backend_marked was a buggy/wrong concept, removed it
[libfirm] / ir / be / ia32 / ia32_emitter.c
index e0a2f98..4716f55 100644 (file)
@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 1995-2007 University of Karlsruhe.  All right reserved.
+ * Copyright (C) 1995-2008 University of Karlsruhe.  All right reserved.
  *
  * This file is part of libFirm.
  *
@@ -23,9 +23,7 @@
  * @author      Christian Wuerdig, Matthias Braun
  * @version     $Id$
  */
-#ifdef HAVE_CONFIG_H
 #include "config.h"
-#endif
 
 #include <limits.h>
 
 #include "irargs_t.h"
 #include "irprog_t.h"
 #include "iredges_t.h"
+#include "irtools.h"
 #include "execfreq.h"
 #include "error.h"
 #include "raw_bitset.h"
+#include "dbginfo.h"
+#include "lc_opts.h"
 
-#include "../besched_t.h"
-#include "../benode_t.h"
+#include "../besched.h"
+#include "../benode.h"
 #include "../beabi.h"
 #include "../be_dbgout.h"
 #include "../beemitter.h"
 #include "../begnuas.h"
-#include "../beirg_t.h"
+#include "../beirg.h"
+#include "../be_dbgout.h"
 
 #include "ia32_emitter.h"
 #include "gen_ia32_emitter.h"
@@ -57,6 +59,7 @@
 #include "ia32_nodes_attr.h"
 #include "ia32_new_nodes.h"
 #include "ia32_map_regs.h"
+#include "ia32_architecture.h"
 #include "bearch_ia32_t.h"
 
 DEBUG_ONLY(static firm_dbg_module_t *dbg = NULL;)
@@ -65,9 +68,59 @@ DEBUG_ONLY(static firm_dbg_module_t *dbg = NULL;)
 
 #define SNPRINTF_BUF_LEN 128
 
-static const arch_env_t *arch_env;
 static const ia32_isa_t *isa;
 static ia32_code_gen_t  *cg;
+static char              pic_base_label[128];
+static ir_label_t        exc_label_id;
+static int               mark_spill_reload = 0;
+static int               do_pic;
+
+/** Return the next block in Block schedule */
+static ir_node *get_prev_block_sched(const ir_node *block)
+{
+       return get_irn_link(block);
+}
+
+/** Checks if the current block is a fall-through target. */
+static int is_fallthrough(const ir_node *cfgpred)
+{
+       ir_node *pred;
+
+       if (!is_Proj(cfgpred))
+               return 1;
+       pred = get_Proj_pred(cfgpred);
+       if (is_ia32_SwitchJmp(pred))
+               return 0;
+
+       return 1;
+}
+
+/**
+ * returns non-zero if the given block needs a label
+ * because of being a jump-target (and not a fall-through)
+ */
+static int block_needs_label(const ir_node *block)
+{
+       int need_label = 1;
+       int  n_cfgpreds = get_Block_n_cfgpreds(block);
+
+       if (has_Block_entity(block))
+               return 1;
+
+       if (n_cfgpreds == 0) {
+               need_label = 0;
+       } else if (n_cfgpreds == 1) {
+               ir_node *cfgpred       = get_Block_cfgpred(block, 0);
+               ir_node *cfgpred_block = get_nodes_block(cfgpred);
+
+               if (get_prev_block_sched(block) == cfgpred_block
+                               && is_fallthrough(cfgpred)) {
+                       need_label = 0;
+               }
+       }
+
+       return need_label;
+}
 
 /**
  * Returns the register at in position pos.
@@ -83,19 +136,16 @@ static const arch_register_t *get_in_reg(const ir_node *irn, int pos)
           in register we need. */
        op = get_irn_n(irn, pos);
 
-       reg = arch_get_irn_register(arch_env, op);
+       reg = arch_get_irn_register(op);
 
        assert(reg && "no in register found");
 
-       if(reg == &ia32_gp_regs[REG_GP_NOREG])
+       if (reg == &ia32_gp_regs[REG_GP_NOREG])
                panic("trying to emit noreg for %+F input %d", irn, pos);
 
        /* in case of unknown register: just return a valid register */
        if (reg == &ia32_gp_regs[REG_GP_UKNWN]) {
-               const arch_register_req_t *req;
-
-               /* ask for the requirements */
-               req = arch_get_register_req(arch_env, irn, pos);
+               const arch_register_req_t *req = arch_get_register_req(irn, pos);
 
                if (arch_register_req_is(req, limited)) {
                        /* in case of limited requirements: get the first allowed register */
@@ -125,9 +175,9 @@ static const arch_register_t *get_out_reg(const ir_node *irn, int pos)
 
        if (get_irn_mode(irn) != mode_T) {
                assert(pos == 0);
-               reg = arch_get_irn_register(arch_env, irn);
+               reg = arch_get_irn_register(irn);
        } else if (is_ia32_irn(irn)) {
-               reg = get_ia32_out_reg(irn, pos);
+               reg = arch_irn_get_register(irn, pos);
        } else {
                const ir_edge_t *edge;
 
@@ -135,7 +185,7 @@ static const arch_register_t *get_out_reg(const ir_node *irn, int pos)
                        proj = get_edge_src_irn(edge);
                        assert(is_Proj(proj) && "non-Proj from mode_T node");
                        if (get_Proj_proj(proj) == pos) {
-                               reg = arch_get_irn_register(arch_env, proj);
+                               reg = arch_get_irn_register(proj);
                                break;
                        }
                }
@@ -166,6 +216,9 @@ static char *get_unique_label(char *buf, size_t buflen, const char *prefix)
  * |_|                                       |_|
  *************************************************************/
 
+/**
+ * Emit the name of the 8bit low register
+ */
 static void emit_8bit_register(const arch_register_t *reg)
 {
        const char *reg_name = arch_register_get_name(reg);
@@ -175,6 +228,18 @@ static void emit_8bit_register(const arch_register_t *reg)
        be_emit_char('l');
 }
 
+/**
+ * Emit the name of the 8bit high register
+ */
+static void emit_8bit_register_high(const arch_register_t *reg)
+{
+       const char *reg_name = arch_register_get_name(reg);
+
+       be_emit_char('%');
+       be_emit_char(reg_name[1]);
+       be_emit_char('h');
+}
+
 static void emit_16bit_register(const arch_register_t *reg)
 {
        const char *reg_name = ia32_get_mapped_reg_name(isa->regs_16bit, reg);
@@ -183,21 +248,23 @@ static void emit_16bit_register(const arch_register_t *reg)
        be_emit_string(reg_name);
 }
 
+/**
+ * emit a register, possible shortened by a mode
+ *
+ * @param reg   the register
+ * @param mode  the mode of the register or NULL for full register
+ */
 static void emit_register(const arch_register_t *reg, const ir_mode *mode)
 {
        const char *reg_name;
 
-       if(mode != NULL) {
+       if (mode != NULL) {
                int size = get_mode_size_bits(mode);
-               if(size == 8) {
-                       emit_8bit_register(reg);
-                       return;
-               } else if(size == 16) {
-                       emit_16bit_register(reg);
-                       return;
-               } else {
-                       assert(size == 32);
+               switch (size) {
+                       case  8: emit_8bit_register(reg);  return;
+                       case 16: emit_16bit_register(reg); return;
                }
+               assert(mode_is_float(mode) || size == 32);
        }
 
        reg_name = arch_register_get_name(reg);
@@ -208,18 +275,58 @@ static void emit_register(const arch_register_t *reg, const ir_mode *mode)
 
 void ia32_emit_source_register(const ir_node *node, int pos)
 {
-       const arch_register_t *reg      = get_in_reg(node, pos);
+       const arch_register_t *reg = get_in_reg(node, pos);
 
        emit_register(reg, NULL);
 }
 
-static void emit_ia32_Immediate(const ir_node *node);
+static void ia32_emit_entity(ir_entity *entity, int no_pic_adjust)
+{
+       be_gas_emit_entity(entity);
+
+       if (get_entity_owner(entity) == get_tls_type()) {
+               if (get_entity_linkage(entity) & IR_LINKAGE_EXTERN) {
+                       be_emit_cstring("@INDNTPOFF");
+               } else {
+                       be_emit_cstring("@NTPOFF");
+               }
+       }
+
+       if (do_pic && !no_pic_adjust) {
+               be_emit_char('-');
+               be_emit_string(pic_base_label);
+       }
+}
+
+static void emit_ia32_Immediate_no_prefix(const ir_node *node)
+{
+       const ia32_immediate_attr_t *attr = get_ia32_immediate_attr_const(node);
+
+       if (attr->symconst != NULL) {
+               if (attr->sc_sign)
+                       be_emit_char('-');
+               ia32_emit_entity(attr->symconst, attr->no_pic_adjust);
+       }
+       if (attr->symconst == NULL || attr->offset != 0) {
+               if (attr->symconst != NULL) {
+                       be_emit_irprintf("%+d", attr->offset);
+               } else {
+                       be_emit_irprintf("0x%X", attr->offset);
+               }
+       }
+}
+
+static void emit_ia32_Immediate(const ir_node *node)
+{
+       be_emit_char('$');
+       emit_ia32_Immediate_no_prefix(node);
+}
 
 void ia32_emit_8bit_source_register_or_immediate(const ir_node *node, int pos)
 {
        const arch_register_t *reg;
-       ir_node               *in = get_irn_n(node, pos);
-       if(is_ia32_Immediate(in)) {
+       const ir_node         *in = get_irn_n(node, pos);
+       if (is_ia32_Immediate(in)) {
                emit_ia32_Immediate(in);
                return;
        }
@@ -228,6 +335,25 @@ void ia32_emit_8bit_source_register_or_immediate(const ir_node *node, int pos)
        emit_8bit_register(reg);
 }
 
+void ia32_emit_8bit_high_source_register(const ir_node *node, int pos)
+{
+       const arch_register_t *reg = get_in_reg(node, pos);
+       emit_8bit_register_high(reg);
+}
+
+void ia32_emit_16bit_source_register_or_immediate(const ir_node *node, int pos)
+{
+       const arch_register_t *reg;
+       const ir_node         *in = get_irn_n(node, pos);
+       if (is_ia32_Immediate(in)) {
+               emit_ia32_Immediate(in);
+               return;
+       }
+
+       reg = get_in_reg(node, pos);
+       emit_16bit_register(reg);
+}
+
 void ia32_emit_dest_register(const ir_node *node, int pos)
 {
        const arch_register_t *reg  = get_out_reg(node, pos);
@@ -235,6 +361,13 @@ void ia32_emit_dest_register(const ir_node *node, int pos)
        emit_register(reg, NULL);
 }
 
+void ia32_emit_dest_register_size(const ir_node *node, int pos)
+{
+       const arch_register_t *reg  = get_out_reg(node, pos);
+
+       emit_register(reg, get_ia32_ls_mode(node));
+}
+
 void ia32_emit_8bit_dest_register(const ir_node *node, int pos)
 {
        const arch_register_t *reg  = get_out_reg(node, pos);
@@ -253,30 +386,22 @@ void ia32_emit_x87_register(const ir_node *node, int pos)
 
 static void ia32_emit_mode_suffix_mode(const ir_mode *mode)
 {
-       if(mode_is_float(mode)) {
-               switch(get_mode_size_bits(mode)) {
-               case 32: be_emit_char('s'); return;
-               case 64: be_emit_char('l'); return;
-               case 80: be_emit_char('t'); return;
-               }
-       } else {
-               assert(mode_is_int(mode) || mode_is_reference(mode));
-               switch(get_mode_size_bits(mode)) {
+       assert(mode_is_int(mode) || mode_is_reference(mode));
+       switch (get_mode_size_bits(mode)) {
+               case 8:  be_emit_char('b');     return;
+               case 16: be_emit_char('w');     return;
+               case 32: be_emit_char('l');     return;
+               /* gas docu says q is the suffix but gcc, objdump and icc use ll
+                * apparently */
                case 64: be_emit_cstring("ll"); return;
-                                /* gas docu says q is the suffix but gcc, objdump and icc use
-                                       ll apparently */
-               case 32: be_emit_char('l'); return;
-               case 16: be_emit_char('w'); return;
-               case 8:  be_emit_char('b'); return;
-               }
        }
-       panic("Can't output mode_suffix for %+F\n", mode);
+       panic("Can't output mode_suffix for %+F", mode);
 }
 
 void ia32_emit_mode_suffix(const ir_node *node)
 {
        ir_mode *mode = get_ia32_ls_mode(node);
-       if(mode == NULL)
+       if (mode == NULL)
                mode = mode_Iu;
 
        ia32_emit_mode_suffix_mode(mode);
@@ -284,24 +409,43 @@ void ia32_emit_mode_suffix(const ir_node *node)
 
 void ia32_emit_x87_mode_suffix(const ir_node *node)
 {
-       ir_mode *mode = get_ia32_ls_mode(node);
-       if(mode != NULL)
-               ia32_emit_mode_suffix_mode(mode);
+       ir_mode *mode;
+
+       /* we only need to emit the mode on address mode */
+       if (get_ia32_op_type(node) == ia32_Normal)
+               return;
+
+       mode = get_ia32_ls_mode(node);
+       assert(mode != NULL);
+
+       if (mode_is_float(mode)) {
+               switch (get_mode_size_bits(mode)) {
+                       case 32: be_emit_char('s'); return;
+                       case 64: be_emit_char('l'); return;
+                       case 80:
+                       case 96: be_emit_char('t'); return;
+               }
+       } else {
+               assert(mode_is_int(mode));
+               switch (get_mode_size_bits(mode)) {
+                       case 16: be_emit_char('s');     return;
+                       case 32: be_emit_char('l');     return;
+                       /* gas docu says q is the suffix but gcc, objdump and icc use ll
+                        * apparently */
+                       case 64: be_emit_cstring("ll"); return;
+               }
+       }
+       panic("Can't output mode_suffix for %+F", mode);
 }
 
-static
-char get_xmm_mode_suffix(ir_mode *mode)
+static char get_xmm_mode_suffix(ir_mode *mode)
 {
        assert(mode_is_float(mode));
        switch(get_mode_size_bits(mode)) {
-       case 32:
-               return 's';
-       case 64:
-               return 'd';
-       default:
-               assert(0);
+       case 32: return 's';
+       case 64: return 'd';
+       default: panic("Invalid XMM mode");
        }
-       return '%';
 }
 
 void ia32_emit_xmm_mode_suffix(const ir_node *node)
@@ -319,60 +463,19 @@ void ia32_emit_xmm_mode_suffix_s(const ir_node *node)
        be_emit_char(get_xmm_mode_suffix(mode));
 }
 
-void ia32_emit_extend_suffix(const ir_mode *mode)
+void ia32_emit_extend_suffix(const ir_node *node)
 {
-       if(get_mode_size_bits(mode) == 32)
+       ir_mode *mode = get_ia32_ls_mode(node);
+       if (get_mode_size_bits(mode) == 32)
                return;
-       if(mode_is_signed(mode)) {
-               be_emit_char('s');
-       } else {
-               be_emit_char('z');
-       }
-}
-
-static
-void ia32_emit_function_object(const char *name)
-{
-       switch (be_gas_flavour) {
-       case GAS_FLAVOUR_NORMAL:
-               be_emit_cstring("\t.type\t");
-               be_emit_string(name);
-               be_emit_cstring(", @function\n");
-               be_emit_write_line();
-               break;
-       case GAS_FLAVOUR_MINGW:
-               be_emit_cstring("\t.def\t");
-               be_emit_string(name);
-               be_emit_cstring(";\t.scl\t2;\t.type\t32;\t.endef\n");
-               be_emit_write_line();
-               break;
-       default:
-               break;
-       }
-}
-
-static
-void ia32_emit_function_size(const char *name)
-{
-       switch (be_gas_flavour) {
-       case GAS_FLAVOUR_NORMAL:
-               be_emit_cstring("\t.size\t");
-               be_emit_string(name);
-               be_emit_cstring(", .-");
-               be_emit_string(name);
-               be_emit_char('\n');
-               be_emit_write_line();
-               break;
-       default:
-               break;
-       }
+       be_emit_char(mode_is_signed(mode) ? 's' : 'z');
+       ia32_emit_mode_suffix_mode(mode);
 }
 
-
 void ia32_emit_source_register_or_immediate(const ir_node *node, int pos)
 {
        ir_node *in = get_irn_n(node, pos);
-       if(is_ia32_Immediate(in)) {
+       if (is_ia32_Immediate(in)) {
                emit_ia32_Immediate(in);
        } else {
                const ir_mode         *mode = get_ia32_ls_mode(node);
@@ -382,123 +485,104 @@ void ia32_emit_source_register_or_immediate(const ir_node *node, int pos)
 }
 
 /**
- * Emits registers and/or address mode of a binary operation.
+ * Returns the target block for a control flow node.
  */
-void ia32_emit_binop(const ir_node *node) {
-       const ir_node         *right_op  = get_irn_n(node, n_ia32_binary_right);
-       const ir_mode         *mode      = get_ia32_ls_mode(node);
-       const arch_register_t *reg_left;
-
-       switch(get_ia32_op_type(node)) {
-       case ia32_Normal:
-               reg_left = get_in_reg(node, n_ia32_binary_left);
-               if(is_ia32_Immediate(right_op)) {
-                       emit_ia32_Immediate(right_op);
-                       be_emit_cstring(", ");
-                       emit_register(reg_left, mode);
-                       break;
-               } else {
-                       const arch_register_t *reg_right
-                               = get_in_reg(node, n_ia32_binary_right);
-                       emit_register(reg_right, mode);
-                       be_emit_cstring(", ");
-                       emit_register(reg_left, mode);
-               }
-               break;
-       case ia32_AddrModeS:
-               if(is_ia32_Immediate(right_op)) {
-                       emit_ia32_Immediate(right_op);
-                       be_emit_cstring(", ");
-                       ia32_emit_am(node);
-               } else {
-                       reg_left = get_in_reg(node, n_ia32_binary_left);
-                       ia32_emit_am(node);
-                       be_emit_cstring(", ");
-                       emit_register(reg_left, mode);
-               }
-               break;
-       case ia32_AddrModeD:
-               panic("DestMode can't be output by %%binop anymore");
-               break;
-       default:
-               assert(0 && "unsupported op type");
-       }
+static ir_node *get_cfop_target_block(const ir_node *irn)
+{
+       assert(get_irn_mode(irn) == mode_X);
+       return get_irn_link(irn);
 }
 
 /**
- * Emits registers and/or address mode of a binary operation.
+ * Emits a block label for the given block.
  */
-void ia32_emit_x87_binop(const ir_node *node) {
-       switch(get_ia32_op_type(node)) {
-               case ia32_Normal:
-                       {
-                               const ia32_x87_attr_t *x87_attr = get_ia32_x87_attr_const(node);
-                               const arch_register_t *in1      = x87_attr->x87[0];
-                               const arch_register_t *in2      = x87_attr->x87[1];
-                               const arch_register_t *out      = x87_attr->x87[2];
-                               const arch_register_t *in;
-
-                               in  = out ? ((out == in2) ? in1 : in2) : in2;
-                               out = out ? out : in1;
-
-                               be_emit_char('%');
-                               be_emit_string(arch_register_get_name(in));
-                               be_emit_cstring(", %");
-                               be_emit_string(arch_register_get_name(out));
-                       }
-                       break;
-               case ia32_AddrModeS:
-                       ia32_emit_am(node);
-                       break;
-               case ia32_AddrModeD:
-               default:
-                       assert(0 && "unsupported op type");
+static void ia32_emit_block_name(const ir_node *block)
+{
+       if (has_Block_entity(block)) {
+               ir_entity *entity = get_Block_entity(block);
+               be_gas_emit_entity(entity);
+       } else {
+               be_emit_cstring(BLOCK_PREFIX);
+               be_emit_irprintf("%ld", get_irn_node_nr(block));
        }
 }
 
-void ia32_emit_am_or_dest_register(const ir_node *node,
-                                   int pos) {
-       if(get_ia32_op_type(node) == ia32_Normal) {
-               ia32_emit_dest_register(node, pos);
-       } else {
-               assert(get_ia32_op_type(node) == ia32_AddrModeD);
-               ia32_emit_am(node);
-       }
+/**
+ * Emits the target label for a control flow node.
+ */
+static void ia32_emit_cfop_target(const ir_node *node)
+{
+       ir_node *block = get_cfop_target_block(node);
+       ia32_emit_block_name(block);
 }
 
+/*
+ * positive conditions for signed compares
+ */
+static const char *const cmp2condition_s[] = {
+       NULL, /* always false */
+       "e",  /* == */
+       "l",  /* <  */
+       "le", /* <= */
+       "g",  /* >  */
+       "ge", /* >= */
+       "ne", /* != */
+       NULL  /* always true */
+};
+
+/*
+ * positive conditions for unsigned compares
+ */
+static const char *const cmp2condition_u[] = {
+       NULL, /* always false */
+       "e",  /* == */
+       "b",  /* <  */
+       "be", /* <= */
+       "a",  /* >  */
+       "ae", /* >= */
+       "ne", /* != */
+       NULL  /* always true */
+};
+
 /**
- * Emits registers and/or address mode of a unary operation.
+ * Emit the suffix for a compare instruction.
  */
-void ia32_emit_unop(const ir_node *node, int pos) {
-       const ir_node *op;
+static void ia32_emit_cmp_suffix(int pnc)
+{
+       const char *str;
 
-       switch(get_ia32_op_type(node)) {
-       case ia32_Normal:
-               op = get_irn_n(node, pos);
-               if (is_ia32_Immediate(op)) {
-                       emit_ia32_Immediate(op);
-               } else {
-                       ia32_emit_source_register(node, pos);
-               }
-               break;
-       case ia32_AddrModeS:
-       case ia32_AddrModeD:
-               ia32_emit_am(node);
-               break;
-       default:
-               assert(0 && "unsupported op type");
+       if (pnc == ia32_pn_Cmp_parity) {
+               be_emit_char('p');
+               return;
        }
+
+       if (pnc & ia32_pn_Cmp_float || pnc & ia32_pn_Cmp_unsigned) {
+               str = cmp2condition_u[pnc & 7];
+       } else {
+               str = cmp2condition_s[pnc & 7];
+       }
+
+       be_emit_string(str);
 }
 
+typedef enum ia32_emit_mod_t {
+       EMIT_RESPECT_LS   = 1U << 0,
+       EMIT_ALTERNATE_AM = 1U << 1,
+       EMIT_LONG         = 1U << 2,
+       EMIT_HIGH_REG     = 1U << 3,
+       EMIT_LOW_REG      = 1U << 4
+} ia32_emit_mod_t;
+
 /**
  * Emits address mode.
  */
-void ia32_emit_am(const ir_node *node) {
+void ia32_emit_am(const ir_node *node)
+{
        ir_entity *ent       = get_ia32_am_sc(node);
        int        offs      = get_ia32_am_offs_int(node);
-       ir_node   *base      = get_irn_n(node, 0);
+       ir_node   *base      = get_irn_n(node, n_ia32_base);
        int        has_base  = !is_ia32_NoReg_GP(base);
-       ir_node   *index     = get_irn_n(node, 1);
+       ir_node   *index     = get_irn_n(node, n_ia32_index);
        int        has_index = !is_ia32_NoReg_GP(index);
 
        /* just to be sure... */
@@ -506,25 +590,15 @@ void ia32_emit_am(const ir_node *node) {
 
        /* emit offset */
        if (ent != NULL) {
-               ident *id;
-
-               set_entity_backend_marked(ent, 1);
-               id = get_entity_ld_ident(ent);
+               const ia32_attr_t *attr = get_ia32_attr_const(node);
                if (is_ia32_am_sc_sign(node))
                        be_emit_char('-');
-               be_emit_ident(id);
-
-               if(get_entity_owner(ent) == get_tls_type()) {
-                       if (get_entity_visibility(ent) == visibility_external_allocated) {
-                               be_emit_cstring("@INDNTPOFF");
-                       } else {
-                               be_emit_cstring("@NTPOFF");
-                       }
-               }
+               ia32_emit_entity(ent, attr->data.am_sc_no_pic_adjust);
        }
 
-       if(offs != 0) {
-               if(ent != NULL) {
+       /* also handle special case if nothing is set */
+       if (offs != 0 || (ent == NULL && !has_base && !has_index)) {
+               if (ent != NULL) {
                        be_emit_irprintf("%+d", offs);
                } else {
                        be_emit_irprintf("%d", offs);
@@ -549,224 +623,413 @@ void ia32_emit_am(const ir_node *node) {
 
                        scale = get_ia32_am_scale(node);
                        if (scale > 0) {
-                               be_emit_irprintf(",%d", 1 << get_ia32_am_scale(node));
+                               be_emit_irprintf(",%d", 1 << scale);
                        }
                }
                be_emit_char(')');
        }
-
-       /* special case if nothing is set */
-       if(ent == NULL && offs == 0 && !has_base && !has_index) {
-               be_emit_char('0');
-       }
 }
 
-/*************************************************
- *                 _ _                         _
- *                (_) |                       | |
- *   ___ _ __ ___  _| |_    ___ ___  _ __   __| |
- *  / _ \ '_ ` _ \| | __|  / __/ _ \| '_ \ / _` |
- * |  __/ | | | | | | |_  | (_| (_) | | | | (_| |
- *  \___|_| |_| |_|_|\__|  \___\___/|_| |_|\__,_|
+/**
+ * fmt  parameter               output
+ * ---- ----------------------  ---------------------------------------------
+ * %%                           %
+ * %AM  <node>                  address mode of the node
+ * %AR  const arch_register_t*  address mode of the node or register
+ * %ASx <node>                  address mode of the node or source register x
+ * %Dx  <node>                  destination register x
+ * %I   <node>                  immediate of the node
+ * %L   <node>                  control flow target of the node
+ * %M   <node>                  mode suffix of the node
+ * %P   int                     condition code
+ * %R   const arch_register_t*  register
+ * %Sx  <node>                  source register x
+ * %s   const char*             string
+ * %u   unsigned int            unsigned int
+ * %d   signed int              signed int
  *
- *************************************************/
-
-#undef IA32_DO_EMIT
-#define IA32_DO_EMIT(irn) ia32_fprintf_format(F, irn, cmd_buf, cmnt_buf)
-
-/*
- * coding of conditions
- */
-struct cmp2conditon_t {
-       const char *name;
-       int         num;
-};
-
-/*
- * positive conditions for signed compares
+ * x starts at 0
+ * # modifier for %ASx, %D, %R, and %S uses ls mode of node to alter register width
+ * * modifier does not prefix immediates with $, but AM with *
+ * l modifier for %lu and %ld
+ * > modifier to output high 8bit register (ah, bh)
+ * < modifier to output low 8bit register (al, bl)
  */
-static const struct cmp2conditon_t cmp2condition_s[] = {
-       { NULL,              pn_Cmp_False },  /* always false */
-       { "e",               pn_Cmp_Eq },     /* == */
-       { "l",               pn_Cmp_Lt },     /* < */
-       { "le",              pn_Cmp_Le },     /* <= */
-       { "g",               pn_Cmp_Gt },     /* > */
-       { "ge",              pn_Cmp_Ge },     /* >= */
-       { "ne",              pn_Cmp_Lg },     /* != */
-       { NULL,              pn_Cmp_Leg},     /* always true */
-};
+static void ia32_emitf(const ir_node *node, const char *fmt, ...)
+{
+       va_list ap;
+       va_start(ap, fmt);
 
-/*
- * positive conditions for unsigned compares
- */
-static const struct cmp2conditon_t cmp2condition_u[] = {
-       { NULL,              pn_Cmp_False },  /* always false */
-       { "e",               pn_Cmp_Eq },     /* == */
-       { "b",               pn_Cmp_Lt },     /* < */
-       { "be",              pn_Cmp_Le },     /* <= */
-       { "a",               pn_Cmp_Gt },     /* > */
-       { "ae",              pn_Cmp_Ge },     /* >= */
-       { "ne",              pn_Cmp_Lg },     /* != */
-       { NULL,              pn_Cmp_Leg },   /* always true  */
-};
+       for (;;) {
+               const char      *start = fmt;
+               ia32_emit_mod_t  mod   = 0;
 
-enum {
-       ia32_pn_Cmp_unsigned = 0x1000,
-       ia32_pn_Cmp_float    = 0x2000,
-};
+               while (*fmt != '%' && *fmt != '\n' && *fmt != '\0')
+                       ++fmt;
+               if (fmt != start) {
+                       be_emit_string_len(start, fmt - start);
+               }
 
-/**
- * walks up a tree of copies/perms/spills/reloads to find the original value
- * that is moved around
- */
-static ir_node *find_original_value(ir_node *node)
-{
-       inc_irg_visited(current_ir_graph);
-       while(1) {
-               mark_irn_visited(node);
-               if(be_is_Copy(node)) {
-                       node = be_get_Copy_op(node);
-               } else if(be_is_CopyKeep(node)) {
-                       node = be_get_CopyKeep_op(node);
-               } else if(is_Proj(node)) {
-                       ir_node *pred = get_Proj_pred(node);
-                       if(be_is_Perm(pred)) {
-                               node = get_irn_n(pred, get_Proj_proj(node));
-                       } else if(be_is_MemPerm(pred)) {
-                               node = get_irn_n(pred, get_Proj_proj(node) + 1);
-                       } else if(is_ia32_Load(pred)) {
-                               node = get_irn_n(pred, n_ia32_Load_mem);
-                       } else {
-                               return node;
-                       }
-               } else if(is_ia32_Store(node)) {
-                       node = get_irn_n(node, n_ia32_Store_val);
-               } else if(is_Phi(node)) {
-                       int i, arity;
-                       arity = get_irn_arity(node);
-                       for(i = 0; i < arity; ++i) {
-                               ir_node *in = get_irn_n(node, i);
-                               if(irn_visited(in))
-                                       continue;
-                               node = in;
+               if (*fmt == '\n') {
+                       be_emit_finish_line_gas(node);
+                       ++fmt;
+                       if (*fmt == '\0')
                                break;
-                       }
-                       assert(i < arity);
-               } else {
-                       return node;
+                       continue;
                }
-       }
-}
 
-static int determine_final_pnc(const ir_node *node, int flags_pos,
-                               int pnc)
-{
-       ir_node           *flags = get_irn_n(node, flags_pos);
-       const ia32_attr_t *flags_attr;
-       flags = skip_Proj(flags);
+               if (*fmt == '\0')
+                       break;
 
-       if(is_ia32_Sahf(flags)) {
-               ir_node *cmp = get_irn_n(flags, n_ia32_Sahf_val);
-               if(!(is_ia32_FucomFnstsw(cmp) || is_ia32_FucompFnstsw(cmp)
-                               || is_ia32_FucomppFnstsw(cmp) || is_ia32_FtstFnstsw(cmp))) {
-                       cmp = find_original_value(cmp);
-                       assert(is_ia32_FucomFnstsw(cmp) || is_ia32_FucompFnstsw(cmp)
-                              || is_ia32_FucomppFnstsw(cmp) || is_ia32_FtstFnstsw(cmp));
+               ++fmt;
+               while (1) {
+                       switch(*fmt) {
+                       case '*': mod |= EMIT_ALTERNATE_AM; break;
+                       case '#': mod |= EMIT_RESPECT_LS;   break;
+                       case 'l': mod |= EMIT_LONG;         break;
+                       case '>': mod |= EMIT_HIGH_REG;     break;
+                       case '<': mod |= EMIT_LOW_REG;      break;
+                       default:
+                               goto end_of_mods;
+                       }
+                       ++fmt;
                }
+end_of_mods:
 
-               flags_attr = get_ia32_attr_const(cmp);
-               if(flags_attr->data.cmp_flipped)
-                       pnc = get_mirrored_pnc(pnc);
-               pnc |= ia32_pn_Cmp_float;
-       } else if(is_ia32_Ucomi(flags)) {
-               flags_attr = get_ia32_attr_const(flags);
+               switch (*fmt++) {
+                       case '%':
+                               be_emit_char('%');
+                               break;
 
-               if(flags_attr->data.cmp_flipped)
-                       pnc = get_mirrored_pnc(pnc);
-               pnc |= ia32_pn_Cmp_float;
-       } else {
-               assert(is_ia32_Cmp(flags) || is_ia32_Test(flags)
-                               || is_ia32_Cmp8Bit(flags) || is_ia32_Test8Bit(flags));
-               flags_attr = get_ia32_attr_const(flags);
+                       case 'A': {
+                               switch (*fmt++) {
+emit_AM:
+                                       case 'M':
+                                               if (mod & EMIT_ALTERNATE_AM)
+                                                       be_emit_char('*');
+                                               ia32_emit_am(node);
+                                               break;
+
+                                       case 'R': {
+                                               const arch_register_t *reg = va_arg(ap, const arch_register_t*);
+                                               if (get_ia32_op_type(node) == ia32_AddrModeS) {
+                                                       goto emit_AM;
+                                               } else {
+                                                       if (mod & EMIT_ALTERNATE_AM)
+                                                               be_emit_char('*');
+                                                       emit_register(reg, NULL);
+                                               }
+                                               break;
+                                       }
+
+                                       case 'S':
+                                               if (get_ia32_op_type(node) == ia32_AddrModeS) {
+                                                       ++fmt;
+                                                       goto emit_AM;
+                                               } else {
+                                                       assert(get_ia32_op_type(node) == ia32_Normal);
+                                                       goto emit_S;
+                                               }
+                                               break;
+
+                                       default: goto unknown;
+                               }
+                               break;
+                       }
 
-               if(flags_attr->data.cmp_flipped)
-                       pnc = get_mirrored_pnc(pnc);
-               if(flags_attr->data.cmp_unsigned)
-                       pnc |= ia32_pn_Cmp_unsigned;
+                       case 'D': {
+                               unsigned               pos;
+                               const arch_register_t *reg;
+
+                               if (*fmt < '0' || '9' <= *fmt)
+                                       goto unknown;
+
+                               pos = *fmt++ - '0';
+                               reg = get_out_reg(node, pos);
+                               emit_register(reg, mod & EMIT_RESPECT_LS ? get_ia32_ls_mode(node) : NULL);
+                               break;
+                       }
+
+                       case 'I':
+                               if (!(mod & EMIT_ALTERNATE_AM))
+                                       be_emit_char('$');
+                               emit_ia32_Immediate_no_prefix(node);
+                               break;
+
+                       case 'L':
+                               ia32_emit_cfop_target(node);
+                               break;
+
+                       case 'M': {
+                               ia32_emit_mode_suffix_mode(get_ia32_ls_mode(node));
+                               break;
+                       }
+
+                       case 'P': {
+                               int pnc = va_arg(ap, int);
+                               ia32_emit_cmp_suffix(pnc);
+                               break;
+                       }
+
+                       case 'R': {
+                               const arch_register_t *reg = va_arg(ap, const arch_register_t*);
+                               if (mod & EMIT_HIGH_REG) {
+                                       emit_8bit_register_high(reg);
+                               } else if (mod & EMIT_LOW_REG) {
+                                       emit_8bit_register(reg);
+                               } else {
+                                       emit_register(reg, mod & EMIT_RESPECT_LS ? get_ia32_ls_mode(node) : NULL);
+                               }
+                               break;
+                       }
+
+emit_S:
+                       case 'S': {
+                               unsigned       pos;
+                               const ir_node *in;
+
+                               if (*fmt < '0' || '9' <= *fmt)
+                                       goto unknown;
+
+                               pos = *fmt++ - '0';
+                               in  = get_irn_n(node, pos);
+                               if (is_ia32_Immediate(in)) {
+                                       if (!(mod & EMIT_ALTERNATE_AM))
+                                               be_emit_char('$');
+                                       emit_ia32_Immediate_no_prefix(in);
+                               } else {
+                                       const arch_register_t *reg;
+
+                                       if (mod & EMIT_ALTERNATE_AM)
+                                               be_emit_char('*');
+                                       reg = get_in_reg(node, pos);
+                                       emit_register(reg, mod & EMIT_RESPECT_LS ? get_ia32_ls_mode(node) : NULL);
+                               }
+                               break;
+                       }
+
+                       case 's': {
+                               const char *str = va_arg(ap, const char*);
+                               be_emit_string(str);
+                               break;
+                       }
+
+                       case 'u':
+                               if (mod & EMIT_LONG) {
+                                       unsigned long num = va_arg(ap, unsigned long);
+                                       be_emit_irprintf("%lu", num);
+                               } else {
+                                       unsigned num = va_arg(ap, unsigned);
+                                       be_emit_irprintf("%u", num);
+                               }
+                               break;
+
+                       case 'd':
+                               if (mod & EMIT_LONG) {
+                                       long num = va_arg(ap, long);
+                                       be_emit_irprintf("%ld", num);
+                               } else {
+                                       int num = va_arg(ap, int);
+                                       be_emit_irprintf("%d", num);
+                               }
+                               break;
+
+                       default:
+unknown:
+                               panic("unknown format conversion in ia32_emitf()");
+               }
        }
 
-       return pnc;
+       va_end(ap);
 }
 
-static void ia32_emit_cmp_suffix(int pnc)
+/**
+ * Emits registers and/or address mode of a binary operation.
+ */
+void ia32_emit_binop(const ir_node *node)
 {
-       const char        *str;
-
-       if((pnc & ia32_pn_Cmp_float) || (pnc & ia32_pn_Cmp_unsigned)) {
-               pnc = pnc & 7;
-               assert(cmp2condition_u[pnc].num == pnc);
-               str = cmp2condition_u[pnc].name;
+       if (is_ia32_Immediate(get_irn_n(node, n_ia32_binary_right))) {
+               ia32_emitf(node, "%#S4, %#AS3");
        } else {
-               pnc = pnc & 7;
-               assert(cmp2condition_s[pnc].num == pnc);
-               str = cmp2condition_s[pnc].name;
+               ia32_emitf(node, "%#AS4, %#S3");
        }
-
-       be_emit_string(str);
 }
 
-void ia32_emit_cmp_suffix_node(const ir_node *node,
-                               int flags_pos)
+/**
+ * Emits registers and/or address mode of a binary operation.
+ */
+void ia32_emit_x87_binop(const ir_node *node)
 {
-       pn_Cmp pnc = get_ia32_pncode(node);
+       switch(get_ia32_op_type(node)) {
+               case ia32_Normal:
+                       {
+                               const ia32_x87_attr_t *x87_attr = get_ia32_x87_attr_const(node);
+                               const arch_register_t *in1      = x87_attr->x87[0];
+                               const arch_register_t *in       = x87_attr->x87[1];
+                               const arch_register_t *out      = x87_attr->x87[2];
 
-       pnc = determine_final_pnc(node, flags_pos, pnc);
-       ia32_emit_cmp_suffix(pnc);
+                               if (out == NULL) {
+                                       out = in1;
+                               } else if (out == in) {
+                                       in = in1;
+                               }
+
+                               be_emit_char('%');
+                               be_emit_string(arch_register_get_name(in));
+                               be_emit_cstring(", %");
+                               be_emit_string(arch_register_get_name(out));
+                       }
+                       break;
+               case ia32_AddrModeS:
+                       ia32_emit_am(node);
+                       break;
+               case ia32_AddrModeD:
+               default:
+                       assert(0 && "unsupported op type");
+       }
 }
 
 /**
- * Returns the target block for a control flow node.
+ * Emits registers and/or address mode of a unary operation.
  */
-static
-ir_node *get_cfop_target_block(const ir_node *irn) {
-       return get_irn_link(irn);
+void ia32_emit_unop(const ir_node *node, int pos)
+{
+       char fmt[] = "%ASx";
+       fmt[3] = '0' + pos;
+       ia32_emitf(node, fmt);
 }
 
-/**
- * Emits a block label for the given block.
- */
-static
-void ia32_emit_block_name(const ir_node *block)
+static void emit_ia32_IMul(const ir_node *node)
 {
-       if (has_Block_label(block)) {
-               be_emit_string(be_gas_label_prefix());
-               be_emit_irprintf("%u", (unsigned)get_Block_label(block));
+       ir_node               *left    = get_irn_n(node, n_ia32_IMul_left);
+       const arch_register_t *out_reg = get_out_reg(node, pn_ia32_IMul_res);
+
+       /* do we need the 3-address form? */
+       if (is_ia32_NoReg_GP(left) ||
+                       get_in_reg(node, n_ia32_IMul_left) != out_reg) {
+               ia32_emitf(node, "\timul%M %#S4, %#AS3, %#D0\n");
        } else {
-               be_emit_cstring(BLOCK_PREFIX);
-               be_emit_irprintf("%d", get_irn_node_nr(block));
+               ia32_emitf(node, "\timul%M %#AS4, %#S3\n");
        }
 }
 
 /**
- * Emits the target label for a control flow node.
+ * walks up a tree of copies/perms/spills/reloads to find the original value
+ * that is moved around
  */
-static void ia32_emit_cfop_target(const ir_node *node)
+static ir_node *find_original_value(ir_node *node)
 {
-       ir_node *block = get_cfop_target_block(node);
+       if (irn_visited(node))
+               return NULL;
+
+       mark_irn_visited(node);
+       if (be_is_Copy(node)) {
+               return find_original_value(be_get_Copy_op(node));
+       } else if (be_is_CopyKeep(node)) {
+               return find_original_value(be_get_CopyKeep_op(node));
+       } else if (is_Proj(node)) {
+               ir_node *pred = get_Proj_pred(node);
+               if (be_is_Perm(pred)) {
+                       return find_original_value(get_irn_n(pred, get_Proj_proj(node)));
+               } else if (be_is_MemPerm(pred)) {
+                       return find_original_value(get_irn_n(pred, get_Proj_proj(node) + 1));
+               } else if (is_ia32_Load(pred)) {
+                       return find_original_value(get_irn_n(pred, n_ia32_Load_mem));
+               } else {
+                       return node;
+               }
+       } else if (is_ia32_Store(node)) {
+               return find_original_value(get_irn_n(node, n_ia32_Store_val));
+       } else if (is_Phi(node)) {
+               int i, arity;
+               arity = get_irn_arity(node);
+               for (i = 0; i < arity; ++i) {
+                       ir_node *in  = get_irn_n(node, i);
+                       ir_node *res = find_original_value(in);
 
-       ia32_emit_block_name(block);
+                       if (res != NULL)
+                               return res;
+               }
+               return NULL;
+       } else {
+               return node;
+       }
 }
 
-/** Return the next block in Block schedule */
-static ir_node *next_blk_sched(const ir_node *block)
+static int determine_final_pnc(const ir_node *node, int flags_pos, int pnc)
 {
-       return get_irn_link(block);
+       ir_node           *flags = get_irn_n(node, flags_pos);
+       const ia32_attr_t *flags_attr;
+       flags = skip_Proj(flags);
+
+       if (is_ia32_Sahf(flags)) {
+               ir_node *cmp = get_irn_n(flags, n_ia32_Sahf_val);
+               if (!(is_ia32_FucomFnstsw(cmp) || is_ia32_FucompFnstsw(cmp)
+                               || is_ia32_FucomppFnstsw(cmp) || is_ia32_FtstFnstsw(cmp))) {
+                       inc_irg_visited(current_ir_graph);
+                       cmp = find_original_value(cmp);
+                       assert(cmp != NULL);
+                       assert(is_ia32_FucomFnstsw(cmp) || is_ia32_FucompFnstsw(cmp)
+                              || is_ia32_FucomppFnstsw(cmp) || is_ia32_FtstFnstsw(cmp));
+               }
+
+               flags_attr = get_ia32_attr_const(cmp);
+               if (flags_attr->data.ins_permuted)
+                       pnc = get_mirrored_pnc(pnc);
+               pnc |= ia32_pn_Cmp_float;
+       } else if (is_ia32_Ucomi(flags) || is_ia32_Fucomi(flags)
+                       || is_ia32_Fucompi(flags)) {
+               flags_attr = get_ia32_attr_const(flags);
+
+               if (flags_attr->data.ins_permuted)
+                       pnc = get_mirrored_pnc(pnc);
+               pnc |= ia32_pn_Cmp_float;
+       } else {
+               flags_attr = get_ia32_attr_const(flags);
+
+               if (flags_attr->data.ins_permuted)
+                       pnc = get_mirrored_pnc(pnc);
+               if (flags_attr->data.cmp_unsigned)
+                       pnc |= ia32_pn_Cmp_unsigned;
+       }
+
+       return pnc;
+}
+
+static pn_Cmp ia32_get_negated_pnc(pn_Cmp pnc)
+{
+       ir_mode *mode = pnc & ia32_pn_Cmp_float ? mode_F : mode_Iu;
+       return get_negated_pnc(pnc, mode);
+}
+
+void ia32_emit_cmp_suffix_node(const ir_node *node,
+                               int flags_pos)
+{
+       const ia32_attr_t *attr = get_ia32_attr_const(node);
+
+       pn_Cmp pnc = get_ia32_condcode(node);
+
+       pnc = determine_final_pnc(node, flags_pos, pnc);
+       if (attr->data.ins_permuted)
+               pnc = ia32_get_negated_pnc(pnc);
+
+       ia32_emit_cmp_suffix(pnc);
+}
+
+/**
+ * Emits an exception label for a given node.
+ */
+static void ia32_emit_exc_label(const ir_node *node)
+{
+       be_emit_string(be_gas_insn_label_prefix());
+       be_emit_irprintf("%lu", get_ia32_exc_label_id(node));
 }
 
 /**
  * Returns the Proj with projection number proj and NOT mode_M
  */
-static ir_node *get_proj(const ir_node *node, long proj) {
+static ir_node *get_proj(const ir_node *node, long proj)
+{
        const ir_edge_t *edge;
        ir_node         *src;
 
@@ -785,16 +1048,23 @@ static ir_node *get_proj(const ir_node *node, long proj) {
        return NULL;
 }
 
+static int can_be_fallthrough(const ir_node *node)
+{
+       ir_node *target_block = get_cfop_target_block(node);
+       ir_node *block        = get_nodes_block(node);
+       return get_prev_block_sched(target_block) == block;
+}
+
 /**
  * Emits the jump sequence for a conditional jump (cmp + jmp_true + jmp_false)
  */
 static void emit_ia32_Jcc(const ir_node *node)
 {
+       int            need_parity_label = 0;
        const ir_node *proj_true;
        const ir_node *proj_false;
        const ir_node *block;
-       const ir_node *next_block;
-       pn_Cmp         pnc = get_ia32_pncode(node);
+       pn_Cmp         pnc = get_ia32_condcode(node);
 
        pnc = determine_final_pnc(node, 0, pnc);
 
@@ -806,124 +1076,162 @@ static void emit_ia32_Jcc(const ir_node *node)
        assert(proj_false && "Jcc without false Proj");
 
        block      = get_nodes_block(node);
-       next_block = next_blk_sched(block);
 
-       if (get_cfop_target_block(proj_true) == next_block) {
+       if (can_be_fallthrough(proj_true)) {
                /* exchange both proj's so the second one can be omitted */
                const ir_node *t = proj_true;
 
                proj_true  = proj_false;
                proj_false = t;
-               if(pnc & ia32_pn_Cmp_float) {
-                       pnc = get_negated_pnc(pnc, mode_F);
-               } else {
-                       pnc = get_negated_pnc(pnc, mode_Iu);
-               }
+               pnc        = ia32_get_negated_pnc(pnc);
        }
 
        if (pnc & ia32_pn_Cmp_float) {
                /* Some floating point comparisons require a test of the parity flag,
                 * which indicates that the result is unordered */
-               switch (pnc & 15) {
-                       case pn_Cmp_Uo:
-                               be_emit_cstring("\tjp ");
-                               ia32_emit_cfop_target(proj_true);
-                               be_emit_finish_line_gas(proj_true);
-                               break;
+               switch (pnc & 0x0f) {
+               case pn_Cmp_Uo: {
+                       ia32_emitf(proj_true, "\tjp %L\n");
+                       break;
+               }
 
-                       case pn_Cmp_Leg:
-                               be_emit_cstring("\tjnp ");
-                               ia32_emit_cfop_target(proj_true);
-                               be_emit_finish_line_gas(proj_true);
-                               break;
+               case pn_Cmp_Leg:
+                       ia32_emitf(proj_true, "\tjnp %L\n");
+                       break;
 
-                       case pn_Cmp_Eq:
-                       case pn_Cmp_Lt:
-                       case pn_Cmp_Le:
-                               be_emit_cstring("\tjp ");
-                               ia32_emit_cfop_target(proj_false);
-                               be_emit_finish_line_gas(proj_false);
-                               goto emit_jcc;
+               case pn_Cmp_Eq:
+               case pn_Cmp_Lt:
+               case pn_Cmp_Le:
+                       /* we need a local label if the false proj is a fallthrough
+                        * as the falseblock might have no label emitted then */
+                       if (can_be_fallthrough(proj_false)) {
+                               need_parity_label = 1;
+                               ia32_emitf(proj_false, "\tjp 1f\n");
+                       } else {
+                               ia32_emitf(proj_false, "\tjp %L\n");
+                       }
+                       goto emit_jcc;
 
-                       case pn_Cmp_Ug:
-                       case pn_Cmp_Uge:
-                       case pn_Cmp_Ne:
-                               be_emit_cstring("\tjp ");
-                               ia32_emit_cfop_target(proj_true);
-                               be_emit_finish_line_gas(proj_true);
-                               goto emit_jcc;
+               case pn_Cmp_Ug:
+               case pn_Cmp_Uge:
+               case pn_Cmp_Ne:
+                       ia32_emitf(proj_true, "\tjp %L\n");
+                       goto emit_jcc;
 
-                       default:
-                               goto emit_jcc;
+               default:
+                       goto emit_jcc;
                }
        } else {
 emit_jcc:
-               be_emit_cstring("\tj");
-               ia32_emit_cmp_suffix(pnc);
-               be_emit_char(' ');
-               ia32_emit_cfop_target(proj_true);
-               be_emit_finish_line_gas(proj_true);
+               ia32_emitf(proj_true, "\tj%P %L\n", pnc);
+       }
+
+       if (need_parity_label) {
+               ia32_emitf(NULL, "1:\n");
        }
 
        /* the second Proj might be a fallthrough */
-       if (get_cfop_target_block(proj_false) != next_block) {
-               be_emit_cstring("\tjmp ");
-               ia32_emit_cfop_target(proj_false);
-               be_emit_finish_line_gas(proj_false);
+       if (can_be_fallthrough(proj_false)) {
+               ia32_emitf(proj_false, "\t/* fallthrough to %L */\n");
        } else {
-               be_emit_cstring("\t/* fallthrough to ");
-               ia32_emit_cfop_target(proj_false);
-               be_emit_cstring(" */");
-               be_emit_finish_line_gas(proj_false);
+               ia32_emitf(proj_false, "\tjmp %L\n");
+       }
+}
+
+/**
+ * Emits an ia32 Setcc. This is mostly easy but some floating point compares
+ * are tricky.
+ */
+static void emit_ia32_Setcc(const ir_node *node)
+{
+       const arch_register_t *dreg = get_out_reg(node, pn_ia32_Setcc_res);
+
+       pn_Cmp pnc = get_ia32_condcode(node);
+       pnc        = determine_final_pnc(node, n_ia32_Setcc_eflags, pnc);
+       if (pnc & ia32_pn_Cmp_float) {
+               switch (pnc & 0x0f) {
+               case pn_Cmp_Uo:
+                       ia32_emitf(node, "\tsetp %#R\n", dreg);
+                       return;
+
+               case pn_Cmp_Leg:
+                       ia32_emitf(node, "\tsetnp %#R\n", dreg);
+                       return;
+
+               case pn_Cmp_Eq:
+               case pn_Cmp_Lt:
+               case pn_Cmp_Le:
+                       ia32_emitf(node, "\tset%P %<R\n", pnc, dreg);
+                       ia32_emitf(node, "\tsetnp %>R\n", dreg);
+                       ia32_emitf(node, "\tandb %>R, %<R\n", dreg, dreg);
+                       return;
+
+               case pn_Cmp_Ug:
+               case pn_Cmp_Uge:
+               case pn_Cmp_Ne:
+                       ia32_emitf(node, "\tset%P %<R\n", pnc, dreg);
+                       ia32_emitf(node, "\tsetp %>R\n", dreg);
+                       ia32_emitf(node, "\torb %>R, %<R\n", dreg, dreg);
+                       return;
+
+               default:
+                       break;
+               }
        }
+       ia32_emitf(node, "\tset%P %#R\n", pnc, dreg);
 }
 
-static void emit_ia32_CMov(const ir_node *node)
+static void emit_ia32_CMovcc(const ir_node *node)
 {
-       const arch_register_t *out = arch_get_irn_register(arch_env, node);
+       const ia32_attr_t     *attr         = get_ia32_attr_const(node);
+       const arch_register_t *out          = arch_irn_get_register(node, pn_ia32_res);
+       pn_Cmp                 pnc          = get_ia32_condcode(node);
        const arch_register_t *in_true;
        const arch_register_t *in_false;
-       pn_Cmp                 pnc = get_ia32_pncode(node);
 
-       pnc = determine_final_pnc(node, n_ia32_CMov_eflags, pnc);
+       pnc = determine_final_pnc(node, n_ia32_CMovcc_eflags, pnc);
+       /* although you can't set ins_permuted in the constructor it might still
+          be set by memory operand folding */
+       if (attr->data.ins_permuted)
+               pnc = ia32_get_negated_pnc(pnc);
 
-       in_true  = arch_get_irn_register(arch_env,
-                                        get_irn_n(node, n_ia32_CMov_val_true));
-       in_false = arch_get_irn_register(arch_env,
-                                        get_irn_n(node, n_ia32_CMov_val_false));
+       in_true  = arch_get_irn_register(get_irn_n(node, n_ia32_CMovcc_val_true));
+       in_false = arch_get_irn_register(get_irn_n(node, n_ia32_CMovcc_val_false));
 
        /* should be same constraint fullfilled? */
-       if(out == in_false) {
+       if (out == in_false) {
                /* yes -> nothing to do */
-       } else if(out == in_true) {
+       } else if (out == in_true) {
                const arch_register_t *tmp;
 
-               /* swap left/right and negate pnc */
-               pnc = get_negated_pnc(pnc, mode_Iu);
+               assert(get_ia32_op_type(node) == ia32_Normal);
+
+               pnc = ia32_get_negated_pnc(pnc);
 
                tmp      = in_true;
                in_true  = in_false;
                in_false = tmp;
        } else {
                /* we need a mov */
-               be_emit_cstring("\tmovl ");
-               emit_register(in_false, NULL);
-               be_emit_cstring(", ");
-               emit_register(out, NULL);
-               be_emit_finish_line_gas(node);
+               ia32_emitf(node, "\tmovl %R, %R\n", in_false, out);
        }
 
-       be_emit_cstring("\tcmov");
-       ia32_emit_cmp_suffix(pnc);
-       be_emit_char(' ');
-       if(get_ia32_op_type(node) == ia32_AddrModeS) {
-               ia32_emit_am(node);
-       } else {
-               emit_register(in_true, get_ia32_ls_mode(node));
+       /* TODO: handling of Nans isn't correct yet */
+       if (pnc & ia32_pn_Cmp_float) {
+               switch (pnc & 0x0f) {
+               case pn_Cmp_Uo:
+               case pn_Cmp_Leg:
+               case pn_Cmp_Eq:
+               case pn_Cmp_Lt:
+               case pn_Cmp_Le:
+               case pn_Cmp_Ug:
+               case pn_Cmp_Uge:
+               case pn_Cmp_Ne:
+                       panic("CMov with floatingpoint compare/parity not supported yet");
+               }
        }
-       be_emit_cstring(", ");
-       emit_register(out, get_ia32_ls_mode(node));
-       be_emit_finish_line_gas(node);
+
+       ia32_emitf(node, "\tcmov%P %#AR, %#R\n", pnc, in_true, out);
 }
 
 /*********************************************************
@@ -945,19 +1253,19 @@ typedef struct _branch_t {
 
 /* jump table for switch generation */
 typedef struct _jmp_tbl_t {
-       ir_node  *defProj;         /**< default target */
-       long      min_value;       /**< smallest switch case */
-       long      max_value;       /**< largest switch case */
-       long      num_branches;    /**< number of jumps */
-       char     *label;           /**< label of the jump table */
-       branch_t *branches;        /**< jump array */
+       ir_node  *defProj;                 /**< default target */
+       long      min_value;               /**< smallest switch case */
+       long      max_value;               /**< largest switch case */
+       long      num_branches;            /**< number of jumps */
+       char      label[SNPRINTF_BUF_LEN]; /**< label of the jump table */
+       branch_t *branches;                /**< jump array */
 } jmp_tbl_t;
 
 /**
  * Compare two variables of type branch_t. Used to sort all switch cases
  */
-static
-int ia32_cmp_branch_t(const void *a, const void *b) {
+static int ia32_cmp_branch_t(const void *a, const void *b)
+{
        branch_t *b1 = (branch_t *)a;
        branch_t *b2 = (branch_t *)b;
 
@@ -967,29 +1275,23 @@ int ia32_cmp_branch_t(const void *a, const void *b) {
                return 1;
 }
 
-/**
- * Emits code for a SwitchJmp (creates a jump table if
- * possible otherwise a cmp-jmp cascade). Port from
- * cggg ia32 backend
- */
-static
-void emit_ia32_SwitchJmp(const ir_node *node) {
-       unsigned long       interval;
-       int                 last_value, i;
+static void generate_jump_table(jmp_tbl_t *tbl, const ir_node *node)
+{
+       int                 i;
        long                pnc;
-       jmp_tbl_t           tbl;
+       long                default_pn;
        ir_node            *proj;
        const ir_edge_t    *edge;
 
        /* fill the table structure */
-       tbl.label        = xmalloc(SNPRINTF_BUF_LEN);
-       tbl.label        = get_unique_label(tbl.label, SNPRINTF_BUF_LEN, ".TBL_");
-       tbl.defProj      = NULL;
-       tbl.num_branches = get_irn_n_edges(node);
-       tbl.branches     = xcalloc(tbl.num_branches, sizeof(tbl.branches[0]));
-       tbl.min_value    = INT_MAX;
-       tbl.max_value    = INT_MIN;
-
+       get_unique_label(tbl->label, SNPRINTF_BUF_LEN, ".TBL_");
+       tbl->defProj      = NULL;
+       tbl->num_branches = get_irn_n_edges(node) - 1;
+       tbl->branches     = XMALLOCNZ(branch_t, tbl->num_branches);
+       tbl->min_value    = LONG_MAX;
+       tbl->max_value    = LONG_MIN;
+
+       default_pn = get_ia32_condcode(node);
        i = 0;
        /* go over all proj's and collect them */
        foreach_out_edge(node, edge) {
@@ -998,137 +1300,112 @@ void emit_ia32_SwitchJmp(const ir_node *node) {
 
                pnc = get_Proj_proj(proj);
 
-               /* create branch entry */
-               tbl.branches[i].target = proj;
-               tbl.branches[i].value  = pnc;
-
-               tbl.min_value = pnc < tbl.min_value ? pnc : tbl.min_value;
-               tbl.max_value = pnc > tbl.max_value ? pnc : tbl.max_value;
-
                /* check for default proj */
-               if (pnc == get_ia32_pncode(node)) {
-                       assert(tbl.defProj == NULL && "found two defProjs at SwitchJmp");
-                       tbl.defProj = proj;
+               if (pnc == default_pn) {
+                       assert(tbl->defProj == NULL && "found two default Projs at SwitchJmp");
+                       tbl->defProj = proj;
+               } else {
+                       tbl->min_value = pnc < tbl->min_value ? pnc : tbl->min_value;
+                       tbl->max_value = pnc > tbl->max_value ? pnc : tbl->max_value;
+
+                       /* create branch entry */
+                       tbl->branches[i].target = proj;
+                       tbl->branches[i].value  = pnc;
+                       ++i;
                }
 
-               i++;
        }
+       assert(i == tbl->num_branches);
 
        /* sort the branches by their number */
-       qsort(tbl.branches, tbl.num_branches, sizeof(tbl.branches[0]), ia32_cmp_branch_t);
+       qsort(tbl->branches, tbl->num_branches, sizeof(tbl->branches[0]), ia32_cmp_branch_t);
+}
+
+/**
+ * Emits code for a SwitchJmp (creates a jump table if
+ * possible otherwise a cmp-jmp cascade). Port from
+ * cggg ia32 backend
+ */
+static void emit_ia32_SwitchJmp(const ir_node *node)
+{
+       unsigned long       interval;
+       int                 last_value, i;
+       jmp_tbl_t           tbl;
+
+       /* fill the table structure */
+       generate_jump_table(&tbl, node);
 
        /* two-complement's magic make this work without overflow */
        interval = tbl.max_value - tbl.min_value;
 
        /* emit the table */
-       be_emit_cstring("\tcmpl $");
-       be_emit_irprintf("%u, ", interval);
-       ia32_emit_source_register(node, 0);
-       be_emit_finish_line_gas(node);
-
-       be_emit_cstring("\tja ");
-       ia32_emit_cfop_target(tbl.defProj);
-       be_emit_finish_line_gas(node);
+       ia32_emitf(node,        "\tcmpl $%u, %S0\n", interval);
+       ia32_emitf(tbl.defProj, "\tja %L\n");
 
        if (tbl.num_branches > 1) {
                /* create table */
-               be_emit_cstring("\tjmp *");
-               be_emit_string(tbl.label);
-               be_emit_cstring("(,");
-               ia32_emit_source_register(node, 0);
-               be_emit_cstring(",4)");
-               be_emit_finish_line_gas(node);
+               ia32_emitf(node, "\tjmp *%s(,%S0,4)\n", tbl.label);
 
                be_gas_emit_switch_section(GAS_SECTION_RODATA);
-               be_emit_cstring("\t.align 4\n");
-               be_emit_write_line();
-
-               be_emit_string(tbl.label);
-               be_emit_cstring(":\n");
-               be_emit_write_line();
-
-               be_emit_cstring(".long ");
-               ia32_emit_cfop_target(tbl.branches[0].target);
-               be_emit_finish_line_gas(NULL);
+               ia32_emitf(NULL, "\t.align 4\n");
+               ia32_emitf(NULL, "%s:\n", tbl.label);
 
                last_value = tbl.branches[0].value;
-               for (i = 1; i < tbl.num_branches; ++i) {
-                       while (++last_value < tbl.branches[i].value) {
-                               be_emit_cstring(".long ");
-                               ia32_emit_cfop_target(tbl.defProj);
-                               be_emit_finish_line_gas(NULL);
+               for (i = 0; i != tbl.num_branches; ++i) {
+                       while (last_value != tbl.branches[i].value) {
+                               ia32_emitf(tbl.defProj, ".long %L\n");
+                               ++last_value;
                        }
-                       be_emit_cstring(".long ");
-                       ia32_emit_cfop_target(tbl.branches[i].target);
-                       be_emit_finish_line_gas(NULL);
+                       ia32_emitf(tbl.branches[i].target, ".long %L\n");
+                       ++last_value;
                }
                be_gas_emit_switch_section(GAS_SECTION_TEXT);
        } else {
                /* one jump is enough */
-               be_emit_cstring("\tjmp ");
-               ia32_emit_cfop_target(tbl.branches[0].target);
-               be_emit_finish_line_gas(node);
+               ia32_emitf(tbl.branches[0].target, "\tjmp %L\n");
        }
 
-       if (tbl.label)
-               free(tbl.label);
-       if (tbl.branches)
-               free(tbl.branches);
+       free(tbl.branches);
 }
 
 /**
  * Emits code for a unconditional jump.
  */
-static void emit_Jmp(const ir_node *node)
+static void emit_ia32_Jmp(const ir_node *node)
 {
-       ir_node *block, *next_block;
+       ir_node *block;
 
        /* for now, the code works for scheduled and non-schedules blocks */
        block = get_nodes_block(node);
 
        /* we have a block schedule */
-       next_block = next_blk_sched(block);
-       if (get_cfop_target_block(node) != next_block) {
-               be_emit_cstring("\tjmp ");
-               ia32_emit_cfop_target(node);
+       if (can_be_fallthrough(node)) {
+               ia32_emitf(node, "\t/* fallthrough to %L */\n");
        } else {
-               be_emit_cstring("\t/* fallthrough to ");
-               ia32_emit_cfop_target(node);
-               be_emit_cstring(" */");
-       }
-       be_emit_finish_line_gas(node);
-}
-
-static void emit_ia32_Immediate(const ir_node *node)
-{
-       const ia32_immediate_attr_t *attr = get_ia32_immediate_attr_const(node);
-
-       be_emit_char('$');
-       if(attr->symconst != NULL) {
-               ident *id = get_entity_ld_ident(attr->symconst);
-
-               if(attr->attr.data.am_sc_sign)
-                       be_emit_char('-');
-               be_emit_ident(id);
-       }
-       if(attr->symconst == NULL || attr->offset != 0) {
-               if(attr->symconst != NULL) {
-                       be_emit_irprintf("%+d", attr->offset);
-               } else {
-                       be_emit_irprintf("0x%X", attr->offset);
-               }
+               ia32_emitf(node, "\tjmp %L\n");
        }
 }
 
+/**
+ * Emit an inline assembler operand.
+ *
+ * @param node  the ia32_ASM node
+ * @param s     points to the operand (a %c)
+ *
+ * @return  pointer to the first char in s NOT in the current operand
+ */
 static const char* emit_asm_operand(const ir_node *node, const char *s)
 {
+       const ia32_attr_t     *ia32_attr = get_ia32_attr_const(node);
+       const ia32_asm_attr_t *attr      = CONST_CAST_IA32_ATTR(ia32_asm_attr_t,
+                                                            ia32_attr);
        const arch_register_t *reg;
+       const ia32_asm_reg_t  *asm_regs = attr->register_map;
+       const ia32_asm_reg_t  *asm_reg;
        const char            *reg_name;
        char                   c;
        char                   modifier = 0;
        int                    num      = -1;
-       const ia32_attr_t     *attr;
-       int                    n_outs;
        int                    p;
 
        assert(*s == '%');
@@ -1137,7 +1414,7 @@ static const char* emit_asm_operand(const ir_node *node, const char *s)
        /* parse modifiers */
        switch(c) {
        case 0:
-               ir_fprintf(stderr, "Warning: asm text (%+F) ends with %\n", node);
+               ir_fprintf(stderr, "Warning: asm text (%+F) ends with %%\n", node);
                be_emit_char('%');
                return s + 1;
        case '%':
@@ -1161,15 +1438,16 @@ static const char* emit_asm_operand(const ir_node *node, const char *s)
        case '9':
                break;
        default:
-               ir_fprintf(stderr, "Warning: asm text (%+F) contains unknown modifier "
-                          "'%c' for asm op\n", node, c);
+               ir_fprintf(stderr,
+                               "Warning: asm text (%+F) contains unknown modifier '%c' for asm op\n",
+                               node, c);
                ++s;
                break;
        }
 
        /* parse number */
        sscanf(s, "%d%n", &num, &p);
-       if(num < 0) {
+       if (num < 0) {
                ir_fprintf(stderr, "Warning: Couldn't parse assembler operand (%+F)\n",
                           node);
                return s;
@@ -1177,52 +1455,63 @@ static const char* emit_asm_operand(const ir_node *node, const char *s)
                s += p;
        }
 
+       if (num < 0 || ARR_LEN(asm_regs) <= num) {
+               ir_fprintf(stderr,
+                               "Error: Custom assembler references invalid input/output (%+F)\n",
+                               node);
+               return s;
+       }
+       asm_reg = & asm_regs[num];
+       assert(asm_reg->valid);
+
        /* get register */
-       attr   = get_ia32_attr_const(node);
-       n_outs = ARR_LEN(attr->slots);
-       if(num < n_outs) {
-               reg = get_out_reg(node, num);
+       if (asm_reg->use_input == 0) {
+               reg = get_out_reg(node, asm_reg->inout_pos);
        } else {
-               ir_node *pred;
-               int      in = num - n_outs;
-               if(in >= get_irn_arity(node)) {
-                       ir_fprintf(stderr, "Warning: Invalid input %d specified in asm "
-                                  "op (%+F)\n", num, node);
-                       return s;
-               }
-               pred = get_irn_n(node, in);
+               ir_node *pred = get_irn_n(node, asm_reg->inout_pos);
+
                /* might be an immediate value */
-               if(is_ia32_Immediate(pred)) {
+               if (is_ia32_Immediate(pred)) {
                        emit_ia32_Immediate(pred);
                        return s;
                }
-               reg = get_in_reg(node, in);
+               reg = get_in_reg(node, asm_reg->inout_pos);
        }
-       if(reg == NULL) {
-               ir_fprintf(stderr, "Warning: no register assigned for %d asm op "
-                          "(%+F)\n", num, node);
+       if (reg == NULL) {
+               ir_fprintf(stderr,
+                               "Warning: no register assigned for %d asm op (%+F)\n",
+                               num, node);
                return s;
        }
 
+       if (asm_reg->memory) {
+               be_emit_char('(');
+       }
+
        /* emit it */
-       be_emit_char('%');
-       switch(modifier) {
-       case 0:
-               reg_name = arch_register_get_name(reg);
-               break;
-       case 'b':
-               reg_name = ia32_get_mapped_reg_name(isa->regs_8bit, reg);
-               break;
-       case 'h':
-               reg_name = ia32_get_mapped_reg_name(isa->regs_8bit_high, reg);
-               break;
-       case 'w':
-               reg_name = ia32_get_mapped_reg_name(isa->regs_16bit, reg);
-               break;
-       default:
-               panic("Invalid asm op modifier");
+       if (modifier != 0) {
+               be_emit_char('%');
+               switch(modifier) {
+               case 'b':
+                       reg_name = ia32_get_mapped_reg_name(isa->regs_8bit, reg);
+                       break;
+               case 'h':
+                       reg_name = ia32_get_mapped_reg_name(isa->regs_8bit_high, reg);
+                       break;
+               case 'w':
+                       reg_name = ia32_get_mapped_reg_name(isa->regs_16bit, reg);
+                       break;
+               default:
+                       panic("Invalid asm op modifier");
+               }
+               be_emit_string(reg_name);
+       } else {
+               emit_register(reg, asm_reg->mode);
+       }
+
+       if (asm_reg->memory) {
+               be_emit_char(')');
        }
-       be_emit_string(reg_name);
 
        return s;
 }
@@ -1238,27 +1527,20 @@ static void emit_ia32_Asm(const ir_node *node)
        ident                 *asm_text = attr->asm_text;
        const char            *s        = get_id_str(asm_text);
 
-       be_emit_cstring("# Begin ASM \t");
-       be_emit_finish_line_gas(node);
+       ia32_emitf(node, "#APP\t\n");
 
        if (s[0] != '\t')
                be_emit_char('\t');
 
        while(*s != 0) {
-               if(*s == '%') {
+               if (*s == '%') {
                        s = emit_asm_operand(node, s);
-                       continue;
                } else {
-                       be_emit_char(*s);
+                       be_emit_char(*s++);
                }
-               ++s;
        }
 
-       be_emit_char('\n');
-       be_emit_write_line();
-
-       be_emit_cstring("# End ASM\n");
-       be_emit_write_line();
+       ia32_emitf(NULL, "\n#NO_APP\n");
 }
 
 /**********************************
@@ -1275,26 +1557,12 @@ static void emit_ia32_Asm(const ir_node *node)
 /**
  * Emit movsb/w instructions to make mov count divideable by 4
  */
-static void emit_CopyB_prolog(int rem) {
-       be_emit_cstring("\tcld");
-       be_emit_finish_line_gas(NULL);
-
-       switch(rem) {
-       case 1:
-               be_emit_cstring("\tmovsb");
-               be_emit_finish_line_gas(NULL);
-               break;
-       case 2:
-               be_emit_cstring("\tmovsw");
-               be_emit_finish_line_gas(NULL);
-               break;
-       case 3:
-               be_emit_cstring("\tmovsb");
-               be_emit_finish_line_gas(NULL);
-               be_emit_cstring("\tmovsw");
-               be_emit_finish_line_gas(NULL);
-               break;
-       }
+static void emit_CopyB_prolog(unsigned size)
+{
+       if (size & 1)
+               ia32_emitf(NULL, "\tmovsb\n");
+       if (size & 2)
+               ia32_emitf(NULL, "\tmovsw\n");
 }
 
 /**
@@ -1302,12 +1570,10 @@ static void emit_CopyB_prolog(int rem) {
  */
 static void emit_ia32_CopyB(const ir_node *node)
 {
-       int rem = get_ia32_pncode(node);
-
-       emit_CopyB_prolog(rem);
+       unsigned size = get_ia32_copyb_size(node);
 
-       be_emit_cstring("\trep movsd");
-       be_emit_finish_line_gas(node);
+       emit_CopyB_prolog(size);
+       ia32_emitf(node, "\trep movsd\n");
 }
 
 /**
@@ -1315,14 +1581,13 @@ static void emit_ia32_CopyB(const ir_node *node)
  */
 static void emit_ia32_CopyB_i(const ir_node *node)
 {
-       int size = get_ia32_pncode(node);
+       unsigned size = get_ia32_copyb_size(node);
 
-       emit_CopyB_prolog(size & 0x3);
+       emit_CopyB_prolog(size);
 
        size >>= 2;
        while (size--) {
-               be_emit_cstring("\tmovsd");
-               be_emit_finish_line_gas(NULL);
+               ia32_emitf(NULL, "\tmovsd\n");
        }
 }
 
@@ -1341,63 +1606,29 @@ static void emit_ia32_CopyB_i(const ir_node *node)
 /**
  * Emit code for conversions (I, FP), (FP, I) and (FP, FP).
  */
-static void emit_ia32_Conv_with_FP(const ir_node *node)
+static void emit_ia32_Conv_with_FP(const ir_node *node, const char* conv_f,
+               const char* conv_d)
 {
        ir_mode            *ls_mode = get_ia32_ls_mode(node);
        int                 ls_bits = get_mode_size_bits(ls_mode);
+       const char         *conv    = ls_bits == 32 ? conv_f : conv_d;
 
-       be_emit_cstring("\tcvt");
-
-       if(is_ia32_Conv_I2FP(node)) {
-               if(ls_bits == 32) {
-                       be_emit_cstring("si2ss");
-               } else {
-                       be_emit_cstring("si2sd");
-               }
-       } else if(is_ia32_Conv_FP2I(node)) {
-               if(ls_bits == 32) {
-                       be_emit_cstring("ss2si");
-               } else {
-                       be_emit_cstring("sd2si");
-               }
-       } else {
-               assert(is_ia32_Conv_FP2FP(node));
-               if(ls_bits == 32) {
-                       be_emit_cstring("sd2ss");
-               } else {
-                       be_emit_cstring("ss2sd");
-               }
-       }
-       be_emit_char(' ');
-
-       switch(get_ia32_op_type(node)) {
-               case ia32_Normal:
-                       ia32_emit_source_register(node, n_ia32_unary_op);
-                       break;
-               case ia32_AddrModeS:
-                       ia32_emit_am(node);
-                       break;
-               default:
-                       assert(0 && "unsupported op type for Conv");
-       }
-       be_emit_cstring(", ");
-       ia32_emit_dest_register(node, 0);
-       be_emit_finish_line_gas(node);
+       ia32_emitf(node, "\tcvt%s %AS3, %D0\n", conv);
 }
 
 static void emit_ia32_Conv_I2FP(const ir_node *node)
 {
-       emit_ia32_Conv_with_FP(node);
+       emit_ia32_Conv_with_FP(node, "si2ss", "si2sd");
 }
 
 static void emit_ia32_Conv_FP2I(const ir_node *node)
 {
-       emit_ia32_Conv_with_FP(node);
+       emit_ia32_Conv_with_FP(node, "ss2si", "sd2si");
 }
 
 static void emit_ia32_Conv_FP2FP(const ir_node *node)
 {
-       emit_ia32_Conv_with_FP(node);
+       emit_ia32_Conv_with_FP(node, "sd2ss", "ss2sd");
 }
 
 /**
@@ -1405,70 +1636,24 @@ static void emit_ia32_Conv_FP2FP(const ir_node *node)
  */
 static void emit_ia32_Conv_I2I(const ir_node *node)
 {
-       const char            *sign_suffix;
-       ir_mode               *smaller_mode = get_ia32_ls_mode(node);
-       int                    smaller_bits = get_mode_size_bits(smaller_mode);
-       int                    signed_mode;
-       const arch_register_t *in_reg, *out_reg;
+       ir_mode *smaller_mode = get_ia32_ls_mode(node);
+       int      signed_mode  = mode_is_signed(smaller_mode);
+       const char *sign_suffix;
 
        assert(!mode_is_float(smaller_mode));
-       assert(smaller_bits == 8 || smaller_bits == 16 || smaller_bits == 32);
-
-       signed_mode = mode_is_signed(smaller_mode);
-       if(smaller_bits == 32) {
-               // this should not happen as it's no convert
-               assert(0);
-               sign_suffix = "";
-       } else {
-               sign_suffix = signed_mode ? "s" : "z";
-       }
-
-       out_reg = get_out_reg(node, 0);
 
-       switch(get_ia32_op_type(node)) {
-               case ia32_Normal:
-                       in_reg  = get_in_reg(node, n_ia32_unary_op);
-
-                       if (in_reg  == &ia32_gp_regs[REG_EAX] &&
-                               out_reg == &ia32_gp_regs[REG_EAX] &&
-                               signed_mode &&
-                               smaller_bits == 16)
-                       {
-                               /* argument and result are both in EAX and */
-                               /* signedness is ok: -> use the smaller cwtl opcode */
-                               be_emit_cstring("\tcwtl");
-                       } else {
-                               be_emit_cstring("\tmov");
-                               be_emit_string(sign_suffix);
-                               ia32_emit_mode_suffix_mode(smaller_mode);
-                               be_emit_cstring("l ");
-                               emit_register(in_reg, smaller_mode);
-                               be_emit_cstring(", ");
-                               emit_register(out_reg, NULL);
-                       }
-                       break;
-               case ia32_AddrModeS: {
-                       be_emit_cstring("\tmov");
-                       be_emit_string(sign_suffix);
-                       ia32_emit_mode_suffix_mode(smaller_mode);
-                       be_emit_cstring("l ");
-                       ia32_emit_am(node);
-                       be_emit_cstring(", ");
-                       emit_register(out_reg, NULL);
-                       break;
-               }
-               default:
-                       assert(0 && "unsupported op type for Conv");
-       }
-       be_emit_finish_line_gas(node);
+       sign_suffix = signed_mode ? "s" : "z";
+       ia32_emitf(node, "\tmov%s%Ml %#AS3, %D0\n", sign_suffix);
 }
 
 /**
- * Emits code for an 8Bit Int conversion.
+ * Emits a call
  */
-static void emit_ia32_Conv_I2I8Bit(const ir_node *node)
+static void emit_ia32_Call(const ir_node *node)
 {
-       emit_ia32_Conv_I2I(node);
+       /* Special case: Call must not have its immediates prefixed by $, instead
+        * address mode is prefixed by *. */
+       ia32_emitf(node, "\tcall %*AS3\n");
 }
 
 
@@ -1482,46 +1667,31 @@ static void emit_ia32_Conv_I2I8Bit(const ir_node *node)
  *
  *******************************************/
 
-/**
- * Emits a backend call
- */
-static void emit_be_Call(const ir_node *node)
-{
-       ir_entity *ent = be_Call_get_entity(node);
-
-       be_emit_cstring("\tcall ");
-       if (ent) {
-               set_entity_backend_marked(ent, 1);
-               be_emit_string(get_entity_ld_name(ent));
-       } else {
-               const arch_register_t *reg = get_in_reg(node, be_pos_Call_ptr);
-               be_emit_char('*');
-               emit_register(reg, NULL);
-       }
-       be_emit_finish_line_gas(node);
-}
-
 /**
  * Emits code to increase stack pointer.
  */
 static void emit_be_IncSP(const ir_node *node)
 {
-       int                    offs = be_get_IncSP_offset(node);
-       const arch_register_t *reg  = arch_get_irn_register(arch_env, node);
+       int offs = be_get_IncSP_offset(node);
 
        if (offs == 0)
                return;
 
        if (offs > 0) {
-               be_emit_cstring("\tsubl $");
-               be_emit_irprintf("%u, ", offs);
-               emit_register(reg, NULL);
+               ia32_emitf(node, "\tsubl $%u, %D0\n", offs);
        } else {
-               be_emit_cstring("\taddl $");
-               be_emit_irprintf("%u, ", -offs);
-               emit_register(reg, NULL);
+               ia32_emitf(node, "\taddl $%u, %D0\n", -offs);
        }
-       be_emit_finish_line_gas(node);
+}
+
+static inline bool is_unknown_reg(const arch_register_t *reg)
+{
+       if(reg == &ia32_gp_regs[REG_GP_UKNWN]
+                       || reg == &ia32_xmm_regs[REG_XMM_UKNWN]
+                       || reg == &ia32_vfp_regs[REG_VFP_UKNWN])
+               return true;
+
+       return false;
 }
 
 /**
@@ -1529,32 +1699,23 @@ static void emit_be_IncSP(const ir_node *node)
  */
 static void Copy_emitter(const ir_node *node, const ir_node *op)
 {
-       const arch_register_t *in  = arch_get_irn_register(arch_env, op);
-       const arch_register_t *out = arch_get_irn_register(arch_env, node);
-       ir_mode               *mode;
+       const arch_register_t *in  = arch_get_irn_register(op);
+       const arch_register_t *out = arch_get_irn_register(node);
 
-       if(in == out) {
+       if (in == out) {
                return;
        }
-       if(is_unknown_reg(in))
+       if (is_unknown_reg(in))
                return;
        /* copies of vf nodes aren't real... */
-       if(arch_register_get_class(in) == &ia32_reg_classes[CLASS_ia32_vfp])
+       if (arch_register_get_class(in) == &ia32_reg_classes[CLASS_ia32_vfp])
                return;
 
-       mode = get_irn_mode(node);
-       if (mode == mode_E) {
-               be_emit_cstring("\tmovsd ");
-               emit_register(in, NULL);
-               be_emit_cstring(", ");
-               emit_register(out, NULL);
+       if (get_irn_mode(node) == mode_E) {
+               ia32_emitf(node, "\tmovsd %R, %R\n", in, out);
        } else {
-               be_emit_cstring("\tmovl ");
-               emit_register(in, NULL);
-               be_emit_cstring(", ");
-               emit_register(out, NULL);
+               ia32_emitf(node, "\tmovl %R, %R\n", in, out);
        }
-       be_emit_finish_line_gas(node);
 }
 
 static void emit_be_Copy(const ir_node *node)
@@ -1575,8 +1736,8 @@ static void emit_be_Perm(const ir_node *node)
        const arch_register_t *in0, *in1;
        const arch_register_class_t *cls0, *cls1;
 
-       in0 = arch_get_irn_register(arch_env, get_irn_n(node, 0));
-       in1 = arch_get_irn_register(arch_env, get_irn_n(node, 1));
+       in0 = arch_get_irn_register(get_irn_n(node, 0));
+       in1 = arch_get_irn_register(get_irn_n(node, 1));
 
        cls0 = arch_register_get_class(in0);
        cls1 = arch_register_get_class(in1);
@@ -1584,35 +1745,17 @@ static void emit_be_Perm(const ir_node *node)
        assert(cls0 == cls1 && "Register class mismatch at Perm");
 
        if (cls0 == &ia32_reg_classes[CLASS_ia32_gp]) {
-               be_emit_cstring("\txchg ");
-               emit_register(in1, NULL);
-               be_emit_cstring(", ");
-               emit_register(in0, NULL);
-               be_emit_finish_line_gas(node);
+               ia32_emitf(node, "\txchg %R, %R\n", in1, in0);
        } else if (cls0 == &ia32_reg_classes[CLASS_ia32_xmm]) {
-               be_emit_cstring("\txorpd ");
-               emit_register(in1, NULL);
-               be_emit_cstring(", ");
-               emit_register(in0, NULL);
-               be_emit_finish_line_gas(NULL);
-
-               be_emit_cstring("\txorpd ");
-               emit_register(in0, NULL);
-               be_emit_cstring(", ");
-               emit_register(in1, NULL);
-               be_emit_finish_line_gas(NULL);
-
-               be_emit_cstring("\txorpd ");
-               emit_register(in1, NULL);
-               be_emit_cstring(", ");
-               emit_register(in0, NULL);
-               be_emit_finish_line_gas(node);
+               ia32_emitf(NULL, "\txorpd %R, %R\n", in1, in0);
+               ia32_emitf(NULL, "\txorpd %R, %R\n", in0, in1);
+               ia32_emitf(node, "\txorpd %R, %R\n", in1, in0);
        } else if (cls0 == &ia32_reg_classes[CLASS_ia32_vfp]) {
                /* is a NOP */
        } else if (cls0 == &ia32_reg_classes[CLASS_ia32_st]) {
                /* is a NOP */
        } else {
-               panic("unexpected register class in be_Perm (%+F)\n", node);
+               panic("unexpected register class in be_Perm (%+F)", node);
        }
 }
 
@@ -1621,12 +1764,7 @@ static void emit_be_Perm(const ir_node *node)
  */
 static void emit_ia32_Const(const ir_node *node)
 {
-       be_emit_cstring("\tmovl ");
-       emit_ia32_Immediate(node);
-       be_emit_cstring(", ");
-       ia32_emit_dest_register(node, 0);
-
-       be_emit_finish_line_gas(node);
+       ia32_emitf(node, "\tmovl %I, %D0\n");
 }
 
 /**
@@ -1634,65 +1772,43 @@ static void emit_ia32_Const(const ir_node *node)
  */
 static void emit_ia32_LdTls(const ir_node *node)
 {
-       be_emit_cstring("\tmovl %gs:0, ");
-       ia32_emit_dest_register(node, 0);
-       be_emit_finish_line_gas(node);
+       ia32_emitf(node, "\tmovl %%gs:0, %D0\n");
 }
 
 /* helper function for emit_ia32_Minus64Bit */
 static void emit_mov(const ir_node* node, const arch_register_t *src, const arch_register_t *dst)
 {
-       be_emit_cstring("\tmovl ");
-       emit_register(src, NULL);
-       be_emit_cstring(", ");
-       emit_register(dst, NULL);
-       be_emit_finish_line_gas(node);
+       ia32_emitf(node, "\tmovl %R, %R\n", src, dst);
 }
 
 /* helper function for emit_ia32_Minus64Bit */
 static void emit_neg(const ir_node* node, const arch_register_t *reg)
 {
-       be_emit_cstring("\tnegl ");
-       emit_register(reg, NULL);
-       be_emit_finish_line_gas(node);
+       ia32_emitf(node, "\tnegl %R\n", reg);
 }
 
 /* helper function for emit_ia32_Minus64Bit */
 static void emit_sbb0(const ir_node* node, const arch_register_t *reg)
 {
-       be_emit_cstring("\tsbbl $0, ");
-       emit_register(reg, NULL);
-       be_emit_finish_line_gas(node);
+       ia32_emitf(node, "\tsbbl $0, %R\n", reg);
 }
 
 /* helper function for emit_ia32_Minus64Bit */
 static void emit_sbb(const ir_node* node, const arch_register_t *src, const arch_register_t *dst)
 {
-       be_emit_cstring("\tsbbl ");
-       emit_register(src, NULL);
-       be_emit_cstring(", ");
-       emit_register(dst, NULL);
-       be_emit_finish_line_gas(node);
+       ia32_emitf(node, "\tsbbl %R, %R\n", src, dst);
 }
 
 /* helper function for emit_ia32_Minus64Bit */
 static void emit_xchg(const ir_node* node, const arch_register_t *src, const arch_register_t *dst)
 {
-       be_emit_cstring("\txchgl ");
-       emit_register(src, NULL);
-       be_emit_cstring(", ");
-       emit_register(dst, NULL);
-       be_emit_finish_line_gas(node);
+       ia32_emitf(node, "\txchgl %R, %R\n", src, dst);
 }
 
 /* helper function for emit_ia32_Minus64Bit */
 static void emit_zero(const ir_node* node, const arch_register_t *reg)
 {
-       be_emit_cstring("\txorl ");
-       emit_register(reg, NULL);
-       be_emit_cstring(", ");
-       emit_register(reg, NULL);
-       be_emit_finish_line_gas(node);
+       ia32_emitf(node, "\txorl %R, %R\n", reg, reg);
 }
 
 static void emit_ia32_Minus64Bit(const ir_node *node)
@@ -1749,10 +1865,34 @@ zero_neg:
        emit_sbb( node, in_hi, out_hi);
 }
 
+static void emit_ia32_GetEIP(const ir_node *node)
+{
+       ia32_emitf(node, "\tcall %s\n", pic_base_label);
+       ia32_emitf(NULL, "%s:\n", pic_base_label);
+       ia32_emitf(node, "\tpopl %D0\n");
+}
+
+static void emit_ia32_ClimbFrame(const ir_node *node)
+{
+       const ia32_climbframe_attr_t *attr = get_ia32_climbframe_attr_const(node);
+
+       ia32_emitf(node, "\tmovl %S0, %D0\n");
+       ia32_emitf(node, "\tmovl $%u, %S1\n", attr->count);
+       ia32_emitf(NULL, BLOCK_PREFIX "%ld:\n", get_irn_node_nr(node));
+       ia32_emitf(node, "\tmovl (%D0), %D0\n");
+       ia32_emitf(node, "\tdec %S1\n");
+       ia32_emitf(node, "\tjnz " BLOCK_PREFIX "%ld\n", get_irn_node_nr(node));
+}
+
 static void emit_be_Return(const ir_node *node)
 {
-       be_emit_cstring("\tret");
-       be_emit_finish_line_gas(node);
+       unsigned pop = be_Return_get_pop(node);
+
+       if (pop > 0 || be_Return_get_emit_pop(node)) {
+               ia32_emitf(node, "\tret $%u\n", pop);
+       } else {
+               ia32_emitf(node, "\tret\n");
+       }
 }
 
 static void emit_Nothing(const ir_node *node)
@@ -1775,9 +1915,8 @@ static void emit_Nothing(const ir_node *node)
  * Enters the emitter functions for handled nodes into the generic
  * pointer of an opcode.
  */
-static
-void ia32_register_emitters(void) {
-
+static void ia32_register_emitters(void)
+{
 #define IA32_EMIT2(a,b) op_ia32_##a->ops.generic = (op_func)emit_ia32_##b
 #define IA32_EMIT(a)    IA32_EMIT2(a,a)
 #define EMIT(a)         op_##a->ops.generic = (op_func)emit_##a
@@ -1792,38 +1931,40 @@ void ia32_register_emitters(void) {
        ia32_register_spec_emitters();
 
        /* other ia32 emitter functions */
+       IA32_EMIT2(Conv_I2I8Bit, Conv_I2I);
        IA32_EMIT(Asm);
-       IA32_EMIT(CMov);
-       IA32_EMIT(SwitchJmp);
-       IA32_EMIT(CopyB);
-       IA32_EMIT(CopyB_i);
-       IA32_EMIT(Conv_I2FP);
-       IA32_EMIT(Conv_FP2I);
+       IA32_EMIT(CMovcc);
+       IA32_EMIT(Call);
+       IA32_EMIT(Const);
        IA32_EMIT(Conv_FP2FP);
+       IA32_EMIT(Conv_FP2I);
+       IA32_EMIT(Conv_I2FP);
        IA32_EMIT(Conv_I2I);
-       IA32_EMIT(Conv_I2I8Bit);
-       IA32_EMIT(Const);
+       IA32_EMIT(CopyB);
+       IA32_EMIT(CopyB_i);
+       IA32_EMIT(GetEIP);
+       IA32_EMIT(IMul);
+       IA32_EMIT(Jcc);
+       IA32_EMIT(Setcc);
        IA32_EMIT(LdTls);
        IA32_EMIT(Minus64Bit);
-       IA32_EMIT(Jcc);
+       IA32_EMIT(SwitchJmp);
+       IA32_EMIT(ClimbFrame);
+       IA32_EMIT(Jmp);
 
        /* benode emitter */
-       BE_EMIT(Call);
-       BE_EMIT(IncSP);
        BE_EMIT(Copy);
        BE_EMIT(CopyKeep);
+       BE_EMIT(IncSP);
        BE_EMIT(Perm);
        BE_EMIT(Return);
 
-       BE_IGN(RegParams);
        BE_IGN(Barrier);
        BE_IGN(Keep);
+       BE_IGN(Start);
 
        /* firm emitter */
-       EMIT(Jmp);
-       IGN(Proj);
        IGN(Phi);
-       IGN(Start);
 
 #undef BE_EMIT
 #undef EMIT
@@ -1832,55 +1973,56 @@ void ia32_register_emitters(void) {
 #undef IA32_EMIT
 }
 
-static const char *last_name = NULL;
-static unsigned last_line = -1;
-static unsigned num = -1;
+typedef void (*emit_func_ptr) (const ir_node *);
 
 /**
- * Emit the debug support for node node.
+ * Assign and emit an exception label if the current instruction can fail.
  */
-static void ia32_emit_dbg(const ir_node *node)
+static void ia32_assign_exc_label(ir_node *node)
 {
-       dbg_info *db = get_irn_dbg_info(node);
-       unsigned lineno;
-       const char *fname = be_retrieve_dbg_info(db, &lineno);
-
-       if (! cg->birg->main_env->options->stabs_debug_support)
-               return;
-
-       if (fname) {
-               if (last_name != fname) {
-                       last_line = -1;
-                       be_dbg_include_begin(cg->birg->main_env->db_handle, fname);
-                       last_name = fname;
-               }
-               if (last_line != lineno) {
-                       char name[64];
-
-                       snprintf(name, sizeof(name), ".LM%u", ++num);
-                       last_line = lineno;
-                       be_dbg_line(cg->birg->main_env->db_handle, lineno, name);
-                       be_emit_string(name);
-                       be_emit_cstring(":\n");
-                       be_emit_write_line();
-               }
-       }
+       /* assign a new ID to the instruction */
+       set_ia32_exc_label_id(node, ++exc_label_id);
+       /* print it */
+       ia32_emit_exc_label(node);
+       be_emit_char(':');
+       be_emit_pad_comment();
+       be_emit_cstring("/* exception to Block ");
+       ia32_emit_cfop_target(node);
+       be_emit_cstring(" */\n");
+       be_emit_write_line();
 }
 
-typedef void (*emit_func_ptr) (const ir_node *);
-
 /**
  * Emits code for a node.
  */
-static void ia32_emit_node(const ir_node *node)
+static void ia32_emit_node(ir_node *node)
 {
        ir_op *op = get_irn_op(node);
 
        DBG((dbg, LEVEL_1, "emitting code for %+F\n", node));
 
+       if (is_ia32_irn(node)) {
+               if (get_ia32_exc_label(node)) {
+                       /* emit the exception label of this instruction */
+                       ia32_assign_exc_label(node);
+               }
+               if (mark_spill_reload) {
+                       if (is_ia32_is_spill(node)) {
+                               ia32_emitf(NULL, "\txchg %ebx, %ebx        /* spill mark */\n");
+                       }
+                       if (is_ia32_is_reload(node)) {
+                               ia32_emitf(NULL, "\txchg %edx, %edx        /* reload mark */\n");
+                       }
+                       if (is_ia32_is_remat(node)) {
+                               ia32_emitf(NULL, "\txchg %ecx, %ecx        /* remat mark */\n");
+                       }
+               }
+       }
        if (op->ops.generic) {
                emit_func_ptr func = (emit_func_ptr) op->ops.generic;
-               ia32_emit_dbg(node);
+
+               be_dbg_set_dbg_info(get_irn_dbg_info(node));
+
                (*func) (node);
        } else {
                emit_Nothing(node);
@@ -1894,147 +2036,132 @@ static void ia32_emit_node(const ir_node *node)
  */
 static void ia32_emit_alignment(unsigned align, unsigned skip)
 {
-       be_emit_cstring("\t.p2align ");
-       be_emit_irprintf("%u,,%u\n", align, skip);
-       be_emit_write_line();
-}
-
-/**
- * Emits gas alignment directives for Functions depended on cpu architecture.
- */
-static void ia32_emit_align_func(cpu_support cpu)
-{
-       unsigned align;
-       unsigned maximum_skip;
-
-       switch (cpu) {
-               case arch_i386:
-                       align = 2;
-                       break;
-               case arch_i486:
-                       align = 4;
-                       break;
-               case arch_k6:
-                       align = 5;
-                       break;
-               default:
-                       align = 4;
-       }
-       maximum_skip = (1 << align) - 1;
-       ia32_emit_alignment(align, maximum_skip);
+       ia32_emitf(NULL, "\t.p2align %u,,%u\n", align, skip);
 }
 
 /**
  * Emits gas alignment directives for Labels depended on cpu architecture.
  */
-static void ia32_emit_align_label(cpu_support cpu)
+static void ia32_emit_align_label(void)
 {
-       unsigned align; unsigned maximum_skip;
-
-       switch (cpu) {
-               case arch_i386:
-                       align = 2;
-                       break;
-               case arch_i486:
-                       align = 4;
-                       break;
-               case arch_k6:
-                       align = 5;
-                       break;
-               default:
-                       align = 4;
-       }
-       maximum_skip = (1 << align) - 1;
+       unsigned align        = ia32_cg_config.label_alignment;
+       unsigned maximum_skip = ia32_cg_config.label_alignment_max_skip;
        ia32_emit_alignment(align, maximum_skip);
 }
 
 /**
- * Test wether a block should be aligned.
- * For cpus in the P4/Athlon class it is usefull to align jump labels to
+ * Test whether a block should be aligned.
+ * For cpus in the P4/Athlon class it is useful to align jump labels to
  * 16 bytes. However we should only do that if the alignment nops before the
  * label aren't executed more often than we have jumps to the label.
  */
-static int should_align_block(ir_node *block, ir_node *prev)
+static int should_align_block(const ir_node *block)
 {
        static const double DELTA = .0001;
        ir_exec_freq *exec_freq   = cg->birg->exec_freq;
+       ir_node      *prev        = get_prev_block_sched(block);
        double        block_freq;
        double        prev_freq = 0;  /**< execfreq of the fallthrough block */
        double        jmp_freq  = 0;  /**< execfreq of all non-fallthrough blocks */
-       cpu_support   cpu       = isa->opt_arch;
        int           i, n_cfgpreds;
 
-       if(exec_freq == NULL)
+       if (exec_freq == NULL)
                return 0;
-       if(cpu == arch_i386 || cpu == arch_i486)
+       if (ia32_cg_config.label_alignment_factor <= 0)
                return 0;
 
        block_freq = get_block_execfreq(exec_freq, block);
-       if(block_freq < DELTA)
+       if (block_freq < DELTA)
                return 0;
 
        n_cfgpreds = get_Block_n_cfgpreds(block);
        for(i = 0; i < n_cfgpreds; ++i) {
-               ir_node *pred      = get_Block_cfgpred_block(block, i);
-               double   pred_freq = get_block_execfreq(exec_freq, pred);
+               const ir_node *pred      = get_Block_cfgpred_block(block, i);
+               double         pred_freq = get_block_execfreq(exec_freq, pred);
 
-               if(pred == prev) {
+               if (pred == prev) {
                        prev_freq += pred_freq;
                } else {
                        jmp_freq  += pred_freq;
                }
        }
 
-       if(prev_freq < DELTA && !(jmp_freq < DELTA))
+       if (prev_freq < DELTA && !(jmp_freq < DELTA))
                return 1;
 
        jmp_freq /= prev_freq;
 
-       switch (cpu) {
-               case arch_athlon:
-               case arch_athlon_64:
-               case arch_k6:
-                       return jmp_freq > 3;
-               default:
-                       return jmp_freq > 2;
-       }
+       return jmp_freq > ia32_cg_config.label_alignment_factor;
 }
 
-static void ia32_emit_block_header(ir_node *block, ir_node *prev)
+/**
+ * Emit the block header for a block.
+ *
+ * @param block       the block
+ * @param prev_block  the previous block
+ */
+static void ia32_emit_block_header(ir_node *block)
 {
-       int           n_cfgpreds;
-       int           need_label;
+       ir_graph     *irg = current_ir_graph;
+       int           need_label = block_needs_label(block);
        int           i, arity;
-       ir_exec_freq  *exec_freq = cg->birg->exec_freq;
+       ir_exec_freq *exec_freq = cg->birg->exec_freq;
 
-       n_cfgpreds = get_Block_n_cfgpreds(block);
-       need_label = (n_cfgpreds != 0);
+       if (block == get_irg_end_block(irg))
+               return;
+
+       if (ia32_cg_config.label_alignment > 0) {
+               /* align the current block if:
+                * a) if should be aligned due to its execution frequency
+                * b) there is no fall-through here
+                */
+               if (should_align_block(block)) {
+                       ia32_emit_align_label();
+               } else {
+                       /* if the predecessor block has no fall-through,
+                          we can always align the label. */
+                       int i;
+                       int has_fallthrough = 0;
+
+                       for (i = get_Block_n_cfgpreds(block) - 1; i >= 0; --i) {
+                               ir_node *cfg_pred = get_Block_cfgpred(block, i);
+                               if (can_be_fallthrough(cfg_pred)) {
+                                       has_fallthrough = 1;
+                                       break;
+                               }
+                       }
 
-       if (should_align_block(block, prev)) {
-               assert(need_label);
-               ia32_emit_align_label(isa->opt_arch);
+                       if (!has_fallthrough)
+                               ia32_emit_align_label();
+               }
        }
 
-       if(need_label) {
+       if (need_label) {
                ia32_emit_block_name(block);
                be_emit_char(':');
 
                be_emit_pad_comment();
-               be_emit_cstring("   /* preds:");
+               be_emit_cstring("   /* ");
+       } else {
+               be_emit_cstring("\t/* ");
+               ia32_emit_block_name(block);
+               be_emit_cstring(": ");
+       }
 
-               /* emit list of pred blocks in comment */
-               arity = get_irn_arity(block);
+       be_emit_cstring("preds:");
+
+       /* emit list of pred blocks in comment */
+       arity = get_irn_arity(block);
+       if (arity <= 0) {
+               be_emit_cstring(" none");
+       } else {
                for (i = 0; i < arity; ++i) {
                        ir_node *predblock = get_Block_cfgpred_block(block, i);
                        be_emit_irprintf(" %d", get_irn_node_nr(predblock));
                }
-       } else {
-               be_emit_cstring("\t/* ");
-               ia32_emit_block_name(block);
-               be_emit_cstring(": ");
        }
        if (exec_freq != NULL) {
-               be_emit_irprintf(" freq: %f",
+               be_emit_irprintf(", freq: %f",
                                 get_block_execfreq(exec_freq, block));
        }
        be_emit_cstring(" */\n");
@@ -2045,84 +2172,62 @@ static void ia32_emit_block_header(ir_node *block, ir_node *prev)
  * Walks over the nodes in a block connected by scheduling edges
  * and emits code for each node.
  */
-static void ia32_gen_block(ir_node *block, ir_node *last_block)
+static void ia32_gen_block(ir_node *block)
 {
-       const ir_node *node;
+       ir_node *node;
 
-       ia32_emit_block_header(block, last_block);
+       ia32_emit_block_header(block);
 
        /* emit the contents of the block */
-       ia32_emit_dbg(block);
+       be_dbg_set_dbg_info(get_irn_dbg_info(block));
        sched_foreach(block, node) {
                ia32_emit_node(node);
        }
 }
 
-/**
- * Emits code for function start.
- */
-static void ia32_emit_func_prolog(ir_graph *irg)
-{
-       ir_entity  *irg_ent  = get_irg_entity(irg);
-       const char *irg_name = get_entity_ld_name(irg_ent);
-       cpu_support cpu      = isa->opt_arch;
-       const be_irg_t *birg = cg->birg;
-
-       be_emit_write_line();
-       be_gas_emit_switch_section(GAS_SECTION_TEXT);
-       be_dbg_method_begin(birg->main_env->db_handle, irg_ent, be_abi_get_stack_layout(birg->abi));
-       ia32_emit_align_func(cpu);
-       if (get_entity_visibility(irg_ent) == visibility_external_visible) {
-               be_emit_cstring(".global ");
-               be_emit_string(irg_name);
-               be_emit_char('\n');
-               be_emit_write_line();
-       }
-       ia32_emit_function_object(irg_name);
-       be_emit_string(irg_name);
-       be_emit_cstring(":\n");
-       be_emit_write_line();
-}
-
-/**
- * Emits code for function end
- */
-static void ia32_emit_func_epilog(ir_graph *irg)
-{
-       const char     *irg_name = get_entity_ld_name(get_irg_entity(irg));
-       const be_irg_t *birg     = cg->birg;
-
-       ia32_emit_function_size(irg_name);
-       be_dbg_method_end(birg->main_env->db_handle);
-       be_emit_char('\n');
-       be_emit_write_line();
-}
+typedef struct exc_entry {
+       ir_node *exc_instr;  /** The instruction that can issue an exception. */
+       ir_node *block;      /** The block to call then. */
+} exc_entry;
 
 /**
  * Block-walker:
- * Sets labels for control flow nodes (jump target)
+ * Sets labels for control flow nodes (jump target).
+ * Links control predecessors to there destination blocks.
  */
 static void ia32_gen_labels(ir_node *block, void *data)
 {
+       exc_entry **exc_list = data;
        ir_node *pred;
-       int n = get_Block_n_cfgpreds(block);
-       (void) data;
+       int     n;
 
-       for (n--; n >= 0; n--) {
+       for (n = get_Block_n_cfgpreds(block) - 1; n >= 0; --n) {
                pred = get_Block_cfgpred(block, n);
                set_irn_link(pred, block);
+
+               pred = skip_Proj(pred);
+               if (is_ia32_irn(pred) && get_ia32_exc_label(pred)) {
+                       exc_entry e;
+
+                       e.exc_instr = pred;
+                       e.block     = block;
+                       ARR_APP1(exc_entry, *exc_list, e);
+                       set_irn_link(pred, block);
+               }
        }
 }
 
 /**
- * Emit an exception label if the current instruction can fail.
+ * Compare two exception_entries.
  */
-void ia32_emit_exc_label(const ir_node *node)
+static int cmp_exc_entry(const void *a, const void *b)
 {
-       if (get_ia32_exc_label(node)) {
-               be_emit_irprintf(".EXL%u\n", 0);
-               be_emit_write_line();
-       }
+       const exc_entry *ea = a;
+       const exc_entry *eb = b;
+
+       if (get_ia32_exc_label_id(ea->exc_instr) < get_ia32_exc_label_id(eb->exc_instr))
+               return -1;
+       return +1;
 }
 
 /**
@@ -2130,37 +2235,2126 @@ void ia32_emit_exc_label(const ir_node *node)
  */
 void ia32_gen_routine(ia32_code_gen_t *ia32_cg, ir_graph *irg)
 {
-       ir_node *block;
-       ir_node *last_block = NULL;
+       ir_entity *entity     = get_irg_entity(irg);
+       exc_entry *exc_list   = NEW_ARR_F(exc_entry, 0);
        int i, n;
 
        cg       = ia32_cg;
-       isa      = (const ia32_isa_t*) cg->arch_env->isa;
-       arch_env = cg->arch_env;
+       isa      = cg->isa;
+       do_pic   = cg->birg->main_env->options->pic;
+
+       be_gas_elf_type_char = '@';
 
        ia32_register_emitters();
 
-       ia32_emit_func_prolog(irg);
-       irg_block_walk_graph(irg, ia32_gen_labels, NULL, NULL);
+       get_unique_label(pic_base_label, sizeof(pic_base_label), ".PIC_BASE");
 
+       be_dbg_method_begin(entity, be_abi_get_stack_layout(cg->birg->abi));
+       be_gas_emit_function_prolog(entity, ia32_cg_config.function_alignment);
+
+       /* we use links to point to target blocks */
+       ir_reserve_resources(irg, IR_RESOURCE_IRN_LINK);
+       irg_block_walk_graph(irg, ia32_gen_labels, NULL, &exc_list);
+
+       /* initialize next block links */
        n = ARR_LEN(cg->blk_sched);
-       for (i = 0; i < n;) {
-               ir_node *next_bl;
+       for (i = 0; i < n; ++i) {
+               ir_node *block = cg->blk_sched[i];
+               ir_node *prev  = i > 0 ? cg->blk_sched[i-1] : NULL;
+
+               set_irn_link(block, prev);
+       }
 
-               block   = cg->blk_sched[i];
-               ++i;
-               next_bl = i < n ? cg->blk_sched[i] : NULL;
+       for (i = 0; i < n; ++i) {
+               ir_node *block = cg->blk_sched[i];
 
-               /* set here the link. the emitter expects to find the next block here */
-               set_irn_link(block, next_bl);
-               ia32_gen_block(block, last_block);
-               last_block = block;
+               ia32_gen_block(block);
        }
 
-       ia32_emit_func_epilog(irg);
+       be_gas_emit_function_epilog(entity);
+       be_dbg_method_end();
+       be_emit_char('\n');
+       be_emit_write_line();
+
+       ir_free_resources(irg, IR_RESOURCE_IRN_LINK);
+
+       /* Sort the exception table using the exception label id's.
+          Those are ascending with ascending addresses. */
+       qsort(exc_list, ARR_LEN(exc_list), sizeof(exc_list[0]), cmp_exc_entry);
+       {
+               int i;
+
+               for (i = 0; i < ARR_LEN(exc_list); ++i) {
+                       be_emit_cstring("\t.long ");
+                       ia32_emit_exc_label(exc_list[i].exc_instr);
+                       be_emit_char('\n');
+                       be_emit_cstring("\t.long ");
+                       ia32_emit_block_name(exc_list[i].block);
+                       be_emit_char('\n');
+               }
+       }
+       DEL_ARR_F(exc_list);
 }
 
-void ia32_init_emitter(void)
+static const lc_opt_table_entry_t ia32_emitter_options[] = {
+       LC_OPT_ENT_BOOL("mark_spill_reload",   "mark spills and reloads with ud opcodes", &mark_spill_reload),
+       LC_OPT_LAST
+};
+
+/* ==== Experimental binary emitter ==== */
+
+static unsigned char reg_gp_map[N_ia32_gp_REGS];
+//static unsigned char reg_mmx_map[N_ia32_mmx_REGS];
+//static unsigned char reg_sse_map[N_ia32_xmm_REGS];
+static unsigned char pnc_map_signed[8];
+static unsigned char pnc_map_unsigned[8];
+
+static void build_reg_map(void)
+{
+       reg_gp_map[REG_EAX] = 0x0;
+       reg_gp_map[REG_ECX] = 0x1;
+       reg_gp_map[REG_EDX] = 0x2;
+       reg_gp_map[REG_EBX] = 0x3;
+       reg_gp_map[REG_ESP] = 0x4;
+       reg_gp_map[REG_EBP] = 0x5;
+       reg_gp_map[REG_ESI] = 0x6;
+       reg_gp_map[REG_EDI] = 0x7;
+
+       pnc_map_signed[pn_Cmp_Eq]    = 0x04;
+       pnc_map_signed[pn_Cmp_Lt]    = 0x0C;
+       pnc_map_signed[pn_Cmp_Le]    = 0x0E;
+       pnc_map_signed[pn_Cmp_Gt]    = 0x0F;
+       pnc_map_signed[pn_Cmp_Ge]    = 0x0D;
+       pnc_map_signed[pn_Cmp_Lg]    = 0x05;
+
+       pnc_map_unsigned[pn_Cmp_Eq]    = 0x04;
+       pnc_map_unsigned[pn_Cmp_Lt]    = 0x02;
+       pnc_map_unsigned[pn_Cmp_Le]    = 0x06;
+       pnc_map_unsigned[pn_Cmp_Gt]    = 0x07;
+       pnc_map_unsigned[pn_Cmp_Ge]    = 0x03;
+       pnc_map_unsigned[pn_Cmp_Lg]    = 0x05;
+}
+
+/** Returns the encoding for a pnc field. */
+static unsigned char pnc2cc(int pnc)
+{
+       unsigned char cc;
+       if (pnc == ia32_pn_Cmp_parity) {
+               cc = 0x0A;
+       } else if (pnc & ia32_pn_Cmp_float || pnc & ia32_pn_Cmp_unsigned) {
+               cc = pnc_map_unsigned[pnc & 0x07];
+       } else {
+               cc = pnc_map_signed[pnc & 0x07];
+       }
+       assert(cc != 0);
+       return cc;
+}
+
+/** Sign extension bit values for binops */
+enum SignExt {
+       UNSIGNED_IMM = 0,  /**< unsigned immediate */
+       SIGNEXT_IMM  = 2,  /**< sign extended immediate */
+};
+
+/** The mod encoding of the ModR/M */
+enum Mod {
+       MOD_IND          = 0x00, /**< [reg1] */
+       MOD_IND_BYTE_OFS = 0x40, /**< [reg1 + byte ofs] */
+       MOD_IND_WORD_OFS = 0x80, /**< [reg1 + word ofs] */
+       MOD_REG          = 0xC0  /**< reg1 */
+};
+
+/** create R/M encoding for ModR/M */
+#define ENC_RM(x) (x)
+/** create REG encoding for ModR/M */
+#define ENC_REG(x) ((x) << 3)
+
+/** create encoding for a SIB byte */
+#define ENC_SIB(scale, index, base) ((scale) << 6 | (index) << 3 | (base))
+
+/* Node: The following routines are supposed to append bytes, words, dwords
+   to the output stream.
+   Currently the implementation is stupid in that it still creates output
+   for an "assembler" in the form of .byte, .long
+   We will change this when enough infrastructure is there to create complete
+   machine code in memory/object files */
+
+static void bemit8(const unsigned char byte)
+{
+       be_emit_irprintf("\t.byte 0x%x\n", byte);
+       be_emit_write_line();
+}
+
+static void bemit16(const unsigned short u16)
+{
+       be_emit_irprintf("\t.word 0x%x\n", u16);
+       be_emit_write_line();
+}
+
+static void bemit32(const unsigned u32)
+{
+       be_emit_irprintf("\t.long 0x%x\n", u32);
+       be_emit_write_line();
+}
+
+/**
+ * Emit address of an entity. If @p is_relative is true then a relative
+ * offset from behind the address to the entity is created.
+ */
+static void bemit_entity(ir_entity *entity, bool entity_sign, int offset,
+                         bool is_relative)
+{
+       if (entity == NULL) {
+               bemit32(offset);
+               return;
+       }
+
+       /* the final version should remember the position in the bytestream
+          and patch it with the correct address at linktime... */
+       be_emit_cstring("\t.long ");
+       if (entity_sign)
+               be_emit_char('-');
+       be_gas_emit_entity(entity);
+
+       if (get_entity_owner(entity) == get_tls_type()) {
+               if (get_entity_linkage(entity) & IR_LINKAGE_EXTERN) {
+                       be_emit_cstring("@INDNTPOFF");
+               } else {
+                       be_emit_cstring("@NTPOFF");
+               }
+       }
+
+       if (is_relative) {
+               be_emit_cstring("-.");
+               offset -= 4;
+       }
+
+       if (offset != 0) {
+               be_emit_irprintf("%+d", offset);
+       }
+       be_emit_char('\n');
+       be_emit_write_line();
+}
+
+static void bemit_jmp_destination(const ir_node *dest_block)
 {
+       be_emit_cstring("\t.long ");
+       ia32_emit_block_name(dest_block);
+       be_emit_cstring(" - . - 4\n");
+       be_emit_write_line();
+}
+
+/* end emit routines, all emitters following here should only use the functions
+   above. */
+
+typedef enum reg_modifier {
+       REG_LOW  = 0,
+       REG_HIGH = 1
+} reg_modifier_t;
+
+/** Create a ModR/M byte for src1,src2 registers */
+static void bemit_modrr(const arch_register_t *src1,
+                        const arch_register_t *src2)
+{
+       unsigned char modrm = MOD_REG;
+       modrm |= ENC_RM(reg_gp_map[src1->index]);
+       modrm |= ENC_REG(reg_gp_map[src2->index]);
+       bemit8(modrm);
+}
+
+/** Create a ModR/M8 byte for src1,src2 registers */
+static void bemit_modrr8(reg_modifier_t high_part1, const arch_register_t *src1,
+                                                reg_modifier_t high_part2, const arch_register_t *src2)
+{
+       unsigned char modrm = MOD_REG;
+       modrm |= ENC_RM(reg_gp_map[src1->index] +  (high_part1 == REG_HIGH ? 4 : 0));
+       modrm |= ENC_REG(reg_gp_map[src2->index] + (high_part2 == REG_HIGH ? 4 : 0));
+       bemit8(modrm);
+}
+
+/** Create a ModR/M byte for one register and extension */
+static void bemit_modru(const arch_register_t *reg, unsigned ext)
+{
+       unsigned char modrm = MOD_REG;
+       assert(ext <= 7);
+       modrm |= ENC_RM(reg_gp_map[reg->index]);
+       modrm |= ENC_REG(ext);
+       bemit8(modrm);
+}
+
+/** Create a ModR/M8 byte for one register */
+static void bemit_modrm8(reg_modifier_t high_part, const arch_register_t *reg)
+{
+       unsigned char modrm = MOD_REG;
+       assert(reg_gp_map[reg->index] < 4);
+       modrm |= ENC_RM(reg_gp_map[reg->index] + (high_part == REG_HIGH ? 4 : 0));
+       modrm |= MOD_REG;
+       bemit8(modrm);
+}
+
+/**
+ * Calculate the size of an signed immediate in bytes.
+ *
+ * @param offset  an offset
+ */
+static unsigned get_signed_imm_size(int offset)
+{
+       if (-128 <= offset && offset < 128) {
+               return 1;
+       } else if (-32768 <= offset && offset < 32768) {
+               return 2;
+       } else {
+               return 4;
+       }
+}
+
+/**
+ * Emit an address mode.
+ *
+ * @param reg   content of the reg field: either a register index or an opcode extension
+ * @param node  the node
+ */
+static void bemit_mod_am(unsigned reg, const ir_node *node)
+{
+       ir_entity *ent       = get_ia32_am_sc(node);
+       int        offs      = get_ia32_am_offs_int(node);
+       ir_node   *base      = get_irn_n(node, n_ia32_base);
+       int        has_base  = !is_ia32_NoReg_GP(base);
+       ir_node   *index     = get_irn_n(node, n_ia32_index);
+       int        has_index = !is_ia32_NoReg_GP(index);
+       unsigned   modrm     = 0;
+       unsigned   sib       = 0;
+       unsigned   emitoffs  = 0;
+       bool       emitsib   = false;
+       unsigned   base_enc;
+
+       /* set the mod part depending on displacement */
+       if (ent != NULL) {
+               modrm |= MOD_IND_WORD_OFS;
+               emitoffs = 32;
+       } else if (offs == 0) {
+               modrm |= MOD_IND;
+               emitoffs = 0;
+       } else if (-128 <= offs && offs < 128) {
+               modrm |= MOD_IND_BYTE_OFS;
+               emitoffs = 8;
+       } else {
+               modrm |= MOD_IND_WORD_OFS;
+               emitoffs = 32;
+       }
+
+       if (has_base) {
+               const arch_register_t *base_reg = arch_get_irn_register(base);
+               base_enc = reg_gp_map[base_reg->index];
+       } else {
+               /* Use the EBP encoding + MOD_IND if NO base register. There is
+                * always a 32bit offset present in this case. */
+               modrm    = MOD_IND;
+               base_enc = 0x05;
+               emitoffs = 32;
+       }
+
+       /* Determine if we need a SIB byte. */
+       if (has_index) {
+               const arch_register_t *reg_index = arch_get_irn_register(index);
+               int                    scale     = get_ia32_am_scale(node);
+               assert(scale < 4);
+               /* R/M set to ESP means SIB in 32bit mode. */
+               modrm   |= ENC_RM(0x04);
+               sib      = ENC_SIB(scale, reg_gp_map[reg_index->index], base_enc);
+               emitsib = true;
+       } else if (base_enc == 0x04) {
+               /* for the above reason we are forced to emit a SIB when base is ESP.
+                * Only the base is used, index must be ESP too, which means no index.
+                */
+               modrm   |= ENC_RM(0x04);
+               sib      = ENC_SIB(0, 0x04, 0x04);
+               emitsib  = true;
+       } else {
+               modrm |= ENC_RM(base_enc);
+       }
+
+       /* We are forced to emit an 8bit offset as EBP base without offset is a
+        * special case for SIB without base register. */
+       if (base_enc == 0x05 && emitoffs == 0) {
+               modrm    |= MOD_IND_BYTE_OFS;
+               emitoffs  = 8;
+       }
+
+       modrm |= ENC_REG(reg);
+
+       bemit8(modrm);
+       if (emitsib)
+               bemit8(sib);
+
+       /* emit displacement */
+       if (emitoffs == 8) {
+               bemit8((unsigned) offs);
+       } else if (emitoffs == 32) {
+               bemit_entity(ent, is_ia32_am_sc_sign(node), offs, false);
+       }
+}
+
+/**
+ * Emit a binop with a immediate operand.
+ *
+ * @param node        the node to emit
+ * @param opcode_eax  the opcode for the op eax, imm variant
+ * @param opcode      the opcode for the reg, imm variant
+ * @param ruval       the opcode extension for opcode
+ */
+static void bemit_binop_with_imm(
+       const ir_node *node,
+       unsigned char opcode_ax,
+       unsigned char opcode, unsigned char ruval)
+{
+       /* Use in-reg, because some instructions (cmp, test) have no out-reg. */
+       const ir_node               *op   = get_irn_n(node, n_ia32_binary_right);
+       const ia32_immediate_attr_t *attr = get_ia32_immediate_attr_const(op);
+       unsigned                     size;
+
+       /* Some instructions (test) have no short form with 32bit value + 8bit
+        * immediate. */
+       if (attr->symconst != NULL || opcode & SIGNEXT_IMM) {
+               size = 4;
+       } else {
+               /* check for sign extension */
+               size = get_signed_imm_size(attr->offset);
+       }
+
+       switch (size) {
+       case 1:
+               bemit8(opcode | SIGNEXT_IMM);
+               /* cmp has this special mode */
+               if (get_ia32_op_type(node) == ia32_AddrModeS) {
+                       bemit_mod_am(ruval, node);
+               } else {
+                       const arch_register_t *reg = get_in_reg(node, n_ia32_binary_left);
+                       bemit_modru(reg, ruval);
+               }
+               bemit8((unsigned char)attr->offset);
+               return;
+       case 2:
+       case 4:
+               /* check for eax variant: this variant is shorter for 32bit immediates only */
+               if (get_ia32_op_type(node) == ia32_AddrModeS) {
+                       bemit8(opcode);
+                       bemit_mod_am(ruval, node);
+               } else {
+                       const arch_register_t *reg = get_in_reg(node, n_ia32_binary_left);
+                       if (reg->index == REG_EAX) {
+                               bemit8(opcode_ax);
+                       } else {
+                               bemit8(opcode);
+                               bemit_modru(reg, ruval);
+                       }
+               }
+               bemit_entity(attr->symconst, attr->sc_sign, attr->offset, false);
+               return;
+       }
+       panic("invalid imm size?!?");
+}
+
+/**
+ * Emits a binop.
+ */
+static void bemit_binop_2(const ir_node *node, unsigned code)
+{
+       const arch_register_t *out = get_in_reg(node, n_ia32_binary_left);
+       bemit8(code);
+       if (get_ia32_op_type(node) == ia32_Normal) {
+               const arch_register_t *op2 = get_in_reg(node, n_ia32_binary_right);
+               bemit_modrr(op2, out);
+       } else {
+               bemit_mod_am(reg_gp_map[out->index], node);
+       }
+}
+
+/**
+ * Emit a binop.
+ */
+static void bemit_binop(const ir_node *node, const unsigned char opcodes[4])
+{
+       ir_node *right = get_irn_n(node, n_ia32_binary_right);
+       if (is_ia32_Immediate(right)) {
+               bemit_binop_with_imm(node, opcodes[1], opcodes[2], opcodes[3]);
+       } else {
+               bemit_binop_2(node, opcodes[0]);
+       }
+}
+
+/**
+ * Emit an unop.
+ */
+static void bemit_unop(const ir_node *node, unsigned char code, unsigned char ext, int input)
+{
+       bemit8(code);
+       if (get_ia32_op_type(node) == ia32_Normal) {
+               const arch_register_t *in = get_in_reg(node, input);
+               bemit_modru(in, ext);
+       } else {
+               bemit_mod_am(ext, node);
+       }
+}
+
+static void bemit_unop_reg(const ir_node *node, unsigned char code, int input)
+{
+       const arch_register_t *out = get_out_reg(node, 0);
+       bemit_unop(node, code, reg_gp_map[out->index], input);
+}
+
+static void bemit_unop_mem(const ir_node *node, unsigned char code, unsigned char ext)
+{
+       unsigned size = get_mode_size_bits(get_ia32_ls_mode(node));
+       if (size == 16)
+               bemit8(0x66);
+       bemit8(size == 8 ? code : code + 1);
+       bemit_mod_am(ext, node);
+}
+
+static void bemit_immediate(const ir_node *node, bool relative)
+{
+       const ia32_immediate_attr_t *attr = get_ia32_immediate_attr_const(node);
+       bemit_entity(attr->symconst, attr->sc_sign, attr->offset, relative);
+}
+
+static void bemit_copy(const ir_node *copy)
+{
+       const arch_register_t *in  = get_in_reg(copy, 0);
+       const arch_register_t *out = get_out_reg(copy, 0);
+
+       if (in == out || is_unknown_reg(in))
+               return;
+       /* copies of vf nodes aren't real... */
+       if (arch_register_get_class(in) == &ia32_reg_classes[CLASS_ia32_vfp])
+               return;
+
+       if (get_irn_mode(copy) == mode_E) {
+               panic("NIY");
+       } else {
+               assert(arch_register_get_class(in) == &ia32_reg_classes[CLASS_ia32_gp]);
+               bemit8(0x8B);
+               bemit_modrr(in, out);
+       }
+}
+
+static void bemit_perm(const ir_node *node)
+{
+       const arch_register_t       *in0  = arch_get_irn_register(get_irn_n(node, 0));
+       const arch_register_t       *in1  = arch_get_irn_register(get_irn_n(node, 1));
+       const arch_register_class_t *cls0 = arch_register_get_class(in0);
+
+       assert(cls0 == arch_register_get_class(in1) && "Register class mismatch at Perm");
+
+       if (cls0 == &ia32_reg_classes[CLASS_ia32_gp]) {
+               if (in0->index == REG_EAX) {
+                       bemit8(0x90 + reg_gp_map[in1->index]);
+               } else if (in1->index == REG_EAX) {
+                       bemit8(0x90 + reg_gp_map[in0->index]);
+               } else {
+                       bemit8(0x87);
+                       bemit_modrr(in0, in1);
+               }
+       } else if (cls0 == &ia32_reg_classes[CLASS_ia32_xmm]) {
+               panic("unimplemented"); // TODO implement
+               //ia32_emitf(NULL, "\txorpd %R, %R\n", in1, in0);
+               //ia32_emitf(NULL, "\txorpd %R, %R\n", in0, in1);
+               //ia32_emitf(node, "\txorpd %R, %R\n", in1, in0);
+       } else if (cls0 == &ia32_reg_classes[CLASS_ia32_vfp]) {
+               /* is a NOP */
+       } else if (cls0 == &ia32_reg_classes[CLASS_ia32_st]) {
+               /* is a NOP */
+       } else {
+               panic("unexpected register class in be_Perm (%+F)", node);
+       }
+}
+
+static void bemit_xor0(const ir_node *node)
+{
+       const arch_register_t *out = get_out_reg(node, 0);
+       bemit8(0x31);
+       bemit_modrr(out, out);
+}
+
+static void bemit_mov_const(const ir_node *node)
+{
+       const arch_register_t *out = get_out_reg(node, 0);
+       bemit8(0xB8 + reg_gp_map[out->index]);
+       bemit_immediate(node, false);
+}
+
+/**
+ * Creates a function for a Binop with 3 possible encodings.
+ */
+#define BINOP(op, op0, op1, op2, op2_ext)                                 \
+static void bemit_ ## op(const ir_node *node) {                           \
+       static const unsigned char op ## _codes[] = {op0, op1, op2, op2_ext}; \
+       bemit_binop(node, op ## _codes);                                      \
+}
+
+/*    insn  def  eax,imm   imm */
+BINOP(add,  0x03, 0x05, 0x81, 0)
+BINOP(or,   0x0B, 0x0D, 0x81, 1)
+BINOP(adc,  0x13, 0x15, 0x81, 2)
+BINOP(sbb,  0x1B, 0x1D, 0x81, 3)
+BINOP(and,  0x23, 0x25, 0x81, 4)
+BINOP(sub,  0x2B, 0x2D, 0x81, 5)
+BINOP(xor,  0x33, 0x35, 0x81, 6)
+BINOP(test, 0x85, 0xA9, 0xF7, 0)
+
+#define BINOPMEM(op, ext) \
+static void bemit_##op(const ir_node *node) \
+{ \
+       ir_node *val; \
+       unsigned size = get_mode_size_bits(get_ia32_ls_mode(node)); \
+       if (size == 16) \
+               bemit8(0x66); \
+       val = get_irn_n(node, n_ia32_unary_op); \
+       if (is_ia32_Immediate(val)) { \
+               const ia32_immediate_attr_t *attr   = get_ia32_immediate_attr_const(val); \
+               int                          offset = attr->offset; \
+               if (attr->symconst == NULL && get_signed_imm_size(offset) == 1) { \
+                       bemit8(0x83); \
+                       bemit_mod_am(ext, node); \
+                       bemit8(offset); \
+               } else { \
+                       bemit8(0x81); \
+                       bemit_mod_am(ext, node); \
+                       if (size == 16) { \
+                               bemit16(offset); \
+                       } else { \
+                               bemit_entity(attr->symconst, attr->sc_sign, offset, false); \
+                       } \
+               } \
+       } else { \
+               bemit8(ext << 3 | 1); \
+               bemit_mod_am(reg_gp_map[get_out_reg(val, 0)->index], node); \
+       } \
+} \
+ \
+static void bemit_##op##8bit(const ir_node *node) \
+{ \
+       ir_node *val = get_irn_n(node, n_ia32_unary_op); \
+       if (is_ia32_Immediate(val)) { \
+               bemit8(0x80); \
+               bemit_mod_am(ext, node); \
+               bemit8(get_ia32_immediate_attr_const(val)->offset); \
+       } else { \
+               bemit8(ext << 3); \
+               bemit_mod_am(reg_gp_map[get_out_reg(val, 0)->index], node); \
+       } \
+}
+
+BINOPMEM(addmem,  0)
+BINOPMEM(ormem,   1)
+BINOPMEM(andmem,  4)
+BINOPMEM(submem,  5)
+BINOPMEM(xormem,  6)
+
+
+/**
+ * Creates a function for an Unop with code /ext encoding.
+ */
+#define UNOP(op, code, ext, input)              \
+static void bemit_ ## op(const ir_node *node) { \
+       bemit_unop(node, code, ext, input);         \
+}
+
+UNOP(not,     0xF7, 2, n_ia32_Not_val)
+UNOP(neg,     0xF7, 3, n_ia32_Neg_val)
+UNOP(mul,     0xF7, 4, n_ia32_Mul_right)
+UNOP(imul1op, 0xF7, 5, n_ia32_IMul1OP_right)
+UNOP(div,     0xF7, 6, n_ia32_Div_divisor)
+UNOP(idiv,    0xF7, 7, n_ia32_IDiv_divisor)
+
+/* TODO: am support for IJmp */
+UNOP(ijmp,    0xFF, 4, n_ia32_IJmp_target)
+
+#define SHIFT(op, ext) \
+static void bemit_##op(const ir_node *node) \
+{ \
+       const arch_register_t *out   = get_out_reg(node, 0); \
+       ir_node               *count = get_irn_n(node, 1); \
+       if (is_ia32_Immediate(count)) { \
+               int offset = get_ia32_immediate_attr_const(count)->offset; \
+               if (offset == 1) { \
+                       bemit8(0xD1); \
+                       bemit_modru(out, ext); \
+               } else { \
+                       bemit8(0xC1); \
+                       bemit_modru(out, ext); \
+                       bemit8(offset); \
+               } \
+       } else { \
+               bemit8(0xD3); \
+               bemit_modru(out, ext); \
+       } \
+} \
+ \
+static void bemit_##op##mem(const ir_node *node) \
+{ \
+       ir_node *count; \
+       unsigned size = get_mode_size_bits(get_ia32_ls_mode(node)); \
+       if (size == 16) \
+               bemit8(0x66); \
+       count = get_irn_n(node, 1); \
+       if (is_ia32_Immediate(count)) { \
+               int offset = get_ia32_immediate_attr_const(count)->offset; \
+               if (offset == 1) { \
+                       bemit8(size == 8 ? 0xD0 : 0xD1); \
+                       bemit_mod_am(ext, node); \
+               } else { \
+                       bemit8(size == 8 ? 0xC0 : 0xC1); \
+                       bemit_mod_am(ext, node); \
+                       bemit8(offset); \
+               } \
+       } else { \
+               bemit8(size == 8 ? 0xD2 : 0xD3); \
+               bemit_mod_am(ext, node); \
+       } \
+}
+
+SHIFT(rol, 0)
+SHIFT(ror, 1)
+SHIFT(shl, 4)
+SHIFT(shr, 5)
+SHIFT(sar, 7)
+
+static void bemit_shld(const ir_node *node)
+{
+       const arch_register_t *in  = get_in_reg(node, n_ia32_ShlD_val_low);
+       const arch_register_t *out = get_out_reg(node, pn_ia32_ShlD_res);
+       ir_node *count = get_irn_n(node, n_ia32_ShlD_count);
+       bemit8(0x0F);
+       if (is_ia32_Immediate(count)) {
+               bemit8(0xA4);
+               bemit_modrr(out, in);
+               bemit8(get_ia32_immediate_attr_const(count)->offset);
+       } else {
+               bemit8(0xA5);
+               bemit_modrr(out, in);
+       }
+}
+
+static void bemit_shrd(const ir_node *node)
+{
+       const arch_register_t *in  = get_in_reg(node, n_ia32_ShrD_val_low);
+       const arch_register_t *out = get_out_reg(node, pn_ia32_ShrD_res);
+       ir_node *count = get_irn_n(node, n_ia32_ShrD_count);
+       bemit8(0x0F);
+       if (is_ia32_Immediate(count)) {
+               bemit8(0xAC);
+               bemit_modrr(out, in);
+               bemit8(get_ia32_immediate_attr_const(count)->offset);
+       } else {
+               bemit8(0xAD);
+               bemit_modrr(out, in);
+       }
+}
+
+/**
+ * binary emitter for setcc.
+ */
+static void bemit_setcc(const ir_node *node)
+{
+       const arch_register_t *dreg = get_out_reg(node, pn_ia32_Setcc_res);
+
+       pn_Cmp pnc = get_ia32_condcode(node);
+       pnc        = determine_final_pnc(node, n_ia32_Setcc_eflags, pnc);
+       if (pnc & ia32_pn_Cmp_float) {
+               switch (pnc & 0x0f) {
+               case pn_Cmp_Uo:
+                        /* setp <dreg */
+                       bemit8(0x0F);
+                       bemit8(0x9A);
+                       bemit_modrm8(REG_LOW, dreg);
+                       return;
+
+               case pn_Cmp_Leg:
+                        /* setnp <dreg*/
+                       bemit8(0x0F);
+                       bemit8(0x9B);
+                       bemit_modrm8(REG_LOW, dreg);
+                       return;
+
+               case pn_Cmp_Eq:
+               case pn_Cmp_Lt:
+               case pn_Cmp_Le:
+                        /* set%PNC <dreg */
+                       bemit8(0x0F);
+                       bemit8(0x90 | pnc2cc(pnc));
+                       bemit_modrm8(REG_LOW, dreg);
+
+                       /* setnp >dreg */
+                       bemit8(0x0F);
+                       bemit8(0x9B);
+                       bemit_modrm8(REG_HIGH, dreg);
+
+                       /* andb %>dreg, %<dreg */
+                       bemit8(0x20);
+                       bemit_modrr8(REG_LOW, dreg, REG_HIGH, dreg);
+                       return;
+
+               case pn_Cmp_Ug:
+               case pn_Cmp_Uge:
+               case pn_Cmp_Ne:
+                       /* set%PNC <dreg */
+                       bemit8(0x0F);
+                       bemit8(0x90 | pnc2cc(pnc));
+                       bemit_modrm8(REG_LOW, dreg);
+
+                       /* setp >dreg */
+                       bemit8(0x0F);
+                       bemit8(0x9A);
+                       bemit_modrm8(REG_HIGH, dreg);
+
+                       /* orb %>dreg, %<dreg */
+                       bemit8(0x08);
+                       bemit_modrr8(REG_LOW, dreg, REG_HIGH, dreg);
+                       return;
+
+               default:
+                       break;
+               }
+       }
+       /* set%PNC <dreg */
+       bemit8(0x0F);
+       bemit8(0x90 | pnc2cc(pnc));
+       bemit_modrm8(REG_LOW, dreg);
+}
+
+static void bemit_cmovcc(const ir_node *node)
+{
+       const ia32_attr_t     *attr         = get_ia32_attr_const(node);
+       int                    ins_permuted = attr->data.ins_permuted;
+       const arch_register_t *out          = arch_irn_get_register(node, pn_ia32_res);
+       pn_Cmp                 pnc          = get_ia32_condcode(node);
+       const arch_register_t *in_true;
+       const arch_register_t *in_false;
+
+       pnc = determine_final_pnc(node, n_ia32_CMovcc_eflags, pnc);
+
+       in_true  = arch_get_irn_register(get_irn_n(node, n_ia32_CMovcc_val_true));
+       in_false = arch_get_irn_register(get_irn_n(node, n_ia32_CMovcc_val_false));
+
+       /* should be same constraint fullfilled? */
+       if (out == in_false) {
+               /* yes -> nothing to do */
+       } else if (out == in_true) {
+               assert(get_ia32_op_type(node) == ia32_Normal);
+               ins_permuted = !ins_permuted;
+               in_true      = in_false;
+       } else {
+               /* we need a mov */
+               bemit8(0x8B); // mov %in_false, %out
+               bemit_modrr(in_false, out);
+       }
+
+       if (ins_permuted)
+               pnc = ia32_get_negated_pnc(pnc);
+
+       /* TODO: handling of Nans isn't correct yet */
+
+       bemit8(0x0F);
+       bemit8(0x40 | pnc2cc(pnc));
+       if (get_ia32_op_type(node) == ia32_Normal) {
+               bemit_modrr(in_true, out);
+       } else {
+               bemit_mod_am(reg_gp_map[out->index], node);
+       }
+}
+
+static void bemit_cmp(const ir_node *node)
+{
+       unsigned  ls_size = get_mode_size_bits(get_ia32_ls_mode(node));
+       ir_node  *right;
+
+       if (ls_size == 16)
+               bemit8(0x66);
+
+       right = get_irn_n(node, n_ia32_binary_right);
+       if (is_ia32_Immediate(right)) {
+               /* Use in-reg, because some instructions (cmp, test) have no out-reg. */
+               const ir_node               *op   = get_irn_n(node, n_ia32_binary_right);
+               const ia32_immediate_attr_t *attr = get_ia32_immediate_attr_const(op);
+               unsigned                     size;
+
+               if (attr->symconst != NULL) {
+                       size = 4;
+               } else {
+                       /* check for sign extension */
+                       size = get_signed_imm_size(attr->offset);
+               }
+
+               switch (size) {
+                       case 1:
+                               bemit8(0x81 | SIGNEXT_IMM);
+                               /* cmp has this special mode */
+                               if (get_ia32_op_type(node) == ia32_AddrModeS) {
+                                       bemit_mod_am(7, node);
+                               } else {
+                                       const arch_register_t *reg = get_in_reg(node, n_ia32_binary_left);
+                                       bemit_modru(reg, 7);
+                               }
+                               bemit8((unsigned char)attr->offset);
+                               return;
+                       case 2:
+                       case 4:
+                               /* check for eax variant: this variant is shorter for 32bit immediates only */
+                               if (get_ia32_op_type(node) == ia32_AddrModeS) {
+                                       bemit8(0x81);
+                                       bemit_mod_am(7, node);
+                               } else {
+                                       const arch_register_t *reg = get_in_reg(node, n_ia32_binary_left);
+                                       if (reg->index == REG_EAX) {
+                                               bemit8(0x3D);
+                                       } else {
+                                               bemit8(0x81);
+                                               bemit_modru(reg, 7);
+                                       }
+                               }
+                               if (ls_size == 16) {
+                                       bemit16(attr->offset);
+                               } else {
+                                       bemit_entity(attr->symconst, attr->sc_sign, attr->offset, false);
+                               }
+                               return;
+               }
+               panic("invalid imm size?!?");
+       } else {
+               const arch_register_t *out = get_in_reg(node, n_ia32_binary_left);
+               bemit8(0x3B);
+               if (get_ia32_op_type(node) == ia32_Normal) {
+                       const arch_register_t *op2 = get_in_reg(node, n_ia32_binary_right);
+                       bemit_modrr(op2, out);
+               } else {
+                       bemit_mod_am(reg_gp_map[out->index], node);
+               }
+       }
+}
+
+static void bemit_cmp8bit(const ir_node *node)
+{
+       ir_node *right = get_irn_n(node, n_ia32_binary_right);
+       if (is_ia32_Immediate(right)) {
+               if (get_ia32_op_type(node) == ia32_Normal) {
+                       const arch_register_t *out = get_in_reg(node, n_ia32_Cmp_left);
+                       if (out->index == REG_EAX) {
+                               bemit8(0x3C);
+                       } else {
+                               bemit8(0x80);
+                               bemit_modru(out, 7);
+                       }
+               } else {
+                       bemit8(0x80);
+                       bemit_mod_am(7, node);
+               }
+               bemit8(get_ia32_immediate_attr_const(right)->offset);
+       } else {
+               const arch_register_t *out = get_in_reg(node, n_ia32_Cmp_left);
+               bemit8(0x3A);
+               if (get_ia32_op_type(node) == ia32_Normal) {
+                       const arch_register_t *in = get_in_reg(node, n_ia32_Cmp_right);
+                       bemit_modrr(out, in);
+               } else {
+                       bemit_mod_am(reg_gp_map[out->index], node);
+               }
+       }
+}
+
+static void bemit_test8bit(const ir_node *node)
+{
+       ir_node *right = get_irn_n(node, n_ia32_Test8Bit_right);
+       if (is_ia32_Immediate(right)) {
+               if (get_ia32_op_type(node) == ia32_Normal) {
+                       const arch_register_t *out = get_in_reg(node, n_ia32_Test8Bit_left);
+                       if (out->index == REG_EAX) {
+                               bemit8(0xA8);
+                       } else {
+                               bemit8(0xF6);
+                               bemit_modru(out, 0);
+                       }
+               } else {
+                       bemit8(0xF6);
+                       bemit_mod_am(0, node);
+               }
+               bemit8(get_ia32_immediate_attr_const(right)->offset);
+       } else {
+               const arch_register_t *out = get_in_reg(node, n_ia32_Test8Bit_left);
+               bemit8(0x84);
+               if (get_ia32_op_type(node) == ia32_Normal) {
+                       const arch_register_t *in = get_in_reg(node, n_ia32_Test8Bit_right);
+                       bemit_modrr(out, in);
+               } else {
+                       bemit_mod_am(reg_gp_map[out->index], node);
+               }
+       }
+}
+
+static void bemit_imul(const ir_node *node)
+{
+       ir_node *right = get_irn_n(node, n_ia32_IMul_right);
+       /* Do we need the immediate form? */
+       if (is_ia32_Immediate(right)) {
+               int imm = get_ia32_immediate_attr_const(right)->offset;
+               if (get_signed_imm_size(imm) == 1) {
+                       bemit_unop_reg(node, 0x6B, n_ia32_IMul_left);
+                       bemit8(imm);
+               } else {
+                       bemit_unop_reg(node, 0x69, n_ia32_IMul_left);
+                       bemit32(imm);
+               }
+       } else {
+               bemit8(0x0F);
+               bemit_unop_reg(node, 0xAF, n_ia32_IMul_right);
+       }
+}
+
+static void bemit_dec(const ir_node *node)
+{
+       const arch_register_t *out = get_out_reg(node, pn_ia32_Dec_res);
+       bemit8(0x48 + reg_gp_map[out->index]);
+}
+
+static void bemit_inc(const ir_node *node)
+{
+       const arch_register_t *out = get_out_reg(node, pn_ia32_Inc_res);
+       bemit8(0x40 + reg_gp_map[out->index]);
+}
+
+#define UNOPMEM(op, code, ext) \
+static void bemit_##op(const ir_node *node) \
+{ \
+       bemit_unop_mem(node, code, ext); \
+}
+
+UNOPMEM(notmem, 0xF6, 2)
+UNOPMEM(negmem, 0xF6, 3)
+UNOPMEM(incmem, 0xFE, 0)
+UNOPMEM(decmem, 0xFE, 1)
+
+static void bemit_ldtls(const ir_node *node)
+{
+       const arch_register_t *out = get_out_reg(node, 0);
+
+       bemit8(0x65); // gs:
+       if (out->index == REG_EAX) {
+               bemit8(0xA1); // movl 0, %eax
+       } else {
+               bemit8(0x8B); // movl 0, %reg
+               bemit8(MOD_IND | ENC_REG(reg_gp_map[out->index]) | ENC_RM(0x05));
+       }
+       bemit32(0);
+}
+
+/**
+ * Emit a Lea.
+ */
+static void bemit_lea(const ir_node *node)
+{
+       const arch_register_t *out = get_out_reg(node, 0);
+       bemit8(0x8D);
+       bemit_mod_am(reg_gp_map[out->index], node);
+}
+
+/* helper function for bemit_minus64bit */
+static void bemit_helper_mov(const arch_register_t *src, const arch_register_t *dst)
+{
+       bemit8(0x8B); // movl %src, %dst
+       bemit_modrr(src, dst);
+}
+
+/* helper function for bemit_minus64bit */
+static void bemit_helper_neg(const arch_register_t *reg)
+{
+       bemit8(0xF7); // negl %reg
+       bemit_modru(reg, 3);
+}
+
+/* helper function for bemit_minus64bit */
+static void bemit_helper_sbb0(const arch_register_t *reg)
+{
+       bemit8(0x83); // sbbl $0, %reg
+       bemit_modru(reg, 3);
+       bemit8(0);
+}
+
+/* helper function for bemit_minus64bit */
+static void bemit_helper_sbb(const arch_register_t *src, const arch_register_t *dst)
+{
+       bemit8(0x1B); // sbbl %src, %dst
+       bemit_modrr(src, dst);
+}
+
+/* helper function for bemit_minus64bit */
+static void bemit_helper_xchg(const arch_register_t *src, const arch_register_t *dst)
+{
+       if (src->index == REG_EAX) {
+               bemit8(0x90 + reg_gp_map[dst->index]); // xchgl %eax, %dst
+       } else if (dst->index == REG_EAX) {
+               bemit8(0x90 + reg_gp_map[src->index]); // xchgl %src, %eax
+       } else {
+               bemit8(0x87); // xchgl %src, %dst
+               bemit_modrr(src, dst);
+       }
+}
+
+/* helper function for bemit_minus64bit */
+static void bemit_helper_zero(const arch_register_t *reg)
+{
+       bemit8(0x33); // xorl %reg, %reg
+       bemit_modrr(reg, reg);
+}
+
+static void bemit_minus64bit(const ir_node *node)
+{
+       const arch_register_t *in_lo  = get_in_reg(node, 0);
+       const arch_register_t *in_hi  = get_in_reg(node, 1);
+       const arch_register_t *out_lo = get_out_reg(node, 0);
+       const arch_register_t *out_hi = get_out_reg(node, 1);
+
+       if (out_lo == in_lo) {
+               if (out_hi != in_hi) {
+                       /* a -> a, b -> d */
+                       goto zero_neg;
+               } else {
+                       /* a -> a, b -> b */
+                       goto normal_neg;
+               }
+       } else if (out_lo == in_hi) {
+               if (out_hi == in_lo) {
+                       /* a -> b, b -> a */
+                       bemit_helper_xchg(in_lo, in_hi);
+                       goto normal_neg;
+               } else {
+                       /* a -> b, b -> d */
+                       bemit_helper_mov(in_hi, out_hi);
+                       bemit_helper_mov(in_lo, out_lo);
+                       goto normal_neg;
+               }
+       } else {
+               if (out_hi == in_lo) {
+                       /* a -> c, b -> a */
+                       bemit_helper_mov(in_lo, out_lo);
+                       goto zero_neg;
+               } else if (out_hi == in_hi) {
+                       /* a -> c, b -> b */
+                       bemit_helper_mov(in_lo, out_lo);
+                       goto normal_neg;
+               } else {
+                       /* a -> c, b -> d */
+                       bemit_helper_mov(in_lo, out_lo);
+                       goto zero_neg;
+               }
+       }
+
+normal_neg:
+       bemit_helper_neg( out_hi);
+       bemit_helper_neg( out_lo);
+       bemit_helper_sbb0(out_hi);
+       return;
+
+zero_neg:
+       bemit_helper_zero(out_hi);
+       bemit_helper_neg( out_lo);
+       bemit_helper_sbb( in_hi, out_hi);
+}
+
+/**
+ * Emit a single opcode.
+ */
+#define EMIT_SINGLEOP(op, code)                 \
+static void bemit_ ## op(const ir_node *node) { \
+       (void) node;                                \
+       bemit8(code);                               \
+}
+
+//EMIT_SINGLEOP(daa,  0x27)
+//EMIT_SINGLEOP(das,  0x2F)
+//EMIT_SINGLEOP(aaa,  0x37)
+//EMIT_SINGLEOP(aas,  0x3F)
+//EMIT_SINGLEOP(nop,  0x90)
+EMIT_SINGLEOP(cwtl,  0x98)
+EMIT_SINGLEOP(cltd,  0x99)
+//EMIT_SINGLEOP(fwait, 0x9B)
+EMIT_SINGLEOP(sahf,  0x9E)
+//EMIT_SINGLEOP(popf, 0x9D)
+EMIT_SINGLEOP(leave, 0xC9)
+EMIT_SINGLEOP(int3,  0xCC)
+//EMIT_SINGLEOP(iret, 0xCF)
+//EMIT_SINGLEOP(xlat, 0xD7)
+//EMIT_SINGLEOP(lock, 0xF0)
+EMIT_SINGLEOP(rep,   0xF3)
+//EMIT_SINGLEOP(halt, 0xF4)
+EMIT_SINGLEOP(cmc,   0xF5)
+EMIT_SINGLEOP(stc,   0xF9)
+//EMIT_SINGLEOP(cli,  0xFA)
+//EMIT_SINGLEOP(sti,  0xFB)
+//EMIT_SINGLEOP(std,  0xFD)
+
+/**
+ * Emits a MOV out, [MEM].
+ */
+static void bemit_load(const ir_node *node)
+{
+       const arch_register_t *out = get_out_reg(node, 0);
+
+       if (out->index == REG_EAX) {
+               ir_node   *base      = get_irn_n(node, n_ia32_base);
+               int        has_base  = !is_ia32_NoReg_GP(base);
+               ir_node   *index     = get_irn_n(node, n_ia32_index);
+               int        has_index = !is_ia32_NoReg_GP(index);
+               if (!has_base && !has_index) {
+                       ir_entity *ent  = get_ia32_am_sc(node);
+                       int        offs = get_ia32_am_offs_int(node);
+                       /* load from constant address to EAX can be encoded
+                          as 0xA1 [offset] */
+                       bemit8(0xA1);
+                       bemit_entity(ent, 0, offs, false);
+                       return;
+               }
+       }
+       bemit8(0x8B);
+       bemit_mod_am(reg_gp_map[out->index], node);
+}
+
+/**
+ * Emits a MOV [mem], in.
+ */
+static void bemit_store(const ir_node *node)
+{
+       const ir_node *value = get_irn_n(node, n_ia32_Store_val);
+       unsigned       size  = get_mode_size_bits(get_ia32_ls_mode(node));
+
+       if (is_ia32_Immediate(value)) {
+               if (size == 8) {
+                       bemit8(0xC6);
+                       bemit_mod_am(0, node);
+                       bemit8(get_ia32_immediate_attr_const(value)->offset);
+               } else if (size == 16) {
+                       bemit8(0x66);
+                       bemit8(0xC7);
+                       bemit_mod_am(0, node);
+                       bemit16(get_ia32_immediate_attr_const(value)->offset);
+               } else {
+                       bemit8(0xC7);
+                       bemit_mod_am(0, node);
+                       bemit_immediate(value, false);
+               }
+       } else {
+               const arch_register_t *in = get_in_reg(node, n_ia32_Store_val);
+
+               if (in->index == REG_EAX) {
+                       ir_node   *base      = get_irn_n(node, n_ia32_base);
+                       int        has_base  = !is_ia32_NoReg_GP(base);
+                       ir_node   *index     = get_irn_n(node, n_ia32_index);
+                       int        has_index = !is_ia32_NoReg_GP(index);
+                       if (!has_base && !has_index) {
+                               ir_entity *ent  = get_ia32_am_sc(node);
+                               int        offs = get_ia32_am_offs_int(node);
+                               /* store to constant address from EAX can be encoded as
+                                * 0xA2/0xA3 [offset]*/
+                               if (size == 8) {
+                                       bemit8(0xA2);
+                               } else {
+                                       if (size == 16)
+                                               bemit8(0x66);
+                                       bemit8(0xA3);
+                               }
+                               bemit_entity(ent, 0, offs, false);
+                               return;
+                       }
+               }
+
+               if (size == 8) {
+                       bemit8(0x88);
+               } else {
+                       if (size == 16)
+                               bemit8(0x66);
+                       bemit8(0x89);
+               }
+               bemit_mod_am(reg_gp_map[in->index], node);
+       }
+}
+
+static void bemit_conv_i2i(const ir_node *node)
+{
+       ir_mode  *smaller_mode = get_ia32_ls_mode(node);
+       unsigned  opcode;
+
+       bemit8(0x0F);
+       /*        8 16 bit source
+        * movzx B6 B7
+        * movsx BE BF
+        */
+       opcode = 0xB6;
+       if (mode_is_signed(smaller_mode))           opcode |= 0x08;
+       if (get_mode_size_bits(smaller_mode) == 16) opcode |= 0x01;
+       bemit_unop_reg(node, opcode, n_ia32_Conv_I2I_val);
+}
+
+/**
+ * Emit a Push.
+ */
+static void bemit_push(const ir_node *node)
+{
+       const ir_node *value = get_irn_n(node, n_ia32_Push_val);
+
+       if (is_ia32_Immediate(value)) {
+               const ia32_immediate_attr_t *attr
+                       = get_ia32_immediate_attr_const(value);
+               unsigned size = get_signed_imm_size(attr->offset);
+               if (attr->symconst)
+                       size = 4;
+               switch (size) {
+               case 1:
+                       bemit8(0x6A);
+                       bemit8((unsigned char)attr->offset);
+                       break;
+               case 2:
+               case 4:
+                       bemit8(0x68);
+                       bemit_immediate(value, false);
+                       break;
+               }
+       } else if (is_ia32_NoReg_GP(value)) {
+               bemit8(0xFF);
+               bemit_mod_am(6, node);
+       } else {
+               const arch_register_t *reg = get_in_reg(node, n_ia32_Push_val);
+               bemit8(0x50 + reg_gp_map[reg->index]);
+       }
+}
+
+/**
+ * Emit a Pop.
+ */
+static void bemit_pop(const ir_node *node)
+{
+       const arch_register_t *reg = get_out_reg(node, pn_ia32_Pop_res);
+       bemit8(0x58 + reg_gp_map[reg->index]);
+}
+
+static void bemit_popmem(const ir_node *node)
+{
+       bemit8(0x8F);
+       bemit_mod_am(0, node);
+}
+
+static void bemit_call(const ir_node *node)
+{
+       ir_node *proc = get_irn_n(node, n_ia32_Call_addr);
+
+       if (is_ia32_Immediate(proc)) {
+               bemit8(0xE8);
+               bemit_immediate(proc, true);
+       } else {
+               bemit_unop(node, 0xFF, 2, n_ia32_Call_addr);
+       }
+}
+
+static void bemit_jmp(const ir_node *dest_block)
+{
+       bemit8(0xE9);
+       bemit_jmp_destination(dest_block);
+}
+
+static void bemit_jump(const ir_node *node)
+{
+       if (can_be_fallthrough(node))
+               return;
+
+       bemit_jmp(get_cfop_target_block(node));
+}
+
+static void bemit_jcc(int pnc, const ir_node *dest_block)
+{
+       unsigned char cc = pnc2cc(pnc);
+       bemit8(0x0F);
+       bemit8(0x80 + cc);
+       bemit_jmp_destination(dest_block);
+}
+
+static void bemit_jp(bool odd, const ir_node *dest_block)
+{
+       bemit8(0x0F);
+       bemit8(0x8A + odd);
+       bemit_jmp_destination(dest_block);
+}
+
+static void bemit_ia32_jcc(const ir_node *node)
+{
+       int            pnc = get_ia32_condcode(node);
+       const ir_node *proj_true;
+       const ir_node *proj_false;
+       const ir_node *dest_true;
+       const ir_node *dest_false;
+       const ir_node *block;
+
+       pnc = determine_final_pnc(node, 0, pnc);
+
+       /* get both Projs */
+       proj_true = get_proj(node, pn_ia32_Jcc_true);
+       assert(proj_true && "Jcc without true Proj");
+
+       proj_false = get_proj(node, pn_ia32_Jcc_false);
+       assert(proj_false && "Jcc without false Proj");
+
+       block = get_nodes_block(node);
+
+       if (can_be_fallthrough(proj_true)) {
+               /* exchange both proj's so the second one can be omitted */
+               const ir_node *t = proj_true;
+
+               proj_true  = proj_false;
+               proj_false = t;
+               pnc        = ia32_get_negated_pnc(pnc);
+       }
+
+       dest_true  = get_cfop_target_block(proj_true);
+       dest_false = get_cfop_target_block(proj_false);
+
+       if (pnc & ia32_pn_Cmp_float) {
+               /* Some floating point comparisons require a test of the parity flag,
+                * which indicates that the result is unordered */
+               switch (pnc & 15) {
+                       case pn_Cmp_Uo: {
+                               bemit_jp(false, dest_true);
+                               break;
+                       }
+
+                       case pn_Cmp_Leg:
+                               bemit_jp(true, dest_true);
+                               break;
+
+                       case pn_Cmp_Eq:
+                       case pn_Cmp_Lt:
+                       case pn_Cmp_Le:
+                               /* we need a local label if the false proj is a fallthrough
+                                * as the falseblock might have no label emitted then */
+                               if (can_be_fallthrough(proj_false)) {
+                                       bemit8(0x7A);
+                                       bemit8(0x06);  // jp + 6
+                               } else {
+                                       bemit_jp(false, dest_false);
+                               }
+                               goto emit_jcc;
+
+                       case pn_Cmp_Ug:
+                       case pn_Cmp_Uge:
+                       case pn_Cmp_Ne:
+                               bemit_jp(false, dest_true);
+                               goto emit_jcc;
+
+                       default:
+                               goto emit_jcc;
+               }
+       } else {
+emit_jcc:
+               bemit_jcc(pnc, dest_true);
+       }
+
+       /* the second Proj might be a fallthrough */
+       if (can_be_fallthrough(proj_false)) {
+               /* it's a fallthrough */
+       } else {
+               bemit_jmp(dest_false);
+       }
+}
+
+static void bemit_switchjmp(const ir_node *node)
+{
+       unsigned long          interval;
+       int                    last_value;
+       int                    i;
+       jmp_tbl_t              tbl;
+       const arch_register_t *in;
+
+       /* fill the table structure */
+       generate_jump_table(&tbl, node);
+
+       /* two-complement's magic make this work without overflow */
+       interval = tbl.max_value - tbl.min_value;
+
+       in = get_in_reg(node, 0);
+       /* emit the table */
+       if (get_signed_imm_size(interval) == 1) {
+               bemit8(0x83); // cmpl $imm8, %in
+               bemit_modru(in, 7);
+               bemit8(interval);
+       } else {
+               bemit8(0x81); // cmpl $imm32, %in
+               bemit_modru(in, 7);
+               bemit32(interval);
+       }
+       bemit8(0x0F); // ja tbl.defProj
+       bemit8(0x87);
+       ia32_emitf(tbl.defProj, ".long %L - . - 4\n");
+
+       if (tbl.num_branches > 1) {
+               /* create table */
+               bemit8(0xFF); // jmp *tbl.label(,%in,4)
+               bemit8(MOD_IND | ENC_REG(4) | ENC_RM(0x04));
+               bemit8(ENC_SIB(2, reg_gp_map[in->index], 0x05));
+               be_emit_irprintf("\t.long %s\n", tbl.label);
+
+               be_gas_emit_switch_section(GAS_SECTION_RODATA);
+               be_emit_cstring(".align 4\n");
+               be_emit_irprintf("%s:\n", tbl.label);
+
+               last_value = tbl.branches[0].value;
+               for (i = 0; i != tbl.num_branches; ++i) {
+                       while (last_value != tbl.branches[i].value) {
+                               ia32_emitf(tbl.defProj, ".long %L\n");
+                               ++last_value;
+                       }
+                       ia32_emitf(tbl.branches[i].target, ".long %L\n");
+                       ++last_value;
+               }
+               be_gas_emit_switch_section(GAS_SECTION_TEXT);
+       } else {
+               /* one jump is enough */
+               panic("switch only has one case");
+               //ia32_emitf(tbl.branches[0].target, "\tjmp %L\n");
+       }
+
+       be_emit_write_line();
+
+       free(tbl.branches);
+}
+
+/**
+ * Emits a return.
+ */
+static void bemit_return(const ir_node *node)
+{
+       unsigned pop = be_Return_get_pop(node);
+       if (pop > 0 || be_Return_get_emit_pop(node)) {
+               bemit8(0xC2);
+               assert(pop <= 0xffff);
+               bemit16(pop);
+       } else {
+               bemit8(0xC3);
+       }
+}
+
+static void bemit_subsp(const ir_node *node)
+{
+       const arch_register_t *out;
+       /* sub %in, %esp */
+       bemit_sub(node);
+       /* mov %esp, %out */
+       bemit8(0x8B);
+       out = get_out_reg(node, 1);
+       bemit8(MOD_REG | ENC_REG(reg_gp_map[out->index]) | ENC_RM(0x04));
+}
+
+static void bemit_incsp(const ir_node *node)
+{
+       int                    offs;
+       const arch_register_t *reg;
+       unsigned               size;
+       unsigned               ext;
+
+       offs = be_get_IncSP_offset(node);
+       if (offs == 0)
+               return;
+
+       if (offs > 0) {
+               ext = 5; /* sub */
+       } else {
+               ext = 0; /* add */
+               offs = -offs;
+       }
+
+       size = get_signed_imm_size(offs);
+       bemit8(size == 1 ? 0x83 : 0x81);
+
+       reg  = get_out_reg(node, 0);
+       bemit_modru(reg, ext);
+
+       if (size == 1) {
+               bemit8(offs);
+       } else {
+               bemit32(offs);
+       }
+}
+
+static void bemit_copybi(const ir_node *node)
+{
+       unsigned size = get_ia32_copyb_size(node);
+       if (size & 1)
+               bemit8(0xA4); // movsb
+       if (size & 2) {
+               bemit8(0x66);
+               bemit8(0xA5); // movsw
+       }
+       size >>= 2;
+       while (size--) {
+               bemit8(0xA5); // movsl
+       }
+}
+
+static void bemit_fbinop(const ir_node *node, unsigned code, unsigned code_to)
+{
+       if (get_ia32_op_type(node) == ia32_Normal) {
+               const ia32_x87_attr_t *x87_attr = get_ia32_x87_attr_const(node);
+               const arch_register_t *in1      = x87_attr->x87[0];
+               const arch_register_t *in       = x87_attr->x87[1];
+               const arch_register_t *out      = x87_attr->x87[2];
+
+               if (out == NULL) {
+                       out = in1;
+               } else if (out == in) {
+                       in = in1;
+               }
+
+               if (out->index == 0) {
+                       bemit8(0xD8);
+                       bemit8(MOD_REG | ENC_REG(code) | ENC_RM(in->index));
+               } else {
+                       bemit8(0xDC);
+                       bemit8(MOD_REG | ENC_REG(code_to) | ENC_RM(out->index));
+               }
+       } else {
+               if (get_mode_size_bits(get_ia32_ls_mode(node)) == 32) {
+                       bemit8(0xD8);
+               } else {
+                       bemit8(0xDC);
+               }
+               bemit_mod_am(code, node);
+       }
+}
+
+static void bemit_fbinopp(const ir_node *node, unsigned const code)
+{
+       const ia32_x87_attr_t *x87_attr = get_ia32_x87_attr_const(node);
+       const arch_register_t *out      = x87_attr->x87[2];
+       bemit8(0xDE);
+       bemit8(code + out->index);
+}
+
+static void bemit_fabs(const ir_node *node)
+{
+       (void)node;
+
+       bemit8(0xD9);
+       bemit8(0xE1);
+}
+
+static void bemit_fadd(const ir_node *node)
+{
+       bemit_fbinop(node, 0, 0);
+}
+
+static void bemit_faddp(const ir_node *node)
+{
+       bemit_fbinopp(node, 0xC0);
+}
+
+static void bemit_fchs(const ir_node *node)
+{
+       (void)node;
+
+       bemit8(0xD9);
+       bemit8(0xE0);
+}
+
+static void bemit_fdiv(const ir_node *node)
+{
+       bemit_fbinop(node, 6, 7);
+}
+
+static void bemit_fdivp(const ir_node *node)
+{
+       bemit_fbinopp(node, 0xF8);
+}
+
+static void bemit_fdivr(const ir_node *node)
+{
+       bemit_fbinop(node, 7, 6);
+}
+
+static void bemit_fdivrp(const ir_node *node)
+{
+       bemit_fbinopp(node, 0xF0);
+}
+
+static void bemit_fild(const ir_node *node)
+{
+       switch (get_mode_size_bits(get_ia32_ls_mode(node))) {
+               case 16:
+                       bemit8(0xDF); // filds
+                       bemit_mod_am(0, node);
+                       return;
+
+               case 32:
+                       bemit8(0xDB); // fildl
+                       bemit_mod_am(0, node);
+                       return;
+
+               case 64:
+                       bemit8(0xDF); // fildll
+                       bemit_mod_am(5, node);
+                       return;
+
+               default:
+                       panic("invalid mode size");
+       }
+}
+
+static void bemit_fist(const ir_node *node)
+{
+       switch (get_mode_size_bits(get_ia32_ls_mode(node))) {
+               case 16:
+                       bemit8(0xDF); // fists
+                       break;
+
+               case 32:
+                       bemit8(0xDB); // fistl
+                       break;
+
+               default:
+                       panic("invalid mode size");
+       }
+       bemit_mod_am(2, node);
+}
+
+static void bemit_fistp(const ir_node *node)
+{
+       switch (get_mode_size_bits(get_ia32_ls_mode(node))) {
+               case 16:
+                       bemit8(0xDF); // fistps
+                       bemit_mod_am(3, node);
+                       return;
+
+               case 32:
+                       bemit8(0xDB); // fistpl
+                       bemit_mod_am(3, node);
+                       return;
+
+               case 64:
+                       bemit8(0xDF); // fistpll
+                       bemit_mod_am(7, node);
+                       return;
+
+               default:
+                       panic("invalid mode size");
+       }
+}
+
+static void bemit_fld(const ir_node *node)
+{
+       switch (get_mode_size_bits(get_ia32_ls_mode(node))) {
+               case 32:
+                       bemit8(0xD9); // flds
+                       bemit_mod_am(0, node);
+                       return;
+
+               case 64:
+                       bemit8(0xDD); // fldl
+                       bemit_mod_am(0, node);
+                       return;
+
+               case 80:
+               case 96:
+                       bemit8(0xDB); // fldt
+                       bemit_mod_am(5, node);
+                       return;
+
+               default:
+                       panic("invalid mode size");
+       }
+}
+
+static void bemit_fld1(const ir_node *node)
+{
+       (void)node;
+       bemit8(0xD9);
+       bemit8(0xE8); // fld1
+}
+
+static void bemit_fldcw(const ir_node *node)
+{
+       bemit8(0xD9); // fldcw
+       bemit_mod_am(5, node);
+}
+
+static void bemit_fldz(const ir_node *node)
+{
+       (void)node;
+       bemit8(0xD9);
+       bemit8(0xEE); // fldz
+}
+
+static void bemit_fmul(const ir_node *node)
+{
+       bemit_fbinop(node, 1, 1);
+}
+
+static void bemit_fmulp(const ir_node *node)
+{
+       bemit_fbinopp(node, 0xC8);
+}
+
+static void bemit_fpop(const ir_node *node)
+{
+       const ia32_x87_attr_t *attr = get_ia32_x87_attr_const(node);
+       bemit8(0xDD);
+       bemit8(0xD8 + attr->x87[0]->index);
+}
+
+static void bemit_fpush(const ir_node *node)
+{
+       const ia32_x87_attr_t *attr = get_ia32_x87_attr_const(node);
+       bemit8(0xD9);
+       bemit8(0xC0 + attr->x87[0]->index);
+}
+
+static void bemit_fpushcopy(const ir_node *node)
+{
+       const ia32_x87_attr_t *attr = get_ia32_x87_attr_const(node);
+       bemit8(0xD9);
+       bemit8(0xC0 + attr->x87[0]->index);
+}
+
+static void bemit_fst(const ir_node *node)
+{
+       switch (get_mode_size_bits(get_ia32_ls_mode(node))) {
+               case 32:
+                       bemit8(0xD9); // fsts
+                       break;
+
+               case 64:
+                       bemit8(0xDD); // fstl
+                       break;
+
+               default:
+                       panic("invalid mode size");
+       }
+       bemit_mod_am(2, node);
+}
+
+static void bemit_fstp(const ir_node *node)
+{
+       switch (get_mode_size_bits(get_ia32_ls_mode(node))) {
+               case 32:
+                       bemit8(0xD9); // fstps
+                       bemit_mod_am(3, node);
+                       return;
+
+               case 64:
+                       bemit8(0xDD); // fstpl
+                       bemit_mod_am(3, node);
+                       return;
+
+               case 80:
+               case 96:
+                       bemit8(0xDB); // fstpt
+                       bemit_mod_am(7, node);
+                       return;
+
+               default:
+                       panic("invalid mode size");
+       }
+}
+
+static void bemit_fsub(const ir_node *node)
+{
+       bemit_fbinop(node, 4, 5);
+}
+
+static void bemit_fsubp(const ir_node *node)
+{
+       bemit_fbinopp(node, 0xE8);
+}
+
+static void bemit_fsubr(const ir_node *node)
+{
+       bemit_fbinop(node, 5, 4);
+}
+
+static void bemit_fsubrp(const ir_node *node)
+{
+       bemit_fbinopp(node, 0xE0);
+}
+
+static void bemit_fnstcw(const ir_node *node)
+{
+       bemit8(0xD9); // fnstcw
+       bemit_mod_am(7, node);
+}
+
+static void bemit_fnstsw(void)
+{
+       bemit8(0xDF); // fnstsw %ax
+       bemit8(0xE0);
+}
+
+static void bemit_ftstfnstsw(const ir_node *node)
+{
+       (void)node;
+
+       bemit8(0xD9); // ftst
+       bemit8(0xE4);
+       bemit_fnstsw();
+}
+
+static void bemit_fucomi(const ir_node *node)
+{
+       const ia32_x87_attr_t *attr = get_ia32_x87_attr_const(node);
+       bemit8(0xDB); // fucomi
+       bemit8(0xE8 + attr->x87[1]->index);
+}
+
+static void bemit_fucomip(const ir_node *node)
+{
+       const ia32_x87_attr_t *attr = get_ia32_x87_attr_const(node);
+       bemit8(0xDF); // fucomip
+       bemit8(0xE8 + attr->x87[1]->index);
+}
+
+static void bemit_fucomfnstsw(const ir_node *node)
+{
+       const ia32_x87_attr_t *attr = get_ia32_x87_attr_const(node);
+       bemit8(0xDD); // fucom
+       bemit8(0xE0 + attr->x87[1]->index);
+       bemit_fnstsw();
+}
+
+static void bemit_fucompfnstsw(const ir_node *node)
+{
+       const ia32_x87_attr_t *attr = get_ia32_x87_attr_const(node);
+       bemit8(0xDD); // fucomp
+       bemit8(0xE8 + attr->x87[1]->index);
+       bemit_fnstsw();
+}
+
+static void bemit_fucomppfnstsw(const ir_node *node)
+{
+       (void)node;
+
+       bemit8(0xDA); // fucompp
+       bemit8(0xE9);
+       bemit_fnstsw();
+}
+
+static void bemit_fxch(const ir_node *node)
+{
+       const ia32_x87_attr_t *attr = get_ia32_x87_attr_const(node);
+       bemit8(0xD9);
+       bemit8(0xC8 + attr->x87[0]->index);
+}
+
+/**
+ * The type of a emitter function.
+ */
+typedef void (*emit_func) (const ir_node *);
+
+/**
+ * Set a node emitter. Make it a bit more type safe.
+ */
+static void register_emitter(ir_op *op, emit_func func)
+{
+       op->ops.generic = (op_func) func;
+}
+
+static void ia32_register_binary_emitters(void)
+{
+       /* first clear the generic function pointer for all ops */
+       clear_irp_opcodes_generic_func();
+
+       /* benode emitter */
+       register_emitter(op_be_Copy,            bemit_copy);
+       register_emitter(op_be_CopyKeep,        bemit_copy);
+       register_emitter(op_be_IncSP,           bemit_incsp);
+       register_emitter(op_be_Perm,            bemit_perm);
+       register_emitter(op_be_Return,          bemit_return);
+       register_emitter(op_ia32_Adc,           bemit_adc);
+       register_emitter(op_ia32_Add,           bemit_add);
+       register_emitter(op_ia32_AddMem,        bemit_addmem);
+       register_emitter(op_ia32_AddMem8Bit,    bemit_addmem8bit);
+       register_emitter(op_ia32_And,           bemit_and);
+       register_emitter(op_ia32_AndMem,        bemit_andmem);
+       register_emitter(op_ia32_AndMem8Bit,    bemit_andmem8bit);
+       register_emitter(op_ia32_Breakpoint,    bemit_int3);
+       register_emitter(op_ia32_CMovcc,        bemit_cmovcc);
+       register_emitter(op_ia32_Call,          bemit_call);
+       register_emitter(op_ia32_Cltd,          bemit_cltd);
+       register_emitter(op_ia32_Cmc,           bemit_cmc);
+       register_emitter(op_ia32_Cmp,           bemit_cmp);
+       register_emitter(op_ia32_Cmp8Bit,       bemit_cmp8bit);
+       register_emitter(op_ia32_Const,         bemit_mov_const);
+       register_emitter(op_ia32_Conv_I2I,      bemit_conv_i2i);
+       register_emitter(op_ia32_Conv_I2I8Bit,  bemit_conv_i2i);
+       register_emitter(op_ia32_CopyB_i,       bemit_copybi);
+       register_emitter(op_ia32_Cwtl,          bemit_cwtl);
+       register_emitter(op_ia32_Dec,           bemit_dec);
+       register_emitter(op_ia32_DecMem,        bemit_decmem);
+       register_emitter(op_ia32_Div,           bemit_div);
+       register_emitter(op_ia32_FldCW,         bemit_fldcw);
+       register_emitter(op_ia32_FnstCW,        bemit_fnstcw);
+       register_emitter(op_ia32_FtstFnstsw,    bemit_ftstfnstsw);
+       register_emitter(op_ia32_FucomFnstsw,   bemit_fucomfnstsw);
+       register_emitter(op_ia32_Fucomi,        bemit_fucomi);
+       register_emitter(op_ia32_FucompFnstsw,  bemit_fucompfnstsw);
+       register_emitter(op_ia32_Fucompi,       bemit_fucomip);
+       register_emitter(op_ia32_FucomppFnstsw, bemit_fucomppfnstsw);
+       register_emitter(op_ia32_IDiv,          bemit_idiv);
+       register_emitter(op_ia32_IJmp,          bemit_ijmp);
+       register_emitter(op_ia32_IMul,          bemit_imul);
+       register_emitter(op_ia32_IMul1OP,       bemit_imul1op);
+       register_emitter(op_ia32_Inc,           bemit_inc);
+       register_emitter(op_ia32_IncMem,        bemit_incmem);
+       register_emitter(op_ia32_Jcc,           bemit_ia32_jcc);
+       register_emitter(op_ia32_Jmp,           bemit_jump);
+       register_emitter(op_ia32_LdTls,         bemit_ldtls);
+       register_emitter(op_ia32_Lea,           bemit_lea);
+       register_emitter(op_ia32_Leave,         bemit_leave);
+       register_emitter(op_ia32_Load,          bemit_load);
+       register_emitter(op_ia32_Minus64Bit,    bemit_minus64bit);
+       register_emitter(op_ia32_Mul,           bemit_mul);
+       register_emitter(op_ia32_Neg,           bemit_neg);
+       register_emitter(op_ia32_NegMem,        bemit_negmem);
+       register_emitter(op_ia32_Not,           bemit_not);
+       register_emitter(op_ia32_NotMem,        bemit_notmem);
+       register_emitter(op_ia32_Or,            bemit_or);
+       register_emitter(op_ia32_OrMem,         bemit_ormem);
+       register_emitter(op_ia32_OrMem8Bit,     bemit_ormem8bit);
+       register_emitter(op_ia32_Pop,           bemit_pop);
+       register_emitter(op_ia32_PopEbp,        bemit_pop);
+       register_emitter(op_ia32_PopMem,        bemit_popmem);
+       register_emitter(op_ia32_Push,          bemit_push);
+       register_emitter(op_ia32_RepPrefix,     bemit_rep);
+       register_emitter(op_ia32_Rol,           bemit_rol);
+       register_emitter(op_ia32_RolMem,        bemit_rolmem);
+       register_emitter(op_ia32_Ror,           bemit_ror);
+       register_emitter(op_ia32_RorMem,        bemit_rormem);
+       register_emitter(op_ia32_Sahf,          bemit_sahf);
+       register_emitter(op_ia32_Sar,           bemit_sar);
+       register_emitter(op_ia32_SarMem,        bemit_sarmem);
+       register_emitter(op_ia32_Sbb,           bemit_sbb);
+       register_emitter(op_ia32_Setcc,         bemit_setcc);
+       register_emitter(op_ia32_Shl,           bemit_shl);
+       register_emitter(op_ia32_ShlD,          bemit_shld);
+       register_emitter(op_ia32_ShlMem,        bemit_shlmem);
+       register_emitter(op_ia32_Shr,           bemit_shr);
+       register_emitter(op_ia32_ShrD,          bemit_shrd);
+       register_emitter(op_ia32_ShrMem,        bemit_shrmem);
+       register_emitter(op_ia32_Stc,           bemit_stc);
+       register_emitter(op_ia32_Store,         bemit_store);
+       register_emitter(op_ia32_Store8Bit,     bemit_store);
+       register_emitter(op_ia32_Sub,           bemit_sub);
+       register_emitter(op_ia32_SubMem,        bemit_submem);
+       register_emitter(op_ia32_SubMem8Bit,    bemit_submem8bit);
+       register_emitter(op_ia32_SubSP,         bemit_subsp);
+       register_emitter(op_ia32_SwitchJmp,     bemit_switchjmp);
+       register_emitter(op_ia32_Test,          bemit_test);
+       register_emitter(op_ia32_Test8Bit,      bemit_test8bit);
+       register_emitter(op_ia32_Xor,           bemit_xor);
+       register_emitter(op_ia32_Xor0,          bemit_xor0);
+       register_emitter(op_ia32_XorMem,        bemit_xormem);
+       register_emitter(op_ia32_XorMem8Bit,    bemit_xormem8bit);
+       register_emitter(op_ia32_fabs,          bemit_fabs);
+       register_emitter(op_ia32_fadd,          bemit_fadd);
+       register_emitter(op_ia32_faddp,         bemit_faddp);
+       register_emitter(op_ia32_fchs,          bemit_fchs);
+       register_emitter(op_ia32_fdiv,          bemit_fdiv);
+       register_emitter(op_ia32_fdivp,         bemit_fdivp);
+       register_emitter(op_ia32_fdivr,         bemit_fdivr);
+       register_emitter(op_ia32_fdivrp,        bemit_fdivrp);
+       register_emitter(op_ia32_fild,          bemit_fild);
+       register_emitter(op_ia32_fist,          bemit_fist);
+       register_emitter(op_ia32_fistp,         bemit_fistp);
+       register_emitter(op_ia32_fld,           bemit_fld);
+       register_emitter(op_ia32_fld1,          bemit_fld1);
+       register_emitter(op_ia32_fldz,          bemit_fldz);
+       register_emitter(op_ia32_fmul,          bemit_fmul);
+       register_emitter(op_ia32_fmulp,         bemit_fmulp);
+       register_emitter(op_ia32_fpop,          bemit_fpop);
+       register_emitter(op_ia32_fpush,         bemit_fpush);
+       register_emitter(op_ia32_fpushCopy,     bemit_fpushcopy);
+       register_emitter(op_ia32_fst,           bemit_fst);
+       register_emitter(op_ia32_fstp,          bemit_fstp);
+       register_emitter(op_ia32_fsub,          bemit_fsub);
+       register_emitter(op_ia32_fsubp,         bemit_fsubp);
+       register_emitter(op_ia32_fsubr,         bemit_fsubr);
+       register_emitter(op_ia32_fsubrp,        bemit_fsubrp);
+       register_emitter(op_ia32_fxch,          bemit_fxch);
+
+       /* ignore the following nodes */
+       register_emitter(op_ia32_ProduceVal,   emit_Nothing);
+       register_emitter(op_be_Barrier,        emit_Nothing);
+       register_emitter(op_be_Keep,           emit_Nothing);
+       register_emitter(op_be_Start,          emit_Nothing);
+       register_emitter(op_Phi,               emit_Nothing);
+       register_emitter(op_Start,             emit_Nothing);
+}
+
+static void gen_binary_block(ir_node *block)
+{
+       ir_node *node;
+
+       ia32_emit_block_header(block);
+
+       /* emit the contents of the block */
+       sched_foreach(block, node) {
+               ia32_emit_node(node);
+       }
+}
+
+void ia32_gen_binary_routine(ia32_code_gen_t *ia32_cg, ir_graph *irg)
+{
+       ir_entity *entity     = get_irg_entity(irg);
+       int i, n;
+
+       cg  = ia32_cg;
+       isa = cg->isa;
+
+       ia32_register_binary_emitters();
+
+       be_gas_emit_function_prolog(entity, ia32_cg_config.function_alignment);
+
+       /* we use links to point to target blocks */
+       ir_reserve_resources(irg, IR_RESOURCE_IRN_LINK);
+       irg_block_walk_graph(irg, ia32_gen_labels, NULL, NULL);
+
+       /* initialize next block links */
+       n = ARR_LEN(cg->blk_sched);
+       for (i = 0; i < n; ++i) {
+               ir_node *block = cg->blk_sched[i];
+               ir_node *prev  = i > 0 ? cg->blk_sched[i-1] : NULL;
+
+               set_irn_link(block, prev);
+       }
+
+       for (i = 0; i < n; ++i) {
+               ir_node *block = cg->blk_sched[i];
+               gen_binary_block(block);
+       }
+
+       be_gas_emit_function_epilog(entity);
+       be_dbg_method_end();
+       be_emit_char('\n');
+       be_emit_write_line();
+
+       ir_free_resources(irg, IR_RESOURCE_IRN_LINK);
+}
+
+
+
+
+void ia32_init_emitter(void)
+{
+       lc_opt_entry_t *be_grp;
+       lc_opt_entry_t *ia32_grp;
+
+       be_grp   = lc_opt_get_grp(firm_opt_get_root(), "be");
+       ia32_grp = lc_opt_get_grp(be_grp, "ia32");
+
+       lc_opt_add_table(ia32_grp, ia32_emitter_options);
+
+       build_reg_map();
+
        FIRM_DBG_REGISTER(dbg, "firm.be.ia32.emitter");
 }