3 * File name: ir/ir/irgopt.c
4 * Purpose: Optimizations for a whole ir graph, i.e., a procedure.
5 * Author: Christian Schaefer, Goetz Lindenmaier
6 * Modified by: Sebastian Felis
9 * Copyright: (c) 1998-2003 Universität Karlsruhe
10 * Licence: This file protected by GPL - GNU GENERAL PUBLIC LICENSE.
22 #include "irgraph_t.h"
34 #include "pdeq.h" /* Fuer code placement */
38 #include "irbackedge_t.h"
44 /* Defined in iropt.c */
45 pset *new_identities (void);
46 void del_identities (pset *value_table);
47 void add_identities (pset *value_table, ir_node *node);
49 /*------------------------------------------------------------------*/
50 /* apply optimizations of iropt to all nodes. */
51 /*------------------------------------------------------------------*/
53 static void init_link (ir_node *n, void *env) {
54 set_irn_link(n, NULL);
57 #if 0 /* Old version. Avoids Ids.
58 This is not necessary: we do a postwalk, and get_irn_n
59 removes ids anyways. So it's much cheaper to call the
60 optimization less often and use the exchange() algorithm. */
62 optimize_in_place_wrapper (ir_node *n, void *env) {
64 ir_node *optimized, *old;
66 irn_arity = get_irn_arity(n);
67 for (i = 0; i < irn_arity; i++) {
68 /* get_irn_n skips Id nodes, so comparison old != optimized does not
69 show all optimizations. Therefore always set new predecessor. */
70 old = get_irn_intra_n(n, i);
71 optimized = optimize_in_place_2(old);
72 set_irn_n(n, i, optimized);
75 if (get_irn_op(n) == op_Block) {
76 optimized = optimize_in_place_2(n);
77 if (optimized != n) exchange (n, optimized);
82 optimize_in_place_wrapper (ir_node *n, void *env) {
83 ir_node *optimized = optimize_in_place_2(n);
84 if (optimized != n) exchange (n, optimized);
89 static INLINE void do_local_optimize(ir_node *n) {
90 /* Handle graph state */
91 assert(get_irg_phase_state(current_ir_graph) != phase_building);
92 if (get_opt_global_cse())
93 set_irg_pinned(current_ir_graph, op_pin_state_floats);
94 if (get_irg_outs_state(current_ir_graph) == outs_consistent)
95 set_irg_outs_inconsistent(current_ir_graph);
96 if (get_irg_dom_state(current_ir_graph) == dom_consistent)
97 set_irg_dom_inconsistent(current_ir_graph);
98 set_irg_loopinfo_inconsistent(current_ir_graph);
101 /* Clean the value_table in irg for the cse. */
102 del_identities(current_ir_graph->value_table);
103 current_ir_graph->value_table = new_identities();
105 /* walk over the graph */
106 irg_walk(n, init_link, optimize_in_place_wrapper, NULL);
109 void local_optimize_node(ir_node *n) {
110 ir_graph *rem = current_ir_graph;
111 current_ir_graph = get_irn_irg(n);
113 do_local_optimize(n);
115 current_ir_graph = rem;
120 local_optimize_graph (ir_graph *irg) {
121 ir_graph *rem = current_ir_graph;
122 current_ir_graph = irg;
124 do_local_optimize(irg->end);
126 current_ir_graph = rem;
130 /*------------------------------------------------------------------*/
131 /* Routines for dead node elimination / copying garbage collection */
132 /* of the obstack. */
133 /*------------------------------------------------------------------*/
136 * Remember the new node in the old node by using a field all nodes have.
139 set_new_node (ir_node *old, ir_node *new)
145 * Get this new node, before the old node is forgotton.
147 static INLINE ir_node *
148 get_new_node (ir_node * n)
154 * We use the block_visited flag to mark that we have computed the
155 * number of useful predecessors for this block.
156 * Further we encode the new arity in this flag in the old blocks.
157 * Remembering the arity is useful, as it saves a lot of pointer
158 * accesses. This function is called for all Phi and Block nodes
162 compute_new_arity(ir_node *b) {
163 int i, res, irn_arity;
166 irg_v = get_irg_block_visited(current_ir_graph);
167 block_v = get_Block_block_visited(b);
168 if (block_v >= irg_v) {
169 /* we computed the number of preds for this block and saved it in the
171 return block_v - irg_v;
173 /* compute the number of good predecessors */
174 res = irn_arity = get_irn_arity(b);
175 for (i = 0; i < irn_arity; i++)
176 if (get_irn_opcode(get_irn_n(b, i)) == iro_Bad) res--;
177 /* save it in the flag. */
178 set_Block_block_visited(b, irg_v + res);
183 /* TODO: add an ir_op operation */
184 static INLINE void new_backedge_info(ir_node *n) {
185 switch(get_irn_opcode(n)) {
187 n->attr.block.cg_backedge = NULL;
188 n->attr.block.backedge = new_backedge_arr(current_ir_graph->obst, get_irn_arity(n));
191 n->attr.phi_backedge = new_backedge_arr(current_ir_graph->obst, get_irn_arity(n));
194 n->attr.filter.backedge = new_backedge_arr(current_ir_graph->obst, get_irn_arity(n));
201 * Copies the node to the new obstack. The Ins of the new node point to
202 * the predecessors on the old obstack. For block/phi nodes not all
203 * predecessors might be copied. n->link points to the new node.
204 * For Phi and Block nodes the function allocates in-arrays with an arity
205 * only for useful predecessors. The arity is determined by counting
206 * the non-bad predecessors of the block.
208 * @param n The node to be copied
209 * @param env if non-NULL, the node number attribute will be copied to the new node
212 copy_node (ir_node *n, void *env) {
215 opcode op = get_irn_opcode(n);
216 int copy_node_nr = env != NULL;
218 /* The end node looses it's flexible in array. This doesn't matter,
219 as dead node elimination builds End by hand, inlineing doesn't use
221 /* assert(n->op == op_End || ((_ARR_DESCR(n->in))->cookie != ARR_F_MAGIC)); */
224 /* node copied already */
226 } else if (op == iro_Block) {
228 new_arity = compute_new_arity(n);
229 n->attr.block.graph_arr = NULL;
231 block = get_nodes_block(n);
232 if (get_irn_opcode(n) == iro_Phi) {
233 new_arity = compute_new_arity(block);
235 new_arity = get_irn_arity(n);
238 nn = new_ir_node(get_irn_dbg_info(n),
245 /* Copy the attributes. These might point to additional data. If this
246 was allocated on the old obstack the pointers now are dangling. This
247 frees e.g. the memory of the graph_arr allocated in new_immBlock. */
249 new_backedge_info(nn);
254 /* for easier debugging, we want to copy the node numbers too */
255 nn->node_nr = n->node_nr;
259 /* printf("\n old node: "); DDMSG2(n);
260 printf(" new node: "); DDMSG2(nn); */
264 * Copies new predecessors of old node to new node remembered in link.
265 * Spare the Bad predecessors of Phi and Block nodes.
268 copy_preds (ir_node *n, void *env) {
272 nn = get_new_node(n);
274 /* printf("\n old node: "); DDMSG2(n);
275 printf(" new node: "); DDMSG2(nn);
276 printf(" arities: old: %d, new: %d\n", get_irn_arity(n), get_irn_arity(nn)); */
278 if (get_irn_opcode(n) == iro_Block) {
279 /* Don't copy Bad nodes. */
281 irn_arity = get_irn_arity(n);
282 for (i = 0; i < irn_arity; i++)
283 if (get_irn_opcode(get_irn_n(n, i)) != iro_Bad) {
284 set_irn_n (nn, j, get_new_node(get_irn_n(n, i)));
285 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
288 /* repair the block visited flag from above misuse. Repair it in both
289 graphs so that the old one can still be used. */
290 set_Block_block_visited(nn, 0);
291 set_Block_block_visited(n, 0);
292 /* Local optimization could not merge two subsequent blocks if
293 in array contained Bads. Now it's possible.
294 We don't call optimize_in_place as it requires
295 that the fields in ir_graph are set properly. */
296 if ((get_opt_control_flow_straightening()) &&
297 (get_Block_n_cfgpreds(nn) == 1) &&
298 (get_irn_op(get_Block_cfgpred(nn, 0)) == op_Jmp)) {
299 ir_node *old = get_nodes_block(get_Block_cfgpred(nn, 0));
301 /* Jmp jumps into the block it is in -- deal self cycle. */
302 assert(is_Bad(get_new_node(get_irg_bad(current_ir_graph))));
303 exchange(nn, get_new_node(get_irg_bad(current_ir_graph)));
308 } else if (get_irn_opcode(n) == iro_Phi) {
309 /* Don't copy node if corresponding predecessor in block is Bad.
310 The Block itself should not be Bad. */
311 block = get_nodes_block(n);
312 set_irn_n (nn, -1, get_new_node(block));
314 irn_arity = get_irn_arity(n);
315 for (i = 0; i < irn_arity; i++)
316 if (get_irn_opcode(get_irn_n(block, i)) != iro_Bad) {
317 set_irn_n (nn, j, get_new_node(get_irn_n(n, i)));
318 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
321 /* If the pre walker reached this Phi after the post walker visited the
322 block block_visited is > 0. */
323 set_Block_block_visited(get_nodes_block(n), 0);
324 /* Compacting the Phi's ins might generate Phis with only one
326 if (get_irn_arity(n) == 1)
327 exchange(n, get_irn_n(n, 0));
329 irn_arity = get_irn_arity(n);
330 for (i = -1; i < irn_arity; i++)
331 set_irn_n (nn, i, get_new_node(get_irn_n(n, i)));
333 /* Now the new node is complete. We can add it to the hash table for cse.
334 @@@ inlinening aborts if we identify End. Why? */
335 if(get_irn_op(nn) != op_End)
336 add_identities (current_ir_graph->value_table, nn);
340 * Copies the graph recursively, compacts the keepalive of the end node.
342 * @param copy_node_nr If non-zero, the node number will be copied
345 copy_graph (int copy_node_nr) {
346 ir_node *oe, *ne, *ob, *nb, *om, *nm; /* old end, new end, old bad, new bad, old NoMem, new NoMem */
347 ir_node *ka; /* keep alive */
350 oe = get_irg_end(current_ir_graph);
351 /* copy the end node by hand, allocate dynamic in array! */
352 ne = new_ir_node(get_irn_dbg_info(oe),
359 /* Copy the attributes. Well, there might be some in the future... */
361 set_new_node(oe, ne);
363 /* copy the Bad node */
364 ob = get_irg_bad(current_ir_graph);
365 nb = new_ir_node(get_irn_dbg_info(ob),
372 set_new_node(ob, nb);
374 /* copy the NoMem node */
375 om = get_irg_no_mem(current_ir_graph);
376 nm = new_ir_node(get_irn_dbg_info(om),
383 set_new_node(om, nm);
385 /* copy the live nodes */
386 irg_walk(get_nodes_block(oe), copy_node, copy_preds, (void *)copy_node_nr);
387 /* copy_preds for the end node ... */
388 set_nodes_block(ne, get_new_node(get_nodes_block(oe)));
390 /*- ... and now the keep alives. -*/
391 /* First pick the not marked block nodes and walk them. We must pick these
392 first as else we will oversee blocks reachable from Phis. */
393 irn_arity = get_irn_arity(oe);
394 for (i = 0; i < irn_arity; i++) {
395 ka = get_irn_intra_n(oe, i);
396 if ((get_irn_op(ka) == op_Block) &&
397 (get_irn_visited(ka) < get_irg_visited(current_ir_graph))) {
398 /* We must keep the block alive and copy everything reachable */
399 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
400 irg_walk(ka, copy_node, copy_preds, (void *)copy_node_nr);
401 add_End_keepalive(ne, get_new_node(ka));
405 /* Now pick the Phis. Here we will keep all! */
406 irn_arity = get_irn_arity(oe);
407 for (i = 0; i < irn_arity; i++) {
408 ka = get_irn_intra_n(oe, i);
409 if ((get_irn_op(ka) == op_Phi)) {
410 if (get_irn_visited(ka) < get_irg_visited(current_ir_graph)) {
411 /* We didn't copy the Phi yet. */
412 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
413 irg_walk(ka, copy_node, copy_preds, (void *)copy_node_nr);
415 add_End_keepalive(ne, get_new_node(ka));
419 /* start block sometimes only reached after keep alives */
420 set_nodes_block(nb, get_new_node(get_nodes_block(ob)));
421 set_nodes_block(nm, get_new_node(get_nodes_block(om)));
425 * Copies the graph reachable from current_ir_graph->end to the obstack
426 * in current_ir_graph and fixes the environment.
427 * Then fixes the fields in current_ir_graph containing nodes of the
430 * @param copy_node_nr If non-zero, the node number will be copied
433 copy_graph_env (int copy_node_nr) {
435 /* Not all nodes remembered in current_ir_graph might be reachable
436 from the end node. Assure their link is set to NULL, so that
437 we can test whether new nodes have been computed. */
438 set_irn_link(get_irg_frame (current_ir_graph), NULL);
439 set_irn_link(get_irg_globals (current_ir_graph), NULL);
440 set_irn_link(get_irg_args (current_ir_graph), NULL);
441 set_irn_link(get_irg_initial_mem(current_ir_graph), NULL);
442 set_irn_link(get_irg_no_mem (current_ir_graph), NULL);
444 /* we use the block walk flag for removing Bads from Blocks ins. */
445 inc_irg_block_visited(current_ir_graph);
448 copy_graph(copy_node_nr);
450 /* fix the fields in current_ir_graph */
451 old_end = get_irg_end(current_ir_graph);
452 set_irg_end (current_ir_graph, get_new_node(old_end));
453 set_irg_end_except (current_ir_graph, get_irg_end(current_ir_graph));
454 set_irg_end_reg (current_ir_graph, get_irg_end(current_ir_graph));
456 set_irg_end_block (current_ir_graph, get_new_node(get_irg_end_block(current_ir_graph)));
457 if (get_irn_link(get_irg_frame(current_ir_graph)) == NULL) {
458 copy_node (get_irg_frame(current_ir_graph), (void *)copy_node_nr);
459 copy_preds(get_irg_frame(current_ir_graph), NULL);
461 if (get_irn_link(get_irg_globals(current_ir_graph)) == NULL) {
462 copy_node (get_irg_globals(current_ir_graph), (void *)copy_node_nr);
463 copy_preds(get_irg_globals(current_ir_graph), NULL);
465 if (get_irn_link(get_irg_initial_mem(current_ir_graph)) == NULL) {
466 copy_node (get_irg_initial_mem(current_ir_graph), (void *)copy_node_nr);
467 copy_preds(get_irg_initial_mem(current_ir_graph), NULL);
469 if (get_irn_link(get_irg_args(current_ir_graph)) == NULL) {
470 copy_node (get_irg_args(current_ir_graph), (void *)copy_node_nr);
471 copy_preds(get_irg_args(current_ir_graph), NULL);
473 set_irg_start (current_ir_graph, get_new_node(get_irg_start(current_ir_graph)));
475 set_irg_start_block(current_ir_graph,
476 get_new_node(get_irg_start_block(current_ir_graph)));
477 set_irg_frame (current_ir_graph, get_new_node(get_irg_frame(current_ir_graph)));
478 set_irg_globals (current_ir_graph, get_new_node(get_irg_globals(current_ir_graph)));
479 set_irg_initial_mem(current_ir_graph, get_new_node(get_irg_initial_mem(current_ir_graph)));
480 set_irg_args (current_ir_graph, get_new_node(get_irg_args(current_ir_graph)));
482 if (get_irn_link(get_irg_bad(current_ir_graph)) == NULL) {
483 copy_node(get_irg_bad(current_ir_graph), (void *)copy_node_nr);
484 copy_preds(get_irg_bad(current_ir_graph), NULL);
486 set_irg_bad(current_ir_graph, get_new_node(get_irg_bad(current_ir_graph)));
488 if (get_irn_link(get_irg_no_mem(current_ir_graph)) == NULL) {
489 copy_node(get_irg_no_mem(current_ir_graph), (void *)copy_node_nr);
490 copy_preds(get_irg_no_mem(current_ir_graph), NULL);
492 set_irg_no_mem(current_ir_graph, get_new_node(get_irg_no_mem(current_ir_graph)));
496 * Copies all reachable nodes to a new obstack. Removes bad inputs
497 * from block nodes and the corresponding inputs from Phi nodes.
498 * Merges single exit blocks with single entry blocks and removes
500 * Adds all new nodes to a new hash table for cse. Does not
501 * perform cse, so the hash table might contain common subexpressions.
504 dead_node_elimination(ir_graph *irg) {
506 int rem_ipview = get_interprocedural_view();
507 struct obstack *graveyard_obst = NULL;
508 struct obstack *rebirth_obst = NULL;
510 /* inform statistics that we started a dead-node elimination run */
511 stat_dead_node_elim_start(irg);
513 /* Remember external state of current_ir_graph. */
514 rem = current_ir_graph;
515 current_ir_graph = irg;
516 set_interprocedural_view(false);
518 /* Handle graph state */
519 assert(get_irg_phase_state(current_ir_graph) != phase_building);
520 free_callee_info(current_ir_graph);
521 free_outs(current_ir_graph);
522 /* @@@ so far we loose loops when copying */
523 free_loop_information(current_ir_graph);
525 if (get_opt_optimize() && get_opt_dead_node_elimination()) {
527 /* A quiet place, where the old obstack can rest in peace,
528 until it will be cremated. */
529 graveyard_obst = irg->obst;
531 /* A new obstack, where the reachable nodes will be copied to. */
532 rebirth_obst = (struct obstack *) xmalloc (sizeof (struct obstack));
533 current_ir_graph->obst = rebirth_obst;
534 obstack_init (current_ir_graph->obst);
536 /* We also need a new hash table for cse */
537 del_identities (irg->value_table);
538 irg->value_table = new_identities ();
540 /* Copy the graph from the old to the new obstack */
543 /* Free memory from old unoptimized obstack */
544 obstack_free(graveyard_obst, 0); /* First empty the obstack ... */
545 xfree (graveyard_obst); /* ... then free it. */
548 /* inform statistics that the run is over */
549 stat_dead_node_elim_stop(irg);
551 current_ir_graph = rem;
552 set_interprocedural_view(rem_ipview);
556 * Relink bad predeseccors of a block and store the old in array to the
557 * link field. This function is called by relink_bad_predecessors().
558 * The array of link field starts with the block operand at position 0.
559 * If block has bad predecessors, create a new in array without bad preds.
560 * Otherwise let in array untouched.
562 static void relink_bad_block_predecessors(ir_node *n, void *env) {
563 ir_node **new_in, *irn;
564 int i, new_irn_n, old_irn_arity, new_irn_arity = 0;
566 /* if link field of block is NULL, look for bad predecessors otherwise
567 this is allready done */
568 if (get_irn_op(n) == op_Block &&
569 get_irn_link(n) == NULL) {
571 /* save old predecessors in link field (position 0 is the block operand)*/
572 set_irn_link(n, (void *)get_irn_in(n));
574 /* count predecessors without bad nodes */
575 old_irn_arity = get_irn_arity(n);
576 for (i = 0; i < old_irn_arity; i++)
577 if (!is_Bad(get_irn_n(n, i))) new_irn_arity++;
579 /* arity changing: set new predecessors without bad nodes */
580 if (new_irn_arity < old_irn_arity) {
581 /* get new predecessor array without Block predecessor */
582 new_in = NEW_ARR_D (ir_node *, current_ir_graph->obst, (new_irn_arity+1));
584 /* set new predeseccors in array */
587 for (i = 1; i < old_irn_arity; i++) {
588 irn = get_irn_n(n, i);
589 if (!is_Bad(irn)) new_in[new_irn_n++] = irn;
592 } /* ir node has bad predecessors */
594 } /* Block is not relinked */
598 * Relinks Bad predecesors from Bocks and Phis called by walker
599 * remove_bad_predecesors(). If n is a Block, call
600 * relink_bad_block_redecessors(). If n is a Phinode, call also the relinking
601 * function of Phi's Block. If this block has bad predecessors, relink preds
604 static void relink_bad_predecessors(ir_node *n, void *env) {
605 ir_node *block, **old_in;
606 int i, old_irn_arity, new_irn_arity;
608 /* relink bad predeseccors of a block */
609 if (get_irn_op(n) == op_Block)
610 relink_bad_block_predecessors(n, env);
612 /* If Phi node relink its block and its predecessors */
613 if (get_irn_op(n) == op_Phi) {
615 /* Relink predeseccors of phi's block */
616 block = get_nodes_block(n);
617 if (get_irn_link(block) == NULL)
618 relink_bad_block_predecessors(block, env);
620 old_in = (ir_node **)get_irn_link(block); /* Of Phi's Block */
621 old_irn_arity = ARR_LEN(old_in);
623 /* Relink Phi predeseccors if count of predeseccors changed */
624 if (old_irn_arity != ARR_LEN(get_irn_in(block))) {
625 /* set new predeseccors in array
626 n->in[0] remains the same block */
628 for(i = 1; i < old_irn_arity; i++)
629 if (!is_Bad((ir_node *)old_in[i])) n->in[new_irn_arity++] = n->in[i];
631 ARR_SETLEN(ir_node *, n->in, new_irn_arity);
634 } /* n is a Phi node */
638 * Removes Bad Bad predecesors from Blocks and the corresponding
639 * inputs to Phi nodes as in dead_node_elimination but without
641 * On walking up set the link field to NULL, on walking down call
642 * relink_bad_predecessors() (This function stores the old in array
643 * to the link field and sets a new in array if arity of predecessors
646 void remove_bad_predecessors(ir_graph *irg) {
647 irg_walk_graph(irg, init_link, relink_bad_predecessors, NULL);
651 /*--------------------------------------------------------------------*/
652 /* Funcionality for inlining */
653 /*--------------------------------------------------------------------*/
656 * Copy node for inlineing. Updates attributes that change when
657 * inlineing but not for dead node elimination.
659 * Copies the node by calling copy_node and then updates the entity if
660 * it's a local one. env must be a pointer of the frame type of the
661 * inlined procedure. The new entities must be in the link field of
665 copy_node_inline (ir_node *n, void *env) {
667 type *frame_tp = (type *)env;
670 if (get_irn_op(n) == op_Sel) {
671 new = get_new_node (n);
672 assert(get_irn_op(new) == op_Sel);
673 if (get_entity_owner(get_Sel_entity(n)) == frame_tp) {
674 set_Sel_entity(new, get_entity_link(get_Sel_entity(n)));
676 } else if (get_irn_op(n) == op_Block) {
677 new = get_new_node (n);
678 new->attr.block.irg = current_ir_graph;
682 static void find_addr(ir_node *node, void *env)
684 if (get_irn_opcode(node) == iro_Proj) {
685 if (get_Proj_proj(node) == pn_Start_P_value_arg_base)
691 * currently, we cannot inline two cases:
692 * - call with compound arguments
693 * - graphs that take the address of a parameter
695 * check these conditions here
697 static int can_inline(ir_node *call, ir_graph *called_graph)
699 type *call_type = get_Call_type(call);
700 int params, ress, i, res;
701 assert(is_method_type(call_type));
703 params = get_method_n_params(call_type);
704 ress = get_method_n_ress(call_type);
707 for (i = 0; i < params; ++i) {
708 type *p_type = get_method_param_type(call_type, i);
710 if (is_compound_type(p_type))
715 for (i = 0; i < ress; ++i) {
716 type *r_type = get_method_res_type(call_type, i);
718 if (is_compound_type(r_type))
723 irg_walk_graph(called_graph, find_addr, NULL, &res);
728 int inline_method(ir_node *call, ir_graph *called_graph) {
730 ir_node *post_call, *post_bl;
732 ir_node *end, *end_bl;
736 int arity, n_ret, n_exc, n_res, i, j, rem_opt, irn_arity;
739 irg_inline_property prop = get_irg_inline_property(called_graph);
741 if ( (prop != irg_inline_forced) &&
742 (!get_opt_optimize() || !get_opt_inline() || (prop == irg_inline_forbidden))) return 0;
744 /* Do not inline variadic functions. */
745 if (get_method_variadicity(get_entity_type(get_irg_entity(called_graph))) == variadicity_variadic)
748 assert(get_method_n_params(get_entity_type(get_irg_entity(called_graph))) ==
749 get_method_n_params(get_Call_type(call)));
752 * currently, we cannot inline two cases:
753 * - call with compound arguments
754 * - graphs that take the address of a parameter
756 if (! can_inline(call, called_graph))
759 /* -- Turn off optimizations, this can cause problems when allocating new nodes. -- */
760 rem_opt = get_opt_optimize();
763 /* Handle graph state */
764 assert(get_irg_phase_state(current_ir_graph) != phase_building);
765 assert(get_irg_pinned(current_ir_graph) == op_pin_state_pinned);
766 assert(get_irg_pinned(called_graph) == op_pin_state_pinned);
767 if (get_irg_outs_state(current_ir_graph) == outs_consistent)
768 set_irg_outs_inconsistent(current_ir_graph);
769 set_irg_loopinfo_inconsistent(current_ir_graph);
771 /* -- Check preconditions -- */
772 assert(get_irn_op(call) == op_Call);
773 /* @@@ does not work for InterfaceIII.java after cgana
774 assert(get_Call_type(call) == get_entity_type(get_irg_entity(called_graph)));
775 assert(smaller_type(get_entity_type(get_irg_entity(called_graph)),
776 get_Call_type(call)));
778 assert(get_type_tpop(get_Call_type(call)) == type_method);
779 if (called_graph == current_ir_graph) {
780 set_optimize(rem_opt);
784 /* here we know we WILL inline, so inform the statistics */
785 stat_inline(call, called_graph);
787 /* -- Decide how to handle exception control flow: Is there a handler
788 for the Call node, or do we branch directly to End on an exception?
790 0 There is a handler.
792 2 Exception handling not represented in Firm. -- */
794 ir_node *proj, *Mproj = NULL, *Xproj = NULL;
795 for (proj = (ir_node *)get_irn_link(call); proj; proj = (ir_node *)get_irn_link(proj)) {
796 assert(get_irn_op(proj) == op_Proj);
797 if (get_Proj_proj(proj) == pn_Call_X_except) Xproj = proj;
798 if (get_Proj_proj(proj) == pn_Call_M_except) Mproj = proj;
800 if (Mproj) { assert(Xproj); exc_handling = 0; } /* Mproj */
801 else if (Xproj) { exc_handling = 1; } /* !Mproj && Xproj */
802 else { exc_handling = 2; } /* !Mproj && !Xproj */
807 the procedure and later replaces the Start node of the called graph.
808 Post_call is the old Call node and collects the results of the called
809 graph. Both will end up being a tuple. -- */
810 post_bl = get_nodes_block(call);
811 set_irg_current_block(current_ir_graph, post_bl);
812 /* XxMxPxP of Start + parameter of Call */
813 in[pn_Start_X_initial_exec] = new_Jmp();
814 in[pn_Start_M] = get_Call_mem(call);
815 in[pn_Start_P_frame_base] = get_irg_frame(current_ir_graph);
816 in[pn_Start_P_globals] = get_irg_globals(current_ir_graph);
817 in[pn_Start_T_args] = new_Tuple(get_Call_n_params(call), get_Call_param_arr(call));
818 /* in[pn_Start_P_value_arg_base] = ??? */
819 pre_call = new_Tuple(5, in);
823 The new block gets the ins of the old block, pre_call and all its
824 predecessors and all Phi nodes. -- */
825 part_block(pre_call);
827 /* -- Prepare state for dead node elimination -- */
828 /* Visited flags in calling irg must be >= flag in called irg.
829 Else walker and arity computation will not work. */
830 if (get_irg_visited(current_ir_graph) <= get_irg_visited(called_graph))
831 set_irg_visited(current_ir_graph, get_irg_visited(called_graph)+1);
832 if (get_irg_block_visited(current_ir_graph)< get_irg_block_visited(called_graph))
833 set_irg_block_visited(current_ir_graph, get_irg_block_visited(called_graph));
834 /* Set pre_call as new Start node in link field of the start node of
835 calling graph and pre_calls block as new block for the start block
837 Further mark these nodes so that they are not visited by the
839 set_irn_link(get_irg_start(called_graph), pre_call);
840 set_irn_visited(get_irg_start(called_graph), get_irg_visited(current_ir_graph));
841 set_irn_link(get_irg_start_block(called_graph), get_nodes_block(pre_call));
842 set_irn_visited(get_irg_start_block(called_graph), get_irg_visited(current_ir_graph));
843 set_irn_link(get_irg_bad(called_graph), get_irg_bad(current_ir_graph));
844 set_irn_visited(get_irg_bad(called_graph), get_irg_visited(current_ir_graph));
846 /* Initialize for compaction of in arrays */
847 inc_irg_block_visited(current_ir_graph);
849 /* -- Replicate local entities of the called_graph -- */
850 /* copy the entities. */
851 called_frame = get_irg_frame_type(called_graph);
852 for (i = 0; i < get_class_n_members(called_frame); i++) {
853 entity *new_ent, *old_ent;
854 old_ent = get_class_member(called_frame, i);
855 new_ent = copy_entity_own(old_ent, get_cur_frame_type());
856 set_entity_link(old_ent, new_ent);
859 /* visited is > than that of called graph. With this trick visited will
860 remain unchanged so that an outer walker, e.g., searching the call nodes
861 to inline, calling this inline will not visit the inlined nodes. */
862 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
864 /* -- Performing dead node elimination inlines the graph -- */
865 /* Copies the nodes to the obstack of current_ir_graph. Updates links to new
867 /* @@@ endless loops are not copied!! -- they should be, I think... */
868 irg_walk(get_irg_end(called_graph), copy_node_inline, copy_preds,
869 get_irg_frame_type(called_graph));
871 /* Repair called_graph */
872 set_irg_visited(called_graph, get_irg_visited(current_ir_graph));
873 set_irg_block_visited(called_graph, get_irg_block_visited(current_ir_graph));
874 set_Block_block_visited(get_irg_start_block(called_graph), 0);
876 /* -- Merge the end of the inlined procedure with the call site -- */
877 /* We will turn the old Call node into a Tuple with the following
880 0: Phi of all Memories of Return statements.
881 1: Jmp from new Block that merges the control flow from all exception
882 predecessors of the old end block.
883 2: Tuple of all arguments.
884 3: Phi of Exception memories.
885 In case the old Call directly branches to End on an exception we don't
886 need the block merging all exceptions nor the Phi of the exception
890 /* -- Precompute some values -- */
891 end_bl = get_new_node(get_irg_end_block(called_graph));
892 end = get_new_node(get_irg_end(called_graph));
893 arity = get_irn_arity(end_bl); /* arity = n_exc + n_ret */
894 n_res = get_method_n_ress(get_Call_type(call));
896 res_pred = (ir_node **) malloc (n_res * sizeof (ir_node *));
897 cf_pred = (ir_node **) malloc (arity * sizeof (ir_node *));
899 set_irg_current_block(current_ir_graph, post_bl); /* just to make sure */
901 /* -- archive keepalives -- */
902 irn_arity = get_irn_arity(end);
903 for (i = 0; i < irn_arity; i++)
904 add_End_keepalive(get_irg_end(current_ir_graph), get_irn_n(end, i));
906 /* The new end node will die. We need not free as the in array is on the obstack:
907 copy_node only generated 'D' arrays. */
909 /* -- Replace Return nodes by Jump nodes. -- */
911 for (i = 0; i < arity; i++) {
913 ret = get_irn_n(end_bl, i);
914 if (get_irn_op(ret) == op_Return) {
915 cf_pred[n_ret] = new_r_Jmp(current_ir_graph, get_nodes_block(ret));
919 set_irn_in(post_bl, n_ret, cf_pred);
921 /* -- Build a Tuple for all results of the method.
922 Add Phi node if there was more than one Return. -- */
923 turn_into_tuple(post_call, 4);
924 /* First the Memory-Phi */
926 for (i = 0; i < arity; i++) {
927 ret = get_irn_n(end_bl, i);
928 if (get_irn_op(ret) == op_Return) {
929 cf_pred[n_ret] = get_Return_mem(ret);
933 phi = new_Phi(n_ret, cf_pred, mode_M);
934 set_Tuple_pred(call, pn_Call_M_regular, phi);
935 /* Conserve Phi-list for further inlinings -- but might be optimized */
936 if (get_nodes_block(phi) == post_bl) {
937 set_irn_link(phi, get_irn_link(post_bl));
938 set_irn_link(post_bl, phi);
940 /* Now the real results */
942 for (j = 0; j < n_res; j++) {
944 for (i = 0; i < arity; i++) {
945 ret = get_irn_n(end_bl, i);
946 if (get_irn_op(ret) == op_Return) {
947 cf_pred[n_ret] = get_Return_res(ret, j);
952 phi = new_Phi(n_ret, cf_pred, get_irn_mode(cf_pred[0]));
956 /* Conserve Phi-list for further inlinings -- but might be optimized */
957 if (get_nodes_block(phi) == post_bl) {
958 set_irn_link(phi, get_irn_link(post_bl));
959 set_irn_link(post_bl, phi);
962 set_Tuple_pred(call, pn_Call_T_result, new_Tuple(n_res, res_pred));
964 set_Tuple_pred(call, pn_Call_T_result, new_Bad());
966 /* Finally the exception control flow.
967 We have two (three) possible situations:
968 First if the Call branches to an exception handler: We need to add a Phi node to
969 collect the memory containing the exception objects. Further we need
970 to add another block to get a correct representation of this Phi. To
971 this block we add a Jmp that resolves into the X output of the Call
972 when the Call is turned into a tuple.
973 Second the Call branches to End, the exception is not handled. Just
974 add all inlined exception branches to the End node.
975 Third: there is no Exception edge at all. Handle as case two. */
976 if (exc_handling == 0) {
978 for (i = 0; i < arity; i++) {
980 ret = get_irn_n(end_bl, i);
981 if (is_fragile_op(skip_Proj(ret)) || (get_irn_op(skip_Proj(ret)) == op_Raise)) {
982 cf_pred[n_exc] = ret;
987 new_Block(n_exc, cf_pred); /* watch it: current_block is changed! */
988 set_Tuple_pred(call, pn_Call_X_except, new_Jmp());
989 /* The Phi for the memories with the exception objects */
991 for (i = 0; i < arity; i++) {
993 ret = skip_Proj(get_irn_n(end_bl, i));
994 if (get_irn_op(ret) == op_Call) {
995 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 3);
997 } else if (is_fragile_op(ret)) {
998 /* We rely that all cfops have the memory output at the same position. */
999 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 0);
1001 } else if (get_irn_op(ret) == op_Raise) {
1002 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 1);
1006 set_Tuple_pred(call, pn_Call_M_except, new_Phi(n_exc, cf_pred, mode_M));
1008 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1009 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1012 ir_node *main_end_bl;
1013 int main_end_bl_arity;
1014 ir_node **end_preds;
1016 /* assert(exc_handling == 1 || no exceptions. ) */
1018 for (i = 0; i < arity; i++) {
1019 ir_node *ret = get_irn_n(end_bl, i);
1021 if (is_fragile_op(skip_Proj(ret)) || (get_irn_op(skip_Proj(ret)) == op_Raise)) {
1022 cf_pred[n_exc] = ret;
1026 main_end_bl = get_irg_end_block(current_ir_graph);
1027 main_end_bl_arity = get_irn_arity(main_end_bl);
1028 end_preds = (ir_node **) malloc ((n_exc + main_end_bl_arity) * sizeof (ir_node *));
1030 for (i = 0; i < main_end_bl_arity; ++i)
1031 end_preds[i] = get_irn_n(main_end_bl, i);
1032 for (i = 0; i < n_exc; ++i)
1033 end_preds[main_end_bl_arity + i] = cf_pred[i];
1034 set_irn_in(main_end_bl, n_exc + main_end_bl_arity, end_preds);
1035 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1036 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1042 #if 0 /* old. now better, correcter, faster implementation. */
1044 /* -- If the exception control flow from the inlined Call directly
1045 branched to the end block we now have the following control
1046 flow predecessor pattern: ProjX -> Tuple -> Jmp. We must
1047 remove the Jmp along with it's empty block and add Jmp's
1048 predecessors as predecessors of this end block. No problem if
1049 there is no exception, because then branches Bad to End which
1051 @@@ can't we know this beforehand: by getting the Proj(1) from
1052 the Call link list and checking whether it goes to Proj. */
1053 /* find the problematic predecessor of the end block. */
1054 end_bl = get_irg_end_block(current_ir_graph);
1055 for (i = 0; i < get_Block_n_cfgpreds(end_bl); i++) {
1056 cf_op = get_Block_cfgpred(end_bl, i);
1057 if (get_irn_op(cf_op) == op_Proj) {
1058 cf_op = get_Proj_pred(cf_op);
1059 if ((get_irn_op(cf_op) == op_Tuple) && (cf_op == call)) {
1060 /* There are unoptimized tuples from inlineing before when no exc */
1061 assert(get_Proj_proj(get_Block_cfgpred(end_bl, i)) == pn_Call_X_except);
1062 cf_op = get_Tuple_pred(cf_op, pn_Call_X_except);
1063 assert(get_irn_op(cf_op) == op_Jmp);
1069 if (i < get_Block_n_cfgpreds(end_bl)) {
1070 bl = get_nodes_block(cf_op);
1071 arity = get_Block_n_cfgpreds(end_bl) + get_Block_n_cfgpreds(bl) - 1;
1072 cf_pred = (ir_node **) malloc (arity * sizeof (ir_node *));
1073 for (j = 0; j < i; j++)
1074 cf_pred[j] = get_Block_cfgpred(end_bl, j);
1075 for (j = j; j < i + get_Block_n_cfgpreds(bl); j++)
1076 cf_pred[j] = get_Block_cfgpred(bl, j-i);
1077 for (j = j; j < arity; j++)
1078 cf_pred[j] = get_Block_cfgpred(end_bl, j-get_Block_n_cfgpreds(bl) +1);
1079 set_irn_in(end_bl, arity, cf_pred);
1081 /* Remove the exception pred from post-call Tuple. */
1082 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1087 /* -- Turn cse back on. -- */
1088 set_optimize(rem_opt);
1093 /********************************************************************/
1094 /* Apply inlineing to small methods. */
1095 /********************************************************************/
1097 /* It makes no sense to inline too many calls in one procedure. Anyways,
1098 I didn't get a version with NEW_ARR_F to run. */
1099 #define MAX_INLINE 1024
1102 * environment for inlining small irgs
1104 typedef struct _inline_env_t {
1106 ir_node *calls[MAX_INLINE];
1110 * Returns the irg called from a Call node. If the irg is not
1111 * known, NULL is returned.
1113 static ir_graph *get_call_called_irg(ir_node *call) {
1115 ir_graph *called_irg = NULL;
1117 assert(get_irn_op(call) == op_Call);
1119 addr = get_Call_ptr(call);
1120 if ((get_irn_op(addr) == op_SymConst) && (get_SymConst_kind (addr) == symconst_addr_ent)) {
1121 called_irg = get_entity_irg(get_SymConst_entity(addr));
1127 static void collect_calls(ir_node *call, void *env) {
1130 if (get_irn_op(call) != op_Call) return;
1132 addr = get_Call_ptr(call);
1134 if (get_irn_op(addr) == op_SymConst) {
1135 if (get_SymConst_kind(addr) == symconst_addr_ent) {
1136 ir_graph *called_irg = get_entity_irg(get_SymConst_entity(addr));
1137 inline_env_t *ienv = (inline_env_t *)env;
1138 if (called_irg && ienv->pos < MAX_INLINE) {
1139 /* The Call node calls a locally defined method. Remember to inline. */
1140 ienv->calls[ienv->pos++] = call;
1147 * Inlines all small methods at call sites where the called address comes
1148 * from a Const node that references the entity representing the called
1150 * The size argument is a rough measure for the code size of the method:
1151 * Methods where the obstack containing the firm graph is smaller than
1154 void inline_small_irgs(ir_graph *irg, int size) {
1156 ir_graph *rem = current_ir_graph;
1157 inline_env_t env /* = {0, NULL}*/;
1159 if (!(get_opt_optimize() && get_opt_inline())) return;
1161 current_ir_graph = irg;
1162 /* Handle graph state */
1163 assert(get_irg_phase_state(current_ir_graph) != phase_building);
1164 free_callee_info(current_ir_graph);
1166 /* Find Call nodes to inline.
1167 (We can not inline during a walk of the graph, as inlineing the same
1168 method several times changes the visited flag of the walked graph:
1169 after the first inlineing visited of the callee equals visited of
1170 the caller. With the next inlineing both are increased.) */
1172 irg_walk(get_irg_end(irg), NULL, collect_calls, &env);
1174 if ((env.pos > 0) && (env.pos < MAX_INLINE)) {
1175 /* There are calls to inline */
1176 collect_phiprojs(irg);
1177 for (i = 0; i < env.pos; i++) {
1179 callee = get_entity_irg(get_SymConst_entity(get_Call_ptr(env.calls[i])));
1180 if (((_obstack_memory_used(callee->obst) - obstack_room(callee->obst)) < size) ||
1181 (get_irg_inline_property(callee) == irg_inline_forced)) {
1182 inline_method(env.calls[i], callee);
1187 current_ir_graph = rem;
1191 * Environment for inlining irgs.
1194 int n_nodes; /**< Nodes in graph except Id, Tuple, Proj, Start, End */
1195 int n_nodes_orig; /**< for statistics */
1196 eset *call_nodes; /**< All call nodes in this graph */
1198 int n_call_nodes_orig; /**< for statistics */
1199 int n_callers; /**< Number of known graphs that call this graphs. */
1200 int n_callers_orig; /**< for statistics */
1203 static inline_irg_env *new_inline_irg_env(void) {
1204 inline_irg_env *env = malloc(sizeof(inline_irg_env));
1205 env->n_nodes = -2; /* uncount Start, End */
1206 env->n_nodes_orig = -2; /* uncount Start, End */
1207 env->call_nodes = eset_create();
1208 env->n_call_nodes = 0;
1209 env->n_call_nodes_orig = 0;
1211 env->n_callers_orig = 0;
1215 static void free_inline_irg_env(inline_irg_env *env) {
1216 eset_destroy(env->call_nodes);
1220 static void collect_calls2(ir_node *call, void *env) {
1221 inline_irg_env *x = (inline_irg_env *)env;
1222 ir_op *op = get_irn_op(call);
1225 /* count nodes in irg */
1226 if (op != op_Proj && op != op_Tuple && op != op_Sync) {
1231 if (op != op_Call) return;
1233 /* collect all call nodes */
1234 eset_insert(x->call_nodes, (void *)call);
1236 x->n_call_nodes_orig++;
1238 /* count all static callers */
1239 callee = get_call_called_irg(call);
1241 ((inline_irg_env *)get_irg_link(callee))->n_callers++;
1242 ((inline_irg_env *)get_irg_link(callee))->n_callers_orig++;
1246 INLINE static int is_leave(ir_graph *irg) {
1247 return (((inline_irg_env *)get_irg_link(irg))->n_call_nodes == 0);
1250 INLINE static int is_smaller(ir_graph *callee, int size) {
1251 return (((inline_irg_env *)get_irg_link(callee))->n_nodes < size);
1256 * Inlines small leave methods at call sites where the called address comes
1257 * from a Const node that references the entity representing the called
1259 * The size argument is a rough measure for the code size of the method:
1260 * Methods where the obstack containing the firm graph is smaller than
1263 void inline_leave_functions(int maxsize, int leavesize, int size) {
1264 inline_irg_env *env;
1265 int i, n_irgs = get_irp_n_irgs();
1266 ir_graph *rem = current_ir_graph;
1269 if (!(get_opt_optimize() && get_opt_inline())) return;
1271 /* extend all irgs by a temporary data structure for inlineing. */
1272 for (i = 0; i < n_irgs; ++i)
1273 set_irg_link(get_irp_irg(i), new_inline_irg_env());
1275 /* Precompute information in temporary data structure. */
1276 for (i = 0; i < n_irgs; ++i) {
1277 current_ir_graph = get_irp_irg(i);
1278 assert(get_irg_phase_state(current_ir_graph) != phase_building);
1279 free_callee_info(current_ir_graph);
1281 irg_walk(get_irg_end(current_ir_graph), NULL, collect_calls2,
1282 get_irg_link(current_ir_graph));
1285 /* -- and now inline. -- */
1287 /* Inline leaves recursively -- we might construct new leaves. */
1288 while (did_inline) {
1291 for (i = 0; i < n_irgs; ++i) {
1293 int phiproj_computed = 0;
1295 current_ir_graph = get_irp_irg(i);
1296 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1298 for (call = eset_first(env->call_nodes); call; call = eset_next(env->call_nodes)) {
1299 if (get_irn_op(call) == op_Tuple) continue; /* We already inlined. */
1300 ir_graph *callee = get_call_called_irg(call);
1302 if (env->n_nodes > maxsize) continue; // break;
1304 if (callee && (is_leave(callee) && is_smaller(callee, leavesize))) {
1305 if (!phiproj_computed) {
1306 phiproj_computed = 1;
1307 collect_phiprojs(current_ir_graph);
1309 did_inline = inline_method(call, callee);
1312 /* Do some statistics */
1313 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1314 env->n_call_nodes --;
1315 env->n_nodes += callee_env->n_nodes;
1316 callee_env->n_callers--;
1323 /* inline other small functions. */
1324 for (i = 0; i < n_irgs; ++i) {
1327 int phiproj_computed = 0;
1329 current_ir_graph = get_irp_irg(i);
1330 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1332 /* we can not walk and change a set, nor remove from it.
1334 walkset = env->call_nodes;
1335 env->call_nodes = eset_create();
1336 for (call = eset_first(walkset); call; call = eset_next(walkset)) {
1337 if (get_irn_op(call) == op_Tuple) continue; /* We already inlined. */
1338 ir_graph *callee = get_call_called_irg(call);
1341 ((is_smaller(callee, size) && (env->n_nodes < maxsize)) || /* small function */
1342 (get_irg_inline_property(callee) == irg_inline_forced))) {
1343 if (!phiproj_computed) {
1344 phiproj_computed = 1;
1345 collect_phiprojs(current_ir_graph);
1347 if (inline_method(call, callee)) {
1348 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1349 env->n_call_nodes--;
1350 eset_insert_all(env->call_nodes, callee_env->call_nodes); /* @@@ ??? This are the wrong nodes !? Not the copied ones. */
1351 env->n_call_nodes += callee_env->n_call_nodes;
1352 env->n_nodes += callee_env->n_nodes;
1353 callee_env->n_callers--;
1356 eset_insert(env->call_nodes, call);
1359 eset_destroy(walkset);
1362 for (i = 0; i < n_irgs; ++i) {
1363 current_ir_graph = get_irp_irg(i);
1365 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1366 if ((env->n_call_nodes_orig != env->n_call_nodes) ||
1367 (env->n_callers_orig != env->n_callers))
1368 printf("Nodes:%3d ->%3d, calls:%3d ->%3d, callers:%3d ->%3d, -- %s\n",
1369 env->n_nodes_orig, env->n_nodes, env->n_call_nodes_orig, env->n_call_nodes,
1370 env->n_callers_orig, env->n_callers,
1371 get_entity_name(get_irg_entity(current_ir_graph)));
1373 free_inline_irg_env((inline_irg_env *)get_irg_link(current_ir_graph));
1376 current_ir_graph = rem;
1379 /*******************************************************************/
1380 /* Code Placement. Pins all floating nodes to a block where they */
1381 /* will be executed only if needed. */
1382 /*******************************************************************/
1385 * Find the earliest correct block for N. --- Place N into the
1386 * same Block as its dominance-deepest Input.
1389 place_floats_early(ir_node *n, pdeq *worklist)
1391 int i, start, irn_arity;
1393 /* we must not run into an infinite loop */
1394 assert (irn_not_visited(n));
1395 mark_irn_visited(n);
1397 /* Place floating nodes. */
1398 if (get_irn_pinned(n) == op_pin_state_floats) {
1400 ir_node *b = new_Bad(); /* The block to place this node in */
1401 int bad_recursion = is_Bad(get_nodes_block(n));
1403 assert(get_irn_op(n) != op_Block);
1405 if ((get_irn_op(n) == op_Const) ||
1406 (get_irn_op(n) == op_SymConst) ||
1408 (get_irn_op(n) == op_Unknown)) {
1409 /* These nodes will not be placed by the loop below. */
1410 b = get_irg_start_block(current_ir_graph);
1414 /* find the block for this node. */
1415 irn_arity = get_irn_arity(n);
1416 for (i = 0; i < irn_arity; i++) {
1417 ir_node *dep = get_irn_n(n, i);
1420 if ((irn_not_visited(dep))
1421 && (get_irn_pinned(dep) == op_pin_state_floats)) {
1422 place_floats_early(dep, worklist);
1426 * A node in the Bad block must stay in the bad block,
1427 * so don't compute a new block for it.
1432 /* Because all loops contain at least one op_pin_state_pinned node, now all
1433 our inputs are either op_pin_state_pinned or place_early has already
1434 been finished on them. We do not have any unfinished inputs! */
1435 dep_block = get_nodes_block(dep);
1436 if ((!is_Bad(dep_block)) &&
1437 (get_Block_dom_depth(dep_block) > depth)) {
1439 depth = get_Block_dom_depth(dep_block);
1441 /* Avoid that the node is placed in the Start block */
1442 if ((depth == 1) && (get_Block_dom_depth(get_nodes_block(n)) > 1)) {
1443 b = get_Block_cfg_out(get_irg_start_block(current_ir_graph), 0);
1444 assert(b != get_irg_start_block(current_ir_graph));
1448 set_nodes_block(n, b);
1451 /* Add predecessors of non floating nodes on worklist. */
1452 start = (get_irn_op(n) == op_Block) ? 0 : -1;
1453 irn_arity = get_irn_arity(n);
1454 for (i = start; i < irn_arity; i++) {
1455 ir_node *pred = get_irn_n(n, i);
1456 if (irn_not_visited(pred)) {
1457 pdeq_putr (worklist, pred);
1463 * Floating nodes form subgraphs that begin at nodes as Const, Load,
1464 * Start, Call and that end at op_pin_state_pinned nodes as Store, Call. Place_early
1465 * places all floating nodes reachable from its argument through floating
1466 * nodes and adds all beginnings at op_pin_state_pinned nodes to the worklist.
1468 static INLINE void place_early(pdeq *worklist) {
1470 inc_irg_visited(current_ir_graph);
1472 /* this inits the worklist */
1473 place_floats_early(get_irg_end(current_ir_graph), worklist);
1475 /* Work the content of the worklist. */
1476 while (!pdeq_empty (worklist)) {
1477 ir_node *n = pdeq_getl (worklist);
1478 if (irn_not_visited(n)) place_floats_early(n, worklist);
1481 set_irg_outs_inconsistent(current_ir_graph);
1482 current_ir_graph->op_pin_state_pinned = op_pin_state_pinned;
1486 /** Deepest common dominance ancestor of DCA and CONSUMER of PRODUCER.
1487 * I.e., DCA is the block where we might place PRODUCER.
1488 * A data flow edge points from producer to consumer.
1491 consumer_dom_dca (ir_node *dca, ir_node *consumer, ir_node *producer)
1493 ir_node *block = NULL;
1495 /* Compute the latest block into which we can place a node so that it is
1497 if (get_irn_op(consumer) == op_Phi) {
1498 /* our consumer is a Phi-node, the effective use is in all those
1499 blocks through which the Phi-node reaches producer */
1501 ir_node *phi_block = get_nodes_block(consumer);
1502 irn_arity = get_irn_arity(consumer);
1504 for (i = 0; i < irn_arity; i++) {
1505 if (get_irn_n(consumer, i) == producer) {
1506 block = get_nodes_block(get_Block_cfgpred(phi_block, i));
1510 assert(is_no_Block(consumer));
1511 block = get_nodes_block(consumer);
1514 /* Compute the deepest common ancestor of block and dca. */
1516 if (!dca) return block;
1517 while (get_Block_dom_depth(block) > get_Block_dom_depth(dca))
1518 block = get_Block_idom(block);
1519 while (get_Block_dom_depth(dca) > get_Block_dom_depth(block)) {
1520 dca = get_Block_idom(dca);
1522 while (block != dca)
1523 { block = get_Block_idom(block); dca = get_Block_idom(dca); }
1528 static INLINE int get_irn_loop_depth(ir_node *n) {
1529 return get_loop_depth(get_irn_loop(n));
1533 * Move n to a block with less loop depth than it's current block. The
1534 * new block must be dominated by early.
1537 move_out_of_loops (ir_node *n, ir_node *early)
1539 ir_node *best, *dca;
1543 /* Find the region deepest in the dominator tree dominating
1544 dca with the least loop nesting depth, but still dominated
1545 by our early placement. */
1546 dca = get_nodes_block(n);
1548 while (dca != early) {
1549 dca = get_Block_idom(dca);
1550 if (!dca || is_Bad(dca)) break; /* may be Bad if not reachable from Start */
1551 if (get_irn_loop_depth(dca) < get_irn_loop_depth(best)) {
1555 if (best != get_nodes_block(n)) {
1557 printf("Moving out of loop: "); DDMN(n);
1558 printf(" Outermost block: "); DDMN(early);
1559 printf(" Best block: "); DDMN(best);
1560 printf(" Innermost block: "); DDMN(get_nodes_block(n));
1562 set_nodes_block(n, best);
1567 * Find the latest legal block for N and place N into the
1568 * `optimal' Block between the latest and earliest legal block.
1569 * The `optimal' block is the dominance-deepest block of those
1570 * with the least loop-nesting-depth. This places N out of as many
1571 * loops as possible and then makes it as control dependant as
1575 place_floats_late(ir_node *n, pdeq *worklist)
1580 assert (irn_not_visited(n)); /* no multiple placement */
1582 mark_irn_visited(n);
1584 /* no need to place block nodes, control nodes are already placed. */
1585 if ((get_irn_op(n) != op_Block) &&
1587 (get_irn_mode(n) != mode_X)) {
1588 /* Remember the early placement of this block to move it
1589 out of loop no further than the early placement. */
1590 early = get_nodes_block(n);
1592 /* Do not move code not reachable from Start. For
1593 * these we could not compute dominator information. */
1594 if (is_Bad(early) || get_Block_dom_depth(early) == -1)
1597 /* Assure that our users are all placed, except the Phi-nodes.
1598 --- Each data flow cycle contains at least one Phi-node. We
1599 have to break the `user has to be placed before the
1600 producer' dependence cycle and the Phi-nodes are the
1601 place to do so, because we need to base our placement on the
1602 final region of our users, which is OK with Phi-nodes, as they
1603 are op_pin_state_pinned, and they never have to be placed after a
1604 producer of one of their inputs in the same block anyway. */
1605 for (i = 0; i < get_irn_n_outs(n); i++) {
1606 ir_node *succ = get_irn_out(n, i);
1607 if (irn_not_visited(succ) && (get_irn_op(succ) != op_Phi))
1608 place_floats_late(succ, worklist);
1611 /* We have to determine the final block of this node... except for
1613 if ((get_irn_pinned(n) == op_pin_state_floats) &&
1614 (get_irn_op(n) != op_Const) &&
1615 (get_irn_op(n) != op_SymConst)) {
1616 ir_node *dca = NULL; /* deepest common ancestor in the
1617 dominator tree of all nodes'
1618 blocks depending on us; our final
1619 placement has to dominate DCA. */
1620 for (i = 0; i < get_irn_n_outs(n); i++) {
1621 ir_node *out = get_irn_out(n, i);
1622 /* ignore if out is in dead code */
1623 ir_node *outbl = get_nodes_block(out);
1624 if (is_Bad(outbl) || get_Block_dom_depth(outbl) == -1)
1626 dca = consumer_dom_dca (dca, out, n);
1629 set_nodes_block(n, dca);
1631 move_out_of_loops (n, early);
1633 /* else all outs are in dead code */
1637 /* Add predecessors of all non-floating nodes on list. (Those of floating
1638 nodes are placeded already and therefore are marked.) */
1639 for (i = 0; i < get_irn_n_outs(n); i++) {
1640 if (irn_not_visited(get_irn_out(n, i))) {
1641 pdeq_putr (worklist, get_irn_out(n, i));
1646 static INLINE void place_late(pdeq *worklist) {
1648 inc_irg_visited(current_ir_graph);
1650 /* This fills the worklist initially. */
1651 place_floats_late(get_irg_start_block(current_ir_graph), worklist);
1653 /* And now empty the worklist again... */
1654 while (!pdeq_empty (worklist)) {
1655 ir_node *n = pdeq_getl (worklist);
1656 if (irn_not_visited(n)) place_floats_late(n, worklist);
1660 void place_code(ir_graph *irg) {
1662 ir_graph *rem = current_ir_graph;
1664 current_ir_graph = irg;
1666 if (!(get_opt_optimize() && get_opt_global_cse())) return;
1668 /* Handle graph state */
1669 assert(get_irg_phase_state(irg) != phase_building);
1670 if (get_irg_dom_state(irg) != dom_consistent)
1673 if (1 || get_irg_loopinfo_state(irg) != loopinfo_consistent) {
1674 free_loop_information(irg);
1675 construct_backedges(irg);
1678 /* Place all floating nodes as early as possible. This guarantees
1679 a legal code placement. */
1680 worklist = new_pdeq();
1681 place_early(worklist);
1683 /* place_early invalidates the outs, place_late needs them. */
1685 /* Now move the nodes down in the dominator tree. This reduces the
1686 unnecessary executions of the node. */
1687 place_late(worklist);
1689 set_irg_outs_inconsistent(current_ir_graph);
1690 set_irg_loopinfo_inconsistent(current_ir_graph);
1692 current_ir_graph = rem;
1696 * Called by walker of remove_critical_cf_edges().
1698 * Place an empty block to an edge between a blocks of multiple
1699 * predecessors and a block of multiple successors.
1702 * @param env Environment of walker. This field is unused and has
1705 static void walk_critical_cf_edges(ir_node *n, void *env) {
1707 ir_node *pre, *block, **in, *jmp;
1709 /* Block has multiple predecessors */
1710 if ((op_Block == get_irn_op(n)) &&
1711 (get_irn_arity(n) > 1)) {
1712 arity = get_irn_arity(n);
1714 if (n == get_irg_end_block(current_ir_graph))
1715 return; /* No use to add a block here. */
1717 for (i=0; i<arity; i++) {
1718 pre = get_irn_n(n, i);
1719 /* Predecessor has multiple successors. Insert new flow edge */
1720 if ((NULL != pre) &&
1721 (op_Proj == get_irn_op(pre)) &&
1722 op_Raise != get_irn_op(skip_Proj(pre))) {
1724 /* set predecessor array for new block */
1725 in = NEW_ARR_D (ir_node *, current_ir_graph->obst, 1);
1726 /* set predecessor of new block */
1728 block = new_Block(1, in);
1729 /* insert new jmp node to new block */
1730 set_cur_block(block);
1733 /* set successor of new block */
1734 set_irn_n(n, i, jmp);
1736 } /* predecessor has multiple successors */
1737 } /* for all predecessors */
1738 } /* n is a block */
1741 void remove_critical_cf_edges(ir_graph *irg) {
1742 if (get_opt_critical_edges())
1743 irg_walk_graph(irg, NULL, walk_critical_cf_edges, NULL);