gas/input-scrub.c

   1 /* input_scrub.c - Break up input buffers into whole numbers of lines.
   2    Copyright 1987, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
   3    2000, 2001, 2003, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012
   4    Free Software Foundation, Inc.
   5
   6    This file is part of GAS, the GNU Assembler.
   7
   8    GAS is free software; you can redistribute it and/or modify
   9    it under the terms of the GNU General Public License as published by
  10    the Free Software Foundation; either version 3, or (at your option)
  11    any later version.
  12
  13    GAS is distributed in the hope that it will be useful,
  14    but WITHOUT ANY WARRANTY; without even the implied warranty of
  15    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16    GNU General Public License for more details.
  17
  18    You should have received a copy of the GNU General Public License
  19    along with GAS; see the file COPYING.  If not, write to the Free
  20    Software Foundation, 51 Franklin Street - Fifth Floor, Boston, MA
  21    02110-1301, USA.  */
  22
  23 #include "as.h"
  24 #include "filenames.h"
  25 #include "input-file.h"
  26 #include "sb.h"
  27 #include "listing.h"
  28
  29 /*
  30  * O/S independent module to supply buffers of sanitised source code
  31  * to rest of assembler.  We get sanitised input data of arbitrary length.
  32  * We break these buffers on line boundaries, recombine pieces that
  33  * were broken across buffers, and return a buffer of full lines to
  34  * the caller.
  35  * The last partial line begins the next buffer we build and return to caller.
  36  * The buffer returned to caller is preceded by BEFORE_STRING and followed
  37  * by AFTER_STRING, as sentinels. The last character before AFTER_STRING
  38  * is a newline.
  39  * Also looks after line numbers, for e.g. error messages.
  40  */
  41
  42 /*
  43  * We don't care how filthy our buffers are, but our callers assume
  44  * that the following sanitation has already been done.
  45  *
  46  * No comments, reduce a comment to a space.
  47  * Reduce a tab to a space unless it is 1st char of line.
  48  * All multiple tabs and spaces collapsed into 1 char. Tab only
  49  *   legal if 1st char of line.
  50  * # line file statements converted to .line x;.file y; statements.
  51  * Escaped newlines at end of line: remove them but add as many newlines
  52  *   to end of statement as you removed in the middle, to synch line numbers.
  53  */
  54 \f
  55 #define BEFORE_STRING ("\n")
  56 #define AFTER_STRING ("\0")     /* memcpy of 0 chars might choke.  */
  57 #define BEFORE_SIZE (1)
  58 #define AFTER_SIZE  (1)
  59
  60 #ifndef TC_EOL_IN_INSN
  61 #define TC_EOL_IN_INSN(P) 0
  62 #endif
  63
  64 static char *buffer_start;      /*->1st char of full buffer area.  */
  65 static char *partial_where;     /*->after last full line in buffer.  */
  66 static int partial_size;        /* >=0. Number of chars in partial line in buffer.  */
  67
  68 /* Because we need AFTER_STRING just after last full line, it clobbers
  69    1st part of partial line. So we preserve 1st part of partial line
  70    here.  */
  71 static char save_source[AFTER_SIZE];
  72
  73 /* What is the largest size buffer that input_file_give_next_buffer()
  74    could return to us?  */
  75 static unsigned int buffer_length;
  76
  77 /* The index into an sb structure we are reading from.  -1 if none.  */
  78 static size_t sb_index = -1;
  79
  80 /* If we are reading from an sb structure, this is it.  */
  81 static sb from_sb;
  82
  83 /* Should we do a conditional check on from_sb? */
  84 static int from_sb_is_expansion = 1;
  85
  86 /* The number of nested sb structures we have included.  */
  87 int macro_nest;
  88
  89 /* We can have more than one source file open at once, though the info for all
  90    but the latest one are saved off in a struct input_save.  These files remain
  91    open, so we are limited by the number of open files allowed by the
  92    underlying OS. We may also sequentially read more than one source file in an
  93    assembly.  */
  94
  95 /* We must track the physical file and line number for error messages. We also
  96    track a "logical" file and line number corresponding to (C?)  compiler
  97    source line numbers.  Whenever we open a file we must fill in
  98    physical_input_file. So if it is NULL we have not opened any files yet.  */
  99
 100 static char *physical_input_file;
 101 static char *logical_input_file;
 102
 103 /* 1-origin line number in a source file.  */
 104 /* A line ends in '\n' or eof.  */
 105 static unsigned int physical_input_line;
 106 static int logical_input_line;
 107
 108 /* Struct used to save the state of the input handler during include files */
 109 struct input_save {
 110   char *              buffer_start;
 111   char *              partial_where;
 112   int                 partial_size;
 113   char                save_source[AFTER_SIZE];
 114   size_t              buffer_length;
 115   char *              physical_input_file;
 116   char *              logical_input_file;
 117   unsigned int        physical_input_line;
 118   int                 logical_input_line;
 119   size_t              sb_index;
 120   sb                  from_sb;
 121   int                 from_sb_is_expansion; /* Should we do a conditional check?  */
 122   struct input_save * next_saved_file;  /* Chain of input_saves.  */
 123   char *              input_file_save;  /* Saved state of input routines.  */
 124   char *              saved_position;   /* Caller's saved position in buf.  */
 125 };
 126
 127 static struct input_save *input_scrub_push (char *saved_position);
 128 static char *input_scrub_pop (struct input_save *arg);
 129
 130 /* Saved information about the file that .include'd this one.  When we hit EOF,
 131    we automatically pop to that file.  */
 132
 133 static struct input_save *next_saved_file;
 134
 135 /* Push the state of input reading and scrubbing so that we can #include.
 136    The return value is a 'void *' (fudged for old compilers) to a save
 137    area, which can be restored by passing it to input_scrub_pop().  */
 138
 139 static struct input_save *
 140 input_scrub_push (char *saved_position)
 141 {
 142   register struct input_save *saved;
 143
 144   saved = (struct input_save *) xmalloc (sizeof *saved);
 145
 146   saved->saved_position = saved_position;
 147   saved->buffer_start = buffer_start;
 148   saved->partial_where = partial_where;
 149   saved->partial_size = partial_size;
 150   saved->buffer_length = buffer_length;
 151   saved->physical_input_file = physical_input_file;
 152   saved->logical_input_file = logical_input_file;
 153   saved->physical_input_line = physical_input_line;
 154   saved->logical_input_line = logical_input_line;
 155   saved->sb_index = sb_index;
 156   saved->from_sb = from_sb;
 157   saved->from_sb_is_expansion = from_sb_is_expansion;
 158   memcpy (saved->save_source, save_source, sizeof (save_source));
 159   saved->next_saved_file = next_saved_file;
 160   saved->input_file_save = input_file_push ();
 161
 162   input_file_begin ();          /* Reinitialize! */
 163   logical_input_line = -1;
 164   logical_input_file = (char *) NULL;
 165   buffer_length = input_file_buffer_size ();
 166   sb_index = -1;
 167
 168   buffer_start = (char *) xmalloc ((BEFORE_SIZE + buffer_length
 169                                     + buffer_length + AFTER_SIZE));
 170   memcpy (buffer_start, BEFORE_STRING, (int) BEFORE_SIZE);
 171
 172   return saved;
 173 }
 174
 175 static char *
 176 input_scrub_pop (struct input_save *saved)
 177 {
 178   char *saved_position;
 179
 180   input_scrub_end ();           /* Finish off old buffer */
 181
 182   input_file_pop (saved->input_file_save);
 183   saved_position = saved->saved_position;
 184   buffer_start = saved->buffer_start;
 185   buffer_length = saved->buffer_length;
 186   physical_input_file = saved->physical_input_file;
 187   logical_input_file = saved->logical_input_file;
 188   physical_input_line = saved->physical_input_line;
 189   logical_input_line = saved->logical_input_line;
 190   sb_index = saved->sb_index;
 191   from_sb = saved->from_sb;
 192   from_sb_is_expansion = saved->from_sb_is_expansion;
 193   partial_where = saved->partial_where;
 194   partial_size = saved->partial_size;
 195   next_saved_file = saved->next_saved_file;
 196   memcpy (save_source, saved->save_source, sizeof (save_source));
 197
 198   free (saved);
 199   return saved_position;
 200 }
 201 \f
 202 void
 203 input_scrub_begin (void)
 204 {
 205   know (strlen (BEFORE_STRING) == BEFORE_SIZE);
 206   know (strlen (AFTER_STRING) == AFTER_SIZE
 207         || (AFTER_STRING[0] == '\0' && AFTER_SIZE == 1));
 208
 209   input_file_begin ();
 210
 211   buffer_length = input_file_buffer_size ();
 212
 213   buffer_start = (char *) xmalloc ((BEFORE_SIZE + buffer_length
 214                                     + buffer_length + AFTER_SIZE));
 215   memcpy (buffer_start, BEFORE_STRING, (int) BEFORE_SIZE);
 216
 217   /* Line number things.  */
 218   logical_input_line = -1;
 219   logical_input_file = (char *) NULL;
 220   physical_input_file = NULL;   /* No file read yet.  */
 221   next_saved_file = NULL;       /* At EOF, don't pop to any other file */
 222   do_scrub_begin (flag_m68k_mri);
 223 }
 224
 225 void
 226 input_scrub_end (void)
 227 {
 228   if (buffer_start)
 229     {
 230       free (buffer_start);
 231       buffer_start = 0;
 232       input_file_end ();
 233     }
 234 }
 235
 236 /* Start reading input from a new file.
 237    Return start of caller's part of buffer.  */
 238
 239 char *
 240 input_scrub_new_file (char *filename)
 241 {
 242   input_file_open (filename, !flag_no_comments);
 243   physical_input_file = filename[0] ? filename : _("{standard input}");
 244   physical_input_line = 0;
 245
 246   partial_size = 0;
 247   return (buffer_start + BEFORE_SIZE);
 248 }
 249
 250 /* Include a file from the current file.  Save our state, cause it to
 251    be restored on EOF, and begin handling a new file.  Same result as
 252    input_scrub_new_file.  */
 253
 254 char *
 255 input_scrub_include_file (char *filename, char *position)
 256 {
 257   next_saved_file = input_scrub_push (position);
 258   return input_scrub_new_file (filename);
 259 }
 260
 261 /* Start getting input from an sb structure.  This is used when
 262    expanding a macro.  */
 263
 264 void
 265 input_scrub_include_sb (sb *from, char *position, int is_expansion)
 266 {
 267   int newline;
 268
 269   if (macro_nest > max_macro_nest)
 270     as_fatal (_("macros nested too deeply"));
 271   ++macro_nest;
 272
 273 #ifdef md_macro_start
 274   if (is_expansion)
 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   from_sb_is_expansion = is_expansion;
 286   if (newline)
 287     {
 288       /* Add the sentinel required by read.c.  */
 289       sb_add_char (&from_sb, '\n');
 290     }
 291   sb_scrub_and_add_sb (&from_sb, from);
 292
 293   /* Make sure the parser looks at defined contents when it scans for
 294      e.g. end-of-line at the end of a macro.  */
 295   sb_terminate (&from_sb);
 296
 297   sb_index = 1;
 298
 299   /* These variables are reset by input_scrub_push.  Restore them
 300      since we are, after all, still at the same point in the file.  */
 301   logical_input_line = next_saved_file->logical_input_line;
 302   logical_input_file = next_saved_file->logical_input_file;
 303 }
 304
 305 void
 306 input_scrub_close (void)
 307 {
 308   input_file_close ();
 309   physical_input_line = 0;
 310   logical_input_line = -1;
 311 }
 312
 313 char *
 314 input_scrub_next_buffer (char **bufp)
 315 {
 316   register char *limit;         /*->just after last char of buffer.  */
 317
 318   if (sb_index != (size_t) -1)
 319     {
 320       if (sb_index >= from_sb.len)
 321         {
 322           sb_kill (&from_sb);
 323           if (from_sb_is_expansion)
 324             {
 325               cond_finish_check (macro_nest);
 326 #ifdef md_macro_end
 327               /* Allow the target to clean up per-macro expansion
 328                  data.  */
 329               md_macro_end ();
 330 #endif
 331             }
 332           --macro_nest;
 333           partial_where = NULL;
 334           if (next_saved_file != NULL)
 335             *bufp = input_scrub_pop (next_saved_file);
 336           return partial_where;
 337         }
 338
 339       partial_where = from_sb.ptr + from_sb.len;
 340       partial_size = 0;
 341       *bufp = from_sb.ptr + sb_index;
 342       sb_index = from_sb.len;
 343       return partial_where;
 344     }
 345
 346   *bufp = buffer_start + BEFORE_SIZE;
 347
 348   if (partial_size)
 349     {
 350       memmove (buffer_start + BEFORE_SIZE, partial_where,
 351                (unsigned int) partial_size);
 352       memcpy (buffer_start + BEFORE_SIZE, save_source, AFTER_SIZE);
 353     }
 354   limit = input_file_give_next_buffer (buffer_start
 355                                        + BEFORE_SIZE
 356                                        + partial_size);
 357   if (limit)
 358     {
 359       register char *p;         /* Find last newline.  */
 360       /* Terminate the buffer to avoid confusing TC_EOL_IN_INSN.  */
 361       *limit = '\0';
 362       for (p = limit - 1; *p != '\n' || TC_EOL_IN_INSN (p); --p)
 363         ;
 364       ++p;
 365
 366       while (p <= buffer_start + BEFORE_SIZE)
 367         {
 368           int limoff;
 369
 370           limoff = limit - buffer_start;
 371           buffer_length += input_file_buffer_size ();
 372           buffer_start = (char *) xrealloc (buffer_start,
 373                                             (BEFORE_SIZE
 374                                              + 2 * buffer_length
 375                                              + AFTER_SIZE));
 376           *bufp = buffer_start + BEFORE_SIZE;
 377           limit = input_file_give_next_buffer (buffer_start + limoff);
 378
 379           if (limit == NULL)
 380             {
 381               as_warn (_("partial line at end of file ignored"));
 382               partial_where = NULL;
 383               if (next_saved_file)
 384                 *bufp = input_scrub_pop (next_saved_file);
 385               return NULL;
 386             }
 387
 388           /* Terminate the buffer to avoid confusing TC_EOL_IN_INSN.  */
 389           *limit = '\0';
 390           for (p = limit - 1; *p != '\n' || TC_EOL_IN_INSN (p); --p)
 391             ;
 392           ++p;
 393         }
 394
 395       partial_where = p;
 396       partial_size = limit - p;
 397       memcpy (save_source, partial_where, (int) AFTER_SIZE);
 398       memcpy (partial_where, AFTER_STRING, (int) AFTER_SIZE);
 399     }
 400   else
 401     {
 402       partial_where = 0;
 403       if (partial_size > 0)
 404         {
 405           as_warn (_("partial line at end of file ignored"));
 406         }
 407
 408       /* Tell the listing we've finished the file.  */
 409       LISTING_EOF ();
 410
 411       /* If we should pop to another file at EOF, do it.  */
 412       if (next_saved_file)
 413         {
 414           *bufp = input_scrub_pop (next_saved_file);    /* Pop state */
 415           /* partial_where is now correct to return, since we popped it.  */
 416         }
 417     }
 418   return (partial_where);
 419 }
 420 \f
 421 /* The remaining part of this file deals with line numbers, error
 422    messages and so on.  Return TRUE if we opened any file.  */
 423
 424 int
 425 seen_at_least_1_file (void)
 426 {
 427   return (physical_input_file != NULL);
 428 }
 429
 430 void
 431 bump_line_counters (void)
 432 {
 433   if (sb_index == (size_t) -1)
 434     {
 435       ++physical_input_line;
 436       if (logical_input_line >= 0)
 437         ++logical_input_line;
 438     }
 439 }
 440 \f
 441 /* Tells us what the new logical line number and file are.
 442    If the line_number is -1, we don't change the current logical line
 443    number.  If it is -2, we decrement the logical line number (this is
 444    to support the .appfile pseudo-op inserted into the stream by
 445    do_scrub_chars).
 446    If the fname is NULL, we don't change the current logical file name.
 447    Returns nonzero if the filename actually changes.  */
 448
 449 int
 450 new_logical_line_flags (char *fname, /* DON'T destroy it!  We point to it!  */
 451                         int line_number,
 452                         int flags)
 453 {
 454   switch (flags)
 455     {
 456     case 0:
 457       break;
 458     case 1:
 459       if (line_number != -1)
 460         abort ();
 461       break;
 462     case 1 << 1:
 463     case 1 << 2:
 464       /* FIXME: we could check that include nesting is correct.  */
 465       break;
 466     default:
 467       abort ();
 468     }
 469
 470   if (line_number >= 0)
 471     logical_input_line = line_number;
 472   else if (line_number == -1 && fname && !*fname && (flags & (1 << 2)))
 473     {
 474       logical_input_file = physical_input_file;
 475       logical_input_line = physical_input_line;
 476       fname = NULL;
 477     }
 478
 479   if (fname
 480       && (logical_input_file == NULL
 481           || filename_cmp (logical_input_file, fname)))
 482     {
 483       logical_input_file = fname;
 484       return 1;
 485     }
 486   else
 487     return 0;
 488 }
 489
 490 int
 491 new_logical_line (char *fname, int line_number)
 492 {
 493   return new_logical_line_flags (fname, line_number, 0);
 494 }
 495
 496 \f
 497 /* Return the current file name and line number.
 498    namep should be char * const *, but there are compilers which screw
 499    up declarations like that, and it's easier to avoid it.  */
 500
 501 void
 502 as_where (char **namep, unsigned int *linep)
 503 {
 504   if (logical_input_file != NULL
 505       && (linep == NULL || logical_input_line >= 0))
 506     {
 507       *namep = logical_input_file;
 508       if (linep != NULL)
 509         *linep = logical_input_line;
 510     }
 511   else if (physical_input_file != NULL)
 512     {
 513       *namep = physical_input_file;
 514       if (linep != NULL)
 515         *linep = physical_input_line;
 516     }
 517   else
 518     {
 519       *namep = 0;
 520       if (linep != NULL)
 521         *linep = 0;
 522     }
 523 }