gas/input-scrub.c

   1 /* input_scrub.c - Break up input buffers into whole numbers of lines.
   2    Copyright (C) 1987-2024 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
 198   /* When expanding an #APP / #NO_APP block, original lines are re-
 199      processed, so whatever they did to physical file/line needs
 200      retaining.  If logical file/line weren't changed, the logical
 201      line number will want bumping by a corresponding value.  */
 202   if (from_sb_expansion != expanding_app)
 203     {
 204       if (logical_input_file == 0 && logical_input_line == -1u
 205           && saved->logical_input_line != -1u)
 206         saved->logical_input_line
 207           += physical_input_line - saved->physical_input_line;
 208       physical_input_file = saved->physical_input_file;
 209       physical_input_line = saved->physical_input_line;
 210     }
 211   logical_input_file = saved->logical_input_file;
 212   logical_input_line = saved->logical_input_line;
 213
 214   is_linefile = saved->is_linefile;
 215   sb_index = saved->sb_index;
 216   from_sb = saved->from_sb;
 217   from_sb_expansion = saved->from_sb_expansion;
 218   partial_where = saved->partial_where;
 219   partial_size = saved->partial_size;
 220   next_saved_file = saved->next_saved_file;
 221   memcpy (save_source, saved->save_source, sizeof (save_source));
 222
 223   free (saved);
 224   return saved_position;
 225 }
 226 \f
 227 void
 228 input_scrub_begin (void)
 229 {
 230   know (strlen (BEFORE_STRING) == BEFORE_SIZE);
 231   know (strlen (AFTER_STRING) == AFTER_SIZE
 232         || (AFTER_STRING[0] == '\0' && AFTER_SIZE == 1));
 233
 234   physical_input_file = NULL;   /* No file read yet.  */
 235   next_saved_file = NULL;       /* At EOF, don't pop to any other file */
 236   macro_nest = 0;
 237   input_scrub_reinit ();
 238   do_scrub_begin (flag_m68k_mri);
 239 }
 240
 241 void
 242 input_scrub_end (void)
 243 {
 244   if (buffer_start)
 245     {
 246       free (buffer_start);
 247       buffer_start = 0;
 248       input_file_end ();
 249     }
 250 }
 251
 252 /* Start reading input from a new file.
 253    Return start of caller's part of buffer.  */
 254
 255 char *
 256 input_scrub_new_file (const char *filename)
 257 {
 258   input_file_open (filename, !flag_no_comments);
 259   physical_input_file = filename[0] ? filename : _("{standard input}");
 260   physical_input_line = 0;
 261
 262   partial_size = 0;
 263   return (buffer_start + BEFORE_SIZE);
 264 }
 265
 266 /* Include a file from the current file.  Save our state, cause it to
 267    be restored on EOF, and begin handling a new file.  Same result as
 268    input_scrub_new_file.  */
 269
 270 char *
 271 input_scrub_include_file (const char *filename, char *position)
 272 {
 273   next_saved_file = input_scrub_push (position);
 274   from_sb_expansion = expanding_none;
 275   return input_scrub_new_file (filename);
 276 }
 277
 278 /* Start getting input from an sb structure.  This is used when
 279    expanding a macro.  */
 280
 281 void
 282 input_scrub_include_sb (sb *from, char *position, enum expansion expansion)
 283 {
 284   int newline;
 285
 286   if (expansion != expanding_app)
 287     {
 288       if (macro_nest > max_macro_nest)
 289         as_fatal (_("macros nested too deeply"));
 290       ++macro_nest;
 291     }
 292
 293 #ifdef md_macro_start
 294   if (expansion == expanding_macro)
 295     {
 296       md_macro_start ();
 297     }
 298 #endif
 299
 300   next_saved_file = input_scrub_push (position);
 301
 302   /* Allocate sufficient space: from->len plus optional newline
 303      plus two ".linefile " directives, plus a little more for other
 304      expansion.  */
 305   newline = from->len >= 1 && from->ptr[0] != '\n';
 306   sb_build (&from_sb, from->len + newline + 2 * sizeof (".linefile") + 30);
 307   from_sb_expansion = expansion;
 308   if (newline)
 309     {
 310       /* Add the sentinel required by read.c.  */
 311       sb_add_char (&from_sb, '\n');
 312     }
 313   sb_scrub_and_add_sb (&from_sb, from);
 314
 315   /* Make sure the parser looks at defined contents when it scans for
 316      e.g. end-of-line at the end of a macro.  */
 317   sb_terminate (&from_sb);
 318
 319   sb_index = 1;
 320
 321   /* These variables are reset by input_scrub_push.  Restore them
 322      since we are, after all, still at the same point in the file.  */
 323   logical_input_line = next_saved_file->logical_input_line;
 324   logical_input_file = next_saved_file->logical_input_file;
 325 }
 326
 327 void
 328 input_scrub_close (void)
 329 {
 330   input_file_close ();
 331   physical_input_line = 0;
 332   logical_input_line = -1u;
 333 }
 334
 335 char *
 336 input_scrub_next_buffer (char **bufp)
 337 {
 338   char *limit;          /*->just after last char of buffer.  */
 339
 340   if (sb_index != (size_t) -1)
 341     {
 342       if (sb_index >= from_sb.len)
 343         {
 344           sb_kill (&from_sb);
 345           if (from_sb_expansion == expanding_macro)
 346             {
 347               cond_finish_check (macro_nest);
 348 #ifdef md_macro_end
 349               /* Allow the target to clean up per-macro expansion
 350                  data.  */
 351               md_macro_end ();
 352 #endif
 353             }
 354           if (from_sb_expansion != expanding_app)
 355             --macro_nest;
 356           partial_where = NULL;
 357           partial_size = 0;
 358           if (next_saved_file != NULL)
 359             *bufp = input_scrub_pop (next_saved_file);
 360           return partial_where;
 361         }
 362
 363       partial_where = from_sb.ptr + from_sb.len;
 364       partial_size = 0;
 365       *bufp = from_sb.ptr + sb_index;
 366       sb_index = from_sb.len;
 367       return partial_where;
 368     }
 369
 370   if (partial_size)
 371     {
 372       memmove (buffer_start + BEFORE_SIZE, partial_where, partial_size);
 373       memcpy (buffer_start + BEFORE_SIZE, save_source, AFTER_SIZE);
 374     }
 375
 376   while (1)
 377     {
 378       char *p;
 379       char *start = buffer_start + BEFORE_SIZE + partial_size;
 380
 381       *bufp = buffer_start + BEFORE_SIZE;
 382       limit = input_file_give_next_buffer (start);
 383       if (!limit)
 384         {
 385           if (!partial_size)
 386             /* End of this file.  */
 387             break;
 388
 389           as_warn (_("end of file not at end of a line; newline inserted"));
 390           p = buffer_start + BEFORE_SIZE + partial_size;
 391           *p++ = '\n';
 392           limit = p;
 393         }
 394       else
 395         {
 396           /* Terminate the buffer to avoid confusing TC_EOL_IN_INSN.  */
 397           *limit = '\0';
 398
 399           /* Find last newline.  */
 400           for (p = limit - 1; *p != '\n' || TC_EOL_IN_INSN (p); --p)
 401             if (p < start)
 402               goto read_more;
 403           ++p;
 404         }
 405
 406       /* We found a newline in the newly read chars.  */
 407       partial_where = p;
 408       partial_size = limit - p;
 409
 410       /* Save the fragment after that last newline.  */
 411       memcpy (save_source, partial_where, (int) AFTER_SIZE);
 412       memcpy (partial_where, AFTER_STRING, (int) AFTER_SIZE);
 413       return partial_where;
 414
 415     read_more:
 416       /* Didn't find a newline.  Read more text.  */
 417       partial_size = limit - (buffer_start + BEFORE_SIZE);
 418       if (buffer_length - input_file_buffer_size () < partial_size)
 419         {
 420           /* Increase the buffer when it doesn't have room for the
 421              next block of input.  */
 422           buffer_length *= 2;
 423           buffer_start = XRESIZEVEC (char, buffer_start,
 424                                      (buffer_length
 425                                       + BEFORE_SIZE + AFTER_SIZE + 1));
 426         }
 427     }
 428
 429   /* Tell the listing we've finished the file.  */
 430   LISTING_EOF ();
 431
 432   /* If we should pop to another file at EOF, do it.  */
 433   partial_where = NULL;
 434   if (next_saved_file)
 435     *bufp = input_scrub_pop (next_saved_file);
 436
 437   return partial_where;
 438 }
 439 \f
 440 /* The remaining part of this file deals with line numbers, error
 441    messages and so on.  Return TRUE if we opened any file.  */
 442
 443 int
 444 seen_at_least_1_file (void)
 445 {
 446   return (physical_input_file != NULL);
 447 }
 448
 449 void
 450 bump_line_counters (void)
 451 {
 452   if (sb_index == (size_t) -1 || from_sb_expansion == expanding_app)
 453     ++physical_input_line;
 454
 455   if (logical_input_line != -1u)
 456     ++logical_input_line;
 457 }
 458 \f
 459 /* Tells us what the new logical line number and file are.
 460    If the line_number is -1, we don't change the current logical line
 461    number.
 462    If fname is NULL, we don't change the current logical file name, unless
 463    bit 3 of flags is set.
 464    Returns nonzero if the filename actually changes.  */
 465
 466 void
 467 new_logical_line_flags (const char *fname, /* DON'T destroy it!  We point to it!  */
 468                         int line_number,
 469                         int flags)
 470 {
 471   switch (flags)
 472     {
 473     case 0:
 474       break;
 475     case 1:
 476       if (line_number != -1)
 477         abort ();
 478       break;
 479     case 1 << 1:
 480     case 1 << 2:
 481       /* FIXME: we could check that include nesting is correct.  */
 482       break;
 483     case 1 << 3:
 484       if (line_number < 0 || fname != NULL)
 485         abort ();
 486       if (next_saved_file == NULL)
 487         fname = physical_input_file;
 488       else if (next_saved_file->logical_input_file)
 489         fname = next_saved_file->logical_input_file;
 490       else
 491         fname = next_saved_file->physical_input_file;
 492       break;
 493     default:
 494       abort ();
 495     }
 496
 497   is_linefile = flags != 1 && (flags != 0 || fname);
 498
 499   if (line_number >= 0)
 500     logical_input_line = line_number;
 501   else if (line_number == -1 && fname && !*fname && (flags & (1 << 2)))
 502     {
 503       logical_input_file = physical_input_file;
 504       logical_input_line = physical_input_line;
 505       fname = NULL;
 506     }
 507
 508   if (fname
 509       && (logical_input_file == NULL
 510           || filename_cmp (logical_input_file, fname)))
 511     logical_input_file = fname;
 512 }
 513
 514 void
 515 new_logical_line (const char *fname, int line_number)
 516 {
 517   new_logical_line_flags (fname, line_number, 0);
 518 }
 519
 520 void
 521 as_report_context (void)
 522 {
 523   const struct input_save *saved = next_saved_file;
 524   enum expansion expansion = from_sb_expansion;
 525   int indent = 1;
 526
 527   if (!macro_nest)
 528     return;
 529
 530   do
 531     {
 532       if (expansion != expanding_macro)
 533         /* Nothing.  */;
 534       else if (saved->logical_input_file != NULL
 535                && saved->logical_input_line != -1u)
 536         as_info_where (saved->logical_input_file, saved->logical_input_line,
 537                        indent, _("macro invoked from here"));
 538       else
 539         as_info_where (saved->physical_input_file, saved->physical_input_line,
 540                        indent, _("macro invoked from here"));
 541
 542       expansion = saved->from_sb_expansion;
 543       ++indent;
 544     }
 545   while ((saved = saved->next_saved_file) != NULL);
 546 }
 547 \f
 548 /* Return the current physical input file name and line number, if known  */
 549
 550 const char *
 551 as_where_physical (unsigned int *linep)
 552 {
 553   if (physical_input_file != NULL)
 554     {
 555       if (linep != NULL)
 556         *linep = physical_input_line;
 557       return physical_input_file;
 558     }
 559
 560   if (linep != NULL)
 561     *linep = 0;
 562   return NULL;
 563 }
 564
 565 /* Return the file name and line number at the top most macro
 566    invocation, unless .file / .line were used inside a macro.  */
 567
 568 const char *
 569 as_where (unsigned int *linep)
 570 {
 571   const char *file = as_where_top (linep);
 572
 573   if (macro_nest && is_linefile)
 574     {
 575       const struct input_save *saved = next_saved_file;
 576       enum expansion expansion = from_sb_expansion;
 577
 578       do
 579         {
 580           if (expansion != expanding_macro)
 581             /* Nothing.  */;
 582           else if (saved->logical_input_file != NULL
 583                    && (linep == NULL || saved->logical_input_line != -1u))
 584             {
 585               if (linep != NULL)
 586                 *linep = saved->logical_input_line;
 587               file = saved->logical_input_file;
 588             }
 589           else if (saved->physical_input_file != NULL)
 590             {
 591               if (linep != NULL)
 592                 *linep = saved->physical_input_line;
 593               file = saved->physical_input_file;
 594             }
 595
 596           expansion = saved->from_sb_expansion;
 597         }
 598       while ((saved = saved->next_saved_file) != NULL);
 599     }
 600
 601   return file;
 602 }
 603
 604 /* Return the current file name and line number.  */
 605
 606 const char *
 607 as_where_top (unsigned int *linep)
 608 {
 609   if (logical_input_file != NULL
 610       && (linep == NULL || logical_input_line != -1u))
 611     {
 612       if (linep != NULL)
 613         *linep = logical_input_line;
 614       return logical_input_file;
 615     }
 616
 617   return as_where_physical (linep);
 618 }