gas/input-scrub.c

   1 /* input_scrub.c - Break up input buffers into whole numbers of lines.
   2    Copyright (C) 1987-2022 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 /* Struct used to save the state of the input handler during include files */
 108 struct input_save {
 109   char *              buffer_start;
 110   char *              partial_where;
 111   size_t              partial_size;
 112   char                save_source[AFTER_SIZE];
 113   size_t              buffer_length;
 114   const char *        physical_input_file;
 115   const char *        logical_input_file;
 116   unsigned int        physical_input_line;
 117   unsigned int        logical_input_line;
 118   size_t              sb_index;
 119   sb                  from_sb;
 120   enum expansion      from_sb_expansion; /* Should we do a conditional check?  */
 121   struct input_save * next_saved_file;  /* Chain of input_saves.  */
 122   char *              input_file_save;  /* Saved state of input routines.  */
 123   char *              saved_position;   /* Caller's saved position in buf.  */
 124 };
 125
 126 static struct input_save *input_scrub_push (char *saved_position);
 127 static char *input_scrub_pop (struct input_save *arg);
 128
 129 /* Saved information about the file that .include'd this one.  When we hit EOF,
 130    we automatically pop to that file.  */
 131
 132 static struct input_save *next_saved_file;
 133
 134 /* Initialize input buffering.  */
 135
 136 static void
 137 input_scrub_reinit (void)
 138 {
 139   input_file_begin ();          /* Reinitialize! */
 140   logical_input_line = -1u;
 141   logical_input_file = NULL;
 142
 143   buffer_length = input_file_buffer_size () * 2;
 144   buffer_start = XNEWVEC (char, BEFORE_SIZE + AFTER_SIZE + 1 + buffer_length);
 145   memcpy (buffer_start, BEFORE_STRING, (int) BEFORE_SIZE);
 146 }
 147
 148 /* Push the state of input reading and scrubbing so that we can #include.
 149    The return value is a 'void *' (fudged for old compilers) to a save
 150    area, which can be restored by passing it to input_scrub_pop().  */
 151
 152 static struct input_save *
 153 input_scrub_push (char *saved_position)
 154 {
 155   struct input_save *saved;
 156
 157   saved = XNEW (struct input_save);
 158
 159   saved->saved_position = saved_position;
 160   saved->buffer_start = buffer_start;
 161   saved->partial_where = partial_where;
 162   saved->partial_size = partial_size;
 163   saved->buffer_length = buffer_length;
 164   saved->physical_input_file = physical_input_file;
 165   saved->logical_input_file = logical_input_file;
 166   saved->physical_input_line = physical_input_line;
 167   saved->logical_input_line = logical_input_line;
 168   saved->sb_index = sb_index;
 169   saved->from_sb = from_sb;
 170   saved->from_sb_expansion = from_sb_expansion;
 171   memcpy (saved->save_source, save_source, sizeof (save_source));
 172   saved->next_saved_file = next_saved_file;
 173   saved->input_file_save = input_file_push ();
 174
 175   sb_index = -1;
 176
 177   input_scrub_reinit ();
 178
 179   return saved;
 180 }
 181
 182 static char *
 183 input_scrub_pop (struct input_save *saved)
 184 {
 185   char *saved_position;
 186
 187   input_scrub_end ();           /* Finish off old buffer */
 188
 189   input_file_pop (saved->input_file_save);
 190   saved_position = saved->saved_position;
 191   buffer_start = saved->buffer_start;
 192   buffer_length = saved->buffer_length;
 193   physical_input_file = saved->physical_input_file;
 194   logical_input_file = saved->logical_input_file;
 195   physical_input_line = saved->physical_input_line;
 196   logical_input_line = saved->logical_input_line;
 197   sb_index = saved->sb_index;
 198   from_sb = saved->from_sb;
 199   from_sb_expansion = saved->from_sb_expansion;
 200   partial_where = saved->partial_where;
 201   partial_size = saved->partial_size;
 202   next_saved_file = saved->next_saved_file;
 203   memcpy (save_source, saved->save_source, sizeof (save_source));
 204
 205   free (saved);
 206   return saved_position;
 207 }
 208 \f
 209 void
 210 input_scrub_begin (void)
 211 {
 212   know (strlen (BEFORE_STRING) == BEFORE_SIZE);
 213   know (strlen (AFTER_STRING) == AFTER_SIZE
 214         || (AFTER_STRING[0] == '\0' && AFTER_SIZE == 1));
 215
 216   physical_input_file = NULL;   /* No file read yet.  */
 217   next_saved_file = NULL;       /* At EOF, don't pop to any other file */
 218   input_scrub_reinit ();
 219   do_scrub_begin (flag_m68k_mri);
 220 }
 221
 222 void
 223 input_scrub_end (void)
 224 {
 225   if (buffer_start)
 226     {
 227       free (buffer_start);
 228       buffer_start = 0;
 229       input_file_end ();
 230     }
 231 }
 232
 233 /* Start reading input from a new file.
 234    Return start of caller's part of buffer.  */
 235
 236 char *
 237 input_scrub_new_file (const char *filename)
 238 {
 239   input_file_open (filename, !flag_no_comments);
 240   physical_input_file = filename[0] ? filename : _("{standard input}");
 241   physical_input_line = 0;
 242
 243   partial_size = 0;
 244   return (buffer_start + BEFORE_SIZE);
 245 }
 246
 247 /* Include a file from the current file.  Save our state, cause it to
 248    be restored on EOF, and begin handling a new file.  Same result as
 249    input_scrub_new_file.  */
 250
 251 char *
 252 input_scrub_include_file (const char *filename, char *position)
 253 {
 254   next_saved_file = input_scrub_push (position);
 255   from_sb_expansion = expanding_none;
 256   return input_scrub_new_file (filename);
 257 }
 258
 259 /* Start getting input from an sb structure.  This is used when
 260    expanding a macro.  */
 261
 262 void
 263 input_scrub_include_sb (sb *from, char *position, enum expansion expansion)
 264 {
 265   int newline;
 266
 267   if (macro_nest > max_macro_nest)
 268     as_fatal (_("macros nested too deeply"));
 269   ++macro_nest;
 270
 271   gas_assert (expansion < expanding_nested);
 272
 273 #ifdef md_macro_start
 274   if (expansion == expanding_macro)
 275     {
 276       md_macro_start ();
 277     }
 278 #endif
 279
 280   next_saved_file = input_scrub_push (position);
 281
 282   /* Allocate sufficient space: from->len + optional newline.  */
 283   newline = from->len >= 1 && from->ptr[0] != '\n';
 284   sb_build (&from_sb, from->len + newline);
 285   if (expansion == expanding_repeat && from_sb_expansion >= expanding_macro)
 286     expansion = expanding_nested;
 287   from_sb_expansion = expansion;
 288   if (newline)
 289     {
 290       /* Add the sentinel required by read.c.  */
 291       sb_add_char (&from_sb, '\n');
 292     }
 293   sb_scrub_and_add_sb (&from_sb, from);
 294
 295   /* Make sure the parser looks at defined contents when it scans for
 296      e.g. end-of-line at the end of a macro.  */
 297   sb_terminate (&from_sb);
 298
 299   sb_index = 1;
 300
 301   /* These variables are reset by input_scrub_push.  Restore them
 302      since we are, after all, still at the same point in the file.  */
 303   logical_input_line = next_saved_file->logical_input_line;
 304   logical_input_file = next_saved_file->logical_input_file;
 305 }
 306
 307 void
 308 input_scrub_close (void)
 309 {
 310   input_file_close ();
 311   physical_input_line = 0;
 312   logical_input_line = -1u;
 313 }
 314
 315 char *
 316 input_scrub_next_buffer (char **bufp)
 317 {
 318   char *limit;          /*->just after last char of buffer.  */
 319
 320   if (sb_index != (size_t) -1)
 321     {
 322       if (sb_index >= from_sb.len)
 323         {
 324           sb_kill (&from_sb);
 325           if (from_sb_expansion == expanding_macro)
 326             {
 327               cond_finish_check (macro_nest);
 328 #ifdef md_macro_end
 329               /* Allow the target to clean up per-macro expansion
 330                  data.  */
 331               md_macro_end ();
 332 #endif
 333             }
 334           --macro_nest;
 335           partial_where = NULL;
 336           partial_size = 0;
 337           if (next_saved_file != NULL)
 338             *bufp = input_scrub_pop (next_saved_file);
 339           return partial_where;
 340         }
 341
 342       partial_where = from_sb.ptr + from_sb.len;
 343       partial_size = 0;
 344       *bufp = from_sb.ptr + sb_index;
 345       sb_index = from_sb.len;
 346       return partial_where;
 347     }
 348
 349   if (partial_size)
 350     {
 351       memmove (buffer_start + BEFORE_SIZE, partial_where, partial_size);
 352       memcpy (buffer_start + BEFORE_SIZE, save_source, AFTER_SIZE);
 353     }
 354
 355   while (1)
 356     {
 357       char *p;
 358       char *start = buffer_start + BEFORE_SIZE + partial_size;
 359
 360       *bufp = buffer_start + BEFORE_SIZE;
 361       limit = input_file_give_next_buffer (start);
 362       if (!limit)
 363         {
 364           if (!partial_size)
 365             /* End of this file.  */
 366             break;
 367
 368           as_warn (_("end of file not at end of a line; newline inserted"));
 369           p = buffer_start + BEFORE_SIZE + partial_size;
 370           *p++ = '\n';
 371           limit = p;
 372         }
 373       else
 374         {
 375           /* Terminate the buffer to avoid confusing TC_EOL_IN_INSN.  */
 376           *limit = '\0';
 377
 378           /* Find last newline.  */
 379           for (p = limit - 1; *p != '\n' || TC_EOL_IN_INSN (p); --p)
 380             if (p < start)
 381               goto read_more;
 382           ++p;
 383         }
 384
 385       if (multibyte_handling == multibyte_warn)
 386         (void) scan_for_multibyte_characters ((const unsigned char *) p,
 387                                               (const unsigned char *) limit,
 388                                               true /* Generate warnings */);
 389
 390       /* We found a newline in the newly read chars.  */
 391       partial_where = p;
 392       partial_size = limit - p;
 393
 394       /* Save the fragment after that last newline.  */
 395       memcpy (save_source, partial_where, (int) AFTER_SIZE);
 396       memcpy (partial_where, AFTER_STRING, (int) AFTER_SIZE);
 397       return partial_where;
 398
 399     read_more:
 400       /* Didn't find a newline.  Read more text.  */
 401       partial_size = limit - (buffer_start + BEFORE_SIZE);
 402       if (buffer_length - input_file_buffer_size () < partial_size)
 403         {
 404           /* Increase the buffer when it doesn't have room for the
 405              next block of input.  */
 406           buffer_length *= 2;
 407           buffer_start = XRESIZEVEC (char, buffer_start,
 408                                      (buffer_length
 409                                       + BEFORE_SIZE + AFTER_SIZE + 1));
 410         }
 411     }
 412
 413   /* Tell the listing we've finished the file.  */
 414   LISTING_EOF ();
 415
 416   /* If we should pop to another file at EOF, do it.  */
 417   partial_where = NULL;
 418   if (next_saved_file)
 419     *bufp = input_scrub_pop (next_saved_file);
 420
 421   return partial_where;
 422 }
 423 \f
 424 /* The remaining part of this file deals with line numbers, error
 425    messages and so on.  Return TRUE if we opened any file.  */
 426
 427 int
 428 seen_at_least_1_file (void)
 429 {
 430   return (physical_input_file != NULL);
 431 }
 432
 433 void
 434 bump_line_counters (void)
 435 {
 436   if (sb_index == (size_t) -1)
 437     ++physical_input_line;
 438
 439   /* PR gas/16908 workaround: Don't bump logical line numbers while
 440      expanding macros, unless file (and maybe line; see as_where()) are
 441      used inside the macro.  */
 442   if (logical_input_line != -1u && from_sb_expansion < expanding_macro)
 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 || next_saved_file == NULL)
 472         abort ();
 473       /* PR gas/16908 workaround: Ignore updates when nested inside a macro
 474          expansion.  */
 475       if (from_sb_expansion == expanding_nested)
 476         return;
 477       if (next_saved_file->logical_input_file)
 478         fname = next_saved_file->logical_input_file;
 479       else
 480         fname = next_saved_file->physical_input_file;
 481       break;
 482     default:
 483       abort ();
 484     }
 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   /* When encountering file or line changes inside a macro, arrange for
 501      bump_line_counters() to henceforth increment the logical line number
 502      again, just like it does when expanding repeats.  See as_where() for
 503      why changing file or line alone doesn't alter expansion mode.  */
 504   if (from_sb_expansion == expanding_macro
 505       && logical_input_file != NULL
 506       && logical_input_line != -1u)
 507     from_sb_expansion = expanding_repeat;
 508 }
 509
 510 void
 511 new_logical_line (const char *fname, int line_number)
 512 {
 513   new_logical_line_flags (fname, line_number, 0);
 514 }
 515
 516 \f
 517 /* Return the current physical input file name and line number, if known  */
 518
 519 const char *
 520 as_where_physical (unsigned int *linep)
 521 {
 522   if (physical_input_file != NULL)
 523     {
 524       if (linep != NULL)
 525         *linep = physical_input_line;
 526       return physical_input_file;
 527     }
 528
 529   if (linep != NULL)
 530     *linep = 0;
 531   return NULL;
 532 }
 533
 534 /* Return the current file name and line number.  */
 535
 536 const char *
 537 as_where (unsigned int *linep)
 538 {
 539   if (logical_input_file != NULL
 540       && (linep == NULL || logical_input_line != -1u))
 541     {
 542       if (linep != NULL)
 543         *linep = logical_input_line;
 544       return logical_input_file;
 545     }
 546
 547   return as_where_physical (linep);
 548 }
 549