iconv/loop.c

   1 /* Conversion loop frame work.
   2    Copyright (C) 1998-2014 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4    Contributed by Ulrich Drepper <drepper@cygnus.com>, 1998.
   5
   6    The GNU C 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    The GNU C Library is distributed in the hope that it will be useful,
  12    but WITHOUT ANY WARRANTY; without even the implied warranty of
  13    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14    Lesser General Public License for more details.
  15
  16    You should have received a copy of the GNU Lesser General Public
  17    License along with the GNU C Library; if not, see
  18    <http://www.gnu.org/licenses/>.  */
  19
  20 /* This file provides a frame for the reader loop in all conversion modules.
  21    The actual code must (of course) be provided in the actual module source
  22    code but certain actions can be written down generically, with some
  23    customization options which are these:
  24
  25      MIN_NEEDED_INPUT   minimal number of input bytes needed for the next
  26                         conversion.
  27      MIN_NEEDED_OUTPUT  minimal number of bytes produced by the next round
  28                         of conversion.
  29
  30      MAX_NEEDED_INPUT   you guess it, this is the maximal number of input
  31                         bytes needed.  It defaults to MIN_NEEDED_INPUT
  32      MAX_NEEDED_OUTPUT  likewise for output bytes.
  33
  34      LOOPFCT            name of the function created.  If not specified
  35                         the name is `loop' but this prevents the use
  36                         of multiple functions in the same file.
  37
  38      BODY               this is supposed to expand to the body of the loop.
  39                         The user must provide this.
  40
  41      EXTRA_LOOP_DECLS   extra arguments passed from conversion loop call.
  42
  43      INIT_PARAMS        code to define and initialize variables from params.
  44      UPDATE_PARAMS      code to store result in params.
  45
  46      ONEBYTE_BODY       body of the specialized conversion function for a
  47                         single byte from the current character set to INTERNAL.
  48 */
  49
  50 #include <assert.h>
  51 #include <endian.h>
  52 #include <gconv.h>
  53 #include <stdint.h>
  54 #include <string.h>
  55 #include <wchar.h>
  56 #include <sys/param.h>          /* For MIN.  */
  57 #define __need_size_t
  58 #include <stddef.h>
  59
  60
  61 /* We have to provide support for machines which are not able to handled
  62    unaligned memory accesses.  Some of the character encodings have
  63    representations with a fixed width of 2 or 4 bytes.  But if we cannot
  64    access unaligned memory we still have to read byte-wise.  */
  65 #undef FCTNAME2
  66 #if defined _STRING_ARCH_unaligned || !defined DEFINE_UNALIGNED
  67 /* We can handle unaligned memory access.  */
  68 # define get16(addr) *((const uint16_t *) (addr))
  69 # define get32(addr) *((const uint32_t *) (addr))
  70
  71 /* We need no special support for writing values either.  */
  72 # define put16(addr, val) *((uint16_t *) (addr)) = (val)
  73 # define put32(addr, val) *((uint32_t *) (addr)) = (val)
  74
  75 # define FCTNAME2(name) name
  76 #else
  77 /* Distinguish between big endian and little endian.  */
  78 # if __BYTE_ORDER == __LITTLE_ENDIAN
  79 #  define get16(addr) \
  80      (((const unsigned char *) (addr))[1] << 8                                \
  81       | ((const unsigned char *) (addr))[0])
  82 #  define get32(addr) \
  83      (((((const unsigned char *) (addr))[3] << 8                              \
  84         | ((const unsigned char *) (addr))[2]) << 8                           \
  85        | ((const unsigned char *) (addr))[1]) << 8                            \
  86       | ((const unsigned char *) (addr))[0])
  87
  88 #  define put16(addr, val) \
  89      ({ uint16_t __val = (val);                                               \
  90         ((unsigned char *) (addr))[0] = __val;                                \
  91         ((unsigned char *) (addr))[1] = __val >> 8;                           \
  92         (void) 0; })
  93 #  define put32(addr, val) \
  94      ({ uint32_t __val = (val);                                               \
  95         ((unsigned char *) (addr))[0] = __val;                                \
  96         __val >>= 8;                                                          \
  97         ((unsigned char *) (addr))[1] = __val;                                \
  98         __val >>= 8;                                                          \
  99         ((unsigned char *) (addr))[2] = __val;                                \
 100         __val >>= 8;                                                          \
 101         ((unsigned char *) (addr))[3] = __val;                                \
 102         (void) 0; })
 103 # else
 104 #  define get16(addr) \
 105      (((const unsigned char *) (addr))[0] << 8                                \
 106       | ((const unsigned char *) (addr))[1])
 107 #  define get32(addr) \
 108      (((((const unsigned char *) (addr))[0] << 8                              \
 109         | ((const unsigned char *) (addr))[1]) << 8                           \
 110        | ((const unsigned char *) (addr))[2]) << 8                            \
 111       | ((const unsigned char *) (addr))[3])
 112
 113 #  define put16(addr, val) \
 114      ({ uint16_t __val = (val);                                               \
 115         ((unsigned char *) (addr))[1] = __val;                                \
 116         ((unsigned char *) (addr))[0] = __val >> 8;                           \
 117         (void) 0; })
 118 #  define put32(addr, val) \
 119      ({ uint32_t __val = (val);                                               \
 120         ((unsigned char *) (addr))[3] = __val;                                \
 121         __val >>= 8;                                                          \
 122         ((unsigned char *) (addr))[2] = __val;                                \
 123         __val >>= 8;                                                          \
 124         ((unsigned char *) (addr))[1] = __val;                                \
 125         __val >>= 8;                                                          \
 126         ((unsigned char *) (addr))[0] = __val;                                \
 127         (void) 0; })
 128 # endif
 129
 130 # define FCTNAME2(name) name##_unaligned
 131 #endif
 132 #define FCTNAME(name) FCTNAME2(name)
 133
 134
 135 /* We need at least one byte for the next round.  */
 136 #ifndef MIN_NEEDED_INPUT
 137 # error "MIN_NEEDED_INPUT definition missing"
 138 #elif MIN_NEEDED_INPUT < 1
 139 # error "MIN_NEEDED_INPUT must be >= 1"
 140 #endif
 141
 142 /* Let's see how many bytes we produce.  */
 143 #ifndef MAX_NEEDED_INPUT
 144 # define MAX_NEEDED_INPUT       MIN_NEEDED_INPUT
 145 #endif
 146
 147 /* We produce at least one byte in the next round.  */
 148 #ifndef MIN_NEEDED_OUTPUT
 149 # error "MIN_NEEDED_OUTPUT definition missing"
 150 #elif MIN_NEEDED_OUTPUT < 1
 151 # error "MIN_NEEDED_OUTPUT must be >= 1"
 152 #endif
 153
 154 /* Let's see how many bytes we produce.  */
 155 #ifndef MAX_NEEDED_OUTPUT
 156 # define MAX_NEEDED_OUTPUT      MIN_NEEDED_OUTPUT
 157 #endif
 158
 159 /* Default name for the function.  */
 160 #ifndef LOOPFCT
 161 # define LOOPFCT                loop
 162 #endif
 163
 164 /* Make sure we have a loop body.  */
 165 #ifndef BODY
 166 # error "Definition of BODY missing for function" LOOPFCT
 167 #endif
 168
 169
 170 /* If no arguments have to passed to the loop function define the macro
 171    as empty.  */
 172 #ifndef EXTRA_LOOP_DECLS
 173 # define EXTRA_LOOP_DECLS
 174 #endif
 175
 176 /* Allow using UPDATE_PARAMS in macros where #ifdef UPDATE_PARAMS test
 177    isn't possible.  */
 178 #ifndef UPDATE_PARAMS
 179 # define UPDATE_PARAMS do { } while (0)
 180 #endif
 181 #ifndef REINIT_PARAMS
 182 # define REINIT_PARAMS do { } while (0)
 183 #endif
 184
 185
 186 /* To make it easier for the writers of the modules, we define a macro
 187    to test whether we have to ignore errors.  */
 188 #define ignore_errors_p() \
 189   (irreversible != NULL && (flags & __GCONV_IGNORE_ERRORS))
 190
 191
 192 /* Error handling for the FROM_LOOP direction, with ignoring of errors.
 193    Note that we cannot use the do while (0) trick since `break' and
 194    `continue' must reach certain points.  */
 195 #define STANDARD_FROM_LOOP_ERR_HANDLER(Incr) \
 196   {                                                                           \
 197     result = __GCONV_ILLEGAL_INPUT;                                           \
 198                                                                               \
 199     if (! ignore_errors_p ())                                                 \
 200       break;                                                                  \
 201                                                                               \
 202     /* We ignore the invalid input byte sequence.  */                         \
 203     inptr += (Incr);                                                          \
 204     ++*irreversible;                                                          \
 205     /* But we keep result == __GCONV_ILLEGAL_INPUT, because of the constraint \
 206        that "iconv -c" must give the same exitcode as "iconv".  */            \
 207     continue;                                                                 \
 208   }
 209
 210 /* Error handling for the TO_LOOP direction, with use of transliteration/
 211    transcription functions and ignoring of errors.  Note that we cannot use
 212    the do while (0) trick since `break' and `continue' must reach certain
 213    points.  */
 214 #define STANDARD_TO_LOOP_ERR_HANDLER(Incr) \
 215   {                                                                           \
 216     struct __gconv_trans_data *trans;                                         \
 217                                                                               \
 218     result = __GCONV_ILLEGAL_INPUT;                                           \
 219                                                                               \
 220     if (irreversible == NULL)                                                 \
 221       /* This means we are in call from __gconv_transliterate.  In this       \
 222          case we are not doing any error recovery outself.  */                \
 223       break;                                                                  \
 224                                                                               \
 225     /* If needed, flush any conversion state, so that __gconv_transliterate   \
 226        starts with current shift state.  */                                   \
 227     UPDATE_PARAMS;                                                            \
 228                                                                               \
 229     /* First try the transliteration methods.  */                             \
 230     for (trans = step_data->__trans; trans != NULL; trans = trans->__next)    \
 231       {                                                                       \
 232         result = DL_CALL_FCT (trans->__trans_fct,                             \
 233                               (step, step_data, trans->__data, *inptrp,       \
 234                                &inptr, inend, &outptr, irreversible));        \
 235         if (result != __GCONV_ILLEGAL_INPUT)                                  \
 236           break;                                                              \
 237       }                                                                       \
 238                                                                               \
 239     REINIT_PARAMS;                                                            \
 240                                                                               \
 241     /* If any of them recognized the input continue with the loop.  */        \
 242     if (result != __GCONV_ILLEGAL_INPUT)                                      \
 243       {                                                                       \
 244         if (__builtin_expect (result == __GCONV_FULL_OUTPUT, 0))              \
 245           break;                                                              \
 246                                                                               \
 247         continue;                                                             \
 248       }                                                                       \
 249                                                                               \
 250     /* Next see whether we have to ignore the error.  If not, stop.  */       \
 251     if (! ignore_errors_p ())                                                 \
 252       break;                                                                  \
 253                                                                               \
 254     /* When we come here it means we ignore the character.  */                \
 255     ++*irreversible;                                                          \
 256     inptr += Incr;                                                            \
 257     /* But we keep result == __GCONV_ILLEGAL_INPUT, because of the constraint \
 258        that "iconv -c" must give the same exitcode as "iconv".  */            \
 259     continue;                                                                 \
 260   }
 261
 262
 263 /* Handling of Unicode 3.1 TAG characters.  Unicode recommends
 264    "If language codes are not relevant to the particular processing
 265     operation, then they should be ignored."  This macro is usually
 266    called right before  STANDARD_TO_LOOP_ERR_HANDLER (Incr).  */
 267 #define UNICODE_TAG_HANDLER(Character, Incr) \
 268   {                                                                           \
 269     /* TAG characters are those in the range U+E0000..U+E007F.  */            \
 270     if (((Character) >> 7) == (0xe0000 >> 7))                                 \
 271       {                                                                       \
 272         inptr += Incr;                                                        \
 273         continue;                                                             \
 274       }                                                                       \
 275   }
 276
 277
 278 /* The function returns the status, as defined in gconv.h.  */
 279 static inline int
 280 __attribute ((always_inline))
 281 FCTNAME (LOOPFCT) (struct __gconv_step *step,
 282                    struct __gconv_step_data *step_data,
 283                    const unsigned char **inptrp, const unsigned char *inend,
 284                    unsigned char **outptrp, const unsigned char *outend,
 285                    size_t *irreversible EXTRA_LOOP_DECLS)
 286 {
 287 #ifdef LOOP_NEED_STATE
 288   mbstate_t *state = step_data->__statep;
 289 #endif
 290 #ifdef LOOP_NEED_FLAGS
 291   int flags = step_data->__flags;
 292 #endif
 293 #ifdef LOOP_NEED_DATA
 294   void *data = step->__data;
 295 #endif
 296   int result = __GCONV_EMPTY_INPUT;
 297   const unsigned char *inptr = *inptrp;
 298   unsigned char *outptr = *outptrp;
 299
 300 #ifdef INIT_PARAMS
 301   INIT_PARAMS;
 302 #endif
 303
 304   while (inptr != inend)
 305     {
 306       /* `if' cases for MIN_NEEDED_OUTPUT ==/!= 1 is made to help the
 307          compiler generating better code.  They will be optimized away
 308          since MIN_NEEDED_OUTPUT is always a constant.  */
 309       if (MIN_NEEDED_INPUT > 1
 310           && __builtin_expect (inptr + MIN_NEEDED_INPUT > inend, 0))
 311         {
 312           /* We don't have enough input for another complete input
 313              character.  */
 314           result = __GCONV_INCOMPLETE_INPUT;
 315           break;
 316         }
 317       if ((MIN_NEEDED_OUTPUT != 1
 318            && __builtin_expect (outptr + MIN_NEEDED_OUTPUT > outend, 0))
 319           || (MIN_NEEDED_OUTPUT == 1
 320               && __builtin_expect (outptr >= outend, 0)))
 321         {
 322           /* Overflow in the output buffer.  */
 323           result = __GCONV_FULL_OUTPUT;
 324           break;
 325         }
 326
 327       /* Here comes the body the user provides.  It can stop with
 328          RESULT set to GCONV_INCOMPLETE_INPUT (if the size of the
 329          input characters vary in size), GCONV_ILLEGAL_INPUT, or
 330          GCONV_FULL_OUTPUT (if the output characters vary in size).  */
 331       BODY
 332     }
 333
 334   /* Update the pointers pointed to by the parameters.  */
 335   *inptrp = inptr;
 336   *outptrp = outptr;
 337   UPDATE_PARAMS;
 338
 339   return result;
 340 }
 341
 342
 343 /* Include the file a second time to define the function to handle
 344    unaligned access.  */
 345 #if !defined DEFINE_UNALIGNED && !defined _STRING_ARCH_unaligned \
 346     && MIN_NEEDED_INPUT != 1 && MAX_NEEDED_INPUT % MIN_NEEDED_INPUT == 0 \
 347     && MIN_NEEDED_OUTPUT != 1 && MAX_NEEDED_OUTPUT % MIN_NEEDED_OUTPUT == 0
 348 # undef get16
 349 # undef get32
 350 # undef put16
 351 # undef put32
 352 # undef unaligned
 353
 354 # define DEFINE_UNALIGNED
 355 # include "loop.c"
 356 # undef DEFINE_UNALIGNED
 357 #endif
 358
 359
 360 #if MAX_NEEDED_INPUT > 1
 361 # define SINGLE(fct) SINGLE2 (fct)
 362 # define SINGLE2(fct) fct##_single
 363 static inline int
 364 __attribute ((always_inline))
 365 SINGLE(LOOPFCT) (struct __gconv_step *step,
 366                  struct __gconv_step_data *step_data,
 367                  const unsigned char **inptrp, const unsigned char *inend,
 368                  unsigned char **outptrp, unsigned char *outend,
 369                  size_t *irreversible EXTRA_LOOP_DECLS)
 370 {
 371   mbstate_t *state = step_data->__statep;
 372 #ifdef LOOP_NEED_FLAGS
 373   int flags = step_data->__flags;
 374 #endif
 375 #ifdef LOOP_NEED_DATA
 376   void *data = step->__data;
 377 #endif
 378   int result = __GCONV_OK;
 379   unsigned char bytebuf[MAX_NEEDED_INPUT];
 380   const unsigned char *inptr = *inptrp;
 381   unsigned char *outptr = *outptrp;
 382   size_t inlen;
 383
 384 #ifdef INIT_PARAMS
 385   INIT_PARAMS;
 386 #endif
 387
 388 #ifdef UNPACK_BYTES
 389   UNPACK_BYTES
 390 #else
 391   /* Add the bytes from the state to the input buffer.  */
 392   assert ((state->__count & 7) <= sizeof (state->__value));
 393   for (inlen = 0; inlen < (size_t) (state->__count & 7); ++inlen)
 394     bytebuf[inlen] = state->__value.__wchb[inlen];
 395 #endif
 396
 397   /* Are there enough bytes in the input buffer?  */
 398   if (MIN_NEEDED_INPUT > 1
 399       && __builtin_expect (inptr + (MIN_NEEDED_INPUT - inlen) > inend, 0))
 400     {
 401       *inptrp = inend;
 402 #ifdef STORE_REST
 403       while (inptr < inend)
 404         bytebuf[inlen++] = *inptr++;
 405
 406       inptr = bytebuf;
 407       inptrp = &inptr;
 408       inend = &bytebuf[inlen];
 409
 410       STORE_REST
 411 #else
 412       /* We don't have enough input for another complete input
 413          character.  */
 414       while (inptr < inend)
 415         state->__value.__wchb[inlen++] = *inptr++;
 416 #endif
 417
 418       return __GCONV_INCOMPLETE_INPUT;
 419     }
 420
 421   /* Enough space in output buffer.  */
 422   if ((MIN_NEEDED_OUTPUT != 1 && outptr + MIN_NEEDED_OUTPUT > outend)
 423       || (MIN_NEEDED_OUTPUT == 1 && outptr >= outend))
 424     /* Overflow in the output buffer.  */
 425     return __GCONV_FULL_OUTPUT;
 426
 427   /*  Now add characters from the normal input buffer.  */
 428   do
 429     bytebuf[inlen++] = *inptr++;
 430   while (inlen < MAX_NEEDED_INPUT && inptr < inend);
 431
 432   inptr = bytebuf;
 433   inend = &bytebuf[inlen];
 434
 435   do
 436     {
 437       BODY
 438     }
 439   while (0);
 440
 441   /* Now we either have produced an output character and consumed all the
 442      bytes from the state and at least one more, or the character is still
 443      incomplete, or we have some other error (like illegal input character,
 444      no space in output buffer).  */
 445   if (__builtin_expect (inptr != bytebuf, 1))
 446     {
 447       /* We found a new character.  */
 448       assert (inptr - bytebuf > (state->__count & 7));
 449
 450       *inptrp += inptr - bytebuf - (state->__count & 7);
 451       *outptrp = outptr;
 452
 453       result = __GCONV_OK;
 454
 455       /* Clear the state buffer.  */
 456 #ifdef CLEAR_STATE
 457       CLEAR_STATE;
 458 #else
 459       state->__count &= ~7;
 460 #endif
 461     }
 462   else if (result == __GCONV_INCOMPLETE_INPUT)
 463     {
 464       /* This can only happen if we have less than MAX_NEEDED_INPUT bytes
 465          available.  */
 466       assert (inend != &bytebuf[MAX_NEEDED_INPUT]);
 467
 468       *inptrp += inend - bytebuf - (state->__count & 7);
 469 #ifdef STORE_REST
 470       inptrp = &inptr;
 471
 472       STORE_REST
 473 #else
 474       /* We don't have enough input for another complete input
 475          character.  */
 476       assert (inend - inptr > (state->__count & ~7));
 477       assert (inend - inptr <= sizeof (state->__value));
 478       state->__count = (state->__count & ~7) | (inend - inptr);
 479       inlen = 0;
 480       while (inptr < inend)
 481         state->__value.__wchb[inlen++] = *inptr++;
 482 #endif
 483     }
 484
 485   return result;
 486 }
 487 # undef SINGLE
 488 # undef SINGLE2
 489 #endif
 490
 491
 492 #ifdef ONEBYTE_BODY
 493 /* Define the shortcut function for btowc.  */
 494 static wint_t
 495 gconv_btowc (struct __gconv_step *step, unsigned char c)
 496   ONEBYTE_BODY
 497 # define FROM_ONEBYTE gconv_btowc
 498 #endif
 499
 500
 501 /* We remove the macro definitions so that we can include this file again
 502    for the definition of another function.  */
 503 #undef MIN_NEEDED_INPUT
 504 #undef MAX_NEEDED_INPUT
 505 #undef MIN_NEEDED_OUTPUT
 506 #undef MAX_NEEDED_OUTPUT
 507 #undef LOOPFCT
 508 #undef BODY
 509 #undef LOOPFCT
 510 #undef EXTRA_LOOP_DECLS
 511 #undef INIT_PARAMS
 512 #undef UPDATE_PARAMS
 513 #undef REINIT_PARAMS
 514 #undef ONEBYTE_BODY
 515 #undef UNPACK_BYTES
 516 #undef CLEAR_STATE
 517 #undef LOOP_NEED_STATE
 518 #undef LOOP_NEED_FLAGS
 519 #undef LOOP_NEED_DATA
 520 #undef get16
 521 #undef get32
 522 #undef put16
 523 #undef put32
 524 #undef unaligned