gas/input-scrub.c

   1 /* input_scrub.c - Break up input buffers into whole numbers of lines.
   2    Copyright (C) 1987-2023 Free Software Foundation, Inc.
   3
   4    This file is part of GAS, the GNU Assembler.
   5
   6    GAS is free software; you can redistribute it and/or modify
   7    it under the terms of the GNU General Public License as published by
   8    the Free Software Foundation; either version 3, or (at your option)
   9    any later version.
  10
  11    GAS 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
  14    GNU General Public License for more details.
  15
  16    You should have received a copy of the GNU General Public License
  17    along with GAS; see the file COPYING.  If not, write to the Free
  18    Software Foundation, 51 Franklin Street - Fifth Floor, Boston, MA
  19    02110-1301, USA.  */
  20
  21 #include "as.h"
  22 #include "filenames.h"
  23 #include "input-file.h"
  24 #include "sb.h"
  25 #include "listing.h"
  26
  27 /*
  28  * O/S independent module to supply buffers of sanitised source code
  29  * to rest of assembler.  We get sanitised input data of arbitrary length.
  30  * We break these buffers on line boundaries, recombine pieces that
  31  * were broken across buffers, and return a buffer of full lines to
  32  * the caller.
  33  * The last partial line begins the next buffer we build and return to caller.
  34  * The buffer returned to caller is preceded by BEFORE_STRING and followed
  35  * by AFTER_STRING, as sentinels. The last character before AFTER_STRING
  36  * is a newline.
  37  * Also looks after line numbers, for e.g. error messages.
  38  */
  39
  40 /*
  41  * We don't care how filthy our buffers are, but our callers assume
  42  * that the following sanitation has already been done.
  43  *
  44  * No comments, reduce a comment to a space.
  45  * Reduce a tab to a space unless it is 1st char of line.
  46  * All multiple tabs and spaces collapsed into 1 char. Tab only
  47  *   legal if 1st char of line.
  48  * # line file statements converted to .line x;.file y; statements.
  49  * Escaped newlines at end of line: remove them but add as many newlines
  50  *   to end of statement as you removed in the middle, to synch line numbers.
  51  */
  52 \f
  53 #define BEFORE_STRING ("\n")
  54 #define AFTER_STRING ("\0")     /* memcpy of 0 chars might choke.  */
  55 #define BEFORE_SIZE (1)
  56 #define AFTER_SIZE  (1)
  57
  58 #ifndef TC_EOL_IN_INSN
  59 #define TC_EOL_IN_INSN(P) 0
  60 #endif
  61
  62 static char *buffer_start;      /*->1st char of full buffer area.  */
  63 static char *partial_where;     /*->after last full line in buffer.  */
  64 static size_t partial_size;     /* >=0. Number of chars in partial line in buffer.  */
  65
  66 /* Because we need AFTER_STRING just after last full line, it clobbers
  67    1st part of partial line. So we preserve 1st part of partial line
  68    here.  */
  69 static char save_source[AFTER_SIZE];
  70
  71 /* The size of the input buffer we concatenate
  72    input_file_give_next_buffer chunks into.  Excludes the BEFORE and
  73    AFTER counts.  */
  74 static size_t buffer_length;
  75
  76 /* The index into an sb structure we are reading from.  -1 if none.  */
  77 static size_t sb_index = -1;
  78
  79 /* If we are reading from an sb structure, this is it.  */
  80 static sb from_sb;
  81
  82 /* Should we do a conditional check on from_sb? */
  83 static enum expansion from_sb_expansion = expanding_none;
  84
  85 /* The number of nested sb structures we have included.  */
  86 int macro_nest;
  87
  88 /* We can have more than one source file open at once, though the info for all
  89    but the latest one are saved off in a struct input_save.  These files remain
  90    open, so we are limited by the number of open files allowed by the
  91    underlying OS. We may also sequentially read more than one source file in an
  92    assembly.  */
  93
  94 /* We must track the physical file and line number for error messages. We also
  95    track a "logical" file and line number corresponding to (C?)  compiler
  96    source line numbers.  Whenever we open a file we must fill in
  97    physical_input_file. So if it is NULL we have not opened any files yet.  */
  98
  99 static const char *physical_input_file;
 100 static const char *logical_input_file;
 101
 102 /* 1-origin line number in a source file.  */
 103 /* A line ends in '\n' or eof.  */
 104 static unsigned int physical_input_line;
 105 static unsigned int logical_input_line;
 106
 107 /* Indicator whether the origin of an update was a .linefile directive. */
 108 static bool is_linefile;
 109
 110 /* Struct used to save the state of the input handler during include files */
 111 struct input_save {
 112   char *              buffer_start;
 113   char *              partial_where;
 114   size_t              partial_size;
 115   char                save_source[AFTER_SIZE];
 116   size_t              buffer_length;
 117   const char *        physical_input_file;
 118   const char *        logical_input_file;
 119   unsigned int        physical_input_line;
 120   unsigned int        logical_input_line;
 121   bool                is_linefile;
 122   size_t              sb_index;
 123   sb                  from_sb;
 124   enum expansion      from_sb_expansion; /* Should we do a conditional check?  */
 125   struct input_save * next_saved_file;  /* Chain of input_saves.  */
 126   char *              input_file_save;  /* Saved state of input routines.  */
 127   char *              saved_position;   /* Caller's saved position in buf.  */
 128 };
 129
 130 static struct input_save *input_scrub_push (char *saved_position);
 131 static char *input_scrub_pop (struct input_save *arg);
 132
 133 /* Saved information about the file that .include'd this one.  When we hit EOF,
 134    we automatically pop to that file.  */
 135
 136 static struct input_save *next_saved_file;
 137
 138 /* Initialize input buffering.  */
 139
 140 static void
 141 input_scrub_reinit (void)
 142 {
 143   input_file_begin ();          /* Reinitialize! */
 144   logical_input_line = -1u;
 145   logical_input_file = NULL;
 146   sb_index = -1;
 147
 148   buffer_length = input_file_buffer_size () * 2;
 149   buffer_start = XNEWVEC (char, BEFORE_SIZE + AFTER_SIZE + 1 + buffer_length);
 150   memcpy (buffer_start, BEFORE_STRING, (int) BEFORE_SIZE);
 151 }
 152
 153 /* Push the state of input reading and scrubbing so that we can #include.
 154    The return value is a 'void *' (fudged for old compilers) to a save
 155    area, which can be restored by passing it to input_scrub_pop().  */
 156
 157 static struct input_save *
 158 input_scrub_push (char *saved_position)
 159 {
 160   struct input_save *saved;
 161
 162   saved = XNEW (struct input_save);
 163
 164   saved->saved_position = saved_position;
 165   saved->buffer_start = buffer_start;
 166   saved->partial_where = partial_where;
 167   saved->partial_size = partial_size;
 168   saved->buffer_length = buffer_length;
 169   saved->physical_input_file = physical_input_file;
 170   saved->logical_input_file = logical_input_file;
 171   saved->physical_input_line = physical_input_line;
 172   saved->logical_input_line = logical_input_line;
 173   saved->is_linefile = is_linefile;
 174   saved->sb_index = sb_index;
 175   saved->from_sb = from_sb;
 176   saved->from_sb_expansion = from_sb_expansion;
 177   memcpy (saved->save_source, save_source, sizeof (save_source));
 178   saved->next_saved_file = next_saved_file;
 179   saved->input_file_save = input_file_push ();
 180
 181   input_scrub_reinit ();
 182
 183   return saved;
 184 }
 185
 186 static char *
 187 input_scrub_pop (struct input_save *saved)
 188 {
 189   char *saved_position;
 190
 191   input_scrub_end ();           /* Finish off old buffer */
 192
 193   input_file_pop (saved->input_file_save);
 194   saved_position = saved->saved_position;
 195   buffer_start = saved->buffer_start;
 196   buffer_length = saved->buffer_length;
 197   physical_input_file = saved->physical_input_file;
 198   logical_input_file = saved->logical_input_file;
 199   physical_input_line = saved->physical_input_line;
 200   logical_input_line = saved->logical_input_line;
 201   is_linefile = saved->is_linefile;
 202   sb_index = saved->sb_index;
 203   from_sb = saved->from_sb;
 204   from_sb_expansion = saved->from_sb_expansion;
 205   partial_where = saved->partial_where;
 206   partial_size = saved->partial_size;
 207   next_saved_file = saved->next_saved_file;
 208   memcpy (save_source, saved->save_source, sizeof (save_source));
 209
 210   free (saved);
 211   return saved_position;
 212 }
 213 \f
 214 void
 215 input_scrub_begin (void)
 216 {
 217   know (strlen (BEFORE_STRING) == BEFORE_SIZE);
 218   know (strlen (AFTER_STRING) == AFTER_SIZE
 219         || (AFTER_STRING[0] == '\0' && AFTER_SIZE == 1));
 220
 221   physical_input_file = NULL;   /* No file read yet.  */
 222   next_saved_file = NULL;       /* At EOF, don't pop to any other file */
 223   input_scrub_reinit ();
 224   do_scrub_begin (flag_m68k_mri);
 225 }
 226
 227 void
 228 input_scrub_end (void)
 229 {
 230   if (buffer_start)
 231     {
 232       free (buffer_start);
 233       buffer_start = 0;
 234       input_file_end ();
 235     }
 236 }
 237
 238 /* Start reading input from a new file.
 239    Return start of caller's part of buffer.  */
 240
 241 char *
 242 input_scrub_new_file (const char *filename)
 243 {
 244   input_file_open (filename, !flag_no_comments);
 245   physical_input_file = filename[0] ? filename : _("{standard input}");
 246   physical_input_line = 0;
 247
 248   partial_size = 0;
 249   return (buffer_start + BEFORE_SIZE);
 250 }
 251
 252 /* Include a file from the current file.  Save our state, cause it to
 253    be restored on EOF, and begin handling a new file.  Same result as
 254    input_scrub_new_file.  */
 255
 256 char *
 257 input_scrub_include_file (const char *filename, char *position)
 258 {
 259   next_saved_file = input_scrub_push (position);
 260   from_sb_expansion = expanding_none;
 261   return input_scrub_new_file (filename);
 262 }
 263
 264 /* Start getting input from an sb structure.  This is used when
 265    expanding a macro.  */
 266
 267 void
 268 input_scrub_include_sb (sb *from, char *position, enum expansion expansion)
 269 {
 270   int newline;
 271
 272   if (macro_nest > max_macro_nest)
 273     as_fatal (_("macros nested too deeply"));
 274   ++macro_nest;
 275
 276 #ifdef md_macro_start
 277   if (expansion == expanding_macro)
 278     {
 279       md_macro_start ();
 280     }
 281 #endif
 282
 283   next_saved_file = input_scrub_push (position);
 284
 285   /* Allocate sufficient space: from->len plus optional newline
 286      plus two ".linefile " directives, plus a little more for other
 287      expansion.  */
 288   newline = from->len >= 1 && from->ptr[0] != '\n';
 289   sb_build (&from_sb, from->len + newline + 2 * sizeof (".linefile") + 30);
 290   from_sb_expansion = expansion;
 291   if (newline)
 292     {
 293       /* Add the sentinel required by read.c.  */
 294       sb_add_char (&from_sb, '\n');
 295     }
 296   sb_scrub_and_add_sb (&from_sb, from);
 297
 298   /* Make sure the parser looks at defined contents when it scans for
 299      e.g. end-of-line at the end of a macro.  */
 300   sb_terminate (&from_sb);
 301
 302   sb_index = 1;
 303
 304   /* These variables are reset by input_scrub_push.  Restore them
 305      since we are, after all, still at the same point in the file.  */
 306   logical_input_line = next_saved_file->logical_input_line;
 307   logical_input_file = next_saved_file->logical_input_file;
 308 }
 309
 310 void
 311 input_scrub_close (void)
 312 {
 313   input_file_close ();
 314   physical_input_line = 0;
 315   logical_input_line = -1u;
 316 }
 317
 318 char *
 319 input_scrub_next_buffer (char **bufp)
 320 {
 321   char *limit;          /*->just after last char of buffer.  */
 322
 323   if (sb_index != (size_t) -1)
 324     {
 325       if (sb_index >= from_sb.len)
 326         {
 327           sb_kill (&from_sb);
 328           if (from_sb_expansion == expanding_macro)
 329             {
 330               cond_finish_check (macro_nest);
 331 #ifdef md_macro_end
 332               /* Allow the target to clean up per-macro expansion
 333                  data.  */
 334               md_macro_end ();
 335 #endif
 336             }
 337           --macro_nest;
 338           partial_where = NULL;
 339           partial_size = 0;
 340           if (next_saved_file != NULL)
 341             *bufp = input_scrub_pop (next_saved_file);
 342           return partial_where;
 343         }
 344
 345       partial_where = from_sb.ptr + from_sb.len;
 346       partial_size = 0;
 347       *bufp = from_sb.ptr + sb_index;
 348       sb_index = from_sb.len;
 349       return partial_where;
 350     }
 351
 352   if (partial_size)
 353     {
 354       memmove (buffer_start + BEFORE_SIZE, partial_where, partial_size);
 355       memcpy (buffer_start + BEFORE_SIZE, save_source, AFTER_SIZE);
 356     }
 357
 358   while (1)
 359     {
 360       char *p;
 361       char *start = buffer_start + BEFORE_SIZE + partial_size;
 362
 363       *bufp = buffer_start + BEFORE_SIZE;
 364       limit = input_file_give_next_buffer (start);
 365       if (!limit)
 366         {
 367           if (!partial_size)
 368             /* End of this file.  */
 369             break;
 370
 371           as_warn (_("end of file not at end of a line; newline inserted"));
 372           p = buffer_start + BEFORE_SIZE + partial_size;
 373           *p++ = '\n';
 374           limit = p;
 375         }
 376       else
 377         {
 378           /* Terminate the buffer to avoid confusing TC_EOL_IN_INSN.  */
 379           *limit = '\0';
 380
 381           /* Find last newline.  */
 382           for (p = limit - 1; *p != '\n' || TC_EOL_IN_INSN (p); --p)
 383             if (p < start)
 384               goto read_more;
 385           ++p;
 386         }
 387
 388       if (multibyte_handling == multibyte_warn)
 389         (void) scan_for_multibyte_characters ((const unsigned char *) p,
 390                                               (const unsigned char *) limit,
 391                                               true /* Generate warnings */);
 392
 393       /* We found a newline in the newly read chars.  */
 394       partial_where = p;
 395       partial_size = limit - p;
 396
 397       /* Save the fragment after that last newline.  */
 398       memcpy (save_source, partial_where, (int) AFTER_SIZE);
 399       memcpy (partial_where, AFTER_STRING, (int) AFTER_SIZE);
 400       return partial_where;
 401
 402     read_more:
 403       /* Didn't find a newline.  Read more text.  */
 404       partial_size = limit - (buffer_start + BEFORE_SIZE);
 405       if (buffer_length - input_file_buffer_size () < partial_size)
 406         {
 407           /* Increase the buffer when it doesn't have room for the
 408              next block of input.  */
 409           buffer_length *= 2;
 410           buffer_start = XRESIZEVEC (char, buffer_start,
 411                                      (buffer_length
 412                                       + BEFORE_SIZE + AFTER_SIZE + 1));
 413         }
 414     }
 415
 416   /* Tell the listing we've finished the file.  */
 417   LISTING_EOF ();
 418
 419   /* If we should pop to another file at EOF, do it.  */
 420   partial_where = NULL;
 421   if (next_saved_file)
 422     *bufp = input_scrub_pop (next_saved_file);
 423
 424   return partial_where;
 425 }
 426 \f
 427 /* The remaining part of this file deals with line numbers, error
 428    messages and so on.  Return TRUE if we opened any file.  */
 429
 430 int
 431 seen_at_least_1_file (void)
 432 {
 433   return (physical_input_file != NULL);
 434 }
 435
 436 void
 437 bump_line_counters (void)
 438 {
 439   if (sb_index == (size_t) -1)
 440     ++physical_input_line;
 441
 442   if (logical_input_line != -1u)
 443     ++logical_input_line;
 444 }
 445 \f
 446 /* Tells us what the new logical line number and file are.
 447    If the line_number is -1, we don't change the current logical line
 448    number.
 449    If fname is NULL, we don't change the current logical file name, unless
 450    bit 3 of flags is set.
 451    Returns nonzero if the filename actually changes.  */
 452
 453 void
 454 new_logical_line_flags (const char *fname, /* DON'T destroy it!  We point to it!  */
 455                         int line_number,
 456                         int flags)
 457 {
 458   switch (flags)
 459     {
 460     case 0:
 461       break;
 462     case 1:
 463       if (line_number != -1)
 464         abort ();
 465       break;
 466     case 1 << 1:
 467     case 1 << 2:
 468       /* FIXME: we could check that include nesting is correct.  */
 469       break;
 470     case 1 << 3:
 471       if (line_number < 0 || fname != NULL)
 472         abort ();
 473       if (next_saved_file == NULL)
 474         fname = physical_input_file;
 475       else if (next_saved_file->logical_input_file)
 476         fname = next_saved_file->logical_input_file;
 477       else
 478         fname = next_saved_file->physical_input_file;
 479       break;
 480     default:
 481       abort ();
 482     }
 483
 484   is_linefile = flags != 1 && (flags != 0 || fname);
 485
 486   if (line_number >= 0)
 487     logical_input_line = line_number;
 488   else if (line_number == -1 && fname && !*fname && (flags & (1 << 2)))
 489     {
 490       logical_input_file = physical_input_file;
 491       logical_input_line = physical_input_line;
 492       fname = NULL;
 493     }
 494
 495   if (fname
 496       && (logical_input_file == NULL
 497           || filename_cmp (logical_input_file, fname)))
 498     logical_input_file = fname;
 499 }
 500
 501 void
 502 new_logical_line (const char *fname, int line_number)
 503 {
 504   new_logical_line_flags (fname, line_number, 0);
 505 }
 506
 507 void
 508 as_report_context (void)
 509 {
 510   const struct input_save *saved = next_saved_file;
 511   enum expansion expansion = from_sb_expansion;
 512   int indent = 1;
 513
 514   if (!macro_nest)
 515     return;
 516
 517   do
 518     {
 519       if (expansion != expanding_macro)
 520         /* Nothing.  */;
 521       else if (saved->logical_input_file != NULL
 522                && saved->logical_input_line != -1u)
 523         as_info_where (saved->logical_input_file, saved->logical_input_line,
 524                        indent, _("macro invoked from here"));
 525       else
 526         as_info_where (saved->physical_input_file, saved->physical_input_line,
 527                        indent, _("macro invoked from here"));
 528
 529       expansion = saved->from_sb_expansion;
 530       ++indent;
 531     }
 532   while ((saved = saved->next_saved_file) != NULL);
 533 }
 534 \f
 535 /* Return the current physical input file name and line number, if known  */
 536
 537 const char *
 538 as_where_physical (unsigned int *linep)
 539 {
 540   if (physical_input_file != NULL)
 541     {
 542       if (linep != NULL)
 543         *linep = physical_input_line;
 544       return physical_input_file;
 545     }
 546
 547   if (linep != NULL)
 548     *linep = 0;
 549   return NULL;
 550 }
 551
 552 /* Return the file name and line number at the top most macro
 553    invocation, unless .file / .line were used inside a macro.  */
 554
 555 const char *
 556 as_where (unsigned int *linep)
 557 {
 558   const char *file = as_where_top (linep);
 559
 560   if (macro_nest && is_linefile)
 561     {
 562       const struct input_save *saved = next_saved_file;
 563       enum expansion expansion = from_sb_expansion;
 564
 565       do
 566         {
 567           if (expansion != expanding_macro)
 568             /* Nothing.  */;
 569           else if (saved->logical_input_file != NULL
 570                    && (linep == NULL || saved->logical_input_line != -1u))
 571             {
 572               if (linep != NULL)
 573                 *linep = saved->logical_input_line;
 574               file = saved->logical_input_file;
 575             }
 576           else if (saved->physical_input_file != NULL)
 577             {
 578               if (linep != NULL)
 579                 *linep = saved->physical_input_line;
 580               file = saved->physical_input_file;
 581             }
 582
 583           expansion = saved->from_sb_expansion;
 584         }
 585       while ((saved = saved->next_saved_file) != NULL);
 586     }
 587
 588   return file;
 589 }
 590
 591 /* Return the current file name and line number.  */
 592
 593 const char *
 594 as_where_top (unsigned int *linep)
 595 {
 596   if (logical_input_file != NULL
 597       && (linep == NULL || logical_input_line != -1u))
 598     {
 599       if (linep != NULL)
 600         *linep = logical_input_line;
 601       return logical_input_file;
 602     }
 603
 604   return as_where_physical (linep);
 605 }