2 * Copyright (C) 1995-2007 University of Karlsruhe. All right reserved.
4 * This file is part of libFirm.
6 * This file may be distributed and/or modified under the terms of the
7 * GNU General Public License version 2 as published by the Free Software
8 * Foundation and appearing in the file LICENSE.GPL included in the
9 * packaging of this file.
11 * Licensees holding valid libFirm Professional Edition licenses may use
12 * this file in accordance with the libFirm Commercial License.
13 * Agreement provided with the Software.
15 * This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE
16 * WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR
22 * @brief Optimizations for a whole ir graph, i.e., a procedure.
23 * @author Christian Schaefer, Goetz Lindenmaier, Sebastian Felis,
34 #include "irgraph_t.h"
37 #include "iroptimize.h"
47 #include "pdeq.h" /* Fuer code placement */
52 #include "irbackedge_t.h"
59 #include "iredges_t.h"
62 /*------------------------------------------------------------------*/
63 /* apply optimizations of iropt to all nodes. */
64 /*------------------------------------------------------------------*/
67 * A wrapper around optimize_inplace_2() to be called from a walker.
69 static void optimize_in_place_wrapper (ir_node *n, void *env) {
70 ir_node *optimized = optimize_in_place_2(n);
74 exchange (n, optimized);
79 * Do local optimizations for a node.
81 * @param n the IR-node where to start. Typically the End node
84 * @note current_ir_graph must be set
86 static INLINE void do_local_optimize(ir_node *n) {
87 /* Handle graph state */
88 assert(get_irg_phase_state(current_ir_graph) != phase_building);
90 if (get_opt_global_cse())
91 set_irg_pinned(current_ir_graph, op_pin_state_floats);
92 set_irg_outs_inconsistent(current_ir_graph);
93 set_irg_doms_inconsistent(current_ir_graph);
94 set_irg_loopinfo_inconsistent(current_ir_graph);
96 /* Clean the value_table in irg for the CSE. */
97 del_identities(current_ir_graph->value_table);
98 current_ir_graph->value_table = new_identities();
100 /* walk over the graph */
101 irg_walk(n, firm_clear_link, optimize_in_place_wrapper, NULL);
104 /* Applies local optimizations (see iropt.h) to all nodes reachable from node n */
105 void local_optimize_node(ir_node *n) {
106 ir_graph *rem = current_ir_graph;
107 current_ir_graph = get_irn_irg(n);
109 do_local_optimize(n);
111 current_ir_graph = rem;
115 * Block-Walker: uses dominance depth to mark dead blocks.
117 static void kill_dead_blocks(ir_node *block, void *env) {
120 if (get_Block_dom_depth(block) < 0) {
122 * Note that the new dominance code correctly handles
123 * the End block, i.e. it is always reachable from Start
125 set_Block_dead(block);
129 /* Applies local optimizations (see iropt.h) to all nodes reachable from node n. */
130 void local_optimize_graph(ir_graph *irg) {
131 ir_graph *rem = current_ir_graph;
132 current_ir_graph = irg;
134 if (get_irg_dom_state(irg) == dom_consistent)
135 irg_block_walk_graph(irg, NULL, kill_dead_blocks, NULL);
137 do_local_optimize(get_irg_end(irg));
139 current_ir_graph = rem;
143 * Enqueue all users of a node to a wait queue.
144 * Handles mode_T nodes.
146 static void enqueue_users(ir_node *n, pdeq *waitq) {
147 const ir_edge_t *edge;
149 foreach_out_edge(n, edge) {
150 ir_node *succ = get_edge_src_irn(edge);
152 if (get_irn_link(succ) != waitq) {
153 pdeq_putr(waitq, succ);
154 set_irn_link(succ, waitq);
156 if (get_irn_mode(succ) == mode_T) {
157 /* A mode_T node has Proj's. Because most optimizations
158 run on the Proj's we have to enqueue them also. */
159 enqueue_users(succ, waitq);
165 * Data flow optimization walker.
166 * Optimizes all nodes and enqueue it's users
169 static void opt_walker(ir_node *n, void *env) {
173 optimized = optimize_in_place_2(n);
174 set_irn_link(optimized, NULL);
176 if (optimized != n) {
177 enqueue_users(n, waitq);
178 exchange(n, optimized);
182 /* Applies local optimizations to all nodes in the graph until fixpoint. */
183 void optimize_graph_df(ir_graph *irg) {
184 pdeq *waitq = new_pdeq();
185 int state = edges_activated(irg);
186 ir_graph *rem = current_ir_graph;
190 current_ir_graph = irg;
195 if (get_opt_global_cse())
196 set_irg_pinned(current_ir_graph, op_pin_state_floats);
198 /* Clean the value_table in irg for the CSE. */
199 del_identities(irg->value_table);
200 irg->value_table = new_identities();
202 if (get_irg_dom_state(irg) == dom_consistent)
203 irg_block_walk_graph(irg, NULL, kill_dead_blocks, NULL);
205 /* invalidate info */
206 set_irg_outs_inconsistent(irg);
207 set_irg_doms_inconsistent(irg);
208 set_irg_loopinfo_inconsistent(irg);
210 set_using_irn_link(irg);
212 /* walk over the graph, but don't touch keep-alives */
213 irg_walk(get_irg_end_block(irg), NULL, opt_walker, waitq);
215 end = get_irg_end(irg);
217 /* optimize keep-alives by removing superfluous ones */
218 for (i = get_End_n_keepalives(end) - 1; i >= 0; --i) {
219 ir_node *ka = get_End_keepalive(end, i);
221 if (irn_visited(ka) && !is_irn_keep(ka)) {
222 /* this node can be regularly visited, no need to keep it */
223 set_End_keepalive(end, i, get_irg_bad(irg));
226 /* now walk again and visit all not yet visited nodes */
227 set_irg_visited(current_ir_graph, get_irg_visited(irg) - 1);
228 irg_walk(get_irg_end(irg), NULL, opt_walker, waitq);
230 /* finish the wait queue */
231 while (! pdeq_empty(waitq)) {
232 ir_node *n = pdeq_getl(waitq);
234 opt_walker(n, waitq);
239 clear_using_irn_link(irg);
242 edges_deactivate(irg);
244 current_ir_graph = rem;
248 /*------------------------------------------------------------------*/
249 /* Routines for dead node elimination / copying garbage collection */
250 /* of the obstack. */
251 /*------------------------------------------------------------------*/
254 * Remember the new node in the old node by using a field all nodes have.
256 #define set_new_node(oldn, newn) set_irn_link(oldn, newn)
259 * Get this new node, before the old node is forgotten.
261 #define get_new_node(oldn) get_irn_link(oldn)
264 * Check if a new node was set.
266 #define has_new_node(n) (get_new_node(n) != NULL)
269 * We use the block_visited flag to mark that we have computed the
270 * number of useful predecessors for this block.
271 * Further we encode the new arity in this flag in the old blocks.
272 * Remembering the arity is useful, as it saves a lot of pointer
273 * accesses. This function is called for all Phi and Block nodes
277 compute_new_arity(ir_node *b) {
278 int i, res, irn_arity;
281 irg_v = get_irg_block_visited(current_ir_graph);
282 block_v = get_Block_block_visited(b);
283 if (block_v >= irg_v) {
284 /* we computed the number of preds for this block and saved it in the
286 return block_v - irg_v;
288 /* compute the number of good predecessors */
289 res = irn_arity = get_irn_arity(b);
290 for (i = 0; i < irn_arity; i++)
291 if (is_Bad(get_irn_n(b, i))) res--;
292 /* save it in the flag. */
293 set_Block_block_visited(b, irg_v + res);
299 * Copies the node to the new obstack. The Ins of the new node point to
300 * the predecessors on the old obstack. For block/phi nodes not all
301 * predecessors might be copied. n->link points to the new node.
302 * For Phi and Block nodes the function allocates in-arrays with an arity
303 * only for useful predecessors. The arity is determined by counting
304 * the non-bad predecessors of the block.
306 * @param n The node to be copied
307 * @param env if non-NULL, the node number attribute will be copied to the new node
309 * Note: Also used for loop unrolling.
311 static void copy_node(ir_node *n, void *env) {
314 ir_op *op = get_irn_op(n);
316 /* The end node looses it's flexible in array. This doesn't matter,
317 as dead node elimination builds End by hand, inlineing doesn't use
319 /* assert(op == op_End || ((_ARR_DESCR(n->in))->cookie != ARR_F_MAGIC)); */
322 /* node copied already */
324 } else if (op == op_Block) {
326 new_arity = compute_new_arity(n);
327 n->attr.block.graph_arr = NULL;
329 block = get_nodes_block(n);
331 new_arity = compute_new_arity(block);
333 new_arity = get_irn_arity(n);
336 nn = new_ir_node(get_irn_dbg_info(n),
343 /* Copy the attributes. These might point to additional data. If this
344 was allocated on the old obstack the pointers now are dangling. This
345 frees e.g. the memory of the graph_arr allocated in new_immBlock. */
346 copy_node_attr(n, nn);
350 int copy_node_nr = env != NULL;
352 /* for easier debugging, we want to copy the node numbers too */
353 nn->node_nr = n->node_nr;
359 hook_dead_node_elim_subst(current_ir_graph, n, nn);
363 * Copies new predecessors of old node to new node remembered in link.
364 * Spare the Bad predecessors of Phi and Block nodes.
366 static void copy_preds(ir_node *n, void *env) {
371 nn = get_new_node(n);
374 /* copy the macro block header */
375 ir_node *mbh = get_Block_MacroBlock(n);
378 /* this block is a macroblock header */
379 set_irn_n(nn, -1, nn);
381 /* get the macro block header */
382 set_irn_n(nn, -1, get_new_node(mbh));
385 /* Don't copy Bad nodes. */
387 irn_arity = get_irn_arity(n);
388 for (i = 0; i < irn_arity; i++) {
389 if (! is_Bad(get_irn_n(n, i))) {
390 set_irn_n(nn, j, get_new_node(get_irn_n(n, i)));
391 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
395 /* repair the block visited flag from above misuse. Repair it in both
396 graphs so that the old one can still be used. */
397 set_Block_block_visited(nn, 0);
398 set_Block_block_visited(n, 0);
399 /* Local optimization could not merge two subsequent blocks if
400 in array contained Bads. Now it's possible.
401 We don't call optimize_in_place as it requires
402 that the fields in ir_graph are set properly. */
403 if ((get_opt_control_flow_straightening()) &&
404 (get_Block_n_cfgpreds(nn) == 1) &&
405 is_Jmp(get_Block_cfgpred(nn, 0))) {
406 ir_node *old = get_nodes_block(get_Block_cfgpred(nn, 0));
408 /* Jmp jumps into the block it is in -- deal self cycle. */
409 assert(is_Bad(get_new_node(get_irg_bad(current_ir_graph))));
410 exchange(nn, get_new_node(get_irg_bad(current_ir_graph)));
415 } else if (is_Phi(n)) {
416 /* Don't copy node if corresponding predecessor in block is Bad.
417 The Block itself should not be Bad. */
418 block = get_nodes_block(n);
419 set_irn_n(nn, -1, get_new_node(block));
421 irn_arity = get_irn_arity(n);
422 for (i = 0; i < irn_arity; i++) {
423 if (! is_Bad(get_irn_n(block, i))) {
424 set_irn_n(nn, j, get_new_node(get_irn_n(n, i)));
425 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
429 /* If the pre walker reached this Phi after the post walker visited the
430 block block_visited is > 0. */
431 set_Block_block_visited(get_nodes_block(n), 0);
432 /* Compacting the Phi's ins might generate Phis with only one
434 if (get_irn_arity(nn) == 1)
435 exchange(nn, get_irn_n(nn, 0));
437 irn_arity = get_irn_arity(n);
438 for (i = -1; i < irn_arity; i++)
439 set_irn_n (nn, i, get_new_node(get_irn_n(n, i)));
441 /* Now the new node is complete. We can add it to the hash table for CSE.
442 @@@ inlining aborts if we identify End. Why? */
444 add_identities(current_ir_graph->value_table, nn);
448 * Copies the graph recursively, compacts the keep-alives of the end node.
450 * @param irg the graph to be copied
451 * @param copy_node_nr If non-zero, the node number will be copied
453 static void copy_graph(ir_graph *irg, int copy_node_nr) {
454 ir_node *oe, *ne, *ob, *nb, *om, *nm; /* old end, new end, old bad, new bad, old NoMem, new NoMem */
455 ir_node *ka; /* keep alive */
459 /* Some nodes must be copied by hand, sigh */
460 vfl = get_irg_visited(irg);
461 set_irg_visited(irg, vfl + 1);
463 oe = get_irg_end(irg);
464 mark_irn_visited(oe);
465 /* copy the end node by hand, allocate dynamic in array! */
466 ne = new_ir_node(get_irn_dbg_info(oe),
473 /* Copy the attributes. Well, there might be some in the future... */
474 copy_node_attr(oe, ne);
475 set_new_node(oe, ne);
477 /* copy the Bad node */
478 ob = get_irg_bad(irg);
479 mark_irn_visited(ob);
480 nb = new_ir_node(get_irn_dbg_info(ob),
487 copy_node_attr(ob, nb);
488 set_new_node(ob, nb);
490 /* copy the NoMem node */
491 om = get_irg_no_mem(irg);
492 mark_irn_visited(om);
493 nm = new_ir_node(get_irn_dbg_info(om),
500 copy_node_attr(om, nm);
501 set_new_node(om, nm);
503 /* copy the live nodes */
504 set_irg_visited(irg, vfl);
505 irg_walk(get_nodes_block(oe), copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
507 /* Note: from yet, the visited flag of the graph is equal to vfl + 1 */
509 /* visit the anchors as well */
510 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
511 ir_node *n = get_irg_anchor(irg, i);
513 if (n && (get_irn_visited(n) <= vfl)) {
514 set_irg_visited(irg, vfl);
515 irg_walk(n, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
519 /* copy_preds for the end node ... */
520 set_nodes_block(ne, get_new_node(get_nodes_block(oe)));
522 /*- ... and now the keep alives. -*/
523 /* First pick the not marked block nodes and walk them. We must pick these
524 first as else we will oversee blocks reachable from Phis. */
525 irn_arity = get_End_n_keepalives(oe);
526 for (i = 0; i < irn_arity; i++) {
527 ka = get_End_keepalive(oe, i);
529 if (get_irn_visited(ka) <= vfl) {
530 /* We must keep the block alive and copy everything reachable */
531 set_irg_visited(irg, vfl);
532 irg_walk(ka, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
534 add_End_keepalive(ne, get_new_node(ka));
538 /* Now pick other nodes. Here we will keep all! */
539 irn_arity = get_End_n_keepalives(oe);
540 for (i = 0; i < irn_arity; i++) {
541 ka = get_End_keepalive(oe, i);
543 if (get_irn_visited(ka) <= vfl) {
544 /* We didn't copy the node yet. */
545 set_irg_visited(irg, vfl);
546 irg_walk(ka, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
548 add_End_keepalive(ne, get_new_node(ka));
552 /* start block sometimes only reached after keep alives */
553 set_nodes_block(nb, get_new_node(get_nodes_block(ob)));
554 set_nodes_block(nm, get_new_node(get_nodes_block(om)));
558 * Copies the graph reachable from current_ir_graph->end to the obstack
559 * in current_ir_graph and fixes the environment.
560 * Then fixes the fields in current_ir_graph containing nodes of the
563 * @param copy_node_nr If non-zero, the node number will be copied
566 copy_graph_env(int copy_node_nr) {
567 ir_graph *irg = current_ir_graph;
568 ir_node *old_end, *new_anchor;
571 /* remove end_except and end_reg nodes */
572 old_end = get_irg_end(irg);
573 set_irg_end_except (irg, old_end);
574 set_irg_end_reg (irg, old_end);
576 /* Not all nodes remembered in irg might be reachable
577 from the end node. Assure their link is set to NULL, so that
578 we can test whether new nodes have been computed. */
579 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
580 ir_node *n = get_irg_anchor(irg, i);
582 set_new_node(n, NULL);
584 /* we use the block walk flag for removing Bads from Blocks ins. */
585 inc_irg_block_visited(irg);
588 copy_graph(irg, copy_node_nr);
591 old_end = get_irg_end(irg);
592 new_anchor = new_Anchor(irg);
594 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
595 ir_node *n = get_irg_anchor(irg, i);
597 set_irn_n(new_anchor, i, get_new_node(n));
600 irg->anchor = new_anchor;
602 /* ensure the new anchor is placed in the endblock */
603 set_irn_n(new_anchor, -1, get_irg_end_block(irg));
607 * Copies all reachable nodes to a new obstack. Removes bad inputs
608 * from block nodes and the corresponding inputs from Phi nodes.
609 * Merges single exit blocks with single entry blocks and removes
611 * Adds all new nodes to a new hash table for CSE. Does not
612 * perform CSE, so the hash table might contain common subexpressions.
615 dead_node_elimination(ir_graph *irg) {
616 if (get_opt_optimize() && get_opt_dead_node_elimination()) {
618 int rem_ipview = get_interprocedural_view();
619 struct obstack *graveyard_obst = NULL;
620 struct obstack *rebirth_obst = NULL;
621 assert(! edges_activated(irg) && "dead node elimination requires disabled edges");
623 /* inform statistics that we started a dead-node elimination run */
624 hook_dead_node_elim(irg, 1);
626 /* Remember external state of current_ir_graph. */
627 rem = current_ir_graph;
628 current_ir_graph = irg;
629 set_interprocedural_view(0);
631 assert(get_irg_phase_state(irg) != phase_building);
633 /* Handle graph state */
634 free_callee_info(irg);
638 /* @@@ so far we loose loops when copying */
639 free_loop_information(irg);
641 set_irg_doms_inconsistent(irg);
643 /* A quiet place, where the old obstack can rest in peace,
644 until it will be cremated. */
645 graveyard_obst = irg->obst;
647 /* A new obstack, where the reachable nodes will be copied to. */
648 rebirth_obst = xmalloc(sizeof(*rebirth_obst));
649 irg->obst = rebirth_obst;
650 obstack_init(irg->obst);
651 irg->last_node_idx = 0;
653 /* We also need a new value table for CSE */
654 del_identities(irg->value_table);
655 irg->value_table = new_identities();
657 /* Copy the graph from the old to the new obstack */
658 copy_graph_env(/*copy_node_nr=*/1);
660 /* Free memory from old unoptimized obstack */
661 obstack_free(graveyard_obst, 0); /* First empty the obstack ... */
662 xfree(graveyard_obst); /* ... then free it. */
664 /* inform statistics that the run is over */
665 hook_dead_node_elim(irg, 0);
667 current_ir_graph = rem;
668 set_interprocedural_view(rem_ipview);
673 * Relink bad predecessors of a block and store the old in array to the
674 * link field. This function is called by relink_bad_predecessors().
675 * The array of link field starts with the block operand at position 0.
676 * If block has bad predecessors, create a new in array without bad preds.
677 * Otherwise let in array untouched.
679 static void relink_bad_block_predecessors(ir_node *n, void *env) {
680 ir_node **new_in, *irn;
681 int i, new_irn_n, old_irn_arity, new_irn_arity = 0;
684 /* if link field of block is NULL, look for bad predecessors otherwise
685 this is already done */
686 if (is_Block(n) && get_irn_link(n) == NULL) {
687 /* save old predecessors in link field (position 0 is the block operand)*/
688 set_irn_link(n, get_irn_in(n));
690 /* count predecessors without bad nodes */
691 old_irn_arity = get_irn_arity(n);
692 for (i = 0; i < old_irn_arity; i++)
693 if (!is_Bad(get_irn_n(n, i))) new_irn_arity++;
695 /* arity changing: set new predecessors without bad nodes */
696 if (new_irn_arity < old_irn_arity) {
697 /* Get new predecessor array. We do not resize the array, as we must
698 keep the old one to update Phis. */
699 new_in = NEW_ARR_D (ir_node *, current_ir_graph->obst, (new_irn_arity+1));
701 /* set new predecessors in array */
704 for (i = 0; i < old_irn_arity; i++) {
705 irn = get_irn_n(n, i);
707 new_in[new_irn_n] = irn;
708 is_backedge(n, i) ? set_backedge(n, new_irn_n-1) : set_not_backedge(n, new_irn_n-1);
712 /* ARR_SETLEN(int, n->attr.block.backedge, new_irn_arity); */
713 ARR_SHRINKLEN(n->attr.block.backedge, new_irn_arity);
715 } /* ir node has bad predecessors */
716 } /* Block is not relinked */
720 * Relinks Bad predecessors from Blocks and Phis called by walker
721 * remove_bad_predecesors(). If n is a Block, call
722 * relink_bad_block_redecessors(). If n is a Phi-node, call also the relinking
723 * function of Phi's Block. If this block has bad predecessors, relink preds
726 static void relink_bad_predecessors(ir_node *n, void *env) {
727 ir_node *block, **old_in;
728 int i, old_irn_arity, new_irn_arity;
730 /* relink bad predecessors of a block */
732 relink_bad_block_predecessors(n, env);
734 /* If Phi node relink its block and its predecessors */
736 /* Relink predecessors of phi's block */
737 block = get_nodes_block(n);
738 if (get_irn_link(block) == NULL)
739 relink_bad_block_predecessors(block, env);
741 old_in = (ir_node **)get_irn_link(block); /* Of Phi's Block */
742 old_irn_arity = ARR_LEN(old_in);
744 /* Relink Phi predecessors if count of predecessors changed */
745 if (old_irn_arity != ARR_LEN(get_irn_in(block))) {
746 /* set new predecessors in array
747 n->in[0] remains the same block */
749 for(i = 1; i < old_irn_arity; i++)
750 if (!is_Bad((ir_node *)old_in[i])) {
751 n->in[new_irn_arity] = n->in[i];
752 is_backedge(n, i) ? set_backedge(n, new_irn_arity) : set_not_backedge(n, new_irn_arity);
756 ARR_SETLEN(ir_node *, n->in, new_irn_arity);
757 ARR_SETLEN(int, n->attr.phi_backedge, new_irn_arity);
759 } /* n is a Phi node */
763 * Removes Bad Bad predecessors from Blocks and the corresponding
764 * inputs to Phi nodes as in dead_node_elimination but without
766 * On walking up set the link field to NULL, on walking down call
767 * relink_bad_predecessors() (This function stores the old in array
768 * to the link field and sets a new in array if arity of predecessors
771 void remove_bad_predecessors(ir_graph *irg) {
772 irg_walk_graph(irg, firm_clear_link, relink_bad_predecessors, NULL);
779 __)|_| | \_/ | \_/(/_ |_/\__|__
781 The following stuff implements a facility that automatically patches
782 registered ir_node pointers to the new node when a dead node elimination occurs.
785 struct _survive_dce_t {
789 hook_entry_t dead_node_elim;
790 hook_entry_t dead_node_elim_subst;
793 typedef struct _survive_dce_list_t {
794 struct _survive_dce_list_t *next;
796 } survive_dce_list_t;
798 static void dead_node_hook(void *context, ir_graph *irg, int start) {
799 survive_dce_t *sd = context;
802 /* Create a new map before the dead node elimination is performed. */
804 sd->new_places = pmap_create_ex(pmap_count(sd->places));
806 /* Patch back all nodes if dead node elimination is over and something is to be done. */
807 pmap_destroy(sd->places);
808 sd->places = sd->new_places;
809 sd->new_places = NULL;
814 * Hook called when dead node elimination replaces old by nw.
816 static void dead_node_subst_hook(void *context, ir_graph *irg, ir_node *old, ir_node *nw) {
817 survive_dce_t *sd = context;
818 survive_dce_list_t *list = pmap_get(sd->places, old);
821 /* If the node is to be patched back, write the new address to all registered locations. */
823 survive_dce_list_t *p;
825 for (p = list; p; p = p->next)
828 pmap_insert(sd->new_places, nw, list);
833 * Make a new Survive DCE environment.
835 survive_dce_t *new_survive_dce(void) {
836 survive_dce_t *res = xmalloc(sizeof(res[0]));
837 obstack_init(&res->obst);
838 res->places = pmap_create();
839 res->new_places = NULL;
841 res->dead_node_elim.hook._hook_dead_node_elim = dead_node_hook;
842 res->dead_node_elim.context = res;
843 res->dead_node_elim.next = NULL;
845 res->dead_node_elim_subst.hook._hook_dead_node_elim_subst = dead_node_subst_hook;
846 res->dead_node_elim_subst.context = res;
847 res->dead_node_elim_subst.next = NULL;
849 #ifndef FIRM_ENABLE_HOOKS
850 assert(0 && "need hooks enabled");
853 register_hook(hook_dead_node_elim, &res->dead_node_elim);
854 register_hook(hook_dead_node_elim_subst, &res->dead_node_elim_subst);
859 * Free a Survive DCE environment.
861 void free_survive_dce(survive_dce_t *sd) {
862 obstack_free(&sd->obst, NULL);
863 pmap_destroy(sd->places);
864 unregister_hook(hook_dead_node_elim, &sd->dead_node_elim);
865 unregister_hook(hook_dead_node_elim_subst, &sd->dead_node_elim_subst);
870 * Register a node pointer to be patched upon DCE.
871 * When DCE occurs, the node pointer specified by @p place will be
872 * patched to the new address of the node it is pointing to.
874 * @param sd The Survive DCE environment.
875 * @param place The address of the node pointer.
877 void survive_dce_register_irn(survive_dce_t *sd, ir_node **place) {
878 if (*place != NULL) {
879 ir_node *irn = *place;
880 survive_dce_list_t *curr = pmap_get(sd->places, irn);
881 survive_dce_list_t *nw = obstack_alloc(&sd->obst, sizeof(nw[0]));
886 pmap_insert(sd->places, irn, nw);
890 /*--------------------------------------------------------------------*/
891 /* Functionality for inlining */
892 /*--------------------------------------------------------------------*/
895 * Copy node for inlineing. Updates attributes that change when
896 * inlineing but not for dead node elimination.
898 * Copies the node by calling copy_node() and then updates the entity if
899 * it's a local one. env must be a pointer of the frame type of the
900 * inlined procedure. The new entities must be in the link field of
904 copy_node_inline(ir_node *n, void *env) {
906 ir_type *frame_tp = (ir_type *)env;
910 nn = get_new_node (n);
912 if (get_entity_owner(get_Sel_entity(n)) == frame_tp) {
913 set_Sel_entity(nn, get_entity_link(get_Sel_entity(n)));
915 } else if (is_Block(n)) {
916 nn = get_new_node (n);
917 nn->attr.block.irg = current_ir_graph;
922 * Walker: checks if P_value_arg_base is used.
924 static void find_addr(ir_node *node, void *env) {
925 int *allow_inline = env;
927 is_Start(get_Proj_pred(node)) &&
928 get_Proj_proj(node) == pn_Start_P_value_arg_base) {
934 * Check if we can inline a given call.
935 * Currently, we cannot inline two cases:
936 * - call with compound arguments
937 * - graphs that take the address of a parameter
939 * check these conditions here
941 static int can_inline(ir_node *call, ir_graph *called_graph) {
942 ir_type *call_type = get_Call_type(call);
943 int params, ress, i, res;
944 assert(is_Method_type(call_type));
946 params = get_method_n_params(call_type);
947 ress = get_method_n_ress(call_type);
949 /* check parameters for compound arguments */
950 for (i = 0; i < params; ++i) {
951 ir_type *p_type = get_method_param_type(call_type, i);
953 if (is_compound_type(p_type))
957 /* check results for compound arguments */
958 for (i = 0; i < ress; ++i) {
959 ir_type *r_type = get_method_res_type(call_type, i);
961 if (is_compound_type(r_type))
966 irg_walk_graph(called_graph, find_addr, NULL, &res);
972 exc_handler = 0, /**< There is a handler. */
973 exc_to_end = 1, /**< Branches to End. */
974 exc_no_handler = 2 /**< Exception handling not represented. */
977 /* Inlines a method at the given call site. */
978 int inline_method(ir_node *call, ir_graph *called_graph) {
980 ir_node *post_call, *post_bl;
981 ir_node *in[pn_Start_max];
982 ir_node *end, *end_bl;
986 int arity, n_ret, n_exc, n_res, i, j, rem_opt, irn_arity;
987 enum exc_mode exc_handling;
988 ir_type *called_frame;
989 irg_inline_property prop = get_irg_inline_property(called_graph);
991 if ( (prop < irg_inline_forced) &&
992 (!get_opt_optimize() || !get_opt_inline() || (prop == irg_inline_forbidden))) return 0;
994 /* Do not inline variadic functions. */
995 if (get_method_variadicity(get_entity_type(get_irg_entity(called_graph))) == variadicity_variadic)
998 assert(get_method_n_params(get_entity_type(get_irg_entity(called_graph))) ==
999 get_method_n_params(get_Call_type(call)));
1002 * currently, we cannot inline two cases:
1003 * - call with compound arguments
1004 * - graphs that take the address of a parameter
1006 if (! can_inline(call, called_graph))
1009 /* -- Turn off optimizations, this can cause problems when allocating new nodes. -- */
1010 rem_opt = get_opt_optimize();
1013 /* Handle graph state */
1014 assert(get_irg_phase_state(current_ir_graph) != phase_building);
1015 assert(get_irg_pinned(current_ir_graph) == op_pin_state_pinned);
1016 assert(get_irg_pinned(called_graph) == op_pin_state_pinned);
1017 set_irg_outs_inconsistent(current_ir_graph);
1018 set_irg_extblk_inconsistent(current_ir_graph);
1019 set_irg_doms_inconsistent(current_ir_graph);
1020 set_irg_loopinfo_inconsistent(current_ir_graph);
1021 set_irg_callee_info_state(current_ir_graph, irg_callee_info_inconsistent);
1023 /* -- Check preconditions -- */
1024 assert(is_Call(call));
1025 /* @@@ does not work for InterfaceIII.java after cgana
1026 assert(get_Call_type(call) == get_entity_type(get_irg_entity(called_graph)));
1027 assert(smaller_type(get_entity_type(get_irg_entity(called_graph)),
1028 get_Call_type(call)));
1030 if (called_graph == current_ir_graph) {
1031 set_optimize(rem_opt);
1035 /* here we know we WILL inline, so inform the statistics */
1036 hook_inline(call, called_graph);
1038 /* -- Decide how to handle exception control flow: Is there a handler
1039 for the Call node, or do we branch directly to End on an exception?
1041 0 There is a handler.
1043 2 Exception handling not represented in Firm. -- */
1045 ir_node *proj, *Mproj = NULL, *Xproj = NULL;
1046 for (proj = get_irn_link(call); proj; proj = get_irn_link(proj)) {
1047 long proj_nr = get_Proj_proj(proj);
1048 if (proj_nr == pn_Call_X_except) Xproj = proj;
1049 if (proj_nr == pn_Call_M_except) Mproj = proj;
1051 if (Mproj) { assert(Xproj); exc_handling = exc_handler; } /* Mproj */
1052 else if (Xproj) { exc_handling = exc_to_end; } /* !Mproj && Xproj */
1053 else { exc_handling = exc_no_handler; } /* !Mproj && !Xproj */
1057 the procedure and later replaces the Start node of the called graph.
1058 Post_call is the old Call node and collects the results of the called
1059 graph. Both will end up being a tuple. -- */
1060 post_bl = get_nodes_block(call);
1061 set_irg_current_block(current_ir_graph, post_bl);
1062 /* XxMxPxPxPxT of Start + parameter of Call */
1063 in[pn_Start_X_initial_exec] = new_Jmp();
1064 in[pn_Start_M] = get_Call_mem(call);
1065 in[pn_Start_P_frame_base] = get_irg_frame(current_ir_graph);
1066 in[pn_Start_P_globals] = get_irg_globals(current_ir_graph);
1067 in[pn_Start_P_tls] = get_irg_tls(current_ir_graph);
1068 in[pn_Start_T_args] = new_Tuple(get_Call_n_params(call), get_Call_param_arr(call));
1069 /* in[pn_Start_P_value_arg_base] = ??? */
1070 assert(pn_Start_P_value_arg_base == pn_Start_max - 1 && "pn_Start_P_value_arg_base not supported, fix");
1071 pre_call = new_Tuple(pn_Start_max - 1, in);
1075 The new block gets the ins of the old block, pre_call and all its
1076 predecessors and all Phi nodes. -- */
1077 part_block(pre_call);
1079 /* -- Prepare state for dead node elimination -- */
1080 /* Visited flags in calling irg must be >= flag in called irg.
1081 Else walker and arity computation will not work. */
1082 if (get_irg_visited(current_ir_graph) <= get_irg_visited(called_graph))
1083 set_irg_visited(current_ir_graph, get_irg_visited(called_graph)+1);
1084 if (get_irg_block_visited(current_ir_graph)< get_irg_block_visited(called_graph))
1085 set_irg_block_visited(current_ir_graph, get_irg_block_visited(called_graph));
1086 /* Set pre_call as new Start node in link field of the start node of
1087 calling graph and pre_calls block as new block for the start block
1089 Further mark these nodes so that they are not visited by the
1091 set_irn_link(get_irg_start(called_graph), pre_call);
1092 set_irn_visited(get_irg_start(called_graph), get_irg_visited(current_ir_graph));
1093 set_irn_link(get_irg_start_block(called_graph), get_nodes_block(pre_call));
1094 set_irn_visited(get_irg_start_block(called_graph), get_irg_visited(current_ir_graph));
1095 set_irn_link(get_irg_bad(called_graph), get_irg_bad(current_ir_graph));
1096 set_irn_visited(get_irg_bad(called_graph), get_irg_visited(current_ir_graph));
1098 /* Initialize for compaction of in arrays */
1099 inc_irg_block_visited(current_ir_graph);
1101 /* -- Replicate local entities of the called_graph -- */
1102 /* copy the entities. */
1103 called_frame = get_irg_frame_type(called_graph);
1104 for (i = 0; i < get_class_n_members(called_frame); i++) {
1105 ir_entity *new_ent, *old_ent;
1106 old_ent = get_class_member(called_frame, i);
1107 new_ent = copy_entity_own(old_ent, get_cur_frame_type());
1108 set_entity_link(old_ent, new_ent);
1111 /* visited is > than that of called graph. With this trick visited will
1112 remain unchanged so that an outer walker, e.g., searching the call nodes
1113 to inline, calling this inline will not visit the inlined nodes. */
1114 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
1116 /* -- Performing dead node elimination inlines the graph -- */
1117 /* Copies the nodes to the obstack of current_ir_graph. Updates links to new
1119 irg_walk(get_irg_end(called_graph), copy_node_inline, copy_preds,
1120 get_irg_frame_type(called_graph));
1122 /* Repair called_graph */
1123 set_irg_visited(called_graph, get_irg_visited(current_ir_graph));
1124 set_irg_block_visited(called_graph, get_irg_block_visited(current_ir_graph));
1125 set_Block_block_visited(get_irg_start_block(called_graph), 0);
1127 /* -- Merge the end of the inlined procedure with the call site -- */
1128 /* We will turn the old Call node into a Tuple with the following
1131 0: Phi of all Memories of Return statements.
1132 1: Jmp from new Block that merges the control flow from all exception
1133 predecessors of the old end block.
1134 2: Tuple of all arguments.
1135 3: Phi of Exception memories.
1136 In case the old Call directly branches to End on an exception we don't
1137 need the block merging all exceptions nor the Phi of the exception
1141 /* -- Precompute some values -- */
1142 end_bl = get_new_node(get_irg_end_block(called_graph));
1143 end = get_new_node(get_irg_end(called_graph));
1144 arity = get_irn_arity(end_bl); /* arity = n_exc + n_ret */
1145 n_res = get_method_n_ress(get_Call_type(call));
1147 res_pred = xmalloc(n_res * sizeof(*res_pred));
1148 cf_pred = xmalloc(arity * sizeof(*res_pred));
1150 set_irg_current_block(current_ir_graph, post_bl); /* just to make sure */
1152 /* -- archive keepalives -- */
1153 irn_arity = get_irn_arity(end);
1154 for (i = 0; i < irn_arity; i++) {
1155 ir_node *ka = get_End_keepalive(end, i);
1157 add_End_keepalive(get_irg_end(current_ir_graph), ka);
1160 /* The new end node will die. We need not free as the in array is on the obstack:
1161 copy_node() only generated 'D' arrays. */
1163 /* -- Replace Return nodes by Jump nodes. -- */
1165 for (i = 0; i < arity; i++) {
1167 ret = get_irn_n(end_bl, i);
1168 if (is_Return(ret)) {
1169 cf_pred[n_ret] = new_r_Jmp(current_ir_graph, get_nodes_block(ret));
1173 set_irn_in(post_bl, n_ret, cf_pred);
1175 /* -- Build a Tuple for all results of the method.
1176 Add Phi node if there was more than one Return. -- */
1177 turn_into_tuple(post_call, pn_Call_max);
1178 /* First the Memory-Phi */
1180 for (i = 0; i < arity; i++) {
1181 ret = get_irn_n(end_bl, i);
1182 if (is_Return(ret)) {
1183 cf_pred[n_ret] = get_Return_mem(ret);
1187 phi = new_Phi(n_ret, cf_pred, mode_M);
1188 set_Tuple_pred(call, pn_Call_M_regular, phi);
1189 /* Conserve Phi-list for further inlinings -- but might be optimized */
1190 if (get_nodes_block(phi) == post_bl) {
1191 set_irn_link(phi, get_irn_link(post_bl));
1192 set_irn_link(post_bl, phi);
1194 /* Now the real results */
1196 for (j = 0; j < n_res; j++) {
1198 for (i = 0; i < arity; i++) {
1199 ret = get_irn_n(end_bl, i);
1200 if (is_Return(ret)) {
1201 cf_pred[n_ret] = get_Return_res(ret, j);
1206 phi = new_Phi(n_ret, cf_pred, get_irn_mode(cf_pred[0]));
1210 /* Conserve Phi-list for further inlinings -- but might be optimized */
1211 if (get_nodes_block(phi) == post_bl) {
1212 set_irn_link(phi, get_irn_link(post_bl));
1213 set_irn_link(post_bl, phi);
1216 set_Tuple_pred(call, pn_Call_T_result, new_Tuple(n_res, res_pred));
1218 set_Tuple_pred(call, pn_Call_T_result, new_Bad());
1220 /* handle the regular call */
1221 set_Tuple_pred(call, pn_Call_X_regular, new_Jmp());
1223 /* For now, we cannot inline calls with value_base */
1224 set_Tuple_pred(call, pn_Call_P_value_res_base, new_Bad());
1226 /* Finally the exception control flow.
1227 We have two (three) possible situations:
1228 First if the Call branches to an exception handler: We need to add a Phi node to
1229 collect the memory containing the exception objects. Further we need
1230 to add another block to get a correct representation of this Phi. To
1231 this block we add a Jmp that resolves into the X output of the Call
1232 when the Call is turned into a tuple.
1233 Second the Call branches to End, the exception is not handled. Just
1234 add all inlined exception branches to the End node.
1235 Third: there is no Exception edge at all. Handle as case two. */
1236 if (exc_handling == exc_handler) {
1238 for (i = 0; i < arity; i++) {
1240 ret = get_irn_n(end_bl, i);
1241 irn = skip_Proj(ret);
1242 if (is_fragile_op(irn) || is_Raise(irn)) {
1243 cf_pred[n_exc] = ret;
1248 new_Block(n_exc, cf_pred); /* watch it: current_block is changed! */
1249 set_Tuple_pred(call, pn_Call_X_except, new_Jmp());
1250 /* The Phi for the memories with the exception objects */
1252 for (i = 0; i < arity; i++) {
1254 ret = skip_Proj(get_irn_n(end_bl, i));
1256 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 3);
1258 } else if (is_fragile_op(ret)) {
1259 /* We rely that all cfops have the memory output at the same position. */
1260 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 0);
1262 } else if (is_Raise(ret)) {
1263 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 1);
1267 set_Tuple_pred(call, pn_Call_M_except, new_Phi(n_exc, cf_pred, mode_M));
1269 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1270 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1273 ir_node *main_end_bl;
1274 int main_end_bl_arity;
1275 ir_node **end_preds;
1277 /* assert(exc_handling == 1 || no exceptions. ) */
1279 for (i = 0; i < arity; i++) {
1280 ir_node *ret = get_irn_n(end_bl, i);
1281 ir_node *irn = skip_Proj(ret);
1283 if (is_fragile_op(irn) || is_Raise(irn)) {
1284 cf_pred[n_exc] = ret;
1288 main_end_bl = get_irg_end_block(current_ir_graph);
1289 main_end_bl_arity = get_irn_arity(main_end_bl);
1290 end_preds = xmalloc((n_exc + main_end_bl_arity) * sizeof(*end_preds));
1292 for (i = 0; i < main_end_bl_arity; ++i)
1293 end_preds[i] = get_irn_n(main_end_bl, i);
1294 for (i = 0; i < n_exc; ++i)
1295 end_preds[main_end_bl_arity + i] = cf_pred[i];
1296 set_irn_in(main_end_bl, n_exc + main_end_bl_arity, end_preds);
1297 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1298 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1304 /* -- Turn CSE back on. -- */
1305 set_optimize(rem_opt);
1310 /********************************************************************/
1311 /* Apply inlineing to small methods. */
1312 /********************************************************************/
1314 /** Represents a possible inlinable call in a graph. */
1315 typedef struct _call_entry call_entry;
1316 struct _call_entry {
1317 ir_node *call; /**< the Call */
1318 ir_graph *callee; /**< the callee called here */
1319 call_entry *next; /**< for linking the next one */
1323 * environment for inlining small irgs
1325 typedef struct _inline_env_t {
1326 struct obstack obst; /**< an obstack where call_entries are allocated on. */
1327 call_entry *head; /**< the head of the call entry list */
1328 call_entry *tail; /**< the tail of the call entry list */
1332 * Returns the irg called from a Call node. If the irg is not
1333 * known, NULL is returned.
1335 static ir_graph *get_call_called_irg(ir_node *call) {
1337 ir_graph *called_irg = NULL;
1339 addr = get_Call_ptr(call);
1340 if (is_SymConst(addr) && get_SymConst_kind(addr) == symconst_addr_ent) {
1341 called_irg = get_entity_irg(get_SymConst_entity(addr));
1348 * Walker: Collect all calls to known graphs inside a graph.
1350 static void collect_calls(ir_node *call, void *env) {
1351 if (is_Call(call)) {
1352 ir_graph *called_irg = get_call_called_irg(call);
1354 /* The Call node calls a locally defined method. Remember to inline. */
1355 inline_env_t *ienv = env;
1356 call_entry *entry = obstack_alloc(&ienv->obst, sizeof(*entry));
1358 entry->callee = called_irg;
1361 if (ienv->tail == NULL)
1364 ienv->tail->next = entry;
1371 * Inlines all small methods at call sites where the called address comes
1372 * from a Const node that references the entity representing the called
1374 * The size argument is a rough measure for the code size of the method:
1375 * Methods where the obstack containing the firm graph is smaller than
1378 void inline_small_irgs(ir_graph *irg, int size) {
1379 ir_graph *rem = current_ir_graph;
1382 DEBUG_ONLY(firm_dbg_module_t *dbg;)
1384 if (!(get_opt_optimize() && get_opt_inline())) return;
1386 FIRM_DBG_REGISTER(dbg, "firm.opt.inline");
1388 current_ir_graph = irg;
1389 /* Handle graph state */
1390 assert(get_irg_phase_state(irg) != phase_building);
1391 free_callee_info(irg);
1393 /* Find Call nodes to inline.
1394 (We can not inline during a walk of the graph, as inlineing the same
1395 method several times changes the visited flag of the walked graph:
1396 after the first inlineing visited of the callee equals visited of
1397 the caller. With the next inlineing both are increased.) */
1398 obstack_init(&env.obst);
1399 env.head = env.tail = NULL;
1400 irg_walk_graph(irg, NULL, collect_calls, &env);
1402 if (env.head != NULL) {
1403 /* There are calls to inline */
1404 collect_phiprojs(irg);
1405 for (entry = env.head; entry != NULL; entry = entry->next) {
1406 ir_graph *callee = entry->callee;
1407 if (((_obstack_memory_used(callee->obst) - (int)obstack_room(callee->obst)) < size) ||
1408 (get_irg_inline_property(callee) >= irg_inline_forced)) {
1409 inline_method(entry->call, callee);
1413 obstack_free(&env.obst, NULL);
1414 current_ir_graph = rem;
1418 * Environment for inlining irgs.
1421 int n_nodes; /**< Number of nodes in graph except Id, Tuple, Proj, Start, End. */
1422 int n_nodes_orig; /**< for statistics */
1423 call_entry *call_head; /**< The head of the list of all call nodes in this graph. */
1424 call_entry *call_tail; /**< The tail of the list of all call nodes in this graph .*/
1425 int n_call_nodes; /**< Number of Call nodes in the graph. */
1426 int n_call_nodes_orig; /**< for statistics */
1427 int n_callers; /**< Number of known graphs that call this graphs. */
1428 int n_callers_orig; /**< for statistics */
1429 int got_inline; /**< Set, if at leat one call inside this graph was inlined. */
1433 * Allocate a new environment for inlining.
1435 static inline_irg_env *alloc_inline_irg_env(struct obstack *obst) {
1436 inline_irg_env *env = obstack_alloc(obst, sizeof(*env));
1437 env->n_nodes = -2; /* do not count count Start, End */
1438 env->n_nodes_orig = -2; /* do not count Start, End */
1439 env->call_head = NULL;
1440 env->call_tail = NULL;
1441 env->n_call_nodes = 0;
1442 env->n_call_nodes_orig = 0;
1444 env->n_callers_orig = 0;
1445 env->got_inline = 0;
1449 typedef struct walker_env {
1450 struct obstack *obst; /**< the obstack for allocations. */
1451 inline_irg_env *x; /**< the inline environment */
1452 int ignore_runtime; /**< the ignore runtime flag */
1456 * post-walker: collect all calls in the inline-environment
1457 * of a graph and sum some statistics.
1459 static void collect_calls2(ir_node *call, void *ctx) {
1461 inline_irg_env *x = env->x;
1462 ir_op *op = get_irn_op(call);
1466 /* count meaningful nodes in irg */
1467 if (op != op_Proj && op != op_Tuple && op != op_Sync) {
1472 if (op != op_Call) return;
1474 /* check, if it's a runtime call */
1475 if (env->ignore_runtime) {
1476 ir_node *symc = get_Call_ptr(call);
1478 if (is_SymConst(symc) && get_SymConst_kind(symc) == symconst_addr_ent) {
1479 ir_entity *ent = get_SymConst_entity(symc);
1481 if (get_entity_additional_properties(ent) & mtp_property_runtime)
1486 /* collect all call nodes */
1488 ++x->n_call_nodes_orig;
1490 callee = get_call_called_irg(call);
1492 inline_irg_env *callee_env = get_irg_link(callee);
1493 /* count all static callers */
1494 ++callee_env->n_callers;
1495 ++callee_env->n_callers_orig;
1497 /* link it in the list of possible inlinable entries */
1498 entry = obstack_alloc(env->obst, sizeof(*entry));
1500 entry->callee = callee;
1502 if (x->call_tail == NULL)
1503 x->call_head = entry;
1505 x->call_tail->next = entry;
1506 x->call_tail = entry;
1511 * Returns TRUE if the number of callers in 0 in the irg's environment,
1512 * hence this irg is a leave.
1514 INLINE static int is_leave(ir_graph *irg) {
1515 inline_irg_env *env = get_irg_link(irg);
1516 return env->n_call_nodes == 0;
1520 * Returns TRUE if the number of callers is smaller size in the irg's environment.
1522 INLINE static int is_smaller(ir_graph *callee, int size) {
1523 inline_irg_env *env = get_irg_link(callee);
1524 return env->n_nodes < size;
1528 * Append the nodes of the list src to the nodes of the list in environment dst.
1530 static void append_call_list(struct obstack *obst, inline_irg_env *dst, call_entry *src) {
1531 call_entry *entry, *nentry;
1533 /* Note that the src list points to Call nodes in the inlined graph, but
1534 we need Call nodes in our graph. Luckily the inliner leaves this information
1535 in the link field. */
1536 for (entry = src; entry != NULL; entry = entry->next) {
1537 nentry = obstack_alloc(obst, sizeof(*nentry));
1538 nentry->call = get_irn_link(entry->call);
1539 nentry->callee = entry->callee;
1540 nentry->next = NULL;
1541 dst->call_tail->next = nentry;
1542 dst->call_tail = nentry;
1547 * Inlines small leave methods at call sites where the called address comes
1548 * from a Const node that references the entity representing the called
1550 * The size argument is a rough measure for the code size of the method:
1551 * Methods where the obstack containing the firm graph is smaller than
1554 void inline_leave_functions(int maxsize, int leavesize, int size, int ignore_runtime) {
1555 inline_irg_env *env;
1561 call_entry *entry, *tail;
1562 const call_entry *centry;
1563 struct obstack obst;
1564 DEBUG_ONLY(firm_dbg_module_t *dbg;)
1566 if (!(get_opt_optimize() && get_opt_inline())) return;
1568 FIRM_DBG_REGISTER(dbg, "firm.opt.inline");
1569 rem = current_ir_graph;
1570 obstack_init(&obst);
1572 /* extend all irgs by a temporary data structure for inlining. */
1573 n_irgs = get_irp_n_irgs();
1574 for (i = 0; i < n_irgs; ++i)
1575 set_irg_link(get_irp_irg(i), alloc_inline_irg_env(&obst));
1577 /* Precompute information in temporary data structure. */
1579 wenv.ignore_runtime = ignore_runtime;
1580 for (i = 0; i < n_irgs; ++i) {
1581 ir_graph *irg = get_irp_irg(i);
1583 assert(get_irg_phase_state(irg) != phase_building);
1584 free_callee_info(irg);
1586 wenv.x = get_irg_link(irg);
1587 irg_walk_graph(irg, NULL, collect_calls2, &wenv);
1590 /* -- and now inline. -- */
1592 /* Inline leaves recursively -- we might construct new leaves. */
1596 for (i = 0; i < n_irgs; ++i) {
1598 int phiproj_computed = 0;
1600 current_ir_graph = get_irp_irg(i);
1601 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1604 for (entry = env->call_head; entry != NULL; entry = entry->next) {
1607 if (env->n_nodes > maxsize) break;
1610 callee = entry->callee;
1612 if (is_leave(callee) && is_smaller(callee, leavesize)) {
1613 if (!phiproj_computed) {
1614 phiproj_computed = 1;
1615 collect_phiprojs(current_ir_graph);
1617 did_inline = inline_method(call, callee);
1620 /* Do some statistics */
1621 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1623 env->got_inline = 1;
1624 --env->n_call_nodes;
1625 env->n_nodes += callee_env->n_nodes;
1626 --callee_env->n_callers;
1628 /* remove this call from the list */
1630 tail->next = entry->next;
1632 env->call_head = entry->next;
1638 env->call_tail = tail;
1640 } while (did_inline);
1642 /* inline other small functions. */
1643 for (i = 0; i < n_irgs; ++i) {
1645 int phiproj_computed = 0;
1647 current_ir_graph = get_irp_irg(i);
1648 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1650 /* note that the list of possible calls is updated during the process */
1652 for (entry = env->call_head; entry != NULL; entry = entry->next) {
1656 callee = entry->callee;
1658 if (((is_smaller(callee, size) && (env->n_nodes < maxsize)) || /* small function */
1659 (get_irg_inline_property(callee) >= irg_inline_forced))) {
1660 if (!phiproj_computed) {
1661 phiproj_computed = 1;
1662 collect_phiprojs(current_ir_graph);
1664 if (inline_method(call, callee)) {
1665 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1667 /* callee was inline. Append it's call list. */
1668 env->got_inline = 1;
1669 --env->n_call_nodes;
1670 append_call_list(&obst, env, callee_env->call_head);
1671 env->n_call_nodes += callee_env->n_call_nodes;
1672 env->n_nodes += callee_env->n_nodes;
1673 --callee_env->n_callers;
1675 /* after we have inlined callee, all called methods inside callee
1676 are now called once more */
1677 for (centry = callee_env->call_head; centry != NULL; centry = centry->next) {
1678 inline_irg_env *penv = get_irg_link(centry->callee);
1682 /* remove this call from the list */
1684 tail->next = entry->next;
1686 env->call_head = entry->next;
1692 env->call_tail = tail;
1695 for (i = 0; i < n_irgs; ++i) {
1696 irg = get_irp_irg(i);
1697 env = (inline_irg_env *)get_irg_link(irg);
1699 if (env->got_inline) {
1700 /* this irg got calls inlined */
1701 set_irg_outs_inconsistent(irg);
1702 set_irg_doms_inconsistent(irg);
1704 optimize_graph_df(irg);
1707 if (env->got_inline || (env->n_callers_orig != env->n_callers))
1708 DB((dbg, SET_LEVEL_1, "Nodes:%3d ->%3d, calls:%3d ->%3d, callers:%3d ->%3d, -- %s\n",
1709 env->n_nodes_orig, env->n_nodes, env->n_call_nodes_orig, env->n_call_nodes,
1710 env->n_callers_orig, env->n_callers,
1711 get_entity_name(get_irg_entity(irg))));
1714 obstack_free(&obst, NULL);
1715 current_ir_graph = rem;
1718 /*******************************************************************/
1719 /* Code Placement. Pins all floating nodes to a block where they */
1720 /* will be executed only if needed. */
1721 /*******************************************************************/
1724 * Returns non-zero, is a block is not reachable from Start.
1726 * @param block the block to test
1729 is_Block_unreachable(ir_node *block) {
1730 return is_Block_dead(block) || get_Block_dom_depth(block) < 0;
1734 * Find the earliest correct block for node n. --- Place n into the
1735 * same Block as its dominance-deepest Input.
1737 * We have to avoid calls to get_nodes_block() here
1738 * because the graph is floating.
1740 * move_out_of_loops() expects that place_floats_early() have placed
1741 * all "living" nodes into a living block. That's why we must
1742 * move nodes in dead block with "live" successors into a valid
1744 * We move them just into the same block as it's successor (or
1745 * in case of a Phi into the effective use block). For Phi successors,
1746 * this may still be a dead block, but then there is no real use, as
1747 * the control flow will be dead later.
1749 * @param n the node to be placed
1750 * @param worklist a worklist, predecessors of non-floating nodes are placed here
1753 place_floats_early(ir_node *n, waitq *worklist) {
1756 /* we must not run into an infinite loop */
1757 assert(irn_not_visited(n));
1758 mark_irn_visited(n);
1760 /* Place floating nodes. */
1761 if (get_irn_pinned(n) == op_pin_state_floats) {
1762 ir_node *curr_block = get_nodes_block(n);
1763 int in_dead_block = is_Block_unreachable(curr_block);
1765 ir_node *b = NULL; /* The block to place this node in */
1767 assert(is_no_Block(n));
1769 if (is_irn_start_block_placed(n)) {
1770 /* These nodes will not be placed by the loop below. */
1771 b = get_irg_start_block(current_ir_graph);
1775 /* find the block for this node. */
1776 irn_arity = get_irn_arity(n);
1777 for (i = 0; i < irn_arity; i++) {
1778 ir_node *pred = get_irn_n(n, i);
1779 ir_node *pred_block;
1781 if ((irn_not_visited(pred))
1782 && (get_irn_pinned(pred) == op_pin_state_floats)) {
1785 * If the current node is NOT in a dead block, but one of its
1786 * predecessors is, we must move the predecessor to a live block.
1787 * Such thing can happen, if global CSE chose a node from a dead block.
1788 * We move it simply to our block.
1789 * Note that neither Phi nor End nodes are floating, so we don't
1790 * need to handle them here.
1792 if (! in_dead_block) {
1793 if (get_irn_pinned(pred) == op_pin_state_floats &&
1794 is_Block_unreachable(get_nodes_block(pred)))
1795 set_nodes_block(pred, curr_block);
1797 place_floats_early(pred, worklist);
1801 * A node in the Bad block must stay in the bad block,
1802 * so don't compute a new block for it.
1807 /* Because all loops contain at least one op_pin_state_pinned node, now all
1808 our inputs are either op_pin_state_pinned or place_early() has already
1809 been finished on them. We do not have any unfinished inputs! */
1810 pred_block = get_nodes_block(pred);
1811 if ((!is_Block_dead(pred_block)) &&
1812 (get_Block_dom_depth(pred_block) > depth)) {
1814 depth = get_Block_dom_depth(pred_block);
1816 /* Avoid that the node is placed in the Start block */
1818 get_Block_dom_depth(get_nodes_block(n)) > 1 &&
1819 get_irg_phase_state(current_ir_graph) != phase_backend) {
1820 b = get_Block_cfg_out(get_irg_start_block(current_ir_graph), 0);
1821 assert(b != get_irg_start_block(current_ir_graph));
1826 set_nodes_block(n, b);
1830 * Add predecessors of non floating nodes and non-floating predecessors
1831 * of floating nodes to worklist and fix their blocks if the are in dead block.
1833 irn_arity = get_irn_arity(n);
1837 * Simplest case: End node. Predecessors are keep-alives,
1838 * no need to move out of dead block.
1840 for (i = -1; i < irn_arity; ++i) {
1841 ir_node *pred = get_irn_n(n, i);
1842 if (irn_not_visited(pred))
1843 waitq_put(worklist, pred);
1845 } else if (is_Block(n)) {
1847 * Blocks: Predecessors are control flow, no need to move
1848 * them out of dead block.
1850 for (i = irn_arity - 1; i >= 0; --i) {
1851 ir_node *pred = get_irn_n(n, i);
1852 if (irn_not_visited(pred))
1853 waitq_put(worklist, pred);
1855 } else if (is_Phi(n)) {
1857 ir_node *curr_block = get_nodes_block(n);
1858 int in_dead_block = is_Block_unreachable(curr_block);
1861 * Phi nodes: move nodes from dead blocks into the effective use
1862 * of the Phi-input if the Phi is not in a bad block.
1864 pred = get_nodes_block(n);
1865 if (irn_not_visited(pred))
1866 waitq_put(worklist, pred);
1868 for (i = irn_arity - 1; i >= 0; --i) {
1869 ir_node *pred = get_irn_n(n, i);
1871 if (irn_not_visited(pred)) {
1872 if (! in_dead_block &&
1873 get_irn_pinned(pred) == op_pin_state_floats &&
1874 is_Block_unreachable(get_nodes_block(pred))) {
1875 set_nodes_block(pred, get_Block_cfgpred_block(curr_block, i));
1877 waitq_put(worklist, pred);
1882 ir_node *curr_block = get_nodes_block(n);
1883 int in_dead_block = is_Block_unreachable(curr_block);
1886 * All other nodes: move nodes from dead blocks into the same block.
1888 pred = get_nodes_block(n);
1889 if (irn_not_visited(pred))
1890 waitq_put(worklist, pred);
1892 for (i = irn_arity - 1; i >= 0; --i) {
1893 ir_node *pred = get_irn_n(n, i);
1895 if (irn_not_visited(pred)) {
1896 if (! in_dead_block &&
1897 get_irn_pinned(pred) == op_pin_state_floats &&
1898 is_Block_unreachable(get_nodes_block(pred))) {
1899 set_nodes_block(pred, curr_block);
1901 waitq_put(worklist, pred);
1908 * Floating nodes form subgraphs that begin at nodes as Const, Load,
1909 * Start, Call and that end at op_pin_state_pinned nodes as Store, Call. Place_early
1910 * places all floating nodes reachable from its argument through floating
1911 * nodes and adds all beginnings at op_pin_state_pinned nodes to the worklist.
1913 * @param worklist a worklist, used for the algorithm, empty on in/output
1915 static void place_early(waitq *worklist) {
1917 inc_irg_visited(current_ir_graph);
1919 /* this inits the worklist */
1920 place_floats_early(get_irg_end(current_ir_graph), worklist);
1922 /* Work the content of the worklist. */
1923 while (!waitq_empty(worklist)) {
1924 ir_node *n = waitq_get(worklist);
1925 if (irn_not_visited(n))
1926 place_floats_early(n, worklist);
1929 set_irg_outs_inconsistent(current_ir_graph);
1930 set_irg_pinned(current_ir_graph, op_pin_state_pinned);
1934 * Compute the deepest common ancestor of block and dca.
1936 static ir_node *calc_dca(ir_node *dca, ir_node *block) {
1939 /* we do not want to place nodes in dead blocks */
1940 if (is_Block_dead(block))
1943 /* We found a first legal placement. */
1944 if (!dca) return block;
1946 /* Find a placement that is dominates both, dca and block. */
1947 while (get_Block_dom_depth(block) > get_Block_dom_depth(dca))
1948 block = get_Block_idom(block);
1950 while (get_Block_dom_depth(dca) > get_Block_dom_depth(block)) {
1951 dca = get_Block_idom(dca);
1954 while (block != dca) {
1955 block = get_Block_idom(block); dca = get_Block_idom(dca);
1961 /** Deepest common dominance ancestor of DCA and CONSUMER of PRODUCER.
1962 * I.e., DCA is the block where we might place PRODUCER.
1963 * A data flow edge points from producer to consumer.
1965 static ir_node *consumer_dom_dca(ir_node *dca, ir_node *consumer, ir_node *producer)
1967 /* Compute the last block into which we can place a node so that it is
1969 if (is_Phi(consumer)) {
1970 /* our consumer is a Phi-node, the effective use is in all those
1971 blocks through which the Phi-node reaches producer */
1972 ir_node *phi_block = get_nodes_block(consumer);
1973 int arity = get_irn_arity(consumer);
1976 for (i = 0; i < arity; i++) {
1977 if (get_Phi_pred(consumer, i) == producer) {
1978 ir_node *new_block = get_Block_cfgpred_block(phi_block, i);
1980 if (!is_Block_unreachable(new_block))
1981 dca = calc_dca(dca, new_block);
1985 dca = calc_dca(dca, get_nodes_block(consumer));
1991 /* FIXME: the name clashes here with the function from ana/field_temperature.c
1993 static INLINE int get_irn_loop_depth(ir_node *n) {
1994 return get_loop_depth(get_irn_loop(n));
1998 * Move n to a block with less loop depth than it's current block. The
1999 * new block must be dominated by early.
2001 * @param n the node that should be moved
2002 * @param early the earliest block we can n move to
2004 static void move_out_of_loops(ir_node *n, ir_node *early) {
2005 ir_node *best, *dca;
2009 /* Find the region deepest in the dominator tree dominating
2010 dca with the least loop nesting depth, but still dominated
2011 by our early placement. */
2012 dca = get_nodes_block(n);
2015 while (dca != early) {
2016 dca = get_Block_idom(dca);
2017 if (!dca || is_Bad(dca)) break; /* may be Bad if not reachable from Start */
2018 if (get_irn_loop_depth(dca) < get_irn_loop_depth(best)) {
2022 if (best != get_nodes_block(n)) {
2024 printf("Moving out of loop: "); DDMN(n);
2025 printf(" Outermost block: "); DDMN(early);
2026 printf(" Best block: "); DDMN(best);
2027 printf(" Innermost block: "); DDMN(get_nodes_block(n));
2029 set_nodes_block(n, best);
2033 /* deepest common ancestor in the dominator tree of all nodes'
2034 blocks depending on us; our final placement has to dominate DCA. */
2035 static ir_node *get_deepest_common_ancestor(ir_node *node, ir_node *dca)
2039 for (i = get_irn_n_outs(node) - 1; i >= 0; --i) {
2040 ir_node *succ = get_irn_out(node, i);
2044 * This consumer is the End node, a keep alive edge.
2045 * This is not a real consumer, so we ignore it
2050 if (is_Proj(succ)) {
2051 dca = get_deepest_common_ancestor(succ, dca);
2053 /* ignore if succ is in dead code */
2054 ir_node *succ_blk = get_nodes_block(succ);
2055 if (is_Block_unreachable(succ_blk))
2057 dca = consumer_dom_dca(dca, succ, node);
2064 static void set_projs_block(ir_node *node, ir_node *block)
2068 for (i = get_irn_n_outs(node) - 1; i >= 0; --i) {
2069 ir_node *succ = get_irn_out(node, i);
2071 assert(is_Proj(succ));
2073 if(get_irn_mode(succ) == mode_T) {
2074 set_projs_block(succ, block);
2076 set_nodes_block(succ, block);
2081 * Find the latest legal block for N and place N into the
2082 * `optimal' Block between the latest and earliest legal block.
2083 * The `optimal' block is the dominance-deepest block of those
2084 * with the least loop-nesting-depth. This places N out of as many
2085 * loops as possible and then makes it as control dependent as
2088 * @param n the node to be placed
2089 * @param worklist a worklist, all successors of non-floating nodes are
2092 static void place_floats_late(ir_node *n, pdeq *worklist) {
2096 assert(irn_not_visited(n)); /* no multiple placement */
2098 mark_irn_visited(n);
2100 /* no need to place block nodes, control nodes are already placed. */
2103 (get_irn_mode(n) != mode_X)) {
2104 /* Remember the early_blk placement of this block to move it
2105 out of loop no further than the early_blk placement. */
2106 early_blk = get_nodes_block(n);
2109 * BEWARE: Here we also get code, that is live, but
2110 * was in a dead block. If the node is life, but because
2111 * of CSE in a dead block, we still might need it.
2114 /* Assure that our users are all placed, except the Phi-nodes.
2115 --- Each data flow cycle contains at least one Phi-node. We
2116 have to break the `user has to be placed before the
2117 producer' dependence cycle and the Phi-nodes are the
2118 place to do so, because we need to base our placement on the
2119 final region of our users, which is OK with Phi-nodes, as they
2120 are op_pin_state_pinned, and they never have to be placed after a
2121 producer of one of their inputs in the same block anyway. */
2122 for (i = get_irn_n_outs(n) - 1; i >= 0; --i) {
2123 ir_node *succ = get_irn_out(n, i);
2124 if (irn_not_visited(succ) && !is_Phi(succ))
2125 place_floats_late(succ, worklist);
2128 if (! is_Block_dead(early_blk)) {
2129 /* do only move things that where not dead */
2130 ir_op *op = get_irn_op(n);
2132 /* We have to determine the final block of this node... except for
2133 constants and Projs */
2134 if ((get_irn_pinned(n) == op_pin_state_floats) &&
2136 (op != op_SymConst) &&
2139 /* deepest common ancestor in the dominator tree of all nodes'
2140 blocks depending on us; our final placement has to dominate
2142 ir_node *dca = get_deepest_common_ancestor(n, NULL);
2144 set_nodes_block(n, dca);
2145 move_out_of_loops(n, early_blk);
2146 if(get_irn_mode(n) == mode_T) {
2147 set_projs_block(n, get_nodes_block(n));
2154 /* Add successors of all non-floating nodes on list. (Those of floating
2155 nodes are placed already and therefore are marked.) */
2156 for (i = 0; i < get_irn_n_outs(n); i++) {
2157 ir_node *succ = get_irn_out(n, i);
2158 if (irn_not_visited(get_irn_out(n, i))) {
2159 pdeq_putr(worklist, succ);
2165 * Place floating nodes on the given worklist as late as possible using
2166 * the dominance tree.
2168 * @param worklist the worklist containing the nodes to place
2170 static void place_late(waitq *worklist) {
2172 inc_irg_visited(current_ir_graph);
2174 /* This fills the worklist initially. */
2175 place_floats_late(get_irg_start_block(current_ir_graph), worklist);
2177 /* And now empty the worklist again... */
2178 while (!waitq_empty(worklist)) {
2179 ir_node *n = waitq_get(worklist);
2180 if (irn_not_visited(n))
2181 place_floats_late(n, worklist);
2185 /* Code Placement. */
2186 void place_code(ir_graph *irg) {
2188 ir_graph *rem = current_ir_graph;
2190 current_ir_graph = irg;
2192 /* Handle graph state */
2193 assert(get_irg_phase_state(irg) != phase_building);
2196 if (1 || get_irg_loopinfo_state(irg) != loopinfo_consistent) {
2197 free_loop_information(irg);
2198 construct_cf_backedges(irg);
2201 /* Place all floating nodes as early as possible. This guarantees
2202 a legal code placement. */
2203 worklist = new_waitq();
2204 place_early(worklist);
2206 /* place_early() invalidates the outs, place_late needs them. */
2207 compute_irg_outs(irg);
2209 /* Now move the nodes down in the dominator tree. This reduces the
2210 unnecessary executions of the node. */
2211 place_late(worklist);
2213 set_irg_outs_inconsistent(current_ir_graph);
2214 set_irg_loopinfo_inconsistent(current_ir_graph);
2215 del_waitq(worklist);
2216 current_ir_graph = rem;
2219 typedef struct cf_env {
2220 char changed; /**< flag indicates that the cf graphs has changed. */
2224 * Called by walker of remove_critical_cf_edges().
2226 * Place an empty block to an edge between a blocks of multiple
2227 * predecessors and a block of multiple successors.
2230 * @param env Environment of walker.
2232 static void walk_critical_cf_edges(ir_node *n, void *env) {
2234 ir_node *pre, *block, *jmp;
2236 ir_graph *irg = get_irn_irg(n);
2238 /* Block has multiple predecessors */
2239 arity = get_irn_arity(n);
2241 if (n == get_irg_end_block(irg))
2242 return; /* No use to add a block here. */
2244 for (i = 0; i < arity; ++i) {
2247 pre = get_irn_n(n, i);
2248 cfop = get_irn_op(skip_Proj(pre));
2250 if (is_op_fragile(cfop)) {
2251 if (cfop != op_Raise)
2255 if (is_op_forking(cfop)) {
2256 /* Predecessor has multiple successors. Insert new control flow edge edges. */
2258 /* set predecessor of new block */
2259 block = new_r_Block(irg, 1, &pre);
2260 /* insert new jmp node to new block */
2261 jmp = new_r_Jmp(irg, block);
2262 /* set successor of new block */
2263 set_irn_n(n, i, jmp);
2265 } /* predecessor has multiple successors */
2266 } /* for all predecessors */
2267 } /* n is a multi-entry block */
2270 void remove_critical_cf_edges(ir_graph *irg) {
2275 irg_block_walk_graph(irg, NULL, walk_critical_cf_edges, &env);
2277 /* control flow changed */
2278 set_irg_outs_inconsistent(irg);
2279 set_irg_extblk_inconsistent(irg);
2280 set_irg_doms_inconsistent(irg);
2281 set_irg_loopinfo_inconsistent(irg);