Commit | Line | Data |
---|---|---|
3c9f3681 JB |
1 | /* |
2 | * Helpers for formatting and printing strings | |
3 | * | |
4 | * Copyright 31 August 2008 James Bottomley | |
16c7fa05 | 5 | * Copyright (C) 2013, Intel Corporation |
3c9f3681 | 6 | */ |
b9f28d86 | 7 | #include <linux/bug.h> |
3c9f3681 JB |
8 | #include <linux/kernel.h> |
9 | #include <linux/math64.h> | |
8bc3bcc9 | 10 | #include <linux/export.h> |
16c7fa05 | 11 | #include <linux/ctype.h> |
c8250381 | 12 | #include <linux/errno.h> |
0d044328 | 13 | #include <linux/mm.h> |
b53f27e4 | 14 | #include <linux/slab.h> |
c8250381 | 15 | #include <linux/string.h> |
3c9f3681 JB |
16 | #include <linux/string_helpers.h> |
17 | ||
18 | /** | |
19 | * string_get_size - get the size in the specified units | |
b9f28d86 JB |
20 | * @size: The size to be converted in blocks |
21 | * @blk_size: Size of the block (use 1 for size in bytes) | |
3c9f3681 JB |
22 | * @units: units to use (powers of 1000 or 1024) |
23 | * @buf: buffer to format to | |
24 | * @len: length of buffer | |
25 | * | |
26 | * This function returns a string formatted to 3 significant figures | |
d1214c65 RV |
27 | * giving the size in the required units. @buf should have room for |
28 | * at least 9 bytes and will always be zero terminated. | |
3c9f3681 JB |
29 | * |
30 | */ | |
b9f28d86 | 31 | void string_get_size(u64 size, u64 blk_size, const enum string_size_units units, |
d1214c65 | 32 | char *buf, int len) |
3c9f3681 | 33 | { |
142cda5d | 34 | static const char *const units_10[] = { |
b9f28d86 | 35 | "B", "kB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB" |
142cda5d MK |
36 | }; |
37 | static const char *const units_2[] = { | |
b9f28d86 | 38 | "B", "KiB", "MiB", "GiB", "TiB", "PiB", "EiB", "ZiB", "YiB" |
142cda5d MK |
39 | }; |
40 | static const char *const *const units_str[] = { | |
41 | [STRING_UNITS_10] = units_10, | |
3c9f3681 JB |
42 | [STRING_UNITS_2] = units_2, |
43 | }; | |
68aecfb9 | 44 | static const unsigned int divisor[] = { |
3c9f3681 JB |
45 | [STRING_UNITS_10] = 1000, |
46 | [STRING_UNITS_2] = 1024, | |
47 | }; | |
564b026f JB |
48 | static const unsigned int rounding[] = { 500, 50, 5 }; |
49 | int i = 0, j; | |
50 | u32 remainder = 0, sf_cap; | |
3c9f3681 | 51 | char tmp[8]; |
b9f28d86 | 52 | const char *unit; |
3c9f3681 JB |
53 | |
54 | tmp[0] = '\0'; | |
564b026f JB |
55 | |
56 | if (blk_size == 0) | |
57 | size = 0; | |
58 | if (size == 0) | |
b9f28d86 | 59 | goto out; |
3c9f3681 | 60 | |
564b026f JB |
61 | /* This is Napier's algorithm. Reduce the original block size to |
62 | * | |
63 | * coefficient * divisor[units]^i | |
64 | * | |
65 | * we do the reduction so both coefficients are just under 32 bits so | |
66 | * that multiplying them together won't overflow 64 bits and we keep | |
67 | * as much precision as possible in the numbers. | |
68 | * | |
69 | * Note: it's safe to throw away the remainders here because all the | |
70 | * precision is in the coefficients. | |
71 | */ | |
72 | while (blk_size >> 32) { | |
73 | do_div(blk_size, divisor[units]); | |
b9f28d86 JB |
74 | i++; |
75 | } | |
76 | ||
564b026f JB |
77 | while (size >> 32) { |
78 | do_div(size, divisor[units]); | |
b9f28d86 | 79 | i++; |
b9f28d86 JB |
80 | } |
81 | ||
564b026f JB |
82 | /* now perform the actual multiplication keeping i as the sum of the |
83 | * two logarithms */ | |
b9f28d86 | 84 | size *= blk_size; |
3c9f3681 | 85 | |
564b026f | 86 | /* and logarithmically reduce it until it's just under the divisor */ |
b9f28d86 JB |
87 | while (size >= divisor[units]) { |
88 | remainder = do_div(size, divisor[units]); | |
89 | i++; | |
3c9f3681 JB |
90 | } |
91 | ||
564b026f JB |
92 | /* work out in j how many digits of precision we need from the |
93 | * remainder */ | |
b9f28d86 JB |
94 | sf_cap = size; |
95 | for (j = 0; sf_cap*10 < 1000; j++) | |
96 | sf_cap *= 10; | |
97 | ||
564b026f JB |
98 | if (units == STRING_UNITS_2) { |
99 | /* express the remainder as a decimal. It's currently the | |
100 | * numerator of a fraction whose denominator is | |
101 | * divisor[units], which is 1 << 10 for STRING_UNITS_2 */ | |
b9f28d86 | 102 | remainder *= 1000; |
564b026f JB |
103 | remainder >>= 10; |
104 | } | |
105 | ||
106 | /* add a 5 to the digit below what will be printed to ensure | |
107 | * an arithmetical round up and carry it through to size */ | |
108 | remainder += rounding[j]; | |
109 | if (remainder >= 1000) { | |
110 | remainder -= 1000; | |
111 | size += 1; | |
112 | } | |
113 | ||
114 | if (j) { | |
b9f28d86 JB |
115 | snprintf(tmp, sizeof(tmp), ".%03u", remainder); |
116 | tmp[j+1] = '\0'; | |
117 | } | |
118 | ||
119 | out: | |
120 | if (i >= ARRAY_SIZE(units_2)) | |
121 | unit = "UNK"; | |
122 | else | |
123 | unit = units_str[units][i]; | |
124 | ||
84b9fbed | 125 | snprintf(buf, len, "%u%s %s", (u32)size, |
b9f28d86 | 126 | tmp, unit); |
3c9f3681 JB |
127 | } |
128 | EXPORT_SYMBOL(string_get_size); | |
16c7fa05 AS |
129 | |
130 | static bool unescape_space(char **src, char **dst) | |
131 | { | |
132 | char *p = *dst, *q = *src; | |
133 | ||
134 | switch (*q) { | |
135 | case 'n': | |
136 | *p = '\n'; | |
137 | break; | |
138 | case 'r': | |
139 | *p = '\r'; | |
140 | break; | |
141 | case 't': | |
142 | *p = '\t'; | |
143 | break; | |
144 | case 'v': | |
145 | *p = '\v'; | |
146 | break; | |
147 | case 'f': | |
148 | *p = '\f'; | |
149 | break; | |
150 | default: | |
151 | return false; | |
152 | } | |
153 | *dst += 1; | |
154 | *src += 1; | |
155 | return true; | |
156 | } | |
157 | ||
158 | static bool unescape_octal(char **src, char **dst) | |
159 | { | |
160 | char *p = *dst, *q = *src; | |
161 | u8 num; | |
162 | ||
163 | if (isodigit(*q) == 0) | |
164 | return false; | |
165 | ||
166 | num = (*q++) & 7; | |
167 | while (num < 32 && isodigit(*q) && (q - *src < 3)) { | |
168 | num <<= 3; | |
169 | num += (*q++) & 7; | |
170 | } | |
171 | *p = num; | |
172 | *dst += 1; | |
173 | *src = q; | |
174 | return true; | |
175 | } | |
176 | ||
177 | static bool unescape_hex(char **src, char **dst) | |
178 | { | |
179 | char *p = *dst, *q = *src; | |
180 | int digit; | |
181 | u8 num; | |
182 | ||
183 | if (*q++ != 'x') | |
184 | return false; | |
185 | ||
186 | num = digit = hex_to_bin(*q++); | |
187 | if (digit < 0) | |
188 | return false; | |
189 | ||
190 | digit = hex_to_bin(*q); | |
191 | if (digit >= 0) { | |
192 | q++; | |
193 | num = (num << 4) | digit; | |
194 | } | |
195 | *p = num; | |
196 | *dst += 1; | |
197 | *src = q; | |
198 | return true; | |
199 | } | |
200 | ||
201 | static bool unescape_special(char **src, char **dst) | |
202 | { | |
203 | char *p = *dst, *q = *src; | |
204 | ||
205 | switch (*q) { | |
206 | case '\"': | |
207 | *p = '\"'; | |
208 | break; | |
209 | case '\\': | |
210 | *p = '\\'; | |
211 | break; | |
212 | case 'a': | |
213 | *p = '\a'; | |
214 | break; | |
215 | case 'e': | |
216 | *p = '\e'; | |
217 | break; | |
218 | default: | |
219 | return false; | |
220 | } | |
221 | *dst += 1; | |
222 | *src += 1; | |
223 | return true; | |
224 | } | |
225 | ||
d295634e AS |
226 | /** |
227 | * string_unescape - unquote characters in the given string | |
228 | * @src: source buffer (escaped) | |
229 | * @dst: destination buffer (unescaped) | |
230 | * @size: size of the destination buffer (0 to unlimit) | |
231 | * @flags: combination of the flags (bitwise OR): | |
232 | * %UNESCAPE_SPACE: | |
233 | * '\f' - form feed | |
234 | * '\n' - new line | |
235 | * '\r' - carriage return | |
236 | * '\t' - horizontal tab | |
237 | * '\v' - vertical tab | |
238 | * %UNESCAPE_OCTAL: | |
239 | * '\NNN' - byte with octal value NNN (1 to 3 digits) | |
240 | * %UNESCAPE_HEX: | |
241 | * '\xHH' - byte with hexadecimal value HH (1 to 2 digits) | |
242 | * %UNESCAPE_SPECIAL: | |
243 | * '\"' - double quote | |
244 | * '\\' - backslash | |
245 | * '\a' - alert (BEL) | |
246 | * '\e' - escape | |
247 | * %UNESCAPE_ANY: | |
248 | * all previous together | |
249 | * | |
250 | * Description: | |
251 | * The function unquotes characters in the given string. | |
252 | * | |
253 | * Because the size of the output will be the same as or less than the size of | |
254 | * the input, the transformation may be performed in place. | |
255 | * | |
256 | * Caller must provide valid source and destination pointers. Be aware that | |
257 | * destination buffer will always be NULL-terminated. Source string must be | |
258 | * NULL-terminated as well. | |
259 | * | |
260 | * Return: | |
261 | * The amount of the characters processed to the destination buffer excluding | |
262 | * trailing '\0' is returned. | |
263 | */ | |
16c7fa05 AS |
264 | int string_unescape(char *src, char *dst, size_t size, unsigned int flags) |
265 | { | |
266 | char *out = dst; | |
267 | ||
268 | while (*src && --size) { | |
269 | if (src[0] == '\\' && src[1] != '\0' && size > 1) { | |
270 | src++; | |
271 | size--; | |
272 | ||
273 | if (flags & UNESCAPE_SPACE && | |
274 | unescape_space(&src, &out)) | |
275 | continue; | |
276 | ||
277 | if (flags & UNESCAPE_OCTAL && | |
278 | unescape_octal(&src, &out)) | |
279 | continue; | |
280 | ||
281 | if (flags & UNESCAPE_HEX && | |
282 | unescape_hex(&src, &out)) | |
283 | continue; | |
284 | ||
285 | if (flags & UNESCAPE_SPECIAL && | |
286 | unescape_special(&src, &out)) | |
287 | continue; | |
288 | ||
289 | *out++ = '\\'; | |
290 | } | |
291 | *out++ = *src++; | |
292 | } | |
293 | *out = '\0'; | |
294 | ||
295 | return out - dst; | |
296 | } | |
297 | EXPORT_SYMBOL(string_unescape); | |
c8250381 | 298 | |
3aeddc7d | 299 | static bool escape_passthrough(unsigned char c, char **dst, char *end) |
c8250381 AS |
300 | { |
301 | char *out = *dst; | |
302 | ||
3aeddc7d RV |
303 | if (out < end) |
304 | *out = c; | |
305 | *dst = out + 1; | |
306 | return true; | |
c8250381 AS |
307 | } |
308 | ||
3aeddc7d | 309 | static bool escape_space(unsigned char c, char **dst, char *end) |
c8250381 AS |
310 | { |
311 | char *out = *dst; | |
312 | unsigned char to; | |
313 | ||
c8250381 AS |
314 | switch (c) { |
315 | case '\n': | |
316 | to = 'n'; | |
317 | break; | |
318 | case '\r': | |
319 | to = 'r'; | |
320 | break; | |
321 | case '\t': | |
322 | to = 't'; | |
323 | break; | |
324 | case '\v': | |
325 | to = 'v'; | |
326 | break; | |
327 | case '\f': | |
328 | to = 'f'; | |
329 | break; | |
330 | default: | |
3aeddc7d | 331 | return false; |
c8250381 AS |
332 | } |
333 | ||
3aeddc7d RV |
334 | if (out < end) |
335 | *out = '\\'; | |
336 | ++out; | |
337 | if (out < end) | |
338 | *out = to; | |
339 | ++out; | |
c8250381 | 340 | |
3aeddc7d RV |
341 | *dst = out; |
342 | return true; | |
c8250381 AS |
343 | } |
344 | ||
3aeddc7d | 345 | static bool escape_special(unsigned char c, char **dst, char *end) |
c8250381 AS |
346 | { |
347 | char *out = *dst; | |
348 | unsigned char to; | |
349 | ||
c8250381 AS |
350 | switch (c) { |
351 | case '\\': | |
352 | to = '\\'; | |
353 | break; | |
354 | case '\a': | |
355 | to = 'a'; | |
356 | break; | |
357 | case '\e': | |
358 | to = 'e'; | |
359 | break; | |
360 | default: | |
3aeddc7d | 361 | return false; |
c8250381 AS |
362 | } |
363 | ||
3aeddc7d RV |
364 | if (out < end) |
365 | *out = '\\'; | |
366 | ++out; | |
367 | if (out < end) | |
368 | *out = to; | |
369 | ++out; | |
c8250381 | 370 | |
3aeddc7d RV |
371 | *dst = out; |
372 | return true; | |
c8250381 AS |
373 | } |
374 | ||
3aeddc7d | 375 | static bool escape_null(unsigned char c, char **dst, char *end) |
c8250381 AS |
376 | { |
377 | char *out = *dst; | |
378 | ||
c8250381 | 379 | if (c) |
3aeddc7d | 380 | return false; |
c8250381 | 381 | |
3aeddc7d RV |
382 | if (out < end) |
383 | *out = '\\'; | |
384 | ++out; | |
385 | if (out < end) | |
386 | *out = '0'; | |
387 | ++out; | |
c8250381 | 388 | |
3aeddc7d RV |
389 | *dst = out; |
390 | return true; | |
c8250381 AS |
391 | } |
392 | ||
3aeddc7d | 393 | static bool escape_octal(unsigned char c, char **dst, char *end) |
c8250381 AS |
394 | { |
395 | char *out = *dst; | |
396 | ||
3aeddc7d RV |
397 | if (out < end) |
398 | *out = '\\'; | |
399 | ++out; | |
400 | if (out < end) | |
401 | *out = ((c >> 6) & 0x07) + '0'; | |
402 | ++out; | |
403 | if (out < end) | |
404 | *out = ((c >> 3) & 0x07) + '0'; | |
405 | ++out; | |
406 | if (out < end) | |
407 | *out = ((c >> 0) & 0x07) + '0'; | |
408 | ++out; | |
c8250381 AS |
409 | |
410 | *dst = out; | |
3aeddc7d | 411 | return true; |
c8250381 AS |
412 | } |
413 | ||
3aeddc7d | 414 | static bool escape_hex(unsigned char c, char **dst, char *end) |
c8250381 AS |
415 | { |
416 | char *out = *dst; | |
417 | ||
3aeddc7d RV |
418 | if (out < end) |
419 | *out = '\\'; | |
420 | ++out; | |
421 | if (out < end) | |
422 | *out = 'x'; | |
423 | ++out; | |
424 | if (out < end) | |
425 | *out = hex_asc_hi(c); | |
426 | ++out; | |
427 | if (out < end) | |
428 | *out = hex_asc_lo(c); | |
429 | ++out; | |
c8250381 AS |
430 | |
431 | *dst = out; | |
3aeddc7d | 432 | return true; |
c8250381 AS |
433 | } |
434 | ||
435 | /** | |
436 | * string_escape_mem - quote characters in the given memory buffer | |
437 | * @src: source buffer (unescaped) | |
438 | * @isz: source buffer size | |
439 | * @dst: destination buffer (escaped) | |
440 | * @osz: destination buffer size | |
441 | * @flags: combination of the flags (bitwise OR): | |
d89a3f73 | 442 | * %ESCAPE_SPACE: (special white space, not space itself) |
c8250381 AS |
443 | * '\f' - form feed |
444 | * '\n' - new line | |
445 | * '\r' - carriage return | |
446 | * '\t' - horizontal tab | |
447 | * '\v' - vertical tab | |
448 | * %ESCAPE_SPECIAL: | |
449 | * '\\' - backslash | |
450 | * '\a' - alert (BEL) | |
451 | * '\e' - escape | |
452 | * %ESCAPE_NULL: | |
453 | * '\0' - null | |
454 | * %ESCAPE_OCTAL: | |
455 | * '\NNN' - byte with octal value NNN (3 digits) | |
456 | * %ESCAPE_ANY: | |
457 | * all previous together | |
458 | * %ESCAPE_NP: | |
459 | * escape only non-printable characters (checked by isprint) | |
460 | * %ESCAPE_ANY_NP: | |
461 | * all previous together | |
462 | * %ESCAPE_HEX: | |
463 | * '\xHH' - byte with hexadecimal value HH (2 digits) | |
b40bdb7f KC |
464 | * @only: NULL-terminated string containing characters used to limit |
465 | * the selected escape class. If characters are included in @only | |
d89a3f73 KC |
466 | * that would not normally be escaped by the classes selected |
467 | * in @flags, they will be copied to @dst unescaped. | |
c8250381 AS |
468 | * |
469 | * Description: | |
470 | * The process of escaping byte buffer includes several parts. They are applied | |
471 | * in the following sequence. | |
472 | * 1. The character is matched to the printable class, if asked, and in | |
473 | * case of match it passes through to the output. | |
b40bdb7f | 474 | * 2. The character is not matched to the one from @only string and thus |
d89a3f73 | 475 | * must go as-is to the output. |
c8250381 AS |
476 | * 3. The character is checked if it falls into the class given by @flags. |
477 | * %ESCAPE_OCTAL and %ESCAPE_HEX are going last since they cover any | |
478 | * character. Note that they actually can't go together, otherwise | |
479 | * %ESCAPE_HEX will be ignored. | |
480 | * | |
481 | * Caller must provide valid source and destination pointers. Be aware that | |
482 | * destination buffer will not be NULL-terminated, thus caller have to append | |
483 | * it if needs. | |
484 | * | |
485 | * Return: | |
41416f23 RV |
486 | * The total size of the escaped output that would be generated for |
487 | * the given input and flags. To check whether the output was | |
488 | * truncated, compare the return value to osz. There is room left in | |
489 | * dst for a '\0' terminator if and only if ret < osz. | |
c8250381 | 490 | */ |
41416f23 | 491 | int string_escape_mem(const char *src, size_t isz, char *dst, size_t osz, |
b40bdb7f | 492 | unsigned int flags, const char *only) |
c8250381 | 493 | { |
41416f23 | 494 | char *p = dst; |
3aeddc7d | 495 | char *end = p + osz; |
b40bdb7f | 496 | bool is_dict = only && *only; |
c8250381 AS |
497 | |
498 | while (isz--) { | |
499 | unsigned char c = *src++; | |
500 | ||
501 | /* | |
502 | * Apply rules in the following sequence: | |
503 | * - the character is printable, when @flags has | |
504 | * %ESCAPE_NP bit set | |
b40bdb7f | 505 | * - the @only string is supplied and does not contain a |
c8250381 AS |
506 | * character under question |
507 | * - the character doesn't fall into a class of symbols | |
508 | * defined by given @flags | |
509 | * In these cases we just pass through a character to the | |
510 | * output buffer. | |
511 | */ | |
512 | if ((flags & ESCAPE_NP && isprint(c)) || | |
b40bdb7f | 513 | (is_dict && !strchr(only, c))) { |
c8250381 AS |
514 | /* do nothing */ |
515 | } else { | |
3aeddc7d RV |
516 | if (flags & ESCAPE_SPACE && escape_space(c, &p, end)) |
517 | continue; | |
518 | ||
519 | if (flags & ESCAPE_SPECIAL && escape_special(c, &p, end)) | |
520 | continue; | |
521 | ||
522 | if (flags & ESCAPE_NULL && escape_null(c, &p, end)) | |
523 | continue; | |
c8250381 AS |
524 | |
525 | /* ESCAPE_OCTAL and ESCAPE_HEX always go last */ | |
3aeddc7d | 526 | if (flags & ESCAPE_OCTAL && escape_octal(c, &p, end)) |
c8250381 | 527 | continue; |
3aeddc7d RV |
528 | |
529 | if (flags & ESCAPE_HEX && escape_hex(c, &p, end)) | |
c8250381 | 530 | continue; |
c8250381 AS |
531 | } |
532 | ||
3aeddc7d | 533 | escape_passthrough(c, &p, end); |
c8250381 AS |
534 | } |
535 | ||
41416f23 | 536 | return p - dst; |
c8250381 AS |
537 | } |
538 | EXPORT_SYMBOL(string_escape_mem); | |
b53f27e4 KC |
539 | |
540 | /* | |
541 | * Return an allocated string that has been escaped of special characters | |
542 | * and double quotes, making it safe to log in quotes. | |
543 | */ | |
544 | char *kstrdup_quotable(const char *src, gfp_t gfp) | |
545 | { | |
546 | size_t slen, dlen; | |
547 | char *dst; | |
548 | const int flags = ESCAPE_HEX; | |
549 | const char esc[] = "\f\n\r\t\v\a\e\\\""; | |
550 | ||
551 | if (!src) | |
552 | return NULL; | |
553 | slen = strlen(src); | |
554 | ||
555 | dlen = string_escape_mem(src, slen, NULL, 0, flags, esc); | |
556 | dst = kmalloc(dlen + 1, gfp); | |
557 | if (!dst) | |
558 | return NULL; | |
559 | ||
560 | WARN_ON(string_escape_mem(src, slen, dst, dlen, flags, esc) != dlen); | |
561 | dst[dlen] = '\0'; | |
562 | ||
563 | return dst; | |
564 | } | |
565 | EXPORT_SYMBOL_GPL(kstrdup_quotable); | |
0d044328 KC |
566 | |
567 | /* | |
568 | * Returns allocated NULL-terminated string containing process | |
569 | * command line, with inter-argument NULLs replaced with spaces, | |
570 | * and other special characters escaped. | |
571 | */ | |
572 | char *kstrdup_quotable_cmdline(struct task_struct *task, gfp_t gfp) | |
573 | { | |
574 | char *buffer, *quoted; | |
575 | int i, res; | |
576 | ||
577 | buffer = kmalloc(PAGE_SIZE, GFP_TEMPORARY); | |
578 | if (!buffer) | |
579 | return NULL; | |
580 | ||
581 | res = get_cmdline(task, buffer, PAGE_SIZE - 1); | |
582 | buffer[res] = '\0'; | |
583 | ||
584 | /* Collapse trailing NULLs, leave res pointing to last non-NULL. */ | |
585 | while (--res >= 0 && buffer[res] == '\0') | |
586 | ; | |
587 | ||
588 | /* Replace inter-argument NULLs. */ | |
589 | for (i = 0; i <= res; i++) | |
590 | if (buffer[i] == '\0') | |
591 | buffer[i] = ' '; | |
592 | ||
593 | /* Make sure result is printable. */ | |
594 | quoted = kstrdup_quotable(buffer, gfp); | |
595 | kfree(buffer); | |
596 | return quoted; | |
597 | } | |
598 | EXPORT_SYMBOL_GPL(kstrdup_quotable_cmdline); |