opcodes/openrisc-asm.c

   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, 2005, 2007
   8    Free Software Foundation, Inc.
   9
  10    This file is part of libopcodes.
  11
  12    This library is free software; you can redistribute it and/or modify
  13    it under the terms of the GNU General Public License as published by
  14    the Free Software Foundation; either version 3, or (at your option)
  15    any later version.
  16
  17    It is distributed in the hope that it will be useful, but WITHOUT
  18    ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
  19    or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public
  20    License for more details.
  21
  22    You should have received a copy of the GNU General Public License
  23    along with this program; if not, write to the Free Software Foundation, Inc.,
  24    51 Franklin Street - Fifth Floor, Boston, MA 02110-1301, USA.  */
  25
  26
  27 /* ??? Eventually more and more of this stuff can go to cpu-independent files.
  28    Keep that in mind.  */
  29
  30 #include "sysdep.h"
  31 #include <stdio.h>
  32 #include "ansidecl.h"
  33 #include "bfd.h"
  34 #include "symcat.h"
  35 #include "openrisc-desc.h"
  36 #include "openrisc-opc.h"
  37 #include "opintl.h"
  38 #include "xregex.h"
  39 #include "libiberty.h"
  40 #include "safe-ctype.h"
  41
  42 #undef  min
  43 #define min(a,b) ((a) < (b) ? (a) : (b))
  44 #undef  max
  45 #define max(a,b) ((a) > (b) ? (a) : (b))
  46
  47 static const char * parse_insn_normal
  48   (CGEN_CPU_DESC, const CGEN_INSN *, const char **, CGEN_FIELDS *);
  49 \f
  50 /* -- assembler routines inserted here.  */
  51
  52 /* -- asm.c */
  53
  54 static const char * MISSING_CLOSING_PARENTHESIS = N_("missing `)'");
  55
  56 #define CGEN_VERBOSE_ASSEMBLER_ERRORS
  57
  58 long
  59 openrisc_sign_extend_16bit (long value)
  60 {
  61   return ((value & 0xffff) ^ 0x8000) - 0x8000;
  62 }
  63
  64 /* Handle hi().  */
  65
  66 static const char *
  67 parse_hi16 (CGEN_CPU_DESC cd, const char ** strp, int opindex, long * valuep)
  68 {
  69   const char *errmsg;
  70   enum cgen_parse_operand_result result_type;
  71   unsigned long ret;
  72
  73   if (**strp == '#')
  74     ++*strp;
  75
  76   if (strncasecmp (*strp, "hi(", 3) == 0)
  77     {
  78       bfd_vma value;
  79
  80       *strp += 3;
  81       errmsg = cgen_parse_address (cd, strp, opindex, BFD_RELOC_HI16,
  82                                    & result_type, & value);
  83       if (**strp != ')')
  84         return MISSING_CLOSING_PARENTHESIS;
  85
  86       ++*strp;
  87       if (errmsg == NULL
  88           && result_type == CGEN_PARSE_OPERAND_RESULT_NUMBER)
  89         value >>= 16;
  90       ret = value;
  91     }
  92   else
  93     {
  94       if (**strp == '-')
  95         {
  96           long value;
  97
  98           errmsg = cgen_parse_signed_integer (cd, strp, opindex, &value);
  99           ret = value;
 100         }
 101       else
 102         {
 103           unsigned long value;
 104
 105           errmsg = cgen_parse_unsigned_integer (cd, strp, opindex, &value);
 106           ret = value;
 107         }
 108     }
 109
 110   *valuep = ((ret & 0xffff) ^ 0x8000) - 0x8000;
 111   return errmsg;
 112 }
 113
 114 /* Handle lo().  */
 115
 116 static const char *
 117 parse_lo16 (CGEN_CPU_DESC cd, const char ** strp, int opindex, long * valuep)
 118 {
 119   const char *errmsg;
 120   enum cgen_parse_operand_result result_type;
 121   unsigned long ret;
 122
 123   if (**strp == '#')
 124     ++*strp;
 125
 126   if (strncasecmp (*strp, "lo(", 3) == 0)
 127     {
 128       bfd_vma value;
 129
 130       *strp += 3;
 131       errmsg = cgen_parse_address (cd, strp, opindex, BFD_RELOC_LO16,
 132                                    & result_type, & value);
 133       if (**strp != ')')
 134         return MISSING_CLOSING_PARENTHESIS;
 135
 136       ++*strp;
 137       ret = value;
 138     }
 139   else
 140     {
 141       if (**strp == '-')
 142         {
 143           long value;
 144
 145           errmsg = cgen_parse_signed_integer (cd, strp, opindex, &value);
 146           ret = value;
 147         }
 148       else
 149         {
 150           unsigned long value;
 151
 152           errmsg = cgen_parse_unsigned_integer (cd, strp, opindex, &value);
 153           ret = value;
 154         }
 155     }
 156
 157   *valuep = ((ret & 0xffff) ^ 0x8000) - 0x8000;
 158   return errmsg;
 159 }
 160
 161 /* -- */
 162
 163 const char * openrisc_cgen_parse_operand
 164   (CGEN_CPU_DESC, int, const char **, CGEN_FIELDS *);
 165
 166 /* Main entry point for operand parsing.
 167
 168    This function is basically just a big switch statement.  Earlier versions
 169    used tables to look up the function to use, but
 170    - if the table contains both assembler and disassembler functions then
 171      the disassembler contains much of the assembler and vice-versa,
 172    - there's a lot of inlining possibilities as things grow,
 173    - using a switch statement avoids the function call overhead.
 174
 175    This function could be moved into `parse_insn_normal', but keeping it
 176    separate makes clear the interface between `parse_insn_normal' and each of
 177    the handlers.  */
 178
 179 const char *
 180 openrisc_cgen_parse_operand (CGEN_CPU_DESC cd,
 181                            int opindex,
 182                            const char ** strp,
 183                            CGEN_FIELDS * fields)
 184 {
 185   const char * errmsg = NULL;
 186   /* Used by scalar operands that still need to be parsed.  */
 187   long junk ATTRIBUTE_UNUSED;
 188
 189   switch (opindex)
 190     {
 191     case OPENRISC_OPERAND_ABS_26 :
 192       {
 193         bfd_vma value = 0;
 194         errmsg = cgen_parse_address (cd, strp, OPENRISC_OPERAND_ABS_26, 0, NULL,  & value);
 195         fields->f_abs26 = value;
 196       }
 197       break;
 198     case OPENRISC_OPERAND_DISP_26 :
 199       {
 200         bfd_vma value = 0;
 201         errmsg = cgen_parse_address (cd, strp, OPENRISC_OPERAND_DISP_26, 0, NULL,  & value);
 202         fields->f_disp26 = value;
 203       }
 204       break;
 205     case OPENRISC_OPERAND_HI16 :
 206       errmsg = parse_hi16 (cd, strp, OPENRISC_OPERAND_HI16, (long *) (& fields->f_simm16));
 207       break;
 208     case OPENRISC_OPERAND_LO16 :
 209       errmsg = parse_lo16 (cd, strp, OPENRISC_OPERAND_LO16, (long *) (& fields->f_lo16));
 210       break;
 211     case OPENRISC_OPERAND_OP_F_23 :
 212       errmsg = cgen_parse_unsigned_integer (cd, strp, OPENRISC_OPERAND_OP_F_23, (unsigned long *) (& fields->f_op4));
 213       break;
 214     case OPENRISC_OPERAND_OP_F_3 :
 215       errmsg = cgen_parse_unsigned_integer (cd, strp, OPENRISC_OPERAND_OP_F_3, (unsigned long *) (& fields->f_op5));
 216       break;
 217     case OPENRISC_OPERAND_RA :
 218       errmsg = cgen_parse_keyword (cd, strp, & openrisc_cgen_opval_h_gr, & fields->f_r2);
 219       break;
 220     case OPENRISC_OPERAND_RB :
 221       errmsg = cgen_parse_keyword (cd, strp, & openrisc_cgen_opval_h_gr, & fields->f_r3);
 222       break;
 223     case OPENRISC_OPERAND_RD :
 224       errmsg = cgen_parse_keyword (cd, strp, & openrisc_cgen_opval_h_gr, & fields->f_r1);
 225       break;
 226     case OPENRISC_OPERAND_SIMM_16 :
 227       errmsg = cgen_parse_signed_integer (cd, strp, OPENRISC_OPERAND_SIMM_16, (long *) (& fields->f_simm16));
 228       break;
 229     case OPENRISC_OPERAND_UI16NC :
 230       errmsg = parse_lo16 (cd, strp, OPENRISC_OPERAND_UI16NC, (long *) (& fields->f_i16nc));
 231       break;
 232     case OPENRISC_OPERAND_UIMM_16 :
 233       errmsg = cgen_parse_unsigned_integer (cd, strp, OPENRISC_OPERAND_UIMM_16, (unsigned long *) (& fields->f_uimm16));
 234       break;
 235     case OPENRISC_OPERAND_UIMM_5 :
 236       errmsg = cgen_parse_unsigned_integer (cd, strp, OPENRISC_OPERAND_UIMM_5, (unsigned long *) (& fields->f_uimm5));
 237       break;
 238
 239     default :
 240       /* xgettext:c-format */
 241       fprintf (stderr, _("Unrecognized field %d while parsing.\n"), opindex);
 242       abort ();
 243   }
 244
 245   return errmsg;
 246 }
 247
 248 cgen_parse_fn * const openrisc_cgen_parse_handlers[] =
 249 {
 250   parse_insn_normal,
 251 };
 252
 253 void
 254 openrisc_cgen_init_asm (CGEN_CPU_DESC cd)
 255 {
 256   openrisc_cgen_init_opcode_table (cd);
 257   openrisc_cgen_init_ibld_table (cd);
 258   cd->parse_handlers = & openrisc_cgen_parse_handlers[0];
 259   cd->parse_operand = openrisc_cgen_parse_operand;
 260 #ifdef CGEN_ASM_INIT_HOOK
 261 CGEN_ASM_INIT_HOOK
 262 #endif
 263 }
 264
 265 \f
 266
 267 /* Regex construction routine.
 268
 269    This translates an opcode syntax string into a regex string,
 270    by replacing any non-character syntax element (such as an
 271    opcode) with the pattern '.*'
 272
 273    It then compiles the regex and stores it in the opcode, for
 274    later use by openrisc_cgen_assemble_insn
 275
 276    Returns NULL for success, an error message for failure.  */
 277
 278 char *
 279 openrisc_cgen_build_insn_regex (CGEN_INSN *insn)
 280 {
 281   CGEN_OPCODE *opc = (CGEN_OPCODE *) CGEN_INSN_OPCODE (insn);
 282   const char *mnem = CGEN_INSN_MNEMONIC (insn);
 283   char rxbuf[CGEN_MAX_RX_ELEMENTS];
 284   char *rx = rxbuf;
 285   const CGEN_SYNTAX_CHAR_TYPE *syn;
 286   int reg_err;
 287
 288   syn = CGEN_SYNTAX_STRING (CGEN_OPCODE_SYNTAX (opc));
 289
 290   /* Mnemonics come first in the syntax string.  */
 291   if (! CGEN_SYNTAX_MNEMONIC_P (* syn))
 292     return _("missing mnemonic in syntax string");
 293   ++syn;
 294
 295   /* Generate a case sensitive regular expression that emulates case
 296      insensitive matching in the "C" locale.  We cannot generate a case
 297      insensitive regular expression because in Turkish locales, 'i' and 'I'
 298      are not equal modulo case conversion.  */
 299
 300   /* Copy the literal mnemonic out of the insn.  */
 301   for (; *mnem; mnem++)
 302     {
 303       char c = *mnem;
 304
 305       if (ISALPHA (c))
 306         {
 307           *rx++ = '[';
 308           *rx++ = TOLOWER (c);
 309           *rx++ = TOUPPER (c);
 310           *rx++ = ']';
 311         }
 312       else
 313         *rx++ = c;
 314     }
 315
 316   /* Copy any remaining literals from the syntax string into the rx.  */
 317   for(; * syn != 0 && rx <= rxbuf + (CGEN_MAX_RX_ELEMENTS - 7 - 4); ++syn)
 318     {
 319       if (CGEN_SYNTAX_CHAR_P (* syn))
 320         {
 321           char c = CGEN_SYNTAX_CHAR (* syn);
 322
 323           switch (c)
 324             {
 325               /* Escape any regex metacharacters in the syntax.  */
 326             case '.': case '[': case '\\':
 327             case '*': case '^': case '$':
 328
 329 #ifdef CGEN_ESCAPE_EXTENDED_REGEX
 330             case '?': case '{': case '}':
 331             case '(': case ')': case '*':
 332             case '|': case '+': case ']':
 333 #endif
 334               *rx++ = '\\';
 335               *rx++ = c;
 336               break;
 337
 338             default:
 339               if (ISALPHA (c))
 340                 {
 341                   *rx++ = '[';
 342                   *rx++ = TOLOWER (c);
 343                   *rx++ = TOUPPER (c);
 344                   *rx++ = ']';
 345                 }
 346               else
 347                 *rx++ = c;
 348               break;
 349             }
 350         }
 351       else
 352         {
 353           /* Replace non-syntax fields with globs.  */
 354           *rx++ = '.';
 355           *rx++ = '*';
 356         }
 357     }
 358
 359   /* Trailing whitespace ok.  */
 360   * rx++ = '[';
 361   * rx++ = ' ';
 362   * rx++ = '\t';
 363   * rx++ = ']';
 364   * rx++ = '*';
 365
 366   /* But anchor it after that.  */
 367   * rx++ = '$';
 368   * rx = '\0';
 369
 370   CGEN_INSN_RX (insn) = xmalloc (sizeof (regex_t));
 371   reg_err = regcomp ((regex_t *) CGEN_INSN_RX (insn), rxbuf, REG_NOSUB);
 372
 373   if (reg_err == 0)
 374     return NULL;
 375   else
 376     {
 377       static char msg[80];
 378
 379       regerror (reg_err, (regex_t *) CGEN_INSN_RX (insn), msg, 80);
 380       regfree ((regex_t *) CGEN_INSN_RX (insn));
 381       free (CGEN_INSN_RX (insn));
 382       (CGEN_INSN_RX (insn)) = NULL;
 383       return msg;
 384     }
 385 }
 386
 387 \f
 388 /* Default insn parser.
 389
 390    The syntax string is scanned and operands are parsed and stored in FIELDS.
 391    Relocs are queued as we go via other callbacks.
 392
 393    ??? Note that this is currently an all-or-nothing parser.  If we fail to
 394    parse the instruction, we return 0 and the caller will start over from
 395    the beginning.  Backtracking will be necessary in parsing subexpressions,
 396    but that can be handled there.  Not handling backtracking here may get
 397    expensive in the case of the m68k.  Deal with later.
 398
 399    Returns NULL for success, an error message for failure.  */
 400
 401 static const char *
 402 parse_insn_normal (CGEN_CPU_DESC cd,
 403                    const CGEN_INSN *insn,
 404                    const char **strp,
 405                    CGEN_FIELDS *fields)
 406 {
 407   /* ??? Runtime added insns not handled yet.  */
 408   const CGEN_SYNTAX *syntax = CGEN_INSN_SYNTAX (insn);
 409   const char *str = *strp;
 410   const char *errmsg;
 411   const char *p;
 412   const CGEN_SYNTAX_CHAR_TYPE * syn;
 413 #ifdef CGEN_MNEMONIC_OPERANDS
 414   /* FIXME: wip */
 415   int past_opcode_p;
 416 #endif
 417
 418   /* For now we assume the mnemonic is first (there are no leading operands).
 419      We can parse it without needing to set up operand parsing.
 420      GAS's input scrubber will ensure mnemonics are lowercase, but we may
 421      not be called from GAS.  */
 422   p = CGEN_INSN_MNEMONIC (insn);
 423   while (*p && TOLOWER (*p) == TOLOWER (*str))
 424     ++p, ++str;
 425
 426   if (* p)
 427     return _("unrecognized instruction");
 428
 429 #ifndef CGEN_MNEMONIC_OPERANDS
 430   if (* str && ! ISSPACE (* str))
 431     return _("unrecognized instruction");
 432 #endif
 433
 434   CGEN_INIT_PARSE (cd);
 435   cgen_init_parse_operand (cd);
 436 #ifdef CGEN_MNEMONIC_OPERANDS
 437   past_opcode_p = 0;
 438 #endif
 439
 440   /* We don't check for (*str != '\0') here because we want to parse
 441      any trailing fake arguments in the syntax string.  */
 442   syn = CGEN_SYNTAX_STRING (syntax);
 443
 444   /* Mnemonics come first for now, ensure valid string.  */
 445   if (! CGEN_SYNTAX_MNEMONIC_P (* syn))
 446     abort ();
 447
 448   ++syn;
 449
 450   while (* syn != 0)
 451     {
 452       /* Non operand chars must match exactly.  */
 453       if (CGEN_SYNTAX_CHAR_P (* syn))
 454         {
 455           /* FIXME: While we allow for non-GAS callers above, we assume the
 456              first char after the mnemonic part is a space.  */
 457           /* FIXME: We also take inappropriate advantage of the fact that
 458              GAS's input scrubber will remove extraneous blanks.  */
 459           if (TOLOWER (*str) == TOLOWER (CGEN_SYNTAX_CHAR (* syn)))
 460             {
 461 #ifdef CGEN_MNEMONIC_OPERANDS
 462               if (CGEN_SYNTAX_CHAR(* syn) == ' ')
 463                 past_opcode_p = 1;
 464 #endif
 465               ++ syn;
 466               ++ str;
 467             }
 468           else if (*str)
 469             {
 470               /* Syntax char didn't match.  Can't be this insn.  */
 471               static char msg [80];
 472
 473               /* xgettext:c-format */
 474               sprintf (msg, _("syntax error (expected char `%c', found `%c')"),
 475                        CGEN_SYNTAX_CHAR(*syn), *str);
 476               return msg;
 477             }
 478           else
 479             {
 480               /* Ran out of input.  */
 481               static char msg [80];
 482
 483               /* xgettext:c-format */
 484               sprintf (msg, _("syntax error (expected char `%c', found end of instruction)"),
 485                        CGEN_SYNTAX_CHAR(*syn));
 486               return msg;
 487             }
 488           continue;
 489         }
 490
 491       /* We have an operand of some sort.  */
 492       errmsg = cd->parse_operand (cd, CGEN_SYNTAX_FIELD (*syn),
 493                                           &str, fields);
 494       if (errmsg)
 495         return errmsg;
 496
 497       /* Done with this operand, continue with next one.  */
 498       ++ syn;
 499     }
 500
 501   /* If we're at the end of the syntax string, we're done.  */
 502   if (* syn == 0)
 503     {
 504       /* FIXME: For the moment we assume a valid `str' can only contain
 505          blanks now.  IE: We needn't try again with a longer version of
 506          the insn and it is assumed that longer versions of insns appear
 507          before shorter ones (eg: lsr r2,r3,1 vs lsr r2,r3).  */
 508       while (ISSPACE (* str))
 509         ++ str;
 510
 511       if (* str != '\0')
 512         return _("junk at end of line"); /* FIXME: would like to include `str' */
 513
 514       return NULL;
 515     }
 516
 517   /* We couldn't parse it.  */
 518   return _("unrecognized instruction");
 519 }
 520 \f
 521 /* Main entry point.
 522    This routine is called for each instruction to be assembled.
 523    STR points to the insn to be assembled.
 524    We assume all necessary tables have been initialized.
 525    The assembled instruction, less any fixups, is stored in BUF.
 526    Remember that if CGEN_INT_INSN_P then BUF is an int and thus the value
 527    still needs to be converted to target byte order, otherwise BUF is an array
 528    of bytes in target byte order.
 529    The result is a pointer to the insn's entry in the opcode table,
 530    or NULL if an error occured (an error message will have already been
 531    printed).
 532
 533    Note that when processing (non-alias) macro-insns,
 534    this function recurses.
 535
 536    ??? It's possible to make this cpu-independent.
 537    One would have to deal with a few minor things.
 538    At this point in time doing so would be more of a curiosity than useful
 539    [for example this file isn't _that_ big], but keeping the possibility in
 540    mind helps keep the design clean.  */
 541
 542 const CGEN_INSN *
 543 openrisc_cgen_assemble_insn (CGEN_CPU_DESC cd,
 544                            const char *str,
 545                            CGEN_FIELDS *fields,
 546                            CGEN_INSN_BYTES_PTR buf,
 547                            char **errmsg)
 548 {
 549   const char *start;
 550   CGEN_INSN_LIST *ilist;
 551   const char *parse_errmsg = NULL;
 552   const char *insert_errmsg = NULL;
 553   int recognized_mnemonic = 0;
 554
 555   /* Skip leading white space.  */
 556   while (ISSPACE (* str))
 557     ++ str;
 558
 559   /* The instructions are stored in hashed lists.
 560      Get the first in the list.  */
 561   ilist = CGEN_ASM_LOOKUP_INSN (cd, str);
 562
 563   /* Keep looking until we find a match.  */
 564   start = str;
 565   for ( ; ilist != NULL ; ilist = CGEN_ASM_NEXT_INSN (ilist))
 566     {
 567       const CGEN_INSN *insn = ilist->insn;
 568       recognized_mnemonic = 1;
 569
 570 #ifdef CGEN_VALIDATE_INSN_SUPPORTED
 571       /* Not usually needed as unsupported opcodes
 572          shouldn't be in the hash lists.  */
 573       /* Is this insn supported by the selected cpu?  */
 574       if (! openrisc_cgen_insn_supported (cd, insn))
 575         continue;
 576 #endif
 577       /* If the RELAXED attribute is set, this is an insn that shouldn't be
 578          chosen immediately.  Instead, it is used during assembler/linker
 579          relaxation if possible.  */
 580       if (CGEN_INSN_ATTR_VALUE (insn, CGEN_INSN_RELAXED) != 0)
 581         continue;
 582
 583       str = start;
 584
 585       /* Skip this insn if str doesn't look right lexically.  */
 586       if (CGEN_INSN_RX (insn) != NULL &&
 587           regexec ((regex_t *) CGEN_INSN_RX (insn), str, 0, NULL, 0) == REG_NOMATCH)
 588         continue;
 589
 590       /* Allow parse/insert handlers to obtain length of insn.  */
 591       CGEN_FIELDS_BITSIZE (fields) = CGEN_INSN_BITSIZE (insn);
 592
 593       parse_errmsg = CGEN_PARSE_FN (cd, insn) (cd, insn, & str, fields);
 594       if (parse_errmsg != NULL)
 595         continue;
 596
 597       /* ??? 0 is passed for `pc'.  */
 598       insert_errmsg = CGEN_INSERT_FN (cd, insn) (cd, insn, fields, buf,
 599                                                  (bfd_vma) 0);
 600       if (insert_errmsg != NULL)
 601         continue;
 602
 603       /* It is up to the caller to actually output the insn and any
 604          queued relocs.  */
 605       return insn;
 606     }
 607
 608   {
 609     static char errbuf[150];
 610 #ifdef CGEN_VERBOSE_ASSEMBLER_ERRORS
 611     const char *tmp_errmsg;
 612
 613     /* If requesting verbose error messages, use insert_errmsg.
 614        Failing that, use parse_errmsg.  */
 615     tmp_errmsg = (insert_errmsg ? insert_errmsg :
 616                   parse_errmsg ? parse_errmsg :
 617                   recognized_mnemonic ?
 618                   _("unrecognized form of instruction") :
 619                   _("unrecognized instruction"));
 620
 621     if (strlen (start) > 50)
 622       /* xgettext:c-format */
 623       sprintf (errbuf, "%s `%.50s...'", tmp_errmsg, start);
 624     else
 625       /* xgettext:c-format */
 626       sprintf (errbuf, "%s `%.50s'", tmp_errmsg, start);
 627 #else
 628     if (strlen (start) > 50)
 629       /* xgettext:c-format */
 630       sprintf (errbuf, _("bad instruction `%.50s...'"), start);
 631     else
 632       /* xgettext:c-format */
 633       sprintf (errbuf, _("bad instruction `%.50s'"), start);
 634 #endif
 635
 636     *errmsg = errbuf;
 637     return NULL;
 638   }
 639 }