xref: /dports/devel/avr-gcc/gcc-10.2.0/gcc/expr.c

/* Convert tree expression to rtl instructions, for GNU compiler.
   Copyright (C) 1988-2020 Free Software Foundation, Inc.

This file is part of GCC.

GCC 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 3, or (at your option) any later
version.

GCC 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 GCC; see the file COPYING3.  If not see
<http://www.gnu.org/licenses/>.  */

#include "config.h"
#include "system.h"
#include "coretypes.h"
#include "backend.h"
#include "target.h"
#include "rtl.h"
#include "tree.h"
#include "gimple.h"
#include "predict.h"
#include "memmodel.h"
#include "tm_p.h"
#include "ssa.h"
#include "expmed.h"
#include "optabs.h"
#include "regs.h"
#include "emit-rtl.h"
#include "recog.h"
#include "cgraph.h"
#include "diagnostic.h"
#include "alias.h"
#include "fold-const.h"
#include "stor-layout.h"
#include "attribs.h"
#include "varasm.h"
#include "except.h"
#include "insn-attr.h"
#include "dojump.h"
#include "explow.h"
#include "calls.h"
#include "stmt.h"
/* Include expr.h after insn-config.h so we get HAVE_conditional_move.  */
#include "expr.h"
#include "optabs-tree.h"
#include "libfuncs.h"
#include "reload.h"
#include "langhooks.h"
#include "common/common-target.h"
#include "tree-dfa.h"
#include "tree-ssa-live.h"
#include "tree-outof-ssa.h"
#include "tree-ssa-address.h"
#include "builtins.h"
#include "ccmp.h"
#include "gimple-fold.h"
#include "rtx-vector-builder.h"


/* If this is nonzero, we do not bother generating VOLATILE
   around volatile memory references, and we are willing to
   output indirect addresses.  If cse is to follow, we reject
   indirect addresses so a useful potential cse is generated;
   if it is used only once, instruction combination will produce
   the same indirect address eventually.  */
int cse_not_expected;

static bool block_move_libcall_safe_for_call_parm (void);
static bool emit_block_move_via_pattern (rtx, rtx, rtx, unsigned, unsigned,
					 HOST_WIDE_INT, unsigned HOST_WIDE_INT,
					 unsigned HOST_WIDE_INT,
					 unsigned HOST_WIDE_INT, bool);
static void emit_block_move_via_loop (rtx, rtx, rtx, unsigned);
static void clear_by_pieces (rtx, unsigned HOST_WIDE_INT, unsigned int);
static rtx_insn *compress_float_constant (rtx, rtx);
static rtx get_subtarget (rtx);
static void store_constructor (tree, rtx, int, poly_int64, bool);
static rtx store_field (rtx, poly_int64, poly_int64, poly_uint64, poly_uint64,
			machine_mode, tree, alias_set_type, bool, bool);

static unsigned HOST_WIDE_INT highest_pow2_factor_for_target (const_tree, const_tree);

static int is_aligning_offset (const_tree, const_tree);
static rtx reduce_to_bit_field_precision (rtx, rtx, tree);
static rtx do_store_flag (sepops, rtx, machine_mode);
#ifdef PUSH_ROUNDING
static void emit_single_push_insn (machine_mode, rtx, tree);
#endif
static void do_tablejump (rtx, machine_mode, rtx, rtx, rtx,
			  profile_probability);
static rtx const_vector_from_tree (tree);
static rtx const_scalar_mask_from_tree (scalar_int_mode, tree);
static tree tree_expr_size (const_tree);
static HOST_WIDE_INT int_expr_size (tree);
static void convert_mode_scalar (rtx, rtx, int);


/* This is run to set up which modes can be used
   directly in memory and to initialize the block move optab.  It is run
   at the beginning of compilation and when the target is reinitialized.  */

void
init_expr_target (void)
{
  rtx pat;
  int num_clobbers;
  rtx mem, mem1;
  rtx reg;

  /* Try indexing by frame ptr and try by stack ptr.
     It is known that on the Convex the stack ptr isn't a valid index.
     With luck, one or the other is valid on any machine.  */
  mem = gen_rtx_MEM (word_mode, stack_pointer_rtx);
  mem1 = gen_rtx_MEM (word_mode, frame_pointer_rtx);

  /* A scratch register we can modify in-place below to avoid
     useless RTL allocations.  */
  reg = gen_rtx_REG (word_mode, LAST_VIRTUAL_REGISTER + 1);

  rtx_insn *insn = as_a<rtx_insn *> (rtx_alloc (INSN));
  pat = gen_rtx_SET (NULL_RTX, NULL_RTX);
  PATTERN (insn) = pat;

  for (machine_mode mode = VOIDmode; (int) mode < NUM_MACHINE_MODES;
       mode = (machine_mode) ((int) mode + 1))
    {
      int regno;

      direct_load[(int) mode] = direct_store[(int) mode] = 0;
      PUT_MODE (mem, mode);
      PUT_MODE (mem1, mode);

      /* See if there is some register that can be used in this mode and
	 directly loaded or stored from memory.  */

      if (mode != VOIDmode && mode != BLKmode)
	for (regno = 0; regno < FIRST_PSEUDO_REGISTER
	     && (direct_load[(int) mode] == 0 || direct_store[(int) mode] == 0);
	     regno++)
	  {
	    if (!targetm.hard_regno_mode_ok (regno, mode))
	      continue;

	    set_mode_and_regno (reg, mode, regno);

	    SET_SRC (pat) = mem;
	    SET_DEST (pat) = reg;
	    if (recog (pat, insn, &num_clobbers) >= 0)
	      direct_load[(int) mode] = 1;

	    SET_SRC (pat) = mem1;
	    SET_DEST (pat) = reg;
	    if (recog (pat, insn, &num_clobbers) >= 0)
	      direct_load[(int) mode] = 1;

	    SET_SRC (pat) = reg;
	    SET_DEST (pat) = mem;
	    if (recog (pat, insn, &num_clobbers) >= 0)
	      direct_store[(int) mode] = 1;

	    SET_SRC (pat) = reg;
	    SET_DEST (pat) = mem1;
	    if (recog (pat, insn, &num_clobbers) >= 0)
	      direct_store[(int) mode] = 1;
	  }
    }

  mem = gen_rtx_MEM (VOIDmode, gen_raw_REG (Pmode, LAST_VIRTUAL_REGISTER + 1));

  opt_scalar_float_mode mode_iter;
  FOR_EACH_MODE_IN_CLASS (mode_iter, MODE_FLOAT)
    {
      scalar_float_mode mode = mode_iter.require ();
      scalar_float_mode srcmode;
      FOR_EACH_MODE_UNTIL (srcmode, mode)
	{
	  enum insn_code ic;

	  ic = can_extend_p (mode, srcmode, 0);
	  if (ic == CODE_FOR_nothing)
	    continue;

	  PUT_MODE (mem, srcmode);

	  if (insn_operand_matches (ic, 1, mem))
	    float_extend_from_mem[mode][srcmode] = true;
	}
    }
}

/* This is run at the start of compiling a function.  */

void
init_expr (void)
{
  memset (&crtl->expr, 0, sizeof (crtl->expr));
}

/* Copy data from FROM to TO, where the machine modes are not the same.
   Both modes may be integer, or both may be floating, or both may be
   fixed-point.
   UNSIGNEDP should be nonzero if FROM is an unsigned type.
   This causes zero-extension instead of sign-extension.  */

void
convert_move (rtx to, rtx from, int unsignedp)
{
  machine_mode to_mode = GET_MODE (to);
  machine_mode from_mode = GET_MODE (from);

  gcc_assert (to_mode != BLKmode);
  gcc_assert (from_mode != BLKmode);

  /* If the source and destination are already the same, then there's
     nothing to do.  */
  if (to == from)
    return;

  /* If FROM is a SUBREG that indicates that we have already done at least
     the required extension, strip it.  We don't handle such SUBREGs as
     TO here.  */

  scalar_int_mode to_int_mode;
  if (GET_CODE (from) == SUBREG
      && SUBREG_PROMOTED_VAR_P (from)
      && is_a <scalar_int_mode> (to_mode, &to_int_mode)
      && (GET_MODE_PRECISION (subreg_promoted_mode (from))
	  >= GET_MODE_PRECISION (to_int_mode))
      && SUBREG_CHECK_PROMOTED_SIGN (from, unsignedp))
    {
      from = gen_lowpart (to_int_mode, SUBREG_REG (from));
      from_mode = to_int_mode;
    }

  gcc_assert (GET_CODE (to) != SUBREG || !SUBREG_PROMOTED_VAR_P (to));

  if (to_mode == from_mode
      || (from_mode == VOIDmode && CONSTANT_P (from)))
    {
      emit_move_insn (to, from);
      return;
    }

  if (VECTOR_MODE_P (to_mode) || VECTOR_MODE_P (from_mode))
    {
      if (GET_MODE_UNIT_PRECISION (to_mode)
	  > GET_MODE_UNIT_PRECISION (from_mode))
	{
	  optab op = unsignedp ? zext_optab : sext_optab;
	  insn_code icode = convert_optab_handler (op, to_mode, from_mode);
	  if (icode != CODE_FOR_nothing)
	    {
	      emit_unop_insn (icode, to, from,
			      unsignedp ? ZERO_EXTEND : SIGN_EXTEND);
	      return;
	    }
	}

      if (GET_MODE_UNIT_PRECISION (to_mode)
	  < GET_MODE_UNIT_PRECISION (from_mode))
	{
	  insn_code icode = convert_optab_handler (trunc_optab,
						   to_mode, from_mode);
	  if (icode != CODE_FOR_nothing)
	    {
	      emit_unop_insn (icode, to, from, TRUNCATE);
	      return;
	    }
	}

      gcc_assert (known_eq (GET_MODE_BITSIZE (from_mode),
			    GET_MODE_BITSIZE (to_mode)));

      if (VECTOR_MODE_P (to_mode))
	from = simplify_gen_subreg (to_mode, from, GET_MODE (from), 0);
      else
	to = simplify_gen_subreg (from_mode, to, GET_MODE (to), 0);

      emit_move_insn (to, from);
      return;
    }

  if (GET_CODE (to) == CONCAT && GET_CODE (from) == CONCAT)
    {
      convert_move (XEXP (to, 0), XEXP (from, 0), unsignedp);
      convert_move (XEXP (to, 1), XEXP (from, 1), unsignedp);
      return;
    }

  convert_mode_scalar (to, from, unsignedp);
}

/* Like convert_move, but deals only with scalar modes.  */

static void
convert_mode_scalar (rtx to, rtx from, int unsignedp)
{
  /* Both modes should be scalar types.  */
  scalar_mode from_mode = as_a <scalar_mode> (GET_MODE (from));
  scalar_mode to_mode = as_a <scalar_mode> (GET_MODE (to));
  bool to_real = SCALAR_FLOAT_MODE_P (to_mode);
  bool from_real = SCALAR_FLOAT_MODE_P (from_mode);
  enum insn_code code;
  rtx libcall;

  gcc_assert (to_real == from_real);

  /* rtx code for making an equivalent value.  */
  enum rtx_code equiv_code = (unsignedp < 0 ? UNKNOWN
			      : (unsignedp ? ZERO_EXTEND : SIGN_EXTEND));

  if (to_real)
    {
      rtx value;
      rtx_insn *insns;
      convert_optab tab;

      gcc_assert ((GET_MODE_PRECISION (from_mode)
		   != GET_MODE_PRECISION (to_mode))
		  || (DECIMAL_FLOAT_MODE_P (from_mode)
		      != DECIMAL_FLOAT_MODE_P (to_mode)));

      if (GET_MODE_PRECISION (from_mode) == GET_MODE_PRECISION (to_mode))
	/* Conversion between decimal float and binary float, same size.  */
	tab = DECIMAL_FLOAT_MODE_P (from_mode) ? trunc_optab : sext_optab;
      else if (GET_MODE_PRECISION (from_mode) < GET_MODE_PRECISION (to_mode))
	tab = sext_optab;
      else
	tab = trunc_optab;

      /* Try converting directly if the insn is supported.  */

      code = convert_optab_handler (tab, to_mode, from_mode);
      if (code != CODE_FOR_nothing)
	{
	  emit_unop_insn (code, to, from,
			  tab == sext_optab ? FLOAT_EXTEND : FLOAT_TRUNCATE);
	  return;
	}

      /* Otherwise use a libcall.  */
      libcall = convert_optab_libfunc (tab, to_mode, from_mode);

      /* Is this conversion implemented yet?  */
      gcc_assert (libcall);

      start_sequence ();
      value = emit_library_call_value (libcall, NULL_RTX, LCT_CONST, to_mode,
				       from, from_mode);
      insns = get_insns ();
      end_sequence ();
      emit_libcall_block (insns, to, value,
			  tab == trunc_optab ? gen_rtx_FLOAT_TRUNCATE (to_mode,
								       from)
			  : gen_rtx_FLOAT_EXTEND (to_mode, from));
      return;
    }

  /* Handle pointer conversion.  */			/* SPEE 900220.  */
  /* If the target has a converter from FROM_MODE to TO_MODE, use it.  */
  {
    convert_optab ctab;

    if (GET_MODE_PRECISION (from_mode) > GET_MODE_PRECISION (to_mode))
      ctab = trunc_optab;
    else if (unsignedp)
      ctab = zext_optab;
    else
      ctab = sext_optab;

    if (convert_optab_handler (ctab, to_mode, from_mode)
	!= CODE_FOR_nothing)
      {
	emit_unop_insn (convert_optab_handler (ctab, to_mode, from_mode),
			to, from, UNKNOWN);
	return;
      }
  }

  /* Targets are expected to provide conversion insns between PxImode and
     xImode for all MODE_PARTIAL_INT modes they use, but no others.  */
  if (GET_MODE_CLASS (to_mode) == MODE_PARTIAL_INT)
    {
      scalar_int_mode full_mode
	= smallest_int_mode_for_size (GET_MODE_BITSIZE (to_mode));

      gcc_assert (convert_optab_handler (trunc_optab, to_mode, full_mode)
		  != CODE_FOR_nothing);

      if (full_mode != from_mode)
	from = convert_to_mode (full_mode, from, unsignedp);
      emit_unop_insn (convert_optab_handler (trunc_optab, to_mode, full_mode),
		      to, from, UNKNOWN);
      return;
    }
  if (GET_MODE_CLASS (from_mode) == MODE_PARTIAL_INT)
    {
      rtx new_from;
      scalar_int_mode full_mode
	= smallest_int_mode_for_size (GET_MODE_BITSIZE (from_mode));
      convert_optab ctab = unsignedp ? zext_optab : sext_optab;
      enum insn_code icode;

      icode = convert_optab_handler (ctab, full_mode, from_mode);
      gcc_assert (icode != CODE_FOR_nothing);

      if (to_mode == full_mode)
	{
	  emit_unop_insn (icode, to, from, UNKNOWN);
	  return;
	}

      new_from = gen_reg_rtx (full_mode);
      emit_unop_insn (icode, new_from, from, UNKNOWN);

      /* else proceed to integer conversions below.  */
      from_mode = full_mode;
      from = new_from;
    }

   /* Make sure both are fixed-point modes or both are not.  */
   gcc_assert (ALL_SCALAR_FIXED_POINT_MODE_P (from_mode) ==
	       ALL_SCALAR_FIXED_POINT_MODE_P (to_mode));
   if (ALL_SCALAR_FIXED_POINT_MODE_P (from_mode))
    {
      /* If we widen from_mode to to_mode and they are in the same class,
	 we won't saturate the result.
	 Otherwise, always saturate the result to play safe.  */
      if (GET_MODE_CLASS (from_mode) == GET_MODE_CLASS (to_mode)
	  && GET_MODE_SIZE (from_mode) < GET_MODE_SIZE (to_mode))
	expand_fixed_convert (to, from, 0, 0);
      else
	expand_fixed_convert (to, from, 0, 1);
      return;
    }

  /* Now both modes are integers.  */

  /* Handle expanding beyond a word.  */
  if (GET_MODE_PRECISION (from_mode) < GET_MODE_PRECISION (to_mode)
      && GET_MODE_PRECISION (to_mode) > BITS_PER_WORD)
    {
      rtx_insn *insns;
      rtx lowpart;
      rtx fill_value;
      rtx lowfrom;
      int i;
      scalar_mode lowpart_mode;
      int nwords = CEIL (GET_MODE_SIZE (to_mode), UNITS_PER_WORD);

      /* Try converting directly if the insn is supported.  */
      if ((code = can_extend_p (to_mode, from_mode, unsignedp))
	  != CODE_FOR_nothing)
	{
	  /* If FROM is a SUBREG, put it into a register.  Do this
	     so that we always generate the same set of insns for
	     better cse'ing; if an intermediate assignment occurred,
	     we won't be doing the operation directly on the SUBREG.  */
	  if (optimize > 0 && GET_CODE (from) == SUBREG)
	    from = force_reg (from_mode, from);
	  emit_unop_insn (code, to, from, equiv_code);
	  return;
	}
      /* Next, try converting via full word.  */
      else if (GET_MODE_PRECISION (from_mode) < BITS_PER_WORD
	       && ((code = can_extend_p (to_mode, word_mode, unsignedp))
		   != CODE_FOR_nothing))
	{
	  rtx word_to = gen_reg_rtx (word_mode);
	  if (REG_P (to))
	    {
	      if (reg_overlap_mentioned_p (to, from))
		from = force_reg (from_mode, from);
	      emit_clobber (to);
	    }
	  convert_move (word_to, from, unsignedp);
	  emit_unop_insn (code, to, word_to, equiv_code);
	  return;
	}

      /* No special multiword conversion insn; do it by hand.  */
      start_sequence ();

      /* Since we will turn this into a no conflict block, we must ensure
         the source does not overlap the target so force it into an isolated
         register when maybe so.  Likewise for any MEM input, since the
         conversion sequence might require several references to it and we
         must ensure we're getting the same value every time.  */

      if (MEM_P (from) || reg_overlap_mentioned_p (to, from))
	from = force_reg (from_mode, from);

      /* Get a copy of FROM widened to a word, if necessary.  */
      if (GET_MODE_PRECISION (from_mode) < BITS_PER_WORD)
	lowpart_mode = word_mode;
      else
	lowpart_mode = from_mode;

      lowfrom = convert_to_mode (lowpart_mode, from, unsignedp);

      lowpart = gen_lowpart (lowpart_mode, to);
      emit_move_insn (lowpart, lowfrom);

      /* Compute the value to put in each remaining word.  */
      if (unsignedp)
	fill_value = const0_rtx;
      else
	fill_value = emit_store_flag_force (gen_reg_rtx (word_mode),
					    LT, lowfrom, const0_rtx,
					    lowpart_mode, 0, -1);

      /* Fill the remaining words.  */
      for (i = GET_MODE_SIZE (lowpart_mode) / UNITS_PER_WORD; i < nwords; i++)
	{
	  int index = (WORDS_BIG_ENDIAN ? nwords - i - 1 : i);
	  rtx subword = operand_subword (to, index, 1, to_mode);

	  gcc_assert (subword);

	  if (fill_value != subword)
	    emit_move_insn (subword, fill_value);
	}

      insns = get_insns ();
      end_sequence ();

      emit_insn (insns);
      return;
    }

  /* Truncating multi-word to a word or less.  */
  if (GET_MODE_PRECISION (from_mode) > BITS_PER_WORD
      && GET_MODE_PRECISION (to_mode) <= BITS_PER_WORD)
    {
      if (!((MEM_P (from)
	     && ! MEM_VOLATILE_P (from)
	     && direct_load[(int) to_mode]
	     && ! mode_dependent_address_p (XEXP (from, 0),
					    MEM_ADDR_SPACE (from)))
	    || REG_P (from)
	    || GET_CODE (from) == SUBREG))
	from = force_reg (from_mode, from);
      convert_move (to, gen_lowpart (word_mode, from), 0);
      return;
    }

  /* Now follow all the conversions between integers
     no more than a word long.  */

  /* For truncation, usually we can just refer to FROM in a narrower mode.  */
  if (GET_MODE_BITSIZE (to_mode) < GET_MODE_BITSIZE (from_mode)
      && TRULY_NOOP_TRUNCATION_MODES_P (to_mode, from_mode))
    {
      if (!((MEM_P (from)
	     && ! MEM_VOLATILE_P (from)
	     && direct_load[(int) to_mode]
	     && ! mode_dependent_address_p (XEXP (from, 0),
					    MEM_ADDR_SPACE (from)))
	    || REG_P (from)
	    || GET_CODE (from) == SUBREG))
	from = force_reg (from_mode, from);
      if (REG_P (from) && REGNO (from) < FIRST_PSEUDO_REGISTER
	  && !targetm.hard_regno_mode_ok (REGNO (from), to_mode))
	from = copy_to_reg (from);
      emit_move_insn (to, gen_lowpart (to_mode, from));
      return;
    }

  /* Handle extension.  */
  if (GET_MODE_PRECISION (to_mode) > GET_MODE_PRECISION (from_mode))
    {
      /* Convert directly if that works.  */
      if ((code = can_extend_p (to_mode, from_mode, unsignedp))
	  != CODE_FOR_nothing)
	{
	  emit_unop_insn (code, to, from, equiv_code);
	  return;
	}
      else
	{
	  rtx tmp;
	  int shift_amount;

	  /* Search for a mode to convert via.  */
	  opt_scalar_mode intermediate_iter;
	  FOR_EACH_MODE_FROM (intermediate_iter, from_mode)
	    {
	      scalar_mode intermediate = intermediate_iter.require ();
	      if (((can_extend_p (to_mode, intermediate, unsignedp)
		    != CODE_FOR_nothing)
		   || (GET_MODE_SIZE (to_mode) < GET_MODE_SIZE (intermediate)
		       && TRULY_NOOP_TRUNCATION_MODES_P (to_mode,
							 intermediate)))
		  && (can_extend_p (intermediate, from_mode, unsignedp)
		      != CODE_FOR_nothing))
		{
		  convert_move (to, convert_to_mode (intermediate, from,
						     unsignedp), unsignedp);
		  return;
		}
	    }

	  /* No suitable intermediate mode.
	     Generate what we need with	shifts.  */
	  shift_amount = (GET_MODE_PRECISION (to_mode)
			  - GET_MODE_PRECISION (from_mode));
	  from = gen_lowpart (to_mode, force_reg (from_mode, from));
	  tmp = expand_shift (LSHIFT_EXPR, to_mode, from, shift_amount,
			      to, unsignedp);
	  tmp = expand_shift (RSHIFT_EXPR, to_mode, tmp, shift_amount,
			      to, unsignedp);
	  if (tmp != to)
	    emit_move_insn (to, tmp);
	  return;
	}
    }

  /* Support special truncate insns for certain modes.  */
  if (convert_optab_handler (trunc_optab, to_mode,
			     from_mode) != CODE_FOR_nothing)
    {
      emit_unop_insn (convert_optab_handler (trunc_optab, to_mode, from_mode),
		      to, from, UNKNOWN);
      return;
    }

  /* Handle truncation of volatile memrefs, and so on;
     the things that couldn't be truncated directly,
     and for which there was no special instruction.

     ??? Code above formerly short-circuited this, for most integer
     mode pairs, with a force_reg in from_mode followed by a recursive
     call to this routine.  Appears always to have been wrong.  */
  if (GET_MODE_PRECISION (to_mode) < GET_MODE_PRECISION (from_mode))
    {
      rtx temp = force_reg (to_mode, gen_lowpart (to_mode, from));
      emit_move_insn (to, temp);
      return;
    }

  /* Mode combination is not recognized.  */
  gcc_unreachable ();
}

/* Return an rtx for a value that would result
   from converting X to mode MODE.
   Both X and MODE may be floating, or both integer.
   UNSIGNEDP is nonzero if X is an unsigned value.
   This can be done by referring to a part of X in place
   or by copying to a new temporary with conversion.  */

rtx
convert_to_mode (machine_mode mode, rtx x, int unsignedp)
{
  return convert_modes (mode, VOIDmode, x, unsignedp);
}

/* Return an rtx for a value that would result
   from converting X from mode OLDMODE to mode MODE.
   Both modes may be floating, or both integer.
   UNSIGNEDP is nonzero if X is an unsigned value.

   This can be done by referring to a part of X in place
   or by copying to a new temporary with conversion.

   You can give VOIDmode for OLDMODE, if you are sure X has a nonvoid mode.  */

rtx
convert_modes (machine_mode mode, machine_mode oldmode, rtx x, int unsignedp)
{
  rtx temp;
  scalar_int_mode int_mode;

  /* If FROM is a SUBREG that indicates that we have already done at least
     the required extension, strip it.  */

  if (GET_CODE (x) == SUBREG
      && SUBREG_PROMOTED_VAR_P (x)
      && is_a <scalar_int_mode> (mode, &int_mode)
      && (GET_MODE_PRECISION (subreg_promoted_mode (x))
	  >= GET_MODE_PRECISION (int_mode))
      && SUBREG_CHECK_PROMOTED_SIGN (x, unsignedp))
    x = gen_lowpart (int_mode, SUBREG_REG (x));

  if (GET_MODE (x) != VOIDmode)
    oldmode = GET_MODE (x);

  if (mode == oldmode)
    return x;

  if (CONST_SCALAR_INT_P (x)
      && is_int_mode (mode, &int_mode))
    {
      /* If the caller did not tell us the old mode, then there is not
	 much to do with respect to canonicalization.  We have to
	 assume that all the bits are significant.  */
      if (GET_MODE_CLASS (oldmode) != MODE_INT)
	oldmode = MAX_MODE_INT;
      wide_int w = wide_int::from (rtx_mode_t (x, oldmode),
				   GET_MODE_PRECISION (int_mode),
				   unsignedp ? UNSIGNED : SIGNED);
      return immed_wide_int_const (w, int_mode);
    }

  /* We can do this with a gen_lowpart if both desired and current modes
     are integer, and this is either a constant integer, a register, or a
     non-volatile MEM. */
  scalar_int_mode int_oldmode;
  if (is_int_mode (mode, &int_mode)
      && is_int_mode (oldmode, &int_oldmode)
      && GET_MODE_PRECISION (int_mode) <= GET_MODE_PRECISION (int_oldmode)
      && ((MEM_P (x) && !MEM_VOLATILE_P (x) && direct_load[(int) int_mode])
	  || CONST_POLY_INT_P (x)
          || (REG_P (x)
              && (!HARD_REGISTER_P (x)
		  || targetm.hard_regno_mode_ok (REGNO (x), int_mode))
              && TRULY_NOOP_TRUNCATION_MODES_P (int_mode, GET_MODE (x)))))
   return gen_lowpart (int_mode, x);

  /* Converting from integer constant into mode is always equivalent to an
     subreg operation.  */
  if (VECTOR_MODE_P (mode) && GET_MODE (x) == VOIDmode)
    {
      gcc_assert (known_eq (GET_MODE_BITSIZE (mode),
			    GET_MODE_BITSIZE (oldmode)));
      return simplify_gen_subreg (mode, x, oldmode, 0);
    }

  temp = gen_reg_rtx (mode);
  convert_move (temp, x, unsignedp);
  return temp;
}

/* Return the largest alignment we can use for doing a move (or store)
   of MAX_PIECES.  ALIGN is the largest alignment we could use.  */

static unsigned int
alignment_for_piecewise_move (unsigned int max_pieces, unsigned int align)
{
  scalar_int_mode tmode
    = int_mode_for_size (max_pieces * BITS_PER_UNIT, 1).require ();

  if (align >= GET_MODE_ALIGNMENT (tmode))
    align = GET_MODE_ALIGNMENT (tmode);
  else
    {
      scalar_int_mode xmode = NARROWEST_INT_MODE;
      opt_scalar_int_mode mode_iter;
      FOR_EACH_MODE_IN_CLASS (mode_iter, MODE_INT)
	{
	  tmode = mode_iter.require ();
	  if (GET_MODE_SIZE (tmode) > max_pieces
	      || targetm.slow_unaligned_access (tmode, align))
	    break;
	  xmode = tmode;
	}

      align = MAX (align, GET_MODE_ALIGNMENT (xmode));
    }

  return align;
}

/* Return the widest integer mode that is narrower than SIZE bytes.  */

static scalar_int_mode
widest_int_mode_for_size (unsigned int size)
{
  scalar_int_mode result = NARROWEST_INT_MODE;

  gcc_checking_assert (size > 1);

  opt_scalar_int_mode tmode;
  FOR_EACH_MODE_IN_CLASS (tmode, MODE_INT)
    if (GET_MODE_SIZE (tmode.require ()) < size)
      result = tmode.require ();

  return result;
}

/* Determine whether an operation OP on LEN bytes with alignment ALIGN can
   and should be performed piecewise.  */

static bool
can_do_by_pieces (unsigned HOST_WIDE_INT len, unsigned int align,
		  enum by_pieces_operation op)
{
  return targetm.use_by_pieces_infrastructure_p (len, align, op,
						 optimize_insn_for_speed_p ());
}

/* Determine whether the LEN bytes can be moved by using several move
   instructions.  Return nonzero if a call to move_by_pieces should
   succeed.  */

bool
can_move_by_pieces (unsigned HOST_WIDE_INT len, unsigned int align)
{
  return can_do_by_pieces (len, align, MOVE_BY_PIECES);
}

/* Return number of insns required to perform operation OP by pieces
   for L bytes.  ALIGN (in bits) is maximum alignment we can assume.  */

unsigned HOST_WIDE_INT
by_pieces_ninsns (unsigned HOST_WIDE_INT l, unsigned int align,
		  unsigned int max_size, by_pieces_operation op)
{
  unsigned HOST_WIDE_INT n_insns = 0;

  align = alignment_for_piecewise_move (MOVE_MAX_PIECES, align);

  while (max_size > 1 && l > 0)
    {
      scalar_int_mode mode = widest_int_mode_for_size (max_size);
      enum insn_code icode;

      unsigned int modesize = GET_MODE_SIZE (mode);

      icode = optab_handler (mov_optab, mode);
      if (icode != CODE_FOR_nothing && align >= GET_MODE_ALIGNMENT (mode))
	{
	  unsigned HOST_WIDE_INT n_pieces = l / modesize;
	  l %= modesize;
	  switch (op)
	    {
	    default:
	      n_insns += n_pieces;
	      break;

	    case COMPARE_BY_PIECES:
	      int batch = targetm.compare_by_pieces_branch_ratio (mode);
	      int batch_ops = 4 * batch - 1;
	      unsigned HOST_WIDE_INT full = n_pieces / batch;
	      n_insns += full * batch_ops;
	      if (n_pieces % batch != 0)
		n_insns++;
	      break;

	    }
	}
      max_size = modesize;
    }

  gcc_assert (!l);
  return n_insns;
}

/* Used when performing piecewise block operations, holds information
   about one of the memory objects involved.  The member functions
   can be used to generate code for loading from the object and
   updating the address when iterating.  */

class pieces_addr
{
  /* The object being referenced, a MEM.  Can be NULL_RTX to indicate
     stack pushes.  */
  rtx m_obj;
  /* The address of the object.  Can differ from that seen in the
     MEM rtx if we copied the address to a register.  */
  rtx m_addr;
  /* Nonzero if the address on the object has an autoincrement already,
     signifies whether that was an increment or decrement.  */
  signed char m_addr_inc;
  /* Nonzero if we intend to use autoinc without the address already
     having autoinc form.  We will insert add insns around each memory
     reference, expecting later passes to form autoinc addressing modes.
     The only supported options are predecrement and postincrement.  */
  signed char m_explicit_inc;
  /* True if we have either of the two possible cases of using
     autoincrement.  */
  bool m_auto;
  /* True if this is an address to be used for load operations rather
     than stores.  */
  bool m_is_load;

  /* Optionally, a function to obtain constants for any given offset into
     the objects, and data associated with it.  */
  by_pieces_constfn m_constfn;
  void *m_cfndata;
public:
  pieces_addr (rtx, bool, by_pieces_constfn, void *);
  rtx adjust (scalar_int_mode, HOST_WIDE_INT);
  void increment_address (HOST_WIDE_INT);
  void maybe_predec (HOST_WIDE_INT);
  void maybe_postinc (HOST_WIDE_INT);
  void decide_autoinc (machine_mode, bool, HOST_WIDE_INT);
  int get_addr_inc ()
  {
    return m_addr_inc;
  }
};

/* Initialize a pieces_addr structure from an object OBJ.  IS_LOAD is
   true if the operation to be performed on this object is a load
   rather than a store.  For stores, OBJ can be NULL, in which case we
   assume the operation is a stack push.  For loads, the optional
   CONSTFN and its associated CFNDATA can be used in place of the
   memory load.  */

pieces_addr::pieces_addr (rtx obj, bool is_load, by_pieces_constfn constfn,
			  void *cfndata)
  : m_obj (obj), m_is_load (is_load), m_constfn (constfn), m_cfndata (cfndata)
{
  m_addr_inc = 0;
  m_auto = false;
  if (obj)
    {
      rtx addr = XEXP (obj, 0);
      rtx_code code = GET_CODE (addr);
      m_addr = addr;
      bool dec = code == PRE_DEC || code == POST_DEC;
      bool inc = code == PRE_INC || code == POST_INC;
      m_auto = inc || dec;
      if (m_auto)
	m_addr_inc = dec ? -1 : 1;

      /* While we have always looked for these codes here, the code
	 implementing the memory operation has never handled them.
	 Support could be added later if necessary or beneficial.  */
      gcc_assert (code != PRE_INC && code != POST_DEC);
    }
  else
    {
      m_addr = NULL_RTX;
      if (!is_load)
	{
	  m_auto = true;
	  if (STACK_GROWS_DOWNWARD)
	    m_addr_inc = -1;
	  else
	    m_addr_inc = 1;
	}
      else
	gcc_assert (constfn != NULL);
    }
  m_explicit_inc = 0;
  if (constfn)
    gcc_assert (is_load);
}

/* Decide whether to use autoinc for an address involved in a memory op.
   MODE is the mode of the accesses, REVERSE is true if we've decided to
   perform the operation starting from the end, and LEN is the length of
   the operation.  Don't override an earlier decision to set m_auto.  */

void
pieces_addr::decide_autoinc (machine_mode ARG_UNUSED (mode), bool reverse,
			     HOST_WIDE_INT len)
{
  if (m_auto || m_obj == NULL_RTX)
    return;

  bool use_predec = (m_is_load
		     ? USE_LOAD_PRE_DECREMENT (mode)
		     : USE_STORE_PRE_DECREMENT (mode));
  bool use_postinc = (m_is_load
		      ? USE_LOAD_POST_INCREMENT (mode)
		      : USE_STORE_POST_INCREMENT (mode));
  machine_mode addr_mode = get_address_mode (m_obj);

  if (use_predec && reverse)
    {
      m_addr = copy_to_mode_reg (addr_mode,
				 plus_constant (addr_mode,
						m_addr, len));
      m_auto = true;
      m_explicit_inc = -1;
    }
  else if (use_postinc && !reverse)
    {
      m_addr = copy_to_mode_reg (addr_mode, m_addr);
      m_auto = true;
      m_explicit_inc = 1;
    }
  else if (CONSTANT_P (m_addr))
    m_addr = copy_to_mode_reg (addr_mode, m_addr);
}

/* Adjust the address to refer to the data at OFFSET in MODE.  If we
   are using autoincrement for this address, we don't add the offset,
   but we still modify the MEM's properties.  */

rtx
pieces_addr::adjust (scalar_int_mode mode, HOST_WIDE_INT offset)
{
  if (m_constfn)
    return m_constfn (m_cfndata, offset, mode);
  if (m_obj == NULL_RTX)
    return NULL_RTX;
  if (m_auto)
    return adjust_automodify_address (m_obj, mode, m_addr, offset);
  else
    return adjust_address (m_obj, mode, offset);
}

/* Emit an add instruction to increment the address by SIZE.  */

void
pieces_addr::increment_address (HOST_WIDE_INT size)
{
  rtx amount = gen_int_mode (size, GET_MODE (m_addr));
  emit_insn (gen_add2_insn (m_addr, amount));
}

/* If we are supposed to decrement the address after each access, emit code
   to do so now.  Increment by SIZE (which has should have the correct sign
   already).  */

void
pieces_addr::maybe_predec (HOST_WIDE_INT size)
{
  if (m_explicit_inc >= 0)
    return;
  gcc_assert (HAVE_PRE_DECREMENT);
  increment_address (size);
}

/* If we are supposed to decrement the address after each access, emit code
   to do so now.  Increment by SIZE.  */

void
pieces_addr::maybe_postinc (HOST_WIDE_INT size)
{
  if (m_explicit_inc <= 0)
    return;
  gcc_assert (HAVE_POST_INCREMENT);
  increment_address (size);
}

/* This structure is used by do_op_by_pieces to describe the operation
   to be performed.  */

class op_by_pieces_d
{
 protected:
  pieces_addr m_to, m_from;
  unsigned HOST_WIDE_INT m_len;
  HOST_WIDE_INT m_offset;
  unsigned int m_align;
  unsigned int m_max_size;
  bool m_reverse;

  /* Virtual functions, overriden by derived classes for the specific
     operation.  */
  virtual void generate (rtx, rtx, machine_mode) = 0;
  virtual bool prepare_mode (machine_mode, unsigned int) = 0;
  virtual void finish_mode (machine_mode)
  {
  }

 public:
  op_by_pieces_d (rtx, bool, rtx, bool, by_pieces_constfn, void *,
		  unsigned HOST_WIDE_INT, unsigned int);
  void run ();
};

/* The constructor for an op_by_pieces_d structure.  We require two
   objects named TO and FROM, which are identified as loads or stores
   by TO_LOAD and FROM_LOAD.  If FROM is a load, the optional FROM_CFN
   and its associated FROM_CFN_DATA can be used to replace loads with
   constant values.  LEN describes the length of the operation.  */

op_by_pieces_d::op_by_pieces_d (rtx to, bool to_load,
				rtx from, bool from_load,
				by_pieces_constfn from_cfn,
				void *from_cfn_data,
				unsigned HOST_WIDE_INT len,
				unsigned int align)
  : m_to (to, to_load, NULL, NULL),
    m_from (from, from_load, from_cfn, from_cfn_data),
    m_len (len), m_max_size (MOVE_MAX_PIECES + 1)
{
  int toi = m_to.get_addr_inc ();
  int fromi = m_from.get_addr_inc ();
  if (toi >= 0 && fromi >= 0)
    m_reverse = false;
  else if (toi <= 0 && fromi <= 0)
    m_reverse = true;
  else
    gcc_unreachable ();

  m_offset = m_reverse ? len : 0;
  align = MIN (to ? MEM_ALIGN (to) : align,
	       from ? MEM_ALIGN (from) : align);

  /* If copying requires more than two move insns,
     copy addresses to registers (to make displacements shorter)
     and use post-increment if available.  */
  if (by_pieces_ninsns (len, align, m_max_size, MOVE_BY_PIECES) > 2)
    {
      /* Find the mode of the largest comparison.  */
      scalar_int_mode mode = widest_int_mode_for_size (m_max_size);

      m_from.decide_autoinc (mode, m_reverse, len);
      m_to.decide_autoinc (mode, m_reverse, len);
    }

  align = alignment_for_piecewise_move (MOVE_MAX_PIECES, align);
  m_align = align;
}

/* This function contains the main loop used for expanding a block
   operation.  First move what we can in the largest integer mode,
   then go to successively smaller modes.  For every access, call
   GENFUN with the two operands and the EXTRA_DATA.  */

void
op_by_pieces_d::run ()
{
  while (m_max_size > 1 && m_len > 0)
    {
      scalar_int_mode mode = widest_int_mode_for_size (m_max_size);

      if (prepare_mode (mode, m_align))
	{
	  unsigned int size = GET_MODE_SIZE (mode);
	  rtx to1 = NULL_RTX, from1;

	  while (m_len >= size)
	    {
	      if (m_reverse)
		m_offset -= size;

	      to1 = m_to.adjust (mode, m_offset);
	      from1 = m_from.adjust (mode, m_offset);

	      m_to.maybe_predec (-(HOST_WIDE_INT)size);
	      m_from.maybe_predec (-(HOST_WIDE_INT)size);

	      generate (to1, from1, mode);

	      m_to.maybe_postinc (size);
	      m_from.maybe_postinc (size);

	      if (!m_reverse)
		m_offset += size;

	      m_len -= size;
	    }

	  finish_mode (mode);
	}

      m_max_size = GET_MODE_SIZE (mode);
    }

  /* The code above should have handled everything.  */
  gcc_assert (!m_len);
}

/* Derived class from op_by_pieces_d, providing support for block move
   operations.  */

class move_by_pieces_d : public op_by_pieces_d
{
  insn_gen_fn m_gen_fun;
  void generate (rtx, rtx, machine_mode);
  bool prepare_mode (machine_mode, unsigned int);

 public:
  move_by_pieces_d (rtx to, rtx from, unsigned HOST_WIDE_INT len,
		    unsigned int align)
    : op_by_pieces_d (to, false, from, true, NULL, NULL, len, align)
  {
  }
  rtx finish_retmode (memop_ret);
};

/* Return true if MODE can be used for a set of copies, given an
   alignment ALIGN.  Prepare whatever data is necessary for later
   calls to generate.  */

bool
move_by_pieces_d::prepare_mode (machine_mode mode, unsigned int align)
{
  insn_code icode = optab_handler (mov_optab, mode);
  m_gen_fun = GEN_FCN (icode);
  return icode != CODE_FOR_nothing && align >= GET_MODE_ALIGNMENT (mode);
}

/* A callback used when iterating for a compare_by_pieces_operation.
   OP0 and OP1 are the values that have been loaded and should be
   compared in MODE.  If OP0 is NULL, this means we should generate a
   push; otherwise EXTRA_DATA holds a pointer to a pointer to the insn
   gen function that should be used to generate the mode.  */

void
move_by_pieces_d::generate (rtx op0, rtx op1,
			    machine_mode mode ATTRIBUTE_UNUSED)
{
#ifdef PUSH_ROUNDING
  if (op0 == NULL_RTX)
    {
      emit_single_push_insn (mode, op1, NULL);
      return;
    }
#endif
  emit_insn (m_gen_fun (op0, op1));
}

/* Perform the final adjustment at the end of a string to obtain the
   correct return value for the block operation.
   Return value is based on RETMODE argument.  */

rtx
move_by_pieces_d::finish_retmode (memop_ret retmode)
{
  gcc_assert (!m_reverse);
  if (retmode == RETURN_END_MINUS_ONE)
    {
      m_to.maybe_postinc (-1);
      --m_offset;
    }
  return m_to.adjust (QImode, m_offset);
}

/* Generate several move instructions to copy LEN bytes from block FROM to
   block TO.  (These are MEM rtx's with BLKmode).

   If PUSH_ROUNDING is defined and TO is NULL, emit_single_push_insn is
   used to push FROM to the stack.

   ALIGN is maximum stack alignment we can assume.

   Return value is based on RETMODE argument.  */

rtx
move_by_pieces (rtx to, rtx from, unsigned HOST_WIDE_INT len,
		unsigned int align, memop_ret retmode)
{
#ifndef PUSH_ROUNDING
  if (to == NULL)
    gcc_unreachable ();
#endif

  move_by_pieces_d data (to, from, len, align);

  data.run ();

  if (retmode != RETURN_BEGIN)
    return data.finish_retmode (retmode);
  else
    return to;
}

/* Derived class from op_by_pieces_d, providing support for block move
   operations.  */

class store_by_pieces_d : public op_by_pieces_d
{
  insn_gen_fn m_gen_fun;
  void generate (rtx, rtx, machine_mode);
  bool prepare_mode (machine_mode, unsigned int);

 public:
  store_by_pieces_d (rtx to, by_pieces_constfn cfn, void *cfn_data,
		     unsigned HOST_WIDE_INT len, unsigned int align)
    : op_by_pieces_d (to, false, NULL_RTX, true, cfn, cfn_data, len, align)
  {
  }
  rtx finish_retmode (memop_ret);
};

/* Return true if MODE can be used for a set of stores, given an
   alignment ALIGN.  Prepare whatever data is necessary for later
   calls to generate.  */

bool
store_by_pieces_d::prepare_mode (machine_mode mode, unsigned int align)
{
  insn_code icode = optab_handler (mov_optab, mode);
  m_gen_fun = GEN_FCN (icode);
  return icode != CODE_FOR_nothing && align >= GET_MODE_ALIGNMENT (mode);
}

/* A callback used when iterating for a store_by_pieces_operation.
   OP0 and OP1 are the values that have been loaded and should be
   compared in MODE.  If OP0 is NULL, this means we should generate a
   push; otherwise EXTRA_DATA holds a pointer to a pointer to the insn
   gen function that should be used to generate the mode.  */

void
store_by_pieces_d::generate (rtx op0, rtx op1, machine_mode)
{
  emit_insn (m_gen_fun (op0, op1));
}

/* Perform the final adjustment at the end of a string to obtain the
   correct return value for the block operation.
   Return value is based on RETMODE argument.  */

rtx
store_by_pieces_d::finish_retmode (memop_ret retmode)
{
  gcc_assert (!m_reverse);
  if (retmode == RETURN_END_MINUS_ONE)
    {
      m_to.maybe_postinc (-1);
      --m_offset;
    }
  return m_to.adjust (QImode, m_offset);
}

/* Determine whether the LEN bytes generated by CONSTFUN can be
   stored to memory using several move instructions.  CONSTFUNDATA is
   a pointer which will be passed as argument in every CONSTFUN call.
   ALIGN is maximum alignment we can assume.  MEMSETP is true if this is
   a memset operation and false if it's a copy of a constant string.
   Return nonzero if a call to store_by_pieces should succeed.  */

int
can_store_by_pieces (unsigned HOST_WIDE_INT len,
		     rtx (*constfun) (void *, HOST_WIDE_INT, scalar_int_mode),
		     void *constfundata, unsigned int align, bool memsetp)
{
  unsigned HOST_WIDE_INT l;
  unsigned int max_size;
  HOST_WIDE_INT offset = 0;
  enum insn_code icode;
  int reverse;
  /* cst is set but not used if LEGITIMATE_CONSTANT doesn't use it.  */
  rtx cst ATTRIBUTE_UNUSED;

  if (len == 0)
    return 1;

  if (!targetm.use_by_pieces_infrastructure_p (len, align,
					       memsetp
						 ? SET_BY_PIECES
						 : STORE_BY_PIECES,
					       optimize_insn_for_speed_p ()))
    return 0;

  align = alignment_for_piecewise_move (STORE_MAX_PIECES, align);

  /* We would first store what we can in the largest integer mode, then go to
     successively smaller modes.  */

  for (reverse = 0;
       reverse <= (HAVE_PRE_DECREMENT || HAVE_POST_DECREMENT);
       reverse++)
    {
      l = len;
      max_size = STORE_MAX_PIECES + 1;
      while (max_size > 1 && l > 0)
	{
	  scalar_int_mode mode = widest_int_mode_for_size (max_size);

	  icode = optab_handler (mov_optab, mode);
	  if (icode != CODE_FOR_nothing
	      && align >= GET_MODE_ALIGNMENT (mode))
	    {
	      unsigned int size = GET_MODE_SIZE (mode);

	      while (l >= size)
		{
		  if (reverse)
		    offset -= size;

		  cst = (*constfun) (constfundata, offset, mode);
		  if (!targetm.legitimate_constant_p (mode, cst))
		    return 0;

		  if (!reverse)
		    offset += size;

		  l -= size;
		}
	    }

	  max_size = GET_MODE_SIZE (mode);
	}

      /* The code above should have handled everything.  */
      gcc_assert (!l);
    }

  return 1;
}

/* Generate several move instructions to store LEN bytes generated by
   CONSTFUN to block TO.  (A MEM rtx with BLKmode).  CONSTFUNDATA is a
   pointer which will be passed as argument in every CONSTFUN call.
   ALIGN is maximum alignment we can assume.  MEMSETP is true if this is
   a memset operation and false if it's a copy of a constant string.
   Return value is based on RETMODE argument.  */

rtx
store_by_pieces (rtx to, unsigned HOST_WIDE_INT len,
		 rtx (*constfun) (void *, HOST_WIDE_INT, scalar_int_mode),
		 void *constfundata, unsigned int align, bool memsetp,
		 memop_ret retmode)
{
  if (len == 0)
    {
      gcc_assert (retmode != RETURN_END_MINUS_ONE);
      return to;
    }

  gcc_assert (targetm.use_by_pieces_infrastructure_p
		(len, align,
		 memsetp ? SET_BY_PIECES : STORE_BY_PIECES,
		 optimize_insn_for_speed_p ()));

  store_by_pieces_d data (to, constfun, constfundata, len, align);
  data.run ();

  if (retmode != RETURN_BEGIN)
    return data.finish_retmode (retmode);
  else
    return to;
}

/* Callback routine for clear_by_pieces.
   Return const0_rtx unconditionally.  */

static rtx
clear_by_pieces_1 (void *, HOST_WIDE_INT, scalar_int_mode)
{
  return const0_rtx;
}

/* Generate several move instructions to clear LEN bytes of block TO.  (A MEM
   rtx with BLKmode).  ALIGN is maximum alignment we can assume.  */

static void
clear_by_pieces (rtx to, unsigned HOST_WIDE_INT len, unsigned int align)
{
  if (len == 0)
    return;

  store_by_pieces_d data (to, clear_by_pieces_1, NULL, len, align);
  data.run ();
}

/* Context used by compare_by_pieces_genfn.  It stores the fail label
   to jump to in case of miscomparison, and for branch ratios greater than 1,
   it stores an accumulator and the current and maximum counts before
   emitting another branch.  */

class compare_by_pieces_d : public op_by_pieces_d
{
  rtx_code_label *m_fail_label;
  rtx m_accumulator;
  int m_count, m_batch;

  void generate (rtx, rtx, machine_mode);
  bool prepare_mode (machine_mode, unsigned int);
  void finish_mode (machine_mode);
 public:
  compare_by_pieces_d (rtx op0, rtx op1, by_pieces_constfn op1_cfn,
		       void *op1_cfn_data, HOST_WIDE_INT len, int align,
		       rtx_code_label *fail_label)
    : op_by_pieces_d (op0, true, op1, true, op1_cfn, op1_cfn_data, len, align)
  {
    m_fail_label = fail_label;
  }
};

/* A callback used when iterating for a compare_by_pieces_operation.
   OP0 and OP1 are the values that have been loaded and should be
   compared in MODE.  DATA holds a pointer to the compare_by_pieces_data
   context structure.  */

void
compare_by_pieces_d::generate (rtx op0, rtx op1, machine_mode mode)
{
  if (m_batch > 1)
    {
      rtx temp = expand_binop (mode, sub_optab, op0, op1, NULL_RTX,
			       true, OPTAB_LIB_WIDEN);
      if (m_count != 0)
	temp = expand_binop (mode, ior_optab, m_accumulator, temp, temp,
			     true, OPTAB_LIB_WIDEN);
      m_accumulator = temp;

      if (++m_count < m_batch)
	return;

      m_count = 0;
      op0 = m_accumulator;
      op1 = const0_rtx;
      m_accumulator = NULL_RTX;
    }
  do_compare_rtx_and_jump (op0, op1, NE, true, mode, NULL_RTX, NULL,
			   m_fail_label, profile_probability::uninitialized ());
}

/* Return true if MODE can be used for a set of moves and comparisons,
   given an alignment ALIGN.  Prepare whatever data is necessary for
   later calls to generate.  */

bool
compare_by_pieces_d::prepare_mode (machine_mode mode, unsigned int align)
{
  insn_code icode = optab_handler (mov_optab, mode);
  if (icode == CODE_FOR_nothing
      || align < GET_MODE_ALIGNMENT (mode)
      || !can_compare_p (EQ, mode, ccp_jump))
    return false;
  m_batch = targetm.compare_by_pieces_branch_ratio (mode);
  if (m_batch < 0)
    return false;
  m_accumulator = NULL_RTX;
  m_count = 0;
  return true;
}

/* Called after expanding a series of comparisons in MODE.  If we have
   accumulated results for which we haven't emitted a branch yet, do
   so now.  */

void
compare_by_pieces_d::finish_mode (machine_mode mode)
{
  if (m_accumulator != NULL_RTX)
    do_compare_rtx_and_jump (m_accumulator, const0_rtx, NE, true, mode,
			     NULL_RTX, NULL, m_fail_label,
			     profile_probability::uninitialized ());
}

/* Generate several move instructions to compare LEN bytes from blocks
   ARG0 and ARG1.  (These are MEM rtx's with BLKmode).

   If PUSH_ROUNDING is defined and TO is NULL, emit_single_push_insn is
   used to push FROM to the stack.

   ALIGN is maximum stack alignment we can assume.

   Optionally, the caller can pass a constfn and associated data in A1_CFN
   and A1_CFN_DATA. describing that the second operand being compared is a
   known constant and how to obtain its data.  */

static rtx
compare_by_pieces (rtx arg0, rtx arg1, unsigned HOST_WIDE_INT len,
		   rtx target, unsigned int align,
		   by_pieces_constfn a1_cfn, void *a1_cfn_data)
{
  rtx_code_label *fail_label = gen_label_rtx ();
  rtx_code_label *end_label = gen_label_rtx ();

  if (target == NULL_RTX
      || !REG_P (target) || REGNO (target) < FIRST_PSEUDO_REGISTER)
    target = gen_reg_rtx (TYPE_MODE (integer_type_node));

  compare_by_pieces_d data (arg0, arg1, a1_cfn, a1_cfn_data, len, align,
			    fail_label);

  data.run ();

  emit_move_insn (target, const0_rtx);
  emit_jump (end_label);
  emit_barrier ();
  emit_label (fail_label);
  emit_move_insn (target, const1_rtx);
  emit_label (end_label);

  return target;
}

/* Emit code to move a block Y to a block X.  This may be done with
   string-move instructions, with multiple scalar move instructions,
   or with a library call.

   Both X and Y must be MEM rtx's (perhaps inside VOLATILE) with mode BLKmode.
   SIZE is an rtx that says how long they are.
   ALIGN is the maximum alignment we can assume they have.
   METHOD describes what kind of copy this is, and what mechanisms may be used.
   MIN_SIZE is the minimal size of block to move
   MAX_SIZE is the maximal size of block to move, if it cannot be represented
   in unsigned HOST_WIDE_INT, than it is mask of all ones.

   Return the address of the new block, if memcpy is called and returns it,
   0 otherwise.  */

rtx
emit_block_move_hints (rtx x, rtx y, rtx size, enum block_op_methods method,
		       unsigned int expected_align, HOST_WIDE_INT expected_size,
		       unsigned HOST_WIDE_INT min_size,
		       unsigned HOST_WIDE_INT max_size,
		       unsigned HOST_WIDE_INT probable_max_size,
		       bool bail_out_libcall, bool *is_move_done,
		       bool might_overlap)
{
  int may_use_call;
  rtx retval = 0;
  unsigned int align;

  if (is_move_done)
    *is_move_done = true;

  gcc_assert (size);
  if (CONST_INT_P (size) && INTVAL (size) == 0)
    return 0;

  switch (method)
    {
    case BLOCK_OP_NORMAL:
    case BLOCK_OP_TAILCALL:
      may_use_call = 1;
      break;

    case BLOCK_OP_CALL_PARM:
      may_use_call = block_move_libcall_safe_for_call_parm ();

      /* Make inhibit_defer_pop nonzero around the library call
	 to force it to pop the arguments right away.  */
      NO_DEFER_POP;
      break;

    case BLOCK_OP_NO_LIBCALL:
      may_use_call = 0;
      break;

    case BLOCK_OP_NO_LIBCALL_RET:
      may_use_call = -1;
      break;

    default:
      gcc_unreachable ();
    }

  gcc_assert (MEM_P (x) && MEM_P (y));
  align = MIN (MEM_ALIGN (x), MEM_ALIGN (y));
  gcc_assert (align >= BITS_PER_UNIT);

  /* Make sure we've got BLKmode addresses; store_one_arg can decide that
     block copy is more efficient for other large modes, e.g. DCmode.  */
  x = adjust_address (x, BLKmode, 0);
  y = adjust_address (y, BLKmode, 0);

  /* Set MEM_SIZE as appropriate for this block copy.  The main place this
     can be incorrect is coming from __builtin_memcpy.  */
  poly_int64 const_size;
  if (poly_int_rtx_p (size, &const_size))
    {
      x = shallow_copy_rtx (x);
      y = shallow_copy_rtx (y);
      set_mem_size (x, const_size);
      set_mem_size (y, const_size);
    }

  bool pieces_ok = CONST_INT_P (size)
    && can_move_by_pieces (INTVAL (size), align);
  bool pattern_ok = false;

  if (!pieces_ok || might_overlap)
    {
      pattern_ok
	= emit_block_move_via_pattern (x, y, size, align,
				       expected_align, expected_size,
				       min_size, max_size, probable_max_size,
				       might_overlap);
      if (!pattern_ok && might_overlap)
	{
	  /* Do not try any of the other methods below as they are not safe
	     for overlapping moves.  */
	  *is_move_done = false;
	  return retval;
	}
    }

  if (pattern_ok)
    ;
  else if (pieces_ok)
    move_by_pieces (x, y, INTVAL (size), align, RETURN_BEGIN);
  else if (may_use_call && !might_overlap
	   && ADDR_SPACE_GENERIC_P (MEM_ADDR_SPACE (x))
	   && ADDR_SPACE_GENERIC_P (MEM_ADDR_SPACE (y)))
    {
      if (bail_out_libcall)
	{
	  if (is_move_done)
	    *is_move_done = false;
	  return retval;
	}

      if (may_use_call < 0)
	return pc_rtx;

      retval = emit_block_copy_via_libcall (x, y, size,
					    method == BLOCK_OP_TAILCALL);
    }
  else if (might_overlap)
    *is_move_done = false;
  else
    emit_block_move_via_loop (x, y, size, align);

  if (method == BLOCK_OP_CALL_PARM)
    OK_DEFER_POP;

  return retval;
}

rtx
emit_block_move (rtx x, rtx y, rtx size, enum block_op_methods method)
{
  unsigned HOST_WIDE_INT max, min = 0;
  if (GET_CODE (size) == CONST_INT)
    min = max = UINTVAL (size);
  else
    max = GET_MODE_MASK (GET_MODE (size));
  return emit_block_move_hints (x, y, size, method, 0, -1,
				min, max, max);
}

/* A subroutine of emit_block_move.  Returns true if calling the
   block move libcall will not clobber any parameters which may have
   already been placed on the stack.  */

static bool
block_move_libcall_safe_for_call_parm (void)
{
  tree fn;

  /* If arguments are pushed on the stack, then they're safe.  */
  if (PUSH_ARGS)
    return true;

  /* If registers go on the stack anyway, any argument is sure to clobber
     an outgoing argument.  */
#if defined (REG_PARM_STACK_SPACE)
  fn = builtin_decl_implicit (BUILT_IN_MEMCPY);
  /* Avoid set but not used warning if *REG_PARM_STACK_SPACE doesn't
     depend on its argument.  */
  (void) fn;
  if (OUTGOING_REG_PARM_STACK_SPACE ((!fn ? NULL_TREE : TREE_TYPE (fn)))
      && REG_PARM_STACK_SPACE (fn) != 0)
    return false;
#endif

  /* If any argument goes in memory, then it might clobber an outgoing
     argument.  */
  {
    CUMULATIVE_ARGS args_so_far_v;
    cumulative_args_t args_so_far;
    tree arg;

    fn = builtin_decl_implicit (BUILT_IN_MEMCPY);
    INIT_CUMULATIVE_ARGS (args_so_far_v, TREE_TYPE (fn), NULL_RTX, 0, 3);
    args_so_far = pack_cumulative_args (&args_so_far_v);

    arg = TYPE_ARG_TYPES (TREE_TYPE (fn));
    for ( ; arg != void_list_node ; arg = TREE_CHAIN (arg))
      {
	machine_mode mode = TYPE_MODE (TREE_VALUE (arg));
	function_arg_info arg_info (mode, /*named=*/true);
	rtx tmp = targetm.calls.function_arg (args_so_far, arg_info);
	if (!tmp || !REG_P (tmp))
	  return false;
	if (targetm.calls.arg_partial_bytes (args_so_far, arg_info))
	  return false;
	targetm.calls.function_arg_advance (args_so_far, arg_info);
      }
  }
  return true;
}

/* A subroutine of emit_block_move.  Expand a cpymem or movmem pattern;
   return true if successful.

   X is the destination of the copy or move.
   Y is the source of the copy or move.
   SIZE is the size of the block to be moved.

   MIGHT_OVERLAP indicates this originated with expansion of a
   builtin_memmove() and the source and destination blocks may
   overlap.
  */

static bool
emit_block_move_via_pattern (rtx x, rtx y, rtx size, unsigned int align,
			     unsigned int expected_align,
			     HOST_WIDE_INT expected_size,
			     unsigned HOST_WIDE_INT min_size,
			     unsigned HOST_WIDE_INT max_size,
			     unsigned HOST_WIDE_INT probable_max_size,
			     bool might_overlap)
{
  if (expected_align < align)
    expected_align = align;
  if (expected_size != -1)
    {
      if ((unsigned HOST_WIDE_INT)expected_size > probable_max_size)
	expected_size = probable_max_size;
      if ((unsigned HOST_WIDE_INT)expected_size < min_size)
	expected_size = min_size;
    }

  /* Since this is a move insn, we don't care about volatility.  */
  temporary_volatile_ok v (true);

  /* Try the most limited insn first, because there's no point
     including more than one in the machine description unless
     the more limited one has some advantage.  */

  opt_scalar_int_mode mode_iter;
  FOR_EACH_MODE_IN_CLASS (mode_iter, MODE_INT)
    {
      scalar_int_mode mode = mode_iter.require ();
      enum insn_code code;
      if (might_overlap)
	code = direct_optab_handler (movmem_optab, mode);
      else
	code = direct_optab_handler (cpymem_optab, mode);

      if (code != CODE_FOR_nothing
	  /* We don't need MODE to be narrower than BITS_PER_HOST_WIDE_INT
	     here because if SIZE is less than the mode mask, as it is
	     returned by the macro, it will definitely be less than the
	     actual mode mask.  Since SIZE is within the Pmode address
	     space, we limit MODE to Pmode.  */
	  && ((CONST_INT_P (size)
	       && ((unsigned HOST_WIDE_INT) INTVAL (size)
		   <= (GET_MODE_MASK (mode) >> 1)))
	      || max_size <= (GET_MODE_MASK (mode) >> 1)
	      || GET_MODE_BITSIZE (mode) >= GET_MODE_BITSIZE (Pmode)))
	{
	  class expand_operand ops[9];
	  unsigned int nops;

	  /* ??? When called via emit_block_move_for_call, it'd be
	     nice if there were some way to inform the backend, so
	     that it doesn't fail the expansion because it thinks
	     emitting the libcall would be more efficient.  */
	  nops = insn_data[(int) code].n_generator_args;
	  gcc_assert (nops == 4 || nops == 6 || nops == 8 || nops == 9);

	  create_fixed_operand (&ops[0], x);
	  create_fixed_operand (&ops[1], y);
	  /* The check above guarantees that this size conversion is valid.  */
	  create_convert_operand_to (&ops[2], size, mode, true);
	  create_integer_operand (&ops[3], align / BITS_PER_UNIT);
	  if (nops >= 6)
	    {
	      create_integer_operand (&ops[4], expected_align / BITS_PER_UNIT);
	      create_integer_operand (&ops[5], expected_size);
	    }
	  if (nops >= 8)
	    {
	      create_integer_operand (&ops[6], min_size);
	      /* If we cannot represent the maximal size,
		 make parameter NULL.  */
	      if ((HOST_WIDE_INT) max_size != -1)
	        create_integer_operand (&ops[7], max_size);
	      else
		create_fixed_operand (&ops[7], NULL);
	    }
	  if (nops == 9)
	    {
	      /* If we cannot represent the maximal size,
		 make parameter NULL.  */
	      if ((HOST_WIDE_INT) probable_max_size != -1)
	        create_integer_operand (&ops[8], probable_max_size);
	      else
		create_fixed_operand (&ops[8], NULL);
	    }
	  if (maybe_expand_insn (code, nops, ops))
	    return true;
	}
    }

  return false;
}

/* A subroutine of emit_block_move.  Copy the data via an explicit
   loop.  This is used only when libcalls are forbidden.  */
/* ??? It'd be nice to copy in hunks larger than QImode.  */

static void
emit_block_move_via_loop (rtx x, rtx y, rtx size,
			  unsigned int align ATTRIBUTE_UNUSED)
{
  rtx_code_label *cmp_label, *top_label;
  rtx iter, x_addr, y_addr, tmp;
  machine_mode x_addr_mode = get_address_mode (x);
  machine_mode y_addr_mode = get_address_mode (y);
  machine_mode iter_mode;

  iter_mode = GET_MODE (size);
  if (iter_mode == VOIDmode)
    iter_mode = word_mode;

  top_label = gen_label_rtx ();
  cmp_label = gen_label_rtx ();
  iter = gen_reg_rtx (iter_mode);

  emit_move_insn (iter, const0_rtx);

  x_addr = force_operand (XEXP (x, 0), NULL_RTX);
  y_addr = force_operand (XEXP (y, 0), NULL_RTX);
  do_pending_stack_adjust ();

  emit_jump (cmp_label);
  emit_label (top_label);

  tmp = convert_modes (x_addr_mode, iter_mode, iter, true);
  x_addr = simplify_gen_binary (PLUS, x_addr_mode, x_addr, tmp);

  if (x_addr_mode != y_addr_mode)
    tmp = convert_modes (y_addr_mode, iter_mode, iter, true);
  y_addr = simplify_gen_binary (PLUS, y_addr_mode, y_addr, tmp);

  x = change_address (x, QImode, x_addr);
  y = change_address (y, QImode, y_addr);

  emit_move_insn (x, y);

  tmp = expand_simple_binop (iter_mode, PLUS, iter, const1_rtx, iter,
			     true, OPTAB_LIB_WIDEN);
  if (tmp != iter)
    emit_move_insn (iter, tmp);

  emit_label (cmp_label);

  emit_cmp_and_jump_insns (iter, size, LT, NULL_RTX, iter_mode,
			   true, top_label,
			   profile_probability::guessed_always ()
				.apply_scale (9, 10));
}

/* Expand a call to memcpy or memmove or memcmp, and return the result.
   TAILCALL is true if this is a tail call.  */

rtx
emit_block_op_via_libcall (enum built_in_function fncode, rtx dst, rtx src,
			   rtx size, bool tailcall)
{
  rtx dst_addr, src_addr;
  tree call_expr, dst_tree, src_tree, size_tree;
  machine_mode size_mode;

  /* Since dst and src are passed to a libcall, mark the corresponding
     tree EXPR as addressable.  */
  tree dst_expr = MEM_EXPR (dst);
  tree src_expr = MEM_EXPR (src);
  if (dst_expr)
    mark_addressable (dst_expr);
  if (src_expr)
    mark_addressable (src_expr);

  dst_addr = copy_addr_to_reg (XEXP (dst, 0));
  dst_addr = convert_memory_address (ptr_mode, dst_addr);
  dst_tree = make_tree (ptr_type_node, dst_addr);

  src_addr = copy_addr_to_reg (XEXP (src, 0));
  src_addr = convert_memory_address (ptr_mode, src_addr);
  src_tree = make_tree (ptr_type_node, src_addr);

  size_mode = TYPE_MODE (sizetype);
  size = convert_to_mode (size_mode, size, 1);
  size = copy_to_mode_reg (size_mode, size);
  size_tree = make_tree (sizetype, size);

  /* It is incorrect to use the libcall calling conventions for calls to
     memcpy/memmove/memcmp because they can be provided by the user.  */
  tree fn = builtin_decl_implicit (fncode);
  call_expr = build_call_expr (fn, 3, dst_tree, src_tree, size_tree);
  CALL_EXPR_TAILCALL (call_expr) = tailcall;

  return expand_call (call_expr, NULL_RTX, false);
}

/* Try to expand cmpstrn or cmpmem operation ICODE with the given operands.
   ARG3_TYPE is the type of ARG3_RTX.  Return the result rtx on success,
   otherwise return null.  */

rtx
expand_cmpstrn_or_cmpmem (insn_code icode, rtx target, rtx arg1_rtx,
			  rtx arg2_rtx, tree arg3_type, rtx arg3_rtx,
			  HOST_WIDE_INT align)
{
  machine_mode insn_mode = insn_data[icode].operand[0].mode;

  if (target && (!REG_P (target) || HARD_REGISTER_P (target)))
    target = NULL_RTX;

  class expand_operand ops[5];
  create_output_operand (&ops[0], target, insn_mode);
  create_fixed_operand (&ops[1], arg1_rtx);
  create_fixed_operand (&ops[2], arg2_rtx);
  create_convert_operand_from (&ops[3], arg3_rtx, TYPE_MODE (arg3_type),
			       TYPE_UNSIGNED (arg3_type));
  create_integer_operand (&ops[4], align);
  if (maybe_expand_insn (icode, 5, ops))
    return ops[0].value;
  return NULL_RTX;
}

/* Expand a block compare between X and Y with length LEN using the
   cmpmem optab, placing the result in TARGET.  LEN_TYPE is the type
   of the expression that was used to calculate the length.  ALIGN
   gives the known minimum common alignment.  */

static rtx
emit_block_cmp_via_cmpmem (rtx x, rtx y, rtx len, tree len_type, rtx target,
			   unsigned align)
{
  /* Note: The cmpstrnsi pattern, if it exists, is not suitable for
     implementing memcmp because it will stop if it encounters two
     zero bytes.  */
  insn_code icode = direct_optab_handler (cmpmem_optab, SImode);

  if (icode == CODE_FOR_nothing)
    return NULL_RTX;

  return expand_cmpstrn_or_cmpmem (icode, target, x, y, len_type, len, align);
}

/* Emit code to compare a block Y to a block X.  This may be done with
   string-compare instructions, with multiple scalar instructions,
   or with a library call.

   Both X and Y must be MEM rtx's.  LEN is an rtx that says how long
   they are.  LEN_TYPE is the type of the expression that was used to
   calculate it.

   If EQUALITY_ONLY is true, it means we don't have to return the tri-state
   value of a normal memcmp call, instead we can just compare for equality.
   If FORCE_LIBCALL is true, we should emit a call to memcmp rather than
   returning NULL_RTX.

   Optionally, the caller can pass a constfn and associated data in Y_CFN
   and Y_CFN_DATA. describing that the second operand being compared is a
   known constant and how to obtain its data.
   Return the result of the comparison, or NULL_RTX if we failed to
   perform the operation.  */

rtx
emit_block_cmp_hints (rtx x, rtx y, rtx len, tree len_type, rtx target,
		      bool equality_only, by_pieces_constfn y_cfn,
		      void *y_cfndata)
{
  rtx result = 0;

  if (CONST_INT_P (len) && INTVAL (len) == 0)
    return const0_rtx;

  gcc_assert (MEM_P (x) && MEM_P (y));
  unsigned int align = MIN (MEM_ALIGN (x), MEM_ALIGN (y));
  gcc_assert (align >= BITS_PER_UNIT);

  x = adjust_address (x, BLKmode, 0);
  y = adjust_address (y, BLKmode, 0);

  if (equality_only
      && CONST_INT_P (len)
      && can_do_by_pieces (INTVAL (len), align, COMPARE_BY_PIECES))
    result = compare_by_pieces (x, y, INTVAL (len), target, align,
				y_cfn, y_cfndata);
  else
    result = emit_block_cmp_via_cmpmem (x, y, len, len_type, target, align);

  return result;
}

/* Copy all or part of a value X into registers starting at REGNO.
   The number of registers to be filled is NREGS.  */

void
move_block_to_reg (int regno, rtx x, int nregs, machine_mode mode)
{
  if (nregs == 0)
    return;

  if (CONSTANT_P (x) && !targetm.legitimate_constant_p (mode, x))
    x = validize_mem (force_const_mem (mode, x));

  /* See if the machine can do this with a load multiple insn.  */
  if (targetm.have_load_multiple ())
    {
      rtx_insn *last = get_last_insn ();
      rtx first = gen_rtx_REG (word_mode, regno);
      if (rtx_insn *pat = targetm.gen_load_multiple (first, x,
						     GEN_INT (nregs)))
	{
	  emit_insn (pat);
	  return;
	}
      else
	delete_insns_since (last);
    }

  for (int i = 0; i < nregs; i++)
    emit_move_insn (gen_rtx_REG (word_mode, regno + i),
		    operand_subword_force (x, i, mode));
}

/* Copy all or part of a BLKmode value X out of registers starting at REGNO.
   The number of registers to be filled is NREGS.  */

void
move_block_from_reg (int regno, rtx x, int nregs)
{
  if (nregs == 0)
    return;

  /* See if the machine can do this with a store multiple insn.  */
  if (targetm.have_store_multiple ())
    {
      rtx_insn *last = get_last_insn ();
      rtx first = gen_rtx_REG (word_mode, regno);
      if (rtx_insn *pat = targetm.gen_store_multiple (x, first,
						      GEN_INT (nregs)))
	{
	  emit_insn (pat);
	  return;
	}
      else
	delete_insns_since (last);
    }

  for (int i = 0; i < nregs; i++)
    {
      rtx tem = operand_subword (x, i, 1, BLKmode);

      gcc_assert (tem);

      emit_move_insn (tem, gen_rtx_REG (word_mode, regno + i));
    }
}

/* Generate a PARALLEL rtx for a new non-consecutive group of registers from
   ORIG, where ORIG is a non-consecutive group of registers represented by
   a PARALLEL.  The clone is identical to the original except in that the
   original set of registers is replaced by a new set of pseudo registers.
   The new set has the same modes as the original set.  */

rtx
gen_group_rtx (rtx orig)
{
  int i, length;
  rtx *tmps;

  gcc_assert (GET_CODE (orig) == PARALLEL);

  length = XVECLEN (orig, 0);
  tmps = XALLOCAVEC (rtx, length);

  /* Skip a NULL entry in first slot.  */
  i = XEXP (XVECEXP (orig, 0, 0), 0) ? 0 : 1;

  if (i)
    tmps[0] = 0;

  for (; i < length; i++)
    {
      machine_mode mode = GET_MODE (XEXP (XVECEXP (orig, 0, i), 0));
      rtx offset = XEXP (XVECEXP (orig, 0, i), 1);

      tmps[i] = gen_rtx_EXPR_LIST (VOIDmode, gen_reg_rtx (mode), offset);
    }

  return gen_rtx_PARALLEL (GET_MODE (orig), gen_rtvec_v (length, tmps));
}

/* A subroutine of emit_group_load.  Arguments as for emit_group_load,
   except that values are placed in TMPS[i], and must later be moved
   into corresponding XEXP (XVECEXP (DST, 0, i), 0) element.  */

static void
emit_group_load_1 (rtx *tmps, rtx dst, rtx orig_src, tree type,
		   poly_int64 ssize)
{
  rtx src;
  int start, i;
  machine_mode m = GET_MODE (orig_src);

  gcc_assert (GET_CODE (dst) == PARALLEL);

  if (m != VOIDmode
      && !SCALAR_INT_MODE_P (m)
      && !MEM_P (orig_src)
      && GET_CODE (orig_src) != CONCAT)
    {
      scalar_int_mode imode;
      if (int_mode_for_mode (GET_MODE (orig_src)).exists (&imode))
	{
	  src = gen_reg_rtx (imode);
	  emit_move_insn (gen_lowpart (GET_MODE (orig_src), src), orig_src);
	}
      else
	{
	  src = assign_stack_temp (GET_MODE (orig_src), ssize);
	  emit_move_insn (src, orig_src);
	}
      emit_group_load_1 (tmps, dst, src, type, ssize);
      return;
    }

  /* Check for a NULL entry, used to indicate that the parameter goes
     both on the stack and in registers.  */
  if (XEXP (XVECEXP (dst, 0, 0), 0))
    start = 0;
  else
    start = 1;

  /* Process the pieces.  */
  for (i = start; i < XVECLEN (dst, 0); i++)
    {
      machine_mode mode = GET_MODE (XEXP (XVECEXP (dst, 0, i), 0));
      poly_int64 bytepos = rtx_to_poly_int64 (XEXP (XVECEXP (dst, 0, i), 1));
      poly_int64 bytelen = GET_MODE_SIZE (mode);
      poly_int64 shift = 0;

      /* Handle trailing fragments that run over the size of the struct.
	 It's the target's responsibility to make sure that the fragment
	 cannot be strictly smaller in some cases and strictly larger
	 in others.  */
      gcc_checking_assert (ordered_p (bytepos + bytelen, ssize));
      if (known_size_p (ssize) && maybe_gt (bytepos + bytelen, ssize))
	{
	  /* Arrange to shift the fragment to where it belongs.
	     extract_bit_field loads to the lsb of the reg.  */
	  if (
#ifdef BLOCK_REG_PADDING
	      BLOCK_REG_PADDING (GET_MODE (orig_src), type, i == start)
	      == (BYTES_BIG_ENDIAN ? PAD_UPWARD : PAD_DOWNWARD)
#else
	      BYTES_BIG_ENDIAN
#endif
	      )
	    shift = (bytelen - (ssize - bytepos)) * BITS_PER_UNIT;
	  bytelen = ssize - bytepos;
	  gcc_assert (maybe_gt (bytelen, 0));
	}

      /* If we won't be loading directly from memory, protect the real source
	 from strange tricks we might play; but make sure that the source can
	 be loaded directly into the destination.  */
      src = orig_src;
      if (!MEM_P (orig_src)
	  && (!CONSTANT_P (orig_src)
	      || (GET_MODE (orig_src) != mode
		  && GET_MODE (orig_src) != VOIDmode)))
	{
	  if (GET_MODE (orig_src) == VOIDmode)
	    src = gen_reg_rtx (mode);
	  else
	    src = gen_reg_rtx (GET_MODE (orig_src));

	  emit_move_insn (src, orig_src);
	}

      /* Optimize the access just a bit.  */
      if (MEM_P (src)
	  && (! targetm.slow_unaligned_access (mode, MEM_ALIGN (src))
	      || MEM_ALIGN (src) >= GET_MODE_ALIGNMENT (mode))
	  && multiple_p (bytepos * BITS_PER_UNIT, GET_MODE_ALIGNMENT (mode))
	  && known_eq (bytelen, GET_MODE_SIZE (mode)))
	{
	  tmps[i] = gen_reg_rtx (mode);
	  emit_move_insn (tmps[i], adjust_address (src, mode, bytepos));
	}
      else if (COMPLEX_MODE_P (mode)
	       && GET_MODE (src) == mode
	       && known_eq (bytelen, GET_MODE_SIZE (mode)))
	/* Let emit_move_complex do the bulk of the work.  */
	tmps[i] = src;
      else if (GET_CODE (src) == CONCAT)
	{
	  poly_int64 slen = GET_MODE_SIZE (GET_MODE (src));
	  poly_int64 slen0 = GET_MODE_SIZE (GET_MODE (XEXP (src, 0)));
	  unsigned int elt;
	  poly_int64 subpos;

	  if (can_div_trunc_p (bytepos, slen0, &elt, &subpos)
	      && known_le (subpos + bytelen, slen0))
	    {
	      /* The following assumes that the concatenated objects all
		 have the same size.  In this case, a simple calculation
		 can be used to determine the object and the bit field
		 to be extracted.  */
	      tmps[i] = XEXP (src, elt);
	      if (maybe_ne (subpos, 0)
		  || maybe_ne (subpos + bytelen, slen0)
		  || (!CONSTANT_P (tmps[i])
		      && (!REG_P (tmps[i]) || GET_MODE (tmps[i]) != mode)))
		tmps[i] = extract_bit_field (tmps[i], bytelen * BITS_PER_UNIT,
					     subpos * BITS_PER_UNIT,
					     1, NULL_RTX, mode, mode, false,
					     NULL);
	    }
	  else
	    {
	      rtx mem;

	      gcc_assert (known_eq (bytepos, 0));
	      mem = assign_stack_temp (GET_MODE (src), slen);
	      emit_move_insn (mem, src);
	      tmps[i] = extract_bit_field (mem, bytelen * BITS_PER_UNIT,
					   0, 1, NULL_RTX, mode, mode, false,
					   NULL);
	    }
	}
      /* FIXME: A SIMD parallel will eventually lead to a subreg of a
	 SIMD register, which is currently broken.  While we get GCC
	 to emit proper RTL for these cases, let's dump to memory.  */
      else if (VECTOR_MODE_P (GET_MODE (dst))
	       && REG_P (src))
	{
	  poly_uint64 slen = GET_MODE_SIZE (GET_MODE (src));
	  rtx mem;

	  mem = assign_stack_temp (GET_MODE (src), slen);
	  emit_move_insn (mem, src);
	  tmps[i] = adjust_address (mem, mode, bytepos);
	}
      else if (CONSTANT_P (src) && GET_MODE (dst) != BLKmode
               && XVECLEN (dst, 0) > 1)
        tmps[i] = simplify_gen_subreg (mode, src, GET_MODE (dst), bytepos);
      else if (CONSTANT_P (src))
	{
	  if (known_eq (bytelen, ssize))
	    tmps[i] = src;
	  else
	    {
	      rtx first, second;

	      /* TODO: const_wide_int can have sizes other than this...  */
	      gcc_assert (known_eq (2 * bytelen, ssize));
	      split_double (src, &first, &second);
	      if (i)
		tmps[i] = second;
	      else
		tmps[i] = first;
	    }
	}
      else if (REG_P (src) && GET_MODE (src) == mode)
	tmps[i] = src;
      else
	tmps[i] = extract_bit_field (src, bytelen * BITS_PER_UNIT,
				     bytepos * BITS_PER_UNIT, 1, NULL_RTX,
				     mode, mode, false, NULL);

      if (maybe_ne (shift, 0))
	tmps[i] = expand_shift (LSHIFT_EXPR, mode, tmps[i],
				shift, tmps[i], 0);
    }
}

/* Emit code to move a block SRC of type TYPE to a block DST,
   where DST is non-consecutive registers represented by a PARALLEL.
   SSIZE represents the total size of block ORIG_SRC in bytes, or -1
   if not known.  */

void
emit_group_load (rtx dst, rtx src, tree type, poly_int64 ssize)
{
  rtx *tmps;
  int i;

  tmps = XALLOCAVEC (rtx, XVECLEN (dst, 0));
  emit_group_load_1 (tmps, dst, src, type, ssize);

  /* Copy the extracted pieces into the proper (probable) hard regs.  */
  for (i = 0; i < XVECLEN (dst, 0); i++)
    {
      rtx d = XEXP (XVECEXP (dst, 0, i), 0);
      if (d == NULL)
	continue;
      emit_move_insn (d, tmps[i]);
    }
}

/* Similar, but load SRC into new pseudos in a format that looks like
   PARALLEL.  This can later be fed to emit_group_move to get things
   in the right place.  */

rtx
emit_group_load_into_temps (rtx parallel, rtx src, tree type, poly_int64 ssize)
{
  rtvec vec;
  int i;

  vec = rtvec_alloc (XVECLEN (parallel, 0));
  emit_group_load_1 (&RTVEC_ELT (vec, 0), parallel, src, type, ssize);

  /* Convert the vector to look just like the original PARALLEL, except
     with the computed values.  */
  for (i = 0; i < XVECLEN (parallel, 0); i++)
    {
      rtx e = XVECEXP (parallel, 0, i);
      rtx d = XEXP (e, 0);

      if (d)
	{
	  d = force_reg (GET_MODE (d), RTVEC_ELT (vec, i));
	  e = alloc_EXPR_LIST (REG_NOTE_KIND (e), d, XEXP (e, 1));
	}
      RTVEC_ELT (vec, i) = e;
    }

  return gen_rtx_PARALLEL (GET_MODE (parallel), vec);
}

/* Emit code to move a block SRC to block DST, where SRC and DST are
   non-consecutive groups of registers, each represented by a PARALLEL.  */

void
emit_group_move (rtx dst, rtx src)
{
  int i;

  gcc_assert (GET_CODE (src) == PARALLEL
	      && GET_CODE (dst) == PARALLEL
	      && XVECLEN (src, 0) == XVECLEN (dst, 0));

  /* Skip first entry if NULL.  */
  for (i = XEXP (XVECEXP (src, 0, 0), 0) ? 0 : 1; i < XVECLEN (src, 0); i++)
    emit_move_insn (XEXP (XVECEXP (dst, 0, i), 0),
		    XEXP (XVECEXP (src, 0, i), 0));
}

/* Move a group of registers represented by a PARALLEL into pseudos.  */

rtx
emit_group_move_into_temps (rtx src)
{
  rtvec vec = rtvec_alloc (XVECLEN (src, 0));
  int i;

  for (i = 0; i < XVECLEN (src, 0); i++)
    {
      rtx e = XVECEXP (src, 0, i);
      rtx d = XEXP (e, 0);

      if (d)
	e = alloc_EXPR_LIST (REG_NOTE_KIND (e), copy_to_reg (d), XEXP (e, 1));
      RTVEC_ELT (vec, i) = e;
    }

  return gen_rtx_PARALLEL (GET_MODE (src), vec);
}

/* Emit code to move a block SRC to a block ORIG_DST of type TYPE,
   where SRC is non-consecutive registers represented by a PARALLEL.
   SSIZE represents the total size of block ORIG_DST, or -1 if not
   known.  */

void
emit_group_store (rtx orig_dst, rtx src, tree type ATTRIBUTE_UNUSED,
		  poly_int64 ssize)
{
  rtx *tmps, dst;
  int start, finish, i;
  machine_mode m = GET_MODE (orig_dst);

  gcc_assert (GET_CODE (src) == PARALLEL);

  if (!SCALAR_INT_MODE_P (m)
      && !MEM_P (orig_dst) && GET_CODE (orig_dst) != CONCAT)
    {
      scalar_int_mode imode;
      if (int_mode_for_mode (GET_MODE (orig_dst)).exists (&imode))
	{
	  dst = gen_reg_rtx (imode);
	  emit_group_store (dst, src, type, ssize);
	  dst = gen_lowpart (GET_MODE (orig_dst), dst);
	}
      else
	{
	  dst = assign_stack_temp (GET_MODE (orig_dst), ssize);
	  emit_group_store (dst, src, type, ssize);
	}
      emit_move_insn (orig_dst, dst);
      return;
    }

  /* Check for a NULL entry, used to indicate that the parameter goes
     both on the stack and in registers.  */
  if (XEXP (XVECEXP (src, 0, 0), 0))
    start = 0;
  else
    start = 1;
  finish = XVECLEN (src, 0);

  tmps = XALLOCAVEC (rtx, finish);

  /* Copy the (probable) hard regs into pseudos.  */
  for (i = start; i < finish; i++)
    {
      rtx reg = XEXP (XVECEXP (src, 0, i), 0);
      if (!REG_P (reg) || REGNO (reg) < FIRST_PSEUDO_REGISTER)
	{
	  tmps[i] = gen_reg_rtx (GET_MODE (reg));
	  emit_move_insn (tmps[i], reg);
	}
      else
	tmps[i] = reg;
    }

  /* If we won't be storing directly into memory, protect the real destination
     from strange tricks we might play.  */
  dst = orig_dst;
  if (GET_CODE (dst) == PARALLEL)
    {
      rtx temp;

      /* We can get a PARALLEL dst if there is a conditional expression in
	 a return statement.  In that case, the dst and src are the same,
	 so no action is necessary.  */
      if (rtx_equal_p (dst, src))
	return;

      /* It is unclear if we can ever reach here, but we may as well handle
	 it.  Allocate a temporary, and split this into a store/load to/from
	 the temporary.  */
      temp = assign_stack_temp (GET_MODE (dst), ssize);
      emit_group_store (temp, src, type, ssize);
      emit_group_load (dst, temp, type, ssize);
      return;
    }
  else if (!MEM_P (dst) && GET_CODE (dst) != CONCAT)
    {
      machine_mode outer = GET_MODE (dst);
      machine_mode inner;
      poly_int64 bytepos;
      bool done = false;
      rtx temp;

      if (!REG_P (dst) || REGNO (dst) < FIRST_PSEUDO_REGISTER)
	dst = gen_reg_rtx (outer);

      /* Make life a bit easier for combine.  */
      /* If the first element of the vector is the low part
	 of the destination mode, use a paradoxical subreg to
	 initialize the destination.  */
      if (start < finish)
	{
	  inner = GET_MODE (tmps[start]);
	  bytepos = subreg_lowpart_offset (inner, outer);
	  if (known_eq (rtx_to_poly_int64 (XEXP (XVECEXP (src, 0, start), 1)),
			bytepos))
	    {
	      temp = simplify_gen_subreg (outer, tmps[start],
					  inner, 0);
	      if (temp)
		{
		  emit_move_insn (dst, temp);
		  done = true;
		  start++;
		}
	    }
	}

      /* If the first element wasn't the low part, try the last.  */
      if (!done
	  && start < finish - 1)
	{
	  inner = GET_MODE (tmps[finish - 1]);
	  bytepos = subreg_lowpart_offset (inner, outer);
	  if (known_eq (rtx_to_poly_int64 (XEXP (XVECEXP (src, 0,
							  finish - 1), 1)),
			bytepos))
	    {
	      temp = simplify_gen_subreg (outer, tmps[finish - 1],
					  inner, 0);
	      if (temp)
		{
		  emit_move_insn (dst, temp);
		  done = true;
		  finish--;
		}
	    }
	}

      /* Otherwise, simply initialize the result to zero.  */
      if (!done)
        emit_move_insn (dst, CONST0_RTX (outer));
    }

  /* Process the pieces.  */
  for (i = start; i < finish; i++)
    {
      poly_int64 bytepos = rtx_to_poly_int64 (XEXP (XVECEXP (src, 0, i), 1));
      machine_mode mode = GET_MODE (tmps[i]);
      poly_int64 bytelen = GET_MODE_SIZE (mode);
      poly_uint64 adj_bytelen;
      rtx dest = dst;

      /* Handle trailing fragments that run over the size of the struct.
	 It's the target's responsibility to make sure that the fragment
	 cannot be strictly smaller in some cases and strictly larger
	 in others.  */
      gcc_checking_assert (ordered_p (bytepos + bytelen, ssize));
      if (known_size_p (ssize) && maybe_gt (bytepos + bytelen, ssize))
	adj_bytelen = ssize - bytepos;
      else
	adj_bytelen = bytelen;

      if (GET_CODE (dst) == CONCAT)
	{
	  if (known_le (bytepos + adj_bytelen,
			GET_MODE_SIZE (GET_MODE (XEXP (dst, 0)))))
	    dest = XEXP (dst, 0);
	  else if (known_ge (bytepos, GET_MODE_SIZE (GET_MODE (XEXP (dst, 0)))))
	    {
	      bytepos -= GET_MODE_SIZE (GET_MODE (XEXP (dst, 0)));
	      dest = XEXP (dst, 1);
	    }
	  else
	    {
	      machine_mode dest_mode = GET_MODE (dest);
	      machine_mode tmp_mode = GET_MODE (tmps[i]);

	      gcc_assert (known_eq (bytepos, 0) && XVECLEN (src, 0));

	      if (GET_MODE_ALIGNMENT (dest_mode)
		  >= GET_MODE_ALIGNMENT (tmp_mode))
		{
		  dest = assign_stack_temp (dest_mode,
					    GET_MODE_SIZE (dest_mode));
		  emit_move_insn (adjust_address (dest,
						  tmp_mode,
						  bytepos),
				  tmps[i]);
		  dst = dest;
		}
	      else
		{
		  dest = assign_stack_temp (tmp_mode,
					    GET_MODE_SIZE (tmp_mode));
		  emit_move_insn (dest, tmps[i]);
		  dst = adjust_address (dest, dest_mode, bytepos);
		}
	      break;
	    }
	}

      /* Handle trailing fragments that run over the size of the struct.  */
      if (known_size_p (ssize) && maybe_gt (bytepos + bytelen, ssize))
	{
	  /* store_bit_field always takes its value from the lsb.
	     Move the fragment to the lsb if it's not already there.  */
	  if (
#ifdef BLOCK_REG_PADDING
	      BLOCK_REG_PADDING (GET_MODE (orig_dst), type, i == start)
	      == (BYTES_BIG_ENDIAN ? PAD_UPWARD : PAD_DOWNWARD)
#else
	      BYTES_BIG_ENDIAN
#endif
	      )
	    {
	      poly_int64 shift = (bytelen - (ssize - bytepos)) * BITS_PER_UNIT;
	      tmps[i] = expand_shift (RSHIFT_EXPR, mode, tmps[i],
				      shift, tmps[i], 0);
	    }

	  /* Make sure not to write past the end of the struct.  */
	  store_bit_field (dest,
			   adj_bytelen * BITS_PER_UNIT, bytepos * BITS_PER_UNIT,
			   bytepos * BITS_PER_UNIT, ssize * BITS_PER_UNIT - 1,
			   VOIDmode, tmps[i], false);
	}

      /* Optimize the access just a bit.  */
      else if (MEM_P (dest)
	       && (!targetm.slow_unaligned_access (mode, MEM_ALIGN (dest))
		   || MEM_ALIGN (dest) >= GET_MODE_ALIGNMENT (mode))
	       && multiple_p (bytepos * BITS_PER_UNIT,
			      GET_MODE_ALIGNMENT (mode))
	       && known_eq (bytelen, GET_MODE_SIZE (mode)))
	emit_move_insn (adjust_address (dest, mode, bytepos), tmps[i]);

      else
	store_bit_field (dest, bytelen * BITS_PER_UNIT, bytepos * BITS_PER_UNIT,
			 0, 0, mode, tmps[i], false);
    }

  /* Copy from the pseudo into the (probable) hard reg.  */
  if (orig_dst != dst)
    emit_move_insn (orig_dst, dst);
}

/* Return a form of X that does not use a PARALLEL.  TYPE is the type
   of the value stored in X.  */

rtx
maybe_emit_group_store (rtx x, tree type)
{
  machine_mode mode = TYPE_MODE (type);
  gcc_checking_assert (GET_MODE (x) == VOIDmode || GET_MODE (x) == mode);
  if (GET_CODE (x) == PARALLEL)
    {
      rtx result = gen_reg_rtx (mode);
      emit_group_store (result, x, type, int_size_in_bytes (type));
      return result;
    }
  return x;
}

/* Copy a BLKmode object of TYPE out of a register SRCREG into TARGET.

   This is used on targets that return BLKmode values in registers.  */

static void
copy_blkmode_from_reg (rtx target, rtx srcreg, tree type)
{
  unsigned HOST_WIDE_INT bytes = int_size_in_bytes (type);
  rtx src = NULL, dst = NULL;
  unsigned HOST_WIDE_INT bitsize = MIN (TYPE_ALIGN (type), BITS_PER_WORD);
  unsigned HOST_WIDE_INT bitpos, xbitpos, padding_correction = 0;
  /* No current ABI uses variable-sized modes to pass a BLKmnode type.  */
  fixed_size_mode mode = as_a <fixed_size_mode> (GET_MODE (srcreg));
  fixed_size_mode tmode = as_a <fixed_size_mode> (GET_MODE (target));
  fixed_size_mode copy_mode;

  /* BLKmode registers created in the back-end shouldn't have survived.  */
  gcc_assert (mode != BLKmode);

  /* If the structure doesn't take up a whole number of words, see whether
     SRCREG is padded on the left or on the right.  If it's on the left,
     set PADDING_CORRECTION to the number of bits to skip.

     In most ABIs, the structure will be returned at the least end of
     the register, which translates to right padding on little-endian
     targets and left padding on big-endian targets.  The opposite
     holds if the structure is returned at the most significant
     end of the register.  */
  if (bytes % UNITS_PER_WORD != 0
      && (targetm.calls.return_in_msb (type)
	  ? !BYTES_BIG_ENDIAN
	  : BYTES_BIG_ENDIAN))
    padding_correction
      = (BITS_PER_WORD - ((bytes % UNITS_PER_WORD) * BITS_PER_UNIT));

  /* We can use a single move if we have an exact mode for the size.  */
  else if (MEM_P (target)
	   && (!targetm.slow_unaligned_access (mode, MEM_ALIGN (target))
	       || MEM_ALIGN (target) >= GET_MODE_ALIGNMENT (mode))
	   && bytes == GET_MODE_SIZE (mode))
  {
    emit_move_insn (adjust_address (target, mode, 0), srcreg);
    return;
  }

  /* And if we additionally have the same mode for a register.  */
  else if (REG_P (target)
	   && GET_MODE (target) == mode
	   && bytes == GET_MODE_SIZE (mode))
  {
    emit_move_insn (target, srcreg);
    return;
  }

  /* This code assumes srcreg is at least a full word.  If it isn't, copy it
     into a new pseudo which is a full word.  */
  if (GET_MODE_SIZE (mode) < UNITS_PER_WORD)
    {
      srcreg = convert_to_mode (word_mode, srcreg, TYPE_UNSIGNED (type));
      mode = word_mode;
    }

  /* Copy the structure BITSIZE bits at a time.  If the target lives in
     memory, take care of not reading/writing past its end by selecting
     a copy mode suited to BITSIZE.  This should always be possible given
     how it is computed.

     If the target lives in register, make sure not to select a copy mode
     larger than the mode of the register.

     We could probably emit more efficient code for machines which do not use
     strict alignment, but it doesn't seem worth the effort at the current
     time.  */

  copy_mode = word_mode;
  if (MEM_P (target))
    {
      opt_scalar_int_mode mem_mode = int_mode_for_size (bitsize, 1);
      if (mem_mode.exists ())
	copy_mode = mem_mode.require ();
    }
  else if (REG_P (target) && GET_MODE_BITSIZE (tmode) < BITS_PER_WORD)
    copy_mode = tmode;

  for (bitpos = 0, xbitpos = padding_correction;
       bitpos < bytes * BITS_PER_UNIT;
       bitpos += bitsize, xbitpos += bitsize)
    {
      /* We need a new source operand each time xbitpos is on a
	 word boundary and when xbitpos == padding_correction
	 (the first time through).  */
      if (xbitpos % BITS_PER_WORD == 0 || xbitpos == padding_correction)
	src = operand_subword_force (srcreg, xbitpos / BITS_PER_WORD, mode);

      /* We need a new destination operand each time bitpos is on
	 a word boundary.  */
      if (REG_P (target) && GET_MODE_BITSIZE (tmode) < BITS_PER_WORD)
	dst = target;
      else if (bitpos % BITS_PER_WORD == 0)
	dst = operand_subword (target, bitpos / BITS_PER_WORD, 1, tmode);

      /* Use xbitpos for the source extraction (right justified) and
	 bitpos for the destination store (left justified).  */
      store_bit_field (dst, bitsize, bitpos % BITS_PER_WORD, 0, 0, copy_mode,
		       extract_bit_field (src, bitsize,
					  xbitpos % BITS_PER_WORD, 1,
					  NULL_RTX, copy_mode, copy_mode,
					  false, NULL),
		       false);
    }
}

/* Copy BLKmode value SRC into a register of mode MODE_IN.  Return the
   register if it contains any data, otherwise return null.

   This is used on targets that return BLKmode values in registers.  */

rtx
copy_blkmode_to_reg (machine_mode mode_in, tree src)
{
  int i, n_regs;
  unsigned HOST_WIDE_INT bitpos, xbitpos, padding_correction = 0, bytes;
  unsigned int bitsize;
  rtx *dst_words, dst, x, src_word = NULL_RTX, dst_word = NULL_RTX;
  /* No current ABI uses variable-sized modes to pass a BLKmnode type.  */
  fixed_size_mode mode = as_a <fixed_size_mode> (mode_in);
  fixed_size_mode dst_mode;
  scalar_int_mode min_mode;

  gcc_assert (TYPE_MODE (TREE_TYPE (src)) == BLKmode);

  x = expand_normal (src);

  bytes = arg_int_size_in_bytes (TREE_TYPE (src));
  if (bytes == 0)
    return NULL_RTX;

  /* If the structure doesn't take up a whole number of words, see
     whether the register value should be padded on the left or on
     the right.  Set PADDING_CORRECTION to the number of padding
     bits needed on the left side.

     In most ABIs, the structure will be returned at the least end of
     the register, which translates to right padding on little-endian
     targets and left padding on big-endian targets.  The opposite
     holds if the structure is returned at the most significant
     end of the register.  */
  if (bytes % UNITS_PER_WORD != 0
      && (targetm.calls.return_in_msb (TREE_TYPE (src))
	  ? !BYTES_BIG_ENDIAN
	  : BYTES_BIG_ENDIAN))
    padding_correction = (BITS_PER_WORD - ((bytes % UNITS_PER_WORD)
					   * BITS_PER_UNIT));

  n_regs = (bytes + UNITS_PER_WORD - 1) / UNITS_PER_WORD;
  dst_words = XALLOCAVEC (rtx, n_regs);
  bitsize = MIN (TYPE_ALIGN (TREE_TYPE (src)), BITS_PER_WORD);
  min_mode = smallest_int_mode_for_size (bitsize);

  /* Copy the structure BITSIZE bits at a time.  */
  for (bitpos = 0, xbitpos = padding_correction;
       bitpos < bytes * BITS_PER_UNIT;
       bitpos += bitsize, xbitpos += bitsize)
    {
      /* We need a new destination pseudo each time xbitpos is
	 on a word boundary and when xbitpos == padding_correction
	 (the first time through).  */
      if (xbitpos % BITS_PER_WORD == 0
	  || xbitpos == padding_correction)
	{
	  /* Generate an appropriate register.  */
	  dst_word = gen_reg_rtx (word_mode);
	  dst_words[xbitpos / BITS_PER_WORD] = dst_word;

	  /* Clear the destination before we move anything into it.  */
	  emit_move_insn (dst_word, CONST0_RTX (word_mode));
	}

      /* Find the largest integer mode that can be used to copy all or as
	 many bits as possible of the structure if the target supports larger
	 copies.  There are too many corner cases here w.r.t to alignments on
	 the read/writes.  So if there is any padding just use single byte
	 operations.  */
      opt_scalar_int_mode mode_iter;
      if (padding_correction == 0 && !STRICT_ALIGNMENT)
	{
	  FOR_EACH_MODE_FROM (mode_iter, min_mode)
	    {
	      unsigned int msize = GET_MODE_BITSIZE (mode_iter.require ());
	      if (msize <= ((bytes * BITS_PER_UNIT) - bitpos)
		  && msize <= BITS_PER_WORD)
		bitsize = msize;
	      else
		break;
	    }
	}

      /* We need a new source operand each time bitpos is on a word
	 boundary.  */
      if (bitpos % BITS_PER_WORD == 0)
	src_word = operand_subword_force (x, bitpos / BITS_PER_WORD, BLKmode);

      /* Use bitpos for the source extraction (left justified) and
	 xbitpos for the destination store (right justified).  */
      store_bit_field (dst_word, bitsize, xbitpos % BITS_PER_WORD,
		       0, 0, word_mode,
		       extract_bit_field (src_word, bitsize,
					  bitpos % BITS_PER_WORD, 1,
					  NULL_RTX, word_mode, word_mode,
					  false, NULL),
		       false);
    }

  if (mode == BLKmode)
    {
      /* Find the smallest integer mode large enough to hold the
	 entire structure.  */
      opt_scalar_int_mode mode_iter;
      FOR_EACH_MODE_IN_CLASS (mode_iter, MODE_INT)
	if (GET_MODE_SIZE (mode_iter.require ()) >= bytes)
	  break;

      /* A suitable mode should have been found.  */
      mode = mode_iter.require ();
    }

  if (GET_MODE_SIZE (mode) < GET_MODE_SIZE (word_mode))
    dst_mode = word_mode;
  else
    dst_mode = mode;
  dst = gen_reg_rtx (dst_mode);

  for (i = 0; i < n_regs; i++)
    emit_move_insn (operand_subword (dst, i, 0, dst_mode), dst_words[i]);

  if (mode != dst_mode)
    dst = gen_lowpart (mode, dst);

  return dst;
}

/* Add a USE expression for REG to the (possibly empty) list pointed
   to by CALL_FUSAGE.  REG must denote a hard register.  */

void
use_reg_mode (rtx *call_fusage, rtx reg, machine_mode mode)
{
  gcc_assert (REG_P (reg));

  if (!HARD_REGISTER_P (reg))
    return;

  *call_fusage
    = gen_rtx_EXPR_LIST (mode, gen_rtx_USE (VOIDmode, reg), *call_fusage);
}

/* Add a CLOBBER expression for REG to the (possibly empty) list pointed
   to by CALL_FUSAGE.  REG must denote a hard register.  */

void
clobber_reg_mode (rtx *call_fusage, rtx reg, machine_mode mode)
{
  gcc_assert (REG_P (reg) && REGNO (reg) < FIRST_PSEUDO_REGISTER);

  *call_fusage
    = gen_rtx_EXPR_LIST (mode, gen_rtx_CLOBBER (VOIDmode, reg), *call_fusage);
}

/* Add USE expressions to *CALL_FUSAGE for each of NREGS consecutive regs,
   starting at REGNO.  All of these registers must be hard registers.  */

void
use_regs (rtx *call_fusage, int regno, int nregs)
{
  int i;

  gcc_assert (regno + nregs <= FIRST_PSEUDO_REGISTER);

  for (i = 0; i < nregs; i++)
    use_reg (call_fusage, regno_reg_rtx[regno + i]);
}

/* Add USE expressions to *CALL_FUSAGE for each REG contained in the
   PARALLEL REGS.  This is for calls that pass values in multiple
   non-contiguous locations.  The Irix 6 ABI has examples of this.  */

void
use_group_regs (rtx *call_fusage, rtx regs)
{
  int i;

  for (i = 0; i < XVECLEN (regs, 0); i++)
    {
      rtx reg = XEXP (XVECEXP (regs, 0, i), 0);

      /* A NULL entry means the parameter goes both on the stack and in
	 registers.  This can also be a MEM for targets that pass values
	 partially on the stack and partially in registers.  */
      if (reg != 0 && REG_P (reg))
	use_reg (call_fusage, reg);
    }
}

/* Return the defining gimple statement for SSA_NAME NAME if it is an
   assigment and the code of the expresion on the RHS is CODE.  Return
   NULL otherwise.  */

static gimple *
get_def_for_expr (tree name, enum tree_code code)
{
  gimple *def_stmt;

  if (TREE_CODE (name) != SSA_NAME)
    return NULL;

  def_stmt = get_gimple_for_ssa_name (name);
  if (!def_stmt
      || gimple_assign_rhs_code (def_stmt) != code)
    return NULL;

  return def_stmt;
}

/* Return the defining gimple statement for SSA_NAME NAME if it is an
   assigment and the class of the expresion on the RHS is CLASS.  Return
   NULL otherwise.  */

static gimple *
get_def_for_expr_class (tree name, enum tree_code_class tclass)
{
  gimple *def_stmt;

  if (TREE_CODE (name) != SSA_NAME)
    return NULL;

  def_stmt = get_gimple_for_ssa_name (name);
  if (!def_stmt
      || TREE_CODE_CLASS (gimple_assign_rhs_code (def_stmt)) != tclass)
    return NULL;

  return def_stmt;
}

/* Write zeros through the storage of OBJECT.  If OBJECT has BLKmode, SIZE is
   its length in bytes.  */

rtx
clear_storage_hints (rtx object, rtx size, enum block_op_methods method,
		     unsigned int expected_align, HOST_WIDE_INT expected_size,
		     unsigned HOST_WIDE_INT min_size,
		     unsigned HOST_WIDE_INT max_size,
		     unsigned HOST_WIDE_INT probable_max_size)
{
  machine_mode mode = GET_MODE (object);
  unsigned int align;

  gcc_assert (method == BLOCK_OP_NORMAL || method == BLOCK_OP_TAILCALL);

  /* If OBJECT is not BLKmode and SIZE is the same size as its mode,
     just move a zero.  Otherwise, do this a piece at a time.  */
  poly_int64 size_val;
  if (mode != BLKmode
      && poly_int_rtx_p (size, &size_val)
      && known_eq (size_val, GET_MODE_SIZE (mode)))
    {
      rtx zero = CONST0_RTX (mode);
      if (zero != NULL)
	{
	  emit_move_insn (object, zero);
	  return NULL;
	}

      if (COMPLEX_MODE_P (mode))
	{
	  zero = CONST0_RTX (GET_MODE_INNER (mode));
	  if (zero != NULL)
	    {
	      write_complex_part (object, zero, 0);
	      write_complex_part (object, zero, 1);
	      return NULL;
	    }
	}
    }

  if (size == const0_rtx)
    return NULL;

  align = MEM_ALIGN (object);

  if (CONST_INT_P (size)
      && targetm.use_by_pieces_infrastructure_p (INTVAL (size), align,
						 CLEAR_BY_PIECES,
						 optimize_insn_for_speed_p ()))
    clear_by_pieces (object, INTVAL (size), align);
  else if (set_storage_via_setmem (object, size, const0_rtx, align,
				   expected_align, expected_size,
				   min_size, max_size, probable_max_size))
    ;
  else if (ADDR_SPACE_GENERIC_P (MEM_ADDR_SPACE (object)))
    return set_storage_via_libcall (object, size, const0_rtx,
				    method == BLOCK_OP_TAILCALL);
  else
    gcc_unreachable ();

  return NULL;
}

rtx
clear_storage (rtx object, rtx size, enum block_op_methods method)
{
  unsigned HOST_WIDE_INT max, min = 0;
  if (GET_CODE (size) == CONST_INT)
    min = max = UINTVAL (size);
  else
    max = GET_MODE_MASK (GET_MODE (size));
  return clear_storage_hints (object, size, method, 0, -1, min, max, max);
}


/* A subroutine of clear_storage.  Expand a call to memset.
   Return the return value of memset, 0 otherwise.  */

rtx
set_storage_via_libcall (rtx object, rtx size, rtx val, bool tailcall)
{
  tree call_expr, fn, object_tree, size_tree, val_tree;
  machine_mode size_mode;

  object = copy_addr_to_reg (XEXP (object, 0));
  object_tree = make_tree (ptr_type_node, object);

  if (!CONST_INT_P (val))
    val = convert_to_mode (TYPE_MODE (integer_type_node), val, 1);
  val_tree = make_tree (integer_type_node, val);

  size_mode = TYPE_MODE (sizetype);
  size = convert_to_mode (size_mode, size, 1);
  size = copy_to_mode_reg (size_mode, size);
  size_tree = make_tree (sizetype, size);

  /* It is incorrect to use the libcall calling conventions for calls to
     memset because it can be provided by the user.  */
  fn = builtin_decl_implicit (BUILT_IN_MEMSET);
  call_expr = build_call_expr (fn, 3, object_tree, val_tree, size_tree);
  CALL_EXPR_TAILCALL (call_expr) = tailcall;

  return expand_call (call_expr, NULL_RTX, false);
}

/* Expand a setmem pattern; return true if successful.  */

bool
set_storage_via_setmem (rtx object, rtx size, rtx val, unsigned int align,
			unsigned int expected_align, HOST_WIDE_INT expected_size,
			unsigned HOST_WIDE_INT min_size,
			unsigned HOST_WIDE_INT max_size,
			unsigned HOST_WIDE_INT probable_max_size)
{
  /* Try the most limited insn first, because there's no point
     including more than one in the machine description unless
     the more limited one has some advantage.  */

  if (expected_align < align)
    expected_align = align;
  if (expected_size != -1)
    {
      if ((unsigned HOST_WIDE_INT)expected_size > max_size)
	expected_size = max_size;
      if ((unsigned HOST_WIDE_INT)expected_size < min_size)
	expected_size = min_size;
    }

  opt_scalar_int_mode mode_iter;
  FOR_EACH_MODE_IN_CLASS (mode_iter, MODE_INT)
    {
      scalar_int_mode mode = mode_iter.require ();
      enum insn_code code = direct_optab_handler (setmem_optab, mode);

      if (code != CODE_FOR_nothing
	  /* We don't need MODE to be narrower than BITS_PER_HOST_WIDE_INT
	     here because if SIZE is less than the mode mask, as it is
	     returned by the macro, it will definitely be less than the
	     actual mode mask.  Since SIZE is within the Pmode address
	     space, we limit MODE to Pmode.  */
	  && ((CONST_INT_P (size)
	       && ((unsigned HOST_WIDE_INT) INTVAL (size)
		   <= (GET_MODE_MASK (mode) >> 1)))
	      || max_size <= (GET_MODE_MASK (mode) >> 1)
	      || GET_MODE_BITSIZE (mode) >= GET_MODE_BITSIZE (Pmode)))
	{
	  class expand_operand ops[9];
	  unsigned int nops;

	  nops = insn_data[(int) code].n_generator_args;
	  gcc_assert (nops == 4 || nops == 6 || nops == 8 || nops == 9);

	  create_fixed_operand (&ops[0], object);
	  /* The check above guarantees that this size conversion is valid.  */
	  create_convert_operand_to (&ops[1], size, mode, true);
	  create_convert_operand_from (&ops[2], val, byte_mode, true);
	  create_integer_operand (&ops[3], align / BITS_PER_UNIT);
	  if (nops >= 6)
	    {
	      create_integer_operand (&ops[4], expected_align / BITS_PER_UNIT);
	      create_integer_operand (&ops[5], expected_size);
	    }
	  if (nops >= 8)
	    {
	      create_integer_operand (&ops[6], min_size);
	      /* If we cannot represent the maximal size,
		 make parameter NULL.  */
	      if ((HOST_WIDE_INT) max_size != -1)
	        create_integer_operand (&ops[7], max_size);
	      else
		create_fixed_operand (&ops[7], NULL);
	    }
	  if (nops == 9)
	    {
	      /* If we cannot represent the maximal size,
		 make parameter NULL.  */
	      if ((HOST_WIDE_INT) probable_max_size != -1)
	        create_integer_operand (&ops[8], probable_max_size);
	      else
		create_fixed_operand (&ops[8], NULL);
	    }
	  if (maybe_expand_insn (code, nops, ops))
	    return true;
	}
    }

  return false;
}


/* Write to one of the components of the complex value CPLX.  Write VAL to
   the real part if IMAG_P is false, and the imaginary part if its true.  */

void
write_complex_part (rtx cplx, rtx val, bool imag_p)
{
  machine_mode cmode;
  scalar_mode imode;
  unsigned ibitsize;

  if (GET_CODE (cplx) == CONCAT)
    {
      emit_move_insn (XEXP (cplx, imag_p), val);
      return;
    }

  cmode = GET_MODE (cplx);
  imode = GET_MODE_INNER (cmode);
  ibitsize = GET_MODE_BITSIZE (imode);

  /* For MEMs simplify_gen_subreg may generate an invalid new address
     because, e.g., the original address is considered mode-dependent
     by the target, which restricts simplify_subreg from invoking
     adjust_address_nv.  Instead of preparing fallback support for an
     invalid address, we call adjust_address_nv directly.  */
  if (MEM_P (cplx))
    {
      emit_move_insn (adjust_address_nv (cplx, imode,
					 imag_p ? GET_MODE_SIZE (imode) : 0),
		      val);
      return;
    }

  /* If the sub-object is at least word sized, then we know that subregging
     will work.  This special case is important, since store_bit_field
     wants to operate on integer modes, and there's rarely an OImode to
     correspond to TCmode.  */
  if (ibitsize >= BITS_PER_WORD
      /* For hard regs we have exact predicates.  Assume we can split
	 the original object if it spans an even number of hard regs.
	 This special case is important for SCmode on 64-bit platforms
	 where the natural size of floating-point regs is 32-bit.  */
      || (REG_P (cplx)
	  && REGNO (cplx) < FIRST_PSEUDO_REGISTER
	  && REG_NREGS (cplx) % 2 == 0))
    {
      rtx part = simplify_gen_subreg (imode, cplx, cmode,
				      imag_p ? GET_MODE_SIZE (imode) : 0);
      if (part)
        {
	  emit_move_insn (part, val);
	  return;
	}
      else
	/* simplify_gen_subreg may fail for sub-word MEMs.  */
	gcc_assert (MEM_P (cplx) && ibitsize < BITS_PER_WORD);
    }

  store_bit_field (cplx, ibitsize, imag_p ? ibitsize : 0, 0, 0, imode, val,
		   false);
}

/* Extract one of the components of the complex value CPLX.  Extract the
   real part if IMAG_P is false, and the imaginary part if it's true.  */

rtx
read_complex_part (rtx cplx, bool imag_p)
{
  machine_mode cmode;
  scalar_mode imode;
  unsigned ibitsize;

  if (GET_CODE (cplx) == CONCAT)
    return XEXP (cplx, imag_p);

  cmode = GET_MODE (cplx);
  imode = GET_MODE_INNER (cmode);
  ibitsize = GET_MODE_BITSIZE (imode);

  /* Special case reads from complex constants that got spilled to memory.  */
  if (MEM_P (cplx) && GET_CODE (XEXP (cplx, 0)) == SYMBOL_REF)
    {
      tree decl = SYMBOL_REF_DECL (XEXP (cplx, 0));
      if (decl && TREE_CODE (decl) == COMPLEX_CST)
	{
	  tree part = imag_p ? TREE_IMAGPART (decl) : TREE_REALPART (decl);
	  if (CONSTANT_CLASS_P (part))
	    return expand_expr (part, NULL_RTX, imode, EXPAND_NORMAL);
	}
    }

  /* For MEMs simplify_gen_subreg may generate an invalid new address
     because, e.g., the original address is considered mode-dependent
     by the target, which restricts simplify_subreg from invoking
     adjust_address_nv.  Instead of preparing fallback support for an
     invalid address, we call adjust_address_nv directly.  */
  if (MEM_P (cplx))
    return adjust_address_nv (cplx, imode,
			      imag_p ? GET_MODE_SIZE (imode) : 0);

  /* If the sub-object is at least word sized, then we know that subregging
     will work.  This special case is important, since extract_bit_field
     wants to operate on integer modes, and there's rarely an OImode to
     correspond to TCmode.  */
  if (ibitsize >= BITS_PER_WORD
      /* For hard regs we have exact predicates.  Assume we can split
	 the original object if it spans an even number of hard regs.
	 This special case is important for SCmode on 64-bit platforms
	 where the natural size of floating-point regs is 32-bit.  */
      || (REG_P (cplx)
	  && REGNO (cplx) < FIRST_PSEUDO_REGISTER
	  && REG_NREGS (cplx) % 2 == 0))
    {
      rtx ret = simplify_gen_subreg (imode, cplx, cmode,
				     imag_p ? GET_MODE_SIZE (imode) : 0);
      if (ret)
        return ret;
      else
	/* simplify_gen_subreg may fail for sub-word MEMs.  */
	gcc_assert (MEM_P (cplx) && ibitsize < BITS_PER_WORD);
    }

  return extract_bit_field (cplx, ibitsize, imag_p ? ibitsize : 0,
			    true, NULL_RTX, imode, imode, false, NULL);
}

/* A subroutine of emit_move_insn_1.  Yet another lowpart generator.
   NEW_MODE and OLD_MODE are the same size.  Return NULL if X cannot be
   represented in NEW_MODE.  If FORCE is true, this will never happen, as
   we'll force-create a SUBREG if needed.  */

static rtx
emit_move_change_mode (machine_mode new_mode,
		       machine_mode old_mode, rtx x, bool force)
{
  rtx ret;

  if (push_operand (x, GET_MODE (x)))
    {
      ret = gen_rtx_MEM (new_mode, XEXP (x, 0));
      MEM_COPY_ATTRIBUTES (ret, x);
    }
  else if (MEM_P (x))
    {
      /* We don't have to worry about changing the address since the
	 size in bytes is supposed to be the same.  */
      if (reload_in_progress)
	{
	  /* Copy the MEM to change the mode and move any
	     substitutions from the old MEM to the new one.  */
	  ret = adjust_address_nv (x, new_mode, 0);
	  copy_replacements (x, ret);
	}
      else
	ret = adjust_address (x, new_mode, 0);
    }
  else
    {
      /* Note that we do want simplify_subreg's behavior of validating
	 that the new mode is ok for a hard register.  If we were to use
	 simplify_gen_subreg, we would create the subreg, but would
	 probably run into the target not being able to implement it.  */
      /* Except, of course, when FORCE is true, when this is exactly what
	 we want.  Which is needed for CCmodes on some targets.  */
      if (force)
	ret = simplify_gen_subreg (new_mode, x, old_mode, 0);
      else
	ret = simplify_subreg (new_mode, x, old_mode, 0);
    }

  return ret;
}

/* A subroutine of emit_move_insn_1.  Generate a move from Y into X using
   an integer mode of the same size as MODE.  Returns the instruction
   emitted, or NULL if such a move could not be generated.  */

static rtx_insn *
emit_move_via_integer (machine_mode mode, rtx x, rtx y, bool force)
{
  scalar_int_mode imode;
  enum insn_code code;

  /* There must exist a mode of the exact size we require.  */
  if (!int_mode_for_mode (mode).exists (&imode))
    return NULL;

  /* The target must support moves in this mode.  */
  code = optab_handler (mov_optab, imode);
  if (code == CODE_FOR_nothing)
    return NULL;

  x = emit_move_change_mode (imode, mode, x, force);
  if (x == NULL_RTX)
    return NULL;
  y = emit_move_change_mode (imode, mode, y, force);
  if (y == NULL_RTX)
    return NULL;
  return emit_insn (GEN_FCN (code) (x, y));
}

/* A subroutine of emit_move_insn_1.  X is a push_operand in MODE.
   Return an equivalent MEM that does not use an auto-increment.  */

rtx
emit_move_resolve_push (machine_mode mode, rtx x)
{
  enum rtx_code code = GET_CODE (XEXP (x, 0));
  rtx temp;

  poly_int64 adjust = GET_MODE_SIZE (mode);
#ifdef PUSH_ROUNDING
  adjust = PUSH_ROUNDING (adjust);
#endif
  if (code == PRE_DEC || code == POST_DEC)
    adjust = -adjust;
  else if (code == PRE_MODIFY || code == POST_MODIFY)
    {
      rtx expr = XEXP (XEXP (x, 0), 1);

      gcc_assert (GET_CODE (expr) == PLUS || GET_CODE (expr) == MINUS);
      poly_int64 val = rtx_to_poly_int64 (XEXP (expr, 1));
      if (GET_CODE (expr) == MINUS)
	val = -val;
      gcc_assert (known_eq (adjust, val) || known_eq (adjust, -val));
      adjust = val;
    }

  /* Do not use anti_adjust_stack, since we don't want to update
     stack_pointer_delta.  */
  temp = expand_simple_binop (Pmode, PLUS, stack_pointer_rtx,
			      gen_int_mode (adjust, Pmode), stack_pointer_rtx,
			      0, OPTAB_LIB_WIDEN);
  if (temp != stack_pointer_rtx)
    emit_move_insn (stack_pointer_rtx, temp);

  switch (code)
    {
    case PRE_INC:
    case PRE_DEC:
    case PRE_MODIFY:
      temp = stack_pointer_rtx;
      break;
    case POST_INC:
    case POST_DEC:
    case POST_MODIFY:
      temp = plus_constant (Pmode, stack_pointer_rtx, -adjust);
      break;
    default:
      gcc_unreachable ();
    }

  return replace_equiv_address (x, temp);
}

/* A subroutine of emit_move_complex.  Generate a move from Y into X.
   X is known to satisfy push_operand, and MODE is known to be complex.
   Returns the last instruction emitted.  */

rtx_insn *
emit_move_complex_push (machine_mode mode, rtx x, rtx y)
{
  scalar_mode submode = GET_MODE_INNER (mode);
  bool imag_first;

#ifdef PUSH_ROUNDING
  poly_int64 submodesize = GET_MODE_SIZE (submode);

  /* In case we output to the stack, but the size is smaller than the
     machine can push exactly, we need to use move instructions.  */
  if (maybe_ne (PUSH_ROUNDING (submodesize), submodesize))
    {
      x = emit_move_resolve_push (mode, x);
      return emit_move_insn (x, y);
    }
#endif

  /* Note that the real part always precedes the imag part in memory
     regardless of machine's endianness.  */
  switch (GET_CODE (XEXP (x, 0)))
    {
    case PRE_DEC:
    case POST_DEC:
      imag_first = true;
      break;
    case PRE_INC:
    case POST_INC:
      imag_first = false;
      break;
    default:
      gcc_unreachable ();
    }

  emit_move_insn (gen_rtx_MEM (submode, XEXP (x, 0)),
		  read_complex_part (y, imag_first));
  return emit_move_insn (gen_rtx_MEM (submode, XEXP (x, 0)),
			 read_complex_part (y, !imag_first));
}

/* A subroutine of emit_move_complex.  Perform the move from Y to X
   via two moves of the parts.  Returns the last instruction emitted.  */

rtx_insn *
emit_move_complex_parts (rtx x, rtx y)
{
  /* Show the output dies here.  This is necessary for SUBREGs
     of pseudos since we cannot track their lifetimes correctly;
     hard regs shouldn't appear here except as return values.  */
  if (!reload_completed && !reload_in_progress
      && REG_P (x) && !reg_overlap_mentioned_p (x, y))
    emit_clobber (x);

  write_complex_part (x, read_complex_part (y, false), false);
  write_complex_part (x, read_complex_part (y, true), true);

  return get_last_insn ();
}

/* A subroutine of emit_move_insn_1.  Generate a move from Y into X.
   MODE is known to be complex.  Returns the last instruction emitted.  */

static rtx_insn *
emit_move_complex (machine_mode mode, rtx x, rtx y)
{
  bool try_int;

  /* Need to take special care for pushes, to maintain proper ordering
     of the data, and possibly extra padding.  */
  if (push_operand (x, mode))
    return emit_move_complex_push (mode, x, y);

  /* See if we can coerce the target into moving both values at once, except
     for floating point where we favor moving as parts if this is easy.  */
  if (GET_MODE_CLASS (mode) == MODE_COMPLEX_FLOAT
      && optab_handler (mov_optab, GET_MODE_INNER (mode)) != CODE_FOR_nothing
      && !(REG_P (x)
	   && HARD_REGISTER_P (x)
	   && REG_NREGS (x) == 1)
      && !(REG_P (y)
	   && HARD_REGISTER_P (y)
	   && REG_NREGS (y) == 1))
    try_int = false;
  /* Not possible if the values are inherently not adjacent.  */
  else if (GET_CODE (x) == CONCAT || GET_CODE (y) == CONCAT)
    try_int = false;
  /* Is possible if both are registers (or subregs of registers).  */
  else if (register_operand (x, mode) && register_operand (y, mode))
    try_int = true;
  /* If one of the operands is a memory, and alignment constraints
     are friendly enough, we may be able to do combined memory operations.
     We do not attempt this if Y is a constant because that combination is
     usually better with the by-parts thing below.  */
  else if ((MEM_P (x) ? !CONSTANT_P (y) : MEM_P (y))
	   && (!STRICT_ALIGNMENT
	       || get_mode_alignment (mode) == BIGGEST_ALIGNMENT))
    try_int = true;
  else
    try_int = false;

  if (try_int)
    {
      rtx_insn *ret;

      /* For memory to memory moves, optimal behavior can be had with the
	 existing block move logic.  But use normal expansion if optimizing
	 for size.  */
      if (MEM_P (x) && MEM_P (y))
	{
	  emit_block_move (x, y, gen_int_mode (GET_MODE_SIZE (mode), Pmode),
			   (optimize_insn_for_speed_p()
			    ? BLOCK_OP_NO_LIBCALL : BLOCK_OP_NORMAL));
	  return get_last_insn ();
	}

      ret = emit_move_via_integer (mode, x, y, true);
      if (ret)
	return ret;
    }

  return emit_move_complex_parts (x, y);
}

/* A subroutine of emit_move_insn_1.  Generate a move from Y into X.
   MODE is known to be MODE_CC.  Returns the last instruction emitted.  */

static rtx_insn *
emit_move_ccmode (machine_mode mode, rtx x, rtx y)
{
  rtx_insn *ret;

  /* Assume all MODE_CC modes are equivalent; if we have movcc, use it.  */
  if (mode != CCmode)
    {
      enum insn_code code = optab_handler (mov_optab, CCmode);
      if (code != CODE_FOR_nothing)
	{
	  x = emit_move_change_mode (CCmode, mode, x, true);
	  y = emit_move_change_mode (CCmode, mode, y, true);
	  return emit_insn (GEN_FCN (code) (x, y));
	}
    }

  /* Otherwise, find the MODE_INT mode of the same width.  */
  ret = emit_move_via_integer (mode, x, y, false);
  gcc_assert (ret != NULL);
  return ret;
}

/* Return true if word I of OP lies entirely in the
   undefined bits of a paradoxical subreg.  */

static bool
undefined_operand_subword_p (const_rtx op, int i)
{
  if (GET_CODE (op) != SUBREG)
    return false;
  machine_mode innermostmode = GET_MODE (SUBREG_REG (op));
  poly_int64 offset = i * UNITS_PER_WORD + subreg_memory_offset (op);
  return (known_ge (offset, GET_MODE_SIZE (innermostmode))
	  || known_le (offset, -UNITS_PER_WORD));
}

/* A subroutine of emit_move_insn_1.  Generate a move from Y into X.
   MODE is any multi-word or full-word mode that lacks a move_insn
   pattern.  Note that you will get better code if you define such
   patterns, even if they must turn into multiple assembler instructions.  */

static rtx_insn *
emit_move_multi_word (machine_mode mode, rtx x, rtx y)
{
  rtx_insn *last_insn = 0;
  rtx_insn *seq;
  rtx inner;
  bool need_clobber;
  int i, mode_size;

  /* This function can only handle cases where the number of words is
     known at compile time.  */
  mode_size = GET_MODE_SIZE (mode).to_constant ();
  gcc_assert (mode_size >= UNITS_PER_WORD);

  /* If X is a push on the stack, do the push now and replace
     X with a reference to the stack pointer.  */
  if (push_operand (x, mode))
    x = emit_move_resolve_push (mode, x);

  /* If we are in reload, see if either operand is a MEM whose address
     is scheduled for replacement.  */
  if (reload_in_progress && MEM_P (x)
      && (inner = find_replacement (&XEXP (x, 0))) != XEXP (x, 0))
    x = replace_equiv_address_nv (x, inner);
  if (reload_in_progress && MEM_P (y)
      && (inner = find_replacement (&XEXP (y, 0))) != XEXP (y, 0))
    y = replace_equiv_address_nv (y, inner);

  start_sequence ();

  need_clobber = false;
  for (i = 0; i < CEIL (mode_size, UNITS_PER_WORD); i++)
    {
      /* Do not generate code for a move if it would go entirely
	 to the non-existing bits of a paradoxical subreg.  */
      if (undefined_operand_subword_p (x, i))
	continue;

      rtx xpart = operand_subword (x, i, 1, mode);
      rtx ypart;

      /* Do not generate code for a move if it would come entirely
	 from the undefined bits of a paradoxical subreg.  */
      if (undefined_operand_subword_p (y, i))
	continue;

      ypart = operand_subword (y, i, 1, mode);

      /* If we can't get a part of Y, put Y into memory if it is a
	 constant.  Otherwise, force it into a register.  Then we must
	 be able to get a part of Y.  */
      if (ypart == 0 && CONSTANT_P (y))
	{
	  y = use_anchored_address (force_const_mem (mode, y));
	  ypart = operand_subword (y, i, 1, mode);
	}
      else if (ypart == 0)
	ypart = operand_subword_force (y, i, mode);

      gcc_assert (xpart && ypart);

      need_clobber |= (GET_CODE (xpart) == SUBREG);

      last_insn = emit_move_insn (xpart, ypart);
    }

  seq = get_insns ();
  end_sequence ();

  /* Show the output dies here.  This is necessary for SUBREGs
     of pseudos since we cannot track their lifetimes correctly;
     hard regs shouldn't appear here except as return values.
     We never want to emit such a clobber after reload.  */
  if (x != y
      && ! (reload_in_progress || reload_completed)
      && need_clobber != 0)
    emit_clobber (x);

  emit_insn (seq);

  return last_insn;
}

/* Low level part of emit_move_insn.
   Called just like emit_move_insn, but assumes X and Y
   are basically valid.  */

rtx_insn *
emit_move_insn_1 (rtx x, rtx y)
{
  machine_mode mode = GET_MODE (x);
  enum insn_code code;

  gcc_assert ((unsigned int) mode < (unsigned int) MAX_MACHINE_MODE);

  code = optab_handler (mov_optab, mode);
  if (code != CODE_FOR_nothing)
    return emit_insn (GEN_FCN (code) (x, y));

  /* Expand complex moves by moving real part and imag part.  */
  if (COMPLEX_MODE_P (mode))
    return emit_move_complex (mode, x, y);

  if (GET_MODE_CLASS (mode) == MODE_DECIMAL_FLOAT
      || ALL_FIXED_POINT_MODE_P (mode))
    {
      rtx_insn *result = emit_move_via_integer (mode, x, y, true);

      /* If we can't find an integer mode, use multi words.  */
      if (result)
	return result;
      else
	return emit_move_multi_word (mode, x, y);
    }

  if (GET_MODE_CLASS (mode) == MODE_CC)
    return emit_move_ccmode (mode, x, y);

  /* Try using a move pattern for the corresponding integer mode.  This is
     only safe when simplify_subreg can convert MODE constants into integer
     constants.  At present, it can only do this reliably if the value
     fits within a HOST_WIDE_INT.  */
  if (!CONSTANT_P (y)
      || known_le (GET_MODE_BITSIZE (mode), HOST_BITS_PER_WIDE_INT))
    {
      rtx_insn *ret = emit_move_via_integer (mode, x, y, lra_in_progress);

      if (ret)
	{
	  if (! lra_in_progress || recog (PATTERN (ret), ret, 0) >= 0)
	    return ret;
	}
    }

  return emit_move_multi_word (mode, x, y);
}

/* Generate code to copy Y into X.
   Both Y and X must have the same mode, except that
   Y can be a constant with VOIDmode.
   This mode cannot be BLKmode; use emit_block_move for that.

   Return the last instruction emitted.  */

rtx_insn *
emit_move_insn (rtx x, rtx y)
{
  machine_mode mode = GET_MODE (x);
  rtx y_cst = NULL_RTX;
  rtx_insn *last_insn;
  rtx set;

  gcc_assert (mode != BLKmode
	      && (GET_MODE (y) == mode || GET_MODE (y) == VOIDmode));

  if (CONSTANT_P (y))
    {
      if (optimize
	  && SCALAR_FLOAT_MODE_P (GET_MODE (x))
	  && (last_insn = compress_float_constant (x, y)))
	return last_insn;

      y_cst = y;

      if (!targetm.legitimate_constant_p (mode, y))
	{
	  y = force_const_mem (mode, y);

	  /* If the target's cannot_force_const_mem prevented the spill,
	     assume that the target's move expanders will also take care
	     of the non-legitimate constant.  */
	  if (!y)
	    y = y_cst;
	  else
	    y = use_anchored_address (y);
	}
    }

  /* If X or Y are memory references, verify that their addresses are valid
     for the machine.  */
  if (MEM_P (x)
      && (! memory_address_addr_space_p (GET_MODE (x), XEXP (x, 0),
					 MEM_ADDR_SPACE (x))
	  && ! push_operand (x, GET_MODE (x))))
    x = validize_mem (x);

  if (MEM_P (y)
      && ! memory_address_addr_space_p (GET_MODE (y), XEXP (y, 0),
					MEM_ADDR_SPACE (y)))
    y = validize_mem (y);

  gcc_assert (mode != BLKmode);

  last_insn = emit_move_insn_1 (x, y);

  if (y_cst && REG_P (x)
      && (set = single_set (last_insn)) != NULL_RTX
      && SET_DEST (set) == x
      && ! rtx_equal_p (y_cst, SET_SRC (set)))
    set_unique_reg_note (last_insn, REG_EQUAL, copy_rtx (y_cst));

  return last_insn;
}

/* Generate the body of an instruction to copy Y into X.
   It may be a list of insns, if one insn isn't enough.  */

rtx_insn *
gen_move_insn (rtx x, rtx y)
{
  rtx_insn *seq;

  start_sequence ();
  emit_move_insn_1 (x, y);
  seq = get_insns ();
  end_sequence ();
  return seq;
}

/* If Y is representable exactly in a narrower mode, and the target can
   perform the extension directly from constant or memory, then emit the
   move as an extension.  */

static rtx_insn *
compress_float_constant (rtx x, rtx y)
{
  machine_mode dstmode = GET_MODE (x);
  machine_mode orig_srcmode = GET_MODE (y);
  machine_mode srcmode;
  const REAL_VALUE_TYPE *r;
  int oldcost, newcost;
  bool speed = optimize_insn_for_speed_p ();

  r = CONST_DOUBLE_REAL_VALUE (y);

  if (targetm.legitimate_constant_p (dstmode, y))
    oldcost = set_src_cost (y, orig_srcmode, speed);
  else
    oldcost = set_src_cost (force_const_mem (dstmode, y), dstmode, speed);

  FOR_EACH_MODE_UNTIL (srcmode, orig_srcmode)
    {
      enum insn_code ic;
      rtx trunc_y;
      rtx_insn *last_insn;

      /* Skip if the target can't extend this way.  */
      ic = can_extend_p (dstmode, srcmode, 0);
      if (ic == CODE_FOR_nothing)
	continue;

      /* Skip if the narrowed value isn't exact.  */
      if (! exact_real_truncate (srcmode, r))
	continue;

      trunc_y = const_double_from_real_value (*r, srcmode);

      if (targetm.legitimate_constant_p (srcmode, trunc_y))
	{
	  /* Skip if the target needs extra instructions to perform
	     the extension.  */
	  if (!insn_operand_matches (ic, 1, trunc_y))
	    continue;
	  /* This is valid, but may not be cheaper than the original. */
	  newcost = set_src_cost (gen_rtx_FLOAT_EXTEND (dstmode, trunc_y),
				  dstmode, speed);
	  if (oldcost < newcost)
	    continue;
	}
      else if (float_extend_from_mem[dstmode][srcmode])
	{
	  trunc_y = force_const_mem (srcmode, trunc_y);
	  /* This is valid, but may not be cheaper than the original. */
	  newcost = set_src_cost (gen_rtx_FLOAT_EXTEND (dstmode, trunc_y),
				  dstmode, speed);
	  if (oldcost < newcost)
	    continue;
	  trunc_y = validize_mem (trunc_y);
	}
      else
	continue;

      /* For CSE's benefit, force the compressed constant pool entry
	 into a new pseudo.  This constant may be used in different modes,
	 and if not, combine will put things back together for us.  */
      trunc_y = force_reg (srcmode, trunc_y);

      /* If x is a hard register, perform the extension into a pseudo,
	 so that e.g. stack realignment code is aware of it.  */
      rtx target = x;
      if (REG_P (x) && HARD_REGISTER_P (x))
	target = gen_reg_rtx (dstmode);

      emit_unop_insn (ic, target, trunc_y, UNKNOWN);
      last_insn = get_last_insn ();

      if (REG_P (target))
	set_unique_reg_note (last_insn, REG_EQUAL, y);

      if (target != x)
	return emit_move_insn (x, target);
      return last_insn;
    }

  return NULL;
}

/* Pushing data onto the stack.  */

/* Push a block of length SIZE (perhaps variable)
   and return an rtx to address the beginning of the block.
   The value may be virtual_outgoing_args_rtx.

   EXTRA is the number of bytes of padding to push in addition to SIZE.
   BELOW nonzero means this padding comes at low addresses;
   otherwise, the padding comes at high addresses.  */

rtx
push_block (rtx size, poly_int64 extra, int below)
{
  rtx temp;

  size = convert_modes (Pmode, ptr_mode, size, 1);
  if (CONSTANT_P (size))
    anti_adjust_stack (plus_constant (Pmode, size, extra));
  else if (REG_P (size) && known_eq (extra, 0))
    anti_adjust_stack (size);
  else
    {
      temp = copy_to_mode_reg (Pmode, size);
      if (maybe_ne (extra, 0))
	temp = expand_binop (Pmode, add_optab, temp,
			     gen_int_mode (extra, Pmode),
			     temp, 0, OPTAB_LIB_WIDEN);
      anti_adjust_stack (temp);
    }

  if (STACK_GROWS_DOWNWARD)
    {
      temp = virtual_outgoing_args_rtx;
      if (maybe_ne (extra, 0) && below)
	temp = plus_constant (Pmode, temp, extra);
    }
  else
    {
      poly_int64 csize;
      if (poly_int_rtx_p (size, &csize))
	temp = plus_constant (Pmode, virtual_outgoing_args_rtx,
			      -csize - (below ? 0 : extra));
      else if (maybe_ne (extra, 0) && !below)
	temp = gen_rtx_PLUS (Pmode, virtual_outgoing_args_rtx,
			     negate_rtx (Pmode, plus_constant (Pmode, size,
							       extra)));
      else
	temp = gen_rtx_PLUS (Pmode, virtual_outgoing_args_rtx,
			     negate_rtx (Pmode, size));
    }

  return memory_address (NARROWEST_INT_MODE, temp);
}

/* A utility routine that returns the base of an auto-inc memory, or NULL.  */

static rtx
mem_autoinc_base (rtx mem)
{
  if (MEM_P (mem))
    {
      rtx addr = XEXP (mem, 0);
      if (GET_RTX_CLASS (GET_CODE (addr)) == RTX_AUTOINC)
	return XEXP (addr, 0);
    }
  return NULL;
}

/* A utility routine used here, in reload, and in try_split.  The insns
   after PREV up to and including LAST are known to adjust the stack,
   with a final value of END_ARGS_SIZE.  Iterate backward from LAST
   placing notes as appropriate.  PREV may be NULL, indicating the
   entire insn sequence prior to LAST should be scanned.

   The set of allowed stack pointer modifications is small:
     (1) One or more auto-inc style memory references (aka pushes),
     (2) One or more addition/subtraction with the SP as destination,
     (3) A single move insn with the SP as destination,
     (4) A call_pop insn,
     (5) Noreturn call insns if !ACCUMULATE_OUTGOING_ARGS.

   Insns in the sequence that do not modify the SP are ignored,
   except for noreturn calls.

   The return value is the amount of adjustment that can be trivially
   verified, via immediate operand or auto-inc.  If the adjustment
   cannot be trivially extracted, the return value is HOST_WIDE_INT_MIN.  */

poly_int64
find_args_size_adjust (rtx_insn *insn)
{
  rtx dest, set, pat;
  int i;

  pat = PATTERN (insn);
  set = NULL;

  /* Look for a call_pop pattern.  */
  if (CALL_P (insn))
    {
      /* We have to allow non-call_pop patterns for the case
	 of emit_single_push_insn of a TLS address.  */
      if (GET_CODE (pat) != PARALLEL)
	return 0;

      /* All call_pop have a stack pointer adjust in the parallel.
	 The call itself is always first, and the stack adjust is
	 usually last, so search from the end.  */
      for (i = XVECLEN (pat, 0) - 1; i > 0; --i)
	{
	  set = XVECEXP (pat, 0, i);
	  if (GET_CODE (set) != SET)
	    continue;
	  dest = SET_DEST (set);
	  if (dest == stack_pointer_rtx)
	    break;
	}
      /* We'd better have found the stack pointer adjust.  */
      if (i == 0)
	return 0;
      /* Fall through to process the extracted SET and DEST
	 as if it was a standalone insn.  */
    }
  else if (GET_CODE (pat) == SET)
    set = pat;
  else if ((set = single_set (insn)) != NULL)
    ;
  else if (GET_CODE (pat) == PARALLEL)
    {
      /* ??? Some older ports use a parallel with a stack adjust
	 and a store for a PUSH_ROUNDING pattern, rather than a
	 PRE/POST_MODIFY rtx.  Don't force them to update yet...  */
      /* ??? See h8300 and m68k, pushqi1.  */
      for (i = XVECLEN (pat, 0) - 1; i >= 0; --i)
	{
	  set = XVECEXP (pat, 0, i);
	  if (GET_CODE (set) != SET)
	    continue;
	  dest = SET_DEST (set);
	  if (dest == stack_pointer_rtx)
	    break;

	  /* We do not expect an auto-inc of the sp in the parallel.  */
	  gcc_checking_assert (mem_autoinc_base (dest) != stack_pointer_rtx);
	  gcc_checking_assert (mem_autoinc_base (SET_SRC (set))
			       != stack_pointer_rtx);
	}
      if (i < 0)
	return 0;
    }
  else
    return 0;

  dest = SET_DEST (set);

  /* Look for direct modifications of the stack pointer.  */
  if (REG_P (dest) && REGNO (dest) == STACK_POINTER_REGNUM)
    {
      /* Look for a trivial adjustment, otherwise assume nothing.  */
      /* Note that the SPU restore_stack_block pattern refers to
	 the stack pointer in V4SImode.  Consider that non-trivial.  */
      poly_int64 offset;
      if (SCALAR_INT_MODE_P (GET_MODE (dest))
	  && strip_offset (SET_SRC (set), &offset) == stack_pointer_rtx)
	return offset;
      /* ??? Reload can generate no-op moves, which will be cleaned
	 up later.  Recognize it and continue searching.  */
      else if (rtx_equal_p (dest, SET_SRC (set)))
	return 0;
      else
	return HOST_WIDE_INT_MIN;
    }
  else
    {
      rtx mem, addr;

      /* Otherwise only think about autoinc patterns.  */
      if (mem_autoinc_base (dest) == stack_pointer_rtx)
	{
	  mem = dest;
	  gcc_checking_assert (mem_autoinc_base (SET_SRC (set))
			       != stack_pointer_rtx);
	}
      else if (mem_autoinc_base (SET_SRC (set)) == stack_pointer_rtx)
	mem = SET_SRC (set);
      else
	return 0;

      addr = XEXP (mem, 0);
      switch (GET_CODE (addr))
	{
	case PRE_INC:
	case POST_INC:
	  return GET_MODE_SIZE (GET_MODE (mem));
	case PRE_DEC:
	case POST_DEC:
	  return -GET_MODE_SIZE (GET_MODE (mem));
	case PRE_MODIFY:
	case POST_MODIFY:
	  addr = XEXP (addr, 1);
	  gcc_assert (GET_CODE (addr) == PLUS);
	  gcc_assert (XEXP (addr, 0) == stack_pointer_rtx);
	  return rtx_to_poly_int64 (XEXP (addr, 1));
	default:
	  gcc_unreachable ();
	}
    }
}

poly_int64
fixup_args_size_notes (rtx_insn *prev, rtx_insn *last,
		       poly_int64 end_args_size)
{
  poly_int64 args_size = end_args_size;
  bool saw_unknown = false;
  rtx_insn *insn;

  for (insn = last; insn != prev; insn = PREV_INSN (insn))
    {
      if (!NONDEBUG_INSN_P (insn))
	continue;

      /* We might have existing REG_ARGS_SIZE notes, e.g. when pushing
	 a call argument containing a TLS address that itself requires
	 a call to __tls_get_addr.  The handling of stack_pointer_delta
	 in emit_single_push_insn is supposed to ensure that any such
	 notes are already correct.  */
      rtx note = find_reg_note (insn, REG_ARGS_SIZE, NULL_RTX);
      gcc_assert (!note || known_eq (args_size, get_args_size (note)));

      poly_int64 this_delta = find_args_size_adjust (insn);
      if (known_eq (this_delta, 0))
	{
	  if (!CALL_P (insn)
	      || ACCUMULATE_OUTGOING_ARGS
	      || find_reg_note (insn, REG_NORETURN, NULL_RTX) == NULL_RTX)
	    continue;
	}

      gcc_assert (!saw_unknown);
      if (known_eq (this_delta, HOST_WIDE_INT_MIN))
	saw_unknown = true;

      if (!note)
	add_args_size_note (insn, args_size);
      if (STACK_GROWS_DOWNWARD)
	this_delta = -poly_uint64 (this_delta);

      if (saw_unknown)
	args_size = HOST_WIDE_INT_MIN;
      else
	args_size -= this_delta;
    }

  return args_size;
}

#ifdef PUSH_ROUNDING
/* Emit single push insn.  */

static void
emit_single_push_insn_1 (machine_mode mode, rtx x, tree type)
{
  rtx dest_addr;
  poly_int64 rounded_size = PUSH_ROUNDING (GET_MODE_SIZE (mode));
  rtx dest;
  enum insn_code icode;

  /* If there is push pattern, use it.  Otherwise try old way of throwing
     MEM representing push operation to move expander.  */
  icode = optab_handler (push_optab, mode);
  if (icode != CODE_FOR_nothing)
    {
      class expand_operand ops[1];

      create_input_operand (&ops[0], x, mode);
      if (maybe_expand_insn (icode, 1, ops))
	return;
    }
  if (known_eq (GET_MODE_SIZE (mode), rounded_size))
    dest_addr = gen_rtx_fmt_e (STACK_PUSH_CODE, Pmode, stack_pointer_rtx);
  /* If we are to pad downward, adjust the stack pointer first and
     then store X into the stack location using an offset.  This is
     because emit_move_insn does not know how to pad; it does not have
     access to type.  */
  else if (targetm.calls.function_arg_padding (mode, type) == PAD_DOWNWARD)
    {
      emit_move_insn (stack_pointer_rtx,
		      expand_binop (Pmode,
				    STACK_GROWS_DOWNWARD ? sub_optab
				    : add_optab,
				    stack_pointer_rtx,
				    gen_int_mode (rounded_size, Pmode),
				    NULL_RTX, 0, OPTAB_LIB_WIDEN));

      poly_int64 offset = rounded_size - GET_MODE_SIZE (mode);
      if (STACK_GROWS_DOWNWARD && STACK_PUSH_CODE == POST_DEC)
	/* We have already decremented the stack pointer, so get the
	   previous value.  */
	offset += rounded_size;

      if (!STACK_GROWS_DOWNWARD && STACK_PUSH_CODE == POST_INC)
	/* We have already incremented the stack pointer, so get the
	   previous value.  */
	offset -= rounded_size;

      dest_addr = plus_constant (Pmode, stack_pointer_rtx, offset);
    }
  else
    {
      if (STACK_GROWS_DOWNWARD)
	/* ??? This seems wrong if STACK_PUSH_CODE == POST_DEC.  */
	dest_addr = plus_constant (Pmode, stack_pointer_rtx, -rounded_size);
      else
	/* ??? This seems wrong if STACK_PUSH_CODE == POST_INC.  */
	dest_addr = plus_constant (Pmode, stack_pointer_rtx, rounded_size);

      dest_addr = gen_rtx_PRE_MODIFY (Pmode, stack_pointer_rtx, dest_addr);
    }

  dest = gen_rtx_MEM (mode, dest_addr);

  if (type != 0)
    {
      set_mem_attributes (dest, type, 1);

      if (cfun->tail_call_marked)
	/* Function incoming arguments may overlap with sibling call
	   outgoing arguments and we cannot allow reordering of reads
	   from function arguments with stores to outgoing arguments
	   of sibling calls.  */
	set_mem_alias_set (dest, 0);
    }
  emit_move_insn (dest, x);
}

/* Emit and annotate a single push insn.  */

static void
emit_single_push_insn (machine_mode mode, rtx x, tree type)
{
  poly_int64 delta, old_delta = stack_pointer_delta;
  rtx_insn *prev = get_last_insn ();
  rtx_insn *last;

  emit_single_push_insn_1 (mode, x, type);

  /* Adjust stack_pointer_delta to describe the situation after the push
     we just performed.  Note that we must do this after the push rather
     than before the push in case calculating X needs pushes and pops of
     its own (e.g. if calling __tls_get_addr).  The REG_ARGS_SIZE notes
     for such pushes and pops must not include the effect of the future
     push of X.  */
  stack_pointer_delta += PUSH_ROUNDING (GET_MODE_SIZE (mode));

  last = get_last_insn ();

  /* Notice the common case where we emitted exactly one insn.  */
  if (PREV_INSN (last) == prev)
    {
      add_args_size_note (last, stack_pointer_delta);
      return;
    }

  delta = fixup_args_size_notes (prev, last, stack_pointer_delta);
  gcc_assert (known_eq (delta, HOST_WIDE_INT_MIN)
	      || known_eq (delta, old_delta));
}
#endif

/* If reading SIZE bytes from X will end up reading from
   Y return the number of bytes that overlap.  Return -1
   if there is no overlap or -2 if we can't determine
   (for example when X and Y have different base registers).  */

static int
memory_load_overlap (rtx x, rtx y, HOST_WIDE_INT size)
{
  rtx tmp = plus_constant (Pmode, x, size);
  rtx sub = simplify_gen_binary (MINUS, Pmode, tmp, y);

  if (!CONST_INT_P (sub))
    return -2;

  HOST_WIDE_INT val = INTVAL (sub);

  return IN_RANGE (val, 1, size) ? val : -1;
}

/* Generate code to push X onto the stack, assuming it has mode MODE and
   type TYPE.
   MODE is redundant except when X is a CONST_INT (since they don't
   carry mode info).
   SIZE is an rtx for the size of data to be copied (in bytes),
   needed only if X is BLKmode.
   Return true if successful.  May return false if asked to push a
   partial argument during a sibcall optimization (as specified by
   SIBCALL_P) and the incoming and outgoing pointers cannot be shown
   to not overlap.

   ALIGN (in bits) is maximum alignment we can assume.

   If PARTIAL and REG are both nonzero, then copy that many of the first
   bytes of X into registers starting with REG, and push the rest of X.
   The amount of space pushed is decreased by PARTIAL bytes.
   REG must be a hard register in this case.
   If REG is zero but PARTIAL is not, take any all others actions for an
   argument partially in registers, but do not actually load any
   registers.

   EXTRA is the amount in bytes of extra space to leave next to this arg.
   This is ignored if an argument block has already been allocated.

   On a machine that lacks real push insns, ARGS_ADDR is the address of
   the bottom of the argument block for this call.  We use indexing off there
   to store the arg.  On machines with push insns, ARGS_ADDR is 0 when a
   argument block has not been preallocated.

   ARGS_SO_FAR is the size of args previously pushed for this call.

   REG_PARM_STACK_SPACE is nonzero if functions require stack space
   for arguments passed in registers.  If nonzero, it will be the number
   of bytes required.  */

bool
emit_push_insn (rtx x, machine_mode mode, tree type, rtx size,
		unsigned int align, int partial, rtx reg, poly_int64 extra,
		rtx args_addr, rtx args_so_far, int reg_parm_stack_space,
		rtx alignment_pad, bool sibcall_p)
{
  rtx xinner;
  pad_direction stack_direction
    = STACK_GROWS_DOWNWARD ? PAD_DOWNWARD : PAD_UPWARD;

  /* Decide where to pad the argument: PAD_DOWNWARD for below,
     PAD_UPWARD for above, or PAD_NONE for don't pad it.
     Default is below for small data on big-endian machines; else above.  */
  pad_direction where_pad = targetm.calls.function_arg_padding (mode, type);

  /* Invert direction if stack is post-decrement.
     FIXME: why?  */
  if (STACK_PUSH_CODE == POST_DEC)
    if (where_pad != PAD_NONE)
      where_pad = (where_pad == PAD_DOWNWARD ? PAD_UPWARD : PAD_DOWNWARD);

  xinner = x;

  int nregs = partial / UNITS_PER_WORD;
  rtx *tmp_regs = NULL;
  int overlapping = 0;

  if (mode == BLKmode
      || (STRICT_ALIGNMENT && align < GET_MODE_ALIGNMENT (mode)))
    {
      /* Copy a block into the stack, entirely or partially.  */

      rtx temp;
      int used;
      int offset;
      int skip;

      offset = partial % (PARM_BOUNDARY / BITS_PER_UNIT);
      used = partial - offset;

      if (mode != BLKmode)
	{
	  /* A value is to be stored in an insufficiently aligned
	     stack slot; copy via a suitably aligned slot if
	     necessary.  */
	  size = gen_int_mode (GET_MODE_SIZE (mode), Pmode);
	  if (!MEM_P (xinner))
	    {
	      temp = assign_temp (type, 1, 1);
	      emit_move_insn (temp, xinner);
	      xinner = temp;
	    }
	}

      gcc_assert (size);

      /* USED is now the # of bytes we need not copy to the stack
	 because registers will take care of them.  */

      if (partial != 0)
	xinner = adjust_address (xinner, BLKmode, used);

      /* If the partial register-part of the arg counts in its stack size,
	 skip the part of stack space corresponding to the registers.
	 Otherwise, start copying to the beginning of the stack space,
	 by setting SKIP to 0.  */
      skip = (reg_parm_stack_space == 0) ? 0 : used;

#ifdef PUSH_ROUNDING
      /* Do it with several push insns if that doesn't take lots of insns
	 and if there is no difficulty with push insns that skip bytes
	 on the stack for alignment purposes.  */
      if (args_addr == 0
	  && PUSH_ARGS
	  && CONST_INT_P (size)
	  && skip == 0
	  && MEM_ALIGN (xinner) >= align
	  && can_move_by_pieces ((unsigned) INTVAL (size) - used, align)
	  /* Here we avoid the case of a structure whose weak alignment
	     forces many pushes of a small amount of data,
	     and such small pushes do rounding that causes trouble.  */
	  && ((!targetm.slow_unaligned_access (word_mode, align))
	      || align >= BIGGEST_ALIGNMENT
	      || known_eq (PUSH_ROUNDING (align / BITS_PER_UNIT),
			   align / BITS_PER_UNIT))
	  && known_eq (PUSH_ROUNDING (INTVAL (size)), INTVAL (size)))
	{
	  /* Push padding now if padding above and stack grows down,
	     or if padding below and stack grows up.
	     But if space already allocated, this has already been done.  */
	  if (maybe_ne (extra, 0)
	      && args_addr == 0
	      && where_pad != PAD_NONE
	      && where_pad != stack_direction)
	    anti_adjust_stack (gen_int_mode (extra, Pmode));

	  move_by_pieces (NULL, xinner, INTVAL (size) - used, align,
			  RETURN_BEGIN);
	}
      else
#endif /* PUSH_ROUNDING  */
	{
	  rtx target;

	  /* Otherwise make space on the stack and copy the data
	     to the address of that space.  */

	  /* Deduct words put into registers from the size we must copy.  */
	  if (partial != 0)
	    {
	      if (CONST_INT_P (size))
		size = GEN_INT (INTVAL (size) - used);
	      else
		size = expand_binop (GET_MODE (size), sub_optab, size,
				     gen_int_mode (used, GET_MODE (size)),
				     NULL_RTX, 0, OPTAB_LIB_WIDEN);
	    }

	  /* Get the address of the stack space.
	     In this case, we do not deal with EXTRA separately.
	     A single stack adjust will do.  */
	  poly_int64 const_args_so_far;
	  if (! args_addr)
	    {
	      temp = push_block (size, extra, where_pad == PAD_DOWNWARD);
	      extra = 0;
	    }
	  else if (poly_int_rtx_p (args_so_far, &const_args_so_far))
	    temp = memory_address (BLKmode,
				   plus_constant (Pmode, args_addr,
						  skip + const_args_so_far));
	  else
	    temp = memory_address (BLKmode,
				   plus_constant (Pmode,
						  gen_rtx_PLUS (Pmode,
								args_addr,
								args_so_far),
						  skip));

	  if (!ACCUMULATE_OUTGOING_ARGS)
	    {
	      /* If the source is referenced relative to the stack pointer,
		 copy it to another register to stabilize it.  We do not need
		 to do this if we know that we won't be changing sp.  */

	      if (reg_mentioned_p (virtual_stack_dynamic_rtx, temp)
		  || reg_mentioned_p (virtual_outgoing_args_rtx, temp))
		temp = copy_to_reg (temp);
	    }

	  target = gen_rtx_MEM (BLKmode, temp);

	  /* We do *not* set_mem_attributes here, because incoming arguments
	     may overlap with sibling call outgoing arguments and we cannot
	     allow reordering of reads from function arguments with stores
	     to outgoing arguments of sibling calls.  We do, however, want
	     to record the alignment of the stack slot.  */
	  /* ALIGN may well be better aligned than TYPE, e.g. due to
	     PARM_BOUNDARY.  Assume the caller isn't lying.  */
	  set_mem_align (target, align);

	  /* If part should go in registers and pushing to that part would
	     overwrite some of the values that need to go into regs, load the
	     overlapping values into temporary pseudos to be moved into the hard
	     regs at the end after the stack pushing has completed.
	     We cannot load them directly into the hard regs here because
	     they can be clobbered by the block move expansions.
	     See PR 65358.  */

	  if (partial > 0 && reg != 0 && mode == BLKmode
	      && GET_CODE (reg) != PARALLEL)
	    {
	      overlapping = memory_load_overlap (XEXP (x, 0), temp, partial);
	      if (overlapping > 0)
	        {
		  gcc_assert (overlapping % UNITS_PER_WORD == 0);
		  overlapping /= UNITS_PER_WORD;

		  tmp_regs = XALLOCAVEC (rtx, overlapping);

		  for (int i = 0; i < overlapping; i++)
		    tmp_regs[i] = gen_reg_rtx (word_mode);

		  for (int i = 0; i < overlapping; i++)
		    emit_move_insn (tmp_regs[i],
				    operand_subword_force (target, i, mode));
	        }
	      else if (overlapping == -1)
		overlapping = 0;
	      /* Could not determine whether there is overlap.
	         Fail the sibcall.  */
	      else
		{
		  overlapping = 0;
		  if (sibcall_p)
		    return false;
		}
	    }
	  emit_block_move (target, xinner, size, BLOCK_OP_CALL_PARM);
	}
    }
  else if (partial > 0)
    {
      /* Scalar partly in registers.  This case is only supported
	 for fixed-wdth modes.  */
      int num_words = GET_MODE_SIZE (mode).to_constant ();
      num_words /= UNITS_PER_WORD;
      int i;
      int not_stack;
      /* # bytes of start of argument
	 that we must make space for but need not store.  */
      int offset = partial % (PARM_BOUNDARY / BITS_PER_UNIT);
      int args_offset = INTVAL (args_so_far);
      int skip;

      /* Push padding now if padding above and stack grows down,
	 or if padding below and stack grows up.
	 But if space already allocated, this has already been done.  */
      if (maybe_ne (extra, 0)
	  && args_addr == 0
	  && where_pad != PAD_NONE
	  && where_pad != stack_direction)
	anti_adjust_stack (gen_int_mode (extra, Pmode));

      /* If we make space by pushing it, we might as well push
	 the real data.  Otherwise, we can leave OFFSET nonzero
	 and leave the space uninitialized.  */
      if (args_addr == 0)
	offset = 0;

      /* Now NOT_STACK gets the number of words that we don't need to
	 allocate on the stack.  Convert OFFSET to words too.  */
      not_stack = (partial - offset) / UNITS_PER_WORD;
      offset /= UNITS_PER_WORD;

      /* If the partial register-part of the arg counts in its stack size,
	 skip the part of stack space corresponding to the registers.
	 Otherwise, start copying to the beginning of the stack space,
	 by setting SKIP to 0.  */
      skip = (reg_parm_stack_space == 0) ? 0 : not_stack;

      if (CONSTANT_P (x) && !targetm.legitimate_constant_p (mode, x))
	x = validize_mem (force_const_mem (mode, x));

      /* If X is a hard register in a non-integer mode, copy it into a pseudo;
	 SUBREGs of such registers are not allowed.  */
      if ((REG_P (x) && REGNO (x) < FIRST_PSEUDO_REGISTER
	   && GET_MODE_CLASS (GET_MODE (x)) != MODE_INT))
	x = copy_to_reg (x);

      /* Loop over all the words allocated on the stack for this arg.  */
      /* We can do it by words, because any scalar bigger than a word
	 has a size a multiple of a word.  */
      for (i = num_words - 1; i >= not_stack; i--)
	if (i >= not_stack + offset)
	  if (!emit_push_insn (operand_subword_force (x, i, mode),
			  word_mode, NULL_TREE, NULL_RTX, align, 0, NULL_RTX,
			  0, args_addr,
			  GEN_INT (args_offset + ((i - not_stack + skip)
						  * UNITS_PER_WORD)),
			  reg_parm_stack_space, alignment_pad, sibcall_p))
	    return false;
    }
  else
    {
      rtx addr;
      rtx dest;

      /* Push padding now if padding above and stack grows down,
	 or if padding below and stack grows up.
	 But if space already allocated, this has already been done.  */
      if (maybe_ne (extra, 0)
	  && args_addr == 0
	  && where_pad != PAD_NONE
	  && where_pad != stack_direction)
	anti_adjust_stack (gen_int_mode (extra, Pmode));

#ifdef PUSH_ROUNDING
      if (args_addr == 0 && PUSH_ARGS)
	emit_single_push_insn (mode, x, type);
      else
#endif
	{
	  addr = simplify_gen_binary (PLUS, Pmode, args_addr, args_so_far);
	  dest = gen_rtx_MEM (mode, memory_address (mode, addr));

	  /* We do *not* set_mem_attributes here, because incoming arguments
	     may overlap with sibling call outgoing arguments and we cannot
	     allow reordering of reads from function arguments with stores
	     to outgoing arguments of sibling calls.  We do, however, want
	     to record the alignment of the stack slot.  */
	  /* ALIGN may well be better aligned than TYPE, e.g. due to
	     PARM_BOUNDARY.  Assume the caller isn't lying.  */
	  set_mem_align (dest, align);

	  emit_move_insn (dest, x);
	}
    }

  /* Move the partial arguments into the registers and any overlapping
     values that we moved into the pseudos in tmp_regs.  */
  if (partial > 0 && reg != 0)
    {
      /* Handle calls that pass values in multiple non-contiguous locations.
	 The Irix 6 ABI has examples of this.  */
      if (GET_CODE (reg) == PARALLEL)
	emit_group_load (reg, x, type, -1);
      else
        {
	  gcc_assert (partial % UNITS_PER_WORD == 0);
	  move_block_to_reg (REGNO (reg), x, nregs - overlapping, mode);

	  for (int i = 0; i < overlapping; i++)
	    emit_move_insn (gen_rtx_REG (word_mode, REGNO (reg)
						    + nregs - overlapping + i),
			    tmp_regs[i]);

	}
    }

  if (maybe_ne (extra, 0) && args_addr == 0 && where_pad == stack_direction)
    anti_adjust_stack (gen_int_mode (extra, Pmode));

  if (alignment_pad && args_addr == 0)
    anti_adjust_stack (alignment_pad);

  return true;
}

/* Return X if X can be used as a subtarget in a sequence of arithmetic
   operations.  */

static rtx
get_subtarget (rtx x)
{
  return (optimize
          || x == 0
	   /* Only registers can be subtargets.  */
	   || !REG_P (x)
	   /* Don't use hard regs to avoid extending their life.  */
	   || REGNO (x) < FIRST_PSEUDO_REGISTER
	  ? 0 : x);
}

/* A subroutine of expand_assignment.  Optimize FIELD op= VAL, where
   FIELD is a bitfield.  Returns true if the optimization was successful,
   and there's nothing else to do.  */

static bool
optimize_bitfield_assignment_op (poly_uint64 pbitsize,
				 poly_uint64 pbitpos,
				 poly_uint64 pbitregion_start,
				 poly_uint64 pbitregion_end,
				 machine_mode mode1, rtx str_rtx,
				 tree to, tree src, bool reverse)
{
  /* str_mode is not guaranteed to be a scalar type.  */
  machine_mode str_mode = GET_MODE (str_rtx);
  unsigned int str_bitsize;
  tree op0, op1;
  rtx value, result;
  optab binop;
  gimple *srcstmt;
  enum tree_code code;

  unsigned HOST_WIDE_INT bitsize, bitpos, bitregion_start, bitregion_end;
  if (mode1 != VOIDmode
      || !pbitsize.is_constant (&bitsize)
      || !pbitpos.is_constant (&bitpos)
      || !pbitregion_start.is_constant (&bitregion_start)
      || !pbitregion_end.is_constant (&bitregion_end)
      || bitsize >= BITS_PER_WORD
      || !GET_MODE_BITSIZE (str_mode).is_constant (&str_bitsize)
      || str_bitsize > BITS_PER_WORD
      || TREE_SIDE_EFFECTS (to)
      || TREE_THIS_VOLATILE (to))
    return false;

  STRIP_NOPS (src);
  if (TREE_CODE (src) != SSA_NAME)
    return false;
  if (TREE_CODE (TREE_TYPE (src)) != INTEGER_TYPE)
    return false;

  srcstmt = get_gimple_for_ssa_name (src);
  if (!srcstmt
      || TREE_CODE_CLASS (gimple_assign_rhs_code (srcstmt)) != tcc_binary)
    return false;

  code = gimple_assign_rhs_code (srcstmt);

  op0 = gimple_assign_rhs1 (srcstmt);

  /* If OP0 is an SSA_NAME, then we want to walk the use-def chain
     to find its initialization.  Hopefully the initialization will
     be from a bitfield load.  */
  if (TREE_CODE (op0) == SSA_NAME)
    {
      gimple *op0stmt = get_gimple_for_ssa_name (op0);

      /* We want to eventually have OP0 be the same as TO, which
	 should be a bitfield.  */
      if (!op0stmt
	  || !is_gimple_assign (op0stmt)
	  || gimple_assign_rhs_code (op0stmt) != TREE_CODE (to))
	return false;
      op0 = gimple_assign_rhs1 (op0stmt);
    }

  op1 = gimple_assign_rhs2 (srcstmt);

  if (!operand_equal_p (to, op0, 0))
    return false;

  if (MEM_P (str_rtx))
    {
      unsigned HOST_WIDE_INT offset1;

      if (str_bitsize == 0 || str_bitsize > BITS_PER_WORD)
	str_bitsize = BITS_PER_WORD;

      scalar_int_mode best_mode;
      if (!get_best_mode (bitsize, bitpos, bitregion_start, bitregion_end,
			  MEM_ALIGN (str_rtx), str_bitsize, false, &best_mode))
	return false;
      str_mode = best_mode;
      str_bitsize = GET_MODE_BITSIZE (best_mode);

      offset1 = bitpos;
      bitpos %= str_bitsize;
      offset1 = (offset1 - bitpos) / BITS_PER_UNIT;
      str_rtx = adjust_address (str_rtx, str_mode, offset1);
    }
  else if (!REG_P (str_rtx) && GET_CODE (str_rtx) != SUBREG)
    return false;

  /* If the bit field covers the whole REG/MEM, store_field
     will likely generate better code.  */
  if (bitsize >= str_bitsize)
    return false;

  /* We can't handle fields split across multiple entities.  */
  if (bitpos + bitsize > str_bitsize)
    return false;

  if (reverse ? !BYTES_BIG_ENDIAN : BYTES_BIG_ENDIAN)
    bitpos = str_bitsize - bitpos - bitsize;

  switch (code)
    {
    case PLUS_EXPR:
    case MINUS_EXPR:
      /* For now, just optimize the case of the topmost bitfield
	 where we don't need to do any masking and also
	 1 bit bitfields where xor can be used.
	 We might win by one instruction for the other bitfields
	 too if insv/extv instructions aren't used, so that
	 can be added later.  */
      if ((reverse || bitpos + bitsize != str_bitsize)
	  && (bitsize != 1 || TREE_CODE (op1) != INTEGER_CST))
	break;

      value = expand_expr (op1, NULL_RTX, str_mode, EXPAND_NORMAL);
      value = convert_modes (str_mode,
			     TYPE_MODE (TREE_TYPE (op1)), value,
			     TYPE_UNSIGNED (TREE_TYPE (op1)));

      /* We may be accessing data outside the field, which means
	 we can alias adjacent data.  */
      if (MEM_P (str_rtx))
	{
	  str_rtx = shallow_copy_rtx (str_rtx);
	  set_mem_alias_set (str_rtx, 0);
	  set_mem_expr (str_rtx, 0);
	}

      if (bitsize == 1 && (reverse || bitpos + bitsize != str_bitsize))
	{
	  value = expand_and (str_mode, value, const1_rtx, NULL);
	  binop = xor_optab;
	}
      else
	binop = code == PLUS_EXPR ? add_optab : sub_optab;

      value = expand_shift (LSHIFT_EXPR, str_mode, value, bitpos, NULL_RTX, 1);
      if (reverse)
	value = flip_storage_order (str_mode, value);
      result = expand_binop (str_mode, binop, str_rtx,
			     value, str_rtx, 1, OPTAB_WIDEN);
      if (result != str_rtx)
	emit_move_insn (str_rtx, result);
      return true;

    case BIT_IOR_EXPR:
    case BIT_XOR_EXPR:
      if (TREE_CODE (op1) != INTEGER_CST)
	break;
      value = expand_expr (op1, NULL_RTX, str_mode, EXPAND_NORMAL);
      value = convert_modes (str_mode,
			     TYPE_MODE (TREE_TYPE (op1)), value,
			     TYPE_UNSIGNED (TREE_TYPE (op1)));

      /* We may be accessing data outside the field, which means
	 we can alias adjacent data.  */
      if (MEM_P (str_rtx))
	{
	  str_rtx = shallow_copy_rtx (str_rtx);
	  set_mem_alias_set (str_rtx, 0);
	  set_mem_expr (str_rtx, 0);
	}

      binop = code == BIT_IOR_EXPR ? ior_optab : xor_optab;
      if (bitpos + bitsize != str_bitsize)
	{
	  rtx mask = gen_int_mode ((HOST_WIDE_INT_1U << bitsize) - 1,
				   str_mode);
	  value = expand_and (str_mode, value, mask, NULL_RTX);
	}
      value = expand_shift (LSHIFT_EXPR, str_mode, value, bitpos, NULL_RTX, 1);
      if (reverse)
	value = flip_storage_order (str_mode, value);
      result = expand_binop (str_mode, binop, str_rtx,
			     value, str_rtx, 1, OPTAB_WIDEN);
      if (result != str_rtx)
	emit_move_insn (str_rtx, result);
      return true;

    default:
      break;
    }

  return false;
}

/* In the C++ memory model, consecutive bit fields in a structure are
   considered one memory location.

   Given a COMPONENT_REF EXP at position (BITPOS, OFFSET), this function
   returns the bit range of consecutive bits in which this COMPONENT_REF
   belongs.  The values are returned in *BITSTART and *BITEND.  *BITPOS
   and *OFFSET may be adjusted in the process.

   If the access does not need to be restricted, 0 is returned in both
   *BITSTART and *BITEND.  */

void
get_bit_range (poly_uint64_pod *bitstart, poly_uint64_pod *bitend, tree exp,
	       poly_int64_pod *bitpos, tree *offset)
{
  poly_int64 bitoffset;
  tree field, repr;

  gcc_assert (TREE_CODE (exp) == COMPONENT_REF);

  field = TREE_OPERAND (exp, 1);
  repr = DECL_BIT_FIELD_REPRESENTATIVE (field);
  /* If we do not have a DECL_BIT_FIELD_REPRESENTATIVE there is no
     need to limit the range we can access.  */
  if (!repr)
    {
      *bitstart = *bitend = 0;
      return;
    }

  /* If we have a DECL_BIT_FIELD_REPRESENTATIVE but the enclosing record is
     part of a larger bit field, then the representative does not serve any
     useful purpose.  This can occur in Ada.  */
  if (handled_component_p (TREE_OPERAND (exp, 0)))
    {
      machine_mode rmode;
      poly_int64 rbitsize, rbitpos;
      tree roffset;
      int unsignedp, reversep, volatilep = 0;
      get_inner_reference (TREE_OPERAND (exp, 0), &rbitsize, &rbitpos,
			   &roffset, &rmode, &unsignedp, &reversep,
			   &volatilep);
      if (!multiple_p (rbitpos, BITS_PER_UNIT))
	{
	  *bitstart = *bitend = 0;
	  return;
	}
    }

  /* Compute the adjustment to bitpos from the offset of the field
     relative to the representative.  DECL_FIELD_OFFSET of field and
     repr are the same by construction if they are not constants,
     see finish_bitfield_layout.  */
  poly_uint64 field_offset, repr_offset;
  if (poly_int_tree_p (DECL_FIELD_OFFSET (field), &field_offset)
      && poly_int_tree_p (DECL_FIELD_OFFSET (repr), &repr_offset))
    bitoffset = (field_offset - repr_offset) * BITS_PER_UNIT;
  else
    bitoffset = 0;
  bitoffset += (tree_to_uhwi (DECL_FIELD_BIT_OFFSET (field))
		- tree_to_uhwi (DECL_FIELD_BIT_OFFSET (repr)));

  /* If the adjustment is larger than bitpos, we would have a negative bit
     position for the lower bound and this may wreak havoc later.  Adjust
     offset and bitpos to make the lower bound non-negative in that case.  */
  if (maybe_gt (bitoffset, *bitpos))
    {
      poly_int64 adjust_bits = upper_bound (bitoffset, *bitpos) - *bitpos;
      poly_int64 adjust_bytes = exact_div (adjust_bits, BITS_PER_UNIT);

      *bitpos += adjust_bits;
      if (*offset == NULL_TREE)
	*offset = size_int (-adjust_bytes);
      else
	*offset = size_binop (MINUS_EXPR, *offset, size_int (adjust_bytes));
      *bitstart = 0;
    }
  else
    *bitstart = *bitpos - bitoffset;

  *bitend = *bitstart + tree_to_poly_uint64 (DECL_SIZE (repr)) - 1;
}

/* Returns true if BASE is a DECL that does not reside in memory and
   has non-BLKmode.  DECL_RTL must not be a MEM; if
   DECL_RTL was not set yet, return false.  */

static inline bool
non_mem_decl_p (tree base)
{
  if (!DECL_P (base)
      || TREE_ADDRESSABLE (base)
      || DECL_MODE (base) == BLKmode)
    return false;

  if (!DECL_RTL_SET_P (base))
    return false;

  return (!MEM_P (DECL_RTL (base)));
}

/* Returns true if REF refers to an object that does not
   reside in memory and has non-BLKmode.  */

static inline bool
mem_ref_refers_to_non_mem_p (tree ref)
{
  tree base;

  if (TREE_CODE (ref) == MEM_REF
      || TREE_CODE (ref) == TARGET_MEM_REF)
    {
      tree addr = TREE_OPERAND (ref, 0);

      if (TREE_CODE (addr) != ADDR_EXPR)
	return false;

      base = TREE_OPERAND (addr, 0);
    }
  else
    base = ref;

  return non_mem_decl_p (base);
}

/* Expand an assignment that stores the value of FROM into TO.  If NONTEMPORAL
   is true, try generating a nontemporal store.  */

void
expand_assignment (tree to, tree from, bool nontemporal)
{
  rtx to_rtx = 0;
  rtx result;
  machine_mode mode;
  unsigned int align;
  enum insn_code icode;

  /* Don't crash if the lhs of the assignment was erroneous.  */
  if (TREE_CODE (to) == ERROR_MARK)
    {
      expand_normal (from);
      return;
    }

  /* Optimize away no-op moves without side-effects.  */
  if (operand_equal_p (to, from, 0))
    return;

  /* Handle misaligned stores.  */
  mode = TYPE_MODE (TREE_TYPE (to));
  if ((TREE_CODE (to) == MEM_REF
       || TREE_CODE (to) == TARGET_MEM_REF
       || DECL_P (to))
      && mode != BLKmode
      && !mem_ref_refers_to_non_mem_p (to)
      && ((align = get_object_alignment (to))
	  < GET_MODE_ALIGNMENT (mode))
      && (((icode = optab_handler (movmisalign_optab, mode))
	   != CODE_FOR_nothing)
	  || targetm.slow_unaligned_access (mode, align)))
    {
      rtx reg, mem;

      reg = expand_expr (from, NULL_RTX, VOIDmode, EXPAND_NORMAL);
      reg = force_not_mem (reg);
      mem = expand_expr (to, NULL_RTX, VOIDmode, EXPAND_WRITE);
      if (TREE_CODE (to) == MEM_REF && REF_REVERSE_STORAGE_ORDER (to))
	reg = flip_storage_order (mode, reg);

      if (icode != CODE_FOR_nothing)
	{
	  class expand_operand ops[2];

	  create_fixed_operand (&ops[0], mem);
	  create_input_operand (&ops[1], reg, mode);
	  /* The movmisalign<mode> pattern cannot fail, else the assignment
	     would silently be omitted.  */
	  expand_insn (icode, 2, ops);
	}
      else
	store_bit_field (mem, GET_MODE_BITSIZE (mode), 0, 0, 0, mode, reg,
			 false);
      return;
    }

  /* Assignment of a structure component needs special treatment
     if the structure component's rtx is not simply a MEM.
     Assignment of an array element at a constant index, and assignment of
     an array element in an unaligned packed structure field, has the same
     problem.  Same for (partially) storing into a non-memory object.  */
  if (handled_component_p (to)
      || (TREE_CODE (to) == MEM_REF
	  && (REF_REVERSE_STORAGE_ORDER (to)
	      || mem_ref_refers_to_non_mem_p (to)))
      || TREE_CODE (TREE_TYPE (to)) == ARRAY_TYPE)
    {
      machine_mode mode1;
      poly_int64 bitsize, bitpos;
      poly_uint64 bitregion_start = 0;
      poly_uint64 bitregion_end = 0;
      tree offset;
      int unsignedp, reversep, volatilep = 0;
      tree tem;

      push_temp_slots ();
      tem = get_inner_reference (to, &bitsize, &bitpos, &offset, &mode1,
				 &unsignedp, &reversep, &volatilep);

      /* Make sure bitpos is not negative, it can wreak havoc later.  */
      if (maybe_lt (bitpos, 0))
	{
	  gcc_assert (offset == NULL_TREE);
	  offset = size_int (bits_to_bytes_round_down (bitpos));
	  bitpos = num_trailing_bits (bitpos);
	}

      if (TREE_CODE (to) == COMPONENT_REF
	  && DECL_BIT_FIELD_TYPE (TREE_OPERAND (to, 1)))
	get_bit_range (&bitregion_start, &bitregion_end, to, &bitpos, &offset);
      /* The C++ memory model naturally applies to byte-aligned fields.
	 However, if we do not have a DECL_BIT_FIELD_TYPE but BITPOS or
	 BITSIZE are not byte-aligned, there is no need to limit the range
	 we can access.  This can occur with packed structures in Ada.  */
      else if (maybe_gt (bitsize, 0)
	       && multiple_p (bitsize, BITS_PER_UNIT)
	       && multiple_p (bitpos, BITS_PER_UNIT))
	{
	  bitregion_start = bitpos;
	  bitregion_end = bitpos + bitsize - 1;
	}

      to_rtx = expand_expr (tem, NULL_RTX, VOIDmode, EXPAND_WRITE);

      /* If the field has a mode, we want to access it in the
	 field's mode, not the computed mode.
	 If a MEM has VOIDmode (external with incomplete type),
	 use BLKmode for it instead.  */
      if (MEM_P (to_rtx))
	{
	  if (mode1 != VOIDmode)
	    to_rtx = adjust_address (to_rtx, mode1, 0);
	  else if (GET_MODE (to_rtx) == VOIDmode)
	    to_rtx = adjust_address (to_rtx, BLKmode, 0);
	}

      if (offset != 0)
	{
	  machine_mode address_mode;
	  rtx offset_rtx;

	  if (!MEM_P (to_rtx))
	    {
	      /* We can get constant negative offsets into arrays with broken
		 user code.  Translate this to a trap instead of ICEing.  */
	      gcc_assert (TREE_CODE (offset) == INTEGER_CST);
	      expand_builtin_trap ();
	      to_rtx = gen_rtx_MEM (BLKmode, const0_rtx);
	    }

	  offset_rtx = expand_expr (offset, NULL_RTX, VOIDmode, EXPAND_SUM);
	  address_mode = get_address_mode (to_rtx);
	  if (GET_MODE (offset_rtx) != address_mode)
	    {
		/* We cannot be sure that the RTL in offset_rtx is valid outside
		   of a memory address context, so force it into a register
		   before attempting to convert it to the desired mode.  */
	      offset_rtx = force_operand (offset_rtx, NULL_RTX);
	      offset_rtx = convert_to_mode (address_mode, offset_rtx, 0);
	    }

	  /* If we have an expression in OFFSET_RTX and a non-zero
	     byte offset in BITPOS, adding the byte offset before the
	     OFFSET_RTX results in better intermediate code, which makes
	     later rtl optimization passes perform better.

	     We prefer intermediate code like this:

	     r124:DI=r123:DI+0x18
	     [r124:DI]=r121:DI

	     ... instead of ...

	     r124:DI=r123:DI+0x10
	     [r124:DI+0x8]=r121:DI

	     This is only done for aligned data values, as these can
	     be expected to result in single move instructions.  */
	  poly_int64 bytepos;
	  if (mode1 != VOIDmode
	      && maybe_ne (bitpos, 0)
	      && maybe_gt (bitsize, 0)
	      && multiple_p (bitpos, BITS_PER_UNIT, &bytepos)
	      && multiple_p (bitpos, bitsize)
	      && multiple_p (bitsize, GET_MODE_ALIGNMENT (mode1))
	      && MEM_ALIGN (to_rtx) >= GET_MODE_ALIGNMENT (mode1))
	    {
	      to_rtx = adjust_address (to_rtx, mode1, bytepos);
	      bitregion_start = 0;
	      if (known_ge (bitregion_end, poly_uint64 (bitpos)))
		bitregion_end -= bitpos;
	      bitpos = 0;
	    }

	  to_rtx = offset_address (to_rtx, offset_rtx,
				   highest_pow2_factor_for_target (to,
				   				   offset));
	}

      /* No action is needed if the target is not a memory and the field
	 lies completely outside that target.  This can occur if the source
	 code contains an out-of-bounds access to a small array.  */
      if (!MEM_P (to_rtx)
	  && GET_MODE (to_rtx) != BLKmode
	  && known_ge (bitpos, GET_MODE_PRECISION (GET_MODE (to_rtx))))
	{
	  expand_normal (from);
	  result = NULL;
	}
      /* Handle expand_expr of a complex value returning a CONCAT.  */
      else if (GET_CODE (to_rtx) == CONCAT)
	{
	  machine_mode to_mode = GET_MODE (to_rtx);
	  gcc_checking_assert (COMPLEX_MODE_P (to_mode));
	  poly_int64 mode_bitsize = GET_MODE_BITSIZE (to_mode);
	  unsigned short inner_bitsize = GET_MODE_UNIT_BITSIZE (to_mode);
	  if (TYPE_MODE (TREE_TYPE (from)) == to_mode
	      && known_eq (bitpos, 0)
	      && known_eq (bitsize, mode_bitsize))
	    result = store_expr (from, to_rtx, false, nontemporal, reversep);
	  else if (TYPE_MODE (TREE_TYPE (from)) == GET_MODE_INNER (to_mode)
		   && known_eq (bitsize, inner_bitsize)
		   && (known_eq (bitpos, 0)
		       || known_eq (bitpos, inner_bitsize)))
	    result = store_expr (from, XEXP (to_rtx, maybe_ne (bitpos, 0)),
				 false, nontemporal, reversep);
	  else if (known_le (bitpos + bitsize, inner_bitsize))
	    result = store_field (XEXP (to_rtx, 0), bitsize, bitpos,
				  bitregion_start, bitregion_end,
				  mode1, from, get_alias_set (to),
				  nontemporal, reversep);
	  else if (known_ge (bitpos, inner_bitsize))
	    result = store_field (XEXP (to_rtx, 1), bitsize,
				  bitpos - inner_bitsize,
				  bitregion_start, bitregion_end,
				  mode1, from, get_alias_set (to),
				  nontemporal, reversep);
	  else if (known_eq (bitpos, 0) && known_eq (bitsize, mode_bitsize))
	    {
	      result = expand_normal (from);
	      if (GET_CODE (result) == CONCAT)
		{
		  to_mode = GET_MODE_INNER (to_mode);
		  machine_mode from_mode = GET_MODE_INNER (GET_MODE (result));
		  rtx from_real
		    = simplify_gen_subreg (to_mode, XEXP (result, 0),
					   from_mode, 0);
		  rtx from_imag
		    = simplify_gen_subreg (to_mode, XEXP (result, 1),
					   from_mode, 0);
		  if (!from_real || !from_imag)
		    goto concat_store_slow;
		  emit_move_insn (XEXP (to_rtx, 0), from_real);
		  emit_move_insn (XEXP (to_rtx, 1), from_imag);
		}
	      else
		{
		  machine_mode from_mode
		    = GET_MODE (result) == VOIDmode
		      ? TYPE_MODE (TREE_TYPE (from))
		      : GET_MODE (result);
		  rtx from_rtx;
		  if (MEM_P (result))
		    from_rtx = change_address (result, to_mode, NULL_RTX);
		  else
		    from_rtx
		      = simplify_gen_subreg (to_mode, result, from_mode, 0);
		  if (from_rtx)
		    {
		      emit_move_insn (XEXP (to_rtx, 0),
				      read_complex_part (from_rtx, false));
		      emit_move_insn (XEXP (to_rtx, 1),
				      read_complex_part (from_rtx, true));
		    }
		  else
		    {
		      to_mode = GET_MODE_INNER (to_mode);
		      rtx from_real
			= simplify_gen_subreg (to_mode, result, from_mode, 0);
		      rtx from_imag
			= simplify_gen_subreg (to_mode, result, from_mode,
					       GET_MODE_SIZE (to_mode));
		      if (!from_real || !from_imag)
			goto concat_store_slow;
		      emit_move_insn (XEXP (to_rtx, 0), from_real);
		      emit_move_insn (XEXP (to_rtx, 1), from_imag);
		    }
		}
	    }
	  else
	    {
	    concat_store_slow:;
	      rtx temp = assign_stack_temp (GET_MODE (to_rtx),
					    GET_MODE_SIZE (GET_MODE (to_rtx)));
	      write_complex_part (temp, XEXP (to_rtx, 0), false);
	      write_complex_part (temp, XEXP (to_rtx, 1), true);
	      result = store_field (temp, bitsize, bitpos,
				    bitregion_start, bitregion_end,
				    mode1, from, get_alias_set (to),
				    nontemporal, reversep);
	      emit_move_insn (XEXP (to_rtx, 0), read_complex_part (temp, false));
	      emit_move_insn (XEXP (to_rtx, 1), read_complex_part (temp, true));
	    }
	}
      /* For calls to functions returning variable length structures, if TO_RTX
	 is not a MEM, go through a MEM because we must not create temporaries
	 of the VLA type.  */
      else if (!MEM_P (to_rtx)
	       && TREE_CODE (from) == CALL_EXPR
	       && COMPLETE_TYPE_P (TREE_TYPE (from))
	       && TREE_CODE (TYPE_SIZE (TREE_TYPE (from))) != INTEGER_CST)
	{
	  rtx temp = assign_stack_temp (GET_MODE (to_rtx),
					GET_MODE_SIZE (GET_MODE (to_rtx)));
	  result = store_field (temp, bitsize, bitpos, bitregion_start,
				bitregion_end, mode1, from, get_alias_set (to),
				nontemporal, reversep);
	  emit_move_insn (to_rtx, temp);
	}
      else
	{
	  if (MEM_P (to_rtx))
	    {
	      /* If the field is at offset zero, we could have been given the
		 DECL_RTX of the parent struct.  Don't munge it.  */
	      to_rtx = shallow_copy_rtx (to_rtx);
	      set_mem_attributes_minus_bitpos (to_rtx, to, 0, bitpos);
	      if (volatilep)
		MEM_VOLATILE_P (to_rtx) = 1;
	    }

	  gcc_checking_assert (known_ge (bitpos, 0));
	  if (optimize_bitfield_assignment_op (bitsize, bitpos,
					       bitregion_start, bitregion_end,
					       mode1, to_rtx, to, from,
					       reversep))
	    result = NULL;
	  else
	    result = store_field (to_rtx, bitsize, bitpos,
				  bitregion_start, bitregion_end,
				  mode1, from, get_alias_set (to),
				  nontemporal, reversep);
	}

      if (result)
	preserve_temp_slots (result);
      pop_temp_slots ();
      return;
    }

  /* If the rhs is a function call and its value is not an aggregate,
     call the function before we start to compute the lhs.
     This is needed for correct code for cases such as
     val = setjmp (buf) on machines where reference to val
     requires loading up part of an address in a separate insn.

     Don't do this if TO is a VAR_DECL or PARM_DECL whose DECL_RTL is REG
     since it might be a promoted variable where the zero- or sign- extension
     needs to be done.  Handling this in the normal way is safe because no
     computation is done before the call.  The same is true for SSA names.  */
  if (TREE_CODE (from) == CALL_EXPR && ! aggregate_value_p (from, from)
      && COMPLETE_TYPE_P (TREE_TYPE (from))
      && TREE_CODE (TYPE_SIZE (TREE_TYPE (from))) == INTEGER_CST
      && ! (((VAR_P (to)
	      || TREE_CODE (to) == PARM_DECL
	      || TREE_CODE (to) == RESULT_DECL)
	     && REG_P (DECL_RTL (to)))
	    || TREE_CODE (to) == SSA_NAME))
    {
      rtx value;

      push_temp_slots ();
      value = expand_normal (from);

      if (to_rtx == 0)
	to_rtx = expand_expr (to, NULL_RTX, VOIDmode, EXPAND_WRITE);

      /* Handle calls that return values in multiple non-contiguous locations.
	 The Irix 6 ABI has examples of this.  */
      if (GET_CODE (to_rtx) == PARALLEL)
	{
	  if (GET_CODE (value) == PARALLEL)
	    emit_group_move (to_rtx, value);
	  else
	    emit_group_load (to_rtx, value, TREE_TYPE (from),
			     int_size_in_bytes (TREE_TYPE (from)));
	}
      else if (GET_CODE (value) == PARALLEL)
	emit_group_store (to_rtx, value, TREE_TYPE (from),
			  int_size_in_bytes (TREE_TYPE (from)));
      else if (GET_MODE (to_rtx) == BLKmode)
	{
	  /* Handle calls that return BLKmode values in registers.  */
	  if (REG_P (value))
	    copy_blkmode_from_reg (to_rtx, value, TREE_TYPE (from));
	  else
	    emit_block_move (to_rtx, value, expr_size (from), BLOCK_OP_NORMAL);
	}
      else
	{
	  if (POINTER_TYPE_P (TREE_TYPE (to)))
	    value = convert_memory_address_addr_space
	      (as_a <scalar_int_mode> (GET_MODE (to_rtx)), value,
	       TYPE_ADDR_SPACE (TREE_TYPE (TREE_TYPE (to))));

	  emit_move_insn (to_rtx, value);
	}

      preserve_temp_slots (to_rtx);
      pop_temp_slots ();
      return;
    }

  /* Ordinary treatment.  Expand TO to get a REG or MEM rtx.  */
  to_rtx = expand_expr (to, NULL_RTX, VOIDmode, EXPAND_WRITE);

  /* Don't move directly into a return register.  */
  if (TREE_CODE (to) == RESULT_DECL
      && (REG_P (to_rtx) || GET_CODE (to_rtx) == PARALLEL))
    {
      rtx temp;

      push_temp_slots ();

      /* If the source is itself a return value, it still is in a pseudo at
	 this point so we can move it back to the return register directly.  */
      if (REG_P (to_rtx)
	  && TYPE_MODE (TREE_TYPE (from)) == BLKmode
	  && TREE_CODE (from) != CALL_EXPR)
	temp = copy_blkmode_to_reg (GET_MODE (to_rtx), from);
      else
	temp = expand_expr (from, NULL_RTX, GET_MODE (to_rtx), EXPAND_NORMAL);

      /* Handle calls that return values in multiple non-contiguous locations.
	 The Irix 6 ABI has examples of this.  */
      if (GET_CODE (to_rtx) == PARALLEL)
	{
	  if (GET_CODE (temp) == PARALLEL)
	    emit_group_move (to_rtx, temp);
	  else
	    emit_group_load (to_rtx, temp, TREE_TYPE (from),
			     int_size_in_bytes (TREE_TYPE (from)));
	}
      else if (temp)
	emit_move_insn (to_rtx, temp);

      preserve_temp_slots (to_rtx);
      pop_temp_slots ();
      return;
    }

  /* In case we are returning the contents of an object which overlaps
     the place the value is being stored, use a safe function when copying
     a value through a pointer into a structure value return block.  */
  if (TREE_CODE (to) == RESULT_DECL
      && TREE_CODE (from) == INDIRECT_REF
      && ADDR_SPACE_GENERIC_P
	   (TYPE_ADDR_SPACE (TREE_TYPE (TREE_TYPE (TREE_OPERAND (from, 0)))))
      && refs_may_alias_p (to, from)
      && cfun->returns_struct
      && !cfun->returns_pcc_struct)
    {
      rtx from_rtx, size;

      push_temp_slots ();
      size = expr_size (from);
      from_rtx = expand_normal (from);

      emit_block_move_via_libcall (XEXP (to_rtx, 0), XEXP (from_rtx, 0), size);

      preserve_temp_slots (to_rtx);
      pop_temp_slots ();
      return;
    }

  /* Compute FROM and store the value in the rtx we got.  */

  push_temp_slots ();
  result = store_expr (from, to_rtx, 0, nontemporal, false);
  preserve_temp_slots (result);
  pop_temp_slots ();
  return;
}

/* Emits nontemporal store insn that moves FROM to TO.  Returns true if this
   succeeded, false otherwise.  */

bool
emit_storent_insn (rtx to, rtx from)
{
  class expand_operand ops[2];
  machine_mode mode = GET_MODE (to);
  enum insn_code code = optab_handler (storent_optab, mode);

  if (code == CODE_FOR_nothing)
    return false;

  create_fixed_operand (&ops[0], to);
  create_input_operand (&ops[1], from, mode);
  return maybe_expand_insn (code, 2, ops);
}

/* Helper function for store_expr storing of STRING_CST.  */

static rtx
string_cst_read_str (void *data, HOST_WIDE_INT offset, scalar_int_mode mode)
{
  tree str = (tree) data;

  gcc_assert (offset >= 0);
  if (offset >= TREE_STRING_LENGTH (str))
    return const0_rtx;

  if ((unsigned HOST_WIDE_INT) offset + GET_MODE_SIZE (mode)
      > (unsigned HOST_WIDE_INT) TREE_STRING_LENGTH (str))
    {
      char *p = XALLOCAVEC (char, GET_MODE_SIZE (mode));
      size_t l = TREE_STRING_LENGTH (str) - offset;
      memcpy (p, TREE_STRING_POINTER (str) + offset, l);
      memset (p + l, '\0', GET_MODE_SIZE (mode) - l);
      return c_readstr (p, mode, false);
    }

  return c_readstr (TREE_STRING_POINTER (str) + offset, mode, false);
}

/* Generate code for computing expression EXP,
   and storing the value into TARGET.

   If the mode is BLKmode then we may return TARGET itself.
   It turns out that in BLKmode it doesn't cause a problem.
   because C has no operators that could combine two different
   assignments into the same BLKmode object with different values
   with no sequence point.  Will other languages need this to
   be more thorough?

   If CALL_PARAM_P is nonzero, this is a store into a call param on the
   stack, and block moves may need to be treated specially.

   If NONTEMPORAL is true, try using a nontemporal store instruction.

   If REVERSE is true, the store is to be done in reverse order.  */

rtx
store_expr (tree exp, rtx target, int call_param_p,
	    bool nontemporal, bool reverse)
{
  rtx temp;
  rtx alt_rtl = NULL_RTX;
  location_t loc = curr_insn_location ();

  if (VOID_TYPE_P (TREE_TYPE (exp)))
    {
      /* C++ can generate ?: expressions with a throw expression in one
	 branch and an rvalue in the other. Here, we resolve attempts to
	 store the throw expression's nonexistent result.  */
      gcc_assert (!call_param_p);
      expand_expr (exp, const0_rtx, VOIDmode, EXPAND_NORMAL);
      return NULL_RTX;
    }
  if (TREE_CODE (exp) == COMPOUND_EXPR)
    {
      /* Perform first part of compound expression, then assign from second
	 part.  */
      expand_expr (TREE_OPERAND (exp, 0), const0_rtx, VOIDmode,
		   call_param_p ? EXPAND_STACK_PARM : EXPAND_NORMAL);
      return store_expr (TREE_OPERAND (exp, 1), target,
				     call_param_p, nontemporal, reverse);
    }
  else if (TREE_CODE (exp) == COND_EXPR && GET_MODE (target) == BLKmode)
    {
      /* For conditional expression, get safe form of the target.  Then
	 test the condition, doing the appropriate assignment on either
	 side.  This avoids the creation of unnecessary temporaries.
	 For non-BLKmode, it is more efficient not to do this.  */

      rtx_code_label *lab1 = gen_label_rtx (), *lab2 = gen_label_rtx ();

      do_pending_stack_adjust ();
      NO_DEFER_POP;
      jumpifnot (TREE_OPERAND (exp, 0), lab1,
		 profile_probability::uninitialized ());
      store_expr (TREE_OPERAND (exp, 1), target, call_param_p,
		  nontemporal, reverse);
      emit_jump_insn (targetm.gen_jump (lab2));
      emit_barrier ();
      emit_label (lab1);
      store_expr (TREE_OPERAND (exp, 2), target, call_param_p,
		  nontemporal, reverse);
      emit_label (lab2);
      OK_DEFER_POP;

      return NULL_RTX;
    }
  else if (GET_CODE (target) == SUBREG && SUBREG_PROMOTED_VAR_P (target))
    /* If this is a scalar in a register that is stored in a wider mode
       than the declared mode, compute the result into its declared mode
       and then convert to the wider mode.  Our value is the computed
       expression.  */
    {
      rtx inner_target = 0;
      scalar_int_mode outer_mode = subreg_unpromoted_mode (target);
      scalar_int_mode inner_mode = subreg_promoted_mode (target);

      /* We can do the conversion inside EXP, which will often result
	 in some optimizations.  Do the conversion in two steps: first
	 change the signedness, if needed, then the extend.  But don't
	 do this if the type of EXP is a subtype of something else
	 since then the conversion might involve more than just
	 converting modes.  */
      if (INTEGRAL_TYPE_P (TREE_TYPE (exp))
	  && TREE_TYPE (TREE_TYPE (exp)) == 0
	  && GET_MODE_PRECISION (outer_mode)
	     == TYPE_PRECISION (TREE_TYPE (exp)))
	{
	  if (!SUBREG_CHECK_PROMOTED_SIGN (target,
					  TYPE_UNSIGNED (TREE_TYPE (exp))))
	    {
	      /* Some types, e.g. Fortran's logical*4, won't have a signed
		 version, so use the mode instead.  */
	      tree ntype
		= (signed_or_unsigned_type_for
		   (SUBREG_PROMOTED_SIGN (target), TREE_TYPE (exp)));
	      if (ntype == NULL)
		ntype = lang_hooks.types.type_for_mode
		  (TYPE_MODE (TREE_TYPE (exp)),
		   SUBREG_PROMOTED_SIGN (target));

	      exp = fold_convert_loc (loc, ntype, exp);
	    }

	  exp = fold_convert_loc (loc, lang_hooks.types.type_for_mode
				  (inner_mode, SUBREG_PROMOTED_SIGN (target)),
				  exp);

	  inner_target = SUBREG_REG (target);
	}

      temp = expand_expr (exp, inner_target, VOIDmode,
			  call_param_p ? EXPAND_STACK_PARM : EXPAND_NORMAL);


      /* If TEMP is a VOIDmode constant, use convert_modes to make
	 sure that we properly convert it.  */
      if (CONSTANT_P (temp) && GET_MODE (temp) == VOIDmode)
	{
	  temp = convert_modes (outer_mode, TYPE_MODE (TREE_TYPE (exp)),
				temp, SUBREG_PROMOTED_SIGN (target));
	  temp = convert_modes (inner_mode, outer_mode, temp,
				SUBREG_PROMOTED_SIGN (target));
	}

      convert_move (SUBREG_REG (target), temp,
		    SUBREG_PROMOTED_SIGN (target));

      return NULL_RTX;
    }
  else if ((TREE_CODE (exp) == STRING_CST
	    || (TREE_CODE (exp) == MEM_REF
		&& TREE_CODE (TREE_OPERAND (exp, 0)) == ADDR_EXPR
		&& TREE_CODE (TREE_OPERAND (TREE_OPERAND (exp, 0), 0))
		   == STRING_CST
		&& integer_zerop (TREE_OPERAND (exp, 1))))
	   && !nontemporal && !call_param_p
	   && MEM_P (target))
    {
      /* Optimize initialization of an array with a STRING_CST.  */
      HOST_WIDE_INT exp_len, str_copy_len;
      rtx dest_mem;
      tree str = TREE_CODE (exp) == STRING_CST
		 ? exp : TREE_OPERAND (TREE_OPERAND (exp, 0), 0);

      exp_len = int_expr_size (exp);
      if (exp_len <= 0)
	goto normal_expr;

      if (TREE_STRING_LENGTH (str) <= 0)
	goto normal_expr;

      if (can_store_by_pieces (exp_len, string_cst_read_str, (void *) str,
			       MEM_ALIGN (target), false))
	{
	  store_by_pieces (target, exp_len, string_cst_read_str, (void *) str,
			   MEM_ALIGN (target), false, RETURN_BEGIN);
	  return NULL_RTX;
	}

      str_copy_len = TREE_STRING_LENGTH (str);
      if ((STORE_MAX_PIECES & (STORE_MAX_PIECES - 1)) == 0)
	{
	  str_copy_len += STORE_MAX_PIECES - 1;
	  str_copy_len &= ~(STORE_MAX_PIECES - 1);
	}
      if (str_copy_len >= exp_len)
	goto normal_expr;

      if (!can_store_by_pieces (str_copy_len, string_cst_read_str,
				(void *) str, MEM_ALIGN (target), false))
	goto normal_expr;

      dest_mem = store_by_pieces (target, str_copy_len, string_cst_read_str,
				  (void *) str, MEM_ALIGN (target), false,
				  RETURN_END);
      clear_storage (adjust_address_1 (dest_mem, BLKmode, 0, 1, 1, 0,
				       exp_len - str_copy_len),
		     GEN_INT (exp_len - str_copy_len), BLOCK_OP_NORMAL);
      return NULL_RTX;
    }
  else
    {
      rtx tmp_target;

  normal_expr:
      /* If we want to use a nontemporal or a reverse order store, force the
	 value into a register first.  */
      tmp_target = nontemporal || reverse ? NULL_RTX : target;
      temp = expand_expr_real (exp, tmp_target, GET_MODE (target),
			       (call_param_p
				? EXPAND_STACK_PARM : EXPAND_NORMAL),
			       &alt_rtl, false);
    }

  /* If TEMP is a VOIDmode constant and the mode of the type of EXP is not
     the same as that of TARGET, adjust the constant.  This is needed, for
     example, in case it is a CONST_DOUBLE or CONST_WIDE_INT and we want
     only a word-sized value.  */
  if (CONSTANT_P (temp) && GET_MODE (temp) == VOIDmode
      && TREE_CODE (exp) != ERROR_MARK
      && GET_MODE (target) != TYPE_MODE (TREE_TYPE (exp)))
    {
      if (GET_MODE_CLASS (GET_MODE (target))
	  != GET_MODE_CLASS (TYPE_MODE (TREE_TYPE (exp)))
	  && known_eq (GET_MODE_BITSIZE (GET_MODE (target)),
		       GET_MODE_BITSIZE (TYPE_MODE (TREE_TYPE (exp)))))
	{
	  rtx t = simplify_gen_subreg (GET_MODE (target), temp,
				       TYPE_MODE (TREE_TYPE (exp)), 0);
	  if (t)
	    temp = t;
	}
      if (GET_MODE (temp) == VOIDmode)
	temp = convert_modes (GET_MODE (target), TYPE_MODE (TREE_TYPE (exp)),
			      temp, TYPE_UNSIGNED (TREE_TYPE (exp)));
    }

  /* If value was not generated in the target, store it there.
     Convert the value to TARGET's type first if necessary and emit the
     pending incrementations that have been queued when expanding EXP.
     Note that we cannot emit the whole queue blindly because this will
     effectively disable the POST_INC optimization later.

     If TEMP and TARGET compare equal according to rtx_equal_p, but
     one or both of them are volatile memory refs, we have to distinguish
     two cases:
     - expand_expr has used TARGET.  In this case, we must not generate
       another copy.  This can be detected by TARGET being equal according
       to == .
     - expand_expr has not used TARGET - that means that the source just
       happens to have the same RTX form.  Since temp will have been created
       by expand_expr, it will compare unequal according to == .
       We must generate a copy in this case, to reach the correct number
       of volatile memory references.  */

  if ((! rtx_equal_p (temp, target)
       || (temp != target && (side_effects_p (temp)
			      || side_effects_p (target))))
      && TREE_CODE (exp) != ERROR_MARK
      /* If store_expr stores a DECL whose DECL_RTL(exp) == TARGET,
	 but TARGET is not valid memory reference, TEMP will differ
	 from TARGET although it is really the same location.  */
      && !(alt_rtl
	   && rtx_equal_p (alt_rtl, target)
	   && !side_effects_p (alt_rtl)
	   && !side_effects_p (target))
      /* If there's nothing to copy, don't bother.  Don't call
	 expr_size unless necessary, because some front-ends (C++)
	 expr_size-hook must not be given objects that are not
	 supposed to be bit-copied or bit-initialized.  */
      && expr_size (exp) != const0_rtx)
    {
      if (GET_MODE (temp) != GET_MODE (target) && GET_MODE (temp) != VOIDmode)
	{
	  if (GET_MODE (target) == BLKmode)
	    {
	      /* Handle calls that return BLKmode values in registers.  */
	      if (REG_P (temp) && TREE_CODE (exp) == CALL_EXPR)
		copy_blkmode_from_reg (target, temp, TREE_TYPE (exp));
	      else
		store_bit_field (target,
				 rtx_to_poly_int64 (expr_size (exp))
				 * BITS_PER_UNIT,
				 0, 0, 0, GET_MODE (temp), temp, reverse);
	    }
	  else
	    convert_move (target, temp, TYPE_UNSIGNED (TREE_TYPE (exp)));
	}

      else if (GET_MODE (temp) == BLKmode && TREE_CODE (exp) == STRING_CST)
	{
	  /* Handle copying a string constant into an array.  The string
	     constant may be shorter than the array.  So copy just the string's
	     actual length, and clear the rest.  First get the size of the data
	     type of the string, which is actually the size of the target.  */
	  rtx size = expr_size (exp);

	  if (CONST_INT_P (size)
	      && INTVAL (size) < TREE_STRING_LENGTH (exp))
	    emit_block_move (target, temp, size,
			     (call_param_p
			      ? BLOCK_OP_CALL_PARM : BLOCK_OP_NORMAL));
	  else
	    {
	      machine_mode pointer_mode
		= targetm.addr_space.pointer_mode (MEM_ADDR_SPACE (target));
	      machine_mode address_mode = get_address_mode (target);

	      /* Compute the size of the data to copy from the string.  */
	      tree copy_size
		= size_binop_loc (loc, MIN_EXPR,
				  make_tree (sizetype, size),
				  size_int (TREE_STRING_LENGTH (exp)));
	      rtx copy_size_rtx
		= expand_expr (copy_size, NULL_RTX, VOIDmode,
			       (call_param_p
				? EXPAND_STACK_PARM : EXPAND_NORMAL));
	      rtx_code_label *label = 0;

	      /* Copy that much.  */
	      copy_size_rtx = convert_to_mode (pointer_mode, copy_size_rtx,
					       TYPE_UNSIGNED (sizetype));
	      emit_block_move (target, temp, copy_size_rtx,
			       (call_param_p
				? BLOCK_OP_CALL_PARM : BLOCK_OP_NORMAL));

	      /* Figure out how much is left in TARGET that we have to clear.
		 Do all calculations in pointer_mode.  */
	      poly_int64 const_copy_size;
	      if (poly_int_rtx_p (copy_size_rtx, &const_copy_size))
		{
		  size = plus_constant (address_mode, size, -const_copy_size);
		  target = adjust_address (target, BLKmode, const_copy_size);
		}
	      else
		{
		  size = expand_binop (TYPE_MODE (sizetype), sub_optab, size,
				       copy_size_rtx, NULL_RTX, 0,
				       OPTAB_LIB_WIDEN);

		  if (GET_MODE (copy_size_rtx) != address_mode)
		    copy_size_rtx = convert_to_mode (address_mode,
						     copy_size_rtx,
						     TYPE_UNSIGNED (sizetype));

		  target = offset_address (target, copy_size_rtx,
					   highest_pow2_factor (copy_size));
		  label = gen_label_rtx ();
		  emit_cmp_and_jump_insns (size, const0_rtx, LT, NULL_RTX,
					   GET_MODE (size), 0, label);
		}

	      if (size != const0_rtx)
		clear_storage (target, size, BLOCK_OP_NORMAL);

	      if (label)
		emit_label (label);
	    }
	}
      /* Handle calls that return values in multiple non-contiguous locations.
	 The Irix 6 ABI has examples of this.  */
      else if (GET_CODE (target) == PARALLEL)
	{
	  if (GET_CODE (temp) == PARALLEL)
	    emit_group_move (target, temp);
	  else
	    emit_group_load (target, temp, TREE_TYPE (exp),
			     int_size_in_bytes (TREE_TYPE (exp)));
	}
      else if (GET_CODE (temp) == PARALLEL)
	emit_group_store (target, temp, TREE_TYPE (exp),
			  int_size_in_bytes (TREE_TYPE (exp)));
      else if (GET_MODE (temp) == BLKmode)
	emit_block_move (target, temp, expr_size (exp),
			 (call_param_p
			  ? BLOCK_OP_CALL_PARM : BLOCK_OP_NORMAL));
      /* If we emit a nontemporal store, there is nothing else to do.  */
      else if (nontemporal && emit_storent_insn (target, temp))
	;
      else
	{
	  if (reverse)
	    temp = flip_storage_order (GET_MODE (target), temp);
	  temp = force_operand (temp, target);
	  if (temp != target)
	    emit_move_insn (target, temp);
	}
    }

  return NULL_RTX;
}

/* Return true if field F of structure TYPE is a flexible array.  */

static bool
flexible_array_member_p (const_tree f, const_tree type)
{
  const_tree tf;

  tf = TREE_TYPE (f);
  return (DECL_CHAIN (f) == NULL
	  && TREE_CODE (tf) == ARRAY_TYPE
	  && TYPE_DOMAIN (tf)
	  && TYPE_MIN_VALUE (TYPE_DOMAIN (tf))
	  && integer_zerop (TYPE_MIN_VALUE (TYPE_DOMAIN (tf)))
	  && !TYPE_MAX_VALUE (TYPE_DOMAIN (tf))
	  && int_size_in_bytes (type) >= 0);
}

/* If FOR_CTOR_P, return the number of top-level elements that a constructor
   must have in order for it to completely initialize a value of type TYPE.
   Return -1 if the number isn't known.

   If !FOR_CTOR_P, return an estimate of the number of scalars in TYPE.  */

static HOST_WIDE_INT
count_type_elements (const_tree type, bool for_ctor_p)
{
  switch (TREE_CODE (type))
    {
    case ARRAY_TYPE:
      {
	tree nelts;

	nelts = array_type_nelts (type);
	if (nelts && tree_fits_uhwi_p (nelts))
	  {
	    unsigned HOST_WIDE_INT n;

	    n = tree_to_uhwi (nelts) + 1;
	    if (n == 0 || for_ctor_p)
	      return n;
	    else
	      return n * count_type_elements (TREE_TYPE (type), false);
	  }
	return for_ctor_p ? -1 : 1;
      }

    case RECORD_TYPE:
      {
	unsigned HOST_WIDE_INT n;
	tree f;

	n = 0;
	for (f = TYPE_FIELDS (type); f ; f = DECL_CHAIN (f))
	  if (TREE_CODE (f) == FIELD_DECL)
	    {
	      if (!for_ctor_p)
		n += count_type_elements (TREE_TYPE (f), false);
	      else if (!flexible_array_member_p (f, type))
		/* Don't count flexible arrays, which are not supposed
		   to be initialized.  */
		n += 1;
	    }

	return n;
      }

    case UNION_TYPE:
    case QUAL_UNION_TYPE:
      {
	tree f;
	HOST_WIDE_INT n, m;

	gcc_assert (!for_ctor_p);
	/* Estimate the number of scalars in each field and pick the
	   maximum.  Other estimates would do instead; the idea is simply
	   to make sure that the estimate is not sensitive to the ordering
	   of the fields.  */
	n = 1;
	for (f = TYPE_FIELDS (type); f ; f = DECL_CHAIN (f))
	  if (TREE_CODE (f) == FIELD_DECL)
	    {
	      m = count_type_elements (TREE_TYPE (f), false);
	      /* If the field doesn't span the whole union, add an extra
		 scalar for the rest.  */
	      if (simple_cst_equal (TYPE_SIZE (TREE_TYPE (f)),
				    TYPE_SIZE (type)) != 1)
		m++;
	      if (n < m)
		n = m;
	    }
	return n;
      }

    case COMPLEX_TYPE:
      return 2;

    case VECTOR_TYPE:
      {
	unsigned HOST_WIDE_INT nelts;
	if (TYPE_VECTOR_SUBPARTS (type).is_constant (&nelts))
	  return nelts;
	else
	  return -1;
      }

    case INTEGER_TYPE:
    case REAL_TYPE:
    case FIXED_POINT_TYPE:
    case ENUMERAL_TYPE:
    case BOOLEAN_TYPE:
    case POINTER_TYPE:
    case OFFSET_TYPE:
    case REFERENCE_TYPE:
    case NULLPTR_TYPE:
      return 1;

    case ERROR_MARK:
      return 0;

    case VOID_TYPE:
    case METHOD_TYPE:
    case FUNCTION_TYPE:
    case LANG_TYPE:
    default:
      gcc_unreachable ();
    }
}

/* Helper for categorize_ctor_elements.  Identical interface.  */

static bool
categorize_ctor_elements_1 (const_tree ctor, HOST_WIDE_INT *p_nz_elts,
			    HOST_WIDE_INT *p_unique_nz_elts,
			    HOST_WIDE_INT *p_init_elts, bool *p_complete)
{
  unsigned HOST_WIDE_INT idx;
  HOST_WIDE_INT nz_elts, unique_nz_elts, init_elts, num_fields;
  tree value, purpose, elt_type;

  /* Whether CTOR is a valid constant initializer, in accordance with what
     initializer_constant_valid_p does.  If inferred from the constructor
     elements, true until proven otherwise.  */
  bool const_from_elts_p = constructor_static_from_elts_p (ctor);
  bool const_p = const_from_elts_p ? true : TREE_STATIC (ctor);

  nz_elts = 0;
  unique_nz_elts = 0;
  init_elts = 0;
  num_fields = 0;
  elt_type = NULL_TREE;

  FOR_EACH_CONSTRUCTOR_ELT (CONSTRUCTOR_ELTS (ctor), idx, purpose, value)
    {
      HOST_WIDE_INT mult = 1;

      if (purpose && TREE_CODE (purpose) == RANGE_EXPR)
	{
	  tree lo_index = TREE_OPERAND (purpose, 0);
	  tree hi_index = TREE_OPERAND (purpose, 1);

	  if (tree_fits_uhwi_p (lo_index) && tree_fits_uhwi_p (hi_index))
	    mult = (tree_to_uhwi (hi_index)
		    - tree_to_uhwi (lo_index) + 1);
	}
      num_fields += mult;
      elt_type = TREE_TYPE (value);

      switch (TREE_CODE (value))
	{
	case CONSTRUCTOR:
	  {
	    HOST_WIDE_INT nz = 0, unz = 0, ic = 0;

	    bool const_elt_p = categorize_ctor_elements_1 (value, &nz, &unz,
							   &ic, p_complete);

	    nz_elts += mult * nz;
	    unique_nz_elts += unz;
 	    init_elts += mult * ic;

	    if (const_from_elts_p && const_p)
	      const_p = const_elt_p;
	  }
	  break;

	case INTEGER_CST:
	case REAL_CST:
	case FIXED_CST:
	  if (!initializer_zerop (value))
	    {
	      nz_elts += mult;
	      unique_nz_elts++;
	    }
	  init_elts += mult;
	  break;

	case STRING_CST:
	  nz_elts += mult * TREE_STRING_LENGTH (value);
	  unique_nz_elts += TREE_STRING_LENGTH (value);
	  init_elts += mult * TREE_STRING_LENGTH (value);
	  break;

	case COMPLEX_CST:
	  if (!initializer_zerop (TREE_REALPART (value)))
	    {
	      nz_elts += mult;
	      unique_nz_elts++;
	    }
	  if (!initializer_zerop (TREE_IMAGPART (value)))
	    {
	      nz_elts += mult;
	      unique_nz_elts++;
	    }
	  init_elts += 2 * mult;
	  break;

	case VECTOR_CST:
	  {
	    /* We can only construct constant-length vectors using
	       CONSTRUCTOR.  */
	    unsigned int nunits = VECTOR_CST_NELTS (value).to_constant ();
	    for (unsigned int i = 0; i < nunits; ++i)
	      {
		tree v = VECTOR_CST_ELT (value, i);
		if (!initializer_zerop (v))
		  {
		    nz_elts += mult;
		    unique_nz_elts++;
		  }
		init_elts += mult;
	      }
	  }
	  break;

	default:
	  {
	    HOST_WIDE_INT tc = count_type_elements (elt_type, false);
	    nz_elts += mult * tc;
	    unique_nz_elts += tc;
	    init_elts += mult * tc;

	    if (const_from_elts_p && const_p)
	      const_p
		= initializer_constant_valid_p (value,
						elt_type,
						TYPE_REVERSE_STORAGE_ORDER
						(TREE_TYPE (ctor)))
		  != NULL_TREE;
	  }
	  break;
	}
    }

  if (*p_complete && !complete_ctor_at_level_p (TREE_TYPE (ctor),
						num_fields, elt_type))
    *p_complete = false;

  *p_nz_elts += nz_elts;
  *p_unique_nz_elts += unique_nz_elts;
  *p_init_elts += init_elts;

  return const_p;
}

/* Examine CTOR to discover:
   * how many scalar fields are set to nonzero values,
     and place it in *P_NZ_ELTS;
   * the same, but counting RANGE_EXPRs as multiplier of 1 instead of
     high - low + 1 (this can be useful for callers to determine ctors
     that could be cheaply initialized with - perhaps nested - loops
     compared to copied from huge read-only data),
     and place it in *P_UNIQUE_NZ_ELTS;
   * how many scalar fields in total are in CTOR,
     and place it in *P_ELT_COUNT.
   * whether the constructor is complete -- in the sense that every
     meaningful byte is explicitly given a value --
     and place it in *P_COMPLETE.

   Return whether or not CTOR is a valid static constant initializer, the same
   as "initializer_constant_valid_p (CTOR, TREE_TYPE (CTOR)) != 0".  */

bool
categorize_ctor_elements (const_tree ctor, HOST_WIDE_INT *p_nz_elts,
			  HOST_WIDE_INT *p_unique_nz_elts,
			  HOST_WIDE_INT *p_init_elts, bool *p_complete)
{
  *p_nz_elts = 0;
  *p_unique_nz_elts = 0;
  *p_init_elts = 0;
  *p_complete = true;

  return categorize_ctor_elements_1 (ctor, p_nz_elts, p_unique_nz_elts,
				     p_init_elts, p_complete);
}

/* TYPE is initialized by a constructor with NUM_ELTS elements, the last
   of which had type LAST_TYPE.  Each element was itself a complete
   initializer, in the sense that every meaningful byte was explicitly
   given a value.  Return true if the same is true for the constructor
   as a whole.  */

bool
complete_ctor_at_level_p (const_tree type, HOST_WIDE_INT num_elts,
			  const_tree last_type)
{
  if (TREE_CODE (type) == UNION_TYPE
      || TREE_CODE (type) == QUAL_UNION_TYPE)
    {
      if (num_elts == 0)
	return false;

      gcc_assert (num_elts == 1 && last_type);

      /* ??? We could look at each element of the union, and find the
	 largest element.  Which would avoid comparing the size of the
	 initialized element against any tail padding in the union.
	 Doesn't seem worth the effort...  */
      return simple_cst_equal (TYPE_SIZE (type), TYPE_SIZE (last_type)) == 1;
    }

  return count_type_elements (type, true) == num_elts;
}

/* Return 1 if EXP contains mostly (3/4) zeros.  */

static int
mostly_zeros_p (const_tree exp)
{
  if (TREE_CODE (exp) == CONSTRUCTOR)
    {
      HOST_WIDE_INT nz_elts, unz_elts, init_elts;
      bool complete_p;

      categorize_ctor_elements (exp, &nz_elts, &unz_elts, &init_elts,
				&complete_p);
      return !complete_p || nz_elts < init_elts / 4;
    }

  return initializer_zerop (exp);
}

/* Return 1 if EXP contains all zeros.  */

static int
all_zeros_p (const_tree exp)
{
  if (TREE_CODE (exp) == CONSTRUCTOR)
    {
      HOST_WIDE_INT nz_elts, unz_elts, init_elts;
      bool complete_p;

      categorize_ctor_elements (exp, &nz_elts, &unz_elts, &init_elts,
				&complete_p);
      return nz_elts == 0;
    }

  return initializer_zerop (exp);
}

/* Helper function for store_constructor.
   TARGET, BITSIZE, BITPOS, MODE, EXP are as for store_field.
   CLEARED is as for store_constructor.
   ALIAS_SET is the alias set to use for any stores.
   If REVERSE is true, the store is to be done in reverse order.

   This provides a recursive shortcut back to store_constructor when it isn't
   necessary to go through store_field.  This is so that we can pass through
   the cleared field to let store_constructor know that we may not have to
   clear a substructure if the outer structure has already been cleared.  */

static void
store_constructor_field (rtx target, poly_uint64 bitsize, poly_int64 bitpos,
			 poly_uint64 bitregion_start,
			 poly_uint64 bitregion_end,
			 machine_mode mode,
			 tree exp, int cleared,
			 alias_set_type alias_set, bool reverse)
{
  poly_int64 bytepos;
  poly_uint64 bytesize;
  if (TREE_CODE (exp) == CONSTRUCTOR
      /* We can only call store_constructor recursively if the size and
	 bit position are on a byte boundary.  */
      && multiple_p (bitpos, BITS_PER_UNIT, &bytepos)
      && maybe_ne (bitsize, 0U)
      && multiple_p (bitsize, BITS_PER_UNIT, &bytesize)
      /* If we have a nonzero bitpos for a register target, then we just
	 let store_field do the bitfield handling.  This is unlikely to
	 generate unnecessary clear instructions anyways.  */
      && (known_eq (bitpos, 0) || MEM_P (target)))
    {
      if (MEM_P (target))
	{
	  machine_mode target_mode = GET_MODE (target);
	  if (target_mode != BLKmode
	      && !multiple_p (bitpos, GET_MODE_ALIGNMENT (target_mode)))
	    target_mode = BLKmode;
	  target = adjust_address (target, target_mode, bytepos);
	}


      /* Update the alias set, if required.  */
      if (MEM_P (target) && ! MEM_KEEP_ALIAS_SET_P (target)
	  && MEM_ALIAS_SET (target) != 0)
	{
	  target = copy_rtx (target);
	  set_mem_alias_set (target, alias_set);
	}

      store_constructor (exp, target, cleared, bytesize, reverse);
    }
  else
    store_field (target, bitsize, bitpos, bitregion_start, bitregion_end, mode,
		 exp, alias_set, false, reverse);
}


/* Returns the number of FIELD_DECLs in TYPE.  */

static int
fields_length (const_tree type)
{
  tree t = TYPE_FIELDS (type);
  int count = 0;

  for (; t; t = DECL_CHAIN (t))
    if (TREE_CODE (t) == FIELD_DECL)
      ++count;

  return count;
}


/* Store the value of constructor EXP into the rtx TARGET.
   TARGET is either a REG or a MEM; we know it cannot conflict, since
   safe_from_p has been called.
   CLEARED is true if TARGET is known to have been zero'd.
   SIZE is the number of bytes of TARGET we are allowed to modify: this
   may not be the same as the size of EXP if we are assigning to a field
   which has been packed to exclude padding bits.
   If REVERSE is true, the store is to be done in reverse order.  */

static void
store_constructor (tree exp, rtx target, int cleared, poly_int64 size,
		   bool reverse)
{
  tree type = TREE_TYPE (exp);
  HOST_WIDE_INT exp_size = int_size_in_bytes (type);
  poly_int64 bitregion_end = known_gt (size, 0) ? size * BITS_PER_UNIT - 1 : 0;

  switch (TREE_CODE (type))
    {
    case RECORD_TYPE:
    case UNION_TYPE:
    case QUAL_UNION_TYPE:
      {
	unsigned HOST_WIDE_INT idx;
	tree field, value;

	/* The storage order is specified for every aggregate type.  */
	reverse = TYPE_REVERSE_STORAGE_ORDER (type);

	/* If size is zero or the target is already cleared, do nothing.  */
	if (known_eq (size, 0) || cleared)
	  cleared = 1;
	/* We either clear the aggregate or indicate the value is dead.  */
	else if ((TREE_CODE (type) == UNION_TYPE
		  || TREE_CODE (type) == QUAL_UNION_TYPE)
		 && ! CONSTRUCTOR_ELTS (exp))
	  /* If the constructor is empty, clear the union.  */
	  {
	    clear_storage (target, expr_size (exp), BLOCK_OP_NORMAL);
	    cleared = 1;
	  }

	/* If we are building a static constructor into a register,
	   set the initial value as zero so we can fold the value into
	   a constant.  But if more than one register is involved,
	   this probably loses.  */
	else if (REG_P (target) && TREE_STATIC (exp)
		 && known_le (GET_MODE_SIZE (GET_MODE (target)),
			      REGMODE_NATURAL_SIZE (GET_MODE (target))))
	  {
	    emit_move_insn (target, CONST0_RTX (GET_MODE (target)));
	    cleared = 1;
	  }

        /* If the constructor has fewer fields than the structure or
	   if we are initializing the structure to mostly zeros, clear
	   the whole structure first.  Don't do this if TARGET is a
	   register whose mode size isn't equal to SIZE since
	   clear_storage can't handle this case.  */
	else if (known_size_p (size)
		 && (((int) CONSTRUCTOR_NELTS (exp) != fields_length (type))
		     || mostly_zeros_p (exp))
		 && (!REG_P (target)
		     || known_eq (GET_MODE_SIZE (GET_MODE (target)), size)))
	  {
	    clear_storage (target, gen_int_mode (size, Pmode),
			   BLOCK_OP_NORMAL);
	    cleared = 1;
	  }

	if (REG_P (target) && !cleared)
	  emit_clobber (target);

	/* Store each element of the constructor into the
	   corresponding field of TARGET.  */
	FOR_EACH_CONSTRUCTOR_ELT (CONSTRUCTOR_ELTS (exp), idx, field, value)
	  {
	    machine_mode mode;
	    HOST_WIDE_INT bitsize;
	    HOST_WIDE_INT bitpos = 0;
	    tree offset;
	    rtx to_rtx = target;

	    /* Just ignore missing fields.  We cleared the whole
	       structure, above, if any fields are missing.  */
	    if (field == 0)
	      continue;

	    if (cleared && initializer_zerop (value))
	      continue;

	    if (tree_fits_uhwi_p (DECL_SIZE (field)))
	      bitsize = tree_to_uhwi (DECL_SIZE (field));
	    else
	      gcc_unreachable ();

	    mode = DECL_MODE (field);
	    if (DECL_BIT_FIELD (field))
	      mode = VOIDmode;

	    offset = DECL_FIELD_OFFSET (field);
	    if (tree_fits_shwi_p (offset)
		&& tree_fits_shwi_p (bit_position (field)))
	      {
		bitpos = int_bit_position (field);
		offset = NULL_TREE;
	      }
	    else
	      gcc_unreachable ();

	    /* If this initializes a field that is smaller than a
	       word, at the start of a word, try to widen it to a full
	       word.  This special case allows us to output C++ member
	       function initializations in a form that the optimizers
	       can understand.  */
	    if (WORD_REGISTER_OPERATIONS
		&& REG_P (target)
		&& bitsize < BITS_PER_WORD
		&& bitpos % BITS_PER_WORD == 0
		&& GET_MODE_CLASS (mode) == MODE_INT
		&& TREE_CODE (value) == INTEGER_CST
		&& exp_size >= 0
		&& bitpos + BITS_PER_WORD <= exp_size * BITS_PER_UNIT)
	      {
		type = TREE_TYPE (value);

		if (TYPE_PRECISION (type) < BITS_PER_WORD)
		  {
		    type = lang_hooks.types.type_for_mode
		      (word_mode, TYPE_UNSIGNED (type));
		    value = fold_convert (type, value);
		    /* Make sure the bits beyond the original bitsize are zero
		       so that we can correctly avoid extra zeroing stores in
		       later constructor elements.  */
		    tree bitsize_mask
		      = wide_int_to_tree (type, wi::mask (bitsize, false,
							   BITS_PER_WORD));
		    value = fold_build2 (BIT_AND_EXPR, type, value, bitsize_mask);
		  }

		if (BYTES_BIG_ENDIAN)
		  value
		   = fold_build2 (LSHIFT_EXPR, type, value,
				   build_int_cst (type,
						  BITS_PER_WORD - bitsize));
		bitsize = BITS_PER_WORD;
		mode = word_mode;
	      }

	    if (MEM_P (to_rtx) && !MEM_KEEP_ALIAS_SET_P (to_rtx)
		&& DECL_NONADDRESSABLE_P (field))
	      {
		to_rtx = copy_rtx (to_rtx);
		MEM_KEEP_ALIAS_SET_P (to_rtx) = 1;
	      }

	    store_constructor_field (to_rtx, bitsize, bitpos,
				     0, bitregion_end, mode,
				     value, cleared,
				     get_alias_set (TREE_TYPE (field)),
				     reverse);
	  }
	break;
      }
    case ARRAY_TYPE:
      {
	tree value, index;
	unsigned HOST_WIDE_INT i;
	int need_to_clear;
	tree domain;
	tree elttype = TREE_TYPE (type);
	int const_bounds_p;
	HOST_WIDE_INT minelt = 0;
	HOST_WIDE_INT maxelt = 0;

	/* The storage order is specified for every aggregate type.  */
	reverse = TYPE_REVERSE_STORAGE_ORDER (type);

	domain = TYPE_DOMAIN (type);
	const_bounds_p = (TYPE_MIN_VALUE (domain)
			  && TYPE_MAX_VALUE (domain)
			  && tree_fits_shwi_p (TYPE_MIN_VALUE (domain))
			  && tree_fits_shwi_p (TYPE_MAX_VALUE (domain)));

	/* If we have constant bounds for the range of the type, get them.  */
	if (const_bounds_p)
	  {
	    minelt = tree_to_shwi (TYPE_MIN_VALUE (domain));
	    maxelt = tree_to_shwi (TYPE_MAX_VALUE (domain));
	  }

	/* If the constructor has fewer elements than the array, clear
           the whole array first.  Similarly if this is static
           constructor of a non-BLKmode object.  */
	if (cleared)
	  need_to_clear = 0;
	else if (REG_P (target) && TREE_STATIC (exp))
	  need_to_clear = 1;
	else
	  {
	    unsigned HOST_WIDE_INT idx;
	    HOST_WIDE_INT count = 0, zero_count = 0;
	    need_to_clear = ! const_bounds_p;

	    /* This loop is a more accurate version of the loop in
	       mostly_zeros_p (it handles RANGE_EXPR in an index).  It
	       is also needed to check for missing elements.  */
	    FOR_EACH_CONSTRUCTOR_ELT (CONSTRUCTOR_ELTS (exp), idx, index, value)
	      {
		HOST_WIDE_INT this_node_count;

		if (need_to_clear)
		  break;

		if (index != NULL_TREE && TREE_CODE (index) == RANGE_EXPR)
		  {
		    tree lo_index = TREE_OPERAND (index, 0);
		    tree hi_index = TREE_OPERAND (index, 1);

		    if (! tree_fits_uhwi_p (lo_index)
			|| ! tree_fits_uhwi_p (hi_index))
		      {
			need_to_clear = 1;
			break;
		      }

		    this_node_count = (tree_to_uhwi (hi_index)
				       - tree_to_uhwi (lo_index) + 1);
		  }
		else
		  this_node_count = 1;

		count += this_node_count;
		if (mostly_zeros_p (value))
		  zero_count += this_node_count;
	      }

	    /* Clear the entire array first if there are any missing
	       elements, or if the incidence of zero elements is >=
	       75%.  */
	    if (! need_to_clear
		&& (count < maxelt - minelt + 1
		    || 4 * zero_count >= 3 * count))
	      need_to_clear = 1;
	  }

	if (need_to_clear && maybe_gt (size, 0))
	  {
	    if (REG_P (target))
	      emit_move_insn (target, CONST0_RTX (GET_MODE (target)));
	    else
	      clear_storage (target, gen_int_mode (size, Pmode),
			     BLOCK_OP_NORMAL);
	    cleared = 1;
	  }

	if (!cleared && REG_P (target))
	  /* Inform later passes that the old value is dead.  */
	  emit_clobber (target);

	/* Store each element of the constructor into the
	   corresponding element of TARGET, determined by counting the
	   elements.  */
	FOR_EACH_CONSTRUCTOR_ELT (CONSTRUCTOR_ELTS (exp), i, index, value)
	  {
	    machine_mode mode;
	    poly_int64 bitsize;
	    HOST_WIDE_INT bitpos;
	    rtx xtarget = target;

	    if (cleared && initializer_zerop (value))
	      continue;

	    mode = TYPE_MODE (elttype);
	    if (mode != BLKmode)
	      bitsize = GET_MODE_BITSIZE (mode);
	    else if (!poly_int_tree_p (TYPE_SIZE (elttype), &bitsize))
	      bitsize = -1;

	    if (index != NULL_TREE && TREE_CODE (index) == RANGE_EXPR)
	      {
		tree lo_index = TREE_OPERAND (index, 0);
		tree hi_index = TREE_OPERAND (index, 1);
		rtx index_r, pos_rtx;
		HOST_WIDE_INT lo, hi, count;
		tree position;

		/* If the range is constant and "small", unroll the loop.  */
		if (const_bounds_p
		    && tree_fits_shwi_p (lo_index)
		    && tree_fits_shwi_p (hi_index)
		    && (lo = tree_to_shwi (lo_index),
			hi = tree_to_shwi (hi_index),
			count = hi - lo + 1,
			(!MEM_P (target)
			 || count <= 2
			 || (tree_fits_uhwi_p (TYPE_SIZE (elttype))
			     && (tree_to_uhwi (TYPE_SIZE (elttype)) * count
				 <= 40 * 8)))))
		  {
		    lo -= minelt;  hi -= minelt;
		    for (; lo <= hi; lo++)
		      {
			bitpos = lo * tree_to_shwi (TYPE_SIZE (elttype));

			if (MEM_P (target)
			    && !MEM_KEEP_ALIAS_SET_P (target)
			    && TREE_CODE (type) == ARRAY_TYPE
			    && TYPE_NONALIASED_COMPONENT (type))
			  {
			    target = copy_rtx (target);
			    MEM_KEEP_ALIAS_SET_P (target) = 1;
			  }

			store_constructor_field
			  (target, bitsize, bitpos, 0, bitregion_end,
			   mode, value, cleared,
			   get_alias_set (elttype), reverse);
		      }
		  }
		else
		  {
		    rtx_code_label *loop_start = gen_label_rtx ();
		    rtx_code_label *loop_end = gen_label_rtx ();
		    tree exit_cond;

		    expand_normal (hi_index);

		    index = build_decl (EXPR_LOCATION (exp),
					VAR_DECL, NULL_TREE, domain);
		    index_r = gen_reg_rtx (promote_decl_mode (index, NULL));
		    SET_DECL_RTL (index, index_r);
		    store_expr (lo_index, index_r, 0, false, reverse);

		    /* Build the head of the loop.  */
		    do_pending_stack_adjust ();
		    emit_label (loop_start);

		    /* Assign value to element index.  */
		    position =
		      fold_convert (ssizetype,
				    fold_build2 (MINUS_EXPR,
						 TREE_TYPE (index),
						 index,
						 TYPE_MIN_VALUE (domain)));

		    position =
			size_binop (MULT_EXPR, position,
				    fold_convert (ssizetype,
						  TYPE_SIZE_UNIT (elttype)));

		    pos_rtx = expand_normal (position);
		    xtarget = offset_address (target, pos_rtx,
					      highest_pow2_factor (position));
		    xtarget = adjust_address (xtarget, mode, 0);
		    if (TREE_CODE (value) == CONSTRUCTOR)
		      store_constructor (value, xtarget, cleared,
					 exact_div (bitsize, BITS_PER_UNIT),
					 reverse);
		    else
		      store_expr (value, xtarget, 0, false, reverse);

		    /* Generate a conditional jump to exit the loop.  */
		    exit_cond = build2 (LT_EXPR, integer_type_node,
					index, hi_index);
		    jumpif (exit_cond, loop_end,
			    profile_probability::uninitialized ());

		    /* Update the loop counter, and jump to the head of
		       the loop.  */
		    expand_assignment (index,
				       build2 (PLUS_EXPR, TREE_TYPE (index),
					       index, integer_one_node),
				       false);

		    emit_jump (loop_start);

		    /* Build the end of the loop.  */
		    emit_label (loop_end);
		  }
	      }
	    else if ((index != 0 && ! tree_fits_shwi_p (index))
		     || ! tree_fits_uhwi_p (TYPE_SIZE (elttype)))
	      {
		tree position;

		if (index == 0)
		  index = ssize_int (1);

		if (minelt)
		  index = fold_convert (ssizetype,
					fold_build2 (MINUS_EXPR,
						     TREE_TYPE (index),
						     index,
						     TYPE_MIN_VALUE (domain)));

		position =
		  size_binop (MULT_EXPR, index,
			      fold_convert (ssizetype,
					    TYPE_SIZE_UNIT (elttype)));
		xtarget = offset_address (target,
					  expand_normal (position),
					  highest_pow2_factor (position));
		xtarget = adjust_address (xtarget, mode, 0);
		store_expr (value, xtarget, 0, false, reverse);
	      }
	    else
	      {
		if (index != 0)
		  bitpos = ((tree_to_shwi (index) - minelt)
			    * tree_to_uhwi (TYPE_SIZE (elttype)));
		else
		  bitpos = (i * tree_to_uhwi (TYPE_SIZE (elttype)));

		if (MEM_P (target) && !MEM_KEEP_ALIAS_SET_P (target)
		    && TREE_CODE (type) == ARRAY_TYPE
		    && TYPE_NONALIASED_COMPONENT (type))
		  {
		    target = copy_rtx (target);
		    MEM_KEEP_ALIAS_SET_P (target) = 1;
		  }
		store_constructor_field (target, bitsize, bitpos, 0,
					 bitregion_end, mode, value,
					 cleared, get_alias_set (elttype),
					 reverse);
	      }
	  }
	break;
      }

    case VECTOR_TYPE:
      {
	unsigned HOST_WIDE_INT idx;
	constructor_elt *ce;
	int i;
	int need_to_clear;
	insn_code icode = CODE_FOR_nothing;
	tree elt;
	tree elttype = TREE_TYPE (type);
	int elt_size = tree_to_uhwi (TYPE_SIZE (elttype));
	machine_mode eltmode = TYPE_MODE (elttype);
	HOST_WIDE_INT bitsize;
	HOST_WIDE_INT bitpos;
	rtvec vector = NULL;
	poly_uint64 n_elts;
	unsigned HOST_WIDE_INT const_n_elts;
	alias_set_type alias;
	bool vec_vec_init_p = false;
	machine_mode mode = GET_MODE (target);

	gcc_assert (eltmode != BLKmode);

	/* Try using vec_duplicate_optab for uniform vectors.  */
	if (!TREE_SIDE_EFFECTS (exp)
	    && VECTOR_MODE_P (mode)
	    && eltmode == GET_MODE_INNER (mode)
	    && ((icode = optab_handler (vec_duplicate_optab, mode))
		!= CODE_FOR_nothing)
	    && (elt = uniform_vector_p (exp)))
	  {
	    class expand_operand ops[2];
	    create_output_operand (&ops[0], target, mode);
	    create_input_operand (&ops[1], expand_normal (elt), eltmode);
	    expand_insn (icode, 2, ops);
	    if (!rtx_equal_p (target, ops[0].value))
	      emit_move_insn (target, ops[0].value);
	    break;
	  }

	n_elts = TYPE_VECTOR_SUBPARTS (type);
	if (REG_P (target)
	    && VECTOR_MODE_P (mode)
	    && n_elts.is_constant (&const_n_elts))
	  {
	    machine_mode emode = eltmode;
	    bool vector_typed_elts_p = false;

	    if (CONSTRUCTOR_NELTS (exp)
		&& (TREE_CODE (TREE_TYPE (CONSTRUCTOR_ELT (exp, 0)->value))
		    == VECTOR_TYPE))
	      {
		tree etype = TREE_TYPE (CONSTRUCTOR_ELT (exp, 0)->value);
		gcc_assert (known_eq (CONSTRUCTOR_NELTS (exp)
				      * TYPE_VECTOR_SUBPARTS (etype),
				      n_elts));
		emode = TYPE_MODE (etype);
		vector_typed_elts_p = true;
	      }
	    icode = convert_optab_handler (vec_init_optab, mode, emode);
	    if (icode != CODE_FOR_nothing)
	      {
		unsigned int n = const_n_elts;

		if (vector_typed_elts_p)
		  {
		    n = CONSTRUCTOR_NELTS (exp);
		    vec_vec_init_p = true;
		  }
		vector = rtvec_alloc (n);
		for (unsigned int k = 0; k < n; k++)
		  RTVEC_ELT (vector, k) = CONST0_RTX (emode);
	      }
	  }

	/* If the constructor has fewer elements than the vector,
	   clear the whole array first.  Similarly if this is static
	   constructor of a non-BLKmode object.  */
	if (cleared)
	  need_to_clear = 0;
	else if (REG_P (target) && TREE_STATIC (exp))
	  need_to_clear = 1;
	else
	  {
	    unsigned HOST_WIDE_INT count = 0, zero_count = 0;
	    tree value;

	    FOR_EACH_CONSTRUCTOR_VALUE (CONSTRUCTOR_ELTS (exp), idx, value)
	      {
		tree sz = TYPE_SIZE (TREE_TYPE (value));
		int n_elts_here
		  = tree_to_uhwi (int_const_binop (TRUNC_DIV_EXPR, sz,
						   TYPE_SIZE (elttype)));

		count += n_elts_here;
		if (mostly_zeros_p (value))
		  zero_count += n_elts_here;
	      }

	    /* Clear the entire vector first if there are any missing elements,
	       or if the incidence of zero elements is >= 75%.  */
	    need_to_clear = (maybe_lt (count, n_elts)
			     || 4 * zero_count >= 3 * count);
	  }

	if (need_to_clear && maybe_gt (size, 0) && !vector)
	  {
	    if (REG_P (target))
	      emit_move_insn (target, CONST0_RTX (mode));
	    else
	      clear_storage (target, gen_int_mode (size, Pmode),
			     BLOCK_OP_NORMAL);
	    cleared = 1;
	  }

	/* Inform later passes that the old value is dead.  */
	if (!cleared && !vector && REG_P (target))
	  emit_move_insn (target, CONST0_RTX (mode));

        if (MEM_P (target))
	  alias = MEM_ALIAS_SET (target);
	else
	  alias = get_alias_set (elttype);

        /* Store each element of the constructor into the corresponding
	   element of TARGET, determined by counting the elements.  */
	for (idx = 0, i = 0;
	     vec_safe_iterate (CONSTRUCTOR_ELTS (exp), idx, &ce);
	     idx++, i += bitsize / elt_size)
	  {
	    HOST_WIDE_INT eltpos;
	    tree value = ce->value;

	    bitsize = tree_to_uhwi (TYPE_SIZE (TREE_TYPE (value)));
	    if (cleared && initializer_zerop (value))
	      continue;

	    if (ce->index)
	      eltpos = tree_to_uhwi (ce->index);
	    else
	      eltpos = i;

	    if (vector)
	      {
		if (vec_vec_init_p)
		  {
		    gcc_assert (ce->index == NULL_TREE);
		    gcc_assert (TREE_CODE (TREE_TYPE (value)) == VECTOR_TYPE);
		    eltpos = idx;
		  }
		else
		  gcc_assert (TREE_CODE (TREE_TYPE (value)) != VECTOR_TYPE);
		RTVEC_ELT (vector, eltpos) = expand_normal (value);
	      }
	    else
	      {
		machine_mode value_mode
		  = (TREE_CODE (TREE_TYPE (value)) == VECTOR_TYPE
		     ? TYPE_MODE (TREE_TYPE (value)) : eltmode);
		bitpos = eltpos * elt_size;
		store_constructor_field (target, bitsize, bitpos, 0,
					 bitregion_end, value_mode,
					 value, cleared, alias, reverse);
	      }
	  }

	if (vector)
	  emit_insn (GEN_FCN (icode) (target,
				      gen_rtx_PARALLEL (mode, vector)));
	break;
      }

    default:
      gcc_unreachable ();
    }
}

/* Store the value of EXP (an expression tree)
   into a subfield of TARGET which has mode MODE and occupies
   BITSIZE bits, starting BITPOS bits from the start of TARGET.
   If MODE is VOIDmode, it means that we are storing into a bit-field.

   BITREGION_START is bitpos of the first bitfield in this region.
   BITREGION_END is the bitpos of the ending bitfield in this region.
   These two fields are 0, if the C++ memory model does not apply,
   or we are not interested in keeping track of bitfield regions.

   Always return const0_rtx unless we have something particular to
   return.

   ALIAS_SET is the alias set for the destination.  This value will
   (in general) be different from that for TARGET, since TARGET is a
   reference to the containing structure.

   If NONTEMPORAL is true, try generating a nontemporal store.

   If REVERSE is true, the store is to be done in reverse order.  */

static rtx
store_field (rtx target, poly_int64 bitsize, poly_int64 bitpos,
	     poly_uint64 bitregion_start, poly_uint64 bitregion_end,
	     machine_mode mode, tree exp,
	     alias_set_type alias_set, bool nontemporal,  bool reverse)
{
  if (TREE_CODE (exp) == ERROR_MARK)
    return const0_rtx;

  /* If we have nothing to store, do nothing unless the expression has
     side-effects.  Don't do that for zero sized addressable lhs of
     calls.  */
  if (known_eq (bitsize, 0)
      && (!TREE_ADDRESSABLE (TREE_TYPE (exp))
	  || TREE_CODE (exp) != CALL_EXPR))
    return expand_expr (exp, const0_rtx, VOIDmode, EXPAND_NORMAL);

  if (GET_CODE (target) == CONCAT)
    {
      /* We're storing into a struct containing a single __complex.  */

      gcc_assert (known_eq (bitpos, 0));
      return store_expr (exp, target, 0, nontemporal, reverse);
    }

  /* If the structure is in a register or if the component
     is a bit field, we cannot use addressing to access it.
     Use bit-field techniques or SUBREG to store in it.  */

  poly_int64 decl_bitsize;
  if (mode == VOIDmode
      || (mode != BLKmode && ! direct_store[(int) mode]
	  && GET_MODE_CLASS (mode) != MODE_COMPLEX_INT
	  && GET_MODE_CLASS (mode) != MODE_COMPLEX_FLOAT)
      || REG_P (target)
      || GET_CODE (target) == SUBREG
      /* If the field isn't aligned enough to store as an ordinary memref,
	 store it as a bit field.  */
      || (mode != BLKmode
	  && ((((MEM_ALIGN (target) < GET_MODE_ALIGNMENT (mode))
		|| !multiple_p (bitpos, GET_MODE_ALIGNMENT (mode)))
	       && targetm.slow_unaligned_access (mode, MEM_ALIGN (target)))
	      || !multiple_p (bitpos, BITS_PER_UNIT)))
      || (known_size_p (bitsize)
	  && mode != BLKmode
	  && maybe_gt (GET_MODE_BITSIZE (mode), bitsize))
      /* If the RHS and field are a constant size and the size of the
	 RHS isn't the same size as the bitfield, we must use bitfield
	 operations.  */
      || (known_size_p (bitsize)
	  && poly_int_tree_p (TYPE_SIZE (TREE_TYPE (exp)))
	  && maybe_ne (wi::to_poly_offset (TYPE_SIZE (TREE_TYPE (exp))),
		       bitsize)
	  /* Except for initialization of full bytes from a CONSTRUCTOR, which
	     we will handle specially below.  */
	  && !(TREE_CODE (exp) == CONSTRUCTOR
	       && multiple_p (bitsize, BITS_PER_UNIT))
	  /* And except for bitwise copying of TREE_ADDRESSABLE types,
	     where the FIELD_DECL has the right bitsize, but TREE_TYPE (exp)
	     includes some extra padding.  store_expr / expand_expr will in
	     that case call get_inner_reference that will have the bitsize
	     we check here and thus the block move will not clobber the
	     padding that shouldn't be clobbered.  In the future we could
	     replace the TREE_ADDRESSABLE check with a check that
	     get_base_address needs to live in memory.  */
	  && (!TREE_ADDRESSABLE (TREE_TYPE (exp))
	      || TREE_CODE (exp) != COMPONENT_REF
	      || !multiple_p (bitsize, BITS_PER_UNIT)
	      || !multiple_p (bitpos, BITS_PER_UNIT)
	      || !poly_int_tree_p (DECL_SIZE (TREE_OPERAND (exp, 1)),
				   &decl_bitsize)
	      || maybe_ne (decl_bitsize, bitsize)))
      /* If we are expanding a MEM_REF of a non-BLKmode non-addressable
         decl we must use bitfield operations.  */
      || (known_size_p (bitsize)
	  && TREE_CODE (exp) == MEM_REF
	  && TREE_CODE (TREE_OPERAND (exp, 0)) == ADDR_EXPR
	  && DECL_P (TREE_OPERAND (TREE_OPERAND (exp, 0), 0))
	  && !TREE_ADDRESSABLE (TREE_OPERAND (TREE_OPERAND (exp, 0), 0))
	  && DECL_MODE (TREE_OPERAND (TREE_OPERAND (exp, 0), 0)) != BLKmode))
    {
      rtx temp;
      gimple *nop_def;

      /* If EXP is a NOP_EXPR of precision less than its mode, then that
	 implies a mask operation.  If the precision is the same size as
	 the field we're storing into, that mask is redundant.  This is
	 particularly common with bit field assignments generated by the
	 C front end.  */
      nop_def = get_def_for_expr (exp, NOP_EXPR);
      if (nop_def)
	{
	  tree type = TREE_TYPE (exp);
	  if (INTEGRAL_TYPE_P (type)
	      && maybe_ne (TYPE_PRECISION (type),
			   GET_MODE_BITSIZE (TYPE_MODE (type)))
	      && known_eq (bitsize, TYPE_PRECISION (type)))
	    {
	      tree op = gimple_assign_rhs1 (nop_def);
	      type = TREE_TYPE (op);
	      if (INTEGRAL_TYPE_P (type)
		  && known_ge (TYPE_PRECISION (type), bitsize))
		exp = op;
	    }
	}

      temp = expand_normal (exp);

      /* We don't support variable-sized BLKmode bitfields, since our
	 handling of BLKmode is bound up with the ability to break
	 things into words.  */
      gcc_assert (mode != BLKmode || bitsize.is_constant ());

      /* Handle calls that return values in multiple non-contiguous locations.
	 The Irix 6 ABI has examples of this.  */
      if (GET_CODE (temp) == PARALLEL)
	{
	  HOST_WIDE_INT size = int_size_in_bytes (TREE_TYPE (exp));
	  machine_mode temp_mode = GET_MODE (temp);
	  if (temp_mode == BLKmode || temp_mode == VOIDmode)
	    temp_mode = smallest_int_mode_for_size (size * BITS_PER_UNIT);
	  rtx temp_target = gen_reg_rtx (temp_mode);
	  emit_group_store (temp_target, temp, TREE_TYPE (exp), size);
	  temp = temp_target;
	}

      /* Handle calls that return BLKmode values in registers.  */
      else if (mode == BLKmode && REG_P (temp) && TREE_CODE (exp) == CALL_EXPR)
	{
	  rtx temp_target = gen_reg_rtx (GET_MODE (temp));
	  copy_blkmode_from_reg (temp_target, temp, TREE_TYPE (exp));
	  temp = temp_target;
	}

      /* If the value has aggregate type and an integral mode then, if BITSIZE
	 is narrower than this mode and this is for big-endian data, we first
	 need to put the value into the low-order bits for store_bit_field,
	 except when MODE is BLKmode and BITSIZE larger than the word size
	 (see the handling of fields larger than a word in store_bit_field).
	 Moreover, the field may be not aligned on a byte boundary; in this
	 case, if it has reverse storage order, it needs to be accessed as a
	 scalar field with reverse storage order and we must first put the
	 value into target order.  */
      scalar_int_mode temp_mode;
      if (AGGREGATE_TYPE_P (TREE_TYPE (exp))
	  && is_int_mode (GET_MODE (temp), &temp_mode))
	{
	  HOST_WIDE_INT size = GET_MODE_BITSIZE (temp_mode);

	  reverse = TYPE_REVERSE_STORAGE_ORDER (TREE_TYPE (exp));

	  if (reverse)
	    temp = flip_storage_order (temp_mode, temp);

	  gcc_checking_assert (known_le (bitsize, size));
	  if (maybe_lt (bitsize, size)
	      && reverse ? !BYTES_BIG_ENDIAN : BYTES_BIG_ENDIAN
	      /* Use of to_constant for BLKmode was checked above.  */
	      && !(mode == BLKmode && bitsize.to_constant () > BITS_PER_WORD))
	    temp = expand_shift (RSHIFT_EXPR, temp_mode, temp,
				 size - bitsize, NULL_RTX, 1);
	}

      /* Unless MODE is VOIDmode or BLKmode, convert TEMP to MODE.  */
      if (mode != VOIDmode && mode != BLKmode
	  && mode != TYPE_MODE (TREE_TYPE (exp)))
	temp = convert_modes (mode, TYPE_MODE (TREE_TYPE (exp)), temp, 1);

      /* If the mode of TEMP and TARGET is BLKmode, both must be in memory
	 and BITPOS must be aligned on a byte boundary.  If so, we simply do
	 a block copy.  Likewise for a BLKmode-like TARGET.  */
      if (GET_MODE (temp) == BLKmode
	  && (GET_MODE (target) == BLKmode
	      || (MEM_P (target)
		  && GET_MODE_CLASS (GET_MODE (target)) == MODE_INT
		  && multiple_p (bitpos, BITS_PER_UNIT)
		  && multiple_p (bitsize, BITS_PER_UNIT))))
	{
	  gcc_assert (MEM_P (target) && MEM_P (temp));
	  poly_int64 bytepos = exact_div (bitpos, BITS_PER_UNIT);
	  poly_int64 bytesize = bits_to_bytes_round_up (bitsize);

	  target = adjust_address (target, VOIDmode, bytepos);
	  emit_block_move (target, temp,
			   gen_int_mode (bytesize, Pmode),
			   BLOCK_OP_NORMAL);

	  return const0_rtx;
	}

      /* If the mode of TEMP is still BLKmode and BITSIZE not larger than the
	 word size, we need to load the value (see again store_bit_field).  */
      if (GET_MODE (temp) == BLKmode && known_le (bitsize, BITS_PER_WORD))
	{
	  temp_mode = smallest_int_mode_for_size (bitsize);
	  temp = extract_bit_field (temp, bitsize, 0, 1, NULL_RTX, temp_mode,
				    temp_mode, false, NULL);
	}

      /* Store the value in the bitfield.  */
      gcc_checking_assert (known_ge (bitpos, 0));
      store_bit_field (target, bitsize, bitpos,
		       bitregion_start, bitregion_end,
		       mode, temp, reverse);

      return const0_rtx;
    }
  else
    {
      /* Now build a reference to just the desired component.  */
      rtx to_rtx = adjust_address (target, mode,
				   exact_div (bitpos, BITS_PER_UNIT));

      if (to_rtx == target)
	to_rtx = copy_rtx (to_rtx);

      if (!MEM_KEEP_ALIAS_SET_P (to_rtx) && MEM_ALIAS_SET (to_rtx) != 0)
	set_mem_alias_set (to_rtx, alias_set);

      /* Above we avoided using bitfield operations for storing a CONSTRUCTOR
	 into a target smaller than its type; handle that case now.  */
      if (TREE_CODE (exp) == CONSTRUCTOR && known_size_p (bitsize))
	{
	  poly_int64 bytesize = exact_div (bitsize, BITS_PER_UNIT);
	  store_constructor (exp, to_rtx, 0, bytesize, reverse);
	  return to_rtx;
	}

      return store_expr (exp, to_rtx, 0, nontemporal, reverse);
    }
}

/* Given an expression EXP that may be a COMPONENT_REF, a BIT_FIELD_REF,
   an ARRAY_REF, or an ARRAY_RANGE_REF, look for nested operations of these
   codes and find the ultimate containing object, which we return.

   We set *PBITSIZE to the size in bits that we want, *PBITPOS to the
   bit position, *PUNSIGNEDP to the signedness and *PREVERSEP to the
   storage order of the field.
   If the position of the field is variable, we store a tree
   giving the variable offset (in units) in *POFFSET.
   This offset is in addition to the bit position.
   If the position is not variable, we store 0 in *POFFSET.

   If any of the extraction expressions is volatile,
   we store 1 in *PVOLATILEP.  Otherwise we don't change that.

   If the field is a non-BLKmode bit-field, *PMODE is set to VOIDmode.
   Otherwise, it is a mode that can be used to access the field.

   If the field describes a variable-sized object, *PMODE is set to
   BLKmode and *PBITSIZE is set to -1.  An access cannot be made in
   this case, but the address of the object can be found.  */

tree
get_inner_reference (tree exp, poly_int64_pod *pbitsize,
		     poly_int64_pod *pbitpos, tree *poffset,
		     machine_mode *pmode, int *punsignedp,
		     int *preversep, int *pvolatilep)
{
  tree size_tree = 0;
  machine_mode mode = VOIDmode;
  bool blkmode_bitfield = false;
  tree offset = size_zero_node;
  poly_offset_int bit_offset = 0;

  /* First get the mode, signedness, storage order and size.  We do this from
     just the outermost expression.  */
  *pbitsize = -1;
  if (TREE_CODE (exp) == COMPONENT_REF)
    {
      tree field = TREE_OPERAND (exp, 1);
      size_tree = DECL_SIZE (field);
      if (flag_strict_volatile_bitfields > 0
	  && TREE_THIS_VOLATILE (exp)
	  && DECL_BIT_FIELD_TYPE (field)
	  && DECL_MODE (field) != BLKmode)
	/* Volatile bitfields should be accessed in the mode of the
	     field's type, not the mode computed based on the bit
	     size.  */
	mode = TYPE_MODE (DECL_BIT_FIELD_TYPE (field));
      else if (!DECL_BIT_FIELD (field))
	{
	  mode = DECL_MODE (field);
	  /* For vector fields re-check the target flags, as DECL_MODE
	     could have been set with different target flags than
	     the current function has.  */
	  if (mode == BLKmode
	      && VECTOR_TYPE_P (TREE_TYPE (field))
	      && VECTOR_MODE_P (TYPE_MODE_RAW (TREE_TYPE (field))))
	    mode = TYPE_MODE (TREE_TYPE (field));
	}
      else if (DECL_MODE (field) == BLKmode)
	blkmode_bitfield = true;

      *punsignedp = DECL_UNSIGNED (field);
    }
  else if (TREE_CODE (exp) == BIT_FIELD_REF)
    {
      size_tree = TREE_OPERAND (exp, 1);
      *punsignedp = (! INTEGRAL_TYPE_P (TREE_TYPE (exp))
		     || TYPE_UNSIGNED (TREE_TYPE (exp)));

      /* For vector element types with the correct size of access or for
         vector typed accesses use the mode of the access type.  */
      if ((TREE_CODE (TREE_TYPE (TREE_OPERAND (exp, 0))) == VECTOR_TYPE
	   && TREE_TYPE (exp) == TREE_TYPE (TREE_TYPE (TREE_OPERAND (exp, 0)))
	   && tree_int_cst_equal (size_tree, TYPE_SIZE (TREE_TYPE (exp))))
	  || VECTOR_TYPE_P (TREE_TYPE (exp)))
	mode = TYPE_MODE (TREE_TYPE (exp));
    }
  else
    {
      mode = TYPE_MODE (TREE_TYPE (exp));
      *punsignedp = TYPE_UNSIGNED (TREE_TYPE (exp));

      if (mode == BLKmode)
	size_tree = TYPE_SIZE (TREE_TYPE (exp));
      else
	*pbitsize = GET_MODE_BITSIZE (mode);
    }

  if (size_tree != 0)
    {
      if (! tree_fits_uhwi_p (size_tree))
	mode = BLKmode, *pbitsize = -1;
      else
	*pbitsize = tree_to_uhwi (size_tree);
    }

  *preversep = reverse_storage_order_for_component_p (exp);

  /* Compute cumulative bit-offset for nested component-refs and array-refs,
     and find the ultimate containing object.  */
  while (1)
    {
      switch (TREE_CODE (exp))
	{
	case BIT_FIELD_REF:
	  bit_offset += wi::to_poly_offset (TREE_OPERAND (exp, 2));
	  break;

	case COMPONENT_REF:
	  {
	    tree field = TREE_OPERAND (exp, 1);
	    tree this_offset = component_ref_field_offset (exp);

	    /* If this field hasn't been filled in yet, don't go past it.
	       This should only happen when folding expressions made during
	       type construction.  */
	    if (this_offset == 0)
	      break;

	    offset = size_binop (PLUS_EXPR, offset, this_offset);
	    bit_offset += wi::to_poly_offset (DECL_FIELD_BIT_OFFSET (field));

	    /* ??? Right now we don't do anything with DECL_OFFSET_ALIGN.  */
	  }
	  break;

	case ARRAY_REF:
	case ARRAY_RANGE_REF:
	  {
	    tree index = TREE_OPERAND (exp, 1);
	    tree low_bound = array_ref_low_bound (exp);
	    tree unit_size = array_ref_element_size (exp);

	    /* We assume all arrays have sizes that are a multiple of a byte.
	       First subtract the lower bound, if any, in the type of the
	       index, then convert to sizetype and multiply by the size of
	       the array element.  */
	    if (! integer_zerop (low_bound))
	      index = fold_build2 (MINUS_EXPR, TREE_TYPE (index),
				   index, low_bound);

	    offset = size_binop (PLUS_EXPR, offset,
			         size_binop (MULT_EXPR,
					     fold_convert (sizetype, index),
					     unit_size));
	  }
	  break;

	case REALPART_EXPR:
	  break;

	case IMAGPART_EXPR:
	  bit_offset += *pbitsize;
	  break;

	case VIEW_CONVERT_EXPR:
	  break;

	case MEM_REF:
	  /* Hand back the decl for MEM[&decl, off].  */
	  if (TREE_CODE (TREE_OPERAND (exp, 0)) == ADDR_EXPR)
	    {
	      tree off = TREE_OPERAND (exp, 1);
	      if (!integer_zerop (off))
		{
		  poly_offset_int boff = mem_ref_offset (exp);
		  boff <<= LOG2_BITS_PER_UNIT;
		  bit_offset += boff;
		}
	      exp = TREE_OPERAND (TREE_OPERAND (exp, 0), 0);
	    }
	  goto done;

	default:
	  goto done;
	}

      /* If any reference in the chain is volatile, the effect is volatile.  */
      if (TREE_THIS_VOLATILE (exp))
	*pvolatilep = 1;

      exp = TREE_OPERAND (exp, 0);
    }
 done:

  /* If OFFSET is constant, see if we can return the whole thing as a
     constant bit position.  Make sure to handle overflow during
     this conversion.  */
  if (poly_int_tree_p (offset))
    {
      poly_offset_int tem = wi::sext (wi::to_poly_offset (offset),
				      TYPE_PRECISION (sizetype));
      tem <<= LOG2_BITS_PER_UNIT;
      tem += bit_offset;
      if (tem.to_shwi (pbitpos))
	*poffset = offset = NULL_TREE;
    }

  /* Otherwise, split it up.  */
  if (offset)
    {
      /* Avoid returning a negative bitpos as this may wreak havoc later.  */
      if (!bit_offset.to_shwi (pbitpos) || maybe_lt (*pbitpos, 0))
        {
	  *pbitpos = num_trailing_bits (bit_offset.force_shwi ());
	  poly_offset_int bytes = bits_to_bytes_round_down (bit_offset);
	  offset = size_binop (PLUS_EXPR, offset,
			       build_int_cst (sizetype, bytes.force_shwi ()));
	}

      *poffset = offset;
    }

  /* We can use BLKmode for a byte-aligned BLKmode bitfield.  */
  if (mode == VOIDmode
      && blkmode_bitfield
      && multiple_p (*pbitpos, BITS_PER_UNIT)
      && multiple_p (*pbitsize, BITS_PER_UNIT))
    *pmode = BLKmode;
  else
    *pmode = mode;

  return exp;
}

/* Alignment in bits the TARGET of an assignment may be assumed to have.  */

static unsigned HOST_WIDE_INT
target_align (const_tree target)
{
  /* We might have a chain of nested references with intermediate misaligning
     bitfields components, so need to recurse to find out.  */

  unsigned HOST_WIDE_INT this_align, outer_align;

  switch (TREE_CODE (target))
    {
    case BIT_FIELD_REF:
      return 1;

    case COMPONENT_REF:
      this_align = DECL_ALIGN (TREE_OPERAND (target, 1));
      outer_align = target_align (TREE_OPERAND (target, 0));
      return MIN (this_align, outer_align);

    case ARRAY_REF:
    case ARRAY_RANGE_REF:
      this_align = TYPE_ALIGN (TREE_TYPE (target));
      outer_align = target_align (TREE_OPERAND (target, 0));
      return MIN (this_align, outer_align);

    CASE_CONVERT:
    case NON_LVALUE_EXPR:
    case VIEW_CONVERT_EXPR:
      this_align = TYPE_ALIGN (TREE_TYPE (target));
      outer_align = target_align (TREE_OPERAND (target, 0));
      return MAX (this_align, outer_align);

    default:
      return TYPE_ALIGN (TREE_TYPE (target));
    }
}


/* Given an rtx VALUE that may contain additions and multiplications, return
   an equivalent value that just refers to a register, memory, or constant.
   This is done by generating instructions to perform the arithmetic and
   returning a pseudo-register containing the value.

   The returned value may be a REG, SUBREG, MEM or constant.  */

rtx
force_operand (rtx value, rtx target)
{
  rtx op1, op2;
  /* Use subtarget as the target for operand 0 of a binary operation.  */
  rtx subtarget = get_subtarget (target);
  enum rtx_code code = GET_CODE (value);

  /* Check for subreg applied to an expression produced by loop optimizer.  */
  if (code == SUBREG
      && !REG_P (SUBREG_REG (value))
      && !MEM_P (SUBREG_REG (value)))
    {
      value
	= simplify_gen_subreg (GET_MODE (value),
			       force_reg (GET_MODE (SUBREG_REG (value)),
					  force_operand (SUBREG_REG (value),
							 NULL_RTX)),
			       GET_MODE (SUBREG_REG (value)),
			       SUBREG_BYTE (value));
      code = GET_CODE (value);
    }

  /* Check for a PIC address load.  */
  if ((code == PLUS || code == MINUS)
      && XEXP (value, 0) == pic_offset_table_rtx
      && (GET_CODE (XEXP (value, 1)) == SYMBOL_REF
	  || GET_CODE (XEXP (value, 1)) == LABEL_REF
	  || GET_CODE (XEXP (value, 1)) == CONST))
    {
      if (!subtarget)
	subtarget = gen_reg_rtx (GET_MODE (value));
      emit_move_insn (subtarget, value);
      return subtarget;
    }

  if (ARITHMETIC_P (value))
    {
      op2 = XEXP (value, 1);
      if (!CONSTANT_P (op2) && !(REG_P (op2) && op2 != subtarget))
	subtarget = 0;
      if (code == MINUS && CONST_INT_P (op2))
	{
	  code = PLUS;
	  op2 = negate_rtx (GET_MODE (value), op2);
	}

      /* Check for an addition with OP2 a constant integer and our first
         operand a PLUS of a virtual register and something else.  In that
         case, we want to emit the sum of the virtual register and the
         constant first and then add the other value.  This allows virtual
         register instantiation to simply modify the constant rather than
         creating another one around this addition.  */
      if (code == PLUS && CONST_INT_P (op2)
	  && GET_CODE (XEXP (value, 0)) == PLUS
	  && REG_P (XEXP (XEXP (value, 0), 0))
	  && REGNO (XEXP (XEXP (value, 0), 0)) >= FIRST_VIRTUAL_REGISTER
	  && REGNO (XEXP (XEXP (value, 0), 0)) <= LAST_VIRTUAL_REGISTER)
	{
	  rtx temp = expand_simple_binop (GET_MODE (value), code,
					  XEXP (XEXP (value, 0), 0), op2,
					  subtarget, 0, OPTAB_LIB_WIDEN);
	  return expand_simple_binop (GET_MODE (value), code, temp,
				      force_operand (XEXP (XEXP (value,
								 0), 1), 0),
				      target, 0, OPTAB_LIB_WIDEN);
	}

      op1 = force_operand (XEXP (value, 0), subtarget);
      op2 = force_operand (op2, NULL_RTX);
      switch (code)
	{
	case MULT:
	  return expand_mult (GET_MODE (value), op1, op2, target, 1);
	case DIV:
	  if (!INTEGRAL_MODE_P (GET_MODE (value)))
	    return expand_simple_binop (GET_MODE (value), code, op1, op2,
					target, 1, OPTAB_LIB_WIDEN);
	  else
	    return expand_divmod (0,
				  FLOAT_MODE_P (GET_MODE (value))
				  ? RDIV_EXPR : TRUNC_DIV_EXPR,
				  GET_MODE (value), op1, op2, target, 0);
	case MOD:
	  return expand_divmod (1, TRUNC_MOD_EXPR, GET_MODE (value), op1, op2,
				target, 0);
	case UDIV:
	  return expand_divmod (0, TRUNC_DIV_EXPR, GET_MODE (value), op1, op2,
				target, 1);
	case UMOD:
	  return expand_divmod (1, TRUNC_MOD_EXPR, GET_MODE (value), op1, op2,
				target, 1);
	case ASHIFTRT:
	  return expand_simple_binop (GET_MODE (value), code, op1, op2,
				      target, 0, OPTAB_LIB_WIDEN);
	default:
	  return expand_simple_binop (GET_MODE (value), code, op1, op2,
				      target, 1, OPTAB_LIB_WIDEN);
	}
    }
  if (UNARY_P (value))
    {
      if (!target)
	target = gen_reg_rtx (GET_MODE (value));
      op1 = force_operand (XEXP (value, 0), NULL_RTX);
      switch (code)
	{
	case ZERO_EXTEND:
	case SIGN_EXTEND:
	case TRUNCATE:
	case FLOAT_EXTEND:
	case FLOAT_TRUNCATE:
	  convert_move (target, op1, code == ZERO_EXTEND);
	  return target;

	case FIX:
	case UNSIGNED_FIX:
	  expand_fix (target, op1, code == UNSIGNED_FIX);
	  return target;

	case FLOAT:
	case UNSIGNED_FLOAT:
	  expand_float (target, op1, code == UNSIGNED_FLOAT);
	  return target;

	default:
	  return expand_simple_unop (GET_MODE (value), code, op1, target, 0);
	}
    }

#ifdef INSN_SCHEDULING
  /* On machines that have insn scheduling, we want all memory reference to be
     explicit, so we need to deal with such paradoxical SUBREGs.  */
  if (paradoxical_subreg_p (value) && MEM_P (SUBREG_REG (value)))
    value
      = simplify_gen_subreg (GET_MODE (value),
			     force_reg (GET_MODE (SUBREG_REG (value)),
					force_operand (SUBREG_REG (value),
						       NULL_RTX)),
			     GET_MODE (SUBREG_REG (value)),
			     SUBREG_BYTE (value));
#endif

  return value;
}

/* Subroutine of expand_expr: return nonzero iff there is no way that
   EXP can reference X, which is being modified.  TOP_P is nonzero if this
   call is going to be used to determine whether we need a temporary
   for EXP, as opposed to a recursive call to this function.

   It is always safe for this routine to return zero since it merely
   searches for optimization opportunities.  */

int
safe_from_p (const_rtx x, tree exp, int top_p)
{
  rtx exp_rtl = 0;
  int i, nops;

  if (x == 0
      /* If EXP has varying size, we MUST use a target since we currently
	 have no way of allocating temporaries of variable size
	 (except for arrays that have TYPE_ARRAY_MAX_SIZE set).
	 So we assume here that something at a higher level has prevented a
	 clash.  This is somewhat bogus, but the best we can do.  Only
	 do this when X is BLKmode and when we are at the top level.  */
      || (top_p && TREE_TYPE (exp) != 0 && COMPLETE_TYPE_P (TREE_TYPE (exp))
	  && TREE_CODE (TYPE_SIZE (TREE_TYPE (exp))) != INTEGER_CST
	  && (TREE_CODE (TREE_TYPE (exp)) != ARRAY_TYPE
	      || TYPE_ARRAY_MAX_SIZE (TREE_TYPE (exp)) == NULL_TREE
	      || TREE_CODE (TYPE_ARRAY_MAX_SIZE (TREE_TYPE (exp)))
	      != INTEGER_CST)
	  && GET_MODE (x) == BLKmode)
      /* If X is in the outgoing argument area, it is always safe.  */
      || (MEM_P (x)
	  && (XEXP (x, 0) == virtual_outgoing_args_rtx
	      || (GET_CODE (XEXP (x, 0)) == PLUS
		  && XEXP (XEXP (x, 0), 0) == virtual_outgoing_args_rtx))))
    return 1;

  /* If this is a subreg of a hard register, declare it unsafe, otherwise,
     find the underlying pseudo.  */
  if (GET_CODE (x) == SUBREG)
    {
      x = SUBREG_REG (x);
      if (REG_P (x) && REGNO (x) < FIRST_PSEUDO_REGISTER)
	return 0;
    }

  /* Now look at our tree code and possibly recurse.  */
  switch (TREE_CODE_CLASS (TREE_CODE (exp)))
    {
    case tcc_declaration:
      exp_rtl = DECL_RTL_IF_SET (exp);
      break;

    case tcc_constant:
      return 1;

    case tcc_exceptional:
      if (TREE_CODE (exp) == TREE_LIST)
	{
	  while (1)
	    {
	      if (TREE_VALUE (exp) && !safe_from_p (x, TREE_VALUE (exp), 0))
		return 0;
	      exp = TREE_CHAIN (exp);
	      if (!exp)
		return 1;
	      if (TREE_CODE (exp) != TREE_LIST)
		return safe_from_p (x, exp, 0);
	    }
	}
      else if (TREE_CODE (exp) == CONSTRUCTOR)
	{
	  constructor_elt *ce;
	  unsigned HOST_WIDE_INT idx;

	  FOR_EACH_VEC_SAFE_ELT (CONSTRUCTOR_ELTS (exp), idx, ce)
	    if ((ce->index != NULL_TREE && !safe_from_p (x, ce->index, 0))
		|| !safe_from_p (x, ce->value, 0))
	      return 0;
	  return 1;
	}
      else if (TREE_CODE (exp) == ERROR_MARK)
	return 1;	/* An already-visited SAVE_EXPR? */
      else
	return 0;

    case tcc_statement:
      /* The only case we look at here is the DECL_INITIAL inside a
	 DECL_EXPR.  */
      return (TREE_CODE (exp) != DECL_EXPR
	      || TREE_CODE (DECL_EXPR_DECL (exp)) != VAR_DECL
	      || !DECL_INITIAL (DECL_EXPR_DECL (exp))
	      || safe_from_p (x, DECL_INITIAL (DECL_EXPR_DECL (exp)), 0));

    case tcc_binary:
    case tcc_comparison:
      if (!safe_from_p (x, TREE_OPERAND (exp, 1), 0))
	return 0;
      /* Fall through.  */

    case tcc_unary:
      return safe_from_p (x, TREE_OPERAND (exp, 0), 0);

    case tcc_expression:
    case tcc_reference:
    case tcc_vl_exp:
      /* Now do code-specific tests.  EXP_RTL is set to any rtx we find in
	 the expression.  If it is set, we conflict iff we are that rtx or
	 both are in memory.  Otherwise, we check all operands of the
	 expression recursively.  */

      switch (TREE_CODE (exp))
	{
	case ADDR_EXPR:
	  /* If the operand is static or we are static, we can't conflict.
	     Likewise if we don't conflict with the operand at all.  */
	  if (staticp (TREE_OPERAND (exp, 0))
	      || TREE_STATIC (exp)
	      || safe_from_p (x, TREE_OPERAND (exp, 0), 0))
	    return 1;

	  /* Otherwise, the only way this can conflict is if we are taking
	     the address of a DECL a that address if part of X, which is
	     very rare.  */
	  exp = TREE_OPERAND (exp, 0);
	  if (DECL_P (exp))
	    {
	      if (!DECL_RTL_SET_P (exp)
		  || !MEM_P (DECL_RTL (exp)))
		return 0;
	      else
		exp_rtl = XEXP (DECL_RTL (exp), 0);
	    }
	  break;

	case MEM_REF:
	  if (MEM_P (x)
	      && alias_sets_conflict_p (MEM_ALIAS_SET (x),
					get_alias_set (exp)))
	    return 0;
	  break;

	case CALL_EXPR:
	  /* Assume that the call will clobber all hard registers and
	     all of memory.  */
	  if ((REG_P (x) && REGNO (x) < FIRST_PSEUDO_REGISTER)
	      || MEM_P (x))
	    return 0;
	  break;

	case WITH_CLEANUP_EXPR:
	case CLEANUP_POINT_EXPR:
	  /* Lowered by gimplify.c.  */
	  gcc_unreachable ();

	case SAVE_EXPR:
	  return safe_from_p (x, TREE_OPERAND (exp, 0), 0);

	default:
	  break;
	}

      /* If we have an rtx, we do not need to scan our operands.  */
      if (exp_rtl)
	break;

      nops = TREE_OPERAND_LENGTH (exp);
      for (i = 0; i < nops; i++)
	if (TREE_OPERAND (exp, i) != 0
	    && ! safe_from_p (x, TREE_OPERAND (exp, i), 0))
	  return 0;

      break;

    case tcc_type:
      /* Should never get a type here.  */
      gcc_unreachable ();
    }

  /* If we have an rtl, find any enclosed object.  Then see if we conflict
     with it.  */
  if (exp_rtl)
    {
      if (GET_CODE (exp_rtl) == SUBREG)
	{
	  exp_rtl = SUBREG_REG (exp_rtl);
	  if (REG_P (exp_rtl)
	      && REGNO (exp_rtl) < FIRST_PSEUDO_REGISTER)
	    return 0;
	}

      /* If the rtl is X, then it is not safe.  Otherwise, it is unless both
	 are memory and they conflict.  */
      return ! (rtx_equal_p (x, exp_rtl)
		|| (MEM_P (x) && MEM_P (exp_rtl)
		    && true_dependence (exp_rtl, VOIDmode, x)));
    }

  /* If we reach here, it is safe.  */
  return 1;
}


/* Return the highest power of two that EXP is known to be a multiple of.
   This is used in updating alignment of MEMs in array references.  */

unsigned HOST_WIDE_INT
highest_pow2_factor (const_tree exp)
{
  unsigned HOST_WIDE_INT ret;
  int trailing_zeros = tree_ctz (exp);
  if (trailing_zeros >= HOST_BITS_PER_WIDE_INT)
    return BIGGEST_ALIGNMENT;
  ret = HOST_WIDE_INT_1U << trailing_zeros;
  if (ret > BIGGEST_ALIGNMENT)
    return BIGGEST_ALIGNMENT;
  return ret;
}

/* Similar, except that the alignment requirements of TARGET are
   taken into account.  Assume it is at least as aligned as its
   type, unless it is a COMPONENT_REF in which case the layout of
   the structure gives the alignment.  */

static unsigned HOST_WIDE_INT
highest_pow2_factor_for_target (const_tree target, const_tree exp)
{
  unsigned HOST_WIDE_INT talign = target_align (target) / BITS_PER_UNIT;
  unsigned HOST_WIDE_INT factor = highest_pow2_factor (exp);

  return MAX (factor, talign);
}

/* Convert the tree comparison code TCODE to the rtl one where the
   signedness is UNSIGNEDP.  */

static enum rtx_code
convert_tree_comp_to_rtx (enum tree_code tcode, int unsignedp)
{
  enum rtx_code code;
  switch (tcode)
    {
    case EQ_EXPR:
      code = EQ;
      break;
    case NE_EXPR:
      code = NE;
      break;
    case LT_EXPR:
      code = unsignedp ? LTU : LT;
      break;
    case LE_EXPR:
      code = unsignedp ? LEU : LE;
      break;
    case GT_EXPR:
      code = unsignedp ? GTU : GT;
      break;
    case GE_EXPR:
      code = unsignedp ? GEU : GE;
      break;
    case UNORDERED_EXPR:
      code = UNORDERED;
      break;
    case ORDERED_EXPR:
      code = ORDERED;
      break;
    case UNLT_EXPR:
      code = UNLT;
      break;
    case UNLE_EXPR:
      code = UNLE;
      break;
    case UNGT_EXPR:
      code = UNGT;
      break;
    case UNGE_EXPR:
      code = UNGE;
      break;
    case UNEQ_EXPR:
      code = UNEQ;
      break;
    case LTGT_EXPR:
      code = LTGT;
      break;

    default:
      gcc_unreachable ();
    }
  return code;
}

/* Subroutine of expand_expr.  Expand the two operands of a binary
   expression EXP0 and EXP1 placing the results in OP0 and OP1.
   The value may be stored in TARGET if TARGET is nonzero.  The
   MODIFIER argument is as documented by expand_expr.  */

void
expand_operands (tree exp0, tree exp1, rtx target, rtx *op0, rtx *op1,
		 enum expand_modifier modifier)
{
  if (! safe_from_p (target, exp1, 1))
    target = 0;
  if (operand_equal_p (exp0, exp1, 0))
    {
      *op0 = expand_expr (exp0, target, VOIDmode, modifier);
      *op1 = copy_rtx (*op0);
    }
  else
    {
      *op0 = expand_expr (exp0, target, VOIDmode, modifier);
      *op1 = expand_expr (exp1, NULL_RTX, VOIDmode, modifier);
    }
}


/* Return a MEM that contains constant EXP.  DEFER is as for
   output_constant_def and MODIFIER is as for expand_expr.  */

static rtx
expand_expr_constant (tree exp, int defer, enum expand_modifier modifier)
{
  rtx mem;

  mem = output_constant_def (exp, defer);
  if (modifier != EXPAND_INITIALIZER)
    mem = use_anchored_address (mem);
  return mem;
}

/* A subroutine of expand_expr_addr_expr.  Evaluate the address of EXP.
   The TARGET, TMODE and MODIFIER arguments are as for expand_expr.  */

static rtx
expand_expr_addr_expr_1 (tree exp, rtx target, scalar_int_mode tmode,
		         enum expand_modifier modifier, addr_space_t as)
{
  rtx result, subtarget;
  tree inner, offset;
  poly_int64 bitsize, bitpos;
  int unsignedp, reversep, volatilep = 0;
  machine_mode mode1;

  /* If we are taking the address of a constant and are at the top level,
     we have to use output_constant_def since we can't call force_const_mem
     at top level.  */
  /* ??? This should be considered a front-end bug.  We should not be
     generating ADDR_EXPR of something that isn't an LVALUE.  The only
     exception here is STRING_CST.  */
  if (CONSTANT_CLASS_P (exp))
    {
      result = XEXP (expand_expr_constant (exp, 0, modifier), 0);
      if (modifier < EXPAND_SUM)
	result = force_operand (result, target);
      return result;
    }

  /* Everything must be something allowed by is_gimple_addressable.  */
  switch (TREE_CODE (exp))
    {
    case INDIRECT_REF:
      /* This case will happen via recursion for &a->b.  */
      return expand_expr (TREE_OPERAND (exp, 0), target, tmode, modifier);

    case MEM_REF:
      {
	tree tem = TREE_OPERAND (exp, 0);
	if (!integer_zerop (TREE_OPERAND (exp, 1)))
	  tem = fold_build_pointer_plus (tem, TREE_OPERAND (exp, 1));
	return expand_expr (tem, target, tmode, modifier);
      }

    case TARGET_MEM_REF:
      return addr_for_mem_ref (exp, as, true);

    case CONST_DECL:
      /* Expand the initializer like constants above.  */
      result = XEXP (expand_expr_constant (DECL_INITIAL (exp),
					   0, modifier), 0);
      if (modifier < EXPAND_SUM)
	result = force_operand (result, target);
      return result;

    case REALPART_EXPR:
      /* The real part of the complex number is always first, therefore
	 the address is the same as the address of the parent object.  */
      offset = 0;
      bitpos = 0;
      inner = TREE_OPERAND (exp, 0);
      break;

    case IMAGPART_EXPR:
      /* The imaginary part of the complex number is always second.
	 The expression is therefore always offset by the size of the
	 scalar type.  */
      offset = 0;
      bitpos = GET_MODE_BITSIZE (SCALAR_TYPE_MODE (TREE_TYPE (exp)));
      inner = TREE_OPERAND (exp, 0);
      break;

    case COMPOUND_LITERAL_EXPR:
      /* Allow COMPOUND_LITERAL_EXPR in initializers or coming from
	 initializers, if e.g. rtl_for_decl_init is called on DECL_INITIAL
	 with COMPOUND_LITERAL_EXPRs in it, or ARRAY_REF on a const static
	 array with address of COMPOUND_LITERAL_EXPR in DECL_INITIAL;
	 the initializers aren't gimplified.  */
      if (COMPOUND_LITERAL_EXPR_DECL (exp)
	  && TREE_STATIC (COMPOUND_LITERAL_EXPR_DECL (exp)))
	return expand_expr_addr_expr_1 (COMPOUND_LITERAL_EXPR_DECL (exp),
					target, tmode, modifier, as);
      /* FALLTHRU */
    default:
      /* If the object is a DECL, then expand it for its rtl.  Don't bypass
	 expand_expr, as that can have various side effects; LABEL_DECLs for
	 example, may not have their DECL_RTL set yet.  Expand the rtl of
	 CONSTRUCTORs too, which should yield a memory reference for the
	 constructor's contents.  Assume language specific tree nodes can
	 be expanded in some interesting way.  */
      gcc_assert (TREE_CODE (exp) < LAST_AND_UNUSED_TREE_CODE);
      if (DECL_P (exp)
	  || TREE_CODE (exp) == CONSTRUCTOR
	  || TREE_CODE (exp) == COMPOUND_LITERAL_EXPR)
	{
	  result = expand_expr (exp, target, tmode,
				modifier == EXPAND_INITIALIZER
				? EXPAND_INITIALIZER : EXPAND_CONST_ADDRESS);

	  /* If the DECL isn't in memory, then the DECL wasn't properly
	     marked TREE_ADDRESSABLE, which will be either a front-end
	     or a tree optimizer bug.  */

	  gcc_assert (MEM_P (result));
	  result = XEXP (result, 0);

	  /* ??? Is this needed anymore?  */
	  if (DECL_P (exp))
	    TREE_USED (exp) = 1;

	  if (modifier != EXPAND_INITIALIZER
	      && modifier != EXPAND_CONST_ADDRESS
	      && modifier != EXPAND_SUM)
	    result = force_operand (result, target);
	  return result;
	}

      /* Pass FALSE as the last argument to get_inner_reference although
	 we are expanding to RTL.  The rationale is that we know how to
	 handle "aligning nodes" here: we can just bypass them because
	 they won't change the final object whose address will be returned
	 (they actually exist only for that purpose).  */
      inner = get_inner_reference (exp, &bitsize, &bitpos, &offset, &mode1,
				   &unsignedp, &reversep, &volatilep);
      break;
    }

  /* We must have made progress.  */
  gcc_assert (inner != exp);

  subtarget = offset || maybe_ne (bitpos, 0) ? NULL_RTX : target;
  /* For VIEW_CONVERT_EXPR, where the outer alignment is bigger than
     inner alignment, force the inner to be sufficiently aligned.  */
  if (CONSTANT_CLASS_P (inner)
      && TYPE_ALIGN (TREE_TYPE (inner)) < TYPE_ALIGN (TREE_TYPE (exp)))
    {
      inner = copy_node (inner);
      TREE_TYPE (inner) = copy_node (TREE_TYPE (inner));
      SET_TYPE_ALIGN (TREE_TYPE (inner), TYPE_ALIGN (TREE_TYPE (exp)));
      TYPE_USER_ALIGN (TREE_TYPE (inner)) = 1;
    }
  result = expand_expr_addr_expr_1 (inner, subtarget, tmode, modifier, as);

  if (offset)
    {
      rtx tmp;

      if (modifier != EXPAND_NORMAL)
	result = force_operand (result, NULL);
      tmp = expand_expr (offset, NULL_RTX, tmode,
			 modifier == EXPAND_INITIALIZER
			  ? EXPAND_INITIALIZER : EXPAND_NORMAL);

      /* expand_expr is allowed to return an object in a mode other
	 than TMODE.  If it did, we need to convert.  */
      if (GET_MODE (tmp) != VOIDmode && tmode != GET_MODE (tmp))
	tmp = convert_modes (tmode, GET_MODE (tmp),
			     tmp, TYPE_UNSIGNED (TREE_TYPE (offset)));
      result = convert_memory_address_addr_space (tmode, result, as);
      tmp = convert_memory_address_addr_space (tmode, tmp, as);

      if (modifier == EXPAND_SUM || modifier == EXPAND_INITIALIZER)
	result = simplify_gen_binary (PLUS, tmode, result, tmp);
      else
	{
	  subtarget = maybe_ne (bitpos, 0) ? NULL_RTX : target;
	  result = expand_simple_binop (tmode, PLUS, result, tmp, subtarget,
					1, OPTAB_LIB_WIDEN);
	}
    }

  if (maybe_ne (bitpos, 0))
    {
      /* Someone beforehand should have rejected taking the address
	 of an object that isn't byte-aligned.  */
      poly_int64 bytepos = exact_div (bitpos, BITS_PER_UNIT);
      result = convert_memory_address_addr_space (tmode, result, as);
      result = plus_constant (tmode, result, bytepos);
      if (modifier < EXPAND_SUM)
	result = force_operand (result, target);
    }

  return result;
}

/* A subroutine of expand_expr.  Evaluate EXP, which is an ADDR_EXPR.
   The TARGET, TMODE and MODIFIER arguments are as for expand_expr.  */

static rtx
expand_expr_addr_expr (tree exp, rtx target, machine_mode tmode,
		       enum expand_modifier modifier)
{
  addr_space_t as = ADDR_SPACE_GENERIC;
  scalar_int_mode address_mode = Pmode;
  scalar_int_mode pointer_mode = ptr_mode;
  machine_mode rmode;
  rtx result;

  /* Target mode of VOIDmode says "whatever's natural".  */
  if (tmode == VOIDmode)
    tmode = TYPE_MODE (TREE_TYPE (exp));

  if (POINTER_TYPE_P (TREE_TYPE (exp)))
    {
      as = TYPE_ADDR_SPACE (TREE_TYPE (TREE_TYPE (exp)));
      address_mode = targetm.addr_space.address_mode (as);
      pointer_mode = targetm.addr_space.pointer_mode (as);
    }

  /* We can get called with some Weird Things if the user does silliness
     like "(short) &a".  In that case, convert_memory_address won't do
     the right thing, so ignore the given target mode.  */
  scalar_int_mode new_tmode = (tmode == pointer_mode
			       ? pointer_mode
			       : address_mode);

  result = expand_expr_addr_expr_1 (TREE_OPERAND (exp, 0), target,
				    new_tmode, modifier, as);

  /* Despite expand_expr claims concerning ignoring TMODE when not
     strictly convenient, stuff breaks if we don't honor it.  Note
     that combined with the above, we only do this for pointer modes.  */
  rmode = GET_MODE (result);
  if (rmode == VOIDmode)
    rmode = new_tmode;
  if (rmode != new_tmode)
    result = convert_memory_address_addr_space (new_tmode, result, as);

  return result;
}

/* Generate code for computing CONSTRUCTOR EXP.
   An rtx for the computed value is returned.  If AVOID_TEMP_MEM
   is TRUE, instead of creating a temporary variable in memory
   NULL is returned and the caller needs to handle it differently.  */

static rtx
expand_constructor (tree exp, rtx target, enum expand_modifier modifier,
		    bool avoid_temp_mem)
{
  tree type = TREE_TYPE (exp);
  machine_mode mode = TYPE_MODE (type);

  /* Try to avoid creating a temporary at all.  This is possible
     if all of the initializer is zero.
     FIXME: try to handle all [0..255] initializers we can handle
     with memset.  */
  if (TREE_STATIC (exp)
      && !TREE_ADDRESSABLE (exp)
      && target != 0 && mode == BLKmode
      && all_zeros_p (exp))
    {
      clear_storage (target, expr_size (exp), BLOCK_OP_NORMAL);
      return target;
    }

  /* All elts simple constants => refer to a constant in memory.  But
     if this is a non-BLKmode mode, let it store a field at a time
     since that should make a CONST_INT, CONST_WIDE_INT or
     CONST_DOUBLE when we fold.  Likewise, if we have a target we can
     use, it is best to store directly into the target unless the type
     is large enough that memcpy will be used.  If we are making an
     initializer and all operands are constant, put it in memory as
     well.

     FIXME: Avoid trying to fill vector constructors piece-meal.
     Output them with output_constant_def below unless we're sure
     they're zeros.  This should go away when vector initializers
     are treated like VECTOR_CST instead of arrays.  */
  if ((TREE_STATIC (exp)
       && ((mode == BLKmode
	    && ! (target != 0 && safe_from_p (target, exp, 1)))
	   || TREE_ADDRESSABLE (exp)
	   || (tree_fits_uhwi_p (TYPE_SIZE_UNIT (type))
	       && (! can_move_by_pieces
		   (tree_to_uhwi (TYPE_SIZE_UNIT (type)),
		    TYPE_ALIGN (type)))
	       && ! mostly_zeros_p (exp))))
      || ((modifier == EXPAND_INITIALIZER || modifier == EXPAND_CONST_ADDRESS)
	  && TREE_CONSTANT (exp)))
    {
      rtx constructor;

      if (avoid_temp_mem)
	return NULL_RTX;

      constructor = expand_expr_constant (exp, 1, modifier);

      if (modifier != EXPAND_CONST_ADDRESS
	  && modifier != EXPAND_INITIALIZER
	  && modifier != EXPAND_SUM)
	constructor = validize_mem (constructor);

      return constructor;
    }

  /* Handle calls that pass values in multiple non-contiguous
     locations.  The Irix 6 ABI has examples of this.  */
  if (target == 0 || ! safe_from_p (target, exp, 1)
      || GET_CODE (target) == PARALLEL || modifier == EXPAND_STACK_PARM
      /* Also make a temporary if the store is to volatile memory, to
	 avoid individual accesses to aggregate members.  */
      || (GET_CODE (target) == MEM
	  && MEM_VOLATILE_P (target)
	  && !TREE_ADDRESSABLE (TREE_TYPE (exp))))
    {
      if (avoid_temp_mem)
	return NULL_RTX;

      target = assign_temp (type, TREE_ADDRESSABLE (exp), 1);
    }

  store_constructor (exp, target, 0, int_expr_size (exp), false);
  return target;
}


/* expand_expr: generate code for computing expression EXP.
   An rtx for the computed value is returned.  The value is never null.
   In the case of a void EXP, const0_rtx is returned.

   The value may be stored in TARGET if TARGET is nonzero.
   TARGET is just a suggestion; callers must assume that
   the rtx returned may not be the same as TARGET.

   If TARGET is CONST0_RTX, it means that the value will be ignored.

   If TMODE is not VOIDmode, it suggests generating the
   result in mode TMODE.  But this is done only when convenient.
   Otherwise, TMODE is ignored and the value generated in its natural mode.
   TMODE is just a suggestion; callers must assume that
   the rtx returned may not have mode TMODE.

   Note that TARGET may have neither TMODE nor MODE.  In that case, it
   probably will not be used.

   If MODIFIER is EXPAND_SUM then when EXP is an addition
   we can return an rtx of the form (MULT (REG ...) (CONST_INT ...))
   or a nest of (PLUS ...) and (MINUS ...) where the terms are
   products as above, or REG or MEM, or constant.
   Ordinarily in such cases we would output mul or add instructions
   and then return a pseudo reg containing the sum.

   EXPAND_INITIALIZER is much like EXPAND_SUM except that
   it also marks a label as absolutely required (it can't be dead).
   It also makes a ZERO_EXTEND or SIGN_EXTEND instead of emitting extend insns.
   This is used for outputting expressions used in initializers.

   EXPAND_CONST_ADDRESS says that it is okay to return a MEM
   with a constant address even if that address is not normally legitimate.
   EXPAND_INITIALIZER and EXPAND_SUM also have this effect.

   EXPAND_STACK_PARM is used when expanding to a TARGET on the stack for
   a call parameter.  Such targets require special care as we haven't yet
   marked TARGET so that it's safe from being trashed by libcalls.  We
   don't want to use TARGET for anything but the final result;
   Intermediate values must go elsewhere.   Additionally, calls to
   emit_block_move will be flagged with BLOCK_OP_CALL_PARM.

   If EXP is a VAR_DECL whose DECL_RTL was a MEM with an invalid
   address, and ALT_RTL is non-NULL, then *ALT_RTL is set to the
   DECL_RTL of the VAR_DECL.  *ALT_RTL is also set if EXP is a
   COMPOUND_EXPR whose second argument is such a VAR_DECL, and so on
   recursively.
   If the result can be stored at TARGET, and ALT_RTL is non-NULL,
   then *ALT_RTL is set to TARGET (before legitimziation).

   If INNER_REFERENCE_P is true, we are expanding an inner reference.
   In this case, we don't adjust a returned MEM rtx that wouldn't be
   sufficiently aligned for its mode; instead, it's up to the caller
   to deal with it afterwards.  This is used to make sure that unaligned
   base objects for which out-of-bounds accesses are supported, for
   example record types with trailing arrays, aren't realigned behind
   the back of the caller.
   The normal operating mode is to pass FALSE for this parameter.  */

rtx
expand_expr_real (tree exp, rtx target, machine_mode tmode,
		  enum expand_modifier modifier, rtx *alt_rtl,
		  bool inner_reference_p)
{
  rtx ret;

  /* Handle ERROR_MARK before anybody tries to access its type.  */
  if (TREE_CODE (exp) == ERROR_MARK
      || (TREE_CODE (TREE_TYPE (exp)) == ERROR_MARK))
    {
      ret = CONST0_RTX (tmode);
      return ret ? ret : const0_rtx;
    }

  ret = expand_expr_real_1 (exp, target, tmode, modifier, alt_rtl,
			    inner_reference_p);
  return ret;
}

/* Try to expand the conditional expression which is represented by
   TREEOP0 ? TREEOP1 : TREEOP2 using conditonal moves.  If it succeeds
   return the rtl reg which represents the result.  Otherwise return
   NULL_RTX.  */

static rtx
expand_cond_expr_using_cmove (tree treeop0 ATTRIBUTE_UNUSED,
			      tree treeop1 ATTRIBUTE_UNUSED,
			      tree treeop2 ATTRIBUTE_UNUSED)
{
  rtx insn;
  rtx op00, op01, op1, op2;
  enum rtx_code comparison_code;
  machine_mode comparison_mode;
  gimple *srcstmt;
  rtx temp;
  tree type = TREE_TYPE (treeop1);
  int unsignedp = TYPE_UNSIGNED (type);
  machine_mode mode = TYPE_MODE (type);
  machine_mode orig_mode = mode;
  static bool expanding_cond_expr_using_cmove = false;

  /* Conditional move expansion can end up TERing two operands which,
     when recursively hitting conditional expressions can result in
     exponential behavior if the cmove expansion ultimatively fails.
     It's hardly profitable to TER a cmove into a cmove so avoid doing
     that by failing early if we end up recursing.  */
  if (expanding_cond_expr_using_cmove)
    return NULL_RTX;

  /* If we cannot do a conditional move on the mode, try doing it
     with the promoted mode. */
  if (!can_conditionally_move_p (mode))
    {
      mode = promote_mode (type, mode, &unsignedp);
      if (!can_conditionally_move_p (mode))
	return NULL_RTX;
      temp = assign_temp (type, 0, 0); /* Use promoted mode for temp.  */
    }
  else
    temp = assign_temp (type, 0, 1);

  expanding_cond_expr_using_cmove = true;
  start_sequence ();
  expand_operands (treeop1, treeop2,
		   temp, &op1, &op2, EXPAND_NORMAL);

  if (TREE_CODE (treeop0) == SSA_NAME
      && (srcstmt = get_def_for_expr_class (treeop0, tcc_comparison)))
    {
      type = TREE_TYPE (gimple_assign_rhs1 (srcstmt));
      enum tree_code cmpcode = gimple_assign_rhs_code (srcstmt);
      op00 = expand_normal (gimple_assign_rhs1 (srcstmt));
      op01 = expand_normal (gimple_assign_rhs2 (srcstmt));
      comparison_mode = TYPE_MODE (type);
      unsignedp = TYPE_UNSIGNED (type);
      comparison_code = convert_tree_comp_to_rtx (cmpcode, unsignedp);
    }
  else if (COMPARISON_CLASS_P (treeop0))
    {
      type = TREE_TYPE (TREE_OPERAND (treeop0, 0));
      enum tree_code cmpcode = TREE_CODE (treeop0);
      op00 = expand_normal (TREE_OPERAND (treeop0, 0));
      op01 = expand_normal (TREE_OPERAND (treeop0, 1));
      unsignedp = TYPE_UNSIGNED (type);
      comparison_mode = TYPE_MODE (type);
      comparison_code = convert_tree_comp_to_rtx (cmpcode, unsignedp);
    }
  else
    {
      op00 = expand_normal (treeop0);
      op01 = const0_rtx;
      comparison_code = NE;
      comparison_mode = GET_MODE (op00);
      if (comparison_mode == VOIDmode)
	comparison_mode = TYPE_MODE (TREE_TYPE (treeop0));
    }
  expanding_cond_expr_using_cmove = false;

  if (GET_MODE (op1) != mode)
    op1 = gen_lowpart (mode, op1);

  if (GET_MODE (op2) != mode)
    op2 = gen_lowpart (mode, op2);

  /* Try to emit the conditional move.  */
  insn = emit_conditional_move (temp, comparison_code,
				op00, op01, comparison_mode,
				op1, op2, mode,
				unsignedp);

  /* If we could do the conditional move, emit the sequence,
     and return.  */
  if (insn)
    {
      rtx_insn *seq = get_insns ();
      end_sequence ();
      emit_insn (seq);
      return convert_modes (orig_mode, mode, temp, 0);
    }

  /* Otherwise discard the sequence and fall back to code with
     branches.  */
  end_sequence ();
  return NULL_RTX;
}

/* A helper function for expand_expr_real_2 to be used with a
   misaligned mem_ref TEMP.  Assume an unsigned type if UNSIGNEDP
   is nonzero, with alignment ALIGN in bits.
   Store the value at TARGET if possible (if TARGET is nonzero).
   Regardless of TARGET, we return the rtx for where the value is placed.
   If the result can be stored at TARGET, and ALT_RTL is non-NULL,
   then *ALT_RTL is set to TARGET (before legitimziation).  */

static rtx
expand_misaligned_mem_ref (rtx temp, machine_mode mode, int unsignedp,
			   unsigned int align, rtx target, rtx *alt_rtl)
{
  enum insn_code icode;

  if ((icode = optab_handler (movmisalign_optab, mode))
      != CODE_FOR_nothing)
    {
      class expand_operand ops[2];

      /* We've already validated the memory, and we're creating a
	 new pseudo destination.  The predicates really can't fail,
	 nor can the generator.  */
      create_output_operand (&ops[0], NULL_RTX, mode);
      create_fixed_operand (&ops[1], temp);
      expand_insn (icode, 2, ops);
      temp = ops[0].value;
    }
  else if (targetm.slow_unaligned_access (mode, align))
    temp = extract_bit_field (temp, GET_MODE_BITSIZE (mode),
			      0, unsignedp, target,
			      mode, mode, false, alt_rtl);
  return temp;
}

rtx
expand_expr_real_2 (sepops ops, rtx target, machine_mode tmode,
		    enum expand_modifier modifier)
{
  rtx op0, op1, op2, temp;
  rtx_code_label *lab;
  tree type;
  int unsignedp;
  machine_mode mode;
  scalar_int_mode int_mode;
  enum tree_code code = ops->code;
  optab this_optab;
  rtx subtarget, original_target;
  int ignore;
  bool reduce_bit_field;
  location_t loc = ops->location;
  tree treeop0, treeop1, treeop2;
#define REDUCE_BIT_FIELD(expr)	(reduce_bit_field			  \
				 ? reduce_to_bit_field_precision ((expr), \
								  target, \
								  type)	  \
				 : (expr))

  type = ops->type;
  mode = TYPE_MODE (type);
  unsignedp = TYPE_UNSIGNED (type);

  treeop0 = ops->op0;
  treeop1 = ops->op1;
  treeop2 = ops->op2;

  /* We should be called only on simple (binary or unary) expressions,
     exactly those that are valid in gimple expressions that aren't
     GIMPLE_SINGLE_RHS (or invalid).  */
  gcc_assert (get_gimple_rhs_class (code) == GIMPLE_UNARY_RHS
	      || get_gimple_rhs_class (code) == GIMPLE_BINARY_RHS
	      || get_gimple_rhs_class (code) == GIMPLE_TERNARY_RHS);

  ignore = (target == const0_rtx
	    || ((CONVERT_EXPR_CODE_P (code)
		 || code == COND_EXPR || code == VIEW_CONVERT_EXPR)
		&& TREE_CODE (type) == VOID_TYPE));

  /* We should be called only if we need the result.  */
  gcc_assert (!ignore);

  /* An operation in what may be a bit-field type needs the
     result to be reduced to the precision of the bit-field type,
     which is narrower than that of the type's mode.  */
  reduce_bit_field = (INTEGRAL_TYPE_P (type)
		      && !type_has_mode_precision_p (type));

  if (reduce_bit_field && modifier == EXPAND_STACK_PARM)
    target = 0;

  /* Use subtarget as the target for operand 0 of a binary operation.  */
  subtarget = get_subtarget (target);
  original_target = target;

  switch (code)
    {
    case NON_LVALUE_EXPR:
    case PAREN_EXPR:
    CASE_CONVERT:
      if (treeop0 == error_mark_node)
	return const0_rtx;

      if (TREE_CODE (type) == UNION_TYPE)
	{
	  tree valtype = TREE_TYPE (treeop0);

	  /* If both input and output are BLKmode, this conversion isn't doing
	     anything except possibly changing memory attribute.  */
	  if (mode == BLKmode && TYPE_MODE (valtype) == BLKmode)
	    {
	      rtx result = expand_expr (treeop0, target, tmode,
					modifier);

	      result = copy_rtx (result);
	      set_mem_attributes (result, type, 0);
	      return result;
	    }

	  if (target == 0)
	    {
	      if (TYPE_MODE (type) != BLKmode)
		target = gen_reg_rtx (TYPE_MODE (type));
	      else
		target = assign_temp (type, 1, 1);
	    }

	  if (MEM_P (target))
	    /* Store data into beginning of memory target.  */
	    store_expr (treeop0,
			adjust_address (target, TYPE_MODE (valtype), 0),
			modifier == EXPAND_STACK_PARM,
			false, TYPE_REVERSE_STORAGE_ORDER (type));

	  else
	    {
	      gcc_assert (REG_P (target)
			  && !TYPE_REVERSE_STORAGE_ORDER (type));

	      /* Store this field into a union of the proper type.  */
	      poly_uint64 op0_size
		= tree_to_poly_uint64 (TYPE_SIZE (TREE_TYPE (treeop0)));
	      poly_uint64 union_size = GET_MODE_BITSIZE (mode);
	      store_field (target,
			   /* The conversion must be constructed so that
			      we know at compile time how many bits
			      to preserve.  */
			   ordered_min (op0_size, union_size),
			   0, 0, 0, TYPE_MODE (valtype), treeop0, 0,
			   false, false);
	    }

	  /* Return the entire union.  */
	  return target;
	}

      if (mode == TYPE_MODE (TREE_TYPE (treeop0)))
	{
	  op0 = expand_expr (treeop0, target, VOIDmode,
			     modifier);

	  /* If the signedness of the conversion differs and OP0 is
	     a promoted SUBREG, clear that indication since we now
	     have to do the proper extension.  */
	  if (TYPE_UNSIGNED (TREE_TYPE (treeop0)) != unsignedp
	      && GET_CODE (op0) == SUBREG)
	    SUBREG_PROMOTED_VAR_P (op0) = 0;

	  return REDUCE_BIT_FIELD (op0);
	}

      op0 = expand_expr (treeop0, NULL_RTX, mode,
			 modifier == EXPAND_SUM ? EXPAND_NORMAL : modifier);
      if (GET_MODE (op0) == mode)
	;

      /* If OP0 is a constant, just convert it into the proper mode.  */
      else if (CONSTANT_P (op0))
	{
	  tree inner_type = TREE_TYPE (treeop0);
	  machine_mode inner_mode = GET_MODE (op0);

	  if (inner_mode == VOIDmode)
	    inner_mode = TYPE_MODE (inner_type);

	  if (modifier == EXPAND_INITIALIZER)
	    op0 = lowpart_subreg (mode, op0, inner_mode);
	  else
	    op0=  convert_modes (mode, inner_mode, op0,
				 TYPE_UNSIGNED (inner_type));
	}

      else if (modifier == EXPAND_INITIALIZER)
	op0 = gen_rtx_fmt_e (TYPE_UNSIGNED (TREE_TYPE (treeop0))
			     ? ZERO_EXTEND : SIGN_EXTEND, mode, op0);

      else if (target == 0)
	op0 = convert_to_mode (mode, op0,
			       TYPE_UNSIGNED (TREE_TYPE
					      (treeop0)));
      else
	{
	  convert_move (target, op0,
			TYPE_UNSIGNED (TREE_TYPE (treeop0)));
	  op0 = target;
	}

      return REDUCE_BIT_FIELD (op0);

    case ADDR_SPACE_CONVERT_EXPR:
      {
	tree treeop0_type = TREE_TYPE (treeop0);

	gcc_assert (POINTER_TYPE_P (type));
	gcc_assert (POINTER_TYPE_P (treeop0_type));

	addr_space_t as_to = TYPE_ADDR_SPACE (TREE_TYPE (type));
	addr_space_t as_from = TYPE_ADDR_SPACE (TREE_TYPE (treeop0_type));

        /* Conversions between pointers to the same address space should
	   have been implemented via CONVERT_EXPR / NOP_EXPR.  */
	gcc_assert (as_to != as_from);

	op0 = expand_expr (treeop0, NULL_RTX, VOIDmode, modifier);

        /* Ask target code to handle conversion between pointers
	   to overlapping address spaces.  */
	if (targetm.addr_space.subset_p (as_to, as_from)
	    || targetm.addr_space.subset_p (as_from, as_to))
	  {
	    op0 = targetm.addr_space.convert (op0, treeop0_type, type);
	  }
        else
          {
	    /* For disjoint address spaces, converting anything but a null
	       pointer invokes undefined behavior.  We truncate or extend the
	       value as if we'd converted via integers, which handles 0 as
	       required, and all others as the programmer likely expects.  */
#ifndef POINTERS_EXTEND_UNSIGNED
	    const int POINTERS_EXTEND_UNSIGNED = 1;
#endif
	    op0 = convert_modes (mode, TYPE_MODE (treeop0_type),
				 op0, POINTERS_EXTEND_UNSIGNED);
	  }
	gcc_assert (op0);
	return op0;
      }

    case POINTER_PLUS_EXPR:
      /* Even though the sizetype mode and the pointer's mode can be different
         expand is able to handle this correctly and get the correct result out
         of the PLUS_EXPR code.  */
      /* Make sure to sign-extend the sizetype offset in a POINTER_PLUS_EXPR
         if sizetype precision is smaller than pointer precision.  */
      if (TYPE_PRECISION (sizetype) < TYPE_PRECISION (type))
	treeop1 = fold_convert_loc (loc, type,
				    fold_convert_loc (loc, ssizetype,
						      treeop1));
      /* If sizetype precision is larger than pointer precision, truncate the
	 offset to have matching modes.  */
      else if (TYPE_PRECISION (sizetype) > TYPE_PRECISION (type))
	treeop1 = fold_convert_loc (loc, type, treeop1);
      /* FALLTHRU */

    case PLUS_EXPR:
      /* If we are adding a constant, a VAR_DECL that is sp, fp, or ap, and
	 something else, make sure we add the register to the constant and
	 then to the other thing.  This case can occur during strength
	 reduction and doing it this way will produce better code if the
	 frame pointer or argument pointer is eliminated.

	 fold-const.c will ensure that the constant is always in the inner
	 PLUS_EXPR, so the only case we need to do anything about is if
	 sp, ap, or fp is our second argument, in which case we must swap
	 the innermost first argument and our second argument.  */

      if (TREE_CODE (treeop0) == PLUS_EXPR
	  && TREE_CODE (TREE_OPERAND (treeop0, 1)) == INTEGER_CST
	  && VAR_P (treeop1)
	  && (DECL_RTL (treeop1) == frame_pointer_rtx
	      || DECL_RTL (treeop1) == stack_pointer_rtx
	      || DECL_RTL (treeop1) == arg_pointer_rtx))
	{
	  gcc_unreachable ();
	}

      /* If the result is to be ptr_mode and we are adding an integer to
	 something, we might be forming a constant.  So try to use
	 plus_constant.  If it produces a sum and we can't accept it,
	 use force_operand.  This allows P = &ARR[const] to generate
	 efficient code on machines where a SYMBOL_REF is not a valid
	 address.

	 If this is an EXPAND_SUM call, always return the sum.  */
      if (modifier == EXPAND_SUM || modifier == EXPAND_INITIALIZER
	  || (mode == ptr_mode && (unsignedp || ! flag_trapv)))
	{
	  if (modifier == EXPAND_STACK_PARM)
	    target = 0;
	  if (TREE_CODE (treeop0) == INTEGER_CST
	      && HWI_COMPUTABLE_MODE_P (mode)
	      && TREE_CONSTANT (treeop1))
	    {
	      rtx constant_part;
	      HOST_WIDE_INT wc;
	      machine_mode wmode = TYPE_MODE (TREE_TYPE (treeop1));

	      op1 = expand_expr (treeop1, subtarget, VOIDmode,
				 EXPAND_SUM);
	      /* Use wi::shwi to ensure that the constant is
		 truncated according to the mode of OP1, then sign extended
		 to a HOST_WIDE_INT.  Using the constant directly can result
		 in non-canonical RTL in a 64x32 cross compile.  */
	      wc = TREE_INT_CST_LOW (treeop0);
	      constant_part =
		immed_wide_int_const (wi::shwi (wc, wmode), wmode);
	      op1 = plus_constant (mode, op1, INTVAL (constant_part));
	      if (modifier != EXPAND_SUM && modifier != EXPAND_INITIALIZER)
		op1 = force_operand (op1, target);
	      return REDUCE_BIT_FIELD (op1);
	    }

	  else if (TREE_CODE (treeop1) == INTEGER_CST
		   && HWI_COMPUTABLE_MODE_P (mode)
		   && TREE_CONSTANT (treeop0))
	    {
	      rtx constant_part;
	      HOST_WIDE_INT wc;
	      machine_mode wmode = TYPE_MODE (TREE_TYPE (treeop0));

	      op0 = expand_expr (treeop0, subtarget, VOIDmode,
				 (modifier == EXPAND_INITIALIZER
				 ? EXPAND_INITIALIZER : EXPAND_SUM));
	      if (! CONSTANT_P (op0))
		{
		  op1 = expand_expr (treeop1, NULL_RTX,
				     VOIDmode, modifier);
		  /* Return a PLUS if modifier says it's OK.  */
		  if (modifier == EXPAND_SUM
		      || modifier == EXPAND_INITIALIZER)
		    return simplify_gen_binary (PLUS, mode, op0, op1);
		  goto binop2;
		}
	      /* Use wi::shwi to ensure that the constant is
		 truncated according to the mode of OP1, then sign extended
		 to a HOST_WIDE_INT.  Using the constant directly can result
		 in non-canonical RTL in a 64x32 cross compile.  */
	      wc = TREE_INT_CST_LOW (treeop1);
	      constant_part
		= immed_wide_int_const (wi::shwi (wc, wmode), wmode);
	      op0 = plus_constant (mode, op0, INTVAL (constant_part));
	      if (modifier != EXPAND_SUM && modifier != EXPAND_INITIALIZER)
		op0 = force_operand (op0, target);
	      return REDUCE_BIT_FIELD (op0);
	    }
	}

      /* Use TER to expand pointer addition of a negated value
	 as pointer subtraction.  */
      if ((POINTER_TYPE_P (TREE_TYPE (treeop0))
	   || (TREE_CODE (TREE_TYPE (treeop0)) == VECTOR_TYPE
	       && POINTER_TYPE_P (TREE_TYPE (TREE_TYPE (treeop0)))))
	  && TREE_CODE (treeop1) == SSA_NAME
	  && TYPE_MODE (TREE_TYPE (treeop0))
	     == TYPE_MODE (TREE_TYPE (treeop1)))
	{
	  gimple *def = get_def_for_expr (treeop1, NEGATE_EXPR);
	  if (def)
	    {
	      treeop1 = gimple_assign_rhs1 (def);
	      code = MINUS_EXPR;
	      goto do_minus;
	    }
	}

      /* No sense saving up arithmetic to be done
	 if it's all in the wrong mode to form part of an address.
	 And force_operand won't know whether to sign-extend or
	 zero-extend.  */
      if (modifier != EXPAND_INITIALIZER
	  && (modifier != EXPAND_SUM || mode != ptr_mode))
	{
	  expand_operands (treeop0, treeop1,
			   subtarget, &op0, &op1, modifier);
	  if (op0 == const0_rtx)
	    return op1;
	  if (op1 == const0_rtx)
	    return op0;
	  goto binop2;
	}

      expand_operands (treeop0, treeop1,
		       subtarget, &op0, &op1, modifier);
      return REDUCE_BIT_FIELD (simplify_gen_binary (PLUS, mode, op0, op1));

    case MINUS_EXPR:
    case POINTER_DIFF_EXPR:
    do_minus:
      /* For initializers, we are allowed to return a MINUS of two
	 symbolic constants.  Here we handle all cases when both operands
	 are constant.  */
      /* Handle difference of two symbolic constants,
	 for the sake of an initializer.  */
      if ((modifier == EXPAND_SUM || modifier == EXPAND_INITIALIZER)
	  && really_constant_p (treeop0)
	  && really_constant_p (treeop1))
	{
	  expand_operands (treeop0, treeop1,
			   NULL_RTX, &op0, &op1, modifier);
	  return simplify_gen_binary (MINUS, mode, op0, op1);
	}

      /* No sense saving up arithmetic to be done
	 if it's all in the wrong mode to form part of an address.
	 And force_operand won't know whether to sign-extend or
	 zero-extend.  */
      if (modifier != EXPAND_INITIALIZER
	  && (modifier != EXPAND_SUM || mode != ptr_mode))
	goto binop;

      expand_operands (treeop0, treeop1,
		       subtarget, &op0, &op1, modifier);

      /* Convert A - const to A + (-const).  */
      if (CONST_INT_P (op1))
	{
	  op1 = negate_rtx (mode, op1);
	  return REDUCE_BIT_FIELD (simplify_gen_binary (PLUS, mode, op0, op1));
	}

      goto binop2;

    case WIDEN_MULT_PLUS_EXPR:
    case WIDEN_MULT_MINUS_EXPR:
      expand_operands (treeop0, treeop1, NULL_RTX, &op0, &op1, EXPAND_NORMAL);
      op2 = expand_normal (treeop2);
      target = expand_widen_pattern_expr (ops, op0, op1, op2,
					  target, unsignedp);
      return target;

    case WIDEN_MULT_EXPR:
      /* If first operand is constant, swap them.
	 Thus the following special case checks need only
	 check the second operand.  */
      if (TREE_CODE (treeop0) == INTEGER_CST)
	std::swap (treeop0, treeop1);

      /* First, check if we have a multiplication of one signed and one
	 unsigned operand.  */
      if (TREE_CODE (treeop1) != INTEGER_CST
	  && (TYPE_UNSIGNED (TREE_TYPE (treeop0))
	      != TYPE_UNSIGNED (TREE_TYPE (treeop1))))
	{
	  machine_mode innermode = TYPE_MODE (TREE_TYPE (treeop0));
	  this_optab = usmul_widen_optab;
	  if (find_widening_optab_handler (this_optab, mode, innermode)
		!= CODE_FOR_nothing)
	    {
	      if (TYPE_UNSIGNED (TREE_TYPE (treeop0)))
		expand_operands (treeop0, treeop1, NULL_RTX, &op0, &op1,
				 EXPAND_NORMAL);
	      else
		expand_operands (treeop0, treeop1, NULL_RTX, &op1, &op0,
				 EXPAND_NORMAL);
	      /* op0 and op1 might still be constant, despite the above
		 != INTEGER_CST check.  Handle it.  */
	      if (GET_MODE (op0) == VOIDmode && GET_MODE (op1) == VOIDmode)
		{
		  op0 = convert_modes (mode, innermode, op0, true);
		  op1 = convert_modes (mode, innermode, op1, false);
		  return REDUCE_BIT_FIELD (expand_mult (mode, op0, op1,
							target, unsignedp));
		}
	      goto binop3;
	    }
	}
      /* Check for a multiplication with matching signedness.  */
      else if ((TREE_CODE (treeop1) == INTEGER_CST
		&& int_fits_type_p (treeop1, TREE_TYPE (treeop0)))
	       || (TYPE_UNSIGNED (TREE_TYPE (treeop1))
		   == TYPE_UNSIGNED (TREE_TYPE (treeop0))))
	{
	  tree op0type = TREE_TYPE (treeop0);
	  machine_mode innermode = TYPE_MODE (op0type);
	  bool zextend_p = TYPE_UNSIGNED (op0type);
	  optab other_optab = zextend_p ? smul_widen_optab : umul_widen_optab;
	  this_optab = zextend_p ? umul_widen_optab : smul_widen_optab;

	  if (TREE_CODE (treeop0) != INTEGER_CST)
	    {
	      if (find_widening_optab_handler (this_optab, mode, innermode)
		  != CODE_FOR_nothing)
		{
		  expand_operands (treeop0, treeop1, NULL_RTX, &op0, &op1,
				   EXPAND_NORMAL);
		  /* op0 and op1 might still be constant, despite the above
		     != INTEGER_CST check.  Handle it.  */
		  if (GET_MODE (op0) == VOIDmode && GET_MODE (op1) == VOIDmode)
		    {
		     widen_mult_const:
		      op0 = convert_modes (mode, innermode, op0, zextend_p);
		      op1
			= convert_modes (mode, innermode, op1,
					 TYPE_UNSIGNED (TREE_TYPE (treeop1)));
		      return REDUCE_BIT_FIELD (expand_mult (mode, op0, op1,
							    target,
							    unsignedp));
		    }
		  temp = expand_widening_mult (mode, op0, op1, target,
					       unsignedp, this_optab);
		  return REDUCE_BIT_FIELD (temp);
		}
	      if (find_widening_optab_handler (other_optab, mode, innermode)
		  != CODE_FOR_nothing
		  && innermode == word_mode)
		{
		  rtx htem, hipart;
		  op0 = expand_normal (treeop0);
		  op1 = expand_normal (treeop1);
		  /* op0 and op1 might be constants, despite the above
		     != INTEGER_CST check.  Handle it.  */
		  if (GET_MODE (op0) == VOIDmode && GET_MODE (op1) == VOIDmode)
		    goto widen_mult_const;
		  temp = expand_binop (mode, other_optab, op0, op1, target,
				       unsignedp, OPTAB_LIB_WIDEN);
		  hipart = gen_highpart (word_mode, temp);
		  htem = expand_mult_highpart_adjust (word_mode, hipart,
						      op0, op1, hipart,
						      zextend_p);
		  if (htem != hipart)
		    emit_move_insn (hipart, htem);
		  return REDUCE_BIT_FIELD (temp);
		}
	    }
	}
      treeop0 = fold_build1 (CONVERT_EXPR, type, treeop0);
      treeop1 = fold_build1 (CONVERT_EXPR, type, treeop1);
      expand_operands (treeop0, treeop1, subtarget, &op0, &op1, EXPAND_NORMAL);
      return REDUCE_BIT_FIELD (expand_mult (mode, op0, op1, target, unsignedp));

    case MULT_EXPR:
      /* If this is a fixed-point operation, then we cannot use the code
	 below because "expand_mult" doesn't support sat/no-sat fixed-point
         multiplications.   */
      if (ALL_FIXED_POINT_MODE_P (mode))
	goto binop;

      /* If first operand is constant, swap them.
	 Thus the following special case checks need only
	 check the second operand.  */
      if (TREE_CODE (treeop0) == INTEGER_CST)
	std::swap (treeop0, treeop1);

      /* Attempt to return something suitable for generating an
	 indexed address, for machines that support that.  */

      if (modifier == EXPAND_SUM && mode == ptr_mode
	  && tree_fits_shwi_p (treeop1))
	{
	  tree exp1 = treeop1;

	  op0 = expand_expr (treeop0, subtarget, VOIDmode,
			     EXPAND_SUM);

	  if (!REG_P (op0))
	    op0 = force_operand (op0, NULL_RTX);
	  if (!REG_P (op0))
	    op0 = copy_to_mode_reg (mode, op0);

	  return REDUCE_BIT_FIELD (gen_rtx_MULT (mode, op0,
			       gen_int_mode (tree_to_shwi (exp1),
					     TYPE_MODE (TREE_TYPE (exp1)))));
	}

      if (modifier == EXPAND_STACK_PARM)
	target = 0;

      expand_operands (treeop0, treeop1, subtarget, &op0, &op1, EXPAND_NORMAL);
      return REDUCE_BIT_FIELD (expand_mult (mode, op0, op1, target, unsignedp));

    case TRUNC_MOD_EXPR:
    case FLOOR_MOD_EXPR:
    case CEIL_MOD_EXPR:
    case ROUND_MOD_EXPR:

    case TRUNC_DIV_EXPR:
    case FLOOR_DIV_EXPR:
    case CEIL_DIV_EXPR:
    case ROUND_DIV_EXPR:
    case EXACT_DIV_EXPR:
     {
       /* If this is a fixed-point operation, then we cannot use the code
	  below because "expand_divmod" doesn't support sat/no-sat fixed-point
	  divisions.   */
       if (ALL_FIXED_POINT_MODE_P (mode))
	 goto binop;

       if (modifier == EXPAND_STACK_PARM)
	 target = 0;
       /* Possible optimization: compute the dividend with EXPAND_SUM
	  then if the divisor is constant can optimize the case
	  where some terms of the dividend have coeffs divisible by it.  */
       expand_operands (treeop0, treeop1,
			subtarget, &op0, &op1, EXPAND_NORMAL);
       bool mod_p = code == TRUNC_MOD_EXPR || code == FLOOR_MOD_EXPR
		    || code == CEIL_MOD_EXPR || code == ROUND_MOD_EXPR;
       if (SCALAR_INT_MODE_P (mode)
	   && optimize >= 2
	   && get_range_pos_neg (treeop0) == 1
	   && get_range_pos_neg (treeop1) == 1)
	 {
	   /* If both arguments are known to be positive when interpreted
	      as signed, we can expand it as both signed and unsigned
	      division or modulo.  Choose the cheaper sequence in that case.  */
	   bool speed_p = optimize_insn_for_speed_p ();
	   do_pending_stack_adjust ();
	   start_sequence ();
	   rtx uns_ret = expand_divmod (mod_p, code, mode, op0, op1, target, 1);
	   rtx_insn *uns_insns = get_insns ();
	   end_sequence ();
	   start_sequence ();
	   rtx sgn_ret = expand_divmod (mod_p, code, mode, op0, op1, target, 0);
	   rtx_insn *sgn_insns = get_insns ();
	   end_sequence ();
	   unsigned uns_cost = seq_cost (uns_insns, speed_p);
	   unsigned sgn_cost = seq_cost (sgn_insns, speed_p);

	   /* If costs are the same then use as tie breaker the other
	      other factor.  */
	   if (uns_cost == sgn_cost)
	     {
		uns_cost = seq_cost (uns_insns, !speed_p);
		sgn_cost = seq_cost (sgn_insns, !speed_p);
	     }

	   if (uns_cost < sgn_cost || (uns_cost == sgn_cost && unsignedp))
	     {
	       emit_insn (uns_insns);
	       return uns_ret;
	     }
	   emit_insn (sgn_insns);
	   return sgn_ret;
	 }
       return expand_divmod (mod_p, code, mode, op0, op1, target, unsignedp);
     }
    case RDIV_EXPR:
      goto binop;

    case MULT_HIGHPART_EXPR:
      expand_operands (treeop0, treeop1, subtarget, &op0, &op1, EXPAND_NORMAL);
      temp = expand_mult_highpart (mode, op0, op1, target, unsignedp);
      gcc_assert (temp);
      return temp;

    case FIXED_CONVERT_EXPR:
      op0 = expand_normal (treeop0);
      if (target == 0 || modifier == EXPAND_STACK_PARM)
	target = gen_reg_rtx (mode);

      if ((TREE_CODE (TREE_TYPE (treeop0)) == INTEGER_TYPE
	   && TYPE_UNSIGNED (TREE_TYPE (treeop0)))
          || (TREE_CODE (type) == INTEGER_TYPE && TYPE_UNSIGNED (type)))
	expand_fixed_convert (target, op0, 1, TYPE_SATURATING (type));
      else
	expand_fixed_convert (target, op0, 0, TYPE_SATURATING (type));
      return target;

    case FIX_TRUNC_EXPR:
      op0 = expand_normal (treeop0);
      if (target == 0 || modifier == EXPAND_STACK_PARM)
	target = gen_reg_rtx (mode);
      expand_fix (target, op0, unsignedp);
      return target;

    case FLOAT_EXPR:
      op0 = expand_normal (treeop0);
      if (target == 0 || modifier == EXPAND_STACK_PARM)
	target = gen_reg_rtx (mode);
      /* expand_float can't figure out what to do if FROM has VOIDmode.
	 So give it the correct mode.  With -O, cse will optimize this.  */
      if (GET_MODE (op0) == VOIDmode)
	op0 = copy_to_mode_reg (TYPE_MODE (TREE_TYPE (treeop0)),
				op0);
      expand_float (target, op0,
		    TYPE_UNSIGNED (TREE_TYPE (treeop0)));
      return target;

    case NEGATE_EXPR:
      op0 = expand_expr (treeop0, subtarget,
			 VOIDmode, EXPAND_NORMAL);
      if (modifier == EXPAND_STACK_PARM)
	target = 0;
      temp = expand_unop (mode,
      			  optab_for_tree_code (NEGATE_EXPR, type,
					       optab_default),
			  op0, target, 0);
      gcc_assert (temp);
      return REDUCE_BIT_FIELD (temp);

    case ABS_EXPR:
    case ABSU_EXPR:
      op0 = expand_expr (treeop0, subtarget,
			 VOIDmode, EXPAND_NORMAL);
      if (modifier == EXPAND_STACK_PARM)
	target = 0;

      /* ABS_EXPR is not valid for complex arguments.  */
      gcc_assert (GET_MODE_CLASS (mode) != MODE_COMPLEX_INT
		  && GET_MODE_CLASS (mode) != MODE_COMPLEX_FLOAT);

      /* Unsigned abs is simply the operand.  Testing here means we don't
	 risk generating incorrect code below.  */
      if (TYPE_UNSIGNED (TREE_TYPE (treeop0)))
	return op0;

      return expand_abs (mode, op0, target, unsignedp,
			 safe_from_p (target, treeop0, 1));

    case MAX_EXPR:
    case MIN_EXPR:
      target = original_target;
      if (target == 0
	  || modifier == EXPAND_STACK_PARM
	  || (MEM_P (target) && MEM_VOLATILE_P (target))
	  || GET_MODE (target) != mode
	  || (REG_P (target)
	      && REGNO (target) < FIRST_PSEUDO_REGISTER))
	target = gen_reg_rtx (mode);
      expand_operands (treeop0, treeop1,
		       target, &op0, &op1, EXPAND_NORMAL);

      /* First try to do it with a special MIN or MAX instruction.
	 If that does not win, use a conditional jump to select the proper
	 value.  */
      this_optab = optab_for_tree_code (code, type, optab_default);
      temp = expand_binop (mode, this_optab, op0, op1, target, unsignedp,
			   OPTAB_WIDEN);
      if (temp != 0)
	return temp;

      /* For vector MIN <x, y>, expand it a VEC_COND_EXPR <x <= y, x, y>
	 and similarly for MAX <x, y>.  */
      if (VECTOR_TYPE_P (type))
	{
	  tree t0 = make_tree (type, op0);
	  tree t1 = make_tree (type, op1);
	  tree comparison = build2 (code == MIN_EXPR ? LE_EXPR : GE_EXPR,
				    type, t0, t1);
	  return expand_vec_cond_expr (type, comparison, t0, t1,
				       original_target);
	}

      /* At this point, a MEM target is no longer useful; we will get better
	 code without it.  */

      if (! REG_P (target))
	target = gen_reg_rtx (mode);

      /* If op1 was placed in target, swap op0 and op1.  */
      if (target != op0 && target == op1)
	std::swap (op0, op1);

      /* We generate better code and avoid problems with op1 mentioning
	 target by forcing op1 into a pseudo if it isn't a constant.  */
      if (! CONSTANT_P (op1))
	op1 = force_reg (mode, op1);

      {
	enum rtx_code comparison_code;
	rtx cmpop1 = op1;

	if (code == MAX_EXPR)
	  comparison_code = unsignedp ? GEU : GE;
	else
	  comparison_code = unsignedp ? LEU : LE;

	/* Canonicalize to comparisons against 0.  */
	if (op1 == const1_rtx)
	  {
	    /* Converting (a >= 1 ? a : 1) into (a > 0 ? a : 1)
	       or (a != 0 ? a : 1) for unsigned.
	       For MIN we are safe converting (a <= 1 ? a : 1)
	       into (a <= 0 ? a : 1)  */
	    cmpop1 = const0_rtx;
	    if (code == MAX_EXPR)
	      comparison_code = unsignedp ? NE : GT;
	  }
	if (op1 == constm1_rtx && !unsignedp)
	  {
	    /* Converting (a >= -1 ? a : -1) into (a >= 0 ? a : -1)
	       and (a <= -1 ? a : -1) into (a < 0 ? a : -1) */
	    cmpop1 = const0_rtx;
	    if (code == MIN_EXPR)
	      comparison_code = LT;
	  }

	/* Use a conditional move if possible.  */
	if (can_conditionally_move_p (mode))
	  {
	    rtx insn;

	    start_sequence ();

	    /* Try to emit the conditional move.  */
	    insn = emit_conditional_move (target, comparison_code,
					  op0, cmpop1, mode,
					  op0, op1, mode,
					  unsignedp);

	    /* If we could do the conditional move, emit the sequence,
	       and return.  */
	    if (insn)
	      {
		rtx_insn *seq = get_insns ();
		end_sequence ();
		emit_insn (seq);
		return target;
	      }

	    /* Otherwise discard the sequence and fall back to code with
	       branches.  */
	    end_sequence ();
	  }

	if (target != op0)
	  emit_move_insn (target, op0);

	lab = gen_label_rtx ();
	do_compare_rtx_and_jump (target, cmpop1, comparison_code,
				 unsignedp, mode, NULL_RTX, NULL, lab,
				 profile_probability::uninitialized ());
      }
      emit_move_insn (target, op1);
      emit_label (lab);
      return target;

    case BIT_NOT_EXPR:
      op0 = expand_expr (treeop0, subtarget,
			 VOIDmode, EXPAND_NORMAL);
      if (modifier == EXPAND_STACK_PARM)
	target = 0;
      /* In case we have to reduce the result to bitfield precision
	 for unsigned bitfield expand this as XOR with a proper constant
	 instead.  */
      if (reduce_bit_field && TYPE_UNSIGNED (type))
	{
	  int_mode = SCALAR_INT_TYPE_MODE (type);
	  wide_int mask = wi::mask (TYPE_PRECISION (type),
				    false, GET_MODE_PRECISION (int_mode));

	  temp = expand_binop (int_mode, xor_optab, op0,
			       immed_wide_int_const (mask, int_mode),
			       target, 1, OPTAB_LIB_WIDEN);
	}
      else
	temp = expand_unop (mode, one_cmpl_optab, op0, target, 1);
      gcc_assert (temp);
      return temp;

      /* ??? Can optimize bitwise operations with one arg constant.
	 Can optimize (a bitwise1 n) bitwise2 (a bitwise3 b)
	 and (a bitwise1 b) bitwise2 b (etc)
	 but that is probably not worth while.  */

    case BIT_AND_EXPR:
    case BIT_IOR_EXPR:
    case BIT_XOR_EXPR:
      goto binop;

    case LROTATE_EXPR:
    case RROTATE_EXPR:
      gcc_assert (VECTOR_MODE_P (TYPE_MODE (type))
		  || type_has_mode_precision_p (type));
      /* fall through */

    case LSHIFT_EXPR:
    case RSHIFT_EXPR:
      {
	/* If this is a fixed-point operation, then we cannot use the code
	   below because "expand_shift" doesn't support sat/no-sat fixed-point
	   shifts.  */
	if (ALL_FIXED_POINT_MODE_P (mode))
	  goto binop;

	if (! safe_from_p (subtarget, treeop1, 1))
	  subtarget = 0;
	if (modifier == EXPAND_STACK_PARM)
	  target = 0;
	op0 = expand_expr (treeop0, subtarget,
			   VOIDmode, EXPAND_NORMAL);

	/* Left shift optimization when shifting across word_size boundary.

	   If mode == GET_MODE_WIDER_MODE (word_mode), then normally
	   there isn't native instruction to support this wide mode
	   left shift.  Given below scenario:

	    Type A = (Type) B  << C

	    |<		 T	    >|
	    | dest_high  |  dest_low |

			 | word_size |

	   If the shift amount C caused we shift B to across the word
	   size boundary, i.e part of B shifted into high half of
	   destination register, and part of B remains in the low
	   half, then GCC will use the following left shift expand
	   logic:

	   1. Initialize dest_low to B.
	   2. Initialize every bit of dest_high to the sign bit of B.
	   3. Logic left shift dest_low by C bit to finalize dest_low.
	      The value of dest_low before this shift is kept in a temp D.
	   4. Logic left shift dest_high by C.
	   5. Logic right shift D by (word_size - C).
	   6. Or the result of 4 and 5 to finalize dest_high.

	   While, by checking gimple statements, if operand B is
	   coming from signed extension, then we can simplify above
	   expand logic into:

	      1. dest_high = src_low >> (word_size - C).
	      2. dest_low = src_low << C.

	   We can use one arithmetic right shift to finish all the
	   purpose of steps 2, 4, 5, 6, thus we reduce the steps
	   needed from 6 into 2.

	   The case is similar for zero extension, except that we
	   initialize dest_high to zero rather than copies of the sign
	   bit from B.  Furthermore, we need to use a logical right shift
	   in this case.

	   The choice of sign-extension versus zero-extension is
	   determined entirely by whether or not B is signed and is
	   independent of the current setting of unsignedp.  */

	temp = NULL_RTX;
	if (code == LSHIFT_EXPR
	    && target
	    && REG_P (target)
	    && GET_MODE_2XWIDER_MODE (word_mode).exists (&int_mode)
	    && mode == int_mode
	    && TREE_CONSTANT (treeop1)
	    && TREE_CODE (treeop0) == SSA_NAME)
	  {
	    gimple *def = SSA_NAME_DEF_STMT (treeop0);
	    if (is_gimple_assign (def)
		&& gimple_assign_rhs_code (def) == NOP_EXPR)
	      {
		scalar_int_mode rmode = SCALAR_INT_TYPE_MODE
		  (TREE_TYPE (gimple_assign_rhs1 (def)));

		if (GET_MODE_SIZE (rmode) < GET_MODE_SIZE (int_mode)
		    && TREE_INT_CST_LOW (treeop1) < GET_MODE_BITSIZE (word_mode)
		    && ((TREE_INT_CST_LOW (treeop1) + GET_MODE_BITSIZE (rmode))
			>= GET_MODE_BITSIZE (word_mode)))
		  {
		    rtx_insn *seq, *seq_old;
		    poly_uint64 high_off = subreg_highpart_offset (word_mode,
								   int_mode);
		    bool extend_unsigned
		      = TYPE_UNSIGNED (TREE_TYPE (gimple_assign_rhs1 (def)));
		    rtx low = lowpart_subreg (word_mode, op0, int_mode);
		    rtx dest_low = lowpart_subreg (word_mode, target, int_mode);
		    rtx dest_high = simplify_gen_subreg (word_mode, target,
							 int_mode, high_off);
		    HOST_WIDE_INT ramount = (BITS_PER_WORD
					     - TREE_INT_CST_LOW (treeop1));
		    tree rshift = build_int_cst (TREE_TYPE (treeop1), ramount);

		    start_sequence ();
		    /* dest_high = src_low >> (word_size - C).  */
		    temp = expand_variable_shift (RSHIFT_EXPR, word_mode, low,
						  rshift, dest_high,
						  extend_unsigned);
		    if (temp != dest_high)
		      emit_move_insn (dest_high, temp);

		    /* dest_low = src_low << C.  */
		    temp = expand_variable_shift (LSHIFT_EXPR, word_mode, low,
						  treeop1, dest_low, unsignedp);
		    if (temp != dest_low)
		      emit_move_insn (dest_low, temp);

		    seq = get_insns ();
		    end_sequence ();
		    temp = target ;

		    if (have_insn_for (ASHIFT, int_mode))
		      {
			bool speed_p = optimize_insn_for_speed_p ();
			start_sequence ();
			rtx ret_old = expand_variable_shift (code, int_mode,
							     op0, treeop1,
							     target,
							     unsignedp);

			seq_old = get_insns ();
			end_sequence ();
			if (seq_cost (seq, speed_p)
			    >= seq_cost (seq_old, speed_p))
			  {
			    seq = seq_old;
			    temp = ret_old;
			  }
		      }
		      emit_insn (seq);
		  }
	      }
	  }

	if (temp == NULL_RTX)
	  temp = expand_variable_shift (code, mode, op0, treeop1, target,
					unsignedp);
	if (code == LSHIFT_EXPR)
	  temp = REDUCE_BIT_FIELD (temp);
	return temp;
      }

      /* Could determine the answer when only additive constants differ.  Also,
	 the addition of one can be handled by changing the condition.  */
    case LT_EXPR:
    case LE_EXPR:
    case GT_EXPR:
    case GE_EXPR:
    case EQ_EXPR:
    case NE_EXPR:
    case UNORDERED_EXPR:
    case ORDERED_EXPR:
    case UNLT_EXPR:
    case UNLE_EXPR:
    case UNGT_EXPR:
    case UNGE_EXPR:
    case UNEQ_EXPR:
    case LTGT_EXPR:
      {
	temp = do_store_flag (ops,
			      modifier != EXPAND_STACK_PARM ? target : NULL_RTX,
			      tmode != VOIDmode ? tmode : mode);
	if (temp)
	  return temp;

	/* Use a compare and a jump for BLKmode comparisons, or for function
	   type comparisons is have_canonicalize_funcptr_for_compare.  */

	if ((target == 0
	     || modifier == EXPAND_STACK_PARM
	     || ! safe_from_p (target, treeop0, 1)
	     || ! safe_from_p (target, treeop1, 1)
	     /* Make sure we don't have a hard reg (such as function's return
		value) live across basic blocks, if not optimizing.  */
	     || (!optimize && REG_P (target)
		 && REGNO (target) < FIRST_PSEUDO_REGISTER)))
	  target = gen_reg_rtx (tmode != VOIDmode ? tmode : mode);

	emit_move_insn (target, const0_rtx);

	rtx_code_label *lab1 = gen_label_rtx ();
	jumpifnot_1 (code, treeop0, treeop1, lab1,
		     profile_probability::uninitialized ());

	if (TYPE_PRECISION (type) == 1 && !TYPE_UNSIGNED (type))
	  emit_move_insn (target, constm1_rtx);
	else
	  emit_move_insn (target, const1_rtx);

	emit_label (lab1);
	return target;
      }
    case COMPLEX_EXPR:
      /* Get the rtx code of the operands.  */
      op0 = expand_normal (treeop0);
      op1 = expand_normal (treeop1);

      if (!target)
	target = gen_reg_rtx (TYPE_MODE (type));
      else
	/* If target overlaps with op1, then either we need to force
	   op1 into a pseudo (if target also overlaps with op0),
	   or write the complex parts in reverse order.  */
	switch (GET_CODE (target))
	  {
	  case CONCAT:
	    if (reg_overlap_mentioned_p (XEXP (target, 0), op1))
	      {
		if (reg_overlap_mentioned_p (XEXP (target, 1), op0))
		  {
		  complex_expr_force_op1:
		    temp = gen_reg_rtx (GET_MODE_INNER (GET_MODE (target)));
		    emit_move_insn (temp, op1);
		    op1 = temp;
		    break;
		  }
	      complex_expr_swap_order:
		/* Move the imaginary (op1) and real (op0) parts to their
		   location.  */
		write_complex_part (target, op1, true);
		write_complex_part (target, op0, false);

		return target;
	      }
	    break;
	  case MEM:
	    temp = adjust_address_nv (target,
				      GET_MODE_INNER (GET_MODE (target)), 0);
	    if (reg_overlap_mentioned_p (temp, op1))
	      {
		scalar_mode imode = GET_MODE_INNER (GET_MODE (target));
		temp = adjust_address_nv (target, imode,
					  GET_MODE_SIZE (imode));
		if (reg_overlap_mentioned_p (temp, op0))
		  goto complex_expr_force_op1;
		goto complex_expr_swap_order;
	      }
	    break;
	  default:
	    if (reg_overlap_mentioned_p (target, op1))
	      {
		if (reg_overlap_mentioned_p (target, op0))
		  goto complex_expr_force_op1;
		goto complex_expr_swap_order;
	      }
	    break;
	  }

      /* Move the real (op0) and imaginary (op1) parts to their location.  */
      write_complex_part (target, op0, false);
      write_complex_part (target, op1, true);

      return target;

    case WIDEN_SUM_EXPR:
      {
        tree oprnd0 = treeop0;
        tree oprnd1 = treeop1;

        expand_operands (oprnd0, oprnd1, NULL_RTX, &op0, &op1, EXPAND_NORMAL);
        target = expand_widen_pattern_expr (ops, op0, NULL_RTX, op1,
                                            target, unsignedp);
        return target;
      }

    case VEC_UNPACK_HI_EXPR:
    case VEC_UNPACK_LO_EXPR:
    case VEC_UNPACK_FIX_TRUNC_HI_EXPR:
    case VEC_UNPACK_FIX_TRUNC_LO_EXPR:
      {
	op0 = expand_normal (treeop0);
	temp = expand_widen_pattern_expr (ops, op0, NULL_RTX, NULL_RTX,
					  target, unsignedp);
	gcc_assert (temp);
	return temp;
      }

    case VEC_UNPACK_FLOAT_HI_EXPR:
    case VEC_UNPACK_FLOAT_LO_EXPR:
      {
	op0 = expand_normal (treeop0);
	/* The signedness is determined from input operand.  */
	temp = expand_widen_pattern_expr
	  (ops, op0, NULL_RTX, NULL_RTX,
	   target, TYPE_UNSIGNED (TREE_TYPE (treeop0)));

	gcc_assert (temp);
	return temp;
      }

    case VEC_WIDEN_MULT_HI_EXPR:
    case VEC_WIDEN_MULT_LO_EXPR:
    case VEC_WIDEN_MULT_EVEN_EXPR:
    case VEC_WIDEN_MULT_ODD_EXPR:
    case VEC_WIDEN_LSHIFT_HI_EXPR:
    case VEC_WIDEN_LSHIFT_LO_EXPR:
      expand_operands (treeop0, treeop1, NULL_RTX, &op0, &op1, EXPAND_NORMAL);
      target = expand_widen_pattern_expr (ops, op0, op1, NULL_RTX,
					  target, unsignedp);
      gcc_assert (target);
      return target;

    case VEC_PACK_SAT_EXPR:
    case VEC_PACK_FIX_TRUNC_EXPR:
      mode = TYPE_MODE (TREE_TYPE (treeop0));
      goto binop;

    case VEC_PACK_TRUNC_EXPR:
      if (VECTOR_BOOLEAN_TYPE_P (type)
	  && VECTOR_BOOLEAN_TYPE_P (TREE_TYPE (treeop0))
	  && mode == TYPE_MODE (TREE_TYPE (treeop0))
	  && SCALAR_INT_MODE_P (mode))
	{
	  class expand_operand eops[4];
	  machine_mode imode = TYPE_MODE (TREE_TYPE (treeop0));
	  expand_operands (treeop0, treeop1,
			   subtarget, &op0, &op1, EXPAND_NORMAL);
	  this_optab = vec_pack_sbool_trunc_optab;
	  enum insn_code icode = optab_handler (this_optab, imode);
	  create_output_operand (&eops[0], target, mode);
	  create_convert_operand_from (&eops[1], op0, imode, false);
	  create_convert_operand_from (&eops[2], op1, imode, false);
	  temp = GEN_INT (TYPE_VECTOR_SUBPARTS (type).to_constant ());
	  create_input_operand (&eops[3], temp, imode);
	  expand_insn (icode, 4, eops);
	  return eops[0].value;
	}
      mode = TYPE_MODE (TREE_TYPE (treeop0));
      goto binop;

    case VEC_PACK_FLOAT_EXPR:
      mode = TYPE_MODE (TREE_TYPE (treeop0));
      expand_operands (treeop0, treeop1,
		       subtarget, &op0, &op1, EXPAND_NORMAL);
      this_optab = optab_for_tree_code (code, TREE_TYPE (treeop0),
					optab_default);
      target = expand_binop (mode, this_optab, op0, op1, target,
			     TYPE_UNSIGNED (TREE_TYPE (treeop0)),
			     OPTAB_LIB_WIDEN);
      gcc_assert (target);
      return target;

    case VEC_PERM_EXPR:
      {
	expand_operands (treeop0, treeop1, target, &op0, &op1, EXPAND_NORMAL);
	vec_perm_builder sel;
	if (TREE_CODE (treeop2) == VECTOR_CST
	    && tree_to_vec_perm_builder (&sel, treeop2))
	  {
	    machine_mode sel_mode = TYPE_MODE (TREE_TYPE (treeop2));
	    temp = expand_vec_perm_const (mode, op0, op1, sel,
					  sel_mode, target);
	  }
	else
	  {
	    op2 = expand_normal (treeop2);
	    temp = expand_vec_perm_var (mode, op0, op1, op2, target);
	  }
	gcc_assert (temp);
	return temp;
      }

    case DOT_PROD_EXPR:
      {
	tree oprnd0 = treeop0;
	tree oprnd1 = treeop1;
	tree oprnd2 = treeop2;

	expand_operands (oprnd0, oprnd1, NULL_RTX, &op0, &op1, EXPAND_NORMAL);
	op2 = expand_normal (oprnd2);
	target = expand_widen_pattern_expr (ops, op0, op1, op2,
					    target, unsignedp);
	return target;
      }

      case SAD_EXPR:
      {
	tree oprnd0 = treeop0;
	tree oprnd1 = treeop1;
	tree oprnd2 = treeop2;

	expand_operands (oprnd0, oprnd1, NULL_RTX, &op0, &op1, EXPAND_NORMAL);
	op2 = expand_normal (oprnd2);
	target = expand_widen_pattern_expr (ops, op0, op1, op2,
					    target, unsignedp);
	return target;
      }

    case REALIGN_LOAD_EXPR:
      {
        tree oprnd0 = treeop0;
        tree oprnd1 = treeop1;
        tree oprnd2 = treeop2;

        this_optab = optab_for_tree_code (code, type, optab_default);
        expand_operands (oprnd0, oprnd1, NULL_RTX, &op0, &op1, EXPAND_NORMAL);
        op2 = expand_normal (oprnd2);
        temp = expand_ternary_op (mode, this_optab, op0, op1, op2,
				  target, unsignedp);
        gcc_assert (temp);
        return temp;
      }

    case COND_EXPR:
      {
	/* A COND_EXPR with its type being VOID_TYPE represents a
	   conditional jump and is handled in
	   expand_gimple_cond_expr.  */
	gcc_assert (!VOID_TYPE_P (type));

	/* Note that COND_EXPRs whose type is a structure or union
	   are required to be constructed to contain assignments of
	   a temporary variable, so that we can evaluate them here
	   for side effect only.  If type is void, we must do likewise.  */

	gcc_assert (!TREE_ADDRESSABLE (type)
		    && !ignore
		    && TREE_TYPE (treeop1) != void_type_node
		    && TREE_TYPE (treeop2) != void_type_node);

	temp = expand_cond_expr_using_cmove (treeop0, treeop1, treeop2);
	if (temp)
	  return temp;

	/* If we are not to produce a result, we have no target.  Otherwise,
	   if a target was specified use it; it will not be used as an
	   intermediate target unless it is safe.  If no target, use a
	   temporary.  */

	if (modifier != EXPAND_STACK_PARM
	    && original_target
	    && safe_from_p (original_target, treeop0, 1)
	    && GET_MODE (original_target) == mode
	    && !MEM_P (original_target))
	  temp = original_target;
	else
	  temp = assign_temp (type, 0, 1);

	do_pending_stack_adjust ();
	NO_DEFER_POP;
	rtx_code_label *lab0 = gen_label_rtx ();
	rtx_code_label *lab1 = gen_label_rtx ();
	jumpifnot (treeop0, lab0,
		   profile_probability::uninitialized ());
	store_expr (treeop1, temp,
		    modifier == EXPAND_STACK_PARM,
		    false, false);

	emit_jump_insn (targetm.gen_jump (lab1));
	emit_barrier ();
	emit_label (lab0);
	store_expr (treeop2, temp,
		    modifier == EXPAND_STACK_PARM,
		    false, false);

	emit_label (lab1);
	OK_DEFER_POP;
	return temp;
      }

    case VEC_COND_EXPR:
      target = expand_vec_cond_expr (type, treeop0, treeop1, treeop2, target);
      return target;

    case VEC_DUPLICATE_EXPR:
      op0 = expand_expr (treeop0, NULL_RTX, VOIDmode, modifier);
      target = expand_vector_broadcast (mode, op0);
      gcc_assert (target);
      return target;

    case VEC_SERIES_EXPR:
      expand_operands (treeop0, treeop1, NULL_RTX, &op0, &op1, modifier);
      return expand_vec_series_expr (mode, op0, op1, target);

    case BIT_INSERT_EXPR:
      {
	unsigned bitpos = tree_to_uhwi (treeop2);
	unsigned bitsize;
	if (INTEGRAL_TYPE_P (TREE_TYPE (treeop1)))
	  bitsize = TYPE_PRECISION (TREE_TYPE (treeop1));
	else
	  bitsize = tree_to_uhwi (TYPE_SIZE (TREE_TYPE (treeop1)));
	op0 = expand_normal (treeop0);
	op1 = expand_normal (treeop1);
	rtx dst = gen_reg_rtx (mode);
	emit_move_insn (dst, op0);
	store_bit_field (dst, bitsize, bitpos, 0, 0,
			 TYPE_MODE (TREE_TYPE (treeop1)), op1, false);
	return dst;
      }

    default:
      gcc_unreachable ();
    }

  /* Here to do an ordinary binary operator.  */
 binop:
  expand_operands (treeop0, treeop1,
		   subtarget, &op0, &op1, EXPAND_NORMAL);
 binop2:
  this_optab = optab_for_tree_code (code, type, optab_default);
 binop3:
  if (modifier == EXPAND_STACK_PARM)
    target = 0;
  temp = expand_binop (mode, this_optab, op0, op1, target,
		       unsignedp, OPTAB_LIB_WIDEN);
  gcc_assert (temp);
  /* Bitwise operations do not need bitfield reduction as we expect their
     operands being properly truncated.  */
  if (code == BIT_XOR_EXPR
      || code == BIT_AND_EXPR
      || code == BIT_IOR_EXPR)
    return temp;
  return REDUCE_BIT_FIELD (temp);
}
#undef REDUCE_BIT_FIELD


/* Return TRUE if expression STMT is suitable for replacement.
   Never consider memory loads as replaceable, because those don't ever lead
   into constant expressions.  */

static bool
stmt_is_replaceable_p (gimple *stmt)
{
  if (ssa_is_replaceable_p (stmt))
    {
      /* Don't move around loads.  */
      if (!gimple_assign_single_p (stmt)
	  || is_gimple_val (gimple_assign_rhs1 (stmt)))
	return true;
    }
  return false;
}

rtx
expand_expr_real_1 (tree exp, rtx target, machine_mode tmode,
		    enum expand_modifier modifier, rtx *alt_rtl,
		    bool inner_reference_p)
{
  rtx op0, op1, temp, decl_rtl;
  tree type;
  int unsignedp;
  machine_mode mode, dmode;
  enum tree_code code = TREE_CODE (exp);
  rtx subtarget, original_target;
  int ignore;
  tree context;
  bool reduce_bit_field;
  location_t loc = EXPR_LOCATION (exp);
  struct separate_ops ops;
  tree treeop0, treeop1, treeop2;
  tree ssa_name = NULL_TREE;
  gimple *g;

  type = TREE_TYPE (exp);
  mode = TYPE_MODE (type);
  unsignedp = TYPE_UNSIGNED (type);

  treeop0 = treeop1 = treeop2 = NULL_TREE;
  if (!VL_EXP_CLASS_P (exp))
    switch (TREE_CODE_LENGTH (code))
      {
	default:
	case 3: treeop2 = TREE_OPERAND (exp, 2); /* FALLTHRU */
	case 2: treeop1 = TREE_OPERAND (exp, 1); /* FALLTHRU */
	case 1: treeop0 = TREE_OPERAND (exp, 0); /* FALLTHRU */
	case 0: break;
      }
  ops.code = code;
  ops.type = type;
  ops.op0 = treeop0;
  ops.op1 = treeop1;
  ops.op2 = treeop2;
  ops.location = loc;

  ignore = (target == const0_rtx
	    || ((CONVERT_EXPR_CODE_P (code)
		 || code == COND_EXPR || code == VIEW_CONVERT_EXPR)
		&& TREE_CODE (type) == VOID_TYPE));

  /* An operation in what may be a bit-field type needs the
     result to be reduced to the precision of the bit-field type,
     which is narrower than that of the type's mode.  */
  reduce_bit_field = (!ignore
		      && INTEGRAL_TYPE_P (type)
		      && !type_has_mode_precision_p (type));

  /* If we are going to ignore this result, we need only do something
     if there is a side-effect somewhere in the expression.  If there
     is, short-circuit the most common cases here.  Note that we must
     not call expand_expr with anything but const0_rtx in case this
     is an initial expansion of a size that contains a PLACEHOLDER_EXPR.  */

  if (ignore)
    {
      if (! TREE_SIDE_EFFECTS (exp))
	return const0_rtx;

      /* Ensure we reference a volatile object even if value is ignored, but
	 don't do this if all we are doing is taking its address.  */
      if (TREE_THIS_VOLATILE (exp)
	  && TREE_CODE (exp) != FUNCTION_DECL
	  && mode != VOIDmode && mode != BLKmode
	  && modifier != EXPAND_CONST_ADDRESS)
	{
	  temp = expand_expr (exp, NULL_RTX, VOIDmode, modifier);
	  if (MEM_P (temp))
	    copy_to_reg (temp);
	  return const0_rtx;
	}

      if (TREE_CODE_CLASS (code) == tcc_unary
	  || code == BIT_FIELD_REF
	  || code == COMPONENT_REF
	  || code == INDIRECT_REF)
	return expand_expr (treeop0, const0_rtx, VOIDmode,
			    modifier);

      else if (TREE_CODE_CLASS (code) == tcc_binary
	       || TREE_CODE_CLASS (code) == tcc_comparison
	       || code == ARRAY_REF || code == ARRAY_RANGE_REF)
	{
	  expand_expr (treeop0, const0_rtx, VOIDmode, modifier);
	  expand_expr (treeop1, const0_rtx, VOIDmode, modifier);
	  return const0_rtx;
	}

      target = 0;
    }

  if (reduce_bit_field && modifier == EXPAND_STACK_PARM)
    target = 0;

  /* Use subtarget as the target for operand 0 of a binary operation.  */
  subtarget = get_subtarget (target);
  original_target = target;

  switch (code)
    {
    case LABEL_DECL:
      {
	tree function = decl_function_context (exp);

	temp = label_rtx (exp);
	temp = gen_rtx_LABEL_REF (Pmode, temp);

	if (function != current_function_decl
	    && function != 0)
	  LABEL_REF_NONLOCAL_P (temp) = 1;

	temp = gen_rtx_MEM (FUNCTION_MODE, temp);
	return temp;
      }

    case SSA_NAME:
      /* ??? ivopts calls expander, without any preparation from
         out-of-ssa.  So fake instructions as if this was an access to the
	 base variable.  This unnecessarily allocates a pseudo, see how we can
	 reuse it, if partition base vars have it set already.  */
      if (!currently_expanding_to_rtl)
	{
	  tree var = SSA_NAME_VAR (exp);
	  if (var && DECL_RTL_SET_P (var))
	    return DECL_RTL (var);
	  return gen_raw_REG (TYPE_MODE (TREE_TYPE (exp)),
			      LAST_VIRTUAL_REGISTER + 1);
	}

      g = get_gimple_for_ssa_name (exp);
      /* For EXPAND_INITIALIZER try harder to get something simpler.  */
      if (g == NULL
	  && modifier == EXPAND_INITIALIZER
	  && !SSA_NAME_IS_DEFAULT_DEF (exp)
	  && (optimize || !SSA_NAME_VAR (exp)
	      || DECL_IGNORED_P (SSA_NAME_VAR (exp)))
	  && stmt_is_replaceable_p (SSA_NAME_DEF_STMT (exp)))
	g = SSA_NAME_DEF_STMT (exp);
      if (g)
	{
	  rtx r;
	  location_t saved_loc = curr_insn_location ();
	  loc = gimple_location (g);
	  if (loc != UNKNOWN_LOCATION)
	    set_curr_insn_location (loc);
	  ops.code = gimple_assign_rhs_code (g);
          switch (get_gimple_rhs_class (ops.code))
	    {
	    case GIMPLE_TERNARY_RHS:
	      ops.op2 = gimple_assign_rhs3 (g);
	      /* Fallthru */
	    case GIMPLE_BINARY_RHS:
	      ops.op1 = gimple_assign_rhs2 (g);

	      /* Try to expand conditonal compare.  */
	      if (targetm.gen_ccmp_first)
		{
		  gcc_checking_assert (targetm.gen_ccmp_next != NULL);
		  r = expand_ccmp_expr (g, mode);
		  if (r)
		    break;
		}
	      /* Fallthru */
	    case GIMPLE_UNARY_RHS:
	      ops.op0 = gimple_assign_rhs1 (g);
	      ops.type = TREE_TYPE (gimple_assign_lhs (g));
	      ops.location = loc;
	      r = expand_expr_real_2 (&ops, target, tmode, modifier);
	      break;
	    case GIMPLE_SINGLE_RHS:
	      {
		r = expand_expr_real (gimple_assign_rhs1 (g), target,
				      tmode, modifier, alt_rtl,
				      inner_reference_p);
		break;
	      }
	    default:
	      gcc_unreachable ();
	    }
	  set_curr_insn_location (saved_loc);
	  if (REG_P (r) && !REG_EXPR (r))
	    set_reg_attrs_for_decl_rtl (SSA_NAME_VAR (exp), r);
	  return r;
	}

      ssa_name = exp;
      decl_rtl = get_rtx_for_ssa_name (ssa_name);
      exp = SSA_NAME_VAR (ssa_name);
      goto expand_decl_rtl;

    case PARM_DECL:
    case VAR_DECL:
      /* If a static var's type was incomplete when the decl was written,
	 but the type is complete now, lay out the decl now.  */
      if (DECL_SIZE (exp) == 0
	  && COMPLETE_OR_UNBOUND_ARRAY_TYPE_P (TREE_TYPE (exp))
	  && (TREE_STATIC (exp) || DECL_EXTERNAL (exp)))
	layout_decl (exp, 0);

      /* fall through */

    case FUNCTION_DECL:
    case RESULT_DECL:
      decl_rtl = DECL_RTL (exp);
    expand_decl_rtl:
      gcc_assert (decl_rtl);

      /* DECL_MODE might change when TYPE_MODE depends on attribute target
	 settings for VECTOR_TYPE_P that might switch for the function.  */
      if (currently_expanding_to_rtl
	  && code == VAR_DECL && MEM_P (decl_rtl)
	  && VECTOR_TYPE_P (type) && exp && DECL_MODE (exp) != mode)
	decl_rtl = change_address (decl_rtl, TYPE_MODE (type), 0);
      else
	decl_rtl = copy_rtx (decl_rtl);

      /* Record writes to register variables.  */
      if (modifier == EXPAND_WRITE
	  && REG_P (decl_rtl)
	  && HARD_REGISTER_P (decl_rtl))
        add_to_hard_reg_set (&crtl->asm_clobbers,
			     GET_MODE (decl_rtl), REGNO (decl_rtl));

      /* Ensure variable marked as used even if it doesn't go through
	 a parser.  If it hasn't be used yet, write out an external
	 definition.  */
      if (exp)
	TREE_USED (exp) = 1;

      /* Show we haven't gotten RTL for this yet.  */
      temp = 0;

      /* Variables inherited from containing functions should have
	 been lowered by this point.  */
      if (exp)
	context = decl_function_context (exp);
      gcc_assert (!exp
		  || SCOPE_FILE_SCOPE_P (context)
		  || context == current_function_decl
		  || TREE_STATIC (exp)
		  || DECL_EXTERNAL (exp)
		  /* ??? C++ creates functions that are not TREE_STATIC.  */
		  || TREE_CODE (exp) == FUNCTION_DECL);

      /* This is the case of an array whose size is to be determined
	 from its initializer, while the initializer is still being parsed.
	 ??? We aren't parsing while expanding anymore.  */

      if (MEM_P (decl_rtl) && REG_P (XEXP (decl_rtl, 0)))
	temp = validize_mem (decl_rtl);

      /* If DECL_RTL is memory, we are in the normal case and the
	 address is not valid, get the address into a register.  */

      else if (MEM_P (decl_rtl) && modifier != EXPAND_INITIALIZER)
	{
	  if (alt_rtl)
	    *alt_rtl = decl_rtl;
	  decl_rtl = use_anchored_address (decl_rtl);
	  if (modifier != EXPAND_CONST_ADDRESS
	      && modifier != EXPAND_SUM
	      && !memory_address_addr_space_p (exp ? DECL_MODE (exp)
					       : GET_MODE (decl_rtl),
					       XEXP (decl_rtl, 0),
					       MEM_ADDR_SPACE (decl_rtl)))
	    temp = replace_equiv_address (decl_rtl,
					  copy_rtx (XEXP (decl_rtl, 0)));
	}

      /* If we got something, return it.  But first, set the alignment
	 if the address is a register.  */
      if (temp != 0)
	{
	  if (exp && MEM_P (temp) && REG_P (XEXP (temp, 0)))
	    mark_reg_pointer (XEXP (temp, 0), DECL_ALIGN (exp));
	}
      else if (MEM_P (decl_rtl))
	temp = decl_rtl;

      if (temp != 0)
	{
	  if (MEM_P (temp)
	      && modifier != EXPAND_WRITE
	      && modifier != EXPAND_MEMORY
	      && modifier != EXPAND_INITIALIZER
	      && modifier != EXPAND_CONST_ADDRESS
	      && modifier != EXPAND_SUM
	      && !inner_reference_p
	      && mode != BLKmode
	      && MEM_ALIGN (temp) < GET_MODE_ALIGNMENT (mode))
	    temp = expand_misaligned_mem_ref (temp, mode, unsignedp,
					      MEM_ALIGN (temp), NULL_RTX, NULL);

	  return temp;
	}

      if (exp)
	dmode = DECL_MODE (exp);
      else
	dmode = TYPE_MODE (TREE_TYPE (ssa_name));

      /* If the mode of DECL_RTL does not match that of the decl,
	 there are two cases: we are dealing with a BLKmode value
	 that is returned in a register, or we are dealing with
	 a promoted value.  In the latter case, return a SUBREG
	 of the wanted mode, but mark it so that we know that it
	 was already extended.  */
      if (REG_P (decl_rtl)
	  && dmode != BLKmode
	  && GET_MODE (decl_rtl) != dmode)
	{
	  machine_mode pmode;

	  /* Get the signedness to be used for this variable.  Ensure we get
	     the same mode we got when the variable was declared.  */
	  if (code != SSA_NAME)
	    pmode = promote_decl_mode (exp, &unsignedp);
	  else if ((g = SSA_NAME_DEF_STMT (ssa_name))
		   && gimple_code (g) == GIMPLE_CALL
		   && !gimple_call_internal_p (g))
	    pmode = promote_function_mode (type, mode, &unsignedp,
					   gimple_call_fntype (g),
					   2);
	  else
	    pmode = promote_ssa_mode (ssa_name, &unsignedp);
	  gcc_assert (GET_MODE (decl_rtl) == pmode);

	  temp = gen_lowpart_SUBREG (mode, decl_rtl);
	  SUBREG_PROMOTED_VAR_P (temp) = 1;
	  SUBREG_PROMOTED_SET (temp, unsignedp);
	  return temp;
	}

      return decl_rtl;

    case INTEGER_CST:
      {
	/* Given that TYPE_PRECISION (type) is not always equal to
	   GET_MODE_PRECISION (TYPE_MODE (type)), we need to extend from
	   the former to the latter according to the signedness of the
	   type.  */
	scalar_int_mode int_mode = SCALAR_INT_TYPE_MODE (type);
	temp = immed_wide_int_const
	  (wi::to_wide (exp, GET_MODE_PRECISION (int_mode)), int_mode);
	return temp;
      }

    case VECTOR_CST:
      {
	tree tmp = NULL_TREE;
	if (VECTOR_MODE_P (mode))
	  return const_vector_from_tree (exp);
	scalar_int_mode int_mode;
	if (is_int_mode (mode, &int_mode))
	  {
	    if (VECTOR_BOOLEAN_TYPE_P (TREE_TYPE (exp)))
	      return const_scalar_mask_from_tree (int_mode, exp);
	    else
	      {
		tree type_for_mode
		  = lang_hooks.types.type_for_mode (int_mode, 1);
		if (type_for_mode)
		  tmp = fold_unary_loc (loc, VIEW_CONVERT_EXPR,
					type_for_mode, exp);
	      }
	  }
	if (!tmp)
	  {
	    vec<constructor_elt, va_gc> *v;
	    /* Constructors need to be fixed-length.  FIXME.  */
	    unsigned int nunits = VECTOR_CST_NELTS (exp).to_constant ();
	    vec_alloc (v, nunits);
	    for (unsigned int i = 0; i < nunits; ++i)
	      CONSTRUCTOR_APPEND_ELT (v, NULL_TREE, VECTOR_CST_ELT (exp, i));
	    tmp = build_constructor (type, v);
	  }
	return expand_expr (tmp, ignore ? const0_rtx : target,
			    tmode, modifier);
      }

    case CONST_DECL:
      if (modifier == EXPAND_WRITE)
	{
	  /* Writing into CONST_DECL is always invalid, but handle it
	     gracefully.  */
	  addr_space_t as = TYPE_ADDR_SPACE (TREE_TYPE (exp));
	  scalar_int_mode address_mode = targetm.addr_space.address_mode (as);
	  op0 = expand_expr_addr_expr_1 (exp, NULL_RTX, address_mode,
					 EXPAND_NORMAL, as);
	  op0 = memory_address_addr_space (mode, op0, as);
	  temp = gen_rtx_MEM (mode, op0);
	  set_mem_addr_space (temp, as);
	  return temp;
	}
      return expand_expr (DECL_INITIAL (exp), target, VOIDmode, modifier);

    case REAL_CST:
      /* If optimized, generate immediate CONST_DOUBLE
	 which will be turned into memory by reload if necessary.

	 We used to force a register so that loop.c could see it.  But
	 this does not allow gen_* patterns to perform optimizations with
	 the constants.  It also produces two insns in cases like "x = 1.0;".
	 On most machines, floating-point constants are not permitted in
	 many insns, so we'd end up copying it to a register in any case.

	 Now, we do the copying in expand_binop, if appropriate.  */
      return const_double_from_real_value (TREE_REAL_CST (exp),
					   TYPE_MODE (TREE_TYPE (exp)));

    case FIXED_CST:
      return CONST_FIXED_FROM_FIXED_VALUE (TREE_FIXED_CST (exp),
					   TYPE_MODE (TREE_TYPE (exp)));

    case COMPLEX_CST:
      /* Handle evaluating a complex constant in a CONCAT target.  */
      if (original_target && GET_CODE (original_target) == CONCAT)
	{
	  rtx rtarg, itarg;

	  mode = TYPE_MODE (TREE_TYPE (TREE_TYPE (exp)));
	  rtarg = XEXP (original_target, 0);
	  itarg = XEXP (original_target, 1);

	  /* Move the real and imaginary parts separately.  */
	  op0 = expand_expr (TREE_REALPART (exp), rtarg, mode, EXPAND_NORMAL);
	  op1 = expand_expr (TREE_IMAGPART (exp), itarg, mode, EXPAND_NORMAL);

	  if (op0 != rtarg)
	    emit_move_insn (rtarg, op0);
	  if (op1 != itarg)
	    emit_move_insn (itarg, op1);

	  return original_target;
	}

      /* fall through */

    case STRING_CST:
      temp = expand_expr_constant (exp, 1, modifier);

      /* temp contains a constant address.
	 On RISC machines where a constant address isn't valid,
	 make some insns to get that address into a register.  */
      if (modifier != EXPAND_CONST_ADDRESS
	  && modifier != EXPAND_INITIALIZER
	  && modifier != EXPAND_SUM
	  && ! memory_address_addr_space_p (mode, XEXP (temp, 0),
					    MEM_ADDR_SPACE (temp)))
	return replace_equiv_address (temp,
				      copy_rtx (XEXP (temp, 0)));
      return temp;

    case POLY_INT_CST:
      return immed_wide_int_const (poly_int_cst_value (exp), mode);

    case SAVE_EXPR:
      {
	tree val = treeop0;
	rtx ret = expand_expr_real_1 (val, target, tmode, modifier, alt_rtl,
				      inner_reference_p);

	if (!SAVE_EXPR_RESOLVED_P (exp))
	  {
	    /* We can indeed still hit this case, typically via builtin
	       expanders calling save_expr immediately before expanding
	       something.  Assume this means that we only have to deal
	       with non-BLKmode values.  */
	    gcc_assert (GET_MODE (ret) != BLKmode);

	    val = build_decl (curr_insn_location (),
			      VAR_DECL, NULL, TREE_TYPE (exp));
	    DECL_ARTIFICIAL (val) = 1;
	    DECL_IGNORED_P (val) = 1;
	    treeop0 = val;
	    TREE_OPERAND (exp, 0) = treeop0;
	    SAVE_EXPR_RESOLVED_P (exp) = 1;

	    if (!CONSTANT_P (ret))
	      ret = copy_to_reg (ret);
	    SET_DECL_RTL (val, ret);
	  }

        return ret;
      }


    case CONSTRUCTOR:
      /* If we don't need the result, just ensure we evaluate any
	 subexpressions.  */
      if (ignore)
	{
	  unsigned HOST_WIDE_INT idx;
	  tree value;

	  FOR_EACH_CONSTRUCTOR_VALUE (CONSTRUCTOR_ELTS (exp), idx, value)
	    expand_expr (value, const0_rtx, VOIDmode, EXPAND_NORMAL);

	  return const0_rtx;
	}

      return expand_constructor (exp, target, modifier, false);

    case TARGET_MEM_REF:
      {
	addr_space_t as
	  = TYPE_ADDR_SPACE (TREE_TYPE (TREE_TYPE (TREE_OPERAND (exp, 0))));
	unsigned int align;

	op0 = addr_for_mem_ref (exp, as, true);
	op0 = memory_address_addr_space (mode, op0, as);
	temp = gen_rtx_MEM (mode, op0);
	set_mem_attributes (temp, exp, 0);
	set_mem_addr_space (temp, as);
	align = get_object_alignment (exp);
	if (modifier != EXPAND_WRITE
	    && modifier != EXPAND_MEMORY
	    && mode != BLKmode
	    && align < GET_MODE_ALIGNMENT (mode))
	  temp = expand_misaligned_mem_ref (temp, mode, unsignedp,
					    align, NULL_RTX, NULL);
	return temp;
      }

    case MEM_REF:
      {
	const bool reverse = REF_REVERSE_STORAGE_ORDER (exp);
	addr_space_t as
	  = TYPE_ADDR_SPACE (TREE_TYPE (TREE_TYPE (TREE_OPERAND (exp, 0))));
	machine_mode address_mode;
	tree base = TREE_OPERAND (exp, 0);
	gimple *def_stmt;
	unsigned align;
	/* Handle expansion of non-aliased memory with non-BLKmode.  That
	   might end up in a register.  */
	if (mem_ref_refers_to_non_mem_p (exp))
	  {
	    poly_int64 offset = mem_ref_offset (exp).force_shwi ();
	    base = TREE_OPERAND (base, 0);
	    poly_uint64 type_size;
	    if (known_eq (offset, 0)
	        && !reverse
		&& poly_int_tree_p (TYPE_SIZE (type), &type_size)
		&& known_eq (GET_MODE_BITSIZE (DECL_MODE (base)), type_size))
	      return expand_expr (build1 (VIEW_CONVERT_EXPR, type, base),
				  target, tmode, modifier);
	    if (TYPE_MODE (type) == BLKmode)
	      {
		temp = assign_stack_temp (DECL_MODE (base),
					  GET_MODE_SIZE (DECL_MODE (base)));
		store_expr (base, temp, 0, false, false);
		temp = adjust_address (temp, BLKmode, offset);
		set_mem_size (temp, int_size_in_bytes (type));
		return temp;
	      }
	    exp = build3 (BIT_FIELD_REF, type, base, TYPE_SIZE (type),
			  bitsize_int (offset * BITS_PER_UNIT));
	    REF_REVERSE_STORAGE_ORDER (exp) = reverse;
	    return expand_expr (exp, target, tmode, modifier);
	  }
	address_mode = targetm.addr_space.address_mode (as);
	if ((def_stmt = get_def_for_expr (base, BIT_AND_EXPR)))
	  {
	    tree mask = gimple_assign_rhs2 (def_stmt);
	    base = build2 (BIT_AND_EXPR, TREE_TYPE (base),
			   gimple_assign_rhs1 (def_stmt), mask);
	    TREE_OPERAND (exp, 0) = base;
	  }
	align = get_object_alignment (exp);
	op0 = expand_expr (base, NULL_RTX, VOIDmode, EXPAND_SUM);
	op0 = memory_address_addr_space (mode, op0, as);
	if (!integer_zerop (TREE_OPERAND (exp, 1)))
	  {
	    rtx off = immed_wide_int_const (mem_ref_offset (exp), address_mode);
	    op0 = simplify_gen_binary (PLUS, address_mode, op0, off);
	    op0 = memory_address_addr_space (mode, op0, as);
	  }
	temp = gen_rtx_MEM (mode, op0);
	set_mem_attributes (temp, exp, 0);
	set_mem_addr_space (temp, as);
	if (TREE_THIS_VOLATILE (exp))
	  MEM_VOLATILE_P (temp) = 1;
	if (modifier != EXPAND_WRITE
	    && modifier != EXPAND_MEMORY
	    && !inner_reference_p
	    && mode != BLKmode
	    && align < GET_MODE_ALIGNMENT (mode))
	  temp = expand_misaligned_mem_ref (temp, mode, unsignedp, align,
					    modifier == EXPAND_STACK_PARM
					    ? NULL_RTX : target, alt_rtl);
	if (reverse
	    && modifier != EXPAND_MEMORY
	    && modifier != EXPAND_WRITE)
	  temp = flip_storage_order (mode, temp);
	return temp;
      }

    case ARRAY_REF:

      {
	tree array = treeop0;
	tree index = treeop1;
	tree init;

	/* Fold an expression like: "foo"[2].
	   This is not done in fold so it won't happen inside &.
	   Don't fold if this is for wide characters since it's too
	   difficult to do correctly and this is a very rare case.  */

	if (modifier != EXPAND_CONST_ADDRESS
	    && modifier != EXPAND_INITIALIZER
	    && modifier != EXPAND_MEMORY)
	  {
	    tree t = fold_read_from_constant_string (exp);

	    if (t)
	      return expand_expr (t, target, tmode, modifier);
	  }

	/* If this is a constant index into a constant array,
	   just get the value from the array.  Handle both the cases when
	   we have an explicit constructor and when our operand is a variable
	   that was declared const.  */

	if (modifier != EXPAND_CONST_ADDRESS
	    && modifier != EXPAND_INITIALIZER
	    && modifier != EXPAND_MEMORY
	    && TREE_CODE (array) == CONSTRUCTOR
	    && ! TREE_SIDE_EFFECTS (array)
	    && TREE_CODE (index) == INTEGER_CST)
	  {
	    unsigned HOST_WIDE_INT ix;
	    tree field, value;

	    FOR_EACH_CONSTRUCTOR_ELT (CONSTRUCTOR_ELTS (array), ix,
				      field, value)
	      if (tree_int_cst_equal (field, index))
		{
		  if (!TREE_SIDE_EFFECTS (value))
		    return expand_expr (fold (value), target, tmode, modifier);
		  break;
		}
	  }

	else if (optimize >= 1
		 && modifier != EXPAND_CONST_ADDRESS
		 && modifier != EXPAND_INITIALIZER
		 && modifier != EXPAND_MEMORY
		 && TREE_READONLY (array) && ! TREE_SIDE_EFFECTS (array)
		 && TREE_CODE (index) == INTEGER_CST
		 && (VAR_P (array) || TREE_CODE (array) == CONST_DECL)
		 && (init = ctor_for_folding (array)) != error_mark_node)
	  {
	    if (init == NULL_TREE)
	      {
		tree value = build_zero_cst (type);
		if (TREE_CODE (value) == CONSTRUCTOR)
		  {
		    /* If VALUE is a CONSTRUCTOR, this optimization is only
		       useful if this doesn't store the CONSTRUCTOR into
		       memory.  If it does, it is more efficient to just
		       load the data from the array directly.  */
		    rtx ret = expand_constructor (value, target,
						  modifier, true);
		    if (ret == NULL_RTX)
		      value = NULL_TREE;
		  }

		if (value)
		  return expand_expr (value, target, tmode, modifier);
	      }
	    else if (TREE_CODE (init) == CONSTRUCTOR)
	      {
		unsigned HOST_WIDE_INT ix;
		tree field, value;

		FOR_EACH_CONSTRUCTOR_ELT (CONSTRUCTOR_ELTS (init), ix,
					  field, value)
		  if (tree_int_cst_equal (field, index))
		    {
		      if (TREE_SIDE_EFFECTS (value))
			break;

		      if (TREE_CODE (value) == CONSTRUCTOR)
			{
			  /* If VALUE is a CONSTRUCTOR, this
			     optimization is only useful if
			     this doesn't store the CONSTRUCTOR
			     into memory.  If it does, it is more
			     efficient to just load the data from
			     the array directly.  */
			  rtx ret = expand_constructor (value, target,
							modifier, true);
			  if (ret == NULL_RTX)
			    break;
			}

		      return
		        expand_expr (fold (value), target, tmode, modifier);
		    }
	      }
	    else if (TREE_CODE (init) == STRING_CST)
	      {
		tree low_bound = array_ref_low_bound (exp);
		tree index1 = fold_convert_loc (loc, sizetype, treeop1);

		/* Optimize the special case of a zero lower bound.

		   We convert the lower bound to sizetype to avoid problems
		   with constant folding.  E.g. suppose the lower bound is
		   1 and its mode is QI.  Without the conversion
		      (ARRAY + (INDEX - (unsigned char)1))
		   becomes
		      (ARRAY + (-(unsigned char)1) + INDEX)
		   which becomes
		      (ARRAY + 255 + INDEX).  Oops!  */
		if (!integer_zerop (low_bound))
		  index1 = size_diffop_loc (loc, index1,
					    fold_convert_loc (loc, sizetype,
							      low_bound));

		if (tree_fits_uhwi_p (index1)
		    && compare_tree_int (index1, TREE_STRING_LENGTH (init)) < 0)
		  {
		    tree char_type = TREE_TYPE (TREE_TYPE (init));
		    scalar_int_mode char_mode;

		    if (is_int_mode (TYPE_MODE (char_type), &char_mode)
			&& GET_MODE_SIZE (char_mode) == 1)
		      return gen_int_mode (TREE_STRING_POINTER (init)
					   [TREE_INT_CST_LOW (index1)],
					   char_mode);
		  }
	      }
	  }
      }
      goto normal_inner_ref;

    case COMPONENT_REF:
      /* If the operand is a CONSTRUCTOR, we can just extract the
	 appropriate field if it is present.  */
      if (TREE_CODE (treeop0) == CONSTRUCTOR)
	{
	  unsigned HOST_WIDE_INT idx;
	  tree field, value;
	  scalar_int_mode field_mode;

	  FOR_EACH_CONSTRUCTOR_ELT (CONSTRUCTOR_ELTS (treeop0),
				    idx, field, value)
	    if (field == treeop1
		/* We can normally use the value of the field in the
		   CONSTRUCTOR.  However, if this is a bitfield in
		   an integral mode that we can fit in a HOST_WIDE_INT,
		   we must mask only the number of bits in the bitfield,
		   since this is done implicitly by the constructor.  If
		   the bitfield does not meet either of those conditions,
		   we can't do this optimization.  */
		&& (! DECL_BIT_FIELD (field)
		    || (is_int_mode (DECL_MODE (field), &field_mode)
			&& (GET_MODE_PRECISION (field_mode)
			    <= HOST_BITS_PER_WIDE_INT))))
	      {
		if (DECL_BIT_FIELD (field)
		    && modifier == EXPAND_STACK_PARM)
		  target = 0;
		op0 = expand_expr (value, target, tmode, modifier);
		if (DECL_BIT_FIELD (field))
		  {
		    HOST_WIDE_INT bitsize = TREE_INT_CST_LOW (DECL_SIZE (field));
		    scalar_int_mode imode
		      = SCALAR_INT_TYPE_MODE (TREE_TYPE (field));

		    if (TYPE_UNSIGNED (TREE_TYPE (field)))
		      {
			op1 = gen_int_mode ((HOST_WIDE_INT_1 << bitsize) - 1,
					    imode);
			op0 = expand_and (imode, op0, op1, target);
		      }
		    else
		      {
			int count = GET_MODE_PRECISION (imode) - bitsize;

			op0 = expand_shift (LSHIFT_EXPR, imode, op0, count,
					    target, 0);
			op0 = expand_shift (RSHIFT_EXPR, imode, op0, count,
					    target, 0);
		      }
		  }

		return op0;
	      }
	}
      goto normal_inner_ref;

    case BIT_FIELD_REF:
    case ARRAY_RANGE_REF:
    normal_inner_ref:
      {
	machine_mode mode1, mode2;
	poly_int64 bitsize, bitpos, bytepos;
	tree offset;
	int reversep, volatilep = 0, must_force_mem;
	tree tem
	  = get_inner_reference (exp, &bitsize, &bitpos, &offset, &mode1,
				 &unsignedp, &reversep, &volatilep);
	rtx orig_op0, memloc;
	bool clear_mem_expr = false;

	/* If we got back the original object, something is wrong.  Perhaps
	   we are evaluating an expression too early.  In any event, don't
	   infinitely recurse.  */
	gcc_assert (tem != exp);

	/* If TEM's type is a union of variable size, pass TARGET to the inner
	   computation, since it will need a temporary and TARGET is known
	   to have to do.  This occurs in unchecked conversion in Ada.  */
	orig_op0 = op0
	  = expand_expr_real (tem,
			      (TREE_CODE (TREE_TYPE (tem)) == UNION_TYPE
			       && COMPLETE_TYPE_P (TREE_TYPE (tem))
			       && (TREE_CODE (TYPE_SIZE (TREE_TYPE (tem)))
				   != INTEGER_CST)
			       && modifier != EXPAND_STACK_PARM
			       ? target : NULL_RTX),
			      VOIDmode,
			      modifier == EXPAND_SUM ? EXPAND_NORMAL : modifier,
			      NULL, true);

	/* If the field has a mode, we want to access it in the
	   field's mode, not the computed mode.
	   If a MEM has VOIDmode (external with incomplete type),
	   use BLKmode for it instead.  */
	if (MEM_P (op0))
	  {
	    if (mode1 != VOIDmode)
	      op0 = adjust_address (op0, mode1, 0);
	    else if (GET_MODE (op0) == VOIDmode)
	      op0 = adjust_address (op0, BLKmode, 0);
	  }

	mode2
	  = CONSTANT_P (op0) ? TYPE_MODE (TREE_TYPE (tem)) : GET_MODE (op0);

	/* Make sure bitpos is not negative, it can wreak havoc later.  */
	if (maybe_lt (bitpos, 0))
	  {
	    gcc_checking_assert (offset == NULL_TREE);
	    offset = size_int (bits_to_bytes_round_down (bitpos));
	    bitpos = num_trailing_bits (bitpos);
	  }

	/* If we have either an offset, a BLKmode result, or a reference
	   outside the underlying object, we must force it to memory.
	   Such a case can occur in Ada if we have unchecked conversion
	   of an expression from a scalar type to an aggregate type or
	   for an ARRAY_RANGE_REF whose type is BLKmode, or if we were
	   passed a partially uninitialized object or a view-conversion
	   to a larger size.  */
	must_force_mem = (offset
			  || mode1 == BLKmode
			  || (mode == BLKmode
			      && !int_mode_for_size (bitsize, 1).exists ())
			  || maybe_gt (bitpos + bitsize,
				       GET_MODE_BITSIZE (mode2)));

	/* Handle CONCAT first.  */
	if (GET_CODE (op0) == CONCAT && !must_force_mem)
	  {
	    if (known_eq (bitpos, 0)
		&& known_eq (bitsize, GET_MODE_BITSIZE (GET_MODE (op0)))
		&& COMPLEX_MODE_P (mode1)
		&& COMPLEX_MODE_P (GET_MODE (op0))
		&& (GET_MODE_PRECISION (GET_MODE_INNER (mode1))
		    == GET_MODE_PRECISION (GET_MODE_INNER (GET_MODE (op0)))))
	      {
		if (reversep)
		  op0 = flip_storage_order (GET_MODE (op0), op0);
		if (mode1 != GET_MODE (op0))
		  {
		    rtx parts[2];
		    for (int i = 0; i < 2; i++)
		      {
			rtx op = read_complex_part (op0, i != 0);
			if (GET_CODE (op) == SUBREG)
			  op = force_reg (GET_MODE (op), op);
			temp = gen_lowpart_common (GET_MODE_INNER (mode1), op);
			if (temp)
			  op = temp;
			else
			  {
			    if (!REG_P (op) && !MEM_P (op))
			      op = force_reg (GET_MODE (op), op);
			    op = gen_lowpart (GET_MODE_INNER (mode1), op);
			  }
			parts[i] = op;
		      }
		    op0 = gen_rtx_CONCAT (mode1, parts[0], parts[1]);
		  }
		return op0;
	      }
	    if (known_eq (bitpos, 0)
		&& known_eq (bitsize,
			     GET_MODE_BITSIZE (GET_MODE (XEXP (op0, 0))))
		&& maybe_ne (bitsize, 0))
	      {
		op0 = XEXP (op0, 0);
		mode2 = GET_MODE (op0);
	      }
	    else if (known_eq (bitpos,
			       GET_MODE_BITSIZE (GET_MODE (XEXP (op0, 0))))
		     && known_eq (bitsize,
				  GET_MODE_BITSIZE (GET_MODE (XEXP (op0, 1))))
		     && maybe_ne (bitpos, 0)
		     && maybe_ne (bitsize, 0))
	      {
		op0 = XEXP (op0, 1);
		bitpos = 0;
		mode2 = GET_MODE (op0);
	      }
	    else
	      /* Otherwise force into memory.  */
	      must_force_mem = 1;
	  }

	/* If this is a constant, put it in a register if it is a legitimate
	   constant and we don't need a memory reference.  */
	if (CONSTANT_P (op0)
	    && mode2 != BLKmode
	    && targetm.legitimate_constant_p (mode2, op0)
	    && !must_force_mem)
	  op0 = force_reg (mode2, op0);

	/* Otherwise, if this is a constant, try to force it to the constant
	   pool.  Note that back-ends, e.g. MIPS, may refuse to do so if it
	   is a legitimate constant.  */
	else if (CONSTANT_P (op0) && (memloc = force_const_mem (mode2, op0)))
	  op0 = validize_mem (memloc);

	/* Otherwise, if this is a constant or the object is not in memory
	   and need be, put it there.  */
	else if (CONSTANT_P (op0) || (!MEM_P (op0) && must_force_mem))
	  {
	    memloc = assign_temp (TREE_TYPE (tem), 1, 1);
	    emit_move_insn (memloc, op0);
	    op0 = memloc;
	    clear_mem_expr = true;
	  }

	if (offset)
	  {
	    machine_mode address_mode;
	    rtx offset_rtx = expand_expr (offset, NULL_RTX, VOIDmode,
					  EXPAND_SUM);

	    gcc_assert (MEM_P (op0));

	    address_mode = get_address_mode (op0);
	    if (GET_MODE (offset_rtx) != address_mode)
	      {
		/* We cannot be sure that the RTL in offset_rtx is valid outside
		   of a memory address context, so force it into a register
		   before attempting to convert it to the desired mode.  */
		offset_rtx = force_operand (offset_rtx, NULL_RTX);
		offset_rtx = convert_to_mode (address_mode, offset_rtx, 0);
	      }

	    /* See the comment in expand_assignment for the rationale.  */
	    if (mode1 != VOIDmode
		&& maybe_ne (bitpos, 0)
		&& maybe_gt (bitsize, 0)
		&& multiple_p (bitpos, BITS_PER_UNIT, &bytepos)
		&& multiple_p (bitpos, bitsize)
		&& multiple_p (bitsize, GET_MODE_ALIGNMENT (mode1))
		&& MEM_ALIGN (op0) >= GET_MODE_ALIGNMENT (mode1))
	      {
		op0 = adjust_address (op0, mode1, bytepos);
		bitpos = 0;
	      }

	    op0 = offset_address (op0, offset_rtx,
				  highest_pow2_factor (offset));
	  }

	/* If OFFSET is making OP0 more aligned than BIGGEST_ALIGNMENT,
	   record its alignment as BIGGEST_ALIGNMENT.  */
	if (MEM_P (op0)
	    && known_eq (bitpos, 0)
	    && offset != 0
	    && is_aligning_offset (offset, tem))
	  set_mem_align (op0, BIGGEST_ALIGNMENT);

	/* Don't forget about volatility even if this is a bitfield.  */
	if (MEM_P (op0) && volatilep && ! MEM_VOLATILE_P (op0))
	  {
	    if (op0 == orig_op0)
	      op0 = copy_rtx (op0);

	    MEM_VOLATILE_P (op0) = 1;
	  }

	if (MEM_P (op0) && TREE_CODE (tem) == FUNCTION_DECL)
	  {
	    if (op0 == orig_op0)
	      op0 = copy_rtx (op0);

	    set_mem_align (op0, BITS_PER_UNIT);
	  }

	/* In cases where an aligned union has an unaligned object
	   as a field, we might be extracting a BLKmode value from
	   an integer-mode (e.g., SImode) object.  Handle this case
	   by doing the extract into an object as wide as the field
	   (which we know to be the width of a basic mode), then
	   storing into memory, and changing the mode to BLKmode.  */
	if (mode1 == VOIDmode
	    || REG_P (op0) || GET_CODE (op0) == SUBREG
	    || (mode1 != BLKmode && ! direct_load[(int) mode1]
		&& GET_MODE_CLASS (mode) != MODE_COMPLEX_INT
		&& GET_MODE_CLASS (mode) != MODE_COMPLEX_FLOAT
		&& modifier != EXPAND_CONST_ADDRESS
		&& modifier != EXPAND_INITIALIZER
		&& modifier != EXPAND_MEMORY)
	    /* If the bitfield is volatile and the bitsize
	       is narrower than the access size of the bitfield,
	       we need to extract bitfields from the access.  */
	    || (volatilep && TREE_CODE (exp) == COMPONENT_REF
		&& DECL_BIT_FIELD_TYPE (TREE_OPERAND (exp, 1))
		&& mode1 != BLKmode
		&& maybe_lt (bitsize, GET_MODE_SIZE (mode1) * BITS_PER_UNIT))
	    /* If the field isn't aligned enough to fetch as a memref,
	       fetch it as a bit field.  */
	    || (mode1 != BLKmode
		&& (((MEM_P (op0)
		      ? MEM_ALIGN (op0) < GET_MODE_ALIGNMENT (mode1)
			|| !multiple_p (bitpos, GET_MODE_ALIGNMENT (mode1))
		      : TYPE_ALIGN (TREE_TYPE (tem)) < GET_MODE_ALIGNMENT (mode)
			|| !multiple_p (bitpos, GET_MODE_ALIGNMENT (mode)))
		     && modifier != EXPAND_MEMORY
		     && ((modifier == EXPAND_CONST_ADDRESS
			  || modifier == EXPAND_INITIALIZER)
			 ? STRICT_ALIGNMENT
			 : targetm.slow_unaligned_access (mode1,
							  MEM_ALIGN (op0))))
		    || !multiple_p (bitpos, BITS_PER_UNIT)))
	    /* If the type and the field are a constant size and the
	       size of the type isn't the same size as the bitfield,
	       we must use bitfield operations.  */
	    || (known_size_p (bitsize)
		&& TYPE_SIZE (TREE_TYPE (exp))
		&& poly_int_tree_p (TYPE_SIZE (TREE_TYPE (exp)))
		&& maybe_ne (wi::to_poly_offset (TYPE_SIZE (TREE_TYPE (exp))),
			     bitsize)))
	  {
	    machine_mode ext_mode = mode;

	    if (ext_mode == BLKmode
		&& ! (target != 0 && MEM_P (op0)
		      && MEM_P (target)
		      && multiple_p (bitpos, BITS_PER_UNIT)))
	      ext_mode = int_mode_for_size (bitsize, 1).else_blk ();

	    if (ext_mode == BLKmode)
	      {
		if (target == 0)
		  target = assign_temp (type, 1, 1);

		/* ??? Unlike the similar test a few lines below, this one is
		   very likely obsolete.  */
		if (known_eq (bitsize, 0))
		  return target;

		/* In this case, BITPOS must start at a byte boundary and
		   TARGET, if specified, must be a MEM.  */
		gcc_assert (MEM_P (op0)
			    && (!target || MEM_P (target)));

		bytepos = exact_div (bitpos, BITS_PER_UNIT);
		poly_int64 bytesize = bits_to_bytes_round_up (bitsize);
		emit_block_move (target,
				 adjust_address (op0, VOIDmode, bytepos),
				 gen_int_mode (bytesize, Pmode),
				 (modifier == EXPAND_STACK_PARM
				  ? BLOCK_OP_CALL_PARM : BLOCK_OP_NORMAL));

		return target;
	      }

	    /* If we have nothing to extract, the result will be 0 for targets
	       with SHIFT_COUNT_TRUNCATED == 0 and garbage otherwise.  Always
	       return 0 for the sake of consistency, as reading a zero-sized
	       bitfield is valid in Ada and the value is fully specified.  */
	    if (known_eq (bitsize, 0))
	      return const0_rtx;

	    op0 = validize_mem (op0);

	    if (MEM_P (op0) && REG_P (XEXP (op0, 0)))
	      mark_reg_pointer (XEXP (op0, 0), MEM_ALIGN (op0));

	    /* If the result has aggregate type and the extraction is done in
	       an integral mode, then the field may be not aligned on a byte
	       boundary; in this case, if it has reverse storage order, it
	       needs to be extracted as a scalar field with reverse storage
	       order and put back into memory order afterwards.  */
	    if (AGGREGATE_TYPE_P (type)
		&& GET_MODE_CLASS (ext_mode) == MODE_INT)
	      reversep = TYPE_REVERSE_STORAGE_ORDER (type);

	    gcc_checking_assert (known_ge (bitpos, 0));
	    op0 = extract_bit_field (op0, bitsize, bitpos, unsignedp,
				     (modifier == EXPAND_STACK_PARM
				      ? NULL_RTX : target),
				     ext_mode, ext_mode, reversep, alt_rtl);

	    /* If the result has aggregate type and the mode of OP0 is an
	       integral mode then, if BITSIZE is narrower than this mode
	       and this is for big-endian data, we must put the field
	       into the high-order bits.  And we must also put it back
	       into memory order if it has been previously reversed.  */
	    scalar_int_mode op0_mode;
	    if (AGGREGATE_TYPE_P (type)
		&& is_int_mode (GET_MODE (op0), &op0_mode))
	      {
		HOST_WIDE_INT size = GET_MODE_BITSIZE (op0_mode);

		gcc_checking_assert (known_le (bitsize, size));
		if (maybe_lt (bitsize, size)
		    && reversep ? !BYTES_BIG_ENDIAN : BYTES_BIG_ENDIAN)
		  op0 = expand_shift (LSHIFT_EXPR, op0_mode, op0,
				      size - bitsize, op0, 1);

		if (reversep)
		  op0 = flip_storage_order (op0_mode, op0);
	      }

	    /* If the result type is BLKmode, store the data into a temporary
	       of the appropriate type, but with the mode corresponding to the
	       mode for the data we have (op0's mode).  */
	    if (mode == BLKmode)
	      {
		rtx new_rtx
		  = assign_stack_temp_for_type (ext_mode,
						GET_MODE_BITSIZE (ext_mode),
						type);
		emit_move_insn (new_rtx, op0);
		op0 = copy_rtx (new_rtx);
		PUT_MODE (op0, BLKmode);
	      }

	    return op0;
	  }

	/* If the result is BLKmode, use that to access the object
	   now as well.  */
	if (mode == BLKmode)
	  mode1 = BLKmode;

	/* Get a reference to just this component.  */
	bytepos = bits_to_bytes_round_down (bitpos);
	if (modifier == EXPAND_CONST_ADDRESS
	    || modifier == EXPAND_SUM || modifier == EXPAND_INITIALIZER)
	  op0 = adjust_address_nv (op0, mode1, bytepos);
	else
	  op0 = adjust_address (op0, mode1, bytepos);

	if (op0 == orig_op0)
	  op0 = copy_rtx (op0);

	/* Don't set memory attributes if the base expression is
	   SSA_NAME that got expanded as a MEM or a CONSTANT.  In that case,
	   we should just honor its original memory attributes.  */
	if (!(TREE_CODE (tem) == SSA_NAME
	      && (MEM_P (orig_op0) || CONSTANT_P (orig_op0))))
	  set_mem_attributes (op0, exp, 0);

	if (REG_P (XEXP (op0, 0)))
	  mark_reg_pointer (XEXP (op0, 0), MEM_ALIGN (op0));

	/* If op0 is a temporary because the original expressions was forced
	   to memory, clear MEM_EXPR so that the original expression cannot
	   be marked as addressable through MEM_EXPR of the temporary.  */
	if (clear_mem_expr)
	  set_mem_expr (op0, NULL_TREE);

	MEM_VOLATILE_P (op0) |= volatilep;

        if (reversep
	    && modifier != EXPAND_MEMORY
	    && modifier != EXPAND_WRITE)
	  op0 = flip_storage_order (mode1, op0);

	if (mode == mode1 || mode1 == BLKmode || mode1 == tmode
	    || modifier == EXPAND_CONST_ADDRESS
	    || modifier == EXPAND_INITIALIZER)
	  return op0;

	if (target == 0)
	  target = gen_reg_rtx (tmode != VOIDmode ? tmode : mode);

	convert_move (target, op0, unsignedp);
	return target;
      }

    case OBJ_TYPE_REF:
      return expand_expr (OBJ_TYPE_REF_EXPR (exp), target, tmode, modifier);

    case CALL_EXPR:
      /* All valid uses of __builtin_va_arg_pack () are removed during
	 inlining.  */
      if (CALL_EXPR_VA_ARG_PACK (exp))
	error ("%Kinvalid use of %<__builtin_va_arg_pack ()%>", exp);
      {
	tree fndecl = get_callee_fndecl (exp), attr;

	if (fndecl
	    /* Don't diagnose the error attribute in thunks, those are
	       artificially created.  */
	    && !CALL_FROM_THUNK_P (exp)
	    && (attr = lookup_attribute ("error",
					 DECL_ATTRIBUTES (fndecl))) != NULL)
	  {
	    const char *ident = lang_hooks.decl_printable_name (fndecl, 1);
	    error ("%Kcall to %qs declared with attribute error: %s", exp,
		   identifier_to_locale (ident),
		   TREE_STRING_POINTER (TREE_VALUE (TREE_VALUE (attr))));
	  }
	if (fndecl
	    /* Don't diagnose the warning attribute in thunks, those are
	       artificially created.  */
	    && !CALL_FROM_THUNK_P (exp)
	    && (attr = lookup_attribute ("warning",
					 DECL_ATTRIBUTES (fndecl))) != NULL)
	  {
	    const char *ident = lang_hooks.decl_printable_name (fndecl, 1);
	    warning_at (tree_nonartificial_location (exp),
			OPT_Wattribute_warning,
			"%Kcall to %qs declared with attribute warning: %s",
			exp, identifier_to_locale (ident),
			TREE_STRING_POINTER (TREE_VALUE (TREE_VALUE (attr))));
	  }

	/* Check for a built-in function.  */
	if (fndecl && fndecl_built_in_p (fndecl))
	  {
	    gcc_assert (DECL_BUILT_IN_CLASS (fndecl) != BUILT_IN_FRONTEND);
	    return expand_builtin (exp, target, subtarget, tmode, ignore);
	  }
      }
      return expand_call (exp, target, ignore);

    case VIEW_CONVERT_EXPR:
      op0 = NULL_RTX;

      /* If we are converting to BLKmode, try to avoid an intermediate
	 temporary by fetching an inner memory reference.  */
      if (mode == BLKmode
	  && poly_int_tree_p (TYPE_SIZE (type))
	  && TYPE_MODE (TREE_TYPE (treeop0)) != BLKmode
	  && handled_component_p (treeop0))
      {
	machine_mode mode1;
	poly_int64 bitsize, bitpos, bytepos;
	tree offset;
	int reversep, volatilep = 0;
	tree tem
	  = get_inner_reference (treeop0, &bitsize, &bitpos, &offset, &mode1,
				 &unsignedp, &reversep, &volatilep);

	/* ??? We should work harder and deal with non-zero offsets.  */
	if (!offset
	    && multiple_p (bitpos, BITS_PER_UNIT, &bytepos)
	    && !reversep
	    && known_size_p (bitsize)
	    && known_eq (wi::to_poly_offset (TYPE_SIZE (type)), bitsize))
	  {
	    /* See the normal_inner_ref case for the rationale.  */
	    rtx orig_op0
	      = expand_expr_real (tem,
				  (TREE_CODE (TREE_TYPE (tem)) == UNION_TYPE
				   && (TREE_CODE (TYPE_SIZE (TREE_TYPE (tem)))
				       != INTEGER_CST)
				   && modifier != EXPAND_STACK_PARM
				   ? target : NULL_RTX),
				  VOIDmode,
				  modifier == EXPAND_SUM ? EXPAND_NORMAL : modifier,
				  NULL, true);

	    if (MEM_P (orig_op0))
	      {
		op0 = orig_op0;

		/* Get a reference to just this component.  */
		if (modifier == EXPAND_CONST_ADDRESS
		    || modifier == EXPAND_SUM
		    || modifier == EXPAND_INITIALIZER)
		  op0 = adjust_address_nv (op0, mode, bytepos);
		else
		  op0 = adjust_address (op0, mode, bytepos);

		if (op0 == orig_op0)
		  op0 = copy_rtx (op0);

		set_mem_attributes (op0, treeop0, 0);
		if (REG_P (XEXP (op0, 0)))
		  mark_reg_pointer (XEXP (op0, 0), MEM_ALIGN (op0));

		MEM_VOLATILE_P (op0) |= volatilep;
	      }
	  }
      }

      if (!op0)
	op0 = expand_expr_real (treeop0, NULL_RTX, VOIDmode, modifier,
				NULL, inner_reference_p);

      /* If the input and output modes are both the same, we are done.  */
      if (mode == GET_MODE (op0))
	;
      /* If neither mode is BLKmode, and both modes are the same size
	 then we can use gen_lowpart.  */
      else if (mode != BLKmode
	       && GET_MODE (op0) != BLKmode
	       && known_eq (GET_MODE_PRECISION (mode),
			    GET_MODE_PRECISION (GET_MODE (op0)))
	       && !COMPLEX_MODE_P (GET_MODE (op0)))
	{
	  if (GET_CODE (op0) == SUBREG)
	    op0 = force_reg (GET_MODE (op0), op0);
	  temp = gen_lowpart_common (mode, op0);
	  if (temp)
	    op0 = temp;
	  else
	    {
	      if (!REG_P (op0) && !MEM_P (op0))
		op0 = force_reg (GET_MODE (op0), op0);
	      op0 = gen_lowpart (mode, op0);
	    }
	}
      /* If both types are integral, convert from one mode to the other.  */
      else if (INTEGRAL_TYPE_P (type) && INTEGRAL_TYPE_P (TREE_TYPE (treeop0)))
	op0 = convert_modes (mode, GET_MODE (op0), op0,
			     TYPE_UNSIGNED (TREE_TYPE (treeop0)));
      /* If the output type is a bit-field type, do an extraction.  */
      else if (reduce_bit_field)
	return extract_bit_field (op0, TYPE_PRECISION (type), 0,
				  TYPE_UNSIGNED (type), NULL_RTX,
				  mode, mode, false, NULL);
      /* As a last resort, spill op0 to memory, and reload it in a
	 different mode.  */
      else if (!MEM_P (op0))
	{
	  /* If the operand is not a MEM, force it into memory.  Since we
	     are going to be changing the mode of the MEM, don't call
	     force_const_mem for constants because we don't allow pool
	     constants to change mode.  */
	  tree inner_type = TREE_TYPE (treeop0);

	  gcc_assert (!TREE_ADDRESSABLE (exp));

	  if (target == 0 || GET_MODE (target) != TYPE_MODE (inner_type))
	    target
	      = assign_stack_temp_for_type
		(TYPE_MODE (inner_type),
		 GET_MODE_SIZE (TYPE_MODE (inner_type)), inner_type);

	  emit_move_insn (target, op0);
	  op0 = target;
	}

      /* If OP0 is (now) a MEM, we need to deal with alignment issues.  If the
	 output type is such that the operand is known to be aligned, indicate
	 that it is.  Otherwise, we need only be concerned about alignment for
	 non-BLKmode results.  */
      if (MEM_P (op0))
	{
	  enum insn_code icode;

	  if (modifier != EXPAND_WRITE
	      && modifier != EXPAND_MEMORY
	      && !inner_reference_p
	      && mode != BLKmode
	      && MEM_ALIGN (op0) < GET_MODE_ALIGNMENT (mode))
	    {
	      /* If the target does have special handling for unaligned
		 loads of mode then use them.  */
	      if ((icode = optab_handler (movmisalign_optab, mode))
		  != CODE_FOR_nothing)
		{
		  rtx reg;

		  op0 = adjust_address (op0, mode, 0);
		  /* We've already validated the memory, and we're creating a
		     new pseudo destination.  The predicates really can't
		     fail.  */
		  reg = gen_reg_rtx (mode);

		  /* Nor can the insn generator.  */
		  rtx_insn *insn = GEN_FCN (icode) (reg, op0);
		  emit_insn (insn);
		  return reg;
		}
	      else if (STRICT_ALIGNMENT)
		{
		  poly_uint64 mode_size = GET_MODE_SIZE (mode);
		  poly_uint64 temp_size = mode_size;
		  if (GET_MODE (op0) != BLKmode)
		    temp_size = upper_bound (temp_size,
					     GET_MODE_SIZE (GET_MODE (op0)));
		  rtx new_rtx
		    = assign_stack_temp_for_type (mode, temp_size, type);
		  rtx new_with_op0_mode
		    = adjust_address (new_rtx, GET_MODE (op0), 0);

		  gcc_assert (!TREE_ADDRESSABLE (exp));

		  if (GET_MODE (op0) == BLKmode)
		    {
		      rtx size_rtx = gen_int_mode (mode_size, Pmode);
		      emit_block_move (new_with_op0_mode, op0, size_rtx,
				       (modifier == EXPAND_STACK_PARM
					? BLOCK_OP_CALL_PARM
					: BLOCK_OP_NORMAL));
		    }
		  else
		    emit_move_insn (new_with_op0_mode, op0);

		  op0 = new_rtx;
		}
	    }

	  op0 = adjust_address (op0, mode, 0);
	}

      return op0;

    case MODIFY_EXPR:
      {
	tree lhs = treeop0;
	tree rhs = treeop1;
	gcc_assert (ignore);

	/* Check for |= or &= of a bitfield of size one into another bitfield
	   of size 1.  In this case, (unless we need the result of the
	   assignment) we can do this more efficiently with a
	   test followed by an assignment, if necessary.

	   ??? At this point, we can't get a BIT_FIELD_REF here.  But if
	   things change so we do, this code should be enhanced to
	   support it.  */
	if (TREE_CODE (lhs) == COMPONENT_REF
	    && (TREE_CODE (rhs) == BIT_IOR_EXPR
		|| TREE_CODE (rhs) == BIT_AND_EXPR)
	    && TREE_OPERAND (rhs, 0) == lhs
	    && TREE_CODE (TREE_OPERAND (rhs, 1)) == COMPONENT_REF
	    && integer_onep (DECL_SIZE (TREE_OPERAND (lhs, 1)))
	    && integer_onep (DECL_SIZE (TREE_OPERAND (TREE_OPERAND (rhs, 1), 1))))
	  {
	    rtx_code_label *label = gen_label_rtx ();
	    int value = TREE_CODE (rhs) == BIT_IOR_EXPR;
	    profile_probability prob = profile_probability::uninitialized ();
 	    if (value)
 	      jumpifnot (TREE_OPERAND (rhs, 1), label, prob);
 	    else
 	      jumpif (TREE_OPERAND (rhs, 1), label, prob);
	    expand_assignment (lhs, build_int_cst (TREE_TYPE (rhs), value),
			       false);
	    do_pending_stack_adjust ();
	    emit_label (label);
	    return const0_rtx;
	  }

	expand_assignment (lhs, rhs, false);
	return const0_rtx;
      }

    case ADDR_EXPR:
      return expand_expr_addr_expr (exp, target, tmode, modifier);

    case REALPART_EXPR:
      op0 = expand_normal (treeop0);
      return read_complex_part (op0, false);

    case IMAGPART_EXPR:
      op0 = expand_normal (treeop0);
      return read_complex_part (op0, true);

    case RETURN_EXPR:
    case LABEL_EXPR:
    case GOTO_EXPR:
    case SWITCH_EXPR:
    case ASM_EXPR:
      /* Expanded in cfgexpand.c.  */
      gcc_unreachable ();

    case TRY_CATCH_EXPR:
    case CATCH_EXPR:
    case EH_FILTER_EXPR:
    case TRY_FINALLY_EXPR:
    case EH_ELSE_EXPR:
      /* Lowered by tree-eh.c.  */
      gcc_unreachable ();

    case WITH_CLEANUP_EXPR:
    case CLEANUP_POINT_EXPR:
    case TARGET_EXPR:
    case CASE_LABEL_EXPR:
    case VA_ARG_EXPR:
    case BIND_EXPR:
    case INIT_EXPR:
    case CONJ_EXPR:
    case COMPOUND_EXPR:
    case PREINCREMENT_EXPR:
    case PREDECREMENT_EXPR:
    case POSTINCREMENT_EXPR:
    case POSTDECREMENT_EXPR:
    case LOOP_EXPR:
    case EXIT_EXPR:
    case COMPOUND_LITERAL_EXPR:
      /* Lowered by gimplify.c.  */
      gcc_unreachable ();

    case FDESC_EXPR:
      /* Function descriptors are not valid except for as
	 initialization constants, and should not be expanded.  */
      gcc_unreachable ();

    case WITH_SIZE_EXPR:
      /* WITH_SIZE_EXPR expands to its first argument.  The caller should
	 have pulled out the size to use in whatever context it needed.  */
      return expand_expr_real (treeop0, original_target, tmode,
			       modifier, alt_rtl, inner_reference_p);

    default:
      return expand_expr_real_2 (&ops, target, tmode, modifier);
    }
}

/* Subroutine of above: reduce EXP to the precision of TYPE (in the
   signedness of TYPE), possibly returning the result in TARGET.
   TYPE is known to be a partial integer type.  */
static rtx
reduce_to_bit_field_precision (rtx exp, rtx target, tree type)
{
  HOST_WIDE_INT prec = TYPE_PRECISION (type);
  if (target && GET_MODE (target) != GET_MODE (exp))
    target = 0;
  /* For constant values, reduce using build_int_cst_type. */
  poly_int64 const_exp;
  if (poly_int_rtx_p (exp, &const_exp))
    {
      tree t = build_int_cst_type (type, const_exp);
      return expand_expr (t, target, VOIDmode, EXPAND_NORMAL);
    }
  else if (TYPE_UNSIGNED (type))
    {
      scalar_int_mode mode = as_a <scalar_int_mode> (GET_MODE (exp));
      rtx mask = immed_wide_int_const
	(wi::mask (prec, false, GET_MODE_PRECISION (mode)), mode);
      return expand_and (mode, exp, mask, target);
    }
  else
    {
      scalar_int_mode mode = as_a <scalar_int_mode> (GET_MODE (exp));
      int count = GET_MODE_PRECISION (mode) - prec;
      exp = expand_shift (LSHIFT_EXPR, mode, exp, count, target, 0);
      return expand_shift (RSHIFT_EXPR, mode, exp, count, target, 0);
    }
}

/* Subroutine of above: returns 1 if OFFSET corresponds to an offset that
   when applied to the address of EXP produces an address known to be
   aligned more than BIGGEST_ALIGNMENT.  */

static int
is_aligning_offset (const_tree offset, const_tree exp)
{
  /* Strip off any conversions.  */
  while (CONVERT_EXPR_P (offset))
    offset = TREE_OPERAND (offset, 0);

  /* We must now have a BIT_AND_EXPR with a constant that is one less than
     power of 2 and which is larger than BIGGEST_ALIGNMENT.  */
  if (TREE_CODE (offset) != BIT_AND_EXPR
      || !tree_fits_uhwi_p (TREE_OPERAND (offset, 1))
      || compare_tree_int (TREE_OPERAND (offset, 1),
			   BIGGEST_ALIGNMENT / BITS_PER_UNIT) <= 0
      || !pow2p_hwi (tree_to_uhwi (TREE_OPERAND (offset, 1)) + 1))
    return 0;

  /* Look at the first operand of BIT_AND_EXPR and strip any conversion.
     It must be NEGATE_EXPR.  Then strip any more conversions.  */
  offset = TREE_OPERAND (offset, 0);
  while (CONVERT_EXPR_P (offset))
    offset = TREE_OPERAND (offset, 0);

  if (TREE_CODE (offset) != NEGATE_EXPR)
    return 0;

  offset = TREE_OPERAND (offset, 0);
  while (CONVERT_EXPR_P (offset))
    offset = TREE_OPERAND (offset, 0);

  /* This must now be the address of EXP.  */
  return TREE_CODE (offset) == ADDR_EXPR && TREE_OPERAND (offset, 0) == exp;
}

/* Return the tree node if an ARG corresponds to a string constant or zero
   if it doesn't.  If we return nonzero, set *PTR_OFFSET to the (possibly
   non-constant) offset in bytes within the string that ARG is accessing.
   If MEM_SIZE is non-zero the storage size of the memory is returned.
   If DECL is non-zero the constant declaration is returned if available.  */

tree
string_constant (tree arg, tree *ptr_offset, tree *mem_size, tree *decl)
{
  tree dummy = NULL_TREE;;
  if (!mem_size)
    mem_size = &dummy;

  /* Store the type of the original expression before conversions
     via NOP_EXPR or POINTER_PLUS_EXPR to other types have been
     removed.  */
  tree argtype = TREE_TYPE (arg);

  tree array;
  STRIP_NOPS (arg);

  /* Non-constant index into the character array in an ARRAY_REF
     expression or null.  */
  tree varidx = NULL_TREE;

  poly_int64 base_off = 0;

  if (TREE_CODE (arg) == ADDR_EXPR)
    {
      arg = TREE_OPERAND (arg, 0);
      tree ref = arg;
      if (TREE_CODE (arg) == ARRAY_REF)
	{
	  tree idx = TREE_OPERAND (arg, 1);
	  if (TREE_CODE (idx) != INTEGER_CST)
	    {
	      /* From a pointer (but not array) argument extract the variable
		 index to prevent get_addr_base_and_unit_offset() from failing
		 due to it.  Use it later to compute the non-constant offset
		 into the string and return it to the caller.  */
	      varidx = idx;
	      ref = TREE_OPERAND (arg, 0);

	      if (TREE_CODE (TREE_TYPE (arg)) == ARRAY_TYPE)
		return NULL_TREE;

	      if (!integer_zerop (array_ref_low_bound (arg)))
		return NULL_TREE;

	      if (!integer_onep (array_ref_element_size (arg)))
		return NULL_TREE;
	    }
	}
      array = get_addr_base_and_unit_offset (ref, &base_off);
      if (!array
	  || (TREE_CODE (array) != VAR_DECL
	      && TREE_CODE (array) != CONST_DECL
	      && TREE_CODE (array) != STRING_CST))
	return NULL_TREE;
    }
  else if (TREE_CODE (arg) == PLUS_EXPR || TREE_CODE (arg) == POINTER_PLUS_EXPR)
    {
      tree arg0 = TREE_OPERAND (arg, 0);
      tree arg1 = TREE_OPERAND (arg, 1);

      tree offset;
      tree str = string_constant (arg0, &offset, mem_size, decl);
      if (!str)
	{
	   str = string_constant (arg1, &offset, mem_size, decl);
	   arg1 = arg0;
	}

      if (str)
	{
	  /* Avoid pointers to arrays (see bug 86622).  */
	  if (POINTER_TYPE_P (TREE_TYPE (arg))
	      && TREE_CODE (TREE_TYPE (TREE_TYPE (arg))) == ARRAY_TYPE
	      && !(decl && !*decl)
	      && !(decl && tree_fits_uhwi_p (DECL_SIZE_UNIT (*decl))
		   && tree_fits_uhwi_p (*mem_size)
		   && tree_int_cst_equal (*mem_size, DECL_SIZE_UNIT (*decl))))
	    return NULL_TREE;

	  tree type = TREE_TYPE (offset);
	  arg1 = fold_convert (type, arg1);
	  *ptr_offset = fold_build2 (PLUS_EXPR, type, offset, arg1);
	  return str;
	}
      return NULL_TREE;
    }
  else if (TREE_CODE (arg) == SSA_NAME)
    {
      gimple *stmt = SSA_NAME_DEF_STMT (arg);
      if (!is_gimple_assign (stmt))
	return NULL_TREE;

      tree rhs1 = gimple_assign_rhs1 (stmt);
      tree_code code = gimple_assign_rhs_code (stmt);
      if (code == ADDR_EXPR)
	return string_constant (rhs1, ptr_offset, mem_size, decl);
      else if (code != POINTER_PLUS_EXPR)
	return NULL_TREE;

      tree offset;
      if (tree str = string_constant (rhs1, &offset, mem_size, decl))
	{
	  /* Avoid pointers to arrays (see bug 86622).  */
	  if (POINTER_TYPE_P (TREE_TYPE (rhs1))
	      && TREE_CODE (TREE_TYPE (TREE_TYPE (rhs1))) == ARRAY_TYPE
	      && !(decl && !*decl)
	      && !(decl && tree_fits_uhwi_p (DECL_SIZE_UNIT (*decl))
		   && tree_fits_uhwi_p (*mem_size)
		   && tree_int_cst_equal (*mem_size, DECL_SIZE_UNIT (*decl))))
	    return NULL_TREE;

	  tree rhs2 = gimple_assign_rhs2 (stmt);
	  tree type = TREE_TYPE (offset);
	  rhs2 = fold_convert (type, rhs2);
	  *ptr_offset = fold_build2 (PLUS_EXPR, type, offset, rhs2);
	  return str;
	}
      return NULL_TREE;
    }
  else if (DECL_P (arg))
    array = arg;
  else
    return NULL_TREE;

  tree offset = wide_int_to_tree (sizetype, base_off);
  if (varidx)
    {
      if (TREE_CODE (TREE_TYPE (array)) != ARRAY_TYPE)
	return NULL_TREE;

      gcc_assert (TREE_CODE (arg) == ARRAY_REF);
      tree chartype = TREE_TYPE (TREE_TYPE (TREE_OPERAND (arg, 0)));
      if (TREE_CODE (chartype) != INTEGER_TYPE)
	return NULL;

      offset = fold_convert (sizetype, varidx);
    }

  if (TREE_CODE (array) == STRING_CST)
    {
      *ptr_offset = fold_convert (sizetype, offset);
      *mem_size = TYPE_SIZE_UNIT (TREE_TYPE (array));
      if (decl)
	*decl = NULL_TREE;
      gcc_checking_assert (tree_to_shwi (TYPE_SIZE_UNIT (TREE_TYPE (array)))
			   >= TREE_STRING_LENGTH (array));
      return array;
    }

  if (!VAR_P (array) && TREE_CODE (array) != CONST_DECL)
    return NULL_TREE;

  tree init = ctor_for_folding (array);

  /* Handle variables initialized with string literals.  */
  if (!init || init == error_mark_node)
    return NULL_TREE;
  if (TREE_CODE (init) == CONSTRUCTOR)
    {
      /* Convert the 64-bit constant offset to a wider type to avoid
	 overflow.  */
      offset_int wioff;
      if (!base_off.is_constant (&wioff))
	return NULL_TREE;

      wioff *= BITS_PER_UNIT;
      if (!wi::fits_uhwi_p (wioff))
	return NULL_TREE;

      base_off = wioff.to_uhwi ();
      unsigned HOST_WIDE_INT fieldoff = 0;
      init = fold_ctor_reference (TREE_TYPE (arg), init, base_off, 0, array,
				  &fieldoff);
      HOST_WIDE_INT cstoff;
      if (!base_off.is_constant (&cstoff))
	return NULL_TREE;

      cstoff = (cstoff - fieldoff) / BITS_PER_UNIT;
      tree off = build_int_cst (sizetype, cstoff);
      if (varidx)
	offset = fold_build2 (PLUS_EXPR, TREE_TYPE (offset), offset, off);
      else
	offset = off;
    }

  if (!init)
    return NULL_TREE;

  *ptr_offset = offset;

  tree inittype = TREE_TYPE (init);

  if (TREE_CODE (init) == INTEGER_CST
      && (TREE_CODE (TREE_TYPE (array)) == INTEGER_TYPE
	  || TYPE_MAIN_VARIANT (inittype) == char_type_node))
    {
      /* For a reference to (address of) a single constant character,
	 store the native representation of the character in CHARBUF.
	 If the reference is to an element of an array or a member
	 of a struct, only consider narrow characters until ctors
	 for wide character arrays are transformed to STRING_CSTs
	 like those for narrow arrays.  */
      unsigned char charbuf[MAX_BITSIZE_MODE_ANY_MODE / BITS_PER_UNIT];
      int len = native_encode_expr (init, charbuf, sizeof charbuf, 0);
      if (len > 0)
	{
	  /* Construct a string literal with elements of INITTYPE and
	     the representation above.  Then strip
	     the ADDR_EXPR (ARRAY_REF (...)) around the STRING_CST.  */
	  init = build_string_literal (len, (char *)charbuf, inittype);
	  init = TREE_OPERAND (TREE_OPERAND (init, 0), 0);
	}
    }

  tree initsize = TYPE_SIZE_UNIT (inittype);

  if (TREE_CODE (init) == CONSTRUCTOR && initializer_zerop (init))
    {
      /* Fold an empty/zero constructor for an implicitly initialized
	 object or subobject into the empty string.  */

      /* Determine the character type from that of the original
	 expression.  */
      tree chartype = argtype;
      if (POINTER_TYPE_P (chartype))
	chartype = TREE_TYPE (chartype);
      while (TREE_CODE (chartype) == ARRAY_TYPE)
	chartype = TREE_TYPE (chartype);
      /* Convert a char array to an empty STRING_CST having an array
	 of the expected type.  */
      if (!initsize)
	  initsize = integer_zero_node;

      unsigned HOST_WIDE_INT size = tree_to_uhwi (initsize);
      init = build_string_literal (size ? 1 : 0, "", chartype, size);
      init = TREE_OPERAND (init, 0);
      init = TREE_OPERAND (init, 0);

      *ptr_offset = integer_zero_node;
    }

  if (decl)
    *decl = array;

  if (TREE_CODE (init) != STRING_CST)
    return NULL_TREE;

  *mem_size = initsize;

  gcc_checking_assert (tree_to_shwi (initsize) >= TREE_STRING_LENGTH (init));

  return init;
}

/* Compute the modular multiplicative inverse of A modulo M
   using extended Euclid's algorithm.  Assumes A and M are coprime.  */
static wide_int
mod_inv (const wide_int &a, const wide_int &b)
{
  /* Verify the assumption.  */
  gcc_checking_assert (wi::eq_p (wi::gcd (a, b), 1));

  unsigned int p = a.get_precision () + 1;
  gcc_checking_assert (b.get_precision () + 1 == p);
  wide_int c = wide_int::from (a, p, UNSIGNED);
  wide_int d = wide_int::from (b, p, UNSIGNED);
  wide_int x0 = wide_int::from (0, p, UNSIGNED);
  wide_int x1 = wide_int::from (1, p, UNSIGNED);

  if (wi::eq_p (b, 1))
    return wide_int::from (1, p, UNSIGNED);

  while (wi::gt_p (c, 1, UNSIGNED))
    {
      wide_int t = d;
      wide_int q = wi::divmod_trunc (c, d, UNSIGNED, &d);
      c = t;
      wide_int s = x0;
      x0 = wi::sub (x1, wi::mul (q, x0));
      x1 = s;
    }
  if (wi::lt_p (x1, 0, SIGNED))
    x1 += d;
  return x1;
}

/* Optimize x % C1 == C2 for signed modulo if C1 is a power of two and C2
   is non-zero and C3 ((1<<(prec-1)) | (C1 - 1)):
   for C2 > 0 to x & C3 == C2
   for C2 < 0 to x & C3 == (C2 & C3).  */
enum tree_code
maybe_optimize_pow2p_mod_cmp (enum tree_code code, tree *arg0, tree *arg1)
{
  gimple *stmt = get_def_for_expr (*arg0, TRUNC_MOD_EXPR);
  tree treeop0 = gimple_assign_rhs1 (stmt);
  tree treeop1 = gimple_assign_rhs2 (stmt);
  tree type = TREE_TYPE (*arg0);
  scalar_int_mode mode;
  if (!is_a <scalar_int_mode> (TYPE_MODE (type), &mode))
    return code;
  if (GET_MODE_BITSIZE (mode) != TYPE_PRECISION (type)
      || TYPE_PRECISION (type) <= 1
      || TYPE_UNSIGNED (type)
      /* Signed x % c == 0 should have been optimized into unsigned modulo
	 earlier.  */
      || integer_zerop (*arg1)
      /* If c is known to be non-negative, modulo will be expanded as unsigned
	 modulo.  */
      || get_range_pos_neg (treeop0) == 1)
    return code;

  /* x % c == d where d < 0 && d <= -c should be always false.  */
  if (tree_int_cst_sgn (*arg1) == -1
      && -wi::to_widest (treeop1) >= wi::to_widest (*arg1))
    return code;

  int prec = TYPE_PRECISION (type);
  wide_int w = wi::to_wide (treeop1) - 1;
  w |= wi::shifted_mask (0, prec - 1, true, prec);
  tree c3 = wide_int_to_tree (type, w);
  tree c4 = *arg1;
  if (tree_int_cst_sgn (*arg1) == -1)
    c4 = wide_int_to_tree (type, w & wi::to_wide (*arg1));

  rtx op0 = expand_normal (treeop0);
  treeop0 = make_tree (TREE_TYPE (treeop0), op0);

  bool speed_p = optimize_insn_for_speed_p ();

  do_pending_stack_adjust ();

  location_t loc = gimple_location (stmt);
  struct separate_ops ops;
  ops.code = TRUNC_MOD_EXPR;
  ops.location = loc;
  ops.type = TREE_TYPE (treeop0);
  ops.op0 = treeop0;
  ops.op1 = treeop1;
  ops.op2 = NULL_TREE;
  start_sequence ();
  rtx mor = expand_expr_real_2 (&ops, NULL_RTX, TYPE_MODE (ops.type),
				EXPAND_NORMAL);
  rtx_insn *moinsns = get_insns ();
  end_sequence ();

  unsigned mocost = seq_cost (moinsns, speed_p);
  mocost += rtx_cost (mor, mode, EQ, 0, speed_p);
  mocost += rtx_cost (expand_normal (*arg1), mode, EQ, 1, speed_p);

  ops.code = BIT_AND_EXPR;
  ops.location = loc;
  ops.type = TREE_TYPE (treeop0);
  ops.op0 = treeop0;
  ops.op1 = c3;
  ops.op2 = NULL_TREE;
  start_sequence ();
  rtx mur = expand_expr_real_2 (&ops, NULL_RTX, TYPE_MODE (ops.type),
				EXPAND_NORMAL);
  rtx_insn *muinsns = get_insns ();
  end_sequence ();

  unsigned mucost = seq_cost (muinsns, speed_p);
  mucost += rtx_cost (mur, mode, EQ, 0, speed_p);
  mucost += rtx_cost (expand_normal (c4), mode, EQ, 1, speed_p);

  if (mocost <= mucost)
    {
      emit_insn (moinsns);
      *arg0 = make_tree (TREE_TYPE (*arg0), mor);
      return code;
    }

  emit_insn (muinsns);
  *arg0 = make_tree (TREE_TYPE (*arg0), mur);
  *arg1 = c4;
  return code;
}

/* Attempt to optimize unsigned (X % C1) == C2 (or (X % C1) != C2).
   If C1 is odd to:
   (X - C2) * C3 <= C4 (or >), where
   C3 is modular multiplicative inverse of C1 and 1<<prec and
   C4 is ((1<<prec) - 1) / C1 or ((1<<prec) - 1) / C1 - 1 (the latter
   if C2 > ((1<<prec) - 1) % C1).
   If C1 is even, S = ctz (C1) and C2 is 0, use
   ((X * C3) r>> S) <= C4, where C3 is modular multiplicative
   inverse of C1>>S and 1<<prec and C4 is (((1<<prec) - 1) / (C1>>S)) >> S.

   For signed (X % C1) == 0 if C1 is odd to (all operations in it
   unsigned):
   (X * C3) + C4 <= 2 * C4, where
   C3 is modular multiplicative inverse of (unsigned) C1 and 1<<prec and
   C4 is ((1<<(prec - 1) - 1) / C1).
   If C1 is even, S = ctz(C1), use
   ((X * C3) + C4) r>> S <= (C4 >> (S - 1))
   where C3 is modular multiplicative inverse of (unsigned)(C1>>S) and 1<<prec
   and C4 is ((1<<(prec - 1) - 1) / (C1>>S)) & (-1<<S).

   See the Hacker's Delight book, section 10-17.  */
enum tree_code
maybe_optimize_mod_cmp (enum tree_code code, tree *arg0, tree *arg1)
{
  gcc_checking_assert (code == EQ_EXPR || code == NE_EXPR);
  gcc_checking_assert (TREE_CODE (*arg1) == INTEGER_CST);

  if (optimize < 2)
    return code;

  gimple *stmt = get_def_for_expr (*arg0, TRUNC_MOD_EXPR);
  if (stmt == NULL)
    return code;

  tree treeop0 = gimple_assign_rhs1 (stmt);
  tree treeop1 = gimple_assign_rhs2 (stmt);
  if (TREE_CODE (treeop0) != SSA_NAME
      || TREE_CODE (treeop1) != INTEGER_CST
      /* Don't optimize the undefined behavior case x % 0;
	 x % 1 should have been optimized into zero, punt if
	 it makes it here for whatever reason;
	 x % -c should have been optimized into x % c.  */
      || compare_tree_int (treeop1, 2) <= 0
      /* Likewise x % c == d where d >= c should be always false.  */
      || tree_int_cst_le (treeop1, *arg1))
    return code;

  /* Unsigned x % pow2 is handled right already, for signed
     modulo handle it in maybe_optimize_pow2p_mod_cmp.  */
  if (integer_pow2p (treeop1))
    return maybe_optimize_pow2p_mod_cmp (code, arg0, arg1);

  tree type = TREE_TYPE (*arg0);
  scalar_int_mode mode;
  if (!is_a <scalar_int_mode> (TYPE_MODE (type), &mode))
    return code;
  if (GET_MODE_BITSIZE (mode) != TYPE_PRECISION (type)
      || TYPE_PRECISION (type) <= 1)
    return code;

  signop sgn = UNSIGNED;
  /* If both operands are known to have the sign bit clear, handle
     even the signed modulo case as unsigned.  treeop1 is always
     positive >= 2, checked above.  */
  if (!TYPE_UNSIGNED (type) && get_range_pos_neg (treeop0) != 1)
    sgn = SIGNED;

  if (!TYPE_UNSIGNED (type))
    {
      if (tree_int_cst_sgn (*arg1) == -1)
	return code;
      type = unsigned_type_for (type);
      if (!type || TYPE_MODE (type) != TYPE_MODE (TREE_TYPE (*arg0)))
	return code;
    }

  int prec = TYPE_PRECISION (type);
  wide_int w = wi::to_wide (treeop1);
  int shift = wi::ctz (w);
  /* Unsigned (X % C1) == C2 is equivalent to (X - C2) % C1 == 0 if
     C2 <= -1U % C1, because for any Z >= 0U - C2 in that case (Z % C1) != 0.
     If C1 is odd, we can handle all cases by subtracting
     C4 below.  We could handle even the even C1 and C2 > -1U % C1 cases
     e.g. by testing for overflow on the subtraction, punt on that for now
     though.  */
  if ((sgn == SIGNED || shift) && !integer_zerop (*arg1))
    {
      if (sgn == SIGNED)
	return code;
      wide_int x = wi::umod_trunc (wi::mask (prec, false, prec), w);
      if (wi::gtu_p (wi::to_wide (*arg1), x))
	return code;
    }

  imm_use_iterator imm_iter;
  use_operand_p use_p;
  FOR_EACH_IMM_USE_FAST (use_p, imm_iter, treeop0)
    {
      gimple *use_stmt = USE_STMT (use_p);
      /* Punt if treeop0 is used in the same bb in a division
	 or another modulo with the same divisor.  We should expect
	 the division and modulo combined together.  */
      if (use_stmt == stmt
	  || gimple_bb (use_stmt) != gimple_bb (stmt))
	continue;
      if (!is_gimple_assign (use_stmt)
	  || (gimple_assign_rhs_code (use_stmt) != TRUNC_DIV_EXPR
	      && gimple_assign_rhs_code (use_stmt) != TRUNC_MOD_EXPR))
	continue;
      if (gimple_assign_rhs1 (use_stmt) != treeop0
	  || !operand_equal_p (gimple_assign_rhs2 (use_stmt), treeop1, 0))
	continue;
      return code;
    }

  w = wi::lrshift (w, shift);
  wide_int a = wide_int::from (w, prec + 1, UNSIGNED);
  wide_int b = wi::shifted_mask (prec, 1, false, prec + 1);
  wide_int m = wide_int::from (mod_inv (a, b), prec, UNSIGNED);
  tree c3 = wide_int_to_tree (type, m);
  tree c5 = NULL_TREE;
  wide_int d, e;
  if (sgn == UNSIGNED)
    {
      d = wi::divmod_trunc (wi::mask (prec, false, prec), w, UNSIGNED, &e);
      /* Use <= floor ((1<<prec) - 1) / C1 only if C2 <= ((1<<prec) - 1) % C1,
	 otherwise use < or subtract one from C4.  E.g. for
	 x % 3U == 0 we transform this into x * 0xaaaaaaab <= 0x55555555, but
	 x % 3U == 1 already needs to be
	 (x - 1) * 0xaaaaaaabU <= 0x55555554.  */
      if (!shift && wi::gtu_p (wi::to_wide (*arg1), e))
	d -= 1;
      if (shift)
	d = wi::lrshift (d, shift);
    }
  else
    {
      e = wi::udiv_trunc (wi::mask (prec - 1, false, prec), w);
      if (!shift)
	d = wi::lshift (e, 1);
      else
	{
	  e = wi::bit_and (e, wi::mask (shift, true, prec));
	  d = wi::lrshift (e, shift - 1);
	}
      c5 = wide_int_to_tree (type, e);
    }
  tree c4 = wide_int_to_tree (type, d);

  rtx op0 = expand_normal (treeop0);
  treeop0 = make_tree (TREE_TYPE (treeop0), op0);

  bool speed_p = optimize_insn_for_speed_p ();

  do_pending_stack_adjust ();

  location_t loc = gimple_location (stmt);
  struct separate_ops ops;
  ops.code = TRUNC_MOD_EXPR;
  ops.location = loc;
  ops.type = TREE_TYPE (treeop0);
  ops.op0 = treeop0;
  ops.op1 = treeop1;
  ops.op2 = NULL_TREE;
  start_sequence ();
  rtx mor = expand_expr_real_2 (&ops, NULL_RTX, TYPE_MODE (ops.type),
				EXPAND_NORMAL);
  rtx_insn *moinsns = get_insns ();
  end_sequence ();

  unsigned mocost = seq_cost (moinsns, speed_p);
  mocost += rtx_cost (mor, mode, EQ, 0, speed_p);
  mocost += rtx_cost (expand_normal (*arg1), mode, EQ, 1, speed_p);

  tree t = fold_convert_loc (loc, type, treeop0);
  if (!integer_zerop (*arg1))
    t = fold_build2_loc (loc, MINUS_EXPR, type, t, fold_convert (type, *arg1));
  t = fold_build2_loc (loc, MULT_EXPR, type, t, c3);
  if (sgn == SIGNED)
    t = fold_build2_loc (loc, PLUS_EXPR, type, t, c5);
  if (shift)
    {
      tree s = build_int_cst (NULL_TREE, shift);
      t = fold_build2_loc (loc, RROTATE_EXPR, type, t, s);
    }

  start_sequence ();
  rtx mur = expand_normal (t);
  rtx_insn *muinsns = get_insns ();
  end_sequence ();

  unsigned mucost = seq_cost (muinsns, speed_p);
  mucost += rtx_cost (mur, mode, LE, 0, speed_p);
  mucost += rtx_cost (expand_normal (c4), mode, LE, 1, speed_p);

  if (mocost <= mucost)
    {
      emit_insn (moinsns);
      *arg0 = make_tree (TREE_TYPE (*arg0), mor);
      return code;
    }

  emit_insn (muinsns);
  *arg0 = make_tree (type, mur);
  *arg1 = c4;
  return code == EQ_EXPR ? LE_EXPR : GT_EXPR;
}

/* Generate code to calculate OPS, and exploded expression
   using a store-flag instruction and return an rtx for the result.
   OPS reflects a comparison.

   If TARGET is nonzero, store the result there if convenient.

   Return zero if there is no suitable set-flag instruction
   available on this machine.

   Once expand_expr has been called on the arguments of the comparison,
   we are committed to doing the store flag, since it is not safe to
   re-evaluate the expression.  We emit the store-flag insn by calling
   emit_store_flag, but only expand the arguments if we have a reason
   to believe that emit_store_flag will be successful.  If we think that
   it will, but it isn't, we have to simulate the store-flag with a
   set/jump/set sequence.  */

static rtx
do_store_flag (sepops ops, rtx target, machine_mode mode)
{
  enum rtx_code code;
  tree arg0, arg1, type;
  machine_mode operand_mode;
  int unsignedp;
  rtx op0, op1;
  rtx subtarget = target;
  location_t loc = ops->location;

  arg0 = ops->op0;
  arg1 = ops->op1;

  /* Don't crash if the comparison was erroneous.  */
  if (arg0 == error_mark_node || arg1 == error_mark_node)
    return const0_rtx;

  type = TREE_TYPE (arg0);
  operand_mode = TYPE_MODE (type);
  unsignedp = TYPE_UNSIGNED (type);

  /* We won't bother with BLKmode store-flag operations because it would mean
     passing a lot of information to emit_store_flag.  */
  if (operand_mode == BLKmode)
    return 0;

  /* We won't bother with store-flag operations involving function pointers
     when function pointers must be canonicalized before comparisons.  */
  if (targetm.have_canonicalize_funcptr_for_compare ()
      && ((POINTER_TYPE_P (TREE_TYPE (arg0))
	   && FUNC_OR_METHOD_TYPE_P (TREE_TYPE (TREE_TYPE (arg0))))
	  || (POINTER_TYPE_P (TREE_TYPE (arg1))
	      && FUNC_OR_METHOD_TYPE_P (TREE_TYPE (TREE_TYPE (arg1))))))
    return 0;

  STRIP_NOPS (arg0);
  STRIP_NOPS (arg1);

  /* For vector typed comparisons emit code to generate the desired
     all-ones or all-zeros mask.  Conveniently use the VEC_COND_EXPR
     expander for this.  */
  if (TREE_CODE (ops->type) == VECTOR_TYPE)
    {
      tree ifexp = build2 (ops->code, ops->type, arg0, arg1);
      if (VECTOR_BOOLEAN_TYPE_P (ops->type)
	  && expand_vec_cmp_expr_p (TREE_TYPE (arg0), ops->type, ops->code))
	return expand_vec_cmp_expr (ops->type, ifexp, target);
      else
	{
	  tree if_true = constant_boolean_node (true, ops->type);
	  tree if_false = constant_boolean_node (false, ops->type);
	  return expand_vec_cond_expr (ops->type, ifexp, if_true,
				       if_false, target);
	}
    }

  /* Optimize (x % C1) == C2 or (x % C1) != C2 if it is beneficial
     into (x - C2) * C3 < C4.  */
  if ((ops->code == EQ_EXPR || ops->code == NE_EXPR)
      && TREE_CODE (arg0) == SSA_NAME
      && TREE_CODE (arg1) == INTEGER_CST)
    {
      enum tree_code new_code = maybe_optimize_mod_cmp (ops->code,
							&arg0, &arg1);
      if (new_code != ops->code)
	{
	  struct separate_ops nops = *ops;
	  nops.code = ops->code = new_code;
	  nops.op0 = arg0;
	  nops.op1 = arg1;
	  nops.type = TREE_TYPE (arg0);
	  return do_store_flag (&nops, target, mode);
	}
    }

  /* Get the rtx comparison code to use.  We know that EXP is a comparison
     operation of some type.  Some comparisons against 1 and -1 can be
     converted to comparisons with zero.  Do so here so that the tests
     below will be aware that we have a comparison with zero.   These
     tests will not catch constants in the first operand, but constants
     are rarely passed as the first operand.  */

  switch (ops->code)
    {
    case EQ_EXPR:
      code = EQ;
      break;
    case NE_EXPR:
      code = NE;
      break;
    case LT_EXPR:
      if (integer_onep (arg1))
	arg1 = integer_zero_node, code = unsignedp ? LEU : LE;
      else
	code = unsignedp ? LTU : LT;
      break;
    case LE_EXPR:
      if (! unsignedp && integer_all_onesp (arg1))
	arg1 = integer_zero_node, code = LT;
      else
	code = unsignedp ? LEU : LE;
      break;
    case GT_EXPR:
      if (! unsignedp && integer_all_onesp (arg1))
	arg1 = integer_zero_node, code = GE;
      else
	code = unsignedp ? GTU : GT;
      break;
    case GE_EXPR:
      if (integer_onep (arg1))
	arg1 = integer_zero_node, code = unsignedp ? GTU : GT;
      else
	code = unsignedp ? GEU : GE;
      break;

    case UNORDERED_EXPR:
      code = UNORDERED;
      break;
    case ORDERED_EXPR:
      code = ORDERED;
      break;
    case UNLT_EXPR:
      code = UNLT;
      break;
    case UNLE_EXPR:
      code = UNLE;
      break;
    case UNGT_EXPR:
      code = UNGT;
      break;
    case UNGE_EXPR:
      code = UNGE;
      break;
    case UNEQ_EXPR:
      code = UNEQ;
      break;
    case LTGT_EXPR:
      code = LTGT;
      break;

    default:
      gcc_unreachable ();
    }

  /* Put a constant second.  */
  if (TREE_CODE (arg0) == REAL_CST || TREE_CODE (arg0) == INTEGER_CST
      || TREE_CODE (arg0) == FIXED_CST)
    {
      std::swap (arg0, arg1);
      code = swap_condition (code);
    }

  /* If this is an equality or inequality test of a single bit, we can
     do this by shifting the bit being tested to the low-order bit and
     masking the result with the constant 1.  If the condition was EQ,
     we xor it with 1.  This does not require an scc insn and is faster
     than an scc insn even if we have it.

     The code to make this transformation was moved into fold_single_bit_test,
     so we just call into the folder and expand its result.  */

  if ((code == NE || code == EQ)
      && integer_zerop (arg1)
      && (TYPE_PRECISION (ops->type) != 1 || TYPE_UNSIGNED (ops->type)))
    {
      gimple *srcstmt = get_def_for_expr (arg0, BIT_AND_EXPR);
      if (srcstmt
	  && integer_pow2p (gimple_assign_rhs2 (srcstmt)))
	{
	  enum tree_code tcode = code == NE ? NE_EXPR : EQ_EXPR;
	  type = lang_hooks.types.type_for_mode (mode, unsignedp);
	  tree temp = fold_build2_loc (loc, BIT_AND_EXPR, TREE_TYPE (arg1),
				       gimple_assign_rhs1 (srcstmt),
				       gimple_assign_rhs2 (srcstmt));
	  temp = fold_single_bit_test (loc, tcode, temp, arg1, type);
	  if (temp)
	    return expand_expr (temp, target, VOIDmode, EXPAND_NORMAL);
	}
    }

  if (! get_subtarget (target)
      || GET_MODE (subtarget) != operand_mode)
    subtarget = 0;

  expand_operands (arg0, arg1, subtarget, &op0, &op1, EXPAND_NORMAL);

  if (target == 0)
    target = gen_reg_rtx (mode);

  /* Try a cstore if possible.  */
  return emit_store_flag_force (target, code, op0, op1,
				operand_mode, unsignedp,
				(TYPE_PRECISION (ops->type) == 1
				 && !TYPE_UNSIGNED (ops->type)) ? -1 : 1);
}

/* Attempt to generate a casesi instruction.  Returns 1 if successful,
   0 otherwise (i.e. if there is no casesi instruction).

   DEFAULT_PROBABILITY is the probability of jumping to the default
   label.  */
int
try_casesi (tree index_type, tree index_expr, tree minval, tree range,
	    rtx table_label, rtx default_label, rtx fallback_label,
            profile_probability default_probability)
{
  class expand_operand ops[5];
  scalar_int_mode index_mode = SImode;
  rtx op1, op2, index;

  if (! targetm.have_casesi ())
    return 0;

  /* The index must be some form of integer.  Convert it to SImode.  */
  scalar_int_mode omode = SCALAR_INT_TYPE_MODE (index_type);
  if (GET_MODE_BITSIZE (omode) > GET_MODE_BITSIZE (index_mode))
    {
      rtx rangertx = expand_normal (range);

      /* We must handle the endpoints in the original mode.  */
      index_expr = build2 (MINUS_EXPR, index_type,
			   index_expr, minval);
      minval = integer_zero_node;
      index = expand_normal (index_expr);
      if (default_label)
        emit_cmp_and_jump_insns (rangertx, index, LTU, NULL_RTX,
				 omode, 1, default_label,
                                 default_probability);
      /* Now we can safely truncate.  */
      index = convert_to_mode (index_mode, index, 0);
    }
  else
    {
      if (omode != index_mode)
	{
	  index_type = lang_hooks.types.type_for_mode (index_mode, 0);
	  index_expr = fold_convert (index_type, index_expr);
	}

      index = expand_normal (index_expr);
    }

  do_pending_stack_adjust ();

  op1 = expand_normal (minval);
  op2 = expand_normal (range);

  create_input_operand (&ops[0], index, index_mode);
  create_convert_operand_from_type (&ops[1], op1, TREE_TYPE (minval));
  create_convert_operand_from_type (&ops[2], op2, TREE_TYPE (range));
  create_fixed_operand (&ops[3], table_label);
  create_fixed_operand (&ops[4], (default_label
				  ? default_label
				  : fallback_label));
  expand_jump_insn (targetm.code_for_casesi, 5, ops);
  return 1;
}

/* Attempt to generate a tablejump instruction; same concept.  */
/* Subroutine of the next function.

   INDEX is the value being switched on, with the lowest value
   in the table already subtracted.
   MODE is its expected mode (needed if INDEX is constant).
   RANGE is the length of the jump table.
   TABLE_LABEL is a CODE_LABEL rtx for the table itself.

   DEFAULT_LABEL is a CODE_LABEL rtx to jump to if the
   index value is out of range.
   DEFAULT_PROBABILITY is the probability of jumping to
   the default label.  */

static void
do_tablejump (rtx index, machine_mode mode, rtx range, rtx table_label,
	      rtx default_label, profile_probability default_probability)
{
  rtx temp, vector;

  if (INTVAL (range) > cfun->cfg->max_jumptable_ents)
    cfun->cfg->max_jumptable_ents = INTVAL (range);

  /* Do an unsigned comparison (in the proper mode) between the index
     expression and the value which represents the length of the range.
     Since we just finished subtracting the lower bound of the range
     from the index expression, this comparison allows us to simultaneously
     check that the original index expression value is both greater than
     or equal to the minimum value of the range and less than or equal to
     the maximum value of the range.  */

  if (default_label)
    emit_cmp_and_jump_insns (index, range, GTU, NULL_RTX, mode, 1,
			     default_label, default_probability);

  /* If index is in range, it must fit in Pmode.
     Convert to Pmode so we can index with it.  */
  if (mode != Pmode)
    {
      unsigned int width;

      /* We know the value of INDEX is between 0 and RANGE.  If we have a
	 sign-extended subreg, and RANGE does not have the sign bit set, then
	 we have a value that is valid for both sign and zero extension.  In
	 this case, we get better code if we sign extend.  */
      if (GET_CODE (index) == SUBREG
	  && SUBREG_PROMOTED_VAR_P (index)
	  && SUBREG_PROMOTED_SIGNED_P (index)
	  && ((width = GET_MODE_PRECISION (as_a <scalar_int_mode> (mode)))
	      <= HOST_BITS_PER_WIDE_INT)
	  && ! (UINTVAL (range) & (HOST_WIDE_INT_1U << (width - 1))))
	index = convert_to_mode (Pmode, index, 0);
      else
	index = convert_to_mode (Pmode, index, 1);
    }

  /* Don't let a MEM slip through, because then INDEX that comes
     out of PIC_CASE_VECTOR_ADDRESS won't be a valid address,
     and break_out_memory_refs will go to work on it and mess it up.  */
#ifdef PIC_CASE_VECTOR_ADDRESS
  if (flag_pic && !REG_P (index))
    index = copy_to_mode_reg (Pmode, index);
#endif

  /* ??? The only correct use of CASE_VECTOR_MODE is the one inside the
     GET_MODE_SIZE, because this indicates how large insns are.  The other
     uses should all be Pmode, because they are addresses.  This code
     could fail if addresses and insns are not the same size.  */
  index = simplify_gen_binary (MULT, Pmode, index,
			       gen_int_mode (GET_MODE_SIZE (CASE_VECTOR_MODE),
					     Pmode));
  index = simplify_gen_binary (PLUS, Pmode, index,
			       gen_rtx_LABEL_REF (Pmode, table_label));

#ifdef PIC_CASE_VECTOR_ADDRESS
  if (flag_pic)
    index = PIC_CASE_VECTOR_ADDRESS (index);
  else
#endif
    index = memory_address (CASE_VECTOR_MODE, index);
  temp = gen_reg_rtx (CASE_VECTOR_MODE);
  vector = gen_const_mem (CASE_VECTOR_MODE, index);
  convert_move (temp, vector, 0);

  emit_jump_insn (targetm.gen_tablejump (temp, table_label));

  /* If we are generating PIC code or if the table is PC-relative, the
     table and JUMP_INSN must be adjacent, so don't output a BARRIER.  */
  if (! CASE_VECTOR_PC_RELATIVE && ! flag_pic)
    emit_barrier ();
}

int
try_tablejump (tree index_type, tree index_expr, tree minval, tree range,
	       rtx table_label, rtx default_label,
	       profile_probability default_probability)
{
  rtx index;

  if (! targetm.have_tablejump ())
    return 0;

  index_expr = fold_build2 (MINUS_EXPR, index_type,
			    fold_convert (index_type, index_expr),
			    fold_convert (index_type, minval));
  index = expand_normal (index_expr);
  do_pending_stack_adjust ();

  do_tablejump (index, TYPE_MODE (index_type),
		convert_modes (TYPE_MODE (index_type),
			       TYPE_MODE (TREE_TYPE (range)),
			       expand_normal (range),
			       TYPE_UNSIGNED (TREE_TYPE (range))),
		table_label, default_label, default_probability);
  return 1;
}

/* Return a CONST_VECTOR rtx representing vector mask for
   a VECTOR_CST of booleans.  */
static rtx
const_vector_mask_from_tree (tree exp)
{
  machine_mode mode = TYPE_MODE (TREE_TYPE (exp));
  machine_mode inner = GET_MODE_INNER (mode);

  rtx_vector_builder builder (mode, VECTOR_CST_NPATTERNS (exp),
			      VECTOR_CST_NELTS_PER_PATTERN (exp));
  unsigned int count = builder.encoded_nelts ();
  for (unsigned int i = 0; i < count; ++i)
    {
      tree elt = VECTOR_CST_ELT (exp, i);
      gcc_assert (TREE_CODE (elt) == INTEGER_CST);
      if (integer_zerop (elt))
	builder.quick_push (CONST0_RTX (inner));
      else if (integer_onep (elt)
	       || integer_minus_onep (elt))
	builder.quick_push (CONSTM1_RTX (inner));
      else
	gcc_unreachable ();
    }
  return builder.build ();
}

/* EXP is a VECTOR_CST in which each element is either all-zeros or all-ones.
   Return a constant scalar rtx of mode MODE in which bit X is set if element
   X of EXP is nonzero.  */
static rtx
const_scalar_mask_from_tree (scalar_int_mode mode, tree exp)
{
  wide_int res = wi::zero (GET_MODE_PRECISION (mode));
  tree elt;

  /* The result has a fixed number of bits so the input must too.  */
  unsigned int nunits = VECTOR_CST_NELTS (exp).to_constant ();
  for (unsigned int i = 0; i < nunits; ++i)
    {
      elt = VECTOR_CST_ELT (exp, i);
      gcc_assert (TREE_CODE (elt) == INTEGER_CST);
      if (integer_all_onesp (elt))
	res = wi::set_bit (res, i);
      else
	gcc_assert (integer_zerop (elt));
    }

  return immed_wide_int_const (res, mode);
}

/* Return a CONST_VECTOR rtx for a VECTOR_CST tree.  */
static rtx
const_vector_from_tree (tree exp)
{
  machine_mode mode = TYPE_MODE (TREE_TYPE (exp));

  if (initializer_zerop (exp))
    return CONST0_RTX (mode);

  if (VECTOR_BOOLEAN_TYPE_P (TREE_TYPE (exp)))
    return const_vector_mask_from_tree (exp);

  machine_mode inner = GET_MODE_INNER (mode);

  rtx_vector_builder builder (mode, VECTOR_CST_NPATTERNS (exp),
			      VECTOR_CST_NELTS_PER_PATTERN (exp));
  unsigned int count = builder.encoded_nelts ();
  for (unsigned int i = 0; i < count; ++i)
    {
      tree elt = VECTOR_CST_ELT (exp, i);
      if (TREE_CODE (elt) == REAL_CST)
	builder.quick_push (const_double_from_real_value (TREE_REAL_CST (elt),
							  inner));
      else if (TREE_CODE (elt) == FIXED_CST)
	builder.quick_push (CONST_FIXED_FROM_FIXED_VALUE (TREE_FIXED_CST (elt),
							  inner));
      else
	builder.quick_push (immed_wide_int_const (wi::to_poly_wide (elt),
						  inner));
    }
  return builder.build ();
}

/* Build a decl for a personality function given a language prefix.  */

tree
build_personality_function (const char *lang)
{
  const char *unwind_and_version;
  tree decl, type;
  char *name;

  switch (targetm_common.except_unwind_info (&global_options))
    {
    case UI_NONE:
      return NULL;
    case UI_SJLJ:
      unwind_and_version = "_sj0";
      break;
    case UI_DWARF2:
    case UI_TARGET:
      unwind_and_version = "_v0";
      break;
    case UI_SEH:
      unwind_and_version = "_seh0";
      break;
    default:
      gcc_unreachable ();
    }

  name = ACONCAT (("__", lang, "_personality", unwind_and_version, NULL));

  type = build_function_type_list (unsigned_type_node,
				   integer_type_node, integer_type_node,
				   long_long_unsigned_type_node,
				   ptr_type_node, ptr_type_node, NULL_TREE);
  decl = build_decl (UNKNOWN_LOCATION, FUNCTION_DECL,
		     get_identifier (name), type);
  DECL_ARTIFICIAL (decl) = 1;
  DECL_EXTERNAL (decl) = 1;
  TREE_PUBLIC (decl) = 1;

  /* Zap the nonsensical SYMBOL_REF_DECL for this.  What we're left with
     are the flags assigned by targetm.encode_section_info.  */
  SET_SYMBOL_REF_DECL (XEXP (DECL_RTL (decl), 0), NULL);

  return decl;
}

/* Extracts the personality function of DECL and returns the corresponding
   libfunc.  */

rtx
get_personality_function (tree decl)
{
  tree personality = DECL_FUNCTION_PERSONALITY (decl);
  enum eh_personality_kind pk;

  pk = function_needs_eh_personality (DECL_STRUCT_FUNCTION (decl));
  if (pk == eh_personality_none)
    return NULL;

  if (!personality
      && pk == eh_personality_any)
    personality = lang_hooks.eh_personality ();

  if (pk == eh_personality_lang)
    gcc_assert (personality != NULL_TREE);

  return XEXP (DECL_RTL (personality), 0);
}

/* Returns a tree for the size of EXP in bytes.  */

static tree
tree_expr_size (const_tree exp)
{
  if (DECL_P (exp)
      && DECL_SIZE_UNIT (exp) != 0)
    return DECL_SIZE_UNIT (exp);
  else
    return size_in_bytes (TREE_TYPE (exp));
}

/* Return an rtx for the size in bytes of the value of EXP.  */

rtx
expr_size (tree exp)
{
  tree size;

  if (TREE_CODE (exp) == WITH_SIZE_EXPR)
    size = TREE_OPERAND (exp, 1);
  else
    {
      size = tree_expr_size (exp);
      gcc_assert (size);
      gcc_assert (size == SUBSTITUTE_PLACEHOLDER_IN_EXPR (size, exp));
    }

  return expand_expr (size, NULL_RTX, TYPE_MODE (sizetype), EXPAND_NORMAL);
}

/* Return a wide integer for the size in bytes of the value of EXP, or -1
   if the size can vary or is larger than an integer.  */

static HOST_WIDE_INT
int_expr_size (tree exp)
{
  tree size;

  if (TREE_CODE (exp) == WITH_SIZE_EXPR)
    size = TREE_OPERAND (exp, 1);
  else
    {
      size = tree_expr_size (exp);
      gcc_assert (size);
    }

  if (size == 0 || !tree_fits_shwi_p (size))
    return -1;

  return tree_to_shwi (size);
}