Wed Nov 4 18:46:47 1998 Dave Brolley <brolley@cygnus.com>

[deliverable/binutils-gdb.git] / opcodes / fr30-asm.c

/* Assembler interface for targets using CGEN. -*- C -*-
   CGEN: Cpu tools GENerator

THIS FILE IS USED TO GENERATE fr30-asm.c.

Copyright (C) 1996, 1997, 1998 Free Software Foundation, Inc.

This file is part of the GNU Binutils and GDB, the GNU debugger.

This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2, or (at your option)
any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software Foundation, Inc.,
59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */

#include "sysdep.h"
#include <ctype.h>
#include <stdio.h>
#include "ansidecl.h"
#include "bfd.h"
#include "symcat.h"
#include "fr30-opc.h"
#include "opintl.h"

#undef min
#define min(a,b) ((a) < (b) ? (a) : (b))
#undef max
#define max(a,b) ((a) > (b) ? (a) : (b))

#undef INLINE
#ifdef __GNUC__
#define INLINE __inline__
#else
#define INLINE
#endif

static const char * insert_normal
     PARAMS ((CGEN_OPCODE_DESC, long, unsigned int, int, int, int,
              CGEN_INSN_BYTES_PTR));
static const char * parse_insn_normal
     PARAMS ((CGEN_OPCODE_DESC, const CGEN_INSN *,
              const char **, CGEN_FIELDS *));
static const char * insert_insn_normal
     PARAMS ((CGEN_OPCODE_DESC, const CGEN_INSN *,
              CGEN_FIELDS *, CGEN_INSN_BYTES_PTR, bfd_vma));
\f
/* -- assembler routines inserted here */

/* Main entry point for operand parsing.

   This function is basically just a big switch statement.  Earlier versions
   used tables to look up the function to use, but
   - if the table contains both assembler and disassembler functions then
     the disassembler contains much of the assembler and vice-versa,
   - there's a lot of inlining possibilities as things grow,
   - using a switch statement avoids the function call overhead.

   This function could be moved into `parse_insn_normal', but keeping it
   separate makes clear the interface between `parse_insn_normal' and each of
   the handlers.
*/

const char *
fr30_cgen_parse_operand (od, opindex, strp, fields)
     CGEN_OPCODE_DESC od;
     int opindex;
     const char ** strp;
     CGEN_FIELDS * fields;
{
  const char * errmsg;

  switch (opindex)
    {
    case FR30_OPERAND_RI :
      errmsg = cgen_parse_keyword (od, strp, & fr30_cgen_opval_h_gr, & fields->f_Ri);
      break;
    case FR30_OPERAND_RJ :
      errmsg = cgen_parse_keyword (od, strp, & fr30_cgen_opval_h_gr, & fields->f_Rj);
      break;

    default :
      /* xgettext:c-format */
      fprintf (stderr, _("Unrecognized field %d while parsing.\n"), opindex);
      abort ();
  }

  return errmsg;
}

/* Main entry point for operand insertion.

   This function is basically just a big switch statement.  Earlier versions
   used tables to look up the function to use, but
   - if the table contains both assembler and disassembler functions then
     the disassembler contains much of the assembler and vice-versa,
   - there's a lot of inlining possibilities as things grow,
   - using a switch statement avoids the function call overhead.

   This function could be moved into `parse_insn_normal', but keeping it
   separate makes clear the interface between `parse_insn_normal' and each of
   the handlers.  It's also needed by GAS to insert operands that couldn't be
   resolved during parsing.
*/

const char *
fr30_cgen_insert_operand (od, opindex, fields, buffer, pc)
     CGEN_OPCODE_DESC od;
     int opindex;
     CGEN_FIELDS * fields;
     CGEN_INSN_BYTES_PTR buffer;
     bfd_vma pc;
{
  const char * errmsg;

  switch (opindex)
    {
    case FR30_OPERAND_RI :
      errmsg = insert_normal (od, fields->f_Ri, 0|(1<<CGEN_OPERAND_UNSIGNED), 12, 4, CGEN_FIELDS_BITSIZE (fields), buffer);
      break;
    case FR30_OPERAND_RJ :
      errmsg = insert_normal (od, fields->f_Rj, 0|(1<<CGEN_OPERAND_UNSIGNED), 8, 4, CGEN_FIELDS_BITSIZE (fields), buffer);
      break;

    default :
      /* xgettext:c-format */
      fprintf (stderr, _("Unrecognized field %d while building insn.\n"),
               opindex);
      abort ();
  }

  return errmsg;
}

cgen_parse_fn * const fr30_cgen_parse_handlers[] = 
{
  0, /* default */
  parse_insn_normal,
};

cgen_insert_fn * const fr30_cgen_insert_handlers[] = 
{
  0, /* default */
  insert_insn_normal,
};

void
fr30_cgen_init_asm (od)
     CGEN_OPCODE_DESC od;
{
}

\f
#if ! CGEN_INT_INSN_P

/* Subroutine of insert_normal.  */

static INLINE void
insert_1 (od, value, start, length, word_length, bufp)
     CGEN_OPCODE_DESC od;
     unsigned long value;
     int start,length,word_length;
     unsigned char *bufp;
{
  unsigned long x,mask;
  int shift;
  int big_p = CGEN_OPCODE_INSN_ENDIAN (od) == CGEN_ENDIAN_BIG;

  switch (word_length)
    {
    case 8:
      x = *bufp;
      break;
    case 16:
      if (big_p)
        x = bfd_getb16 (bufp);
      else
        x = bfd_getl16 (bufp);
      break;
    case 24:
      /* ??? This may need reworking as these cases don't necessarily
         want the first byte and the last two bytes handled like this.  */
      if (big_p)
        x = (bfd_getb8 (bufp) << 16) | bfd_getb16 (bufp + 1);
      else
        x = bfd_getl16 (bufp) | (bfd_getb8 (bufp + 2) << 16);
      break;
    case 32:
      if (big_p)
        x = bfd_getb32 (bufp);
      else
        x = bfd_getl32 (bufp);
      break;
    default :
      abort ();
    }

  /* Written this way to avoid undefined behaviour.  */
  mask = (((1L << (length - 1)) - 1) << 1) | 1;
  if (CGEN_INSN_LSB0_P)
    shift = start;
  else
    shift = (word_length - (start + length));
  x = (x & ~(mask << shift)) | ((value & mask) << shift);

  switch (word_length)
    {
    case 8:
      *bufp = x;
      break;
    case 16:
      if (big_p)
        bfd_putb16 (x, bufp);
      else
        bfd_putl16 (x, bufp);
      break;
    case 24:
      /* ??? This may need reworking as these cases don't necessarily
         want the first byte and the last two bytes handled like this.  */
      if (big_p)
        {
          bfd_putb8 (x >> 16, bufp);
          bfd_putb16 (x, bufp + 1);
        }
      else
        {
          bfd_putl16 (x, bufp);
          bfd_putb8 (x >> 16, bufp + 2);
        }
      break;
    case 32:
      if (big_p)
        bfd_putb32 (x, bufp);
      else
        bfd_putl32 (x, bufp);
      break;
    default :
      abort ();
    }
}

#endif /* ! CGEN_INT_INSN_P */

/* Default insertion routine.

   ATTRS is a mask of the boolean attributes.
   START is the starting bit number, architecture origin.
   LENGTH is the length of VALUE in bits.
   TOTAL_LENGTH is the total length of the insn.

   The result is an error message or NULL if success.  */

/* ??? May need to know word length in order to properly place values as
   an insn may be made of multiple words and the current bit number handling
   may be insufficient.  Word length is an architectural attribute and thus
   methinks the way to go [if needed] is to fetch this value from OD or
   define a macro in <arch>-opc.h rather than adding an extra argument -
   after all that's how endianness is handled.  */
/* ??? This duplicates functionality with bfd's howto table and
   bfd_install_relocation.  */
/* ??? For architectures where insns can be representable as ints,
   store insn in `field' struct and add registers, etc. while parsing?  */
/* ??? This doesn't handle bfd_vma's.  Create another function when
   necessary.  */

static const char *
insert_normal (od, value, attrs, start, length, total_length, buffer)
     CGEN_OPCODE_DESC od;
     long value;
     unsigned int attrs;
     int start;
     int length;
     int total_length;
     CGEN_INSN_BYTES_PTR buffer;
{
  static char errbuf[100];
  /* Written this way to avoid undefined behaviour.  */
  unsigned long mask = (((1L << (length - 1)) - 1) << 1) | 1;

  /* If LENGTH is zero, this operand doesn't contribute to the value.  */
  if (length == 0)
    return NULL;

  /* Ensure VALUE will fit.  */
  if ((attrs & CGEN_ATTR_MASK (CGEN_OPERAND_UNSIGNED)) != 0)
    {
      unsigned long maxval = mask;
      if ((unsigned long) value > maxval)
        {
          /* xgettext:c-format */
          sprintf (errbuf,
                   _("operand out of range (%lu not between 0 and %lu)"),
                   value, maxval);
          return errbuf;
        }
    }
  else
    {
      long minval = - (1L << (length - 1));
      long maxval = (1L << (length - 1)) - 1;
      if (value < minval || value > maxval)
        {
          sprintf
            /* xgettext:c-format */
            (errbuf, _("operand out of range (%ld not between %ld and %ld)"),
             value, minval, maxval);
          return errbuf;
        }
    }

#if CGEN_INT_INSN_P

  if (total_length > 32)
    abort ();
  {
    int shift;

    if (CGEN_INSN_LSB0_P)
      shift = start;
    else
      shift = total_length - (start + length);
    *buffer = (*buffer & ~(mask << shift)) | ((value & mask) << shift);
  }

#else

  /* FIXME: unfinished and untested */

/* ??? To be defined in <arch>-opc.h as necessary.  */
#ifndef CGEN_WORD_ENDIAN
#define CGEN_WORD_ENDIAN(od) CGEN_OPCODE_ENDIAN (od)
#endif
#ifndef CGEN_INSN_WORD_ENDIAN
#define CGEN_INSN_WORD_ENDIAN(od) CGEN_WORD_ENDIAN (od)
#endif

  /* The hard case is probably too slow for the normal cases.
     It's certainly more difficult to understand than the normal case.
     Thus this is split into two.  Keep it that way.  The hard case is defined
     to be when a field straddles a (loosely defined) word boundary
     (??? which may require target specific help to determine).  */

#if 0 /*wip*/

#define HARD_CASE_P 0 /* FIXME:wip */

  if (HARD_CASE_P)
    {
      unsigned char *bufp = (unsigned char *) buffer;
      int insn_length_left = total_length;

      if (CGEN_INSN_LSB0_P)
        {
          int word_offset = (CGEN_INSN_WORD_ENDIAN (od) == CGEN_ENDIAN_BIG
                             ? ...
                             : start / CGEN_BASE_INSN_BITSIZE);
          bufp += word_offset * (CGEN_BASE_INSN_BITSIZE / 8);
          if (CGEN_INSN_WORD_ENDIAN (od) == CGEN_ENDIAN_BIG)
          else
            start -= word_offset * CGEN_BASE_INSN_BITSIZE;
        }
      else
        {
          int word_offset = (CGEN_INSN_WORD_ENDIAN (od) == CGEN_ENDIAN_BIG
                             ? start / CGEN_BASE_INSN_BITSIZE
                             : ...);
          bufp += word_offset * (CGEN_BASE_INSN_BITSIZE / 8);
          if (CGEN_INSN_WORD_ENDIAN (od) == CGEN_ENDIAN_BIG)
            start -= word_offset * CGEN_BASE_INSN_BITSIZE;
          else
        }

      /* Loop so we handle a field straddling an insn word boundary
         (remember, "insn word boundary" is loosely defined here).  */

      while (length > 0)
        {
          int this_pass_length = length;
          int this_pass_start = start;
          int this_pass_word_length = min (insn_length_left,
                                           (CGEN_BASE_INSN_BITSIZE == 8
                                            ? 32
                                            : CGEN_BASE_INSN_BITSIZE));

          insert_1 (od, value, attrs,
                    this_pass_start, this_pass_length, this_pass_word_length,
                    bufp);

          length -= this_pass_length;
          insn_length_left -= this_pass_word_length;
          if (???)
            {
              value >>= ???;
              start += ???;
            }
          else
            {
              value >>= ???;
              start += ???;
            }
          bufp += this_pass_word_length / 8;
        }
    }
  else
#endif /* 0 */
    {
      unsigned char *bufp = (unsigned char *) buffer;

      if (length > 32)
        abort ();

      /* Adjust start,total_length,bufp to point to the pseudo-word that holds
         the value.  For example in a 48 bit insn where the value to insert
         (say an immediate value) is the last 16 bits then word_length here
         would be 16.  To handle a 24 bit insn with an 18 bit immediate,
         insert_1 handles 24 bits (using a combination of bfd_get8,16).  */

      if (total_length > 32)
        {
          int needed_width = start % 8 + length;
          int fetch_length = (needed_width <= 8 ? 8
                              : needed_width <= 16 ? 16
                              : 32);

          if (CGEN_INSN_LSB0_P)
            {
              if (CGEN_INSN_WORD_ENDIAN (od) == CGEN_ENDIAN_BIG)
                {
                  abort (); /* wip */
                }
              else
                {
                  int offset = start & ~7;

                  bufp += offset / 8;
                  start -= offset;
                  total_length -= offset;
                }
            }
          else
            {
              if (CGEN_INSN_WORD_ENDIAN (od) == CGEN_ENDIAN_BIG)
                {
                  int offset = start & ~7;

                  bufp += offset / 8;
                  start -= offset;
                  total_length -= offset;
                }
              else
                {
                  abort (); /* wip */
                }
            }
        }

      insert_1 (od, value, start, length, total_length, bufp);
    }

#endif /* ! CGEN_INT_INSN_P */

  return NULL;
}
\f
/* Default insn parser.

   The syntax string is scanned and operands are parsed and stored in FIELDS.
   Relocs are queued as we go via other callbacks.

   ??? Note that this is currently an all-or-nothing parser.  If we fail to
   parse the instruction, we return 0 and the caller will start over from
   the beginning.  Backtracking will be necessary in parsing subexpressions,
   but that can be handled there.  Not handling backtracking here may get
   expensive in the case of the m68k.  Deal with later.

   Returns NULL for success, an error message for failure.
*/

static const char *
parse_insn_normal (od, insn, strp, fields)
     CGEN_OPCODE_DESC od;
     const CGEN_INSN * insn;
     const char ** strp;
     CGEN_FIELDS * fields;
{
  const CGEN_SYNTAX * syntax = CGEN_INSN_SYNTAX (insn);
  const char * str = *strp;
  const char * errmsg;
  const char * p;
  const unsigned char * syn;
#ifdef CGEN_MNEMONIC_OPERANDS
  /* FIXME: wip */
  int past_opcode_p;
#endif

  /* For now we assume the mnemonic is first (there are no leading operands).
     We can parse it without needing to set up operand parsing.
     GAS's input scrubber will ensure mnemonics are lowercase, but we may
     not be called from GAS.  */
  p = CGEN_INSN_MNEMONIC (insn);
  while (*p && tolower (*p) == tolower (*str))
    ++p, ++str;
  
  if (* p || (* str && !isspace (* str)))
    return _("unrecognized instruction");

  CGEN_INIT_PARSE (od);
  cgen_init_parse_operand (od);
#ifdef CGEN_MNEMONIC_OPERANDS
  past_opcode_p = 0;
#endif

  /* We don't check for (*str != '\0') here because we want to parse
     any trailing fake arguments in the syntax string.  */
  syn = CGEN_SYNTAX_STRING (syntax);

  /* Mnemonics come first for now, ensure valid string.  */
  if (! CGEN_SYNTAX_MNEMONIC_P (* syn))
    abort ();

  ++syn;

  while (* syn != 0)
    {
      /* Non operand chars must match exactly.  */
      if (CGEN_SYNTAX_CHAR_P (* syn))
        {
          if (*str == CGEN_SYNTAX_CHAR (* syn))
            {
#ifdef CGEN_MNEMONIC_OPERANDS
              if (* syn == ' ')
                past_opcode_p = 1;
#endif
              ++ syn;
              ++ str;
            }
          else
            {
              /* Syntax char didn't match.  Can't be this insn.  */
              /* FIXME: would like to return something like
                 "expected char `c'" */
              return _("syntax error");
            }
          continue;
        }

      /* We have an operand of some sort.  */
      errmsg = fr30_cgen_parse_operand (od, CGEN_SYNTAX_FIELD (*syn),
                                          &str, fields);
      if (errmsg)
        return errmsg;

      /* Done with this operand, continue with next one.  */
      ++ syn;
    }

  /* If we're at the end of the syntax string, we're done.  */
  if (* syn == '\0')
    {
      /* FIXME: For the moment we assume a valid `str' can only contain
         blanks now.  IE: We needn't try again with a longer version of
         the insn and it is assumed that longer versions of insns appear
         before shorter ones (eg: lsr r2,r3,1 vs lsr r2,r3).  */
      while (isspace (* str))
        ++ str;

      if (* str != '\0')
        return _("junk at end of line"); /* FIXME: would like to include `str' */

      return NULL;
    }

  /* We couldn't parse it.  */
  return _("unrecognized instruction");
}

/* Default insn builder (insert handler).
   The instruction is recorded in CGEN_INT_INSN_P byte order
   (meaning that if CGEN_INT_INSN_P BUFFER is an int * and thus the value is
   recorded in host byte order, otherwise BUFFER is an array of bytes and the
   value is recorded in target byte order).
   The result is an error message or NULL if success.  */

static const char *
insert_insn_normal (od, insn, fields, buffer, pc)
     CGEN_OPCODE_DESC od;
     const CGEN_INSN * insn;
     CGEN_FIELDS * fields;
     CGEN_INSN_BYTES_PTR buffer;
     bfd_vma pc;
{
  const CGEN_SYNTAX * syntax = CGEN_INSN_SYNTAX (insn);
  unsigned long value;
  const unsigned char * syn;

  CGEN_INIT_INSERT (od);
  value = CGEN_INSN_VALUE (insn);

  /* If we're recording insns as numbers (rather than a string of bytes),
     target byte order handling is deferred until later.  */

#if CGEN_INT_INSN_P

  *buffer = value;

#else

  cgen_insn_put_value (od, buffer, min (CGEN_BASE_INSN_BITSIZE,
                                        CGEN_FIELDS_BITSIZE (fields)),
                       value);

#endif /* ! CGEN_INT_INSN_P */

  /* ??? Rather than scanning the syntax string again, we could store
     in `fields' a null terminated list of the fields that are present.  */

  for (syn = CGEN_SYNTAX_STRING (syntax); * syn != '\0'; ++ syn)
    {
      const char *errmsg;

      if (CGEN_SYNTAX_CHAR_P (* syn))
        continue;

      errmsg = fr30_cgen_insert_operand (od, CGEN_SYNTAX_FIELD (*syn),
                                           fields, buffer, pc);
      if (errmsg)
        return errmsg;
    }

  return NULL;
}
\f
/* Main entry point.
   This routine is called for each instruction to be assembled.
   STR points to the insn to be assembled.
   We assume all necessary tables have been initialized.
   The assembled instruction, less any fixups, is stored in BUF.
   Remember that if CGEN_INT_INSN_P then BUF is an int and thus the value
   still needs to be converted to target byte order, otherwise BUF is an array
   of bytes in target byte order.
   The result is a pointer to the insn's entry in the opcode table,
   or NULL if an error occured (an error message will have already been
   printed).

   Note that when processing (non-alias) macro-insns,
   this function recurses.  */

const CGEN_INSN *
fr30_cgen_assemble_insn (od, str, fields, buf, errmsg)
     CGEN_OPCODE_DESC od;
     const char * str;
     CGEN_FIELDS * fields;
     CGEN_INSN_BYTES_PTR buf;
     char ** errmsg;
{
  const char * start;
  CGEN_INSN_LIST * ilist;

  /* Skip leading white space.  */
  while (isspace (* str))
    ++ str;

  /* The instructions are stored in hashed lists.
     Get the first in the list.  */
  ilist = CGEN_ASM_LOOKUP_INSN (od, str);

  /* Keep looking until we find a match.  */

  start = str;
  for ( ; ilist != NULL ; ilist = CGEN_ASM_NEXT_INSN (ilist))
    {
      const CGEN_INSN *insn = ilist->insn;

#if 0 /* not needed as unsupported opcodes shouldn't be in the hash lists */
      /* Is this insn supported by the selected cpu?  */
      if (! fr30_cgen_insn_supported (od, insn))
        continue;
#endif

      /* If the RELAX attribute is set, this is an insn that shouldn't be
         chosen immediately.  Instead, it is used during assembler/linker
         relaxation if possible.  */
      if (CGEN_INSN_ATTR (insn, CGEN_INSN_RELAX) != 0)
        continue;

      str = start;

      /* Record a default length for the insn.  This will get set to the
         correct value while parsing.  */
      /* FIXME: wip */
      CGEN_FIELDS_BITSIZE (fields) = CGEN_INSN_BITSIZE (insn);

      if (! CGEN_PARSE_FN (insn) (od, insn, & str, fields))
        {
          /* ??? 0 is passed for `pc' */
          if (CGEN_INSERT_FN (insn) (od, insn, fields, buf, (bfd_vma) 0) != NULL)
            continue;
          /* It is up to the caller to actually output the insn and any
             queued relocs.  */
          return insn;
        }

      /* Try the next entry.  */
    }

  /* FIXME: We can return a better error message than this.
     Need to track why it failed and pick the right one.  */
  {
    static char errbuf[100];
    if (strlen (start) > 50)
      /* xgettext:c-format */
      sprintf (errbuf, _("bad instruction `%.50s...'"), start);
    else 
      /* xgettext:c-format */
      sprintf (errbuf, _("bad instruction `%.50s'"), start);
      
    *errmsg = errbuf;
    return NULL;
  }
}
\f
#if 0 /* This calls back to GAS which we can't do without care.  */

/* Record each member of OPVALS in the assembler's symbol table.
   This lets GAS parse registers for us.
   ??? Interesting idea but not currently used.  */

/* Record each member of OPVALS in the assembler's symbol table.
   FIXME: Not currently used.  */

void
fr30_cgen_asm_hash_keywords (od, opvals)
     CGEN_OPCODE_DESC od;
     CGEN_KEYWORD * opvals;
{
  CGEN_KEYWORD_SEARCH search = cgen_keyword_search_init (opvals, NULL);
  const CGEN_KEYWORD_ENTRY * ke;

  while ((ke = cgen_keyword_search_next (& search)) != NULL)
    {
#if 0 /* Unnecessary, should be done in the search routine.  */
      if (! fr30_cgen_opval_supported (ke))
        continue;
#endif
      cgen_asm_record_register (od, ke->name, ke->value);
    }
}

#endif /* 0 */