865 lines · c
1// SPDX-License-Identifier: GPL-2.02/*3 * linux/lib/string.c4 *5 * Copyright (C) 1991, 1992 Linus Torvalds6 */7 8/*9 * This file should be used only for "library" routines that may have10 * alternative implementations on specific architectures (generally11 * found in <asm-xx/string.h>), or get overloaded by FORTIFY_SOURCE.12 * (Specifically, this file is built with __NO_FORTIFY.)13 *14 * Other helper functions should live in string_helpers.c.15 */16 17#define __NO_FORTIFY18#include <linux/bits.h>19#include <linux/bug.h>20#include <linux/ctype.h>21#include <linux/errno.h>22#include <linux/limits.h>23#include <linux/linkage.h>24#include <linux/stddef.h>25#include <linux/string.h>26#include <linux/types.h>27 28#include <asm/page.h>29#include <asm/rwonce.h>30#include <linux/unaligned.h>31#include <asm/word-at-a-time.h>32 33#ifndef __HAVE_ARCH_STRNCASECMP34/**35 * strncasecmp - Case insensitive, length-limited string comparison36 * @s1: One string37 * @s2: The other string38 * @len: the maximum number of characters to compare39 */40int strncasecmp(const char *s1, const char *s2, size_t len)41{42 /* Yes, Virginia, it had better be unsigned */43 unsigned char c1, c2;44 45 if (!len)46 return 0;47 48 do {49 c1 = *s1++;50 c2 = *s2++;51 if (!c1 || !c2)52 break;53 if (c1 == c2)54 continue;55 c1 = tolower(c1);56 c2 = tolower(c2);57 if (c1 != c2)58 break;59 } while (--len);60 return (int)c1 - (int)c2;61}62EXPORT_SYMBOL(strncasecmp);63#endif64 65#ifndef __HAVE_ARCH_STRCASECMP66int strcasecmp(const char *s1, const char *s2)67{68 int c1, c2;69 70 do {71 c1 = tolower(*s1++);72 c2 = tolower(*s2++);73 } while (c1 == c2 && c1 != 0);74 return c1 - c2;75}76EXPORT_SYMBOL(strcasecmp);77#endif78 79#ifndef __HAVE_ARCH_STRCPY80char *strcpy(char *dest, const char *src)81{82 char *tmp = dest;83 84 while ((*dest++ = *src++) != '\0')85 /* nothing */;86 return tmp;87}88EXPORT_SYMBOL(strcpy);89#endif90 91#ifndef __HAVE_ARCH_STRNCPY92char *strncpy(char *dest, const char *src, size_t count)93{94 char *tmp = dest;95 96 while (count) {97 if ((*tmp = *src) != 0)98 src++;99 tmp++;100 count--;101 }102 return dest;103}104EXPORT_SYMBOL(strncpy);105#endif106 107ssize_t sized_strscpy(char *dest, const char *src, size_t count)108{109 const struct word_at_a_time constants = WORD_AT_A_TIME_CONSTANTS;110 size_t max = count;111 long res = 0;112 113 if (count == 0 || WARN_ON_ONCE(count > INT_MAX))114 return -E2BIG;115 116#ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS117 /*118 * If src is unaligned, don't cross a page boundary,119 * since we don't know if the next page is mapped.120 */121 if ((long)src & (sizeof(long) - 1)) {122 size_t limit = PAGE_SIZE - ((long)src & (PAGE_SIZE - 1));123 if (limit < max)124 max = limit;125 }126#else127 /* If src or dest is unaligned, don't do word-at-a-time. */128 if (((long) dest | (long) src) & (sizeof(long) - 1))129 max = 0;130#endif131 132 /*133 * read_word_at_a_time() below may read uninitialized bytes after the134 * trailing zero and use them in comparisons. Disable this optimization135 * under KMSAN to prevent false positive reports.136 */137 if (IS_ENABLED(CONFIG_KMSAN))138 max = 0;139 140 while (max >= sizeof(unsigned long)) {141 unsigned long c, data;142 143 c = read_word_at_a_time(src+res);144 if (has_zero(c, &data, &constants)) {145 data = prep_zero_mask(c, data, &constants);146 data = create_zero_mask(data);147 *(unsigned long *)(dest+res) = c & zero_bytemask(data);148 return res + find_zero(data);149 }150 *(unsigned long *)(dest+res) = c;151 res += sizeof(unsigned long);152 count -= sizeof(unsigned long);153 max -= sizeof(unsigned long);154 }155 156 while (count) {157 char c;158 159 c = src[res];160 dest[res] = c;161 if (!c)162 return res;163 res++;164 count--;165 }166 167 /* Hit buffer length without finding a NUL; force NUL-termination. */168 if (res)169 dest[res-1] = '\0';170 171 return -E2BIG;172}173EXPORT_SYMBOL(sized_strscpy);174 175/**176 * stpcpy - copy a string from src to dest returning a pointer to the new end177 * of dest, including src's %NUL-terminator. May overrun dest.178 * @dest: pointer to end of string being copied into. Must be large enough179 * to receive copy.180 * @src: pointer to the beginning of string being copied from. Must not overlap181 * dest.182 *183 * stpcpy differs from strcpy in a key way: the return value is a pointer184 * to the new %NUL-terminating character in @dest. (For strcpy, the return185 * value is a pointer to the start of @dest). This interface is considered186 * unsafe as it doesn't perform bounds checking of the inputs. As such it's187 * not recommended for usage. Instead, its definition is provided in case188 * the compiler lowers other libcalls to stpcpy.189 */190char *stpcpy(char *__restrict__ dest, const char *__restrict__ src);191char *stpcpy(char *__restrict__ dest, const char *__restrict__ src)192{193 while ((*dest++ = *src++) != '\0')194 /* nothing */;195 return --dest;196}197EXPORT_SYMBOL(stpcpy);198 199#ifndef __HAVE_ARCH_STRCAT200char *strcat(char *dest, const char *src)201{202 char *tmp = dest;203 204 while (*dest)205 dest++;206 while ((*dest++ = *src++) != '\0')207 ;208 return tmp;209}210EXPORT_SYMBOL(strcat);211#endif212 213#ifndef __HAVE_ARCH_STRNCAT214char *strncat(char *dest, const char *src, size_t count)215{216 char *tmp = dest;217 218 if (count) {219 while (*dest)220 dest++;221 while ((*dest++ = *src++) != 0) {222 if (--count == 0) {223 *dest = '\0';224 break;225 }226 }227 }228 return tmp;229}230EXPORT_SYMBOL(strncat);231#endif232 233#ifndef __HAVE_ARCH_STRLCAT234size_t strlcat(char *dest, const char *src, size_t count)235{236 size_t dsize = strlen(dest);237 size_t len = strlen(src);238 size_t res = dsize + len;239 240 /* This would be a bug */241 BUG_ON(dsize >= count);242 243 dest += dsize;244 count -= dsize;245 if (len >= count)246 len = count-1;247 __builtin_memcpy(dest, src, len);248 dest[len] = 0;249 return res;250}251EXPORT_SYMBOL(strlcat);252#endif253 254#ifndef __HAVE_ARCH_STRCMP255/**256 * strcmp - Compare two strings257 * @cs: One string258 * @ct: Another string259 */260int strcmp(const char *cs, const char *ct)261{262 unsigned char c1, c2;263 264 while (1) {265 c1 = *cs++;266 c2 = *ct++;267 if (c1 != c2)268 return c1 < c2 ? -1 : 1;269 if (!c1)270 break;271 }272 return 0;273}274EXPORT_SYMBOL(strcmp);275#endif276 277#ifndef __HAVE_ARCH_STRNCMP278/**279 * strncmp - Compare two length-limited strings280 * @cs: One string281 * @ct: Another string282 * @count: The maximum number of bytes to compare283 */284int strncmp(const char *cs, const char *ct, size_t count)285{286 unsigned char c1, c2;287 288 while (count) {289 c1 = *cs++;290 c2 = *ct++;291 if (c1 != c2)292 return c1 < c2 ? -1 : 1;293 if (!c1)294 break;295 count--;296 }297 return 0;298}299EXPORT_SYMBOL(strncmp);300#endif301 302#ifndef __HAVE_ARCH_STRCHR303/**304 * strchr - Find the first occurrence of a character in a string305 * @s: The string to be searched306 * @c: The character to search for307 *308 * Note that the %NUL-terminator is considered part of the string, and can309 * be searched for.310 */311char *strchr(const char *s, int c)312{313 for (; *s != (char)c; ++s)314 if (*s == '\0')315 return NULL;316 return (char *)s;317}318EXPORT_SYMBOL(strchr);319#endif320 321#ifndef __HAVE_ARCH_STRCHRNUL322/**323 * strchrnul - Find and return a character in a string, or end of string324 * @s: The string to be searched325 * @c: The character to search for326 *327 * Returns pointer to first occurrence of 'c' in s. If c is not found, then328 * return a pointer to the null byte at the end of s.329 */330char *strchrnul(const char *s, int c)331{332 while (*s && *s != (char)c)333 s++;334 return (char *)s;335}336EXPORT_SYMBOL(strchrnul);337#endif338 339/**340 * strnchrnul - Find and return a character in a length limited string,341 * or end of string342 * @s: The string to be searched343 * @count: The number of characters to be searched344 * @c: The character to search for345 *346 * Returns pointer to the first occurrence of 'c' in s. If c is not found,347 * then return a pointer to the last character of the string.348 */349char *strnchrnul(const char *s, size_t count, int c)350{351 while (count-- && *s && *s != (char)c)352 s++;353 return (char *)s;354}355 356#ifndef __HAVE_ARCH_STRRCHR357/**358 * strrchr - Find the last occurrence of a character in a string359 * @s: The string to be searched360 * @c: The character to search for361 */362char *strrchr(const char *s, int c)363{364 const char *last = NULL;365 do {366 if (*s == (char)c)367 last = s;368 } while (*s++);369 return (char *)last;370}371EXPORT_SYMBOL(strrchr);372#endif373 374#ifndef __HAVE_ARCH_STRNCHR375/**376 * strnchr - Find a character in a length limited string377 * @s: The string to be searched378 * @count: The number of characters to be searched379 * @c: The character to search for380 *381 * Note that the %NUL-terminator is considered part of the string, and can382 * be searched for.383 */384char *strnchr(const char *s, size_t count, int c)385{386 while (count--) {387 if (*s == (char)c)388 return (char *)s;389 if (*s++ == '\0')390 break;391 }392 return NULL;393}394EXPORT_SYMBOL(strnchr);395#endif396 397#ifndef __HAVE_ARCH_STRLEN398size_t strlen(const char *s)399{400 const char *sc;401 402 for (sc = s; *sc != '\0'; ++sc)403 /* nothing */;404 return sc - s;405}406EXPORT_SYMBOL(strlen);407#endif408 409#ifndef __HAVE_ARCH_STRNLEN410size_t strnlen(const char *s, size_t count)411{412 const char *sc;413 414 for (sc = s; count-- && *sc != '\0'; ++sc)415 /* nothing */;416 return sc - s;417}418EXPORT_SYMBOL(strnlen);419#endif420 421#ifndef __HAVE_ARCH_STRSPN422/**423 * strspn - Calculate the length of the initial substring of @s which only contain letters in @accept424 * @s: The string to be searched425 * @accept: The string to search for426 */427size_t strspn(const char *s, const char *accept)428{429 const char *p;430 431 for (p = s; *p != '\0'; ++p) {432 if (!strchr(accept, *p))433 break;434 }435 return p - s;436}437EXPORT_SYMBOL(strspn);438#endif439 440#ifndef __HAVE_ARCH_STRCSPN441/**442 * strcspn - Calculate the length of the initial substring of @s which does not contain letters in @reject443 * @s: The string to be searched444 * @reject: The string to avoid445 */446size_t strcspn(const char *s, const char *reject)447{448 const char *p;449 450 for (p = s; *p != '\0'; ++p) {451 if (strchr(reject, *p))452 break;453 }454 return p - s;455}456EXPORT_SYMBOL(strcspn);457#endif458 459#ifndef __HAVE_ARCH_STRPBRK460/**461 * strpbrk - Find the first occurrence of a set of characters462 * @cs: The string to be searched463 * @ct: The characters to search for464 */465char *strpbrk(const char *cs, const char *ct)466{467 const char *sc;468 469 for (sc = cs; *sc != '\0'; ++sc) {470 if (strchr(ct, *sc))471 return (char *)sc;472 }473 return NULL;474}475EXPORT_SYMBOL(strpbrk);476#endif477 478#ifndef __HAVE_ARCH_STRSEP479/**480 * strsep - Split a string into tokens481 * @s: The string to be searched482 * @ct: The characters to search for483 *484 * strsep() updates @s to point after the token, ready for the next call.485 *486 * It returns empty tokens, too, behaving exactly like the libc function487 * of that name. In fact, it was stolen from glibc2 and de-fancy-fied.488 * Same semantics, slimmer shape. ;)489 */490char *strsep(char **s, const char *ct)491{492 char *sbegin = *s;493 char *end;494 495 if (sbegin == NULL)496 return NULL;497 498 end = strpbrk(sbegin, ct);499 if (end)500 *end++ = '\0';501 *s = end;502 return sbegin;503}504EXPORT_SYMBOL(strsep);505#endif506 507#ifndef __HAVE_ARCH_MEMSET508/**509 * memset - Fill a region of memory with the given value510 * @s: Pointer to the start of the area.511 * @c: The byte to fill the area with512 * @count: The size of the area.513 *514 * Do not use memset() to access IO space, use memset_io() instead.515 */516void *memset(void *s, int c, size_t count)517{518 char *xs = s;519 520 while (count--)521 *xs++ = c;522 return s;523}524EXPORT_SYMBOL(memset);525#endif526 527#ifndef __HAVE_ARCH_MEMSET16528/**529 * memset16() - Fill a memory area with a uint16_t530 * @s: Pointer to the start of the area.531 * @v: The value to fill the area with532 * @count: The number of values to store533 *534 * Differs from memset() in that it fills with a uint16_t instead535 * of a byte. Remember that @count is the number of uint16_ts to536 * store, not the number of bytes.537 */538void *memset16(uint16_t *s, uint16_t v, size_t count)539{540 uint16_t *xs = s;541 542 while (count--)543 *xs++ = v;544 return s;545}546EXPORT_SYMBOL(memset16);547#endif548 549#ifndef __HAVE_ARCH_MEMSET32550/**551 * memset32() - Fill a memory area with a uint32_t552 * @s: Pointer to the start of the area.553 * @v: The value to fill the area with554 * @count: The number of values to store555 *556 * Differs from memset() in that it fills with a uint32_t instead557 * of a byte. Remember that @count is the number of uint32_ts to558 * store, not the number of bytes.559 */560void *memset32(uint32_t *s, uint32_t v, size_t count)561{562 uint32_t *xs = s;563 564 while (count--)565 *xs++ = v;566 return s;567}568EXPORT_SYMBOL(memset32);569#endif570 571#ifndef __HAVE_ARCH_MEMSET64572/**573 * memset64() - Fill a memory area with a uint64_t574 * @s: Pointer to the start of the area.575 * @v: The value to fill the area with576 * @count: The number of values to store577 *578 * Differs from memset() in that it fills with a uint64_t instead579 * of a byte. Remember that @count is the number of uint64_ts to580 * store, not the number of bytes.581 */582void *memset64(uint64_t *s, uint64_t v, size_t count)583{584 uint64_t *xs = s;585 586 while (count--)587 *xs++ = v;588 return s;589}590EXPORT_SYMBOL(memset64);591#endif592 593#ifndef __HAVE_ARCH_MEMCPY594/**595 * memcpy - Copy one area of memory to another596 * @dest: Where to copy to597 * @src: Where to copy from598 * @count: The size of the area.599 *600 * You should not use this function to access IO space, use memcpy_toio()601 * or memcpy_fromio() instead.602 */603void *memcpy(void *dest, const void *src, size_t count)604{605 char *tmp = dest;606 const char *s = src;607 608 while (count--)609 *tmp++ = *s++;610 return dest;611}612EXPORT_SYMBOL(memcpy);613#endif614 615#ifndef __HAVE_ARCH_MEMMOVE616/**617 * memmove - Copy one area of memory to another618 * @dest: Where to copy to619 * @src: Where to copy from620 * @count: The size of the area.621 *622 * Unlike memcpy(), memmove() copes with overlapping areas.623 */624void *memmove(void *dest, const void *src, size_t count)625{626 char *tmp;627 const char *s;628 629 if (dest <= src) {630 tmp = dest;631 s = src;632 while (count--)633 *tmp++ = *s++;634 } else {635 tmp = dest;636 tmp += count;637 s = src;638 s += count;639 while (count--)640 *--tmp = *--s;641 }642 return dest;643}644EXPORT_SYMBOL(memmove);645#endif646 647#ifndef __HAVE_ARCH_MEMCMP648/**649 * memcmp - Compare two areas of memory650 * @cs: One area of memory651 * @ct: Another area of memory652 * @count: The size of the area.653 */654#undef memcmp655__visible int memcmp(const void *cs, const void *ct, size_t count)656{657 const unsigned char *su1, *su2;658 int res = 0;659 660#ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS661 if (count >= sizeof(unsigned long)) {662 const unsigned long *u1 = cs;663 const unsigned long *u2 = ct;664 do {665 if (get_unaligned(u1) != get_unaligned(u2))666 break;667 u1++;668 u2++;669 count -= sizeof(unsigned long);670 } while (count >= sizeof(unsigned long));671 cs = u1;672 ct = u2;673 }674#endif675 for (su1 = cs, su2 = ct; 0 < count; ++su1, ++su2, count--)676 if ((res = *su1 - *su2) != 0)677 break;678 return res;679}680EXPORT_SYMBOL(memcmp);681#endif682 683#ifndef __HAVE_ARCH_BCMP684/**685 * bcmp - returns 0 if and only if the buffers have identical contents.686 * @a: pointer to first buffer.687 * @b: pointer to second buffer.688 * @len: size of buffers.689 *690 * The sign or magnitude of a non-zero return value has no particular691 * meaning, and architectures may implement their own more efficient bcmp(). So692 * while this particular implementation is a simple (tail) call to memcmp, do693 * not rely on anything but whether the return value is zero or non-zero.694 */695int bcmp(const void *a, const void *b, size_t len)696{697 return memcmp(a, b, len);698}699EXPORT_SYMBOL(bcmp);700#endif701 702#ifndef __HAVE_ARCH_MEMSCAN703/**704 * memscan - Find a character in an area of memory.705 * @addr: The memory area706 * @c: The byte to search for707 * @size: The size of the area.708 *709 * returns the address of the first occurrence of @c, or 1 byte past710 * the area if @c is not found711 */712void *memscan(void *addr, int c, size_t size)713{714 unsigned char *p = addr;715 716 while (size) {717 if (*p == (unsigned char)c)718 return (void *)p;719 p++;720 size--;721 }722 return (void *)p;723}724EXPORT_SYMBOL(memscan);725#endif726 727#ifndef __HAVE_ARCH_STRSTR728/**729 * strstr - Find the first substring in a %NUL terminated string730 * @s1: The string to be searched731 * @s2: The string to search for732 */733char *strstr(const char *s1, const char *s2)734{735 size_t l1, l2;736 737 l2 = strlen(s2);738 if (!l2)739 return (char *)s1;740 l1 = strlen(s1);741 while (l1 >= l2) {742 l1--;743 if (!memcmp(s1, s2, l2))744 return (char *)s1;745 s1++;746 }747 return NULL;748}749EXPORT_SYMBOL(strstr);750#endif751 752#ifndef __HAVE_ARCH_STRNSTR753/**754 * strnstr - Find the first substring in a length-limited string755 * @s1: The string to be searched756 * @s2: The string to search for757 * @len: the maximum number of characters to search758 */759char *strnstr(const char *s1, const char *s2, size_t len)760{761 size_t l2;762 763 l2 = strlen(s2);764 if (!l2)765 return (char *)s1;766 while (len >= l2) {767 len--;768 if (!memcmp(s1, s2, l2))769 return (char *)s1;770 s1++;771 }772 return NULL;773}774EXPORT_SYMBOL(strnstr);775#endif776 777#ifndef __HAVE_ARCH_MEMCHR778/**779 * memchr - Find a character in an area of memory.780 * @s: The memory area781 * @c: The byte to search for782 * @n: The size of the area.783 *784 * returns the address of the first occurrence of @c, or %NULL785 * if @c is not found786 */787void *memchr(const void *s, int c, size_t n)788{789 const unsigned char *p = s;790 while (n-- != 0) {791 if ((unsigned char)c == *p++) {792 return (void *)(p - 1);793 }794 }795 return NULL;796}797EXPORT_SYMBOL(memchr);798#endif799 800static void *check_bytes8(const u8 *start, u8 value, unsigned int bytes)801{802 while (bytes) {803 if (*start != value)804 return (void *)start;805 start++;806 bytes--;807 }808 return NULL;809}810 811/**812 * memchr_inv - Find an unmatching character in an area of memory.813 * @start: The memory area814 * @c: Find a character other than c815 * @bytes: The size of the area.816 *817 * returns the address of the first character other than @c, or %NULL818 * if the whole buffer contains just @c.819 */820void *memchr_inv(const void *start, int c, size_t bytes)821{822 u8 value = c;823 u64 value64;824 unsigned int words, prefix;825 826 if (bytes <= 16)827 return check_bytes8(start, value, bytes);828 829 value64 = value;830#if defined(CONFIG_ARCH_HAS_FAST_MULTIPLIER) && BITS_PER_LONG == 64831 value64 *= 0x0101010101010101ULL;832#elif defined(CONFIG_ARCH_HAS_FAST_MULTIPLIER)833 value64 *= 0x01010101;834 value64 |= value64 << 32;835#else836 value64 |= value64 << 8;837 value64 |= value64 << 16;838 value64 |= value64 << 32;839#endif840 841 prefix = (unsigned long)start % 8;842 if (prefix) {843 u8 *r;844 845 prefix = 8 - prefix;846 r = check_bytes8(start, value, prefix);847 if (r)848 return r;849 start += prefix;850 bytes -= prefix;851 }852 853 words = bytes / 8;854 855 while (words) {856 if (*(u64 *)start != value64)857 return check_bytes8(start, value, 8);858 start += 8;859 words--;860 }861 862 return check_bytes8(start, value, bytes % 8);863}864EXPORT_SYMBOL(memchr_inv);865