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, 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.
209 copy_node (ir_node *n, void *env) {
212 opcode op = get_irn_opcode(n);
213 /* The end node looses it's flexible in array. This doesn't matter,
214 as dead node elimination builds End by hand, inlineing doesn't use
216 /* assert(n->op == op_End || ((_ARR_DESCR(n->in))->cookie != ARR_F_MAGIC)); */
219 /* node copied already */
221 } else if (op == iro_Block) {
223 new_arity = compute_new_arity(n);
224 n->attr.block.graph_arr = NULL;
226 block = get_nodes_Block(n);
227 if (get_irn_opcode(n) == iro_Phi) {
228 new_arity = compute_new_arity(block);
230 new_arity = get_irn_arity(n);
233 nn = new_ir_node(get_irn_dbg_info(n),
240 /* Copy the attributes. These might point to additional data. If this
241 was allocated on the old obstack the pointers now are dangling. This
242 frees e.g. the memory of the graph_arr allocated in new_immBlock. */
244 new_backedge_info(nn);
247 /* printf("\n old node: "); DDMSG2(n);
248 printf(" new node: "); DDMSG2(nn); */
253 * Copies new predecessors of old node to new node remembered in link.
254 * Spare the Bad predecessors of Phi and Block nodes.
257 copy_preds (ir_node *n, void *env) {
261 nn = get_new_node(n);
263 /* printf("\n old node: "); DDMSG2(n);
264 printf(" new node: "); DDMSG2(nn);
265 printf(" arities: old: %d, new: %d\n", get_irn_arity(n), get_irn_arity(nn)); */
267 if (get_irn_opcode(n) == iro_Block) {
268 /* Don't copy Bad nodes. */
270 irn_arity = get_irn_arity(n);
271 for (i = 0; i < irn_arity; i++)
272 if (get_irn_opcode(get_irn_n(n, i)) != iro_Bad) {
273 set_irn_n (nn, j, get_new_node(get_irn_n(n, i)));
274 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
277 /* repair the block visited flag from above misuse. Repair it in both
278 graphs so that the old one can still be used. */
279 set_Block_block_visited(nn, 0);
280 set_Block_block_visited(n, 0);
281 /* Local optimization could not merge two subsequent blocks if
282 in array contained Bads. Now it's possible.
283 We don't call optimize_in_place as it requires
284 that the fields in ir_graph are set properly. */
285 if ((get_opt_control_flow_straightening()) &&
286 (get_Block_n_cfgpreds(nn) == 1) &&
287 (get_irn_op(get_Block_cfgpred(nn, 0)) == op_Jmp)) {
288 ir_node *old = get_nodes_Block(get_Block_cfgpred(nn, 0));
290 /* Jmp jumps into the block it is in -- deal self cycle. */
291 assert(is_Bad(get_new_node(get_irg_bad(current_ir_graph))));
292 exchange(nn, get_new_node(get_irg_bad(current_ir_graph)));
297 } else if (get_irn_opcode(n) == iro_Phi) {
298 /* Don't copy node if corresponding predecessor in block is Bad.
299 The Block itself should not be Bad. */
300 block = get_nodes_Block(n);
301 set_irn_n (nn, -1, get_new_node(block));
303 irn_arity = get_irn_arity(n);
304 for (i = 0; i < irn_arity; i++)
305 if (get_irn_opcode(get_irn_n(block, i)) != iro_Bad) {
306 set_irn_n (nn, j, get_new_node(get_irn_n(n, i)));
307 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
310 /* If the pre walker reached this Phi after the post walker visited the
311 block block_visited is > 0. */
312 set_Block_block_visited(get_nodes_Block(n), 0);
313 /* Compacting the Phi's ins might generate Phis with only one
315 if (get_irn_arity(n) == 1)
316 exchange(n, get_irn_n(n, 0));
318 irn_arity = get_irn_arity(n);
319 for (i = -1; i < irn_arity; i++)
320 set_irn_n (nn, i, get_new_node(get_irn_n(n, i)));
322 /* Now the new node is complete. We can add it to the hash table for cse.
323 @@@ inlinening aborts if we identify End. Why? */
324 if(get_irn_op(nn) != op_End)
325 add_identities (current_ir_graph->value_table, nn);
329 * Copies the graph recursively, compacts the keepalive of the end node.
333 ir_node *oe, *ne, *ob, *nb; /* old end, new end, old bad, new bad */
334 ir_node *ka; /* keep alive */
337 oe = get_irg_end(current_ir_graph);
338 /* copy the end node by hand, allocate dynamic in array! */
339 ne = new_ir_node(get_irn_dbg_info(oe),
346 /* Copy the attributes. Well, there might be some in the future... */
348 set_new_node(oe, ne);
350 ob = get_irg_bad(current_ir_graph);
351 nb = new_ir_node(get_irn_dbg_info(ob),
358 set_new_node(ob, nb);
360 /* copy the live nodes */
361 irg_walk(get_nodes_Block(oe), copy_node, copy_preds, NULL);
362 /* copy_preds for the end node ... */
363 set_nodes_Block(ne, get_new_node(get_nodes_Block(oe)));
364 set_nodes_Block(nb, get_new_node(get_nodes_Block(ob)));
366 /*- ... and now the keep alives. -*/
367 /* First pick the not marked block nodes and walk them. We must pick these
368 first as else we will oversee blocks reachable from Phis. */
369 irn_arity = get_irn_arity(oe);
370 for (i = 0; i < irn_arity; i++) {
371 ka = get_irn_intra_n(oe, i);
372 if ((get_irn_op(ka) == op_Block) &&
373 (get_irn_visited(ka) < get_irg_visited(current_ir_graph))) {
374 /* We must keep the block alive and copy everything reachable */
375 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
376 irg_walk(ka, copy_node, copy_preds, NULL);
377 add_End_keepalive(ne, get_new_node(ka));
381 /* Now pick the Phis. Here we will keep all! */
382 irn_arity = get_irn_arity(oe);
383 for (i = 0; i < irn_arity; i++) {
384 ka = get_irn_intra_n(oe, i);
385 if ((get_irn_op(ka) == op_Phi)) {
386 if (get_irn_visited(ka) < get_irg_visited(current_ir_graph)) {
387 /* We didn't copy the Phi yet. */
388 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
389 irg_walk(ka, copy_node, copy_preds, NULL);
391 add_End_keepalive(ne, get_new_node(ka));
397 * Copies the graph reachable from current_ir_graph->end to the obstack
398 * in current_ir_graph and fixes the environment.
399 * Then fixes the fields in current_ir_graph containing nodes of the
403 copy_graph_env (void) {
405 /* Not all nodes remembered in current_ir_graph might be reachable
406 from the end node. Assure their link is set to NULL, so that
407 we can test whether new nodes have been computed. */
408 set_irn_link(get_irg_frame (current_ir_graph), NULL);
409 set_irn_link(get_irg_globals (current_ir_graph), NULL);
410 set_irn_link(get_irg_args (current_ir_graph), NULL);
411 set_irn_link(get_irg_initial_mem(current_ir_graph), NULL);
413 /* we use the block walk flag for removing Bads from Blocks ins. */
414 inc_irg_block_visited(current_ir_graph);
419 /* fix the fields in current_ir_graph */
420 old_end = get_irg_end(current_ir_graph);
421 set_irg_end (current_ir_graph, get_new_node(old_end));
422 set_irg_end_except (current_ir_graph, get_irg_end(current_ir_graph));
423 set_irg_end_reg (current_ir_graph, get_irg_end(current_ir_graph));
425 set_irg_end_block (current_ir_graph, get_new_node(get_irg_end_block(current_ir_graph)));
426 if (get_irn_link(get_irg_frame(current_ir_graph)) == NULL) {
427 copy_node (get_irg_frame(current_ir_graph), NULL);
428 copy_preds(get_irg_frame(current_ir_graph), NULL);
430 if (get_irn_link(get_irg_globals(current_ir_graph)) == NULL) {
431 copy_node (get_irg_globals(current_ir_graph), NULL);
432 copy_preds(get_irg_globals(current_ir_graph), NULL);
434 if (get_irn_link(get_irg_initial_mem(current_ir_graph)) == NULL) {
435 copy_node (get_irg_initial_mem(current_ir_graph), NULL);
436 copy_preds(get_irg_initial_mem(current_ir_graph), NULL);
438 if (get_irn_link(get_irg_args(current_ir_graph)) == NULL) {
439 copy_node (get_irg_args(current_ir_graph), NULL);
440 copy_preds(get_irg_args(current_ir_graph), NULL);
442 set_irg_start (current_ir_graph, get_new_node(get_irg_start(current_ir_graph)));
444 set_irg_start_block(current_ir_graph,
445 get_new_node(get_irg_start_block(current_ir_graph)));
446 set_irg_frame (current_ir_graph, get_new_node(get_irg_frame(current_ir_graph)));
447 set_irg_globals (current_ir_graph, get_new_node(get_irg_globals(current_ir_graph)));
448 set_irg_initial_mem(current_ir_graph, get_new_node(get_irg_initial_mem(current_ir_graph)));
449 set_irg_args (current_ir_graph, get_new_node(get_irg_args(current_ir_graph)));
451 if (get_irn_link(get_irg_bad(current_ir_graph)) == NULL) {
452 copy_node(get_irg_bad(current_ir_graph), NULL);
453 copy_preds(get_irg_bad(current_ir_graph), NULL);
455 set_irg_bad(current_ir_graph, get_new_node(get_irg_bad(current_ir_graph)));
459 * Copies all reachable nodes to a new obstack. Removes bad inputs
460 * from block nodes and the corresponding inputs from Phi nodes.
461 * Merges single exit blocks with single entry blocks and removes
463 * Adds all new nodes to a new hash table for cse. Does not
464 * perform cse, so the hash table might contain common subexpressions.
467 dead_node_elimination(ir_graph *irg) {
469 int rem_ipview = interprocedural_view;
470 struct obstack *graveyard_obst = NULL;
471 struct obstack *rebirth_obst = NULL;
473 /* inform statistics that we started a dead-node elimination run */
474 stat_dead_node_elim_start(irg);
476 /* Remember external state of current_ir_graph. */
477 rem = current_ir_graph;
478 current_ir_graph = irg;
479 interprocedural_view = 0;
481 /* Handle graph state */
482 assert(get_irg_phase_state(current_ir_graph) != phase_building);
483 free_callee_info(current_ir_graph);
484 free_outs(current_ir_graph);
485 /* @@@ so far we loose loops when copying */
486 free_loop_information(current_ir_graph);
488 if (get_opt_optimize() && get_opt_dead_node_elimination()) {
490 /* A quiet place, where the old obstack can rest in peace,
491 until it will be cremated. */
492 graveyard_obst = irg->obst;
494 /* A new obstack, where the reachable nodes will be copied to. */
495 rebirth_obst = (struct obstack *) xmalloc (sizeof (struct obstack));
496 current_ir_graph->obst = rebirth_obst;
497 obstack_init (current_ir_graph->obst);
499 /* We also need a new hash table for cse */
500 del_identities (irg->value_table);
501 irg->value_table = new_identities ();
503 /* Copy the graph from the old to the new obstack */
506 /* Free memory from old unoptimized obstack */
507 obstack_free(graveyard_obst, 0); /* First empty the obstack ... */
508 xfree (graveyard_obst); /* ... then free it. */
511 /* inform statistics that the run is over */
512 stat_dead_node_elim_stop(irg);
514 current_ir_graph = rem;
515 interprocedural_view = rem_ipview;
519 * Relink bad predeseccors of a block and store the old in array to the
520 * link field. This function is called by relink_bad_predecessors().
521 * The array of link field starts with the block operand at position 0.
522 * If block has bad predecessors, create a new in array without bad preds.
523 * Otherwise let in array untouched.
525 static void relink_bad_block_predecessors(ir_node *n, void *env) {
526 ir_node **new_in, *irn;
527 int i, new_irn_n, old_irn_arity, new_irn_arity = 0;
529 /* if link field of block is NULL, look for bad predecessors otherwise
530 this is allready done */
531 if (get_irn_op(n) == op_Block &&
532 get_irn_link(n) == NULL) {
534 /* save old predecessors in link field (position 0 is the block operand)*/
535 set_irn_link(n, (void *)get_irn_in(n));
537 /* count predecessors without bad nodes */
538 old_irn_arity = get_irn_arity(n);
539 for (i = 0; i < old_irn_arity; i++)
540 if (!is_Bad(get_irn_n(n, i))) new_irn_arity++;
542 /* arity changing: set new predecessors without bad nodes */
543 if (new_irn_arity < old_irn_arity) {
544 /* get new predecessor array without Block predecessor */
545 new_in = NEW_ARR_D (ir_node *, current_ir_graph->obst, (new_irn_arity+1));
547 /* set new predeseccors in array */
550 for (i = 1; i < old_irn_arity; i++) {
551 irn = get_irn_n(n, i);
552 if (!is_Bad(irn)) new_in[new_irn_n++] = irn;
555 } /* ir node has bad predecessors */
557 } /* Block is not relinked */
561 * Relinks Bad predecesors from Bocks and Phis called by walker
562 * remove_bad_predecesors(). If n is a Block, call
563 * relink_bad_block_redecessors(). If n is a Phinode, call also the relinking
564 * function of Phi's Block. If this block has bad predecessors, relink preds
567 static void relink_bad_predecessors(ir_node *n, void *env) {
568 ir_node *block, **old_in;
569 int i, old_irn_arity, new_irn_arity;
571 /* relink bad predeseccors of a block */
572 if (get_irn_op(n) == op_Block)
573 relink_bad_block_predecessors(n, env);
575 /* If Phi node relink its block and its predecessors */
576 if (get_irn_op(n) == op_Phi) {
578 /* Relink predeseccors of phi's block */
579 block = get_nodes_Block(n);
580 if (get_irn_link(block) == NULL)
581 relink_bad_block_predecessors(block, env);
583 old_in = (ir_node **)get_irn_link(block); /* Of Phi's Block */
584 old_irn_arity = ARR_LEN(old_in);
586 /* Relink Phi predeseccors if count of predeseccors changed */
587 if (old_irn_arity != ARR_LEN(get_irn_in(block))) {
588 /* set new predeseccors in array
589 n->in[0] remains the same block */
591 for(i = 1; i < old_irn_arity; i++)
592 if (!is_Bad((ir_node *)old_in[i])) n->in[new_irn_arity++] = n->in[i];
594 ARR_SETLEN(ir_node *, n->in, new_irn_arity);
597 } /* n is a Phi node */
601 * Removes Bad Bad predecesors from Blocks and the corresponding
602 * inputs to Phi nodes as in dead_node_elimination but without
604 * On walking up set the link field to NULL, on walking down call
605 * relink_bad_predecessors() (This function stores the old in array
606 * to the link field and sets a new in array if arity of predecessors
609 void remove_bad_predecessors(ir_graph *irg) {
610 irg_walk_graph(irg, init_link, relink_bad_predecessors, NULL);
614 /*--------------------------------------------------------------------*/
615 /* Funcionality for inlining */
616 /*--------------------------------------------------------------------*/
619 * Copy node for inlineing. Updates attributes that change when
620 * inlineing but not for dead node elimination.
622 * Copies the node by calling copy_node and then updates the entity if
623 * it's a local one. env must be a pointer of the frame type of the
624 * inlined procedure. The new entities must be in the link field of
628 copy_node_inline (ir_node *n, void *env) {
630 type *frame_tp = (type *)env;
633 if (get_irn_op(n) == op_Sel) {
634 new = get_new_node (n);
635 assert(get_irn_op(new) == op_Sel);
636 if (get_entity_owner(get_Sel_entity(n)) == frame_tp) {
637 set_Sel_entity(new, get_entity_link(get_Sel_entity(n)));
639 } else if (get_irn_op(n) == op_Block) {
640 new = get_new_node (n);
641 new->attr.block.irg = current_ir_graph;
645 static void find_addr(ir_node *node, void *env)
647 if (get_irn_opcode(node) == iro_Proj) {
648 if (get_Proj_proj(node) == pn_Start_P_value_arg_base)
654 * currently, we cannot inline two cases:
655 * - call with compound arguments
656 * - graphs that take the address of a parameter
658 * check these condition here
660 static int can_inline(ir_node *call, ir_graph *called_graph)
662 type *call_type = get_Call_type(call);
663 int params, ress, i, res;
664 assert(is_method_type(call_type));
666 params = get_method_n_params(call_type);
667 ress = get_method_n_ress(call_type);
670 for (i = 0; i < params; ++i) {
671 type *p_type = get_method_param_type(call_type, i);
673 if (is_compound_type(p_type))
678 for (i = 0; i < ress; ++i) {
679 type *r_type = get_method_res_type(call_type, i);
681 if (is_compound_type(r_type))
686 irg_walk_graph(called_graph, find_addr, NULL, &res);
691 int inline_method(ir_node *call, ir_graph *called_graph) {
693 ir_node *post_call, *post_bl;
695 ir_node *end, *end_bl;
699 int arity, n_ret, n_exc, n_res, i, j, rem_opt, irn_arity;
702 irg_inline_property prop = get_irg_inline_property(called_graph);
704 if ( (prop != irg_inline_forced) && (!get_opt_optimize() || !get_opt_inline() ||
705 (prop == irg_inline_forbidden))) return 0;
708 * currently, we cannot inline two cases:
709 * - call with compound arguments
710 * - graphs that take the address of a parameter
712 if (! can_inline(call, called_graph))
715 /* -- Turn off optimizations, this can cause problems when allocating new nodes. -- */
716 rem_opt = get_opt_optimize();
719 /* Handle graph state */
720 assert(get_irg_phase_state(current_ir_graph) != phase_building);
721 assert(get_irg_pinned(current_ir_graph) == pinned);
722 assert(get_irg_pinned(called_graph) == pinned);
723 if (get_irg_outs_state(current_ir_graph) == outs_consistent)
724 set_irg_outs_inconsistent(current_ir_graph);
725 set_irg_loopinfo_inconsistent(current_ir_graph);
727 /* -- Check preconditions -- */
728 assert(get_irn_op(call) == op_Call);
729 /* @@@ does not work for InterfaceIII.java after cgana
730 assert(get_Call_type(call) == get_entity_type(get_irg_ent(called_graph)));
731 assert(smaller_type(get_entity_type(get_irg_ent(called_graph)),
732 get_Call_type(call)));
734 assert(get_type_tpop(get_Call_type(call)) == type_method);
735 if (called_graph == current_ir_graph) {
736 set_optimize(rem_opt);
740 /* here we know we WILL inline, so inform the statistics */
741 stat_inline(call, called_graph);
743 /* -- Decide how to handle exception control flow: Is there a handler
744 for the Call node, or do we branch directly to End on an exception?
745 exc_handling: 0 There is a handler.
747 2 Exception handling not represented in Firm. -- */
749 ir_node *proj, *Mproj = NULL, *Xproj = NULL;
750 for (proj = (ir_node *)get_irn_link(call); proj; proj = (ir_node *)get_irn_link(proj)) {
751 assert(get_irn_op(proj) == op_Proj);
752 if (get_Proj_proj(proj) == pn_Call_X_except) Xproj = proj;
753 if (get_Proj_proj(proj) == pn_Call_M_except) Mproj = proj;
755 if (Mproj) { assert(Xproj); exc_handling = 0; } /* Mproj */
756 else if (Xproj) { exc_handling = 1; } /* !Mproj && Xproj */
757 else { exc_handling = 2; } /* !Mproj && !Xproj */
762 the procedure and later replaces the Start node of the called graph.
763 Post_call is the old Call node and collects the results of the called
764 graph. Both will end up being a tuple. -- */
765 post_bl = get_nodes_Block(call);
766 set_irg_current_block(current_ir_graph, post_bl);
767 /* XxMxPxP of Start + parameter of Call */
768 in[pn_Start_X_initial_exec] = new_Jmp();
769 in[pn_Start_M] = get_Call_mem(call);
770 in[pn_Start_P_frame_base] = get_irg_frame(current_ir_graph);
771 in[pn_Start_P_globals] = get_irg_globals(current_ir_graph);
772 in[pn_Start_T_args] = new_Tuple(get_Call_n_params(call), get_Call_param_arr(call));
773 /* in[pn_Start_P_value_arg_base] = ??? */
774 pre_call = new_Tuple(5, in);
778 The new block gets the ins of the old block, pre_call and all its
779 predecessors and all Phi nodes. -- */
780 part_block(pre_call);
782 /* -- Prepare state for dead node elimination -- */
783 /* Visited flags in calling irg must be >= flag in called irg.
784 Else walker and arity computation will not work. */
785 if (get_irg_visited(current_ir_graph) <= get_irg_visited(called_graph))
786 set_irg_visited(current_ir_graph, get_irg_visited(called_graph)+1);
787 if (get_irg_block_visited(current_ir_graph)< get_irg_block_visited(called_graph))
788 set_irg_block_visited(current_ir_graph, get_irg_block_visited(called_graph));
789 /* Set pre_call as new Start node in link field of the start node of
790 calling graph and pre_calls block as new block for the start block
792 Further mark these nodes so that they are not visited by the
794 set_irn_link(get_irg_start(called_graph), pre_call);
795 set_irn_visited(get_irg_start(called_graph), get_irg_visited(current_ir_graph));
796 set_irn_link(get_irg_start_block(called_graph), get_nodes_Block(pre_call));
797 set_irn_visited(get_irg_start_block(called_graph), get_irg_visited(current_ir_graph));
798 set_irn_link(get_irg_bad(called_graph), get_irg_bad(current_ir_graph));
799 set_irn_visited(get_irg_bad(called_graph), get_irg_visited(current_ir_graph));
801 /* Initialize for compaction of in arrays */
802 inc_irg_block_visited(current_ir_graph);
804 /* -- Replicate local entities of the called_graph -- */
805 /* copy the entities. */
806 called_frame = get_irg_frame_type(called_graph);
807 for (i = 0; i < get_class_n_members(called_frame); i++) {
808 entity *new_ent, *old_ent;
809 old_ent = get_class_member(called_frame, i);
810 new_ent = copy_entity_own(old_ent, get_cur_frame_type());
811 set_entity_link(old_ent, new_ent);
814 /* visited is > than that of called graph. With this trick visited will
815 remain unchanged so that an outer walker, e.g., searching the call nodes
816 to inline, calling this inline will not visit the inlined nodes. */
817 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
819 /* -- Performing dead node elimination inlines the graph -- */
820 /* Copies the nodes to the obstack of current_ir_graph. Updates links to new
822 /* @@@ endless loops are not copied!! -- they should be, I think... */
823 irg_walk(get_irg_end(called_graph), copy_node_inline, copy_preds,
824 get_irg_frame_type(called_graph));
826 /* Repair called_graph */
827 set_irg_visited(called_graph, get_irg_visited(current_ir_graph));
828 set_irg_block_visited(called_graph, get_irg_block_visited(current_ir_graph));
829 set_Block_block_visited(get_irg_start_block(called_graph), 0);
831 /* -- Merge the end of the inlined procedure with the call site -- */
832 /* We will turn the old Call node into a Tuple with the following
835 0: Phi of all Memories of Return statements.
836 1: Jmp from new Block that merges the control flow from all exception
837 predecessors of the old end block.
838 2: Tuple of all arguments.
839 3: Phi of Exception memories.
840 In case the old Call directly branches to End on an exception we don't
841 need the block merging all exceptions nor the Phi of the exception
845 /* -- Precompute some values -- */
846 end_bl = get_new_node(get_irg_end_block(called_graph));
847 end = get_new_node(get_irg_end(called_graph));
848 arity = get_irn_arity(end_bl); /* arity = n_exc + n_ret */
849 n_res = get_method_n_ress(get_Call_type(call));
851 res_pred = (ir_node **) malloc (n_res * sizeof (ir_node *));
852 cf_pred = (ir_node **) malloc (arity * sizeof (ir_node *));
854 set_irg_current_block(current_ir_graph, post_bl); /* just to make sure */
856 /* -- archive keepalives -- */
857 irn_arity = get_irn_arity(end);
858 for (i = 0; i < irn_arity; i++)
859 add_End_keepalive(get_irg_end(current_ir_graph), get_irn_n(end, i));
861 /* The new end node will die. We need not free as the in array is on the obstack:
862 copy_node only generated 'D' arrays. */
864 /* -- Replace Return nodes by Jump nodes. -- */
866 for (i = 0; i < arity; i++) {
868 ret = get_irn_n(end_bl, i);
869 if (get_irn_op(ret) == op_Return) {
870 cf_pred[n_ret] = new_r_Jmp(current_ir_graph, get_nodes_Block(ret));
874 set_irn_in(post_bl, n_ret, cf_pred);
876 /* -- Build a Tuple for all results of the method.
877 Add Phi node if there was more than one Return. -- */
878 turn_into_tuple(post_call, 4);
879 /* First the Memory-Phi */
881 for (i = 0; i < arity; i++) {
882 ret = get_irn_n(end_bl, i);
883 if (get_irn_op(ret) == op_Return) {
884 cf_pred[n_ret] = get_Return_mem(ret);
888 phi = new_Phi(n_ret, cf_pred, mode_M);
889 set_Tuple_pred(call, pn_Call_M_regular, phi);
890 /* Conserve Phi-list for further inlinings -- but might be optimized */
891 if (get_nodes_Block(phi) == post_bl) {
892 set_irn_link(phi, get_irn_link(post_bl));
893 set_irn_link(post_bl, phi);
895 /* Now the real results */
897 for (j = 0; j < n_res; j++) {
899 for (i = 0; i < arity; i++) {
900 ret = get_irn_n(end_bl, i);
901 if (get_irn_op(ret) == op_Return) {
902 cf_pred[n_ret] = get_Return_res(ret, j);
907 phi = new_Phi(n_ret, cf_pred, get_irn_mode(cf_pred[0]));
911 /* Conserve Phi-list for further inlinings -- but might be optimized */
912 if (get_nodes_Block(phi) == post_bl) {
913 set_irn_link(phi, get_irn_link(post_bl));
914 set_irn_link(post_bl, phi);
917 set_Tuple_pred(call, pn_Call_T_result, new_Tuple(n_res, res_pred));
919 set_Tuple_pred(call, pn_Call_T_result, new_Bad());
921 /* Finally the exception control flow.
922 We have two (three) possible situations:
923 First if the Call branches to an exception handler: We need to add a Phi node to
924 collect the memory containing the exception objects. Further we need
925 to add another block to get a correct representation of this Phi. To
926 this block we add a Jmp that resolves into the X output of the Call
927 when the Call is turned into a tuple.
928 Second the Call branches to End, the exception is not handled. Just
929 add all inlined exception branches to the End node.
930 Third: there is no Exception edge at all. Handle as case two. */
931 if (exc_handling == 0) {
933 for (i = 0; i < arity; i++) {
935 ret = get_irn_n(end_bl, i);
936 if (is_fragile_op(skip_Proj(ret)) || (get_irn_op(skip_Proj(ret)) == op_Raise)) {
937 cf_pred[n_exc] = ret;
942 new_Block(n_exc, cf_pred); /* watch it: current_block is changed! */
943 set_Tuple_pred(call, pn_Call_X_except, new_Jmp());
944 /* The Phi for the memories with the exception objects */
946 for (i = 0; i < arity; i++) {
948 ret = skip_Proj(get_irn_n(end_bl, i));
949 if (get_irn_op(ret) == op_Call) {
950 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_Block(ret), ret, mode_M, 3);
952 } else if (is_fragile_op(ret)) {
953 /* We rely that all cfops have the memory output at the same position. */
954 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_Block(ret), ret, mode_M, 0);
956 } else if (get_irn_op(ret) == op_Raise) {
957 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_Block(ret), ret, mode_M, 1);
961 set_Tuple_pred(call, pn_Call_M_except, new_Phi(n_exc, cf_pred, mode_M));
963 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
964 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
967 ir_node *main_end_bl;
968 int main_end_bl_arity;
971 /* assert(exc_handling == 1 || no exceptions. ) */
973 for (i = 0; i < arity; i++) {
974 ir_node *ret = get_irn_n(end_bl, i);
976 if (is_fragile_op(skip_Proj(ret)) || (get_irn_op(skip_Proj(ret)) == op_Raise)) {
977 cf_pred[n_exc] = ret;
981 main_end_bl = get_irg_end_block(current_ir_graph);
982 main_end_bl_arity = get_irn_arity(main_end_bl);
983 end_preds = (ir_node **) malloc ((n_exc + main_end_bl_arity) * sizeof (ir_node *));
985 for (i = 0; i < main_end_bl_arity; ++i)
986 end_preds[i] = get_irn_n(main_end_bl, i);
987 for (i = 0; i < n_exc; ++i)
988 end_preds[main_end_bl_arity + i] = cf_pred[i];
989 set_irn_in(main_end_bl, n_exc + main_end_bl_arity, end_preds);
990 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
991 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
997 #if 0 /* old. now better, correcter, faster implementation. */
999 /* -- If the exception control flow from the inlined Call directly
1000 branched to the end block we now have the following control
1001 flow predecessor pattern: ProjX -> Tuple -> Jmp. We must
1002 remove the Jmp along with it's empty block and add Jmp's
1003 predecessors as predecessors of this end block. No problem if
1004 there is no exception, because then branches Bad to End which
1006 @@@ can't we know this beforehand: by getting the Proj(1) from
1007 the Call link list and checking whether it goes to Proj. */
1008 /* find the problematic predecessor of the end block. */
1009 end_bl = get_irg_end_block(current_ir_graph);
1010 for (i = 0; i < get_Block_n_cfgpreds(end_bl); i++) {
1011 cf_op = get_Block_cfgpred(end_bl, i);
1012 if (get_irn_op(cf_op) == op_Proj) {
1013 cf_op = get_Proj_pred(cf_op);
1014 if ((get_irn_op(cf_op) == op_Tuple) && (cf_op == call)) {
1015 /* There are unoptimized tuples from inlineing before when no exc */
1016 assert(get_Proj_proj(get_Block_cfgpred(end_bl, i)) == pn_Call_X_except);
1017 cf_op = get_Tuple_pred(cf_op, pn_Call_X_except);
1018 assert(get_irn_op(cf_op) == op_Jmp);
1024 if (i < get_Block_n_cfgpreds(end_bl)) {
1025 bl = get_nodes_Block(cf_op);
1026 arity = get_Block_n_cfgpreds(end_bl) + get_Block_n_cfgpreds(bl) - 1;
1027 cf_pred = (ir_node **) malloc (arity * sizeof (ir_node *));
1028 for (j = 0; j < i; j++)
1029 cf_pred[j] = get_Block_cfgpred(end_bl, j);
1030 for (j = j; j < i + get_Block_n_cfgpreds(bl); j++)
1031 cf_pred[j] = get_Block_cfgpred(bl, j-i);
1032 for (j = j; j < arity; j++)
1033 cf_pred[j] = get_Block_cfgpred(end_bl, j-get_Block_n_cfgpreds(bl) +1);
1034 set_irn_in(end_bl, arity, cf_pred);
1036 /* Remove the exception pred from post-call Tuple. */
1037 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1042 /* -- Turn cse back on. -- */
1043 set_optimize(rem_opt);
1048 /********************************************************************/
1049 /* Apply inlineing to small methods. */
1050 /********************************************************************/
1052 /* It makes no sense to inline too many calls in one procedure. Anyways,
1053 I didn't get a version with NEW_ARR_F to run. */
1054 #define MAX_INLINE 1024
1057 * environment for inlining small irgs
1059 typedef struct _inline_env_t {
1061 ir_node *calls[MAX_INLINE];
1065 * Returns the irg called from a Call node. If the irg is not
1066 * known, NULL is returned.
1068 static ir_graph *get_call_called_irg(ir_node *call) {
1071 ir_graph *called_irg = NULL;
1073 assert(get_irn_op(call) == op_Call);
1075 addr = get_Call_ptr(call);
1076 if (get_irn_op(addr) == op_Const) {
1077 /* Check whether the constant is the pointer to a compiled entity. */
1078 tv = get_Const_tarval(addr);
1079 if (get_tarval_entity(tv))
1080 called_irg = get_entity_irg(get_tarval_entity(tv));
1081 } else if (get_irn_op(addr) == op_SymConst && get_SymConst_kind(addr) == symconst_addr_ent) {
1082 called_irg = get_entity_irg(get_SymConst_entity(addr));
1087 static void collect_calls(ir_node *call, void *env) {
1088 inline_env_t *ienv = env;
1091 ir_graph *called_irg;
1093 if (get_irn_op(call) != op_Call) return;
1095 addr = get_Call_ptr(call);
1096 if (get_irn_op(addr) == op_Const) {
1097 /* Check whether the constant is the pointer to a compiled entity. */
1098 tv = get_Const_tarval(addr);
1099 if (get_tarval_entity(tv)) {
1100 called_irg = get_entity_irg(get_tarval_entity(tv));
1101 if (called_irg && ienv->pos < MAX_INLINE) {
1102 /* The Call node calls a locally defined method. Remember to inline. */
1103 ienv->calls[ienv->pos++] = call;
1110 * Inlines all small methods at call sites where the called address comes
1111 * from a Const node that references the entity representing the called
1113 * The size argument is a rough measure for the code size of the method:
1114 * Methods where the obstack containing the firm graph is smaller than
1117 void inline_small_irgs(ir_graph *irg, int size) {
1119 ir_graph *rem = current_ir_graph;
1122 if (!(get_opt_optimize() && get_opt_inline())) return;
1124 current_ir_graph = irg;
1125 /* Handle graph state */
1126 assert(get_irg_phase_state(current_ir_graph) != phase_building);
1127 free_callee_info(current_ir_graph);
1129 /* Find Call nodes to inline.
1130 (We can not inline during a walk of the graph, as inlineing the same
1131 method several times changes the visited flag of the walked graph:
1132 after the first inlineing visited of the callee equals visited of
1133 the caller. With the next inlineing both are increased.) */
1135 irg_walk(get_irg_end(irg), NULL, collect_calls, &env);
1137 if ((env.pos > 0) && (env.pos < MAX_INLINE)) {
1138 /* There are calls to inline */
1139 collect_phiprojs(irg);
1140 for (i = 0; i < env.pos; i++) {
1143 tv = get_Const_tarval(get_Call_ptr(env.calls[i]));
1144 callee = get_entity_irg(get_tarval_entity(tv));
1145 if (((_obstack_memory_used(callee->obst) - obstack_room(callee->obst)) < size) ||
1146 (get_irg_inline_property(callee) == irg_inline_forced)) {
1147 inline_method(env.calls[i], callee);
1152 current_ir_graph = rem;
1156 * Environment for inlining irgs.
1159 int n_nodes; /**< Nodes in graph except Id, Tuple, Proj, Start, End */
1160 int n_nodes_orig; /**< for statistics */
1161 eset *call_nodes; /**< All call nodes in this graph */
1163 int n_call_nodes_orig; /**< for statistics */
1164 int n_callers; /**< Number of known graphs that call this graphs. */
1165 int n_callers_orig; /**< for statistics */
1168 static inline_irg_env *new_inline_irg_env(void) {
1169 inline_irg_env *env = malloc(sizeof(inline_irg_env));
1170 env->n_nodes = -2; /* uncount Start, End */
1171 env->n_nodes_orig = -2; /* uncount Start, End */
1172 env->call_nodes = eset_create();
1173 env->n_call_nodes = 0;
1174 env->n_call_nodes_orig = 0;
1176 env->n_callers_orig = 0;
1180 static void free_inline_irg_env(inline_irg_env *env) {
1181 eset_destroy(env->call_nodes);
1185 static void collect_calls2(ir_node *call, void *env) {
1186 inline_irg_env *x = (inline_irg_env *)env;
1187 ir_op *op = get_irn_op(call);
1190 /* count nodes in irg */
1191 if (op != op_Proj && op != op_Tuple && op != op_Sync) {
1196 if (op != op_Call) return;
1198 /* collect all call nodes */
1199 eset_insert(x->call_nodes, (void *)call);
1201 x->n_call_nodes_orig++;
1203 /* count all static callers */
1204 callee = get_call_called_irg(call);
1206 ((inline_irg_env *)get_irg_link(callee))->n_callers++;
1207 ((inline_irg_env *)get_irg_link(callee))->n_callers_orig++;
1211 INLINE static int is_leave(ir_graph *irg) {
1212 return (((inline_irg_env *)get_irg_link(irg))->n_call_nodes == 0);
1215 INLINE static int is_smaller(ir_graph *callee, int size) {
1216 return (((inline_irg_env *)get_irg_link(callee))->n_nodes < size);
1221 * Inlines small leave methods at call sites where the called address comes
1222 * from a Const node that references the entity representing the called
1224 * The size argument is a rough measure for the code size of the method:
1225 * Methods where the obstack containing the firm graph is smaller than
1228 void inline_leave_functions(int maxsize, int leavesize, int size) {
1229 inline_irg_env *env;
1230 int i, n_irgs = get_irp_n_irgs();
1231 ir_graph *rem = current_ir_graph;
1234 if (!(get_opt_optimize() && get_opt_inline())) return;
1236 /* extend all irgs by a temporary data structure for inlineing. */
1237 for (i = 0; i < n_irgs; ++i)
1238 set_irg_link(get_irp_irg(i), new_inline_irg_env());
1240 /* Precompute information in temporary data structure. */
1241 for (i = 0; i < n_irgs; ++i) {
1242 current_ir_graph = get_irp_irg(i);
1243 assert(get_irg_phase_state(current_ir_graph) != phase_building);
1244 free_callee_info(current_ir_graph);
1246 irg_walk(get_irg_end(current_ir_graph), NULL, collect_calls2,
1247 get_irg_link(current_ir_graph));
1251 Inline leaves recursively -- we might construct new leaves. */
1252 /* int itercnt = 1; */
1253 while (did_inline) {
1254 /* printf("iteration %d\n", itercnt++); */
1256 for (i = 0; i < n_irgs; ++i) {
1259 int phiproj_computed = 0;
1261 current_ir_graph = get_irp_irg(i);
1262 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1264 /* we can not walk and change a set, nor remove from it.
1266 walkset = env->call_nodes;
1267 env->call_nodes = eset_create();
1268 for (call = eset_first(walkset); call; call = eset_next(walkset)) {
1269 inline_irg_env *callee_env;
1270 ir_graph *callee = get_call_called_irg(call);
1272 if (env->n_nodes > maxsize) break;
1274 ((is_leave(callee) && is_smaller(callee, leavesize)) ||
1275 (get_irg_inline_property(callee) == irg_inline_forced))) {
1276 if (!phiproj_computed) {
1277 phiproj_computed = 1;
1278 collect_phiprojs(current_ir_graph);
1280 callee_env = (inline_irg_env *)get_irg_link(callee);
1281 /* printf(" %s: Inlineing %s.\n", get_entity_name(get_irg_entity(current_ir_graph)), */
1282 /* get_entity_name(get_irg_entity(callee))); */
1283 if (inline_method(call, callee)) {
1285 env->n_call_nodes--;
1286 eset_insert_all(env->call_nodes, callee_env->call_nodes);
1287 env->n_call_nodes += callee_env->n_call_nodes;
1288 env->n_nodes += callee_env->n_nodes;
1289 callee_env->n_callers--;
1292 eset_insert(env->call_nodes, call);
1295 eset_destroy(walkset);
1299 /* printf("Non leaves\n"); */
1300 /* inline other small functions. */
1301 for (i = 0; i < n_irgs; ++i) {
1304 int phiproj_computed = 0;
1306 current_ir_graph = get_irp_irg(i);
1307 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1309 /* we can not walk and change a set, nor remove from it.
1311 walkset = env->call_nodes;
1312 env->call_nodes = eset_create();
1313 for (call = eset_first(walkset); call; call = eset_next(walkset)) {
1314 inline_irg_env *callee_env;
1315 ir_graph *callee = get_call_called_irg(call);
1317 if (env->n_nodes > maxsize) break;
1318 if (callee && is_smaller(callee, size)) {
1319 if (!phiproj_computed) {
1320 phiproj_computed = 1;
1321 collect_phiprojs(current_ir_graph);
1323 callee_env = (inline_irg_env *)get_irg_link(callee);
1324 /* printf(" %s: Inlineing %s.\n", get_entity_name(get_irg_entity(current_ir_graph)), */
1325 /* get_entity_name(get_irg_entity(callee))); */
1326 if (inline_method(call, callee)) {
1328 env->n_call_nodes--;
1329 eset_insert_all(env->call_nodes, callee_env->call_nodes);
1330 env->n_call_nodes += callee_env->n_call_nodes;
1331 env->n_nodes += callee_env->n_nodes;
1332 callee_env->n_callers--;
1335 eset_insert(env->call_nodes, call);
1338 eset_destroy(walkset);
1341 for (i = 0; i < n_irgs; ++i) {
1342 current_ir_graph = get_irp_irg(i);
1344 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1345 if ((env->n_call_nodes_orig != env->n_call_nodes) ||
1346 (env->n_callers_orig != env->n_callers))
1347 printf("Nodes:%3d ->%3d, calls:%3d ->%3d, callers:%3d ->%3d, -- %s\n",
1348 env->n_nodes_orig, env->n_nodes, env->n_call_nodes_orig, env->n_call_nodes,
1349 env->n_callers_orig, env->n_callers,
1350 get_entity_name(get_irg_entity(current_ir_graph)));
1352 free_inline_irg_env((inline_irg_env *)get_irg_link(current_ir_graph));
1355 current_ir_graph = rem;
1358 /*******************************************************************/
1359 /* Code Placement. Pins all floating nodes to a block where they */
1360 /* will be executed only if needed. */
1361 /*******************************************************************/
1364 * Find the earliest correct block for N. --- Place N into the
1365 * same Block as its dominance-deepest Input.
1368 place_floats_early(ir_node *n, pdeq *worklist)
1370 int i, start, irn_arity;
1372 /* we must not run into an infinite loop */
1373 assert (irn_not_visited(n));
1374 mark_irn_visited(n);
1376 /* Place floating nodes. */
1377 if (get_op_pinned(get_irn_op(n)) == floats) {
1379 ir_node *b = new_Bad(); /* The block to place this node in */
1380 int bad_recursion = is_Bad(get_nodes_block(n));
1382 assert(get_irn_op(n) != op_Block);
1384 if ((get_irn_op(n) == op_Const) ||
1385 (get_irn_op(n) == op_SymConst) ||
1387 (get_irn_op(n) == op_Unknown)) {
1388 /* These nodes will not be placed by the loop below. */
1389 b = get_irg_start_block(current_ir_graph);
1393 /* find the block for this node. */
1394 irn_arity = get_irn_arity(n);
1395 for (i = 0; i < irn_arity; i++) {
1396 ir_node *dep = get_irn_n(n, i);
1399 if ((irn_not_visited(dep))
1400 && (get_op_pinned(get_irn_op(dep)) == floats)) {
1401 place_floats_early(dep, worklist);
1405 * A node in the Bad block must stay in the bad block,
1406 * so don't compute a new block for it.
1411 /* Because all loops contain at least one pinned node, now all
1412 our inputs are either pinned or place_early has already
1413 been finished on them. We do not have any unfinished inputs! */
1414 dep_block = get_nodes_Block(dep);
1415 if ((!is_Bad(dep_block)) &&
1416 (get_Block_dom_depth(dep_block) > depth)) {
1418 depth = get_Block_dom_depth(dep_block);
1420 /* Avoid that the node is placed in the Start block */
1421 if ((depth == 1) && (get_Block_dom_depth(get_nodes_Block(n)) > 1)) {
1422 b = get_Block_cfg_out(get_irg_start_block(current_ir_graph), 0);
1423 assert(b != get_irg_start_block(current_ir_graph));
1427 set_nodes_Block(n, b);
1430 /* Add predecessors of non floating nodes on worklist. */
1431 start = (get_irn_op(n) == op_Block) ? 0 : -1;
1432 irn_arity = get_irn_arity(n);
1433 for (i = start; i < irn_arity; i++) {
1434 ir_node *pred = get_irn_n(n, i);
1435 if (irn_not_visited(pred)) {
1436 pdeq_putr (worklist, pred);
1442 * Floating nodes form subgraphs that begin at nodes as Const, Load,
1443 * Start, Call and that end at pinned nodes as Store, Call. Place_early
1444 * places all floating nodes reachable from its argument through floating
1445 * nodes and adds all beginnings at pinned nodes to the worklist.
1447 static INLINE void place_early(pdeq* worklist) {
1449 inc_irg_visited(current_ir_graph);
1451 /* this inits the worklist */
1452 place_floats_early(get_irg_end(current_ir_graph), worklist);
1454 /* Work the content of the worklist. */
1455 while (!pdeq_empty (worklist)) {
1456 ir_node *n = pdeq_getl (worklist);
1457 if (irn_not_visited(n)) place_floats_early(n, worklist);
1460 set_irg_outs_inconsistent(current_ir_graph);
1461 current_ir_graph->pinned = pinned;
1465 /** deepest common dominance ancestor of DCA and CONSUMER of PRODUCER. */
1467 consumer_dom_dca (ir_node *dca, ir_node *consumer, ir_node *producer)
1469 ir_node *block = NULL;
1471 /* Compute the latest block into which we can place a node so that it is
1473 if (get_irn_op(consumer) == op_Phi) {
1474 /* our consumer is a Phi-node, the effective use is in all those
1475 blocks through which the Phi-node reaches producer */
1477 ir_node *phi_block = get_nodes_Block(consumer);
1478 irn_arity = get_irn_arity(consumer);
1479 for (i = 0; i < irn_arity; i++) {
1480 if (get_irn_n(consumer, i) == producer) {
1481 block = get_nodes_Block(get_Block_cfgpred(phi_block, i));
1485 assert(is_no_Block(consumer));
1486 block = get_nodes_Block(consumer);
1489 /* Compute the deepest common ancestor of block and dca. */
1491 if (!dca) return block;
1492 while (get_Block_dom_depth(block) > get_Block_dom_depth(dca))
1493 block = get_Block_idom(block);
1494 while (get_Block_dom_depth(dca) > get_Block_dom_depth(block))
1495 dca = get_Block_idom(dca);
1496 while (block != dca)
1497 { block = get_Block_idom(block); dca = get_Block_idom(dca); }
1502 static INLINE int get_irn_loop_depth(ir_node *n) {
1503 return get_loop_depth(get_irn_loop(n));
1507 * Move n to a block with less loop depth than it's current block. The
1508 * new block must be dominated by early.
1511 move_out_of_loops (ir_node *n, ir_node *early)
1513 ir_node *best, *dca;
1517 /* Find the region deepest in the dominator tree dominating
1518 dca with the least loop nesting depth, but still dominated
1519 by our early placement. */
1520 dca = get_nodes_Block(n);
1522 while (dca != early) {
1523 dca = get_Block_idom(dca);
1524 if (!dca) break; /* should we put assert(dca)? */
1525 if (get_irn_loop_depth(dca) < get_irn_loop_depth(best)) {
1529 if (best != get_nodes_Block(n)) {
1531 printf("Moving out of loop: "); DDMN(n);
1532 printf(" Outermost block: "); DDMN(early);
1533 printf(" Best block: "); DDMN(best);
1534 printf(" Innermost block: "); DDMN(get_nodes_Block(n));
1536 set_nodes_Block(n, best);
1541 * Find the latest legal block for N and place N into the
1542 * `optimal' Block between the latest and earliest legal block.
1543 * The `optimal' block is the dominance-deepest block of those
1544 * with the least loop-nesting-depth. This places N out of as many
1545 * loops as possible and then makes it as control dependant as
1549 place_floats_late(ir_node *n, pdeq *worklist)
1554 assert (irn_not_visited(n)); /* no multiple placement */
1556 /* no need to place block nodes, control nodes are already placed. */
1557 if ((get_irn_op(n) != op_Block) &&
1559 (get_irn_mode(n) != mode_X)) {
1560 /* Remember the early placement of this block to move it
1561 out of loop no further than the early placement. */
1562 early = get_nodes_Block(n);
1563 /* Assure that our users are all placed, except the Phi-nodes.
1564 --- Each data flow cycle contains at least one Phi-node. We
1565 have to break the `user has to be placed before the
1566 producer' dependence cycle and the Phi-nodes are the
1567 place to do so, because we need to base our placement on the
1568 final region of our users, which is OK with Phi-nodes, as they
1569 are pinned, and they never have to be placed after a
1570 producer of one of their inputs in the same block anyway. */
1571 for (i = 0; i < get_irn_n_outs(n); i++) {
1572 ir_node *succ = get_irn_out(n, i);
1573 if (irn_not_visited(succ) && (get_irn_op(succ) != op_Phi))
1574 place_floats_late(succ, worklist);
1577 /* We have to determine the final block of this node... except for
1579 if ((get_op_pinned(get_irn_op(n)) == floats) &&
1580 (get_irn_op(n) != op_Const) &&
1581 (get_irn_op(n) != op_SymConst)) {
1582 ir_node *dca = NULL; /* deepest common ancestor in the
1583 dominator tree of all nodes'
1584 blocks depending on us; our final
1585 placement has to dominate DCA. */
1586 for (i = 0; i < get_irn_n_outs(n); i++) {
1587 dca = consumer_dom_dca (dca, get_irn_out(n, i), n);
1589 set_nodes_Block(n, dca);
1591 move_out_of_loops (n, early);
1595 mark_irn_visited(n);
1597 /* Add predecessors of all non-floating nodes on list. (Those of floating
1598 nodes are placeded already and therefore are marked.) */
1599 for (i = 0; i < get_irn_n_outs(n); i++) {
1600 if (irn_not_visited(get_irn_out(n, i))) {
1601 pdeq_putr (worklist, get_irn_out(n, i));
1606 static INLINE void place_late(pdeq *worklist) {
1608 inc_irg_visited(current_ir_graph);
1610 /* This fills the worklist initially. */
1611 place_floats_late(get_irg_start_block(current_ir_graph), worklist);
1612 /* And now empty the worklist again... */
1613 while (!pdeq_empty (worklist)) {
1614 ir_node *n = pdeq_getl (worklist);
1615 if (irn_not_visited(n)) place_floats_late(n, worklist);
1619 void place_code(ir_graph *irg) {
1621 ir_graph *rem = current_ir_graph;
1623 current_ir_graph = irg;
1625 if (!(get_opt_optimize() && get_opt_global_cse())) return;
1627 /* Handle graph state */
1628 assert(get_irg_phase_state(irg) != phase_building);
1629 if (get_irg_dom_state(irg) != dom_consistent)
1632 if (1 || get_irg_loopinfo_state(irg) != loopinfo_consistent) {
1633 free_loop_information(irg);
1634 construct_backedges(irg);
1637 /* Place all floating nodes as early as possible. This guarantees
1638 a legal code placement. */
1639 worklist = new_pdeq();
1640 place_early(worklist);
1642 /* place_early invalidates the outs, place_late needs them. */
1644 /* Now move the nodes down in the dominator tree. This reduces the
1645 unnecessary executions of the node. */
1646 place_late(worklist);
1648 set_irg_outs_inconsistent(current_ir_graph);
1649 set_irg_loopinfo_inconsistent(current_ir_graph);
1651 current_ir_graph = rem;
1656 /********************************************************************/
1657 /* Control flow optimization. */
1658 /* Removes Bad control flow predecessors and empty blocks. A block */
1659 /* is empty if it contains only a Jmp node. */
1660 /* Blocks can only be removed if they are not needed for the */
1661 /* semantics of Phi nodes. */
1662 /********************************************************************/
1665 * Removes Tuples from Block control flow predecessors.
1666 * Optimizes blocks with equivalent_node().
1667 * Replaces n by Bad if n is unreachable control flow.
1669 static void merge_blocks(ir_node *n, void *env) {
1671 set_irn_link(n, NULL);
1673 if (get_irn_op(n) == op_Block) {
1675 for (i = 0; i < get_Block_n_cfgpreds(n); i++)
1676 /* GL @@@ : is this possible? if (get_opt_normalize()) -- added, all tests go through.
1677 A different order of optimizations might cause problems. */
1678 if (get_opt_normalize())
1679 set_Block_cfgpred(n, i, skip_Tuple(get_Block_cfgpred(n, i)));
1680 } else if (get_opt_optimize() && (get_irn_mode(n) == mode_X)) {
1681 /* We will soon visit a block. Optimize it before visiting! */
1682 ir_node *b = get_nodes_Block(n);
1683 ir_node *new_node = equivalent_node(b);
1684 while (irn_not_visited(b) && (!is_Bad(new_node)) && (new_node != b)) {
1685 /* We would have to run gigo if new is bad, so we
1686 promote it directly below. */
1687 assert(((b == new_node) ||
1688 get_opt_control_flow_straightening() ||
1689 get_opt_control_flow_weak_simplification()) &&
1690 ("strange flag setting"));
1691 exchange (b, new_node);
1693 new_node = equivalent_node(b);
1695 if (is_Bad(new_node) && get_opt_normalize()) exchange(n, new_Bad());
1700 * Collects all Phi nodes in link list of Block.
1701 * Marks all blocks "block_visited" if they contain a node other
1704 * Colelct Conds and its Projs in the cond_list
1706 static void collect_nodes(ir_node *n, void *env) {
1707 ir_node **cond_list = env;
1709 if (is_no_Block(n)) {
1710 ir_node *b = get_nodes_Block(n);
1712 switch (get_irn_opcode(n)) {
1715 /* Collect Phi nodes to compact ins along with block's ins. */
1716 set_irn_link(n, get_irn_link(b));
1722 ir_node *value = get_Cond_selector(n);
1723 tarval *tv = computed_value(value);
1725 if (tv != tarval_bad && mode_is_int(get_tarval_mode(tv))) {
1726 set_irn_link(n, *cond_list);
1729 else { /* mark the cond, we cannot optimize it */
1730 set_irn_link(n, NULL);
1737 ir_node *cond = get_Proj_pred(n);
1739 if (get_irn_op(cond) == op_Cond) {
1740 void *link = get_irn_link(cond);
1743 /* the cond node can be optimized, collect Proj */
1744 set_irn_link(n, link);
1745 set_irn_link(cond, n);
1755 if ((get_irn_op(n) != op_Jmp) && !is_Bad(b)) { /* Check for non empty block. */
1756 mark_Block_block_visited(b);
1761 /** Returns true if pred is predecessor of block. */
1762 static int is_pred_of(ir_node *pred, ir_node *b) {
1764 for (i = 0; i < get_Block_n_cfgpreds(b); i++) {
1765 ir_node *b_pred = get_nodes_Block(get_Block_cfgpred(b, i));
1766 if (b_pred == pred) return 1;
1771 static int test_whether_dispensable(ir_node *b, int pos) {
1772 int i, j, n_preds = 1;
1773 int dispensable = 1;
1774 ir_node *cfop = get_Block_cfgpred(b, pos);
1775 ir_node *pred = get_nodes_Block(cfop);
1777 if (get_Block_block_visited(pred) + 1
1778 < get_irg_block_visited(current_ir_graph)) {
1779 if (!get_opt_optimize() || !get_opt_control_flow_strong_simplification()) {
1780 /* Mark block so that is will not be removed. */
1781 set_Block_block_visited(pred, get_irg_block_visited(current_ir_graph)-1);
1784 /* Seems to be empty. */
1785 if (!get_irn_link(b)) {
1786 /* There are no Phi nodes ==> dispensable. */
1787 n_preds = get_Block_n_cfgpreds(pred);
1789 /* b's pred blocks and pred's pred blocks must be pairwise disjunct.
1790 Work preds < pos as if they were already removed. */
1791 for (i = 0; i < pos; i++) {
1792 ir_node *b_pred = get_nodes_Block(get_Block_cfgpred(b, i));
1793 if (get_Block_block_visited(b_pred) + 1
1794 < get_irg_block_visited(current_ir_graph)) {
1795 for (j = 0; j < get_Block_n_cfgpreds(b_pred); j++) {
1796 ir_node *b_pred_pred = get_nodes_Block(get_Block_cfgpred(b_pred, j));
1797 if (is_pred_of(b_pred_pred, pred)) dispensable = 0;
1800 if (is_pred_of(b_pred, pred)) dispensable = 0;
1803 for (i = pos +1; i < get_Block_n_cfgpreds(b); i++) {
1804 ir_node *b_pred = get_nodes_Block(get_Block_cfgpred(b, i));
1805 if (is_pred_of(b_pred, pred)) dispensable = 0;
1808 set_Block_block_visited(pred, get_irg_block_visited(current_ir_graph)-1);
1811 n_preds = get_Block_n_cfgpreds(pred);
1819 static void optimize_blocks(ir_node *b, void *env) {
1820 int i, j, k, max_preds, n_preds;
1821 ir_node *pred, *phi;
1824 /* Count the number of predecessor if this block is merged with pred blocks
1827 for (i = 0; i < get_Block_n_cfgpreds(b); i++) {
1828 max_preds += test_whether_dispensable(b, i);
1830 in = (ir_node **) malloc(max_preds * sizeof(ir_node *));
1833 printf(" working on "); DDMN(b);
1834 for (i = 0; i < get_Block_n_cfgpreds(b); i++) {
1835 pred = get_nodes_Block(get_Block_cfgpred(b, i));
1836 if (is_Bad(get_Block_cfgpred(b, i))) {
1837 printf(" removing Bad %i\n ", i);
1838 } else if (get_Block_block_visited(pred) +1
1839 < get_irg_block_visited(current_ir_graph)) {
1840 printf(" removing pred %i ", i); DDMN(pred);
1841 } else { printf(" Nothing to do for "); DDMN(pred); }
1843 * end Debug output -*/
1845 /*- Fix the Phi nodes -*/
1846 phi = get_irn_link(b);
1848 assert(get_irn_op(phi) == op_Phi);
1849 /* Find the new predecessors for the Phi */
1851 for (i = 0; i < get_Block_n_cfgpreds(b); i++) {
1852 pred = get_nodes_Block(get_Block_cfgpred(b, i));
1853 if (is_Bad(get_Block_cfgpred(b, i))) {
1855 } else if (get_Block_block_visited(pred) +1
1856 < get_irg_block_visited(current_ir_graph)) {
1857 /* It's an empty block and not yet visited. */
1858 ir_node *phi_pred = get_Phi_pred(phi, i);
1859 for (j = 0; j < get_Block_n_cfgpreds(pred); j++) {
1860 if (get_nodes_Block(phi_pred) == pred) {
1861 assert(get_irn_op(phi_pred) == op_Phi); /* Block is empty!! */
1862 in[n_preds] = get_Phi_pred(phi_pred, j);
1864 in[n_preds] = phi_pred;
1868 /* The Phi_pred node is replaced now if it is a Phi.
1869 In Schleifen kann offenbar der entfernte Phi Knoten legal verwendet werden.
1870 Daher muss der Phiknoten durch den neuen ersetzt werden.
1871 Weiter muss der alte Phiknoten entfernt werden (durch ersetzen oder
1872 durch einen Bad) damit er aus den keep_alive verschwinden kann.
1873 Man sollte also, falls keine Schleife vorliegt, exchange mit new_Bad
1875 if (get_nodes_Block(phi_pred) == pred) {
1876 /* remove the Phi as it might be kept alive. Further there
1877 might be other users. */
1878 exchange(phi_pred, phi); /* geht, ist aber doch semantisch falsch! Warum?? */
1881 in[n_preds] = get_Phi_pred(phi, i);
1886 set_irn_in(phi, n_preds, in);
1888 phi = get_irn_link(phi);
1891 /*- This happens only if merge between loop backedge and single loop entry. -*/
1892 for (k = 0; k < get_Block_n_cfgpreds(b); k++) {
1893 pred = get_nodes_Block(get_Block_cfgpred(b, k));
1894 if (get_Block_block_visited(pred)+1 < get_irg_block_visited(current_ir_graph)) {
1895 phi = get_irn_link(pred);
1897 if (get_irn_op(phi) == op_Phi) {
1898 set_nodes_Block(phi, b);
1901 for (i = 0; i < k; i++) {
1902 pred = get_nodes_Block(get_Block_cfgpred(b, i));
1903 if (is_Bad(get_Block_cfgpred(b, i))) {
1905 } else if (get_Block_block_visited(pred) +1
1906 < get_irg_block_visited(current_ir_graph)) {
1907 /* It's an empty block and not yet visited. */
1908 for (j = 0; j < get_Block_n_cfgpreds(pred); j++) {
1909 /* @@@ Hier brauche ich Schleifeninformation!!! Kontrollflusskante
1910 muss Rueckwaertskante sein! (An allen vier in[n_preds] = phi
1911 Anweisungen.) Trotzdem tuts bisher!! */
1920 for (i = 0; i < get_Phi_n_preds(phi); i++) {
1921 in[n_preds] = get_Phi_pred(phi, i);
1924 for (i = k+1; i < get_Block_n_cfgpreds(b); i++) {
1925 pred = get_nodes_Block(get_Block_cfgpred(b, i));
1926 if (is_Bad(get_Block_cfgpred(b, i))) {
1928 } else if (get_Block_block_visited(pred) +1
1929 < get_irg_block_visited(current_ir_graph)) {
1930 /* It's an empty block and not yet visited. */
1931 for (j = 0; j < get_Block_n_cfgpreds(pred); j++) {
1940 set_irn_in(phi, n_preds, in);
1942 phi = get_irn_link(phi);
1947 /*- Fix the block -*/
1949 for (i = 0; i < get_Block_n_cfgpreds(b); i++) {
1950 pred = get_nodes_Block(get_Block_cfgpred(b, i));
1951 if (is_Bad(get_Block_cfgpred(b, i))) {
1953 } else if (get_Block_block_visited(pred) +1
1954 < get_irg_block_visited(current_ir_graph)) {
1955 /* It's an empty block and not yet visited. */
1956 assert(get_Block_n_cfgpreds(b) > 1);
1957 /* Else it should be optimized by equivalent_node. */
1958 for (j = 0; j < get_Block_n_cfgpreds(pred); j++) {
1959 in[n_preds] = get_Block_cfgpred(pred, j);
1962 /* Remove block as it might be kept alive. */
1963 exchange(pred, b/*new_Bad()*/);
1965 in[n_preds] = get_Block_cfgpred(b, i);
1969 set_irn_in(b, n_preds, in);
1974 * an impossible ir_node * != NULL
1976 static const char _anchor;
1977 #define ANCHOR ((ir_node *)&_anchor)
1980 * optimize constant cond with projs
1982 * If we come here, we should have 3 situations:
1984 * 1.) Only one Proj exists: This should be the default Proj, convert to jump
1985 * 2.) Two Proj's exists: One is the taken, the other the not-taken default proj
1986 * 3.) More than 2 Proj's: local_optimize() did not run or internal error do nothing
1988 static void optimize_const_conds(ir_node *cond_list)
1990 ir_node *cond, *next;
1992 for (cond = cond_list; cond != ANCHOR; cond = next) {
1993 ir_node *block, *jmp;
1994 ir_node *proj = NULL, *def_proj = NULL;
1997 for (next = get_irn_link(cond); next != ANCHOR; next = get_irn_link(next)) {
1998 if (get_irn_op(next) == op_Cond)
2001 assert(get_irn_op(next) == op_Proj && "something neither Cond not Proj in cond-list");
2002 assert(get_Proj_pred(next) == cond && "cond-list corrupt");
2004 if (get_Proj_proj(next) == get_Cond_defaultProj(cond))
2014 block = get_nodes_block(def_proj);
2015 jmp = new_rd_Jmp(get_irn_dbg_info(def_proj), current_ir_graph, block);
2016 exchange(def_proj, jmp);
2018 else if (num == 2) {
2019 assert(def_proj && proj);
2021 exchange(def_proj, new_Bad());
2023 block = get_nodes_block(proj);
2024 jmp = new_rd_Jmp(get_irn_dbg_info(proj), current_ir_graph, block);
2025 exchange(proj, jmp);
2030 void optimize_cf(ir_graph *irg) {
2033 ir_node *end = get_irg_end(irg);
2034 ir_graph *rem = current_ir_graph;
2035 current_ir_graph = irg;
2036 ir_node *cond_list = ANCHOR;
2038 /* Handle graph state */
2039 assert(get_irg_phase_state(irg) != phase_building);
2040 if (get_irg_outs_state(current_ir_graph) == outs_consistent)
2041 set_irg_outs_inconsistent(current_ir_graph);
2042 if (get_irg_dom_state(current_ir_graph) == dom_consistent)
2043 set_irg_dom_inconsistent(current_ir_graph);
2045 /* Use block visited flag to mark non-empty blocks. */
2046 inc_irg_block_visited(irg);
2047 irg_walk(end, merge_blocks, collect_nodes, &cond_list);
2049 /* now, optimize switches */
2050 if (get_opt_unreachable_code())
2051 optimize_const_conds(cond_list);
2053 /* Optimize the standard code. */
2054 irg_block_walk(get_irg_end_block(irg), optimize_blocks, NULL, NULL);
2056 /* Walk all keep alives, optimize them if block, add to new in-array
2057 for end if useful. */
2058 in = NEW_ARR_F (ir_node *, 1);
2059 in[0] = get_nodes_Block(end);
2060 inc_irg_visited(current_ir_graph);
2061 for(i = 0; i < get_End_n_keepalives(end); i++) {
2062 ir_node *ka = get_End_keepalive(end, i);
2063 if (irn_not_visited(ka)) {
2064 if ((get_irn_op(ka) == op_Block) && Block_not_block_visited(ka)) {
2065 set_irg_block_visited(current_ir_graph, /* Don't walk all the way to Start. */
2066 get_irg_block_visited(current_ir_graph)-1);
2067 irg_block_walk(ka, optimize_blocks, NULL, NULL);
2068 mark_irn_visited(ka);
2069 ARR_APP1 (ir_node *, in, ka);
2070 } else if (get_irn_op(ka) == op_Phi) {
2071 mark_irn_visited(ka);
2072 ARR_APP1 (ir_node *, in, ka);
2076 /* DEL_ARR_F(end->in); GL @@@ tut nicht ! */
2079 current_ir_graph = rem;
2084 * Called by walker of remove_critical_cf_edges().
2086 * Place an empty block to an edge between a blocks of multiple
2087 * predecessors and a block of multiple successors.
2090 * @param env Environment of walker. This field is unused and has
2093 static void walk_critical_cf_edges(ir_node *n, void *env) {
2095 ir_node *pre, *block, **in, *jmp;
2097 arity = get_irn_arity(n);
2099 /* Block has multiple predecessors */
2100 if (op_Block == get_irn_op(n) && arity > 1) {
2102 if (n == get_irg_end_block(current_ir_graph))
2103 return; /* No use to add a block here. */
2105 for (i = 0; i < arity; ++i) {
2106 pre = get_irn_n(n, i);
2107 /* Predecessor has multiple successors. Insert new flow edge */
2108 if ((NULL != pre) &&
2109 (op_Proj == get_irn_op(pre)) &&
2110 op_Raise != get_irn_op(skip_Proj(pre))) {
2112 /* set predecessor array for new block */
2113 in = NEW_ARR_D (ir_node *, current_ir_graph->obst, 1);
2114 /* set predecessor of new block */
2116 block = new_Block(1, in);
2117 /* insert new jmp node to new block */
2118 switch_block(block);
2121 /* set successor of new block */
2122 set_irn_n(n, i, jmp);
2124 } /* predecessor has multiple successors */
2125 } /* for all predecessors */
2126 } /* n is a block */
2129 void remove_critical_cf_edges(ir_graph *irg) {
2130 if (get_opt_critical_edges())
2131 irg_walk_graph(irg, NULL, walk_critical_cf_edges, NULL);