[873] | 1 | /*
|
---|
| 2 | Unix SMB/CIFS implementation.
|
---|
| 3 | Character set conversion Extensions
|
---|
| 4 | Copyright (C) Igor Vergeichik <iverg@mail.ru> 2001
|
---|
| 5 | Copyright (C) Andrew Tridgell 2001
|
---|
| 6 | Copyright (C) Simo Sorce 2001
|
---|
| 7 | Copyright (C) Martin Pool 2003
|
---|
| 8 |
|
---|
| 9 | This program is free software; you can redistribute it and/or modify
|
---|
| 10 | it under the terms of the GNU General Public License as published by
|
---|
| 11 | the Free Software Foundation; either version 3 of the License, or
|
---|
| 12 | (at your option) any later version.
|
---|
| 13 |
|
---|
| 14 | This program is distributed in the hope that it will be useful,
|
---|
| 15 | but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 16 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 17 | GNU General Public License for more details.
|
---|
| 18 |
|
---|
| 19 | You should have received a copy of the GNU General Public License
|
---|
| 20 | along with this program. If not, see <http://www.gnu.org/licenses/>.
|
---|
| 21 |
|
---|
| 22 | */
|
---|
| 23 | #include "includes.h"
|
---|
| 24 |
|
---|
| 25 | /* We can parameterize this if someone complains.... JRA. */
|
---|
| 26 |
|
---|
| 27 | char lp_failed_convert_char(void)
|
---|
| 28 | {
|
---|
| 29 | return '_';
|
---|
| 30 | }
|
---|
| 31 |
|
---|
| 32 | /**
|
---|
| 33 | * @file
|
---|
| 34 | *
|
---|
| 35 | * @brief Character-set conversion routines built on our iconv.
|
---|
| 36 | *
|
---|
| 37 | * @note Samba's internal character set (at least in the 3.0 series)
|
---|
| 38 | * is always the same as the one for the Unix filesystem. It is
|
---|
| 39 | * <b>not</b> necessarily UTF-8 and may be different on machines that
|
---|
| 40 | * need i18n filenames to be compatible with Unix software. It does
|
---|
| 41 | * have to be a superset of ASCII. All multibyte sequences must start
|
---|
| 42 | * with a byte with the high bit set.
|
---|
| 43 | *
|
---|
| 44 | * @sa lib/iconv.c
|
---|
| 45 | */
|
---|
| 46 |
|
---|
| 47 |
|
---|
| 48 | static bool conv_silent; /* Should we do a debug if the conversion fails ? */
|
---|
| 49 | static bool initialized;
|
---|
| 50 |
|
---|
| 51 | void lazy_initialize_conv(void)
|
---|
| 52 | {
|
---|
| 53 | if (!initialized) {
|
---|
| 54 | load_case_tables_library();
|
---|
| 55 | init_iconv();
|
---|
| 56 | initialized = true;
|
---|
| 57 | }
|
---|
| 58 | }
|
---|
| 59 |
|
---|
| 60 | /**
|
---|
| 61 | * Destroy global objects allocated by init_iconv()
|
---|
| 62 | **/
|
---|
| 63 | void gfree_charcnv(void)
|
---|
| 64 | {
|
---|
| 65 | TALLOC_FREE(global_iconv_convenience);
|
---|
| 66 | initialized = false;
|
---|
| 67 | }
|
---|
| 68 |
|
---|
| 69 | /**
|
---|
| 70 | * Initialize iconv conversion descriptors.
|
---|
| 71 | *
|
---|
| 72 | * This is called the first time it is needed, and also called again
|
---|
| 73 | * every time the configuration is reloaded, because the charset or
|
---|
| 74 | * codepage might have changed.
|
---|
| 75 | **/
|
---|
| 76 | void init_iconv(void)
|
---|
| 77 | {
|
---|
| 78 | global_iconv_convenience = smb_iconv_convenience_reinit(NULL, lp_dos_charset(),
|
---|
| 79 | lp_unix_charset(), lp_display_charset(),
|
---|
| 80 | true, global_iconv_convenience);
|
---|
| 81 | }
|
---|
| 82 |
|
---|
| 83 | /**
|
---|
| 84 | * Convert string from one encoding to another, making error checking etc
|
---|
| 85 | * Slow path version - uses (slow) iconv.
|
---|
| 86 | *
|
---|
| 87 | * @param src pointer to source string (multibyte or singlebyte)
|
---|
| 88 | * @param srclen length of the source string in bytes
|
---|
| 89 | * @param dest pointer to destination string (multibyte or singlebyte)
|
---|
| 90 | * @param destlen maximal length allowed for string
|
---|
| 91 | * @param allow_bad_conv determines if a "best effort" conversion is acceptable (never returns errors)
|
---|
| 92 | * @returns the number of bytes occupied in the destination
|
---|
| 93 | *
|
---|
| 94 | * Ensure the srclen contains the terminating zero.
|
---|
| 95 | *
|
---|
| 96 | **/
|
---|
| 97 |
|
---|
| 98 | static size_t convert_string_internal(charset_t from, charset_t to,
|
---|
| 99 | void const *src, size_t srclen,
|
---|
| 100 | void *dest, size_t destlen, bool allow_bad_conv)
|
---|
| 101 | {
|
---|
| 102 | size_t i_len, o_len;
|
---|
| 103 | size_t retval;
|
---|
| 104 | const char* inbuf = (const char*)src;
|
---|
| 105 | char* outbuf = (char*)dest;
|
---|
| 106 | smb_iconv_t descriptor;
|
---|
| 107 | struct smb_iconv_convenience *ic;
|
---|
| 108 |
|
---|
| 109 | lazy_initialize_conv();
|
---|
| 110 | ic = get_iconv_convenience();
|
---|
| 111 | descriptor = get_conv_handle(ic, from, to);
|
---|
| 112 |
|
---|
| 113 | if (srclen == (size_t)-1) {
|
---|
| 114 | if (from == CH_UTF16LE || from == CH_UTF16BE) {
|
---|
| 115 | srclen = (strlen_w((const smb_ucs2_t *)src)+1) * 2;
|
---|
| 116 | } else {
|
---|
| 117 | srclen = strlen((const char *)src)+1;
|
---|
| 118 | }
|
---|
| 119 | }
|
---|
| 120 |
|
---|
| 121 |
|
---|
| 122 | if (descriptor == (smb_iconv_t)-1 || descriptor == (smb_iconv_t)0) {
|
---|
| 123 | if (!conv_silent)
|
---|
| 124 | DEBUG(0,("convert_string_internal: Conversion not supported.\n"));
|
---|
| 125 | return (size_t)-1;
|
---|
| 126 | }
|
---|
| 127 |
|
---|
| 128 | i_len=srclen;
|
---|
| 129 | o_len=destlen;
|
---|
| 130 |
|
---|
| 131 | again:
|
---|
| 132 |
|
---|
| 133 | retval = smb_iconv(descriptor, &inbuf, &i_len, &outbuf, &o_len);
|
---|
| 134 | if(retval==(size_t)-1) {
|
---|
| 135 | const char *reason="unknown error";
|
---|
| 136 | switch(errno) {
|
---|
| 137 | case EINVAL:
|
---|
| 138 | reason="Incomplete multibyte sequence";
|
---|
| 139 | if (!conv_silent)
|
---|
| 140 | DEBUG(3,("convert_string_internal: Conversion error: %s(%s)\n",reason,inbuf));
|
---|
| 141 | if (allow_bad_conv)
|
---|
| 142 | goto use_as_is;
|
---|
| 143 | return (size_t)-1;
|
---|
| 144 | case E2BIG:
|
---|
| 145 | reason="No more room";
|
---|
| 146 | if (!conv_silent) {
|
---|
| 147 | if (from == CH_UNIX) {
|
---|
| 148 | DEBUG(3,("E2BIG: convert_string(%s,%s): srclen=%u destlen=%u - '%s'\n",
|
---|
| 149 | charset_name(ic, from), charset_name(ic, to),
|
---|
| 150 | (unsigned int)srclen, (unsigned int)destlen, (const char *)src));
|
---|
| 151 | } else {
|
---|
| 152 | DEBUG(3,("E2BIG: convert_string(%s,%s): srclen=%u destlen=%u\n",
|
---|
| 153 | charset_name(ic, from), charset_name(ic, to),
|
---|
| 154 | (unsigned int)srclen, (unsigned int)destlen));
|
---|
| 155 | }
|
---|
| 156 | }
|
---|
| 157 | break;
|
---|
| 158 | case EILSEQ:
|
---|
| 159 | reason="Illegal multibyte sequence";
|
---|
| 160 | if (!conv_silent)
|
---|
| 161 | DEBUG(3,("convert_string_internal: Conversion error: %s(%s)\n",reason,inbuf));
|
---|
| 162 | if (allow_bad_conv)
|
---|
| 163 | goto use_as_is;
|
---|
| 164 |
|
---|
| 165 | return (size_t)-1;
|
---|
| 166 | default:
|
---|
| 167 | if (!conv_silent)
|
---|
| 168 | DEBUG(0,("convert_string_internal: Conversion error: %s(%s)\n",reason,inbuf));
|
---|
| 169 | return (size_t)-1;
|
---|
| 170 | }
|
---|
| 171 | /* smb_panic(reason); */
|
---|
| 172 | }
|
---|
| 173 | return destlen-o_len;
|
---|
| 174 |
|
---|
| 175 | use_as_is:
|
---|
| 176 |
|
---|
| 177 | /*
|
---|
| 178 | * Conversion not supported. This is actually an error, but there are so
|
---|
| 179 | * many misconfigured iconv systems and smb.conf's out there we can't just
|
---|
| 180 | * fail. Do a very bad conversion instead.... JRA.
|
---|
| 181 | */
|
---|
| 182 |
|
---|
| 183 | {
|
---|
| 184 | if (o_len == 0 || i_len == 0)
|
---|
| 185 | return destlen - o_len;
|
---|
| 186 |
|
---|
| 187 | if (((from == CH_UTF16LE)||(from == CH_UTF16BE)) &&
|
---|
| 188 | ((to != CH_UTF16LE)||(to != CH_UTF16BE))) {
|
---|
| 189 | /* Can't convert from utf16 any endian to multibyte.
|
---|
| 190 | Replace with the default fail char.
|
---|
| 191 | */
|
---|
| 192 | if (i_len < 2)
|
---|
| 193 | return destlen - o_len;
|
---|
| 194 | if (i_len >= 2) {
|
---|
| 195 | *outbuf = lp_failed_convert_char();
|
---|
| 196 |
|
---|
| 197 | outbuf++;
|
---|
| 198 | o_len--;
|
---|
| 199 |
|
---|
| 200 | inbuf += 2;
|
---|
| 201 | i_len -= 2;
|
---|
| 202 | }
|
---|
| 203 |
|
---|
| 204 | if (o_len == 0 || i_len == 0)
|
---|
| 205 | return destlen - o_len;
|
---|
| 206 |
|
---|
| 207 | /* Keep trying with the next char... */
|
---|
| 208 | goto again;
|
---|
| 209 |
|
---|
| 210 | } else if (from != CH_UTF16LE && from != CH_UTF16BE && to == CH_UTF16LE) {
|
---|
| 211 | /* Can't convert to UTF16LE - just widen by adding the
|
---|
| 212 | default fail char then zero.
|
---|
| 213 | */
|
---|
| 214 | if (o_len < 2)
|
---|
| 215 | return destlen - o_len;
|
---|
| 216 |
|
---|
| 217 | outbuf[0] = lp_failed_convert_char();
|
---|
| 218 | outbuf[1] = '\0';
|
---|
| 219 |
|
---|
| 220 | inbuf++;
|
---|
| 221 | i_len--;
|
---|
| 222 |
|
---|
| 223 | outbuf += 2;
|
---|
| 224 | o_len -= 2;
|
---|
| 225 |
|
---|
| 226 | if (o_len == 0 || i_len == 0)
|
---|
| 227 | return destlen - o_len;
|
---|
| 228 |
|
---|
| 229 | /* Keep trying with the next char... */
|
---|
| 230 | goto again;
|
---|
| 231 |
|
---|
| 232 | } else if (from != CH_UTF16LE && from != CH_UTF16BE &&
|
---|
| 233 | to != CH_UTF16LE && to != CH_UTF16BE) {
|
---|
| 234 | /* Failed multibyte to multibyte. Just copy the default fail char and
|
---|
| 235 | try again. */
|
---|
| 236 | outbuf[0] = lp_failed_convert_char();
|
---|
| 237 |
|
---|
| 238 | inbuf++;
|
---|
| 239 | i_len--;
|
---|
| 240 |
|
---|
| 241 | outbuf++;
|
---|
| 242 | o_len--;
|
---|
| 243 |
|
---|
| 244 | if (o_len == 0 || i_len == 0)
|
---|
| 245 | return destlen - o_len;
|
---|
| 246 |
|
---|
| 247 | /* Keep trying with the next char... */
|
---|
| 248 | goto again;
|
---|
| 249 |
|
---|
| 250 | } else {
|
---|
| 251 | /* Keep compiler happy.... */
|
---|
| 252 | return destlen - o_len;
|
---|
| 253 | }
|
---|
| 254 | }
|
---|
| 255 | }
|
---|
| 256 |
|
---|
| 257 | /**
|
---|
| 258 | * Convert string from one encoding to another, making error checking etc
|
---|
| 259 | * Fast path version - handles ASCII first.
|
---|
| 260 | *
|
---|
| 261 | * @param src pointer to source string (multibyte or singlebyte)
|
---|
| 262 | * @param srclen length of the source string in bytes, or -1 for nul terminated.
|
---|
| 263 | * @param dest pointer to destination string (multibyte or singlebyte)
|
---|
| 264 | * @param destlen maximal length allowed for string - *NEVER* -1.
|
---|
| 265 | * @param allow_bad_conv determines if a "best effort" conversion is acceptable (never returns errors)
|
---|
| 266 | * @returns the number of bytes occupied in the destination
|
---|
| 267 | *
|
---|
| 268 | * Ensure the srclen contains the terminating zero.
|
---|
| 269 | *
|
---|
| 270 | * This function has been hand-tuned to provide a fast path.
|
---|
| 271 | * Don't change unless you really know what you are doing. JRA.
|
---|
| 272 | **/
|
---|
| 273 |
|
---|
| 274 | size_t convert_string(charset_t from, charset_t to,
|
---|
| 275 | void const *src, size_t srclen,
|
---|
| 276 | void *dest, size_t destlen, bool allow_bad_conv)
|
---|
| 277 | {
|
---|
| 278 | /*
|
---|
| 279 | * NB. We deliberately don't do a strlen here if srclen == -1.
|
---|
| 280 | * This is very expensive over millions of calls and is taken
|
---|
| 281 | * care of in the slow path in convert_string_internal. JRA.
|
---|
| 282 | */
|
---|
| 283 |
|
---|
| 284 | #ifdef DEVELOPER
|
---|
| 285 | SMB_ASSERT(destlen != (size_t)-1);
|
---|
| 286 | #endif
|
---|
| 287 |
|
---|
| 288 | if (srclen == 0)
|
---|
| 289 | return 0;
|
---|
| 290 |
|
---|
| 291 | if (from != CH_UTF16LE && from != CH_UTF16BE && to != CH_UTF16LE && to != CH_UTF16BE) {
|
---|
| 292 | const unsigned char *p = (const unsigned char *)src;
|
---|
| 293 | unsigned char *q = (unsigned char *)dest;
|
---|
| 294 | size_t slen = srclen;
|
---|
| 295 | size_t dlen = destlen;
|
---|
| 296 | unsigned char lastp = '\0';
|
---|
| 297 | size_t retval = 0;
|
---|
| 298 |
|
---|
| 299 | /* If all characters are ascii, fast path here. */
|
---|
| 300 | while (slen && dlen) {
|
---|
| 301 | if ((lastp = *p) <= 0x7f) {
|
---|
| 302 | *q++ = *p++;
|
---|
| 303 | if (slen != (size_t)-1) {
|
---|
| 304 | slen--;
|
---|
| 305 | }
|
---|
| 306 | dlen--;
|
---|
| 307 | retval++;
|
---|
| 308 | if (!lastp)
|
---|
| 309 | break;
|
---|
| 310 | } else {
|
---|
| 311 | #ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
|
---|
| 312 | goto general_case;
|
---|
| 313 | #else
|
---|
| 314 | size_t ret = convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
|
---|
| 315 | if (ret == (size_t)-1) {
|
---|
| 316 | return ret;
|
---|
| 317 | }
|
---|
| 318 | return retval + ret;
|
---|
| 319 | #endif
|
---|
| 320 | }
|
---|
| 321 | }
|
---|
| 322 | if (!dlen) {
|
---|
| 323 | /* Even if we fast path we should note if we ran out of room. */
|
---|
| 324 | if (((slen != (size_t)-1) && slen) ||
|
---|
| 325 | ((slen == (size_t)-1) && lastp)) {
|
---|
| 326 | errno = E2BIG;
|
---|
| 327 | }
|
---|
| 328 | }
|
---|
| 329 | return retval;
|
---|
| 330 | } else if (from == CH_UTF16LE && to != CH_UTF16LE) {
|
---|
| 331 | const unsigned char *p = (const unsigned char *)src;
|
---|
| 332 | unsigned char *q = (unsigned char *)dest;
|
---|
| 333 | size_t retval = 0;
|
---|
| 334 | size_t slen = srclen;
|
---|
| 335 | size_t dlen = destlen;
|
---|
| 336 | unsigned char lastp = '\0';
|
---|
| 337 |
|
---|
| 338 | /* If all characters are ascii, fast path here. */
|
---|
| 339 | while (((slen == (size_t)-1) || (slen >= 2)) && dlen) {
|
---|
| 340 | if (((lastp = *p) <= 0x7f) && (p[1] == 0)) {
|
---|
| 341 | *q++ = *p;
|
---|
| 342 | if (slen != (size_t)-1) {
|
---|
| 343 | slen -= 2;
|
---|
| 344 | }
|
---|
| 345 | p += 2;
|
---|
| 346 | dlen--;
|
---|
| 347 | retval++;
|
---|
| 348 | if (!lastp)
|
---|
| 349 | break;
|
---|
| 350 | } else {
|
---|
| 351 | #ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
|
---|
| 352 | goto general_case;
|
---|
| 353 | #else
|
---|
| 354 | size_t ret = convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
|
---|
| 355 | if (ret == (size_t)-1) {
|
---|
| 356 | return ret;
|
---|
| 357 | }
|
---|
| 358 | return retval + ret;
|
---|
| 359 | #endif
|
---|
| 360 | }
|
---|
| 361 | }
|
---|
| 362 | if (!dlen) {
|
---|
| 363 | /* Even if we fast path we should note if we ran out of room. */
|
---|
| 364 | if (((slen != (size_t)-1) && slen) ||
|
---|
| 365 | ((slen == (size_t)-1) && lastp)) {
|
---|
| 366 | errno = E2BIG;
|
---|
| 367 | }
|
---|
| 368 | }
|
---|
| 369 | return retval;
|
---|
| 370 | } else if (from != CH_UTF16LE && from != CH_UTF16BE && to == CH_UTF16LE) {
|
---|
| 371 | const unsigned char *p = (const unsigned char *)src;
|
---|
| 372 | unsigned char *q = (unsigned char *)dest;
|
---|
| 373 | size_t retval = 0;
|
---|
| 374 | size_t slen = srclen;
|
---|
| 375 | size_t dlen = destlen;
|
---|
| 376 | unsigned char lastp = '\0';
|
---|
| 377 |
|
---|
| 378 | /* If all characters are ascii, fast path here. */
|
---|
| 379 | while (slen && (dlen >= 2)) {
|
---|
| 380 | if ((lastp = *p) <= 0x7F) {
|
---|
| 381 | *q++ = *p++;
|
---|
| 382 | *q++ = '\0';
|
---|
| 383 | if (slen != (size_t)-1) {
|
---|
| 384 | slen--;
|
---|
| 385 | }
|
---|
| 386 | dlen -= 2;
|
---|
| 387 | retval += 2;
|
---|
| 388 | if (!lastp)
|
---|
| 389 | break;
|
---|
| 390 | } else {
|
---|
| 391 | #ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
|
---|
| 392 | goto general_case;
|
---|
| 393 | #else
|
---|
| 394 | size_t ret = convert_string_internal(from, to, p, slen, q, dlen, allow_bad_conv);
|
---|
| 395 | if (ret == (size_t)-1) {
|
---|
| 396 | return ret;
|
---|
| 397 | }
|
---|
| 398 | return retval + ret;
|
---|
| 399 | #endif
|
---|
| 400 | }
|
---|
| 401 | }
|
---|
| 402 | if (!dlen) {
|
---|
| 403 | /* Even if we fast path we should note if we ran out of room. */
|
---|
| 404 | if (((slen != (size_t)-1) && slen) ||
|
---|
| 405 | ((slen == (size_t)-1) && lastp)) {
|
---|
| 406 | errno = E2BIG;
|
---|
| 407 | }
|
---|
| 408 | }
|
---|
| 409 | return retval;
|
---|
| 410 | }
|
---|
| 411 |
|
---|
| 412 | #ifdef BROKEN_UNICODE_COMPOSE_CHARACTERS
|
---|
| 413 | general_case:
|
---|
| 414 | #endif
|
---|
| 415 | return convert_string_internal(from, to, src, srclen, dest, destlen, allow_bad_conv);
|
---|
| 416 | }
|
---|
| 417 |
|
---|
| 418 | /**
|
---|
| 419 | * Convert between character sets, allocating a new buffer using talloc for the result.
|
---|
| 420 | *
|
---|
| 421 | * @param srclen length of source buffer.
|
---|
| 422 | * @param dest always set at least to NULL
|
---|
| 423 | * @parm converted_size set to the number of bytes occupied by the string in
|
---|
| 424 | * the destination on success.
|
---|
| 425 | * @note -1 is not accepted for srclen.
|
---|
| 426 | *
|
---|
| 427 | * @return true if new buffer was correctly allocated, and string was
|
---|
| 428 | * converted.
|
---|
| 429 | *
|
---|
| 430 | * Ensure the srclen contains the terminating zero.
|
---|
| 431 | *
|
---|
| 432 | * I hate the goto's in this function. It's embarressing.....
|
---|
| 433 | * There has to be a cleaner way to do this. JRA.
|
---|
| 434 | */
|
---|
| 435 | bool convert_string_talloc(TALLOC_CTX *ctx, charset_t from, charset_t to,
|
---|
| 436 | void const *src, size_t srclen, void *dst,
|
---|
| 437 | size_t *converted_size, bool allow_bad_conv)
|
---|
| 438 |
|
---|
| 439 | {
|
---|
| 440 | size_t i_len, o_len, destlen = (srclen * 3) / 2;
|
---|
| 441 | size_t retval;
|
---|
| 442 | const char *inbuf = (const char *)src;
|
---|
| 443 | char *outbuf = NULL, *ob = NULL;
|
---|
| 444 | smb_iconv_t descriptor;
|
---|
| 445 | void **dest = (void **)dst;
|
---|
| 446 | struct smb_iconv_convenience *ic;
|
---|
| 447 |
|
---|
| 448 | *dest = NULL;
|
---|
| 449 |
|
---|
| 450 | if (!converted_size) {
|
---|
| 451 | errno = EINVAL;
|
---|
| 452 | return false;
|
---|
| 453 | }
|
---|
| 454 |
|
---|
| 455 | if (src == NULL || srclen == (size_t)-1) {
|
---|
| 456 | errno = EINVAL;
|
---|
| 457 | return false;
|
---|
| 458 | }
|
---|
| 459 |
|
---|
| 460 | if (srclen == 0) {
|
---|
| 461 | /* We really should treat this as an error, but
|
---|
| 462 | there are too many callers that need this to
|
---|
| 463 | return a NULL terminated string in the correct
|
---|
| 464 | character set. */
|
---|
| 465 | if (to == CH_UTF16LE|| to == CH_UTF16BE || to == CH_UTF16MUNGED) {
|
---|
| 466 | destlen = 2;
|
---|
| 467 | } else {
|
---|
| 468 | destlen = 1;
|
---|
| 469 | }
|
---|
| 470 | ob = talloc_zero_array(ctx, char, destlen);
|
---|
| 471 | if (ob == NULL) {
|
---|
| 472 | errno = ENOMEM;
|
---|
| 473 | return false;
|
---|
| 474 | }
|
---|
| 475 | *converted_size = destlen;
|
---|
| 476 | *dest = ob;
|
---|
| 477 | return true;
|
---|
| 478 | }
|
---|
| 479 |
|
---|
| 480 | lazy_initialize_conv();
|
---|
| 481 | ic = get_iconv_convenience();
|
---|
| 482 | descriptor = get_conv_handle(ic, from, to);
|
---|
| 483 |
|
---|
| 484 | if (descriptor == (smb_iconv_t)-1 || descriptor == (smb_iconv_t)0) {
|
---|
| 485 | if (!conv_silent)
|
---|
| 486 | DEBUG(0,("convert_string_talloc: Conversion not supported.\n"));
|
---|
| 487 | errno = EOPNOTSUPP;
|
---|
| 488 | return false;
|
---|
| 489 | }
|
---|
| 490 |
|
---|
| 491 | convert:
|
---|
| 492 |
|
---|
| 493 | /* +2 is for ucs2 null termination. */
|
---|
| 494 | if ((destlen*2)+2 < destlen) {
|
---|
| 495 | /* wrapped ! abort. */
|
---|
| 496 | if (!conv_silent)
|
---|
| 497 | DEBUG(0, ("convert_string_talloc: destlen wrapped !\n"));
|
---|
| 498 | TALLOC_FREE(outbuf);
|
---|
| 499 | errno = EOPNOTSUPP;
|
---|
| 500 | return false;
|
---|
| 501 | } else {
|
---|
| 502 | destlen = destlen * 2;
|
---|
| 503 | }
|
---|
| 504 |
|
---|
| 505 | /* +2 is for ucs2 null termination. */
|
---|
| 506 | ob = (char *)TALLOC_REALLOC(ctx, ob, destlen + 2);
|
---|
| 507 |
|
---|
| 508 | if (!ob) {
|
---|
| 509 | DEBUG(0, ("convert_string_talloc: realloc failed!\n"));
|
---|
| 510 | errno = ENOMEM;
|
---|
| 511 | return false;
|
---|
| 512 | }
|
---|
| 513 | outbuf = ob;
|
---|
| 514 | i_len = srclen;
|
---|
| 515 | o_len = destlen;
|
---|
| 516 |
|
---|
| 517 | again:
|
---|
| 518 |
|
---|
| 519 | retval = smb_iconv(descriptor,
|
---|
| 520 | &inbuf, &i_len,
|
---|
| 521 | &outbuf, &o_len);
|
---|
| 522 | if(retval == (size_t)-1) {
|
---|
| 523 | const char *reason="unknown error";
|
---|
| 524 | switch(errno) {
|
---|
| 525 | case EINVAL:
|
---|
| 526 | reason="Incomplete multibyte sequence";
|
---|
| 527 | if (!conv_silent)
|
---|
| 528 | DEBUG(3,("convert_string_talloc: Conversion error: %s(%s)\n",reason,inbuf));
|
---|
| 529 | if (allow_bad_conv)
|
---|
| 530 | goto use_as_is;
|
---|
| 531 | break;
|
---|
| 532 | case E2BIG:
|
---|
| 533 | goto convert;
|
---|
| 534 | case EILSEQ:
|
---|
| 535 | reason="Illegal multibyte sequence";
|
---|
| 536 | if (!conv_silent)
|
---|
| 537 | DEBUG(3,("convert_string_talloc: Conversion error: %s(%s)\n",reason,inbuf));
|
---|
| 538 | if (allow_bad_conv)
|
---|
| 539 | goto use_as_is;
|
---|
| 540 | break;
|
---|
| 541 | }
|
---|
| 542 | if (!conv_silent)
|
---|
| 543 | DEBUG(0,("Conversion error: %s(%s)\n",reason,inbuf));
|
---|
| 544 | /* smb_panic(reason); */
|
---|
| 545 | TALLOC_FREE(ob);
|
---|
| 546 | return false;
|
---|
| 547 | }
|
---|
| 548 |
|
---|
| 549 | out:
|
---|
| 550 |
|
---|
| 551 | destlen = destlen - o_len;
|
---|
| 552 | /* Don't shrink unless we're reclaiming a lot of
|
---|
| 553 | * space. This is in the hot codepath and these
|
---|
| 554 | * reallocs *cost*. JRA.
|
---|
| 555 | */
|
---|
| 556 | if (o_len > 1024) {
|
---|
| 557 | /* We're shrinking here so we know the +2 is safe from wrap. */
|
---|
| 558 | ob = (char *)TALLOC_REALLOC(ctx,ob,destlen + 2);
|
---|
| 559 | }
|
---|
| 560 |
|
---|
| 561 | if (destlen && !ob) {
|
---|
| 562 | DEBUG(0, ("convert_string_talloc: out of memory!\n"));
|
---|
| 563 | errno = ENOMEM;
|
---|
| 564 | return false;
|
---|
| 565 | }
|
---|
| 566 |
|
---|
| 567 | *dest = ob;
|
---|
| 568 |
|
---|
| 569 | /* Must ucs2 null terminate in the extra space we allocated. */
|
---|
| 570 | ob[destlen] = '\0';
|
---|
| 571 | ob[destlen+1] = '\0';
|
---|
| 572 |
|
---|
| 573 | /* Ensure we can never return a *converted_size of zero. */
|
---|
| 574 | if (destlen == 0) {
|
---|
| 575 | /* This can happen from a bad iconv "use_as_is:" call. */
|
---|
| 576 | if (to == CH_UTF16LE|| to == CH_UTF16BE || to == CH_UTF16MUNGED) {
|
---|
| 577 | destlen = 2;
|
---|
| 578 | } else {
|
---|
| 579 | destlen = 1;
|
---|
| 580 | }
|
---|
| 581 | }
|
---|
| 582 |
|
---|
| 583 | *converted_size = destlen;
|
---|
| 584 | return true;
|
---|
| 585 |
|
---|
| 586 | use_as_is:
|
---|
| 587 |
|
---|
| 588 | /*
|
---|
| 589 | * Conversion not supported. This is actually an error, but there are so
|
---|
| 590 | * many misconfigured iconv systems and smb.conf's out there we can't just
|
---|
| 591 | * fail. Do a very bad conversion instead.... JRA.
|
---|
| 592 | */
|
---|
| 593 |
|
---|
| 594 | {
|
---|
| 595 | if (o_len == 0 || i_len == 0)
|
---|
| 596 | goto out;
|
---|
| 597 |
|
---|
| 598 | if (((from == CH_UTF16LE)||(from == CH_UTF16BE)) &&
|
---|
| 599 | ((to != CH_UTF16LE)||(to != CH_UTF16BE))) {
|
---|
| 600 | /* Can't convert from utf16 any endian to multibyte.
|
---|
| 601 | Replace with the default fail char.
|
---|
| 602 | */
|
---|
| 603 |
|
---|
| 604 | if (i_len < 2)
|
---|
| 605 | goto out;
|
---|
| 606 |
|
---|
| 607 | if (i_len >= 2) {
|
---|
| 608 | *outbuf = lp_failed_convert_char();
|
---|
| 609 |
|
---|
| 610 | outbuf++;
|
---|
| 611 | o_len--;
|
---|
| 612 |
|
---|
| 613 | inbuf += 2;
|
---|
| 614 | i_len -= 2;
|
---|
| 615 | }
|
---|
| 616 |
|
---|
| 617 | if (o_len == 0 || i_len == 0)
|
---|
| 618 | goto out;
|
---|
| 619 |
|
---|
| 620 | /* Keep trying with the next char... */
|
---|
| 621 | goto again;
|
---|
| 622 |
|
---|
| 623 | } else if (from != CH_UTF16LE && from != CH_UTF16BE && to == CH_UTF16LE) {
|
---|
| 624 | /* Can't convert to UTF16LE - just widen by adding the
|
---|
| 625 | default fail char then zero.
|
---|
| 626 | */
|
---|
| 627 | if (o_len < 2)
|
---|
| 628 | goto out;
|
---|
| 629 |
|
---|
| 630 | outbuf[0] = lp_failed_convert_char();
|
---|
| 631 | outbuf[1] = '\0';
|
---|
| 632 |
|
---|
| 633 | inbuf++;
|
---|
| 634 | i_len--;
|
---|
| 635 |
|
---|
| 636 | outbuf += 2;
|
---|
| 637 | o_len -= 2;
|
---|
| 638 |
|
---|
| 639 | if (o_len == 0 || i_len == 0)
|
---|
| 640 | goto out;
|
---|
| 641 |
|
---|
| 642 | /* Keep trying with the next char... */
|
---|
| 643 | goto again;
|
---|
| 644 |
|
---|
| 645 | } else if (from != CH_UTF16LE && from != CH_UTF16BE &&
|
---|
| 646 | to != CH_UTF16LE && to != CH_UTF16BE) {
|
---|
| 647 | /* Failed multibyte to multibyte. Just copy the default fail char and
|
---|
| 648 | try again. */
|
---|
| 649 | outbuf[0] = lp_failed_convert_char();
|
---|
| 650 |
|
---|
| 651 | inbuf++;
|
---|
| 652 | i_len--;
|
---|
| 653 |
|
---|
| 654 | outbuf++;
|
---|
| 655 | o_len--;
|
---|
| 656 |
|
---|
| 657 | if (o_len == 0 || i_len == 0)
|
---|
| 658 | goto out;
|
---|
| 659 |
|
---|
| 660 | /* Keep trying with the next char... */
|
---|
| 661 | goto again;
|
---|
| 662 |
|
---|
| 663 | } else {
|
---|
| 664 | /* Keep compiler happy.... */
|
---|
| 665 | goto out;
|
---|
| 666 | }
|
---|
| 667 | }
|
---|
| 668 | }
|
---|
| 669 |
|
---|
| 670 | size_t unix_strupper(const char *src, size_t srclen, char *dest, size_t destlen)
|
---|
| 671 | {
|
---|
| 672 | size_t size;
|
---|
| 673 | smb_ucs2_t *buffer;
|
---|
| 674 |
|
---|
| 675 | if (!push_ucs2_talloc(talloc_tos(), &buffer, src, &size)) {
|
---|
| 676 | return (size_t)-1;
|
---|
| 677 | }
|
---|
| 678 |
|
---|
| 679 | if (!strupper_w(buffer) && (dest == src)) {
|
---|
| 680 | TALLOC_FREE(buffer);
|
---|
| 681 | return srclen;
|
---|
| 682 | }
|
---|
| 683 |
|
---|
| 684 | size = convert_string(CH_UTF16LE, CH_UNIX, buffer, size, dest, destlen, True);
|
---|
| 685 | TALLOC_FREE(buffer);
|
---|
| 686 | return size;
|
---|
| 687 | }
|
---|
| 688 |
|
---|
| 689 | /**
|
---|
| 690 | talloc_strdup() a unix string to upper case.
|
---|
| 691 | **/
|
---|
| 692 |
|
---|
| 693 | char *talloc_strdup_upper(TALLOC_CTX *ctx, const char *s)
|
---|
| 694 | {
|
---|
| 695 | char *out_buffer = talloc_strdup(ctx,s);
|
---|
| 696 | const unsigned char *p = (const unsigned char *)s;
|
---|
| 697 | unsigned char *q = (unsigned char *)out_buffer;
|
---|
| 698 |
|
---|
| 699 | if (!q) {
|
---|
| 700 | return NULL;
|
---|
| 701 | }
|
---|
| 702 |
|
---|
| 703 | /* this is quite a common operation, so we want it to be
|
---|
| 704 | fast. We optimise for the ascii case, knowing that all our
|
---|
| 705 | supported multi-byte character sets are ascii-compatible
|
---|
| 706 | (ie. they match for the first 128 chars) */
|
---|
| 707 |
|
---|
| 708 | while (*p) {
|
---|
| 709 | if (*p & 0x80)
|
---|
| 710 | break;
|
---|
| 711 | *q++ = toupper_ascii_fast(*p);
|
---|
| 712 | p++;
|
---|
| 713 | }
|
---|
| 714 |
|
---|
| 715 | if (*p) {
|
---|
| 716 | /* MB case. */
|
---|
| 717 | size_t converted_size, converted_size2;
|
---|
| 718 | smb_ucs2_t *ubuf = NULL;
|
---|
| 719 |
|
---|
| 720 | /* We're not using the ascii buffer above. */
|
---|
| 721 | TALLOC_FREE(out_buffer);
|
---|
| 722 |
|
---|
| 723 | if (!convert_string_talloc(ctx, CH_UNIX, CH_UTF16LE, s,
|
---|
| 724 | strlen(s)+1, (void *)&ubuf,
|
---|
| 725 | &converted_size, True))
|
---|
| 726 | {
|
---|
| 727 | return NULL;
|
---|
| 728 | }
|
---|
| 729 |
|
---|
| 730 | strupper_w(ubuf);
|
---|
| 731 |
|
---|
| 732 | if (!convert_string_talloc(ctx, CH_UTF16LE, CH_UNIX, ubuf,
|
---|
| 733 | converted_size, (void *)&out_buffer,
|
---|
| 734 | &converted_size2, True))
|
---|
| 735 | {
|
---|
| 736 | TALLOC_FREE(ubuf);
|
---|
| 737 | return NULL;
|
---|
| 738 | }
|
---|
| 739 |
|
---|
| 740 | /* Don't need the intermediate buffer
|
---|
| 741 | * anymore.
|
---|
| 742 | */
|
---|
| 743 | TALLOC_FREE(ubuf);
|
---|
| 744 | }
|
---|
| 745 |
|
---|
| 746 | return out_buffer;
|
---|
| 747 | }
|
---|
| 748 |
|
---|
| 749 | char *strupper_talloc(TALLOC_CTX *ctx, const char *s) {
|
---|
| 750 | return talloc_strdup_upper(ctx, s);
|
---|
| 751 | }
|
---|
| 752 |
|
---|
| 753 |
|
---|
| 754 | size_t unix_strlower(const char *src, size_t srclen, char *dest, size_t destlen)
|
---|
| 755 | {
|
---|
| 756 | size_t size;
|
---|
| 757 | smb_ucs2_t *buffer = NULL;
|
---|
| 758 |
|
---|
| 759 | if (!convert_string_talloc(talloc_tos(), CH_UNIX, CH_UTF16LE, src, srclen,
|
---|
| 760 | (void **)(void *)&buffer, &size,
|
---|
| 761 | True))
|
---|
| 762 | {
|
---|
| 763 | smb_panic("failed to create UCS2 buffer");
|
---|
| 764 | }
|
---|
| 765 | if (!strlower_w(buffer) && (dest == src)) {
|
---|
| 766 | TALLOC_FREE(buffer);
|
---|
| 767 | return srclen;
|
---|
| 768 | }
|
---|
| 769 | size = convert_string(CH_UTF16LE, CH_UNIX, buffer, size, dest, destlen, True);
|
---|
| 770 | TALLOC_FREE(buffer);
|
---|
| 771 | return size;
|
---|
| 772 | }
|
---|
| 773 |
|
---|
| 774 |
|
---|
| 775 | char *talloc_strdup_lower(TALLOC_CTX *ctx, const char *s)
|
---|
| 776 | {
|
---|
| 777 | size_t converted_size;
|
---|
| 778 | smb_ucs2_t *buffer = NULL;
|
---|
| 779 | char *out_buffer;
|
---|
| 780 |
|
---|
| 781 | if (!push_ucs2_talloc(ctx, &buffer, s, &converted_size)) {
|
---|
| 782 | return NULL;
|
---|
| 783 | }
|
---|
| 784 |
|
---|
| 785 | strlower_w(buffer);
|
---|
| 786 |
|
---|
| 787 | if (!pull_ucs2_talloc(ctx, &out_buffer, buffer, &converted_size)) {
|
---|
| 788 | TALLOC_FREE(buffer);
|
---|
| 789 | return NULL;
|
---|
| 790 | }
|
---|
| 791 |
|
---|
| 792 | TALLOC_FREE(buffer);
|
---|
| 793 |
|
---|
| 794 | return out_buffer;
|
---|
| 795 | }
|
---|
| 796 |
|
---|
| 797 | char *strlower_talloc(TALLOC_CTX *ctx, const char *s) {
|
---|
| 798 | return talloc_strdup_lower(ctx, s);
|
---|
| 799 | }
|
---|
| 800 |
|
---|
| 801 | size_t ucs2_align(const void *base_ptr, const void *p, int flags)
|
---|
| 802 | {
|
---|
| 803 | if (flags & (STR_NOALIGN|STR_ASCII))
|
---|
| 804 | return 0;
|
---|
| 805 | return PTR_DIFF(p, base_ptr) & 1;
|
---|
| 806 | }
|
---|
| 807 |
|
---|
| 808 |
|
---|
| 809 | /**
|
---|
| 810 | * Copy a string from a char* unix src to a dos codepage string destination.
|
---|
| 811 | *
|
---|
| 812 | * @return the number of bytes occupied by the string in the destination.
|
---|
| 813 | *
|
---|
| 814 | * @param flags can include
|
---|
| 815 | * <dl>
|
---|
| 816 | * <dt>STR_TERMINATE</dt> <dd>means include the null termination</dd>
|
---|
| 817 | * <dt>STR_UPPER</dt> <dd>means uppercase in the destination</dd>
|
---|
| 818 | * </dl>
|
---|
| 819 | *
|
---|
| 820 | * @param dest_len the maximum length in bytes allowed in the
|
---|
| 821 | * destination.
|
---|
| 822 | **/
|
---|
| 823 | size_t push_ascii(void *dest, const char *src, size_t dest_len, int flags)
|
---|
| 824 | {
|
---|
[874] | 825 | size_t src_len = 0;
|
---|
[873] | 826 | char *tmpbuf = NULL;
|
---|
| 827 | size_t ret;
|
---|
| 828 |
|
---|
| 829 | /* No longer allow a length of -1. */
|
---|
| 830 | if (dest_len == (size_t)-1) {
|
---|
| 831 | smb_panic("push_ascii - dest_len == -1");
|
---|
| 832 | }
|
---|
| 833 |
|
---|
| 834 | if (flags & STR_UPPER) {
|
---|
| 835 | tmpbuf = SMB_STRDUP(src);
|
---|
| 836 | if (!tmpbuf) {
|
---|
| 837 | smb_panic("malloc fail");
|
---|
| 838 | }
|
---|
| 839 | strupper_m(tmpbuf);
|
---|
| 840 | src = tmpbuf;
|
---|
| 841 | }
|
---|
| 842 |
|
---|
[874] | 843 | src_len = strlen(src);
|
---|
[873] | 844 | if (flags & (STR_TERMINATE | STR_TERMINATE_ASCII)) {
|
---|
| 845 | src_len++;
|
---|
| 846 | }
|
---|
| 847 |
|
---|
| 848 | ret = convert_string(CH_UNIX, CH_DOS, src, src_len, dest, dest_len, True);
|
---|
[874] | 849 |
|
---|
| 850 | SAFE_FREE(tmpbuf);
|
---|
| 851 | if (ret == (size_t)-1) {
|
---|
| 852 | if ((flags & (STR_TERMINATE | STR_TERMINATE_ASCII))
|
---|
| 853 | && dest_len > 0) {
|
---|
| 854 | ((char *)dest)[0] = '\0';
|
---|
| 855 | }
|
---|
| 856 | return 0;
|
---|
[873] | 857 | }
|
---|
| 858 | return ret;
|
---|
| 859 | }
|
---|
| 860 |
|
---|
| 861 | size_t push_ascii_fstring(void *dest, const char *src)
|
---|
| 862 | {
|
---|
| 863 | return push_ascii(dest, src, sizeof(fstring), STR_TERMINATE);
|
---|
| 864 | }
|
---|
| 865 |
|
---|
| 866 | /********************************************************************
|
---|
| 867 | Push an nstring - ensure null terminated. Written by
|
---|
| 868 | moriyama@miraclelinux.com (MORIYAMA Masayuki).
|
---|
| 869 | ********************************************************************/
|
---|
| 870 |
|
---|
| 871 | size_t push_ascii_nstring(void *dest, const char *src)
|
---|
| 872 | {
|
---|
| 873 | size_t i, buffer_len, dest_len;
|
---|
| 874 | smb_ucs2_t *buffer;
|
---|
| 875 |
|
---|
| 876 | conv_silent = True;
|
---|
| 877 | if (!push_ucs2_talloc(talloc_tos(), &buffer, src, &buffer_len)) {
|
---|
| 878 | smb_panic("failed to create UCS2 buffer");
|
---|
| 879 | }
|
---|
| 880 |
|
---|
| 881 | /* We're using buffer_len below to count ucs2 characters, not bytes. */
|
---|
| 882 | buffer_len /= sizeof(smb_ucs2_t);
|
---|
| 883 |
|
---|
| 884 | dest_len = 0;
|
---|
| 885 | for (i = 0; buffer[i] != 0 && (i < buffer_len); i++) {
|
---|
| 886 | unsigned char mb[10];
|
---|
| 887 | /* Convert one smb_ucs2_t character at a time. */
|
---|
| 888 | size_t mb_len = convert_string(CH_UTF16LE, CH_DOS, buffer+i, sizeof(smb_ucs2_t), mb, sizeof(mb), False);
|
---|
| 889 | if ((mb_len != (size_t)-1) && (dest_len + mb_len <= MAX_NETBIOSNAME_LEN - 1)) {
|
---|
| 890 | memcpy((char *)dest + dest_len, mb, mb_len);
|
---|
| 891 | dest_len += mb_len;
|
---|
| 892 | } else {
|
---|
| 893 | errno = E2BIG;
|
---|
| 894 | break;
|
---|
| 895 | }
|
---|
| 896 | }
|
---|
| 897 | ((char *)dest)[dest_len] = '\0';
|
---|
| 898 |
|
---|
| 899 | conv_silent = False;
|
---|
| 900 | TALLOC_FREE(buffer);
|
---|
| 901 | return dest_len;
|
---|
| 902 | }
|
---|
| 903 |
|
---|
| 904 | /********************************************************************
|
---|
| 905 | Push and malloc an ascii string. src and dest null terminated.
|
---|
| 906 | ********************************************************************/
|
---|
| 907 |
|
---|
| 908 | bool push_ascii_talloc(TALLOC_CTX *mem_ctx, char **dest, const char *src, size_t *converted_size)
|
---|
| 909 | {
|
---|
| 910 | size_t src_len = strlen(src)+1;
|
---|
| 911 |
|
---|
| 912 | *dest = NULL;
|
---|
| 913 | return convert_string_talloc(mem_ctx, CH_UNIX, CH_DOS, src, src_len,
|
---|
| 914 | (void **)dest, converted_size, True);
|
---|
| 915 | }
|
---|
| 916 |
|
---|
| 917 | /**
|
---|
| 918 | * Copy a string from a dos codepage source to a unix char* destination.
|
---|
| 919 | *
|
---|
| 920 | * The resulting string in "dest" is always null terminated.
|
---|
| 921 | *
|
---|
| 922 | * @param flags can have:
|
---|
| 923 | * <dl>
|
---|
| 924 | * <dt>STR_TERMINATE</dt>
|
---|
| 925 | * <dd>STR_TERMINATE means the string in @p src
|
---|
| 926 | * is null terminated, and src_len is ignored.</dd>
|
---|
| 927 | * </dl>
|
---|
| 928 | *
|
---|
| 929 | * @param src_len is the length of the source area in bytes.
|
---|
| 930 | * @returns the number of bytes occupied by the string in @p src.
|
---|
| 931 | **/
|
---|
| 932 | size_t pull_ascii(char *dest, const void *src, size_t dest_len, size_t src_len, int flags)
|
---|
| 933 | {
|
---|
| 934 | size_t ret;
|
---|
| 935 |
|
---|
| 936 | if (dest_len == (size_t)-1) {
|
---|
| 937 | /* No longer allow dest_len of -1. */
|
---|
| 938 | smb_panic("pull_ascii - invalid dest_len of -1");
|
---|
| 939 | }
|
---|
| 940 |
|
---|
| 941 | if (flags & STR_TERMINATE) {
|
---|
| 942 | if (src_len == (size_t)-1) {
|
---|
| 943 | src_len = strlen((const char *)src) + 1;
|
---|
| 944 | } else {
|
---|
| 945 | size_t len = strnlen((const char *)src, src_len);
|
---|
| 946 | if (len < src_len)
|
---|
| 947 | len++;
|
---|
| 948 | src_len = len;
|
---|
| 949 | }
|
---|
| 950 | }
|
---|
| 951 |
|
---|
| 952 | ret = convert_string(CH_DOS, CH_UNIX, src, src_len, dest, dest_len, True);
|
---|
| 953 | if (ret == (size_t)-1) {
|
---|
| 954 | ret = 0;
|
---|
| 955 | dest_len = 0;
|
---|
| 956 | }
|
---|
| 957 |
|
---|
| 958 | if (dest_len && ret) {
|
---|
| 959 | /* Did we already process the terminating zero ? */
|
---|
| 960 | if (dest[MIN(ret-1, dest_len-1)] != 0) {
|
---|
| 961 | dest[MIN(ret, dest_len-1)] = 0;
|
---|
| 962 | }
|
---|
| 963 | } else {
|
---|
| 964 | dest[0] = 0;
|
---|
| 965 | }
|
---|
| 966 |
|
---|
| 967 | return src_len;
|
---|
| 968 | }
|
---|
| 969 |
|
---|
| 970 | /**
|
---|
| 971 | * Copy a string from a dos codepage source to a unix char* destination.
|
---|
| 972 | * Talloc version.
|
---|
| 973 | *
|
---|
| 974 | * The resulting string in "dest" is always null terminated.
|
---|
| 975 | *
|
---|
| 976 | * @param flags can have:
|
---|
| 977 | * <dl>
|
---|
| 978 | * <dt>STR_TERMINATE</dt>
|
---|
| 979 | * <dd>STR_TERMINATE means the string in @p src
|
---|
| 980 | * is null terminated, and src_len is ignored.</dd>
|
---|
| 981 | * </dl>
|
---|
| 982 | *
|
---|
| 983 | * @param src_len is the length of the source area in bytes.
|
---|
| 984 | * @returns the number of bytes occupied by the string in @p src.
|
---|
| 985 | **/
|
---|
| 986 |
|
---|
| 987 | static size_t pull_ascii_base_talloc(TALLOC_CTX *ctx,
|
---|
| 988 | char **ppdest,
|
---|
| 989 | const void *src,
|
---|
| 990 | size_t src_len,
|
---|
| 991 | int flags)
|
---|
| 992 | {
|
---|
| 993 | char *dest = NULL;
|
---|
| 994 | size_t dest_len;
|
---|
| 995 |
|
---|
| 996 | *ppdest = NULL;
|
---|
| 997 |
|
---|
| 998 | if (!src_len) {
|
---|
| 999 | return 0;
|
---|
| 1000 | }
|
---|
| 1001 |
|
---|
| 1002 | if (flags & STR_TERMINATE) {
|
---|
| 1003 | if (src_len == (size_t)-1) {
|
---|
| 1004 | src_len = strlen((const char *)src) + 1;
|
---|
| 1005 | } else {
|
---|
| 1006 | size_t len = strnlen((const char *)src, src_len);
|
---|
| 1007 | if (len < src_len)
|
---|
| 1008 | len++;
|
---|
| 1009 | src_len = len;
|
---|
| 1010 | }
|
---|
| 1011 | /* Ensure we don't use an insane length from the client. */
|
---|
| 1012 | if (src_len >= 1024*1024) {
|
---|
| 1013 | char *msg = talloc_asprintf(ctx,
|
---|
| 1014 | "Bad src length (%u) in "
|
---|
| 1015 | "pull_ascii_base_talloc",
|
---|
| 1016 | (unsigned int)src_len);
|
---|
| 1017 | smb_panic(msg);
|
---|
| 1018 | }
|
---|
| 1019 | } else {
|
---|
| 1020 | /* Can't have an unlimited length
|
---|
| 1021 | * non STR_TERMINATE'd.
|
---|
| 1022 | */
|
---|
| 1023 | if (src_len == (size_t)-1) {
|
---|
| 1024 | errno = EINVAL;
|
---|
| 1025 | return 0;
|
---|
| 1026 | }
|
---|
| 1027 | }
|
---|
| 1028 |
|
---|
| 1029 | /* src_len != -1 here. */
|
---|
| 1030 |
|
---|
| 1031 | if (!convert_string_talloc(ctx, CH_DOS, CH_UNIX, src, src_len, &dest,
|
---|
| 1032 | &dest_len, True)) {
|
---|
| 1033 | dest_len = 0;
|
---|
| 1034 | }
|
---|
| 1035 |
|
---|
| 1036 | if (dest_len && dest) {
|
---|
| 1037 | /* Did we already process the terminating zero ? */
|
---|
| 1038 | if (dest[dest_len-1] != 0) {
|
---|
| 1039 | size_t size = talloc_get_size(dest);
|
---|
| 1040 | /* Have we got space to append the '\0' ? */
|
---|
| 1041 | if (size <= dest_len) {
|
---|
| 1042 | /* No, realloc. */
|
---|
| 1043 | dest = TALLOC_REALLOC_ARRAY(ctx, dest, char,
|
---|
| 1044 | dest_len+1);
|
---|
| 1045 | if (!dest) {
|
---|
| 1046 | /* talloc fail. */
|
---|
| 1047 | dest_len = (size_t)-1;
|
---|
| 1048 | return 0;
|
---|
| 1049 | }
|
---|
| 1050 | }
|
---|
| 1051 | /* Yay - space ! */
|
---|
| 1052 | dest[dest_len] = '\0';
|
---|
| 1053 | dest_len++;
|
---|
| 1054 | }
|
---|
| 1055 | } else if (dest) {
|
---|
| 1056 | dest[0] = 0;
|
---|
| 1057 | }
|
---|
| 1058 |
|
---|
| 1059 | *ppdest = dest;
|
---|
| 1060 | return src_len;
|
---|
| 1061 | }
|
---|
| 1062 |
|
---|
| 1063 | size_t pull_ascii_fstring(char *dest, const void *src)
|
---|
| 1064 | {
|
---|
| 1065 | return pull_ascii(dest, src, sizeof(fstring), -1, STR_TERMINATE);
|
---|
| 1066 | }
|
---|
| 1067 |
|
---|
| 1068 | /* When pulling an nstring it can expand into a larger size (dos cp -> utf8). Cope with this. */
|
---|
| 1069 |
|
---|
| 1070 | size_t pull_ascii_nstring(char *dest, size_t dest_len, const void *src)
|
---|
| 1071 | {
|
---|
| 1072 | return pull_ascii(dest, src, dest_len, sizeof(nstring)-1, STR_TERMINATE);
|
---|
| 1073 | }
|
---|
| 1074 |
|
---|
| 1075 | /**
|
---|
| 1076 | * Copy a string from a char* src to a unicode destination.
|
---|
| 1077 | *
|
---|
| 1078 | * @returns the number of bytes occupied by the string in the destination.
|
---|
| 1079 | *
|
---|
| 1080 | * @param flags can have:
|
---|
| 1081 | *
|
---|
| 1082 | * <dl>
|
---|
| 1083 | * <dt>STR_TERMINATE <dd>means include the null termination.
|
---|
| 1084 | * <dt>STR_UPPER <dd>means uppercase in the destination.
|
---|
| 1085 | * <dt>STR_NOALIGN <dd>means don't do alignment.
|
---|
| 1086 | * </dl>
|
---|
| 1087 | *
|
---|
| 1088 | * @param dest_len is the maximum length allowed in the
|
---|
| 1089 | * destination.
|
---|
| 1090 | **/
|
---|
| 1091 |
|
---|
| 1092 | size_t push_ucs2(const void *base_ptr, void *dest, const char *src, size_t dest_len, int flags)
|
---|
| 1093 | {
|
---|
| 1094 | size_t len=0;
|
---|
| 1095 | size_t src_len;
|
---|
| 1096 | size_t ret;
|
---|
| 1097 |
|
---|
| 1098 | if (dest_len == (size_t)-1) {
|
---|
| 1099 | /* No longer allow dest_len of -1. */
|
---|
| 1100 | smb_panic("push_ucs2 - invalid dest_len of -1");
|
---|
| 1101 | }
|
---|
| 1102 |
|
---|
| 1103 | if (flags & STR_TERMINATE)
|
---|
| 1104 | src_len = (size_t)-1;
|
---|
| 1105 | else
|
---|
| 1106 | src_len = strlen(src);
|
---|
| 1107 |
|
---|
| 1108 | if (ucs2_align(base_ptr, dest, flags)) {
|
---|
| 1109 | *(char *)dest = 0;
|
---|
| 1110 | dest = (void *)((char *)dest + 1);
|
---|
| 1111 | if (dest_len)
|
---|
| 1112 | dest_len--;
|
---|
| 1113 | len++;
|
---|
| 1114 | }
|
---|
| 1115 |
|
---|
| 1116 | /* ucs2 is always a multiple of 2 bytes */
|
---|
| 1117 | dest_len &= ~1;
|
---|
| 1118 |
|
---|
| 1119 | ret = convert_string(CH_UNIX, CH_UTF16LE, src, src_len, dest, dest_len, True);
|
---|
| 1120 | if (ret == (size_t)-1) {
|
---|
| 1121 | if ((flags & STR_TERMINATE) &&
|
---|
| 1122 | dest &&
|
---|
| 1123 | dest_len) {
|
---|
| 1124 | *(char *)dest = 0;
|
---|
| 1125 | }
|
---|
| 1126 | return len;
|
---|
| 1127 | }
|
---|
| 1128 |
|
---|
| 1129 | len += ret;
|
---|
| 1130 |
|
---|
| 1131 | if (flags & STR_UPPER) {
|
---|
| 1132 | smb_ucs2_t *dest_ucs2 = (smb_ucs2_t *)dest;
|
---|
| 1133 | size_t i;
|
---|
| 1134 |
|
---|
| 1135 | /* We check for i < (ret / 2) below as the dest string isn't null
|
---|
| 1136 | terminated if STR_TERMINATE isn't set. */
|
---|
| 1137 |
|
---|
| 1138 | for (i = 0; i < (ret / 2) && i < (dest_len / 2) && dest_ucs2[i]; i++) {
|
---|
| 1139 | smb_ucs2_t v = toupper_w(dest_ucs2[i]);
|
---|
| 1140 | if (v != dest_ucs2[i]) {
|
---|
| 1141 | dest_ucs2[i] = v;
|
---|
| 1142 | }
|
---|
| 1143 | }
|
---|
| 1144 | }
|
---|
| 1145 |
|
---|
| 1146 | return len;
|
---|
| 1147 | }
|
---|
| 1148 |
|
---|
| 1149 |
|
---|
| 1150 | /**
|
---|
| 1151 | * Copy a string from a unix char* src to a UCS2 destination,
|
---|
| 1152 | * allocating a buffer using talloc().
|
---|
| 1153 | *
|
---|
| 1154 | * @param dest always set at least to NULL
|
---|
| 1155 | * @parm converted_size set to the number of bytes occupied by the string in
|
---|
| 1156 | * the destination on success.
|
---|
| 1157 | *
|
---|
| 1158 | * @return true if new buffer was correctly allocated, and string was
|
---|
| 1159 | * converted.
|
---|
| 1160 | **/
|
---|
| 1161 | bool push_ucs2_talloc(TALLOC_CTX *ctx, smb_ucs2_t **dest, const char *src,
|
---|
| 1162 | size_t *converted_size)
|
---|
| 1163 | {
|
---|
| 1164 | size_t src_len = strlen(src)+1;
|
---|
| 1165 |
|
---|
| 1166 | *dest = NULL;
|
---|
| 1167 | return convert_string_talloc(ctx, CH_UNIX, CH_UTF16LE, src, src_len,
|
---|
| 1168 | (void **)dest, converted_size, True);
|
---|
| 1169 | }
|
---|
| 1170 |
|
---|
| 1171 |
|
---|
| 1172 | /**
|
---|
| 1173 | Copy a string from a char* src to a UTF-8 destination.
|
---|
| 1174 | Return the number of bytes occupied by the string in the destination
|
---|
| 1175 | Flags can have:
|
---|
| 1176 | STR_TERMINATE means include the null termination
|
---|
| 1177 | STR_UPPER means uppercase in the destination
|
---|
| 1178 | dest_len is the maximum length allowed in the destination. If dest_len
|
---|
| 1179 | is -1 then no maxiumum is used.
|
---|
| 1180 | **/
|
---|
| 1181 |
|
---|
| 1182 | static size_t push_utf8(void *dest, const char *src, size_t dest_len, int flags)
|
---|
| 1183 | {
|
---|
| 1184 | size_t src_len = 0;
|
---|
| 1185 | size_t ret;
|
---|
| 1186 | char *tmpbuf = NULL;
|
---|
| 1187 |
|
---|
| 1188 | if (dest_len == (size_t)-1) {
|
---|
| 1189 | /* No longer allow dest_len of -1. */
|
---|
| 1190 | smb_panic("push_utf8 - invalid dest_len of -1");
|
---|
| 1191 | }
|
---|
| 1192 |
|
---|
| 1193 | if (flags & STR_UPPER) {
|
---|
| 1194 | tmpbuf = strupper_talloc(talloc_tos(), src);
|
---|
| 1195 | if (!tmpbuf) {
|
---|
| 1196 | return (size_t)-1;
|
---|
| 1197 | }
|
---|
| 1198 | src = tmpbuf;
|
---|
| 1199 | src_len = strlen(src);
|
---|
| 1200 | }
|
---|
| 1201 |
|
---|
| 1202 | src_len = strlen(src);
|
---|
| 1203 | if (flags & STR_TERMINATE) {
|
---|
| 1204 | src_len++;
|
---|
| 1205 | }
|
---|
| 1206 |
|
---|
| 1207 | ret = convert_string(CH_UNIX, CH_UTF8, src, src_len, dest, dest_len, True);
|
---|
| 1208 | TALLOC_FREE(tmpbuf);
|
---|
| 1209 | return ret;
|
---|
| 1210 | }
|
---|
| 1211 |
|
---|
| 1212 | size_t push_utf8_fstring(void *dest, const char *src)
|
---|
| 1213 | {
|
---|
| 1214 | return push_utf8(dest, src, sizeof(fstring), STR_TERMINATE);
|
---|
| 1215 | }
|
---|
| 1216 |
|
---|
| 1217 | /**
|
---|
| 1218 | * Copy a string from a unix char* src to a UTF-8 destination, allocating a buffer using talloc
|
---|
| 1219 | *
|
---|
| 1220 | * @param dest always set at least to NULL
|
---|
| 1221 | * @parm converted_size set to the number of bytes occupied by the string in
|
---|
| 1222 | * the destination on success.
|
---|
| 1223 | *
|
---|
| 1224 | * @return true if new buffer was correctly allocated, and string was
|
---|
| 1225 | * converted.
|
---|
| 1226 | **/
|
---|
| 1227 |
|
---|
| 1228 | bool push_utf8_talloc(TALLOC_CTX *ctx, char **dest, const char *src,
|
---|
| 1229 | size_t *converted_size)
|
---|
| 1230 | {
|
---|
| 1231 | size_t src_len = strlen(src)+1;
|
---|
| 1232 |
|
---|
| 1233 | *dest = NULL;
|
---|
| 1234 | return convert_string_talloc(ctx, CH_UNIX, CH_UTF8, src, src_len,
|
---|
| 1235 | (void**)dest, converted_size, True);
|
---|
| 1236 | }
|
---|
| 1237 |
|
---|
| 1238 | /**
|
---|
| 1239 | Copy a string from a ucs2 source to a unix char* destination.
|
---|
| 1240 | Flags can have:
|
---|
| 1241 | STR_TERMINATE means the string in src is null terminated.
|
---|
| 1242 | STR_NOALIGN means don't try to align.
|
---|
| 1243 | if STR_TERMINATE is set then src_len is ignored if it is -1.
|
---|
| 1244 | src_len is the length of the source area in bytes
|
---|
| 1245 | Return the number of bytes occupied by the string in src.
|
---|
| 1246 | The resulting string in "dest" is always null terminated.
|
---|
| 1247 | **/
|
---|
| 1248 |
|
---|
| 1249 | size_t pull_ucs2(const void *base_ptr, char *dest, const void *src, size_t dest_len, size_t src_len, int flags)
|
---|
| 1250 | {
|
---|
| 1251 | size_t ret;
|
---|
| 1252 | size_t ucs2_align_len = 0;
|
---|
| 1253 |
|
---|
| 1254 | if (dest_len == (size_t)-1) {
|
---|
| 1255 | /* No longer allow dest_len of -1. */
|
---|
| 1256 | smb_panic("pull_ucs2 - invalid dest_len of -1");
|
---|
| 1257 | }
|
---|
| 1258 |
|
---|
| 1259 | if (!src_len) {
|
---|
| 1260 | if (dest && dest_len > 0) {
|
---|
| 1261 | dest[0] = '\0';
|
---|
| 1262 | }
|
---|
| 1263 | return 0;
|
---|
| 1264 | }
|
---|
| 1265 |
|
---|
| 1266 | if (ucs2_align(base_ptr, src, flags)) {
|
---|
| 1267 | src = (const void *)((const char *)src + 1);
|
---|
| 1268 | if (src_len != (size_t)-1)
|
---|
| 1269 | src_len--;
|
---|
| 1270 | ucs2_align_len = 1;
|
---|
| 1271 | }
|
---|
| 1272 |
|
---|
| 1273 | if (flags & STR_TERMINATE) {
|
---|
| 1274 | /* src_len -1 is the default for null terminated strings. */
|
---|
| 1275 | if (src_len != (size_t)-1) {
|
---|
| 1276 | size_t len = strnlen_w((const smb_ucs2_t *)src,
|
---|
| 1277 | src_len/2);
|
---|
| 1278 | if (len < src_len/2)
|
---|
| 1279 | len++;
|
---|
| 1280 | src_len = len*2;
|
---|
| 1281 | }
|
---|
| 1282 | }
|
---|
| 1283 |
|
---|
| 1284 | /* ucs2 is always a multiple of 2 bytes */
|
---|
| 1285 | if (src_len != (size_t)-1)
|
---|
| 1286 | src_len &= ~1;
|
---|
| 1287 |
|
---|
| 1288 | ret = convert_string(CH_UTF16LE, CH_UNIX, src, src_len, dest, dest_len, True);
|
---|
| 1289 | if (ret == (size_t)-1) {
|
---|
| 1290 | ret = 0;
|
---|
| 1291 | dest_len = 0;
|
---|
| 1292 | }
|
---|
| 1293 |
|
---|
| 1294 | if (src_len == (size_t)-1)
|
---|
| 1295 | src_len = ret*2;
|
---|
| 1296 |
|
---|
| 1297 | if (dest_len && ret) {
|
---|
| 1298 | /* Did we already process the terminating zero ? */
|
---|
| 1299 | if (dest[MIN(ret-1, dest_len-1)] != 0) {
|
---|
| 1300 | dest[MIN(ret, dest_len-1)] = 0;
|
---|
| 1301 | }
|
---|
| 1302 | } else {
|
---|
| 1303 | dest[0] = 0;
|
---|
| 1304 | }
|
---|
| 1305 |
|
---|
| 1306 | return src_len + ucs2_align_len;
|
---|
| 1307 | }
|
---|
| 1308 |
|
---|
| 1309 | /**
|
---|
| 1310 | Copy a string from a ucs2 source to a unix char* destination.
|
---|
| 1311 | Talloc version with a base pointer.
|
---|
| 1312 | Uses malloc if TALLOC_CTX is NULL (this is a bad interface and
|
---|
| 1313 | needs fixing. JRA).
|
---|
| 1314 | Flags can have:
|
---|
| 1315 | STR_TERMINATE means the string in src is null terminated.
|
---|
| 1316 | STR_NOALIGN means don't try to align.
|
---|
| 1317 | if STR_TERMINATE is set then src_len is ignored if it is -1.
|
---|
| 1318 | src_len is the length of the source area in bytes
|
---|
| 1319 | Return the number of bytes occupied by the string in src.
|
---|
| 1320 | The resulting string in "dest" is always null terminated.
|
---|
| 1321 | **/
|
---|
| 1322 |
|
---|
| 1323 | size_t pull_ucs2_base_talloc(TALLOC_CTX *ctx,
|
---|
| 1324 | const void *base_ptr,
|
---|
| 1325 | char **ppdest,
|
---|
| 1326 | const void *src,
|
---|
| 1327 | size_t src_len,
|
---|
| 1328 | int flags)
|
---|
| 1329 | {
|
---|
| 1330 | char *dest;
|
---|
| 1331 | size_t dest_len;
|
---|
| 1332 | size_t ucs2_align_len = 0;
|
---|
| 1333 |
|
---|
| 1334 | *ppdest = NULL;
|
---|
| 1335 |
|
---|
| 1336 | #ifdef DEVELOPER
|
---|
| 1337 | /* Ensure we never use the braindead "malloc" varient. */
|
---|
| 1338 | if (ctx == NULL) {
|
---|
| 1339 | smb_panic("NULL talloc CTX in pull_ucs2_base_talloc\n");
|
---|
| 1340 | }
|
---|
| 1341 | #endif
|
---|
| 1342 |
|
---|
| 1343 | if (!src_len) {
|
---|
| 1344 | return 0;
|
---|
| 1345 | }
|
---|
| 1346 |
|
---|
| 1347 | if (ucs2_align(base_ptr, src, flags)) {
|
---|
| 1348 | src = (const void *)((const char *)src + 1);
|
---|
| 1349 | if (src_len != (size_t)-1)
|
---|
| 1350 | src_len--;
|
---|
| 1351 | ucs2_align_len = 1;
|
---|
| 1352 | }
|
---|
| 1353 |
|
---|
| 1354 | if (flags & STR_TERMINATE) {
|
---|
| 1355 | /* src_len -1 is the default for null terminated strings. */
|
---|
| 1356 | if (src_len != (size_t)-1) {
|
---|
| 1357 | size_t len = strnlen_w((const smb_ucs2_t *)src,
|
---|
| 1358 | src_len/2);
|
---|
| 1359 | if (len < src_len/2)
|
---|
| 1360 | len++;
|
---|
| 1361 | src_len = len*2;
|
---|
| 1362 | } else {
|
---|
| 1363 | /*
|
---|
| 1364 | * src_len == -1 - alloc interface won't take this
|
---|
| 1365 | * so we must calculate.
|
---|
| 1366 | */
|
---|
| 1367 | src_len = (strlen_w((const smb_ucs2_t *)src)+1)*sizeof(smb_ucs2_t);
|
---|
| 1368 | }
|
---|
| 1369 | /* Ensure we don't use an insane length from the client. */
|
---|
| 1370 | if (src_len >= 1024*1024) {
|
---|
| 1371 | smb_panic("Bad src length in pull_ucs2_base_talloc\n");
|
---|
| 1372 | }
|
---|
| 1373 | } else {
|
---|
| 1374 | /* Can't have an unlimited length
|
---|
| 1375 | * non STR_TERMINATE'd.
|
---|
| 1376 | */
|
---|
| 1377 | if (src_len == (size_t)-1) {
|
---|
| 1378 | errno = EINVAL;
|
---|
| 1379 | return 0;
|
---|
| 1380 | }
|
---|
| 1381 | }
|
---|
| 1382 |
|
---|
| 1383 | /* src_len != -1 here. */
|
---|
| 1384 |
|
---|
| 1385 | /* ucs2 is always a multiple of 2 bytes */
|
---|
| 1386 | src_len &= ~1;
|
---|
| 1387 |
|
---|
| 1388 | if (!convert_string_talloc(ctx, CH_UTF16LE, CH_UNIX, src, src_len,
|
---|
| 1389 | (void *)&dest, &dest_len, True)) {
|
---|
| 1390 | dest_len = 0;
|
---|
| 1391 | }
|
---|
| 1392 |
|
---|
| 1393 | if (dest_len) {
|
---|
| 1394 | /* Did we already process the terminating zero ? */
|
---|
| 1395 | if (dest[dest_len-1] != 0) {
|
---|
| 1396 | size_t size = talloc_get_size(dest);
|
---|
| 1397 | /* Have we got space to append the '\0' ? */
|
---|
| 1398 | if (size <= dest_len) {
|
---|
| 1399 | /* No, realloc. */
|
---|
| 1400 | dest = TALLOC_REALLOC_ARRAY(ctx, dest, char,
|
---|
| 1401 | dest_len+1);
|
---|
| 1402 | if (!dest) {
|
---|
| 1403 | /* talloc fail. */
|
---|
| 1404 | dest_len = (size_t)-1;
|
---|
| 1405 | return 0;
|
---|
| 1406 | }
|
---|
| 1407 | }
|
---|
| 1408 | /* Yay - space ! */
|
---|
| 1409 | dest[dest_len] = '\0';
|
---|
| 1410 | dest_len++;
|
---|
| 1411 | }
|
---|
| 1412 | } else if (dest) {
|
---|
| 1413 | dest[0] = 0;
|
---|
| 1414 | }
|
---|
| 1415 |
|
---|
| 1416 | *ppdest = dest;
|
---|
| 1417 | return src_len + ucs2_align_len;
|
---|
| 1418 | }
|
---|
| 1419 |
|
---|
| 1420 | size_t pull_ucs2_fstring(char *dest, const void *src)
|
---|
| 1421 | {
|
---|
| 1422 | return pull_ucs2(NULL, dest, src, sizeof(fstring), -1, STR_TERMINATE);
|
---|
| 1423 | }
|
---|
| 1424 |
|
---|
| 1425 | /**
|
---|
| 1426 | * Copy a string from a UCS2 src to a unix char * destination, allocating a buffer using talloc
|
---|
| 1427 | *
|
---|
| 1428 | * @param dest always set at least to NULL
|
---|
| 1429 | * @parm converted_size set to the number of bytes occupied by the string in
|
---|
| 1430 | * the destination on success.
|
---|
| 1431 | *
|
---|
| 1432 | * @return true if new buffer was correctly allocated, and string was
|
---|
| 1433 | * converted.
|
---|
| 1434 | **/
|
---|
| 1435 |
|
---|
| 1436 | bool pull_ucs2_talloc(TALLOC_CTX *ctx, char **dest, const smb_ucs2_t *src,
|
---|
| 1437 | size_t *converted_size)
|
---|
| 1438 | {
|
---|
| 1439 | size_t src_len = (strlen_w(src)+1) * sizeof(smb_ucs2_t);
|
---|
| 1440 |
|
---|
| 1441 | *dest = NULL;
|
---|
| 1442 | return convert_string_talloc(ctx, CH_UTF16LE, CH_UNIX, src, src_len,
|
---|
| 1443 | (void **)dest, converted_size, True);
|
---|
| 1444 | }
|
---|
| 1445 |
|
---|
| 1446 | /**
|
---|
| 1447 | * Copy a string from a UTF-8 src to a unix char * destination, allocating a buffer using talloc
|
---|
| 1448 | *
|
---|
| 1449 | * @param dest always set at least to NULL
|
---|
| 1450 | * @parm converted_size set to the number of bytes occupied by the string in
|
---|
| 1451 | * the destination on success.
|
---|
| 1452 | *
|
---|
| 1453 | * @return true if new buffer was correctly allocated, and string was
|
---|
| 1454 | * converted.
|
---|
| 1455 | **/
|
---|
| 1456 |
|
---|
| 1457 | bool pull_utf8_talloc(TALLOC_CTX *ctx, char **dest, const char *src,
|
---|
| 1458 | size_t *converted_size)
|
---|
| 1459 | {
|
---|
| 1460 | size_t src_len = strlen(src)+1;
|
---|
| 1461 |
|
---|
| 1462 | *dest = NULL;
|
---|
| 1463 | return convert_string_talloc(ctx, CH_UTF8, CH_UNIX, src, src_len,
|
---|
| 1464 | (void **)dest, converted_size, True);
|
---|
| 1465 | }
|
---|
| 1466 |
|
---|
| 1467 |
|
---|
| 1468 | /**
|
---|
| 1469 | * Copy a string from a DOS src to a unix char * destination, allocating a buffer using talloc
|
---|
| 1470 | *
|
---|
| 1471 | * @param dest always set at least to NULL
|
---|
| 1472 | * @parm converted_size set to the number of bytes occupied by the string in
|
---|
| 1473 | * the destination on success.
|
---|
| 1474 | *
|
---|
| 1475 | * @return true if new buffer was correctly allocated, and string was
|
---|
| 1476 | * converted.
|
---|
| 1477 | **/
|
---|
| 1478 |
|
---|
| 1479 | bool pull_ascii_talloc(TALLOC_CTX *ctx, char **dest, const char *src,
|
---|
| 1480 | size_t *converted_size)
|
---|
| 1481 | {
|
---|
| 1482 | size_t src_len = strlen(src)+1;
|
---|
| 1483 |
|
---|
| 1484 | *dest = NULL;
|
---|
| 1485 | return convert_string_talloc(ctx, CH_DOS, CH_UNIX, src, src_len,
|
---|
| 1486 | (void **)dest, converted_size, True);
|
---|
| 1487 | }
|
---|
| 1488 |
|
---|
| 1489 | /**
|
---|
| 1490 | Copy a string from a char* src to a unicode or ascii
|
---|
| 1491 | dos codepage destination choosing unicode or ascii based on the
|
---|
| 1492 | flags supplied
|
---|
| 1493 | Return the number of bytes occupied by the string in the destination.
|
---|
| 1494 | flags can have:
|
---|
| 1495 | STR_TERMINATE means include the null termination.
|
---|
| 1496 | STR_UPPER means uppercase in the destination.
|
---|
| 1497 | STR_ASCII use ascii even with unicode packet.
|
---|
| 1498 | STR_NOALIGN means don't do alignment.
|
---|
| 1499 | dest_len is the maximum length allowed in the destination. If dest_len
|
---|
| 1500 | is -1 then no maxiumum is used.
|
---|
| 1501 | **/
|
---|
| 1502 |
|
---|
| 1503 | size_t push_string_check_fn(const char *function, unsigned int line,
|
---|
| 1504 | void *dest, const char *src,
|
---|
| 1505 | size_t dest_len, int flags)
|
---|
| 1506 | {
|
---|
| 1507 | #ifdef DEVELOPER
|
---|
| 1508 | /* We really need to zero fill here, not clobber
|
---|
| 1509 | * region, as we want to ensure that valgrind thinks
|
---|
| 1510 | * all of the outgoing buffer has been written to
|
---|
| 1511 | * so a send() or write() won't trap an error.
|
---|
| 1512 | * JRA.
|
---|
| 1513 | */
|
---|
| 1514 | #if 0
|
---|
| 1515 | clobber_region(function, line, dest, dest_len);
|
---|
| 1516 | #else
|
---|
| 1517 | memset(dest, '\0', dest_len);
|
---|
| 1518 | #endif
|
---|
| 1519 | #endif
|
---|
| 1520 |
|
---|
| 1521 | if (!(flags & STR_ASCII) && (flags & STR_UNICODE)) {
|
---|
| 1522 | return push_ucs2(NULL, dest, src, dest_len, flags);
|
---|
| 1523 | }
|
---|
| 1524 | return push_ascii(dest, src, dest_len, flags);
|
---|
| 1525 | }
|
---|
| 1526 |
|
---|
| 1527 |
|
---|
| 1528 | /**
|
---|
| 1529 | Copy a string from a char* src to a unicode or ascii
|
---|
| 1530 | dos codepage destination choosing unicode or ascii based on the
|
---|
| 1531 | flags in the SMB buffer starting at base_ptr.
|
---|
| 1532 | Return the number of bytes occupied by the string in the destination.
|
---|
| 1533 | flags can have:
|
---|
| 1534 | STR_TERMINATE means include the null termination.
|
---|
| 1535 | STR_UPPER means uppercase in the destination.
|
---|
| 1536 | STR_ASCII use ascii even with unicode packet.
|
---|
| 1537 | STR_NOALIGN means don't do alignment.
|
---|
| 1538 | dest_len is the maximum length allowed in the destination. If dest_len
|
---|
| 1539 | is -1 then no maxiumum is used.
|
---|
| 1540 | **/
|
---|
| 1541 |
|
---|
| 1542 | size_t push_string_base(const char *function, unsigned int line,
|
---|
| 1543 | const char *base, uint16 flags2,
|
---|
| 1544 | void *dest, const char *src,
|
---|
| 1545 | size_t dest_len, int flags)
|
---|
| 1546 | {
|
---|
| 1547 | #ifdef DEVELOPER
|
---|
| 1548 | /* We really need to zero fill here, not clobber
|
---|
| 1549 | * region, as we want to ensure that valgrind thinks
|
---|
| 1550 | * all of the outgoing buffer has been written to
|
---|
| 1551 | * so a send() or write() won't trap an error.
|
---|
| 1552 | * JRA.
|
---|
| 1553 | */
|
---|
| 1554 | #if 0
|
---|
| 1555 | clobber_region(function, line, dest, dest_len);
|
---|
| 1556 | #else
|
---|
| 1557 | memset(dest, '\0', dest_len);
|
---|
| 1558 | #endif
|
---|
| 1559 | #endif
|
---|
| 1560 |
|
---|
| 1561 | if (!(flags & STR_ASCII) && \
|
---|
| 1562 | ((flags & STR_UNICODE || \
|
---|
| 1563 | (flags2 & FLAGS2_UNICODE_STRINGS)))) {
|
---|
| 1564 | return push_ucs2(base, dest, src, dest_len, flags);
|
---|
| 1565 | }
|
---|
| 1566 | return push_ascii(dest, src, dest_len, flags);
|
---|
| 1567 | }
|
---|
| 1568 |
|
---|
| 1569 | /**
|
---|
| 1570 | Copy a string from a char* src to a unicode or ascii
|
---|
| 1571 | dos codepage destination choosing unicode or ascii based on the
|
---|
| 1572 | flags supplied
|
---|
| 1573 | Return the number of bytes occupied by the string in the destination.
|
---|
| 1574 | flags can have:
|
---|
| 1575 | STR_TERMINATE means include the null termination.
|
---|
| 1576 | STR_UPPER means uppercase in the destination.
|
---|
| 1577 | STR_ASCII use ascii even with unicode packet.
|
---|
| 1578 | STR_NOALIGN means don't do alignment.
|
---|
| 1579 | dest_len is the maximum length allowed in the destination. If dest_len
|
---|
| 1580 | is -1 then no maxiumum is used.
|
---|
| 1581 | **/
|
---|
| 1582 |
|
---|
| 1583 | ssize_t push_string(void *dest, const char *src, size_t dest_len, int flags)
|
---|
| 1584 | {
|
---|
| 1585 | size_t ret;
|
---|
| 1586 | #ifdef DEVELOPER
|
---|
| 1587 | /* We really need to zero fill here, not clobber
|
---|
| 1588 | * region, as we want to ensure that valgrind thinks
|
---|
| 1589 | * all of the outgoing buffer has been written to
|
---|
| 1590 | * so a send() or write() won't trap an error.
|
---|
| 1591 | * JRA.
|
---|
| 1592 | */
|
---|
| 1593 | memset(dest, '\0', dest_len);
|
---|
| 1594 | #endif
|
---|
| 1595 |
|
---|
| 1596 | if (!(flags & STR_ASCII) && \
|
---|
| 1597 | (flags & STR_UNICODE)) {
|
---|
| 1598 | ret = push_ucs2(NULL, dest, src, dest_len, flags);
|
---|
| 1599 | } else {
|
---|
| 1600 | ret = push_ascii(dest, src, dest_len, flags);
|
---|
| 1601 | }
|
---|
| 1602 | if (ret == (size_t)-1) {
|
---|
| 1603 | return -1;
|
---|
| 1604 | }
|
---|
| 1605 | return ret;
|
---|
| 1606 | }
|
---|
| 1607 |
|
---|
| 1608 | /**
|
---|
| 1609 | Copy a string from a unicode or ascii source (depending on
|
---|
| 1610 | the packet flags) to a char* destination.
|
---|
| 1611 | Flags can have:
|
---|
| 1612 | STR_TERMINATE means the string in src is null terminated.
|
---|
| 1613 | STR_UNICODE means to force as unicode.
|
---|
| 1614 | STR_ASCII use ascii even with unicode packet.
|
---|
| 1615 | STR_NOALIGN means don't do alignment.
|
---|
| 1616 | if STR_TERMINATE is set then src_len is ignored is it is -1
|
---|
| 1617 | src_len is the length of the source area in bytes.
|
---|
| 1618 | Return the number of bytes occupied by the string in src.
|
---|
| 1619 | The resulting string in "dest" is always null terminated.
|
---|
| 1620 | **/
|
---|
| 1621 |
|
---|
| 1622 | size_t pull_string_fn(const char *function,
|
---|
| 1623 | unsigned int line,
|
---|
| 1624 | const void *base_ptr,
|
---|
| 1625 | uint16 smb_flags2,
|
---|
| 1626 | char *dest,
|
---|
| 1627 | const void *src,
|
---|
| 1628 | size_t dest_len,
|
---|
| 1629 | size_t src_len,
|
---|
| 1630 | int flags)
|
---|
| 1631 | {
|
---|
| 1632 | #ifdef DEVELOPER
|
---|
| 1633 | clobber_region(function, line, dest, dest_len);
|
---|
| 1634 | #endif
|
---|
| 1635 |
|
---|
| 1636 | if ((base_ptr == NULL) && ((flags & (STR_ASCII|STR_UNICODE)) == 0)) {
|
---|
| 1637 | smb_panic("No base ptr to get flg2 and neither ASCII nor "
|
---|
| 1638 | "UNICODE defined");
|
---|
| 1639 | }
|
---|
| 1640 |
|
---|
| 1641 | if (!(flags & STR_ASCII) && \
|
---|
| 1642 | ((flags & STR_UNICODE || \
|
---|
| 1643 | (smb_flags2 & FLAGS2_UNICODE_STRINGS)))) {
|
---|
| 1644 | return pull_ucs2(base_ptr, dest, src, dest_len, src_len, flags);
|
---|
| 1645 | }
|
---|
| 1646 | return pull_ascii(dest, src, dest_len, src_len, flags);
|
---|
| 1647 | }
|
---|
| 1648 |
|
---|
| 1649 | /**
|
---|
| 1650 | Copy a string from a unicode or ascii source (depending on
|
---|
| 1651 | the packet flags) to a char* destination.
|
---|
| 1652 | Variant that uses talloc.
|
---|
| 1653 | Flags can have:
|
---|
| 1654 | STR_TERMINATE means the string in src is null terminated.
|
---|
| 1655 | STR_UNICODE means to force as unicode.
|
---|
| 1656 | STR_ASCII use ascii even with unicode packet.
|
---|
| 1657 | STR_NOALIGN means don't do alignment.
|
---|
| 1658 | if STR_TERMINATE is set then src_len is ignored is it is -1
|
---|
| 1659 | src_len is the length of the source area in bytes.
|
---|
| 1660 | Return the number of bytes occupied by the string in src.
|
---|
| 1661 | The resulting string in "dest" is always null terminated.
|
---|
| 1662 | **/
|
---|
| 1663 |
|
---|
| 1664 | size_t pull_string_talloc_fn(const char *function,
|
---|
| 1665 | unsigned int line,
|
---|
| 1666 | TALLOC_CTX *ctx,
|
---|
| 1667 | const void *base_ptr,
|
---|
| 1668 | uint16 smb_flags2,
|
---|
| 1669 | char **ppdest,
|
---|
| 1670 | const void *src,
|
---|
| 1671 | size_t src_len,
|
---|
| 1672 | int flags)
|
---|
| 1673 | {
|
---|
| 1674 | if ((base_ptr == NULL) && ((flags & (STR_ASCII|STR_UNICODE)) == 0)) {
|
---|
| 1675 | smb_panic("No base ptr to get flg2 and neither ASCII nor "
|
---|
| 1676 | "UNICODE defined");
|
---|
| 1677 | }
|
---|
| 1678 |
|
---|
| 1679 | if (!(flags & STR_ASCII) && \
|
---|
| 1680 | ((flags & STR_UNICODE || \
|
---|
| 1681 | (smb_flags2 & FLAGS2_UNICODE_STRINGS)))) {
|
---|
| 1682 | return pull_ucs2_base_talloc(ctx,
|
---|
| 1683 | base_ptr,
|
---|
| 1684 | ppdest,
|
---|
| 1685 | src,
|
---|
| 1686 | src_len,
|
---|
| 1687 | flags);
|
---|
| 1688 | }
|
---|
| 1689 | return pull_ascii_base_talloc(ctx,
|
---|
| 1690 | ppdest,
|
---|
| 1691 | src,
|
---|
| 1692 | src_len,
|
---|
| 1693 | flags);
|
---|
| 1694 | }
|
---|
| 1695 |
|
---|
| 1696 |
|
---|
| 1697 | size_t align_string(const void *base_ptr, const char *p, int flags)
|
---|
| 1698 | {
|
---|
| 1699 | if (!(flags & STR_ASCII) && \
|
---|
| 1700 | ((flags & STR_UNICODE || \
|
---|
| 1701 | (SVAL(base_ptr, smb_flg2) & FLAGS2_UNICODE_STRINGS)))) {
|
---|
| 1702 | return ucs2_align(base_ptr, p, flags);
|
---|
| 1703 | }
|
---|
| 1704 | return 0;
|
---|
| 1705 | }
|
---|
| 1706 |
|
---|