src/libudev-util.c

   1 /*
   2  * libudev - interface to udev device information
   3  *
   4  * Copyright (C) 2008-2011 Kay Sievers <kay.sievers@vrfy.org>
   5  *
   6  * This library is free software; you can redistribute it and/or
   7  * modify it under the terms of the GNU Lesser General Public
   8  * License as published by the Free Software Foundation; either
   9  * version 2.1 of the License, or (at your option) any later version.
  10  */
  11
  12 #include <stdio.h>
  13 #include <stdlib.h>
  14 #include <stddef.h>
  15 #include <unistd.h>
  16 #include <errno.h>
  17 #include <string.h>
  18 #include <dirent.h>
  19 #include <ctype.h>
  20 #include <fcntl.h>
  21 #include <time.h>
  22 #include <sys/stat.h>
  23
  24 #include "libudev.h"
  25 #include "libudev-private.h"
  26
  27 /**
  28  * SECTION:libudev-util
  29  * @short_description: utils
  30  */
  31
  32 ssize_t util_get_sys_core_link_value(struct udev *udev, const char *slink, const char *syspath, char *value, size_t size)
  33 {
  34         char path[UTIL_PATH_SIZE];
  35         char target[UTIL_PATH_SIZE];
  36         ssize_t len;
  37         const char *pos;
  38
  39         util_strscpyl(path, sizeof(path), syspath, "/", slink, NULL);
  40         len = readlink(path, target, sizeof(target));
  41         if (len <= 0 || len == (ssize_t)sizeof(target))
  42                 return -1;
  43         target[len] = '\0';
  44         pos = strrchr(target, '/');
  45         if (pos == NULL)
  46                 return -1;
  47         pos = &pos[1];
  48         dbg(udev, "resolved link to: '%s'\n", pos);
  49         return util_strscpy(value, size, pos);
  50 }
  51
  52 int util_resolve_sys_link(struct udev *udev, char *syspath, size_t size)
  53 {
  54         char link_target[UTIL_PATH_SIZE];
  55
  56         ssize_t len;
  57         int i;
  58         int back;
  59         char *base;
  60
  61         len = readlink(syspath, link_target, sizeof(link_target));
  62         if (len <= 0 || len == (ssize_t)sizeof(link_target))
  63                 return -1;
  64         link_target[len] = '\0';
  65         dbg(udev, "path link '%s' points to '%s'\n", syspath, link_target);
  66
  67         for (back = 0; strncmp(&link_target[back * 3], "../", 3) == 0; back++)
  68                 ;
  69         dbg(udev, "base '%s', tail '%s', back %i\n", syspath, &link_target[back * 3], back);
  70         for (i = 0; i <= back; i++) {
  71                 base = strrchr(syspath, '/');
  72                 if (base == NULL)
  73                         return -1;
  74                 base[0] = '\0';
  75         }
  76         dbg(udev, "after moving back '%s'\n", syspath);
  77         util_strscpyl(base, size - (base - syspath), "/", &link_target[back * 3], NULL);
  78         return 0;
  79 }
  80
  81 int util_log_priority(const char *priority)
  82 {
  83         char *endptr;
  84         int prio;
  85
  86         prio = strtol(priority, &endptr, 10);
  87         if (endptr[0] == '\0' || isspace(endptr[0]))
  88                 return prio;
  89         if (strncmp(priority, "err", 3) == 0)
  90                 return LOG_ERR;
  91         if (strncmp(priority, "info", 4) == 0)
  92                 return LOG_INFO;
  93         if (strncmp(priority, "debug", 5) == 0)
  94                 return LOG_DEBUG;
  95         return 0;
  96 }
  97
  98 size_t util_path_encode(const char *src, char *dest, size_t size)
  99 {
 100         size_t i, j;
 101
 102         for (i = 0, j = 0; src[i] != '\0'; i++) {
 103                 if (src[i] == '/') {
 104                         if (j+4 >= size) {
 105                                 j = 0;
 106                                 break;
 107                         }
 108                         memcpy(&dest[j], "\\x2f", 4);
 109                         j += 4;
 110                 } else if (src[i] == '\\') {
 111                         if (j+4 >= size) {
 112                                 j = 0;
 113                                 break;
 114                         }
 115                         memcpy(&dest[j], "\\x5c", 4);
 116                         j += 4;
 117                 } else {
 118                         if (j+1 >= size) {
 119                                 j = 0;
 120                                 break;
 121                         }
 122                         dest[j] = src[i];
 123                         j++;
 124                 }
 125         }
 126         dest[j] = '\0';
 127         return j;
 128 }
 129
 130 size_t util_path_decode(char *s)
 131 {
 132         size_t i, j;
 133
 134         for (i = 0, j = 0; s[i] != '\0'; j++) {
 135                 if (memcmp(&s[i], "\\x2f", 4) == 0) {
 136                         s[j] = '/';
 137                         i += 4;
 138                 } else if (memcmp(&s[i], "\\x5c", 4) == 0) {
 139                         s[j] = '\\';
 140                         i += 4;
 141                 } else {
 142                         s[j] = s[i];
 143                         i++;
 144                 }
 145         }
 146         s[j] = '\0';
 147         return j;
 148 }
 149
 150 void util_remove_trailing_chars(char *path, char c)
 151 {
 152         size_t len;
 153
 154         if (path == NULL)
 155                 return;
 156         len = strlen(path);
 157         while (len > 0 && path[len-1] == c)
 158                 path[--len] = '\0';
 159 }
 160
 161 /*
 162  * Concatenates strings. In any case, terminates in _all_ cases with '\0'
 163  * and moves the @dest pointer forward to the added '\0'. Returns the
 164  * remaining size, and 0 if the string was truncated.
 165  */
 166 size_t util_strpcpy(char **dest, size_t size, const char *src)
 167 {
 168         size_t len;
 169
 170         len = strlen(src);
 171         if (len >= size) {
 172                 if (size > 1)
 173                         *dest = mempcpy(*dest, src, size-1);
 174                 size = 0;
 175                 *dest[0] = '\0';
 176         } else {
 177                 if (len > 0) {
 178                         *dest = mempcpy(*dest, src, len);
 179                         size -= len;
 180                 }
 181                 *dest[0] = '\0';
 182         }
 183         return size;
 184 }
 185
 186 /* concatenates list of strings, moves dest forward */
 187 size_t util_strpcpyl(char **dest, size_t size, const char *src, ...)
 188 {
 189         va_list va;
 190
 191         va_start(va, src);
 192         do {
 193                 size = util_strpcpy(dest, size, src);
 194                 src = va_arg(va, char *);
 195         } while (src != NULL);
 196         va_end(va);
 197
 198         return size;
 199 }
 200
 201 /* copies string */
 202 size_t util_strscpy(char *dest, size_t size, const char *src)
 203 {
 204         char *s;
 205
 206         s = dest;
 207         return util_strpcpy(&s, size, src);
 208 }
 209
 210 /* concatenates list of strings */
 211 size_t util_strscpyl(char *dest, size_t size, const char *src, ...)
 212 {
 213         va_list va;
 214         char *s;
 215
 216         va_start(va, src);
 217         s = dest;
 218         do {
 219                 size = util_strpcpy(&s, size, src);
 220                 src = va_arg(va, char *);
 221         } while (src != NULL);
 222         va_end(va);
 223
 224         return size;
 225 }
 226
 227 /* count of characters used to encode one unicode char */
 228 static int utf8_encoded_expected_len(const char *str)
 229 {
 230         unsigned char c = (unsigned char)str[0];
 231
 232         if (c < 0x80)
 233                 return 1;
 234         if ((c & 0xe0) == 0xc0)
 235                 return 2;
 236         if ((c & 0xf0) == 0xe0)
 237                 return 3;
 238         if ((c & 0xf8) == 0xf0)
 239                 return 4;
 240         if ((c & 0xfc) == 0xf8)
 241                 return 5;
 242         if ((c & 0xfe) == 0xfc)
 243                 return 6;
 244         return 0;
 245 }
 246
 247 /* decode one unicode char */
 248 static int utf8_encoded_to_unichar(const char *str)
 249 {
 250         int unichar;
 251         int len;
 252         int i;
 253
 254         len = utf8_encoded_expected_len(str);
 255         switch (len) {
 256         case 1:
 257                 return (int)str[0];
 258         case 2:
 259                 unichar = str[0] & 0x1f;
 260                 break;
 261         case 3:
 262                 unichar = (int)str[0] & 0x0f;
 263                 break;
 264         case 4:
 265                 unichar = (int)str[0] & 0x07;
 266                 break;
 267         case 5:
 268                 unichar = (int)str[0] & 0x03;
 269                 break;
 270         case 6:
 271                 unichar = (int)str[0] & 0x01;
 272                 break;
 273         default:
 274                 return -1;
 275         }
 276
 277         for (i = 1; i < len; i++) {
 278                 if (((int)str[i] & 0xc0) != 0x80)
 279                         return -1;
 280                 unichar <<= 6;
 281                 unichar |= (int)str[i] & 0x3f;
 282         }
 283
 284         return unichar;
 285 }
 286
 287 /* expected size used to encode one unicode char */
 288 static int utf8_unichar_to_encoded_len(int unichar)
 289 {
 290         if (unichar < 0x80)
 291                 return 1;
 292         if (unichar < 0x800)
 293                 return 2;
 294         if (unichar < 0x10000)
 295                 return 3;
 296         if (unichar < 0x200000)
 297                 return 4;
 298         if (unichar < 0x4000000)
 299                 return 5;
 300         return 6;
 301 }
 302
 303 /* check if unicode char has a valid numeric range */
 304 static int utf8_unichar_valid_range(int unichar)
 305 {
 306         if (unichar > 0x10ffff)
 307                 return 0;
 308         if ((unichar & 0xfffff800) == 0xd800)
 309                 return 0;
 310         if ((unichar > 0xfdcf) && (unichar < 0xfdf0))
 311                 return 0;
 312         if ((unichar & 0xffff) == 0xffff)
 313                 return 0;
 314         return 1;
 315 }
 316
 317 /* validate one encoded unicode char and return its length */
 318 static int utf8_encoded_valid_unichar(const char *str)
 319 {
 320         int len;
 321         int unichar;
 322         int i;
 323
 324         len = utf8_encoded_expected_len(str);
 325         if (len == 0)
 326                 return -1;
 327
 328         /* ascii is valid */
 329         if (len == 1)
 330                 return 1;
 331
 332         /* check if expected encoded chars are available */
 333         for (i = 0; i < len; i++)
 334                 if ((str[i] & 0x80) != 0x80)
 335                         return -1;
 336
 337         unichar = utf8_encoded_to_unichar(str);
 338
 339         /* check if encoded length matches encoded value */
 340         if (utf8_unichar_to_encoded_len(unichar) != len)
 341                 return -1;
 342
 343         /* check if value has valid range */
 344         if (!utf8_unichar_valid_range(unichar))
 345                 return -1;
 346
 347         return len;
 348 }
 349
 350 int util_replace_whitespace(const char *str, char *to, size_t len)
 351 {
 352         size_t i, j;
 353
 354         /* strip trailing whitespace */
 355         len = strnlen(str, len);
 356         while (len && isspace(str[len-1]))
 357                 len--;
 358
 359         /* strip leading whitespace */
 360         i = 0;
 361         while (isspace(str[i]) && (i < len))
 362                 i++;
 363
 364         j = 0;
 365         while (i < len) {
 366                 /* substitute multiple whitespace with a single '_' */
 367                 if (isspace(str[i])) {
 368                         while (isspace(str[i]))
 369                                 i++;
 370                         to[j++] = '_';
 371                 }
 372                 to[j++] = str[i++];
 373         }
 374         to[j] = '\0';
 375         return 0;
 376 }
 377
 378 static int is_whitelisted(char c, const char *white)
 379 {
 380         if ((c >= '0' && c <= '9') ||
 381             (c >= 'A' && c <= 'Z') ||
 382             (c >= 'a' && c <= 'z') ||
 383             strchr("#+-.:=@_", c) != NULL ||
 384             (white != NULL && strchr(white, c) != NULL))
 385                 return 1;
 386         return 0;
 387 }
 388
 389 /* allow chars in whitelist, plain ascii, hex-escaping and valid utf8 */
 390 int util_replace_chars(char *str, const char *white)
 391 {
 392         size_t i = 0;
 393         int replaced = 0;
 394
 395         while (str[i] != '\0') {
 396                 int len;
 397
 398                 if (is_whitelisted(str[i], white)) {
 399                         i++;
 400                         continue;
 401                 }
 402
 403                 /* accept hex encoding */
 404                 if (str[i] == '\\' && str[i+1] == 'x') {
 405                         i += 2;
 406                         continue;
 407                 }
 408
 409                 /* accept valid utf8 */
 410                 len = utf8_encoded_valid_unichar(&str[i]);
 411                 if (len > 1) {
 412                         i += len;
 413                         continue;
 414                 }
 415
 416                 /* if space is allowed, replace whitespace with ordinary space */
 417                 if (isspace(str[i]) && white != NULL && strchr(white, ' ') != NULL) {
 418                         str[i] = ' ';
 419                         i++;
 420                         replaced++;
 421                         continue;
 422                 }
 423
 424                 /* everything else is replaced with '_' */
 425                 str[i] = '_';
 426                 i++;
 427                 replaced++;
 428         }
 429         return replaced;
 430 }
 431
 432 /**
 433  * udev_util_encode_string:
 434  * @str: input string to be encoded
 435  * @str_enc: output string to store the encoded input string
 436  * @len: maximum size of the output string, which may be
 437  *       four times as long as the input string
 438  *
 439  * Encode all potentially unsafe characters of a string to the
 440  * corresponding 2 char hex value prefixed by '\x'.
 441  *
 442  * Returns: 0 if the entire string was copied, non-zero otherwise.
 443  **/
 444 UDEV_EXPORT int udev_util_encode_string(const char *str, char *str_enc, size_t len)
 445 {
 446         size_t i, j;
 447
 448         if (str == NULL || str_enc == NULL)
 449                 return -1;
 450
 451         for (i = 0, j = 0; str[i] != '\0'; i++) {
 452                 int seqlen;
 453
 454                 seqlen = utf8_encoded_valid_unichar(&str[i]);
 455                 if (seqlen > 1) {
 456                         if (len-j < (size_t)seqlen)
 457                                 goto err;
 458                         memcpy(&str_enc[j], &str[i], seqlen);
 459                         j += seqlen;
 460                         i += (seqlen-1);
 461                 } else if (str[i] == '\\' || !is_whitelisted(str[i], NULL)) {
 462                         if (len-j < 4)
 463                                 goto err;
 464                         sprintf(&str_enc[j], "\\x%02x", (unsigned char) str[i]);
 465                         j += 4;
 466                 } else {
 467                         if (len-j < 1)
 468                                 goto err;
 469                         str_enc[j] = str[i];
 470                         j++;
 471                 }
 472         }
 473         if (len-j < 1)
 474                 goto err;
 475         str_enc[j] = '\0';
 476         return 0;
 477 err:
 478         return -1;
 479 }
 480
 481 /*
 482  * http://sites.google.com/site/murmurhash/
 483  *
 484  * All code is released to the public domain. For business purposes,
 485  * Murmurhash is under the MIT license.
 486  *
 487  */
 488 static unsigned int murmur_hash2(const char *key, int len, unsigned int seed)
 489 {
 490         /*
 491          *  'm' and 'r' are mixing constants generated offline.
 492          *  They're not really 'magic', they just happen to work well.
 493          */
 494         const unsigned int m = 0x5bd1e995;
 495         const int r = 24;
 496
 497         /* initialize the hash to a 'random' value */
 498         unsigned int h = seed ^ len;
 499
 500         /* mix 4 bytes at a time into the hash */
 501         const unsigned char * data = (const unsigned char *)key;
 502
 503         while(len >= 4) {
 504                 unsigned int k = *(unsigned int *)data;
 505
 506                 k *= m;
 507                 k ^= k >> r;
 508                 k *= m;
 509                 h *= m;
 510                 h ^= k;
 511
 512                 data += 4;
 513                 len -= 4;
 514         }
 515
 516         /* handle the last few bytes of the input array */
 517         switch(len) {
 518         case 3:
 519                 h ^= data[2] << 16;
 520         case 2:
 521                 h ^= data[1] << 8;
 522         case 1:
 523                 h ^= data[0];
 524                 h *= m;
 525         };
 526
 527         /* do a few final mixes of the hash to ensure the last few bytes are well-incorporated */
 528         h ^= h >> 13;
 529         h *= m;
 530         h ^= h >> 15;
 531
 532         return h;
 533 }
 534
 535 unsigned int util_string_hash32(const char *str)
 536 {
 537         return murmur_hash2(str, strlen(str), 0);
 538 }
 539
 540 /* get a bunch of bit numbers out of the hash, and set the bits in our bit field */
 541 uint64_t util_string_bloom64(const char *str)
 542 {
 543         uint64_t bits = 0;
 544         unsigned int hash = util_string_hash32(str);
 545
 546         bits |= 1LLU << (hash & 63);
 547         bits |= 1LLU << ((hash >> 6) & 63);
 548         bits |= 1LLU << ((hash >> 12) & 63);
 549         bits |= 1LLU << ((hash >> 18) & 63);
 550         return bits;
 551 }
 552
 553 #define USEC_PER_SEC  1000000ULL
 554 #define NSEC_PER_USEC 1000ULL
 555 unsigned long long ts_usec(const struct timespec *ts)
 556 {
 557         return (unsigned long long) ts->tv_sec * USEC_PER_SEC +
 558                (unsigned long long) ts->tv_nsec / NSEC_PER_USEC;
 559 }
 560
 561 unsigned long long now_usec(void)
 562 {
 563         struct timespec ts;
 564
 565         if (clock_gettime(CLOCK_MONOTONIC, &ts) != 0)
 566                 return 0;
 567         return ts_usec(&ts);
 568 }