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   sb_index = -1;
 143
 144   buffer_length = input_file_buffer_size () * 2;
 145   buffer_start = XNEWVEC (char, BEFORE_SIZE + AFTER_SIZE + 1 + buffer_length);
 146   memcpy (buffer_start, BEFORE_STRING, (int) BEFORE_SIZE);
 147 }
 148
 149 /* Push the state of input reading and scrubbing so that we can #include.
 150    The return value is a 'void *' (fudged for old compilers) to a save
 151    area, which can be restored by passing it to input_scrub_pop().  */
 152
 153 static struct input_save *
 154 input_scrub_push (char *saved_position)
 155 {
 156   struct input_save *saved;
 157
 158   saved = XNEW (struct input_save);
 159
 160   saved->saved_position = saved_position;
 161   saved->buffer_start = buffer_start;
 162   saved->partial_where = partial_where;
 163   saved->partial_size = partial_size;
 164   saved->buffer_length = buffer_length;
 165   saved->physical_input_file = physical_input_file;
 166   saved->logical_input_file = logical_input_file;
 167   saved->physical_input_line = physical_input_line;
 168   saved->logical_input_line = logical_input_line;
 169   saved->sb_index = sb_index;
 170   saved->from_sb = from_sb;
 171   saved->from_sb_expansion = from_sb_expansion;
 172   memcpy (saved->save_source, save_source, sizeof (save_source));
 173   saved->next_saved_file = next_saved_file;
 174   saved->input_file_save = input_file_push ();
 175
 176   input_scrub_reinit ();
 177
 178   return saved;
 179 }
 180
 181 static char *
 182 input_scrub_pop (struct input_save *saved)
 183 {
 184   char *saved_position;
 185
 186   input_scrub_end ();           /* Finish off old buffer */
 187
 188   input_file_pop (saved->input_file_save);
 189   saved_position = saved->saved_position;
 190   buffer_start = saved->buffer_start;
 191   buffer_length = saved->buffer_length;
 192   physical_input_file = saved->physical_input_file;
 193   logical_input_file = saved->logical_input_file;
 194   physical_input_line = saved->physical_input_line;
 195   logical_input_line = saved->logical_input_line;
 196   sb_index = saved->sb_index;
 197   from_sb = saved->from_sb;
 198   from_sb_expansion = saved->from_sb_expansion;
 199   partial_where = saved->partial_where;
 200   partial_size = saved->partial_size;
 201   next_saved_file = saved->next_saved_file;
 202   memcpy (save_source, saved->save_source, sizeof (save_source));
 203
 204   free (saved);
 205   return saved_position;
 206 }
 207 \f
 208 void
 209 input_scrub_begin (void)
 210 {
 211   know (strlen (BEFORE_STRING) == BEFORE_SIZE);
 212   know (strlen (AFTER_STRING) == AFTER_SIZE
 213         || (AFTER_STRING[0] == '\0' && AFTER_SIZE == 1));
 214
 215   physical_input_file = NULL;   /* No file read yet.  */
 216   next_saved_file = NULL;       /* At EOF, don't pop to any other file */
 217   input_scrub_reinit ();
 218   do_scrub_begin (flag_m68k_mri);
 219 }
 220
 221 void
 222 input_scrub_end (void)
 223 {
 224   if (buffer_start)
 225     {
 226       free (buffer_start);
 227       buffer_start = 0;
 228       input_file_end ();
 229     }
 230 }
 231
 232 /* Start reading input from a new file.
 233    Return start of caller's part of buffer.  */
 234
 235 char *
 236 input_scrub_new_file (const char *filename)
 237 {
 238   input_file_open (filename, !flag_no_comments);
 239   physical_input_file = filename[0] ? filename : _("{standard input}");
 240   physical_input_line = 0;
 241
 242   partial_size = 0;
 243   return (buffer_start + BEFORE_SIZE);
 244 }
 245
 246 /* Include a file from the current file.  Save our state, cause it to
 247    be restored on EOF, and begin handling a new file.  Same result as
 248    input_scrub_new_file.  */
 249
 250 char *
 251 input_scrub_include_file (const char *filename, char *position)
 252 {
 253   next_saved_file = input_scrub_push (position);
 254   from_sb_expansion = expanding_none;
 255   return input_scrub_new_file (filename);
 256 }
 257
 258 /* Start getting input from an sb structure.  This is used when
 259    expanding a macro.  */
 260
 261 void
 262 input_scrub_include_sb (sb *from, char *position, enum expansion expansion)
 263 {
 264   int newline;
 265
 266   if (macro_nest > max_macro_nest)
 267     as_fatal (_("macros nested too deeply"));
 268   ++macro_nest;
 269
 270   gas_assert (expansion < expanding_nested);
 271
 272 #ifdef md_macro_start
 273   if (expansion == expanding_macro)
 274     {
 275       md_macro_start ();
 276     }
 277 #endif
 278
 279   next_saved_file = input_scrub_push (position);
 280
 281   /* Allocate sufficient space: from->len plus optional newline
 282      plus two ".linefile " directives, plus a little more for other
 283      expansion.  */
 284   newline = from->len >= 1 && from->ptr[0] != '\n';
 285   sb_build (&from_sb, from->len + newline + 2 * sizeof (".linefile") + 30);
 286   if (expansion == expanding_repeat && from_sb_expansion >= expanding_macro)
 287     expansion = expanding_nested;
 288   from_sb_expansion = expansion;
 289   if (newline)
 290     {
 291       /* Add the sentinel required by read.c.  */
 292       sb_add_char (&from_sb, '\n');
 293     }
 294   sb_scrub_and_add_sb (&from_sb, from);
 295
 296   /* Make sure the parser looks at defined contents when it scans for
 297      e.g. end-of-line at the end of a macro.  */
 298   sb_terminate (&from_sb);
 299
 300   sb_index = 1;
 301
 302   /* These variables are reset by input_scrub_push.  Restore them
 303      since we are, after all, still at the same point in the file.  */
 304   logical_input_line = next_saved_file->logical_input_line;
 305   logical_input_file = next_saved_file->logical_input_file;
 306 }
 307
 308 void
 309 input_scrub_close (void)
 310 {
 311   input_file_close ();
 312   physical_input_line = 0;
 313   logical_input_line = -1u;
 314 }
 315
 316 char *
 317 input_scrub_next_buffer (char **bufp)
 318 {
 319   char *limit;          /*->just after last char of buffer.  */
 320
 321   if (sb_index != (size_t) -1)
 322     {
 323       if (sb_index >= from_sb.len)
 324         {
 325           sb_kill (&from_sb);
 326           if (from_sb_expansion == expanding_macro)
 327             {
 328               cond_finish_check (macro_nest);
 329 #ifdef md_macro_end
 330               /* Allow the target to clean up per-macro expansion
 331                  data.  */
 332               md_macro_end ();
 333 #endif
 334             }
 335           --macro_nest;
 336           partial_where = NULL;
 337           partial_size = 0;
 338           if (next_saved_file != NULL)
 339             *bufp = input_scrub_pop (next_saved_file);
 340           return partial_where;
 341         }
 342
 343       partial_where = from_sb.ptr + from_sb.len;
 344       partial_size = 0;
 345       *bufp = from_sb.ptr + sb_index;
 346       sb_index = from_sb.len;
 347       return partial_where;
 348     }
 349
 350   if (partial_size)
 351     {
 352       memmove (buffer_start + BEFORE_SIZE, partial_where, partial_size);
 353       memcpy (buffer_start + BEFORE_SIZE, save_source, AFTER_SIZE);
 354     }
 355
 356   while (1)
 357     {
 358       char *p;
 359       char *start = buffer_start + BEFORE_SIZE + partial_size;
 360
 361       *bufp = buffer_start + BEFORE_SIZE;
 362       limit = input_file_give_next_buffer (start);
 363       if (!limit)
 364         {
 365           if (!partial_size)
 366             /* End of this file.  */
 367             break;
 368
 369           as_warn (_("end of file not at end of a line; newline inserted"));
 370           p = buffer_start + BEFORE_SIZE + partial_size;
 371           *p++ = '\n';
 372           limit = p;
 373         }
 374       else
 375         {
 376           /* Terminate the buffer to avoid confusing TC_EOL_IN_INSN.  */
 377           *limit = '\0';
 378
 379           /* Find last newline.  */
 380           for (p = limit - 1; *p != '\n' || TC_EOL_IN_INSN (p); --p)
 381             if (p < start)
 382               goto read_more;
 383           ++p;
 384         }
 385
 386       if (multibyte_handling == multibyte_warn)
 387         (void) scan_for_multibyte_characters ((const unsigned char *) p,
 388                                               (const unsigned char *) limit,
 389                                               true /* Generate warnings */);
 390
 391       /* We found a newline in the newly read chars.  */
 392       partial_where = p;
 393       partial_size = limit - p;
 394
 395       /* Save the fragment after that last newline.  */
 396       memcpy (save_source, partial_where, (int) AFTER_SIZE);
 397       memcpy (partial_where, AFTER_STRING, (int) AFTER_SIZE);
 398       return partial_where;
 399
 400     read_more:
 401       /* Didn't find a newline.  Read more text.  */
 402       partial_size = limit - (buffer_start + BEFORE_SIZE);
 403       if (buffer_length - input_file_buffer_size () < partial_size)
 404         {
 405           /* Increase the buffer when it doesn't have room for the
 406              next block of input.  */
 407           buffer_length *= 2;
 408           buffer_start = XRESIZEVEC (char, buffer_start,
 409                                      (buffer_length
 410                                       + BEFORE_SIZE + AFTER_SIZE + 1));
 411         }
 412     }
 413
 414   /* Tell the listing we've finished the file.  */
 415   LISTING_EOF ();
 416
 417   /* If we should pop to another file at EOF, do it.  */
 418   partial_where = NULL;
 419   if (next_saved_file)
 420     *bufp = input_scrub_pop (next_saved_file);
 421
 422   return partial_where;
 423 }
 424 \f
 425 /* The remaining part of this file deals with line numbers, error
 426    messages and so on.  Return TRUE if we opened any file.  */
 427
 428 int
 429 seen_at_least_1_file (void)
 430 {
 431   return (physical_input_file != NULL);
 432 }
 433
 434 void
 435 bump_line_counters (void)
 436 {
 437   if (sb_index == (size_t) -1)
 438     ++physical_input_line;
 439
 440   /* PR gas/16908 workaround: Don't bump logical line numbers while
 441      expanding macros, unless file (and maybe line; see as_where()) are
 442      used inside the macro.  */
 443   if (logical_input_line != -1u && from_sb_expansion < expanding_macro)
 444     ++logical_input_line;
 445 }
 446 \f
 447 /* Tells us what the new logical line number and file are.
 448    If the line_number is -1, we don't change the current logical line
 449    number.
 450    If fname is NULL, we don't change the current logical file name, unless
 451    bit 3 of flags is set.
 452    Returns nonzero if the filename actually changes.  */
 453
 454 void
 455 new_logical_line_flags (const char *fname, /* DON'T destroy it!  We point to it!  */
 456                         int line_number,
 457                         int flags)
 458 {
 459   switch (flags)
 460     {
 461     case 0:
 462       break;
 463     case 1:
 464       if (line_number != -1)
 465         abort ();
 466       break;
 467     case 1 << 1:
 468     case 1 << 2:
 469       /* FIXME: we could check that include nesting is correct.  */
 470       break;
 471     case 1 << 3:
 472       if (line_number < 0 || fname != NULL)
 473         abort ();
 474       /* PR gas/16908 workaround: Ignore updates when nested inside a macro
 475          expansion.  */
 476       if (from_sb_expansion == expanding_nested)
 477         return;
 478       if (next_saved_file == NULL)
 479         fname = physical_input_file;
 480       else if (next_saved_file->logical_input_file)
 481         fname = next_saved_file->logical_input_file;
 482       else
 483         fname = next_saved_file->physical_input_file;
 484       break;
 485     default:
 486       abort ();
 487     }
 488
 489   if (line_number >= 0)
 490     logical_input_line = line_number;
 491   else if (line_number == -1 && fname && !*fname && (flags & (1 << 2)))
 492     {
 493       logical_input_file = physical_input_file;
 494       logical_input_line = physical_input_line;
 495       fname = NULL;
 496     }
 497
 498   if (fname
 499       && (logical_input_file == NULL
 500           || filename_cmp (logical_input_file, fname)))
 501     logical_input_file = fname;
 502
 503   /* When encountering file or line changes inside a macro, arrange for
 504      bump_line_counters() to henceforth increment the logical line number
 505      again, just like it does when expanding repeats.  See as_where() for
 506      why changing file or line alone doesn't alter expansion mode.  */
 507   if (from_sb_expansion == expanding_macro
 508       && logical_input_file != NULL
 509       && logical_input_line != -1u)
 510     from_sb_expansion = expanding_repeat;
 511 }
 512
 513 void
 514 new_logical_line (const char *fname, int line_number)
 515 {
 516   new_logical_line_flags (fname, line_number, 0);
 517 }
 518
 519 \f
 520 /* Return the current physical input file name and line number, if known  */
 521
 522 const char *
 523 as_where_physical (unsigned int *linep)
 524 {
 525   if (physical_input_file != NULL)
 526     {
 527       if (linep != NULL)
 528         *linep = physical_input_line;
 529       return physical_input_file;
 530     }
 531
 532   if (linep != NULL)
 533     *linep = 0;
 534   return NULL;
 535 }
 536
 537 /* Return the current file name and line number.  */
 538
 539 const char *
 540 as_where (unsigned int *linep)
 541 {
 542   if (logical_input_file != NULL
 543       && (linep == NULL || logical_input_line != -1u))
 544     {
 545       if (linep != NULL)
 546         *linep = logical_input_line;
 547       return logical_input_file;
 548     }
 549
 550   return as_where_physical (linep);
 551 }
 552