable or non-ascii characters
548 * append characters from @only to be escaped by the given classes
550 * %ESCAPE_APPEND would help to pass additional characters to the escaped, when
551 * one of %ESCAPE_NP, %ESCAPE_NA, or %ESCAPE_NAP is provided.
553 * One notable caveat, the %ESCAPE_NAP, %ESCAPE_NP and %ESCAPE_NA have the
554 * higher priority than the rest of the flags (%ESCAPE_NAP is the highest).
555 * It doesn't make much sense to use either of them without %ESCAPE_OCTAL
556 * or %ESCAPE_HEX, because they cover most of the other character classes.
557 * %ESCAPE_NAP can utilize %ESCAPE_SPACE or %ESCAPE_SPECIAL in addition to
561 * The total size of the escaped output that would be generated for
562 * the given input and flags. To check whether the output was
563 * truncated, compare the return value to osz. There is room left in
564 * dst for a '\0' terminator if and only if ret < osz.
566 int string_escape_mem(const char *src, size_t isz, char *dst, size_t osz,
567 unsigned int flags, const char *only)
571 bool is_dict = only && *only;
572 bool is_append = flags & ESCAPE_APPEND;
575 unsigned char c = *src++;
576 bool in_dict = is_dict && strchr(only, c);
579 * Apply rules in the following sequence:
580 * - the @only string is supplied and does not contain a
581 * character under question
582 * - the character is printable and ASCII, when @flags has
583 * %ESCAPE_NAP bit set
584 * - the character is printable, when @flags has
586 * - the character is ASCII, when @flags has
588 * - the character doesn't fall into a class of symbols
589 * defined by given @flags
590 * In these cases we just pass through a character to the
593 * When %ESCAPE_APPEND is passed, the characters from @only
594 * have been excluded from the %ESCAPE_NAP, %ESCAPE_NP, and
597 if (!(is_append || in_dict) && is_dict &&
598 escape_passthrough(c, &p, end))
601 if (!(is_append && in_dict) && isascii(c) && isprint(c) &&
602 flags & ESCAPE_NAP && escape_passthrough(c, &p, end))
605 if (!(is_append && in_dict) && isprint(c) &&
606 flags & ESCAPE_NP && escape_passthrough(c, &p, end))
609 if (!(is_append && in_dict) && isascii(c) &&
610 flags & ESCAPE_NA && escape_passthrough(c, &p, end))
613 if (flags & ESCAPE_SPACE && escape_space(c, &p, end))
616 if (flags & ESCAPE_SPECIAL && escape_special(c, &p, end))
619 if (flags & ESCAPE_NULL && escape_null(c, &p, end))
622 /* ESCAPE_OCTAL and ESCAPE_HEX always go last */
623 if (flags & ESCAPE_OCTAL && escape_octal(c, &p, end))
626 if (flags & ESCAPE_HEX && escape_hex(c, &p, end))
629 escape_passthrough(c, &p, end);
634 EXPORT_SYMBOL(string_escape_mem);
637 * Return an allocated string that has been escaped of special characters
638 * and double quotes, making it safe to log in quotes.
640 char *kstrdup_quotable(const char *src, gfp_t gfp)
644 const int flags = ESCAPE_HEX;
645 const char esc[] = "\f\n\r\t\v\a\e\\\"";
651 dlen = string_escape_mem(src, slen, NULL, 0, flags, esc);
652 dst = kmalloc(dlen + 1, gfp);
656 WARN_ON(string_escape_mem(src, slen, dst, dlen, flags, esc) != dlen);
661 EXPORT_SYMBOL_GPL(kstrdup_quotable);
664 * Returns allocated NULL-terminated string containing process
665 * command line, with inter-argument NULLs replaced with spaces,
666 * and other special characters escaped.
668 char *kstrdup_quotable_cmdline(struct task_struct *task, gfp_t gfp)
670 char *buffer, *quoted;
673 buffer = kmalloc(PAGE_SIZE, GFP_KERNEL);
677 res = get_cmdline(task, buffer, PAGE_SIZE - 1);
680 /* Collapse trailing NULLs, leave res pointing to last non-NULL. */
681 while (--res >= 0 && buffer[res] == '\0')
684 /* Replace inter-argument NULLs. */
685 for (i = 0; i <= res; i++)
686 if (buffer[i] == '\0')
689 /* Make sure result is printable. */
690 quoted = kstrdup_quotable(buffer, gfp);
694 EXPORT_SYMBOL_GPL(kstrdup_quotable_cmdline);
697 * Returns allocated NULL-terminated string containing pathname,
698 * with special characters escaped, able to be safely logged. If
699 * there is an error, the leading character will be "<".
701 char *kstrdup_quotable_file(struct file *file, gfp_t gfp)
703 char *temp, *pathname;
706 return kstrdup("<unknown>", gfp);
708 /* We add 11 spaces for ' (deleted)' to be appended */
709 temp = kmalloc(PATH_MAX + 11, GFP_KERNEL);
711 return kstrdup("<no_memory>", gfp);
713 pathname = file_path(file, temp, PATH_MAX + 11);
714 if (IS_ERR(pathname))
715 pathname = kstrdup("<too_long>", gfp);
717 pathname = kstrdup_quotable(pathname, gfp);
722 EXPORT_SYMBOL_GPL(kstrdup_quotable_file);
725 * Returns duplicate string in which the @old characters are replaced by @new.
727 char *kstrdup_and_replace(const char *src, char old, char new, gfp_t gfp)
731 dst = kstrdup(src, gfp);
735 return strreplace(dst, old, new);
737 EXPORT_SYMBOL_GPL(kstrdup_and_replace);
740 * kasprintf_strarray - allocate and fill array of sequential strings
741 * @gfp: flags for the slab allocator
742 * @prefix: prefix to be used
743 * @n: amount of lines to be allocated and filled
745 * Allocates and fills @n strings using pattern "%s-%zu", where prefix
746 * is provided by caller. The caller is responsible to free them with
747 * kfree_strarray() after use.
749 * Returns array of strings or NULL when memory can't be allocated.
751 char **kasprintf_strarray(gfp_t gfp, const char *prefix, size_t n)
756 names = kcalloc(n + 1, sizeof(char *), gfp);
760 for (i = 0; i < n; i++) {
761 names[i] = kasprintf(gfp, "%s-%zu", prefix, i);
763 kfree_strarray(names, i);
770 EXPORT_SYMBOL_GPL(kasprintf_strarray);
773 * kfree_strarray - free a number of dynamically allocated strings contained
774 * in an array and the array itself
776 * @array: Dynamically allocated array of strings to free.
777 * @n: Number of strings (starting from the beginning of the array) to free.
779 * Passing a non-NULL @array and @n == 0 as well as NULL @array are valid
780 * use-cases. If @array is NULL, the function does nothing.
782 void kfree_strarray(char **array, size_t n)
789 for (i = 0; i < n; i++)
793 EXPORT_SYMBOL_GPL(kfree_strarray);
800 static void devm_kfree_strarray(struct device *dev, void *res)
802 struct strarray *array = res;
804 kfree_strarray(array->array, array->n);
807 char **devm_kasprintf_strarray(struct device *dev, const char *prefix, size_t n)
809 struct strarray *ptr;
811 ptr = devres_alloc(devm_kfree_strarray, sizeof(*ptr), GFP_KERNEL);
813 return ERR_PTR(-ENOMEM);
815 ptr->array = kasprintf_strarray(GFP_KERNEL, prefix, n);
818 return ERR_PTR(-ENOMEM);
822 devres_add(dev, ptr);
826 EXPORT_SYMBOL_GPL(devm_kasprintf_strarray);
829 * strscpy_pad() - Copy a C-string into a sized buffer
830 * @dest: Where to copy the string to
831 * @src: Where to copy the string from
832 * @count: Size of destination buffer
834 * Copy the string, or as much of it as fits, into the dest buffer. The
835 * behavior is undefined if the string buffers overlap. The destination
836 * buffer is always %NUL terminated, unless it's zero-sized.
838 * If the source string is shorter than the destination buffer, zeros
839 * the tail of the destination buffer.
841 * For full explanation of why you may want to consider using the
842 * 'strscpy' functions please see the function docstring for strscpy().
845 * * The number of characters copied (not including the trailing %NUL)
846 * * -E2BIG if count is 0 or @src was truncated.
848 ssize_t strscpy_pad(char *dest, const char *src, size_t count)
852 written = strscpy(dest, src, count);
853 if (written < 0 || written == count - 1)
856 memset(dest + written + 1, 0, count - written - 1);
860 EXPORT_SYMBOL(strscpy_pad);
863 * skip_spaces - Removes leading whitespace from @str.
864 * @str: The string to be stripped.
866 * Returns a pointer to the first non-whitespace character in @str.
868 char *skip_spaces(const char *str)
870 while (isspace(*str))
874 EXPORT_SYMBOL(skip_spaces);
877 * strim - Removes leading and trailing whitespace from @s.
878 * @s: The string to be stripped.
880 * Note that the first trailing whitespace is replaced with a %NUL-terminator
881 * in the given string @s. Returns a pointer to the first non-whitespace
894 while (end >= s && isspace(*end))
898 return skip_spaces(s);
900 EXPORT_SYMBOL(strim);
903 * sysfs_streq - return true if strings are equal, modulo trailing newline
905 * @s2: another string
907 * This routine returns true iff two strings are equal, treating both
908 * NUL and newline-then-NUL as equivalent string terminations. It's
909 * geared for use with sysfs input strings, which generally terminate
910 * with newlines but are compared against values without newlines.
912 bool sysfs_streq(const char *s1, const char *s2)
914 while (*s1 && *s1 == *s2) {
921 if (!*s1 && *s2 == '\n' && !s2[1])
923 if (*s1 == '\n' && !s1[1] && !*s2)
927 EXPORT_SYMBOL(sysfs_streq);
930 * match_string - matches given string in an array
931 * @array: array of strings
932 * @n: number of strings in the array or -1 for NULL terminated arrays
933 * @string: string to match with
935 * This routine will look for a string in an array of strings up to the
936 * n-th element in the array or until the first NULL element.
938 * Historically the value of -1 for @n, was used to search in arrays that
939 * are NULL terminated. However, the function does not make a distinction
940 * when finishing the search: either @n elements have been compared OR
941 * the first NULL element was found.
944 * index of a @string in the @array if matches, or %-EINVAL otherwise.
946 int match_string(const char * const *array, size_t n, const char *string)
951 for (index = 0; index < n; index++) {
955 if (!strcmp(item, string))
961 EXPORT_SYMBOL(match_string);
964 * __sysfs_match_string - matches given string in an array
965 * @array: array of strings
966 * @n: number of strings in the array or -1 for NULL terminated arrays
967 * @str: string to match with
969 * Returns index of @str in the @array or -EINVAL, just like match_string().
970 * Uses sysfs_streq instead of strcmp for matching.
972 * This routine will look for a string in an array of strings up to the
973 * n-th element in the array or until the first NULL element.
975 * Historically the value of -1 for @n, was used to search in arrays that
976 * are NULL terminated. However, the function does not make a distinction
977 * when finishing the search: either @n elements have been compared OR
978 * the first NULL element was found.
980 int __sysfs_match_string(const char * const *array, size_t n, const char *str)
985 for (index = 0; index < n; index++) {
989 if (sysfs_streq(item, str))
995 EXPORT_SYMBOL(__sysfs_match_string);
998 * strreplace - Replace all occurrences of character in string.
999 * @str: The string to operate on.
1000 * @old: The character being replaced.
1001 * @new: The character @old is replaced with.
1003 * Replaces the each @old character with a @new one in the given string @str.
1005 * Return: pointer to the string @str itself.
1007 char *strreplace(char *str, char old, char new)
1016 EXPORT_SYMBOL(strreplace);
1019 * memcpy_and_pad - Copy one buffer to another with padding
1020 * @dest: Where to copy to
1021 * @dest_len: The destination buffer size
1022 * @src: Where to copy from
1023 * @count: The number of bytes to copy
1024 * @pad: Character to use for padding if space is left in destination.
1026 void memcpy_and_pad(void *dest, size_t dest_len, const void *src, size_t count,
1029 if (dest_len > count) {
1030 memcpy(dest, src, count);
1031 memset(dest + count, pad, dest_len - count);
1033 memcpy(dest, src, dest_len);
1036 EXPORT_SYMBOL(memcpy_and_pad);
1038 #ifdef CONFIG_FORTIFY_SOURCE
1039 /* These are placeholders for fortify compile-time warnings. */
1040 void __read_overflow2_field(size_t avail, size_t wanted) { }
1041 EXPORT_SYMBOL(__read_overflow2_field);
1042 void __write_overflow_field(size_t avail, size_t wanted) { }
1043 EXPORT_SYMBOL(__write_overflow_field);
1045 void fortify_panic(const char *name)
1047 pr_emerg("detected buffer overflow in %s\n", name);
1050 EXPORT_SYMBOL(fortify_panic);
1051 #endif /* CONFIG_FORTIFY_SOURCE */