-/* Copyright (C) 1998 - 2000 by Universitaet Karlsruhe
-** All rights reserved.
-**
-** Author: Boris Boesler
-**
-** traverse an ir graph
-** - execute the pre function before recursion
-** - execute the post function after recursion
+/*
+ * Project: libFIRM
+ * File name: ir/ir/irgwalk.c
+ * Purpose:
+ * Author: Boris Boesler
+ * Modified by: Goetz Lindenmaier
+ * Created:
+ * CVS-ID: $Id$
+ * Copyright: (c) 1999-2003 Universität Karlsruhe
+ * Licence: This file protected by GPL - GNU GENERAL PUBLIC LICENSE.
+ */
+
+/**
+* traverse an ir graph
+* - execute the pre function before recursion
+* - execute the post function after recursion
*/
-/* $Id$ */
#ifdef HAVE_CONFIG_H
# include <config.h>
#endif
-# include <stdlib.h>
+#include <stdlib.h>
-# include "irnode.h"
-# include "irgraph.h" /* visited flag */
-# include "irprog.h"
-# include "irgwalk.h"
+#include "irnode_t.h"
+#include "irgraph_t.h" /* visited flag */
+#include "irprog.h"
+#include "irgwalk.h"
+#include "typewalk.h"
+#include "firmstat.h"
+#include "ircgcons.h"
-# include "eset.h"
-# include "array.h"
+#include "eset.h"
+#include "array.h"
/* walk over an interprocedural graph (callgraph). Visits only graphs in irg_set. */
-static void irg_walk_cg(ir_node * node, int visited, eset * irg_set, irg_walk_func pre, irg_walk_func post, void * env) {
+static void irg_walk_cg(ir_node * node, int visited, eset * irg_set,
+ irg_walk_func *pre, irg_walk_func *post, void * env) {
int i;
ir_graph * rem = current_ir_graph;
ir_node * pred;
- assert(node);
+ assert(node && node->kind == k_ir_node);
if (get_irn_visited(node) >= visited) return;
set_irn_visited(node, visited);
+ if (pre) pre(node, env);
+
pred = skip_Proj(node);
if (get_irn_op(pred) == op_CallBegin
|| get_irn_op(pred) == op_EndReg
current_ir_graph = get_irn_irg(pred);
}
- if (pre) pre(node, env);
-
- if (is_no_Block(node)) irg_walk_cg(get_nodes_Block(node), visited, irg_set, pre, post, env);
+ if (is_no_Block(node)) { /* not block */
+ irg_walk_cg(get_nodes_block(node), visited, irg_set, pre, post, env);
+ }
if (get_irn_op(node) == op_Block) { /* block */
for (i = get_irn_arity(node) - 1; i >= 0; --i) {
ir_node * exec = get_irn_n(node, i);
ir_node * pred = skip_Proj(exec);
if ((get_irn_op(pred) != op_CallBegin
- && get_irn_op(pred) != op_EndReg
- && get_irn_op(pred) != op_EndExcept)
- || eset_contains(irg_set, get_irn_irg(pred))) {
- irg_walk_cg(exec, visited, irg_set, pre, post, env);
+ && get_irn_op(pred) != op_EndReg
+ && get_irn_op(pred) != op_EndExcept)
+ || eset_contains(irg_set, get_irn_irg(pred))) {
+ irg_walk_cg(exec, visited, irg_set, pre, post, env);
}
}
- } else if (get_irn_op(node) == op_Filter) {
+ } else if (get_irn_op(node) == op_Filter) { /* filter */
for (i = get_irn_arity(node) - 1; i >= 0; --i) {
ir_node * pred = get_irn_n(node, i);
if (get_irn_op(pred) == op_Unknown || get_irn_op(pred) == op_Bad) {
- irg_walk_cg(pred, visited, irg_set, pre, post, env);
+ irg_walk_cg(pred, visited, irg_set, pre, post, env);
} else {
- ir_node * exec;
- exec = skip_Proj(get_Block_cfgpred(get_nodes_Block(node), i));
- assert(get_irn_op(exec) == op_CallBegin
- || get_irn_op(exec) == op_EndReg
- || get_irn_op(exec) == op_EndExcept);
- if (eset_contains(irg_set, get_irn_irg(exec))) {
- current_ir_graph = get_irn_irg(exec);
- irg_walk_cg(pred, visited, irg_set, pre, post, env);
- current_ir_graph = rem;
- }
+ ir_node * exec;
+ exec = skip_Proj(get_Block_cfgpred(get_nodes_block(node), i));
+
+ if (op_Bad == get_irn_op (exec)) {
+ continue;
+ }
+
+ assert(get_irn_op(exec) == op_CallBegin
+ || get_irn_op(exec) == op_EndReg
+ || get_irn_op(exec) == op_EndExcept);
+ if (eset_contains(irg_set, get_irn_irg(exec))) {
+ current_ir_graph = get_irn_irg(exec);
+ irg_walk_cg(pred, visited, irg_set, pre, post, env);
+ current_ir_graph = rem;
+ }
}
}
- } else {
+ } else { /* everything else */
for (i = get_irn_arity(node) - 1; i >= 0; --i) {
irg_walk_cg(get_irn_n(node, i), visited, irg_set, pre, post, env);
}
entity * ent = get_Call_callee(node, i);
ir_graph * irg = ent ? get_entity_irg(ent) : NULL;
if (irg && !eset_contains(irg_set, irg)) {
- eset_insert(irg_set, irg);
- irg_walk_graph(irg, (irg_walk_func) collect_irgs, NULL, irg_set);
+ eset_insert(irg_set, irg);
+ irg_walk_graph(irg, (irg_walk_func *) collect_irgs, NULL, irg_set);
}
}
}
}
-
-void irg_walk_2(ir_node *node, irg_walk_func pre, irg_walk_func post, void * env)
-{
+/**
+ * specialized version of irg_walk_2, called if only pre callback exists
+ */
+static void
+irg_walk_2_pre(ir_node *node, irg_walk_func *pre, void * env) {
int i;
+ set_irn_visited(node, current_ir_graph->visited);
+ pre(node, env);
- assert(node);
- /* printf(" - "); DDMSG2(node); */
+ if (node->op != op_Block) {
+ ir_node *pred = get_irn_n(node, -1);
+ if (pred->visited < current_ir_graph->visited)
+ irg_walk_2_pre(pred, pre, env);
+ }
+ for (i = get_irn_arity(node) - 1; i >= 0; --i) {
+ ir_node *pred = get_irn_n(node, i);
+ if (pred->visited < current_ir_graph->visited)
+ irg_walk_2_pre(pred, pre, env);
+ }
+}
- if (get_irn_visited(node) < get_irg_visited(current_ir_graph)) {
+/**
+ * specialized version of irg_walk_2, called if only post callback exists
+ */
+static void
+irg_walk_2_post(ir_node *node, irg_walk_func *post, void * env) {
+ int i;
+ set_irn_visited(node, current_ir_graph->visited);
-/* printf(" -> "); DDMSG2(node); */
- set_irn_visited(node, get_irg_visited(current_ir_graph));
+ if (node->op != op_Block) {
+ ir_node *pred = get_irn_n(node, -1);
+ if (pred->visited < current_ir_graph->visited)
+ irg_walk_2_post(pred, post, env);
+ }
+ for (i = get_irn_arity(node) - 1; i >= 0; --i) {
+ ir_node *pred = get_irn_n(node, i);
+ if (pred->visited < current_ir_graph->visited)
+ irg_walk_2_post(pred, post, env);
+ }
- if (pre) {
- pre(node, env);
- }
+ post(node, env);
+}
- if (is_no_Block(node)) {
- irg_walk_2(get_nodes_Block(node), pre, post, env);
- }
- for (i = get_irn_arity(node) - 1; i >= 0; --i) {
- /* printf(" "); DDMSG2(node);
- printf(" "); DDMSG2(get_irn_n(node, i)); */
+/**
+ * specialized version of irg_walk_2, called if pre and post callbacks exist
+ */
+static void
+irg_walk_2_both(ir_node *node, irg_walk_func *pre, irg_walk_func *post, void * env) {
+ int i;
+ set_irn_visited(node, current_ir_graph->visited);
- irg_walk_2(get_irn_n(node, i), pre, post, env);
- }
+ pre(node, env);
-/* printf(" <- "); DDMSG2(node); */
- if (post)
- post(node, env);
+ if (node->op != op_Block) {
+ ir_node *pred = get_irn_n(node, -1);
+ if (pred->visited < current_ir_graph->visited)
+ irg_walk_2_both(pred, pre, post, env);
}
- return;
+ for (i = get_irn_arity(node) - 1; i >= 0; --i) {
+ ir_node *pred = get_irn_n(node, i);
+ if (pred->visited < current_ir_graph->visited)
+ irg_walk_2_both(pred, pre, post, env);
+ }
+
+ post(node, env);
}
+/**
+ * Intraprozedural graph walker.
+ */
+static void
+irg_walk_2(ir_node *node, irg_walk_func *pre, irg_walk_func *post, void * env)
+{
+ if (node->visited < current_ir_graph->visited) {
+ if (!post) irg_walk_2_pre (node, pre, env);
+ else if (!pre) irg_walk_2_post(node, post, env);
+ else irg_walk_2_both(node, pre, post, env);
+ }
+}
-void irg_walk(ir_node *node, irg_walk_func pre, irg_walk_func post, void *env)
+/*
+ * generic graph walker
+ */
+void irg_walk(ir_node *node, irg_walk_func *pre, irg_walk_func *post, void *env)
{
- assert(node);
+ assert(node && node->kind==k_ir_node);
+
if (interprocedural_view) {
eset * irg_set = eset_create();
int visited;
ir_graph * irg;
+ assert(get_irp_ip_view_state() == ip_view_valid);
+
interprocedural_view = false;
eset_insert(irg_set, current_ir_graph);
- irg_walk(node, (irg_walk_func) collect_irgs, NULL, irg_set);
+ irg_walk(node, (irg_walk_func *) collect_irgs, NULL, irg_set);
interprocedural_view = true;
visited = get_max_irg_visited() + 1;
- irg_walk_cg(node, visited, irg_set, pre, post, env);
for (irg = eset_first(irg_set); irg; irg = eset_next(irg_set)) {
set_irg_visited(irg, visited);
}
+ irg_walk_cg(node, visited, irg_set, pre, post, env);
eset_destroy(irg_set);
} else {
inc_irg_visited(current_ir_graph);
}
-void irg_walk_graph(ir_graph *irg, irg_walk_func pre, irg_walk_func post, void *env) {
+void irg_walk_graph(ir_graph *irg, irg_walk_func *pre, irg_walk_func *post, void *env) {
ir_graph * rem = current_ir_graph;
+
+ stat_irg_walk(irg, (void *)pre, (void *)post);
current_ir_graph = irg;
irg_walk(get_irg_end(irg), pre, post, env);
current_ir_graph = rem;
}
-
-/***************************************************************************/
-
/* Executes irg_walk(end, pre, post, env) for all irgraphs in irprog.
Sets current_ir_graph properly for each walk. Conserves current
current_ir_graph. */
-void all_irg_walk(irg_walk_func pre, irg_walk_func post, void *env) {
+void all_irg_walk(irg_walk_func *pre, irg_walk_func *post, void *env) {
int i;
ir_graph *irg, *rem;
current_ir_graph = rem;
}
+/***************************************************************************/
+
+/* Returns current_ir_graph and sets it to the irg of predecessor index
+ of node n. */
+static INLINE ir_graph *
+switch_irg (ir_node *n, int index) {
+ ir_graph *old_current = current_ir_graph;
+
+ if (interprocedural_view) {
+ /* Only Filter and Block nodes can have predecessors in other graphs. */
+ if (get_irn_op(n) == op_Filter)
+ n = get_nodes_block(n);
+ if (get_irn_op(n) == op_Block) {
+ ir_node *cfop = skip_Proj(get_Block_cfgpred(n, index));
+ if (is_ip_cfop(cfop)) {
+ current_ir_graph = get_irn_irg(cfop);
+ }
+ }
+ }
+
+ return old_current;
+}
+
+static void
+cg_walk_2(ir_node *node, irg_walk_func *pre, irg_walk_func *post, void * env)
+{
+ int i;
+ ir_graph *rem = NULL;
+ assert(node);
+
+ if (get_irn_visited(node) < get_irg_visited(current_ir_graph)) {
+ set_irn_visited(node, get_irg_visited(current_ir_graph));
+
+ if (pre) pre(node, env);
+
+ if (is_no_Block(node))
+ cg_walk_2(get_nodes_block(node), pre, post, env);
+ for (i = get_irn_arity(node) - 1; i >= 0; --i) {
+ rem = switch_irg(node, i); /* @@@ AS: Is this wrong? We do have to
+ switch to the irg of the predecessor, don't we? */
+ cg_walk_2(get_irn_n(node, i), pre, post, env);
+ current_ir_graph = rem;
+ }
+
+ if (post) post(node, env);
+ }
+ return;
+}
+
+
+/* Walks all irgs in interprocedural view. Visits each node only once. */
+void cg_walk(irg_walk_func *pre, irg_walk_func *post, void *env) {
+ int i;
+ ir_graph *rem = current_ir_graph;
+ int rem_view = interprocedural_view;
+
+ interprocedural_view = true;
+
+ inc_max_irg_visited();
+ /* Fix all irg_visited flags */
+ for (i = 0; i < get_irp_n_irgs(); i++)
+ set_irg_visited(get_irp_irg(i), get_max_irg_visited());
+
+ /* Walk starting at unreachable procedures. Only these
+ * have End blocks visible in interprocedural view. */
+ for (i = 0; i < get_irp_n_irgs(); i++) {
+ ir_node *sb;
+ current_ir_graph = get_irp_irg(i);
+
+ sb = get_irg_start_block(current_ir_graph);
+
+ if ((get_Block_n_cfgpreds(sb) > 1) ||
+ (get_nodes_block(get_Block_cfgpred(sb, 0)) != sb)) continue;
+
+ cg_walk_2(get_irg_end(current_ir_graph), pre, post, env);
+ }
+
+ /* Check whether we walked all procedures: there could be procedures
+ with cyclic calls but no call from the outside. */
+ for (i = 0; i < get_irp_n_irgs(); i++) {
+ ir_node *sb;
+ current_ir_graph = get_irp_irg(i);
+
+ /* Test start block: if inner procedure end and end block are not
+ * visible and therefore not marked. */
+ sb = get_irg_start_block(current_ir_graph);
+ if (get_irn_visited(sb) < get_irg_visited(current_ir_graph)) {
+ cg_walk_2(sb, pre, post, env);
+ }
+ }
+
+ /* Walk all endless loops in inner procedures.
+ * We recognize an inner procedure if the End node is not visited. */
+ for (i = 0; i < get_irp_n_irgs(); i++) {
+ ir_node *e;
+ current_ir_graph = get_irp_irg(i);
+ e = get_irg_end(current_ir_graph);
+ if (get_irn_visited(e) < get_irg_visited(current_ir_graph)) {
+ int j;
+ /* Don't visit the End node. */
+ for (j = 0; j < get_End_n_keepalives(e); j++)
+ cg_walk_2(get_End_keepalive(e, j), pre, post, env);
+ }
+ }
+
+ interprocedural_view = rem_view;
+ current_ir_graph = rem;
+}
+
+
/***************************************************************************/
/* Walks back from n until it finds a real cf op. */
-ir_node *get_cf_op(ir_node *n) {
+static ir_node *get_cf_op(ir_node *n) {
ir_node *pred;
n = skip_nop(n);
n = skip_Tuple(n);
pred = skip_Proj(n);
if (!(is_cfop(pred) || is_fragile_op(pred) ||
- (get_irn_op(pred) == op_Bad)))
+ (get_irn_op(pred) == op_Bad)))
n = get_cf_op(n);
return skip_Proj(n);
}
-void irg_block_walk_2(ir_node *node, irg_walk_func pre, irg_walk_func post, void *env)
+static void irg_block_walk_2(ir_node *node, irg_walk_func *pre, irg_walk_func *post, void *env)
{
int i;
- assert(get_irn_opcode(node) == iro_Block);
-
if(get_Block_block_visited(node) < get_irg_block_visited(current_ir_graph)) {
set_Block_block_visited(node, get_irg_block_visited(current_ir_graph));
- if(pre)
- pre(node, env);
+ if(pre) pre(node, env);
for(i = get_Block_n_cfgpreds(node) - 1; i >= 0; --i) {
-
/* find the corresponding predecessor block. */
ir_node *pred = get_cf_op(get_Block_cfgpred(node, i));
- /* GL: I'm not sure whether this assert must go through. There
- could be Id chains?? Tuple/Proj .... */
-
- assert(is_cfop(pred)
- || is_fragile_op(pred)
- || (get_irn_op(pred) == op_Bad));
-
- pred = get_nodes_Block(pred);
+ pred = get_nodes_block(pred);
if(get_irn_opcode(pred) == iro_Block) {
- /* recursion */
- irg_block_walk_2(pred, pre, post, env);
+ /* recursion */
+ irg_block_walk_2(pred, pre, post, env);
}
else {
- assert(get_irn_opcode(pred) == iro_Bad);
+ assert(get_irn_opcode(pred) == iro_Bad);
}
}
- if(post)
- post(node, env);
+ if(post) post(node, env);
}
return;
}
/* walks only over Block nodes in the graph. Has it's own visited
flag, so that it can be interleaved with the other walker. */
-void irg_block_walk(ir_node *node, irg_walk_func pre, irg_walk_func post, void *env)
+void irg_block_walk(ir_node *node, irg_walk_func *pre, irg_walk_func *post, void *env)
{
ir_node *block, *pred;
int i;
+ stat_irg_block_walk(current_ir_graph, node, (void *)pre, (void *)post);
+
assert(node);
assert(!interprocedural_view); /* interprocedural_view not implemented, because it
- * interleaves with irg_walk */
+ * interleaves with irg_walk */
inc_irg_block_visited(current_ir_graph);
- if (is_no_Block(node)) block = get_nodes_Block(node); else block = node;
+ if (is_no_Block(node)) block = get_nodes_block(node); else block = node;
assert(get_irn_opcode(block) == iro_Block);
irg_block_walk_2(block, pre, post, env);
/* keepalive: the endless loops ... */
- if (get_irn_op(node) == op_End)
- for (i = 0; i < get_irn_arity(node); i++) {
+ if (get_irn_op(node) == op_End) {
+ int arity = get_irn_arity(node);
+ for (i = 0; i < arity; i++) {
pred = get_irn_n(node, i);
if (get_irn_op(pred) == op_Block)
- irg_block_walk_2(pred, pre, post, env);
+ irg_block_walk_2(pred, pre, post, env);
}
+ }
return;
}
-void irg_block_walk_graph(ir_graph *irg, irg_walk_func pre, irg_walk_func post, void *env) {
+void irg_block_walk_graph(ir_graph *irg, irg_walk_func *pre,
+ irg_walk_func *post, void *env) {
ir_graph * rem = current_ir_graph;
current_ir_graph = irg;
irg_block_walk(get_irg_end(irg), pre, post, env);
current_ir_graph = rem;
}
+/********************************************************************/
+
+typedef struct walk_env {
+ irg_walk_func *pre;
+ irg_walk_func *post;
+ void *env;
+} walk_env;
+
+/* Walk to all constant expressions in this entity. */
+static void walk_entity(entity *ent, void *env)
+{
+ walk_env *my_env = (walk_env *)env;
+
+ if (get_entity_variability(ent) != variability_uninitialized) {
+ if (is_atomic_entity(ent)) {
+ irg_walk(get_atomic_ent_value(ent), my_env->pre, my_env->post, my_env->env);
+ }
+ else {
+ int i, n = get_compound_ent_n_values(ent);
+
+ for (i = 0; i < n; i++)
+ irg_walk(get_compound_ent_value(ent, i), my_env->pre, my_env->post, my_env->env);
+ }
+ }
+}
+
+/* Walks over all code in const_code_irg. */
+void walk_const_code(irg_walk_func *pre, irg_walk_func *post, void *env) {
+ int i, j;
+ walk_env my_env;
+
+ ir_graph *rem = current_ir_graph;
+ current_ir_graph = get_const_code_irg();
+ inc_irg_visited(current_ir_graph);
+
+ my_env.pre = pre;
+ my_env.post = post;
+ my_env.env = env;
+
+ /* Walk all types that can contain constant entities. */
+ walk_types_entities(get_glob_type(), &walk_entity, &my_env);
+ for (i = 0; i < get_irp_n_types(); i++)
+ walk_types_entities(get_irp_type(i), &walk_entity, &my_env);
+ for (i = 0; i < get_irp_n_irgs(); i++)
+ walk_types_entities(get_irg_frame_type(get_irp_irg(i)), &walk_entity, &my_env);
+
+ /* Walk constant array bounds. */
+ for (i = 0; i < get_irp_n_types(); i++) {
+ type *tp = get_irp_type(i);
+ if (is_array_type(tp)) {
+ for (j = 0; j < get_array_n_dimensions(tp); j++) {
+ ir_node *n;
+ n = get_array_lower_bound(tp, j);
+ if (n) irg_walk(n, pre, post, env);
+ n = get_array_upper_bound(tp, j);
+ if (n) irg_walk(n, pre, post, env);
+ }
+ }
+ }
+
+ current_ir_graph = rem;
+}
+
+
+/********************************************************************/
+/** Walking support for interprocedural analysis **/
+/** **/
+/** @@@ Don't use, not operational yet, doesn't grok recursions!! **/
+/** @@@ Header for irgwalk.h, here until it works. **/
+/** **/
+/** Interprocedural walking should not walk all predecessors of **/
+/** all nodes. When leaving a procedure the walker should only **/
+/** follow the edge corresponding to the most recent entry of the **/
+/** procedure. The following functions use an internal stack to **/
+/** remember the current call site of a procedure. **/
+/** They also set current_ir_graph correctly. **/
+/** **/
+/** Usage example: **/
+/** **/
+/** void init_ip_walk (); **/
+/** work_on_graph(some_end_node); **/
+/** void finish_ip_walk(); **/
+/** **/
+/** work_on_graph(ir_node *n) { **/
+/** for (i = 0; i < get_irn_arity(n); i++) { **/
+/** if (...) continue; **/
+/** ir_node *m = get_irn_ip_pred(n, i); **/
+/** if !m continue; **/
+/** work_on_graph(m); **/
+/** return_recur(n, i); **/
+/** } **/
+/** } **/
+/********************************************************************/
+
+/* Call for i in {0|-1 ... get_irn_arity(n)}.
+ If n is a conventional node returns the same node as get_irn_n(n, i).
+ If the predecessors of n are in the callee of the procedure n belongs
+ to, returns get_irn_n(n, i) if this node is in the callee on the top
+ of the stack, else returns NULL.
+ If the predecessors of n are in a procedure called by the procedure n
+ belongs to pushes the caller on the caller stack in the callee.
+ Sets current_ir_graph to the graph the node returned is in. */
+ir_node *get_irn_ip_pred(ir_node *n, int pos);
+
+/* If get_irn_ip_pred() returned a node (not NULL) this must be
+ called to clear up the stacks.
+ Sets current_ir_graph to the graph n is in. */
+void return_recur(ir_node *n, int pos);
/********************************************************************/
/* Initialization routines ******************************************/
-void init_ip_walk () {
+void init_ip_walk (void) {
int i;
for (i = 0; i < get_irp_n_irgs(); i++)
new_callsite_stack(get_irp_irg(i));
}
-void finish_ip_walk() {
+void finish_ip_walk(void) {
int i;
for (i = 0; i < get_irp_n_irgs(); i++)
free_callsite_stack(get_irp_irg(i));
/* Find the cf_pred refering to pos. */
ir_node *block = n;
ir_node *cf_pred;
- if (get_irn_opcode(n) == iro_Filter) block = get_nodes_Block(n);
+ if (get_irn_opcode(n) == iro_Filter) block = get_nodes_block(n);
cf_pred = skip_Proj(get_irn_n(block, pos));
/* Check whether we enter or leave a procedure and act according. */
if ((get_irn_op(cf_pred) == op_EndReg) ||
- (get_irn_op(cf_pred) == op_EndExcept))
+ (get_irn_op(cf_pred) == op_EndExcept))
enter_procedure(block, cf_pred, pos);
if (get_irn_op(cf_pred) == op_CallBegin)
if (!leave_procedure(block, cf_pred, pos)) return NULL;
proj = get_Block_cg_cfgpred(block, pos);
assert(is_Proj(proj));
callee = get_entity_irg(get_Call_callee(get_CallBegin_call(cf_pred),
- get_Proj_proj(proj)));
+ get_Proj_proj(proj)));
current_ir_graph = callee;
push_callsite(callee, cf_pred);
}
/* Find the cf_pred refering to pos. */
block = n;
- if (get_irn_opcode(n) == iro_Filter) block = get_nodes_Block(n);
+ if (get_irn_opcode(n) == iro_Filter) block = get_nodes_block(n);
cf_pred = skip_Proj(get_irn_n(block, pos));
/* Check whether we re_enter or re_leave a procedure and act according. */
projects / libfirm / blobdiff