opcodes/cgen-asm.in

   1 /* Assembler interface for targets using CGEN. -*- C -*-
   2    CGEN: Cpu tools GENerator
   3
   4 THIS FILE IS MACHINE GENERATED WITH CGEN.
   5 - the resultant file is machine generated, cgen-asm.in isn't
   6
   7 Copyright 1996, 1997, 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
   8
   9 This file is part of the GNU Binutils and GDB, the GNU debugger.
  10
  11 This program is free software; you can redistribute it and/or modify
  12 it under the terms of the GNU General Public License as published by
  13 the Free Software Foundation; either version 2, or (at your option)
  14 any later version.
  15
  16 This program is distributed in the hope that it will be useful,
  17 but WITHOUT ANY WARRANTY; without even the implied warranty of
  18 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19 GNU General Public License for more details.
  20
  21 You should have received a copy of the GNU General Public License
  22 along with this program; if not, write to the Free Software Foundation, Inc.,
  23 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  24
  25 /* ??? Eventually more and more of this stuff can go to cpu-independent files.
  26    Keep that in mind.  */
  27
  28 #include "sysdep.h"
  29 #include <stdio.h>
  30 #include "ansidecl.h"
  31 #include "bfd.h"
  32 #include "symcat.h"
  33 #include "@prefix@-desc.h"
  34 #include "@prefix@-opc.h"
  35 #include "opintl.h"
  36 #include "xregex.h"
  37 #include "libiberty.h"
  38 #include "safe-ctype.h"
  39
  40 #undef  min
  41 #define min(a,b) ((a) < (b) ? (a) : (b))
  42 #undef  max
  43 #define max(a,b) ((a) > (b) ? (a) : (b))
  44
  45 static const char * parse_insn_normal
  46      PARAMS ((CGEN_CPU_DESC, const CGEN_INSN *, const char **, CGEN_FIELDS *));
  47 \f
  48 /* -- assembler routines inserted here.  */
  49 \f
  50
  51 /* Regex construction routine.
  52
  53    This translates an opcode syntax string into a regex string,
  54    by replacing any non-character syntax element (such as an
  55    opcode) with the pattern '.*'
  56
  57    It then compiles the regex and stores it in the opcode, for
  58    later use by @arch@_cgen_assemble_insn
  59
  60    Returns NULL for success, an error message for failure.  */
  61
  62 char *
  63 @arch@_cgen_build_insn_regex (insn)
  64      CGEN_INSN *insn;
  65 {
  66   CGEN_OPCODE *opc = (CGEN_OPCODE *) CGEN_INSN_OPCODE (insn);
  67   const char *mnem = CGEN_INSN_MNEMONIC (insn);
  68   char rxbuf[CGEN_MAX_RX_ELEMENTS];
  69   char *rx = rxbuf;
  70   const CGEN_SYNTAX_CHAR_TYPE *syn;
  71   int reg_err;
  72
  73   syn = CGEN_SYNTAX_STRING (CGEN_OPCODE_SYNTAX (opc));
  74
  75   /* Mnemonics come first in the syntax string.  */
  76   if (! CGEN_SYNTAX_MNEMONIC_P (* syn))
  77     return _("missing mnemonic in syntax string");
  78   ++syn;
  79
  80   /* Generate a case sensitive regular expression that emulates case
  81      insensitive matching in the "C" locale.  We cannot generate a case
  82      insensitive regular expression because in Turkish locales, 'i' and 'I'
  83      are not equal modulo case conversion.  */
  84
  85   /* Copy the literal mnemonic out of the insn.  */
  86   for (; *mnem; mnem++)
  87     {
  88       char c = *mnem;
  89
  90       if (ISALPHA (c))
  91         {
  92           *rx++ = '[';
  93           *rx++ = TOLOWER (c);
  94           *rx++ = TOUPPER (c);
  95           *rx++ = ']';
  96         }
  97       else
  98         *rx++ = c;
  99     }
 100
 101   /* Copy any remaining literals from the syntax string into the rx.  */
 102   for(; * syn != 0 && rx <= rxbuf + (CGEN_MAX_RX_ELEMENTS - 7 - 4); ++syn)
 103     {
 104       if (CGEN_SYNTAX_CHAR_P (* syn))
 105         {
 106           char c = CGEN_SYNTAX_CHAR (* syn);
 107
 108           switch (c)
 109             {
 110               /* Escape any regex metacharacters in the syntax.  */
 111             case '.': case '[': case '\\':
 112             case '*': case '^': case '$':
 113
 114 #ifdef CGEN_ESCAPE_EXTENDED_REGEX
 115             case '?': case '{': case '}':
 116             case '(': case ')': case '*':
 117             case '|': case '+': case ']':
 118 #endif
 119               *rx++ = '\\';
 120               *rx++ = c;
 121               break;
 122
 123             default:
 124               if (ISALPHA (c))
 125                 {
 126                   *rx++ = '[';
 127                   *rx++ = TOLOWER (c);
 128                   *rx++ = TOUPPER (c);
 129                   *rx++ = ']';
 130                 }
 131               else
 132                 *rx++ = c;
 133               break;
 134             }
 135         }
 136       else
 137         {
 138           /* Replace non-syntax fields with globs.  */
 139           *rx++ = '.';
 140           *rx++ = '*';
 141         }
 142     }
 143
 144   /* Trailing whitespace ok.  */
 145   * rx++ = '[';
 146   * rx++ = ' ';
 147   * rx++ = '\t';
 148   * rx++ = ']';
 149   * rx++ = '*';
 150
 151   /* But anchor it after that.  */
 152   * rx++ = '$';
 153   * rx = '\0';
 154
 155   CGEN_INSN_RX (insn) = xmalloc (sizeof (regex_t));
 156   reg_err = regcomp ((regex_t *) CGEN_INSN_RX (insn), rxbuf, REG_NOSUB);
 157
 158   if (reg_err == 0)
 159     return NULL;
 160   else
 161     {
 162       static char msg[80];
 163
 164       regerror (reg_err, (regex_t *) CGEN_INSN_RX (insn), msg, 80);
 165       regfree ((regex_t *) CGEN_INSN_RX (insn));
 166       free (CGEN_INSN_RX (insn));
 167       (CGEN_INSN_RX (insn)) = NULL;
 168       return msg;
 169     }
 170 }
 171
 172 \f
 173 /* Default insn parser.
 174
 175    The syntax string is scanned and operands are parsed and stored in FIELDS.
 176    Relocs are queued as we go via other callbacks.
 177
 178    ??? Note that this is currently an all-or-nothing parser.  If we fail to
 179    parse the instruction, we return 0 and the caller will start over from
 180    the beginning.  Backtracking will be necessary in parsing subexpressions,
 181    but that can be handled there.  Not handling backtracking here may get
 182    expensive in the case of the m68k.  Deal with later.
 183
 184    Returns NULL for success, an error message for failure.  */
 185
 186 static const char *
 187 parse_insn_normal (cd, insn, strp, fields)
 188      CGEN_CPU_DESC cd;
 189      const CGEN_INSN *insn;
 190      const char **strp;
 191      CGEN_FIELDS *fields;
 192 {
 193   /* ??? Runtime added insns not handled yet.  */
 194   const CGEN_SYNTAX *syntax = CGEN_INSN_SYNTAX (insn);
 195   const char *str = *strp;
 196   const char *errmsg;
 197   const char *p;
 198   const CGEN_SYNTAX_CHAR_TYPE * syn;
 199 #ifdef CGEN_MNEMONIC_OPERANDS
 200   /* FIXME: wip */
 201   int past_opcode_p;
 202 #endif
 203
 204   /* For now we assume the mnemonic is first (there are no leading operands).
 205      We can parse it without needing to set up operand parsing.
 206      GAS's input scrubber will ensure mnemonics are lowercase, but we may
 207      not be called from GAS.  */
 208   p = CGEN_INSN_MNEMONIC (insn);
 209   while (*p && TOLOWER (*p) == TOLOWER (*str))
 210     ++p, ++str;
 211
 212   if (* p)
 213     return _("unrecognized instruction");
 214
 215 #ifndef CGEN_MNEMONIC_OPERANDS
 216   if (* str && ! ISSPACE (* str))
 217     return _("unrecognized instruction");
 218 #endif
 219
 220   CGEN_INIT_PARSE (cd);
 221   cgen_init_parse_operand (cd);
 222 #ifdef CGEN_MNEMONIC_OPERANDS
 223   past_opcode_p = 0;
 224 #endif
 225
 226   /* We don't check for (*str != '\0') here because we want to parse
 227      any trailing fake arguments in the syntax string.  */
 228   syn = CGEN_SYNTAX_STRING (syntax);
 229
 230   /* Mnemonics come first for now, ensure valid string.  */
 231   if (! CGEN_SYNTAX_MNEMONIC_P (* syn))
 232     abort ();
 233
 234   ++syn;
 235
 236   while (* syn != 0)
 237     {
 238       /* Non operand chars must match exactly.  */
 239       if (CGEN_SYNTAX_CHAR_P (* syn))
 240         {
 241           /* FIXME: While we allow for non-GAS callers above, we assume the
 242              first char after the mnemonic part is a space.  */
 243           /* FIXME: We also take inappropriate advantage of the fact that
 244              GAS's input scrubber will remove extraneous blanks.  */
 245           if (TOLOWER (*str) == TOLOWER (CGEN_SYNTAX_CHAR (* syn)))
 246             {
 247 #ifdef CGEN_MNEMONIC_OPERANDS
 248               if (CGEN_SYNTAX_CHAR(* syn) == ' ')
 249                 past_opcode_p = 1;
 250 #endif
 251               ++ syn;
 252               ++ str;
 253             }
 254           else if (*str)
 255             {
 256               /* Syntax char didn't match.  Can't be this insn.  */
 257               static char msg [80];
 258
 259               /* xgettext:c-format */
 260               sprintf (msg, _("syntax error (expected char `%c', found `%c')"),
 261                        CGEN_SYNTAX_CHAR(*syn), *str);
 262               return msg;
 263             }
 264           else
 265             {
 266               /* Ran out of input.  */
 267               static char msg [80];
 268
 269               /* xgettext:c-format */
 270               sprintf (msg, _("syntax error (expected char `%c', found end of instruction)"),
 271                        CGEN_SYNTAX_CHAR(*syn));
 272               return msg;
 273             }
 274           continue;
 275         }
 276
 277       /* We have an operand of some sort.  */
 278       errmsg = cd->parse_operand (cd, CGEN_SYNTAX_FIELD (*syn),
 279                                           &str, fields);
 280       if (errmsg)
 281         return errmsg;
 282
 283       /* Done with this operand, continue with next one.  */
 284       ++ syn;
 285     }
 286
 287   /* If we're at the end of the syntax string, we're done.  */
 288   if (* syn == 0)
 289     {
 290       /* FIXME: For the moment we assume a valid `str' can only contain
 291          blanks now.  IE: We needn't try again with a longer version of
 292          the insn and it is assumed that longer versions of insns appear
 293          before shorter ones (eg: lsr r2,r3,1 vs lsr r2,r3).  */
 294       while (ISSPACE (* str))
 295         ++ str;
 296
 297       if (* str != '\0')
 298         return _("junk at end of line"); /* FIXME: would like to include `str' */
 299
 300       return NULL;
 301     }
 302
 303   /* We couldn't parse it.  */
 304   return _("unrecognized instruction");
 305 }
 306 \f
 307 /* Main entry point.
 308    This routine is called for each instruction to be assembled.
 309    STR points to the insn to be assembled.
 310    We assume all necessary tables have been initialized.
 311    The assembled instruction, less any fixups, is stored in BUF.
 312    Remember that if CGEN_INT_INSN_P then BUF is an int and thus the value
 313    still needs to be converted to target byte order, otherwise BUF is an array
 314    of bytes in target byte order.
 315    The result is a pointer to the insn's entry in the opcode table,
 316    or NULL if an error occured (an error message will have already been
 317    printed).
 318
 319    Note that when processing (non-alias) macro-insns,
 320    this function recurses.
 321
 322    ??? It's possible to make this cpu-independent.
 323    One would have to deal with a few minor things.
 324    At this point in time doing so would be more of a curiosity than useful
 325    [for example this file isn't _that_ big], but keeping the possibility in
 326    mind helps keep the design clean.  */
 327
 328 const CGEN_INSN *
 329 @arch@_cgen_assemble_insn (cd, str, fields, buf, errmsg)
 330      CGEN_CPU_DESC cd;
 331      const char *str;
 332      CGEN_FIELDS *fields;
 333      CGEN_INSN_BYTES_PTR buf;
 334      char **errmsg;
 335 {
 336   const char *start;
 337   CGEN_INSN_LIST *ilist;
 338   const char *parse_errmsg = NULL;
 339   const char *insert_errmsg = NULL;
 340   int recognized_mnemonic = 0;
 341
 342   /* Skip leading white space.  */
 343   while (ISSPACE (* str))
 344     ++ str;
 345
 346   /* The instructions are stored in hashed lists.
 347      Get the first in the list.  */
 348   ilist = CGEN_ASM_LOOKUP_INSN (cd, str);
 349
 350   /* Keep looking until we find a match.  */
 351   start = str;
 352   for ( ; ilist != NULL ; ilist = CGEN_ASM_NEXT_INSN (ilist))
 353     {
 354       const CGEN_INSN *insn = ilist->insn;
 355       recognized_mnemonic = 1;
 356
 357 #ifdef CGEN_VALIDATE_INSN_SUPPORTED
 358       /* Not usually needed as unsupported opcodes
 359          shouldn't be in the hash lists.  */
 360       /* Is this insn supported by the selected cpu?  */
 361       if (! @arch@_cgen_insn_supported (cd, insn))
 362         continue;
 363 #endif
 364       /* If the RELAX attribute is set, this is an insn that shouldn't be
 365          chosen immediately.  Instead, it is used during assembler/linker
 366          relaxation if possible.  */
 367       if (CGEN_INSN_ATTR_VALUE (insn, CGEN_INSN_RELAX) != 0)
 368         continue;
 369
 370       str = start;
 371
 372       /* Skip this insn if str doesn't look right lexically.  */
 373       if (CGEN_INSN_RX (insn) != NULL &&
 374           regexec ((regex_t *) CGEN_INSN_RX (insn), str, 0, NULL, 0) == REG_NOMATCH)
 375         continue;
 376
 377       /* Allow parse/insert handlers to obtain length of insn.  */
 378       CGEN_FIELDS_BITSIZE (fields) = CGEN_INSN_BITSIZE (insn);
 379
 380       parse_errmsg = CGEN_PARSE_FN (cd, insn) (cd, insn, & str, fields);
 381       if (parse_errmsg != NULL)
 382         continue;
 383
 384       /* ??? 0 is passed for `pc'.  */
 385       insert_errmsg = CGEN_INSERT_FN (cd, insn) (cd, insn, fields, buf,
 386                                                  (bfd_vma) 0);
 387       if (insert_errmsg != NULL)
 388         continue;
 389
 390       /* It is up to the caller to actually output the insn and any
 391          queued relocs.  */
 392       return insn;
 393     }
 394
 395   {
 396     static char errbuf[150];
 397 #ifdef CGEN_VERBOSE_ASSEMBLER_ERRORS
 398     const char *tmp_errmsg;
 399
 400     /* If requesting verbose error messages, use insert_errmsg.
 401        Failing that, use parse_errmsg.  */
 402     tmp_errmsg = (insert_errmsg ? insert_errmsg :
 403                   parse_errmsg ? parse_errmsg :
 404                   recognized_mnemonic ?
 405                   _("unrecognized form of instruction") :
 406                   _("unrecognized instruction"));
 407
 408     if (strlen (start) > 50)
 409       /* xgettext:c-format */
 410       sprintf (errbuf, "%s `%.50s...'", tmp_errmsg, start);
 411     else
 412       /* xgettext:c-format */
 413       sprintf (errbuf, "%s `%.50s'", tmp_errmsg, start);
 414 #else
 415     if (strlen (start) > 50)
 416       /* xgettext:c-format */
 417       sprintf (errbuf, _("bad instruction `%.50s...'"), start);
 418     else
 419       /* xgettext:c-format */
 420       sprintf (errbuf, _("bad instruction `%.50s'"), start);
 421 #endif
 422
 423     *errmsg = errbuf;
 424     return NULL;
 425   }
 426 }
 427 \f
 428 #if 0 /* This calls back to GAS which we can't do without care.  */
 429
 430 /* Record each member of OPVALS in the assembler's symbol table.
 431    This lets GAS parse registers for us.
 432    ??? Interesting idea but not currently used.  */
 433
 434 /* Record each member of OPVALS in the assembler's symbol table.
 435    FIXME: Not currently used.  */
 436
 437 void
 438 @arch@_cgen_asm_hash_keywords (cd, opvals)
 439      CGEN_CPU_DESC cd;
 440      CGEN_KEYWORD *opvals;
 441 {
 442   CGEN_KEYWORD_SEARCH search = cgen_keyword_search_init (opvals, NULL);
 443   const CGEN_KEYWORD_ENTRY * ke;
 444
 445   while ((ke = cgen_keyword_search_next (& search)) != NULL)
 446     {
 447 #if 0 /* Unnecessary, should be done in the search routine.  */
 448       if (! @arch@_cgen_opval_supported (ke))
 449         continue;
 450 #endif
 451       cgen_asm_record_register (cd, ke->name, ke->value);
 452     }
 453 }
 454
 455 #endif /* 0 */