correctly implement memop handling

[libfirm] / ir / opt / loop.c

/*
 * Copyright (C) 1995-2011 University of Karlsruhe.  All right reserved.
 *
 * This file is part of libFirm.
 *
 * This file may be distributed and/or modified under the terms of the
 * GNU General Public License version 2 as published by the Free Software
 * Foundation and appearing in the file LICENSE.GPL included in the
 * packaging of this file.
 *
 * Licensees holding valid libFirm Professional Edition licenses may use
 * this file in accordance with the libFirm Commercial License.
 * Agreement provided with the Software.
 *
 * This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE
 * WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR
 * PURPOSE.
 */

/**
 * @file
 * @author   Christian Helmer
 * @brief    loop inversion and loop unrolling
 *
 */
#include "config.h"

#include <stdbool.h>

#include "iroptimize.h"
#include "opt_init.h"
#include "irnode.h"
#include "debug.h"
#include "error.h"

#include "ircons.h"
#include "irgopt.h"
#include "irgmod.h"
#include "irgwalk.h"
#include "irouts.h"
#include "iredges.h"
#include "irtools.h"
#include "array_t.h"
#include "beutil.h"
#include "irpass.h"
#include "irdom.h"
#include "opt_manage.h"

#include <math.h>
#include "irbackedge_t.h"
#include "irnodemap.h"
#include "irloop_t.h"

DEBUG_ONLY(static firm_dbg_module_t *dbg;)

/**
 * Convenience macro for iterating over every phi node of the given block.
 * Requires phi list per block.
 */
#define for_each_phi(block, phi) \
        for ((phi) = get_Block_phis( (block) ); (phi) ; (phi) = get_Phi_next((phi)))

#define for_each_phi_safe(head, phi, next) \
        for ((phi) = (head), (next) = (head) ? get_Phi_next((head)) : NULL; \
                        (phi) ; (phi) = (next), (next) = (next) ? get_Phi_next((next)) : NULL)

/* Currently processed loop. */
static ir_loop *cur_loop;

/* Flag for kind of unrolling. */
typedef enum {
        constant,
        invariant
} unrolling_kind_flag;

/* Condition for performing visiting a node during copy_walk. */
typedef bool walker_condition(const ir_node *);

/* Node and position of a predecessor. */
typedef struct entry_edge {
        ir_node *node;
        int pos;
        ir_node *pred;
} entry_edge;

/* Node info for unrolling. */
typedef struct unrolling_node_info {
        ir_node **copies;
} unrolling_node_info;

/* Outs of the nodes head. */
static entry_edge *cur_head_outs;

/* Information about the loop head */
static ir_node *loop_head       = NULL;
static bool     loop_head_valid = true;

/* List of all inner loops, that are processed. */
static ir_loop **loops;

/* Stats */
typedef struct loop_stats_t {
        unsigned loops;
        unsigned inverted;
        unsigned too_large;
        unsigned too_large_adapted;
        unsigned cc_limit_reached;
        unsigned calls_limit;

        unsigned u_simple_counting_loop;
        unsigned constant_unroll;
        unsigned invariant_unroll;

        unsigned unhandled;
} loop_stats_t;

static loop_stats_t stats;

/* Set stats to sero */
static void reset_stats(void)
{
        memset(&stats, 0, sizeof(loop_stats_t));
}

/* Print stats */
static void print_stats(void)
{
        DB((dbg, LEVEL_2, "---------------------------------------\n"));
        DB((dbg, LEVEL_2, "loops             :   %d\n",stats.loops));
        DB((dbg, LEVEL_2, "inverted          :   %d\n",stats.inverted));
        DB((dbg, LEVEL_2, "too_large         :   %d\n",stats.too_large));
        DB((dbg, LEVEL_2, "too_large_adapted :   %d\n",stats.too_large_adapted));
        DB((dbg, LEVEL_2, "cc_limit_reached  :   %d\n",stats.cc_limit_reached));
        DB((dbg, LEVEL_2, "calls_limit       :   %d\n",stats.calls_limit));
        DB((dbg, LEVEL_2, "u_simple_counting :   %d\n",stats.u_simple_counting_loop));
        DB((dbg, LEVEL_2, "constant_unroll   :   %d\n",stats.constant_unroll));
        DB((dbg, LEVEL_2, "invariant_unroll  :   %d\n",stats.invariant_unroll));
        DB((dbg, LEVEL_2, "=======================================\n"));
}

/* Commandline parameters */
typedef struct loop_opt_params_t {
unsigned max_loop_size;     /* Maximum number of nodes  [nodes]*/
int      depth_adaption;    /* Loop nest depth adaption [percent] */
unsigned allowed_calls;     /* Number of calls allowed [number] */
bool     count_phi;         /* Count phi nodes */
bool     count_proj;        /* Count projections */

unsigned max_cc_size;       /* Maximum condition chain size [nodes] */
unsigned max_branches;

unsigned max_unrolled_loop_size;    /* [nodes] */
bool     allow_const_unrolling;
bool     allow_invar_unrolling;
unsigned invar_unrolling_min_size;  /* [nodes] */

} loop_opt_params_t;

static loop_opt_params_t opt_params;

/* Loop analysis informations */
typedef struct loop_info_t {
        unsigned nodes;        /* node count */
        unsigned ld_st;        /* load and store nodes */
        unsigned branches;     /* number of conditions */
        unsigned calls;        /* number of calls */
        unsigned cf_outs;      /* number of cf edges which leave the loop */
        entry_edge cf_out;     /* single loop leaving cf edge */
        int be_src_pos;        /* position of the single own backedge in the head */

        /* for inversion */
        unsigned cc_size;      /* nodes in the condition chain */

        /* for unrolling */
        unsigned max_unroll;   /* Number of unrolls satisfying max_loop_size */
        unsigned exit_cond;    /* 1 if condition==true exits the loop.  */
        unsigned latest_value:1;    /* 1 if condition is checked against latest counter value */
        unsigned needs_backedge:1;  /* 0 if loop is completely unrolled */
        unsigned decreasing:1;      /* Step operation is_Sub, or step is<0 */

        /* IV informations of a simple loop */
        ir_node *start_val;
        ir_node *step;
        ir_node *end_val;
        ir_node *iteration_phi;
        ir_node *add;

        ir_tarval *count_tar;               /* Number of loop iterations */

        ir_node *duff_cond;                 /* Duff mod */
        unrolling_kind_flag unroll_kind;    /* constant or invariant unrolling */
} loop_info_t;

/* Information about the current loop */
static loop_info_t loop_info;

/* Outs of the condition chain (loop inversion). */
static ir_node **cc_blocks;
/* df/cf edges with def in the condition chain */
static entry_edge *cond_chain_entries;
/* Array of df loops found in the condition chain. */
static entry_edge *head_df_loop;
/* Number of blocks in cc */
static unsigned inversion_blocks_in_cc;


/* Cf/df edges leaving the loop.
 * Called entries here, as they are used to enter the loop with walkers. */
static entry_edge *loop_entries;
/* Number of unrolls to perform */
static int unroll_nr;
/* Phase is used to keep copies of nodes. */
static ir_nodemap     map;
static struct obstack obst;

/* Loop operations.  */
typedef enum loop_op_t {
        loop_op_inversion,
        loop_op_unrolling,
        loop_op_peeling
} loop_op_t;

/* Saves which loop operation to do until after basic tests. */
static loop_op_t loop_op;

/* Returns the maximum nodes for the given nest depth */
static unsigned get_max_nodes_adapted(unsigned depth)
{
        double perc = 100.0 + (double)opt_params.depth_adaption;
        double factor = pow(perc / 100.0, depth);

        return (int)((double)opt_params.max_loop_size * factor);
}

/* Reset nodes link. For use with a walker. */
static void reset_link(ir_node *node, void *env)
{
        (void)env;
        set_irn_link(node, NULL);
}

/* Returns 0 if the node or block is not in cur_loop. */
static bool is_in_loop(const ir_node *node)
{
        return get_irn_loop(get_block_const(node)) == cur_loop;
}

/* Returns 0 if the given edge is not a backedge
 * with its pred in the cur_loop. */
static bool is_own_backedge(const ir_node *n, int pos)
{
        return is_backedge(n, pos) && is_in_loop(get_irn_n(n, pos));
}

/* Finds loop head and some loop_info as calls or else if necessary. */
static void get_loop_info(ir_node *node, void *env)
{
        bool node_in_loop = is_in_loop(node);
        int i, arity;
        (void)env;

        /* collect some loop information */
        if (node_in_loop) {
                if (is_Phi(node) && opt_params.count_phi)
                        ++loop_info.nodes;
                else if (is_Proj(node) && opt_params.count_proj)
                        ++loop_info.nodes;
                else if (!is_Confirm(node) && !is_Const(node) && !is_SymConst(node))
                        ++loop_info.nodes;

                if (is_Load(node) || is_Store(node))
                        ++loop_info.ld_st;

                if (is_Call(node))
                        ++loop_info.calls;
        }

        arity = get_irn_arity(node);
        for (i = 0; i < arity; i++) {
                ir_node *pred         = get_irn_n(node, i);
                bool     pred_in_loop = is_in_loop(pred);

                if (is_Block(node) && !node_in_loop && pred_in_loop) {
                        entry_edge entry;
                        entry.node = node;
                        entry.pos = i;
                        entry.pred = pred;
                        /* Count cf outs */
                        ++loop_info.cf_outs;
                        loop_info.cf_out = entry;
                }

                /* Find the loops head/the blocks with cfpred outside of the loop */
                if (is_Block(node)) {
                        const ir_edge_t *edge;
                        unsigned outs_n = 0;

                        /* Count innerloop branches */
                        foreach_out_edge_kind(node, edge, EDGE_KIND_BLOCK) {
                                ir_node *succ = get_edge_src_irn(edge);
                                if (is_Block(succ) && is_in_loop(succ))
                                        ++outs_n;
                        }
                        if (outs_n > 1)
                                ++loop_info.branches;

                        if (node_in_loop && !pred_in_loop && loop_head_valid) {
                                ir_node *cfgpred = get_Block_cfgpred(node, i);

                                if (!is_in_loop(cfgpred)) {
                                        DB((dbg, LEVEL_5, "potential head %+F because inloop and pred %+F not inloop\n",
                                                                node, pred));
                                        /* another head? We do not touch this. */
                                        if (loop_head && loop_head != node) {
                                                loop_head_valid = false;
                                        } else {
                                                loop_head = node;
                                        }
                                }
                        }
                }
        }
}

/* Finds all edges with users outside of the loop
 * and definition inside the loop. */
static void get_loop_entries(ir_node *node, void *env)
{
        unsigned node_in_loop, pred_in_loop;
        int i, arity;
        (void) env;

        arity = get_irn_arity(node);
        for (i = 0; i < arity; ++i) {
                ir_node *pred = get_irn_n(node, i);

                pred_in_loop = is_in_loop(pred);
                node_in_loop = is_in_loop(node);

                if (pred_in_loop && !node_in_loop) {
                        entry_edge entry;
                        entry.node = node;
                        entry.pos = i;
                        entry.pred = pred;
                        ARR_APP1(entry_edge, loop_entries, entry);
                }
        }
}

/* ssa */
static ir_node *ssa_second_def;
static ir_node *ssa_second_def_block;

/**
 * Walks the graph bottom up, searching for definitions and creates phis.
 */
static ir_node *search_def_and_create_phis(ir_node *block, ir_mode *mode, int first)
{
        int i;
        int n_cfgpreds;
        ir_graph *irg = get_irn_irg(block);
        ir_node *phi;
        ir_node **in;

        DB((dbg, LEVEL_5, "ssa search_def_and_create_phis: block %N\n", block));

        /* Prevents creation of phi that would be bad anyway.
         * Dead and bad blocks. */
        if (get_irn_arity(block) < 1 || is_Bad(block)) {
                DB((dbg, LEVEL_5, "ssa bad %N\n", block));
                return new_r_Bad(irg, mode);
        }

        if (block == ssa_second_def_block && !first) {
                DB((dbg, LEVEL_5, "ssa found second definition: use second def %N\n", ssa_second_def));
                return ssa_second_def;
        }

        /* already processed this block? */
        if (irn_visited(block)) {
                ir_node *value = (ir_node *) get_irn_link(block);
                DB((dbg, LEVEL_5, "ssa already visited: use linked %N\n", value));
                return value;
        }

        assert(block != get_irg_start_block(irg));

        /* a Block with only 1 predecessor needs no Phi */
        n_cfgpreds = get_Block_n_cfgpreds(block);
        if (n_cfgpreds == 1) {
                ir_node *pred_block = get_Block_cfgpred_block(block, 0);
                ir_node *value;

                DB((dbg, LEVEL_5, "ssa 1 pred: walk pred %N\n", pred_block));

                value = search_def_and_create_phis(pred_block, mode, 0);
                set_irn_link(block, value);
                mark_irn_visited(block);

                return value;
        }

        /* create a new Phi */
        NEW_ARR_A(ir_node*, in, n_cfgpreds);
        for (i = 0; i < n_cfgpreds; ++i)
                in[i] = new_r_Dummy(irg, mode);

        phi = new_r_Phi(block, n_cfgpreds, in, mode);
        /* Important: always keep block phi list up to date. */
        add_Block_phi(block, phi);
        DB((dbg, LEVEL_5, "ssa phi creation: link new phi %N to block %N\n", phi, block));
        set_irn_link(block, phi);
        mark_irn_visited(block);

        /* set Phi predecessors */
        for (i = 0; i < n_cfgpreds; ++i) {
                ir_node *pred_val;
                ir_node *pred_block = get_Block_cfgpred_block(block, i);
                assert(pred_block != NULL);
                pred_val = search_def_and_create_phis(pred_block, mode, 0);

                assert(pred_val != NULL);

                DB((dbg, LEVEL_5, "ssa phi pred:phi %N, pred %N\n", phi, pred_val));
                set_irn_n(phi, i, pred_val);
        }

        return phi;
}


/**
 * Given a set of values this function constructs SSA-form for the users of the
 * first value (the users are determined through the out-edges of the value).
 * Works without using the dominance tree.
 */
static void construct_ssa(ir_node *orig_block, ir_node *orig_val,
                ir_node *second_block, ir_node *second_val)
{
        ir_graph *irg;
        ir_mode *mode;
        const ir_edge_t *edge;
        const ir_edge_t *next;

        assert(orig_block && orig_val && second_block && second_val &&
                        "no parameter of construct_ssa may be NULL");

        if (orig_val == second_val)
                return;

        irg = get_irn_irg(orig_val);

        ir_reserve_resources(irg, IR_RESOURCE_IRN_VISITED);
        inc_irg_visited(irg);

        mode = get_irn_mode(orig_val);
        set_irn_link(orig_block, orig_val);
        mark_irn_visited(orig_block);

        ssa_second_def_block = second_block;
        ssa_second_def       = second_val;

        /* Only fix the users of the first, i.e. the original node */
        foreach_out_edge_safe(orig_val, edge, next) {
                ir_node *user = get_edge_src_irn(edge);
                int j = get_edge_src_pos(edge);
                ir_node *user_block = get_nodes_block(user);
                ir_node *newval;

                /* ignore keeps */
                if (is_End(user))
                        continue;

                DB((dbg, LEVEL_5, "original user %N\n", user));

                if (is_Phi(user)) {
                        ir_node *pred_block = get_Block_cfgpred_block(user_block, j);
                        newval = search_def_and_create_phis(pred_block, mode, 1);
                } else {
                        newval = search_def_and_create_phis(user_block, mode, 1);
                }
                if (newval != user && !is_Bad(newval))
                        set_irn_n(user, j, newval);
        }

        ir_free_resources(irg, IR_RESOURCE_IRN_VISITED);
}


/***** Unrolling Helper Functions *****/

/* Assign the copy with index nr to node n */
static void set_unroll_copy(ir_node *n, int nr, ir_node *cp)
{
        unrolling_node_info *info;
        assert(nr != 0 && "0 reserved");

        info = (unrolling_node_info*)ir_nodemap_get(&map, n);
        if (! info) {
                ir_node **arr = NEW_ARR_D(ir_node*, &obst, unroll_nr);
                memset(arr, 0, unroll_nr * sizeof(ir_node*));

                info = OALLOCZ(&obst, unrolling_node_info);
                info->copies = arr;
                ir_nodemap_insert(&map, n, info);
        }
        /* Original node */
        info->copies[0] = n;

        info->copies[nr] = cp;
}

/* Returns a nodes copy if it exists, else NULL. */
static ir_node *get_unroll_copy(ir_node *n, int nr)
{
        ir_node             *cp;
        unrolling_node_info *info = (unrolling_node_info *)ir_nodemap_get(&map, n);
        if (! info)
                return NULL;

        cp = info->copies[nr];
        return cp;
}


/***** Inversion Helper Functions *****/

/* Sets copy cp of node n. */
static void set_inversion_copy(ir_node *n, ir_node *cp)
{
        ir_nodemap_insert(&map, n, cp);
}

/* Getter of copy of n for inversion */
static ir_node *get_inversion_copy(ir_node *n)
{
        ir_node *cp = (ir_node *)ir_nodemap_get(&map, n);
        return cp;
}

/* Resets block mark for given node. For use with walker */
static void reset_block_mark(ir_node *node, void * env)
{
        (void) env;

        if (is_Block(node))
                set_Block_mark(node, 0);
}

/* Returns mark of node, or its block if node is not a block.
 * Used in this context to determine if node is in the condition chain. */
static bool is_nodes_block_marked(const ir_node* node)
{
        return get_Block_mark(get_block_const(node));
}

/* Extends a nodes ins by node new.
 * NOTE: This is slow if a node n needs to be extended more than once. */
static void extend_irn(ir_node *n, ir_node *newnode, bool new_is_backedge)
{
        int i;
        int arity = get_irn_arity(n);
        int new_arity = arity + 1;
        ir_node **ins = XMALLOCN(ir_node*, new_arity);
        bool     *bes = XMALLOCN(bool, new_arity);

        /* save bes */
        /* Bes are important!
         * Another way would be recreating the looptree,
         * but after that we cannot distinguish already processed loops
         * from not yet processed ones. */
        if (is_Block(n)) {
                for(i = 0; i < arity; ++i) {
                        bes[i] = is_backedge(n, i);
                }
                bes[i] = new_is_backedge;
        }

        for(i = 0; i < arity; ++i) {
                ins[i] = get_irn_n(n, i);
        }
        ins[i] = newnode;

        set_irn_in(n, new_arity, ins);

        /* restore bes  */
        if (is_Block(n)) {
                for(i = 0; i < new_arity; ++i) {
                        if (bes[i])
                                set_backedge(n, i);
                }
        }
}

/* Extends a block by a copy of its pred at pos,
 * fixing also the phis in the same way. */
static void extend_ins_by_copy(ir_node *block, int pos)
{
        ir_node *new_in;
        ir_node *phi;
        ir_node *pred;
        assert(is_Block(block));

        /* Extend block by copy of definition at pos */
        pred = get_irn_n(block, pos);
        new_in = get_inversion_copy(pred);
        DB((dbg, LEVEL_5, "Extend block %N by %N cp of %N\n", block, new_in, pred));
        extend_irn(block, new_in, false);

        /* Extend block phis by copy of definition at pos */
        for_each_phi(block, phi) {
                ir_node *pred, *cp;

                pred = get_irn_n(phi, pos);
                cp = get_inversion_copy(pred);
                /* If the phis in is not in the condition chain (eg. a constant),
                 * there is no copy. */
                if (cp == NULL)
                        new_in = pred;
                else
                        new_in = cp;

                DB((dbg, LEVEL_5, "Extend phi %N by %N cp of %N\n", phi, new_in, pred));
                extend_irn(phi, new_in, false);
        }
}

/* Returns the number of blocks backedges. With or without alien bes. */
static int get_backedge_n(ir_node *block, bool with_alien)
{
        int i;
        int be_n = 0;
        int arity = get_irn_arity(block);

        assert(is_Block(block));

        for (i = 0; i < arity; ++i) {
                ir_node *pred = get_irn_n(block, i);
                if (is_backedge(block, i) && (with_alien || is_in_loop(pred)))
                        ++be_n;
        }
        return be_n;
}

/* Returns a raw copy of the given node.
 * Attributes are kept/set according to the needs of loop inversion. */
static ir_node *copy_node(ir_node *node)
{
        int i, arity;
        ir_node *cp;

        cp = exact_copy(node);
        arity = get_irn_arity(node);

        /* Keep backedge info */
        for (i = 0; i < arity; ++i) {
                if (is_backedge(node, i))
                        set_backedge(cp, i);
        }

        if (is_Block(cp)) {
                set_Block_mark(cp, 0);
        }

        return cp;
}


/**
 * This walker copies all walked nodes.
 * If the walk_condition is true for a node, it is copied.
 * All nodes node_info->copy have to be NULL prior to every walk.
 * Order of ins is important for later usage.
 */
static void copy_walk(ir_node *node, walker_condition *walk_condition,
                      ir_loop *set_loop)
{
        int i;
        int arity;
        ir_node *cp;
        ir_node **cpin;
        ir_graph *irg = current_ir_graph;

        /**
         * break condition and cycle resolver, creating temporary node copies
         */
        if (get_irn_visited(node) >= get_irg_visited(irg)) {
                /* Here we rely on nodestate's copy being initialized with NULL */
                DB((dbg, LEVEL_5, "copy_walk: We have already visited %N\n", node));
                if (get_inversion_copy(node) == NULL) {
                        cp = copy_node(node);
                        set_inversion_copy(node, cp);

                        DB((dbg, LEVEL_5, "The TEMP copy of %N is created %N\n", node, cp));
                }
                return;
        }

        /* Walk */
        mark_irn_visited(node);

        if (!is_Block(node)) {
                ir_node *pred = get_nodes_block(node);
                if (walk_condition(pred))
                        DB((dbg, LEVEL_5, "walk block %N\n", pred));
                copy_walk(pred, walk_condition, set_loop);
        }

        arity = get_irn_arity(node);

        NEW_ARR_A(ir_node *, cpin, arity);

        for (i = 0; i < arity; ++i) {
                ir_node *pred = get_irn_n(node, i);

                if (walk_condition(pred)) {
                        DB((dbg, LEVEL_5, "walk node %N\n", pred));
                        copy_walk(pred, walk_condition, set_loop);
                        cpin[i] = get_inversion_copy(pred);
                        DB((dbg, LEVEL_5, "copy of %N gets new in %N which is copy of %N\n",
                                                node, get_inversion_copy(pred), pred));
                } else {
                        cpin[i] = pred;
                }
        }

        /* copy node / finalize temp node */
        if (get_inversion_copy(node) == NULL) {
                /* No temporary copy existent */
                cp = copy_node(node);
                set_inversion_copy(node, cp);
                DB((dbg, LEVEL_5, "The FINAL copy of %N is CREATED %N\n", node, cp));
        } else {
                /* temporary copy is existent but without correct ins */
                cp = get_inversion_copy(node);
                DB((dbg, LEVEL_5, "The FINAL copy of %N is EXISTENT %N\n", node, cp));
        }

        if (!is_Block(node)) {
                ir_node *cpblock = get_inversion_copy(get_nodes_block(node));

                set_nodes_block(cp, cpblock );
                if (is_Phi(cp))
                        add_Block_phi(cpblock, cp);
        }

        /* Keeps phi list of temporary node. */
        set_irn_in(cp, ARR_LEN(cpin), cpin);
}

/**
 * This walker copies all walked nodes.
 * If the walk_condition is true for a node, it is copied.
 * All nodes node_info->copy have to be NULL prior to every walk.
 * Order of ins is important for later usage.
 * Takes copy_index, to phase-link copy at specific index.
 */
static void copy_walk_n(ir_node *node, walker_condition *walk_condition,
                        int copy_index)
{
        int i;
        int arity;
        ir_node *cp;
        ir_node **cpin;

        /**
         * break condition and cycle resolver, creating temporary node copies
         */
        if (irn_visited(node)) {
                /* Here we rely on nodestate's copy being initialized with NULL */
                DB((dbg, LEVEL_5, "copy_walk: We have already visited %N\n", node));
                if (get_unroll_copy(node, copy_index) == NULL) {
                        ir_node *u;
                        u = copy_node(node);
                        set_unroll_copy(node, copy_index, u);
                        DB((dbg, LEVEL_5, "The TEMP unknown of %N is created %N\n", node, u));
                }
                return;
        }

        /* Walk */
        mark_irn_visited(node);

        if (!is_Block(node)) {
                ir_node *block = get_nodes_block(node);
                if (walk_condition(block))
                        DB((dbg, LEVEL_5, "walk block %N\n", block));
                copy_walk_n(block, walk_condition, copy_index);
        }

        arity = get_irn_arity(node);
        NEW_ARR_A(ir_node *, cpin, arity);

        for (i = 0; i < arity; ++i) {
                ir_node *pred = get_irn_n(node, i);

                if (walk_condition(pred)) {
                        DB((dbg, LEVEL_5, "walk node %N\n", pred));
                        copy_walk_n(pred, walk_condition, copy_index);
                        cpin[i] = get_unroll_copy(pred, copy_index);
                } else {
                        cpin[i] = pred;
                }
        }

        /* copy node / finalize temp node */
        cp = get_unroll_copy(node, copy_index);
        if (cp == NULL || is_Unknown(cp)) {
                cp = copy_node(node);
                set_unroll_copy(node, copy_index, cp);
                DB((dbg, LEVEL_5, "The FINAL copy of %N is CREATED %N\n", node, cp));
        } else {
                /* temporary copy is existent but without correct ins */
                cp = get_unroll_copy(node, copy_index);
                DB((dbg, LEVEL_5, "The FINAL copy of %N is EXISTENT %N\n", node, cp));
        }

        if (!is_Block(node)) {
                ir_node *cpblock = get_unroll_copy(get_nodes_block(node), copy_index);

                set_nodes_block(cp, cpblock );
                if (is_Phi(cp))
                        add_Block_phi(cpblock, cp);
        }

        /* Keeps phi list of temporary node. */
        set_irn_in(cp, ARR_LEN(cpin), cpin);
}

/* Removes alle Blocks with non marked predecessors from the condition chain. */
static void unmark_not_allowed_cc_blocks(void)
{
        size_t blocks = ARR_LEN(cc_blocks);
        size_t i;

        for(i = 0; i < blocks; ++i) {
                ir_node *block = cc_blocks[i];
                int a;
                int arity = get_irn_arity(block);

                /* Head is an exception. */
                if (block == loop_head)
                        continue;

                for(a = 0; a < arity; ++a) {
                        if (! is_nodes_block_marked(get_irn_n(block, a))) {
                                set_Block_mark(block, 0);
                                --inversion_blocks_in_cc;
                                DB((dbg, LEVEL_5, "Removed %N from cc (blocks in cc %d)\n",
                                                block, inversion_blocks_in_cc));

                                break;
                        }
                }
        }
}

/* Unmarks all cc blocks using cc_blocks except head.
 * TODO: invert head for unrolling? */
static void unmark_cc_blocks(void)
{
        size_t blocks = ARR_LEN(cc_blocks);
        size_t i;

        for(i = 0; i < blocks; ++i) {
                ir_node *block = cc_blocks[i];

                /* TODO Head is an exception. */
                /*if (block != loop_head)*/
                set_Block_mark(block, 0);
        }
        /*inversion_blocks_in_cc = 1;*/
        inversion_blocks_in_cc = 0;

        /* invalidate */
        loop_info.cc_size = 0;
}

/**
 * Populates head_entries with (node, pred_pos) tuple
 * whereas the node's pred at pred_pos is in the cc but not the node itself.
 * Also finds df loops inside the cc.
 * Head and condition chain blocks have been marked previously.
 */
static void get_head_outs(ir_node *node, void *env)
{
        int i;
        int arity = get_irn_arity(node);
        (void) env;

        for (i = 0; i < arity; ++i) {
                if (!is_nodes_block_marked(node) && is_nodes_block_marked(get_irn_n(node, i))) {
                        entry_edge entry;
                        entry.node = node;
                        entry.pos = i;
                        /* Saving also predecessor seems redundant, but becomes
                         * necessary when changing position of it, before
                         * dereferencing it.*/
                        entry.pred = get_irn_n(node, i);
                        ARR_APP1(entry_edge, cur_head_outs, entry);
                }
        }

        arity = get_irn_arity(loop_head);

        /* Find df loops inside the cc */
        if (is_Phi(node) && get_nodes_block(node) == loop_head) {
                for (i = 0; i < arity; ++i) {
                        if (is_own_backedge(loop_head, i)) {
                                if (is_nodes_block_marked(get_irn_n(node, i))) {
                                        entry_edge entry;
                                        entry.node = node;
                                        entry.pos = i;
                                        entry.pred = get_irn_n(node, i);
                                        ARR_APP1(entry_edge, head_df_loop, entry);
                                        DB((dbg, LEVEL_5, "Found incc assignment node %N @%d is pred %N, graph %N %N\n",
                                                        node, i, entry.pred, current_ir_graph, get_irg_start_block(current_ir_graph)));
                                }
                        }
                }
        }
}

/**
 * Find condition chains, and add them to be inverted.
 * A block belongs to the chain if a condition branches out of the loop.
 * (Some blocks need to be removed once again.)
 * Returns 1 if the given block belongs to the condition chain.
 */
static void find_condition_chain(ir_node *block)
{
        const    ir_edge_t *edge;
        bool     mark     = false;
        bool     has_be   = false;
        bool     jmp_only = true;
        unsigned nodes_n  = 0;

        mark_irn_visited(block);

        DB((dbg, LEVEL_5, "condition_chains for block %N\n", block));

        /* Get node count */
        foreach_out_edge_kind(block, edge, EDGE_KIND_NORMAL) {
                ++nodes_n;
        }

        /* Check if node count would exceed maximum cc size.
         * TODO
         * This is not optimal, as we search depth-first and break here,
         * continuing with another subtree. */
        if (loop_info.cc_size + nodes_n > opt_params.max_cc_size) {
                set_Block_mark(block, 0);
                return;
        }

        /* Check if block only has a jmp instruction. */
        foreach_out_edge(block, edge) {
                ir_node *src = get_edge_src_irn(edge);

                if (!is_Block(src) && !is_Jmp(src)) {
                        jmp_only = false;
                }
        }

        /* Check cf outs if one is leaving the loop,
         * or if this node has a backedge. */
        foreach_block_succ(block, edge) {
                ir_node *src = get_edge_src_irn(edge);
                int pos = get_edge_src_pos(edge);

                if (!is_in_loop(src))
                        mark = true;

                /* Inverting blocks with backedge outs leads to a cf edge
                 * from the inverted head, into the inverted head (skipping the body).
                 * As the body becomes the new loop head,
                 * this would introduce another loop in the existing loop.
                 * This loop inversion cannot cope with this case. */
                if (is_backedge(src, pos)) {
                        has_be = true;
                        break;
                }
        }

        /* We need all predecessors to already belong to the condition chain.
         * Example of wrong case:  * == in cc
         *
         *     Head*             ,--.
         *    /|   \            B   |
         *   / A*  B           /    |
         *  / /\   /          ?     |
         *   /   C*      =>      D  |
         *      /  D           Head |
         *     /               A  \_|
         *                      C
         */
        /* Collect blocks containing only a Jmp.
         * Do not collect blocks with backedge outs. */
        if ((jmp_only || mark) && !has_be) {
                set_Block_mark(block, 1);
                ++inversion_blocks_in_cc;
                loop_info.cc_size += nodes_n;
                DB((dbg, LEVEL_5, "block %N is part of condition chain\n", block));
                ARR_APP1(ir_node *, cc_blocks, block);
        } else {
                set_Block_mark(block, 0);
        }

        foreach_block_succ(block, edge) {
                ir_node *src = get_edge_src_irn( edge );

                if (is_in_loop(src) && ! irn_visited(src))
                        find_condition_chain(src);
        }
}

/**
 * Rewires the copied condition chain. Removes backedges
 * as this condition chain is prior to the loop.
 * Copy of loop_head must have phi list and old (unfixed) backedge info of the loop head.
 * (loop_head is already fixed, we cannot rely on it.)
 */
static void fix_copy_inversion(void)
{
        ir_node *new_head;
        ir_node **ins;
        ir_node **phis;
        ir_node *phi, *next;
        ir_node *head_cp = get_inversion_copy(loop_head);
        ir_graph *irg    = get_irn_irg(head_cp);
        int arity        = get_irn_arity(head_cp);
        int backedges    = get_backedge_n(head_cp, false);
        int new_arity    = arity - backedges;
        int pos;
        int i;

        NEW_ARR_A(ir_node *, ins, new_arity);

        pos = 0;
        /* Remove block backedges */
        for(i = 0; i < arity; ++i) {
                if (!is_backedge(head_cp, i))
                        ins[pos++] = get_irn_n(head_cp, i);
        }

        new_head = new_r_Block(irg, new_arity, ins);

        phis = NEW_ARR_F(ir_node *, 0);

        for_each_phi_safe(get_Block_phis(head_cp), phi, next) {
                ir_node *new_phi;
                NEW_ARR_A(ir_node *, ins, new_arity);
                pos = 0;
                for(i = 0; i < arity; ++i) {
                        if (!is_backedge(head_cp, i))
                                ins[pos++] = get_irn_n(phi, i);
                }
                new_phi = new_rd_Phi(get_irn_dbg_info(phi),
                                new_head, new_arity, ins,
                                get_irn_mode(phi));
                ARR_APP1(ir_node *, phis, new_phi);
        }

        pos = 0;
        for_each_phi_safe(get_Block_phis(head_cp), phi, next) {
                exchange(phi, phis[pos++]);
        }

        exchange(head_cp, new_head);

        DEL_ARR_F(phis);
}


/* Puts the original condition chain at the end of the loop,
 * subsequently to the body.
 * Relies on block phi list and correct backedges.
 */
static void fix_head_inversion(void)
{
        ir_node *new_head;
        ir_node **ins;
        ir_node *phi, *next;
        ir_node **phis;
        ir_graph *irg = get_irn_irg(loop_head);
        int arity     = get_irn_arity(loop_head);
        int backedges = get_backedge_n(loop_head, false);
        int new_arity = backedges;
        int pos;
        int i;

        NEW_ARR_A(ir_node *, ins, new_arity);

        pos = 0;
        /* Keep only backedges */
        for(i = 0; i < arity; ++i) {
                if (is_own_backedge(loop_head, i))
                        ins[pos++] = get_irn_n(loop_head, i);
        }

        new_head = new_r_Block(irg, new_arity, ins);

        phis = NEW_ARR_F(ir_node *, 0);

        for_each_phi(loop_head, phi) {
                ir_node *new_phi;
                DB((dbg, LEVEL_5, "Fixing phi %N of loop head\n", phi));

                NEW_ARR_A(ir_node *, ins, new_arity);

                pos = 0;
                for (i = 0; i < arity; ++i) {
                        ir_node *pred = get_irn_n(phi, i);

                        if (is_own_backedge(loop_head, i)) {
                                /* If assignment is in the condition chain,
                                 * we need to create a phi in the new loop head.
                                 * This can only happen for df, not cf. See find_condition_chains. */
                                /*if (is_nodes_block_marked(pred)) {
                                        ins[pos++] = pred;
                                } else {*/
                                ins[pos++] = pred;

                        }
                }

                new_phi = new_rd_Phi(get_irn_dbg_info(phi),
                        new_head, new_arity, ins,
                        get_irn_mode(phi));

                ARR_APP1(ir_node *, phis, new_phi);

                DB((dbg, LEVEL_5, "fix inverted head should exch %N by %N (pos %d)\n", phi, new_phi, pos ));
        }

        pos = 0;
        for_each_phi_safe(get_Block_phis(loop_head), phi, next) {
                DB((dbg, LEVEL_5, "fix inverted exch phi %N by %N\n", phi, phis[pos]));
                if (phis[pos] != phi)
                        exchange(phi, phis[pos++]);
        }

        DEL_ARR_F(phis);

        DB((dbg, LEVEL_5, "fix inverted head exch head block %N by %N\n", loop_head, new_head));
        exchange(loop_head, new_head);
}

/* Does the loop inversion.  */
static void inversion_walk(ir_graph *irg, entry_edge *head_entries)
{
        size_t i;

        /*
         * The order of rewiring bottom-up is crucial.
         * Any change of the order leads to lost information that would be needed later.
         */

        ir_reserve_resources(irg, IR_RESOURCE_IRN_VISITED);

        /* 1. clone condition chain */
        inc_irg_visited(irg);

        for (i = 0; i < ARR_LEN(head_entries); ++i) {
                entry_edge entry = head_entries[i];
                ir_node *pred = get_irn_n(entry.node, entry.pos);

                DB((dbg, LEVEL_5, "\nInit walk block %N\n", pred));

                copy_walk(pred, is_nodes_block_marked, cur_loop);
        }

        ir_free_resources(irg, IR_RESOURCE_IRN_VISITED);

        /* 2. Extends the head control flow successors ins
         *    with the definitions of the copied head node. */
        for (i = 0; i < ARR_LEN(head_entries); ++i) {
                entry_edge head_out = head_entries[i];

                if (is_Block(head_out.node))
                        extend_ins_by_copy(head_out.node, head_out.pos);
        }

        /* 3. construct_ssa for users of definitions in the condition chain,
         *    as there is now a second definition. */
        for (i = 0; i < ARR_LEN(head_entries); ++i) {
                entry_edge head_out = head_entries[i];

                /* Ignore keepalives */
                if (is_End(head_out.node))
                        continue;

                /* Construct ssa for assignments in the condition chain. */
                if (!is_Block(head_out.node)) {
                        ir_node *pred, *cppred, *block, *cpblock;

                        pred = head_out.pred;
                        cppred = get_inversion_copy(pred);
                        block = get_nodes_block(pred);
                        cpblock = get_nodes_block(cppred);
                        construct_ssa(block, pred, cpblock, cppred);
                }
        }

        /*
         * If there is an assignment in the condition chain
         * with a user also in the condition chain,
         * the dominance frontier is in the new loop head.
         * The dataflow loop is completely in the condition chain.
         * Goal:
         *  To be wired: >|
         *
         *  | ,--.   |
         * Phi_cp |  | copied condition chain
         * >| |   |  |
         * >| ?__/   |
         * >| ,-.
         *  Phi* |   | new loop head with newly created phi.
         *   |   |
         *  Phi  |   | original, inverted condition chain
         *   |   |   |
         *   ?__/    |
         *
         */
        for (i = 0; i < ARR_LEN(head_df_loop); ++i) {
                entry_edge head_out = head_df_loop[i];

                /* Construct ssa for assignments in the condition chain. */
                ir_node *pred, *cppred, *block, *cpblock;

                pred = head_out.pred;
                cppred = get_inversion_copy(pred);
                assert(cppred && pred);
                block = get_nodes_block(pred);
                cpblock = get_nodes_block(cppred);
                construct_ssa(block, pred, cpblock, cppred);
        }

        /* 4. Remove the ins which are no backedges from the original condition chain
         *    as the cc is now subsequent to the body. */
        fix_head_inversion();

        /* 5. Remove the backedges of the copied condition chain,
         *    because it is going to be the new 'head' in advance to the loop. */
        fix_copy_inversion();

}

/* Performs loop inversion of cur_loop if possible and reasonable. */
static void loop_inversion(ir_graph *irg)
{
        int      loop_depth;
        unsigned max_loop_nodes = opt_params.max_loop_size;
        unsigned max_loop_nodes_adapted;
        int      depth_adaption = opt_params.depth_adaption;

        bool do_inversion = true;

        /* Depth of 0 is the procedure and 1 a topmost loop. */
        loop_depth = get_loop_depth(cur_loop) - 1;

        /* Calculating in per mil. */
        max_loop_nodes_adapted = get_max_nodes_adapted(loop_depth);

        DB((dbg, LEVEL_1, "max_nodes: %d\nmax_nodes_adapted %d at depth of %d (adaption %d)\n",
                        max_loop_nodes, max_loop_nodes_adapted, loop_depth, depth_adaption));

        if (loop_info.nodes == 0)
                return;

        if (loop_info.nodes > max_loop_nodes) {
                /* Only for stats */
                DB((dbg, LEVEL_1, "Nodes %d > allowed nodes %d\n",
                        loop_info.nodes, loop_depth, max_loop_nodes));
                ++stats.too_large;
                /* no RETURN */
                /* Adaption might change it */
        }

        /* Limit processing to loops smaller than given parameter. */
        if (loop_info.nodes > max_loop_nodes_adapted) {
                DB((dbg, LEVEL_1, "Nodes %d > allowed nodes (depth %d adapted) %d\n",
                        loop_info.nodes, loop_depth, max_loop_nodes_adapted));
                ++stats.too_large_adapted;
                return;
        }

        if (loop_info.calls > opt_params.allowed_calls) {
                DB((dbg, LEVEL_1, "Calls %d > allowed calls %d\n",
                        loop_info.calls, opt_params.allowed_calls));
                ++stats.calls_limit;
                return;
        }

        /*inversion_head_node_limit = INT_MAX;*/
        ir_reserve_resources(irg, IR_RESOURCE_BLOCK_MARK);

        /* Reset block marks.
         * We use block marks to flag blocks of the original condition chain. */
        irg_walk_graph(irg, reset_block_mark, NULL, NULL);

        /*loop_info.blocks = get_loop_n_blocks(cur_loop);*/
        cond_chain_entries = NEW_ARR_F(entry_edge, 0);
        head_df_loop = NEW_ARR_F(entry_edge, 0);

        /*head_inversion_node_count = 0;*/
        inversion_blocks_in_cc = 0;

        /* Use phase to keep copy of nodes from the condition chain. */
        ir_nodemap_init(&map, irg);
        obstack_init(&obst);

        /* Search for condition chains and temporarily save the blocks in an array. */
        cc_blocks = NEW_ARR_F(ir_node *, 0);
        inc_irg_visited(irg);
        find_condition_chain(loop_head);

        unmark_not_allowed_cc_blocks();
        DEL_ARR_F(cc_blocks);

        /* Condition chain too large.
         * Loop should better be small enough to fit into the cache. */
        /* TODO Of course, we should take a small enough cc in the first place,
         * which is not that simple. (bin packing)  */
        if (loop_info.cc_size > opt_params.max_cc_size) {
                ++stats.cc_limit_reached;

                do_inversion = false;

                /* Unmark cc blocks except the head.
                 * Invert head only for possible unrolling. */
                unmark_cc_blocks();
        }

        /* We also catch endless loops here,
         * because they do not have a condition chain. */
        if (inversion_blocks_in_cc < 1) {
                do_inversion = false;
                DB((dbg, LEVEL_3,
                        "Loop contains %d (less than 1) invertible blocks => No Inversion done.\n",
                        inversion_blocks_in_cc));
        }

        if (do_inversion) {
                cur_head_outs = NEW_ARR_F(entry_edge, 0);

                /* Get all edges pointing into the condition chain. */
                irg_walk_graph(irg, get_head_outs, NULL, NULL);

                /* Do the inversion */
                inversion_walk(irg, cur_head_outs);

                DEL_ARR_F(cur_head_outs);

                /* Duplicated blocks changed doms */
                clear_irg_state(irg, IR_GRAPH_STATE_CONSISTENT_DOMINANCE
                                   | IR_GRAPH_STATE_CONSISTENT_LOOPINFO);

                ++stats.inverted;
        }

        /* free */
        obstack_free(&obst, NULL);
        ir_nodemap_destroy(&map);
        DEL_ARR_F(cond_chain_entries);
        DEL_ARR_F(head_df_loop);

        ir_free_resources(irg, IR_RESOURCE_BLOCK_MARK);
}

/* Fix the original loop_heads ins for invariant unrolling case. */
static void unrolling_fix_loop_head_inv(void)
{
        ir_node *ins[2];
        ir_node *phi;
        ir_node *proj = new_Proj(loop_info.duff_cond, mode_X, 0);
        ir_node *head_pred = get_irn_n(loop_head, loop_info.be_src_pos);
        ir_node *loop_condition = get_unroll_copy(head_pred, unroll_nr - 1);

        /* Original loop_heads ins are:
         * duff block and the own backedge */

        ins[0] = loop_condition;
        ins[1] = proj;
        set_irn_in(loop_head, 2, ins);
        DB((dbg, LEVEL_4, "Rewire ins of block loophead %N to pred %N and duffs entry %N \n" , loop_head, ins[0], ins[1]));

        for_each_phi(loop_head, phi) {
                ir_node *pred = get_irn_n(phi, loop_info.be_src_pos);
                /* TODO we think it is a phi, but for Mergesort it is not the case.*/

                ir_node *last_pred = get_unroll_copy(pred, unroll_nr - 1);

                ins[0] = last_pred;
                ins[1] = (ir_node*)get_irn_link(phi);
                set_irn_in(phi, 2, ins);
                DB((dbg, LEVEL_4, "Rewire ins of loophead phi %N to pred %N and duffs entry %N \n" , phi, ins[0], ins[1]));
        }
}

/* Removes previously created phis with only 1 in. */
static void correct_phis(ir_node *node, void *env)
{
        (void)env;

        if (is_Phi(node) && get_irn_arity(node) == 1) {
                ir_node *exch;
                ir_node *in[1];

                in[0] = get_irn_n(node, 0);

                exch = new_rd_Phi(get_irn_dbg_info(node),
                    get_nodes_block(node), 1, in,
                        get_irn_mode(node));

                exchange(node, exch);
        }
}

/* Unrolling: Rewire floating copies. */
static void place_copies(int copies)
{
        ir_node *loophead = loop_head;
        size_t i;
        int c;
        int be_src_pos = loop_info.be_src_pos;

        /* Serialize loops by fixing their head ins.
         * Processed are the copies.
         * The original loop is done after that, to keep backedge infos. */
        for (c = 0; c < copies; ++c) {
                ir_node *upper = get_unroll_copy(loophead, c);
                ir_node *lower = get_unroll_copy(loophead, c + 1);
                ir_node *phi;
                ir_node *topmost_be_block = get_nodes_block(get_irn_n(loophead, be_src_pos));

                /* Important: get the preds first and then their copy. */
                ir_node *upper_be_block = get_unroll_copy(topmost_be_block, c);
                ir_node *new_jmp = new_r_Jmp(upper_be_block);
                DB((dbg, LEVEL_5, " place_copies upper %N lower %N\n", upper, lower));

                DB((dbg, LEVEL_5, "topmost be block %N \n", topmost_be_block));

                if (loop_info.unroll_kind == constant) {
                        ir_node *ins[1];
                        ins[0] = new_jmp;
                        set_irn_in(lower, 1, ins);

                        for_each_phi(loophead, phi) {
                                ir_node *topmost_def = get_irn_n(phi, be_src_pos);
                                ir_node *upper_def = get_unroll_copy(topmost_def, c);
                                ir_node *lower_phi = get_unroll_copy(phi, c + 1);

                                /* It is possible, that the value used
                                 * in the OWN backedge path is NOT defined in this loop. */
                                if (is_in_loop(topmost_def))
                                        ins[0] = upper_def;
                                else
                                        ins[0] = topmost_def;

                                set_irn_in(lower_phi, 1, ins);
                                /* Need to replace phis with 1 in later. */
                        }
                } else {
                        /* Invariant case */
                        /* Every node has 2 ins. One from the duff blocks
                         * and one from the previously unrolled loop. */
                        ir_node *ins[2];
                        /* Calculate corresponding projection of mod result for this copy c */
                        ir_node *proj = new_Proj(loop_info.duff_cond, mode_X, unroll_nr - c - 1);
                        DB((dbg, LEVEL_4, "New duff proj %N\n" , proj));

                        ins[0] = new_jmp;
                        ins[1] = proj;
                        set_irn_in(lower, 2, ins);
                        DB((dbg, LEVEL_4, "Rewire ins of Block %N to pred %N and duffs entry %N \n" , lower, ins[0], ins[1]));

                        for_each_phi(loophead, phi) {
                                ir_node *topmost_phi_pred = get_irn_n(phi, be_src_pos);
                                ir_node *upper_phi_pred;
                                ir_node *lower_phi;
                                ir_node *duff_phi;

                                lower_phi = get_unroll_copy(phi, c + 1);
                                duff_phi = (ir_node*)get_irn_link(phi);
                                DB((dbg, LEVEL_4, "DD Link of %N is %N\n" , phi, duff_phi));

                                /*  */
                                if (is_in_loop(topmost_phi_pred)) {
                                        upper_phi_pred = get_unroll_copy(topmost_phi_pred, c);
                                } else {
                                        upper_phi_pred = topmost_phi_pred;
                                }

                                ins[0] = upper_phi_pred;
                                ins[1] = duff_phi;
                                set_irn_in(lower_phi, 2, ins);
                                DB((dbg, LEVEL_4, "Rewire ins of %N to pred %N and duffs entry %N \n" , lower_phi, ins[0], ins[1]));
                        }
                }
        }

        /* Reconnect last copy. */
        for (i = 0; i < ARR_LEN(loop_entries); ++i) {
                entry_edge edge = loop_entries[i];
                /* Last copy is at the bottom */
                ir_node *new_pred = get_unroll_copy(edge.pred, copies);
                set_irn_n(edge.node, edge.pos, new_pred);
        }

        /* Fix original loops head.
         * Done in the end, as ins and be info were needed before. */
        if (loop_info.unroll_kind == constant) {
                ir_node *phi;
                ir_node *head_pred = get_irn_n(loop_head, be_src_pos);
                ir_node *loop_condition = get_unroll_copy(head_pred, unroll_nr - 1);

                set_irn_n(loop_head, loop_info.be_src_pos, loop_condition);

                for_each_phi(loop_head, phi) {
                        ir_node *pred = get_irn_n(phi, be_src_pos);
                        ir_node *last_pred;

                        /* It is possible, that the value used
                         * in the OWN backedge path is NOT assigned in this loop. */
                        if (is_in_loop(pred))
                                last_pred = get_unroll_copy(pred, copies);
                        else
                                last_pred = pred;
                        set_irn_n(phi, be_src_pos, last_pred);
                }

        } else {
                unrolling_fix_loop_head_inv();
        }
}

/* Copies the cur_loop several times. */
static void copy_loop(entry_edge *cur_loop_outs, int copies)
{
        int c;

        ir_reserve_resources(current_ir_graph, IR_RESOURCE_IRN_VISITED);

        for (c = 0; c < copies; ++c) {
                size_t i;

                inc_irg_visited(current_ir_graph);

                DB((dbg, LEVEL_5, "         ### Copy_loop  copy nr: %d ###\n", c));
                for (i = 0; i < ARR_LEN(cur_loop_outs); ++i) {
                        entry_edge entry = cur_loop_outs[i];
                        ir_node *pred = get_irn_n(entry.node, entry.pos);

                        copy_walk_n(pred, is_in_loop, c + 1);
                }
        }

        ir_free_resources(current_ir_graph, IR_RESOURCE_IRN_VISITED);
}


/* Creates a new phi from the given phi node omitting own bes,
 * using be_block as supplier of backedge informations. */
static ir_node *clone_phis_sans_bes(ir_node *phi, ir_node *be_block, ir_node *dest_block)
{
        ir_node **ins;
        int arity = get_irn_arity(phi);
        int i, c = 0;
        ir_node *newphi;

        assert(get_irn_arity(phi) == get_irn_arity(be_block));
        assert(is_Phi(phi));

        ins = NEW_ARR_F(ir_node *, arity);
        for (i = 0; i < arity; ++i) {
                if (! is_own_backedge(be_block, i)) {
                        ins[c] = get_irn_n(phi, i);
                        ++c;
                }
        }

        newphi = new_r_Phi(dest_block, c, ins, get_irn_mode(phi));

        set_irn_link(phi, newphi);
        DB((dbg, LEVEL_4, "Linking for duffs device %N to %N\n", phi, newphi));

        return newphi;
}

/* Creates a new block from the given block node omitting own bes,
 * using be_block as supplier of backedge informations. */
static ir_node *clone_block_sans_bes(ir_node *node, ir_node *be_block)
{
        int arity = get_irn_arity(node);
        int i, c = 0;
        ir_node **ins;

        assert(get_irn_arity(node) == get_irn_arity(be_block));
        assert(is_Block(node));

        NEW_ARR_A(ir_node *, ins, arity);
        for (i = 0; i < arity; ++i) {
                if (! is_own_backedge(be_block, i)) {
                        ins[c] = get_irn_n(node, i);
                        ++c;
                }
        }

        return new_Block(c, ins);
}

/* Creates a structure to calculate absolute value of node op.
 * Returns mux node with absolute value. */
static ir_node *new_Abs(ir_node *op, ir_mode *mode)
{
  ir_graph *irg      = get_irn_irg(op);
  ir_node  *block    = get_nodes_block(op);
  ir_node  *zero     = new_r_Const(irg, get_mode_null(mode));
  ir_node  *cmp      = new_r_Cmp(block, op, zero, ir_relation_less);
  ir_node  *minus_op = new_r_Minus(block, op, mode);
  ir_node  *mux      = new_r_Mux(block, cmp, op, minus_op, mode);

  return mux;
}


/* Creates blocks for duffs device, using previously obtained
 * informations about the iv.
 * TODO split */
static void create_duffs_block(void)
{
        ir_mode *mode;

        ir_node *block1, *count_block, *duff_block;
        ir_node *ems, *ems_mod, *ems_div, *ems_mod_proj, *cmp_null,
                *ems_mode_cond, *x_true, *x_false, *const_null;
        ir_node *true_val, *false_val;
        ir_node *ins[2];

        ir_node *duff_mod, *proj, *cond;

        ir_node *count, *correction, *unroll_c;
        ir_node *cmp_bad_count, *good_count, *bad_count, *count_phi, *bad_count_neg;
        ir_node *phi;

        mode = get_irn_mode(loop_info.end_val);
        const_null = new_Const(get_mode_null(mode));

        /* TODO naming
         * 1. Calculate first approach to count.
         *    Condition: (end - start) % step == 0 */
        block1 = clone_block_sans_bes(loop_head, loop_head);
        DB((dbg, LEVEL_4, "Duff block 1 %N\n", block1));

        /* Create loop entry phis in first duff block
         * as it becomes the loops preheader */
        for_each_phi(loop_head, phi) {
                /* Returns phis pred if phi would have arity 1*/
                ir_node *new_phi = clone_phis_sans_bes(phi, loop_head, block1);

                DB((dbg, LEVEL_4, "HEAD %N phi %N\n", loop_head, phi));
                DB((dbg, LEVEL_4, "BLOCK1 %N phi %N\n", block1, new_phi));
        }

        ems = new_r_Sub(block1, loop_info.end_val, loop_info.start_val,
                get_irn_mode(loop_info.end_val));
                DB((dbg, LEVEL_4, "BLOCK1 sub %N\n", ems));


        ems = new_Sub(loop_info.end_val, loop_info.start_val,
                get_irn_mode(loop_info.end_val));

        DB((dbg, LEVEL_4, "mod ins %N %N\n", ems, loop_info.step));
        ems_mod = new_r_Mod(block1,
                new_NoMem(),
                ems,
                loop_info.step,
                mode,
                op_pin_state_pinned);
        ems_div = new_r_Div(block1,
                new_NoMem(),
                ems,
                loop_info.step,
                mode,
                op_pin_state_pinned);

        DB((dbg, LEVEL_4, "New module node %N\n", ems_mod));

        ems_mod_proj = new_r_Proj(ems_mod, mode_Iu, pn_Mod_res);
        cmp_null = new_r_Cmp(block1, ems_mod_proj, const_null, ir_relation_less);
        ems_mode_cond = new_r_Cond(block1, cmp_null);

        /* ems % step == 0 */
        x_true = new_r_Proj(ems_mode_cond, mode_X, pn_Cond_true);
        /* ems % step != 0 */
        x_false = new_r_Proj(ems_mode_cond, mode_X, pn_Cond_false);

        /* 2. Second block.
         * Assures, duffs device receives a valid count.
         * Condition:
         *     decreasing: count < 0
         *     increasing: count > 0
         */
        ins[0] = x_true;
        ins[1] = x_false;

        count_block = new_Block(2, ins);
        DB((dbg, LEVEL_4, "Duff block 2 %N\n", count_block));


        /* Increase loop-taken-count depending on the loop condition
         * uses the latest iv to compare to. */
        if (loop_info.latest_value == 1) {
                /* ems % step == 0 :  +0 */
                true_val = new_Const(get_mode_null(mode));
                /* ems % step != 0 :  +1 */
                false_val = new_Const(get_mode_one(mode));
        } else {
                ir_tarval *tv_two = new_tarval_from_long(2, mode);
                /* ems % step == 0 :  +1 */
                true_val = new_Const(get_mode_one(mode));
                /* ems % step != 0 :  +2 */
                false_val = new_Const(tv_two);
        }

        ins[0] = true_val;
        ins[1] = false_val;

        correction = new_r_Phi(count_block, 2, ins, mode);

        count = new_r_Proj(ems_div, mode, pn_Div_res);

        /* (end - start) / step  +  correction */
        count = new_Add(count, correction, mode);

        /* We preconditioned the loop to be tail-controlled.
         * So, if count is something 'wrong' like 0,
         * negative/positive (depending on step direction),
         * we may take the loop once (tail-contr.) and leave it
         * to the existing condition, to break; */

        /* Depending on step direction, we have to check for > or < 0 */
        if (loop_info.decreasing == 1) {
                cmp_bad_count = new_r_Cmp(count_block, count, const_null,
                                          ir_relation_less);
        } else {
                cmp_bad_count = new_r_Cmp(count_block, count, const_null,
                                          ir_relation_greater);
        }

        bad_count_neg = new_r_Cond(count_block, cmp_bad_count);
        good_count = new_Proj(bad_count_neg, mode_X, pn_Cond_true);
        bad_count = new_Proj(ems_mode_cond, mode_X, pn_Cond_false);

        /* 3. Duff Block
         *    Contains module to decide which loop to start from. */

        ins[0] = good_count;
        ins[1] = bad_count;
        duff_block = new_Block(2, ins);
        DB((dbg, LEVEL_4, "Duff block 3 %N\n", duff_block));

        /* Get absolute value */
        ins[0] = new_Abs(count, mode);
        /* Manually feed the aforementioned count = 1 (bad case)*/
        ins[1] = new_Const(get_mode_one(mode));
        count_phi = new_r_Phi(duff_block, 2, ins, mode);

        unroll_c = new_Const(new_tarval_from_long((long)unroll_nr, mode));

        /* count % unroll_nr */
        duff_mod = new_r_Mod(duff_block,
                new_NoMem(),
                count_phi,
                unroll_c,
                mode,
                op_pin_state_pinned);


        proj = new_Proj(duff_mod, mode, pn_Mod_res);
        /* condition does NOT create itself in the block of the proj! */
        cond = new_r_Cond(duff_block, proj);

        loop_info.duff_cond = cond;
}

/* Returns 1 if given node is not in loop,
 * or if it is a phi of the loop head with only loop invariant defs.
 */
static unsigned is_loop_invariant_def(ir_node *node)
{
        int i;

        if (! is_in_loop(node)) {
                DB((dbg, LEVEL_4, "Not in loop %N\n", node));
                /* || is_Const(node) || is_SymConst(node)) {*/
                return 1;
        }

        /* If this is a phi of the loophead shared by more than 1 loop,
         * we need to check if all defs are not in the loop.  */
        if (is_Phi(node)) {
                ir_node *block;
                block = get_nodes_block(node);

                /* To prevent unexpected situations. */
                if (block != loop_head) {
                        return 0;
                }

                for (i = 0; i < get_irn_arity(node); ++i) {
                        /* Check if all bes are just loopbacks. */
                        if (is_own_backedge(block, i) && get_irn_n(node, i) != node)
                                return 0;
                }
                DB((dbg, LEVEL_4, "invar %N\n", node));
                return 1;
        }
        DB((dbg, LEVEL_4, "Not invar %N\n", node));

        return 0;
}

/* Returns 1 if one pred of node is invariant and the other is not.
 * invar_pred and other are set analogously. */
static unsigned get_invariant_pred(ir_node *node, ir_node **invar_pred, ir_node **other)
{
        ir_node *pred0 = get_irn_n(node, 0);
        ir_node *pred1 = get_irn_n(node, 1);

        *invar_pred = NULL;
        *other = NULL;

        if (is_loop_invariant_def(pred0)) {
                DB((dbg, LEVEL_4, "pred0 invar %N\n", pred0));
                *invar_pred = pred0;
                *other = pred1;
        }

        if (is_loop_invariant_def(pred1)) {
                DB((dbg, LEVEL_4, "pred1 invar %N\n", pred1));

                if (*invar_pred != NULL) {
                        /* RETURN. We do not want both preds to be invariant. */
                        return 0;
                }

                *other = pred0;
                *invar_pred = pred1;
                return 1;
        } else {
                DB((dbg, LEVEL_4, "pred1 not invar %N\n", pred1));

                if (*invar_pred != NULL)
                        return 1;
                else
                        return 0;
        }
}

/* Starts from a phi that may belong to an iv.
 * If an add forms a loop with iteration_phi,
 * and add uses a constant, 1 is returned
 * and 'start' as well as 'add' are sane. */
static unsigned get_start_and_add(ir_node *iteration_phi, unrolling_kind_flag role)
{
        int i;
        ir_node *found_add = loop_info.add;
        int arity = get_irn_arity(iteration_phi);

        DB((dbg, LEVEL_4, "Find start and add from %N\n", iteration_phi));

        for (i = 0; i < arity; ++i) {

                /* Find start_val which needs to be pred of the iteration_phi.
                 * If start_val already known, sanity check. */
                if (!is_backedge(get_nodes_block(loop_info.iteration_phi), i)) {
                        ir_node *found_start_val = get_irn_n(loop_info.iteration_phi, i);

                        DB((dbg, LEVEL_4, "found_start_val %N\n", found_start_val));

                        /* We already found a start_val it has to be always the same. */
                        if (loop_info.start_val && found_start_val != loop_info.start_val)
                                return 0;

                        if ((role == constant) && !(is_SymConst(found_start_val) || is_Const(found_start_val)))
                                        return 0;
                        else if((role == constant) && !(is_loop_invariant_def(found_start_val)))
                                        return 0;

                        loop_info.start_val = found_start_val;
                }

                /* The phi has to be in the loop head.
                 * Follow all own backedges. Every value supplied from these preds of the phi
                 * needs to origin from the same add. */
                if (is_own_backedge(get_nodes_block(loop_info.iteration_phi), i)) {
                        ir_node *new_found = get_irn_n(loop_info.iteration_phi,i);

                        DB((dbg, LEVEL_4, "is add? %N\n", new_found));

                        if (! (is_Add(new_found) || is_Sub(new_found)) || (found_add && found_add != new_found))
                                return 0;
                        else
                                found_add = new_found;
                }
        }

        loop_info.add = found_add;

        return 1;
}


/* Returns 1 if one pred of node is a const value and the other is not.
 * const_pred and other are set analogously. */
static unsigned get_const_pred(ir_node *node, ir_node **const_pred, ir_node **other)
{
        ir_node *pred0 = get_irn_n(node, 0);
        ir_node *pred1 = get_irn_n(node, 1);

        DB((dbg, LEVEL_4, "Checking for constant pred of %N\n", node));

        *const_pred = NULL;
        *other = NULL;

        /*DB((dbg, LEVEL_4, "is %N const\n", pred0));*/
        if (is_Const(pred0) || is_SymConst(pred0)) {
                *const_pred = pred0;
                *other = pred1;
        }

        /*DB((dbg, LEVEL_4, "is %N const\n", pred1));*/
        if (is_Const(pred1) || is_SymConst(pred1)) {
                if (*const_pred != NULL) {
                        /* RETURN. We do not want both preds to be constant. */
                        return 0;
                }

                *other = pred0;
                *const_pred = pred1;
        }

        if (*const_pred == NULL)
                return 0;
        else
                return 1;
}

/* Returns 1 if loop exits within 2 steps of the iv.
 * Norm_proj means we do not exit the loop.*/
static unsigned simulate_next(ir_tarval **count_tar,
                ir_tarval *stepped, ir_tarval *step_tar, ir_tarval *end_tar,
                ir_relation norm_proj)
{
        ir_tarval *next;

        DB((dbg, LEVEL_4, "Loop taken if (stepped)%ld %s (end)%ld ",
                                get_tarval_long(stepped),
                                get_relation_string((norm_proj)),
                                get_tarval_long(end_tar)));
        DB((dbg, LEVEL_4, "comparing latest value %d\n", loop_info.latest_value));

        /* If current iv does not stay in the loop,
         * this run satisfied the exit condition. */
        if (! (tarval_cmp(stepped, end_tar) & norm_proj))
                return 1;

        DB((dbg, LEVEL_4, "Result: (stepped)%ld IS %s (end)%ld\n",
                                get_tarval_long(stepped),
                                get_relation_string(tarval_cmp(stepped, end_tar)),
                                get_tarval_long(end_tar)));

        /* next step */
        if (is_Add(loop_info.add))
                next = tarval_add(stepped, step_tar);
        else
                /* sub */
                next = tarval_sub(stepped, step_tar, get_irn_mode(loop_info.end_val));

        DB((dbg, LEVEL_4, "Loop taken if %ld %s %ld ",
                                get_tarval_long(next),
                                get_relation_string(norm_proj),
                                get_tarval_long(end_tar)));
        DB((dbg, LEVEL_4, "comparing latest value %d\n", loop_info.latest_value));

        /* Increase steps. */
        *count_tar = tarval_add(*count_tar, get_tarval_one(get_tarval_mode(*count_tar)));

        /* Next has to fail the loop condition, or we will never exit. */
        if (! (tarval_cmp(next, end_tar) & norm_proj))
                return 1;
        else
                return 0;
}

/* Check if loop meets requirements for a 'simple loop':
 * - Exactly one cf out
 * - Allowed calls
 * - Max nodes after unrolling
 * - tail-controlled
 * - exactly one be
 * - cmp
 * Returns Projection of cmp node or NULL; */
static ir_node *is_simple_loop(void)
{
        int arity, i;
        ir_node *loop_block, *exit_block, *projx, *cond, *cmp;

        /* Maximum of one condition, and no endless loops. */
        if (loop_info.cf_outs != 1)
                return NULL;

        DB((dbg, LEVEL_4, "1 loop exit\n"));

        /* Calculate maximum unroll_nr keeping node count below limit. */
        loop_info.max_unroll = (int)((double)opt_params.max_unrolled_loop_size / (double)loop_info.nodes);
        if (loop_info.max_unroll < 2) {
                ++stats.too_large;
                return NULL;
        }

        DB((dbg, LEVEL_4, "maximum unroll factor %u, to not exceed node limit \n",
                opt_params.max_unrolled_loop_size));

        arity = get_irn_arity(loop_head);
        /* RETURN if we have more than 1 be. */
        /* Get my backedges without alien bes. */
        loop_block = NULL;
        for (i = 0; i < arity; ++i) {
                ir_node *pred = get_irn_n(loop_head, i);
                if (is_own_backedge(loop_head, i)) {
                        if (loop_block)
                                /* Our simple loops may have only one backedge. */
                                return NULL;
                        else {
                                loop_block = get_nodes_block(pred);
                                loop_info.be_src_pos = i;
                        }
                }
        }

        DB((dbg, LEVEL_4, "loop has 1 own backedge.\n"));

        exit_block = get_nodes_block(loop_info.cf_out.pred);
        /* The loop has to be tail-controlled.
         * This can be changed/improved,
         * but we would need a duff iv. */
        if (exit_block != loop_block)
                return NULL;

        DB((dbg, LEVEL_4, "tail-controlled loop.\n"));

        /* find value on which loop exit depends */
        projx = loop_info.cf_out.pred;
        cond = get_irn_n(projx, 0);
        cmp = get_irn_n(cond, 0);

        if (!is_Cmp(cmp))
                return NULL;

        DB((dbg, LEVEL_5, "projection is %s\n", get_relation_string(get_Proj_proj(projx))));

        switch(get_Proj_proj(projx)) {
                case pn_Cond_false:
                        loop_info.exit_cond = 0;
                        break;
                case pn_Cond_true:
                        loop_info.exit_cond = 1;
                        break;
                default:
                        panic("Cond Proj_proj other than true/false");
        }

        DB((dbg, LEVEL_4, "Valid Cmp.\n"));
        return cmp;
}

/* Returns 1 if all nodes are mode_Iu or mode_Is. */
static unsigned are_mode_I(ir_node *n1, ir_node* n2, ir_node *n3)
{
        ir_mode *m1 = get_irn_mode(n1);
        ir_mode *m2 = get_irn_mode(n2);
        ir_mode *m3 = get_irn_mode(n3);

        if ((m1 == mode_Iu && m2 == mode_Iu && m3 == mode_Iu) ||
            (m1 == mode_Is && m2 == mode_Is && m3 == mode_Is))
                return 1;
        else
                return 0;
}

/* Checks if cur_loop is a simple tail-controlled counting loop
 * with start and end value loop invariant, step constant. */
static unsigned get_unroll_decision_invariant(void)
{

        ir_node   *projres, *loop_condition, *iteration_path;
        unsigned   success;
        ir_tarval *step_tar;
        ir_mode   *mode;


        /* RETURN if loop is not 'simple' */
        projres = is_simple_loop();
        if (projres == NULL)
                return 0;

        /* Use a minimal size for the invariant unrolled loop,
     * as duffs device produces overhead */
        if (loop_info.nodes < opt_params.invar_unrolling_min_size)
                return 0;

        loop_condition = get_irn_n(projres, 0);

        success = get_invariant_pred(loop_condition, &loop_info.end_val, &iteration_path);
        DB((dbg, LEVEL_4, "pred invar %d\n", success));

        if (! success)
                return 0;

        DB((dbg, LEVEL_4, "Invariant End_val %N, other %N\n", loop_info.end_val, iteration_path));

        /* We may find the add or the phi first.
         * Until now we only have end_val. */
        if (is_Add(iteration_path) || is_Sub(iteration_path)) {

                loop_info.add = iteration_path;
                DB((dbg, LEVEL_4, "Case 1: Got add %N (maybe not sane)\n", loop_info.add));

                /* Preds of the add should be step and the iteration_phi */
                success = get_const_pred(loop_info.add, &loop_info.step, &loop_info.iteration_phi);
                if (! success)
                        return 0;

                DB((dbg, LEVEL_4, "Got step %N\n", loop_info.step));

                if (! is_Phi(loop_info.iteration_phi))
                        return 0;

                DB((dbg, LEVEL_4, "Got phi %N\n", loop_info.iteration_phi));

                /* Find start_val.
                 * Does necessary sanity check of add, if it is already set.  */
                success = get_start_and_add(loop_info.iteration_phi, invariant);
                if (! success)
                        return 0;

                DB((dbg, LEVEL_4, "Got start A  %N\n", loop_info.start_val));

        } else if (is_Phi(iteration_path)) {
                ir_node *new_iteration_phi;

                loop_info.iteration_phi = iteration_path;
                DB((dbg, LEVEL_4, "Case 2: Got phi %N\n", loop_info.iteration_phi));

                /* Find start_val and add-node.
                 * Does necessary sanity check of add, if it is already set.  */
                success = get_start_and_add(loop_info.iteration_phi, invariant);
                if (! success)
                        return 0;

                DB((dbg, LEVEL_4, "Got start B %N\n", loop_info.start_val));
                DB((dbg, LEVEL_4, "Got add or sub %N\n", loop_info.add));

                success = get_const_pred(loop_info.add, &loop_info.step, &new_iteration_phi);
                if (! success)
                        return 0;

                DB((dbg, LEVEL_4, "Got step (B) %N\n", loop_info.step));

                if (loop_info.iteration_phi != new_iteration_phi)
                        return 0;

        } else {
                return 0;
        }

        mode = get_irn_mode(loop_info.end_val);

        DB((dbg, LEVEL_4, "start %N, end %N, step %N\n",
                                loop_info.start_val, loop_info.end_val, loop_info.step));

        if (mode != mode_Is && mode != mode_Iu)
                return 0;

        /* TODO necessary? */
        if (!are_mode_I(loop_info.start_val, loop_info.step, loop_info.end_val))
                return 0;

        DB((dbg, LEVEL_4, "mode integer\n"));

        step_tar = get_Const_tarval(loop_info.step);

        if (tarval_is_null(step_tar)) {
                /* TODO Might be worth a warning. */
                return 0;
        }

        DB((dbg, LEVEL_4, "step is not 0\n"));

        create_duffs_block();

        return loop_info.max_unroll;
}

/* Returns unroll factor,
 * given maximum unroll factor and number of loop passes. */
static unsigned get_preferred_factor_constant(ir_tarval *count_tar)
{
        ir_tarval *tar_6, *tar_5, *tar_4, *tar_3, *tar_2;
        unsigned prefer;
        ir_mode *mode = get_irn_mode(loop_info.end_val);

        tar_6 = new_tarval_from_long(6, mode);
        tar_5 = new_tarval_from_long(5, mode);
        tar_4 = new_tarval_from_long(4, mode);
        tar_3 = new_tarval_from_long(3, mode);
        tar_2 = new_tarval_from_long(2, mode);

        /* loop passes % {6, 5, 4, 3, 2} == 0  */
        if (tarval_is_null(tarval_mod(count_tar, tar_6)))
                prefer = 6;
        else if (tarval_is_null(tarval_mod(count_tar, tar_5)))
                prefer = 5;
        else if (tarval_is_null(tarval_mod(count_tar, tar_4)))
                prefer = 4;
        else if (tarval_is_null(tarval_mod(count_tar, tar_3)))
                prefer = 3;
        else if (tarval_is_null(tarval_mod(count_tar, tar_2)))
                prefer = 2;
        else {
                /* gcd(max_unroll, count_tar) */
                int a = loop_info.max_unroll;
                int b = (int)get_tarval_long(count_tar);
                int c;

                DB((dbg, LEVEL_4, "gcd of max_unroll %d and count_tar %d: ", a, b));

                do {
                c = a % b;
                a = b; b = c;
                } while( c != 0);

                DB((dbg, LEVEL_4, "%d\n", a));
                return a;
        }

        DB((dbg, LEVEL_4, "preferred unroll factor %d\n", prefer));

        /*
         * If our preference is greater than the allowed unroll factor
         * we either might reduce the preferred factor and prevent a duffs device block,
         * or create a duffs device block, from which in this case (constants only)
         * we know the startloop at compiletime.
         * The latter yields the following graphs.
         * but for code generation we would want to use graph A.
         * The graphs are equivalent. So, we can only reduce the preferred factor.
         * A)                   B)
         *     PreHead             PreHead
         *        |      ,--.         |   ,--.
         *         \ Loop1   \        Loop2   \
         *          \  |     |       /  |     |
         *           Loop2   /      / Loop1   /
         *           |   `--'      |      `--'
         */

        if (prefer <= loop_info.max_unroll)
                return prefer;
        else {
                switch(prefer) {
                        case 6:
                                if (loop_info.max_unroll >= 3)
                                        return 3;
                                else if (loop_info.max_unroll >= 2)
                                        return 2;
                                else
                                        return 0;

                        case 4:
                                if (loop_info.max_unroll >= 2)
                                        return 2;
                                else
                                        return 0;

                        default:
                                return 0;
                }
        }
}

/* Check if cur_loop is a simple counting loop.
 * Start, step and end are constants.
 * TODO The whole constant case should use procedures similar to
 * the invariant case, as they are more versatile. */
/* TODO split. */
static unsigned get_unroll_decision_constant(void)
{
        ir_node     *cmp, *iteration_path;
        unsigned     success, is_latest_val;
        ir_tarval   *start_tar, *end_tar, *step_tar, *diff_tar, *count_tar;
        ir_tarval   *stepped;
        ir_relation  proj_proj, norm_proj;
        ir_mode     *mode;

        /* RETURN if loop is not 'simple' */
        cmp = is_simple_loop();
        if (cmp == NULL)
                return 0;

        /* One in of the loop condition needs to be loop invariant. => end_val
         * The other in is assigned by an add. => add
         * The add uses a loop invariant value => step
         * and a phi with a loop invariant start_val and the add node as ins.

           ^   ^
           |   | .-,
           |   Phi |
                \  |   |
          ^  Add   |
           \  | \__|
            cond
             /\
        */

        success = get_const_pred(cmp, &loop_info.end_val, &iteration_path);
        if (! success)
                return 0;

        DB((dbg, LEVEL_4, "End_val %N, other %N\n", loop_info.end_val, iteration_path));

        /* We may find the add or the phi first.
         * Until now we only have end_val. */
        if (is_Add(iteration_path) || is_Sub(iteration_path)) {

                /* We test against the latest value of the iv. */
                is_latest_val = 1;

                loop_info.add = iteration_path;
                DB((dbg, LEVEL_4, "Case 2: Got add %N (maybe not sane)\n", loop_info.add));

                /* Preds of the add should be step and the iteration_phi */
                success = get_const_pred(loop_info.add, &loop_info.step, &loop_info.iteration_phi);
                if (! success)
                        return 0;

                DB((dbg, LEVEL_4, "Got step %N\n", loop_info.step));

                if (! is_Phi(loop_info.iteration_phi))
                        return 0;

                DB((dbg, LEVEL_4, "Got phi %N\n", loop_info.iteration_phi));

                /* Find start_val.
                 * Does necessary sanity check of add, if it is already set.  */
                success = get_start_and_add(loop_info.iteration_phi, constant);
                if (! success)
                        return 0;

                DB((dbg, LEVEL_4, "Got start %N\n", loop_info.start_val));

        } else if (is_Phi(iteration_path)) {
                ir_node *new_iteration_phi;

                /* We compare with the value the iv had entering this run. */
                is_latest_val = 0;

                loop_info.iteration_phi = iteration_path;
                DB((dbg, LEVEL_4, "Case 1: Got phi %N \n", loop_info.iteration_phi));

                /* Find start_val and add-node.
                 * Does necessary sanity check of add, if it is already set.  */
                success = get_start_and_add(loop_info.iteration_phi, constant);
                if (! success)
                        return 0;

                DB((dbg, LEVEL_4, "Got start %N\n", loop_info.start_val));
                DB((dbg, LEVEL_4, "Got add or sub %N\n", loop_info.add));

                success = get_const_pred(loop_info.add, &loop_info.step, &new_iteration_phi);
                if (! success)
                        return 0;

                DB((dbg, LEVEL_4, "Got step %N\n", loop_info.step));

                if (loop_info.iteration_phi != new_iteration_phi)
                        return 0;

        } else {
                /* RETURN */
                return 0;
        }

        mode = get_irn_mode(loop_info.end_val);

        DB((dbg, LEVEL_4, "start %N, end %N, step %N\n",
                                loop_info.start_val, loop_info.end_val, loop_info.step));

        if (mode != mode_Is && mode != mode_Iu)
                return 0;

        /* TODO necessary? */
        if (!are_mode_I(loop_info.start_val, loop_info.step, loop_info.end_val))
                return 0;

        DB((dbg, LEVEL_4, "mode integer\n"));

        end_tar = get_Const_tarval(loop_info.end_val);
        start_tar = get_Const_tarval(loop_info.start_val);
        step_tar = get_Const_tarval(loop_info.step);

        if (tarval_is_null(step_tar))
                /* TODO Might be worth a warning. */
                return 0;

        DB((dbg, LEVEL_4, "step is not 0\n"));

        if ((!tarval_is_negative(step_tar)) ^ (!is_Sub(loop_info.add)))
                loop_info.decreasing = 1;

        diff_tar = tarval_sub(end_tar, start_tar, mode);

        /* We need at least count_tar steps to be close to end_val, maybe more.
         * No way, that we have gone too many steps.
         * This represents the 'latest value'.
         * (If condition checks against latest value, is checked later) */
        count_tar = tarval_div(diff_tar, step_tar);

        /* Iv will not pass end_val (except overflows).
         * Nothing done, as it would yield to no advantage. */
        if (tarval_is_negative(count_tar)) {
                DB((dbg, LEVEL_4, "Loop is endless or never taken."));
                /* TODO Might be worth a warning. */
                return 0;
        }

        ++stats.u_simple_counting_loop;

        loop_info.latest_value = is_latest_val;

        /* TODO split here
        if (! is_simple_counting_loop(&count_tar))
                return 0;
        */

        /* stepped can be negative, if step < 0 */
        stepped = tarval_mul(count_tar, step_tar);

        /* step as close to end_val as possible, */
        /* |stepped| <= |end_tar|, and dist(stepped, end_tar) is smaller than a step. */
        if (is_Sub(loop_info.add))
                stepped = tarval_sub(start_tar, stepped, mode_Is);
        else
                stepped = tarval_add(start_tar, stepped);

        DB((dbg, LEVEL_4, "stepped to %ld\n", get_tarval_long(stepped)));

        proj_proj = get_Cmp_relation(cmp);
        /* Assure that norm_proj is the stay-in-loop case. */
        if (loop_info.exit_cond == 1)
                norm_proj = get_negated_relation(proj_proj);
        else
                norm_proj = proj_proj;

        DB((dbg, LEVEL_4, "normalized projection %s\n", get_relation_string(norm_proj)));
        /* Executed at most once (stay in counting loop if a Eq b) */
        if (norm_proj == ir_relation_equal)
                /* TODO Might be worth a warning. */
                return 0;

        /* calculates next values and increases count_tar according to it */
        success = simulate_next(&count_tar, stepped, step_tar, end_tar, norm_proj);
        if (! success)
                return 0;

        /* We run loop once more, if we compare to the
         * not yet in-/decreased iv. */
        if (is_latest_val == 0) {
                DB((dbg, LEVEL_4, "condition uses not latest iv value\n"));
                count_tar = tarval_add(count_tar, get_tarval_one(mode));
        }

        DB((dbg, LEVEL_4, "loop taken %ld times\n", get_tarval_long(count_tar)));

        /* Assure the loop is taken at least 1 time. */
        if (tarval_is_null(count_tar)) {
                /* TODO Might be worth a warning. */
                return 0;
        }

        loop_info.count_tar = count_tar;
        return get_preferred_factor_constant(count_tar);
}

/**
 * Loop unrolling
 */
static void unroll_loop(void)
{

        if (! (loop_info.nodes > 0))
                return;

        if (loop_info.nodes > opt_params.max_unrolled_loop_size) {
                DB((dbg, LEVEL_2, "Nodes %d > allowed nodes %d\n",
                        loop_info.nodes, opt_params.max_unrolled_loop_size));
                ++stats.too_large;
                return;
        }

        if (loop_info.calls > 0) {
                DB((dbg, LEVEL_2, "Calls %d > allowed calls 0\n",
                        loop_info.calls));
                ++stats.calls_limit;
                return;
        }

        unroll_nr = 0;

        /* get_unroll_decision_constant and invariant are completely
         * independent for flexibility.
         * Some checks may be performed twice. */

        /* constant case? */
        if (opt_params.allow_const_unrolling)
                unroll_nr = get_unroll_decision_constant();
        if (unroll_nr > 1) {
                loop_info.unroll_kind = constant;

        } else {
                /* invariant case? */
                if (opt_params.allow_invar_unrolling)
                        unroll_nr = get_unroll_decision_invariant();
                if (unroll_nr > 1)
                        loop_info.unroll_kind = invariant;
        }

        DB((dbg, LEVEL_2, " *** Unrolling %d times ***\n", unroll_nr));

        if (unroll_nr > 1) {
                loop_entries = NEW_ARR_F(entry_edge, 0);

                /* Get loop outs */
                irg_walk_graph(current_ir_graph, get_loop_entries, NULL, NULL);

                if (loop_info.unroll_kind == constant) {
                        if ((int)get_tarval_long(loop_info.count_tar) == unroll_nr)
                                loop_info.needs_backedge = 0;
                        else
                                loop_info.needs_backedge = 1;
                } else {
                        loop_info.needs_backedge = 1;
                }

                /* Use phase to keep copy of nodes from the condition chain. */
                ir_nodemap_init(&map, current_ir_graph);
                obstack_init(&obst);

                /* Copies the loop */
                copy_loop(loop_entries, unroll_nr - 1);

                /* Line up the floating copies. */
                place_copies(unroll_nr - 1);

                /* Remove phis with 1 in
                 * If there were no nested phis, this would not be necessary.
                 * Avoiding the creation in the first place
                 * leads to complex special cases. */
                irg_walk_graph(current_ir_graph, correct_phis, NULL, NULL);

                if (loop_info.unroll_kind == constant)
                        ++stats.constant_unroll;
                else
                        ++stats.invariant_unroll;

                clear_irg_state(current_ir_graph, IR_GRAPH_STATE_CONSISTENT_DOMINANCE);

                DEL_ARR_F(loop_entries);
                obstack_free(&obst, NULL);
                ir_nodemap_destroy(&map);
        }

}

/* Analyzes the loop, and checks if size is within allowed range.
 * Decides if loop will be processed. */
static void init_analyze(ir_graph *irg, ir_loop *loop)
{
        cur_loop = loop;

        loop_head       = NULL;
        loop_head_valid = true;

        /* Reset loop info */
        memset(&loop_info, 0, sizeof(loop_info_t));

        DB((dbg, LEVEL_1, "    >>>> current loop %ld <<<\n",
            get_loop_loop_nr(loop)));

        /* Collect loop informations: head, node counts. */
        irg_walk_graph(irg, get_loop_info, NULL, NULL);

        /* RETURN if there is no valid head */
        if (!loop_head || !loop_head_valid) {
                DB((dbg, LEVEL_1,   "No valid loop head. Nothing done.\n"));
                return;
        } else {
                DB((dbg, LEVEL_1,   "Loophead: %N\n", loop_head));
        }

        if (loop_info.branches > opt_params.max_branches) {
                DB((dbg, LEVEL_1, "Branches %d > allowed branches %d\n",
                        loop_info.branches, opt_params.max_branches));
                ++stats.calls_limit;
                return;
        }

        switch (loop_op) {
                case loop_op_inversion:
                        loop_inversion(irg);
                        break;

                case loop_op_unrolling:
                        unroll_loop();
                        break;

                default:
                        panic("Loop optimization not implemented.");
        }
        DB((dbg, LEVEL_1, "       <<<< end of loop with node %ld >>>>\n",
            get_loop_loop_nr(loop)));
}

/* Find innermost loops and add them to loops. */
static void find_innermost_loop(ir_loop *loop)
{
        bool   had_sons   = false;
        size_t n_elements = get_loop_n_elements(loop);
        size_t e;

        for (e = 0; e < n_elements; ++e) {
                loop_element element = get_loop_element(loop, e);
                if (*element.kind == k_ir_loop) {
                        find_innermost_loop(element.son);
                        had_sons = true;
                }
        }

        if (!had_sons) {
                ARR_APP1(ir_loop*, loops, loop);
        }
}

static void set_loop_params(void)
{
    opt_params.max_loop_size = 100;
    opt_params.depth_adaption = -50;
    opt_params.count_phi = true;
    opt_params.count_proj = false;
    opt_params.allowed_calls = 0;

    opt_params.max_cc_size = 5;


    opt_params.allow_const_unrolling = true;
    opt_params.allow_invar_unrolling = false;

    opt_params.invar_unrolling_min_size = 20;
    opt_params.max_unrolled_loop_size = 400;
    opt_params.max_branches = 9999;
}

/* Assure preconditions are met and go through all loops. */
void loop_optimization(ir_graph *irg)
{
        ir_loop *loop;
        size_t   i;
        size_t   n_elements;

        set_loop_params();

        /* Reset stats for this procedure */
        reset_stats();

        /* Preconditions */
        set_current_ir_graph(irg);

        ir_reserve_resources(irg, IR_RESOURCE_IRN_LINK | IR_RESOURCE_PHI_LIST);
        collect_phiprojs(irg);

        loop = get_irg_loop(irg);

        loops = NEW_ARR_F(ir_loop *, 0);
        /* List all inner loops */
        n_elements = get_loop_n_elements(loop);
        for (i = 0; i < n_elements; ++i) {
                loop_element element = get_loop_element(loop, i);
                if (*element.kind != k_ir_loop)
                        continue;
                find_innermost_loop(element.son);
        }

        /* Set all links to NULL */
        irg_walk_graph(irg, reset_link, NULL, NULL);

        for (i = 0; i < ARR_LEN(loops); ++i) {
                ir_loop *loop = loops[i];

                ++stats.loops;

                /* Analyze and handle loop */
                init_analyze(irg, loop);

                /* Copied blocks do not have their phi list yet */
                collect_phiprojs(irg);

                /* Set links to NULL
                 * TODO Still necessary? */
                irg_walk_graph(irg, reset_link, NULL, NULL);
        }

        print_stats();

        DEL_ARR_F(loops);
        ir_free_resources(irg, IR_RESOURCE_IRN_LINK | IR_RESOURCE_PHI_LIST);
}

static ir_graph_state_t perform_loop_unrolling(ir_graph *irg)
{
        loop_op = loop_op_unrolling;
        loop_optimization(irg);
        return 0;
}

static ir_graph_state_t perform_loop_inversion(ir_graph *irg)
{
        loop_op = loop_op_inversion;
        loop_optimization(irg);
        return 0;
}

static ir_graph_state_t perform_loop_peeling(ir_graph *irg)
{
        loop_op = loop_op_peeling;
        loop_optimization(irg);
        return 0;
}

static optdesc_t opt_unroll_loops = {
        "unroll-loops",
        IR_GRAPH_STATE_CONSISTENT_OUT_EDGES | IR_GRAPH_STATE_CONSISTENT_OUTS | IR_GRAPH_STATE_CONSISTENT_LOOPINFO,
        perform_loop_unrolling,
};

static optdesc_t opt_invert_loops = {
        "invert-loops",
        IR_GRAPH_STATE_CONSISTENT_OUT_EDGES | IR_GRAPH_STATE_CONSISTENT_OUTS | IR_GRAPH_STATE_CONSISTENT_LOOPINFO,
        perform_loop_inversion,
};

static optdesc_t opt_peel_loops = {
        "peel-loops",
        IR_GRAPH_STATE_CONSISTENT_OUT_EDGES | IR_GRAPH_STATE_CONSISTENT_OUTS | IR_GRAPH_STATE_CONSISTENT_LOOPINFO,
        perform_loop_peeling,
};

void do_loop_unrolling(ir_graph *irg)
{ perform_irg_optimization(irg, &opt_unroll_loops); }

void do_loop_inversion(ir_graph *irg)
{ perform_irg_optimization(irg, &opt_invert_loops); }

void do_loop_peeling(ir_graph *irg)
{ perform_irg_optimization(irg, &opt_peel_loops); }

ir_graph_pass_t *loop_inversion_pass(const char *name)
{
        return def_graph_pass(name ? name : "loop_inversion", do_loop_inversion);
}

ir_graph_pass_t *loop_unroll_pass(const char *name)
{
        return def_graph_pass(name ? name : "loop_unroll", do_loop_unrolling);
}

ir_graph_pass_t *loop_peeling_pass(const char *name)
{
        return def_graph_pass(name ? name : "loop_peeling", do_loop_peeling);
}

void firm_init_loop_opt(void)
{
        FIRM_DBG_REGISTER(dbg, "firm.opt.loop");
}