iconv/skeleton.c

   1 /* Skeleton for a conversion module.
   2    Copyright (C) 1998-2016 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 can be included to provide definitions of several things
  21    many modules have in common.  It can be customized using the following
  22    macros:
  23
  24      DEFINE_INIT        define the default initializer.  This requires the
  25                         following symbol to be defined.
  26
  27      CHARSET_NAME       string with official name of the coded character
  28                         set (in all-caps)
  29
  30      DEFINE_FINI        define the default destructor function.
  31
  32      MIN_NEEDED_FROM    minimal number of bytes needed for the from-charset.
  33      MIN_NEEDED_TO      likewise for the to-charset.
  34
  35      MAX_NEEDED_FROM    maximal number of bytes needed for the from-charset.
  36                         This macro is optional, it defaults to MIN_NEEDED_FROM.
  37      MAX_NEEDED_TO      likewise for the to-charset.
  38
  39      FROM_LOOP_MIN_NEEDED_FROM
  40      FROM_LOOP_MAX_NEEDED_FROM
  41                         minimal/maximal number of bytes needed on input
  42                         of one round through the FROM_LOOP.  Defaults
  43                         to MIN_NEEDED_FROM and MAX_NEEDED_FROM, respectively.
  44      FROM_LOOP_MIN_NEEDED_TO
  45      FROM_LOOP_MAX_NEEDED_TO
  46                         minimal/maximal number of bytes needed on output
  47                         of one round through the FROM_LOOP.  Defaults
  48                         to MIN_NEEDED_TO and MAX_NEEDED_TO, respectively.
  49      TO_LOOP_MIN_NEEDED_FROM
  50      TO_LOOP_MAX_NEEDED_FROM
  51                         minimal/maximal number of bytes needed on input
  52                         of one round through the TO_LOOP.  Defaults
  53                         to MIN_NEEDED_TO and MAX_NEEDED_TO, respectively.
  54      TO_LOOP_MIN_NEEDED_TO
  55      TO_LOOP_MAX_NEEDED_TO
  56                         minimal/maximal number of bytes needed on output
  57                         of one round through the TO_LOOP.  Defaults
  58                         to MIN_NEEDED_FROM and MAX_NEEDED_FROM, respectively.
  59
  60      FROM_DIRECTION     this macro is supposed to return a value != 0
  61                         if we convert from the current character set,
  62                         otherwise it return 0.
  63
  64      EMIT_SHIFT_TO_INIT this symbol is optional.  If it is defined it
  65                         defines some code which writes out a sequence
  66                         of bytes which bring the current state into
  67                         the initial state.
  68
  69      FROM_LOOP          name of the function implementing the conversion
  70                         from the current character set.
  71      TO_LOOP            likewise for the other direction
  72
  73      ONE_DIRECTION      optional.  If defined to 1, only one conversion
  74                         direction is defined instead of two.  In this
  75                         case, FROM_DIRECTION should be defined to 1, and
  76                         FROM_LOOP and TO_LOOP should have the same value.
  77
  78      SAVE_RESET_STATE   in case of an error we must reset the state for
  79                         the rerun so this macro must be defined for
  80                         stateful encodings.  It takes an argument which
  81                         is nonzero when saving.
  82
  83      RESET_INPUT_BUFFER If the input character sets allow this the macro
  84                         can be defined to reset the input buffer pointers
  85                         to cover only those characters up to the error.
  86
  87      FUNCTION_NAME      if not set the conversion function is named `gconv'.
  88
  89      PREPARE_LOOP       optional code preparing the conversion loop.  Can
  90                         contain variable definitions.
  91      END_LOOP           also optional, may be used to store information
  92
  93      EXTRA_LOOP_ARGS    optional macro specifying extra arguments passed
  94                         to loop function.
  95
  96      STORE_REST         optional, needed only when MAX_NEEDED_FROM > 4.
  97                         This macro stores the seen but unconverted input bytes
  98                         in the state.
  99
 100      FROM_ONEBYTE       optional.  If defined, should be the name of a
 101                         specialized conversion function for a single byte
 102                         from the current character set to INTERNAL.  This
 103                         function has prototype
 104                            wint_t
 105                            FROM_ONEBYTE (struct __gconv_step *, unsigned char);
 106                         and does a special conversion:
 107                         - The input is a single byte.
 108                         - The output is a single uint32_t.
 109                         - The state before the conversion is the initial state;
 110                           the state after the conversion is irrelevant.
 111                         - No transliteration.
 112                         - __invocation_counter = 0.
 113                         - __internal_use = 1.
 114                         - do_flush = 0.
 115
 116    Modules can use mbstate_t to store conversion state as follows:
 117
 118    * Bits 2..0 of '__count' contain the number of lookahead input bytes
 119      stored in __value.__wchb.  Always zero if the converter never
 120      returns __GCONV_INCOMPLETE_INPUT.
 121
 122    * Bits 31..3 of '__count' are module dependent shift state.
 123
 124    * __value: When STORE_REST/UNPACK_BYTES aren't defined and when the
 125      converter has returned __GCONV_INCOMPLETE_INPUT, this contains
 126      at most 4 lookahead bytes. Converters with an mb_cur_max > 4
 127      (currently only UTF-8) must find a way to store their state
 128      in __value.__wch and define STORE_REST/UNPACK_BYTES appropriately.
 129
 130    When __value contains lookahead, __count must not be zero, because
 131    the converter is not in the initial state then, and mbsinit() --
 132    defined as a (__count == 0) test -- must reflect this.
 133  */
 134
 135 #include <assert.h>
 136 #include <gconv.h>
 137 #include <string.h>
 138 #define __need_size_t
 139 #define __need_NULL
 140 #include <stddef.h>
 141
 142 #ifndef STATIC_GCONV
 143 # include <dlfcn.h>
 144 #endif
 145
 146 #include <sysdep.h>
 147 #include <stdint.h>
 148
 149 #ifndef DL_CALL_FCT
 150 # define DL_CALL_FCT(fct, args) fct args
 151 #endif
 152
 153 /* The direction objects.  */
 154 #if DEFINE_INIT
 155 # ifndef FROM_DIRECTION
 156 #  define FROM_DIRECTION_VAL NULL
 157 #  define TO_DIRECTION_VAL ((void *) ~((uintptr_t) 0))
 158 #  define FROM_DIRECTION (step->__data == FROM_DIRECTION_VAL)
 159 # endif
 160 #else
 161 # ifndef FROM_DIRECTION
 162 #  error "FROM_DIRECTION must be provided if non-default init is used"
 163 # endif
 164 #endif
 165
 166 /* How many bytes are needed at most for the from-charset.  */
 167 #ifndef MAX_NEEDED_FROM
 168 # define MAX_NEEDED_FROM        MIN_NEEDED_FROM
 169 #endif
 170
 171 /* Same for the to-charset.  */
 172 #ifndef MAX_NEEDED_TO
 173 # define MAX_NEEDED_TO          MIN_NEEDED_TO
 174 #endif
 175
 176 /* Defaults for the per-direction min/max constants.  */
 177 #ifndef FROM_LOOP_MIN_NEEDED_FROM
 178 # define FROM_LOOP_MIN_NEEDED_FROM      MIN_NEEDED_FROM
 179 #endif
 180 #ifndef FROM_LOOP_MAX_NEEDED_FROM
 181 # define FROM_LOOP_MAX_NEEDED_FROM      MAX_NEEDED_FROM
 182 #endif
 183 #ifndef FROM_LOOP_MIN_NEEDED_TO
 184 # define FROM_LOOP_MIN_NEEDED_TO        MIN_NEEDED_TO
 185 #endif
 186 #ifndef FROM_LOOP_MAX_NEEDED_TO
 187 # define FROM_LOOP_MAX_NEEDED_TO        MAX_NEEDED_TO
 188 #endif
 189 #ifndef TO_LOOP_MIN_NEEDED_FROM
 190 # define TO_LOOP_MIN_NEEDED_FROM        MIN_NEEDED_TO
 191 #endif
 192 #ifndef TO_LOOP_MAX_NEEDED_FROM
 193 # define TO_LOOP_MAX_NEEDED_FROM        MAX_NEEDED_TO
 194 #endif
 195 #ifndef TO_LOOP_MIN_NEEDED_TO
 196 # define TO_LOOP_MIN_NEEDED_TO          MIN_NEEDED_FROM
 197 #endif
 198 #ifndef TO_LOOP_MAX_NEEDED_TO
 199 # define TO_LOOP_MAX_NEEDED_TO          MAX_NEEDED_FROM
 200 #endif
 201
 202
 203 /* Define macros which can access unaligned buffers.  These macros are
 204    supposed to be used only in code outside the inner loops.  For the inner
 205    loops we have other definitions which allow optimized access.  */
 206 #if _STRING_ARCH_unaligned
 207 /* We can handle unaligned memory access.  */
 208 # define get16u(addr) *((const uint16_t *) (addr))
 209 # define get32u(addr) *((const uint32_t *) (addr))
 210
 211 /* We need no special support for writing values either.  */
 212 # define put16u(addr, val) *((uint16_t *) (addr)) = (val)
 213 # define put32u(addr, val) *((uint32_t *) (addr)) = (val)
 214 #else
 215 /* Distinguish between big endian and little endian.  */
 216 # if __BYTE_ORDER == __LITTLE_ENDIAN
 217 #  define get16u(addr) \
 218      (((const unsigned char *) (addr))[1] << 8                                \
 219       | ((const unsigned char *) (addr))[0])
 220 #  define get32u(addr) \
 221      (((((const unsigned char *) (addr))[3] << 8                              \
 222         | ((const unsigned char *) (addr))[2]) << 8                           \
 223        | ((const unsigned char *) (addr))[1]) << 8                            \
 224       | ((const unsigned char *) (addr))[0])
 225
 226 #  define put16u(addr, val) \
 227      ({ uint16_t __val = (val);                                               \
 228         ((unsigned char *) (addr))[0] = __val;                                \
 229         ((unsigned char *) (addr))[1] = __val >> 8;                           \
 230         (void) 0; })
 231 #  define put32u(addr, val) \
 232      ({ uint32_t __val = (val);                                               \
 233         ((unsigned char *) (addr))[0] = __val;                                \
 234         __val >>= 8;                                                          \
 235         ((unsigned char *) (addr))[1] = __val;                                \
 236         __val >>= 8;                                                          \
 237         ((unsigned char *) (addr))[2] = __val;                                \
 238         __val >>= 8;                                                          \
 239         ((unsigned char *) (addr))[3] = __val;                                \
 240         (void) 0; })
 241 # else
 242 #  define get16u(addr) \
 243      (((const unsigned char *) (addr))[0] << 8                                \
 244       | ((const unsigned char *) (addr))[1])
 245 #  define get32u(addr) \
 246      (((((const unsigned char *) (addr))[0] << 8                              \
 247         | ((const unsigned char *) (addr))[1]) << 8                           \
 248        | ((const unsigned char *) (addr))[2]) << 8                            \
 249       | ((const unsigned char *) (addr))[3])
 250
 251 #  define put16u(addr, val) \
 252      ({ uint16_t __val = (val);                                               \
 253         ((unsigned char *) (addr))[1] = __val;                                \
 254         ((unsigned char *) (addr))[0] = __val >> 8;                           \
 255         (void) 0; })
 256 #  define put32u(addr, val) \
 257      ({ uint32_t __val = (val);                                               \
 258         ((unsigned char *) (addr))[3] = __val;                                \
 259         __val >>= 8;                                                          \
 260         ((unsigned char *) (addr))[2] = __val;                                \
 261         __val >>= 8;                                                          \
 262         ((unsigned char *) (addr))[1] = __val;                                \
 263         __val >>= 8;                                                          \
 264         ((unsigned char *) (addr))[0] = __val;                                \
 265         (void) 0; })
 266 # endif
 267 #endif
 268
 269
 270 /* For conversions from a fixed width character set to another fixed width
 271    character set we can define RESET_INPUT_BUFFER in a very fast way.  */
 272 #if !defined RESET_INPUT_BUFFER && !defined SAVE_RESET_STATE
 273 # if FROM_LOOP_MIN_NEEDED_FROM == FROM_LOOP_MAX_NEEDED_FROM \
 274      && FROM_LOOP_MIN_NEEDED_TO == FROM_LOOP_MAX_NEEDED_TO \
 275      && TO_LOOP_MIN_NEEDED_FROM == TO_LOOP_MAX_NEEDED_FROM \
 276      && TO_LOOP_MIN_NEEDED_TO == TO_LOOP_MAX_NEEDED_TO
 277 /* We have to use these `if's here since the compiler cannot know that
 278    (outbuf - outerr) is always divisible by FROM/TO_LOOP_MIN_NEEDED_TO.
 279    The ?:1 avoids division by zero warnings that gcc 3.2 emits even for
 280    obviously unreachable code.  */
 281 #  define RESET_INPUT_BUFFER \
 282   if (FROM_DIRECTION)                                                         \
 283     {                                                                         \
 284       if (FROM_LOOP_MIN_NEEDED_FROM % FROM_LOOP_MIN_NEEDED_TO == 0)           \
 285         *inptrp -= (outbuf - outerr)                                          \
 286                    * (FROM_LOOP_MIN_NEEDED_FROM / FROM_LOOP_MIN_NEEDED_TO);   \
 287       else if (FROM_LOOP_MIN_NEEDED_TO % FROM_LOOP_MIN_NEEDED_FROM == 0)      \
 288         *inptrp -= (outbuf - outerr)                                          \
 289                    / (FROM_LOOP_MIN_NEEDED_TO / FROM_LOOP_MIN_NEEDED_FROM     \
 290                       ? : 1);                                                 \
 291       else                                                                    \
 292         *inptrp -= ((outbuf - outerr) / FROM_LOOP_MIN_NEEDED_TO)              \
 293                    * FROM_LOOP_MIN_NEEDED_FROM;                               \
 294     }                                                                         \
 295   else                                                                        \
 296     {                                                                         \
 297       if (TO_LOOP_MIN_NEEDED_FROM % TO_LOOP_MIN_NEEDED_TO == 0)               \
 298         *inptrp -= (outbuf - outerr)                                          \
 299                    * (TO_LOOP_MIN_NEEDED_FROM / TO_LOOP_MIN_NEEDED_TO);       \
 300       else if (TO_LOOP_MIN_NEEDED_TO % TO_LOOP_MIN_NEEDED_FROM == 0)          \
 301         *inptrp -= (outbuf - outerr)                                          \
 302                    / (TO_LOOP_MIN_NEEDED_TO / TO_LOOP_MIN_NEEDED_FROM ? : 1); \
 303       else                                                                    \
 304         *inptrp -= ((outbuf - outerr) / TO_LOOP_MIN_NEEDED_TO)                \
 305                    * TO_LOOP_MIN_NEEDED_FROM;                                 \
 306     }
 307 # endif
 308 #endif
 309
 310
 311 /* The default init function.  It simply matches the name and initializes
 312    the step data to point to one of the objects above.  */
 313 #if DEFINE_INIT
 314 # ifndef CHARSET_NAME
 315 #  error "CHARSET_NAME not defined"
 316 # endif
 317
 318 extern int gconv_init (struct __gconv_step *step);
 319 int
 320 gconv_init (struct __gconv_step *step)
 321 {
 322   /* Determine which direction.  */
 323   if (strcmp (step->__from_name, CHARSET_NAME) == 0)
 324     {
 325       step->__data = FROM_DIRECTION_VAL;
 326
 327       step->__min_needed_from = FROM_LOOP_MIN_NEEDED_FROM;
 328       step->__max_needed_from = FROM_LOOP_MAX_NEEDED_FROM;
 329       step->__min_needed_to = FROM_LOOP_MIN_NEEDED_TO;
 330       step->__max_needed_to = FROM_LOOP_MAX_NEEDED_TO;
 331
 332 #ifdef FROM_ONEBYTE
 333       step->__btowc_fct = FROM_ONEBYTE;
 334 #endif
 335     }
 336   else if (__builtin_expect (strcmp (step->__to_name, CHARSET_NAME), 0) == 0)
 337     {
 338       step->__data = TO_DIRECTION_VAL;
 339
 340       step->__min_needed_from = TO_LOOP_MIN_NEEDED_FROM;
 341       step->__max_needed_from = TO_LOOP_MAX_NEEDED_FROM;
 342       step->__min_needed_to = TO_LOOP_MIN_NEEDED_TO;
 343       step->__max_needed_to = TO_LOOP_MAX_NEEDED_TO;
 344     }
 345   else
 346     return __GCONV_NOCONV;
 347
 348 #ifdef SAVE_RESET_STATE
 349   step->__stateful = 1;
 350 #else
 351   step->__stateful = 0;
 352 #endif
 353
 354   return __GCONV_OK;
 355 }
 356 #endif
 357
 358
 359 /* The default destructor function does nothing in the moment and so
 360    we don't define it at all.  But we still provide the macro just in
 361    case we need it some day.  */
 362 #if DEFINE_FINI
 363 #endif
 364
 365
 366 /* If no arguments have to passed to the loop function define the macro
 367    as empty.  */
 368 #ifndef EXTRA_LOOP_ARGS
 369 # define EXTRA_LOOP_ARGS
 370 #endif
 371
 372
 373 /* This is the actual conversion function.  */
 374 #ifndef FUNCTION_NAME
 375 # define FUNCTION_NAME  gconv
 376 #endif
 377
 378 /* The macros are used to access the function to convert single characters.  */
 379 #define SINGLE(fct) SINGLE2 (fct)
 380 #define SINGLE2(fct) fct##_single
 381
 382
 383 extern int FUNCTION_NAME (struct __gconv_step *step,
 384                           struct __gconv_step_data *data,
 385                           const unsigned char **inptrp,
 386                           const unsigned char *inend,
 387                           unsigned char **outbufstart, size_t *irreversible,
 388                           int do_flush, int consume_incomplete);
 389 int
 390 FUNCTION_NAME (struct __gconv_step *step, struct __gconv_step_data *data,
 391                const unsigned char **inptrp, const unsigned char *inend,
 392                unsigned char **outbufstart, size_t *irreversible, int do_flush,
 393                int consume_incomplete)
 394 {
 395   struct __gconv_step *next_step = step + 1;
 396   struct __gconv_step_data *next_data = data + 1;
 397   __gconv_fct fct = NULL;
 398   int status;
 399
 400   if ((data->__flags & __GCONV_IS_LAST) == 0)
 401     {
 402       fct = next_step->__fct;
 403 #ifdef PTR_DEMANGLE
 404       if (next_step->__shlib_handle != NULL)
 405         PTR_DEMANGLE (fct);
 406 #endif
 407     }
 408
 409   /* If the function is called with no input this means we have to reset
 410      to the initial state.  The possibly partly converted input is
 411      dropped.  */
 412   if (__glibc_unlikely (do_flush))
 413     {
 414       /* This should never happen during error handling.  */
 415       assert (outbufstart == NULL);
 416
 417       status = __GCONV_OK;
 418
 419 #ifdef EMIT_SHIFT_TO_INIT
 420       if (do_flush == 1)
 421         {
 422           /* We preserve the initial values of the pointer variables.  */
 423           unsigned char *outbuf = data->__outbuf;
 424           unsigned char *outstart = outbuf;
 425           unsigned char *outend = data->__outbufend;
 426
 427 # ifdef PREPARE_LOOP
 428           PREPARE_LOOP
 429 # endif
 430
 431 # ifdef SAVE_RESET_STATE
 432           SAVE_RESET_STATE (1);
 433 # endif
 434
 435           /* Emit the escape sequence to reset the state.  */
 436           EMIT_SHIFT_TO_INIT;
 437
 438           /* Call the steps down the chain if there are any but only if we
 439              successfully emitted the escape sequence.  This should only
 440              fail if the output buffer is full.  If the input is invalid
 441              it should be discarded since the user wants to start from a
 442              clean state.  */
 443           if (status == __GCONV_OK)
 444             {
 445               if (data->__flags & __GCONV_IS_LAST)
 446                 /* Store information about how many bytes are available.  */
 447                 data->__outbuf = outbuf;
 448               else
 449                 {
 450                   /* Write out all output which was produced.  */
 451                   if (outbuf > outstart)
 452                     {
 453                       const unsigned char *outerr = outstart;
 454                       int result;
 455
 456                       result = DL_CALL_FCT (fct, (next_step, next_data,
 457                                                   &outerr, outbuf, NULL,
 458                                                   irreversible, 0,
 459                                                   consume_incomplete));
 460
 461                       if (result != __GCONV_EMPTY_INPUT)
 462                         {
 463                           if (__glibc_unlikely (outerr != outbuf))
 464                             {
 465                               /* We have a problem.  Undo the conversion.  */
 466                               outbuf = outstart;
 467
 468                               /* Restore the state.  */
 469 # ifdef SAVE_RESET_STATE
 470                               SAVE_RESET_STATE (0);
 471 # endif
 472                             }
 473
 474                           /* Change the status.  */
 475                           status = result;
 476                         }
 477                     }
 478
 479                   if (status == __GCONV_OK)
 480                     /* Now flush the remaining steps.  */
 481                     status = DL_CALL_FCT (fct, (next_step, next_data, NULL,
 482                                                 NULL, NULL, irreversible, 1,
 483                                                 consume_incomplete));
 484                 }
 485             }
 486         }
 487       else
 488 #endif
 489         {
 490           /* Clear the state object.  There might be bytes in there from
 491              previous calls with CONSUME_INCOMPLETE == 1.  But don't emit
 492              escape sequences.  */
 493           memset (data->__statep, '\0', sizeof (*data->__statep));
 494
 495           if (! (data->__flags & __GCONV_IS_LAST))
 496             /* Now flush the remaining steps.  */
 497             status = DL_CALL_FCT (fct, (next_step, next_data, NULL, NULL,
 498                                         NULL, irreversible, do_flush,
 499                                         consume_incomplete));
 500         }
 501     }
 502   else
 503     {
 504       /* We preserve the initial values of the pointer variables,
 505          but only some conversion modules need it.  */
 506       const unsigned char *inptr __attribute__ ((__unused__)) = *inptrp;
 507       unsigned char *outbuf = (__builtin_expect (outbufstart == NULL, 1)
 508                                ? data->__outbuf : *outbufstart);
 509       unsigned char *outend = data->__outbufend;
 510       unsigned char *outstart;
 511       /* This variable is used to count the number of characters we
 512          actually converted.  */
 513       size_t lirreversible = 0;
 514       size_t *lirreversiblep = irreversible ? &lirreversible : NULL;
 515
 516       /* The following assumes that encodings, which have a variable length
 517          what might unalign a buffer even though it is an aligned in the
 518          beginning, either don't have the minimal number of bytes as a divisor
 519          of the maximum length or have a minimum length of 1.  This is true
 520          for all known and supported encodings.
 521          We use && instead of || to combine the subexpression for the FROM
 522          encoding and for the TO encoding, because usually one of them is
 523          INTERNAL, for which the subexpression evaluates to 1, but INTERNAL
 524          buffers are always aligned correctly.  */
 525 #define POSSIBLY_UNALIGNED \
 526   (!_STRING_ARCH_unaligned                                                    \
 527    && (((FROM_LOOP_MIN_NEEDED_FROM != 1                                       \
 528          && FROM_LOOP_MAX_NEEDED_FROM % FROM_LOOP_MIN_NEEDED_FROM == 0)       \
 529         && (FROM_LOOP_MIN_NEEDED_TO != 1                                      \
 530             && FROM_LOOP_MAX_NEEDED_TO % FROM_LOOP_MIN_NEEDED_TO == 0))       \
 531        || ((TO_LOOP_MIN_NEEDED_FROM != 1                                      \
 532             && TO_LOOP_MAX_NEEDED_FROM % TO_LOOP_MIN_NEEDED_FROM == 0)        \
 533            && (TO_LOOP_MIN_NEEDED_TO != 1                                     \
 534                && TO_LOOP_MAX_NEEDED_TO % TO_LOOP_MIN_NEEDED_TO == 0))))
 535 #if POSSIBLY_UNALIGNED
 536       int unaligned;
 537 # define GEN_unaligned(name) GEN_unaligned2 (name)
 538 # define GEN_unaligned2(name) name##_unaligned
 539 #else
 540 # define unaligned 0
 541 #endif
 542
 543 #ifdef PREPARE_LOOP
 544       PREPARE_LOOP
 545 #endif
 546
 547 #if FROM_LOOP_MAX_NEEDED_FROM > 1 || TO_LOOP_MAX_NEEDED_FROM > 1
 548       /* If the function is used to implement the mb*towc*() or wc*tomb*()
 549          functions we must test whether any bytes from the last call are
 550          stored in the `state' object.  */
 551       if (((FROM_LOOP_MAX_NEEDED_FROM > 1 && TO_LOOP_MAX_NEEDED_FROM > 1)
 552            || (FROM_LOOP_MAX_NEEDED_FROM > 1 && FROM_DIRECTION)
 553            || (TO_LOOP_MAX_NEEDED_FROM > 1 && !FROM_DIRECTION))
 554           && consume_incomplete && (data->__statep->__count & 7) != 0)
 555         {
 556           /* Yep, we have some bytes left over.  Process them now.
 557              But this must not happen while we are called from an
 558              error handler.  */
 559           assert (outbufstart == NULL);
 560
 561 # if FROM_LOOP_MAX_NEEDED_FROM > 1
 562           if (TO_LOOP_MAX_NEEDED_FROM == 1 || FROM_DIRECTION)
 563             status = SINGLE(FROM_LOOP) (step, data, inptrp, inend, &outbuf,
 564                                         outend, lirreversiblep
 565                                         EXTRA_LOOP_ARGS);
 566 # endif
 567 # if !ONE_DIRECTION
 568 #  if FROM_LOOP_MAX_NEEDED_FROM > 1 && TO_LOOP_MAX_NEEDED_FROM > 1
 569           else
 570 #  endif
 571 #  if TO_LOOP_MAX_NEEDED_FROM > 1
 572             status = SINGLE(TO_LOOP) (step, data, inptrp, inend, &outbuf,
 573                                       outend, lirreversiblep EXTRA_LOOP_ARGS);
 574 #  endif
 575 # endif
 576
 577           if (__builtin_expect (status, __GCONV_OK) != __GCONV_OK)
 578             return status;
 579         }
 580 #endif
 581
 582 #if POSSIBLY_UNALIGNED
 583       unaligned =
 584         ((FROM_DIRECTION
 585           && ((uintptr_t) inptr % FROM_LOOP_MIN_NEEDED_FROM != 0
 586               || ((data->__flags & __GCONV_IS_LAST)
 587                   && (uintptr_t) outbuf % FROM_LOOP_MIN_NEEDED_TO != 0)))
 588          || (!FROM_DIRECTION
 589              && (((data->__flags & __GCONV_IS_LAST)
 590                   && (uintptr_t) outbuf % TO_LOOP_MIN_NEEDED_TO != 0)
 591                  || (uintptr_t) inptr % TO_LOOP_MIN_NEEDED_FROM != 0)));
 592 #endif
 593
 594       while (1)
 595         {
 596           /* Remember the start value for this round.  */
 597           inptr = *inptrp;
 598           /* The outbuf buffer is empty.  */
 599           outstart = outbuf;
 600
 601 #ifdef SAVE_RESET_STATE
 602           SAVE_RESET_STATE (1);
 603 #endif
 604
 605           if (__glibc_likely (!unaligned))
 606             {
 607               if (FROM_DIRECTION)
 608                 /* Run the conversion loop.  */
 609                 status = FROM_LOOP (step, data, inptrp, inend, &outbuf, outend,
 610                                     lirreversiblep EXTRA_LOOP_ARGS);
 611               else
 612                 /* Run the conversion loop.  */
 613                 status = TO_LOOP (step, data, inptrp, inend, &outbuf, outend,
 614                                   lirreversiblep EXTRA_LOOP_ARGS);
 615             }
 616 #if POSSIBLY_UNALIGNED
 617           else
 618             {
 619               if (FROM_DIRECTION)
 620                 /* Run the conversion loop.  */
 621                 status = GEN_unaligned (FROM_LOOP) (step, data, inptrp, inend,
 622                                                     &outbuf, outend,
 623                                                     lirreversiblep
 624                                                     EXTRA_LOOP_ARGS);
 625               else
 626                 /* Run the conversion loop.  */
 627                 status = GEN_unaligned (TO_LOOP) (step, data, inptrp, inend,
 628                                                   &outbuf, outend,
 629                                                   lirreversiblep
 630                                                   EXTRA_LOOP_ARGS);
 631             }
 632 #endif
 633
 634           /* If we were called as part of an error handling module we
 635              don't do anything else here.  */
 636           if (__glibc_unlikely (outbufstart != NULL))
 637             {
 638               *outbufstart = outbuf;
 639               return status;
 640             }
 641
 642           /* We finished one use of the loops.  */
 643           ++data->__invocation_counter;
 644
 645           /* If this is the last step leave the loop, there is nothing
 646              we can do.  */
 647           if (__glibc_unlikely (data->__flags & __GCONV_IS_LAST))
 648             {
 649               /* Store information about how many bytes are available.  */
 650               data->__outbuf = outbuf;
 651
 652               /* Remember how many non-identical characters we
 653                  converted in an irreversible way.  */
 654               *irreversible += lirreversible;
 655
 656               break;
 657             }
 658
 659           /* Write out all output which was produced.  */
 660           if (__glibc_likely (outbuf > outstart))
 661             {
 662               const unsigned char *outerr = data->__outbuf;
 663               int result;
 664
 665               result = DL_CALL_FCT (fct, (next_step, next_data, &outerr,
 666                                           outbuf, NULL, irreversible, 0,
 667                                           consume_incomplete));
 668
 669               if (result != __GCONV_EMPTY_INPUT)
 670                 {
 671                   if (__glibc_unlikely (outerr != outbuf))
 672                     {
 673 #ifdef RESET_INPUT_BUFFER
 674                       RESET_INPUT_BUFFER;
 675 #else
 676                       /* We have a problem in one of the functions below.
 677                          Undo the conversion upto the error point.  */
 678                       size_t nstatus __attribute__ ((unused));
 679
 680                       /* Reload the pointers.  */
 681                       *inptrp = inptr;
 682                       outbuf = outstart;
 683
 684                       /* Restore the state.  */
 685 # ifdef SAVE_RESET_STATE
 686                       SAVE_RESET_STATE (0);
 687 # endif
 688
 689                       if (__glibc_likely (!unaligned))
 690                         {
 691                           if (FROM_DIRECTION)
 692                             /* Run the conversion loop.  */
 693                             nstatus = FROM_LOOP (step, data, inptrp, inend,
 694                                                  &outbuf, outerr,
 695                                                  lirreversiblep
 696                                                  EXTRA_LOOP_ARGS);
 697                           else
 698                             /* Run the conversion loop.  */
 699                             nstatus = TO_LOOP (step, data, inptrp, inend,
 700                                                &outbuf, outerr,
 701                                                lirreversiblep
 702                                                EXTRA_LOOP_ARGS);
 703                         }
 704 # if POSSIBLY_UNALIGNED
 705                       else
 706                         {
 707                           if (FROM_DIRECTION)
 708                             /* Run the conversion loop.  */
 709                             nstatus = GEN_unaligned (FROM_LOOP) (step, data,
 710                                                                  inptrp, inend,
 711                                                                  &outbuf,
 712                                                                  outerr,
 713                                                                  lirreversiblep
 714                                                                  EXTRA_LOOP_ARGS);
 715                           else
 716                             /* Run the conversion loop.  */
 717                             nstatus = GEN_unaligned (TO_LOOP) (step, data,
 718                                                                inptrp, inend,
 719                                                                &outbuf, outerr,
 720                                                                lirreversiblep
 721                                                                EXTRA_LOOP_ARGS);
 722                         }
 723 # endif
 724
 725                       /* We must run out of output buffer space in this
 726                          rerun.  */
 727                       assert (outbuf == outerr);
 728                       assert (nstatus == __GCONV_FULL_OUTPUT);
 729
 730                       /* If we haven't consumed a single byte decrement
 731                          the invocation counter.  */
 732                       if (__glibc_unlikely (outbuf == outstart))
 733                         --data->__invocation_counter;
 734 #endif  /* reset input buffer */
 735                     }
 736
 737                   /* Change the status.  */
 738                   status = result;
 739                 }
 740               else
 741                 /* All the output is consumed, we can make another run
 742                    if everything was ok.  */
 743                 if (status == __GCONV_FULL_OUTPUT)
 744                   {
 745                     status = __GCONV_OK;
 746                     outbuf = data->__outbuf;
 747                   }
 748             }
 749
 750           if (status != __GCONV_OK)
 751             break;
 752
 753           /* Reset the output buffer pointer for the next round.  */
 754           outbuf = data->__outbuf;
 755         }
 756
 757 #ifdef END_LOOP
 758       END_LOOP
 759 #endif
 760
 761       /* If we are supposed to consume all character store now all of the
 762          remaining characters in the `state' object.  */
 763 #if FROM_LOOP_MAX_NEEDED_FROM > 1 || TO_LOOP_MAX_NEEDED_FROM > 1
 764       if (((FROM_LOOP_MAX_NEEDED_FROM > 1 && TO_LOOP_MAX_NEEDED_FROM > 1)
 765            || (FROM_LOOP_MAX_NEEDED_FROM > 1 && FROM_DIRECTION)
 766            || (TO_LOOP_MAX_NEEDED_FROM > 1 && !FROM_DIRECTION))
 767           && __builtin_expect (consume_incomplete, 0)
 768           && status == __GCONV_INCOMPLETE_INPUT)
 769         {
 770 # ifdef STORE_REST
 771           mbstate_t *state = data->__statep;
 772
 773           STORE_REST
 774 # else
 775           /* Make sure the remaining bytes fit into the state objects
 776              buffer.  */
 777           assert (inend - *inptrp < 4);
 778
 779           size_t cnt;
 780           for (cnt = 0; *inptrp < inend; ++cnt)
 781             data->__statep->__value.__wchb[cnt] = *(*inptrp)++;
 782           data->__statep->__count &= ~7;
 783           data->__statep->__count |= cnt;
 784 # endif
 785         }
 786 #endif
 787 #undef unaligned
 788 #undef POSSIBLY_UNALIGNED
 789     }
 790
 791   return status;
 792 }
 793
 794 #undef DEFINE_INIT
 795 #undef CHARSET_NAME
 796 #undef DEFINE_FINI
 797 #undef MIN_NEEDED_FROM
 798 #undef MIN_NEEDED_TO
 799 #undef MAX_NEEDED_FROM
 800 #undef MAX_NEEDED_TO
 801 #undef FROM_LOOP_MIN_NEEDED_FROM
 802 #undef FROM_LOOP_MAX_NEEDED_FROM
 803 #undef FROM_LOOP_MIN_NEEDED_TO
 804 #undef FROM_LOOP_MAX_NEEDED_TO
 805 #undef TO_LOOP_MIN_NEEDED_FROM
 806 #undef TO_LOOP_MAX_NEEDED_FROM
 807 #undef TO_LOOP_MIN_NEEDED_TO
 808 #undef TO_LOOP_MAX_NEEDED_TO
 809 #undef FROM_DIRECTION
 810 #undef EMIT_SHIFT_TO_INIT
 811 #undef FROM_LOOP
 812 #undef TO_LOOP
 813 #undef ONE_DIRECTION
 814 #undef SAVE_RESET_STATE
 815 #undef RESET_INPUT_BUFFER
 816 #undef FUNCTION_NAME
 817 #undef PREPARE_LOOP
 818 #undef END_LOOP
 819 #undef EXTRA_LOOP_ARGS
 820 #undef STORE_REST
 821 #undef FROM_ONEBYTE