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);
317 /* The end node looses it's flexible in array. This doesn't matter,
318 as dead node elimination builds End by hand, inlineing doesn't use
320 /* assert(op == op_End || ((_ARR_DESCR(n->in))->cookie != ARR_F_MAGIC)); */
323 /* node copied already */
325 } else if (op == op_Block) {
327 new_arity = compute_new_arity(n);
328 n->attr.block.graph_arr = NULL;
330 block = get_nodes_block(n);
332 new_arity = compute_new_arity(block);
334 new_arity = get_irn_arity(n);
337 nn = new_ir_node(get_irn_dbg_info(n),
344 /* Copy the attributes. These might point to additional data. If this
345 was allocated on the old obstack the pointers now are dangling. This
346 frees e.g. the memory of the graph_arr allocated in new_immBlock. */
347 copy_node_attr(n, nn);
351 int copy_node_nr = env != NULL;
353 /* for easier debugging, we want to copy the node numbers too */
354 nn->node_nr = n->node_nr;
360 hook_dead_node_elim_subst(current_ir_graph, n, nn);
364 * Copies new predecessors of old node to new node remembered in link.
365 * Spare the Bad predecessors of Phi and Block nodes.
367 static void copy_preds(ir_node *n, void *env) {
372 nn = get_new_node(n);
375 /* copy the macro block header */
376 ir_node *mbh = get_Block_MacroBlock(n);
379 /* this block is a macroblock header */
380 set_irn_n(nn, -1, nn);
382 /* get the macro block header */
383 set_irn_n(nn, -1, get_new_node(mbh));
386 /* Don't copy Bad nodes. */
388 irn_arity = get_irn_arity(n);
389 for (i = 0; i < irn_arity; i++) {
390 if (! is_Bad(get_irn_n(n, i))) {
391 set_irn_n(nn, j, get_new_node(get_irn_n(n, i)));
392 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
396 /* repair the block visited flag from above misuse. Repair it in both
397 graphs so that the old one can still be used. */
398 set_Block_block_visited(nn, 0);
399 set_Block_block_visited(n, 0);
400 /* Local optimization could not merge two subsequent blocks if
401 in array contained Bads. Now it's possible.
402 We don't call optimize_in_place as it requires
403 that the fields in ir_graph are set properly. */
404 if ((get_opt_control_flow_straightening()) &&
405 (get_Block_n_cfgpreds(nn) == 1) &&
406 is_Jmp(get_Block_cfgpred(nn, 0))) {
407 ir_node *old = get_nodes_block(get_Block_cfgpred(nn, 0));
409 /* Jmp jumps into the block it is in -- deal self cycle. */
410 assert(is_Bad(get_new_node(get_irg_bad(current_ir_graph))));
411 exchange(nn, get_new_node(get_irg_bad(current_ir_graph)));
416 } else if (is_Phi(n) && get_irn_arity(n) > 0) {
417 /* Don't copy node if corresponding predecessor in block is Bad.
418 The Block itself should not be Bad. */
419 block = get_nodes_block(n);
420 set_irn_n(nn, -1, get_new_node(block));
422 irn_arity = get_irn_arity(n);
423 for (i = 0; i < irn_arity; i++) {
424 if (! is_Bad(get_irn_n(block, i))) {
425 set_irn_n(nn, j, get_new_node(get_irn_n(n, i)));
426 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
430 /* If the pre walker reached this Phi after the post walker visited the
431 block block_visited is > 0. */
432 set_Block_block_visited(get_nodes_block(n), 0);
433 /* Compacting the Phi's ins might generate Phis with only one
435 if (get_irn_arity(nn) == 1)
436 exchange(nn, get_irn_n(nn, 0));
438 irn_arity = get_irn_arity(n);
439 for (i = -1; i < irn_arity; i++)
440 set_irn_n (nn, i, get_new_node(get_irn_n(n, i)));
442 /* Now the new node is complete. We can add it to the hash table for CSE.
443 @@@ inlining aborts if we identify End. Why? */
445 add_identities(current_ir_graph->value_table, nn);
449 * Copies the graph recursively, compacts the keep-alives of the end node.
451 * @param irg the graph to be copied
452 * @param copy_node_nr If non-zero, the node number will be copied
454 static void copy_graph(ir_graph *irg, int copy_node_nr) {
455 ir_node *oe, *ne, *ob, *nb, *om, *nm; /* old end, new end, old bad, new bad, old NoMem, new NoMem */
456 ir_node *ka; /* keep alive */
460 /* Some nodes must be copied by hand, sigh */
461 vfl = get_irg_visited(irg);
462 set_irg_visited(irg, vfl + 1);
464 oe = get_irg_end(irg);
465 mark_irn_visited(oe);
466 /* copy the end node by hand, allocate dynamic in array! */
467 ne = new_ir_node(get_irn_dbg_info(oe),
474 /* Copy the attributes. Well, there might be some in the future... */
475 copy_node_attr(oe, ne);
476 set_new_node(oe, ne);
478 /* copy the Bad node */
479 ob = get_irg_bad(irg);
480 mark_irn_visited(ob);
481 nb = new_ir_node(get_irn_dbg_info(ob),
488 copy_node_attr(ob, nb);
489 set_new_node(ob, nb);
491 /* copy the NoMem node */
492 om = get_irg_no_mem(irg);
493 mark_irn_visited(om);
494 nm = new_ir_node(get_irn_dbg_info(om),
501 copy_node_attr(om, nm);
502 set_new_node(om, nm);
504 /* copy the live nodes */
505 set_irg_visited(irg, vfl);
506 irg_walk(get_nodes_block(oe), copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
508 /* Note: from yet, the visited flag of the graph is equal to vfl + 1 */
510 /* visit the anchors as well */
511 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
512 ir_node *n = get_irg_anchor(irg, i);
514 if (n && (get_irn_visited(n) <= vfl)) {
515 set_irg_visited(irg, vfl);
516 irg_walk(n, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
520 /* copy_preds for the end node ... */
521 set_nodes_block(ne, get_new_node(get_nodes_block(oe)));
523 /*- ... and now the keep alives. -*/
524 /* First pick the not marked block nodes and walk them. We must pick these
525 first as else we will oversee blocks reachable from Phis. */
526 irn_arity = get_End_n_keepalives(oe);
527 for (i = 0; i < irn_arity; i++) {
528 ka = get_End_keepalive(oe, i);
530 if (get_irn_visited(ka) <= vfl) {
531 /* We must keep the block alive and copy everything reachable */
532 set_irg_visited(irg, vfl);
533 irg_walk(ka, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
535 add_End_keepalive(ne, get_new_node(ka));
539 /* Now pick other nodes. Here we will keep all! */
540 irn_arity = get_End_n_keepalives(oe);
541 for (i = 0; i < irn_arity; i++) {
542 ka = get_End_keepalive(oe, i);
544 if (get_irn_visited(ka) <= vfl) {
545 /* We didn't copy the node yet. */
546 set_irg_visited(irg, vfl);
547 irg_walk(ka, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
549 add_End_keepalive(ne, get_new_node(ka));
553 /* start block sometimes only reached after keep alives */
554 set_nodes_block(nb, get_new_node(get_nodes_block(ob)));
555 set_nodes_block(nm, get_new_node(get_nodes_block(om)));
559 * Copies the graph reachable from current_ir_graph->end to the obstack
560 * in current_ir_graph and fixes the environment.
561 * Then fixes the fields in current_ir_graph containing nodes of the
564 * @param copy_node_nr If non-zero, the node number will be copied
567 copy_graph_env(int copy_node_nr) {
568 ir_graph *irg = current_ir_graph;
569 ir_node *old_end, *new_anchor;
572 /* remove end_except and end_reg nodes */
573 old_end = get_irg_end(irg);
574 set_irg_end_except (irg, old_end);
575 set_irg_end_reg (irg, old_end);
577 /* Not all nodes remembered in irg might be reachable
578 from the end node. Assure their link is set to NULL, so that
579 we can test whether new nodes have been computed. */
580 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
581 ir_node *n = get_irg_anchor(irg, i);
583 set_new_node(n, NULL);
585 /* we use the block walk flag for removing Bads from Blocks ins. */
586 inc_irg_block_visited(irg);
589 copy_graph(irg, copy_node_nr);
592 old_end = get_irg_end(irg);
593 new_anchor = new_Anchor(irg);
595 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
596 ir_node *n = get_irg_anchor(irg, i);
598 set_irn_n(new_anchor, i, get_new_node(n));
601 irg->anchor = new_anchor;
603 /* ensure the new anchor is placed in the endblock */
604 set_irn_n(new_anchor, -1, get_irg_end_block(irg));
608 * Copies all reachable nodes to a new obstack. Removes bad inputs
609 * from block nodes and the corresponding inputs from Phi nodes.
610 * Merges single exit blocks with single entry blocks and removes
612 * Adds all new nodes to a new hash table for CSE. Does not
613 * perform CSE, so the hash table might contain common subexpressions.
616 dead_node_elimination(ir_graph *irg) {
617 if (get_opt_optimize() && get_opt_dead_node_elimination()) {
619 #ifdef INTERPROCEDURAL_VIEW
620 int rem_ipview = get_interprocedural_view();
622 struct obstack *graveyard_obst = NULL;
623 struct obstack *rebirth_obst = NULL;
624 assert(! edges_activated(irg) && "dead node elimination requires disabled edges");
626 /* inform statistics that we started a dead-node elimination run */
627 hook_dead_node_elim(irg, 1);
629 /* Remember external state of current_ir_graph. */
630 rem = current_ir_graph;
631 current_ir_graph = irg;
632 #ifdef INTERPROCEDURAL_VIEW
633 set_interprocedural_view(0);
636 assert(get_irg_phase_state(irg) != phase_building);
638 /* Handle graph state */
639 free_callee_info(irg);
643 /* @@@ so far we loose loops when copying */
644 free_loop_information(irg);
646 set_irg_doms_inconsistent(irg);
648 /* A quiet place, where the old obstack can rest in peace,
649 until it will be cremated. */
650 graveyard_obst = irg->obst;
652 /* A new obstack, where the reachable nodes will be copied to. */
653 rebirth_obst = xmalloc(sizeof(*rebirth_obst));
654 irg->obst = rebirth_obst;
655 obstack_init(irg->obst);
656 irg->last_node_idx = 0;
658 /* We also need a new value table for CSE */
659 del_identities(irg->value_table);
660 irg->value_table = new_identities();
662 /* Copy the graph from the old to the new obstack */
663 copy_graph_env(/*copy_node_nr=*/1);
665 /* Free memory from old unoptimized obstack */
666 obstack_free(graveyard_obst, 0); /* First empty the obstack ... */
667 xfree(graveyard_obst); /* ... then free it. */
669 /* inform statistics that the run is over */
670 hook_dead_node_elim(irg, 0);
672 current_ir_graph = rem;
673 #ifdef INTERPROCEDURAL_VIEW
674 set_interprocedural_view(rem_ipview);
680 * Relink bad predecessors of a block and store the old in array to the
681 * link field. This function is called by relink_bad_predecessors().
682 * The array of link field starts with the block operand at position 0.
683 * If block has bad predecessors, create a new in array without bad preds.
684 * Otherwise let in array untouched.
686 static void relink_bad_block_predecessors(ir_node *n, void *env) {
687 ir_node **new_in, *irn;
688 int i, new_irn_n, old_irn_arity, new_irn_arity = 0;
691 /* if link field of block is NULL, look for bad predecessors otherwise
692 this is already done */
693 if (is_Block(n) && get_irn_link(n) == NULL) {
694 /* save old predecessors in link field (position 0 is the block operand)*/
695 set_irn_link(n, get_irn_in(n));
697 /* count predecessors without bad nodes */
698 old_irn_arity = get_irn_arity(n);
699 for (i = 0; i < old_irn_arity; i++)
700 if (!is_Bad(get_irn_n(n, i))) new_irn_arity++;
702 /* arity changing: set new predecessors without bad nodes */
703 if (new_irn_arity < old_irn_arity) {
704 /* Get new predecessor array. We do not resize the array, as we must
705 keep the old one to update Phis. */
706 new_in = NEW_ARR_D (ir_node *, current_ir_graph->obst, (new_irn_arity+1));
708 /* set new predecessors in array */
711 for (i = 0; i < old_irn_arity; i++) {
712 irn = get_irn_n(n, i);
714 new_in[new_irn_n] = irn;
715 is_backedge(n, i) ? set_backedge(n, new_irn_n-1) : set_not_backedge(n, new_irn_n-1);
719 /* ARR_SETLEN(int, n->attr.block.backedge, new_irn_arity); */
720 ARR_SHRINKLEN(n->attr.block.backedge, new_irn_arity);
722 } /* ir node has bad predecessors */
723 } /* Block is not relinked */
727 * Relinks Bad predecessors from Blocks and Phis called by walker
728 * remove_bad_predecesors(). If n is a Block, call
729 * relink_bad_block_redecessors(). If n is a Phi-node, call also the relinking
730 * function of Phi's Block. If this block has bad predecessors, relink preds
733 static void relink_bad_predecessors(ir_node *n, void *env) {
734 ir_node *block, **old_in;
735 int i, old_irn_arity, new_irn_arity;
737 /* relink bad predecessors of a block */
739 relink_bad_block_predecessors(n, env);
741 /* If Phi node relink its block and its predecessors */
743 /* Relink predecessors of phi's block */
744 block = get_nodes_block(n);
745 if (get_irn_link(block) == NULL)
746 relink_bad_block_predecessors(block, env);
748 old_in = (ir_node **)get_irn_link(block); /* Of Phi's Block */
749 old_irn_arity = ARR_LEN(old_in);
751 /* Relink Phi predecessors if count of predecessors changed */
752 if (old_irn_arity != ARR_LEN(get_irn_in(block))) {
753 /* set new predecessors in array
754 n->in[0] remains the same block */
756 for(i = 1; i < old_irn_arity; i++)
757 if (!is_Bad((ir_node *)old_in[i])) {
758 n->in[new_irn_arity] = n->in[i];
759 is_backedge(n, i) ? set_backedge(n, new_irn_arity) : set_not_backedge(n, new_irn_arity);
763 ARR_SETLEN(ir_node *, n->in, new_irn_arity);
764 ARR_SETLEN(int, n->attr.phi_backedge, new_irn_arity);
766 } /* n is a Phi node */
770 * Removes Bad Bad predecessors from Blocks and the corresponding
771 * inputs to Phi nodes as in dead_node_elimination but without
773 * On walking up set the link field to NULL, on walking down call
774 * relink_bad_predecessors() (This function stores the old in array
775 * to the link field and sets a new in array if arity of predecessors
778 void remove_bad_predecessors(ir_graph *irg) {
779 irg_walk_graph(irg, firm_clear_link, relink_bad_predecessors, NULL);
786 __)|_| | \_/ | \_/(/_ |_/\__|__
788 The following stuff implements a facility that automatically patches
789 registered ir_node pointers to the new node when a dead node elimination occurs.
792 struct _survive_dce_t {
796 hook_entry_t dead_node_elim;
797 hook_entry_t dead_node_elim_subst;
800 typedef struct _survive_dce_list_t {
801 struct _survive_dce_list_t *next;
803 } survive_dce_list_t;
805 static void dead_node_hook(void *context, ir_graph *irg, int start) {
806 survive_dce_t *sd = context;
809 /* Create a new map before the dead node elimination is performed. */
811 sd->new_places = pmap_create_ex(pmap_count(sd->places));
813 /* Patch back all nodes if dead node elimination is over and something is to be done. */
814 pmap_destroy(sd->places);
815 sd->places = sd->new_places;
816 sd->new_places = NULL;
821 * Hook called when dead node elimination replaces old by nw.
823 static void dead_node_subst_hook(void *context, ir_graph *irg, ir_node *old, ir_node *nw) {
824 survive_dce_t *sd = context;
825 survive_dce_list_t *list = pmap_get(sd->places, old);
828 /* If the node is to be patched back, write the new address to all registered locations. */
830 survive_dce_list_t *p;
832 for (p = list; p; p = p->next)
835 pmap_insert(sd->new_places, nw, list);
840 * Make a new Survive DCE environment.
842 survive_dce_t *new_survive_dce(void) {
843 survive_dce_t *res = xmalloc(sizeof(res[0]));
844 obstack_init(&res->obst);
845 res->places = pmap_create();
846 res->new_places = NULL;
848 res->dead_node_elim.hook._hook_dead_node_elim = dead_node_hook;
849 res->dead_node_elim.context = res;
850 res->dead_node_elim.next = NULL;
852 res->dead_node_elim_subst.hook._hook_dead_node_elim_subst = dead_node_subst_hook;
853 res->dead_node_elim_subst.context = res;
854 res->dead_node_elim_subst.next = NULL;
856 #ifndef FIRM_ENABLE_HOOKS
857 assert(0 && "need hooks enabled");
860 register_hook(hook_dead_node_elim, &res->dead_node_elim);
861 register_hook(hook_dead_node_elim_subst, &res->dead_node_elim_subst);
866 * Free a Survive DCE environment.
868 void free_survive_dce(survive_dce_t *sd) {
869 obstack_free(&sd->obst, NULL);
870 pmap_destroy(sd->places);
871 unregister_hook(hook_dead_node_elim, &sd->dead_node_elim);
872 unregister_hook(hook_dead_node_elim_subst, &sd->dead_node_elim_subst);
877 * Register a node pointer to be patched upon DCE.
878 * When DCE occurs, the node pointer specified by @p place will be
879 * patched to the new address of the node it is pointing to.
881 * @param sd The Survive DCE environment.
882 * @param place The address of the node pointer.
884 void survive_dce_register_irn(survive_dce_t *sd, ir_node **place) {
885 if (*place != NULL) {
886 ir_node *irn = *place;
887 survive_dce_list_t *curr = pmap_get(sd->places, irn);
888 survive_dce_list_t *nw = obstack_alloc(&sd->obst, sizeof(nw[0]));
893 pmap_insert(sd->places, irn, nw);
897 /*--------------------------------------------------------------------*/
898 /* Functionality for inlining */
899 /*--------------------------------------------------------------------*/
902 * Copy node for inlineing. Updates attributes that change when
903 * inlineing but not for dead node elimination.
905 * Copies the node by calling copy_node() and then updates the entity if
906 * it's a local one. env must be a pointer of the frame type of the
907 * inlined procedure. The new entities must be in the link field of
911 copy_node_inline(ir_node *n, void *env) {
913 ir_type *frame_tp = (ir_type *)env;
917 nn = get_new_node (n);
919 if (get_entity_owner(get_Sel_entity(n)) == frame_tp) {
920 set_Sel_entity(nn, get_entity_link(get_Sel_entity(n)));
922 } else if (is_Block(n)) {
923 nn = get_new_node (n);
924 nn->attr.block.irg = current_ir_graph;
929 * Walker: checks if P_value_arg_base is used.
931 static void find_addr(ir_node *node, void *env) {
932 int *allow_inline = env;
934 is_Start(get_Proj_pred(node)) &&
935 get_Proj_proj(node) == pn_Start_P_value_arg_base) {
941 * Check if we can inline a given call.
942 * Currently, we cannot inline two cases:
943 * - call with compound arguments
944 * - graphs that take the address of a parameter
946 * check these conditions here
948 static int can_inline(ir_node *call, ir_graph *called_graph) {
949 ir_type *call_type = get_Call_type(call);
950 int params, ress, i, res;
951 assert(is_Method_type(call_type));
953 params = get_method_n_params(call_type);
954 ress = get_method_n_ress(call_type);
956 /* check parameters for compound arguments */
957 for (i = 0; i < params; ++i) {
958 ir_type *p_type = get_method_param_type(call_type, i);
960 if (is_compound_type(p_type))
964 /* check results for compound arguments */
965 for (i = 0; i < ress; ++i) {
966 ir_type *r_type = get_method_res_type(call_type, i);
968 if (is_compound_type(r_type))
973 irg_walk_graph(called_graph, find_addr, NULL, &res);
979 exc_handler = 0, /**< There is a handler. */
980 exc_to_end = 1, /**< Branches to End. */
981 exc_no_handler = 2 /**< Exception handling not represented. */
984 /* Inlines a method at the given call site. */
985 int inline_method(ir_node *call, ir_graph *called_graph) {
987 ir_node *post_call, *post_bl;
988 ir_node *in[pn_Start_max];
989 ir_node *end, *end_bl;
993 int arity, n_ret, n_exc, n_res, i, j, rem_opt, irn_arity;
994 enum exc_mode exc_handling;
995 ir_type *called_frame;
996 irg_inline_property prop = get_irg_inline_property(called_graph);
998 if ( (prop < irg_inline_forced) &&
999 (!get_opt_optimize() || !get_opt_inline() || (prop == irg_inline_forbidden))) return 0;
1001 /* Do not inline variadic functions. */
1002 if (get_method_variadicity(get_entity_type(get_irg_entity(called_graph))) == variadicity_variadic)
1005 assert(get_method_n_params(get_entity_type(get_irg_entity(called_graph))) ==
1006 get_method_n_params(get_Call_type(call)));
1009 * currently, we cannot inline two cases:
1010 * - call with compound arguments
1011 * - graphs that take the address of a parameter
1013 if (! can_inline(call, called_graph))
1016 /* -- Turn off optimizations, this can cause problems when allocating new nodes. -- */
1017 rem_opt = get_opt_optimize();
1020 /* Handle graph state */
1021 assert(get_irg_phase_state(current_ir_graph) != phase_building);
1022 assert(get_irg_pinned(current_ir_graph) == op_pin_state_pinned);
1023 assert(get_irg_pinned(called_graph) == op_pin_state_pinned);
1024 set_irg_outs_inconsistent(current_ir_graph);
1025 set_irg_extblk_inconsistent(current_ir_graph);
1026 set_irg_doms_inconsistent(current_ir_graph);
1027 set_irg_loopinfo_inconsistent(current_ir_graph);
1028 set_irg_callee_info_state(current_ir_graph, irg_callee_info_inconsistent);
1030 /* -- Check preconditions -- */
1031 assert(is_Call(call));
1032 /* @@@ does not work for InterfaceIII.java after cgana
1033 assert(get_Call_type(call) == get_entity_type(get_irg_entity(called_graph)));
1034 assert(smaller_type(get_entity_type(get_irg_entity(called_graph)),
1035 get_Call_type(call)));
1037 if (called_graph == current_ir_graph) {
1038 set_optimize(rem_opt);
1042 /* here we know we WILL inline, so inform the statistics */
1043 hook_inline(call, called_graph);
1045 /* -- Decide how to handle exception control flow: Is there a handler
1046 for the Call node, or do we branch directly to End on an exception?
1048 0 There is a handler.
1050 2 Exception handling not represented in Firm. -- */
1052 ir_node *proj, *Mproj = NULL, *Xproj = NULL;
1053 for (proj = get_irn_link(call); proj; proj = get_irn_link(proj)) {
1054 long proj_nr = get_Proj_proj(proj);
1055 if (proj_nr == pn_Call_X_except) Xproj = proj;
1056 if (proj_nr == pn_Call_M_except) Mproj = proj;
1058 if (Mproj) { assert(Xproj); exc_handling = exc_handler; } /* Mproj */
1059 else if (Xproj) { exc_handling = exc_to_end; } /* !Mproj && Xproj */
1060 else { exc_handling = exc_no_handler; } /* !Mproj && !Xproj */
1064 the procedure and later replaces the Start node of the called graph.
1065 Post_call is the old Call node and collects the results of the called
1066 graph. Both will end up being a tuple. -- */
1067 post_bl = get_nodes_block(call);
1068 set_irg_current_block(current_ir_graph, post_bl);
1069 /* XxMxPxPxPxT of Start + parameter of Call */
1070 in[pn_Start_X_initial_exec] = new_Jmp();
1071 in[pn_Start_M] = get_Call_mem(call);
1072 in[pn_Start_P_frame_base] = get_irg_frame(current_ir_graph);
1073 in[pn_Start_P_globals] = get_irg_globals(current_ir_graph);
1074 in[pn_Start_P_tls] = get_irg_tls(current_ir_graph);
1075 in[pn_Start_T_args] = new_Tuple(get_Call_n_params(call), get_Call_param_arr(call));
1076 /* in[pn_Start_P_value_arg_base] = ??? */
1077 assert(pn_Start_P_value_arg_base == pn_Start_max - 1 && "pn_Start_P_value_arg_base not supported, fix");
1078 pre_call = new_Tuple(pn_Start_max - 1, in);
1082 The new block gets the ins of the old block, pre_call and all its
1083 predecessors and all Phi nodes. -- */
1084 part_block(pre_call);
1086 /* -- Prepare state for dead node elimination -- */
1087 /* Visited flags in calling irg must be >= flag in called irg.
1088 Else walker and arity computation will not work. */
1089 if (get_irg_visited(current_ir_graph) <= get_irg_visited(called_graph))
1090 set_irg_visited(current_ir_graph, get_irg_visited(called_graph)+1);
1091 if (get_irg_block_visited(current_ir_graph)< get_irg_block_visited(called_graph))
1092 set_irg_block_visited(current_ir_graph, get_irg_block_visited(called_graph));
1093 /* Set pre_call as new Start node in link field of the start node of
1094 calling graph and pre_calls block as new block for the start block
1096 Further mark these nodes so that they are not visited by the
1098 set_irn_link(get_irg_start(called_graph), pre_call);
1099 set_irn_visited(get_irg_start(called_graph), get_irg_visited(current_ir_graph));
1100 set_irn_link(get_irg_start_block(called_graph), get_nodes_block(pre_call));
1101 set_irn_visited(get_irg_start_block(called_graph), get_irg_visited(current_ir_graph));
1102 set_irn_link(get_irg_bad(called_graph), get_irg_bad(current_ir_graph));
1103 set_irn_visited(get_irg_bad(called_graph), get_irg_visited(current_ir_graph));
1105 /* Initialize for compaction of in arrays */
1106 inc_irg_block_visited(current_ir_graph);
1108 /* -- Replicate local entities of the called_graph -- */
1109 /* copy the entities. */
1110 called_frame = get_irg_frame_type(called_graph);
1111 for (i = 0; i < get_class_n_members(called_frame); i++) {
1112 ir_entity *new_ent, *old_ent;
1113 old_ent = get_class_member(called_frame, i);
1114 new_ent = copy_entity_own(old_ent, get_cur_frame_type());
1115 set_entity_link(old_ent, new_ent);
1118 /* visited is > than that of called graph. With this trick visited will
1119 remain unchanged so that an outer walker, e.g., searching the call nodes
1120 to inline, calling this inline will not visit the inlined nodes. */
1121 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
1123 /* -- Performing dead node elimination inlines the graph -- */
1124 /* Copies the nodes to the obstack of current_ir_graph. Updates links to new
1126 irg_walk(get_irg_end(called_graph), copy_node_inline, copy_preds,
1127 get_irg_frame_type(called_graph));
1129 /* Repair called_graph */
1130 set_irg_visited(called_graph, get_irg_visited(current_ir_graph));
1131 set_irg_block_visited(called_graph, get_irg_block_visited(current_ir_graph));
1132 set_Block_block_visited(get_irg_start_block(called_graph), 0);
1134 /* -- Merge the end of the inlined procedure with the call site -- */
1135 /* We will turn the old Call node into a Tuple with the following
1138 0: Phi of all Memories of Return statements.
1139 1: Jmp from new Block that merges the control flow from all exception
1140 predecessors of the old end block.
1141 2: Tuple of all arguments.
1142 3: Phi of Exception memories.
1143 In case the old Call directly branches to End on an exception we don't
1144 need the block merging all exceptions nor the Phi of the exception
1148 /* -- Precompute some values -- */
1149 end_bl = get_new_node(get_irg_end_block(called_graph));
1150 end = get_new_node(get_irg_end(called_graph));
1151 arity = get_irn_arity(end_bl); /* arity = n_exc + n_ret */
1152 n_res = get_method_n_ress(get_Call_type(call));
1154 res_pred = xmalloc(n_res * sizeof(*res_pred));
1155 cf_pred = xmalloc(arity * sizeof(*res_pred));
1157 set_irg_current_block(current_ir_graph, post_bl); /* just to make sure */
1159 /* -- archive keepalives -- */
1160 irn_arity = get_irn_arity(end);
1161 for (i = 0; i < irn_arity; i++) {
1162 ir_node *ka = get_End_keepalive(end, i);
1164 add_End_keepalive(get_irg_end(current_ir_graph), ka);
1167 /* The new end node will die. We need not free as the in array is on the obstack:
1168 copy_node() only generated 'D' arrays. */
1170 /* -- Replace Return nodes by Jump nodes. -- */
1172 for (i = 0; i < arity; i++) {
1174 ret = get_irn_n(end_bl, i);
1175 if (is_Return(ret)) {
1176 cf_pred[n_ret] = new_r_Jmp(current_ir_graph, get_nodes_block(ret));
1180 set_irn_in(post_bl, n_ret, cf_pred);
1182 /* -- Build a Tuple for all results of the method.
1183 Add Phi node if there was more than one Return. -- */
1184 turn_into_tuple(post_call, pn_Call_max);
1185 /* First the Memory-Phi */
1187 for (i = 0; i < arity; i++) {
1188 ret = get_irn_n(end_bl, i);
1189 if (is_Return(ret)) {
1190 cf_pred[n_ret] = get_Return_mem(ret);
1194 phi = new_Phi(n_ret, cf_pred, mode_M);
1195 set_Tuple_pred(call, pn_Call_M_regular, phi);
1196 /* Conserve Phi-list for further inlinings -- but might be optimized */
1197 if (get_nodes_block(phi) == post_bl) {
1198 set_irn_link(phi, get_irn_link(post_bl));
1199 set_irn_link(post_bl, phi);
1201 /* Now the real results */
1203 for (j = 0; j < n_res; j++) {
1205 for (i = 0; i < arity; i++) {
1206 ret = get_irn_n(end_bl, i);
1207 if (is_Return(ret)) {
1208 cf_pred[n_ret] = get_Return_res(ret, j);
1213 phi = new_Phi(n_ret, cf_pred, get_irn_mode(cf_pred[0]));
1217 /* Conserve Phi-list for further inlinings -- but might be optimized */
1218 if (get_nodes_block(phi) == post_bl) {
1219 set_irn_link(phi, get_irn_link(post_bl));
1220 set_irn_link(post_bl, phi);
1223 set_Tuple_pred(call, pn_Call_T_result, new_Tuple(n_res, res_pred));
1225 set_Tuple_pred(call, pn_Call_T_result, new_Bad());
1227 /* handle the regular call */
1228 set_Tuple_pred(call, pn_Call_X_regular, new_Jmp());
1230 /* For now, we cannot inline calls with value_base */
1231 set_Tuple_pred(call, pn_Call_P_value_res_base, new_Bad());
1233 /* Finally the exception control flow.
1234 We have two (three) possible situations:
1235 First if the Call branches to an exception handler: We need to add a Phi node to
1236 collect the memory containing the exception objects. Further we need
1237 to add another block to get a correct representation of this Phi. To
1238 this block we add a Jmp that resolves into the X output of the Call
1239 when the Call is turned into a tuple.
1240 Second the Call branches to End, the exception is not handled. Just
1241 add all inlined exception branches to the End node.
1242 Third: there is no Exception edge at all. Handle as case two. */
1243 if (exc_handling == exc_handler) {
1245 for (i = 0; i < arity; i++) {
1247 ret = get_irn_n(end_bl, i);
1248 irn = skip_Proj(ret);
1249 if (is_fragile_op(irn) || is_Raise(irn)) {
1250 cf_pred[n_exc] = ret;
1255 new_Block(n_exc, cf_pred); /* watch it: current_block is changed! */
1256 set_Tuple_pred(call, pn_Call_X_except, new_Jmp());
1257 /* The Phi for the memories with the exception objects */
1259 for (i = 0; i < arity; i++) {
1261 ret = skip_Proj(get_irn_n(end_bl, i));
1263 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 3);
1265 } else if (is_fragile_op(ret)) {
1266 /* We rely that all cfops have the memory output at the same position. */
1267 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 0);
1269 } else if (is_Raise(ret)) {
1270 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 1);
1274 set_Tuple_pred(call, pn_Call_M_except, new_Phi(n_exc, cf_pred, mode_M));
1276 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1277 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1280 ir_node *main_end_bl;
1281 int main_end_bl_arity;
1282 ir_node **end_preds;
1284 /* assert(exc_handling == 1 || no exceptions. ) */
1286 for (i = 0; i < arity; i++) {
1287 ir_node *ret = get_irn_n(end_bl, i);
1288 ir_node *irn = skip_Proj(ret);
1290 if (is_fragile_op(irn) || is_Raise(irn)) {
1291 cf_pred[n_exc] = ret;
1295 main_end_bl = get_irg_end_block(current_ir_graph);
1296 main_end_bl_arity = get_irn_arity(main_end_bl);
1297 end_preds = xmalloc((n_exc + main_end_bl_arity) * sizeof(*end_preds));
1299 for (i = 0; i < main_end_bl_arity; ++i)
1300 end_preds[i] = get_irn_n(main_end_bl, i);
1301 for (i = 0; i < n_exc; ++i)
1302 end_preds[main_end_bl_arity + i] = cf_pred[i];
1303 set_irn_in(main_end_bl, n_exc + main_end_bl_arity, end_preds);
1304 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1305 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1311 /* -- Turn CSE back on. -- */
1312 set_optimize(rem_opt);
1317 /********************************************************************/
1318 /* Apply inlineing to small methods. */
1319 /********************************************************************/
1321 /** Represents a possible inlinable call in a graph. */
1322 typedef struct _call_entry call_entry;
1323 struct _call_entry {
1324 ir_node *call; /**< the Call */
1325 ir_graph *callee; /**< the callee called here */
1326 call_entry *next; /**< for linking the next one */
1330 * environment for inlining small irgs
1332 typedef struct _inline_env_t {
1333 struct obstack obst; /**< an obstack where call_entries are allocated on. */
1334 call_entry *head; /**< the head of the call entry list */
1335 call_entry *tail; /**< the tail of the call entry list */
1339 * Returns the irg called from a Call node. If the irg is not
1340 * known, NULL is returned.
1342 static ir_graph *get_call_called_irg(ir_node *call) {
1344 ir_graph *called_irg = NULL;
1346 addr = get_Call_ptr(call);
1347 if (is_SymConst(addr) && get_SymConst_kind(addr) == symconst_addr_ent) {
1348 called_irg = get_entity_irg(get_SymConst_entity(addr));
1355 * Walker: Collect all calls to known graphs inside a graph.
1357 static void collect_calls(ir_node *call, void *env) {
1358 if (is_Call(call)) {
1359 ir_graph *called_irg = get_call_called_irg(call);
1361 /* The Call node calls a locally defined method. Remember to inline. */
1362 inline_env_t *ienv = env;
1363 call_entry *entry = obstack_alloc(&ienv->obst, sizeof(*entry));
1365 entry->callee = called_irg;
1368 if (ienv->tail == NULL)
1371 ienv->tail->next = entry;
1378 * Inlines all small methods at call sites where the called address comes
1379 * from a Const node that references the entity representing the called
1381 * The size argument is a rough measure for the code size of the method:
1382 * Methods where the obstack containing the firm graph is smaller than
1385 void inline_small_irgs(ir_graph *irg, int size) {
1386 ir_graph *rem = current_ir_graph;
1389 DEBUG_ONLY(firm_dbg_module_t *dbg;)
1391 if (!(get_opt_optimize() && get_opt_inline())) return;
1393 FIRM_DBG_REGISTER(dbg, "firm.opt.inline");
1395 current_ir_graph = irg;
1396 /* Handle graph state */
1397 assert(get_irg_phase_state(irg) != phase_building);
1398 free_callee_info(irg);
1400 /* Find Call nodes to inline.
1401 (We can not inline during a walk of the graph, as inlineing the same
1402 method several times changes the visited flag of the walked graph:
1403 after the first inlineing visited of the callee equals visited of
1404 the caller. With the next inlineing both are increased.) */
1405 obstack_init(&env.obst);
1406 env.head = env.tail = NULL;
1407 irg_walk_graph(irg, NULL, collect_calls, &env);
1409 if (env.head != NULL) {
1410 /* There are calls to inline */
1411 collect_phiprojs(irg);
1412 for (entry = env.head; entry != NULL; entry = entry->next) {
1413 ir_graph *callee = entry->callee;
1414 if (((_obstack_memory_used(callee->obst) - (int)obstack_room(callee->obst)) < size) ||
1415 (get_irg_inline_property(callee) >= irg_inline_forced)) {
1416 inline_method(entry->call, callee);
1420 obstack_free(&env.obst, NULL);
1421 current_ir_graph = rem;
1425 * Environment for inlining irgs.
1428 int n_nodes; /**< Number of nodes in graph except Id, Tuple, Proj, Start, End. */
1429 int n_nodes_orig; /**< for statistics */
1430 call_entry *call_head; /**< The head of the list of all call nodes in this graph. */
1431 call_entry *call_tail; /**< The tail of the list of all call nodes in this graph .*/
1432 int n_call_nodes; /**< Number of Call nodes in the graph. */
1433 int n_call_nodes_orig; /**< for statistics */
1434 int n_callers; /**< Number of known graphs that call this graphs. */
1435 int n_callers_orig; /**< for statistics */
1436 int got_inline; /**< Set, if at leat one call inside this graph was inlined. */
1440 * Allocate a new environment for inlining.
1442 static inline_irg_env *alloc_inline_irg_env(struct obstack *obst) {
1443 inline_irg_env *env = obstack_alloc(obst, sizeof(*env));
1444 env->n_nodes = -2; /* do not count count Start, End */
1445 env->n_nodes_orig = -2; /* do not count Start, End */
1446 env->call_head = NULL;
1447 env->call_tail = NULL;
1448 env->n_call_nodes = 0;
1449 env->n_call_nodes_orig = 0;
1451 env->n_callers_orig = 0;
1452 env->got_inline = 0;
1456 typedef struct walker_env {
1457 struct obstack *obst; /**< the obstack for allocations. */
1458 inline_irg_env *x; /**< the inline environment */
1459 int ignore_runtime; /**< the ignore runtime flag */
1463 * post-walker: collect all calls in the inline-environment
1464 * of a graph and sum some statistics.
1466 static void collect_calls2(ir_node *call, void *ctx) {
1468 inline_irg_env *x = env->x;
1469 ir_op *op = get_irn_op(call);
1473 /* count meaningful nodes in irg */
1474 if (op != op_Proj && op != op_Tuple && op != op_Sync) {
1479 if (op != op_Call) return;
1481 /* check, if it's a runtime call */
1482 if (env->ignore_runtime) {
1483 ir_node *symc = get_Call_ptr(call);
1485 if (is_SymConst(symc) && get_SymConst_kind(symc) == symconst_addr_ent) {
1486 ir_entity *ent = get_SymConst_entity(symc);
1488 if (get_entity_additional_properties(ent) & mtp_property_runtime)
1493 /* collect all call nodes */
1495 ++x->n_call_nodes_orig;
1497 callee = get_call_called_irg(call);
1499 inline_irg_env *callee_env = get_irg_link(callee);
1500 /* count all static callers */
1501 ++callee_env->n_callers;
1502 ++callee_env->n_callers_orig;
1504 /* link it in the list of possible inlinable entries */
1505 entry = obstack_alloc(env->obst, sizeof(*entry));
1507 entry->callee = callee;
1509 if (x->call_tail == NULL)
1510 x->call_head = entry;
1512 x->call_tail->next = entry;
1513 x->call_tail = entry;
1518 * Returns TRUE if the number of callers is 0 in the irg's environment,
1519 * hence this irg is a leave.
1521 INLINE static int is_leave(ir_graph *irg) {
1522 inline_irg_env *env = get_irg_link(irg);
1523 return env->n_call_nodes == 0;
1527 * Returns TRUE if the number of nodes in the callee is
1528 * smaller then size in the irg's environment.
1530 INLINE static int is_smaller(ir_graph *callee, int size) {
1531 inline_irg_env *env = get_irg_link(callee);
1532 return env->n_nodes < size;
1536 * Append the nodes of the list src to the nodes of the list in environment dst.
1538 static void append_call_list(struct obstack *obst, inline_irg_env *dst, call_entry *src) {
1539 call_entry *entry, *nentry;
1541 /* Note that the src list points to Call nodes in the inlined graph, but
1542 we need Call nodes in our graph. Luckily the inliner leaves this information
1543 in the link field. */
1544 for (entry = src; entry != NULL; entry = entry->next) {
1545 nentry = obstack_alloc(obst, sizeof(*nentry));
1546 nentry->call = get_irn_link(entry->call);
1547 nentry->callee = entry->callee;
1548 nentry->next = NULL;
1549 dst->call_tail->next = nentry;
1550 dst->call_tail = nentry;
1555 * Inlines small leave methods at call sites where the called address comes
1556 * from a Const node that references the entity representing the called
1558 * The size argument is a rough measure for the code size of the method:
1559 * Methods where the obstack containing the firm graph is smaller than
1562 void inline_leave_functions(int maxsize, int leavesize, int size, int ignore_runtime) {
1563 inline_irg_env *env;
1569 call_entry *entry, *tail;
1570 const call_entry *centry;
1571 struct obstack obst;
1572 DEBUG_ONLY(firm_dbg_module_t *dbg;)
1574 if (!(get_opt_optimize() && get_opt_inline())) return;
1576 FIRM_DBG_REGISTER(dbg, "firm.opt.inline");
1577 rem = current_ir_graph;
1578 obstack_init(&obst);
1580 /* extend all irgs by a temporary data structure for inlining. */
1581 n_irgs = get_irp_n_irgs();
1582 for (i = 0; i < n_irgs; ++i)
1583 set_irg_link(get_irp_irg(i), alloc_inline_irg_env(&obst));
1585 /* Precompute information in temporary data structure. */
1587 wenv.ignore_runtime = ignore_runtime;
1588 for (i = 0; i < n_irgs; ++i) {
1589 ir_graph *irg = get_irp_irg(i);
1591 assert(get_irg_phase_state(irg) != phase_building);
1592 free_callee_info(irg);
1594 wenv.x = get_irg_link(irg);
1595 irg_walk_graph(irg, NULL, collect_calls2, &wenv);
1598 /* -- and now inline. -- */
1600 /* Inline leaves recursively -- we might construct new leaves. */
1604 for (i = 0; i < n_irgs; ++i) {
1606 int phiproj_computed = 0;
1608 current_ir_graph = get_irp_irg(i);
1609 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1612 for (entry = env->call_head; entry != NULL; entry = entry->next) {
1615 if (env->n_nodes > maxsize) break;
1618 callee = entry->callee;
1620 if (is_leave(callee) && is_smaller(callee, leavesize)) {
1621 if (!phiproj_computed) {
1622 phiproj_computed = 1;
1623 collect_phiprojs(current_ir_graph);
1625 did_inline = inline_method(call, callee);
1628 /* Do some statistics */
1629 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1631 env->got_inline = 1;
1632 --env->n_call_nodes;
1633 env->n_nodes += callee_env->n_nodes;
1634 --callee_env->n_callers;
1636 /* remove this call from the list */
1638 tail->next = entry->next;
1640 env->call_head = entry->next;
1646 env->call_tail = tail;
1648 } while (did_inline);
1650 /* inline other small functions. */
1651 for (i = 0; i < n_irgs; ++i) {
1653 int phiproj_computed = 0;
1655 current_ir_graph = get_irp_irg(i);
1656 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1658 /* note that the list of possible calls is updated during the process */
1660 for (entry = env->call_head; entry != NULL; entry = entry->next) {
1664 callee = entry->callee;
1666 if (((is_smaller(callee, size) && (env->n_nodes < maxsize)) || /* small function */
1667 (get_irg_inline_property(callee) >= irg_inline_forced))) {
1668 if (!phiproj_computed) {
1669 phiproj_computed = 1;
1670 collect_phiprojs(current_ir_graph);
1672 if (inline_method(call, callee)) {
1673 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1675 /* callee was inline. Append it's call list. */
1676 env->got_inline = 1;
1677 --env->n_call_nodes;
1678 append_call_list(&obst, env, callee_env->call_head);
1679 env->n_call_nodes += callee_env->n_call_nodes;
1680 env->n_nodes += callee_env->n_nodes;
1681 --callee_env->n_callers;
1683 /* after we have inlined callee, all called methods inside callee
1684 are now called once more */
1685 for (centry = callee_env->call_head; centry != NULL; centry = centry->next) {
1686 inline_irg_env *penv = get_irg_link(centry->callee);
1690 /* remove this call from the list */
1692 tail->next = entry->next;
1694 env->call_head = entry->next;
1700 env->call_tail = tail;
1703 for (i = 0; i < n_irgs; ++i) {
1704 irg = get_irp_irg(i);
1705 env = (inline_irg_env *)get_irg_link(irg);
1707 if (env->got_inline) {
1708 /* this irg got calls inlined */
1709 set_irg_outs_inconsistent(irg);
1710 set_irg_doms_inconsistent(irg);
1712 optimize_graph_df(irg);
1715 if (env->got_inline || (env->n_callers_orig != env->n_callers)) {
1716 DB((dbg, SET_LEVEL_1, "Nodes:%3d ->%3d, calls:%3d ->%3d, callers:%3d ->%3d, -- %s\n",
1717 env->n_nodes_orig, env->n_nodes, env->n_call_nodes_orig, env->n_call_nodes,
1718 env->n_callers_orig, env->n_callers,
1719 get_entity_name(get_irg_entity(irg))));
1723 obstack_free(&obst, NULL);
1724 current_ir_graph = rem;
1727 /*******************************************************************/
1728 /* Code Placement. Pins all floating nodes to a block where they */
1729 /* will be executed only if needed. */
1730 /*******************************************************************/
1733 * Returns non-zero, is a block is not reachable from Start.
1735 * @param block the block to test
1738 is_Block_unreachable(ir_node *block) {
1739 return is_Block_dead(block) || get_Block_dom_depth(block) < 0;
1743 * Find the earliest correct block for node n. --- Place n into the
1744 * same Block as its dominance-deepest Input.
1746 * We have to avoid calls to get_nodes_block() here
1747 * because the graph is floating.
1749 * move_out_of_loops() expects that place_floats_early() have placed
1750 * all "living" nodes into a living block. That's why we must
1751 * move nodes in dead block with "live" successors into a valid
1753 * We move them just into the same block as it's successor (or
1754 * in case of a Phi into the effective use block). For Phi successors,
1755 * this may still be a dead block, but then there is no real use, as
1756 * the control flow will be dead later.
1758 * @param n the node to be placed
1759 * @param worklist a worklist, predecessors of non-floating nodes are placed here
1762 place_floats_early(ir_node *n, waitq *worklist) {
1765 /* we must not run into an infinite loop */
1766 assert(irn_not_visited(n));
1767 mark_irn_visited(n);
1769 /* Place floating nodes. */
1770 if (get_irn_pinned(n) == op_pin_state_floats) {
1771 ir_node *curr_block = get_nodes_block(n);
1772 int in_dead_block = is_Block_unreachable(curr_block);
1774 ir_node *b = NULL; /* The block to place this node in */
1776 assert(is_no_Block(n));
1778 if (is_irn_start_block_placed(n)) {
1779 /* These nodes will not be placed by the loop below. */
1780 b = get_irg_start_block(current_ir_graph);
1784 /* find the block for this node. */
1785 irn_arity = get_irn_arity(n);
1786 for (i = 0; i < irn_arity; i++) {
1787 ir_node *pred = get_irn_n(n, i);
1788 ir_node *pred_block;
1790 if ((irn_not_visited(pred))
1791 && (get_irn_pinned(pred) == op_pin_state_floats)) {
1794 * If the current node is NOT in a dead block, but one of its
1795 * predecessors is, we must move the predecessor to a live block.
1796 * Such thing can happen, if global CSE chose a node from a dead block.
1797 * We move it simply to our block.
1798 * Note that neither Phi nor End nodes are floating, so we don't
1799 * need to handle them here.
1801 if (! in_dead_block) {
1802 if (get_irn_pinned(pred) == op_pin_state_floats &&
1803 is_Block_unreachable(get_nodes_block(pred)))
1804 set_nodes_block(pred, curr_block);
1806 place_floats_early(pred, worklist);
1810 * A node in the Bad block must stay in the bad block,
1811 * so don't compute a new block for it.
1816 /* Because all loops contain at least one op_pin_state_pinned node, now all
1817 our inputs are either op_pin_state_pinned or place_early() has already
1818 been finished on them. We do not have any unfinished inputs! */
1819 pred_block = get_nodes_block(pred);
1820 if ((!is_Block_dead(pred_block)) &&
1821 (get_Block_dom_depth(pred_block) > depth)) {
1823 depth = get_Block_dom_depth(pred_block);
1825 /* Avoid that the node is placed in the Start block */
1827 get_Block_dom_depth(get_nodes_block(n)) > 1 &&
1828 get_irg_phase_state(current_ir_graph) != phase_backend) {
1829 b = get_Block_cfg_out(get_irg_start_block(current_ir_graph), 0);
1830 assert(b != get_irg_start_block(current_ir_graph));
1835 set_nodes_block(n, b);
1839 * Add predecessors of non floating nodes and non-floating predecessors
1840 * of floating nodes to worklist and fix their blocks if the are in dead block.
1842 irn_arity = get_irn_arity(n);
1846 * Simplest case: End node. Predecessors are keep-alives,
1847 * no need to move out of dead block.
1849 for (i = -1; i < irn_arity; ++i) {
1850 ir_node *pred = get_irn_n(n, i);
1851 if (irn_not_visited(pred))
1852 waitq_put(worklist, pred);
1854 } else if (is_Block(n)) {
1856 * Blocks: Predecessors are control flow, no need to move
1857 * them out of dead block.
1859 for (i = irn_arity - 1; i >= 0; --i) {
1860 ir_node *pred = get_irn_n(n, i);
1861 if (irn_not_visited(pred))
1862 waitq_put(worklist, pred);
1864 } else if (is_Phi(n)) {
1866 ir_node *curr_block = get_nodes_block(n);
1867 int in_dead_block = is_Block_unreachable(curr_block);
1870 * Phi nodes: move nodes from dead blocks into the effective use
1871 * of the Phi-input if the Phi is not in a bad block.
1873 pred = get_nodes_block(n);
1874 if (irn_not_visited(pred))
1875 waitq_put(worklist, pred);
1877 for (i = irn_arity - 1; i >= 0; --i) {
1878 ir_node *pred = get_irn_n(n, i);
1880 if (irn_not_visited(pred)) {
1881 if (! in_dead_block &&
1882 get_irn_pinned(pred) == op_pin_state_floats &&
1883 is_Block_unreachable(get_nodes_block(pred))) {
1884 set_nodes_block(pred, get_Block_cfgpred_block(curr_block, i));
1886 waitq_put(worklist, pred);
1891 ir_node *curr_block = get_nodes_block(n);
1892 int in_dead_block = is_Block_unreachable(curr_block);
1895 * All other nodes: move nodes from dead blocks into the same block.
1897 pred = get_nodes_block(n);
1898 if (irn_not_visited(pred))
1899 waitq_put(worklist, pred);
1901 for (i = irn_arity - 1; i >= 0; --i) {
1902 ir_node *pred = get_irn_n(n, i);
1904 if (irn_not_visited(pred)) {
1905 if (! in_dead_block &&
1906 get_irn_pinned(pred) == op_pin_state_floats &&
1907 is_Block_unreachable(get_nodes_block(pred))) {
1908 set_nodes_block(pred, curr_block);
1910 waitq_put(worklist, pred);
1917 * Floating nodes form subgraphs that begin at nodes as Const, Load,
1918 * Start, Call and that end at op_pin_state_pinned nodes as Store, Call. Place_early
1919 * places all floating nodes reachable from its argument through floating
1920 * nodes and adds all beginnings at op_pin_state_pinned nodes to the worklist.
1922 * @param worklist a worklist, used for the algorithm, empty on in/output
1924 static void place_early(waitq *worklist) {
1926 inc_irg_visited(current_ir_graph);
1928 /* this inits the worklist */
1929 place_floats_early(get_irg_end(current_ir_graph), worklist);
1931 /* Work the content of the worklist. */
1932 while (!waitq_empty(worklist)) {
1933 ir_node *n = waitq_get(worklist);
1934 if (irn_not_visited(n))
1935 place_floats_early(n, worklist);
1938 set_irg_outs_inconsistent(current_ir_graph);
1939 set_irg_pinned(current_ir_graph, op_pin_state_pinned);
1943 * Compute the deepest common ancestor of block and dca.
1945 static ir_node *calc_dca(ir_node *dca, ir_node *block) {
1948 /* we do not want to place nodes in dead blocks */
1949 if (is_Block_dead(block))
1952 /* We found a first legal placement. */
1953 if (!dca) return block;
1955 /* Find a placement that is dominates both, dca and block. */
1956 while (get_Block_dom_depth(block) > get_Block_dom_depth(dca))
1957 block = get_Block_idom(block);
1959 while (get_Block_dom_depth(dca) > get_Block_dom_depth(block)) {
1960 dca = get_Block_idom(dca);
1963 while (block != dca) {
1964 block = get_Block_idom(block); dca = get_Block_idom(dca);
1970 /** Deepest common dominance ancestor of DCA and CONSUMER of PRODUCER.
1971 * I.e., DCA is the block where we might place PRODUCER.
1972 * A data flow edge points from producer to consumer.
1974 static ir_node *consumer_dom_dca(ir_node *dca, ir_node *consumer, ir_node *producer)
1976 /* Compute the last block into which we can place a node so that it is
1978 if (is_Phi(consumer)) {
1979 /* our consumer is a Phi-node, the effective use is in all those
1980 blocks through which the Phi-node reaches producer */
1981 ir_node *phi_block = get_nodes_block(consumer);
1982 int arity = get_irn_arity(consumer);
1985 for (i = 0; i < arity; i++) {
1986 if (get_Phi_pred(consumer, i) == producer) {
1987 ir_node *new_block = get_Block_cfgpred_block(phi_block, i);
1989 if (!is_Block_unreachable(new_block))
1990 dca = calc_dca(dca, new_block);
1994 dca = calc_dca(dca, get_nodes_block(consumer));
2000 /* FIXME: the name clashes here with the function from ana/field_temperature.c
2002 static INLINE int get_irn_loop_depth(ir_node *n) {
2003 return get_loop_depth(get_irn_loop(n));
2007 * Move n to a block with less loop depth than it's current block. The
2008 * new block must be dominated by early.
2010 * @param n the node that should be moved
2011 * @param early the earliest block we can n move to
2013 static void move_out_of_loops(ir_node *n, ir_node *early) {
2014 ir_node *best, *dca;
2018 /* Find the region deepest in the dominator tree dominating
2019 dca with the least loop nesting depth, but still dominated
2020 by our early placement. */
2021 dca = get_nodes_block(n);
2024 while (dca != early) {
2025 dca = get_Block_idom(dca);
2026 if (!dca || is_Bad(dca)) break; /* may be Bad if not reachable from Start */
2027 if (get_irn_loop_depth(dca) < get_irn_loop_depth(best)) {
2031 if (best != get_nodes_block(n)) {
2033 printf("Moving out of loop: "); DDMN(n);
2034 printf(" Outermost block: "); DDMN(early);
2035 printf(" Best block: "); DDMN(best);
2036 printf(" Innermost block: "); DDMN(get_nodes_block(n));
2038 set_nodes_block(n, best);
2042 /* deepest common ancestor in the dominator tree of all nodes'
2043 blocks depending on us; our final placement has to dominate DCA. */
2044 static ir_node *get_deepest_common_ancestor(ir_node *node, ir_node *dca)
2048 for (i = get_irn_n_outs(node) - 1; i >= 0; --i) {
2049 ir_node *succ = get_irn_out(node, i);
2053 * This consumer is the End node, a keep alive edge.
2054 * This is not a real consumer, so we ignore it
2059 if (is_Proj(succ)) {
2060 dca = get_deepest_common_ancestor(succ, dca);
2062 /* ignore if succ is in dead code */
2063 ir_node *succ_blk = get_nodes_block(succ);
2064 if (is_Block_unreachable(succ_blk))
2066 dca = consumer_dom_dca(dca, succ, node);
2073 static void set_projs_block(ir_node *node, ir_node *block)
2077 for (i = get_irn_n_outs(node) - 1; i >= 0; --i) {
2078 ir_node *succ = get_irn_out(node, i);
2080 assert(is_Proj(succ));
2082 if(get_irn_mode(succ) == mode_T) {
2083 set_projs_block(succ, block);
2085 set_nodes_block(succ, block);
2090 * Find the latest legal block for N and place N into the
2091 * `optimal' Block between the latest and earliest legal block.
2092 * The `optimal' block is the dominance-deepest block of those
2093 * with the least loop-nesting-depth. This places N out of as many
2094 * loops as possible and then makes it as control dependent as
2097 * @param n the node to be placed
2098 * @param worklist a worklist, all successors of non-floating nodes are
2101 static void place_floats_late(ir_node *n, pdeq *worklist) {
2105 assert(irn_not_visited(n)); /* no multiple placement */
2107 mark_irn_visited(n);
2109 /* no need to place block nodes, control nodes are already placed. */
2112 (get_irn_mode(n) != mode_X)) {
2113 /* Remember the early_blk placement of this block to move it
2114 out of loop no further than the early_blk placement. */
2115 early_blk = get_nodes_block(n);
2118 * BEWARE: Here we also get code, that is live, but
2119 * was in a dead block. If the node is life, but because
2120 * of CSE in a dead block, we still might need it.
2123 /* Assure that our users are all placed, except the Phi-nodes.
2124 --- Each data flow cycle contains at least one Phi-node. We
2125 have to break the `user has to be placed before the
2126 producer' dependence cycle and the Phi-nodes are the
2127 place to do so, because we need to base our placement on the
2128 final region of our users, which is OK with Phi-nodes, as they
2129 are op_pin_state_pinned, and they never have to be placed after a
2130 producer of one of their inputs in the same block anyway. */
2131 for (i = get_irn_n_outs(n) - 1; i >= 0; --i) {
2132 ir_node *succ = get_irn_out(n, i);
2133 if (irn_not_visited(succ) && !is_Phi(succ))
2134 place_floats_late(succ, worklist);
2137 if (! is_Block_dead(early_blk)) {
2138 /* do only move things that where not dead */
2139 ir_op *op = get_irn_op(n);
2141 /* We have to determine the final block of this node... except for
2142 constants and Projs */
2143 if ((get_irn_pinned(n) == op_pin_state_floats) &&
2145 (op != op_SymConst) &&
2148 /* deepest common ancestor in the dominator tree of all nodes'
2149 blocks depending on us; our final placement has to dominate
2151 ir_node *dca = get_deepest_common_ancestor(n, NULL);
2153 set_nodes_block(n, dca);
2154 move_out_of_loops(n, early_blk);
2155 if(get_irn_mode(n) == mode_T) {
2156 set_projs_block(n, get_nodes_block(n));
2163 /* Add successors of all non-floating nodes on list. (Those of floating
2164 nodes are placed already and therefore are marked.) */
2165 for (i = 0; i < get_irn_n_outs(n); i++) {
2166 ir_node *succ = get_irn_out(n, i);
2167 if (irn_not_visited(get_irn_out(n, i))) {
2168 pdeq_putr(worklist, succ);
2174 * Place floating nodes on the given worklist as late as possible using
2175 * the dominance tree.
2177 * @param worklist the worklist containing the nodes to place
2179 static void place_late(waitq *worklist) {
2181 inc_irg_visited(current_ir_graph);
2183 /* This fills the worklist initially. */
2184 place_floats_late(get_irg_start_block(current_ir_graph), worklist);
2186 /* And now empty the worklist again... */
2187 while (!waitq_empty(worklist)) {
2188 ir_node *n = waitq_get(worklist);
2189 if (irn_not_visited(n))
2190 place_floats_late(n, worklist);
2194 /* Code Placement. */
2195 void place_code(ir_graph *irg) {
2197 ir_graph *rem = current_ir_graph;
2199 current_ir_graph = irg;
2201 /* Handle graph state */
2202 assert(get_irg_phase_state(irg) != phase_building);
2205 if (1 || get_irg_loopinfo_state(irg) != loopinfo_consistent) {
2206 free_loop_information(irg);
2207 construct_cf_backedges(irg);
2210 /* Place all floating nodes as early as possible. This guarantees
2211 a legal code placement. */
2212 worklist = new_waitq();
2213 place_early(worklist);
2215 /* place_early() invalidates the outs, place_late needs them. */
2216 compute_irg_outs(irg);
2218 /* Now move the nodes down in the dominator tree. This reduces the
2219 unnecessary executions of the node. */
2220 place_late(worklist);
2222 set_irg_outs_inconsistent(current_ir_graph);
2223 set_irg_loopinfo_inconsistent(current_ir_graph);
2224 del_waitq(worklist);
2225 current_ir_graph = rem;
2228 typedef struct cf_env {
2229 char changed; /**< flag indicates that the cf graphs has changed. */
2233 * Called by walker of remove_critical_cf_edges().
2235 * Place an empty block to an edge between a blocks of multiple
2236 * predecessors and a block of multiple successors.
2239 * @param env Environment of walker.
2241 static void walk_critical_cf_edges(ir_node *n, void *env) {
2243 ir_node *pre, *block, *jmp;
2245 ir_graph *irg = get_irn_irg(n);
2247 /* Block has multiple predecessors */
2248 arity = get_irn_arity(n);
2250 if (n == get_irg_end_block(irg))
2251 return; /* No use to add a block here. */
2253 for (i = 0; i < arity; ++i) {
2256 pre = get_irn_n(n, i);
2257 cfop = get_irn_op(skip_Proj(pre));
2259 if (is_op_fragile(cfop)) {
2260 if (cfop != op_Raise)
2264 /* we don't want place nodes in the start block, so handle it like forking */
2265 if (is_op_forking(cfop) || cfop == op_Start) {
2266 /* Predecessor has multiple successors. Insert new control flow edge edges. */
2268 /* set predecessor of new block */
2269 block = new_r_Block(irg, 1, &pre);
2270 /* insert new jmp node to new block */
2271 jmp = new_r_Jmp(irg, block);
2272 /* set successor of new block */
2273 set_irn_n(n, i, jmp);
2275 } /* predecessor has multiple successors */
2276 } /* for all predecessors */
2277 } /* n is a multi-entry block */
2280 void remove_critical_cf_edges(ir_graph *irg) {
2285 irg_block_walk_graph(irg, NULL, walk_critical_cf_edges, &env);
2287 /* control flow changed */
2288 set_irg_outs_inconsistent(irg);
2289 set_irg_extblk_inconsistent(irg);
2290 set_irg_doms_inconsistent(irg);
2291 set_irg_loopinfo_inconsistent(irg);