path: root/gdb/x86-64-tdep.c

/* Target-dependent code for the x86-64 for GDB, the GNU debugger.

   Copyright 2001, 2002, 2003 Free Software Foundation, Inc.
   Contributed by Jiri Smid, SuSE Labs.

   This file is part of GDB.

   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 of the License, 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 "defs.h"
#include "arch-utils.h"
#include "block.h"
#include "dummy-frame.h"
#include "frame.h"
#include "frame-base.h"
#include "frame-unwind.h"
#include "inferior.h"
#include "gdbcmd.h"
#include "gdbcore.h"
#include "objfiles.h"
#include "regcache.h"
#include "symfile.h"

#include "gdb_assert.h"

#include "x86-64-tdep.h"
#include "i387-tdep.h"

/* Register information.  */

struct x86_64_register_info
{
  char *name;
  struct type **type;
};

static struct x86_64_register_info x86_64_register_info[] =
{
  { "rax", &builtin_type_int64 },
  { "rbx", &builtin_type_int64 },
  { "rcx", &builtin_type_int64 },
  { "rdx", &builtin_type_int64 },
  { "rsi", &builtin_type_int64 },
  { "rdi", &builtin_type_int64 },
  { "rbp", &builtin_type_void_data_ptr },
  { "rsp", &builtin_type_void_data_ptr },

  /* %r8 is indeed register number 8.  */
  { "r8", &builtin_type_int64 },
  { "r9", &builtin_type_int64 },
  { "r10", &builtin_type_int64 },
  { "r11", &builtin_type_int64 },
  { "r12", &builtin_type_int64 },
  { "r13", &builtin_type_int64 },
  { "r14", &builtin_type_int64 },
  { "r15", &builtin_type_int64 },
  { "rip", &builtin_type_void_func_ptr },
  { "eflags", &builtin_type_int32 },
  { "ds", &builtin_type_int32 },
  { "es", &builtin_type_int32 },
  { "fs", &builtin_type_int32 },
  { "gs", &builtin_type_int32 },

  /* %st0 is register number 22.  */
  { "st0", &builtin_type_i387_ext },
  { "st1", &builtin_type_i387_ext },
  { "st2", &builtin_type_i387_ext },
  { "st3", &builtin_type_i387_ext },
  { "st4", &builtin_type_i387_ext },
  { "st5", &builtin_type_i387_ext },
  { "st6", &builtin_type_i387_ext },
  { "st7", &builtin_type_i387_ext },
  { "fctrl", &builtin_type_int32 },
  { "fstat", &builtin_type_int32 },
  { "ftag", &builtin_type_int32 },
  { "fiseg", &builtin_type_int32 },
  { "fioff", &builtin_type_int32 },
  { "foseg", &builtin_type_int32 },
  { "fooff", &builtin_type_int32 },
  { "fop", &builtin_type_int32 },

  /* %xmm0 is register number 38.  */
  { "xmm0", &builtin_type_v4sf },
  { "xmm1", &builtin_type_v4sf },
  { "xmm2", &builtin_type_v4sf },
  { "xmm3", &builtin_type_v4sf },
  { "xmm4", &builtin_type_v4sf },
  { "xmm5", &builtin_type_v4sf },
  { "xmm6", &builtin_type_v4sf },
  { "xmm7", &builtin_type_v4sf },
  { "xmm8", &builtin_type_v4sf },
  { "xmm9", &builtin_type_v4sf },
  { "xmm10", &builtin_type_v4sf },
  { "xmm11", &builtin_type_v4sf },
  { "xmm12", &builtin_type_v4sf },
  { "xmm13", &builtin_type_v4sf },
  { "xmm14", &builtin_type_v4sf },
  { "xmm15", &builtin_type_v4sf },
  { "mxcsr", &builtin_type_int32 }
};

/* Total number of registers.  */
#define X86_64_NUM_REGS \
  (sizeof (x86_64_register_info) / sizeof (x86_64_register_info[0]))

/* Return the name of register REGNUM.  */

static const char *
x86_64_register_name (int regnum)
{
  if (regnum >= 0 && regnum < X86_64_NUM_REGS)
    return x86_64_register_info[regnum].name;

  return NULL;
}

/* Return the GDB type object for the "standard" data type of data in
   register REGNUM. */

static struct type *
x86_64_register_type (struct gdbarch *gdbarch, int regnum)
{
  gdb_assert (regnum >= 0 && regnum < X86_64_NUM_REGS);

  return *x86_64_register_info[regnum].type;
}

/* DWARF Register Number Mapping as defined in the System V psABI,
   section 3.6.  */

static int x86_64_dwarf_regmap[] =
{
  /* General Purpose Registers RAX, RDX, RCX, RBX, RSI, RDI.  */
  X86_64_RAX_REGNUM, X86_64_RDX_REGNUM, 2, 1,
  4, X86_64_RDI_REGNUM,

  /* Frame Pointer Register RBP.  */
  X86_64_RBP_REGNUM,

  /* Stack Pointer Register RSP.  */
  X86_64_RSP_REGNUM,

  /* Extended Integer Registers 8 - 15.  */
  8, 9, 10, 11, 12, 13, 14, 15,

  /* Return Address RA.  Not mapped.  */
  -1,

  /* SSE Registers 0 - 7.  */
  X86_64_XMM0_REGNUM + 0, X86_64_XMM1_REGNUM,
  X86_64_XMM0_REGNUM + 2, X86_64_XMM0_REGNUM + 3,
  X86_64_XMM0_REGNUM + 4, X86_64_XMM0_REGNUM + 5,
  X86_64_XMM0_REGNUM + 6, X86_64_XMM0_REGNUM + 7,

  /* Extended SSE Registers 8 - 15.  */
  X86_64_XMM0_REGNUM + 8, X86_64_XMM0_REGNUM + 9,
  X86_64_XMM0_REGNUM + 10, X86_64_XMM0_REGNUM + 11,
  X86_64_XMM0_REGNUM + 12, X86_64_XMM0_REGNUM + 13,
  X86_64_XMM0_REGNUM + 14, X86_64_XMM0_REGNUM + 15,

  /* Floating Point Registers 0-7.  */
  X86_64_ST0_REGNUM + 0, X86_64_ST0_REGNUM + 1,
  X86_64_ST0_REGNUM + 2, X86_64_ST0_REGNUM + 3,
  X86_64_ST0_REGNUM + 4, X86_64_ST0_REGNUM + 5,
  X86_64_ST0_REGNUM + 6, X86_64_ST0_REGNUM + 7
};

static const int x86_64_dwarf_regmap_len =
  (sizeof (x86_64_dwarf_regmap) / sizeof (x86_64_dwarf_regmap[0]));

/* Convert DWARF register number REG to the appropriate register
   number used by GDB.  */

static int
x86_64_dwarf_reg_to_regnum (int reg)
{
  int regnum = -1;

  if (reg >= 0 || reg < x86_64_dwarf_regmap_len)
    regnum = x86_64_dwarf_regmap[reg];

  if (regnum == -1)
    warning ("Unmapped DWARF Register #%d encountered\n", reg);

  return regnum;
}

/* Return nonzero if a value of type TYPE stored in register REGNUM
   needs any special handling.  */

static int
x86_64_convert_register_p (int regnum, struct type *type)
{
  return i386_fp_regnum_p (regnum);
}


/* The returning of values is done according to the special algorithm.
   Some types are returned in registers an some (big structures) in
   memory.  See the System V psABI for details.  */

#define MAX_CLASSES 4

enum x86_64_reg_class
{
  X86_64_NO_CLASS,
  X86_64_INTEGER_CLASS,
  X86_64_INTEGERSI_CLASS,
  X86_64_SSE_CLASS,
  X86_64_SSESF_CLASS,
  X86_64_SSEDF_CLASS,
  X86_64_SSEUP_CLASS,
  X86_64_X87_CLASS,
  X86_64_X87UP_CLASS,
  X86_64_MEMORY_CLASS
};

/* Return the union class of CLASS1 and CLASS2.
   See the System V psABI for details.  */

static enum x86_64_reg_class
merge_classes (enum x86_64_reg_class class1, enum x86_64_reg_class class2)
{
  /* Rule (a): If both classes are equal, this is the resulting class.  */
  if (class1 == class2)
    return class1;

  /* Rule (b): If one of the classes is NO_CLASS, the resulting class
     is the other class.  */
  if (class1 == X86_64_NO_CLASS)
    return class2;
  if (class2 == X86_64_NO_CLASS)
    return class1;

  /* Rule (c): If one of the classes is MEMORY, the result is MEMORY.  */
  if (class1 == X86_64_MEMORY_CLASS || class2 == X86_64_MEMORY_CLASS)
    return X86_64_MEMORY_CLASS;

  /* Rule (d): If one of the classes is INTEGER, the result is INTEGER.  */
  if ((class1 == X86_64_INTEGERSI_CLASS && class2 == X86_64_SSESF_CLASS)
      || (class2 == X86_64_INTEGERSI_CLASS && class1 == X86_64_SSESF_CLASS))
    return X86_64_INTEGERSI_CLASS;
  if (class1 == X86_64_INTEGER_CLASS || class1 == X86_64_INTEGERSI_CLASS
      || class2 == X86_64_INTEGER_CLASS || class2 == X86_64_INTEGERSI_CLASS)
    return X86_64_INTEGER_CLASS;

  /* Rule (e): If one of the classes is X87 or X87UP class, MEMORY is
     used as class.  */
  if (class1 == X86_64_X87_CLASS || class1 == X86_64_X87UP_CLASS
      || class2 == X86_64_X87_CLASS || class2 == X86_64_X87UP_CLASS)
    return X86_64_MEMORY_CLASS;

  /* Rule (f): Otherwise class SSE is used.  */
  return X86_64_SSE_CLASS;
}

/* Classify the argument type.  CLASSES will be filled by the register
   class used to pass each word of the operand.  The number of words
   is returned.  In case the parameter should be passed in memory, 0
   is returned.  As a special case for zero sized containers,
   classes[0] will be NO_CLASS and 1 is returned.

   See the System V psABI for details.  */

static int
classify_argument (struct type *type,
		   enum x86_64_reg_class classes[MAX_CLASSES], int bit_offset)
{
  int bytes = TYPE_LENGTH (type);
  int words = (bytes + 8 - 1) / 8;

  switch (TYPE_CODE (type))
    {
    case TYPE_CODE_ARRAY:
    case TYPE_CODE_STRUCT:
    case TYPE_CODE_UNION:
      {
	int i;
	enum x86_64_reg_class subclasses[MAX_CLASSES];

	/* On x86-64 we pass structures larger than 16 bytes on the stack.  */
	if (bytes > 16)
	  return 0;

	for (i = 0; i < words; i++)
	  classes[i] = X86_64_NO_CLASS;

	/* Zero sized arrays or structures are NO_CLASS.  We return 0
	   to signalize memory class, so handle it as special case.  */
	if (!words)
	  {
	    classes[0] = X86_64_NO_CLASS;
	    return 1;
	  }
	switch (TYPE_CODE (type))
	  {
	  case TYPE_CODE_STRUCT:
	    {
	      int j;
	      for (j = 0; j < TYPE_NFIELDS (type); ++j)
		{
		  int num = classify_argument (TYPE_FIELDS (type)[j].type,
					       subclasses,
					       (TYPE_FIELDS (type)[j].loc.
						bitpos + bit_offset) % 256);
		  if (!num)
		    return 0;
		  for (i = 0; i < num; i++)
		    {
		      int pos =
			(TYPE_FIELDS (type)[j].loc.bitpos +
			 bit_offset) / 8 / 8;
		      classes[i + pos] =
			merge_classes (subclasses[i], classes[i + pos]);
		    }
		}
	    }
	    break;
	  case TYPE_CODE_ARRAY:
	    {
	      int num;

	      num = classify_argument (TYPE_TARGET_TYPE (type),
				       subclasses, bit_offset);
	      if (!num)
		return 0;

	      /* The partial classes are now full classes.  */
	      if (subclasses[0] == X86_64_SSESF_CLASS && bytes != 4)
		subclasses[0] = X86_64_SSE_CLASS;
	      if (subclasses[0] == X86_64_INTEGERSI_CLASS && bytes != 4)
		subclasses[0] = X86_64_INTEGER_CLASS;

	      for (i = 0; i < words; i++)
		classes[i] = subclasses[i % num];
	    }
	    break;
	  case TYPE_CODE_UNION:
	    {
	      int j;
	      {
		for (j = 0; j < TYPE_NFIELDS (type); ++j)
		  {
		    int num;
		    num = classify_argument (TYPE_FIELDS (type)[j].type,
					     subclasses, bit_offset);
		    if (!num)
		      return 0;
		    for (i = 0; i < num; i++)
		      classes[i] = merge_classes (subclasses[i], classes[i]);
		  }
	      }
	    }
	    break;
	  default:
	    break;
	  }
	/* Final merger cleanup.  */
	for (i = 0; i < words; i++)
	  {
	    /* If one class is MEMORY, everything should be passed in
	       memory.  */
	    if (classes[i] == X86_64_MEMORY_CLASS)
	      return 0;

	    /* The X86_64_SSEUP_CLASS should be always preceeded by
	       X86_64_SSE_CLASS.  */
	    if (classes[i] == X86_64_SSEUP_CLASS
		&& (i == 0 || classes[i - 1] != X86_64_SSE_CLASS))
	      classes[i] = X86_64_SSE_CLASS;

	    /* X86_64_X87UP_CLASS should be preceeded by X86_64_X87_CLASS.  */
	    if (classes[i] == X86_64_X87UP_CLASS
		&& (i == 0 || classes[i - 1] != X86_64_X87_CLASS))
	      classes[i] = X86_64_SSE_CLASS;
	  }
	return words;
      }
      break;
    case TYPE_CODE_FLT:
      switch (bytes)
	{
	case 4:
	  if (!(bit_offset % 64))
	    classes[0] = X86_64_SSESF_CLASS;
	  else
	    classes[0] = X86_64_SSE_CLASS;
	  return 1;
	case 8:
	  classes[0] = X86_64_SSEDF_CLASS;
	  return 1;
	case 16:
	  classes[0] = X86_64_X87_CLASS;
	  classes[1] = X86_64_X87UP_CLASS;
	  return 2;
	}
      break;
    case TYPE_CODE_ENUM:
    case TYPE_CODE_REF:
    case TYPE_CODE_INT:
    case TYPE_CODE_PTR:
      switch (bytes)
	{
	case 1:
	case 2:
	case 4:
	case 8:
	  if (bytes * 8 + bit_offset <= 32)
	    classes[0] = X86_64_INTEGERSI_CLASS;
	  else
	    classes[0] = X86_64_INTEGER_CLASS;
	  return 1;
	case 16:
	  classes[0] = classes[1] = X86_64_INTEGER_CLASS;
	  return 2;
	default:
	  break;
	}
    case TYPE_CODE_VOID:
      return 0;
    default:			/* Avoid warning.  */
      break;
    }
  internal_error (__FILE__, __LINE__,
		  "classify_argument: unknown argument type");
}

/* Examine the argument and set *INT_NREGS and *SSE_NREGS to the
   number of registers required based on the information passed in
   CLASSES.  Return 0 if parameter should be passed in memory.  */

static int
examine_argument (enum x86_64_reg_class classes[MAX_CLASSES],
		  int n, int *int_nregs, int *sse_nregs)
{
  *int_nregs = 0;
  *sse_nregs = 0;
  if (!n)
    return 0;
  for (n--; n >= 0; n--)
    switch (classes[n])
      {
      case X86_64_INTEGER_CLASS:
      case X86_64_INTEGERSI_CLASS:
	(*int_nregs)++;
	break;
      case X86_64_SSE_CLASS:
      case X86_64_SSESF_CLASS:
      case X86_64_SSEDF_CLASS:
	(*sse_nregs)++;
	break;
      case X86_64_NO_CLASS:
      case X86_64_SSEUP_CLASS:
      case X86_64_X87_CLASS:
      case X86_64_X87UP_CLASS:
	break;
      case X86_64_MEMORY_CLASS:
	internal_error (__FILE__, __LINE__,
			"examine_argument: unexpected memory class");
      }
  return 1;
}

#define RET_INT_REGS 2
#define RET_SSE_REGS 2

/* Check if the structure in value_type is returned in registers or in
   memory. If this function returns 1, GDB will call
   STORE_STRUCT_RETURN and EXTRACT_STRUCT_VALUE_ADDRESS else
   STORE_RETURN_VALUE and EXTRACT_RETURN_VALUE will be used.  */

static int
x86_64_use_struct_convention (int gcc_p, struct type *value_type)
{
  enum x86_64_reg_class class[MAX_CLASSES];
  int n = classify_argument (value_type, class, 0);
  int needed_intregs;
  int needed_sseregs;

  return (!n ||
	  !examine_argument (class, n, &needed_intregs, &needed_sseregs) ||
	  needed_intregs > RET_INT_REGS || needed_sseregs > RET_SSE_REGS);
}

/* Extract from an array REGBUF containing the (raw) register state, a
   function return value of TYPE, and copy that, in virtual format,
   into VALBUF.  */

static void
x86_64_extract_return_value (struct type *type, struct regcache *regcache,
			     void *valbuf)
{
  enum x86_64_reg_class class[MAX_CLASSES];
  int n = classify_argument (type, class, 0);
  int needed_intregs;
  int needed_sseregs;
  int intreg = 0;
  int ssereg = 0;
  int offset = 0;
  int ret_int_r[RET_INT_REGS] = { X86_64_RAX_REGNUM, X86_64_RDX_REGNUM };
  int ret_sse_r[RET_SSE_REGS] = { X86_64_XMM0_REGNUM, X86_64_XMM1_REGNUM };

  if (!n ||
      !examine_argument (class, n, &needed_intregs, &needed_sseregs) ||
      needed_intregs > RET_INT_REGS || needed_sseregs > RET_SSE_REGS)
    {				/* memory class */
      CORE_ADDR addr;
      regcache_cooked_read (regcache, X86_64_RAX_REGNUM, &addr);
      read_memory (addr, valbuf, TYPE_LENGTH (type));
      return;
    }
  else
    {
      int i;
      for (i = 0; i < n; i++)
	{
	  switch (class[i])
	    {
	    case X86_64_NO_CLASS:
	      break;
	    case X86_64_INTEGER_CLASS:
	      regcache_cooked_read (regcache, ret_int_r[(intreg + 1) / 2],
				    (char *) valbuf + offset);
	      offset += 8;
	      intreg += 2;
	      break;
	    case X86_64_INTEGERSI_CLASS:
	      regcache_cooked_read_part (regcache, ret_int_r[intreg / 2],
					 0, 4, (char *) valbuf + offset);
	      offset += 8;
	      intreg++;
	      break;
	    case X86_64_SSEDF_CLASS:
	    case X86_64_SSESF_CLASS:
	    case X86_64_SSE_CLASS:
	      regcache_cooked_read_part (regcache,
					 ret_sse_r[(ssereg + 1) / 2], 0, 8,
					 (char *) valbuf + offset);
	      offset += 8;
	      ssereg += 2;
	      break;
	    case X86_64_SSEUP_CLASS:
	      regcache_cooked_read_part (regcache, ret_sse_r[ssereg / 2],
					 0, 8, (char *) valbuf + offset);
	      offset += 8;
	      ssereg++;
	      break;
	    case X86_64_X87_CLASS:
	      regcache_cooked_read_part (regcache, X86_64_ST0_REGNUM,
					 0, 8, (char *) valbuf + offset);
	      offset += 8;
	      break;
	    case X86_64_X87UP_CLASS:
	      regcache_cooked_read_part (regcache, X86_64_ST0_REGNUM,
					 8, 2, (char *) valbuf + offset);
	      offset += 8;
	      break;
	    case X86_64_MEMORY_CLASS:
	    default:
	      internal_error (__FILE__, __LINE__,
			      "Unexpected argument class");
	    }
	}
    }
}

#define INT_REGS 6
#define SSE_REGS 8

static CORE_ADDR
x86_64_push_arguments (struct regcache *regcache, int nargs,
		       struct value **args, CORE_ADDR sp)
{
  int intreg = 0;
  int ssereg = 0;
  /* For varargs functions we have to pass the total number of SSE
     registers used in %rax.  So, let's count this number.  */
  int total_sse_args = 0;
  /* Once an SSE/int argument is passed on the stack, all subsequent
     arguments are passed there.  */
  int sse_stack = 0;
  int int_stack = 0;
  unsigned total_sp;
  int i;
  char buf[8];
  static int int_parameter_registers[INT_REGS] =
  {
    X86_64_RDI_REGNUM, 4,	/* %rdi, %rsi */
    X86_64_RDX_REGNUM, 2,	/* %rdx, %rcx */
    8, 9			/* %r8, %r9 */
  };
  /* %xmm0 - %xmm7 */
  static int sse_parameter_registers[SSE_REGS] =
  {
    X86_64_XMM0_REGNUM + 0, X86_64_XMM1_REGNUM,
    X86_64_XMM0_REGNUM + 2, X86_64_XMM0_REGNUM + 3,
    X86_64_XMM0_REGNUM + 4, X86_64_XMM0_REGNUM + 5,
    X86_64_XMM0_REGNUM + 6, X86_64_XMM0_REGNUM + 7,
  };
  int stack_values_count = 0;
  int *stack_values;
  stack_values = alloca (nargs * sizeof (int));

  for (i = 0; i < nargs; i++)
    {
      enum x86_64_reg_class class[MAX_CLASSES];
      int n = classify_argument (args[i]->type, class, 0);
      int needed_intregs;
      int needed_sseregs;

      if (!n ||
	  !examine_argument (class, n, &needed_intregs, &needed_sseregs))
	{			/* memory class */
	  stack_values[stack_values_count++] = i;
	}
      else
	{
	  int j;
	  int offset = 0;

	  if (intreg / 2 + needed_intregs > INT_REGS)
	    int_stack = 1;
	  if (ssereg / 2 + needed_sseregs > SSE_REGS)
	    sse_stack = 1;
	  if (!sse_stack)
	    total_sse_args += needed_sseregs;

	  for (j = 0; j < n; j++)
	    {
	      switch (class[j])
		{
		case X86_64_NO_CLASS:
		  break;
		case X86_64_INTEGER_CLASS:
		  if (int_stack)
		    stack_values[stack_values_count++] = i;
		  else
		    {
		      regcache_cooked_write
			(regcache, int_parameter_registers[(intreg + 1) / 2],
			 VALUE_CONTENTS_ALL (args[i]) + offset);
		      offset += 8;
		      intreg += 2;
		    }
		  break;
		case X86_64_INTEGERSI_CLASS:
		  if (int_stack)
		    stack_values[stack_values_count++] = i;
		  else
		    {
		      LONGEST val = extract_signed_integer
			(VALUE_CONTENTS_ALL (args[i]) + offset, 4);
		      regcache_cooked_write_signed
			(regcache, int_parameter_registers[intreg / 2], val);
		      
		      offset += 8;
		      intreg++;
		    }
		  break;
		case X86_64_SSEDF_CLASS:
		case X86_64_SSESF_CLASS:
		case X86_64_SSE_CLASS:
		  if (sse_stack)
		    stack_values[stack_values_count++] = i;
		  else
		    {
		      regcache_cooked_write
			(regcache, sse_parameter_registers[(ssereg + 1) / 2],
			 VALUE_CONTENTS_ALL (args[i]) + offset);
		      offset += 8;
		      ssereg += 2;
		    }
		  break;
		case X86_64_SSEUP_CLASS:
		  if (sse_stack)
		    stack_values[stack_values_count++] = i;
		  else
		    {
		      regcache_cooked_write
			(regcache, sse_parameter_registers[ssereg / 2],
			 VALUE_CONTENTS_ALL (args[i]) + offset);
		      offset += 8;
		      ssereg++;
		    }
		  break;
		case X86_64_X87_CLASS:
		case X86_64_MEMORY_CLASS:
		  stack_values[stack_values_count++] = i;
		  break;
		case X86_64_X87UP_CLASS:
		  break;
		default:
		  internal_error (__FILE__, __LINE__,
				  "Unexpected argument class");
		}
	      intreg += intreg % 2;
	      ssereg += ssereg % 2;
	    }
	}
    }

  /* We have to make sure that the stack is 16-byte aligned after the
     setup.  Let's calculate size of arguments first, align stack and
     then fill in the arguments.  */
  total_sp = 0;
  for (i = 0; i < stack_values_count; i++)
    {
      struct value *arg = args[stack_values[i]];
      int len = TYPE_LENGTH (VALUE_ENCLOSING_TYPE (arg));
      total_sp += (len + 7) & ~7;
    }
  /* total_sp is now a multiple of 8, if it is not a multiple of 16,
     change the stack pointer so that it will be afterwards correctly
     aligned.  */
  if (total_sp & 15)
    sp -= 8;
    
  /* Push any remaining arguments onto the stack.  */
  while (--stack_values_count >= 0)
    {
      struct value *arg = args[stack_values[stack_values_count]];
      int len = TYPE_LENGTH (VALUE_ENCLOSING_TYPE (arg));

      /* Make sure the stack is 8-byte-aligned.  */
      sp -= (len + 7) & ~7;
      write_memory (sp, VALUE_CONTENTS_ALL (arg), len);
    }

  /* Write number of SSE type arguments to RAX to take care of varargs
     functions.  */
  store_unsigned_integer (buf, 8, total_sse_args);
  regcache_cooked_write (regcache, X86_64_RAX_REGNUM, buf);

  return sp;
}

/* Write into the appropriate registers a function return value stored
   in VALBUF of type TYPE, given in virtual format.  */

static void
x86_64_store_return_value (struct type *type, struct regcache *regcache,
			   const void *valbuf)
{
  int len = TYPE_LENGTH (type);

  /* First handle long doubles.  */
  if (TYPE_CODE_FLT == TYPE_CODE (type)  && len == 16)
    {
      ULONGEST fstat;
      char buf[FPU_REG_RAW_SIZE];

      /* Returning floating-point values is a bit tricky.  Apart from
         storing the return value in %st(0), we have to simulate the
         state of the FPU at function return point.  */

      /* Convert the value found in VALBUF to the extended
	 floating-point format used by the FPU.  This is probably
	 not exactly how it would happen on the target itself, but
	 it is the best we can do.  */
      convert_typed_floating (valbuf, type, buf, builtin_type_i387_ext);
      regcache_raw_write (regcache, X86_64_ST0_REGNUM, buf);

      /* Set the top of the floating-point register stack to 7.  The
         actual value doesn't really matter, but 7 is what a normal
         function return would end up with if the program started out
         with a freshly initialized FPU.  */
      regcache_raw_read_unsigned (regcache, FSTAT_REGNUM, &fstat);
      fstat |= (7 << 11);
      regcache_raw_write_unsigned (regcache, FSTAT_REGNUM, fstat);

      /* Mark %st(1) through %st(7) as empty.  Since we set the top of
         the floating-point register stack to 7, the appropriate value
         for the tag word is 0x3fff.  */
      regcache_raw_write_unsigned (regcache, FTAG_REGNUM, 0x3fff);
    }
  else if (TYPE_CODE_FLT == TYPE_CODE (type))
    {
      /* Handle double and float variables.  */
      regcache_cooked_write_part (regcache, X86_64_XMM0_REGNUM,
				  0, len, valbuf);
    }
  /* XXX: What about complex floating point types?  */
  else
    {
      int low_size = REGISTER_RAW_SIZE (0);
      int high_size = REGISTER_RAW_SIZE (1);

      if (len <= low_size)
        regcache_cooked_write_part (regcache, 0, 0, len, valbuf);
      else if (len <= (low_size + high_size))
	{
 	  regcache_cooked_write_part (regcache, 0, 0, low_size, valbuf);
 	  regcache_cooked_write_part (regcache, 1, 0,
 				      len - low_size,
 				      (const char *) valbuf + low_size);
	}
      else
	internal_error (__FILE__, __LINE__,
			"Cannot store return value of %d bytes long.", len);
    }
}


static CORE_ADDR
x86_64_push_dummy_call (struct gdbarch *gdbarch, CORE_ADDR func_addr,
			struct regcache *regcache, CORE_ADDR bp_addr,
			int nargs, struct value **args,	CORE_ADDR sp,
			int struct_return, CORE_ADDR struct_addr)
{
  char buf[8];

  /* Pass arguments.  */
  sp = x86_64_push_arguments (regcache, nargs, args, sp);

  /* Pass "hidden" argument".  */
  if (struct_return)
    {
      store_unsigned_integer (buf, 8, struct_addr);
      regcache_cooked_write (regcache, X86_64_RDI_REGNUM, buf);
    }

  /* Store return address.  */
  sp -= 8;
  store_unsigned_integer (buf, 8, bp_addr);
  write_memory (sp, buf, 8);

  /* Finally, update the stack pointer...  */
  store_unsigned_integer (buf, 8, sp);
  regcache_cooked_write (regcache, X86_64_RSP_REGNUM, buf);

  /* ...and fake a frame pointer.  */
  regcache_cooked_write (regcache, X86_64_RBP_REGNUM, buf);

  return sp + 16;
}


/* The maximum number of saved registers.  This should include %rip.  */
#define X86_64_NUM_SAVED_REGS	X86_64_NUM_GREGS

struct x86_64_frame_cache
{
  /* Base address.  */
  CORE_ADDR base;
  CORE_ADDR sp_offset;
  CORE_ADDR pc;

  /* Saved registers.  */
  CORE_ADDR saved_regs[X86_64_NUM_SAVED_REGS];
  CORE_ADDR saved_sp;

  /* Do we have a frame?  */
  int frameless_p;
};

/* Allocate and initialize a frame cache.  */

static struct x86_64_frame_cache *
x86_64_alloc_frame_cache (void)
{
  struct x86_64_frame_cache *cache;
  int i;

  cache = FRAME_OBSTACK_ZALLOC (struct x86_64_frame_cache);

  /* Base address.  */
  cache->base = 0;
  cache->sp_offset = -8;
  cache->pc = 0;

  /* Saved registers.  We initialize these to -1 since zero is a valid
     offset (that's where %rbp is supposed to be stored).  */
  for (i = 0; i < X86_64_NUM_SAVED_REGS; i++)
    cache->saved_regs[i] = -1;
  cache->saved_sp = 0;

  /* Frameless until proven otherwise.  */
  cache->frameless_p = 1;

  return cache;
}

/* Do a limited analysis of the prologue at PC and update CACHE
   accordingly.  Bail out early if CURRENT_PC is reached.  Return the
   address where the analysis stopped.

   We will handle only functions beginning with:

      pushq %rbp        0x55
      movq %rsp, %rbp   0x48 0x89 0xe5

   Any function that doesn't start with this sequence will be assumed
   to have no prologue and thus no valid frame pointer in %rbp.  */

static CORE_ADDR
x86_64_analyze_prologue (CORE_ADDR pc, CORE_ADDR current_pc,
			 struct x86_64_frame_cache *cache)
{
  static unsigned char proto[3] = { 0x48, 0x89, 0xe5 };
  unsigned char buf[3];
  unsigned char op;

  if (current_pc <= pc)
    return current_pc;

  op = read_memory_unsigned_integer (pc, 1);

  if (op == 0x55)		/* pushq %rbp */
    {
      /* Take into account that we've executed the `pushq %rbp' that
         starts this instruction sequence.  */
      cache->saved_regs[X86_64_RBP_REGNUM] = 0;
      cache->sp_offset += 8;

      /* If that's all, return now.  */
      if (current_pc <= pc + 1)
        return current_pc;

      /* Check for `movq %rsp, %rbp'.  */
      read_memory (pc + 1, buf, 3);
      if (memcmp (buf, proto, 3) != 0)
	return pc + 1;

      /* OK, we actually have a frame.  */
      cache->frameless_p = 0;
      return pc + 4;
    }

  return pc;
}

/* Return PC of first real instruction.  */

static CORE_ADDR
x86_64_skip_prologue (CORE_ADDR start_pc)
{
  struct x86_64_frame_cache cache;
  CORE_ADDR pc;

  pc = x86_64_analyze_prologue (start_pc, 0xffffffffffffffff, &cache);
  if (cache.frameless_p)
    return start_pc;

  return pc;
}


/* Normal frames.  */

static struct x86_64_frame_cache *
x86_64_frame_cache (struct frame_info *next_frame, void **this_cache)
{
  struct x86_64_frame_cache *cache;
  char buf[8];
  int i;

  if (*this_cache)
    return *this_cache;

  cache = x86_64_alloc_frame_cache ();
  *this_cache = cache;

  frame_unwind_register (next_frame, X86_64_RBP_REGNUM, buf);
  cache->base = extract_unsigned_integer (buf, 8);
  if (cache->base == 0)
    return cache;

  /* For normal frames, %rip is stored at 8(%rbp).  */
  cache->saved_regs[X86_64_RIP_REGNUM] = 8;

  cache->pc = frame_func_unwind (next_frame);
  if (cache->pc != 0)
    x86_64_analyze_prologue (cache->pc, frame_pc_unwind (next_frame), cache);

  if (cache->frameless_p)
    {
      /* We didn't find a valid frame, which means that CACHE->base
	 currently holds the frame pointer for our calling frame.  If
	 we're at the start of a function, or somewhere half-way its
	 prologue, the function's frame probably hasn't been fully
	 setup yet.  Try to reconstruct the base address for the stack
	 frame by looking at the stack pointer.  For truly "frameless"
	 functions this might work too.  */

      frame_unwind_register (next_frame, X86_64_RSP_REGNUM, buf);
      cache->base = extract_unsigned_integer (buf, 8) + cache->sp_offset;
    }

  /* Now that we have the base address for the stack frame we can
     calculate the value of %rsp in the calling frame.  */
  cache->saved_sp = cache->base + 16;

  /* Adjust all the saved registers such that they contain addresses
     instead of offsets.  */
  for (i = 0; i < X86_64_NUM_SAVED_REGS; i++)
    if (cache->saved_regs[i] != -1)
      cache->saved_regs[i] += cache->base;

  return cache;
}

static void
x86_64_frame_this_id (struct frame_info *next_frame, void **this_cache,
		      struct frame_id *this_id)
{
  struct x86_64_frame_cache *cache =
    x86_64_frame_cache (next_frame, this_cache);

  /* This marks the outermost frame.  */
  if (cache->base == 0)
    return;

  (*this_id) = frame_id_build (cache->base + 16, cache->pc);
}

static void
x86_64_frame_prev_register (struct frame_info *next_frame, void **this_cache,
			    int regnum, int *optimizedp,
			    enum lval_type *lvalp, CORE_ADDR *addrp,
			    int *realnump, void *valuep)
{
  struct x86_64_frame_cache *cache =
    x86_64_frame_cache (next_frame, this_cache);

  gdb_assert (regnum >= 0);

  if (regnum == SP_REGNUM && cache->saved_sp)
    {
      *optimizedp = 0;
      *lvalp = not_lval;
      *addrp = 0;
      *realnump = -1;
      if (valuep)
	{
	  /* Store the value.  */
	  store_unsigned_integer (valuep, 8, cache->saved_sp);
	}
      return;
    }

  if (regnum < X86_64_NUM_SAVED_REGS && cache->saved_regs[regnum] != -1)
    {
      *optimizedp = 0;
      *lvalp = lval_memory;
      *addrp = cache->saved_regs[regnum];
      *realnump = -1;
      if (valuep)
	{
	  /* Read the value in from memory.  */
	  read_memory (*addrp, valuep,
		       register_size (current_gdbarch, regnum));
	}
      return;
    }

  frame_register_unwind (next_frame, regnum,
			 optimizedp, lvalp, addrp, realnump, valuep);
}

static const struct frame_unwind x86_64_frame_unwind =
{
  NORMAL_FRAME,
  x86_64_frame_this_id,
  x86_64_frame_prev_register
};

static const struct frame_unwind *
x86_64_frame_sniffer (struct frame_info *next_frame)
{
  return &x86_64_frame_unwind;
}


/* Signal trampolines.  */

/* FIXME: kettenis/20030419: Perhaps, we can unify the 32-bit and
   64-bit variants.  This would require using identical frame caches
   on both platforms.  */

static struct x86_64_frame_cache *
x86_64_sigtramp_frame_cache (struct frame_info *next_frame, void **this_cache)
{
  struct x86_64_frame_cache *cache;
  struct gdbarch_tdep *tdep = gdbarch_tdep (current_gdbarch);
  CORE_ADDR addr;
  char buf[8];
  int i;

  if (*this_cache)
    return *this_cache;

  cache = x86_64_alloc_frame_cache ();

  frame_unwind_register (next_frame, X86_64_RSP_REGNUM, buf);
  cache->base = extract_unsigned_integer (buf, 8) - 8;

  addr = tdep->sigcontext_addr (next_frame);
  gdb_assert (tdep->sc_reg_offset);
  gdb_assert (tdep->sc_num_regs <= X86_64_NUM_SAVED_REGS);
  for (i = 0; i < tdep->sc_num_regs; i++)
    if (tdep->sc_reg_offset[i] != -1)
      cache->saved_regs[i] = addr + tdep->sc_reg_offset[i];

  *this_cache = cache;
  return cache;
}

static void
x86_64_sigtramp_frame_this_id (struct frame_info *next_frame,
			       void **this_cache, struct frame_id *this_id)
{
  struct x86_64_frame_cache *cache =
    x86_64_sigtramp_frame_cache (next_frame, this_cache);

  (*this_id) = frame_id_build (cache->base + 16, frame_pc_unwind (next_frame));
}

static void
x86_64_sigtramp_frame_prev_register (struct frame_info *next_frame,
				     void **this_cache,
				     int regnum, int *optimizedp,
				     enum lval_type *lvalp, CORE_ADDR *addrp,
				     int *realnump, void *valuep)
{
  /* Make sure we've initialized the cache.  */
  x86_64_sigtramp_frame_cache (next_frame, this_cache);

  x86_64_frame_prev_register (next_frame, this_cache, regnum,
			      optimizedp, lvalp, addrp, realnump, valuep);
}

static const struct frame_unwind x86_64_sigtramp_frame_unwind =
{
  SIGTRAMP_FRAME,
  x86_64_sigtramp_frame_this_id,
  x86_64_sigtramp_frame_prev_register
};

static const struct frame_unwind *
x86_64_sigtramp_frame_sniffer (struct frame_info *next_frame)
{
  CORE_ADDR pc = frame_pc_unwind (next_frame);
  char *name;

  find_pc_partial_function (pc, &name, NULL, NULL);
  if (PC_IN_SIGTRAMP (pc, name))
    {
      gdb_assert (gdbarch_tdep (current_gdbarch)->sigcontext_addr);

      return &x86_64_sigtramp_frame_unwind;
    }

  return NULL;
}


static CORE_ADDR
x86_64_frame_base_address (struct frame_info *next_frame, void **this_cache)
{
  struct x86_64_frame_cache *cache =
    x86_64_frame_cache (next_frame, this_cache);

  return cache->base;
}

static const struct frame_base x86_64_frame_base =
{
  &x86_64_frame_unwind,
  x86_64_frame_base_address,
  x86_64_frame_base_address,
  x86_64_frame_base_address
};

static struct frame_id
x86_64_unwind_dummy_id (struct gdbarch *gdbarch, struct frame_info *next_frame)
{
  char buf[8];
  CORE_ADDR fp;

  frame_unwind_register (next_frame, X86_64_RBP_REGNUM, buf);
  fp = extract_unsigned_integer (buf, 8);

  return frame_id_build (fp + 16, frame_pc_unwind (next_frame));
}

/* 16 byte align the SP per frame requirements.  */

static CORE_ADDR
x86_64_frame_align (struct gdbarch *gdbarch, CORE_ADDR sp)
{
  return sp & -(CORE_ADDR)16;
}

void
x86_64_init_abi (struct gdbarch_info info, struct gdbarch *gdbarch)
{
  struct gdbarch_tdep *tdep = gdbarch_tdep (gdbarch);

  /* The x86-64 has 16 SSE registers.  */
  tdep->num_xmm_regs = 16;

  /* This is what all the fuss is about.  */
  set_gdbarch_long_bit (gdbarch, 64);
  set_gdbarch_long_long_bit (gdbarch, 64);
  set_gdbarch_ptr_bit (gdbarch, 64);

  /* In contrast to the i386, on the x86-64 a `long double' actually
     takes up 128 bits, even though it's still based on the i387
     extended floating-point format which has only 80 significant bits.  */
  set_gdbarch_long_double_bit (gdbarch, 128);

  set_gdbarch_num_regs (gdbarch, X86_64_NUM_REGS);
  set_gdbarch_register_name (gdbarch, x86_64_register_name);
  set_gdbarch_register_type (gdbarch, x86_64_register_type);

  /* Register numbers of various important registers.  */
  set_gdbarch_sp_regnum (gdbarch, X86_64_RSP_REGNUM); /* %rsp */
  set_gdbarch_pc_regnum (gdbarch, X86_64_RIP_REGNUM); /* %rip */
  set_gdbarch_ps_regnum (gdbarch, X86_64_EFLAGS_REGNUM); /* %eflags */
  set_gdbarch_fp0_regnum (gdbarch, X86_64_ST0_REGNUM); /* %st(0) */

  /* The "default" register numbering scheme for the x86-64 is
     referred to as the "DWARF Register Number Mapping" in the System
     V psABI.  The preferred debugging format for all known x86-64
     targets is actually DWARF2, and GCC doesn't seem to support DWARF
     (that is DWARF-1), but we provide the same mapping just in case.
     This mapping is also used for stabs, which GCC does support.  */
  set_gdbarch_stab_reg_to_regnum (gdbarch, x86_64_dwarf_reg_to_regnum);
  set_gdbarch_dwarf_reg_to_regnum (gdbarch, x86_64_dwarf_reg_to_regnum);
  set_gdbarch_dwarf2_reg_to_regnum (gdbarch, x86_64_dwarf_reg_to_regnum);

  /* We don't override SDB_REG_RO_REGNUM, since COFF doesn't seem to
     be in use on any of the supported x86-64 targets.  */

  /* Call dummy code.  */
  set_gdbarch_push_dummy_call (gdbarch, x86_64_push_dummy_call);
  set_gdbarch_frame_align (gdbarch, x86_64_frame_align);
  set_gdbarch_frame_red_zone_size (gdbarch, 128);

  set_gdbarch_convert_register_p (gdbarch, x86_64_convert_register_p);
  set_gdbarch_register_to_value (gdbarch, i387_register_to_value);
  set_gdbarch_value_to_register (gdbarch, i387_value_to_register);

  set_gdbarch_extract_return_value (gdbarch, x86_64_extract_return_value);
  set_gdbarch_store_return_value (gdbarch, x86_64_store_return_value);
  /* Override, since this is handled by x86_64_extract_return_value.  */
  set_gdbarch_extract_struct_value_address (gdbarch, NULL);
  set_gdbarch_use_struct_convention (gdbarch, x86_64_use_struct_convention);

  set_gdbarch_skip_prologue (gdbarch, x86_64_skip_prologue);

  /* Avoid wiring in the MMX registers for now.  */
  set_gdbarch_num_pseudo_regs (gdbarch, 0);

  set_gdbarch_unwind_dummy_id (gdbarch, x86_64_unwind_dummy_id);

  /* FIXME: kettenis/20021026: This is ELF-specific.  Fine for now,
     since all supported x86-64 targets are ELF, but that might change
     in the future.  */
  set_gdbarch_in_solib_call_trampoline (gdbarch, in_plt_section);

  frame_unwind_append_sniffer (gdbarch, x86_64_sigtramp_frame_sniffer);
  frame_unwind_append_sniffer (gdbarch, x86_64_frame_sniffer);
  frame_base_set_default (gdbarch, &x86_64_frame_base);
}


#define I387_FISEG_REGNUM FISEG_REGNUM
#define I387_FOSEG_REGNUM FOSEG_REGNUM

/* The 64-bit FXSAVE format differs from the 32-bit format in the
   sense that the instruction pointer and data pointer are simply
   64-bit offsets into the code segment and the data segment instead
   of a selector offset pair.  The functions below store the upper 32
   bits of these pointers (instead of just the 16-bits of the segment
   selector).  */

/* Fill register REGNUM in GDB's register cache with the appropriate
   floating-point or SSE register value from *FXSAVE.  If REGNUM is
   -1, do this for all registers.  This function masks off any of the
   reserved bits in *FXSAVE.  */

void
x86_64_supply_fxsave (const char *fxsave, int regnum)
{
  i387_supply_fxsave (fxsave, regnum);

  if (fxsave)
    {
      if (regnum == -1 || regnum == I387_FISEG_REGNUM)
	supply_register (I387_FISEG_REGNUM, fxsave + 12);
      if (regnum == -1 || regnum == I387_FOSEG_REGNUM)
      supply_register (I387_FOSEG_REGNUM, fxsave + 20);
    }
}

/* Fill register REGNUM (if it is a floating-point or SSE register) in
   *FXSAVE with the value in GDB's register cache.  If REGNUM is -1, do
   this for all registers.  This function doesn't touch any of the
   reserved bits in *FXSAVE.  */

void
x86_64_fill_fxsave (char *fxsave, int regnum)
{
  i387_fill_fxsave (fxsave, regnum);

  if (regnum == -1 || regnum == I387_FISEG_REGNUM)
    regcache_collect (I387_FISEG_REGNUM, fxsave + 12);
  if (regnum == -1 || regnum == I387_FOSEG_REGNUM)
    regcache_collect (I387_FOSEG_REGNUM, fxsave + 20);
}