projects/linux·blob v7.2-rc5 · refs

864 lines · 18.1 KiB

1// SPDX-License-Identifier: GPL-2.0
2/*
3 * linux/lib/string.c
4 *
5 * Copyright (C) 1991, 1992 Linus Torvalds
6 */
7
8/*
9 * This file should be used only for "library" routines that may have
10 * alternative implementations on specific architectures (generally
11 * 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_FORTIFY
18#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/minmax.h>
25#include <linux/stddef.h>
26#include <linux/string.h>
27#include <linux/types.h>
28
29#include <asm/page.h>
30#include <asm/rwonce.h>
31#include <linux/unaligned.h>
32#include <asm/word-at-a-time.h>
33
34#ifndef __HAVE_ARCH_STRNCASECMP
35/**
36 * strncasecmp - Case insensitive, length-limited string comparison
37 * @s1: One string
38 * @s2: The other string
39 * @len: the maximum number of characters to compare
40 */
41int strncasecmp(const char *s1, const char *s2, size_t len)
42{
43 /* Yes, Virginia, it had better be unsigned */
44 unsigned char c1, c2;
45
46 if (!len)
47 return 0;
48
49 do {
50 c1 = *s1++;
51 c2 = *s2++;
52 if (!c1 || !c2)
53 break;
54 if (c1 == c2)
55 continue;
56 c1 = tolower(c1);
57 c2 = tolower(c2);
58 if (c1 != c2)
59 break;
60 } while (--len);
61 return (int)c1 - (int)c2;
62}
63EXPORT_SYMBOL(strncasecmp);
64#endif
65
66#ifndef __HAVE_ARCH_STRCASECMP
67int strcasecmp(const char *s1, const char *s2)
68{
69 int c1, c2;
70
71 do {
72 c1 = tolower(*s1++);
73 c2 = tolower(*s2++);
74 } while (c1 == c2 && c1 != 0);
75 return c1 - c2;
76}
77EXPORT_SYMBOL(strcasecmp);
78#endif
79
80#ifndef __HAVE_ARCH_STRCPY
81char *strcpy(char *dest, const char *src)
82{
83 char *tmp = dest;
84
85 while ((*dest++ = *src++) != '\0')
86 /* nothing */;
87 return tmp;
88}
89EXPORT_SYMBOL(strcpy);
90#endif
91
92#ifdef __BIG_ENDIAN
93# define ALLBUTLAST_BYTE_MASK (~255ul)
94#else
95# define ALLBUTLAST_BYTE_MASK (~0ul >> 8)
96#endif
97
98ssize_t sized_strscpy(char *dest, const char *src, size_t count)
99{
100 const struct word_at_a_time constants = WORD_AT_A_TIME_CONSTANTS;
101 size_t max = count;
102 long res = 0;
103
104 if (count == 0 || WARN_ON_ONCE(count > INT_MAX))
105 return -E2BIG;
106
107#ifndef CONFIG_DCACHE_WORD_ACCESS
108#ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
109 /*
110 * If src is unaligned, don't cross a page boundary,
111 * since we don't know if the next page is mapped.
112 */
113 if ((long)src & (sizeof(long) - 1))
114 max = min(PAGE_SIZE - ((long)src & (PAGE_SIZE - 1)), max);
115#else
116 /* If src or dest is unaligned, don't do word-at-a-time. */
117 if (((long) dest | (long) src) & (sizeof(long) - 1))
118 max = 0;
119#endif
120#endif
121
122 /*
123 * load_unaligned_zeropad() or read_word_at_a_time() below may read
124 * uninitialized bytes after the trailing zero and use them in
125 * comparisons. Disable this optimization under KMSAN to prevent
126 * false positive reports.
127 */
128 if (IS_ENABLED(CONFIG_KMSAN))
129 max = 0;
130
131 while (max >= sizeof(unsigned long)) {
132 unsigned long c, data;
133
134#ifdef CONFIG_DCACHE_WORD_ACCESS
135 c = load_unaligned_zeropad(src+res);
136#else
137 c = read_word_at_a_time(src+res);
138#endif
139 if (has_zero(c, &data, &constants)) {
140 data = prep_zero_mask(c, data, &constants);
141 data = create_zero_mask(data);
142 *(unsigned long *)(dest+res) = c & zero_bytemask(data);
143 return res + find_zero(data);
144 }
145 count -= sizeof(unsigned long);
146 if (unlikely(!count)) {
147 c &= ALLBUTLAST_BYTE_MASK;
148 *(unsigned long *)(dest+res) = c;
149 return -E2BIG;
150 }
151 *(unsigned long *)(dest+res) = c;
152 res += sizeof(unsigned long);
153 max -= sizeof(unsigned long);
154 }
155
156 while (count > 1) {
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 /* Force NUL-termination. */
168 dest[res] = '\0';
169
170 /* Return E2BIG if the source didn't stop */
171 return src[res] ? -E2BIG : res;
172}
173EXPORT_SYMBOL(sized_strscpy);
174
175/**
176 * stpcpy - copy a string from src to dest returning a pointer to the new end
177 * of dest, including src's %NUL-terminator. May overrun dest.
178 * @dest: pointer to end of string being copied into. Must be large enough
179 * to receive copy.
180 * @src: pointer to the beginning of string being copied from. Must not overlap
181 * dest.
182 *
183 * stpcpy differs from strcpy in a key way: the return value is a pointer
184 * to the new %NUL-terminating character in @dest. (For strcpy, the return
185 * value is a pointer to the start of @dest). This interface is considered
186 * unsafe as it doesn't perform bounds checking of the inputs. As such it's
187 * not recommended for usage. Instead, its definition is provided in case
188 * 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_STRCAT
200char *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#endif
212
213#ifndef __HAVE_ARCH_STRNCAT
214char *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#endif
232
233#ifndef __HAVE_ARCH_STRLCAT
234size_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#endif
253
254#ifndef __HAVE_ARCH_STRCMP
255/**
256 * strcmp - Compare two strings
257 * @cs: One string
258 * @ct: Another string
259 */
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#endif
276
277#ifndef __HAVE_ARCH_STRNCMP
278/**
279 * strncmp - Compare two length-limited strings
280 * @cs: One string
281 * @ct: Another string
282 * @count: The maximum number of bytes to compare
283 */
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#endif
301
302#ifndef __HAVE_ARCH_STRCHR
303/**
304 * strchr - Find the first occurrence of a character in a string
305 * @s: The string to be searched
306 * @c: The character to search for
307 *
308 * Note that the %NUL-terminator is considered part of the string, and can
309 * 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#endif
320
321#ifndef __HAVE_ARCH_STRCHRNUL
322/**
323 * strchrnul - Find and return a character in a string, or end of string
324 * @s: The string to be searched
325 * @c: The character to search for
326 *
327 * Returns pointer to first occurrence of 'c' in s. If c is not found, then
328 * 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#endif
338
339/**
340 * strnchrnul - Find and return a character in a length limited string,
341 * or end of string
342 * @s: The string to be searched
343 * @count: The number of characters to be searched
344 * @c: The character to search for
345 *
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_STRRCHR
357/**
358 * strrchr - Find the last occurrence of a character in a string
359 * @s: The string to be searched
360 * @c: The character to search for
361 */
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#endif
373
374#ifndef __HAVE_ARCH_STRNCHR
375/**
376 * strnchr - Find a character in a length limited string
377 * @s: The string to be searched
378 * @count: The number of characters to be searched
379 * @c: The character to search for
380 *
381 * Note that the %NUL-terminator is considered part of the string, and can
382 * 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#endif
396
397#ifndef __HAVE_ARCH_STRLEN
398size_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#endif
408
409#ifndef __HAVE_ARCH_STRNLEN
410size_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#endif
420
421#ifndef __HAVE_ARCH_STRSPN
422/**
423 * strspn - Calculate the length of the initial substring of @s which only contain letters in @accept
424 * @s: The string to be searched
425 * @accept: The string to search for
426 */
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#endif
439
440#ifndef __HAVE_ARCH_STRCSPN
441/**
442 * strcspn - Calculate the length of the initial substring of @s which does not contain letters in @reject
443 * @s: The string to be searched
444 * @reject: The string to avoid
445 */
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#endif
458
459#ifndef __HAVE_ARCH_STRPBRK
460/**
461 * strpbrk - Find the first occurrence of a set of characters
462 * @cs: The string to be searched
463 * @ct: The characters to search for
464 */
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#endif
477
478#ifndef __HAVE_ARCH_STRSEP
479/**
480 * strsep - Split a string into tokens
481 * @s: The string to be searched
482 * @ct: The characters to search for
483 *
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 function
487 * 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#endif
506
507#ifndef __HAVE_ARCH_MEMSET
508/**
509 * memset - Fill a region of memory with the given value
510 * @s: Pointer to the start of the area.
511 * @c: The byte to fill the area with
512 * @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#endif
526
527#ifndef __HAVE_ARCH_MEMSET16
528/**
529 * memset16() - Fill a memory area with a uint16_t
530 * @s: Pointer to the start of the area.
531 * @v: The value to fill the area with
532 * @count: The number of values to store
533 *
534 * Differs from memset() in that it fills with a uint16_t instead
535 * of a byte. Remember that @count is the number of uint16_ts to
536 * 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#endif
548
549#ifndef __HAVE_ARCH_MEMSET32
550/**
551 * memset32() - Fill a memory area with a uint32_t
552 * @s: Pointer to the start of the area.
553 * @v: The value to fill the area with
554 * @count: The number of values to store
555 *
556 * Differs from memset() in that it fills with a uint32_t instead
557 * of a byte. Remember that @count is the number of uint32_ts to
558 * 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#endif
570
571#ifndef __HAVE_ARCH_MEMSET64
572/**
573 * memset64() - Fill a memory area with a uint64_t
574 * @s: Pointer to the start of the area.
575 * @v: The value to fill the area with
576 * @count: The number of values to store
577 *
578 * Differs from memset() in that it fills with a uint64_t instead
579 * of a byte. Remember that @count is the number of uint64_ts to
580 * 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#endif
592
593#ifndef __HAVE_ARCH_MEMCPY
594/**
595 * memcpy - Copy one area of memory to another
596 * @dest: Where to copy to
597 * @src: Where to copy from
598 * @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#endif
614
615#ifndef __HAVE_ARCH_MEMMOVE
616/**
617 * memmove - Copy one area of memory to another
618 * @dest: Where to copy to
619 * @src: Where to copy from
620 * @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#endif
646
647#ifndef __HAVE_ARCH_MEMCMP
648/**
649 * memcmp - Compare two areas of memory
650 * @cs: One area of memory
651 * @ct: Another area of memory
652 * @count: The size of the area.
653 */
654#undef memcmp
655__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_ACCESS
661 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#endif
675 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#endif
682
683#ifndef __HAVE_ARCH_BCMP
684/**
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 particular
691 * meaning, and architectures may implement their own more efficient bcmp(). So
692 * while this particular implementation is a simple (tail) call to memcmp, do
693 * 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#endif
701
702#ifndef __HAVE_ARCH_MEMSCAN
703/**
704 * memscan - Find a character in an area of memory.
705 * @addr: The memory area
706 * @c: The byte to search for
707 * @size: The size of the area.
708 *
709 * returns the address of the first occurrence of @c, or 1 byte past
710 * the area if @c is not found
711 */
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#endif
726
727#ifndef __HAVE_ARCH_STRSTR
728/**
729 * strstr - Find the first substring in a %NUL terminated string
730 * @s1: The string to be searched
731 * @s2: The string to search for
732 */
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#endif
751
752#ifndef __HAVE_ARCH_STRNSTR
753/**
754 * strnstr - Find the first substring in a length-limited string
755 * @s1: The string to be searched
756 * @s2: The string to search for
757 * @len: the maximum number of characters to search
758 */
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#endif
776
777#ifndef __HAVE_ARCH_MEMCHR
778/**
779 * memchr - Find a character in an area of memory.
780 * @s: The memory area
781 * @c: The byte to search for
782 * @n: The size of the area.
783 *
784 * returns the address of the first occurrence of @c, or %NULL
785 * if @c is not found
786 */
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#endif
799
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 area
814 * @c: Find a character other than c
815 * @bytes: The size of the area.
816 *
817 * returns the address of the first character other than @c, or %NULL
818 * 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 == 64
831 value64 *= 0x0101010101010101ULL;
832#elif defined(CONFIG_ARCH_HAS_FAST_MULTIPLIER)
833 value64 *= 0x01010101;
834 value64 |= value64 << 32;
835#else
836 value64 |= value64 << 8;
837 value64 |= value64 << 16;
838 value64 |= value64 << 32;
839#endif
840
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);