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
);
141 int parse_int_array(const char *buf
, size_t count
, int **array
)
145 get_options(buf
, 0, &nints
);
149 ints
= kcalloc(nints
+ 1, sizeof(*ints
), GFP_KERNEL
);
153 get_options(buf
, nints
+ 1, ints
);
158 EXPORT_SYMBOL(parse_int_array
);
161 * parse_int_array_user - Split string into a sequence of integers
162 * @from: The user space buffer to read from
163 * @count: The maximum number of bytes to read
164 * @array: Returned pointer to sequence of integers
166 * On success @array is allocated and initialized with a sequence of
167 * integers extracted from the @from plus an additional element that
168 * begins the sequence and specifies the integers count.
170 * Caller takes responsibility for freeing @array when it is no longer
173 int parse_int_array_user(const char __user
*from
, size_t count
, int **array
)
178 buf
= memdup_user_nul(from
, count
);
182 ret
= parse_int_array(buf
, count
, array
);
186 EXPORT_SYMBOL(parse_int_array_user
);
188 static bool unescape_space(char **src
, char **dst
)
190 char *p
= *dst
, *q
= *src
;
216 static bool unescape_octal(char **src
, char **dst
)
218 char *p
= *dst
, *q
= *src
;
221 if (isodigit(*q
) == 0)
225 while (num
< 32 && isodigit(*q
) && (q
- *src
< 3)) {
235 static bool unescape_hex(char **src
, char **dst
)
237 char *p
= *dst
, *q
= *src
;
244 num
= digit
= hex_to_bin(*q
++);
248 digit
= hex_to_bin(*q
);
251 num
= (num
<< 4) | digit
;
259 static bool unescape_special(char **src
, char **dst
)
261 char *p
= *dst
, *q
= *src
;
285 * string_unescape - unquote characters in the given string
286 * @src: source buffer (escaped)
287 * @dst: destination buffer (unescaped)
288 * @size: size of the destination buffer (0 to unlimit)
289 * @flags: combination of the flags.
292 * The function unquotes characters in the given string.
294 * Because the size of the output will be the same as or less than the size of
295 * the input, the transformation may be performed in place.
297 * Caller must provide valid source and destination pointers. Be aware that
298 * destination buffer will always be NULL-terminated. Source string must be
299 * NULL-terminated as well. The supported flags are::
304 * '\r' - carriage return
305 * '\t' - horizontal tab
306 * '\v' - vertical tab
308 * '\NNN' - byte with octal value NNN (1 to 3 digits)
310 * '\xHH' - byte with hexadecimal value HH (1 to 2 digits)
312 * '\"' - double quote
317 * all previous together
320 * The amount of the characters processed to the destination buffer excluding
321 * trailing '\0' is returned.
323 int string_unescape(char *src
, char *dst
, size_t size
, unsigned int flags
)
330 while (*src
&& --size
) {
331 if (src
[0] == '\\' && src
[1] != '\0' && size
> 1) {
335 if (flags
& UNESCAPE_SPACE
&&
336 unescape_space(&src
, &out
))
339 if (flags
& UNESCAPE_OCTAL
&&
340 unescape_octal(&src
, &out
))
343 if (flags
& UNESCAPE_HEX
&&
344 unescape_hex(&src
, &out
))
347 if (flags
& UNESCAPE_SPECIAL
&&
348 unescape_special(&src
, &out
))
359 EXPORT_SYMBOL(string_unescape
);
361 static bool escape_passthrough(unsigned char c
, char **dst
, char *end
)
371 static bool escape_space(unsigned char c
, char **dst
, char *end
)
407 static bool escape_special(unsigned char c
, char **dst
, char *end
)
440 static bool escape_null(unsigned char c
, char **dst
, char *end
)
458 static bool escape_octal(unsigned char c
, char **dst
, char *end
)
466 *out
= ((c
>> 6) & 0x07) + '0';
469 *out
= ((c
>> 3) & 0x07) + '0';
472 *out
= ((c
>> 0) & 0x07) + '0';
479 static bool escape_hex(unsigned char c
, char **dst
, char *end
)
490 *out
= hex_asc_hi(c
);
493 *out
= hex_asc_lo(c
);
501 * string_escape_mem - quote characters in the given memory buffer
502 * @src: source buffer (unescaped)
503 * @isz: source buffer size
504 * @dst: destination buffer (escaped)
505 * @osz: destination buffer size
506 * @flags: combination of the flags
507 * @only: NULL-terminated string containing characters used to limit
508 * the selected escape class. If characters are included in @only
509 * that would not normally be escaped by the classes selected
510 * in @flags, they will be copied to @dst unescaped.
513 * The process of escaping byte buffer includes several parts. They are applied
514 * in the following sequence.
516 * 1. The character is not matched to the one from @only string and thus
517 * must go as-is to the output.
518 * 2. The character is matched to the printable and ASCII classes, if asked,
519 * and in case of match it passes through to the output.
520 * 3. The character is matched to the printable or ASCII class, if asked,
521 * and in case of match it passes through to the output.
522 * 4. The character is checked if it falls into the class given by @flags.
523 * %ESCAPE_OCTAL and %ESCAPE_HEX are going last since they cover any
524 * character. Note that they actually can't go together, otherwise
525 * %ESCAPE_HEX will be ignored.
527 * Caller must provide valid source and destination pointers. Be aware that
528 * destination buffer will not be NULL-terminated, thus caller have to append
529 * it if needs. The supported flags are::
531 * %ESCAPE_SPACE: (special white space, not space itself)
534 * '\r' - carriage return
535 * '\t' - horizontal tab
536 * '\v' - vertical tab
538 * '\"' - double quote
545 * '\NNN' - byte with octal value NNN (3 digits)
547 * all previous together
549 * escape only non-printable characters, checked by isprint()
551 * all previous together
553 * '\xHH' - byte with hexadecimal value HH (2 digits)
555 * escape only non-ascii characters, checked by isascii()
557 * escape only non-printable or non-ascii characters
559 * append characters from @only to be escaped by the given classes
561 * %ESCAPE_APPEND would help to pass additional characters to the escaped, when
562 * one of %ESCAPE_NP, %ESCAPE_NA, or %ESCAPE_NAP is provided.
564 * One notable caveat, the %ESCAPE_NAP, %ESCAPE_NP and %ESCAPE_NA have the
565 * higher priority than the rest of the flags (%ESCAPE_NAP is the highest).
566 * It doesn't make much sense to use either of them without %ESCAPE_OCTAL
567 * or %ESCAPE_HEX, because they cover most of the other character classes.
568 * %ESCAPE_NAP can utilize %ESCAPE_SPACE or %ESCAPE_SPECIAL in addition to
572 * The total size of the escaped output that would be generated for
573 * the given input and flags. To check whether the output was
574 * truncated, compare the return value to osz. There is room left in
575 * dst for a '\0' terminator if and only if ret < osz.
577 int string_escape_mem(const char *src
, size_t isz
, char *dst
, size_t osz
,
578 unsigned int flags
, const char *only
)
582 bool is_dict
= only
&& *only
;
583 bool is_append
= flags
& ESCAPE_APPEND
;
586 unsigned char c
= *src
++;
587 bool in_dict
= is_dict
&& strchr(only
, c
);
590 * Apply rules in the following sequence:
591 * - the @only string is supplied and does not contain a
592 * character under question
593 * - the character is printable and ASCII, when @flags has
594 * %ESCAPE_NAP bit set
595 * - the character is printable, when @flags has
597 * - the character is ASCII, when @flags has
599 * - the character doesn't fall into a class of symbols
600 * defined by given @flags
601 * In these cases we just pass through a character to the
604 * When %ESCAPE_APPEND is passed, the characters from @only
605 * have been excluded from the %ESCAPE_NAP, %ESCAPE_NP, and
608 if (!(is_append
|| in_dict
) && is_dict
&&
609 escape_passthrough(c
, &p
, end
))
612 if (!(is_append
&& in_dict
) && isascii(c
) && isprint(c
) &&
613 flags
& ESCAPE_NAP
&& escape_passthrough(c
, &p
, end
))
616 if (!(is_append
&& in_dict
) && isprint(c
) &&
617 flags
& ESCAPE_NP
&& escape_passthrough(c
, &p
, end
))
620 if (!(is_append
&& in_dict
) && isascii(c
) &&
621 flags
& ESCAPE_NA
&& escape_passthrough(c
, &p
, end
))
624 if (flags
& ESCAPE_SPACE
&& escape_space(c
, &p
, end
))
627 if (flags
& ESCAPE_SPECIAL
&& escape_special(c
, &p
, end
))
630 if (flags
& ESCAPE_NULL
&& escape_null(c
, &p
, end
))
633 /* ESCAPE_OCTAL and ESCAPE_HEX always go last */
634 if (flags
& ESCAPE_OCTAL
&& escape_octal(c
, &p
, end
))
637 if (flags
& ESCAPE_HEX
&& escape_hex(c
, &p
, end
))
640 escape_passthrough(c
, &p
, end
);
645 EXPORT_SYMBOL(string_escape_mem
);
648 * Return an allocated string that has been escaped of special characters
649 * and double quotes, making it safe to log in quotes.
651 char *kstrdup_quotable(const char *src
, gfp_t gfp
)
655 const int flags
= ESCAPE_HEX
;
656 const char esc
[] = "\f\n\r\t\v\a\e\\\"";
662 dlen
= string_escape_mem(src
, slen
, NULL
, 0, flags
, esc
);
663 dst
= kmalloc(dlen
+ 1, gfp
);
667 WARN_ON(string_escape_mem(src
, slen
, dst
, dlen
, flags
, esc
) != dlen
);
672 EXPORT_SYMBOL_GPL(kstrdup_quotable
);
675 * Returns allocated NULL-terminated string containing process
676 * command line, with inter-argument NULLs replaced with spaces,
677 * and other special characters escaped.
679 char *kstrdup_quotable_cmdline(struct task_struct
*task
, gfp_t gfp
)
681 char *buffer
, *quoted
;
684 buffer
= kmalloc(PAGE_SIZE
, GFP_KERNEL
);
688 res
= get_cmdline(task
, buffer
, PAGE_SIZE
- 1);
691 /* Collapse trailing NULLs, leave res pointing to last non-NULL. */
692 while (--res
>= 0 && buffer
[res
] == '\0')
695 /* Replace inter-argument NULLs. */
696 for (i
= 0; i
<= res
; i
++)
697 if (buffer
[i
] == '\0')
700 /* Make sure result is printable. */
701 quoted
= kstrdup_quotable(buffer
, gfp
);
705 EXPORT_SYMBOL_GPL(kstrdup_quotable_cmdline
);
708 * Returns allocated NULL-terminated string containing pathname,
709 * with special characters escaped, able to be safely logged. If
710 * there is an error, the leading character will be "<".
712 char *kstrdup_quotable_file(struct file
*file
, gfp_t gfp
)
714 char *temp
, *pathname
;
717 return kstrdup("<unknown>", gfp
);
719 /* We add 11 spaces for ' (deleted)' to be appended */
720 temp
= kmalloc(PATH_MAX
+ 11, GFP_KERNEL
);
722 return kstrdup("<no_memory>", gfp
);
724 pathname
= file_path(file
, temp
, PATH_MAX
+ 11);
725 if (IS_ERR(pathname
))
726 pathname
= kstrdup("<too_long>", gfp
);
728 pathname
= kstrdup_quotable(pathname
, gfp
);
733 EXPORT_SYMBOL_GPL(kstrdup_quotable_file
);
736 * Returns duplicate string in which the @old characters are replaced by @new.
738 char *kstrdup_and_replace(const char *src
, char old
, char new, gfp_t gfp
)
742 dst
= kstrdup(src
, gfp
);
746 return strreplace(dst
, old
, new);
748 EXPORT_SYMBOL_GPL(kstrdup_and_replace
);
751 * kasprintf_strarray - allocate and fill array of sequential strings
752 * @gfp: flags for the slab allocator
753 * @prefix: prefix to be used
754 * @n: amount of lines to be allocated and filled
756 * Allocates and fills @n strings using pattern "%s-%zu", where prefix
757 * is provided by caller. The caller is responsible to free them with
758 * kfree_strarray() after use.
760 * Returns array of strings or NULL when memory can't be allocated.
762 char **kasprintf_strarray(gfp_t gfp
, const char *prefix
, size_t n
)
767 names
= kcalloc(n
+ 1, sizeof(char *), gfp
);
771 for (i
= 0; i
< n
; i
++) {
772 names
[i
] = kasprintf(gfp
, "%s-%zu", prefix
, i
);
774 kfree_strarray(names
, i
);
781 EXPORT_SYMBOL_GPL(kasprintf_strarray
);
784 * kfree_strarray - free a number of dynamically allocated strings contained
785 * in an array and the array itself
787 * @array: Dynamically allocated array of strings to free.
788 * @n: Number of strings (starting from the beginning of the array) to free.
790 * Passing a non-NULL @array and @n == 0 as well as NULL @array are valid
791 * use-cases. If @array is NULL, the function does nothing.
793 void kfree_strarray(char **array
, size_t n
)
800 for (i
= 0; i
< n
; i
++)
804 EXPORT_SYMBOL_GPL(kfree_strarray
);
811 static void devm_kfree_strarray(struct device
*dev
, void *res
)
813 struct strarray
*array
= res
;
815 kfree_strarray(array
->array
, array
->n
);
818 char **devm_kasprintf_strarray(struct device
*dev
, const char *prefix
, size_t n
)
820 struct strarray
*ptr
;
822 ptr
= devres_alloc(devm_kfree_strarray
, sizeof(*ptr
), GFP_KERNEL
);
824 return ERR_PTR(-ENOMEM
);
826 ptr
->array
= kasprintf_strarray(GFP_KERNEL
, prefix
, n
);
829 return ERR_PTR(-ENOMEM
);
833 devres_add(dev
, ptr
);
837 EXPORT_SYMBOL_GPL(devm_kasprintf_strarray
);
840 * skip_spaces - Removes leading whitespace from @str.
841 * @str: The string to be stripped.
843 * Returns a pointer to the first non-whitespace character in @str.
845 char *skip_spaces(const char *str
)
847 while (isspace(*str
))
851 EXPORT_SYMBOL(skip_spaces
);
854 * strim - Removes leading and trailing whitespace from @s.
855 * @s: The string to be stripped.
857 * Note that the first trailing whitespace is replaced with a %NUL-terminator
858 * in the given string @s. Returns a pointer to the first non-whitespace
871 while (end
>= s
&& isspace(*end
))
875 return skip_spaces(s
);
877 EXPORT_SYMBOL(strim
);
880 * sysfs_streq - return true if strings are equal, modulo trailing newline
882 * @s2: another string
884 * This routine returns true iff two strings are equal, treating both
885 * NUL and newline-then-NUL as equivalent string terminations. It's
886 * geared for use with sysfs input strings, which generally terminate
887 * with newlines but are compared against values without newlines.
889 bool sysfs_streq(const char *s1
, const char *s2
)
891 while (*s1
&& *s1
== *s2
) {
898 if (!*s1
&& *s2
== '\n' && !s2
[1])
900 if (*s1
== '\n' && !s1
[1] && !*s2
)
904 EXPORT_SYMBOL(sysfs_streq
);
907 * match_string - matches given string in an array
908 * @array: array of strings
909 * @n: number of strings in the array or -1 for NULL terminated arrays
910 * @string: string to match with
912 * This routine will look for a string in an array of strings up to the
913 * n-th element in the array or until the first NULL element.
915 * Historically the value of -1 for @n, was used to search in arrays that
916 * are NULL terminated. However, the function does not make a distinction
917 * when finishing the search: either @n elements have been compared OR
918 * the first NULL element was found.
921 * index of a @string in the @array if matches, or %-EINVAL otherwise.
923 int match_string(const char * const *array
, size_t n
, const char *string
)
928 for (index
= 0; index
< n
; index
++) {
932 if (!strcmp(item
, string
))
938 EXPORT_SYMBOL(match_string
);
941 * __sysfs_match_string - matches given string in an array
942 * @array: array of strings
943 * @n: number of strings in the array or -1 for NULL terminated arrays
944 * @str: string to match with
946 * Returns index of @str in the @array or -EINVAL, just like match_string().
947 * Uses sysfs_streq instead of strcmp for matching.
949 * This routine will look for a string in an array of strings up to the
950 * n-th element in the array or until the first NULL element.
952 * Historically the value of -1 for @n, was used to search in arrays that
953 * are NULL terminated. However, the function does not make a distinction
954 * when finishing the search: either @n elements have been compared OR
955 * the first NULL element was found.
957 int __sysfs_match_string(const char * const *array
, size_t n
, const char *str
)
962 for (index
= 0; index
< n
; index
++) {
966 if (sysfs_streq(item
, str
))
972 EXPORT_SYMBOL(__sysfs_match_string
);
975 * strreplace - Replace all occurrences of character in string.
976 * @str: The string to operate on.
977 * @old: The character being replaced.
978 * @new: The character @old is replaced with.
980 * Replaces the each @old character with a @new one in the given string @str.
982 * Return: pointer to the string @str itself.
984 char *strreplace(char *str
, char old
, char new)
993 EXPORT_SYMBOL(strreplace
);
996 * memcpy_and_pad - Copy one buffer to another with padding
997 * @dest: Where to copy to
998 * @dest_len: The destination buffer size
999 * @src: Where to copy from
1000 * @count: The number of bytes to copy
1001 * @pad: Character to use for padding if space is left in destination.
1003 void memcpy_and_pad(void *dest
, size_t dest_len
, const void *src
, size_t count
,
1006 if (dest_len
> count
) {
1007 memcpy(dest
, src
, count
);
1008 memset(dest
+ count
, pad
, dest_len
- count
);
1010 memcpy(dest
, src
, dest_len
);
1013 EXPORT_SYMBOL(memcpy_and_pad
);
1015 #ifdef CONFIG_FORTIFY_SOURCE
1016 /* These are placeholders for fortify compile-time warnings. */
1017 void __read_overflow2_field(size_t avail
, size_t wanted
) { }
1018 EXPORT_SYMBOL(__read_overflow2_field
);
1019 void __write_overflow_field(size_t avail
, size_t wanted
) { }
1020 EXPORT_SYMBOL(__write_overflow_field
);
1022 static const char * const fortify_func_name
[] = {
1023 #define MAKE_FORTIFY_FUNC_NAME(func) [MAKE_FORTIFY_FUNC(func)] = #func
1024 EACH_FORTIFY_FUNC(MAKE_FORTIFY_FUNC_NAME
)
1025 #undef MAKE_FORTIFY_FUNC_NAME
1028 void __fortify_report(const u8 reason
, const size_t avail
, const size_t size
)
1030 const u8 func
= FORTIFY_REASON_FUNC(reason
);
1031 const bool write
= FORTIFY_REASON_DIR(reason
);
1034 name
= fortify_func_name
[umin(func
, FORTIFY_FUNC_UNKNOWN
)];
1035 WARN(1, "%s: detected buffer overflow: %zu byte %s of buffer size %zu\n",
1036 name
, size
, str_read_write(!write
), avail
);
1038 EXPORT_SYMBOL(__fortify_report
);
1040 void __fortify_panic(const u8 reason
, const size_t avail
, const size_t size
)
1042 __fortify_report(reason
, avail
, size
);
1045 EXPORT_SYMBOL(__fortify_panic
);
1046 #endif /* CONFIG_FORTIFY_SOURCE */