gas/input-scrub.c

   1 /* input_scrub.c - Break up input buffers into whole numbers of lines.
   2    Copyright (C) 1987-2016 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 int 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 /* What is the largest size buffer that input_file_give_next_buffer()
  72    could return to us?  */
  73 static unsigned int buffer_length;
  74
  75 /* The index into an sb structure we are reading from.  -1 if none.  */
  76 static size_t sb_index = -1;
  77
  78 /* If we are reading from an sb structure, this is it.  */
  79 static sb from_sb;
  80
  81 /* Should we do a conditional check on from_sb? */
  82 static int from_sb_is_expansion = 1;
  83
  84 /* The number of nested sb structures we have included.  */
  85 int macro_nest;
  86
  87 /* We can have more than one source file open at once, though the info for all
  88    but the latest one are saved off in a struct input_save.  These files remain
  89    open, so we are limited by the number of open files allowed by the
  90    underlying OS. We may also sequentially read more than one source file in an
  91    assembly.  */
  92
  93 /* We must track the physical file and line number for error messages. We also
  94    track a "logical" file and line number corresponding to (C?)  compiler
  95    source line numbers.  Whenever we open a file we must fill in
  96    physical_input_file. So if it is NULL we have not opened any files yet.  */
  97
  98 static const char *physical_input_file;
  99 static const char *logical_input_file;
 100
 101 /* 1-origin line number in a source file.  */
 102 /* A line ends in '\n' or eof.  */
 103 static unsigned int physical_input_line;
 104 static int logical_input_line;
 105
 106 /* Struct used to save the state of the input handler during include files */
 107 struct input_save {
 108   char *              buffer_start;
 109   char *              partial_where;
 110   int                 partial_size;
 111   char                save_source[AFTER_SIZE];
 112   size_t              buffer_length;
 113   const char *              physical_input_file;
 114   const char *              logical_input_file;
 115   unsigned int        physical_input_line;
 116   int                 logical_input_line;
 117   size_t              sb_index;
 118   sb                  from_sb;
 119   int                 from_sb_is_expansion; /* Should we do a conditional check?  */
 120   struct input_save * next_saved_file;  /* Chain of input_saves.  */
 121   char *              input_file_save;  /* Saved state of input routines.  */
 122   char *              saved_position;   /* Caller's saved position in buf.  */
 123 };
 124
 125 static struct input_save *input_scrub_push (char *saved_position);
 126 static char *input_scrub_pop (struct input_save *arg);
 127
 128 /* Saved information about the file that .include'd this one.  When we hit EOF,
 129    we automatically pop to that file.  */
 130
 131 static struct input_save *next_saved_file;
 132
 133 /* Push the state of input reading and scrubbing so that we can #include.
 134    The return value is a 'void *' (fudged for old compilers) to a save
 135    area, which can be restored by passing it to input_scrub_pop().  */
 136
 137 static struct input_save *
 138 input_scrub_push (char *saved_position)
 139 {
 140   struct input_save *saved;
 141
 142   saved = XNEW (struct input_save);
 143
 144   saved->saved_position = saved_position;
 145   saved->buffer_start = buffer_start;
 146   saved->partial_where = partial_where;
 147   saved->partial_size = partial_size;
 148   saved->buffer_length = buffer_length;
 149   saved->physical_input_file = physical_input_file;
 150   saved->logical_input_file = logical_input_file;
 151   saved->physical_input_line = physical_input_line;
 152   saved->logical_input_line = logical_input_line;
 153   saved->sb_index = sb_index;
 154   saved->from_sb = from_sb;
 155   saved->from_sb_is_expansion = from_sb_is_expansion;
 156   memcpy (saved->save_source, save_source, sizeof (save_source));
 157   saved->next_saved_file = next_saved_file;
 158   saved->input_file_save = input_file_push ();
 159
 160   input_file_begin ();          /* Reinitialize! */
 161   logical_input_line = -1;
 162   logical_input_file = NULL;
 163   buffer_length = input_file_buffer_size ();
 164   sb_index = -1;
 165
 166   buffer_start = XNEWVEC (char, (BEFORE_SIZE + buffer_length
 167                                  + buffer_length + AFTER_SIZE + 1));
 168   memcpy (buffer_start, BEFORE_STRING, (int) BEFORE_SIZE);
 169
 170   return saved;
 171 }
 172
 173 static char *
 174 input_scrub_pop (struct input_save *saved)
 175 {
 176   char *saved_position;
 177
 178   input_scrub_end ();           /* Finish off old buffer */
 179
 180   input_file_pop (saved->input_file_save);
 181   saved_position = saved->saved_position;
 182   buffer_start = saved->buffer_start;
 183   buffer_length = saved->buffer_length;
 184   physical_input_file = saved->physical_input_file;
 185   logical_input_file = saved->logical_input_file;
 186   physical_input_line = saved->physical_input_line;
 187   logical_input_line = saved->logical_input_line;
 188   sb_index = saved->sb_index;
 189   from_sb = saved->from_sb;
 190   from_sb_is_expansion = saved->from_sb_is_expansion;
 191   partial_where = saved->partial_where;
 192   partial_size = saved->partial_size;
 193   next_saved_file = saved->next_saved_file;
 194   memcpy (save_source, saved->save_source, sizeof (save_source));
 195
 196   free (saved);
 197   return saved_position;
 198 }
 199 \f
 200 void
 201 input_scrub_begin (void)
 202 {
 203   know (strlen (BEFORE_STRING) == BEFORE_SIZE);
 204   know (strlen (AFTER_STRING) == AFTER_SIZE
 205         || (AFTER_STRING[0] == '\0' && AFTER_SIZE == 1));
 206
 207   input_file_begin ();
 208
 209   buffer_length = input_file_buffer_size ();
 210
 211   buffer_start = XNEWVEC (char, (BEFORE_SIZE + buffer_length
 212                                  + buffer_length + AFTER_SIZE + 1));
 213   memcpy (buffer_start, BEFORE_STRING, (int) BEFORE_SIZE);
 214
 215   /* Line number things.  */
 216   logical_input_line = -1;
 217   logical_input_file = NULL;
 218   physical_input_file = NULL;   /* No file read yet.  */
 219   next_saved_file = NULL;       /* At EOF, don't pop to any other file */
 220   do_scrub_begin (flag_m68k_mri);
 221 }
 222
 223 void
 224 input_scrub_end (void)
 225 {
 226   if (buffer_start)
 227     {
 228       free (buffer_start);
 229       buffer_start = 0;
 230       input_file_end ();
 231     }
 232 }
 233
 234 /* Start reading input from a new file.
 235    Return start of caller's part of buffer.  */
 236
 237 char *
 238 input_scrub_new_file (const char *filename)
 239 {
 240   input_file_open (filename, !flag_no_comments);
 241   physical_input_file = filename[0] ? filename : _("{standard input}");
 242   physical_input_line = 0;
 243
 244   partial_size = 0;
 245   return (buffer_start + BEFORE_SIZE);
 246 }
 247
 248 /* Include a file from the current file.  Save our state, cause it to
 249    be restored on EOF, and begin handling a new file.  Same result as
 250    input_scrub_new_file.  */
 251
 252 char *
 253 input_scrub_include_file (const char *filename, char *position)
 254 {
 255   next_saved_file = input_scrub_push (position);
 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, int is_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 #ifdef md_macro_start
 272   if (is_expansion)
 273     {
 274       md_macro_start ();
 275     }
 276 #endif
 277
 278   next_saved_file = input_scrub_push (position);
 279
 280   /* Allocate sufficient space: from->len + optional newline.  */
 281   newline = from->len >= 1 && from->ptr[0] != '\n';
 282   sb_build (&from_sb, from->len + newline);
 283   from_sb_is_expansion = is_expansion;
 284   if (newline)
 285     {
 286       /* Add the sentinel required by read.c.  */
 287       sb_add_char (&from_sb, '\n');
 288     }
 289   sb_scrub_and_add_sb (&from_sb, from);
 290
 291   /* Make sure the parser looks at defined contents when it scans for
 292      e.g. end-of-line at the end of a macro.  */
 293   sb_terminate (&from_sb);
 294
 295   sb_index = 1;
 296
 297   /* These variables are reset by input_scrub_push.  Restore them
 298      since we are, after all, still at the same point in the file.  */
 299   logical_input_line = next_saved_file->logical_input_line;
 300   logical_input_file = next_saved_file->logical_input_file;
 301 }
 302
 303 void
 304 input_scrub_close (void)
 305 {
 306   input_file_close ();
 307   physical_input_line = 0;
 308   logical_input_line = -1;
 309 }
 310
 311 char *
 312 input_scrub_next_buffer (char **bufp)
 313 {
 314   char *limit;          /*->just after last char of buffer.  */
 315
 316   if (sb_index != (size_t) -1)
 317     {
 318       if (sb_index >= from_sb.len)
 319         {
 320           sb_kill (&from_sb);
 321           if (from_sb_is_expansion)
 322             {
 323               cond_finish_check (macro_nest);
 324 #ifdef md_macro_end
 325               /* Allow the target to clean up per-macro expansion
 326                  data.  */
 327               md_macro_end ();
 328 #endif
 329             }
 330           --macro_nest;
 331           partial_where = NULL;
 332           partial_size = 0;
 333           if (next_saved_file != NULL)
 334             *bufp = input_scrub_pop (next_saved_file);
 335           return partial_where;
 336         }
 337
 338       partial_where = from_sb.ptr + from_sb.len;
 339       partial_size = 0;
 340       *bufp = from_sb.ptr + sb_index;
 341       sb_index = from_sb.len;
 342       return partial_where;
 343     }
 344
 345   if (partial_size)
 346     {
 347       memmove (buffer_start + BEFORE_SIZE, partial_where,
 348                (unsigned int) partial_size);
 349       memcpy (buffer_start + BEFORE_SIZE, save_source, AFTER_SIZE);
 350     }
 351
 352   while (1)
 353     {
 354       char *p;
 355
 356       *bufp = buffer_start + BEFORE_SIZE;
 357       limit = input_file_give_next_buffer (buffer_start
 358                                            + BEFORE_SIZE
 359                                            + partial_size);
 360       if (!limit)
 361         {
 362           if (partial_size == 0)
 363             break;
 364
 365           as_warn (_("end of file not at end of a line; newline inserted"));
 366           p = buffer_start + BEFORE_SIZE + partial_size;
 367           *p++ = '\n';
 368           limit = p;
 369         }
 370       else
 371         {
 372           /* Terminate the buffer to avoid confusing TC_EOL_IN_INSN.  */
 373           *limit = '\0';
 374
 375           /* Find last newline.  */
 376           for (p = limit - 1; *p != '\n' || TC_EOL_IN_INSN (p); --p)
 377             ;
 378           ++p;
 379         }
 380
 381       if (p != buffer_start + BEFORE_SIZE)
 382         {
 383           partial_where = p;
 384           partial_size = limit - p;
 385           memcpy (save_source, partial_where, (int) AFTER_SIZE);
 386           memcpy (partial_where, AFTER_STRING, (int) AFTER_SIZE);
 387           return partial_where;
 388         }
 389
 390       partial_size = limit - (buffer_start + BEFORE_SIZE);
 391       buffer_length += input_file_buffer_size ();
 392       buffer_start = XRESIZEVEC (char, buffer_start,
 393                                  (BEFORE_SIZE
 394                                   + 2 * buffer_length
 395                                   + AFTER_SIZE + 1));
 396     }
 397
 398   /* Tell the listing we've finished the file.  */
 399   LISTING_EOF ();
 400
 401   /* If we should pop to another file at EOF, do it.  */
 402   partial_where = NULL;
 403   if (next_saved_file)
 404     *bufp = input_scrub_pop (next_saved_file);
 405
 406   return partial_where;
 407 }
 408 \f
 409 /* The remaining part of this file deals with line numbers, error
 410    messages and so on.  Return TRUE if we opened any file.  */
 411
 412 int
 413 seen_at_least_1_file (void)
 414 {
 415   return (physical_input_file != NULL);
 416 }
 417
 418 void
 419 bump_line_counters (void)
 420 {
 421   if (sb_index == (size_t) -1)
 422     {
 423       ++physical_input_line;
 424       if (logical_input_line >= 0)
 425         ++logical_input_line;
 426     }
 427 }
 428 \f
 429 /* Tells us what the new logical line number and file are.
 430    If the line_number is -1, we don't change the current logical line
 431    number.  If it is -2, we decrement the logical line number (this is
 432    to support the .appfile pseudo-op inserted into the stream by
 433    do_scrub_chars).
 434    If the fname is NULL, we don't change the current logical file name.
 435    Returns nonzero if the filename actually changes.  */
 436
 437 int
 438 new_logical_line_flags (const char *fname, /* DON'T destroy it!  We point to it!  */
 439                         int line_number,
 440                         int flags)
 441 {
 442   switch (flags)
 443     {
 444     case 0:
 445       break;
 446     case 1:
 447       if (line_number != -1)
 448         abort ();
 449       break;
 450     case 1 << 1:
 451     case 1 << 2:
 452       /* FIXME: we could check that include nesting is correct.  */
 453       break;
 454     default:
 455       abort ();
 456     }
 457
 458   if (line_number >= 0)
 459     logical_input_line = line_number;
 460   else if (line_number == -1 && fname && !*fname && (flags & (1 << 2)))
 461     {
 462       logical_input_file = physical_input_file;
 463       logical_input_line = physical_input_line;
 464       fname = NULL;
 465     }
 466
 467   if (fname
 468       && (logical_input_file == NULL
 469           || filename_cmp (logical_input_file, fname)))
 470     {
 471       logical_input_file = fname;
 472       return 1;
 473     }
 474   else
 475     return 0;
 476 }
 477
 478 int
 479 new_logical_line (const char *fname, int line_number)
 480 {
 481   return new_logical_line_flags (fname, line_number, 0);
 482 }
 483
 484 \f
 485 /* Return the current file name and line number.  */
 486
 487 const char *
 488 as_where (unsigned int *linep)
 489 {
 490   if (logical_input_file != NULL
 491       && (linep == NULL || logical_input_line >= 0))
 492     {
 493       if (linep != NULL)
 494         *linep = logical_input_line;
 495       return logical_input_file;
 496     }
 497   else if (physical_input_file != NULL)
 498     {
 499       if (linep != NULL)
 500         *linep = physical_input_line;
 501       return physical_input_file;
 502     }
 503   else
 504     {
 505       if (linep != NULL)
 506         *linep = 0;
 507       return NULL;
 508     }
 509 }