iconv/loop.c

   1 /* Conversion loop frame work.
   2    Copyright (C) 1998-2024 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4
   5    The GNU C Library is free software; you can redistribute it and/or
   6    modify it under the terms of the GNU Lesser General Public
   7    License as published by the Free Software Foundation; either
   8    version 2.1 of the License, or (at your option) any later version.
   9
  10    The GNU C Library is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public
  16    License along with the GNU C Library; if not, see
  17    <https://www.gnu.org/licenses/>.  */
  18
  19 /* This file provides a frame for the reader loop in all conversion modules.
  20    The actual code must (of course) be provided in the actual module source
  21    code but certain actions can be written down generically, with some
  22    customization options which are these:
  23
  24      MIN_NEEDED_INPUT   minimal number of input bytes needed for the next
  25                         conversion.
  26      MIN_NEEDED_OUTPUT  minimal number of bytes produced by the next round
  27                         of conversion.
  28
  29      MAX_NEEDED_INPUT   you guess it, this is the maximal number of input
  30                         bytes needed.  It defaults to MIN_NEEDED_INPUT
  31      MAX_NEEDED_OUTPUT  likewise for output bytes.
  32
  33      LOOPFCT            name of the function created.  If not specified
  34                         the name is `loop' but this prevents the use
  35                         of multiple functions in the same file.
  36
  37      BODY               this is supposed to expand to the body of the loop.
  38                         The user must provide this.
  39
  40      EXTRA_LOOP_DECLS   extra arguments passed from conversion loop call.
  41
  42      INIT_PARAMS        code to define and initialize variables from params.
  43      UPDATE_PARAMS      code to store result in params.
  44
  45      ONEBYTE_BODY       body of the specialized conversion function for a
  46                         single byte from the current character set to INTERNAL.
  47 */
  48
  49 #include <assert.h>
  50 #include <endian.h>
  51 #include <iconv/gconv_int.h>
  52 #include <stdint.h>
  53 #include <string.h>
  54 #include <wchar.h>
  55 #include <sys/param.h>          /* For MIN.  */
  56 #define __need_size_t
  57 #include <stddef.h>
  58 #include <libc-diag.h>
  59
  60 #undef FCTNAME2
  61 #define FCTNAME(name) name
  62
  63
  64 /* We need at least one byte for the next round.  */
  65 #ifndef MIN_NEEDED_INPUT
  66 # error "MIN_NEEDED_INPUT definition missing"
  67 #elif MIN_NEEDED_INPUT < 1
  68 # error "MIN_NEEDED_INPUT must be >= 1"
  69 #endif
  70
  71 /* Let's see how many bytes we produce.  */
  72 #ifndef MAX_NEEDED_INPUT
  73 # define MAX_NEEDED_INPUT       MIN_NEEDED_INPUT
  74 #endif
  75
  76 /* We produce at least one byte in the next round.  */
  77 #ifndef MIN_NEEDED_OUTPUT
  78 # error "MIN_NEEDED_OUTPUT definition missing"
  79 #elif MIN_NEEDED_OUTPUT < 1
  80 # error "MIN_NEEDED_OUTPUT must be >= 1"
  81 #endif
  82
  83 /* Let's see how many bytes we produce.  */
  84 #ifndef MAX_NEEDED_OUTPUT
  85 # define MAX_NEEDED_OUTPUT      MIN_NEEDED_OUTPUT
  86 #endif
  87
  88 /* Default name for the function.  */
  89 #ifndef LOOPFCT
  90 # define LOOPFCT                loop
  91 #endif
  92
  93 /* Make sure we have a loop body.  */
  94 #ifndef BODY
  95 # error "Definition of BODY missing for function" LOOPFCT
  96 #endif
  97
  98
  99 /* If no arguments have to passed to the loop function define the macro
 100    as empty.  */
 101 #ifndef EXTRA_LOOP_DECLS
 102 # define EXTRA_LOOP_DECLS
 103 #endif
 104
 105 /* Allow using UPDATE_PARAMS in macros where #ifdef UPDATE_PARAMS test
 106    isn't possible.  */
 107 #ifndef UPDATE_PARAMS
 108 # define UPDATE_PARAMS do { } while (0)
 109 #endif
 110 #ifndef REINIT_PARAMS
 111 # define REINIT_PARAMS do { } while (0)
 112 #endif
 113
 114
 115 /* To make it easier for the writers of the modules, we define a macro
 116    to test whether we have to ignore errors.  */
 117 #define ignore_errors_p() \
 118   (irreversible != NULL && (flags & __GCONV_IGNORE_ERRORS))
 119
 120
 121 /* Error handling for the FROM_LOOP direction, with ignoring of errors.
 122    Note that we cannot use the do while (0) trick since `break' and
 123    `continue' must reach certain points.  */
 124 #define STANDARD_FROM_LOOP_ERR_HANDLER(Incr) \
 125   {                                                                           \
 126     result = __GCONV_ILLEGAL_INPUT;                                           \
 127                                                                               \
 128     if (! ignore_errors_p ())                                                 \
 129       break;                                                                  \
 130                                                                               \
 131     /* We ignore the invalid input byte sequence.  */                         \
 132     inptr += (Incr);                                                          \
 133     ++*irreversible;                                                          \
 134     /* But we keep result == __GCONV_ILLEGAL_INPUT, because of the constraint \
 135        that "iconv -c" must give the same exitcode as "iconv".  */            \
 136     continue;                                                                 \
 137   }
 138
 139 /* Error handling for the TO_LOOP direction, with use of transliteration/
 140    transcription functions and ignoring of errors.  Note that we cannot use
 141    the do while (0) trick since `break' and `continue' must reach certain
 142    points.  */
 143 #define STANDARD_TO_LOOP_ERR_HANDLER(Incr) \
 144   {                                                                           \
 145     result = __GCONV_ILLEGAL_INPUT;                                           \
 146                                                                               \
 147     if (irreversible == NULL)                                                 \
 148       /* This means we are in call from __gconv_transliterate.  In this       \
 149          case we are not doing any error recovery outself.  */                \
 150       break;                                                                  \
 151                                                                               \
 152     /* If needed, flush any conversion state, so that __gconv_transliterate   \
 153        starts with current shift state.  */                                   \
 154     UPDATE_PARAMS;                                                            \
 155                                                                               \
 156     /* First try the transliteration methods.  */                             \
 157     if ((step_data->__flags & __GCONV_TRANSLIT) != 0)                         \
 158       result = __gconv_transliterate                                          \
 159         (step, step_data, *inptrp,                                            \
 160          &inptr, inend, &outptr, irreversible);                       \
 161                                                                               \
 162     REINIT_PARAMS;                                                            \
 163                                                                               \
 164     /* If any of them recognized the input continue with the loop.  */        \
 165     if (result != __GCONV_ILLEGAL_INPUT)                                      \
 166       {                                                                       \
 167         if (__glibc_unlikely (result == __GCONV_FULL_OUTPUT))                 \
 168           break;                                                              \
 169                                                                               \
 170         continue;                                                             \
 171       }                                                                       \
 172                                                                               \
 173     /* Next see whether we have to ignore the error.  If not, stop.  */       \
 174     if (! ignore_errors_p ())                                                 \
 175       break;                                                                  \
 176                                                                               \
 177     /* When we come here it means we ignore the character.  */                \
 178     ++*irreversible;                                                          \
 179     inptr += Incr;                                                            \
 180     /* But we keep result == __GCONV_ILLEGAL_INPUT, because of the constraint \
 181        that "iconv -c" must give the same exitcode as "iconv".  */            \
 182     continue;                                                                 \
 183   }
 184
 185
 186 /* With GCC 7 when compiling with -Os for 32-bit s390 the compiler
 187    warns that the variable 'ch', in the definition of BODY in
 188    sysdeps/s390/multiarch/8bit-generic.c, may be used uninitialized in
 189    the call to UNICODE_TAG_HANDLER in that macro.  This variable is
 190    actually always initialized before use, in the prior loop if INDEX
 191    is nonzero and in the following 'if' if INDEX is zero.  That code
 192    has a comment referencing this diagnostic disabling; updates in one
 193    place may require updates in the other.  */
 194 DIAG_PUSH_NEEDS_COMMENT;
 195 DIAG_IGNORE_Os_NEEDS_COMMENT (7, "-Wmaybe-uninitialized");
 196 /* Handling of Unicode 3.1 TAG characters.  Unicode recommends
 197    "If language codes are not relevant to the particular processing
 198     operation, then they should be ignored."  This macro is usually
 199    called right before  STANDARD_TO_LOOP_ERR_HANDLER (Incr).  */
 200 #define UNICODE_TAG_HANDLER(Character, Incr) \
 201   {                                                                           \
 202     /* TAG characters are those in the range U+E0000..U+E007F.  */            \
 203     if (((Character) >> 7) == (0xe0000 >> 7))                                 \
 204       {                                                                       \
 205         inptr += Incr;                                                        \
 206         continue;                                                             \
 207       }                                                                       \
 208   }
 209 DIAG_POP_NEEDS_COMMENT;
 210
 211
 212 /* The function returns the status, as defined in gconv.h.  */
 213 static inline int
 214 __attribute ((always_inline))
 215 FCTNAME (LOOPFCT) (struct __gconv_step *step,
 216                    struct __gconv_step_data *step_data,
 217                    const unsigned char **inptrp, const unsigned char *inend,
 218                    unsigned char **outptrp, const unsigned char *outend,
 219                    size_t *irreversible EXTRA_LOOP_DECLS)
 220 {
 221 #ifdef LOOP_NEED_STATE
 222   mbstate_t *state = step_data->__statep;
 223 #endif
 224 #ifdef LOOP_NEED_FLAGS
 225   int flags = step_data->__flags;
 226 #endif
 227 #ifdef LOOP_NEED_DATA
 228   void *data = step->__data;
 229 #endif
 230   int result = __GCONV_EMPTY_INPUT;
 231   const unsigned char *inptr = *inptrp;
 232   unsigned char *outptr = *outptrp;
 233
 234 #ifdef INIT_PARAMS
 235   INIT_PARAMS;
 236 #endif
 237
 238   while (inptr != inend)
 239     {
 240       /* `if' cases for MIN_NEEDED_OUTPUT ==/!= 1 is made to help the
 241          compiler generating better code.  They will be optimized away
 242          since MIN_NEEDED_OUTPUT is always a constant.  */
 243       if (MIN_NEEDED_INPUT > 1
 244           && __builtin_expect (inptr + MIN_NEEDED_INPUT > inend, 0))
 245         {
 246           /* We don't have enough input for another complete input
 247              character.  */
 248           result = __GCONV_INCOMPLETE_INPUT;
 249           break;
 250         }
 251       if ((MIN_NEEDED_OUTPUT != 1
 252            && __builtin_expect (outptr + MIN_NEEDED_OUTPUT > outend, 0))
 253           || (MIN_NEEDED_OUTPUT == 1
 254               && __builtin_expect (outptr >= outend, 0)))
 255         {
 256           /* Overflow in the output buffer.  */
 257           result = __GCONV_FULL_OUTPUT;
 258           break;
 259         }
 260
 261       /* Here comes the body the user provides.  It can stop with
 262          RESULT set to GCONV_INCOMPLETE_INPUT (if the size of the
 263          input characters vary in size), GCONV_ILLEGAL_INPUT, or
 264          GCONV_FULL_OUTPUT (if the output characters vary in size).  */
 265       BODY
 266     }
 267
 268   /* Update the pointers pointed to by the parameters.  */
 269   *inptrp = inptr;
 270   *outptrp = outptr;
 271   UPDATE_PARAMS;
 272
 273   return result;
 274 }
 275
 276
 277 #if MAX_NEEDED_INPUT > 1
 278 # define SINGLE(fct) SINGLE2 (fct)
 279 # define SINGLE2(fct) fct##_single
 280 static inline int
 281 __attribute ((always_inline))
 282 SINGLE(LOOPFCT) (struct __gconv_step *step,
 283                  struct __gconv_step_data *step_data,
 284                  const unsigned char **inptrp, const unsigned char *inend,
 285                  unsigned char **outptrp, unsigned char *outend,
 286                  size_t *irreversible EXTRA_LOOP_DECLS)
 287 {
 288   mbstate_t *state = step_data->__statep;
 289 # ifdef LOOP_NEED_FLAGS
 290   int flags = step_data->__flags;
 291 # endif
 292 # ifdef LOOP_NEED_DATA
 293   void *data = step->__data;
 294 # endif
 295   int result = __GCONV_OK;
 296   unsigned char bytebuf[MAX_NEEDED_INPUT];
 297   const unsigned char *inptr = *inptrp;
 298   unsigned char *outptr = *outptrp;
 299   size_t inlen;
 300
 301 # ifdef INIT_PARAMS
 302   INIT_PARAMS;
 303 # endif
 304
 305 # ifdef UNPACK_BYTES
 306   UNPACK_BYTES
 307 # else
 308   /* Add the bytes from the state to the input buffer.  */
 309   assert ((state->__count & 7) <= sizeof (state->__value));
 310   for (inlen = 0; inlen < (size_t) (state->__count & 7); ++inlen)
 311     bytebuf[inlen] = state->__value.__wchb[inlen];
 312 # endif
 313
 314   /* Are there enough bytes in the input buffer?  */
 315   if (MIN_NEEDED_INPUT > 1
 316       && __builtin_expect (inptr + (MIN_NEEDED_INPUT - inlen) > inend, 0))
 317     {
 318       *inptrp = inend;
 319 # ifdef STORE_REST
 320
 321       /* Building with -O3 GCC emits a `array subscript is above array
 322          bounds' warning.  GCC BZ #64739 has been opened for this.  */
 323       DIAG_PUSH_NEEDS_COMMENT;
 324       DIAG_IGNORE_NEEDS_COMMENT (4.9, "-Warray-bounds");
 325       while (inptr < inend)
 326         bytebuf[inlen++] = *inptr++;
 327       DIAG_POP_NEEDS_COMMENT;
 328
 329       inptr = bytebuf;
 330       inptrp = &inptr;
 331       inend = &bytebuf[inlen];
 332
 333       STORE_REST
 334 # else
 335       /* We don't have enough input for another complete input
 336          character.  */
 337       size_t inlen_after = inlen + (inend - inptr);
 338       assert (inlen_after <= sizeof (state->__value.__wchb));
 339       for (; inlen < inlen_after; inlen++)
 340         state->__value.__wchb[inlen] = *inptr++;
 341 # endif
 342
 343       return __GCONV_INCOMPLETE_INPUT;
 344     }
 345
 346   /* Enough space in output buffer.  */
 347   if ((MIN_NEEDED_OUTPUT != 1 && outptr + MIN_NEEDED_OUTPUT > outend)
 348       || (MIN_NEEDED_OUTPUT == 1 && outptr >= outend))
 349     /* Overflow in the output buffer.  */
 350     return __GCONV_FULL_OUTPUT;
 351
 352   /*  Now add characters from the normal input buffer.  */
 353   if (inlen >= MAX_NEEDED_INPUT || inptr >= inend)
 354     /* Avoid a -Wstringop-overflow= warning when this loop is
 355        unrolled.  The compiler cannot otherwise see that this is
 356        unreachable because it depends on (state->__count & 7) not
 357        being too large after a previous conversion step.
 358        Starting with GCC 12, we also have mark the inptr >= inend
 359        case as unreachable to omit the warning.  Note that this SINGLE
 360        function is only used to implement the mb*towc*() or wc*tomb*()
 361        functions.  Those functions use inptr and inend pointing to a
 362        variable on stack, compute the inend pointer or explicitly check
 363        the arguments which always leads to inptr < inend.  */
 364     __builtin_unreachable ();
 365   do
 366     bytebuf[inlen++] = *inptr++;
 367   while (inlen < MAX_NEEDED_INPUT && inptr < inend);
 368
 369   inptr = bytebuf;
 370   inend = &bytebuf[inlen];
 371
 372   do
 373     {
 374       BODY
 375     }
 376   while (0);
 377
 378   /* Now we either have produced an output character and consumed all the
 379      bytes from the state and at least one more, or the character is still
 380      incomplete, or we have some other error (like illegal input character,
 381      no space in output buffer).  */
 382   if (__glibc_likely (inptr != bytebuf))
 383     {
 384       /* We found a new character.  */
 385       assert (inptr - bytebuf > (state->__count & 7));
 386
 387       *inptrp += inptr - bytebuf - (state->__count & 7);
 388       *outptrp = outptr;
 389
 390       result = __GCONV_OK;
 391
 392       /* Clear the state buffer.  */
 393 # ifdef CLEAR_STATE
 394       CLEAR_STATE;
 395 # else
 396       state->__count &= ~7;
 397 # endif
 398     }
 399   else if (result == __GCONV_INCOMPLETE_INPUT)
 400     {
 401       /* This can only happen if we have less than MAX_NEEDED_INPUT bytes
 402          available.  */
 403       assert (inend != &bytebuf[MAX_NEEDED_INPUT]);
 404
 405       *inptrp += inend - bytebuf - (state->__count & 7);
 406 # ifdef STORE_REST
 407       inptrp = &inptr;
 408
 409       STORE_REST
 410 # else
 411       /* We don't have enough input for another complete input
 412          character.  */
 413       assert (inend - inptr > (state->__count & ~7));
 414       assert (inend - inptr <= sizeof (state->__value.__wchb));
 415       state->__count = (state->__count & ~7) | (inend - inptr);
 416       for (inlen = 0; inlen < inend - inptr; inlen++)
 417         state->__value.__wchb[inlen] = inptr[inlen];
 418       inptr = inend;
 419 # endif
 420     }
 421
 422   return result;
 423 }
 424 # undef SINGLE
 425 # undef SINGLE2
 426
 427
 428 # ifdef ONEBYTE_BODY
 429 /* Define the shortcut function for btowc.  */
 430 static wint_t
 431 gconv_btowc (struct __gconv_step *step, unsigned char c)
 432   ONEBYTE_BODY
 433 #  define FROM_ONEBYTE gconv_btowc
 434 # endif
 435
 436 #endif
 437
 438 /* We remove the macro definitions so that we can include this file again
 439    for the definition of another function.  */
 440 #undef MIN_NEEDED_INPUT
 441 #undef MAX_NEEDED_INPUT
 442 #undef MIN_NEEDED_OUTPUT
 443 #undef MAX_NEEDED_OUTPUT
 444 #undef LOOPFCT
 445 #undef BODY
 446 #undef LOOPFCT
 447 #undef EXTRA_LOOP_DECLS
 448 #undef INIT_PARAMS
 449 #undef UPDATE_PARAMS
 450 #undef REINIT_PARAMS
 451 #undef ONEBYTE_BODY
 452 #undef UNPACK_BYTES
 453 #undef CLEAR_STATE
 454 #undef LOOP_NEED_STATE
 455 #undef LOOP_NEED_FLAGS
 456 #undef LOOP_NEED_DATA