i386.md (sse_mov?fcc*): New patterns and splitters.

[gcc.git] / gcc / config / i386 / i386.c

/* Subroutines used for code generation on IA-32.
   Copyright (C) 1988, 1992, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001
   Free Software Foundation, Inc.

This file is part of GNU CC.

GNU CC 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.

GNU CC 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 GNU CC; see the file COPYING.  If not, write to
the Free Software Foundation, 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA.  */

#include "config.h"
#include <setjmp.h>
#include "system.h"
#include "rtl.h"
#include "tree.h"
#include "tm_p.h"
#include "regs.h"
#include "hard-reg-set.h"
#include "real.h"
#include "insn-config.h"
#include "conditions.h"
#include "insn-flags.h"
#include "output.h"
#include "insn-attr.h"
#include "flags.h"
#include "except.h"
#include "function.h"
#include "recog.h"
#include "expr.h"
#include "toplev.h"
#include "basic-block.h"
#include "ggc.h"

#ifndef CHECK_STACK_LIMIT
#define CHECK_STACK_LIMIT -1
#endif

/* Processor costs (relative to an add) */
struct processor_costs i386_cost = {	/* 386 specific costs */
  1,					/* cost of an add instruction */
  1,					/* cost of a lea instruction */
  3,					/* variable shift costs */
  2,					/* constant shift costs */
  6,					/* cost of starting a multiply */
  1,					/* cost of multiply per each bit set */
  23,					/* cost of a divide/mod */
  15,					/* "large" insn */
  3,					/* MOVE_RATIO */
  4,					/* cost for loading QImode using movzbl */
  {2, 4, 2},				/* cost of loading integer registers
					   in QImode, HImode and SImode.
					   Relative to reg-reg move (2).  */
  {2, 4, 2},				/* cost of storing integer registers */
  2,					/* cost of reg,reg fld/fst */
  {8, 8, 8},				/* cost of loading fp registers
					   in SFmode, DFmode and XFmode */
  {8, 8, 8},				/* cost of loading integer registers */
  2,					/* cost of moving MMX register */
  {4, 8},				/* cost of loading MMX registers
					   in SImode and DImode */
  {4, 8},				/* cost of storing MMX registers
					   in SImode and DImode */
  2,					/* cost of moving SSE register */
  {4, 8, 16},				/* cost of loading SSE registers
					   in SImode, DImode and TImode */
  {4, 8, 16},				/* cost of storing SSE registers
					   in SImode, DImode and TImode */
  3,					/* MMX or SSE register to integer */
};

struct processor_costs i486_cost = {	/* 486 specific costs */
  1,					/* cost of an add instruction */
  1,					/* cost of a lea instruction */
  3,					/* variable shift costs */
  2,					/* constant shift costs */
  12,					/* cost of starting a multiply */
  1,					/* cost of multiply per each bit set */
  40,					/* cost of a divide/mod */
  15,					/* "large" insn */
  3,					/* MOVE_RATIO */
  4,					/* cost for loading QImode using movzbl */
  {2, 4, 2},				/* cost of loading integer registers
					   in QImode, HImode and SImode.
					   Relative to reg-reg move (2).  */
  {2, 4, 2},				/* cost of storing integer registers */
  2,					/* cost of reg,reg fld/fst */
  {8, 8, 8},				/* cost of loading fp registers
					   in SFmode, DFmode and XFmode */
  {8, 8, 8},				/* cost of loading integer registers */
  2,					/* cost of moving MMX register */
  {4, 8},				/* cost of loading MMX registers
					   in SImode and DImode */
  {4, 8},				/* cost of storing MMX registers
					   in SImode and DImode */
  2,					/* cost of moving SSE register */
  {4, 8, 16},				/* cost of loading SSE registers
					   in SImode, DImode and TImode */
  {4, 8, 16},				/* cost of storing SSE registers
					   in SImode, DImode and TImode */
  3					/* MMX or SSE register to integer */
};

struct processor_costs pentium_cost = {
  1,					/* cost of an add instruction */
  1,					/* cost of a lea instruction */
  4,					/* variable shift costs */
  1,					/* constant shift costs */
  11,					/* cost of starting a multiply */
  0,					/* cost of multiply per each bit set */
  25,					/* cost of a divide/mod */
  8,					/* "large" insn */
  6,					/* MOVE_RATIO */
  6,					/* cost for loading QImode using movzbl */
  {2, 4, 2},				/* cost of loading integer registers
					   in QImode, HImode and SImode.
					   Relative to reg-reg move (2).  */
  {2, 4, 2},				/* cost of storing integer registers */
  2,					/* cost of reg,reg fld/fst */
  {2, 2, 6},				/* cost of loading fp registers
					   in SFmode, DFmode and XFmode */
  {4, 4, 6},				/* cost of loading integer registers */
  8,					/* cost of moving MMX register */
  {8, 8},				/* cost of loading MMX registers
					   in SImode and DImode */
  {8, 8},				/* cost of storing MMX registers
					   in SImode and DImode */
  2,					/* cost of moving SSE register */
  {4, 8, 16},				/* cost of loading SSE registers
					   in SImode, DImode and TImode */
  {4, 8, 16},				/* cost of storing SSE registers
					   in SImode, DImode and TImode */
  3					/* MMX or SSE register to integer */
};

struct processor_costs pentiumpro_cost = {
  1,					/* cost of an add instruction */
  1,					/* cost of a lea instruction */
  1,					/* variable shift costs */
  1,					/* constant shift costs */
  4,					/* cost of starting a multiply */
  0,					/* cost of multiply per each bit set */
  17,					/* cost of a divide/mod */
  8,					/* "large" insn */
  6,					/* MOVE_RATIO */
  2,					/* cost for loading QImode using movzbl */
  {4, 4, 4},				/* cost of loading integer registers
					   in QImode, HImode and SImode.
					   Relative to reg-reg move (2).  */
  {2, 2, 2},				/* cost of storing integer registers */
  2,					/* cost of reg,reg fld/fst */
  {2, 2, 6},				/* cost of loading fp registers
					   in SFmode, DFmode and XFmode */
  {4, 4, 6},				/* cost of loading integer registers */
  2,					/* cost of moving MMX register */
  {2, 2},				/* cost of loading MMX registers
					   in SImode and DImode */
  {2, 2},				/* cost of storing MMX registers
					   in SImode and DImode */
  2,					/* cost of moving SSE register */
  {2, 2, 8},				/* cost of loading SSE registers
					   in SImode, DImode and TImode */
  {2, 2, 8},				/* cost of storing SSE registers
					   in SImode, DImode and TImode */
  3					/* MMX or SSE register to integer */
};

struct processor_costs k6_cost = {
  1,					/* cost of an add instruction */
  2,					/* cost of a lea instruction */
  1,					/* variable shift costs */
  1,					/* constant shift costs */
  3,					/* cost of starting a multiply */
  0,					/* cost of multiply per each bit set */
  18,					/* cost of a divide/mod */
  8,					/* "large" insn */
  4,					/* MOVE_RATIO */
  3,					/* cost for loading QImode using movzbl */
  {4, 5, 4},				/* cost of loading integer registers
					   in QImode, HImode and SImode.
					   Relative to reg-reg move (2).  */
  {2, 3, 2},				/* cost of storing integer registers */
  4,					/* cost of reg,reg fld/fst */
  {6, 6, 6},				/* cost of loading fp registers
					   in SFmode, DFmode and XFmode */
  {4, 4, 4},				/* cost of loading integer registers */
  2,					/* cost of moving MMX register */
  {2, 2},				/* cost of loading MMX registers
					   in SImode and DImode */
  {2, 2},				/* cost of storing MMX registers
					   in SImode and DImode */
  2,					/* cost of moving SSE register */
  {2, 2, 8},				/* cost of loading SSE registers
					   in SImode, DImode and TImode */
  {2, 2, 8},				/* cost of storing SSE registers
					   in SImode, DImode and TImode */
  6					/* MMX or SSE register to integer */
};

struct processor_costs athlon_cost = {
  1,					/* cost of an add instruction */
  2,					/* cost of a lea instruction */
  1,					/* variable shift costs */
  1,					/* constant shift costs */
  5,					/* cost of starting a multiply */
  0,					/* cost of multiply per each bit set */
  42,					/* cost of a divide/mod */
  8,					/* "large" insn */
  9,					/* MOVE_RATIO */
  4,					/* cost for loading QImode using movzbl */
  {4, 5, 4},				/* cost of loading integer registers
					   in QImode, HImode and SImode.
					   Relative to reg-reg move (2).  */
  {2, 3, 2},				/* cost of storing integer registers */
  4,					/* cost of reg,reg fld/fst */
  {6, 6, 20},				/* cost of loading fp registers
					   in SFmode, DFmode and XFmode */
  {4, 4, 16},				/* cost of loading integer registers */
  2,					/* cost of moving MMX register */
  {2, 2},				/* cost of loading MMX registers
					   in SImode and DImode */
  {2, 2},				/* cost of storing MMX registers
					   in SImode and DImode */
  2,					/* cost of moving SSE register */
  {2, 2, 8},				/* cost of loading SSE registers
					   in SImode, DImode and TImode */
  {2, 2, 8},				/* cost of storing SSE registers
					   in SImode, DImode and TImode */
  6					/* MMX or SSE register to integer */
};

struct processor_costs *ix86_cost = &pentium_cost;

/* Processor feature/optimization bitmasks.  */
#define m_386 (1<<PROCESSOR_I386)
#define m_486 (1<<PROCESSOR_I486)
#define m_PENT (1<<PROCESSOR_PENTIUM)
#define m_PPRO (1<<PROCESSOR_PENTIUMPRO)
#define m_K6  (1<<PROCESSOR_K6)
#define m_ATHLON  (1<<PROCESSOR_ATHLON)

const int x86_use_leave = m_386 | m_K6 | m_ATHLON;
const int x86_push_memory = m_386 | m_K6 | m_ATHLON;
const int x86_zero_extend_with_and = m_486 | m_PENT;
const int x86_movx = m_ATHLON | m_PPRO /* m_386 | m_K6 */;
const int x86_double_with_add = ~m_386;
const int x86_use_bit_test = m_386;
const int x86_unroll_strlen = m_486 | m_PENT | m_PPRO | m_ATHLON | m_K6;
const int x86_use_q_reg = m_PENT | m_PPRO | m_K6;
const int x86_use_any_reg = m_486;
const int x86_cmove = m_PPRO | m_ATHLON;
const int x86_deep_branch = m_PPRO | m_K6 | m_ATHLON;
const int x86_use_sahf = m_PPRO | m_K6;
const int x86_partial_reg_stall = m_PPRO;
const int x86_use_loop = m_K6;
const int x86_use_fiop = ~(m_PPRO | m_ATHLON | m_PENT);
const int x86_use_mov0 = m_K6;
const int x86_use_cltd = ~(m_PENT | m_K6);
const int x86_read_modify_write = ~m_PENT;
const int x86_read_modify = ~(m_PENT | m_PPRO);
const int x86_split_long_moves = m_PPRO;
const int x86_promote_QImode = m_K6 | m_PENT | m_386 | m_486;
const int x86_single_stringop = m_386;
const int x86_qimode_math = ~(0);
const int x86_promote_qi_regs = 0;
const int x86_himode_math = ~(m_PPRO);
const int x86_promote_hi_regs = m_PPRO;
const int x86_sub_esp_4 = m_ATHLON | m_PPRO;
const int x86_sub_esp_8 = m_ATHLON | m_PPRO | m_386 | m_486;
const int x86_add_esp_4 = m_ATHLON | m_K6;
const int x86_add_esp_8 = m_ATHLON | m_PPRO | m_K6 | m_386 | m_486;
const int x86_integer_DFmode_moves = ~m_ATHLON;
const int x86_partial_reg_dependency = m_ATHLON;
const int x86_memory_mismatch_stall = m_ATHLON;

#define AT_BP(mode) (gen_rtx_MEM ((mode), hard_frame_pointer_rtx))

const char * const hi_reg_name[] = HI_REGISTER_NAMES;
const char * const qi_reg_name[] = QI_REGISTER_NAMES;
const char * const qi_high_reg_name[] = QI_HIGH_REGISTER_NAMES;

/* Array of the smallest class containing reg number REGNO, indexed by
   REGNO.  Used by REGNO_REG_CLASS in i386.h.  */

enum reg_class const regclass_map[FIRST_PSEUDO_REGISTER] =
{
  /* ax, dx, cx, bx */
  AREG, DREG, CREG, BREG,
  /* si, di, bp, sp */
  SIREG, DIREG, NON_Q_REGS, NON_Q_REGS,
  /* FP registers */
  FP_TOP_REG, FP_SECOND_REG, FLOAT_REGS, FLOAT_REGS,
  FLOAT_REGS, FLOAT_REGS, FLOAT_REGS, FLOAT_REGS,
  /* arg pointer */
  NON_Q_REGS,
  /* flags, fpsr, dirflag, frame */
  NO_REGS, NO_REGS, NO_REGS, NON_Q_REGS,
  SSE_REGS, SSE_REGS, SSE_REGS, SSE_REGS, SSE_REGS, SSE_REGS,
  SSE_REGS, SSE_REGS,
  MMX_REGS, MMX_REGS, MMX_REGS, MMX_REGS, MMX_REGS, MMX_REGS,
  MMX_REGS, MMX_REGS
};

/* The "default" register map.  */

int const dbx_register_map[FIRST_PSEUDO_REGISTER] =
{
  0, 2, 1, 3, 6, 7, 4, 5,		/* general regs */
  12, 13, 14, 15, 16, 17, 18, 19,	/* fp regs */
  -1, -1, -1, -1,			/* arg, flags, fpsr, dir */
  21, 22, 23, 24, 25, 26, 27, 28,	/* SSE */
  29, 30, 31, 32, 33, 34, 35, 36,       /* MMX */
};

/* Define the register numbers to be used in Dwarf debugging information.
   The SVR4 reference port C compiler uses the following register numbers
   in its Dwarf output code:
	0 for %eax (gcc regno = 0)
	1 for %ecx (gcc regno = 2)
	2 for %edx (gcc regno = 1)
	3 for %ebx (gcc regno = 3)
	4 for %esp (gcc regno = 7)
	5 for %ebp (gcc regno = 6)
	6 for %esi (gcc regno = 4)
	7 for %edi (gcc regno = 5)
   The following three DWARF register numbers are never generated by
   the SVR4 C compiler or by the GNU compilers, but SDB on x86/svr4
   believes these numbers have these meanings.
	8  for %eip    (no gcc equivalent)
	9  for %eflags (gcc regno = 17)
	10 for %trapno (no gcc equivalent)
   It is not at all clear how we should number the FP stack registers
   for the x86 architecture.  If the version of SDB on x86/svr4 were
   a bit less brain dead with respect to floating-point then we would
   have a precedent to follow with respect to DWARF register numbers
   for x86 FP registers, but the SDB on x86/svr4 is so completely
   broken with respect to FP registers that it is hardly worth thinking
   of it as something to strive for compatibility with.
   The version of x86/svr4 SDB I have at the moment does (partially)
   seem to believe that DWARF register number 11 is associated with
   the x86 register %st(0), but that's about all.  Higher DWARF
   register numbers don't seem to be associated with anything in
   particular, and even for DWARF regno 11, SDB only seems to under-
   stand that it should say that a variable lives in %st(0) (when
   asked via an `=' command) if we said it was in DWARF regno 11,
   but SDB still prints garbage when asked for the value of the
   variable in question (via a `/' command).
   (Also note that the labels SDB prints for various FP stack regs
   when doing an `x' command are all wrong.)
   Note that these problems generally don't affect the native SVR4
   C compiler because it doesn't allow the use of -O with -g and
   because when it is *not* optimizing, it allocates a memory
   location for each floating-point variable, and the memory
   location is what gets described in the DWARF AT_location
   attribute for the variable in question.
   Regardless of the severe mental illness of the x86/svr4 SDB, we
   do something sensible here and we use the following DWARF
   register numbers.  Note that these are all stack-top-relative
   numbers.
	11 for %st(0) (gcc regno = 8)
	12 for %st(1) (gcc regno = 9)
	13 for %st(2) (gcc regno = 10)
	14 for %st(3) (gcc regno = 11)
	15 for %st(4) (gcc regno = 12)
	16 for %st(5) (gcc regno = 13)
	17 for %st(6) (gcc regno = 14)
	18 for %st(7) (gcc regno = 15)
*/
int const svr4_dbx_register_map[FIRST_PSEUDO_REGISTER] =
{
  0, 2, 1, 3, 6, 7, 5, 4,		/* general regs */
  11, 12, 13, 14, 15, 16, 17, 18,	/* fp regs */
  -1, 9, -1, -1,			/* arg, flags, fpsr, dir */
  21, 22, 23, 24, 25, 26, 27, 28,	/* SSE registers */
  29, 30, 31, 32, 33, 34, 35, 36,	/* MMX registers */
};

/* Test and compare insns in i386.md store the information needed to
   generate branch and scc insns here.  */

struct rtx_def *ix86_compare_op0 = NULL_RTX;
struct rtx_def *ix86_compare_op1 = NULL_RTX;

#define MAX_386_STACK_LOCALS 2

/* Define the structure for the machine field in struct function.  */
struct machine_function
{
  rtx stack_locals[(int) MAX_MACHINE_MODE][MAX_386_STACK_LOCALS];
  int accesses_prev_frame;
};

#define ix86_stack_locals (cfun->machine->stack_locals)

/* Structure describing stack frame layout.
   Stack grows downward:

   [arguments]
					      <- ARG_POINTER
   saved pc

   saved frame pointer if frame_pointer_needed
					      <- HARD_FRAME_POINTER
   [saved regs]

   [padding1]          \
		        )
   [va_arg registers]  (
		        > to_allocate	      <- FRAME_POINTER
   [frame]	       (
		        )
   [padding2]	       /
  */
struct ix86_frame
{
  int nregs;
  int padding1;
  HOST_WIDE_INT frame;
  int padding2;
  int outgoing_arguments_size;

  HOST_WIDE_INT to_allocate;
  /* The offsets relative to ARG_POINTER.  */
  HOST_WIDE_INT frame_pointer_offset;
  HOST_WIDE_INT hard_frame_pointer_offset;
  HOST_WIDE_INT stack_pointer_offset;
};

/* which cpu are we scheduling for */
enum processor_type ix86_cpu;

/* which instruction set architecture to use.  */
int ix86_arch;

/* Strings to hold which cpu and instruction set architecture  to use.  */
const char *ix86_cpu_string;		/* for -mcpu=<xxx> */
const char *ix86_arch_string;		/* for -march=<xxx> */

/* # of registers to use to pass arguments.  */
const char *ix86_regparm_string;

/* ix86_regparm_string as a number */
int ix86_regparm;

/* Alignment to use for loops and jumps:  */

/* Power of two alignment for loops.  */
const char *ix86_align_loops_string;

/* Power of two alignment for non-loop jumps.  */
const char *ix86_align_jumps_string;

/* Power of two alignment for stack boundary in bytes.  */
const char *ix86_preferred_stack_boundary_string;

/* Preferred alignment for stack boundary in bits.  */
int ix86_preferred_stack_boundary;

/* Values 1-5: see jump.c */
int ix86_branch_cost;
const char *ix86_branch_cost_string;

/* Power of two alignment for functions.  */
int ix86_align_funcs;
const char *ix86_align_funcs_string;

/* Power of two alignment for loops.  */
int ix86_align_loops;

/* Power of two alignment for non-loop jumps.  */
int ix86_align_jumps;
\f
static void output_pic_addr_const PARAMS ((FILE *, rtx, int));
static void put_condition_code PARAMS ((enum rtx_code, enum machine_mode,
				       int, int, FILE *));
static rtx ix86_expand_int_compare PARAMS ((enum rtx_code, rtx, rtx));
static enum rtx_code ix86_prepare_fp_compare_args PARAMS ((enum rtx_code,
							   rtx *, rtx *));
static rtx gen_push PARAMS ((rtx));
static int memory_address_length PARAMS ((rtx addr));
static int ix86_flags_dependant PARAMS ((rtx, rtx, enum attr_type));
static int ix86_agi_dependant PARAMS ((rtx, rtx, enum attr_type));
static int ix86_safe_length PARAMS ((rtx));
static enum attr_memory ix86_safe_memory PARAMS ((rtx));
static enum attr_pent_pair ix86_safe_pent_pair PARAMS ((rtx));
static enum attr_ppro_uops ix86_safe_ppro_uops PARAMS ((rtx));
static void ix86_dump_ppro_packet PARAMS ((FILE *));
static void ix86_reorder_insn PARAMS ((rtx *, rtx *));
static rtx * ix86_pent_find_pair PARAMS ((rtx *, rtx *, enum attr_pent_pair,
					 rtx));
static void ix86_init_machine_status PARAMS ((struct function *));
static void ix86_mark_machine_status PARAMS ((struct function *));
static void ix86_free_machine_status PARAMS ((struct function *));
static int ix86_split_to_parts PARAMS ((rtx, rtx *, enum machine_mode));
static int ix86_safe_length_prefix PARAMS ((rtx));
static int ix86_nsaved_regs PARAMS((void));
static void ix86_emit_save_regs PARAMS((void));
static void ix86_emit_restore_regs_using_mov PARAMS ((rtx, int));
static void ix86_emit_epilogue_esp_adjustment PARAMS((int));
static void ix86_set_move_mem_attrs_1 PARAMS ((rtx, rtx, rtx, rtx, rtx));
static void ix86_sched_reorder_pentium PARAMS((rtx *, rtx *));
static void ix86_sched_reorder_ppro PARAMS((rtx *, rtx *));
static HOST_WIDE_INT ix86_GOT_alias_set PARAMS ((void));

struct ix86_address
{
  rtx base, index, disp;
  HOST_WIDE_INT scale;
};

static int ix86_decompose_address PARAMS ((rtx, struct ix86_address *));

struct builtin_description;
static rtx ix86_expand_sse_comi PARAMS ((struct builtin_description *, tree,
					 rtx));
static rtx ix86_expand_sse_compare PARAMS ((struct builtin_description *, tree,
					    rtx));
static rtx ix86_expand_unop1_builtin PARAMS ((enum insn_code, tree, rtx));
static rtx ix86_expand_unop_builtin PARAMS ((enum insn_code, tree, rtx, int));
static rtx ix86_expand_binop_builtin PARAMS ((enum insn_code, tree, rtx));
static rtx ix86_expand_store_builtin PARAMS ((enum insn_code, tree, int));
static rtx safe_vector_operand PARAMS ((rtx, enum machine_mode));
static enum rtx_code ix86_fp_compare_code_to_integer PARAMS ((enum rtx_code));
static void ix86_fp_comparison_codes PARAMS ((enum rtx_code code,
					      enum rtx_code *,
					      enum rtx_code *,
					      enum rtx_code *));
static rtx ix86_expand_fp_compare PARAMS ((enum rtx_code, rtx, rtx, rtx,
					  rtx *, rtx *));
static int ix86_fp_comparison_arithmetics_cost PARAMS ((enum rtx_code code));
static int ix86_fp_comparison_fcomi_cost PARAMS ((enum rtx_code code));
static int ix86_fp_comparison_sahf_cost PARAMS ((enum rtx_code code));
static int ix86_fp_comparison_cost PARAMS ((enum rtx_code code));
static int ix86_save_reg PARAMS ((int));
static void ix86_compute_frame_layout PARAMS ((struct ix86_frame *));
\f
/* Sometimes certain combinations of command options do not make
   sense on a particular target machine.  You can define a macro
   `OVERRIDE_OPTIONS' to take account of this.  This macro, if
   defined, is executed once just after all the command options have
   been parsed.

   Don't use this macro to turn on various extra optimizations for
   `-O'.  That is what `OPTIMIZATION_OPTIONS' is for.  */

void
override_options ()
{
  int i;
  /* Comes from final.c -- no real reason to change it.  */
#define MAX_CODE_ALIGN 16

  static struct ptt
    {
      struct processor_costs *cost;	/* Processor costs */
      int target_enable;		/* Target flags to enable.  */
      int target_disable;		/* Target flags to disable.  */
      int align_loop;			/* Default alignments.  */
      int align_jump;
      int align_func;
      int branch_cost;
    }
  const processor_target_table[PROCESSOR_max] =
    {
      {&i386_cost, 0, 0, 2, 2, 2, 1},
      {&i486_cost, 0, 0, 4, 4, 4, 1},
      {&pentium_cost, 0, 0, -4, -4, -4, 1},
      {&pentiumpro_cost, 0, 0, 4, -4, 4, 1},
      {&k6_cost, 0, 0, -5, -5, 4, 1},
      {&athlon_cost, 0, 0, 4, -4, 4, 1}
    };

  static struct pta
    {
      const char *name;		/* processor name or nickname.  */
      enum processor_type processor;
    }
  const processor_alias_table[] =
    {
      {"i386", PROCESSOR_I386},
      {"i486", PROCESSOR_I486},
      {"i586", PROCESSOR_PENTIUM},
      {"pentium", PROCESSOR_PENTIUM},
      {"i686", PROCESSOR_PENTIUMPRO},
      {"pentiumpro", PROCESSOR_PENTIUMPRO},
      {"k6", PROCESSOR_K6},
      {"athlon", PROCESSOR_ATHLON},
    };

  int const pta_size = sizeof (processor_alias_table) / sizeof (struct pta);

#ifdef SUBTARGET_OVERRIDE_OPTIONS
  SUBTARGET_OVERRIDE_OPTIONS;
#endif

  ix86_arch = PROCESSOR_I386;
  ix86_cpu = (enum processor_type) TARGET_CPU_DEFAULT;

  if (ix86_arch_string != 0)
    {
      for (i = 0; i < pta_size; i++)
	if (! strcmp (ix86_arch_string, processor_alias_table[i].name))
	  {
	    ix86_arch = processor_alias_table[i].processor;
	    /* Default cpu tuning to the architecture.  */
	    ix86_cpu = ix86_arch;
	    break;
	  }

      if (i == pta_size)
	error ("bad value (%s) for -march= switch", ix86_arch_string);
    }

  if (ix86_cpu_string != 0)
    {
      for (i = 0; i < pta_size; i++)
	if (! strcmp (ix86_cpu_string, processor_alias_table[i].name))
	  {
	    ix86_cpu = processor_alias_table[i].processor;
	    break;
	  }
      if (i == pta_size)
	error ("bad value (%s) for -mcpu= switch", ix86_cpu_string);
    }

  ix86_cost = processor_target_table[ix86_cpu].cost;
  target_flags |= processor_target_table[ix86_cpu].target_enable;
  target_flags &= ~processor_target_table[ix86_cpu].target_disable;

  /* Arrange to set up i386_stack_locals for all functions.  */
  init_machine_status = ix86_init_machine_status;
  mark_machine_status = ix86_mark_machine_status;
  free_machine_status = ix86_free_machine_status;

  /* Validate -mregparm= value.  */
  if (ix86_regparm_string)
    {
      i = atoi (ix86_regparm_string);
      if (i < 0 || i > REGPARM_MAX)
	error ("-mregparm=%d is not between 0 and %d", i, REGPARM_MAX);
      else
	ix86_regparm = i;
    }

  /* Validate -malign-loops= value, or provide default.  */
  ix86_align_loops = processor_target_table[ix86_cpu].align_loop;
  if (ix86_align_loops_string)
    {
      i = atoi (ix86_align_loops_string);
      if (i < 0 || i > MAX_CODE_ALIGN)
	error ("-malign-loops=%d is not between 0 and %d", i, MAX_CODE_ALIGN);
      else
	ix86_align_loops = i;
    }

  /* Validate -malign-jumps= value, or provide default.  */
  ix86_align_jumps = processor_target_table[ix86_cpu].align_jump;
  if (ix86_align_jumps_string)
    {
      i = atoi (ix86_align_jumps_string);
      if (i < 0 || i > MAX_CODE_ALIGN)
	error ("-malign-jumps=%d is not between 0 and %d", i, MAX_CODE_ALIGN);
      else
	ix86_align_jumps = i;
    }

  /* Validate -malign-functions= value, or provide default.  */
  ix86_align_funcs = processor_target_table[ix86_cpu].align_func;
  if (ix86_align_funcs_string)
    {
      i = atoi (ix86_align_funcs_string);
      if (i < 0 || i > MAX_CODE_ALIGN)
	error ("-malign-functions=%d is not between 0 and %d",
	       i, MAX_CODE_ALIGN);
      else
	ix86_align_funcs = i;
    }

  /* Validate -mpreferred-stack-boundary= value, or provide default.
     The default of 128 bits is for Pentium III's SSE __m128.  */
  ix86_preferred_stack_boundary = 128;
  if (ix86_preferred_stack_boundary_string)
    {
      i = atoi (ix86_preferred_stack_boundary_string);
      if (i < 2 || i > 31)
	error ("-mpreferred-stack-boundary=%d is not between 2 and 31", i);
      else
	ix86_preferred_stack_boundary = (1 << i) * BITS_PER_UNIT;
    }

  /* Validate -mbranch-cost= value, or provide default.  */
  ix86_branch_cost = processor_target_table[ix86_cpu].branch_cost;
  if (ix86_branch_cost_string)
    {
      i = atoi (ix86_branch_cost_string);
      if (i < 0 || i > 5)
	error ("-mbranch-cost=%d is not between 0 and 5", i);
      else
	ix86_branch_cost = i;
    }

  /* Keep nonleaf frame pointers.  */
  if (TARGET_OMIT_LEAF_FRAME_POINTER)
    flag_omit_frame_pointer = 1;

  /* If we're doing fast math, we don't care about comparison order
     wrt NaNs.  This lets us use a shorter comparison sequence.  */
  if (flag_fast_math)
    target_flags &= ~MASK_IEEE_FP;

  /* It makes no sense to ask for just SSE builtins, so MMX is also turned
     on by -msse.  */
  if (TARGET_SSE)
    target_flags |= MASK_MMX;
}
\f
void
optimization_options (level, size)
     int level;
     int size ATTRIBUTE_UNUSED;
{
  /* For -O2 and beyond, turn off -fschedule-insns by default.  It tends to
     make the problem with not enough registers even worse.  */
#ifdef INSN_SCHEDULING
  if (level > 1)
    flag_schedule_insns = 0;
#endif
}
\f
/* Return nonzero if IDENTIFIER with arguments ARGS is a valid machine specific
   attribute for DECL.  The attributes in ATTRIBUTES have previously been
   assigned to DECL.  */

int
ix86_valid_decl_attribute_p (decl, attributes, identifier, args)
     tree decl ATTRIBUTE_UNUSED;
     tree attributes ATTRIBUTE_UNUSED;
     tree identifier ATTRIBUTE_UNUSED;
     tree args ATTRIBUTE_UNUSED;
{
  return 0;
}

/* Return nonzero if IDENTIFIER with arguments ARGS is a valid machine specific
   attribute for TYPE.  The attributes in ATTRIBUTES have previously been
   assigned to TYPE.  */

int
ix86_valid_type_attribute_p (type, attributes, identifier, args)
     tree type;
     tree attributes ATTRIBUTE_UNUSED;
     tree identifier;
     tree args;
{
  if (TREE_CODE (type) != FUNCTION_TYPE
      && TREE_CODE (type) != METHOD_TYPE
      && TREE_CODE (type) != FIELD_DECL
      && TREE_CODE (type) != TYPE_DECL)
    return 0;

  /* Stdcall attribute says callee is responsible for popping arguments
     if they are not variable.  */
  if (is_attribute_p ("stdcall", identifier))
    return (args == NULL_TREE);

  /* Cdecl attribute says the callee is a normal C declaration.  */
  if (is_attribute_p ("cdecl", identifier))
    return (args == NULL_TREE);

  /* Regparm attribute specifies how many integer arguments are to be
     passed in registers.  */
  if (is_attribute_p ("regparm", identifier))
    {
      tree cst;

      if (! args || TREE_CODE (args) != TREE_LIST
	  || TREE_CHAIN (args) != NULL_TREE
	  || TREE_VALUE (args) == NULL_TREE)
	return 0;

      cst = TREE_VALUE (args);
      if (TREE_CODE (cst) != INTEGER_CST)
	return 0;

      if (compare_tree_int (cst, REGPARM_MAX) > 0)
	return 0;

      return 1;
    }

  return 0;
}

/* Return 0 if the attributes for two types are incompatible, 1 if they
   are compatible, and 2 if they are nearly compatible (which causes a
   warning to be generated).  */

int
ix86_comp_type_attributes (type1, type2)
     tree type1;
     tree type2;
{
  /* Check for mismatch of non-default calling convention.  */
  const char *rtdstr = TARGET_RTD ? "cdecl" : "stdcall";

  if (TREE_CODE (type1) != FUNCTION_TYPE)
    return 1;

  /* Check for mismatched return types (cdecl vs stdcall).  */
  if (!lookup_attribute (rtdstr, TYPE_ATTRIBUTES (type1))
      != !lookup_attribute (rtdstr, TYPE_ATTRIBUTES (type2)))
    return 0;
  return 1;
}
\f
/* Value is the number of bytes of arguments automatically
   popped when returning from a subroutine call.
   FUNDECL is the declaration node of the function (as a tree),
   FUNTYPE is the data type of the function (as a tree),
   or for a library call it is an identifier node for the subroutine name.
   SIZE is the number of bytes of arguments passed on the stack.

   On the 80386, the RTD insn may be used to pop them if the number
     of args is fixed, but if the number is variable then the caller
     must pop them all.  RTD can't be used for library calls now
     because the library is compiled with the Unix compiler.
   Use of RTD is a selectable option, since it is incompatible with
   standard Unix calling sequences.  If the option is not selected,
   the caller must always pop the args.

   The attribute stdcall is equivalent to RTD on a per module basis.  */

int
ix86_return_pops_args (fundecl, funtype, size)
     tree fundecl;
     tree funtype;
     int size;
{
  int rtd = TARGET_RTD && (!fundecl || TREE_CODE (fundecl) != IDENTIFIER_NODE);

    /* Cdecl functions override -mrtd, and never pop the stack.  */
  if (! lookup_attribute ("cdecl", TYPE_ATTRIBUTES (funtype))) {

    /* Stdcall functions will pop the stack if not variable args.  */
    if (lookup_attribute ("stdcall", TYPE_ATTRIBUTES (funtype)))
      rtd = 1;

    if (rtd
        && (TYPE_ARG_TYPES (funtype) == NULL_TREE
	    || (TREE_VALUE (tree_last (TYPE_ARG_TYPES (funtype)))
		== void_type_node)))
      return size;
  }

  /* Lose any fake structure return argument.  */
  if (aggregate_value_p (TREE_TYPE (funtype)))
    return GET_MODE_SIZE (Pmode);

    return 0;
}
\f
/* Argument support functions.  */

/* Initialize a variable CUM of type CUMULATIVE_ARGS
   for a call to a function whose data type is FNTYPE.
   For a library call, FNTYPE is 0.  */

void
init_cumulative_args (cum, fntype, libname)
     CUMULATIVE_ARGS *cum;	/* Argument info to initialize */
     tree fntype;		/* tree ptr for function decl */
     rtx libname;		/* SYMBOL_REF of library name or 0 */
{
  static CUMULATIVE_ARGS zero_cum;
  tree param, next_param;

  if (TARGET_DEBUG_ARG)
    {
      fprintf (stderr, "\ninit_cumulative_args (");
      if (fntype)
	fprintf (stderr, "fntype code = %s, ret code = %s",
		 tree_code_name[(int) TREE_CODE (fntype)],
		 tree_code_name[(int) TREE_CODE (TREE_TYPE (fntype))]);
      else
	fprintf (stderr, "no fntype");

      if (libname)
	fprintf (stderr, ", libname = %s", XSTR (libname, 0));
    }

  *cum = zero_cum;

  /* Set up the number of registers to use for passing arguments.  */
  cum->nregs = ix86_regparm;
  if (fntype)
    {
      tree attr = lookup_attribute ("regparm", TYPE_ATTRIBUTES (fntype));

      if (attr)
	cum->nregs = TREE_INT_CST_LOW (TREE_VALUE (TREE_VALUE (attr)));
    }

  /* Determine if this function has variable arguments.  This is
     indicated by the last argument being 'void_type_mode' if there
     are no variable arguments.  If there are variable arguments, then
     we won't pass anything in registers */

  if (cum->nregs)
    {
      for (param = (fntype) ? TYPE_ARG_TYPES (fntype) : 0;
	   param != 0; param = next_param)
	{
	  next_param = TREE_CHAIN (param);
	  if (next_param == 0 && TREE_VALUE (param) != void_type_node)
	    cum->nregs = 0;
	}
    }

  if (TARGET_DEBUG_ARG)
    fprintf (stderr, ", nregs=%d )\n", cum->nregs);

  return;
}

/* Update the data in CUM to advance over an argument
   of mode MODE and data type TYPE.
   (TYPE is null for libcalls where that information may not be available.)  */

void
function_arg_advance (cum, mode, type, named)
     CUMULATIVE_ARGS *cum;	/* current arg information */
     enum machine_mode mode;	/* current arg mode */
     tree type;			/* type of the argument or 0 if lib support */
     int named;			/* whether or not the argument was named */
{
  int bytes =
    (mode == BLKmode) ? int_size_in_bytes (type) : (int) GET_MODE_SIZE (mode);
  int words = (bytes + UNITS_PER_WORD - 1) / UNITS_PER_WORD;

  if (TARGET_DEBUG_ARG)
    fprintf (stderr,
	     "function_adv (sz=%d, wds=%2d, nregs=%d, mode=%s, named=%d)\n\n",
	     words, cum->words, cum->nregs, GET_MODE_NAME (mode), named);
  if (TARGET_SSE && mode == TImode)
    {
      cum->sse_words += words;
      cum->sse_nregs -= 1;
      cum->sse_regno += 1;
      if (cum->sse_nregs <= 0)
	{
	  cum->sse_nregs = 0;
	  cum->sse_regno = 0;
	}
    }
  else         
    {
      cum->words += words;
      cum->nregs -= words;
      cum->regno += words;

      if (cum->nregs <= 0)
	{
	  cum->nregs = 0;
	  cum->regno = 0;
	}
    }
  return;
}

/* Define where to put the arguments to a function.
   Value is zero to push the argument on the stack,
   or a hard register in which to store the argument.

   MODE is the argument's machine mode.
   TYPE is the data type of the argument (as a tree).
    This is null for libcalls where that information may
    not be available.
   CUM is a variable of type CUMULATIVE_ARGS which gives info about
    the preceding args and about the function being called.
   NAMED is nonzero if this argument is a named parameter
    (otherwise it is an extra parameter matching an ellipsis).  */

struct rtx_def *
function_arg (cum, mode, type, named)
     CUMULATIVE_ARGS *cum;	/* current arg information */
     enum machine_mode mode;	/* current arg mode */
     tree type;			/* type of the argument or 0 if lib support */
     int named;			/* != 0 for normal args, == 0 for ... args */
{
  rtx ret   = NULL_RTX;
  int bytes =
    (mode == BLKmode) ? int_size_in_bytes (type) : (int) GET_MODE_SIZE (mode);
  int words = (bytes + UNITS_PER_WORD - 1) / UNITS_PER_WORD;

  switch (mode)
    {
      /* For now, pass fp/complex values on the stack.  */
    default:
      break;

    case BLKmode:
    case DImode:
    case SImode:
    case HImode:
    case QImode:
      if (words <= cum->nregs)
	ret = gen_rtx_REG (mode, cum->regno);
      break;
    case TImode:
      if (cum->sse_nregs)
        ret = gen_rtx_REG (mode, cum->sse_regno);
      break;    
    }

  if (TARGET_DEBUG_ARG)
    {
      fprintf (stderr,
	       "function_arg (size=%d, wds=%2d, nregs=%d, mode=%4s, named=%d",
	       words, cum->words, cum->nregs, GET_MODE_NAME (mode), named);

      if (ret)
	fprintf (stderr, ", reg=%%e%s", reg_names[ REGNO(ret) ]);
      else
	fprintf (stderr, ", stack");

      fprintf (stderr, " )\n");
    }

  return ret;
}
\f

/* Return nonzero if OP is (const_int 1), else return zero.  */

int
const_int_1_operand (op, mode)
     rtx op;
     enum machine_mode mode ATTRIBUTE_UNUSED;
{
  return (GET_CODE (op) == CONST_INT && INTVAL (op) == 1);
}

/* Returns 1 if OP is either a symbol reference or a sum of a symbol
   reference and a constant.  */

int
symbolic_operand (op, mode)
     register rtx op;
     enum machine_mode mode ATTRIBUTE_UNUSED;
{
  switch (GET_CODE (op))
    {
    case SYMBOL_REF:
    case LABEL_REF:
      return 1;

    case CONST:
      op = XEXP (op, 0);
      if (GET_CODE (op) == SYMBOL_REF
	  || GET_CODE (op) == LABEL_REF
	  || (GET_CODE (op) == UNSPEC
	      && XINT (op, 1) >= 6
	      && XINT (op, 1) <= 7))
	return 1;
      if (GET_CODE (op) != PLUS
	  || GET_CODE (XEXP (op, 1)) != CONST_INT)
	return 0;

      op = XEXP (op, 0);
      if (GET_CODE (op) == SYMBOL_REF
	  || GET_CODE (op) == LABEL_REF)
	return 1;
      /* Only @GOTOFF gets offsets.  */
      if (GET_CODE (op) != UNSPEC
	  || XINT (op, 1) != 7)
	return 0;

      op = XVECEXP (op, 0, 0);
      if (GET_CODE (op) == SYMBOL_REF
	  || GET_CODE (op) == LABEL_REF)
	return 1;
      return 0;

    default:
      return 0;
    }
}

/* Return true if the operand contains a @GOT or @GOTOFF reference.  */

int
pic_symbolic_operand (op, mode)
     register rtx op;
     enum machine_mode mode ATTRIBUTE_UNUSED;
{
  if (GET_CODE (op) == CONST)
    {
      op = XEXP (op, 0);
      if (GET_CODE (op) == UNSPEC)
	return 1;
      if (GET_CODE (op) != PLUS
	  || GET_CODE (XEXP (op, 1)) != CONST_INT)
	return 0;
      op = XEXP (op, 0);
      if (GET_CODE (op) == UNSPEC)
	return 1;
    }
  return 0;
}

/* Test for a valid operand for a call instruction.  Don't allow the
   arg pointer register or virtual regs since they may decay into
   reg + const, which the patterns can't handle.  */

int
call_insn_operand (op, mode)
     rtx op;
     enum machine_mode mode ATTRIBUTE_UNUSED;
{
  /* Disallow indirect through a virtual register.  This leads to
     compiler aborts when trying to eliminate them.  */
  if (GET_CODE (op) == REG
      && (op == arg_pointer_rtx
	  || op == frame_pointer_rtx
	  || (REGNO (op) >= FIRST_PSEUDO_REGISTER
	      && REGNO (op) <= LAST_VIRTUAL_REGISTER)))
    return 0;

  /* Disallow `call 1234'.  Due to varying assembler lameness this
     gets either rejected or translated to `call .+1234'.  */
  if (GET_CODE (op) == CONST_INT)
    return 0;

  /* Explicitly allow SYMBOL_REF even if pic.  */
  if (GET_CODE (op) == SYMBOL_REF)
    return 1;

  /* Half-pic doesn't allow anything but registers and constants.
     We've just taken care of the later.  */
  if (HALF_PIC_P ())
    return register_operand (op, Pmode);

  /* Otherwise we can allow any general_operand in the address.  */
  return general_operand (op, Pmode);
}

int
constant_call_address_operand (op, mode)
     rtx op;
     enum machine_mode mode ATTRIBUTE_UNUSED;
{
  if (GET_CODE (op) == CONST
      && GET_CODE (XEXP (op, 0)) == PLUS
      && GET_CODE (XEXP (XEXP (op, 0), 1)) == CONST_INT)
    op = XEXP (XEXP (op, 0), 0);
  return GET_CODE (op) == SYMBOL_REF;
}

/* Match exactly zero and one.  */

int
const0_operand (op, mode)
     register rtx op;
     enum machine_mode mode;
{
  return op == CONST0_RTX (mode);
}

int
const1_operand (op, mode)
     register rtx op;
     enum machine_mode mode ATTRIBUTE_UNUSED;
{
  return op == const1_rtx;
}

/* Match 2, 4, or 8.  Used for leal multiplicands.  */

int
const248_operand (op, mode)
     register rtx op;
     enum machine_mode mode ATTRIBUTE_UNUSED;
{
  return (GET_CODE (op) == CONST_INT
	  && (INTVAL (op) == 2 || INTVAL (op) == 4 || INTVAL (op) == 8));
}

/* True if this is a constant appropriate for an increment or decremenmt.  */

int
incdec_operand (op, mode)
     register rtx op;
     enum machine_mode mode;
{
  if (op == const1_rtx || op == constm1_rtx)
    return 1;
  if (GET_CODE (op) != CONST_INT)
    return 0;
  if (mode == SImode && INTVAL (op) == (HOST_WIDE_INT) 0xffffffff)
    return 1;
  if (mode == HImode && INTVAL (op) == (HOST_WIDE_INT) 0xffff)
    return 1;
  if (mode == QImode && INTVAL (op) == (HOST_WIDE_INT) 0xff)
    return 1;
  return 0;
}

/* Return false if this is the stack pointer, or any other fake
   register eliminable to the stack pointer.  Otherwise, this is
   a register operand.

   This is used to prevent esp from being used as an index reg.
   Which would only happen in pathological cases.  */

int
reg_no_sp_operand (op, mode)
     register rtx op;
     enum machine_mode mode;
{
  rtx t = op;
  if (GET_CODE (t) == SUBREG)
    t = SUBREG_REG (t);
  if (t == stack_pointer_rtx || t == arg_pointer_rtx || t == frame_pointer_rtx)
    return 0;

  return register_operand (op, mode);
}

int
mmx_reg_operand (op, mode)
     register rtx op;
     enum machine_mode mode ATTRIBUTE_UNUSED;
{
  return MMX_REG_P (op);
}

/* Return false if this is any eliminable register.  Otherwise
   general_operand.  */

int
general_no_elim_operand (op, mode)
     register rtx op;
     enum machine_mode mode;
{
  rtx t = op;
  if (GET_CODE (t) == SUBREG)
    t = SUBREG_REG (t);
  if (t == arg_pointer_rtx || t == frame_pointer_rtx
      || t == virtual_incoming_args_rtx || t == virtual_stack_vars_rtx
      || t == virtual_stack_dynamic_rtx)
    return 0;

  return general_operand (op, mode);
}

/* Return false if this is any eliminable register.  Otherwise
   register_operand or const_int.  */

int
nonmemory_no_elim_operand (op, mode)
     register rtx op;
     enum machine_mode mode;
{
  rtx t = op;
  if (GET_CODE (t) == SUBREG)
    t = SUBREG_REG (t);
  if (t == arg_pointer_rtx || t == frame_pointer_rtx
      || t == virtual_incoming_args_rtx || t == virtual_stack_vars_rtx
      || t == virtual_stack_dynamic_rtx)
    return 0;

  return GET_CODE (op) == CONST_INT || register_operand (op, mode);
}

/* Return true if op is a Q_REGS class register.  */

int
q_regs_operand (op, mode)
     register rtx op;
     enum machine_mode mode;
{
  if (mode != VOIDmode && GET_MODE (op) != mode)
    return 0;
  if (GET_CODE (op) == SUBREG)
    op = SUBREG_REG (op);
  return QI_REG_P (op);
}

/* Return true if op is a NON_Q_REGS class register.  */

int
non_q_regs_operand (op, mode)
     register rtx op;
     enum machine_mode mode;
{
  if (mode != VOIDmode && GET_MODE (op) != mode)
    return 0;
  if (GET_CODE (op) == SUBREG)
    op = SUBREG_REG (op);
  return NON_QI_REG_P (op);
}

/* Return 1 if OP is a comparison that can be used in the CMPSS/CMPPS
   insns.  */
int
sse_comparison_operator (op, mode)
     rtx op;
     enum machine_mode mode ATTRIBUTE_UNUSED;
{
  enum rtx_code code = GET_CODE (op);
  switch (code)
    {
    /* Operations supported directly.  */
    case EQ:
    case LT:
    case LE:
    case UNORDERED:
    case NE:
    case UNGE:
    case UNGT:
    case ORDERED:
      return 1;
    /* These are equivalent to ones above in non-IEEE comparisons.  */
    case UNEQ:
    case UNLT:
    case UNLE:
    case LTGT:
    case GE:
    case GT:
      return !TARGET_IEEE_FP;
    default:
      return 0;
    }
}
/* Return 1 if OP is a valid comparison operator in valid mode.  */
int
ix86_comparison_operator (op, mode)
     register rtx op;
     enum machine_mode mode;
{
  enum machine_mode inmode;
  enum rtx_code code = GET_CODE (op);
  if (mode != VOIDmode && GET_MODE (op) != mode)
    return 0;
  if (GET_RTX_CLASS (code) != '<')
    return 0;
  inmode = GET_MODE (XEXP (op, 0));

  if (inmode == CCFPmode || inmode == CCFPUmode)
    {
      enum rtx_code second_code, bypass_code;
      ix86_fp_comparison_codes (code, &bypass_code, &code, &second_code);
      return (bypass_code == NIL && second_code == NIL);
    }
  switch (code)
    {
    case EQ: case NE:
      return 1;
    case LT: case GE:
      if (inmode == CCmode || inmode == CCGCmode
	  || inmode == CCGOCmode || inmode == CCNOmode)
	return 1;
      return 0;
    case LTU: case GTU: case LEU: case ORDERED: case UNORDERED: case GEU:
      if (inmode == CCmode)
	return 1;
      return 0;
    case GT: case LE:
      if (inmode == CCmode || inmode == CCGCmode || inmode == CCNOmode)
	return 1;
      return 0;
    default:
      return 0;
    }
}

/* Return 1 if OP is a comparison operator that can be issued by fcmov.  */

int
fcmov_comparison_operator (op, mode)
    register rtx op;
    enum machine_mode mode;
{
  enum machine_mode inmode;
  enum rtx_code code = GET_CODE (op);
  if (mode != VOIDmode && GET_MODE (op) != mode)
    return 0;
  if (GET_RTX_CLASS (code) != '<')
    return 0;
  inmode = GET_MODE (XEXP (op, 0));
  if (inmode == CCFPmode || inmode == CCFPUmode)
    {
      enum rtx_code second_code, bypass_code;
      ix86_fp_comparison_codes (code, &bypass_code, &code, &second_code);
      if (bypass_code != NIL || second_code != NIL)
	return 0;
      code = ix86_fp_compare_code_to_integer (code);
    }
  /* i387 supports just limited amount of conditional codes.  */
  switch (code)
    {
    case LTU: case GTU: case LEU: case GEU:
      if (inmode == CCmode || inmode == CCFPmode || inmode == CCFPUmode)
	return 1;
      return 0;
    case ORDERED: case UNORDERED:
    case EQ: case NE:
      return 1;
    default:
      return 0;
    }
}

/* Return 1 if OP is a binary operator that can be promoted to wider mode.  */

int
promotable_binary_operator (op, mode)
     register rtx op;
     enum machine_mode mode ATTRIBUTE_UNUSED;
{
  switch (GET_CODE (op))
    {
    case MULT:
      /* Modern CPUs have same latency for HImode and SImode multiply,
         but 386 and 486 do HImode multiply faster.  */
      return ix86_cpu > PROCESSOR_I486;
    case PLUS:
    case AND:
    case IOR:
    case XOR:
    case ASHIFT:
      return 1;
    default:
      return 0;
    }
}

/* Nearly general operand, but accept any const_double, since we wish
   to be able to drop them into memory rather than have them get pulled
   into registers.  */

int
cmp_fp_expander_operand (op, mode)
     register rtx op;
     enum machine_mode mode;
{
  if (mode != VOIDmode && mode != GET_MODE (op))
    return 0;
  if (GET_CODE (op) == CONST_DOUBLE)
    return 1;
  return general_operand (op, mode);
}

/* Match an SI or HImode register for a zero_extract.  */

int
ext_register_operand (op, mode)
     register rtx op;
     enum machine_mode mode ATTRIBUTE_UNUSED;
{
  if (GET_MODE (op) != SImode && GET_MODE (op) != HImode)
    return 0;
  return register_operand (op, VOIDmode);
}

/* Return 1 if this is a valid binary floating-point operation.
   OP is the expression matched, and MODE is its mode.  */

int
binary_fp_operator (op, mode)
    register rtx op;
    enum machine_mode mode;
{
  if (mode != VOIDmode && mode != GET_MODE (op))
    return 0;

  switch (GET_CODE (op))
    {
    case PLUS:
    case MINUS:
    case MULT:
    case DIV:
      return GET_MODE_CLASS (GET_MODE (op)) == MODE_FLOAT;

    default:
      return 0;
    }
}

int
mult_operator(op, mode)
    register rtx op;
    enum machine_mode mode ATTRIBUTE_UNUSED;
{
  return GET_CODE (op) == MULT;
}

int
div_operator(op, mode)
    register rtx op;
    enum machine_mode mode ATTRIBUTE_UNUSED;
{
  return GET_CODE (op) == DIV;
}

int
arith_or_logical_operator (op, mode)
      rtx op;
      enum machine_mode mode;
{
  return ((mode == VOIDmode || GET_MODE (op) == mode)
          && (GET_RTX_CLASS (GET_CODE (op)) == 'c'
              || GET_RTX_CLASS (GET_CODE (op)) == '2'));
}

/* Returns 1 if OP is memory operand with a displacement.  */

int
memory_displacement_operand (op, mode)
     register rtx op;
     enum machine_mode mode;
{
  struct ix86_address parts;

  if (! memory_operand (op, mode))
    return 0;

  if (! ix86_decompose_address (XEXP (op, 0), &parts))
    abort ();

  return parts.disp != NULL_RTX;
}

/* To avoid problems when jump re-emits comparisons like testqi_ext_ccno_0,
   re-recognize the operand to avoid a copy_to_mode_reg that will fail.

   ??? It seems likely that this will only work because cmpsi is an
   expander, and no actual insns use this.  */

int
cmpsi_operand (op, mode)
      rtx op;
      enum machine_mode mode;
{
  if (general_operand (op, mode))
    return 1;

  if (GET_CODE (op) == AND
      && GET_MODE (op) == SImode
      && GET_CODE (XEXP (op, 0)) == ZERO_EXTRACT
      && GET_CODE (XEXP (XEXP (op, 0), 1)) == CONST_INT
      && GET_CODE (XEXP (XEXP (op, 0), 2)) == CONST_INT
      && INTVAL (XEXP (XEXP (op, 0), 1)) == 8
      && INTVAL (XEXP (XEXP (op, 0), 2)) == 8
      && GET_CODE (XEXP (op, 1)) == CONST_INT)
    return 1;

  return 0;
}

/* Returns 1 if OP is memory operand that can not be represented by the
   modRM array.  */

int
long_memory_operand (op, mode)
     register rtx op;
     enum machine_mode mode;
{
  if (! memory_operand (op, mode))
    return 0;

  return memory_address_length (op) != 0;
}

/* Return nonzero if the rtx is known aligned.  */

int
aligned_operand (op, mode)
     rtx op;
     enum machine_mode mode;
{
  struct ix86_address parts;

  if (!general_operand (op, mode))
    return 0;

  /* Registers and immediate operands are always "aligned".  */
  if (GET_CODE (op) != MEM)
    return 1;

  /* Don't even try to do any aligned optimizations with volatiles.  */
  if (MEM_VOLATILE_P (op))
    return 0;

  op = XEXP (op, 0);

  /* Pushes and pops are only valid on the stack pointer.  */
  if (GET_CODE (op) == PRE_DEC
      || GET_CODE (op) == POST_INC)
    return 1;

  /* Decode the address.  */
  if (! ix86_decompose_address (op, &parts))
    abort ();

  /* Look for some component that isn't known to be aligned.  */
  if (parts.index)
    {
      if (parts.scale < 4
	  && REGNO_POINTER_ALIGN (REGNO (parts.index)) < 32)
	return 0;
    }
  if (parts.base)
    {
      if (REGNO_POINTER_ALIGN (REGNO (parts.base)) < 32)
	return 0;
    }
  if (parts.disp)
    {
      if (GET_CODE (parts.disp) != CONST_INT
	  || (INTVAL (parts.disp) & 3) != 0)
	return 0;
    }

  /* Didn't find one -- this must be an aligned address.  */
  return 1;
}
\f
/* Return true if the constant is something that can be loaded with
   a special instruction.  Only handle 0.0 and 1.0; others are less
   worthwhile.  */

int
standard_80387_constant_p (x)
     rtx x;
{
  if (GET_CODE (x) != CONST_DOUBLE || !FLOAT_MODE_P (GET_MODE (x)))
    return -1;
  /* Note that on the 80387, other constants, such as pi, that we should support
     too.  On some machines, these are much slower to load as standard constant,
     than to load from doubles in memory.  */
  if (x == CONST0_RTX (GET_MODE (x)))
    return 1;
  if (x == CONST1_RTX (GET_MODE (x)))
    return 2;
  return 0;
}

/* Return 1 if X is FP constant we can load to SSE register w/o using memory.
 */
int
standard_sse_constant_p (x)
     rtx x;
{
  if (GET_CODE (x) != CONST_DOUBLE)
    return -1;
  return (x == CONST0_RTX (GET_MODE (x)));
}

/* Returns 1 if OP contains a symbol reference */

int
symbolic_reference_mentioned_p (op)
     rtx op;
{
  register const char *fmt;
  register int i;

  if (GET_CODE (op) == SYMBOL_REF || GET_CODE (op) == LABEL_REF)
    return 1;

  fmt = GET_RTX_FORMAT (GET_CODE (op));
  for (i = GET_RTX_LENGTH (GET_CODE (op)) - 1; i >= 0; i--)
    {
      if (fmt[i] == 'E')
	{
	  register int j;

	  for (j = XVECLEN (op, i) - 1; j >= 0; j--)
	    if (symbolic_reference_mentioned_p (XVECEXP (op, i, j)))
	      return 1;
	}

      else if (fmt[i] == 'e' && symbolic_reference_mentioned_p (XEXP (op, i)))
	return 1;
    }

  return 0;
}

/* Return 1 if it is appropriate to emit `ret' instructions in the
   body of a function.  Do this only if the epilogue is simple, needing a
   couple of insns.  Prior to reloading, we can't tell how many registers
   must be saved, so return 0 then.  Return 0 if there is no frame
   marker to de-allocate.

   If NON_SAVING_SETJMP is defined and true, then it is not possible
   for the epilogue to be simple, so return 0.  This is a special case
   since NON_SAVING_SETJMP will not cause regs_ever_live to change
   until final, but jump_optimize may need to know sooner if a
   `return' is OK.  */

int
ix86_can_use_return_insn_p ()
{
  struct ix86_frame frame;

#ifdef NON_SAVING_SETJMP
  if (NON_SAVING_SETJMP && current_function_calls_setjmp)
    return 0;
#endif
#ifdef FUNCTION_BLOCK_PROFILER_EXIT
  if (profile_block_flag == 2)
    return 0;
#endif

  if (! reload_completed || frame_pointer_needed)
    return 0;

  /* Don't allow more than 32 pop, since that's all we can do
     with one instruction.  */
  if (current_function_pops_args
      && current_function_args_size >= 32768)
    return 0;

  ix86_compute_frame_layout (&frame);
  return frame.to_allocate == 0 && frame.nregs == 0;
}

/* Value should be nonzero if functions must have frame pointers.
   Zero means the frame pointer need not be set up (and parms may
   be accessed via the stack pointer) in functions that seem suitable.  */

int
ix86_frame_pointer_required ()
{
  /* If we accessed previous frames, then the generated code expects
     to be able to access the saved ebp value in our frame.  */
  if (cfun->machine->accesses_prev_frame)
    return 1;
  
  /* Several x86 os'es need a frame pointer for other reasons,
     usually pertaining to setjmp.  */
  if (SUBTARGET_FRAME_POINTER_REQUIRED)
    return 1;

  /* In override_options, TARGET_OMIT_LEAF_FRAME_POINTER turns off
     the frame pointer by default.  Turn it back on now if we've not
     got a leaf function.  */
  if (TARGET_OMIT_LEAF_FRAME_POINTER && ! leaf_function_p ())
    return 1;

  return 0;
}

/* Record that the current function accesses previous call frames.  */

void
ix86_setup_frame_addresses ()
{
  cfun->machine->accesses_prev_frame = 1;
}
\f
static char pic_label_name[32];

/* This function generates code for -fpic that loads %ebx with
   the return address of the caller and then returns.  */

void
ix86_asm_file_end (file)
     FILE *file;
{
  rtx xops[2];

  if (! TARGET_DEEP_BRANCH_PREDICTION || pic_label_name[0] == 0)
    return;

  /* ??? Binutils 2.10 and earlier has a linkonce elimination bug related
     to updating relocations to a section being discarded such that this
     doesn't work.  Ought to detect this at configure time.  */
#if 0 && defined (ASM_OUTPUT_SECTION_NAME)
  /* The trick here is to create a linkonce section containing the
     pic label thunk, but to refer to it with an internal label.
     Because the label is internal, we don't have inter-dso name
     binding issues on hosts that don't support ".hidden".

     In order to use these macros, however, we must create a fake
     function decl.  */
  {
    tree decl = build_decl (FUNCTION_DECL,
			    get_identifier ("i686.get_pc_thunk"),
			    error_mark_node);
    DECL_ONE_ONLY (decl) = 1;
    UNIQUE_SECTION (decl, 0);
    named_section (decl, NULL, 0);
  }
#else
  text_section ();
#endif

  /* This used to call ASM_DECLARE_FUNCTION_NAME() but since it's an
     internal (non-global) label that's being emitted, it didn't make
     sense to have .type information for local labels.   This caused
     the SCO OpenServer 5.0.4 ELF assembler grief (why are you giving
     me debug info for a label that you're declaring non-global?) this
     was changed to call ASM_OUTPUT_LABEL() instead.  */

  ASM_OUTPUT_LABEL (file, pic_label_name);

  xops[0] = pic_offset_table_rtx;
  xops[1] = gen_rtx_MEM (SImode, stack_pointer_rtx);
  output_asm_insn ("mov{l}\t{%1, %0|%0, %1}", xops);
  output_asm_insn ("ret", xops);
}

void
load_pic_register ()
{
  rtx gotsym, pclab;

  gotsym = gen_rtx_SYMBOL_REF (Pmode, "_GLOBAL_OFFSET_TABLE_");

  if (TARGET_DEEP_BRANCH_PREDICTION)
    {
      if (! pic_label_name[0])
	ASM_GENERATE_INTERNAL_LABEL (pic_label_name, "LPR", 0);
      pclab = gen_rtx_MEM (QImode, gen_rtx_SYMBOL_REF (Pmode, pic_label_name));
    }
  else
    {
      pclab = gen_rtx_LABEL_REF (VOIDmode, gen_label_rtx ());
    }

  emit_insn (gen_prologue_get_pc (pic_offset_table_rtx, pclab));

  if (! TARGET_DEEP_BRANCH_PREDICTION)
    emit_insn (gen_popsi1 (pic_offset_table_rtx));

  emit_insn (gen_prologue_set_got (pic_offset_table_rtx, gotsym, pclab));
}

/* Generate an SImode "push" pattern for input ARG.  */

static rtx
gen_push (arg)
     rtx arg;
{
  return gen_rtx_SET (VOIDmode,
		      gen_rtx_MEM (SImode,
				   gen_rtx_PRE_DEC (SImode,
						    stack_pointer_rtx)),
		      arg);
}

/* Return 1 if we need to save REGNO.  */
static int
ix86_save_reg (regno)
	int regno;
{
  int pic_reg_used = flag_pic && (current_function_uses_pic_offset_table
				  || current_function_uses_const_pool);
  return ((regs_ever_live[regno] && !call_used_regs[regno]
	   && !fixed_regs[regno]
	   && (regno != HARD_FRAME_POINTER_REGNUM || !frame_pointer_needed))
	  || (regno == PIC_OFFSET_TABLE_REGNUM && pic_reg_used));

}

/* Return number of registers to be saved on the stack.  */

static int
ix86_nsaved_regs ()
{
  int nregs = 0;
  int regno;

  for (regno = FIRST_PSEUDO_REGISTER - 1; regno >= 0; regno--)
    if (ix86_save_reg (regno))
      nregs++;
  return nregs;
}

/* Return the offset between two registers, one to be eliminated, and the other
   its replacement, at the start of a routine.  */

HOST_WIDE_INT
ix86_initial_elimination_offset (from, to)
     int from;
     int to;
{
  struct ix86_frame frame;
  ix86_compute_frame_layout (&frame);

  if (from == ARG_POINTER_REGNUM && to == HARD_FRAME_POINTER_REGNUM)
    return frame.hard_frame_pointer_offset;
  else if (from == FRAME_POINTER_REGNUM
	   && to == HARD_FRAME_POINTER_REGNUM)
    return frame.hard_frame_pointer_offset - frame.frame_pointer_offset;
  else
    {
      if (to != STACK_POINTER_REGNUM)
	abort ();
      else if (from == ARG_POINTER_REGNUM)
	return frame.stack_pointer_offset;
      else if (from != FRAME_POINTER_REGNUM)
	abort ();
      else
	return frame.stack_pointer_offset - frame.frame_pointer_offset;
    }
}

/* Fill structure ix86_frame about frame of currently computed function.  */

static void
ix86_compute_frame_layout (frame)
     struct ix86_frame *frame;
{
  HOST_WIDE_INT total_size;
  int stack_alignment_needed = cfun->stack_alignment_needed / BITS_PER_UNIT;
  int offset;
  int preferred_alignment = cfun->preferred_stack_boundary / BITS_PER_UNIT;
  HOST_WIDE_INT size = get_frame_size ();

  frame->nregs = ix86_nsaved_regs ();
  total_size = size;

  /* Skip return value and save base pointer.  */
  offset = frame_pointer_needed ? UNITS_PER_WORD * 2 : UNITS_PER_WORD;

  frame->hard_frame_pointer_offset = offset;

  /* Do some sanity checking of stack_alignment_needed and
     preferred_alignment, since i386 port is the only using those features
     that may break easilly.  */

  if (size && !stack_alignment_needed)
    abort ();
  if (preferred_alignment < STACK_BOUNDARY / BITS_PER_UNIT)
    abort ();
  if (preferred_alignment > PREFERRED_STACK_BOUNDARY / BITS_PER_UNIT)
    abort ();
  if (stack_alignment_needed > PREFERRED_STACK_BOUNDARY / BITS_PER_UNIT)
    abort ();

  if (stack_alignment_needed < STACK_BOUNDARY / BITS_PER_UNIT)
    stack_alignment_needed = STACK_BOUNDARY / BITS_PER_UNIT;

  /* Register save area */
  offset += frame->nregs * UNITS_PER_WORD;

  /* Align start of frame for local function.  */
  frame->padding1 = ((offset + stack_alignment_needed - 1)
		     & -stack_alignment_needed) - offset;

  offset += frame->padding1;

  /* Frame pointer points here.  */
  frame->frame_pointer_offset = offset;

  offset += size;

  /* Add outgoing arguments area.  */
  if (ACCUMULATE_OUTGOING_ARGS)
    {
      offset += current_function_outgoing_args_size;
      frame->outgoing_arguments_size = current_function_outgoing_args_size;
    }
  else
    frame->outgoing_arguments_size = 0;

  /* Align stack boundary.  */
  frame->padding2 = ((offset + preferred_alignment - 1)
		     & -preferred_alignment) - offset;

  offset += frame->padding2;

  /* We've reached end of stack frame.  */
  frame->stack_pointer_offset = offset;

  /* Size prologue needs to allocate.  */
  frame->to_allocate =
    (size + frame->padding1 + frame->padding2
     + frame->outgoing_arguments_size);

#if 0
  fprintf (stderr, "nregs: %i\n", frame->nregs);
  fprintf (stderr, "size: %i\n", size);
  fprintf (stderr, "alignment1: %i\n", stack_alignment_needed);
  fprintf (stderr, "padding1: %i\n", frame->padding1);
  fprintf (stderr, "padding2: %i\n", frame->padding2);
  fprintf (stderr, "to_allocate: %i\n", frame->to_allocate);
  fprintf (stderr, "frame_pointer_offset: %i\n", frame->frame_pointer_offset);
  fprintf (stderr, "hard_frame_pointer_offset: %i\n",
	   frame->hard_frame_pointer_offset);
  fprintf (stderr, "stack_pointer_offset: %i\n", frame->stack_pointer_offset);
#endif
}

/* Emit code to save registers in the prologue.  */

static void
ix86_emit_save_regs ()
{
  register int regno;
  rtx insn;

  for (regno = FIRST_PSEUDO_REGISTER - 1; regno >= 0; regno--)
    if (ix86_save_reg (regno))
      {
	insn = emit_insn (gen_push (gen_rtx_REG (SImode, regno)));
	RTX_FRAME_RELATED_P (insn) = 1;
      }
}

/* Expand the prologue into a bunch of separate insns.  */

void
ix86_expand_prologue ()
{
  rtx insn;
  int pic_reg_used = flag_pic && (current_function_uses_pic_offset_table
				  || current_function_uses_const_pool);
  struct ix86_frame frame;

  ix86_compute_frame_layout (&frame);

  /* Note: AT&T enter does NOT have reversed args.  Enter is probably
     slower on all targets.  Also sdb doesn't like it.  */

  if (frame_pointer_needed)
    {
      insn = emit_insn (gen_push (hard_frame_pointer_rtx));
      RTX_FRAME_RELATED_P (insn) = 1;

      insn = emit_move_insn (hard_frame_pointer_rtx, stack_pointer_rtx);
      RTX_FRAME_RELATED_P (insn) = 1;
    }

  ix86_emit_save_regs ();

  if (frame.to_allocate == 0)
    ;
  else if (! TARGET_STACK_PROBE || frame.to_allocate < CHECK_STACK_LIMIT)
    {
      if (frame_pointer_needed)
	insn = emit_insn (gen_pro_epilogue_adjust_stack
			  (stack_pointer_rtx, stack_pointer_rtx,
		           GEN_INT (-frame.to_allocate), hard_frame_pointer_rtx));
      else
        insn = emit_insn (gen_addsi3 (stack_pointer_rtx, stack_pointer_rtx,
				      GEN_INT (-frame.to_allocate)));
      RTX_FRAME_RELATED_P (insn) = 1;
    }
  else
    {
      /* ??? Is this only valid for Win32?  */

      rtx arg0, sym;

      arg0 = gen_rtx_REG (SImode, 0);
      emit_move_insn (arg0, GEN_INT (frame.to_allocate));

      sym = gen_rtx_MEM (FUNCTION_MODE,
			 gen_rtx_SYMBOL_REF (Pmode, "_alloca"));
      insn = emit_call_insn (gen_call (sym, const0_rtx));

      CALL_INSN_FUNCTION_USAGE (insn)
	= gen_rtx_EXPR_LIST (VOIDmode, gen_rtx_USE (VOIDmode, arg0),
			     CALL_INSN_FUNCTION_USAGE (insn));
    }

#ifdef SUBTARGET_PROLOGUE
  SUBTARGET_PROLOGUE;
#endif

  if (pic_reg_used)
    load_pic_register ();

  /* If we are profiling, make sure no instructions are scheduled before
     the call to mcount.  However, if -fpic, the above call will have
     done that.  */
  if ((profile_flag || profile_block_flag) && ! pic_reg_used)
    emit_insn (gen_blockage ());
}

/* Emit code to add TSIZE to esp value.  Use POP instruction when
   profitable.  */

static void
ix86_emit_epilogue_esp_adjustment (tsize)
     int tsize;
{
  /* If a frame pointer is present, we must be sure to tie the sp
     to the fp so that we don't mis-schedule.  */
  if (frame_pointer_needed)
    emit_insn (gen_pro_epilogue_adjust_stack (stack_pointer_rtx,
					      stack_pointer_rtx,
					      GEN_INT (tsize),
					      hard_frame_pointer_rtx));
  else
    emit_insn (gen_addsi3 (stack_pointer_rtx, stack_pointer_rtx,
			   GEN_INT (tsize)));
}

/* Emit code to restore saved registers using MOV insns.  First register
   is restored from POINTER + OFFSET.  */
static void
ix86_emit_restore_regs_using_mov (pointer, offset)
	rtx pointer;
	int offset;
{
  int regno;

  for (regno = 0; regno < FIRST_PSEUDO_REGISTER; regno++)
    if (ix86_save_reg (regno))
      {
	emit_move_insn (gen_rtx_REG (Pmode, regno),
			adj_offsettable_operand (gen_rtx_MEM (Pmode,
							      pointer),
						 offset));
	offset += UNITS_PER_WORD;
      }
}

/* Restore function stack, frame, and registers.  */

void
ix86_expand_epilogue (emit_return)
     int emit_return;
{
  int regno;
  int sp_valid = !frame_pointer_needed || current_function_sp_is_unchanging;
  struct ix86_frame frame;
  HOST_WIDE_INT offset;

  ix86_compute_frame_layout (&frame);

  /* Calculate start of saved registers relative to ebp.  */
  offset = -frame.nregs * UNITS_PER_WORD;

#ifdef FUNCTION_BLOCK_PROFILER_EXIT
  if (profile_block_flag == 2)
    {
      FUNCTION_BLOCK_PROFILER_EXIT;
    }
#endif

  /* If we're only restoring one register and sp is not valid then
     using a move instruction to restore the register since it's
     less work than reloading sp and popping the register.

     The default code result in stack adjustment using add/lea instruction,
     while this code results in LEAVE instruction (or discrete equivalent),
     so it is profitable in some other cases as well.  Especially when there
     are no registers to restore.  We also use this code when TARGET_USE_LEAVE
     and there is exactly one register to pop. This heruistic may need some
     tuning in future.  */
  if ((!sp_valid && frame.nregs <= 1)
      || (frame_pointer_needed && !frame.nregs && frame.to_allocate)
      || (frame_pointer_needed && TARGET_USE_LEAVE && !optimize_size
	  && frame.nregs == 1))
    {
      /* Restore registers.  We can use ebp or esp to address the memory
	 locations.  If both are available, default to ebp, since offsets
	 are known to be small.  Only exception is esp pointing directly to the
	 end of block of saved registers, where we may simplify addressing
	 mode.  */

      if (!frame_pointer_needed || (sp_valid && !frame.to_allocate))
	ix86_emit_restore_regs_using_mov (stack_pointer_rtx, frame.to_allocate);
      else
	ix86_emit_restore_regs_using_mov (hard_frame_pointer_rtx, offset);

      if (!frame_pointer_needed)
	ix86_emit_epilogue_esp_adjustment (frame.to_allocate
					   + frame.nregs * UNITS_PER_WORD);
      /* If not an i386, mov & pop is faster than "leave".  */
      else if (TARGET_USE_LEAVE || optimize_size)
	emit_insn (gen_leave ());
      else
	{
	  emit_insn (gen_pro_epilogue_adjust_stack (stack_pointer_rtx,
						    hard_frame_pointer_rtx,
						    const0_rtx,
						    hard_frame_pointer_rtx));
	  emit_insn (gen_popsi1 (hard_frame_pointer_rtx));
	}
    }
  else
    {
      /* First step is to deallocate the stack frame so that we can
	 pop the registers.  */
      if (!sp_valid)
	{
	  if (!frame_pointer_needed)
	    abort ();
          emit_insn (gen_pro_epilogue_adjust_stack (stack_pointer_rtx,
						    hard_frame_pointer_rtx,
						    GEN_INT (offset),
						    hard_frame_pointer_rtx));
	}
      else if (frame.to_allocate)
	ix86_emit_epilogue_esp_adjustment (frame.to_allocate);

      for (regno = 0; regno < FIRST_PSEUDO_REGISTER; regno++)
	if (ix86_save_reg (regno))
	  emit_insn (gen_popsi1 (gen_rtx_REG (SImode, regno)));
      if (frame_pointer_needed)
	emit_insn (gen_popsi1 (hard_frame_pointer_rtx));
    }

  /* Sibcall epilogues don't want a return instruction.  */
  if (! emit_return)
    return;

  if (current_function_pops_args && current_function_args_size)
    {
      rtx popc = GEN_INT (current_function_pops_args);

      /* i386 can only pop 64K bytes.  If asked to pop more, pop
	 return address, do explicit add, and jump indirectly to the
	 caller.  */

      if (current_function_pops_args >= 65536)
	{
	  rtx ecx = gen_rtx_REG (SImode, 2);

	  emit_insn (gen_popsi1 (ecx));
	  emit_insn (gen_addsi3 (stack_pointer_rtx, stack_pointer_rtx, popc));
	  emit_jump_insn (gen_return_indirect_internal (ecx));
	}
      else
	emit_jump_insn (gen_return_pop_internal (popc));
    }
  else
    emit_jump_insn (gen_return_internal ());
}
\f
/* Extract the parts of an RTL expression that is a valid memory address
   for an instruction.  Return false if the structure of the address is
   grossly off.  */

static int
ix86_decompose_address (addr, out)
     register rtx addr;
     struct ix86_address *out;
{
  rtx base = NULL_RTX;
  rtx index = NULL_RTX;
  rtx disp = NULL_RTX;
  HOST_WIDE_INT scale = 1;
  rtx scale_rtx = NULL_RTX;

  if (GET_CODE (addr) == REG || GET_CODE (addr) == SUBREG)
    base = addr;
  else if (GET_CODE (addr) == PLUS)
    {
      rtx op0 = XEXP (addr, 0);
      rtx op1 = XEXP (addr, 1);
      enum rtx_code code0 = GET_CODE (op0);
      enum rtx_code code1 = GET_CODE (op1);

      if (code0 == REG || code0 == SUBREG)
	{
	  if (code1 == REG || code1 == SUBREG)
	    index = op0, base = op1;	/* index + base */
	  else
	    base = op0, disp = op1;	/* base + displacement */
	}
      else if (code0 == MULT)
	{
	  index = XEXP (op0, 0);
	  scale_rtx = XEXP (op0, 1);
	  if (code1 == REG || code1 == SUBREG)
	    base = op1;			/* index*scale + base */
	  else
	    disp = op1;			/* index*scale + disp */
	}
      else if (code0 == PLUS && GET_CODE (XEXP (op0, 0)) == MULT)
	{
	  index = XEXP (XEXP (op0, 0), 0);	/* index*scale + base + disp */
	  scale_rtx = XEXP (XEXP (op0, 0), 1);
	  base = XEXP (op0, 1);
	  disp = op1;
	}
      else if (code0 == PLUS)
	{
	  index = XEXP (op0, 0);	/* index + base + disp */
	  base = XEXP (op0, 1);
	  disp = op1;
	}
      else
	return FALSE;
    }
  else if (GET_CODE (addr) == MULT)
    {
      index = XEXP (addr, 0);		/* index*scale */
      scale_rtx = XEXP (addr, 1);
    }
  else if (GET_CODE (addr) == ASHIFT)
    {
      rtx tmp;

      /* We're called for lea too, which implements ashift on occasion.  */
      index = XEXP (addr, 0);
      tmp = XEXP (addr, 1);
      if (GET_CODE (tmp) != CONST_INT)
	return FALSE;
      scale = INTVAL (tmp);
      if ((unsigned HOST_WIDE_INT) scale > 3)
	return FALSE;
      scale = 1 << scale;
    }
  else
    disp = addr;			/* displacement */

  /* Extract the integral value of scale.  */
  if (scale_rtx)
    {
      if (GET_CODE (scale_rtx) != CONST_INT)
	return FALSE;
      scale = INTVAL (scale_rtx);
    }

  /* Allow arg pointer and stack pointer as index if there is not scaling */
  if (base && index && scale == 1
      && (index == arg_pointer_rtx || index == frame_pointer_rtx
          || index == stack_pointer_rtx))
    {
      rtx tmp = base;
      base = index;
      index = tmp;
    }

  /* Special case: %ebp cannot be encoded as a base without a displacement.  */
  if ((base == hard_frame_pointer_rtx
       || base == frame_pointer_rtx
       || base == arg_pointer_rtx) && !disp)
    disp = const0_rtx;

  /* Special case: on K6, [%esi] makes the instruction vector decoded.
     Avoid this by transforming to [%esi+0].  */
  if (ix86_cpu == PROCESSOR_K6 && !optimize_size
      && base && !index && !disp
      && REG_P (base)
      && REGNO_REG_CLASS (REGNO (base)) == SIREG)
    disp = const0_rtx;

  /* Special case: encode reg+reg instead of reg*2.  */
  if (!base && index && scale && scale == 2)
    base = index, scale = 1;

  /* Special case: scaling cannot be encoded without base or displacement.  */
  if (!base && !disp && index && scale != 1)
    disp = const0_rtx;

  out->base = base;
  out->index = index;
  out->disp = disp;
  out->scale = scale;

  return TRUE;
}
\f
/* Return cost of the memory address x.
   For i386, it is better to use a complex address than let gcc copy
   the address into a reg and make a new pseudo.  But not if the address
   requires to two regs - that would mean more pseudos with longer
   lifetimes.  */
int
ix86_address_cost (x)
     rtx x;
{
  struct ix86_address parts;
  int cost = 1;

  if (!ix86_decompose_address (x, &parts))
    abort ();

  /* More complex memory references are better.  */
  if (parts.disp && parts.disp != const0_rtx)
    cost--;

  /* Attempt to minimize number of registers in the address.  */
  if ((parts.base
       && (!REG_P (parts.base) || REGNO (parts.base) >= FIRST_PSEUDO_REGISTER))
      || (parts.index
	  && (!REG_P (parts.index)
	      || REGNO (parts.index) >= FIRST_PSEUDO_REGISTER)))
    cost++;

  if (parts.base
      && (!REG_P (parts.base) || REGNO (parts.base) >= FIRST_PSEUDO_REGISTER)
      && parts.index
      && (!REG_P (parts.index) || REGNO (parts.index) >= FIRST_PSEUDO_REGISTER)
      && parts.base != parts.index)
    cost++;

  /* AMD-K6 don't like addresses with ModR/M set to 00_xxx_100b,
     since it's predecode logic can't detect the length of instructions
     and it degenerates to vector decoded.  Increase cost of such
     addresses here.  The penalty is minimally 2 cycles.  It may be worthwhile
     to split such addresses or even refuse such addresses at all.

     Following addressing modes are affected:
      [base+scale*index]
      [scale*index+disp]
      [base+index]

     The first and last case  may be avoidable by explicitly coding the zero in
     memory address, but I don't have AMD-K6 machine handy to check this
     theory.  */

  if (TARGET_K6
      && ((!parts.disp && parts.base && parts.index && parts.scale != 1)
	  || (parts.disp && !parts.base && parts.index && parts.scale != 1)
	  || (!parts.disp && parts.base && parts.index && parts.scale == 1)))
    cost += 10;

  return cost;
}
\f
/* If X is a machine specific address (i.e. a symbol or label being
   referenced as a displacement from the GOT implemented using an
   UNSPEC), then return the base term.  Otherwise return X.  */

rtx
ix86_find_base_term (x)
     rtx x;
{
  rtx term;

  if (GET_CODE (x) != PLUS
      || XEXP (x, 0) != pic_offset_table_rtx
      || GET_CODE (XEXP (x, 1)) != CONST)
    return x;

  term = XEXP (XEXP (x, 1), 0);

  if (GET_CODE (term) == PLUS && GET_CODE (XEXP (term, 1)) == CONST_INT)
    term = XEXP (term, 0);

  if (GET_CODE (term) != UNSPEC
      || XVECLEN (term, 0) != 1
      || XINT (term, 1) !=  7)
    return x;

  term = XVECEXP (term, 0, 0);

  if (GET_CODE (term) != SYMBOL_REF
      && GET_CODE (term) != LABEL_REF)
    return x;

  return term;
}
\f
/* Determine if a given CONST RTX is a valid memory displacement
   in PIC mode.  */

int
legitimate_pic_address_disp_p (disp)
     register rtx disp;
{
  if (GET_CODE (disp) != CONST)
    return 0;
  disp = XEXP (disp, 0);

  if (GET_CODE (disp) == PLUS)
    {
      if (GET_CODE (XEXP (disp, 1)) != CONST_INT)
	return 0;
      disp = XEXP (disp, 0);
    }

  if (GET_CODE (disp) != UNSPEC
      || XVECLEN (disp, 0) != 1)
    return 0;

  /* Must be @GOT or @GOTOFF.  */
  if (XINT (disp, 1) != 6
      && XINT (disp, 1) != 7)
    return 0;

  if (GET_CODE (XVECEXP (disp, 0, 0)) != SYMBOL_REF
      && GET_CODE (XVECEXP (disp, 0, 0)) != LABEL_REF)
    return 0;

  return 1;
}

/* GO_IF_LEGITIMATE_ADDRESS recognizes an RTL expression that is a valid
   memory address for an instruction.  The MODE argument is the machine mode
   for the MEM expression that wants to use this address.

   It only recognizes address in canonical form.  LEGITIMIZE_ADDRESS should
   convert common non-canonical forms to canonical form so that they will
   be recognized.  */

int
legitimate_address_p (mode, addr, strict)
     enum machine_mode mode;
     register rtx addr;
     int strict;
{
  struct ix86_address parts;
  rtx base, index, disp;
  HOST_WIDE_INT scale;
  const char *reason = NULL;
  rtx reason_rtx = NULL_RTX;

  if (TARGET_DEBUG_ADDR)
    {
      fprintf (stderr,
	       "\n======\nGO_IF_LEGITIMATE_ADDRESS, mode = %s, strict = %d\n",
	       GET_MODE_NAME (mode), strict);
      debug_rtx (addr);
    }

  if (! ix86_decompose_address (addr, &parts))
    {
      reason = "decomposition failed";
      goto report_error;
    }

  base = parts.base;
  index = parts.index;
  disp = parts.disp;
  scale = parts.scale;

  /* Validate base register.

     Don't allow SUBREG's here, it can lead to spill failures when the base
     is one word out of a two word structure, which is represented internally
     as a DImode int.  */

  if (base)
    {
      reason_rtx = base;

      if (GET_CODE (base) != REG)
	{
	  reason = "base is not a register";
	  goto report_error;
	}

      if (GET_MODE (base) != Pmode)
	{
	  reason = "base is not in Pmode";
	  goto report_error;
	}

      if ((strict && ! REG_OK_FOR_BASE_STRICT_P (base))
	  || (! strict && ! REG_OK_FOR_BASE_NONSTRICT_P (base)))
	{
	  reason = "base is not valid";
	  goto report_error;
	}
    }

  /* Validate index register.

     Don't allow SUBREG's here, it can lead to spill failures when the index
     is one word out of a two word structure, which is represented internally
     as a DImode int.  */

  if (index)
    {
      reason_rtx = index;

      if (GET_CODE (index) != REG)
	{
	  reason = "index is not a register";
	  goto report_error;
	}

      if (GET_MODE (index) != Pmode)
	{
	  reason = "index is not in Pmode";
	  goto report_error;
	}

      if ((strict && ! REG_OK_FOR_INDEX_STRICT_P (index))
	  || (! strict && ! REG_OK_FOR_INDEX_NONSTRICT_P (index)))
	{
	  reason = "index is not valid";
	  goto report_error;
	}
    }

  /* Validate scale factor.  */
  if (scale != 1)
    {
      reason_rtx = GEN_INT (scale);
      if (!index)
	{
	  reason = "scale without index";
	  goto report_error;
	}

      if (scale != 2 && scale != 4 && scale != 8)
	{
	  reason = "scale is not a valid multiplier";
	  goto report_error;
	}
    }

  /* Validate displacement.  */
  if (disp)
    {
      reason_rtx = disp;

      if (!CONSTANT_ADDRESS_P (disp))
	{
	  reason = "displacement is not constant";
	  goto report_error;
	}

      if (GET_CODE (disp) == CONST_DOUBLE)
	{
	  reason = "displacement is a const_double";
	  goto report_error;
	}

      if (flag_pic && SYMBOLIC_CONST (disp))
	{
	  if (! legitimate_pic_address_disp_p (disp))
	    {
	      reason = "displacement is an invalid pic construct";
	      goto report_error;
	    }

          /* This code used to verify that a symbolic pic displacement
	     includes the pic_offset_table_rtx register.

	     While this is good idea, unfortunately these constructs may
	     be created by "adds using lea" optimization for incorrect
	     code like:

	     int a;
	     int foo(int i)
	       {
	         return *(&a+i);
	       }

	     This code is nonsensical, but results in addressing
	     GOT table with pic_offset_table_rtx base.  We can't
	     just refuse it easilly, since it gets matched by
	     "addsi3" pattern, that later gets split to lea in the
	     case output register differs from input.  While this
	     can be handled by separate addsi pattern for this case
	     that never results in lea, this seems to be easier and
	     correct fix for crash to disable this test.  */
	}
      else if (HALF_PIC_P ())
	{
	  if (! HALF_PIC_ADDRESS_P (disp)
	      || (base != NULL_RTX || index != NULL_RTX))
	    {
	      reason = "displacement is an invalid half-pic reference";
	      goto report_error;
	    }
	}
    }

  /* Everything looks valid.  */
  if (TARGET_DEBUG_ADDR)
    fprintf (stderr, "Success.\n");
  return TRUE;

report_error:
  if (TARGET_DEBUG_ADDR)
    {
      fprintf (stderr, "Error: %s\n", reason);
      debug_rtx (reason_rtx);
    }
  return FALSE;
}
\f
/* Return an unique alias set for the GOT.  */

static HOST_WIDE_INT
ix86_GOT_alias_set ()
{
    static HOST_WIDE_INT set = -1;
    if (set == -1)
      set = new_alias_set ();
    return set;
}

/* Return a legitimate reference for ORIG (an address) using the
   register REG.  If REG is 0, a new pseudo is generated.

   There are two types of references that must be handled:

   1. Global data references must load the address from the GOT, via
      the PIC reg.  An insn is emitted to do this load, and the reg is
      returned.

   2. Static data references, constant pool addresses, and code labels
      compute the address as an offset from the GOT, whose base is in
      the PIC reg.  Static data objects have SYMBOL_REF_FLAG set to
      differentiate them from global data objects.  The returned
      address is the PIC reg + an unspec constant.

   GO_IF_LEGITIMATE_ADDRESS rejects symbolic references unless the PIC
   reg also appears in the address.  */

rtx
legitimize_pic_address (orig, reg)
     rtx orig;
     rtx reg;
{
  rtx addr = orig;
  rtx new = orig;
  rtx base;

  if (GET_CODE (addr) == LABEL_REF
      || (GET_CODE (addr) == SYMBOL_REF
	  && (CONSTANT_POOL_ADDRESS_P (addr)
	      || SYMBOL_REF_FLAG (addr))))
    {
      /* This symbol may be referenced via a displacement from the PIC
	 base address (@GOTOFF).  */

      current_function_uses_pic_offset_table = 1;
      new = gen_rtx_UNSPEC (Pmode, gen_rtvec (1, addr), 7);
      new = gen_rtx_CONST (Pmode, new);
      new = gen_rtx_PLUS (Pmode, pic_offset_table_rtx, new);

      if (reg != 0)
	{
	  emit_move_insn (reg, new);
	  new = reg;
	}
    }
  else if (GET_CODE (addr) == SYMBOL_REF)
    {
      /* This symbol must be referenced via a load from the
	 Global Offset Table (@GOT).  */

      current_function_uses_pic_offset_table = 1;
      new = gen_rtx_UNSPEC (Pmode, gen_rtvec (1, addr), 6);
      new = gen_rtx_CONST (Pmode, new);
      new = gen_rtx_PLUS (Pmode, pic_offset_table_rtx, new);
      new = gen_rtx_MEM (Pmode, new);
      RTX_UNCHANGING_P (new) = 1;
      MEM_ALIAS_SET (new) = ix86_GOT_alias_set ();

      if (reg == 0)
	reg = gen_reg_rtx (Pmode);
      emit_move_insn (reg, new);
      new = reg;
    }
  else
    {
      if (GET_CODE (addr) == CONST)
	{
	  addr = XEXP (addr, 0);
	  if (GET_CODE (addr) == UNSPEC)
	    {
	      /* Check that the unspec is one of the ones we generate?  */
	    }
	  else if (GET_CODE (addr) != PLUS)
	    abort ();
	}
      if (GET_CODE (addr) == PLUS)
	{
	  rtx op0 = XEXP (addr, 0), op1 = XEXP (addr, 1);

	  /* Check first to see if this is a constant offset from a @GOTOFF
	     symbol reference.  */
	  if ((GET_CODE (op0) == LABEL_REF
	       || (GET_CODE (op0) == SYMBOL_REF
		   && (CONSTANT_POOL_ADDRESS_P (op0)
		       || SYMBOL_REF_FLAG (op0))))
	      && GET_CODE (op1) == CONST_INT)
	    {
	      current_function_uses_pic_offset_table = 1;
	      new = gen_rtx_UNSPEC (Pmode, gen_rtvec (1, op0), 7);
	      new = gen_rtx_PLUS (Pmode, new, op1);
	      new = gen_rtx_CONST (Pmode, new);
	      new = gen_rtx_PLUS (Pmode, pic_offset_table_rtx, new);

	      if (reg != 0)
		{
		  emit_move_insn (reg, new);
		  new = reg;
		}
	    }
	  else
	    {
	      base = legitimize_pic_address (XEXP (addr, 0), reg);
	      new  = legitimize_pic_address (XEXP (addr, 1),
					     base == reg ? NULL_RTX : reg);

	      if (GET_CODE (new) == CONST_INT)
		new = plus_constant (base, INTVAL (new));
	      else
		{
		  if (GET_CODE (new) == PLUS && CONSTANT_P (XEXP (new, 1)))
		    {
		      base = gen_rtx_PLUS (Pmode, base, XEXP (new, 0));
		      new = XEXP (new, 1);
		    }
		  new = gen_rtx_PLUS (Pmode, base, new);
		}
	    }
	}
    }
  return new;
}
\f
/* Try machine-dependent ways of modifying an illegitimate address
   to be legitimate.  If we find one, return the new, valid address.
   This macro is used in only one place: `memory_address' in explow.c.

   OLDX is the address as it was before break_out_memory_refs was called.
   In some cases it is useful to look at this to decide what needs to be done.

   MODE and WIN are passed so that this macro can use
   GO_IF_LEGITIMATE_ADDRESS.

   It is always safe for this macro to do nothing.  It exists to recognize
   opportunities to optimize the output.

   For the 80386, we handle X+REG by loading X into a register R and
   using R+REG.  R will go in a general reg and indexing will be used.
   However, if REG is a broken-out memory address or multiplication,
   nothing needs to be done because REG can certainly go in a general reg.

   When -fpic is used, special handling is needed for symbolic references.
   See comments by legitimize_pic_address in i386.c for details.  */

rtx
legitimize_address (x, oldx, mode)
     register rtx x;
     register rtx oldx ATTRIBUTE_UNUSED;
     enum machine_mode mode;
{
  int changed = 0;
  unsigned log;

  if (TARGET_DEBUG_ADDR)
    {
      fprintf (stderr, "\n==========\nLEGITIMIZE_ADDRESS, mode = %s\n",
	       GET_MODE_NAME (mode));
      debug_rtx (x);
    }

  if (flag_pic && SYMBOLIC_CONST (x))
    return legitimize_pic_address (x, 0);

  /* Canonicalize shifts by 0, 1, 2, 3 into multiply */
  if (GET_CODE (x) == ASHIFT
      && GET_CODE (XEXP (x, 1)) == CONST_INT
      && (log = (unsigned)exact_log2 (INTVAL (XEXP (x, 1)))) < 4)
    {
      changed = 1;
      x = gen_rtx_MULT (Pmode, force_reg (Pmode, XEXP (x, 0)),
			GEN_INT (1 << log));
    }

  if (GET_CODE (x) == PLUS)
    {
      /* Canonicalize shifts by 0, 1, 2, 3 into multiply.  */

      if (GET_CODE (XEXP (x, 0)) == ASHIFT
	  && GET_CODE (XEXP (XEXP (x, 0), 1)) == CONST_INT
	  && (log = (unsigned)exact_log2 (INTVAL (XEXP (XEXP (x, 0), 1)))) < 4)
	{
	  changed = 1;
	  XEXP (x, 0) = gen_rtx_MULT (Pmode,
				      force_reg (Pmode, XEXP (XEXP (x, 0), 0)),
				      GEN_INT (1 << log));
	}

      if (GET_CODE (XEXP (x, 1)) == ASHIFT
	  && GET_CODE (XEXP (XEXP (x, 1), 1)) == CONST_INT
	  && (log = (unsigned)exact_log2 (INTVAL (XEXP (XEXP (x, 1), 1)))) < 4)
	{
	  changed = 1;
	  XEXP (x, 1) = gen_rtx_MULT (Pmode,
				      force_reg (Pmode, XEXP (XEXP (x, 1), 0)),
				      GEN_INT (1 << log));
	}

      /* Put multiply first if it isn't already.  */
      if (GET_CODE (XEXP (x, 1)) == MULT)
	{
	  rtx tmp = XEXP (x, 0);
	  XEXP (x, 0) = XEXP (x, 1);
	  XEXP (x, 1) = tmp;
	  changed = 1;
	}

      /* Canonicalize (plus (mult (reg) (const)) (plus (reg) (const)))
	 into (plus (plus (mult (reg) (const)) (reg)) (const)).  This can be
	 created by virtual register instantiation, register elimination, and
	 similar optimizations.  */
      if (GET_CODE (XEXP (x, 0)) == MULT && GET_CODE (XEXP (x, 1)) == PLUS)
	{
	  changed = 1;
	  x = gen_rtx_PLUS (Pmode,
			    gen_rtx_PLUS (Pmode, XEXP (x, 0),
					  XEXP (XEXP (x, 1), 0)),
			    XEXP (XEXP (x, 1), 1));
	}

      /* Canonicalize
	 (plus (plus (mult (reg) (const)) (plus (reg) (const))) const)
	 into (plus (plus (mult (reg) (const)) (reg)) (const)).  */
      else if (GET_CODE (x) == PLUS && GET_CODE (XEXP (x, 0)) == PLUS
	       && GET_CODE (XEXP (XEXP (x, 0), 0)) == MULT
	       && GET_CODE (XEXP (XEXP (x, 0), 1)) == PLUS
	       && CONSTANT_P (XEXP (x, 1)))
	{
	  rtx constant;
	  rtx other = NULL_RTX;

	  if (GET_CODE (XEXP (x, 1)) == CONST_INT)
	    {
	      constant = XEXP (x, 1);
	      other = XEXP (XEXP (XEXP (x, 0), 1), 1);
	    }
	  else if (GET_CODE (XEXP (XEXP (XEXP (x, 0), 1), 1)) == CONST_INT)
	    {
	      constant = XEXP (XEXP (XEXP (x, 0), 1), 1);
	      other = XEXP (x, 1);
	    }
	  else
	    constant = 0;

	  if (constant)
	    {
	      changed = 1;
	      x = gen_rtx_PLUS (Pmode,
				gen_rtx_PLUS (Pmode, XEXP (XEXP (x, 0), 0),
					      XEXP (XEXP (XEXP (x, 0), 1), 0)),
				plus_constant (other, INTVAL (constant)));
	    }
	}

      if (changed && legitimate_address_p (mode, x, FALSE))
	return x;

      if (GET_CODE (XEXP (x, 0)) == MULT)
	{
	  changed = 1;
	  XEXP (x, 0) = force_operand (XEXP (x, 0), 0);
	}

      if (GET_CODE (XEXP (x, 1)) == MULT)
	{
	  changed = 1;
	  XEXP (x, 1) = force_operand (XEXP (x, 1), 0);
	}

      if (changed
	  && GET_CODE (XEXP (x, 1)) == REG
	  && GET_CODE (XEXP (x, 0)) == REG)
	return x;

      if (flag_pic && SYMBOLIC_CONST (XEXP (x, 1)))
	{
	  changed = 1;
	  x = legitimize_pic_address (x, 0);
	}

      if (changed && legitimate_address_p (mode, x, FALSE))
	return x;

      if (GET_CODE (XEXP (x, 0)) == REG)
	{
	  register rtx temp = gen_reg_rtx (Pmode);
	  register rtx val  = force_operand (XEXP (x, 1), temp);
	  if (val != temp)
	    emit_move_insn (temp, val);

	  XEXP (x, 1) = temp;
	  return x;
	}

      else if (GET_CODE (XEXP (x, 1)) == REG)
	{
	  register rtx temp = gen_reg_rtx (Pmode);
	  register rtx val  = force_operand (XEXP (x, 0), temp);
	  if (val != temp)
	    emit_move_insn (temp, val);

	  XEXP (x, 0) = temp;
	  return x;
	}
    }

  return x;
}
\f
/* Print an integer constant expression in assembler syntax.  Addition
   and subtraction are the only arithmetic that may appear in these
   expressions.  FILE is the stdio stream to write to, X is the rtx, and
   CODE is the operand print code from the output string.  */

static void
output_pic_addr_const (file, x, code)
     FILE *file;
     rtx x;
     int code;
{
  char buf[256];

  switch (GET_CODE (x))
    {
    case PC:
      if (flag_pic)
	putc ('.', file);
      else
	abort ();
      break;

    case SYMBOL_REF:
      assemble_name (file, XSTR (x, 0));
      if (code == 'P' && ! SYMBOL_REF_FLAG (x))
	fputs ("@PLT", file);
      break;

    case LABEL_REF:
      x = XEXP (x, 0);
      /* FALLTHRU */
    case CODE_LABEL:
      ASM_GENERATE_INTERNAL_LABEL (buf, "L", CODE_LABEL_NUMBER (x));
      assemble_name (asm_out_file, buf);
      break;

    case CONST_INT:
      fprintf (file, HOST_WIDE_INT_PRINT_DEC, INTVAL (x));
      break;

    case CONST:
      /* This used to output parentheses around the expression,
	 but that does not work on the 386 (either ATT or BSD assembler).  */
      output_pic_addr_const (file, XEXP (x, 0), code);
      break;

    case CONST_DOUBLE:
      if (GET_MODE (x) == VOIDmode)
	{
	  /* We can use %d if the number is <32 bits and positive.  */
	  if (CONST_DOUBLE_HIGH (x) || CONST_DOUBLE_LOW (x) < 0)
	    fprintf (file, "0x%lx%08lx",
		     (unsigned long) CONST_DOUBLE_HIGH (x),
		     (unsigned long) CONST_DOUBLE_LOW (x));
	  else
	    fprintf (file, HOST_WIDE_INT_PRINT_DEC, CONST_DOUBLE_LOW (x));
	}
      else
	/* We can't handle floating point constants;
	   PRINT_OPERAND must handle them.  */
	output_operand_lossage ("floating constant misused");
      break;

    case PLUS:
      /* Some assemblers need integer constants to appear first.  */
      if (GET_CODE (XEXP (x, 0)) == CONST_INT)
	{
	  output_pic_addr_const (file, XEXP (x, 0), code);
	  putc ('+', file);
	  output_pic_addr_const (file, XEXP (x, 1), code);
	}
      else if (GET_CODE (XEXP (x, 1)) == CONST_INT)
	{
	  output_pic_addr_const (file, XEXP (x, 1), code);
	  putc ('+', file);
	  output_pic_addr_const (file, XEXP (x, 0), code);
	}
      else
	abort ();
      break;

    case MINUS:
      putc (ASSEMBLER_DIALECT ? '(' : '[', file);
      output_pic_addr_const (file, XEXP (x, 0), code);
      putc ('-', file);
      output_pic_addr_const (file, XEXP (x, 1), code);
      putc (ASSEMBLER_DIALECT ? ')' : ']', file);
      break;

     case UNSPEC:
       if (XVECLEN (x, 0) != 1)
	abort ();
       output_pic_addr_const (file, XVECEXP (x, 0, 0), code);
       switch (XINT (x, 1))
	{
	case 6:
	  fputs ("@GOT", file);
	  break;
	case 7:
	  fputs ("@GOTOFF", file);
	  break;
	case 8:
	  fputs ("@PLT", file);
	  break;
	default:
	  output_operand_lossage ("invalid UNSPEC as operand");
	  break;
	}
       break;

    default:
      output_operand_lossage ("invalid expression as operand");
    }
}

/* This is called from dwarfout.c via ASM_OUTPUT_DWARF_ADDR_CONST.
   We need to handle our special PIC relocations.  */

void
i386_dwarf_output_addr_const (file, x)
     FILE *file;
     rtx x;
{
  fprintf (file, "%s", INT_ASM_OP);
  if (flag_pic)
    output_pic_addr_const (file, x, '\0');
  else
    output_addr_const (file, x);
  fputc ('\n', file);
}

/* In the name of slightly smaller debug output, and to cater to
   general assembler losage, recognize PIC+GOTOFF and turn it back
   into a direct symbol reference.  */

rtx
i386_simplify_dwarf_addr (orig_x)
     rtx orig_x;
{
  rtx x = orig_x;

  if (GET_CODE (x) != PLUS
      || GET_CODE (XEXP (x, 0)) != REG
      || GET_CODE (XEXP (x, 1)) != CONST)
    return orig_x;

  x = XEXP (XEXP (x, 1), 0);
  if (GET_CODE (x) == UNSPEC
      && (XINT (x, 1) == 6
	  || XINT (x, 1) == 7))
    return XVECEXP (x, 0, 0);

  if (GET_CODE (x) == PLUS
      && GET_CODE (XEXP (x, 0)) == UNSPEC
      && GET_CODE (XEXP (x, 1)) == CONST_INT
      && (XINT (XEXP (x, 0), 1) == 6
	  || XINT (XEXP (x, 0), 1) == 7))
    return gen_rtx_PLUS (VOIDmode, XVECEXP (XEXP (x, 0), 0, 0), XEXP (x, 1));

  return orig_x;
}
\f
static void
put_condition_code (code, mode, reverse, fp, file)
     enum rtx_code code;
     enum machine_mode mode;
     int reverse, fp;
     FILE *file;
{
  const char *suffix;

  if (mode == CCFPmode || mode == CCFPUmode)
    {
      enum rtx_code second_code, bypass_code;
      ix86_fp_comparison_codes (code, &bypass_code, &code, &second_code);
      if (bypass_code != NIL || second_code != NIL)
	abort();
      code = ix86_fp_compare_code_to_integer (code);
      mode = CCmode;
    }
  if (reverse)
    code = reverse_condition (code);

  switch (code)
    {
    case EQ:
      suffix = "e";
      break;
    case NE:
      suffix = "ne";
      break;
    case GT:
      if (mode != CCmode && mode != CCNOmode && mode != CCGCmode)
	abort ();
      suffix = "g";
      break;
    case GTU:
      /* ??? Use "nbe" instead of "a" for fcmov losage on some assemblers.
	 Those same assemblers have the same but opposite losage on cmov.  */
      if (mode != CCmode)
	abort ();
      suffix = fp ? "nbe" : "a";
      break;
    case LT:
      if (mode == CCNOmode || mode == CCGOCmode)
	suffix = "s";
      else if (mode == CCmode || mode == CCGCmode)
	suffix = "l";
      else
	abort ();
      break;
    case LTU:
      if (mode != CCmode)
	abort ();
      suffix = "b";
      break;
    case GE:
      if (mode == CCNOmode || mode == CCGOCmode)
	suffix = "ns";
      else if (mode == CCmode || mode == CCGCmode)
	suffix = "ge";
      else
	abort ();
      break;
    case GEU:
      /* ??? As above.  */
      if (mode != CCmode)
	abort ();
      suffix = fp ? "nb" : "ae";
      break;
    case LE:
      if (mode != CCmode && mode != CCGCmode && mode != CCNOmode)
	abort ();
      suffix = "le";
      break;
    case LEU:
      if (mode != CCmode)
	abort ();
      suffix = "be";
      break;
    case UNORDERED:
      suffix = fp ? "u" : "p";
      break;
    case ORDERED:
      suffix = fp ? "nu" : "np";
      break;
    default:
      abort ();
    }
  fputs (suffix, file);
}

void
print_reg (x, code, file)
     rtx x;
     int code;
     FILE *file;
{
  if (REGNO (x) == ARG_POINTER_REGNUM
      || REGNO (x) == FRAME_POINTER_REGNUM
      || REGNO (x) == FLAGS_REG
      || REGNO (x) == FPSR_REG)
    abort ();

  if (ASSEMBLER_DIALECT == 0 || USER_LABEL_PREFIX[0] == 0)
    putc ('%', file);

  if (code == 'w')
    code = 2;
  else if (code == 'b')
    code = 1;
  else if (code == 'k')
    code = 4;
  else if (code == 'y')
    code = 3;
  else if (code == 'h')
    code = 0;
  else if (code == 'm' || MMX_REG_P (x))
    code = 5;
  else
    code = GET_MODE_SIZE (GET_MODE (x));

  switch (code)
    {
    case 5:
      fputs (hi_reg_name[REGNO (x)], file);
      break;
    case 3:
      if (STACK_TOP_P (x))
	{
	  fputs ("st(0)", file);
	  break;
	}
      /* FALLTHRU */
    case 4:
    case 8:
    case 12:
      if (! ANY_FP_REG_P (x))
	putc ('e', file);
      /* FALLTHRU */
    case 16:
    case 2:
      fputs (hi_reg_name[REGNO (x)], file);
      break;
    case 1:
      fputs (qi_reg_name[REGNO (x)], file);
      break;
    case 0:
      fputs (qi_high_reg_name[REGNO (x)], file);
      break;
    default:
      abort ();
    }
}

/* Meaning of CODE:
   L,W,B,Q,S,T -- print the opcode suffix for specified size of operand.
   C -- print opcode suffix for set/cmov insn.
   c -- like C, but print reversed condition
   R -- print the prefix for register names.
   z -- print the opcode suffix for the size of the current operand.
   * -- print a star (in certain assembler syntax)
   A -- print an absolute memory reference.
   w -- print the operand as if it's a "word" (HImode) even if it isn't.
   s -- print a shift double count, followed by the assemblers argument
	delimiter.
   b -- print the QImode name of the register for the indicated operand.
	%b0 would print %al if operands[0] is reg 0.
   w --  likewise, print the HImode name of the register.
   k --  likewise, print the SImode name of the register.
   h --  print the QImode name for a "high" register, either ah, bh, ch or dh.
   y --  print "st(0)" instead of "st" as a register.
   m --  print "st(n)" as an mmx register.
   D -- print condition for SSE cmp instruction.
 */

void
print_operand (file, x, code)
     FILE *file;
     rtx x;
     int code;
{
  if (code)
    {
      switch (code)
	{
	case '*':
	  if (ASSEMBLER_DIALECT == 0)
	    putc ('*', file);
	  return;

	case 'A':
	  if (ASSEMBLER_DIALECT == 0)
	    putc ('*', file);
	  else if (ASSEMBLER_DIALECT == 1)
	    {
	      /* Intel syntax. For absolute addresses, registers should not
		 be surrounded by braces.  */
	      if (GET_CODE (x) != REG)
		{
		  putc ('[', file);
		  PRINT_OPERAND (file, x, 0);
		  putc (']', file);
		  return;
		}
	    }

	  PRINT_OPERAND (file, x, 0);
	  return;


	case 'L':
	  if (ASSEMBLER_DIALECT == 0)
	    putc ('l', file);
	  return;

	case 'W':
	  if (ASSEMBLER_DIALECT == 0)
	    putc ('w', file);
	  return;

	case 'B':
	  if (ASSEMBLER_DIALECT == 0)
	    putc ('b', file);
	  return;

	case 'Q':
	  if (ASSEMBLER_DIALECT == 0)
	    putc ('l', file);
	  return;

	case 'S':
	  if (ASSEMBLER_DIALECT == 0)
	    putc ('s', file);
	  return;

	case 'T':
	  if (ASSEMBLER_DIALECT == 0)
	    putc ('t', file);
	  return;

	case 'z':
	  /* 387 opcodes don't get size suffixes if the operands are
	     registers.  */

	  if (STACK_REG_P (x))
	    return;

	  /* this is the size of op from size of operand */
	  switch (GET_MODE_SIZE (GET_MODE (x)))
	    {
	    case 2:
#ifdef HAVE_GAS_FILDS_FISTS
	      putc ('s', file);
#endif
	      return;

	    case 4:
	      if (GET_MODE (x) == SFmode)
		{
		  putc ('s', file);
		  return;
		}
	      else
		putc ('l', file);
	      return;

	    case 12:
	    case 16:
	      putc ('t', file);
	      return;

	    case 8:
	      if (GET_MODE_CLASS (GET_MODE (x)) == MODE_INT)
		{
#ifdef GAS_MNEMONICS
		  putc ('q', file);
#else
		  putc ('l', file);
		  putc ('l', file);
#endif
		}
	      else
	        putc ('l', file);
	      return;

	    default:
	      abort ();
	    }

	case 'b':
	case 'w':
	case 'k':
	case 'h':
	case 'y':
	case 'm':
	case 'X':
	case 'P':
	  break;

	case 's':
	  if (GET_CODE (x) == CONST_INT || ! SHIFT_DOUBLE_OMITS_COUNT)
	    {
	      PRINT_OPERAND (file, x, 0);
	      putc (',', file);
	    }
	  return;

	case 'D':
	  /* Little bit of braindamage here.  The SSE compare instructions
	     does use completely different names for the comparisons that the
	     fp conditional moves.  */
	  switch (GET_CODE (x))
	    {
	    case EQ:
	    case UNEQ:
	      fputs ("eq", file);
	      break;
	    case LT:
	    case UNLT:
	      fputs ("lt", file);
	      break;
	    case LE:
	    case UNLE:
	      fputs ("le", file);
	      break;
	    case UNORDERED:
	      fputs ("unord", file);
	      break;
	    case NE:
	    case LTGT:
	      fputs ("neq", file);
	      break;
	    case UNGE:
	    case GE:
	      fputs ("nlt", file);
	      break;
	    case UNGT:
	    case GT:
	      fputs ("nle", file);
	      break;
	    case ORDERED:
	      fputs ("ord", file);
	      break;
	    default:
	      abort ();
	      break;
	    }
	  return;
	case 'C':
	  put_condition_code (GET_CODE (x), GET_MODE (XEXP (x, 0)), 0, 0, file);
	  return;
	case 'F':
	  put_condition_code (GET_CODE (x), GET_MODE (XEXP (x, 0)), 0, 1, file);
	  return;

	  /* Like above, but reverse condition */
	case 'c':
	  put_condition_code (GET_CODE (x), GET_MODE (XEXP (x, 0)), 1, 0, file);
	  return;
	case 'f':
	  put_condition_code (GET_CODE (x), GET_MODE (XEXP (x, 0)), 1, 1, file);
	  return;

	default:
	  {
	    char str[50];
	    sprintf (str, "invalid operand code `%c'", code);
	    output_operand_lossage (str);
	  }
	}
    }

  if (GET_CODE (x) == REG)
    {
      PRINT_REG (x, code, file);
    }

  else if (GET_CODE (x) == MEM)
    {
      /* No `byte ptr' prefix for call instructions.  */
      if (ASSEMBLER_DIALECT != 0 && code != 'X' && code != 'P')
	{
	  const char * size;
	  switch (GET_MODE_SIZE (GET_MODE (x)))
	    {
	    case 1: size = "BYTE"; break;
	    case 2: size = "WORD"; break;
	    case 4: size = "DWORD"; break;
	    case 8: size = "QWORD"; break;
	    case 12: size = "XWORD"; break;
	    case 16: size = "XMMWORD"; break;
	    default:
	      abort ();
	    }

	  /* Check for explicit size override (codes 'b', 'w' and 'k')  */
	  if (code == 'b')
	    size = "BYTE";
	  else if (code == 'w')
	    size = "WORD";
	  else if (code == 'k')
	    size = "DWORD";

	  fputs (size, file);
	  fputs (" PTR ", file);
	}

      x = XEXP (x, 0);
      if (flag_pic && CONSTANT_ADDRESS_P (x))
	output_pic_addr_const (file, x, code);
      else
	output_address (x);
    }

  else if (GET_CODE (x) == CONST_DOUBLE && GET_MODE (x) == SFmode)
    {
      REAL_VALUE_TYPE r;
      long l;

      REAL_VALUE_FROM_CONST_DOUBLE (r, x);
      REAL_VALUE_TO_TARGET_SINGLE (r, l);

      if (ASSEMBLER_DIALECT == 0)
	putc ('$', file);
      fprintf (file, "0x%lx", l);
    }

 /* These float cases don't actually occur as immediate operands.  */
 else if (GET_CODE (x) == CONST_DOUBLE && GET_MODE (x) == DFmode)
    {
      REAL_VALUE_TYPE r;
      char dstr[30];

      REAL_VALUE_FROM_CONST_DOUBLE (r, x);
      REAL_VALUE_TO_DECIMAL (r, "%.22e", dstr);
      fprintf (file, "%s", dstr);
    }

  else if (GET_CODE (x) == CONST_DOUBLE
	   && (GET_MODE (x) == XFmode || GET_MODE (x) == TFmode))
    {
      REAL_VALUE_TYPE r;
      char dstr[30];

      REAL_VALUE_FROM_CONST_DOUBLE (r, x);
      REAL_VALUE_TO_DECIMAL (r, "%.22e", dstr);
      fprintf (file, "%s", dstr);
    }
  else
    {
      if (code != 'P')
	{
	  if (GET_CODE (x) == CONST_INT || GET_CODE (x) == CONST_DOUBLE)
	    {
	      if (ASSEMBLER_DIALECT == 0)
		putc ('$', file);
	    }
	  else if (GET_CODE (x) == CONST || GET_CODE (x) == SYMBOL_REF
		   || GET_CODE (x) == LABEL_REF)
	    {
	      if (ASSEMBLER_DIALECT == 0)
		putc ('$', file);
	      else
		fputs ("OFFSET FLAT:", file);
	    }
	}
      if (GET_CODE (x) == CONST_INT)
	fprintf (file, HOST_WIDE_INT_PRINT_DEC, INTVAL (x));
      else if (flag_pic)
	output_pic_addr_const (file, x, code);
      else
	output_addr_const (file, x);
    }
}
\f
/* Print a memory operand whose address is ADDR.  */

void
print_operand_address (file, addr)
     FILE *file;
     register rtx addr;
{
  struct ix86_address parts;
  rtx base, index, disp;
  int scale;

  if (! ix86_decompose_address (addr, &parts))
    abort ();

  base = parts.base;
  index = parts.index;
  disp = parts.disp;
  scale = parts.scale;

  if (!base && !index)
    {
      /* Displacement only requires special attention.  */

      if (GET_CODE (disp) == CONST_INT)
	{
	  if (ASSEMBLER_DIALECT != 0)
	    {
	      if (USER_LABEL_PREFIX[0] == 0)
		putc ('%', file);
	      fputs ("ds:", file);
	    }
	  fprintf (file, HOST_WIDE_INT_PRINT_DEC, INTVAL (addr));
	}
      else if (flag_pic)
	output_pic_addr_const (file, addr, 0);
      else
	output_addr_const (file, addr);
    }
  else
    {
      if (ASSEMBLER_DIALECT == 0)
	{
	  if (disp)
	    {
	      if (flag_pic)
		output_pic_addr_const (file, disp, 0);
	      else if (GET_CODE (disp) == LABEL_REF)
		output_asm_label (disp);
	      else
		output_addr_const (file, disp);
	    }

	  putc ('(', file);
	  if (base)
	    PRINT_REG (base, 0, file);
	  if (index)
	    {
	      putc (',', file);
	      PRINT_REG (index, 0, file);
	      if (scale != 1)
		fprintf (file, ",%d", scale);
	    }
	  putc (')', file);
	}
      else
	{
	  rtx offset = NULL_RTX;

	  if (disp)
	    {
	      /* Pull out the offset of a symbol; print any symbol itself.  */
	      if (GET_CODE (disp) == CONST
		  && GET_CODE (XEXP (disp, 0)) == PLUS
		  && GET_CODE (XEXP (XEXP (disp, 0), 1)) == CONST_INT)
		{
		  offset = XEXP (XEXP (disp, 0), 1);
		  disp = gen_rtx_CONST (VOIDmode,
					XEXP (XEXP (disp, 0), 0));
		}

	      if (flag_pic)
		output_pic_addr_const (file, disp, 0);
	      else if (GET_CODE (disp) == LABEL_REF)
		output_asm_label (disp);
	      else if (GET_CODE (disp) == CONST_INT)
		offset = disp;
	      else
		output_addr_const (file, disp);
	    }

	  putc ('[', file);
	  if (base)
	    {
	      PRINT_REG (base, 0, file);
	      if (offset)
		{
		  if (INTVAL (offset) >= 0)
		    putc ('+', file);
		  fprintf (file, HOST_WIDE_INT_PRINT_DEC, INTVAL (offset));
		}
	    }
	  else if (offset)
	    fprintf (file, HOST_WIDE_INT_PRINT_DEC, INTVAL (offset));
	  else
	    putc ('0', file);

	  if (index)
	    {
	      putc ('+', file);
	      PRINT_REG (index, 0, file);
	      if (scale != 1)
		fprintf (file, "*%d", scale);
	    }
	  putc (']', file);
	}
    }
}
\f
/* Split one or more DImode RTL references into pairs of SImode
   references.  The RTL can be REG, offsettable MEM, integer constant, or
   CONST_DOUBLE.  "operands" is a pointer to an array of DImode RTL to
   split and "num" is its length.  lo_half and hi_half are output arrays
   that parallel "operands".  */

void
split_di (operands, num, lo_half, hi_half)
     rtx operands[];
     int num;
     rtx lo_half[], hi_half[];
{
  while (num--)
    {
      rtx op = operands[num];
      if (CONSTANT_P (op))
	split_double (op, &lo_half[num], &hi_half[num]);
      else if (! reload_completed)
	{
	  lo_half[num] = gen_lowpart (SImode, op);
	  hi_half[num] = gen_highpart (SImode, op);
	}
      else if (GET_CODE (op) == REG)
	{
	  lo_half[num] = gen_rtx_REG (SImode, REGNO (op));
	  hi_half[num] = gen_rtx_REG (SImode, REGNO (op) + 1);
	}
      else if (offsettable_memref_p (op))
	{
	  rtx lo_addr = XEXP (op, 0);
	  rtx hi_addr = XEXP (adj_offsettable_operand (op, 4), 0);
	  lo_half[num] = change_address (op, SImode, lo_addr);
	  hi_half[num] = change_address (op, SImode, hi_addr);
	}
      else
	abort ();
    }
}
\f
/* Output code to perform a 387 binary operation in INSN, one of PLUS,
   MINUS, MULT or DIV.  OPERANDS are the insn operands, where operands[3]
   is the expression of the binary operation.  The output may either be
   emitted here, or returned to the caller, like all output_* functions.

   There is no guarantee that the operands are the same mode, as they
   might be within FLOAT or FLOAT_EXTEND expressions.  */

#ifndef SYSV386_COMPAT
/* Set to 1 for compatibility with brain-damaged assemblers.  No-one
   wants to fix the assemblers because that causes incompatibility
   with gcc.  No-one wants to fix gcc because that causes
   incompatibility with assemblers...  You can use the option of
   -DSYSV386_COMPAT=0 if you recompile both gcc and gas this way.  */
#define SYSV386_COMPAT 1
#endif

const char *
output_387_binary_op (insn, operands)
     rtx insn;
     rtx *operands;
{
  static char buf[30];
  const char *p;
  const char *ssep;
  int is_sse = SSE_REG_P (operands[0]) | SSE_REG_P (operands[1]) | SSE_REG_P (operands[2]);

#ifdef ENABLE_CHECKING
  /* Even if we do not want to check the inputs, this documents input
     constraints.  Which helps in understanding the following code.  */
  if (STACK_REG_P (operands[0])
      && ((REG_P (operands[1])
	   && REGNO (operands[0]) == REGNO (operands[1])
	   && (STACK_REG_P (operands[2]) || GET_CODE (operands[2]) == MEM))
	  || (REG_P (operands[2])
	      && REGNO (operands[0]) == REGNO (operands[2])
	      && (STACK_REG_P (operands[1]) || GET_CODE (operands[1]) == MEM)))
      && (STACK_TOP_P (operands[1]) || STACK_TOP_P (operands[2])))
    ; /* ok */
  else if (!is_sse)
    abort ();
#endif

  switch (GET_CODE (operands[3]))
    {
    case PLUS:
      if (GET_MODE_CLASS (GET_MODE (operands[1])) == MODE_INT
	  || GET_MODE_CLASS (GET_MODE (operands[2])) == MODE_INT)
	p = "fiadd";
      else
	p = "fadd";
      ssep = "add";
      break;

    case MINUS:
      if (GET_MODE_CLASS (GET_MODE (operands[1])) == MODE_INT
	  || GET_MODE_CLASS (GET_MODE (operands[2])) == MODE_INT)
	p = "fisub";
      else
	p = "fsub";
      ssep = "sub";
      break;

    case MULT:
      if (GET_MODE_CLASS (GET_MODE (operands[1])) == MODE_INT
	  || GET_MODE_CLASS (GET_MODE (operands[2])) == MODE_INT)
	p = "fimul";
      else
	p = "fmul";
      ssep = "mul";
      break;

    case DIV:
      if (GET_MODE_CLASS (GET_MODE (operands[1])) == MODE_INT
	  || GET_MODE_CLASS (GET_MODE (operands[2])) == MODE_INT)
	p = "fidiv";
      else
	p = "fdiv";
      ssep = "div";
      break;

    default:
      abort ();
    }

  if (is_sse)
   {
      strcpy (buf, ssep);
      if (GET_MODE (operands[0]) == SFmode)
	strcat (buf, "ss\t{%2, %0|%0, %2}");
      else
	strcat (buf, "sd\t{%2, %0|%0, %2}");
      return buf;
   }
  strcpy (buf, p);

  switch (GET_CODE (operands[3]))
    {
    case MULT:
    case PLUS:
      if (REG_P (operands[2]) && REGNO (operands[0]) == REGNO (operands[2]))
	{
	  rtx temp = operands[2];
	  operands[2] = operands[1];
	  operands[1] = temp;
	}

      /* know operands[0] == operands[1].  */

      if (GET_CODE (operands[2]) == MEM)
	{
	  p = "%z2\t%2";
	  break;
	}

      if (find_regno_note (insn, REG_DEAD, REGNO (operands[2])))
	{
	  if (STACK_TOP_P (operands[0]))
	    /* How is it that we are storing to a dead operand[2]?
	       Well, presumably operands[1] is dead too.  We can't
	       store the result to st(0) as st(0) gets popped on this
	       instruction.  Instead store to operands[2] (which I
	       think has to be st(1)).  st(1) will be popped later.
	       gcc <= 2.8.1 didn't have this check and generated
	       assembly code that the Unixware assembler rejected.  */
	    p = "p\t{%0, %2|%2, %0}";	/* st(1) = st(0) op st(1); pop */
	  else
	    p = "p\t{%2, %0|%0, %2}";	/* st(r1) = st(r1) op st(0); pop */
	  break;
	}

      if (STACK_TOP_P (operands[0]))
	p = "\t{%y2, %0|%0, %y2}";	/* st(0) = st(0) op st(r2) */
      else
	p = "\t{%2, %0|%0, %2}";	/* st(r1) = st(r1) op st(0) */
      break;

    case MINUS:
    case DIV:
      if (GET_CODE (operands[1]) == MEM)
	{
	  p = "r%z1\t%1";
	  break;
	}

      if (GET_CODE (operands[2]) == MEM)
	{
	  p = "%z2\t%2";
	  break;
	}

      if (find_regno_note (insn, REG_DEAD, REGNO (operands[2])))
	{
#if SYSV386_COMPAT
	  /* The SystemV/386 SVR3.2 assembler, and probably all AT&T
	     derived assemblers, confusingly reverse the direction of
	     the operation for fsub{r} and fdiv{r} when the
	     destination register is not st(0).  The Intel assembler
	     doesn't have this brain damage.  Read !SYSV386_COMPAT to
	     figure out what the hardware really does.  */
	  if (STACK_TOP_P (operands[0]))
	    p = "{p\t%0, %2|rp\t%2, %0}";
	  else
	    p = "{rp\t%2, %0|p\t%0, %2}";
#else
	  if (STACK_TOP_P (operands[0]))
	    /* As above for fmul/fadd, we can't store to st(0).  */
	    p = "rp\t{%0, %2|%2, %0}";	/* st(1) = st(0) op st(1); pop */
	  else
	    p = "p\t{%2, %0|%0, %2}";	/* st(r1) = st(r1) op st(0); pop */
#endif
	  break;
	}

      if (find_regno_note (insn, REG_DEAD, REGNO (operands[1])))
	{
#if SYSV386_COMPAT
	  if (STACK_TOP_P (operands[0]))
	    p = "{rp\t%0, %1|p\t%1, %0}";
	  else
	    p = "{p\t%1, %0|rp\t%0, %1}";
#else
	  if (STACK_TOP_P (operands[0]))
	    p = "p\t{%0, %1|%1, %0}";	/* st(1) = st(1) op st(0); pop */
	  else
	    p = "rp\t{%1, %0|%0, %1}";	/* st(r2) = st(0) op st(r2); pop */
#endif
	  break;
	}

      if (STACK_TOP_P (operands[0]))
	{
	  if (STACK_TOP_P (operands[1]))
	    p = "\t{%y2, %0|%0, %y2}";	/* st(0) = st(0) op st(r2) */
	  else
	    p = "r\t{%y1, %0|%0, %y1}";	/* st(0) = st(r1) op st(0) */
	  break;
	}
      else if (STACK_TOP_P (operands[1]))
	{
#if SYSV386_COMPAT
	  p = "{\t%1, %0|r\t%0, %1}";
#else
	  p = "r\t{%1, %0|%0, %1}";	/* st(r2) = st(0) op st(r2) */
#endif
	}
      else
	{
#if SYSV386_COMPAT
	  p = "{r\t%2, %0|\t%0, %2}";
#else
	  p = "\t{%2, %0|%0, %2}";	/* st(r1) = st(r1) op st(0) */
#endif
	}
      break;

    default:
      abort ();
    }

  strcat (buf, p);
  return buf;
}

/* Output code for INSN to convert a float to a signed int.  OPERANDS
   are the insn operands.  The output may be [HSD]Imode and the input
   operand may be [SDX]Fmode.  */

const char *
output_fix_trunc (insn, operands)
     rtx insn;
     rtx *operands;
{
  int stack_top_dies = find_regno_note (insn, REG_DEAD, FIRST_STACK_REG) != 0;
  int dimode_p = GET_MODE (operands[0]) == DImode;
  rtx xops[4];

  /* Jump through a hoop or two for DImode, since the hardware has no
     non-popping instruction.  We used to do this a different way, but
     that was somewhat fragile and broke with post-reload splitters.  */
  if (dimode_p && !stack_top_dies)
    output_asm_insn ("fld\t%y1", operands);

  if (! STACK_TOP_P (operands[1]))
    abort ();

  xops[0] = GEN_INT (12);
  xops[1] = adj_offsettable_operand (operands[2], 1);
  xops[1] = change_address (xops[1], QImode, NULL_RTX);

  xops[2] = operands[0];
  if (GET_CODE (operands[0]) != MEM)
    xops[2] = operands[3];

  output_asm_insn ("fnstcw\t%2", operands);
  output_asm_insn ("mov{l}\t{%2, %4|%4, %2}", operands);
  output_asm_insn ("mov{b}\t{%0, %1|%1, %0}", xops);
  output_asm_insn ("fldcw\t%2", operands);
  output_asm_insn ("mov{l}\t{%4, %2|%2, %4}", operands);

  if (stack_top_dies || dimode_p)
    output_asm_insn ("fistp%z2\t%2", xops);
  else
    output_asm_insn ("fist%z2\t%2", xops);

  output_asm_insn ("fldcw\t%2", operands);

  if (GET_CODE (operands[0]) != MEM)
    {
      if (dimode_p)
	{
	  split_di (operands+0, 1, xops+0, xops+1);
	  split_di (operands+3, 1, xops+2, xops+3);
	  output_asm_insn ("mov{l}\t{%2, %0|%0, %2}", xops);
	  output_asm_insn ("mov{l}\t{%3, %1|%1, %3}", xops);
	}
      else if (GET_MODE (operands[0]) == SImode)
	output_asm_insn ("mov{l}\t{%3, %0|%0, %3}", operands);
      else
	output_asm_insn ("mov{w}\t{%3, %0|%0, %3}", operands);
    }

  return "";
}

/* Output code for INSN to compare OPERANDS.  EFLAGS_P is 1 when fcomi
   should be used and 2 when fnstsw should be used.  UNORDERED_P is true
   when fucom should be used.  */

const char *
output_fp_compare (insn, operands, eflags_p, unordered_p)
     rtx insn;
     rtx *operands;
     int eflags_p, unordered_p;
{
  int stack_top_dies;
  rtx cmp_op0 = operands[0];
  rtx cmp_op1 = operands[1];
  int is_sse = SSE_REG_P (operands[0]) | SSE_REG_P (operands[1]);

  if (eflags_p == 2)
    {
      cmp_op0 = cmp_op1;
      cmp_op1 = operands[2];
    }
  if (is_sse)
    {
      if (GET_MODE (operands[0]) == SFmode)
	if (unordered_p)
	  return "ucomiss\t{%1, %0|%0, %1}";
	else
	  return "comiss\t{%1, %0|%0, %y}";
      else
	if (unordered_p)
	  return "ucomisd\t{%1, %0|%0, %1}";
	else
	  return "comisd\t{%1, %0|%0, %y}";
    }

  if (! STACK_TOP_P (cmp_op0))
    abort ();

  stack_top_dies = find_regno_note (insn, REG_DEAD, FIRST_STACK_REG) != 0;

  if (STACK_REG_P (cmp_op1)
      && stack_top_dies
      && find_regno_note (insn, REG_DEAD, REGNO (cmp_op1))
      && REGNO (cmp_op1) != FIRST_STACK_REG)
    {
      /* If both the top of the 387 stack dies, and the other operand
	 is also a stack register that dies, then this must be a
	 `fcompp' float compare */

      if (eflags_p == 1)
	{
	  /* There is no double popping fcomi variant.  Fortunately,
	     eflags is immune from the fstp's cc clobbering.  */
	  if (unordered_p)
	    output_asm_insn ("fucomip\t{%y1, %0|%0, %y1}", operands);
	  else
	    output_asm_insn ("fcomip\t{%y1, %0|%0, %y1}", operands);
	  return "fstp\t%y0";
	}
      else
	{
	  if (eflags_p == 2)
	    {
	      if (unordered_p)
		return "fucompp\n\tfnstsw\t%0";
	      else
		return "fcompp\n\tfnstsw\t%0";
	    }
	  else
	    {
	      if (unordered_p)
		return "fucompp";
	      else
		return "fcompp";
	    }
	}
    }
  else
    {
      /* Encoded here as eflags_p | intmode | unordered_p | stack_top_dies.  */

      static const char * const alt[24] =
      {
	"fcom%z1\t%y1",
	"fcomp%z1\t%y1",
	"fucom%z1\t%y1",
	"fucomp%z1\t%y1",

	"ficom%z1\t%y1",
	"ficomp%z1\t%y1",
	NULL,
	NULL,

	"fcomi\t{%y1, %0|%0, %y1}",
	"fcomip\t{%y1, %0|%0, %y1}",
	"fucomi\t{%y1, %0|%0, %y1}",
	"fucomip\t{%y1, %0|%0, %y1}",

	NULL,
	NULL,
	NULL,
	NULL,

	"fcom%z2\t%y2\n\tfnstsw\t%0",
	"fcomp%z2\t%y2\n\tfnstsw\t%0",
	"fucom%z2\t%y2\n\tfnstsw\t%0",
	"fucomp%z2\t%y2\n\tfnstsw\t%0",

	"ficom%z2\t%y2\n\tfnstsw\t%0",
	"ficomp%z2\t%y2\n\tfnstsw\t%0",
	NULL,
	NULL
      };

      int mask;
      const char *ret;

      mask  = eflags_p << 3;
      mask |= (GET_MODE_CLASS (GET_MODE (operands[1])) == MODE_INT) << 2;
      mask |= unordered_p << 1;
      mask |= stack_top_dies;

      if (mask >= 24)
	abort ();
      ret = alt[mask];
      if (ret == NULL)
	abort ();

      return ret;
    }
}

/* Output assembler code to FILE to initialize basic-block profiling.

   If profile_block_flag == 2

	Output code to call the subroutine `__bb_init_trace_func'
	and pass two parameters to it. The first parameter is
	the address of a block allocated in the object module.
	The second parameter is the number of the first basic block
	of the function.

	The name of the block is a local symbol made with this statement:

	    ASM_GENERATE_INTERNAL_LABEL (BUFFER, "LPBX", 0);

	Of course, since you are writing the definition of
	`ASM_GENERATE_INTERNAL_LABEL' as well as that of this macro, you
	can take a short cut in the definition of this macro and use the
	name that you know will result.

	The number of the first basic block of the function is
	passed to the macro in BLOCK_OR_LABEL.

	If described in a virtual assembler language the code to be
	output looks like:

		parameter1 <- LPBX0
		parameter2 <- BLOCK_OR_LABEL
		call __bb_init_trace_func

    else if profile_block_flag != 0

	Output code to call the subroutine `__bb_init_func'
	and pass one single parameter to it, which is the same
	as the first parameter to `__bb_init_trace_func'.

	The first word of this parameter is a flag which will be nonzero if
	the object module has already been initialized.  So test this word
	first, and do not call `__bb_init_func' if the flag is nonzero.
	Note: When profile_block_flag == 2 the test need not be done
	but `__bb_init_trace_func' *must* be called.

	BLOCK_OR_LABEL may be used to generate a label number as a
	branch destination in case `__bb_init_func' will not be called.

	If described in a virtual assembler language the code to be
	output looks like:

		cmp (LPBX0),0
		jne local_label
		parameter1 <- LPBX0
		call __bb_init_func
	      local_label:
*/

void
ix86_output_function_block_profiler (file, block_or_label)
     FILE *file;
     int block_or_label;
{
  static int num_func = 0;
  rtx xops[8];
  char block_table[80], false_label[80];

  ASM_GENERATE_INTERNAL_LABEL (block_table, "LPBX", 0);

  xops[1] = gen_rtx_SYMBOL_REF (VOIDmode, block_table);
  xops[5] = stack_pointer_rtx;
  xops[7] = gen_rtx_REG (Pmode, 0); /* eax */

  CONSTANT_POOL_ADDRESS_P (xops[1]) = TRUE;

  switch (profile_block_flag)
    {
    case 2:
      xops[2] = GEN_INT (block_or_label);
      xops[3] = gen_rtx_MEM (Pmode,
		     gen_rtx_SYMBOL_REF (VOIDmode, "__bb_init_trace_func"));
      xops[6] = GEN_INT (8);

      output_asm_insn ("push{l}\t%2", xops);
      if (!flag_pic)
	output_asm_insn ("push{l}\t%1", xops);
      else
	{
	  output_asm_insn ("lea{l}\t{%a1, %7|%7, %a1}", xops);
	  output_asm_insn ("push{l}\t%7", xops);
	}
      output_asm_insn ("call\t%P3", xops);
      output_asm_insn ("add{l}\t{%6, %5|%5, %6}", xops);
      break;

    default:
      ASM_GENERATE_INTERNAL_LABEL (false_label, "LPBZ", num_func);

      xops[0] = const0_rtx;
      xops[2] = gen_rtx_MEM (Pmode,
			     gen_rtx_SYMBOL_REF (VOIDmode, false_label));
      xops[3] = gen_rtx_MEM (Pmode,
			     gen_rtx_SYMBOL_REF (VOIDmode, "__bb_init_func"));
      xops[4] = gen_rtx_MEM (Pmode, xops[1]);
      xops[6] = GEN_INT (4);

      CONSTANT_POOL_ADDRESS_P (xops[2]) = TRUE;

      output_asm_insn ("cmp{l}\t{%0, %4|%4, %0}", xops);
      output_asm_insn ("jne\t%2", xops);

      if (!flag_pic)
	output_asm_insn ("push{l}\t%1", xops);
      else
	{
	  output_asm_insn ("lea{l}\t{%a1, %7|%7, %a2}", xops);
	  output_asm_insn ("push{l}\t%7", xops);
	}
      output_asm_insn ("call\t%P3", xops);
      output_asm_insn ("add{l}\t{%6, %5|%5, %6}", xops);
      ASM_OUTPUT_INTERNAL_LABEL (file, "LPBZ", num_func);
      num_func++;
      break;
    }
}

/* Output assembler code to FILE to increment a counter associated
   with basic block number BLOCKNO.

   If profile_block_flag == 2

	Output code to initialize the global structure `__bb' and
	call the function `__bb_trace_func' which will increment the
	counter.

	`__bb' consists of two words. In the first word the number
	of the basic block has to be stored. In the second word
	the address of a block allocated in the object module
	has to be stored.

	The basic block number is given by BLOCKNO.

	The address of the block is given by the label created with

	    ASM_GENERATE_INTERNAL_LABEL (BUFFER, "LPBX", 0);

	by FUNCTION_BLOCK_PROFILER.

	Of course, since you are writing the definition of
	`ASM_GENERATE_INTERNAL_LABEL' as well as that of this macro, you
	can take a short cut in the definition of this macro and use the
	name that you know will result.

	If described in a virtual assembler language the code to be
	output looks like:

		move BLOCKNO -> (__bb)
		move LPBX0 -> (__bb+4)
		call __bb_trace_func

	Note that function `__bb_trace_func' must not change the
	machine state, especially the flag register. To grant
	this, you must output code to save and restore registers
	either in this macro or in the macros MACHINE_STATE_SAVE
	and MACHINE_STATE_RESTORE. The last two macros will be
	used in the function `__bb_trace_func', so you must make
	sure that the function prologue does not change any
	register prior to saving it with MACHINE_STATE_SAVE.

   else if profile_block_flag != 0

	Output code to increment the counter directly.
	Basic blocks are numbered separately from zero within each
	compiled object module. The count associated with block number
	BLOCKNO is at index BLOCKNO in an array of words; the name of
	this array is a local symbol made with this statement:

	    ASM_GENERATE_INTERNAL_LABEL (BUFFER, "LPBX", 2);

	Of course, since you are writing the definition of
	`ASM_GENERATE_INTERNAL_LABEL' as well as that of this macro, you
	can take a short cut in the definition of this macro and use the
	name that you know will result.

	If described in a virtual assembler language the code to be
	output looks like:

		inc (LPBX2+4*BLOCKNO)
*/

void
ix86_output_block_profiler (file, blockno)
     FILE *file ATTRIBUTE_UNUSED;
     int blockno;
{
  rtx xops[8], cnt_rtx;
  char counts[80];
  char *block_table = counts;

  switch (profile_block_flag)
    {
    case 2:
      ASM_GENERATE_INTERNAL_LABEL (block_table, "LPBX", 0);

      xops[1] = gen_rtx_SYMBOL_REF (VOIDmode, block_table);
      xops[2] = GEN_INT (blockno);
      xops[3] = gen_rtx_MEM (Pmode,
			     gen_rtx_SYMBOL_REF (VOIDmode, "__bb_trace_func"));
      xops[4] = gen_rtx_SYMBOL_REF (VOIDmode, "__bb");
      xops[5] = plus_constant (xops[4], 4);
      xops[0] = gen_rtx_MEM (SImode, xops[4]);
      xops[6] = gen_rtx_MEM (SImode, xops[5]);

      CONSTANT_POOL_ADDRESS_P (xops[1]) = TRUE;

      output_asm_insn ("pushf", xops);
      output_asm_insn ("mov{l}\t{%2, %0|%0, %2}", xops);
      if (flag_pic)
	{
	  xops[7] = gen_rtx_REG (Pmode, 0); /* eax */
	  output_asm_insn ("push{l}\t%7", xops);
	  output_asm_insn ("lea{l}\t{%a1, %7|%7, %a1}", xops);
	  output_asm_insn ("mov{l}\t{%7, %6|%6, %7}", xops);
	  output_asm_insn ("pop{l}\t%7", xops);
	}
      else
	output_asm_insn ("mov{l}\t{%1, %6|%6, %1}", xops);
      output_asm_insn ("call\t%P3", xops);
      output_asm_insn ("popf", xops);

      break;

    default:
      ASM_GENERATE_INTERNAL_LABEL (counts, "LPBX", 2);
      cnt_rtx = gen_rtx_SYMBOL_REF (VOIDmode, counts);
      SYMBOL_REF_FLAG (cnt_rtx) = TRUE;

      if (blockno)
	cnt_rtx = plus_constant (cnt_rtx, blockno*4);

      if (flag_pic)
	cnt_rtx = gen_rtx_PLUS (Pmode, pic_offset_table_rtx, cnt_rtx);

      xops[0] = gen_rtx_MEM (SImode, cnt_rtx);
      output_asm_insn ("inc{l}\t%0", xops);

      break;
    }
}
\f
void
ix86_expand_move (mode, operands)
     enum machine_mode mode;
     rtx operands[];
{
  int strict = (reload_in_progress || reload_completed);
  rtx insn;

  if (flag_pic && mode == Pmode && symbolic_operand (operands[1], Pmode))
    {
      /* Emit insns to move operands[1] into operands[0].  */

      if (GET_CODE (operands[0]) == MEM)
	operands[1] = force_reg (Pmode, operands[1]);
      else
	{
	  rtx temp = operands[0];
	  if (GET_CODE (temp) != REG)
	    temp = gen_reg_rtx (Pmode);
	  temp = legitimize_pic_address (operands[1], temp);
	  if (temp == operands[0])
	    return;
	  operands[1] = temp;
	}
    }
  else
    {
      if (GET_CODE (operands[0]) == MEM
	  && (GET_MODE (operands[0]) == QImode
	      || !push_operand (operands[0], mode))
	  && GET_CODE (operands[1]) == MEM)
	operands[1] = force_reg (mode, operands[1]);

      if (push_operand (operands[0], mode)
	  && ! general_no_elim_operand (operands[1], mode))
	operands[1] = copy_to_mode_reg (mode, operands[1]);

      if (FLOAT_MODE_P (mode))
	{
	  /* If we are loading a floating point constant to a register,
	     force the value to memory now, since we'll get better code
	     out the back end.  */

	  if (strict)
	    ;
	  else if (GET_CODE (operands[1]) == CONST_DOUBLE
		   && register_operand (operands[0], mode))
	    operands[1] = validize_mem (force_const_mem (mode, operands[1]));
	}
    }

  insn = gen_rtx_SET (VOIDmode, operands[0], operands[1]);

  emit_insn (insn);
}

/* Attempt to expand a binary operator.  Make the expansion closer to the
   actual machine, then just general_operand, which will allow 3 separate
   memory references (one output, two input) in a single insn.  */

void
ix86_expand_binary_operator (code, mode, operands)
     enum rtx_code code;
     enum machine_mode mode;
     rtx operands[];
{
  int matching_memory;
  rtx src1, src2, dst, op, clob;

  dst = operands[0];
  src1 = operands[1];
  src2 = operands[2];

  /* Recognize <var1> = <value> <op> <var1> for commutative operators */
  if (GET_RTX_CLASS (code) == 'c'
      && (rtx_equal_p (dst, src2)
	  || immediate_operand (src1, mode)))
    {
      rtx temp = src1;
      src1 = src2;
      src2 = temp;
    }

  /* If the destination is memory, and we do not have matching source
     operands, do things in registers.  */
  matching_memory = 0;
  if (GET_CODE (dst) == MEM)
    {
      if (rtx_equal_p (dst, src1))
	matching_memory = 1;
      else if (GET_RTX_CLASS (code) == 'c'
	       && rtx_equal_p (dst, src2))
	matching_memory = 2;
      else
	dst = gen_reg_rtx (mode);
    }

  /* Both source operands cannot be in memory.  */
  if (GET_CODE (src1) == MEM && GET_CODE (src2) == MEM)
    {
      if (matching_memory != 2)
	src2 = force_reg (mode, src2);
      else
	src1 = force_reg (mode, src1);
    }

  /* If the operation is not commutable, source 1 cannot be a constant
     or non-matching memory.  */
  if ((CONSTANT_P (src1)
       || (!matching_memory && GET_CODE (src1) == MEM))
      && GET_RTX_CLASS (code) != 'c')
    src1 = force_reg (mode, src1);

  /* If optimizing, copy to regs to improve CSE */
  if (optimize && ! no_new_pseudos)
    {
      if (GET_CODE (dst) == MEM)
	dst = gen_reg_rtx (mode);
      if (GET_CODE (src1) == MEM)
	src1 = force_reg (mode, src1);
      if (GET_CODE (src2) == MEM)
	src2 = force_reg (mode, src2);
    }

  /* Emit the instruction.  */

  op = gen_rtx_SET (VOIDmode, dst, gen_rtx_fmt_ee (code, mode, src1, src2));
  if (reload_in_progress)
    {
      /* Reload doesn't know about the flags register, and doesn't know that
         it doesn't want to clobber it.  We can only do this with PLUS.  */
      if (code != PLUS)
	abort ();
      emit_insn (op);
    }
  else
    {
      clob = gen_rtx_CLOBBER (VOIDmode, gen_rtx_REG (CCmode, FLAGS_REG));
      emit_insn (gen_rtx_PARALLEL (VOIDmode, gen_rtvec (2, op, clob)));
    }

  /* Fix up the destination if needed.  */
  if (dst != operands[0])
    emit_move_insn (operands[0], dst);
}

/* Return TRUE or FALSE depending on whether the binary operator meets the
   appropriate constraints.  */

int
ix86_binary_operator_ok (code, mode, operands)
     enum rtx_code code;
     enum machine_mode mode ATTRIBUTE_UNUSED;
     rtx operands[3];
{
  /* Both source operands cannot be in memory.  */
  if (GET_CODE (operands[1]) == MEM && GET_CODE (operands[2]) == MEM)
    return 0;
  /* If the operation is not commutable, source 1 cannot be a constant.  */
  if (CONSTANT_P (operands[1]) && GET_RTX_CLASS (code) != 'c')
    return 0;
  /* If the destination is memory, we must have a matching source operand.  */
  if (GET_CODE (operands[0]) == MEM
      && ! (rtx_equal_p (operands[0], operands[1])
	    || (GET_RTX_CLASS (code) == 'c'
		&& rtx_equal_p (operands[0], operands[2]))))
    return 0;
  /* If the operation is not commutable and the source 1 is memory, we must
     have a matching destionation.  */
  if (GET_CODE (operands[1]) == MEM
      && GET_RTX_CLASS (code) != 'c'
      && ! rtx_equal_p (operands[0], operands[1]))
    return 0;
  return 1;
}

/* Attempt to expand a unary operator.  Make the expansion closer to the
   actual machine, then just general_operand, which will allow 2 separate
   memory references (one output, one input) in a single insn.  */

void
ix86_expand_unary_operator (code, mode, operands)
     enum rtx_code code;
     enum machine_mode mode;
     rtx operands[];
{
  int matching_memory;
  rtx src, dst, op, clob;

  dst = operands[0];
  src = operands[1];

  /* If the destination is memory, and we do not have matching source
     operands, do things in registers.  */
  matching_memory = 0;
  if (GET_CODE (dst) == MEM)
    {
      if (rtx_equal_p (dst, src))
	matching_memory = 1;
      else
	dst = gen_reg_rtx (mode);
    }

  /* When source operand is memory, destination must match.  */
  if (!matching_memory && GET_CODE (src) == MEM)
    src = force_reg (mode, src);

  /* If optimizing, copy to regs to improve CSE */
  if (optimize && ! no_new_pseudos)
    {
      if (GET_CODE (dst) == MEM)
	dst = gen_reg_rtx (mode);
      if (GET_CODE (src) == MEM)
	src = force_reg (mode, src);
    }

  /* Emit the instruction.  */

  op = gen_rtx_SET (VOIDmode, dst, gen_rtx_fmt_e (code, mode, src));
  if (reload_in_progress || code == NOT)
    {
      /* Reload doesn't know about the flags register, and doesn't know that
         it doesn't want to clobber it.  */
      if (code != NOT)
        abort ();
      emit_insn (op);
    }
  else
    {
      clob = gen_rtx_CLOBBER (VOIDmode, gen_rtx_REG (CCmode, FLAGS_REG));
      emit_insn (gen_rtx_PARALLEL (VOIDmode, gen_rtvec (2, op, clob)));
    }

  /* Fix up the destination if needed.  */
  if (dst != operands[0])
    emit_move_insn (operands[0], dst);
}

/* Return TRUE or FALSE depending on whether the unary operator meets the
   appropriate constraints.  */

int
ix86_unary_operator_ok (code, mode, operands)
     enum rtx_code code ATTRIBUTE_UNUSED;
     enum machine_mode mode ATTRIBUTE_UNUSED;
     rtx operands[2] ATTRIBUTE_UNUSED;
{
  /* If one of operands is memory, source and destination must match.  */
  if ((GET_CODE (operands[0]) == MEM
       || GET_CODE (operands[1]) == MEM)
      && ! rtx_equal_p (operands[0], operands[1]))
    return FALSE;
  return TRUE;
}

/* Return TRUE or FALSE depending on whether the first SET in INSN
   has source and destination with matching CC modes, and that the
   CC mode is at least as constrained as REQ_MODE.  */

int
ix86_match_ccmode (insn, req_mode)
     rtx insn;
     enum machine_mode req_mode;
{
  rtx set;
  enum machine_mode set_mode;

  set = PATTERN (insn);
  if (GET_CODE (set) == PARALLEL)
    set = XVECEXP (set, 0, 0);
  if (GET_CODE (set) != SET)
    abort ();
  if (GET_CODE (SET_SRC (set)) != COMPARE)
    abort ();

  set_mode = GET_MODE (SET_DEST (set));
  switch (set_mode)
    {
    case CCNOmode:
      if (req_mode != CCNOmode
	  && (req_mode != CCmode
	      || XEXP (SET_SRC (set), 1) != const0_rtx))
	return 0;
      break;
    case CCmode:
      if (req_mode == CCGCmode)
	return 0;
      /* FALLTHRU */
    case CCGCmode:
      if (req_mode == CCGOCmode || req_mode == CCNOmode)
	return 0;
      /* FALLTHRU */
    case CCGOCmode:
      if (req_mode == CCZmode)
	return 0;
      /* FALLTHRU */
    case CCZmode:
      break;

    default:
      abort ();
    }

  return (GET_MODE (SET_SRC (set)) == set_mode);
}

/* Generate insn patterns to do an integer compare of OPERANDS.  */

static rtx
ix86_expand_int_compare (code, op0, op1)
     enum rtx_code code;
     rtx op0, op1;
{
  enum machine_mode cmpmode;
  rtx tmp, flags;

  cmpmode = SELECT_CC_MODE (code, op0, op1);
  flags = gen_rtx_REG (cmpmode, FLAGS_REG);

  /* This is very simple, but making the interface the same as in the
     FP case makes the rest of the code easier.  */
  tmp = gen_rtx_COMPARE (cmpmode, op0, op1);
  emit_insn (gen_rtx_SET (VOIDmode, flags, tmp));

  /* Return the test that should be put into the flags user, i.e.
     the bcc, scc, or cmov instruction.  */
  return gen_rtx_fmt_ee (code, VOIDmode, flags, const0_rtx);
}

/* Figure out whether to use ordered or unordered fp comparisons.
   Return the appropriate mode to use.  */

enum machine_mode
ix86_fp_compare_mode (code)
     enum rtx_code code ATTRIBUTE_UNUSED;
{
  /* ??? In order to make all comparisons reversible, we do all comparisons
     non-trapping when compiling for IEEE.  Once gcc is able to distinguish
     all forms trapping and nontrapping comparisons, we can make inequality
     comparisons trapping again, since it results in better code when using
     FCOM based compares.  */
  return TARGET_IEEE_FP ? CCFPUmode : CCFPmode;
}

enum machine_mode
ix86_cc_mode (code, op0, op1)
     enum rtx_code code;
     rtx op0, op1;
{
  if (GET_MODE_CLASS (GET_MODE (op0)) == MODE_FLOAT)
    return ix86_fp_compare_mode (code);
  switch (code)
    {
      /* Only zero flag is needed.  */
    case EQ:			/* ZF=0 */
    case NE:			/* ZF!=0 */
      return CCZmode;
      /* Codes needing carry flag.  */
    case GEU:			/* CF=0 */
    case GTU:			/* CF=0 & ZF=0 */
    case LTU:			/* CF=1 */
    case LEU:			/* CF=1 | ZF=1 */
      return CCmode;
      /* Codes possibly doable only with sign flag when
         comparing against zero.  */
    case GE:			/* SF=OF   or   SF=0 */
    case LT:			/* SF<>OF  or   SF=1 */
      if (op1 == const0_rtx)
	return CCGOCmode;
      else
	/* For other cases Carry flag is not required.  */
	return CCGCmode;
      /* Codes doable only with sign flag when comparing
         against zero, but we miss jump instruction for it
         so we need to use relational tests agains overflow
         that thus needs to be zero.  */
    case GT:			/* ZF=0 & SF=OF */
    case LE:			/* ZF=1 | SF<>OF */
      if (op1 == const0_rtx)
	return CCNOmode;
      else
	return CCGCmode;
    default:
      abort ();
    }
}

/* Return true if we should use an FCOMI instruction for this fp comparison.  */

int
ix86_use_fcomi_compare (code)
     enum rtx_code code ATTRIBUTE_UNUSED;
{
  enum rtx_code swapped_code = swap_condition (code);
  return ((ix86_fp_comparison_cost (code) == ix86_fp_comparison_fcomi_cost (code))
	  || (ix86_fp_comparison_cost (swapped_code)
	      == ix86_fp_comparison_fcomi_cost (swapped_code)));
}

/* Swap, force into registers, or otherwise massage the two operands
   to a fp comparison.  The operands are updated in place; the new
   comparsion code is returned.  */

static enum rtx_code
ix86_prepare_fp_compare_args (code, pop0, pop1)
     enum rtx_code code;
     rtx *pop0, *pop1;
{
  enum machine_mode fpcmp_mode = ix86_fp_compare_mode (code);
  rtx op0 = *pop0, op1 = *pop1;
  enum machine_mode op_mode = GET_MODE (op0);
  int is_sse = SSE_REG_P (op0) | SSE_REG_P (op1);

  /* All of the unordered compare instructions only work on registers.
     The same is true of the XFmode compare instructions.  The same is
     true of the fcomi compare instructions.  */

  if (!is_sse
      && (fpcmp_mode == CCFPUmode
	  || op_mode == XFmode
	  || op_mode == TFmode
	  || ix86_use_fcomi_compare (code)))
    {
      op0 = force_reg (op_mode, op0);
      op1 = force_reg (op_mode, op1);
    }
  else
    {
      /* %%% We only allow op1 in memory; op0 must be st(0).  So swap
	 things around if they appear profitable, otherwise force op0
	 into a register.  */

      if (standard_80387_constant_p (op0) == 0
	  || (GET_CODE (op0) == MEM
	      && ! (standard_80387_constant_p (op1) == 0
		    || GET_CODE (op1) == MEM)))
	{
	  rtx tmp;
	  tmp = op0, op0 = op1, op1 = tmp;
	  code = swap_condition (code);
	}

      if (GET_CODE (op0) != REG)
	op0 = force_reg (op_mode, op0);

      if (CONSTANT_P (op1))
	{
	  if (standard_80387_constant_p (op1))
	    op1 = force_reg (op_mode, op1);
	  else
	    op1 = validize_mem (force_const_mem (op_mode, op1));
	}
    }

  /* Try to rearrange the comparison to make it cheaper.  */
  if (ix86_fp_comparison_cost (code)
      > ix86_fp_comparison_cost (swap_condition (code))
      && (GET_CODE (op0) == REG || !reload_completed))
    {
      rtx tmp;
      tmp = op0, op0 = op1, op1 = tmp;
      code = swap_condition (code);
      if (GET_CODE (op0) != REG)
	op0 = force_reg (op_mode, op0);
    }

  *pop0 = op0;
  *pop1 = op1;
  return code;
}

/* Convert comparison codes we use to represent FP comparison to integer
   code that will result in proper branch.  Return UNKNOWN if no such code
   is available.  */
static enum rtx_code
ix86_fp_compare_code_to_integer (code)
     enum rtx_code code;
{
  switch (code)
    {
    case GT:
      return GTU;
    case GE:
      return GEU;
    case ORDERED:
    case UNORDERED:
      return code;
      break;
    case UNEQ:
      return EQ;
      break;
    case UNLT:
      return LTU;
      break;
    case UNLE:
      return LEU;
      break;
    case LTGT:
      return NE;
      break;
    default:
      return UNKNOWN;
    }
}

/* Split comparison code CODE into comparisons we can do using branch
   instructions.  BYPASS_CODE is comparison code for branch that will
   branch around FIRST_CODE and SECOND_CODE.  If some of branches
   is not required, set value to NIL.
   We never require more than two branches.  */
static void
ix86_fp_comparison_codes (code, bypass_code, first_code, second_code)
     enum rtx_code code, *bypass_code, *first_code, *second_code;
{
  *first_code = code;
  *bypass_code = NIL;
  *second_code = NIL;

  /* The fcomi comparison sets flags as follows:

     cmp    ZF PF CF
     >      0  0  0
     <      0  0  1
     =      1  0  0
     un     1  1  1 */

  switch (code)
    {
    case GT:			/* GTU - CF=0 & ZF=0 */
    case GE:			/* GEU - CF=0 */
    case ORDERED:		/* PF=0 */
    case UNORDERED:		/* PF=1 */
    case UNEQ:			/* EQ - ZF=1 */
    case UNLT:			/* LTU - CF=1 */
    case UNLE:			/* LEU - CF=1 | ZF=1 */
    case LTGT:			/* EQ - ZF=0 */
      break;
    case LT:			/* LTU - CF=1 - fails on unordered */
      *first_code = UNLT;
      *bypass_code = UNORDERED;
      break;
    case LE:			/* LEU - CF=1 | ZF=1 - fails on unordered */
      *first_code = UNLE;
      *bypass_code = UNORDERED;
      break;
    case EQ:			/* EQ - ZF=1 - fails on unordered */
      *first_code = UNEQ;
      *bypass_code = UNORDERED;
      break;
    case NE:			/* NE - ZF=0 - fails on unordered */
      *first_code = LTGT;
      *second_code = UNORDERED;
      break;
    case UNGE:			/* GEU - CF=0 - fails on unordered */
      *first_code = GE;
      *second_code = UNORDERED;
      break;
    case UNGT:			/* GTU - CF=0 & ZF=0 - fails on unordered */
      *first_code = GT;
      *second_code = UNORDERED;
      break;
    default:
      abort ();
    }
  if (!TARGET_IEEE_FP)
    {
      *second_code = NIL;
      *bypass_code = NIL;
    }
}

/* Return cost of comparison done fcom + arithmetics operations on AX.
   All following functions do use number of instructions as an cost metrics.
   In future this should be tweaked to compute bytes for optimize_size and
   take into account performance of various instructions on various CPUs.  */
static int
ix86_fp_comparison_arithmetics_cost (code)
     enum rtx_code code;
{
  if (!TARGET_IEEE_FP)
    return 4;
  /* The cost of code output by ix86_expand_fp_compare.  */
  switch (code)
    {
    case UNLE:
    case UNLT:
    case LTGT:
    case GT:
    case GE:
    case UNORDERED:
    case ORDERED:
    case UNEQ:
      return 4;
      break;
    case LT:
    case NE:
    case EQ:
    case UNGE:
      return 5;
      break;
    case LE:
    case UNGT:
      return 6;
      break;
    default:
      abort ();
    }
}

/* Return cost of comparison done using fcomi operation.
   See ix86_fp_comparison_arithmetics_cost for the metrics.  */
static int
ix86_fp_comparison_fcomi_cost (code)
     enum rtx_code code;
{
  enum rtx_code bypass_code, first_code, second_code;
  /* Return arbitarily high cost when instruction is not supported - this
     prevents gcc from using it.  */
  if (!TARGET_CMOVE)
    return 1024;
  ix86_fp_comparison_codes (code, &bypass_code, &first_code, &second_code);
  return (bypass_code != NIL || second_code != NIL) + 2;
}

/* Return cost of comparison done using sahf operation.
   See ix86_fp_comparison_arithmetics_cost for the metrics.  */
static int
ix86_fp_comparison_sahf_cost (code)
     enum rtx_code code;
{
  enum rtx_code bypass_code, first_code, second_code;
  /* Return arbitarily high cost when instruction is not preferred - this
     avoids gcc from using it.  */
  if (!TARGET_USE_SAHF && !optimize_size)
    return 1024;
  ix86_fp_comparison_codes (code, &bypass_code, &first_code, &second_code);
  return (bypass_code != NIL || second_code != NIL) + 3;
}

/* Compute cost of the comparison done using any method.
   See ix86_fp_comparison_arithmetics_cost for the metrics.  */
static int
ix86_fp_comparison_cost (code)
     enum rtx_code code;
{
  int fcomi_cost, sahf_cost, arithmetics_cost = 1024;
  int min;

  fcomi_cost = ix86_fp_comparison_fcomi_cost (code);
  sahf_cost = ix86_fp_comparison_sahf_cost (code);

  min = arithmetics_cost = ix86_fp_comparison_arithmetics_cost (code);
  if (min > sahf_cost)
    min = sahf_cost;
  if (min > fcomi_cost)
    min = fcomi_cost;
  return min;
}

/* Generate insn patterns to do a floating point compare of OPERANDS.  */

static rtx
ix86_expand_fp_compare (code, op0, op1, scratch, second_test, bypass_test)
     enum rtx_code code;
     rtx op0, op1, scratch;
     rtx *second_test;
     rtx *bypass_test;
{
  enum machine_mode fpcmp_mode, intcmp_mode;
  rtx tmp, tmp2;
  int cost = ix86_fp_comparison_cost (code);
  enum rtx_code bypass_code, first_code, second_code;

  fpcmp_mode = ix86_fp_compare_mode (code);
  code = ix86_prepare_fp_compare_args (code, &op0, &op1);

  if (second_test)
    *second_test = NULL_RTX;
  if (bypass_test)
    *bypass_test = NULL_RTX;

  ix86_fp_comparison_codes (code, &bypass_code, &first_code, &second_code);

  /* Do fcomi/sahf based test when profitable.  */
  if ((bypass_code == NIL || bypass_test)
      && (second_code == NIL || second_test)
      && ix86_fp_comparison_arithmetics_cost (code) > cost)
    {
      if (TARGET_CMOVE)
	{
	  tmp = gen_rtx_COMPARE (fpcmp_mode, op0, op1);
	  tmp = gen_rtx_SET (VOIDmode, gen_rtx_REG (fpcmp_mode, FLAGS_REG),
			     tmp);
	  emit_insn (tmp);
	}
      else
	{
	  tmp = gen_rtx_COMPARE (fpcmp_mode, op0, op1);
	  tmp2 = gen_rtx_UNSPEC (HImode, gen_rtvec (1, tmp), 9);
	  emit_insn (gen_rtx_SET (VOIDmode, scratch, tmp2));
	  emit_insn (gen_x86_sahf_1 (scratch));
	}

      /* The FP codes work out to act like unsigned.  */
      intcmp_mode = fpcmp_mode;
      code = first_code;
      if (bypass_code != NIL)
	*bypass_test = gen_rtx_fmt_ee (bypass_code, VOIDmode,
				       gen_rtx_REG (intcmp_mode, FLAGS_REG),
				       const0_rtx);
      if (second_code != NIL)
	*second_test = gen_rtx_fmt_ee (second_code, VOIDmode,
				       gen_rtx_REG (intcmp_mode, FLAGS_REG),
				       const0_rtx);
    }
  else
    {
      /* Sadness wrt reg-stack pops killing fpsr -- gotta get fnstsw first.  */
      tmp = gen_rtx_COMPARE (fpcmp_mode, op0, op1);
      tmp2 = gen_rtx_UNSPEC (HImode, gen_rtvec (1, tmp), 9);
      emit_insn (gen_rtx_SET (VOIDmode, scratch, tmp2));

      /* In the unordered case, we have to check C2 for NaN's, which
	 doesn't happen to work out to anything nice combination-wise.
	 So do some bit twiddling on the value we've got in AH to come
	 up with an appropriate set of condition codes.  */

      intcmp_mode = CCNOmode;
      switch (code)
	{
	case GT:
	case UNGT:
	  if (code == GT || !TARGET_IEEE_FP)
	    {
	      emit_insn (gen_testqi_ext_ccno_0 (scratch, GEN_INT (0x45)));
	      code = EQ;
	    }
	  else
	    {
	      emit_insn (gen_andqi_ext_0 (scratch, scratch, GEN_INT (0x45)));
	      emit_insn (gen_addqi_ext_1 (scratch, scratch, constm1_rtx));
	      emit_insn (gen_cmpqi_ext_3 (scratch, GEN_INT (0x44)));
	      intcmp_mode = CCmode;
	      code = GEU;
	    }
	  break;
	case LT:
	case UNLT:
	  if (code == LT && TARGET_IEEE_FP)
	    {
	      emit_insn (gen_andqi_ext_0 (scratch, scratch, GEN_INT (0x45)));
	      emit_insn (gen_cmpqi_ext_3 (scratch, GEN_INT (0x01)));
	      intcmp_mode = CCmode;
	      code = EQ;
	    }
	  else
	    {
	      emit_insn (gen_testqi_ext_ccno_0 (scratch, GEN_INT (0x01)));
	      code = NE;
	    }
	  break;
	case GE:
	case UNGE:
	  if (code == GE || !TARGET_IEEE_FP)
	    {
	      emit_insn (gen_testqi_ext_ccno_0 (scratch, GEN_INT (0x05)));
	      code = EQ;
	    }
	  else
	    {
	      emit_insn (gen_andqi_ext_0 (scratch, scratch, GEN_INT (0x45)));
	      emit_insn (gen_xorqi_cc_ext_1 (scratch, scratch,
					     GEN_INT (0x01)));
	      code = NE;
	    }
	  break;
	case LE:
	case UNLE:
	  if (code == LE && TARGET_IEEE_FP)
	    {
	      emit_insn (gen_andqi_ext_0 (scratch, scratch, GEN_INT (0x45)));
	      emit_insn (gen_addqi_ext_1 (scratch, scratch, constm1_rtx));
	      emit_insn (gen_cmpqi_ext_3 (scratch, GEN_INT (0x40)));
	      intcmp_mode = CCmode;
	      code = LTU;
	    }
	  else
	    {
	      emit_insn (gen_testqi_ext_ccno_0 (scratch, GEN_INT (0x45)));
	      code = NE;
	    }
	  break;
	case EQ:
	case UNEQ:
	  if (code == EQ && TARGET_IEEE_FP)
	    {
	      emit_insn (gen_andqi_ext_0 (scratch, scratch, GEN_INT (0x45)));
	      emit_insn (gen_cmpqi_ext_3 (scratch, GEN_INT (0x40)));
	      intcmp_mode = CCmode;
	      code = EQ;
	    }
	  else
	    {
	      emit_insn (gen_testqi_ext_ccno_0 (scratch, GEN_INT (0x40)));
	      code = NE;
	      break;
	    }
	  break;
	case NE:
	case LTGT:
	  if (code == NE && TARGET_IEEE_FP)
	    {
	      emit_insn (gen_andqi_ext_0 (scratch, scratch, GEN_INT (0x45)));
	      emit_insn (gen_xorqi_cc_ext_1 (scratch, scratch,
					     GEN_INT (0x40)));
	      code = NE;
	    }
	  else
	    {
	      emit_insn (gen_testqi_ext_ccno_0 (scratch, GEN_INT (0x40)));
	      code = EQ;
	    }
	  break;

	case UNORDERED:
	  emit_insn (gen_testqi_ext_ccno_0 (scratch, GEN_INT (0x04)));
	  code = NE;
	  break;
	case ORDERED:
	  emit_insn (gen_testqi_ext_ccno_0 (scratch, GEN_INT (0x04)));
	  code = EQ;
	  break;

	default:
	  abort ();
	}
    }

  /* Return the test that should be put into the flags user, i.e.
     the bcc, scc, or cmov instruction.  */
  return gen_rtx_fmt_ee (code, VOIDmode,
			 gen_rtx_REG (intcmp_mode, FLAGS_REG),
			 const0_rtx);
}

rtx
ix86_expand_compare (code, second_test, bypass_test)
     enum rtx_code code;
     rtx *second_test, *bypass_test;
{
  rtx op0, op1, ret;
  op0 = ix86_compare_op0;
  op1 = ix86_compare_op1;

  if (second_test)
    *second_test = NULL_RTX;
  if (bypass_test)
    *bypass_test = NULL_RTX;

  if (GET_MODE_CLASS (GET_MODE (op0)) == MODE_FLOAT)
    ret = ix86_expand_fp_compare (code, op0, op1, gen_reg_rtx (HImode),
				  second_test, bypass_test);
  else
    ret = ix86_expand_int_compare (code, op0, op1);

  return ret;
}

void
ix86_expand_branch (code, label)
     enum rtx_code code;
     rtx label;
{
  rtx tmp;

  switch (GET_MODE (ix86_compare_op0))
    {
    case QImode:
    case HImode:
    case SImode:
      tmp = ix86_expand_compare (code, NULL, NULL);
      tmp = gen_rtx_IF_THEN_ELSE (VOIDmode, tmp,
				  gen_rtx_LABEL_REF (VOIDmode, label),
				  pc_rtx);
      emit_jump_insn (gen_rtx_SET (VOIDmode, pc_rtx, tmp));
      return;

    case SFmode:
    case DFmode:
    case XFmode:
    case TFmode:
      /* Don't expand the comparison early, so that we get better code
	 when jump or whoever decides to reverse the comparison.  */
      {
	rtvec vec;
	int use_fcomi;

	code = ix86_prepare_fp_compare_args (code, &ix86_compare_op0,
					     &ix86_compare_op1);

	tmp = gen_rtx_fmt_ee (code, VOIDmode,
			      ix86_compare_op0, ix86_compare_op1);
	tmp = gen_rtx_IF_THEN_ELSE (VOIDmode, tmp,
				    gen_rtx_LABEL_REF (VOIDmode, label),
				    pc_rtx);
	tmp = gen_rtx_SET (VOIDmode, pc_rtx, tmp);

	use_fcomi = ix86_use_fcomi_compare (code);
	vec = rtvec_alloc (3 + !use_fcomi);
	RTVEC_ELT (vec, 0) = tmp;
	RTVEC_ELT (vec, 1)
	  = gen_rtx_CLOBBER (VOIDmode, gen_rtx_REG (CCFPmode, 18));
	RTVEC_ELT (vec, 2)
	  = gen_rtx_CLOBBER (VOIDmode, gen_rtx_REG (CCFPmode, 17));
	if (! use_fcomi)
	  RTVEC_ELT (vec, 3)
	    = gen_rtx_CLOBBER (VOIDmode, gen_rtx_SCRATCH (HImode));

        emit_jump_insn (gen_rtx_PARALLEL (VOIDmode, vec));
	return;
      }

    case DImode:
      /* Expand DImode branch into multiple compare+branch.  */
      {
	rtx lo[2], hi[2], label2;
	enum rtx_code code1, code2, code3;

	if (CONSTANT_P (ix86_compare_op0) && ! CONSTANT_P (ix86_compare_op1))
	  {
	    tmp = ix86_compare_op0;
	    ix86_compare_op0 = ix86_compare_op1;
	    ix86_compare_op1 = tmp;
	    code = swap_condition (code);
	  }
	split_di (&ix86_compare_op0, 1, lo+0, hi+0);
	split_di (&ix86_compare_op1, 1, lo+1, hi+1);

	/* When comparing for equality, we can use (hi0^hi1)|(lo0^lo1) to
	   avoid two branches.  This costs one extra insn, so disable when
	   optimizing for size.  */

	if ((code == EQ || code == NE)
	    && (!optimize_size
	        || hi[1] == const0_rtx || lo[1] == const0_rtx))
	  {
	    rtx xor0, xor1;

	    xor1 = hi[0];
	    if (hi[1] != const0_rtx)
	      xor1 = expand_binop (SImode, xor_optab, xor1, hi[1],
				   NULL_RTX, 0, OPTAB_WIDEN);

	    xor0 = lo[0];
	    if (lo[1] != const0_rtx)
	      xor0 = expand_binop (SImode, xor_optab, xor0, lo[1],
				   NULL_RTX, 0, OPTAB_WIDEN);

	    tmp = expand_binop (SImode, ior_optab, xor1, xor0,
				NULL_RTX, 0, OPTAB_WIDEN);

	    ix86_compare_op0 = tmp;
	    ix86_compare_op1 = const0_rtx;
	    ix86_expand_branch (code, label);
	    return;
	  }

	/* Otherwise, if we are doing less-than or greater-or-equal-than,
	   op1 is a constant and the low word is zero, then we can just
	   examine the high word.  */

	if (GET_CODE (hi[1]) == CONST_INT && lo[1] == const0_rtx)
	  switch (code)
	    {
	    case LT: case LTU: case GE: case GEU:
	      ix86_compare_op0 = hi[0];
	      ix86_compare_op1 = hi[1];
	      ix86_expand_branch (code, label);
	      return;
	    default:
	      break;
	    }

	/* Otherwise, we need two or three jumps.  */

	label2 = gen_label_rtx ();

	code1 = code;
	code2 = swap_condition (code);
	code3 = unsigned_condition (code);

	switch (code)
	  {
	  case LT: case GT: case LTU: case GTU:
	    break;

	  case LE:   code1 = LT;  code2 = GT;  break;
	  case GE:   code1 = GT;  code2 = LT;  break;
	  case LEU:  code1 = LTU; code2 = GTU; break;
	  case GEU:  code1 = GTU; code2 = LTU; break;

	  case EQ:   code1 = NIL; code2 = NE;  break;
	  case NE:   code2 = NIL; break;

	  default:
	    abort ();
	  }

	/*
	 * a < b =>
	 *    if (hi(a) < hi(b)) goto true;
	 *    if (hi(a) > hi(b)) goto false;
	 *    if (lo(a) < lo(b)) goto true;
	 *  false:
	 */

	ix86_compare_op0 = hi[0];
	ix86_compare_op1 = hi[1];

	if (code1 != NIL)
	  ix86_expand_branch (code1, label);
	if (code2 != NIL)
	  ix86_expand_branch (code2, label2);

	ix86_compare_op0 = lo[0];
	ix86_compare_op1 = lo[1];
	ix86_expand_branch (code3, label);

	if (code2 != NIL)
	  emit_label (label2);
	return;
      }

    default:
      abort ();
    }
}

/* Split branch based on floating point condition.  */
void
ix86_split_fp_branch (condition, op1, op2, target1, target2, tmp)
     rtx condition, op1, op2, target1, target2, tmp;
{
  rtx second, bypass;
  rtx label = NULL_RTX;
  enum rtx_code code = GET_CODE (condition);

  if (target2 != pc_rtx)
    {
      rtx tmp = target2;
      code = reverse_condition_maybe_unordered (code);
      target2 = target1;
      target1 = tmp;
    }

  condition = ix86_expand_fp_compare (code, op1, op2,
				      tmp, &second, &bypass);
  if (bypass != NULL_RTX)
    {
      label = gen_label_rtx ();
      emit_jump_insn (gen_rtx_SET
		      (VOIDmode, pc_rtx,
		       gen_rtx_IF_THEN_ELSE (VOIDmode,
					     bypass,
					     gen_rtx_LABEL_REF (VOIDmode,
								label),
					     pc_rtx)));
    }
  /* AMD Athlon and probably other CPUs too have fast bypass path between the
     comparison and first branch.  The second branch takes longer to execute
     so place first branch the worse predicable one if possible.  */
  if (second != NULL_RTX
      && (GET_CODE (second) == UNORDERED || GET_CODE (second) == ORDERED))
    {
      rtx tmp = condition;
      condition = second;
      second = tmp;
    }
  emit_jump_insn (gen_rtx_SET
		  (VOIDmode, pc_rtx,
		   gen_rtx_IF_THEN_ELSE (VOIDmode,
					 condition, target1, target2)));
  if (second != NULL_RTX)
    emit_jump_insn (gen_rtx_SET
		    (VOIDmode, pc_rtx,
		     gen_rtx_IF_THEN_ELSE (VOIDmode, second, target1, target2)));
  if (label != NULL_RTX)
    emit_label (label);
}

int
ix86_expand_setcc (code, dest)
     enum rtx_code code;
     rtx dest;
{
  rtx ret, tmp, tmpreg;
  rtx second_test, bypass_test;
  int type;

  if (GET_MODE (ix86_compare_op0) == DImode)
    return 0; /* FAIL */

  /* Three modes of generation:
     0 -- destination does not overlap compare sources:
          clear dest first, emit strict_low_part setcc.
     1 -- destination does overlap compare sources:
          emit subreg setcc, zero extend.
     2 -- destination is in QImode:
          emit setcc only.
  */

  type = 0;

  if (GET_MODE (dest) == QImode)
    type = 2;
  else if (reg_overlap_mentioned_p (dest, ix86_compare_op0)
	   || reg_overlap_mentioned_p (dest, ix86_compare_op1))
    type = 1;

  if (type == 0)
    emit_move_insn (dest, const0_rtx);

  ret = ix86_expand_compare (code, &second_test, &bypass_test);
  PUT_MODE (ret, QImode);

  tmp = dest;
  tmpreg = dest;
  if (type == 0)
    {
      tmp = gen_lowpart (QImode, dest);
      tmpreg = tmp;
      tmp = gen_rtx_STRICT_LOW_PART (VOIDmode, tmp);
    }
  else if (type == 1)
    {
      if (!cse_not_expected)
	tmp = gen_reg_rtx (QImode);
      else
        tmp = gen_lowpart (QImode, dest);
      tmpreg = tmp;
    }

  emit_insn (gen_rtx_SET (VOIDmode, tmp, ret));
  if (bypass_test || second_test)
    {
      rtx test = second_test;
      int bypass = 0;
      rtx tmp2 = gen_reg_rtx (QImode);
      if (bypass_test)
	{
	  if (second_test)
	    abort();
	  test = bypass_test;
	  bypass = 1;
	  PUT_CODE (test, reverse_condition_maybe_unordered (GET_CODE (test)));
	}
      PUT_MODE (test, QImode);
      emit_insn (gen_rtx_SET (VOIDmode, tmp2, test));

      if (bypass)
	emit_insn (gen_andqi3 (tmp, tmpreg, tmp2));
      else
	emit_insn (gen_iorqi3 (tmp, tmpreg, tmp2));
    }

  if (type == 1)
    {
      rtx clob;

      tmp = gen_rtx_ZERO_EXTEND (GET_MODE (dest), tmp);
      tmp = gen_rtx_SET (VOIDmode, dest, tmp);
      clob = gen_rtx_CLOBBER (VOIDmode, gen_rtx_REG (CCmode, FLAGS_REG));
      tmp = gen_rtx_PARALLEL (VOIDmode, gen_rtvec (2, tmp, clob));
      emit_insn (tmp);
    }

  return 1; /* DONE */
}

int
ix86_expand_int_movcc (operands)
     rtx operands[];
{
  enum rtx_code code = GET_CODE (operands[1]), compare_code;
  rtx compare_seq, compare_op;
  rtx second_test, bypass_test;

  /* When the compare code is not LTU or GEU, we can not use sbbl case.
     In case comparsion is done with immediate, we can convert it to LTU or
     GEU by altering the integer.  */

  if ((code == LEU || code == GTU)
      && GET_CODE (ix86_compare_op1) == CONST_INT
      && GET_MODE (operands[0]) != HImode
      && (unsigned int)INTVAL (ix86_compare_op1) != 0xffffffff
      && GET_CODE (operands[2]) == CONST_INT
      && GET_CODE (operands[3]) == CONST_INT)
    {
      if (code == LEU)
	code = LTU;
      else
	code = GEU;
      ix86_compare_op1 = GEN_INT (INTVAL (ix86_compare_op1) + 1);
    }

  start_sequence ();
  compare_op = ix86_expand_compare (code, &second_test, &bypass_test);
  compare_seq = gen_sequence ();
  end_sequence ();

  compare_code = GET_CODE (compare_op);

  /* Don't attempt mode expansion here -- if we had to expand 5 or 6
     HImode insns, we'd be swallowed in word prefix ops.  */

  if (GET_MODE (operands[0]) != HImode
      && GET_CODE (operands[2]) == CONST_INT
      && GET_CODE (operands[3]) == CONST_INT)
    {
      rtx out = operands[0];
      HOST_WIDE_INT ct = INTVAL (operands[2]);
      HOST_WIDE_INT cf = INTVAL (operands[3]);
      HOST_WIDE_INT diff;

      if ((compare_code == LTU || compare_code == GEU)
	  && !second_test && !bypass_test)
	{

	  /* Detect overlap between destination and compare sources.  */
	  rtx tmp = out;

	  /* To simplify rest of code, restrict to the GEU case.  */
	  if (compare_code == LTU)
	    {
	      int tmp = ct;
	      ct = cf;
	      cf = tmp;
	      compare_code = reverse_condition (compare_code);
	      code = reverse_condition (code);
	    }
	  diff = ct - cf;

	  if (reg_overlap_mentioned_p (out, ix86_compare_op0)
	      || reg_overlap_mentioned_p (out, ix86_compare_op1))
	    tmp = gen_reg_rtx (SImode);

	  emit_insn (compare_seq);
	  emit_insn (gen_x86_movsicc_0_m1 (tmp));

	  if (diff == 1)
	    {
	      /*
	       * cmpl op0,op1
	       * sbbl dest,dest
	       * [addl dest, ct]
	       *
	       * Size 5 - 8.
	       */
	      if (ct)
	        emit_insn (gen_addsi3 (out, out, GEN_INT (ct)));
	    }
	  else if (cf == -1)
	    {
	      /*
	       * cmpl op0,op1
	       * sbbl dest,dest
	       * orl $ct, dest
	       *
	       * Size 8.
	       */
	      emit_insn (gen_iorsi3 (out, out, GEN_INT (ct)));
	    }
	  else if (diff == -1 && ct)
	    {
	      /*
	       * cmpl op0,op1
	       * sbbl dest,dest
	       * xorl $-1, dest
	       * [addl dest, cf]
	       *
	       * Size 8 - 11.
	       */
	      emit_insn (gen_one_cmplsi2 (tmp, tmp));
	      if (cf)
	        emit_insn (gen_addsi3 (out, out, GEN_INT (cf)));
	    }
	  else
	    {
	      /*
	       * cmpl op0,op1
	       * sbbl dest,dest
	       * andl cf - ct, dest
	       * [addl dest, ct]
	       *
	       * Size 8 - 11.
	       */
	      emit_insn (gen_andsi3 (out, out, GEN_INT (cf - ct)));
	      if (ct)
	        emit_insn (gen_addsi3 (out, out, GEN_INT (ct)));
	    }

	  if (tmp != out)
	    emit_move_insn (out, tmp);

	  return 1; /* DONE */
	}

      diff = ct - cf;
      if (diff < 0)
	{
	  HOST_WIDE_INT tmp;
	  tmp = ct, ct = cf, cf = tmp;
	  diff = -diff;
	  if (FLOAT_MODE_P (GET_MODE (ix86_compare_op0)))
	    {
	      /* We may be reversing unordered compare to normal compare, that
		 is not valid in general (we may convert non-trapping condition
		 to trapping one), however on i386 we currently emit all
		 comparisons unordered.  */
	      compare_code = reverse_condition_maybe_unordered (compare_code);
	      code = reverse_condition_maybe_unordered (code);
	    }
	  else
	    {
	      compare_code = reverse_condition (compare_code);
	      code = reverse_condition (code);
	    }
	}
      if (diff == 1 || diff == 2 || diff == 4 || diff == 8
	  || diff == 3 || diff == 5 || diff == 9)
	{
	  /*
	   * xorl dest,dest
	   * cmpl op1,op2
	   * setcc dest
	   * lea cf(dest*(ct-cf)),dest
	   *
	   * Size 14.
	   *
	   * This also catches the degenerate setcc-only case.
	   */

	  rtx tmp;
	  int nops;

	  out = emit_store_flag (out, code, ix86_compare_op0,
				 ix86_compare_op1, VOIDmode, 0, 1);

	  nops = 0;
	  if (diff == 1)
	      tmp = out;
	  else
	    {
	      tmp = gen_rtx_MULT (SImode, out, GEN_INT (diff & ~1));
	      nops++;
	      if (diff & 1)
		{
		  tmp = gen_rtx_PLUS (SImode, tmp, out);
		  nops++;
		}
	    }
	  if (cf != 0)
	    {
	      tmp = gen_rtx_PLUS (SImode, tmp, GEN_INT (cf));
	      nops++;
	    }
	  if (tmp != out)
	    {
	      if (nops == 0)
		emit_move_insn (out, tmp);
	      else if (nops == 1)
		{
		  rtx clob;

		  clob = gen_rtx_REG (CCmode, FLAGS_REG);
		  clob = gen_rtx_CLOBBER (VOIDmode, clob);

		  tmp = gen_rtx_SET (VOIDmode, out, tmp);
		  tmp = gen_rtx_PARALLEL (VOIDmode, gen_rtvec (2, tmp, clob));
		  emit_insn (tmp);
		}
	      else
		emit_insn (gen_rtx_SET (VOIDmode, out, tmp));
	    }
	  if (out != operands[0])
	    emit_move_insn (operands[0], out);

	  return 1; /* DONE */
	}

      /*
       * General case:			Jumpful:
       *   xorl dest,dest		cmpl op1, op2
       *   cmpl op1, op2		movl ct, dest
       *   setcc dest			jcc 1f
       *   decl dest			movl cf, dest
       *   andl (cf-ct),dest		1:
       *   addl ct,dest
       *
       * Size 20.			Size 14.
       *
       * This is reasonably steep, but branch mispredict costs are
       * high on modern cpus, so consider failing only if optimizing
       * for space.
       *
       * %%% Parameterize branch_cost on the tuning architecture, then
       * use that.  The 80386 couldn't care less about mispredicts.
       */

      if (!optimize_size && !TARGET_CMOVE)
	{
	  if (ct == 0)
	    {
	      ct = cf;
	      cf = 0;
	      if (FLOAT_MODE_P (GET_MODE (ix86_compare_op0)))
		{
		  /* We may be reversing unordered compare to normal compare,
		     that is not valid in general (we may convert non-trapping
		     condition to trapping one), however on i386 we currently
		     emit all comparisons unordered.  */
		  compare_code = reverse_condition_maybe_unordered (compare_code);
		  code = reverse_condition_maybe_unordered (code);
		}
	      else
		{
		  compare_code = reverse_condition (compare_code);
		  code = reverse_condition (code);
		}
	    }

	  out = emit_store_flag (out, code, ix86_compare_op0,
				 ix86_compare_op1, VOIDmode, 0, 1);

	  emit_insn (gen_addsi3 (out, out, constm1_rtx));
	  emit_insn (gen_andsi3 (out, out, GEN_INT (cf-ct)));
	  if (ct != 0)
	    emit_insn (gen_addsi3 (out, out, GEN_INT (ct)));
	  if (out != operands[0])
	    emit_move_insn (operands[0], out);

	  return 1; /* DONE */
	}
    }

  if (!TARGET_CMOVE)
    {
      /* Try a few things more with specific constants and a variable.  */

      optab op;
      rtx var, orig_out, out, tmp;

      if (optimize_size)
	return 0; /* FAIL */

      /* If one of the two operands is an interesting constant, load a
	 constant with the above and mask it in with a logical operation.  */

      if (GET_CODE (operands[2]) == CONST_INT)
	{
	  var = operands[3];
	  if (INTVAL (operands[2]) == 0)
	    operands[3] = constm1_rtx, op = and_optab;
	  else if (INTVAL (operands[2]) == -1)
	    operands[3] = const0_rtx, op = ior_optab;
	  else
	    return 0; /* FAIL */
	}
      else if (GET_CODE (operands[3]) == CONST_INT)
	{
	  var = operands[2];
	  if (INTVAL (operands[3]) == 0)
	    operands[2] = constm1_rtx, op = and_optab;
	  else if (INTVAL (operands[3]) == -1)
	    operands[2] = const0_rtx, op = ior_optab;
	  else
	    return 0; /* FAIL */
	}
      else
        return 0; /* FAIL */

      orig_out = operands[0];
      tmp = gen_reg_rtx (GET_MODE (orig_out));
      operands[0] = tmp;

      /* Recurse to get the constant loaded.  */
      if (ix86_expand_int_movcc (operands) == 0)
        return 0; /* FAIL */

      /* Mask in the interesting variable.  */
      out = expand_binop (GET_MODE (orig_out), op, var, tmp, orig_out, 0,
			  OPTAB_WIDEN);
      if (out != orig_out)
	emit_move_insn (orig_out, out);

      return 1; /* DONE */
    }

  /*
   * For comparison with above,
   *
   * movl cf,dest
   * movl ct,tmp
   * cmpl op1,op2
   * cmovcc tmp,dest
   *
   * Size 15.
   */

  if (! nonimmediate_operand (operands[2], GET_MODE (operands[0])))
    operands[2] = force_reg (GET_MODE (operands[0]), operands[2]);
  if (! nonimmediate_operand (operands[3], GET_MODE (operands[0])))
    operands[3] = force_reg (GET_MODE (operands[0]), operands[3]);

  if (bypass_test && reg_overlap_mentioned_p (operands[0], operands[3]))
    {
      rtx tmp = gen_reg_rtx (GET_MODE (operands[0]));
      emit_move_insn (tmp, operands[3]);
      operands[3] = tmp;
    }
  if (second_test && reg_overlap_mentioned_p (operands[0], operands[2]))
    {
      rtx tmp = gen_reg_rtx (GET_MODE (operands[0]));
      emit_move_insn (tmp, operands[2]);
      operands[2] = tmp;
    }

  emit_insn (compare_seq);
  emit_insn (gen_rtx_SET (VOIDmode, operands[0],
			  gen_rtx_IF_THEN_ELSE (GET_MODE (operands[0]),
						compare_op, operands[2],
						operands[3])));
  if (bypass_test)
    emit_insn (gen_rtx_SET (VOIDmode, operands[0],
			    gen_rtx_IF_THEN_ELSE (GET_MODE (operands[0]),
				  bypass_test,
				  operands[3],
				  operands[0])));
  if (second_test)
    emit_insn (gen_rtx_SET (VOIDmode, operands[0],
			    gen_rtx_IF_THEN_ELSE (GET_MODE (operands[0]),
				  second_test,
				  operands[2],
				  operands[0])));

  return 1; /* DONE */
}

int
ix86_expand_fp_movcc (operands)
     rtx operands[];
{
  enum rtx_code code;
  rtx tmp;
  rtx compare_op, second_test, bypass_test;

  /* For SF/DFmode conditional moves based on comparisons
     in same mode, we may want to use SSE min/max instructions.  */
  if (((TARGET_SSE && GET_MODE (operands[0]) == SFmode)
       || (TARGET_SSE2 && GET_MODE (operands[0]) == DFmode))
      && GET_MODE (ix86_compare_op0) == GET_MODE (operands[0])
      /* We may be called from the post-reload splitter.  */
      && (!REG_P (operands[0])
	  || SSE_REG_P (operands[0])
	  || REGNO (operands[0]) >= FIRST_PSEUDO_REGISTER))
    {
      rtx op0 = ix86_compare_op0, op1 = ix86_compare_op1;
      code = GET_CODE (operands[1]);

      /* See if we have (cross) match between comparison operands and
         conditional move operands.  */
      if (rtx_equal_p (operands[2], op1))
	{
	  rtx tmp = op0;
	  op0 = op1;
	  op1 = tmp;
	  code = reverse_condition_maybe_unordered (code);
	}
      if (rtx_equal_p (operands[2], op0) && rtx_equal_p (operands[3], op1))
	{
	  /* Check for min operation.  */
	  if (code == LT)
	    {
	       operands[0] = force_reg (GET_MODE (operands[0]), operands[0]);
	       if (memory_operand (op0, VOIDmode))
		 op0 = force_reg (GET_MODE (operands[0]), op0);
	       if (GET_MODE (operands[0]) == SFmode)
		 emit_insn (gen_minsf3 (operands[0], op0, op1));
	       else
		 emit_insn (gen_mindf3 (operands[0], op0, op1));
	       return 1;
	    }
	  /* Check for max operation.  */
	  if (code == GT)
	    {
	       operands[0] = force_reg (GET_MODE (operands[0]), operands[0]);
	       if (memory_operand (op0, VOIDmode))
		 op0 = force_reg (GET_MODE (operands[0]), op0);
	       if (GET_MODE (operands[0]) == SFmode)
		 emit_insn (gen_maxsf3 (operands[0], op0, op1));
	       else
		 emit_insn (gen_maxdf3 (operands[0], op0, op1));
	       return 1;
	    }
	}
      /* Manage condition to be sse_comparison_operator.  In case we are
	 in non-ieee mode, try to canonicalize the destination operand
	 to be first in the comparison - this helps reload to avoid extra
	 moves.  */
      if (!sse_comparison_operator (operands[1], VOIDmode)
	  || (rtx_equal_p (operands[0], ix86_compare_op1) && !TARGET_IEEE_FP))
	{
	  rtx tmp = ix86_compare_op0;
	  ix86_compare_op0 = ix86_compare_op1;
	  ix86_compare_op1 = tmp;
	  operands[1] = gen_rtx_fmt_ee (swap_condition (GET_CODE (operands[1])),
					VOIDmode, ix86_compare_op0,
					ix86_compare_op1);
	}
      /* Similary try to manage result to be first operand of conditional
	 move. */
      if (rtx_equal_p (operands[0], operands[3]))
	{
	  rtx tmp = operands[2];
	  operands[2] = operands[3];
	  operands[2] = tmp;
	  operands[1] = gen_rtx_fmt_ee (reverse_condition_maybe_unordered
					  (GET_CODE (operands[1])),
					VOIDmode, ix86_compare_op0,
					ix86_compare_op1);
	}
      if (GET_MODE (operands[0]) == SFmode)
	emit_insn (gen_sse_movsfcc (operands[0], operands[1],
				    operands[2], operands[3],
				    ix86_compare_op0, ix86_compare_op1));
      else
	emit_insn (gen_sse_movdfcc (operands[0], operands[1],
				    operands[2], operands[3],
				    ix86_compare_op0, ix86_compare_op1));
      return 1;
    }

  /* The floating point conditional move instructions don't directly
     support conditions resulting from a signed integer comparison.  */

  code = GET_CODE (operands[1]);
  compare_op = ix86_expand_compare (code, &second_test, &bypass_test);

  /* The floating point conditional move instructions don't directly
     support signed integer comparisons.  */

  if (!fcmov_comparison_operator (compare_op, VOIDmode))
    {
      if (second_test != NULL || bypass_test != NULL)
	abort();
      tmp = gen_reg_rtx (QImode);
      ix86_expand_setcc (code, tmp);
      code = NE;
      ix86_compare_op0 = tmp;
      ix86_compare_op1 = const0_rtx;
      compare_op = ix86_expand_compare (code,  &second_test, &bypass_test);
    }
  if (bypass_test && reg_overlap_mentioned_p (operands[0], operands[3]))
    {
      tmp = gen_reg_rtx (GET_MODE (operands[0]));
      emit_move_insn (tmp, operands[3]);
      operands[3] = tmp;
    }
  if (second_test && reg_overlap_mentioned_p (operands[0], operands[2]))
    {
      tmp = gen_reg_rtx (GET_MODE (operands[0]));
      emit_move_insn (tmp, operands[2]);
      operands[2] = tmp;
    }

  emit_insn (gen_rtx_SET (VOIDmode, operands[0],
			  gen_rtx_IF_THEN_ELSE (GET_MODE (operands[0]),
				compare_op,
				operands[2],
				operands[3])));
  if (bypass_test)
    emit_insn (gen_rtx_SET (VOIDmode, operands[0],
			    gen_rtx_IF_THEN_ELSE (GET_MODE (operands[0]),
				  bypass_test,
				  operands[3],
				  operands[0])));
  if (second_test)
    emit_insn (gen_rtx_SET (VOIDmode, operands[0],
			    gen_rtx_IF_THEN_ELSE (GET_MODE (operands[0]),
				  second_test,
				  operands[2],
				  operands[0])));

  return 1;
}

/* Split operands 0 and 1 into SImode parts.  Similar to split_di, but
   works for floating pointer parameters and nonoffsetable memories.
   For pushes, it returns just stack offsets; the values will be saved
   in the right order.  Maximally three parts are generated.  */

static int
ix86_split_to_parts (operand, parts, mode)
     rtx operand;
     rtx *parts;
     enum machine_mode mode;
{
  int size = mode == TFmode ? 3 : GET_MODE_SIZE (mode) / 4;

  if (GET_CODE (operand) == REG && MMX_REGNO_P (REGNO (operand)))
    abort ();
  if (size < 2 || size > 3)
    abort ();

  /* Optimize constant pool reference to immediates.  This is used by fp moves,
     that force all constants to memory to allow combining.  */

  if (GET_CODE (operand) == MEM
      && GET_CODE (XEXP (operand, 0)) == SYMBOL_REF
      && CONSTANT_POOL_ADDRESS_P (XEXP (operand, 0)))
    operand = get_pool_constant (XEXP (operand, 0));

  if (GET_CODE (operand) == MEM && !offsettable_memref_p (operand))
    {
      /* The only non-offsetable memories we handle are pushes.  */
      if (! push_operand (operand, VOIDmode))
	abort ();

      PUT_MODE (operand, SImode);
      parts[0] = parts[1] = parts[2] = operand;
    }
  else
    {
      if (mode == DImode)
	split_di (&operand, 1, &parts[0], &parts[1]);
      else
	{
	  if (REG_P (operand))
	    {
	      if (!reload_completed)
		abort ();
	      parts[0] = gen_rtx_REG (SImode, REGNO (operand) + 0);
	      parts[1] = gen_rtx_REG (SImode, REGNO (operand) + 1);
	      if (size == 3)
		parts[2] = gen_rtx_REG (SImode, REGNO (operand) + 2);
	    }
	  else if (offsettable_memref_p (operand))
	    {
	      PUT_MODE (operand, SImode);
	      parts[0] = operand;
	      parts[1] = adj_offsettable_operand (operand, 4);
	      if (size == 3)
		parts[2] = adj_offsettable_operand (operand, 8);
	    }
	  else if (GET_CODE (operand) == CONST_DOUBLE)
	    {
	      REAL_VALUE_TYPE r;
	      long l[4];

	      REAL_VALUE_FROM_CONST_DOUBLE (r, operand);
	      switch (mode)
		{
		case XFmode:
		case TFmode:
		  REAL_VALUE_TO_TARGET_LONG_DOUBLE (r, l);
		  parts[2] = GEN_INT (l[2]);
		  break;
		case DFmode:
		  REAL_VALUE_TO_TARGET_DOUBLE (r, l);
		  break;
		default:
		  abort ();
		}
	      parts[1] = GEN_INT (l[1]);
	      parts[0] = GEN_INT (l[0]);
	    }
	  else
	    abort ();
	}
    }

  return size;
}

/* Emit insns to perform a move or push of DI, DF, and XF values.
   Return false when normal moves are needed; true when all required
   insns have been emitted.  Operands 2-4 contain the input values
   int the correct order; operands 5-7 contain the output values.  */

int
ix86_split_long_move (operands1)
     rtx operands1[];
{
  rtx part[2][3];
  rtx operands[2];
  int size;
  int push = 0;
  int collisions = 0;

  /* Make our own copy to avoid clobbering the operands.  */
  operands[0] = copy_rtx (operands1[0]);
  operands[1] = copy_rtx (operands1[1]);

  /* The only non-offsettable memory we handle is push.  */
  if (push_operand (operands[0], VOIDmode))
    push = 1;
  else if (GET_CODE (operands[0]) == MEM
	   && ! offsettable_memref_p (operands[0]))
    abort ();

  size = ix86_split_to_parts (operands[0], part[0], GET_MODE (operands1[0]));
  ix86_split_to_parts (operands[1], part[1], GET_MODE (operands1[0]));

  /* When emitting push, take care for source operands on the stack.  */
  if (push && GET_CODE (operands[1]) == MEM
      && reg_overlap_mentioned_p (stack_pointer_rtx, operands[1]))
    {
      if (size == 3)
	part[1][1] = part[1][2];
      part[1][0] = part[1][1];
    }

  /* We need to do copy in the right order in case an address register
     of the source overlaps the destination.  */
  if (REG_P (part[0][0]) && GET_CODE (part[1][0]) == MEM)
    {
      if (reg_overlap_mentioned_p (part[0][0], XEXP (part[1][0], 0)))
	collisions++;
      if (reg_overlap_mentioned_p (part[0][1], XEXP (part[1][0], 0)))
	collisions++;
      if (size == 3
	  && reg_overlap_mentioned_p (part[0][2], XEXP (part[1][0], 0)))
	collisions++;

      /* Collision in the middle part can be handled by reordering.  */
      if (collisions == 1 && size == 3
	  && reg_overlap_mentioned_p (part[0][1], XEXP (part[1][0], 0)))
	{
	  rtx tmp;
	  tmp = part[0][1]; part[0][1] = part[0][2]; part[0][2] = tmp;
	  tmp = part[1][1]; part[1][1] = part[1][2]; part[1][2] = tmp;
	}

      /* If there are more collisions, we can't handle it by reordering.
	 Do an lea to the last part and use only one colliding move.  */
      else if (collisions > 1)
	{
	  collisions = 1;
	  emit_insn (gen_rtx_SET (VOIDmode, part[0][size - 1],
				  XEXP (part[1][0], 0)));
	  part[1][0] = change_address (part[1][0], SImode, part[0][size - 1]);
	  part[1][1] = adj_offsettable_operand (part[1][0], 4);
	  if (size == 3)
	    part[1][2] = adj_offsettable_operand (part[1][0], 8);
	}
    }

  if (push)
    {
      if (size == 3)
	{
	  /* We use only first 12 bytes of TFmode value, but for pushing we
	     are required to adjust stack as if we were pushing real 16byte
	     value.  */
	  if (GET_MODE (operands1[0]) == TFmode)
	    emit_insn (gen_addsi3 (stack_pointer_rtx, stack_pointer_rtx,
				   GEN_INT (-4)));
	  emit_insn (gen_push (part[1][2]));
	}
      emit_insn (gen_push (part[1][1]));
      emit_insn (gen_push (part[1][0]));
      return 1;
    }

  /* Choose correct order to not overwrite the source before it is copied.  */
  if ((REG_P (part[0][0])
       && REG_P (part[1][1])
       && (REGNO (part[0][0]) == REGNO (part[1][1])
	   || (size == 3
	       && REGNO (part[0][0]) == REGNO (part[1][2]))))
      || (collisions > 0
	  && reg_overlap_mentioned_p (part[0][0], XEXP (part[1][0], 0))))
    {
      if (size == 3)
	{
	  operands1[2] = part[0][2];
	  operands1[3] = part[0][1];
	  operands1[4] = part[0][0];
	  operands1[5] = part[1][2];
	  operands1[6] = part[1][1];
	  operands1[7] = part[1][0];
	}
      else
	{
	  operands1[2] = part[0][1];
	  operands1[3] = part[0][0];
	  operands1[5] = part[1][1];
	  operands1[6] = part[1][0];
	}
    }
  else
    {
      if (size == 3)
	{
	  operands1[2] = part[0][0];
	  operands1[3] = part[0][1];
	  operands1[4] = part[0][2];
	  operands1[5] = part[1][0];
	  operands1[6] = part[1][1];
	  operands1[7] = part[1][2];
	}
      else
	{
	  operands1[2] = part[0][0];
	  operands1[3] = part[0][1];
	  operands1[5] = part[1][0];
	  operands1[6] = part[1][1];
	}
    }

  return 0;
}

void
ix86_split_ashldi (operands, scratch)
     rtx *operands, scratch;
{
  rtx low[2], high[2];
  int count;

  if (GET_CODE (operands[2]) == CONST_INT)
    {
      split_di (operands, 2, low, high);
      count = INTVAL (operands[2]) & 63;

      if (count >= 32)
	{
	  emit_move_insn (high[0], low[1]);
	  emit_move_insn (low[0], const0_rtx);

	  if (count > 32)
	    emit_insn (gen_ashlsi3 (high[0], high[0], GEN_INT (count - 32)));
	}
      else
	{
	  if (!rtx_equal_p (operands[0], operands[1]))
	    emit_move_insn (operands[0], operands[1]);
	  emit_insn (gen_x86_shld_1 (high[0], low[0], GEN_INT (count)));
	  emit_insn (gen_ashlsi3 (low[0], low[0], GEN_INT (count)));
	}
    }
  else
    {
      if (!rtx_equal_p (operands[0], operands[1]))
	emit_move_insn (operands[0], operands[1]);

      split_di (operands, 1, low, high);

      emit_insn (gen_x86_shld_1 (high[0], low[0], operands[2]));
      emit_insn (gen_ashlsi3 (low[0], low[0], operands[2]));

      if (TARGET_CMOVE && (! no_new_pseudos || scratch))
	{
	  if (! no_new_pseudos)
	    scratch = force_reg (SImode, const0_rtx);
	  else
	    emit_move_insn (scratch, const0_rtx);

	  emit_insn (gen_x86_shift_adj_1 (high[0], low[0], operands[2],
					  scratch));
	}
      else
	emit_insn (gen_x86_shift_adj_2 (high[0], low[0], operands[2]));
    }
}

void
ix86_split_ashrdi (operands, scratch)
     rtx *operands, scratch;
{
  rtx low[2], high[2];
  int count;

  if (GET_CODE (operands[2]) == CONST_INT)
    {
      split_di (operands, 2, low, high);
      count = INTVAL (operands[2]) & 63;

      if (count >= 32)
	{
	  emit_move_insn (low[0], high[1]);

	  if (! reload_completed)
	    emit_insn (gen_ashrsi3 (high[0], low[0], GEN_INT (31)));
	  else
	    {
	      emit_move_insn (high[0], low[0]);
	      emit_insn (gen_ashrsi3 (high[0], high[0], GEN_INT (31)));
	    }

	  if (count > 32)
	    emit_insn (gen_ashrsi3 (low[0], low[0], GEN_INT (count - 32)));
	}
      else
	{
	  if (!rtx_equal_p (operands[0], operands[1]))
	    emit_move_insn (operands[0], operands[1]);
	  emit_insn (gen_x86_shrd_1 (low[0], high[0], GEN_INT (count)));
	  emit_insn (gen_ashrsi3 (high[0], high[0], GEN_INT (count)));
	}
    }
  else
    {
      if (!rtx_equal_p (operands[0], operands[1]))
	emit_move_insn (operands[0], operands[1]);

      split_di (operands, 1, low, high);

      emit_insn (gen_x86_shrd_1 (low[0], high[0], operands[2]));
      emit_insn (gen_ashrsi3 (high[0], high[0], operands[2]));

      if (TARGET_CMOVE && (! no_new_pseudos || scratch))
	{
	  if (! no_new_pseudos)
	    scratch = gen_reg_rtx (SImode);
	  emit_move_insn (scratch, high[0]);
	  emit_insn (gen_ashrsi3 (scratch, scratch, GEN_INT (31)));
	  emit_insn (gen_x86_shift_adj_1 (low[0], high[0], operands[2],
					  scratch));
	}
      else
	emit_insn (gen_x86_shift_adj_3 (low[0], high[0], operands[2]));
    }
}

void
ix86_split_lshrdi (operands, scratch)
     rtx *operands, scratch;
{
  rtx low[2], high[2];
  int count;

  if (GET_CODE (operands[2]) == CONST_INT)
    {
      split_di (operands, 2, low, high);
      count = INTVAL (operands[2]) & 63;

      if (count >= 32)
	{
	  emit_move_insn (low[0], high[1]);
	  emit_move_insn (high[0], const0_rtx);

	  if (count > 32)
	    emit_insn (gen_lshrsi3 (low[0], low[0], GEN_INT (count - 32)));
	}
      else
	{
	  if (!rtx_equal_p (operands[0], operands[1]))
	    emit_move_insn (operands[0], operands[1]);
	  emit_insn (gen_x86_shrd_1 (low[0], high[0], GEN_INT (count)));
	  emit_insn (gen_lshrsi3 (high[0], high[0], GEN_INT (count)));
	}
    }
  else
    {
      if (!rtx_equal_p (operands[0], operands[1]))
	emit_move_insn (operands[0], operands[1]);

      split_di (operands, 1, low, high);

      emit_insn (gen_x86_shrd_1 (low[0], high[0], operands[2]));
      emit_insn (gen_lshrsi3 (high[0], high[0], operands[2]));

      /* Heh.  By reversing the arguments, we can reuse this pattern.  */
      if (TARGET_CMOVE && (! no_new_pseudos || scratch))
	{
	  if (! no_new_pseudos)
	    scratch = force_reg (SImode, const0_rtx);
	  else
	    emit_move_insn (scratch, const0_rtx);

	  emit_insn (gen_x86_shift_adj_1 (low[0], high[0], operands[2],
					  scratch));
	}
      else
	emit_insn (gen_x86_shift_adj_2 (low[0], high[0], operands[2]));
    }
}

/* Expand the appropriate insns for doing strlen if not just doing
   repnz; scasb

   out = result, initialized with the start address
   align_rtx = alignment of the address.
   scratch = scratch register, initialized with the startaddress when
	not aligned, otherwise undefined

   This is just the body. It needs the initialisations mentioned above and
   some address computing at the end.  These things are done in i386.md.  */

void
ix86_expand_strlensi_unroll_1 (out, align_rtx, scratch)
     rtx out, align_rtx, scratch;
{
  int align;
  rtx tmp;
  rtx align_2_label = NULL_RTX;
  rtx align_3_label = NULL_RTX;
  rtx align_4_label = gen_label_rtx ();
  rtx end_0_label = gen_label_rtx ();
  rtx mem;
  rtx tmpreg = gen_reg_rtx (SImode);

  align = 0;
  if (GET_CODE (align_rtx) == CONST_INT)
    align = INTVAL (align_rtx);

  /* Loop to check 1..3 bytes for null to get an aligned pointer.  */

  /* Is there a known alignment and is it less than 4?  */
  if (align < 4)
    {
      /* Is there a known alignment and is it not 2? */
      if (align != 2)
	{
	  align_3_label = gen_label_rtx (); /* Label when aligned to 3-byte */
	  align_2_label = gen_label_rtx (); /* Label when aligned to 2-byte */

	  /* Leave just the 3 lower bits.  */
	  align_rtx = expand_binop (SImode, and_optab, scratch, GEN_INT (3),
				    NULL_RTX, 0, OPTAB_WIDEN);

	  emit_cmp_and_jump_insns (align_rtx, const0_rtx, EQ, NULL,
				   SImode, 1, 0, align_4_label);
	  emit_cmp_and_jump_insns (align_rtx, GEN_INT (2), EQ, NULL,
				   SImode, 1, 0, align_2_label);
	  emit_cmp_and_jump_insns (align_rtx, GEN_INT (2), GTU, NULL,
				   SImode, 1, 0, align_3_label);
	}
      else
        {
	  /* Since the alignment is 2, we have to check 2 or 0 bytes;
	     check if is aligned to 4 - byte.  */

	  align_rtx = expand_binop (SImode, and_optab, scratch, GEN_INT (2),
				    NULL_RTX, 0, OPTAB_WIDEN);

	  emit_cmp_and_jump_insns (align_rtx, const0_rtx, EQ, NULL,
				   SImode, 1, 0, align_4_label);
        }

      mem = gen_rtx_MEM (QImode, out);

      /* Now compare the bytes.  */

      /* Compare the first n unaligned byte on a byte per byte basis.  */
      emit_cmp_and_jump_insns (mem, const0_rtx, EQ, NULL,
			       QImode, 1, 0, end_0_label);

      /* Increment the address.  */
      emit_insn (gen_addsi3 (out, out, const1_rtx));

      /* Not needed with an alignment of 2 */
      if (align != 2)
	{
	  emit_label (align_2_label);

	  emit_cmp_and_jump_insns (mem, const0_rtx, EQ, NULL,
				   QImode, 1, 0, end_0_label);

	  emit_insn (gen_addsi3 (out, out, const1_rtx));

	  emit_label (align_3_label);
	}

      emit_cmp_and_jump_insns (mem, const0_rtx, EQ, NULL,
			       QImode, 1, 0, end_0_label);

      emit_insn (gen_addsi3 (out, out, const1_rtx));
    }

  /* Generate loop to check 4 bytes at a time.  It is not a good idea to
     align this loop.  It gives only huge programs, but does not help to
     speed up.  */
  emit_label (align_4_label);

  mem = gen_rtx_MEM (SImode, out);
  emit_move_insn (scratch, mem);
  emit_insn (gen_addsi3 (out, out, GEN_INT (4)));

  /* This formula yields a nonzero result iff one of the bytes is zero.
     This saves three branches inside loop and many cycles.  */

  emit_insn (gen_addsi3 (tmpreg, scratch, GEN_INT (-0x01010101)));
  emit_insn (gen_one_cmplsi2 (scratch, scratch));
  emit_insn (gen_andsi3 (tmpreg, tmpreg, scratch));
  emit_insn (gen_andsi3 (tmpreg, tmpreg, GEN_INT (0x80808080)));
  emit_cmp_and_jump_insns (tmpreg, const0_rtx, EQ, 0,
			   SImode, 1, 0, align_4_label);

  if (TARGET_CMOVE)
    {
       rtx reg = gen_reg_rtx (SImode);
       emit_move_insn (reg, tmpreg);
       emit_insn (gen_lshrsi3 (reg, reg, GEN_INT (16)));

       /* If zero is not in the first two bytes, move two bytes forward.  */
       emit_insn (gen_testsi_ccno_1 (tmpreg, GEN_INT (0x8080)));
       tmp = gen_rtx_REG (CCNOmode, FLAGS_REG);
       tmp = gen_rtx_EQ (VOIDmode, tmp, const0_rtx);
       emit_insn (gen_rtx_SET (VOIDmode, tmpreg,
			       gen_rtx_IF_THEN_ELSE (SImode, tmp,
						     reg,
						     tmpreg)));
       /* Emit lea manually to avoid clobbering of flags.  */
       emit_insn (gen_rtx_SET (SImode, reg,
			       gen_rtx_PLUS (SImode, out, GEN_INT (2))));

       tmp = gen_rtx_REG (CCNOmode, FLAGS_REG);
       tmp = gen_rtx_EQ (VOIDmode, tmp, const0_rtx);
       emit_insn (gen_rtx_SET (VOIDmode, out,
			       gen_rtx_IF_THEN_ELSE (SImode, tmp,
						     reg,
						     out)));

    }
  else
    {
       rtx end_2_label = gen_label_rtx ();
       /* Is zero in the first two bytes? */

       emit_insn (gen_testsi_ccno_1 (tmpreg, GEN_INT (0x8080)));
       tmp = gen_rtx_REG (CCNOmode, FLAGS_REG);
       tmp = gen_rtx_NE (VOIDmode, tmp, const0_rtx);
       tmp = gen_rtx_IF_THEN_ELSE (VOIDmode, tmp,
                            gen_rtx_LABEL_REF (VOIDmode, end_2_label),
                            pc_rtx);
       tmp = emit_jump_insn (gen_rtx_SET (VOIDmode, pc_rtx, tmp));
       JUMP_LABEL (tmp) = end_2_label;

       /* Not in the first two.  Move two bytes forward.  */
       emit_insn (gen_lshrsi3 (tmpreg, tmpreg, GEN_INT (16)));
       emit_insn (gen_addsi3 (out, out, GEN_INT (2)));

       emit_label (end_2_label);

    }

  /* Avoid branch in fixing the byte.  */
  tmpreg = gen_lowpart (QImode, tmpreg);
  emit_insn (gen_addqi3_cc (tmpreg, tmpreg, tmpreg));
  emit_insn (gen_subsi3_carry (out, out, GEN_INT (3)));

  emit_label (end_0_label);
}
\f
/* Clear stack slot assignments remembered from previous functions.
   This is called from INIT_EXPANDERS once before RTL is emitted for each
   function.  */

static void
ix86_init_machine_status (p)
     struct function *p;
{
  p->machine = (struct machine_function *)
    xcalloc (1, sizeof (struct machine_function));
}

/* Mark machine specific bits of P for GC.  */
static void
ix86_mark_machine_status (p)
     struct function *p;
{
  struct machine_function *machine = p->machine;
  enum machine_mode mode;
  int n;

  if (! machine)
    return;

  for (mode = VOIDmode; (int) mode < (int) MAX_MACHINE_MODE;
       mode = (enum machine_mode) ((int) mode + 1))
    for (n = 0; n < MAX_386_STACK_LOCALS; n++)
      ggc_mark_rtx (machine->stack_locals[(int) mode][n]);
}

static void
ix86_free_machine_status (p)
     struct function *p;
{
  free (p->machine);
  p->machine = NULL;
}

/* Return a MEM corresponding to a stack slot with mode MODE.
   Allocate a new slot if necessary.

   The RTL for a function can have several slots available: N is
   which slot to use.  */

rtx
assign_386_stack_local (mode, n)
     enum machine_mode mode;
     int n;
{
  if (n < 0 || n >= MAX_386_STACK_LOCALS)
    abort ();

  if (ix86_stack_locals[(int) mode][n] == NULL_RTX)
    ix86_stack_locals[(int) mode][n]
      = assign_stack_local (mode, GET_MODE_SIZE (mode), 0);

  return ix86_stack_locals[(int) mode][n];
}
\f
/* Calculate the length of the memory address in the instruction
   encoding.  Does not include the one-byte modrm, opcode, or prefix.  */

static int
memory_address_length (addr)
     rtx addr;
{
  struct ix86_address parts;
  rtx base, index, disp;
  int len;

  if (GET_CODE (addr) == PRE_DEC
      || GET_CODE (addr) == POST_INC)
    return 0;

  if (! ix86_decompose_address (addr, &parts))
    abort ();

  base = parts.base;
  index = parts.index;
  disp = parts.disp;
  len = 0;

  /* Register Indirect.  */
  if (base && !index && !disp)
    {
      /* Special cases: ebp and esp need the two-byte modrm form.  */
      if (addr == stack_pointer_rtx
	  || addr == arg_pointer_rtx
	  || addr == frame_pointer_rtx
	  || addr == hard_frame_pointer_rtx)
	len = 1;
    }

  /* Direct Addressing.  */
  else if (disp && !base && !index)
    len = 4;

  else
    {
      /* Find the length of the displacement constant.  */
      if (disp)
	{
	  if (GET_CODE (disp) == CONST_INT
	      && CONST_OK_FOR_LETTER_P (INTVAL (disp), 'K'))
	    len = 1;
	  else
	    len = 4;
	}

      /* An index requires the two-byte modrm form.  */
      if (index)
	len += 1;
    }

  return len;
}

/* Compute default value for "length_immediate" attribute.  When SHORTFORM is set
   expect that insn have 8bit immediate alternative.  */
int
ix86_attr_length_immediate_default (insn, shortform)
     rtx insn;
     int shortform;
{
  int len = 0;
  int i;
  extract_insn_cached (insn);
  for (i = recog_data.n_operands - 1; i >= 0; --i)
    if (CONSTANT_P (recog_data.operand[i]))
      {
	if (len)
	  abort ();
	if (shortform
	    && GET_CODE (recog_data.operand[i]) == CONST_INT
	    && CONST_OK_FOR_LETTER_P (INTVAL (recog_data.operand[i]), 'K'))
	  len = 1;
	else
	  {
	    switch (get_attr_mode (insn))
	      {
		case MODE_QI:
		  len+=1;
		  break;
		case MODE_HI:
		  len+=2;
		  break;
		case MODE_SI:
		  len+=4;
		  break;
		default:
		  fatal_insn ("Unknown insn mode", insn);
	      }
	  }
      }
  return len;
}
/* Compute default value for "length_address" attribute.  */
int
ix86_attr_length_address_default (insn)
     rtx insn;
{
  int i;
  extract_insn_cached (insn);
  for (i = recog_data.n_operands - 1; i >= 0; --i)
    if (GET_CODE (recog_data.operand[i]) == MEM)
      {
	return memory_address_length (XEXP (recog_data.operand[i], 0));
	break;
      }
  return 0;
}
\f
/* Return the maximum number of instructions a cpu can issue.  */

int
ix86_issue_rate ()
{
  switch (ix86_cpu)
    {
    case PROCESSOR_PENTIUM:
    case PROCESSOR_K6:
      return 2;

    case PROCESSOR_PENTIUMPRO:
      return 3;

    default:
      return 1;
    }
}

/* A subroutine of ix86_adjust_cost -- return true iff INSN reads flags set
   by DEP_INSN and nothing set by DEP_INSN.  */

static int
ix86_flags_dependant (insn, dep_insn, insn_type)
     rtx insn, dep_insn;
     enum attr_type insn_type;
{
  rtx set, set2;

  /* Simplify the test for uninteresting insns.  */
  if (insn_type != TYPE_SETCC
      && insn_type != TYPE_ICMOV
      && insn_type != TYPE_FCMOV
      && insn_type != TYPE_IBR)
    return 0;

  if ((set = single_set (dep_insn)) != 0)
    {
      set = SET_DEST (set);
      set2 = NULL_RTX;
    }
  else if (GET_CODE (PATTERN (dep_insn)) == PARALLEL
	   && XVECLEN (PATTERN (dep_insn), 0) == 2
	   && GET_CODE (XVECEXP (PATTERN (dep_insn), 0, 0)) == SET
	   && GET_CODE (XVECEXP (PATTERN (dep_insn), 0, 1)) == SET)
    {
      set = SET_DEST (XVECEXP (PATTERN (dep_insn), 0, 0));
      set2 = SET_DEST (XVECEXP (PATTERN (dep_insn), 0, 0));
    }
  else
    return 0;

  if (GET_CODE (set) != REG || REGNO (set) != FLAGS_REG)
    return 0;

  /* This test is true if the dependant insn reads the flags but
     not any other potentially set register.  */
  if (!reg_overlap_mentioned_p (set, PATTERN (insn)))
    return 0;

  if (set2 && reg_overlap_mentioned_p (set2, PATTERN (insn)))
    return 0;

  return 1;
}

/* A subroutine of ix86_adjust_cost -- return true iff INSN has a memory
   address with operands set by DEP_INSN.  */

static int
ix86_agi_dependant (insn, dep_insn, insn_type)
     rtx insn, dep_insn;
     enum attr_type insn_type;
{
  rtx addr;

  if (insn_type == TYPE_LEA)
    {
      addr = PATTERN (insn);
      if (GET_CODE (addr) == SET)
	;
      else if (GET_CODE (addr) == PARALLEL
	       && GET_CODE (XVECEXP (addr, 0, 0)) == SET)
	addr = XVECEXP (addr, 0, 0);
      else
	abort ();
      addr = SET_SRC (addr);
    }
  else
    {
      int i;
      extract_insn_cached (insn);
      for (i = recog_data.n_operands - 1; i >= 0; --i)
	if (GET_CODE (recog_data.operand[i]) == MEM)
	  {
	    addr = XEXP (recog_data.operand[i], 0);
	    goto found;
	  }
      return 0;
    found:;
    }

  return modified_in_p (addr, dep_insn);
}

int
ix86_adjust_cost (insn, link, dep_insn, cost)
     rtx insn, link, dep_insn;
     int cost;
{
  enum attr_type insn_type, dep_insn_type;
  enum attr_memory memory;
  rtx set, set2;
  int dep_insn_code_number;

  /* Anti and output depenancies have zero cost on all CPUs.  */
  if (REG_NOTE_KIND (link) != 0)
    return 0;

  dep_insn_code_number = recog_memoized (dep_insn);

  /* If we can't recognize the insns, we can't really do anything.  */
  if (dep_insn_code_number < 0 || recog_memoized (insn) < 0)
    return cost;

  insn_type = get_attr_type (insn);
  dep_insn_type = get_attr_type (dep_insn);

  /* Prologue and epilogue allocators can have a false dependency on ebp.
     This results in one cycle extra stall on Pentium prologue scheduling,
     so handle this important case manually.  */
  if (dep_insn_code_number == CODE_FOR_pro_epilogue_adjust_stack
      && dep_insn_type == TYPE_ALU
      && !reg_mentioned_p (stack_pointer_rtx, insn))
    return 0;

  switch (ix86_cpu)
    {
    case PROCESSOR_PENTIUM:
      /* Address Generation Interlock adds a cycle of latency.  */
      if (ix86_agi_dependant (insn, dep_insn, insn_type))
	cost += 1;

      /* ??? Compares pair with jump/setcc.  */
      if (ix86_flags_dependant (insn, dep_insn, insn_type))
	cost = 0;

      /* Floating point stores require value to be ready one cycle ealier.  */
      if (insn_type == TYPE_FMOV
	  && get_attr_memory (insn) == MEMORY_STORE
	  && !ix86_agi_dependant (insn, dep_insn, insn_type))
	cost += 1;
      break;

    case PROCESSOR_PENTIUMPRO:
      /* Since we can't represent delayed latencies of load+operation,
	 increase the cost here for non-imov insns.  */
      if (dep_insn_type != TYPE_IMOV
	  && dep_insn_type != TYPE_FMOV
	  && ((memory = get_attr_memory (dep_insn) == MEMORY_LOAD)
              || memory == MEMORY_BOTH))
	cost += 1;

      /* INT->FP conversion is expensive.  */
      if (get_attr_fp_int_src (dep_insn))
	cost += 5;

      /* There is one cycle extra latency between an FP op and a store.  */
      if (insn_type == TYPE_FMOV
	  && (set = single_set (dep_insn)) != NULL_RTX
	  && (set2 = single_set (insn)) != NULL_RTX
	  && rtx_equal_p (SET_DEST (set), SET_SRC (set2))
	  && GET_CODE (SET_DEST (set2)) == MEM)
	cost += 1;
      break;

    case PROCESSOR_K6:
      /* The esp dependency is resolved before the instruction is really
         finished.  */
      if ((insn_type == TYPE_PUSH || insn_type == TYPE_POP)
	  && (dep_insn_type == TYPE_PUSH || dep_insn_type == TYPE_POP))
	return 1;

      /* Since we can't represent delayed latencies of load+operation,
	 increase the cost here for non-imov insns.  */
      if ((memory = get_attr_memory (dep_insn) == MEMORY_LOAD)
          || memory == MEMORY_BOTH)
	cost += (dep_insn_type != TYPE_IMOV) ? 2 : 1;

      /* INT->FP conversion is expensive.  */
      if (get_attr_fp_int_src (dep_insn))
	cost += 5;
      break;

    case PROCESSOR_ATHLON:
      if ((memory = get_attr_memory (dep_insn)) == MEMORY_LOAD
           || memory == MEMORY_BOTH)
	{
	  if (dep_insn_type == TYPE_IMOV || dep_insn_type == TYPE_FMOV)
	    cost += 2;
	  else
	    cost += 3;
        }

    default:
      break;
    }

  return cost;
}

static union
{
  struct ppro_sched_data
  {
    rtx decode[3];
    int issued_this_cycle;
  } ppro;
} ix86_sched_data;

static int
ix86_safe_length (insn)
     rtx insn;
{
  if (recog_memoized (insn) >= 0)
    return get_attr_length(insn);
  else
    return 128;
}

static int
ix86_safe_length_prefix (insn)
     rtx insn;
{
  if (recog_memoized (insn) >= 0)
    return get_attr_length(insn);
  else
    return 0;
}

static enum attr_memory
ix86_safe_memory (insn)
     rtx insn;
{
  if (recog_memoized (insn) >= 0)
    return get_attr_memory(insn);
  else
    return MEMORY_UNKNOWN;
}

static enum attr_pent_pair
ix86_safe_pent_pair (insn)
     rtx insn;
{
  if (recog_memoized (insn) >= 0)
    return get_attr_pent_pair(insn);
  else
    return PENT_PAIR_NP;
}

static enum attr_ppro_uops
ix86_safe_ppro_uops (insn)
     rtx insn;
{
  if (recog_memoized (insn) >= 0)
    return get_attr_ppro_uops (insn);
  else
    return PPRO_UOPS_MANY;
}

static void
ix86_dump_ppro_packet (dump)
     FILE *dump;
{
  if (ix86_sched_data.ppro.decode[0])
    {
      fprintf (dump, "PPRO packet: %d",
	       INSN_UID (ix86_sched_data.ppro.decode[0]));
      if (ix86_sched_data.ppro.decode[1])
	fprintf (dump, " %d", INSN_UID (ix86_sched_data.ppro.decode[1]));
      if (ix86_sched_data.ppro.decode[2])
	fprintf (dump, " %d", INSN_UID (ix86_sched_data.ppro.decode[2]));
      fputc ('\n', dump);
    }
}

/* We're beginning a new block.  Initialize data structures as necessary.  */

void
ix86_sched_init (dump, sched_verbose)
     FILE *dump ATTRIBUTE_UNUSED;
     int sched_verbose ATTRIBUTE_UNUSED;
{
  memset (&ix86_sched_data, 0, sizeof (ix86_sched_data));
}

/* Shift INSN to SLOT, and shift everything else down.  */

static void
ix86_reorder_insn (insnp, slot)
     rtx *insnp, *slot;
{
  if (insnp != slot)
    {
      rtx insn = *insnp;
      do
	insnp[0] = insnp[1];
      while (++insnp != slot);
      *insnp = insn;
    }
}

/* Find an instruction with given pairability and minimal amount of cycles
   lost by the fact that the CPU waits for both pipelines to finish before
   reading next instructions.  Also take care that both instructions together
   can not exceed 7 bytes.  */

static rtx *
ix86_pent_find_pair (e_ready, ready, type, first)
     rtx *e_ready;
     rtx *ready;
     enum attr_pent_pair type;
     rtx first;
{
  int mincycles, cycles;
  enum attr_pent_pair tmp;
  enum attr_memory memory;
  rtx *insnp, *bestinsnp = NULL;

  if (ix86_safe_length (first) > 7 + ix86_safe_length_prefix (first))
    return NULL;

  memory = ix86_safe_memory (first);
  cycles = result_ready_cost (first);
  mincycles = INT_MAX;

  for (insnp = e_ready; insnp >= ready && mincycles; --insnp)
    if ((tmp = ix86_safe_pent_pair (*insnp)) == type
	&& ix86_safe_length (*insnp) <= 7 + ix86_safe_length_prefix (*insnp))
      {
	enum attr_memory second_memory;
	int secondcycles, currentcycles;

	second_memory = ix86_safe_memory (*insnp);
	secondcycles = result_ready_cost (*insnp);
	currentcycles = abs (cycles - secondcycles);

	if (secondcycles >= 1 && cycles >= 1)
	  {
	    /* Two read/modify/write instructions together takes two
	       cycles longer.  */
	    if (memory == MEMORY_BOTH && second_memory == MEMORY_BOTH)
	      currentcycles += 2;

	    /* Read modify/write instruction followed by read/modify
	       takes one cycle longer.  */
	    if (memory == MEMORY_BOTH && second_memory == MEMORY_LOAD
	        && tmp != PENT_PAIR_UV
	        && ix86_safe_pent_pair (first) != PENT_PAIR_UV)
	      currentcycles += 1;
	  }
	if (currentcycles < mincycles)
	  bestinsnp = insnp, mincycles = currentcycles;
      }

  return bestinsnp;
}

/* Subroutines of ix86_sched_reorder.  */

static void
ix86_sched_reorder_pentium (ready, e_ready)
     rtx *ready;
     rtx *e_ready;
{
  enum attr_pent_pair pair1, pair2;
  rtx *insnp;

  /* This wouldn't be necessary if Haifa knew that static insn ordering
     is important to which pipe an insn is issued to.  So we have to make
     some minor rearrangements.  */

  pair1 = ix86_safe_pent_pair (*e_ready);

  /* If the first insn is non-pairable, let it be.  */
  if (pair1 == PENT_PAIR_NP)
    return;

  pair2 = PENT_PAIR_NP;
  insnp = 0;

  /* If the first insn is UV or PV pairable, search for a PU
     insn to go with.  */
  if (pair1 == PENT_PAIR_UV || pair1 == PENT_PAIR_PV)
    {
      insnp = ix86_pent_find_pair (e_ready-1, ready,
				   PENT_PAIR_PU, *e_ready);
      if (insnp)
	pair2 = PENT_PAIR_PU;
    }

  /* If the first insn is PU or UV pairable, search for a PV
     insn to go with.  */
  if (pair2 == PENT_PAIR_NP
      && (pair1 == PENT_PAIR_PU || pair1 == PENT_PAIR_UV))
    {
      insnp = ix86_pent_find_pair (e_ready-1, ready,
				   PENT_PAIR_PV, *e_ready);
      if (insnp)
	pair2 = PENT_PAIR_PV;
    }

  /* If the first insn is pairable, search for a UV
     insn to go with.  */
  if (pair2 == PENT_PAIR_NP)
    {
      insnp = ix86_pent_find_pair (e_ready-1, ready,
				   PENT_PAIR_UV, *e_ready);
      if (insnp)
	pair2 = PENT_PAIR_UV;
    }

  if (pair2 == PENT_PAIR_NP)
    return;

  /* Found something!  Decide if we need to swap the order.  */
  if (pair1 == PENT_PAIR_PV || pair2 == PENT_PAIR_PU
      || (pair1 == PENT_PAIR_UV && pair2 == PENT_PAIR_UV
	  && ix86_safe_memory (*e_ready) == MEMORY_BOTH
	  && ix86_safe_memory (*insnp) == MEMORY_LOAD))
    ix86_reorder_insn (insnp, e_ready);
  else
    ix86_reorder_insn (insnp, e_ready - 1);
}

static void
ix86_sched_reorder_ppro (ready, e_ready)
     rtx *ready;
     rtx *e_ready;
{
  rtx decode[3];
  enum attr_ppro_uops cur_uops;
  int issued_this_cycle;
  rtx *insnp;
  int i;

  /* At this point .ppro.decode contains the state of the three
     decoders from last "cycle".  That is, those insns that were
     actually independent.  But here we're scheduling for the
     decoder, and we may find things that are decodable in the
     same cycle.  */

  memcpy (decode, ix86_sched_data.ppro.decode, sizeof (decode));
  issued_this_cycle = 0;

  insnp = e_ready;
  cur_uops = ix86_safe_ppro_uops (*insnp);

  /* If the decoders are empty, and we've a complex insn at the
     head of the priority queue, let it issue without complaint.  */
  if (decode[0] == NULL)
    {
      if (cur_uops == PPRO_UOPS_MANY)
	{
	  decode[0] = *insnp;
	  goto ppro_done;
	}

      /* Otherwise, search for a 2-4 uop unsn to issue.  */
      while (cur_uops != PPRO_UOPS_FEW)
	{
	  if (insnp == ready)
	    break;
	  cur_uops = ix86_safe_ppro_uops (*--insnp);
	}

      /* If so, move it to the head of the line.  */
      if (cur_uops == PPRO_UOPS_FEW)
	ix86_reorder_insn (insnp, e_ready);

      /* Issue the head of the queue.  */
      issued_this_cycle = 1;
      decode[0] = *e_ready--;
    }

  /* Look for simple insns to fill in the other two slots.  */
  for (i = 1; i < 3; ++i)
    if (decode[i] == NULL)
      {
	if (ready >= e_ready)
	  goto ppro_done;

	insnp = e_ready;
	cur_uops = ix86_safe_ppro_uops (*insnp);
	while (cur_uops != PPRO_UOPS_ONE)
	  {
	    if (insnp == ready)
	      break;
	    cur_uops = ix86_safe_ppro_uops (*--insnp);
	  }

	/* Found one.  Move it to the head of the queue and issue it.  */
	if (cur_uops == PPRO_UOPS_ONE)
	  {
	    ix86_reorder_insn (insnp, e_ready);
	    decode[i] = *e_ready--;
	    issued_this_cycle++;
	    continue;
	  }

	/* ??? Didn't find one.  Ideally, here we would do a lazy split
	   of 2-uop insns, issue one and queue the other.  */
      }

 ppro_done:
  if (issued_this_cycle == 0)
    issued_this_cycle = 1;
  ix86_sched_data.ppro.issued_this_cycle = issued_this_cycle;
}

/* We are about to being issuing insns for this clock cycle.
   Override the default sort algorithm to better slot instructions.  */
int
ix86_sched_reorder (dump, sched_verbose, ready, n_ready, clock_var)
     FILE *dump ATTRIBUTE_UNUSED;
     int sched_verbose ATTRIBUTE_UNUSED;
     rtx *ready;
     int n_ready;
     int clock_var ATTRIBUTE_UNUSED;
{
  rtx *e_ready = ready + n_ready - 1;

  if (n_ready < 2)
    goto out;

  switch (ix86_cpu)
    {
    default:
      break;

    case PROCESSOR_PENTIUM:
      ix86_sched_reorder_pentium (ready, e_ready);
      break;

    case PROCESSOR_PENTIUMPRO:
      ix86_sched_reorder_ppro (ready, e_ready);
      break;
    }

out:
  return ix86_issue_rate ();
}

/* We are about to issue INSN.  Return the number of insns left on the
   ready queue that can be issued this cycle.  */

int
ix86_variable_issue (dump, sched_verbose, insn, can_issue_more)
     FILE *dump;
     int sched_verbose;
     rtx insn;
     int can_issue_more;
{
  int i;
  switch (ix86_cpu)
    {
    default:
      return can_issue_more - 1;

    case PROCESSOR_PENTIUMPRO:
      {
	enum attr_ppro_uops uops = ix86_safe_ppro_uops (insn);

	if (uops == PPRO_UOPS_MANY)
	  {
	    if (sched_verbose)
	      ix86_dump_ppro_packet (dump);
	    ix86_sched_data.ppro.decode[0] = insn;
	    ix86_sched_data.ppro.decode[1] = NULL;
	    ix86_sched_data.ppro.decode[2] = NULL;
	    if (sched_verbose)
	      ix86_dump_ppro_packet (dump);
	    ix86_sched_data.ppro.decode[0] = NULL;
	  }
	else if (uops == PPRO_UOPS_FEW)
	  {
	    if (sched_verbose)
	      ix86_dump_ppro_packet (dump);
	    ix86_sched_data.ppro.decode[0] = insn;
	    ix86_sched_data.ppro.decode[1] = NULL;
	    ix86_sched_data.ppro.decode[2] = NULL;
	  }
	else
	  {
	    for (i = 0; i < 3; ++i)
	      if (ix86_sched_data.ppro.decode[i] == NULL)
		{
		  ix86_sched_data.ppro.decode[i] = insn;
		  break;
		}
	    if (i == 3)
	      abort ();
	    if (i == 2)
	      {
	        if (sched_verbose)
	          ix86_dump_ppro_packet (dump);
		ix86_sched_data.ppro.decode[0] = NULL;
		ix86_sched_data.ppro.decode[1] = NULL;
		ix86_sched_data.ppro.decode[2] = NULL;
	      }
	  }
      }
      return --ix86_sched_data.ppro.issued_this_cycle;
    }
}
\f
/* Walk through INSNS and look for MEM references whose address is DSTREG or
   SRCREG and set the memory attribute to those of DSTREF and SRCREF, as
   appropriate.  */

void
ix86_set_move_mem_attrs (insns, dstref, srcref, dstreg, srcreg)
     rtx insns;
     rtx dstref, srcref, dstreg, srcreg;
{
  rtx insn;

  for (insn = insns; insn != 0 ; insn = NEXT_INSN (insn))
    if (INSN_P (insn))
      ix86_set_move_mem_attrs_1 (PATTERN (insn), dstref, srcref,
				 dstreg, srcreg);
}

/* Subroutine of above to actually do the updating by recursively walking
   the rtx.  */

static void
ix86_set_move_mem_attrs_1 (x, dstref, srcref, dstreg, srcreg)
     rtx x;
     rtx dstref, srcref, dstreg, srcreg;
{
  enum rtx_code code = GET_CODE (x);
  const char *format_ptr = GET_RTX_FORMAT (code);
  int i, j;

  if (code == MEM && XEXP (x, 0) == dstreg)
    MEM_COPY_ATTRIBUTES (x, dstref);
  else if (code == MEM && XEXP (x, 0) == srcreg)
    MEM_COPY_ATTRIBUTES (x, srcref);

  for (i = 0; i < GET_RTX_LENGTH (code); i++, format_ptr++)
    {
      if (*format_ptr == 'e')
	ix86_set_move_mem_attrs_1 (XEXP (x, i), dstref, srcref,
				   dstreg, srcreg);
      else if (*format_ptr == 'E')
	for (j = XVECLEN (x, i) - 1; j >= 0; j--)
	  ix86_set_move_mem_attrs_1 (XVECEXP (x, i, j), dstref, srcref,
				     dstreg, srcreg);
    }
}
\f
/* Compute the alignment given to a constant that is being placed in memory.
   EXP is the constant and ALIGN is the alignment that the object would
   ordinarily have.
   The value of this function is used instead of that alignment to align
   the object.  */

int
ix86_constant_alignment (exp, align)
     tree exp;
     int align;
{
  if (TREE_CODE (exp) == REAL_CST)
    {
      if (TYPE_MODE (TREE_TYPE (exp)) == DFmode && align < 64)
	return 64;
      else if (ALIGN_MODE_128 (TYPE_MODE (TREE_TYPE (exp))) && align < 128)
	return 128;
    }
  else if (TREE_CODE (exp) == STRING_CST && TREE_STRING_LENGTH (exp) >= 31
	   && align < 256)
    return 256;

  return align;
}

/* Compute the alignment for a static variable.
   TYPE is the data type, and ALIGN is the alignment that
   the object would ordinarily have.  The value of this function is used
   instead of that alignment to align the object.  */

int
ix86_data_alignment (type, align)
     tree type;
     int align;
{
  if (AGGREGATE_TYPE_P (type)
       && TYPE_SIZE (type)
       && TREE_CODE (TYPE_SIZE (type)) == INTEGER_CST
       && (TREE_INT_CST_LOW (TYPE_SIZE (type)) >= 256
	   || TREE_INT_CST_HIGH (TYPE_SIZE (type))) && align < 256)
    return 256;

  if (TREE_CODE (type) == ARRAY_TYPE)
    {
      if (TYPE_MODE (TREE_TYPE (type)) == DFmode && align < 64)
	return 64;
      if (ALIGN_MODE_128 (TYPE_MODE (TREE_TYPE (type))) && align < 128)
	return 128;
    }
  else if (TREE_CODE (type) == COMPLEX_TYPE)
    {

      if (TYPE_MODE (type) == DCmode && align < 64)
	return 64;
      if (TYPE_MODE (type) == XCmode && align < 128)
	return 128;
    }
  else if ((TREE_CODE (type) == RECORD_TYPE
	    || TREE_CODE (type) == UNION_TYPE
	    || TREE_CODE (type) == QUAL_UNION_TYPE)
	   && TYPE_FIELDS (type))
    {
      if (DECL_MODE (TYPE_FIELDS (type)) == DFmode && align < 64)
	return 64;
      if (ALIGN_MODE_128 (DECL_MODE (TYPE_FIELDS (type))) && align < 128)
	return 128;
    }
  else if (TREE_CODE (type) == REAL_TYPE || TREE_CODE (type) == VECTOR_TYPE
	   || TREE_CODE (type) == INTEGER_TYPE)
    {
      if (TYPE_MODE (type) == DFmode && align < 64)
	return 64;
      if (ALIGN_MODE_128 (TYPE_MODE (type)) && align < 128)
	return 128;
    }

  return align;
}

/* Compute the alignment for a local variable.
   TYPE is the data type, and ALIGN is the alignment that
   the object would ordinarily have.  The value of this macro is used
   instead of that alignment to align the object.  */

int
ix86_local_alignment (type, align)
     tree type;
     int align;
{
  if (TREE_CODE (type) == ARRAY_TYPE)
    {
      if (TYPE_MODE (TREE_TYPE (type)) == DFmode && align < 64)
	return 64;
      if (ALIGN_MODE_128 (TYPE_MODE (TREE_TYPE (type))) && align < 128)
	return 128;
    }
  else if (TREE_CODE (type) == COMPLEX_TYPE)
    {
      if (TYPE_MODE (type) == DCmode && align < 64)
	return 64;
      if (TYPE_MODE (type) == XCmode && align < 128)
	return 128;
    }
  else if ((TREE_CODE (type) == RECORD_TYPE
	    || TREE_CODE (type) == UNION_TYPE
	    || TREE_CODE (type) == QUAL_UNION_TYPE)
	   && TYPE_FIELDS (type))
    {
      if (DECL_MODE (TYPE_FIELDS (type)) == DFmode && align < 64)
	return 64;
      if (ALIGN_MODE_128 (DECL_MODE (TYPE_FIELDS (type))) && align < 128)
	return 128;
    }
  else if (TREE_CODE (type) == REAL_TYPE || TREE_CODE (type) == VECTOR_TYPE
	   || TREE_CODE (type) == INTEGER_TYPE)
    {

      if (TYPE_MODE (type) == DFmode && align < 64)
	return 64;
      if (ALIGN_MODE_128 (TYPE_MODE (type)) && align < 128)
	return 128;
    }
  return align;
}

#define def_builtin(NAME, TYPE, CODE) \
  builtin_function ((NAME), (TYPE), (CODE), BUILT_IN_MD, NULL_PTR)
struct builtin_description
{
  enum insn_code icode;
  const char * name;
  enum ix86_builtins code;
  enum rtx_code comparison;
  unsigned int flag;
};

static struct builtin_description bdesc_comi[] =
{
  { CODE_FOR_sse_comi, "__builtin_ia32_comieq", IX86_BUILTIN_COMIEQSS, EQ, 0 },
  { CODE_FOR_sse_comi, "__builtin_ia32_comilt", IX86_BUILTIN_COMILTSS, LT, 0 },
  { CODE_FOR_sse_comi, "__builtin_ia32_comile", IX86_BUILTIN_COMILESS, LE, 0 },
  { CODE_FOR_sse_comi, "__builtin_ia32_comigt", IX86_BUILTIN_COMIGTSS, LT, 1 },
  { CODE_FOR_sse_comi, "__builtin_ia32_comige", IX86_BUILTIN_COMIGESS, LE, 1 },
  { CODE_FOR_sse_comi, "__builtin_ia32_comineq", IX86_BUILTIN_COMINEQSS, NE, 0 },
  { CODE_FOR_sse_ucomi, "__builtin_ia32_ucomieq", IX86_BUILTIN_UCOMIEQSS, EQ, 0 },
  { CODE_FOR_sse_ucomi, "__builtin_ia32_ucomilt", IX86_BUILTIN_UCOMILTSS, LT, 0 },
  { CODE_FOR_sse_ucomi, "__builtin_ia32_ucomile", IX86_BUILTIN_UCOMILESS, LE, 0 },
  { CODE_FOR_sse_ucomi, "__builtin_ia32_ucomigt", IX86_BUILTIN_UCOMIGTSS, LT, 1 },
  { CODE_FOR_sse_ucomi, "__builtin_ia32_ucomige", IX86_BUILTIN_UCOMIGESS, LE, 1 },
  { CODE_FOR_sse_ucomi, "__builtin_ia32_ucomineq", IX86_BUILTIN_UCOMINEQSS, NE, 0 }
};

static struct builtin_description bdesc_2arg[] =
{
  /* SSE */
  { CODE_FOR_addv4sf3, "__builtin_ia32_addps", IX86_BUILTIN_ADDPS, 0, 0 },
  { CODE_FOR_subv4sf3, "__builtin_ia32_subps", IX86_BUILTIN_SUBPS, 0, 0 },
  { CODE_FOR_mulv4sf3, "__builtin_ia32_mulps", IX86_BUILTIN_MULPS, 0, 0 },
  { CODE_FOR_divv4sf3, "__builtin_ia32_divps", IX86_BUILTIN_DIVPS, 0, 0 },
  { CODE_FOR_vmaddv4sf3,  "__builtin_ia32_addss", IX86_BUILTIN_ADDSS, 0, 0 },
  { CODE_FOR_vmsubv4sf3,  "__builtin_ia32_subss", IX86_BUILTIN_SUBSS, 0, 0 },
  { CODE_FOR_vmmulv4sf3,  "__builtin_ia32_mulss", IX86_BUILTIN_MULSS, 0, 0 },
  { CODE_FOR_vmdivv4sf3,  "__builtin_ia32_divss", IX86_BUILTIN_DIVSS, 0, 0 },

  { CODE_FOR_maskcmpv4sf3, "__builtin_ia32_cmpeqps", IX86_BUILTIN_CMPEQPS, EQ, 0 },
  { CODE_FOR_maskcmpv4sf3, "__builtin_ia32_cmpltps", IX86_BUILTIN_CMPLTPS, LT, 0 },
  { CODE_FOR_maskcmpv4sf3, "__builtin_ia32_cmpleps", IX86_BUILTIN_CMPLEPS, LE, 0 },
  { CODE_FOR_maskcmpv4sf3, "__builtin_ia32_cmpgtps", IX86_BUILTIN_CMPGTPS, LT, 1 },
  { CODE_FOR_maskcmpv4sf3, "__builtin_ia32_cmpgeps", IX86_BUILTIN_CMPGEPS, LE, 1 },
  { CODE_FOR_maskcmpv4sf3, "__builtin_ia32_cmpunordps", IX86_BUILTIN_CMPUNORDPS, UNORDERED, 0 },
  { CODE_FOR_maskncmpv4sf3, "__builtin_ia32_cmpneqps", IX86_BUILTIN_CMPNEQPS, EQ, 0 },
  { CODE_FOR_maskncmpv4sf3, "__builtin_ia32_cmpnltps", IX86_BUILTIN_CMPNLTPS, LT, 0 },
  { CODE_FOR_maskncmpv4sf3, "__builtin_ia32_cmpnleps", IX86_BUILTIN_CMPNLEPS, LE, 0 },
  { CODE_FOR_maskncmpv4sf3, "__builtin_ia32_cmpngtps", IX86_BUILTIN_CMPNGTPS, LT, 1 },
  { CODE_FOR_maskncmpv4sf3, "__builtin_ia32_cmpngeps", IX86_BUILTIN_CMPNGEPS, LE, 1 },
  { CODE_FOR_maskncmpv4sf3, "__builtin_ia32_cmpordps", IX86_BUILTIN_CMPORDPS, UNORDERED, 0 },
  { CODE_FOR_vmmaskcmpv4sf3, "__builtin_ia32_cmpeqss", IX86_BUILTIN_CMPEQSS, EQ, 0 },
  { CODE_FOR_vmmaskcmpv4sf3, "__builtin_ia32_cmpltss", IX86_BUILTIN_CMPLTSS, LT, 0 },
  { CODE_FOR_vmmaskcmpv4sf3, "__builtin_ia32_cmpless", IX86_BUILTIN_CMPLESS, LE, 0 },
  { CODE_FOR_vmmaskcmpv4sf3, "__builtin_ia32_cmpgtss", IX86_BUILTIN_CMPGTSS, LT, 1 },
  { CODE_FOR_vmmaskcmpv4sf3, "__builtin_ia32_cmpgess", IX86_BUILTIN_CMPGESS, LE, 1 },
  { CODE_FOR_vmmaskcmpv4sf3, "__builtin_ia32_cmpunordss", IX86_BUILTIN_CMPUNORDSS, UNORDERED, 0 },
  { CODE_FOR_vmmaskncmpv4sf3, "__builtin_ia32_cmpneqss", IX86_BUILTIN_CMPNEQSS, EQ, 0 },
  { CODE_FOR_vmmaskncmpv4sf3, "__builtin_ia32_cmpnltss", IX86_BUILTIN_CMPNLTSS, LT, 0 },
  { CODE_FOR_vmmaskncmpv4sf3, "__builtin_ia32_cmpnless", IX86_BUILTIN_CMPNLESS, LE, 0 },
  { CODE_FOR_vmmaskncmpv4sf3, "__builtin_ia32_cmpngtss", IX86_BUILTIN_CMPNGTSS, LT, 1 },
  { CODE_FOR_vmmaskncmpv4sf3, "__builtin_ia32_cmpngess", IX86_BUILTIN_CMPNGESS, LE, 1 },
  { CODE_FOR_vmmaskncmpv4sf3, "__builtin_ia32_cmpordss", IX86_BUILTIN_CMPORDSS, UNORDERED, 0 },

  { CODE_FOR_sminv4sf3, "__builtin_ia32_minps", IX86_BUILTIN_MINPS, 0, 0 },
  { CODE_FOR_smaxv4sf3, "__builtin_ia32_maxps", IX86_BUILTIN_MAXPS, 0, 0 },
  { CODE_FOR_vmsminv4sf3, "__builtin_ia32_minss", IX86_BUILTIN_MINSS, 0, 0 },
  { CODE_FOR_vmsmaxv4sf3, "__builtin_ia32_maxss", IX86_BUILTIN_MAXSS, 0, 0 },

  { CODE_FOR_sse_andti3, "__builtin_ia32_andps", IX86_BUILTIN_ANDPS, 0, 0 },
  { CODE_FOR_sse_nandti3,  "__builtin_ia32_andnps", IX86_BUILTIN_ANDNPS, 0, 0 },
  { CODE_FOR_sse_iorti3, "__builtin_ia32_orps", IX86_BUILTIN_ORPS, 0, 0 },
  { CODE_FOR_sse_xorti3,  "__builtin_ia32_xorps", IX86_BUILTIN_XORPS, 0, 0 },

  { CODE_FOR_sse_movss,  "__builtin_ia32_movss", IX86_BUILTIN_MOVSS, 0, 0 },
  { CODE_FOR_sse_movhlps,  "__builtin_ia32_movhlps", IX86_BUILTIN_MOVHLPS, 0, 0 },
  { CODE_FOR_sse_movlhps,  "__builtin_ia32_movlhps", IX86_BUILTIN_MOVLHPS, 0, 0 },
  { CODE_FOR_sse_unpckhps, "__builtin_ia32_unpckhps", IX86_BUILTIN_UNPCKHPS, 0, 0 },
  { CODE_FOR_sse_unpcklps, "__builtin_ia32_unpcklps", IX86_BUILTIN_UNPCKLPS, 0, 0 },

  /* MMX */
  { CODE_FOR_addv8qi3, "__builtin_ia32_paddb", IX86_BUILTIN_PADDB, 0, 0 },
  { CODE_FOR_addv4hi3, "__builtin_ia32_paddw", IX86_BUILTIN_PADDW, 0, 0 },
  { CODE_FOR_addv2si3, "__builtin_ia32_paddd", IX86_BUILTIN_PADDD, 0, 0 },
  { CODE_FOR_subv8qi3, "__builtin_ia32_psubb", IX86_BUILTIN_PSUBB, 0, 0 },
  { CODE_FOR_subv4hi3, "__builtin_ia32_psubw", IX86_BUILTIN_PSUBW, 0, 0 },
  { CODE_FOR_subv2si3, "__builtin_ia32_psubd", IX86_BUILTIN_PSUBD, 0, 0 },

  { CODE_FOR_ssaddv8qi3, "__builtin_ia32_paddsb", IX86_BUILTIN_PADDSB, 0, 0 },
  { CODE_FOR_ssaddv4hi3, "__builtin_ia32_paddsw", IX86_BUILTIN_PADDSW, 0, 0 },
  { CODE_FOR_sssubv8qi3, "__builtin_ia32_psubsb", IX86_BUILTIN_PSUBSB, 0, 0 },
  { CODE_FOR_sssubv4hi3, "__builtin_ia32_psubsw", IX86_BUILTIN_PSUBSW, 0, 0 },
  { CODE_FOR_usaddv8qi3, "__builtin_ia32_paddusb", IX86_BUILTIN_PADDUSB, 0, 0 },
  { CODE_FOR_usaddv4hi3, "__builtin_ia32_paddusw", IX86_BUILTIN_PADDUSW, 0, 0 },
  { CODE_FOR_ussubv8qi3, "__builtin_ia32_psubusb", IX86_BUILTIN_PSUBUSB, 0, 0 },
  { CODE_FOR_ussubv4hi3, "__builtin_ia32_psubusw", IX86_BUILTIN_PSUBUSW, 0, 0 },

  { CODE_FOR_mulv4hi3, "__builtin_ia32_pmullw", IX86_BUILTIN_PMULLW, 0, 0 },
  { CODE_FOR_smulv4hi3_highpart, "__builtin_ia32_pmulhw", IX86_BUILTIN_PMULHW, 0, 0 },
  { CODE_FOR_umulv4hi3_highpart, "__builtin_ia32_pmulhuw", IX86_BUILTIN_PMULHUW, 0, 0 },

  { CODE_FOR_mmx_anddi3, "__builtin_ia32_pand", IX86_BUILTIN_PAND, 0, 0 },
  { CODE_FOR_mmx_nanddi3, "__builtin_ia32_pandn", IX86_BUILTIN_PANDN, 0, 0 },
  { CODE_FOR_mmx_iordi3, "__builtin_ia32_por", IX86_BUILTIN_POR, 0, 0 },
  { CODE_FOR_mmx_xordi3, "__builtin_ia32_pxor", IX86_BUILTIN_PXOR, 0, 0 },

  { CODE_FOR_mmx_uavgv8qi3, "__builtin_ia32_pavgb", IX86_BUILTIN_PAVGB, 0, 0 },
  { CODE_FOR_mmx_uavgv4hi3, "__builtin_ia32_pavgw", IX86_BUILTIN_PAVGW, 0, 0 },

  { CODE_FOR_eqv8qi3, "__builtin_ia32_pcmpeqb", IX86_BUILTIN_PCMPEQB, 0, 0 },
  { CODE_FOR_eqv4hi3, "__builtin_ia32_pcmpeqw", IX86_BUILTIN_PCMPEQW, 0, 0 },
  { CODE_FOR_eqv2si3, "__builtin_ia32_pcmpeqd", IX86_BUILTIN_PCMPEQD, 0, 0 },
  { CODE_FOR_gtv8qi3, "__builtin_ia32_pcmpgtb", IX86_BUILTIN_PCMPGTB, 0, 0 },
  { CODE_FOR_gtv4hi3, "__builtin_ia32_pcmpgtw", IX86_BUILTIN_PCMPGTW, 0, 0 },
  { CODE_FOR_gtv2si3, "__builtin_ia32_pcmpgtd", IX86_BUILTIN_PCMPGTD, 0, 0 },

  { CODE_FOR_umaxv8qi3, "__builtin_ia32_pmaxub", IX86_BUILTIN_PMAXUB, 0, 0 },
  { CODE_FOR_smaxv4hi3, "__builtin_ia32_pmaxsw", IX86_BUILTIN_PMAXSW, 0, 0 },
  { CODE_FOR_uminv8qi3, "__builtin_ia32_pminub", IX86_BUILTIN_PMINUB, 0, 0 },
  { CODE_FOR_sminv4hi3, "__builtin_ia32_pminsw", IX86_BUILTIN_PMINSW, 0, 0 },

  { CODE_FOR_mmx_punpckhbw, "__builtin_ia32_punpckhbw", IX86_BUILTIN_PUNPCKHBW, 0, 0 },
  { CODE_FOR_mmx_punpckhwd, "__builtin_ia32_punpckhwd", IX86_BUILTIN_PUNPCKHWD, 0, 0 },
  { CODE_FOR_mmx_punpckhdq, "__builtin_ia32_punpckhdq", IX86_BUILTIN_PUNPCKHDQ, 0, 0 },
  { CODE_FOR_mmx_punpcklbw, "__builtin_ia32_punpcklbw", IX86_BUILTIN_PUNPCKLBW, 0, 0 },
  { CODE_FOR_mmx_punpcklwd, "__builtin_ia32_punpcklwd", IX86_BUILTIN_PUNPCKLWD, 0, 0 },
  { CODE_FOR_mmx_punpckldq, "__builtin_ia32_punpckldq", IX86_BUILTIN_PUNPCKLDQ, 0, 0 },

  /* Special.  */
  { CODE_FOR_mmx_packsswb, 0, IX86_BUILTIN_PACKSSWB, 0, 0 },
  { CODE_FOR_mmx_packssdw, 0, IX86_BUILTIN_PACKSSDW, 0, 0 },
  { CODE_FOR_mmx_packuswb, 0, IX86_BUILTIN_PACKUSWB, 0, 0 },

  { CODE_FOR_cvtpi2ps, 0, IX86_BUILTIN_CVTPI2PS, 0, 0 },
  { CODE_FOR_cvtsi2ss, 0, IX86_BUILTIN_CVTSI2SS, 0, 0 },

  { CODE_FOR_ashlv4hi3, 0, IX86_BUILTIN_PSLLW, 0, 0 },
  { CODE_FOR_ashlv4hi3, 0, IX86_BUILTIN_PSLLWI, 0, 0 },
  { CODE_FOR_ashlv2si3, 0, IX86_BUILTIN_PSLLD, 0, 0 },
  { CODE_FOR_ashlv2si3, 0, IX86_BUILTIN_PSLLDI, 0, 0 },
  { CODE_FOR_mmx_ashldi3, 0, IX86_BUILTIN_PSLLQ, 0, 0 },
  { CODE_FOR_mmx_ashldi3, 0, IX86_BUILTIN_PSLLQI, 0, 0 },

  { CODE_FOR_lshrv4hi3, 0, IX86_BUILTIN_PSRLW, 0, 0 },
  { CODE_FOR_lshrv4hi3, 0, IX86_BUILTIN_PSRLWI, 0, 0 },
  { CODE_FOR_lshrv2si3, 0, IX86_BUILTIN_PSRLD, 0, 0 },
  { CODE_FOR_lshrv2si3, 0, IX86_BUILTIN_PSRLDI, 0, 0 },
  { CODE_FOR_mmx_lshrdi3, 0, IX86_BUILTIN_PSRLQ, 0, 0 },
  { CODE_FOR_mmx_lshrdi3, 0, IX86_BUILTIN_PSRLQI, 0, 0 },

  { CODE_FOR_ashrv4hi3, 0, IX86_BUILTIN_PSRAW, 0, 0 },
  { CODE_FOR_ashrv4hi3, 0, IX86_BUILTIN_PSRAWI, 0, 0 },
  { CODE_FOR_ashrv2si3, 0, IX86_BUILTIN_PSRAD, 0, 0 },
  { CODE_FOR_ashrv2si3, 0, IX86_BUILTIN_PSRADI, 0, 0 },

  { CODE_FOR_mmx_psadbw, 0, IX86_BUILTIN_PSADBW, 0, 0 },
  { CODE_FOR_mmx_pmaddwd, 0, IX86_BUILTIN_PMADDWD, 0, 0 }

};

static struct builtin_description bdesc_1arg[] =
{
  { CODE_FOR_mmx_pmovmskb, 0, IX86_BUILTIN_PMOVMSKB, 0, 0 },
  { CODE_FOR_sse_movmskps, 0, IX86_BUILTIN_MOVMSKPS, 0, 0 },

  { CODE_FOR_sqrtv4sf2, 0, IX86_BUILTIN_SQRTPS, 0, 0 },
  { CODE_FOR_rsqrtv4sf2, 0, IX86_BUILTIN_RSQRTPS, 0, 0 },
  { CODE_FOR_rcpv4sf2, 0, IX86_BUILTIN_RCPPS, 0, 0 },

  { CODE_FOR_cvtps2pi, 0, IX86_BUILTIN_CVTPS2PI, 0, 0 },
  { CODE_FOR_cvtss2si, 0, IX86_BUILTIN_CVTSS2SI, 0, 0 },
  { CODE_FOR_cvttps2pi, 0, IX86_BUILTIN_CVTTPS2PI, 0, 0 },
  { CODE_FOR_cvttss2si, 0, IX86_BUILTIN_CVTTSS2SI, 0, 0 }

};

/* Expand all the target specific builtins.  This is not called if TARGET_MMX
   is zero.  Otherwise, if TARGET_SSE is not set, only expand the MMX
   builtins.  */
void
ix86_init_builtins ()
{
  struct builtin_description * d;
  size_t i;
  tree endlink = void_list_node;

  tree pchar_type_node = build_pointer_type (char_type_node);
  tree pfloat_type_node = build_pointer_type (float_type_node);
  tree pv2si_type_node = build_pointer_type (V2SI_type_node);
  tree pdi_type_node = build_pointer_type (long_long_unsigned_type_node);

  /* Comparisons.  */
  tree int_ftype_v4sf_v4sf
    = build_function_type (integer_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      tree_cons (NULL_TREE,
						 V4SF_type_node,
						 endlink)));
  tree v4si_ftype_v4sf_v4sf
    = build_function_type (V4SI_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      tree_cons (NULL_TREE,
						 V4SF_type_node,
						 endlink)));
  /* MMX/SSE/integer conversions.  */
  tree int_ftype_v4sf_int
    = build_function_type (integer_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      tree_cons (NULL_TREE,
						 integer_type_node,
						 endlink)));
  tree int_ftype_v4sf
    = build_function_type (integer_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      endlink));
  tree int_ftype_v8qi
    = build_function_type (integer_type_node,
			   tree_cons (NULL_TREE, V8QI_type_node,
				      endlink));
  tree int_ftype_v2si
    = build_function_type (integer_type_node,
			   tree_cons (NULL_TREE, V2SI_type_node,
				      endlink));
  tree v2si_ftype_int
    = build_function_type (V2SI_type_node,
			   tree_cons (NULL_TREE, integer_type_node,
				      endlink));
  tree v4sf_ftype_v4sf_int
    = build_function_type (integer_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      tree_cons (NULL_TREE, integer_type_node,
						 endlink)));
  tree v4sf_ftype_v4sf_v2si
    = build_function_type (V4SF_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      tree_cons (NULL_TREE, V2SI_type_node,
						 endlink)));
  tree int_ftype_v4hi_int
    = build_function_type (integer_type_node,
			   tree_cons (NULL_TREE, V4HI_type_node,
				      tree_cons (NULL_TREE, integer_type_node,
						 endlink)));
  tree v4hi_ftype_v4hi_int_int
    = build_function_type (V4HI_type_node,
			   tree_cons (NULL_TREE, V4HI_type_node,
				      tree_cons (NULL_TREE, integer_type_node,
						 tree_cons (NULL_TREE,
							    integer_type_node,
							    endlink))));
  /* Miscellaneous.  */
  tree v8qi_ftype_v4hi_v4hi
    = build_function_type (V8QI_type_node,
			   tree_cons (NULL_TREE, V4HI_type_node,
				      tree_cons (NULL_TREE, V4HI_type_node,
						 endlink)));
  tree v4hi_ftype_v2si_v2si
    = build_function_type (V4HI_type_node,
			   tree_cons (NULL_TREE, V2SI_type_node,
				      tree_cons (NULL_TREE, V2SI_type_node,
						 endlink)));
  tree v4sf_ftype_v4sf_v4sf_int
    = build_function_type (V4SF_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      tree_cons (NULL_TREE, V4SF_type_node,
						 tree_cons (NULL_TREE,
							    integer_type_node,
							    endlink))));
  tree v4hi_ftype_v8qi_v8qi
    = build_function_type (V4HI_type_node,
			   tree_cons (NULL_TREE, V8QI_type_node,
				      tree_cons (NULL_TREE, V8QI_type_node,
						 endlink)));
  tree v2si_ftype_v4hi_v4hi
    = build_function_type (V2SI_type_node,
			   tree_cons (NULL_TREE, V4HI_type_node,
				      tree_cons (NULL_TREE, V4HI_type_node,
						 endlink)));
  tree v4hi_ftype_v4hi_int
    = build_function_type (V4HI_type_node,
			   tree_cons (NULL_TREE, V4HI_type_node,
				      tree_cons (NULL_TREE, integer_type_node,
						 endlink)));
  tree di_ftype_di_int
    = build_function_type (long_long_unsigned_type_node,
			   tree_cons (NULL_TREE, long_long_unsigned_type_node,
				      tree_cons (NULL_TREE, integer_type_node,
						 endlink)));
  tree v8qi_ftype_v8qi_di
    = build_function_type (V8QI_type_node,
			   tree_cons (NULL_TREE, V8QI_type_node,
				      tree_cons (NULL_TREE,
						 long_long_integer_type_node,
						 endlink)));
  tree v4hi_ftype_v4hi_di
    = build_function_type (V4HI_type_node,
			   tree_cons (NULL_TREE, V4HI_type_node,
				      tree_cons (NULL_TREE,
						 long_long_integer_type_node,
						 endlink)));
  tree v2si_ftype_v2si_di
    = build_function_type (V2SI_type_node,
			   tree_cons (NULL_TREE, V2SI_type_node,
				      tree_cons (NULL_TREE,
						 long_long_integer_type_node,
						 endlink)));
  tree void_ftype_void
    = build_function_type (void_type_node, endlink);
  tree void_ftype_pchar_int
    = build_function_type (void_type_node,
			   tree_cons (NULL_TREE, pchar_type_node,
				      tree_cons (NULL_TREE, integer_type_node,
						 endlink)));
  tree void_ftype_unsigned
    = build_function_type (void_type_node,
			   tree_cons (NULL_TREE, unsigned_type_node,
				      endlink));
  tree unsigned_ftype_void
    = build_function_type (unsigned_type_node, endlink);
  tree di_ftype_void
    = build_function_type (long_long_unsigned_type_node, endlink);
  tree ti_ftype_void
    = build_function_type (intTI_type_node, endlink);
  tree v2si_ftype_v4sf
    = build_function_type (V2SI_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      endlink));
  /* Loads/stores.  */
  tree maskmovq_args = tree_cons (NULL_TREE, V8QI_type_node,
				  tree_cons (NULL_TREE, V8QI_type_node,
					     tree_cons (NULL_TREE,
							pchar_type_node,
							endlink)));
  tree void_ftype_v8qi_v8qi_pchar
    = build_function_type (void_type_node, maskmovq_args);
  tree v4sf_ftype_pfloat
    = build_function_type (V4SF_type_node,
			   tree_cons (NULL_TREE, pfloat_type_node,
				      endlink));
  tree v4sf_ftype_float
    = build_function_type (V4SF_type_node,
			   tree_cons (NULL_TREE, float_type_node,
				      endlink));
  tree v4sf_ftype_float_float_float_float
    = build_function_type (V4SF_type_node,
			   tree_cons (NULL_TREE, float_type_node,
				      tree_cons (NULL_TREE, float_type_node,
						 tree_cons (NULL_TREE,
							    float_type_node,
							    tree_cons (NULL_TREE,
								       float_type_node,
								       endlink)))));
  /* @@@ the type is bogus */
  tree v4sf_ftype_v4sf_pv2si
    = build_function_type (V4SF_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      tree_cons (NULL_TREE, pv2si_type_node,
						 endlink)));
  tree v4sf_ftype_pv2si_v4sf
    = build_function_type (V4SF_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      tree_cons (NULL_TREE, pv2si_type_node,
						 endlink)));
  tree void_ftype_pfloat_v4sf
    = build_function_type (void_type_node,
			   tree_cons (NULL_TREE, pfloat_type_node,
				      tree_cons (NULL_TREE, V4SF_type_node,
						 endlink)));
  tree void_ftype_pdi_di
    = build_function_type (void_type_node,
			   tree_cons (NULL_TREE, pdi_type_node,
				      tree_cons (NULL_TREE,
						 long_long_unsigned_type_node,
						 endlink)));
  /* Normal vector unops.  */
  tree v4sf_ftype_v4sf
    = build_function_type (V4SF_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      endlink));

  /* Normal vector binops.  */
  tree v4sf_ftype_v4sf_v4sf
    = build_function_type (V4SF_type_node,
			   tree_cons (NULL_TREE, V4SF_type_node,
				      tree_cons (NULL_TREE, V4SF_type_node,
						 endlink)));
  tree v8qi_ftype_v8qi_v8qi
    = build_function_type (V8QI_type_node,
			   tree_cons (NULL_TREE, V8QI_type_node,
				      tree_cons (NULL_TREE, V8QI_type_node,
						 endlink)));
  tree v4hi_ftype_v4hi_v4hi
    = build_function_type (V4HI_type_node,
			   tree_cons (NULL_TREE, V4HI_type_node,
				      tree_cons (NULL_TREE, V4HI_type_node,
						 endlink)));
  tree v2si_ftype_v2si_v2si
    = build_function_type (V2SI_type_node,
			   tree_cons (NULL_TREE, V2SI_type_node,
				      tree_cons (NULL_TREE, V2SI_type_node,
						 endlink)));
  tree ti_ftype_ti_ti
    = build_function_type (intTI_type_node,
			   tree_cons (NULL_TREE, intTI_type_node,
				      tree_cons (NULL_TREE, intTI_type_node,
						 endlink)));
  tree di_ftype_di_di
    = build_function_type (long_long_unsigned_type_node,
			   tree_cons (NULL_TREE, long_long_unsigned_type_node,
				      tree_cons (NULL_TREE,
						 long_long_unsigned_type_node,
						 endlink)));

  /* Add all builtins that are more or less simple operations on two
     operands.  */
  for (i = 0, d = bdesc_2arg; i < sizeof (bdesc_2arg) / sizeof *d; i++, d++)
    {
      /* Use one of the operands; the target can have a different mode for
	 mask-generating compares.  */
      enum machine_mode mode;
      tree type;

      if (d->name == 0)
	continue;
      mode = insn_data[d->icode].operand[1].mode;

      if (! TARGET_SSE && ! VALID_MMX_REG_MODE (mode))
	continue;

      switch (mode)
	{
	case V4SFmode:
	  type = v4sf_ftype_v4sf_v4sf;
	  break;
	case V8QImode:
	  type = v8qi_ftype_v8qi_v8qi;
	  break;
	case V4HImode:
	  type = v4hi_ftype_v4hi_v4hi;
	  break;
	case V2SImode:
	  type = v2si_ftype_v2si_v2si;
	  break;
	case TImode:
	  type = ti_ftype_ti_ti;
	  break;
	case DImode:
	  type = di_ftype_di_di;
	  break;

	default:
	  abort ();
	}

      /* Override for comparisons.  */
      if (d->icode == CODE_FOR_maskcmpv4sf3
	  || d->icode == CODE_FOR_maskncmpv4sf3
	  || d->icode == CODE_FOR_vmmaskcmpv4sf3
	  || d->icode == CODE_FOR_vmmaskncmpv4sf3)
	type = v4si_ftype_v4sf_v4sf;

      def_builtin (d->name, type, d->code);
    }

  /* Add the remaining MMX insns with somewhat more complicated types.  */
  def_builtin ("__builtin_ia32_m_from_int", v2si_ftype_int, IX86_BUILTIN_M_FROM_INT);
  def_builtin ("__builtin_ia32_m_to_int", int_ftype_v2si, IX86_BUILTIN_M_TO_INT);
  def_builtin ("__builtin_ia32_mmx_zero", di_ftype_void, IX86_BUILTIN_MMX_ZERO);
  def_builtin ("__builtin_ia32_emms", void_ftype_void, IX86_BUILTIN_EMMS);
  def_builtin ("__builtin_ia32_ldmxcsr", void_ftype_unsigned, IX86_BUILTIN_LDMXCSR);
  def_builtin ("__builtin_ia32_stmxcsr", unsigned_ftype_void, IX86_BUILTIN_STMXCSR);
  def_builtin ("__builtin_ia32_psllw", v4hi_ftype_v4hi_di, IX86_BUILTIN_PSLLW);
  def_builtin ("__builtin_ia32_pslld", v2si_ftype_v2si_di, IX86_BUILTIN_PSLLD);
  def_builtin ("__builtin_ia32_psllq", di_ftype_di_di, IX86_BUILTIN_PSLLQ);

  def_builtin ("__builtin_ia32_psrlw", v4hi_ftype_v4hi_di, IX86_BUILTIN_PSRLW);
  def_builtin ("__builtin_ia32_psrld", v2si_ftype_v2si_di, IX86_BUILTIN_PSRLD);
  def_builtin ("__builtin_ia32_psrlq", di_ftype_di_di, IX86_BUILTIN_PSRLQ);

  def_builtin ("__builtin_ia32_psraw", v4hi_ftype_v4hi_di, IX86_BUILTIN_PSRAW);
  def_builtin ("__builtin_ia32_psrad", v2si_ftype_v2si_di, IX86_BUILTIN_PSRAD);

  def_builtin ("__builtin_ia32_pshufw", v4hi_ftype_v4hi_int, IX86_BUILTIN_PSHUFW);
  def_builtin ("__builtin_ia32_pmaddwd", v2si_ftype_v4hi_v4hi, IX86_BUILTIN_PMADDWD);

  /* Everything beyond this point is SSE only.  */
  if (! TARGET_SSE)
    return;

  /* comi/ucomi insns.  */
  for (i = 0, d = bdesc_comi; i < sizeof (bdesc_comi) / sizeof *d; i++, d++)
    def_builtin (d->name, int_ftype_v4sf_v4sf, d->code);

  def_builtin ("__builtin_ia32_packsswb", v8qi_ftype_v4hi_v4hi, IX86_BUILTIN_PACKSSWB);
  def_builtin ("__builtin_ia32_packssdw", v4hi_ftype_v2si_v2si, IX86_BUILTIN_PACKSSDW);
  def_builtin ("__builtin_ia32_packuswb", v8qi_ftype_v4hi_v4hi, IX86_BUILTIN_PACKUSWB);

  def_builtin ("__builtin_ia32_cvtpi2ps", v4sf_ftype_v4sf_v2si, IX86_BUILTIN_CVTPI2PS);
  def_builtin ("__builtin_ia32_cvtps2pi", v2si_ftype_v4sf, IX86_BUILTIN_CVTPS2PI);
  def_builtin ("__builtin_ia32_cvtsi2ss", v4sf_ftype_v4sf_int, IX86_BUILTIN_CVTSI2SS);
  def_builtin ("__builtin_ia32_cvtss2si", int_ftype_v4sf, IX86_BUILTIN_CVTSS2SI);
  def_builtin ("__builtin_ia32_cvttps2pi", v2si_ftype_v4sf, IX86_BUILTIN_CVTTPS2PI);
  def_builtin ("__builtin_ia32_cvttss2si", int_ftype_v4sf, IX86_BUILTIN_CVTTSS2SI);

  def_builtin ("__builtin_ia32_pextrw", int_ftype_v4hi_int, IX86_BUILTIN_PEXTRW);
  def_builtin ("__builtin_ia32_pinsrw", v4hi_ftype_v4hi_int_int, IX86_BUILTIN_PINSRW);

  def_builtin ("__builtin_ia32_maskmovq", void_ftype_v8qi_v8qi_pchar, IX86_BUILTIN_MASKMOVQ);

  def_builtin ("__builtin_ia32_loadaps", v4sf_ftype_pfloat, IX86_BUILTIN_LOADAPS);
  def_builtin ("__builtin_ia32_loadups", v4sf_ftype_pfloat, IX86_BUILTIN_LOADUPS);
  def_builtin ("__builtin_ia32_loadss", v4sf_ftype_pfloat, IX86_BUILTIN_LOADSS);
  def_builtin ("__builtin_ia32_storeaps", void_ftype_pfloat_v4sf, IX86_BUILTIN_STOREAPS);
  def_builtin ("__builtin_ia32_storeups", void_ftype_pfloat_v4sf, IX86_BUILTIN_STOREUPS);
  def_builtin ("__builtin_ia32_storess", void_ftype_pfloat_v4sf, IX86_BUILTIN_STORESS);

  def_builtin ("__builtin_ia32_loadhps", v4sf_ftype_v4sf_pv2si, IX86_BUILTIN_LOADHPS);
  def_builtin ("__builtin_ia32_loadlps", v4sf_ftype_v4sf_pv2si, IX86_BUILTIN_LOADLPS);
  def_builtin ("__builtin_ia32_storehps", v4sf_ftype_pv2si_v4sf, IX86_BUILTIN_STOREHPS);
  def_builtin ("__builtin_ia32_storelps", v4sf_ftype_pv2si_v4sf, IX86_BUILTIN_STORELPS);

  def_builtin ("__builtin_ia32_movmskps", int_ftype_v4sf, IX86_BUILTIN_MOVMSKPS);
  def_builtin ("__builtin_ia32_pmovmskb", int_ftype_v8qi, IX86_BUILTIN_PMOVMSKB);
  def_builtin ("__builtin_ia32_movntps", void_ftype_pfloat_v4sf, IX86_BUILTIN_MOVNTPS);
  def_builtin ("__builtin_ia32_movntq", void_ftype_pdi_di, IX86_BUILTIN_MOVNTQ);

  def_builtin ("__builtin_ia32_sfence", void_ftype_void, IX86_BUILTIN_SFENCE);
  def_builtin ("__builtin_ia32_prefetch", void_ftype_pchar_int, IX86_BUILTIN_PREFETCH);

  def_builtin ("__builtin_ia32_psadbw", v4hi_ftype_v8qi_v8qi, IX86_BUILTIN_PSADBW);

  def_builtin ("__builtin_ia32_rcpps", v4sf_ftype_v4sf, IX86_BUILTIN_RCPPS);
  def_builtin ("__builtin_ia32_rcpss", v4sf_ftype_v4sf, IX86_BUILTIN_RCPSS);
  def_builtin ("__builtin_ia32_rsqrtps", v4sf_ftype_v4sf, IX86_BUILTIN_RSQRTPS);
  def_builtin ("__builtin_ia32_rsqrtss", v4sf_ftype_v4sf, IX86_BUILTIN_RSQRTSS);
  def_builtin ("__builtin_ia32_sqrtps", v4sf_ftype_v4sf, IX86_BUILTIN_SQRTPS);
  def_builtin ("__builtin_ia32_sqrtss", v4sf_ftype_v4sf, IX86_BUILTIN_SQRTSS);

  def_builtin ("__builtin_ia32_shufps", v4sf_ftype_v4sf_v4sf_int, IX86_BUILTIN_SHUFPS);

  /* Composite intrinsics.  */
  def_builtin ("__builtin_ia32_setps1", v4sf_ftype_float, IX86_BUILTIN_SETPS1);
  def_builtin ("__builtin_ia32_setps", v4sf_ftype_float_float_float_float, IX86_BUILTIN_SETPS);
  def_builtin ("__builtin_ia32_setzerops", ti_ftype_void, IX86_BUILTIN_CLRPS);
  def_builtin ("__builtin_ia32_loadps1", v4sf_ftype_pfloat, IX86_BUILTIN_LOADPS1);
  def_builtin ("__builtin_ia32_loadrps", v4sf_ftype_pfloat, IX86_BUILTIN_LOADRPS);
  def_builtin ("__builtin_ia32_storeps1", void_ftype_pfloat_v4sf, IX86_BUILTIN_STOREPS1);
  def_builtin ("__builtin_ia32_storerps", void_ftype_pfloat_v4sf, IX86_BUILTIN_STORERPS);
}

/* Errors in the source file can cause expand_expr to return const0_rtx
   where we expect a vector.  To avoid crashing, use one of the vector
   clear instructions.  */
static rtx
safe_vector_operand (x, mode)
     rtx x;
     enum machine_mode mode;
{
  if (x != const0_rtx)
    return x;
  x = gen_reg_rtx (mode);

  if (VALID_MMX_REG_MODE (mode))
    emit_insn (gen_mmx_clrdi (mode == DImode ? x
			      : gen_rtx_SUBREG (DImode, x, 0)));
  else
    emit_insn (gen_sse_clrti (mode == TImode ? x
			      : gen_rtx_SUBREG (TImode, x, 0)));
  return x;
}

/* Subroutine of ix86_expand_builtin to take care of binop insns.  */

static rtx
ix86_expand_binop_builtin (icode, arglist, target)
     enum insn_code icode;
     tree arglist;
     rtx target;
{
  rtx pat;
  tree arg0 = TREE_VALUE (arglist);
  tree arg1 = TREE_VALUE (TREE_CHAIN (arglist));
  rtx op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
  rtx op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
  enum machine_mode tmode = insn_data[icode].operand[0].mode;
  enum machine_mode mode0 = insn_data[icode].operand[1].mode;
  enum machine_mode mode1 = insn_data[icode].operand[2].mode;

  if (VECTOR_MODE_P (mode0))
    op0 = safe_vector_operand (op0, mode0);
  if (VECTOR_MODE_P (mode1))
    op1 = safe_vector_operand (op1, mode1);

  if (! target
      || GET_MODE (target) != tmode
      || ! (*insn_data[icode].operand[0].predicate) (target, tmode))
    target = gen_reg_rtx (tmode);

  /* In case the insn wants input operands in modes different from
     the result, abort.  */
  if (GET_MODE (op0) != mode0 || GET_MODE (op1) != mode1)
    abort ();

  if (! (*insn_data[icode].operand[1].predicate) (op0, mode0))
    op0 = copy_to_mode_reg (mode0, op0);
  if (! (*insn_data[icode].operand[2].predicate) (op1, mode1))
    op1 = copy_to_mode_reg (mode1, op1);

  pat = GEN_FCN (icode) (target, op0, op1);
  if (! pat)
    return 0;
  emit_insn (pat);
  return target;
}

/* Subroutine of ix86_expand_builtin to take care of stores.  */

static rtx
ix86_expand_store_builtin (icode, arglist, shuffle)
     enum insn_code icode;
     tree arglist;
     int shuffle;
{
  rtx pat;
  tree arg0 = TREE_VALUE (arglist);
  tree arg1 = TREE_VALUE (TREE_CHAIN (arglist));
  rtx op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
  rtx op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
  enum machine_mode mode0 = insn_data[icode].operand[0].mode;
  enum machine_mode mode1 = insn_data[icode].operand[1].mode;

  if (VECTOR_MODE_P (mode1))
    op1 = safe_vector_operand (op1, mode1);

  op0 = gen_rtx_MEM (mode0, copy_to_mode_reg (Pmode, op0));
  if (shuffle >= 0 || ! (*insn_data[icode].operand[1].predicate) (op1, mode1))
    op1 = copy_to_mode_reg (mode1, op1);
  if (shuffle >= 0)
    emit_insn (gen_sse_shufps (op1, op1, op1, GEN_INT (shuffle)));
  pat = GEN_FCN (icode) (op0, op1);
  if (pat)
    emit_insn (pat);
  return 0;
}

/* Subroutine of ix86_expand_builtin to take care of unop insns.  */

static rtx
ix86_expand_unop_builtin (icode, arglist, target, do_load)
     enum insn_code icode;
     tree arglist;
     rtx target;
     int do_load;
{
  rtx pat;
  tree arg0 = TREE_VALUE (arglist);
  rtx op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
  enum machine_mode tmode = insn_data[icode].operand[0].mode;
  enum machine_mode mode0 = insn_data[icode].operand[1].mode;

  if (! target
      || GET_MODE (target) != tmode
      || ! (*insn_data[icode].operand[0].predicate) (target, tmode))
    target = gen_reg_rtx (tmode);
  if (do_load)
    op0 = gen_rtx_MEM (mode0, copy_to_mode_reg (Pmode, op0));
  else
    {
      if (VECTOR_MODE_P (mode0))
	op0 = safe_vector_operand (op0, mode0);

      if (! (*insn_data[icode].operand[1].predicate) (op0, mode0))
	op0 = copy_to_mode_reg (mode0, op0);
    }

  pat = GEN_FCN (icode) (target, op0);
  if (! pat)
    return 0;
  emit_insn (pat);
  return target;
}

/* Subroutine of ix86_expand_builtin to take care of three special unop insns:
   sqrtss, rsqrtss, rcpss.  */

static rtx
ix86_expand_unop1_builtin (icode, arglist, target)
     enum insn_code icode;
     tree arglist;
     rtx target;
{
  rtx pat;
  tree arg0 = TREE_VALUE (arglist);
  rtx op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
  enum machine_mode tmode = insn_data[icode].operand[0].mode;
  enum machine_mode mode0 = insn_data[icode].operand[1].mode;

  if (! target
      || GET_MODE (target) != tmode
      || ! (*insn_data[icode].operand[0].predicate) (target, tmode))
    target = gen_reg_rtx (tmode);

  if (VECTOR_MODE_P (mode0))
    op0 = safe_vector_operand (op0, mode0);

  if (! (*insn_data[icode].operand[1].predicate) (op0, mode0))
    op0 = copy_to_mode_reg (mode0, op0);

  pat = GEN_FCN (icode) (target, op0, op0);
  if (! pat)
    return 0;
  emit_insn (pat);
  return target;
}

/* Subroutine of ix86_expand_builtin to take care of comparison insns.  */

static rtx
ix86_expand_sse_compare (d, arglist, target)
     struct builtin_description *d;
     tree arglist;
     rtx target;
{
  rtx pat;
  tree arg0 = TREE_VALUE (arglist);
  tree arg1 = TREE_VALUE (TREE_CHAIN (arglist));
  rtx op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
  rtx op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
  rtx op2;
  enum machine_mode tmode = insn_data[d->icode].operand[0].mode;
  enum machine_mode mode0 = insn_data[d->icode].operand[1].mode;
  enum machine_mode mode1 = insn_data[d->icode].operand[2].mode;
  enum rtx_code comparison = d->comparison;

  if (VECTOR_MODE_P (mode0))
    op0 = safe_vector_operand (op0, mode0);
  if (VECTOR_MODE_P (mode1))
    op1 = safe_vector_operand (op1, mode1);

  /* Swap operands if we have a comparison that isn't available in
     hardware.  */
  if (d->flag)
    {
      target = gen_reg_rtx (tmode);
      emit_move_insn (target, op1);
      op1 = op0;
      op0 = target;
      comparison = swap_condition (comparison);
    }
  else if (! target
	   || GET_MODE (target) != tmode
	   || ! (*insn_data[d->icode].operand[0].predicate) (target, tmode))
    target = gen_reg_rtx (tmode);

  if (! (*insn_data[d->icode].operand[1].predicate) (op0, mode0))
    op0 = copy_to_mode_reg (mode0, op0);
  if (! (*insn_data[d->icode].operand[2].predicate) (op1, mode1))
    op1 = copy_to_mode_reg (mode1, op1);

  op2 = gen_rtx_fmt_ee (comparison, mode0, op0, op1);
  pat = GEN_FCN (d->icode) (target, op0, op1, op2);
  if (! pat)
    return 0;
  emit_insn (pat);
  return target;
}

/* Subroutine of ix86_expand_builtin to take care of comi insns.  */

static rtx
ix86_expand_sse_comi (d, arglist, target)
     struct builtin_description *d;
     tree arglist;
     rtx target;
{
  rtx pat;
  tree arg0 = TREE_VALUE (arglist);
  tree arg1 = TREE_VALUE (TREE_CHAIN (arglist));
  rtx op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
  rtx op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
  rtx op2;
  enum machine_mode mode0 = insn_data[d->icode].operand[0].mode;
  enum machine_mode mode1 = insn_data[d->icode].operand[1].mode;
  enum rtx_code comparison = d->comparison;

  if (VECTOR_MODE_P (mode0))
    op0 = safe_vector_operand (op0, mode0);
  if (VECTOR_MODE_P (mode1))
    op1 = safe_vector_operand (op1, mode1);

  /* Swap operands if we have a comparison that isn't available in
     hardware.  */
  if (d->flag)
    {
      rtx tmp = op1;
      op1 = op0;
      op0 = tmp;
      comparison = swap_condition (comparison);
    }

  target = gen_reg_rtx (SImode);
  emit_move_insn (target, const0_rtx);
  target = gen_rtx_SUBREG (QImode, target, 0);

  if (! (*insn_data[d->icode].operand[0].predicate) (op0, mode0))
    op0 = copy_to_mode_reg (mode0, op0);
  if (! (*insn_data[d->icode].operand[1].predicate) (op1, mode1))
    op1 = copy_to_mode_reg (mode1, op1);

  op2 = gen_rtx_fmt_ee (comparison, mode0, op0, op1);
  pat = GEN_FCN (d->icode) (op0, op1, op2);
  if (! pat)
    return 0;
  emit_insn (pat);
  emit_insn (gen_setcc_2 (target, op2));

  return target;
}

/* Expand an expression EXP that calls a built-in function,
   with result going to TARGET if that's convenient
   (and in mode MODE if that's convenient).
   SUBTARGET may be used as the target for computing one of EXP's operands.
   IGNORE is nonzero if the value is to be ignored.  */

rtx
ix86_expand_builtin (exp, target, subtarget, mode, ignore)
     tree exp;
     rtx target;
     rtx subtarget ATTRIBUTE_UNUSED;
     enum machine_mode mode ATTRIBUTE_UNUSED;
     int ignore ATTRIBUTE_UNUSED;
{
  struct builtin_description *d;
  size_t i;
  enum insn_code icode;
  tree fndecl = TREE_OPERAND (TREE_OPERAND (exp, 0), 0);
  tree arglist = TREE_OPERAND (exp, 1);
  tree arg0, arg1, arg2, arg3;
  rtx op0, op1, op2, pat;
  enum machine_mode tmode, mode0, mode1, mode2;
  unsigned int fcode = DECL_FUNCTION_CODE (fndecl);

  switch (fcode)
    {
    case IX86_BUILTIN_EMMS:
      emit_insn (gen_emms ());
      return 0;

    case IX86_BUILTIN_SFENCE:
      emit_insn (gen_sfence ());
      return 0;

    case IX86_BUILTIN_M_FROM_INT:
      target = gen_reg_rtx (DImode);
      op0 = expand_expr (TREE_VALUE (arglist), NULL_RTX, VOIDmode, 0);
      emit_move_insn (gen_rtx_SUBREG (SImode, target, 0), op0);
      return target;

    case IX86_BUILTIN_M_TO_INT:
      op0 = expand_expr (TREE_VALUE (arglist), NULL_RTX, VOIDmode, 0);
      op0 = copy_to_mode_reg (DImode, op0);
      target = gen_reg_rtx (SImode);
      emit_move_insn (target, gen_rtx_SUBREG (SImode, op0, 0));
      return target;

    case IX86_BUILTIN_PEXTRW:
      icode = CODE_FOR_mmx_pextrw;
      arg0 = TREE_VALUE (arglist);
      arg1 = TREE_VALUE (TREE_CHAIN (arglist));
      op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
      op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
      tmode = insn_data[icode].operand[0].mode;
      mode0 = insn_data[icode].operand[1].mode;
      mode1 = insn_data[icode].operand[2].mode;

      if (! (*insn_data[icode].operand[1].predicate) (op0, mode0))
	op0 = copy_to_mode_reg (mode0, op0);
      if (! (*insn_data[icode].operand[2].predicate) (op1, mode1))
	{
	  /* @@@ better error message */
	  error ("selector must be an immediate");
	  return const0_rtx;
	}
      if (target == 0
	  || GET_MODE (target) != tmode
	  || ! (*insn_data[icode].operand[0].predicate) (target, tmode))
	target = gen_reg_rtx (tmode);
      pat = GEN_FCN (icode) (target, op0, op1);
      if (! pat)
	return 0;
      emit_insn (pat);
      return target;

    case IX86_BUILTIN_PINSRW:
      icode = CODE_FOR_mmx_pinsrw;
      arg0 = TREE_VALUE (arglist);
      arg1 = TREE_VALUE (TREE_CHAIN (arglist));
      arg2 = TREE_VALUE (TREE_CHAIN (TREE_CHAIN (arglist)));
      op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
      op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
      op2 = expand_expr (arg2, NULL_RTX, VOIDmode, 0);
      tmode = insn_data[icode].operand[0].mode;
      mode0 = insn_data[icode].operand[1].mode;
      mode1 = insn_data[icode].operand[2].mode;
      mode2 = insn_data[icode].operand[3].mode;

      if (! (*insn_data[icode].operand[1].predicate) (op0, mode0))
	op0 = copy_to_mode_reg (mode0, op0);
      if (! (*insn_data[icode].operand[2].predicate) (op1, mode1))
	op1 = copy_to_mode_reg (mode1, op1);
      if (! (*insn_data[icode].operand[3].predicate) (op2, mode2))
	{
	  /* @@@ better error message */
	  error ("selector must be an immediate");
	  return const0_rtx;
	}
      if (target == 0
	  || GET_MODE (target) != tmode
	  || ! (*insn_data[icode].operand[0].predicate) (target, tmode))
	target = gen_reg_rtx (tmode);
      pat = GEN_FCN (icode) (target, op0, op1, op2);
      if (! pat)
	return 0;
      emit_insn (pat);
      return target;

    case IX86_BUILTIN_MASKMOVQ:
      icode = CODE_FOR_mmx_maskmovq;
      /* Note the arg order is different from the operand order.  */
      arg1 = TREE_VALUE (arglist);
      arg2 = TREE_VALUE (TREE_CHAIN (arglist));
      arg0 = TREE_VALUE (TREE_CHAIN (TREE_CHAIN (arglist)));
      op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
      op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
      op2 = expand_expr (arg2, NULL_RTX, VOIDmode, 0);
      mode0 = insn_data[icode].operand[0].mode;
      mode1 = insn_data[icode].operand[1].mode;
      mode2 = insn_data[icode].operand[2].mode;

      if (! (*insn_data[icode].operand[1].predicate) (op0, mode0))
	op0 = copy_to_mode_reg (mode0, op0);
      if (! (*insn_data[icode].operand[1].predicate) (op1, mode1))
	op1 = copy_to_mode_reg (mode1, op1);
      if (! (*insn_data[icode].operand[2].predicate) (op2, mode2))
	op2 = copy_to_mode_reg (mode2, op2);
      pat = GEN_FCN (icode) (op0, op1, op2);
      if (! pat)
	return 0;
      emit_insn (pat);
      return 0;

    case IX86_BUILTIN_SQRTSS:
      return ix86_expand_unop1_builtin (CODE_FOR_vmsqrtv4sf2, arglist, target);
    case IX86_BUILTIN_RSQRTSS:
      return ix86_expand_unop1_builtin (CODE_FOR_vmrsqrtv4sf2, arglist, target);
    case IX86_BUILTIN_RCPSS:
      return ix86_expand_unop1_builtin (CODE_FOR_vmrcpv4sf2, arglist, target);

    case IX86_BUILTIN_LOADAPS:
      return ix86_expand_unop_builtin (CODE_FOR_sse_movaps, arglist, target, 1);

    case IX86_BUILTIN_LOADUPS:
      return ix86_expand_unop_builtin (CODE_FOR_sse_movups, arglist, target, 1);

    case IX86_BUILTIN_STOREAPS:
      return ix86_expand_store_builtin (CODE_FOR_sse_movaps, arglist, -1);
    case IX86_BUILTIN_STOREUPS:
      return ix86_expand_store_builtin (CODE_FOR_sse_movups, arglist, -1);

    case IX86_BUILTIN_LOADSS:
      return ix86_expand_unop_builtin (CODE_FOR_sse_loadss, arglist, target, 1);

    case IX86_BUILTIN_STORESS:
      return ix86_expand_store_builtin (CODE_FOR_sse_storess, arglist, -1);

    case IX86_BUILTIN_LOADHPS:
    case IX86_BUILTIN_LOADLPS:
      icode = (fcode == IX86_BUILTIN_LOADHPS
	       ? CODE_FOR_sse_movhps : CODE_FOR_sse_movlps);
      arg0 = TREE_VALUE (arglist);
      arg1 = TREE_VALUE (TREE_CHAIN (arglist));
      op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
      op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
      tmode = insn_data[icode].operand[0].mode;
      mode0 = insn_data[icode].operand[1].mode;
      mode1 = insn_data[icode].operand[2].mode;

      if (! (*insn_data[icode].operand[1].predicate) (op0, mode0))
	op0 = copy_to_mode_reg (mode0, op0);
      op1 = gen_rtx_MEM (mode1, copy_to_mode_reg (Pmode, op1));
      if (target == 0
	  || GET_MODE (target) != tmode
	  || ! (*insn_data[icode].operand[0].predicate) (target, tmode))
	target = gen_reg_rtx (tmode);
      pat = GEN_FCN (icode) (target, op0, op1);
      if (! pat)
	return 0;
      emit_insn (pat);
      return target;

    case IX86_BUILTIN_STOREHPS:
    case IX86_BUILTIN_STORELPS:
      icode = (fcode == IX86_BUILTIN_STOREHPS
	       ? CODE_FOR_sse_movhps : CODE_FOR_sse_movlps);
      arg0 = TREE_VALUE (arglist);
      arg1 = TREE_VALUE (TREE_CHAIN (arglist));
      op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
      op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
      mode0 = insn_data[icode].operand[1].mode;
      mode1 = insn_data[icode].operand[2].mode;

      op0 = gen_rtx_MEM (mode0, copy_to_mode_reg (Pmode, op0));
      if (! (*insn_data[icode].operand[2].predicate) (op1, mode1))
	op1 = copy_to_mode_reg (mode1, op1);

      pat = GEN_FCN (icode) (op0, op0, op1);
      if (! pat)
	return 0;
      emit_insn (pat);
      return 0;

    case IX86_BUILTIN_MOVNTPS:
      return ix86_expand_store_builtin (CODE_FOR_sse_movntv4sf, arglist, -1);
    case IX86_BUILTIN_MOVNTQ:
      return ix86_expand_store_builtin (CODE_FOR_sse_movntdi, arglist, -1);

    case IX86_BUILTIN_LDMXCSR:
      op0 = expand_expr (TREE_VALUE (arglist), NULL_RTX, VOIDmode, 0);
      target = assign_386_stack_local (SImode, 0);
      emit_move_insn (target, op0);
      emit_insn (gen_ldmxcsr (target));
      return 0;

    case IX86_BUILTIN_STMXCSR:
      target = assign_386_stack_local (SImode, 0);
      emit_insn (gen_stmxcsr (target));
      return copy_to_mode_reg (SImode, target);

    case IX86_BUILTIN_PREFETCH:
      icode = CODE_FOR_prefetch;
      arg0 = TREE_VALUE (arglist);
      arg1 = TREE_VALUE (TREE_CHAIN (arglist));
      op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
      op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
      mode0 = insn_data[icode].operand[0].mode;
      mode1 = insn_data[icode].operand[1].mode;

      if (! (*insn_data[icode].operand[1].predicate) (op1, mode1))
	{
	  /* @@@ better error message */
	  error ("selector must be an immediate");
	  return const0_rtx;
	}

      op0 = copy_to_mode_reg (Pmode, op0);
      pat = GEN_FCN (icode) (op0, op1);
      if (! pat)
	return 0;
      emit_insn (pat);
      return target;

    case IX86_BUILTIN_SHUFPS:
      icode = CODE_FOR_sse_shufps;
      arg0 = TREE_VALUE (arglist);
      arg1 = TREE_VALUE (TREE_CHAIN (arglist));
      arg2 = TREE_VALUE (TREE_CHAIN (TREE_CHAIN (arglist)));
      op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
      op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
      op2 = expand_expr (arg2, NULL_RTX, VOIDmode, 0);
      tmode = insn_data[icode].operand[0].mode;
      mode0 = insn_data[icode].operand[1].mode;
      mode1 = insn_data[icode].operand[2].mode;
      mode2 = insn_data[icode].operand[3].mode;

      if (! (*insn_data[icode].operand[1].predicate) (op0, mode0))
	op0 = copy_to_mode_reg (mode0, op0);
      if (! (*insn_data[icode].operand[2].predicate) (op1, mode1))
	op1 = copy_to_mode_reg (mode1, op1);
      if (! (*insn_data[icode].operand[3].predicate) (op2, mode2))
	{
	  /* @@@ better error message */
	  error ("mask must be an immediate");
	  return const0_rtx;
	}
      if (target == 0
	  || GET_MODE (target) != tmode
	  || ! (*insn_data[icode].operand[0].predicate) (target, tmode))
	target = gen_reg_rtx (tmode);
      pat = GEN_FCN (icode) (target, op0, op1, op2);
      if (! pat)
	return 0;
      emit_insn (pat);
      return target;

    case IX86_BUILTIN_PSHUFW:
      icode = CODE_FOR_mmx_pshufw;
      arg0 = TREE_VALUE (arglist);
      arg1 = TREE_VALUE (TREE_CHAIN (arglist));
      op0 = expand_expr (arg0, NULL_RTX, VOIDmode, 0);
      op1 = expand_expr (arg1, NULL_RTX, VOIDmode, 0);
      tmode = insn_data[icode].operand[0].mode;
      mode0 = insn_data[icode].operand[2].mode;
      mode1 = insn_data[icode].operand[3].mode;

      if (! (*insn_data[icode].operand[1].predicate) (op0, mode0))
	op0 = copy_to_mode_reg (mode0, op0);
      if (! (*insn_data[icode].operand[3].predicate) (op1, mode1))
	{
	  /* @@@ better error message */
	  error ("mask must be an immediate");
	  return const0_rtx;
	}
      if (target == 0
	  || GET_MODE (target) != tmode
	  || ! (*insn_data[icode].operand[0].predicate) (target, tmode))
	target = gen_reg_rtx (tmode);
      pat = GEN_FCN (icode) (target, target, op0, op1);
      if (! pat)
	return 0;
      emit_insn (pat);
      return target;

      /* Composite intrinsics.  */
    case IX86_BUILTIN_SETPS1:
      target = assign_386_stack_local (SFmode, 0);
      arg0 = TREE_VALUE (arglist);
      emit_move_insn (change_address (target, SFmode, XEXP (target, 0)),
		      expand_expr (arg0, NULL_RTX, VOIDmode, 0));
      op0 = gen_reg_rtx (V4SFmode);
      emit_insn (gen_sse_loadss (op0, change_address (target, V4SFmode,
						      XEXP (target, 0))));
      emit_insn (gen_sse_shufps (op0, op0, op0, GEN_INT (0)));
      return op0;

    case IX86_BUILTIN_SETPS:
      target = assign_386_stack_local (V4SFmode, 0);
      op0 = change_address (target, SFmode, XEXP (target, 0));
      arg0 = TREE_VALUE (arglist);
      arg1 = TREE_VALUE (TREE_CHAIN (arglist));
      arg2 = TREE_VALUE (TREE_CHAIN (TREE_CHAIN (arglist)));
      arg3 = TREE_VALUE (TREE_CHAIN (TREE_CHAIN (TREE_CHAIN (arglist))));
      emit_move_insn (op0,
		      expand_expr (arg0, NULL_RTX, VOIDmode, 0));
      emit_move_insn (adj_offsettable_operand (op0, 4),
		      expand_expr (arg1, NULL_RTX, VOIDmode, 0));
      emit_move_insn (adj_offsettable_operand (op0, 8),
		      expand_expr (arg2, NULL_RTX, VOIDmode, 0));
      emit_move_insn (adj_offsettable_operand (op0, 12),
		      expand_expr (arg3, NULL_RTX, VOIDmode, 0));
      op0 = gen_reg_rtx (V4SFmode);
      emit_insn (gen_sse_movaps (op0, target));
      return op0;

    case IX86_BUILTIN_CLRPS:
      target = gen_reg_rtx (TImode);
      emit_insn (gen_sse_clrti (target));
      return target;

    case IX86_BUILTIN_LOADRPS:
      target = ix86_expand_unop_builtin (CODE_FOR_sse_movaps, arglist,
					 gen_reg_rtx (V4SFmode), 1);
      emit_insn (gen_sse_shufps (target, target, target, GEN_INT (0x1b)));
      return target;

    case IX86_BUILTIN_LOADPS1:
      target = ix86_expand_unop_builtin (CODE_FOR_sse_loadss, arglist,
					 gen_reg_rtx (V4SFmode), 1);
      emit_insn (gen_sse_shufps (target, target, target, const0_rtx));
      return target;

    case IX86_BUILTIN_STOREPS1:
      return ix86_expand_store_builtin (CODE_FOR_sse_movaps, arglist, 0);
    case IX86_BUILTIN_STORERPS:
      return ix86_expand_store_builtin (CODE_FOR_sse_movaps, arglist, 0x1B);

    case IX86_BUILTIN_MMX_ZERO:
      target = gen_reg_rtx (DImode);
      emit_insn (gen_mmx_clrdi (target));
      return target;

    default:
      break;
    }

  for (i = 0, d = bdesc_2arg; i < sizeof (bdesc_2arg) / sizeof *d; i++, d++)
    if (d->code == fcode)
      {
	/* Compares are treated specially.  */
	if (d->icode == CODE_FOR_maskcmpv4sf3
	    || d->icode == CODE_FOR_vmmaskcmpv4sf3
	    || d->icode == CODE_FOR_maskncmpv4sf3
	    || d->icode == CODE_FOR_vmmaskncmpv4sf3)
	  return ix86_expand_sse_compare (d, arglist, target);

	return ix86_expand_binop_builtin (d->icode, arglist, target);
      }

  for (i = 0, d = bdesc_1arg; i < sizeof (bdesc_1arg) / sizeof *d; i++, d++)
    if (d->code == fcode)
      return ix86_expand_unop_builtin (d->icode, arglist, target, 0);

  for (i = 0, d = bdesc_comi; i < sizeof (bdesc_comi) / sizeof *d; i++, d++)
    if (d->code == fcode)
      return ix86_expand_sse_comi (d, arglist, target);

  /* @@@ Should really do something sensible here.  */
  return 0;
}

/* Store OPERAND to the memory after reload is completed.  This means
   that we can't easilly use assign_stack_local.  */
rtx
ix86_force_to_memory (mode, operand)
     enum machine_mode mode;
     rtx operand;
{
  if (!reload_completed)
    abort ();
  switch (mode)
    {
      case DImode:
	{
	  rtx operands[2];
	  split_di (&operand, 1, operands, operands+1);
	  emit_insn (
	    gen_rtx_SET (VOIDmode,
			 gen_rtx_MEM (SImode,
				      gen_rtx_PRE_DEC (Pmode,
						       stack_pointer_rtx)),
			 operands[1]));
	  emit_insn (
	    gen_rtx_SET (VOIDmode,
			 gen_rtx_MEM (SImode,
				      gen_rtx_PRE_DEC (Pmode,
						       stack_pointer_rtx)),
			 operands[0]));
	}
	break;
      case HImode:
	/* It is better to store HImodes as SImodes.  */
	if (!TARGET_PARTIAL_REG_STALL)
	  operand = gen_lowpart (SImode, operand);
	/* FALLTHRU */
      case SImode:
	emit_insn (
	  gen_rtx_SET (VOIDmode,
		       gen_rtx_MEM (GET_MODE (operand),
				    gen_rtx_PRE_DEC (SImode,
						     stack_pointer_rtx)),
		       operand));
	break;
      default:
	abort();
    }
  return gen_rtx_MEM (mode, stack_pointer_rtx);
}

/* Free operand from the memory.  */
void
ix86_free_from_memory (mode)
     enum machine_mode mode;
{
  /* Use LEA to deallocate stack space.  In peephole2 it will be converted
     to pop or add instruction if registers are available.  */
  emit_insn (gen_rtx_SET (VOIDmode, stack_pointer_rtx,
			  gen_rtx_PLUS (Pmode, stack_pointer_rtx,
					GEN_INT (mode == DImode
						 ? 8
						 : mode == HImode && TARGET_PARTIAL_REG_STALL
						 ? 2
						 : 4))));
}

/* Put float CONST_DOUBLE in the constant pool instead of fp regs.
   QImode must go into class Q_REGS.
   Narrow ALL_REGS to GENERAL_REGS.  This supports allowing movsf and
   movdf to do mem-to-mem moves through integer regs. */
enum reg_class
ix86_preferred_reload_class (x, class)
     rtx x;
     enum reg_class class;
{
  if (GET_CODE (x) == CONST_DOUBLE && GET_MODE (x) != VOIDmode)
    {
      /* SSE can't load any constant directly yet.  */
      if (SSE_CLASS_P (class))
	return NO_REGS;
      /* Floats can load 0 and 1.  */
      if (MAYBE_FLOAT_CLASS_P (class) && standard_80387_constant_p (x))
	{
	  /* Limit class to non-SSE.  Use GENERAL_REGS if possible.  */
	  if (MAYBE_SSE_CLASS_P (class))
	    return (reg_class_subset_p (class, GENERAL_REGS)
		    ? GENERAL_REGS : FLOAT_REGS);
	  else
	    return class;
	}
      /* General regs can load everything.  */
      if (reg_class_subset_p (class, GENERAL_REGS))
	return GENERAL_REGS;
      /* In case we haven't resolved FLOAT or SSE yet, give up.  */
      if (MAYBE_FLOAT_CLASS_P (class) || MAYBE_SSE_CLASS_P (class))
	return NO_REGS;
    }
  if (MAYBE_MMX_CLASS_P (class) && CONSTANT_P (x))
    return NO_REGS;
  if (GET_MODE (x) == QImode && ! reg_class_subset_p (class, Q_REGS))
    return Q_REGS;
  return class;
}

/* If we are copying between general and FP registers, we need a memory
   location. The same is true for SSE and MMX registers.

   The macro can't work reliably when one of the CLASSES is class containing
   registers from multiple units (SSE, MMX, integer).  We avoid this by never
   combining those units in single alternative in the machine description.
   Ensure that this constraint holds to avoid unexpected surprises.

   When STRICT is false, we are being called from REGISTER_MOVE_COST, so do not
   enforce these sanity checks.  */
int
ix86_secondary_memory_needed (class1, class2, mode, strict)
     enum reg_class class1, class2;
     enum machine_mode mode;
     int strict;
{
  if (MAYBE_FLOAT_CLASS_P (class1) != FLOAT_CLASS_P (class1)
      || MAYBE_FLOAT_CLASS_P (class2) != FLOAT_CLASS_P (class2)
      || MAYBE_SSE_CLASS_P (class1) != SSE_CLASS_P (class1)
      || MAYBE_SSE_CLASS_P (class2) != SSE_CLASS_P (class2)
      || MAYBE_MMX_CLASS_P (class1) != MMX_CLASS_P (class1)
      || MAYBE_MMX_CLASS_P (class2) != MMX_CLASS_P (class2))
    {
      if (strict)
	abort ();
      else
	return 1;
    }
  return (FLOAT_CLASS_P (class1) != FLOAT_CLASS_P (class2)
	  || (SSE_CLASS_P (class1) != SSE_CLASS_P (class2)
	      && (mode) != SImode)
	  || (MMX_CLASS_P (class1) != MMX_CLASS_P (class2)
	      && (mode) != SImode));
}
/* Return the cost of moving data from a register in class CLASS1 to
   one in class CLASS2. 

   It is not required that the cost always equal 2 when FROM is the same as TO;
   on some machines it is expensive to move between registers if they are not
   general registers.  */
int
ix86_register_move_cost (mode, class1, class2)
     enum machine_mode mode;
     enum reg_class class1, class2;
{
  /* In case we require secondary memory, compute cost of the store followed
     by load.  In case of copying from general_purpose_register we may emit
     multiple stores followed by single load causing memory size mismatch
     stall.  Count this as arbitarily high cost of 20.  */
  if (ix86_secondary_memory_needed (class1, class2, mode, 0))
    {
      if (CLASS_MAX_NREGS (class1, mode) > CLASS_MAX_NREGS (class2, mode))
	return 10;
      return (MEMORY_MOVE_COST (mode, class1, 0)
	      + MEMORY_MOVE_COST (mode, class2, 1));
    }
  /* Moves between SSE/MMX and integer unit are expensive.
     ??? We should make this cost CPU specific.  */
  if (MMX_CLASS_P (class1) != MMX_CLASS_P (class2)
      || SSE_CLASS_P (class1) != SSE_CLASS_P (class2))
    return ix86_cost->mmxsse_to_integer;
  if (MAYBE_FLOAT_CLASS_P (class1))
    return ix86_cost->fp_move;
  if (MAYBE_SSE_CLASS_P (class1))
    return ix86_cost->sse_move;
  if (MAYBE_MMX_CLASS_P (class1))
    return ix86_cost->mmx_move;
  return 2;
}

/* Return 1 if hard register REGNO can hold a value of machine-mode MODE.  */
int
ix86_hard_regno_mode_ok (regno, mode)
     int regno;
     enum machine_mode mode;
{
  /* Flags and only flags can only hold CCmode values.  */
  if (CC_REGNO_P (regno))
    return GET_MODE_CLASS (mode) == MODE_CC;
  if (GET_MODE_CLASS (mode) == MODE_CC
      || GET_MODE_CLASS (mode) == MODE_RANDOM
      || GET_MODE_CLASS (mode) == MODE_PARTIAL_INT)
    return 0;
  if (FP_REGNO_P (regno))
    return VALID_FP_MODE_P (mode);
  if (SSE_REGNO_P (regno))
    return VALID_SSE_REG_MODE (mode);
  if (MMX_REGNO_P (regno))
    return VALID_MMX_REG_MODE (mode);
  /* We handle both integer and floats in the general purpose registers.
     In future we should be able to handle vector modes as well.  */
  if (!VALID_INT_MODE_P (mode) && !VALID_FP_MODE_P (mode))
    return 0;
  /* Take care for QImode values - they can be in non-QI regs, but then
     they do cause partial register stalls.  */
  if (regno < 4 || mode != QImode)
    return 1;
  return reload_in_progress || reload_completed || !TARGET_PARTIAL_REG_STALL;
}

/* Return the cost of moving data of mode M between a
   register and memory.  A value of 2 is the default; this cost is
   relative to those in `REGISTER_MOVE_COST'.

   If moving between registers and memory is more expensive than
   between two registers, you should define this macro to express the
   relative cost.  
 
   Model also increased moving costs of QImode registers in non
   Q_REGS classes.
 */
int
ix86_memory_move_cost (mode, class, in)
     enum machine_mode mode;
     enum reg_class class;
     int in;
{
  if (FLOAT_CLASS_P (class))
    {
      int index;
      switch (mode)
	{
	  case SFmode:
	    index = 0;
	    break;
	  case DFmode:
	    index = 1;
	    break;
	  case XFmode:
	  case TFmode:
	    index = 2;
	    break;
	  default:
	    return 100;
	}
      return in ? ix86_cost->fp_load [index] : ix86_cost->fp_store [index];
    }
  if (SSE_CLASS_P (class))
    {
      int index;
      switch (GET_MODE_SIZE (mode))
	{
	  case 4:
	    index = 0;
	    break;
	  case 8:
	    index = 1;
	    break;
	  case 16:
	    index = 2;
	    break;
	  default:
	    return 100;
	}
      return in ? ix86_cost->sse_load [index] : ix86_cost->sse_store [index];
    }
  if (MMX_CLASS_P (class))
    {
      int index;
      switch (GET_MODE_SIZE (mode))
	{
	  case 4:
	    index = 0;
	    break;
	  case 8:
	    index = 1;
	    break;
	  default:
	    return 100;
	}
      return in ? ix86_cost->mmx_load [index] : ix86_cost->mmx_store [index];
    }
  switch (GET_MODE_SIZE (mode))
    {
      case 1:
	if (in)
	  return (Q_CLASS_P (class) ? ix86_cost->int_load[0]
		  : ix86_cost->movzbl_load);
	else
	  return (Q_CLASS_P (class) ? ix86_cost->int_store[0]
		  : ix86_cost->int_store[0] + 4);
	break;
      case 2:
	return in ? ix86_cost->int_load[1] : ix86_cost->int_store[1];
      default:
	/* Compute number of 32bit moves needed.  TFmode is moved as XFmode.  */
	if (mode == TFmode)
	  mode = XFmode;
	return ((in ? ix86_cost->int_load[2] : ix86_cost->int_store[2])
		* (int) GET_MODE_SIZE (mode) / 4);
    }
}