3 * File name: ir/ir/irgopt.c
4 * Purpose: Optimizations for a whole ir graph, i.e., a procedure.
5 * Author: Christian Schaefer, Goetz Lindenmaier
6 * Modified by: Sebastian Felis, Michael Beck
9 * Copyright: (c) 1998-2007 Universität Karlsruhe
10 * Licence: This file protected by GPL - GNU GENERAL PUBLIC LICENSE.
19 #include "irgraph_t.h"
32 #include "pdeq.h" /* Fuer code placement */
37 #include "irbackedge_t.h"
44 #include "iredges_t.h"
47 /*------------------------------------------------------------------*/
48 /* apply optimizations of iropt to all nodes. */
49 /*------------------------------------------------------------------*/
52 * A wrapper around optimize_inplace_2() to be called from a walker.
54 static void optimize_in_place_wrapper (ir_node *n, void *env) {
55 ir_node *optimized = optimize_in_place_2(n);
56 if (optimized != n) exchange (n, optimized);
60 * Do local optimizations for a node.
62 * @param n the IR-node where to start. Typically the End node
65 * @note current_ir_graph must be set
67 static INLINE void do_local_optimize(ir_node *n) {
68 /* Handle graph state */
69 assert(get_irg_phase_state(current_ir_graph) != phase_building);
71 if (get_opt_global_cse())
72 set_irg_pinned(current_ir_graph, op_pin_state_floats);
73 set_irg_outs_inconsistent(current_ir_graph);
74 set_irg_doms_inconsistent(current_ir_graph);
75 set_irg_loopinfo_inconsistent(current_ir_graph);
77 /* Clean the value_table in irg for the CSE. */
78 del_identities(current_ir_graph->value_table);
79 current_ir_graph->value_table = new_identities();
81 /* walk over the graph */
82 irg_walk(n, firm_clear_link, optimize_in_place_wrapper, NULL);
85 /* Applies local optimizations (see iropt.h) to all nodes reachable from node n */
86 void local_optimize_node(ir_node *n) {
87 ir_graph *rem = current_ir_graph;
88 current_ir_graph = get_irn_irg(n);
92 current_ir_graph = rem;
96 * Block-Walker: uses dominance depth to mark dead blocks.
98 static void kill_dead_blocks(ir_node *block, void *env) {
99 if (get_Block_dom_depth(block) < 0) {
101 * Note that the new dominance code correctly handles
102 * the End block, i.e. it is always reachable from Start
104 set_Block_dead(block);
108 /* Applies local optimizations (see iropt.h) to all nodes reachable from node n. */
109 void local_optimize_graph(ir_graph *irg) {
110 ir_graph *rem = current_ir_graph;
111 current_ir_graph = irg;
113 if (get_irg_dom_state(irg) == dom_consistent)
114 irg_block_walk_graph(irg, NULL, kill_dead_blocks, NULL);
116 do_local_optimize(get_irg_end(irg));
118 current_ir_graph = rem;
122 * Enqueue all users of a node to a wait queue.
123 * Handles mode_T nodes.
125 static void enqueue_users(ir_node *n, pdeq *waitq) {
126 const ir_edge_t *edge;
128 foreach_out_edge(n, edge) {
129 ir_node *succ = get_edge_src_irn(edge);
131 if (get_irn_link(succ) != waitq) {
132 pdeq_putr(waitq, succ);
133 set_irn_link(succ, waitq);
135 if (get_irn_mode(succ) == mode_T) {
136 /* A mode_T node has Proj's. Because most optimizations
137 run on the Proj's we have to enqueue them also. */
138 enqueue_users(succ, waitq);
144 * Data flow optimization walker.
145 * Optimizes all nodes and enqueue it's users
148 static void opt_walker(ir_node *n, void *env) {
152 optimized = optimize_in_place_2(n);
153 set_irn_link(optimized, NULL);
155 if (optimized != n) {
156 enqueue_users(n, waitq);
157 exchange(n, optimized);
161 /* Applies local optimizations to all nodes in the graph until fixpoint. */
162 void optimize_graph_df(ir_graph *irg) {
163 pdeq *waitq = new_pdeq();
164 int state = edges_activated(irg);
165 ir_graph *rem = current_ir_graph;
167 current_ir_graph = irg;
172 if (get_opt_global_cse())
173 set_irg_pinned(current_ir_graph, op_pin_state_floats);
175 /* Clean the value_table in irg for the CSE. */
176 del_identities(irg->value_table);
177 irg->value_table = new_identities();
179 if (get_irg_dom_state(irg) == dom_consistent)
180 irg_block_walk_graph(irg, NULL, kill_dead_blocks, NULL);
182 /* invalidate info */
183 set_irg_outs_inconsistent(irg);
184 set_irg_doms_inconsistent(irg);
185 set_irg_loopinfo_inconsistent(irg);
187 set_using_irn_link(irg);
189 /* walk over the graph */
190 irg_walk_graph(irg, NULL, opt_walker, waitq);
192 /* finish the wait queue */
193 while (! pdeq_empty(waitq)) {
194 ir_node *n = pdeq_getl(waitq);
196 opt_walker(n, waitq);
201 clear_using_irn_link(irg);
204 edges_deactivate(irg);
206 current_ir_graph = rem;
210 /*------------------------------------------------------------------*/
211 /* Routines for dead node elimination / copying garbage collection */
212 /* of the obstack. */
213 /*------------------------------------------------------------------*/
216 * Remember the new node in the old node by using a field all nodes have.
218 #define set_new_node(oldn, newn) set_irn_link(oldn, newn)
221 * Get this new node, before the old node is forgotten.
223 #define get_new_node(oldn) get_irn_link(oldn)
226 * Check if a new node was set.
228 #define has_new_node(n) (get_new_node(n) != NULL)
231 * We use the block_visited flag to mark that we have computed the
232 * number of useful predecessors for this block.
233 * Further we encode the new arity in this flag in the old blocks.
234 * Remembering the arity is useful, as it saves a lot of pointer
235 * accesses. This function is called for all Phi and Block nodes
239 compute_new_arity(ir_node *b) {
240 int i, res, irn_arity;
243 irg_v = get_irg_block_visited(current_ir_graph);
244 block_v = get_Block_block_visited(b);
245 if (block_v >= irg_v) {
246 /* we computed the number of preds for this block and saved it in the
248 return block_v - irg_v;
250 /* compute the number of good predecessors */
251 res = irn_arity = get_irn_arity(b);
252 for (i = 0; i < irn_arity; i++)
253 if (get_irn_opcode(get_irn_n(b, i)) == iro_Bad) res--;
254 /* save it in the flag. */
255 set_Block_block_visited(b, irg_v + res);
261 * Copies the node to the new obstack. The Ins of the new node point to
262 * the predecessors on the old obstack. For block/phi nodes not all
263 * predecessors might be copied. n->link points to the new node.
264 * For Phi and Block nodes the function allocates in-arrays with an arity
265 * only for useful predecessors. The arity is determined by counting
266 * the non-bad predecessors of the block.
268 * @param n The node to be copied
269 * @param env if non-NULL, the node number attribute will be copied to the new node
271 * Note: Also used for loop unrolling.
273 static void copy_node(ir_node *n, void *env) {
276 ir_op *op = get_irn_op(n);
278 /* The end node looses it's flexible in array. This doesn't matter,
279 as dead node elimination builds End by hand, inlineing doesn't use
281 /* assert(op == op_End || ((_ARR_DESCR(n->in))->cookie != ARR_F_MAGIC)); */
284 /* node copied already */
286 } else if (op == op_Block) {
288 new_arity = compute_new_arity(n);
289 n->attr.block.graph_arr = NULL;
291 block = get_nodes_block(n);
293 new_arity = compute_new_arity(block);
295 new_arity = get_irn_arity(n);
298 nn = new_ir_node(get_irn_dbg_info(n),
305 /* Copy the attributes. These might point to additional data. If this
306 was allocated on the old obstack the pointers now are dangling. This
307 frees e.g. the memory of the graph_arr allocated in new_immBlock. */
308 copy_node_attr(n, nn);
312 int copy_node_nr = env != NULL;
314 /* for easier debugging, we want to copy the node numbers too */
315 nn->node_nr = n->node_nr;
321 hook_dead_node_elim_subst(current_ir_graph, n, nn);
325 * Copies new predecessors of old node to new node remembered in link.
326 * Spare the Bad predecessors of Phi and Block nodes.
329 copy_preds(ir_node *n, void *env) {
333 nn = get_new_node(n);
336 /* Don't copy Bad nodes. */
338 irn_arity = get_irn_arity(n);
339 for (i = 0; i < irn_arity; i++) {
340 if (! is_Bad(get_irn_n(n, i))) {
341 set_irn_n (nn, j, get_new_node(get_irn_n(n, i)));
342 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
346 /* repair the block visited flag from above misuse. Repair it in both
347 graphs so that the old one can still be used. */
348 set_Block_block_visited(nn, 0);
349 set_Block_block_visited(n, 0);
350 /* Local optimization could not merge two subsequent blocks if
351 in array contained Bads. Now it's possible.
352 We don't call optimize_in_place as it requires
353 that the fields in ir_graph are set properly. */
354 if ((get_opt_control_flow_straightening()) &&
355 (get_Block_n_cfgpreds(nn) == 1) &&
356 (get_irn_op(get_Block_cfgpred(nn, 0)) == op_Jmp)) {
357 ir_node *old = get_nodes_block(get_Block_cfgpred(nn, 0));
359 /* Jmp jumps into the block it is in -- deal self cycle. */
360 assert(is_Bad(get_new_node(get_irg_bad(current_ir_graph))));
361 exchange(nn, get_new_node(get_irg_bad(current_ir_graph)));
366 } else if (get_irn_op(n) == op_Phi) {
367 /* Don't copy node if corresponding predecessor in block is Bad.
368 The Block itself should not be Bad. */
369 block = get_nodes_block(n);
370 set_irn_n(nn, -1, get_new_node(block));
372 irn_arity = get_irn_arity(n);
373 for (i = 0; i < irn_arity; i++) {
374 if (! is_Bad(get_irn_n(block, i))) {
375 set_irn_n(nn, j, get_new_node(get_irn_n(n, i)));
376 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
380 /* If the pre walker reached this Phi after the post walker visited the
381 block block_visited is > 0. */
382 set_Block_block_visited(get_nodes_block(n), 0);
383 /* Compacting the Phi's ins might generate Phis with only one
385 if (get_irn_arity(nn) == 1)
386 exchange(nn, get_irn_n(nn, 0));
388 irn_arity = get_irn_arity(n);
389 for (i = -1; i < irn_arity; i++)
390 set_irn_n (nn, i, get_new_node(get_irn_n(n, i)));
392 /* Now the new node is complete. We can add it to the hash table for CSE.
393 @@@ inlining aborts if we identify End. Why? */
394 if (get_irn_op(nn) != op_End)
395 add_identities(current_ir_graph->value_table, nn);
399 * Copies the graph recursively, compacts the keep-alives of the end node.
401 * @param irg the graph to be copied
402 * @param copy_node_nr If non-zero, the node number will be copied
404 static void copy_graph(ir_graph *irg, int copy_node_nr) {
405 ir_node *oe, *ne, *ob, *nb, *om, *nm; /* old end, new end, old bad, new bad, old NoMem, new NoMem */
406 ir_node *ka; /* keep alive */
410 /* Some nodes must be copied by hand, sigh */
411 vfl = get_irg_visited(irg);
412 set_irg_visited(irg, vfl + 1);
414 oe = get_irg_end(irg);
415 mark_irn_visited(oe);
416 /* copy the end node by hand, allocate dynamic in array! */
417 ne = new_ir_node(get_irn_dbg_info(oe),
424 /* Copy the attributes. Well, there might be some in the future... */
425 copy_node_attr(oe, ne);
426 set_new_node(oe, ne);
428 /* copy the Bad node */
429 ob = get_irg_bad(irg);
430 mark_irn_visited(ob);
431 nb = new_ir_node(get_irn_dbg_info(ob),
438 copy_node_attr(ob, nb);
439 set_new_node(ob, nb);
441 /* copy the NoMem node */
442 om = get_irg_no_mem(irg);
443 mark_irn_visited(om);
444 nm = new_ir_node(get_irn_dbg_info(om),
451 copy_node_attr(om, nm);
452 set_new_node(om, nm);
454 /* copy the live nodes */
455 set_irg_visited(irg, vfl);
456 irg_walk(get_nodes_block(oe), copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
458 /* Note: from yet, the visited flag of the graph is equal to vfl + 1 */
460 /* visit the anchors as well */
461 for (i = anchor_max - 1; i >= 0; --i) {
462 ir_node *n = irg->anchors[i];
464 if (n && (get_irn_visited(n) <= vfl)) {
465 set_irg_visited(irg, vfl);
466 irg_walk(n, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
470 /* copy_preds for the end node ... */
471 set_nodes_block(ne, get_new_node(get_nodes_block(oe)));
473 /*- ... and now the keep alives. -*/
474 /* First pick the not marked block nodes and walk them. We must pick these
475 first as else we will oversee blocks reachable from Phis. */
476 irn_arity = get_End_n_keepalives(oe);
477 for (i = 0; i < irn_arity; i++) {
478 ka = get_End_keepalive(oe, i);
480 if (get_irn_visited(ka) <= vfl) {
481 /* We must keep the block alive and copy everything reachable */
482 set_irg_visited(irg, vfl);
483 irg_walk(ka, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
485 add_End_keepalive(ne, get_new_node(ka));
489 /* Now pick other nodes. Here we will keep all! */
490 irn_arity = get_End_n_keepalives(oe);
491 for (i = 0; i < irn_arity; i++) {
492 ka = get_End_keepalive(oe, i);
494 if (get_irn_visited(ka) <= vfl) {
495 /* We didn't copy the node yet. */
496 set_irg_visited(irg, vfl);
497 irg_walk(ka, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
499 add_End_keepalive(ne, get_new_node(ka));
503 /* start block sometimes only reached after keep alives */
504 set_nodes_block(nb, get_new_node(get_nodes_block(ob)));
505 set_nodes_block(nm, get_new_node(get_nodes_block(om)));
509 * Copies the graph reachable from current_ir_graph->end to the obstack
510 * in current_ir_graph and fixes the environment.
511 * Then fixes the fields in current_ir_graph containing nodes of the
514 * @param copy_node_nr If non-zero, the node number will be copied
517 copy_graph_env(int copy_node_nr) {
518 ir_graph *irg = current_ir_graph;
519 ir_node *old_end, *n;
522 /* remove end_except and end_reg nodes */
523 old_end = get_irg_end(irg);
524 set_irg_end_except (irg, old_end);
525 set_irg_end_reg (irg, old_end);
527 /* Not all nodes remembered in irg might be reachable
528 from the end node. Assure their link is set to NULL, so that
529 we can test whether new nodes have been computed. */
530 for (i = anchor_max - 1; i >= 0; --i) {
532 set_new_node(irg->anchors[i], NULL);
534 /* we use the block walk flag for removing Bads from Blocks ins. */
535 inc_irg_block_visited(irg);
538 copy_graph(irg, copy_node_nr);
540 /* fix the fields in irg */
541 old_end = get_irg_end(irg);
542 for (i = anchor_max - 1; i >= 0; --i) {
545 irg->anchors[i] = get_new_node(n);
551 * Copies all reachable nodes to a new obstack. Removes bad inputs
552 * from block nodes and the corresponding inputs from Phi nodes.
553 * Merges single exit blocks with single entry blocks and removes
555 * Adds all new nodes to a new hash table for CSE. Does not
556 * perform CSE, so the hash table might contain common subexpressions.
559 dead_node_elimination(ir_graph *irg) {
560 if (get_opt_optimize() && get_opt_dead_node_elimination()) {
562 int rem_ipview = get_interprocedural_view();
563 struct obstack *graveyard_obst = NULL;
564 struct obstack *rebirth_obst = NULL;
565 assert(! edges_activated(irg) && "dead node elimination requires disabled edges");
567 /* inform statistics that we started a dead-node elimination run */
568 hook_dead_node_elim(irg, 1);
570 /* Remember external state of current_ir_graph. */
571 rem = current_ir_graph;
572 current_ir_graph = irg;
573 set_interprocedural_view(0);
575 assert(get_irg_phase_state(irg) != phase_building);
577 /* Handle graph state */
578 free_callee_info(irg);
582 /* @@@ so far we loose loops when copying */
583 free_loop_information(irg);
585 set_irg_doms_inconsistent(irg);
587 /* A quiet place, where the old obstack can rest in peace,
588 until it will be cremated. */
589 graveyard_obst = irg->obst;
591 /* A new obstack, where the reachable nodes will be copied to. */
592 rebirth_obst = xmalloc(sizeof(*rebirth_obst));
593 irg->obst = rebirth_obst;
594 obstack_init(irg->obst);
595 irg->last_node_idx = 0;
597 /* We also need a new value table for CSE */
598 del_identities(irg->value_table);
599 irg->value_table = new_identities();
601 /* Copy the graph from the old to the new obstack */
602 copy_graph_env(/*copy_node_nr=*/1);
604 /* Free memory from old unoptimized obstack */
605 obstack_free(graveyard_obst, 0); /* First empty the obstack ... */
606 xfree (graveyard_obst); /* ... then free it. */
608 /* inform statistics that the run is over */
609 hook_dead_node_elim(irg, 0);
611 current_ir_graph = rem;
612 set_interprocedural_view(rem_ipview);
617 * Relink bad predecessors of a block and store the old in array to the
618 * link field. This function is called by relink_bad_predecessors().
619 * The array of link field starts with the block operand at position 0.
620 * If block has bad predecessors, create a new in array without bad preds.
621 * Otherwise let in array untouched.
623 static void relink_bad_block_predecessors(ir_node *n, void *env) {
624 ir_node **new_in, *irn;
625 int i, new_irn_n, old_irn_arity, new_irn_arity = 0;
627 /* if link field of block is NULL, look for bad predecessors otherwise
628 this is already done */
629 if (get_irn_op(n) == op_Block &&
630 get_irn_link(n) == NULL) {
632 /* save old predecessors in link field (position 0 is the block operand)*/
633 set_irn_link(n, get_irn_in(n));
635 /* count predecessors without bad nodes */
636 old_irn_arity = get_irn_arity(n);
637 for (i = 0; i < old_irn_arity; i++)
638 if (!is_Bad(get_irn_n(n, i))) new_irn_arity++;
640 /* arity changing: set new predecessors without bad nodes */
641 if (new_irn_arity < old_irn_arity) {
642 /* Get new predecessor array. We do not resize the array, as we must
643 keep the old one to update Phis. */
644 new_in = NEW_ARR_D (ir_node *, current_ir_graph->obst, (new_irn_arity+1));
646 /* set new predecessors in array */
649 for (i = 0; i < old_irn_arity; i++) {
650 irn = get_irn_n(n, i);
652 new_in[new_irn_n] = irn;
653 is_backedge(n, i) ? set_backedge(n, new_irn_n-1) : set_not_backedge(n, new_irn_n-1);
657 /* ARR_SETLEN(int, n->attr.block.backedge, new_irn_arity); */
658 ARR_SHRINKLEN(n->attr.block.backedge, new_irn_arity);
660 } /* ir node has bad predecessors */
661 } /* Block is not relinked */
665 * Relinks Bad predecessors from Blocks and Phis called by walker
666 * remove_bad_predecesors(). If n is a Block, call
667 * relink_bad_block_redecessors(). If n is a Phi-node, call also the relinking
668 * function of Phi's Block. If this block has bad predecessors, relink preds
671 static void relink_bad_predecessors(ir_node *n, void *env) {
672 ir_node *block, **old_in;
673 int i, old_irn_arity, new_irn_arity;
675 /* relink bad predecessors of a block */
676 if (get_irn_op(n) == op_Block)
677 relink_bad_block_predecessors(n, env);
679 /* If Phi node relink its block and its predecessors */
680 if (get_irn_op(n) == op_Phi) {
682 /* Relink predecessors of phi's block */
683 block = get_nodes_block(n);
684 if (get_irn_link(block) == NULL)
685 relink_bad_block_predecessors(block, env);
687 old_in = (ir_node **)get_irn_link(block); /* Of Phi's Block */
688 old_irn_arity = ARR_LEN(old_in);
690 /* Relink Phi predecessors if count of predecessors changed */
691 if (old_irn_arity != ARR_LEN(get_irn_in(block))) {
692 /* set new predecessors in array
693 n->in[0] remains the same block */
695 for(i = 1; i < old_irn_arity; i++)
696 if (!is_Bad((ir_node *)old_in[i])) {
697 n->in[new_irn_arity] = n->in[i];
698 is_backedge(n, i) ? set_backedge(n, new_irn_arity) : set_not_backedge(n, new_irn_arity);
702 ARR_SETLEN(ir_node *, n->in, new_irn_arity);
703 ARR_SETLEN(int, n->attr.phi_backedge, new_irn_arity);
705 } /* n is a Phi node */
709 * Removes Bad Bad predecessors from Blocks and the corresponding
710 * inputs to Phi nodes as in dead_node_elimination but without
712 * On walking up set the link field to NULL, on walking down call
713 * relink_bad_predecessors() (This function stores the old in array
714 * to the link field and sets a new in array if arity of predecessors
717 void remove_bad_predecessors(ir_graph *irg) {
718 irg_walk_graph(irg, firm_clear_link, relink_bad_predecessors, NULL);
725 __)|_| | \_/ | \_/(/_ |_/\__|__
727 The following stuff implements a facility that automatically patches
728 registered ir_node pointers to the new node when a dead node elimination occurs.
731 struct _survive_dce_t {
735 hook_entry_t dead_node_elim;
736 hook_entry_t dead_node_elim_subst;
739 typedef struct _survive_dce_list_t {
740 struct _survive_dce_list_t *next;
742 } survive_dce_list_t;
744 static void dead_node_hook(void *context, ir_graph *irg, int start) {
745 survive_dce_t *sd = context;
747 /* Create a new map before the dead node elimination is performed. */
749 sd->new_places = pmap_create_ex(pmap_count(sd->places));
751 /* Patch back all nodes if dead node elimination is over and something is to be done. */
752 pmap_destroy(sd->places);
753 sd->places = sd->new_places;
754 sd->new_places = NULL;
759 * Hook called when dead node elimination replaces old by nw.
761 static void dead_node_subst_hook(void *context, ir_graph *irg, ir_node *old, ir_node *nw) {
762 survive_dce_t *sd = context;
763 survive_dce_list_t *list = pmap_get(sd->places, old);
765 /* If the node is to be patched back, write the new address to all registered locations. */
767 survive_dce_list_t *p;
769 for (p = list; p; p = p->next)
772 pmap_insert(sd->new_places, nw, list);
777 * Make a new Survive DCE environment.
779 survive_dce_t *new_survive_dce(void) {
780 survive_dce_t *res = xmalloc(sizeof(res[0]));
781 obstack_init(&res->obst);
782 res->places = pmap_create();
783 res->new_places = NULL;
785 res->dead_node_elim.hook._hook_dead_node_elim = dead_node_hook;
786 res->dead_node_elim.context = res;
787 res->dead_node_elim.next = NULL;
789 res->dead_node_elim_subst.hook._hook_dead_node_elim_subst = dead_node_subst_hook;
790 res->dead_node_elim_subst.context = res;
791 res->dead_node_elim_subst.next = NULL;
793 #ifndef FIRM_ENABLE_HOOKS
794 assert(0 && "need hooks enabled");
797 register_hook(hook_dead_node_elim, &res->dead_node_elim);
798 register_hook(hook_dead_node_elim_subst, &res->dead_node_elim_subst);
803 * Free a Survive DCE environment.
805 void free_survive_dce(survive_dce_t *sd) {
806 obstack_free(&sd->obst, NULL);
807 pmap_destroy(sd->places);
808 unregister_hook(hook_dead_node_elim, &sd->dead_node_elim);
809 unregister_hook(hook_dead_node_elim_subst, &sd->dead_node_elim_subst);
814 * Register a node pointer to be patched upon DCE.
815 * When DCE occurs, the node pointer specified by @p place will be
816 * patched to the new address of the node it is pointing to.
818 * @param sd The Survive DCE environment.
819 * @param place The address of the node pointer.
821 void survive_dce_register_irn(survive_dce_t *sd, ir_node **place) {
822 if (*place != NULL) {
823 ir_node *irn = *place;
824 survive_dce_list_t *curr = pmap_get(sd->places, irn);
825 survive_dce_list_t *nw = obstack_alloc(&sd->obst, sizeof(nw[0]));
830 pmap_insert(sd->places, irn, nw);
834 /*--------------------------------------------------------------------*/
835 /* Functionality for inlining */
836 /*--------------------------------------------------------------------*/
839 * Copy node for inlineing. Updates attributes that change when
840 * inlineing but not for dead node elimination.
842 * Copies the node by calling copy_node() and then updates the entity if
843 * it's a local one. env must be a pointer of the frame type of the
844 * inlined procedure. The new entities must be in the link field of
848 copy_node_inline(ir_node *n, void *env) {
850 ir_type *frame_tp = (ir_type *)env;
853 if (get_irn_op(n) == op_Sel) {
854 nn = get_new_node (n);
856 if (get_entity_owner(get_Sel_entity(n)) == frame_tp) {
857 set_Sel_entity(nn, get_entity_link(get_Sel_entity(n)));
859 } else if (get_irn_op(n) == op_Block) {
860 nn = get_new_node (n);
861 nn->attr.block.irg = current_ir_graph;
866 * Walker: checks if P_value_arg_base is used.
868 static void find_addr(ir_node *node, void *env) {
869 int *allow_inline = env;
870 if (is_Proj(node) && get_irn_op(get_Proj_pred(node)) == op_Start) {
871 if (get_Proj_proj(node) == pn_Start_P_value_arg_base)
877 * Check if we can inline a given call.
878 * Currently, we cannot inline two cases:
879 * - call with compound arguments
880 * - graphs that take the address of a parameter
882 * check these conditions here
884 static int can_inline(ir_node *call, ir_graph *called_graph) {
885 ir_type *call_type = get_Call_type(call);
886 int params, ress, i, res;
887 assert(is_Method_type(call_type));
889 params = get_method_n_params(call_type);
890 ress = get_method_n_ress(call_type);
892 /* check parameters for compound arguments */
893 for (i = 0; i < params; ++i) {
894 ir_type *p_type = get_method_param_type(call_type, i);
896 if (is_compound_type(p_type))
900 /* check results for compound arguments */
901 for (i = 0; i < ress; ++i) {
902 ir_type *r_type = get_method_res_type(call_type, i);
904 if (is_compound_type(r_type))
909 irg_walk_graph(called_graph, find_addr, NULL, &res);
914 /* Inlines a method at the given call site. */
915 int inline_method(ir_node *call, ir_graph *called_graph) {
917 ir_node *post_call, *post_bl;
918 ir_node *in[pn_Start_max];
919 ir_node *end, *end_bl;
923 int arity, n_ret, n_exc, n_res, i, j, rem_opt, irn_arity;
925 ir_type *called_frame;
926 irg_inline_property prop = get_irg_inline_property(called_graph);
928 if ( (prop < irg_inline_forced) &&
929 (!get_opt_optimize() || !get_opt_inline() || (prop == irg_inline_forbidden))) return 0;
931 /* Do not inline variadic functions. */
932 if (get_method_variadicity(get_entity_type(get_irg_entity(called_graph))) == variadicity_variadic)
935 assert(get_method_n_params(get_entity_type(get_irg_entity(called_graph))) ==
936 get_method_n_params(get_Call_type(call)));
939 * currently, we cannot inline two cases:
940 * - call with compound arguments
941 * - graphs that take the address of a parameter
943 if (! can_inline(call, called_graph))
946 /* -- Turn off optimizations, this can cause problems when allocating new nodes. -- */
947 rem_opt = get_opt_optimize();
950 /* Handle graph state */
951 assert(get_irg_phase_state(current_ir_graph) != phase_building);
952 assert(get_irg_pinned(current_ir_graph) == op_pin_state_pinned);
953 assert(get_irg_pinned(called_graph) == op_pin_state_pinned);
954 set_irg_outs_inconsistent(current_ir_graph);
955 set_irg_extblk_inconsistent(current_ir_graph);
956 set_irg_doms_inconsistent(current_ir_graph);
957 set_irg_loopinfo_inconsistent(current_ir_graph);
958 set_irg_callee_info_state(current_ir_graph, irg_callee_info_inconsistent);
960 /* -- Check preconditions -- */
961 assert(is_Call(call));
962 /* @@@ does not work for InterfaceIII.java after cgana
963 assert(get_Call_type(call) == get_entity_type(get_irg_entity(called_graph)));
964 assert(smaller_type(get_entity_type(get_irg_entity(called_graph)),
965 get_Call_type(call)));
967 if (called_graph == current_ir_graph) {
968 set_optimize(rem_opt);
972 /* here we know we WILL inline, so inform the statistics */
973 hook_inline(call, called_graph);
975 /* -- Decide how to handle exception control flow: Is there a handler
976 for the Call node, or do we branch directly to End on an exception?
978 0 There is a handler.
980 2 Exception handling not represented in Firm. -- */
982 ir_node *proj, *Mproj = NULL, *Xproj = NULL;
983 for (proj = get_irn_link(call); proj; proj = get_irn_link(proj)) {
984 assert(is_Proj(proj));
985 if (get_Proj_proj(proj) == pn_Call_X_except) Xproj = proj;
986 if (get_Proj_proj(proj) == pn_Call_M_except) Mproj = proj;
988 if (Mproj) { assert(Xproj); exc_handling = 0; } /* Mproj */
989 else if (Xproj) { exc_handling = 1; } /* !Mproj && Xproj */
990 else { exc_handling = 2; } /* !Mproj && !Xproj */
994 the procedure and later replaces the Start node of the called graph.
995 Post_call is the old Call node and collects the results of the called
996 graph. Both will end up being a tuple. -- */
997 post_bl = get_nodes_block(call);
998 set_irg_current_block(current_ir_graph, post_bl);
999 /* XxMxPxPxPxT of Start + parameter of Call */
1000 in[pn_Start_X_initial_exec] = new_Jmp();
1001 in[pn_Start_M] = get_Call_mem(call);
1002 in[pn_Start_P_frame_base] = get_irg_frame(current_ir_graph);
1003 in[pn_Start_P_globals] = get_irg_globals(current_ir_graph);
1004 in[pn_Start_P_tls] = get_irg_tls(current_ir_graph);
1005 in[pn_Start_T_args] = new_Tuple(get_Call_n_params(call), get_Call_param_arr(call));
1006 /* in[pn_Start_P_value_arg_base] = ??? */
1007 assert(pn_Start_P_value_arg_base == pn_Start_max - 1 && "pn_Start_P_value_arg_base not supported, fix");
1008 pre_call = new_Tuple(pn_Start_max - 1, in);
1012 The new block gets the ins of the old block, pre_call and all its
1013 predecessors and all Phi nodes. -- */
1014 part_block(pre_call);
1016 /* -- Prepare state for dead node elimination -- */
1017 /* Visited flags in calling irg must be >= flag in called irg.
1018 Else walker and arity computation will not work. */
1019 if (get_irg_visited(current_ir_graph) <= get_irg_visited(called_graph))
1020 set_irg_visited(current_ir_graph, get_irg_visited(called_graph)+1);
1021 if (get_irg_block_visited(current_ir_graph)< get_irg_block_visited(called_graph))
1022 set_irg_block_visited(current_ir_graph, get_irg_block_visited(called_graph));
1023 /* Set pre_call as new Start node in link field of the start node of
1024 calling graph and pre_calls block as new block for the start block
1026 Further mark these nodes so that they are not visited by the
1028 set_irn_link(get_irg_start(called_graph), pre_call);
1029 set_irn_visited(get_irg_start(called_graph), get_irg_visited(current_ir_graph));
1030 set_irn_link(get_irg_start_block(called_graph), get_nodes_block(pre_call));
1031 set_irn_visited(get_irg_start_block(called_graph), get_irg_visited(current_ir_graph));
1032 set_irn_link(get_irg_bad(called_graph), get_irg_bad(current_ir_graph));
1033 set_irn_visited(get_irg_bad(called_graph), get_irg_visited(current_ir_graph));
1035 /* Initialize for compaction of in arrays */
1036 inc_irg_block_visited(current_ir_graph);
1038 /* -- Replicate local entities of the called_graph -- */
1039 /* copy the entities. */
1040 called_frame = get_irg_frame_type(called_graph);
1041 for (i = 0; i < get_class_n_members(called_frame); i++) {
1042 ir_entity *new_ent, *old_ent;
1043 old_ent = get_class_member(called_frame, i);
1044 new_ent = copy_entity_own(old_ent, get_cur_frame_type());
1045 set_entity_link(old_ent, new_ent);
1048 /* visited is > than that of called graph. With this trick visited will
1049 remain unchanged so that an outer walker, e.g., searching the call nodes
1050 to inline, calling this inline will not visit the inlined nodes. */
1051 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
1053 /* -- Performing dead node elimination inlines the graph -- */
1054 /* Copies the nodes to the obstack of current_ir_graph. Updates links to new
1056 /* @@@ endless loops are not copied!! -- they should be, I think... */
1057 irg_walk(get_irg_end(called_graph), copy_node_inline, copy_preds,
1058 get_irg_frame_type(called_graph));
1060 /* Repair called_graph */
1061 set_irg_visited(called_graph, get_irg_visited(current_ir_graph));
1062 set_irg_block_visited(called_graph, get_irg_block_visited(current_ir_graph));
1063 set_Block_block_visited(get_irg_start_block(called_graph), 0);
1065 /* -- Merge the end of the inlined procedure with the call site -- */
1066 /* We will turn the old Call node into a Tuple with the following
1069 0: Phi of all Memories of Return statements.
1070 1: Jmp from new Block that merges the control flow from all exception
1071 predecessors of the old end block.
1072 2: Tuple of all arguments.
1073 3: Phi of Exception memories.
1074 In case the old Call directly branches to End on an exception we don't
1075 need the block merging all exceptions nor the Phi of the exception
1079 /* -- Precompute some values -- */
1080 end_bl = get_new_node(get_irg_end_block(called_graph));
1081 end = get_new_node(get_irg_end(called_graph));
1082 arity = get_irn_arity(end_bl); /* arity = n_exc + n_ret */
1083 n_res = get_method_n_ress(get_Call_type(call));
1085 res_pred = xmalloc (n_res * sizeof(*res_pred));
1086 cf_pred = xmalloc (arity * sizeof(*res_pred));
1088 set_irg_current_block(current_ir_graph, post_bl); /* just to make sure */
1090 /* -- archive keepalives -- */
1091 irn_arity = get_irn_arity(end);
1092 for (i = 0; i < irn_arity; i++)
1093 add_End_keepalive(get_irg_end(current_ir_graph), get_irn_n(end, i));
1095 /* The new end node will die. We need not free as the in array is on the obstack:
1096 copy_node() only generated 'D' arrays. */
1098 /* -- Replace Return nodes by Jump nodes. -- */
1100 for (i = 0; i < arity; i++) {
1102 ret = get_irn_n(end_bl, i);
1103 if (is_Return(ret)) {
1104 cf_pred[n_ret] = new_r_Jmp(current_ir_graph, get_nodes_block(ret));
1108 set_irn_in(post_bl, n_ret, cf_pred);
1110 /* -- Build a Tuple for all results of the method.
1111 Add Phi node if there was more than one Return. -- */
1112 turn_into_tuple(post_call, 4);
1113 /* First the Memory-Phi */
1115 for (i = 0; i < arity; i++) {
1116 ret = get_irn_n(end_bl, i);
1117 if (is_Return(ret)) {
1118 cf_pred[n_ret] = get_Return_mem(ret);
1122 phi = new_Phi(n_ret, cf_pred, mode_M);
1123 set_Tuple_pred(call, pn_Call_M_regular, phi);
1124 /* Conserve Phi-list for further inlinings -- but might be optimized */
1125 if (get_nodes_block(phi) == post_bl) {
1126 set_irn_link(phi, get_irn_link(post_bl));
1127 set_irn_link(post_bl, phi);
1129 /* Now the real results */
1131 for (j = 0; j < n_res; j++) {
1133 for (i = 0; i < arity; i++) {
1134 ret = get_irn_n(end_bl, i);
1135 if (get_irn_op(ret) == op_Return) {
1136 cf_pred[n_ret] = get_Return_res(ret, j);
1141 phi = new_Phi(n_ret, cf_pred, get_irn_mode(cf_pred[0]));
1145 /* Conserve Phi-list for further inlinings -- but might be optimized */
1146 if (get_nodes_block(phi) == post_bl) {
1147 set_irn_link(phi, get_irn_link(post_bl));
1148 set_irn_link(post_bl, phi);
1151 set_Tuple_pred(call, pn_Call_T_result, new_Tuple(n_res, res_pred));
1153 set_Tuple_pred(call, pn_Call_T_result, new_Bad());
1155 /* Finally the exception control flow.
1156 We have two (three) possible situations:
1157 First if the Call branches to an exception handler: We need to add a Phi node to
1158 collect the memory containing the exception objects. Further we need
1159 to add another block to get a correct representation of this Phi. To
1160 this block we add a Jmp that resolves into the X output of the Call
1161 when the Call is turned into a tuple.
1162 Second the Call branches to End, the exception is not handled. Just
1163 add all inlined exception branches to the End node.
1164 Third: there is no Exception edge at all. Handle as case two. */
1165 if (exc_handling == 0) {
1167 for (i = 0; i < arity; i++) {
1169 ret = get_irn_n(end_bl, i);
1170 if (is_fragile_op(skip_Proj(ret)) || (get_irn_op(skip_Proj(ret)) == op_Raise)) {
1171 cf_pred[n_exc] = ret;
1176 new_Block(n_exc, cf_pred); /* watch it: current_block is changed! */
1177 set_Tuple_pred(call, pn_Call_X_except, new_Jmp());
1178 /* The Phi for the memories with the exception objects */
1180 for (i = 0; i < arity; i++) {
1182 ret = skip_Proj(get_irn_n(end_bl, i));
1184 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 3);
1186 } else if (is_fragile_op(ret)) {
1187 /* We rely that all cfops have the memory output at the same position. */
1188 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 0);
1190 } else if (get_irn_op(ret) == op_Raise) {
1191 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 1);
1195 set_Tuple_pred(call, pn_Call_M_except, new_Phi(n_exc, cf_pred, mode_M));
1197 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1198 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1201 ir_node *main_end_bl;
1202 int main_end_bl_arity;
1203 ir_node **end_preds;
1205 /* assert(exc_handling == 1 || no exceptions. ) */
1207 for (i = 0; i < arity; i++) {
1208 ir_node *ret = get_irn_n(end_bl, i);
1210 if (is_fragile_op(skip_Proj(ret)) || (get_irn_op(skip_Proj(ret)) == op_Raise)) {
1211 cf_pred[n_exc] = ret;
1215 main_end_bl = get_irg_end_block(current_ir_graph);
1216 main_end_bl_arity = get_irn_arity(main_end_bl);
1217 end_preds = xmalloc ((n_exc + main_end_bl_arity) * sizeof(*end_preds));
1219 for (i = 0; i < main_end_bl_arity; ++i)
1220 end_preds[i] = get_irn_n(main_end_bl, i);
1221 for (i = 0; i < n_exc; ++i)
1222 end_preds[main_end_bl_arity + i] = cf_pred[i];
1223 set_irn_in(main_end_bl, n_exc + main_end_bl_arity, end_preds);
1224 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1225 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1231 /* -- Turn CSE back on. -- */
1232 set_optimize(rem_opt);
1237 /********************************************************************/
1238 /* Apply inlineing to small methods. */
1239 /********************************************************************/
1241 /** Represents a possible inlinable call in a graph. */
1242 typedef struct _call_entry call_entry;
1243 struct _call_entry {
1244 ir_node *call; /**< the Call */
1245 ir_graph *callee; /**< the callee called here */
1246 call_entry *next; /**< for linking the next one */
1250 * environment for inlining small irgs
1252 typedef struct _inline_env_t {
1253 struct obstack obst; /**< an obstack where call_entries are allocated on. */
1254 call_entry *head; /**< the head of the call entry list */
1255 call_entry *tail; /**< the tail of the call entry list */
1259 * Returns the irg called from a Call node. If the irg is not
1260 * known, NULL is returned.
1262 static ir_graph *get_call_called_irg(ir_node *call) {
1264 ir_graph *called_irg = NULL;
1266 addr = get_Call_ptr(call);
1267 if (is_SymConst(addr) && get_SymConst_kind(addr) == symconst_addr_ent) {
1268 called_irg = get_entity_irg(get_SymConst_entity(addr));
1275 * Walker: Collect all calls to known graphs inside a graph.
1277 static void collect_calls(ir_node *call, void *env) {
1278 if (is_Call(call)) {
1279 ir_graph *called_irg = get_call_called_irg(call);
1281 /* The Call node calls a locally defined method. Remember to inline. */
1282 inline_env_t *ienv = env;
1283 call_entry *entry = obstack_alloc(&ienv->obst, sizeof(*entry));
1285 entry->callee = called_irg;
1288 if (ienv->tail == NULL)
1291 ienv->tail->next = entry;
1298 * Inlines all small methods at call sites where the called address comes
1299 * from a Const node that references the entity representing the called
1301 * The size argument is a rough measure for the code size of the method:
1302 * Methods where the obstack containing the firm graph is smaller than
1305 void inline_small_irgs(ir_graph *irg, int size) {
1306 ir_graph *rem = current_ir_graph;
1309 DEBUG_ONLY(firm_dbg_module_t *dbg;)
1311 if (!(get_opt_optimize() && get_opt_inline())) return;
1313 FIRM_DBG_REGISTER(dbg, "firm.opt.inline");
1315 current_ir_graph = irg;
1316 /* Handle graph state */
1317 assert(get_irg_phase_state(irg) != phase_building);
1318 free_callee_info(irg);
1320 /* Find Call nodes to inline.
1321 (We can not inline during a walk of the graph, as inlineing the same
1322 method several times changes the visited flag of the walked graph:
1323 after the first inlineing visited of the callee equals visited of
1324 the caller. With the next inlineing both are increased.) */
1325 obstack_init(&env.obst);
1326 env.head = env.tail = NULL;
1327 irg_walk_graph(irg, NULL, collect_calls, &env);
1329 if (env.head != NULL) {
1330 /* There are calls to inline */
1331 collect_phiprojs(irg);
1332 for (entry = env.head; entry != NULL; entry = entry->next) {
1333 ir_graph *callee = entry->callee;
1334 if (((_obstack_memory_used(callee->obst) - (int)obstack_room(callee->obst)) < size) ||
1335 (get_irg_inline_property(callee) >= irg_inline_forced)) {
1336 inline_method(entry->call, callee);
1340 obstack_free(&env.obst, NULL);
1341 current_ir_graph = rem;
1345 * Environment for inlining irgs.
1348 int n_nodes; /**< Number of nodes in graph except Id, Tuple, Proj, Start, End. */
1349 int n_nodes_orig; /**< for statistics */
1350 call_entry *call_head; /**< The head of the list of all call nodes in this graph. */
1351 call_entry *call_tail; /**< The tail of the list of all call nodes in this graph .*/
1352 int n_call_nodes; /**< Number of Call nodes in the graph. */
1353 int n_call_nodes_orig; /**< for statistics */
1354 int n_callers; /**< Number of known graphs that call this graphs. */
1355 int n_callers_orig; /**< for statistics */
1356 int got_inline; /**< Set, if at leat one call inside this graph was inlined. */
1360 * Allocate a new environment for inlining.
1362 static inline_irg_env *alloc_inline_irg_env(struct obstack *obst) {
1363 inline_irg_env *env = obstack_alloc(obst, sizeof(*env));
1364 env->n_nodes = -2; /* do not count count Start, End */
1365 env->n_nodes_orig = -2; /* do not count Start, End */
1366 env->call_head = NULL;
1367 env->call_tail = NULL;
1368 env->n_call_nodes = 0;
1369 env->n_call_nodes_orig = 0;
1371 env->n_callers_orig = 0;
1372 env->got_inline = 0;
1376 typedef struct walker_env {
1377 struct obstack *obst; /**< the obstack for allocations. */
1378 inline_irg_env *x; /**< the inline environment */
1379 int ignore_runtime; /**< the ignore runtime flag */
1383 * post-walker: collect all calls in the inline-environment
1384 * of a graph and sum some statistics.
1386 static void collect_calls2(ir_node *call, void *ctx) {
1388 inline_irg_env *x = env->x;
1389 ir_op *op = get_irn_op(call);
1393 /* count meaningful nodes in irg */
1394 if (op != op_Proj && op != op_Tuple && op != op_Sync) {
1399 if (op != op_Call) return;
1401 /* check, if it's a runtime call */
1402 if (env->ignore_runtime) {
1403 ir_node *symc = get_Call_ptr(call);
1405 if (is_SymConst(symc) && get_SymConst_kind(symc) == symconst_addr_ent) {
1406 ir_entity *ent = get_SymConst_entity(symc);
1408 if (get_entity_additional_properties(ent) & mtp_property_runtime)
1413 /* collect all call nodes */
1415 ++x->n_call_nodes_orig;
1417 callee = get_call_called_irg(call);
1419 inline_irg_env *callee_env = get_irg_link(callee);
1420 /* count all static callers */
1421 ++callee_env->n_callers;
1422 ++callee_env->n_callers_orig;
1424 /* link it in the list of possible inlinable entries */
1425 entry = obstack_alloc(env->obst, sizeof(*entry));
1427 entry->callee = callee;
1429 if (x->call_tail == NULL)
1430 x->call_head = entry;
1432 x->call_tail->next = entry;
1433 x->call_tail = entry;
1438 * Returns TRUE if the number of callers in 0 in the irg's environment,
1439 * hence this irg is a leave.
1441 INLINE static int is_leave(ir_graph *irg) {
1442 inline_irg_env *env = get_irg_link(irg);
1443 return env->n_call_nodes == 0;
1447 * Returns TRUE if the number of callers is smaller size in the irg's environment.
1449 INLINE static int is_smaller(ir_graph *callee, int size) {
1450 inline_irg_env *env = get_irg_link(callee);
1451 return env->n_nodes < size;
1455 * Append the nodes of the list src to the nodes of the list in environment dst.
1457 static void append_call_list(struct obstack *obst, inline_irg_env *dst, call_entry *src) {
1458 call_entry *entry, *nentry;
1460 /* Note that the src list points to Call nodes in the inlined graph, but
1461 we need Call nodes in our graph. Luckily the inliner leaves this information
1462 in the link field. */
1463 for (entry = src; entry != NULL; entry = entry->next) {
1464 nentry = obstack_alloc(obst, sizeof(*nentry));
1465 nentry->call = get_irn_link(entry->call);
1466 nentry->callee = entry->callee;
1467 nentry->next = NULL;
1468 dst->call_tail->next = nentry;
1469 dst->call_tail = nentry;
1474 * Inlines small leave methods at call sites where the called address comes
1475 * from a Const node that references the entity representing the called
1477 * The size argument is a rough measure for the code size of the method:
1478 * Methods where the obstack containing the firm graph is smaller than
1481 void inline_leave_functions(int maxsize, int leavesize, int size, int ignore_runtime) {
1482 inline_irg_env *env;
1488 call_entry *entry, *tail;
1489 const call_entry *centry;
1490 struct obstack obst;
1491 DEBUG_ONLY(firm_dbg_module_t *dbg;)
1493 if (!(get_opt_optimize() && get_opt_inline())) return;
1495 FIRM_DBG_REGISTER(dbg, "firm.opt.inline");
1496 rem = current_ir_graph;
1497 obstack_init(&obst);
1499 /* extend all irgs by a temporary data structure for inlining. */
1500 n_irgs = get_irp_n_irgs();
1501 for (i = 0; i < n_irgs; ++i)
1502 set_irg_link(get_irp_irg(i), alloc_inline_irg_env(&obst));
1504 /* Precompute information in temporary data structure. */
1506 wenv.ignore_runtime = ignore_runtime;
1507 for (i = 0; i < n_irgs; ++i) {
1508 ir_graph *irg = get_irp_irg(i);
1510 assert(get_irg_phase_state(irg) != phase_building);
1511 free_callee_info(irg);
1513 wenv.x = get_irg_link(irg);
1514 irg_walk_graph(irg, NULL, collect_calls2, &wenv);
1517 /* -- and now inline. -- */
1519 /* Inline leaves recursively -- we might construct new leaves. */
1523 for (i = 0; i < n_irgs; ++i) {
1525 int phiproj_computed = 0;
1527 current_ir_graph = get_irp_irg(i);
1528 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1531 for (entry = env->call_head; entry != NULL; entry = entry->next) {
1534 if (env->n_nodes > maxsize) break;
1537 callee = entry->callee;
1539 if (is_leave(callee) && is_smaller(callee, leavesize)) {
1540 if (!phiproj_computed) {
1541 phiproj_computed = 1;
1542 collect_phiprojs(current_ir_graph);
1544 did_inline = inline_method(call, callee);
1547 /* Do some statistics */
1548 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1550 env->got_inline = 1;
1551 --env->n_call_nodes;
1552 env->n_nodes += callee_env->n_nodes;
1553 --callee_env->n_callers;
1555 /* remove this call from the list */
1557 tail->next = entry->next;
1559 env->call_head = entry->next;
1565 env->call_tail = tail;
1567 } while (did_inline);
1569 /* inline other small functions. */
1570 for (i = 0; i < n_irgs; ++i) {
1572 int phiproj_computed = 0;
1574 current_ir_graph = get_irp_irg(i);
1575 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1577 /* note that the list of possible calls is updated during the process */
1579 for (entry = env->call_head; entry != NULL; entry = entry->next) {
1583 callee = entry->callee;
1585 if (((is_smaller(callee, size) && (env->n_nodes < maxsize)) || /* small function */
1586 (get_irg_inline_property(callee) >= irg_inline_forced))) {
1587 if (!phiproj_computed) {
1588 phiproj_computed = 1;
1589 collect_phiprojs(current_ir_graph);
1591 if (inline_method(call, callee)) {
1592 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1594 /* callee was inline. Append it's call list. */
1595 env->got_inline = 1;
1596 --env->n_call_nodes;
1597 append_call_list(&obst, env, callee_env->call_head);
1598 env->n_call_nodes += callee_env->n_call_nodes;
1599 env->n_nodes += callee_env->n_nodes;
1600 --callee_env->n_callers;
1602 /* after we have inlined callee, all called methods inside callee
1603 are now called once more */
1604 for (centry = callee_env->call_head; centry != NULL; centry = centry->next) {
1605 inline_irg_env *penv = get_irg_link(centry->callee);
1609 /* remove this call from the list */
1611 tail->next = entry->next;
1613 env->call_head = entry->next;
1619 env->call_tail = tail;
1622 for (i = 0; i < n_irgs; ++i) {
1623 irg = get_irp_irg(i);
1624 env = (inline_irg_env *)get_irg_link(irg);
1626 if (env->got_inline) {
1627 /* this irg got calls inlined */
1628 set_irg_outs_inconsistent(irg);
1629 set_irg_doms_inconsistent(irg);
1631 optimize_graph_df(irg);
1634 if (env->got_inline || (env->n_callers_orig != env->n_callers))
1635 DB((dbg, SET_LEVEL_1, "Nodes:%3d ->%3d, calls:%3d ->%3d, callers:%3d ->%3d, -- %s\n",
1636 env->n_nodes_orig, env->n_nodes, env->n_call_nodes_orig, env->n_call_nodes,
1637 env->n_callers_orig, env->n_callers,
1638 get_entity_name(get_irg_entity(irg))));
1641 obstack_free(&obst, NULL);
1642 current_ir_graph = rem;
1645 /*******************************************************************/
1646 /* Code Placement. Pins all floating nodes to a block where they */
1647 /* will be executed only if needed. */
1648 /*******************************************************************/
1651 * Returns non-zero, is a block is not reachable from Start.
1653 * @param block the block to test
1656 is_Block_unreachable(ir_node *block) {
1657 return is_Block_dead(block) || get_Block_dom_depth(block) < 0;
1661 * Find the earliest correct block for node n. --- Place n into the
1662 * same Block as its dominance-deepest Input.
1664 * We have to avoid calls to get_nodes_block() here
1665 * because the graph is floating.
1667 * move_out_of_loops() expects that place_floats_early() have placed
1668 * all "living" nodes into a living block. That's why we must
1669 * move nodes in dead block with "live" successors into a valid
1671 * We move them just into the same block as it's successor (or
1672 * in case of a Phi into the effective use block). For Phi successors,
1673 * this may still be a dead block, but then there is no real use, as
1674 * the control flow will be dead later.
1676 * @param n the node to be placed
1677 * @param worklist a worklist, predecessors of non-floating nodes are placed here
1680 place_floats_early(ir_node *n, waitq *worklist) {
1683 /* we must not run into an infinite loop */
1684 assert(irn_not_visited(n));
1685 mark_irn_visited(n);
1687 /* Place floating nodes. */
1688 if (get_irn_pinned(n) == op_pin_state_floats) {
1689 ir_node *curr_block = get_irn_n(n, -1);
1690 int in_dead_block = is_Block_unreachable(curr_block);
1692 ir_node *b = NULL; /* The block to place this node in */
1694 assert(is_no_Block(n));
1696 if (is_irn_start_block_placed(n)) {
1697 /* These nodes will not be placed by the loop below. */
1698 b = get_irg_start_block(current_ir_graph);
1702 /* find the block for this node. */
1703 irn_arity = get_irn_arity(n);
1704 for (i = 0; i < irn_arity; i++) {
1705 ir_node *pred = get_irn_n(n, i);
1706 ir_node *pred_block;
1708 if ((irn_not_visited(pred))
1709 && (get_irn_pinned(pred) == op_pin_state_floats)) {
1712 * If the current node is NOT in a dead block, but one of its
1713 * predecessors is, we must move the predecessor to a live block.
1714 * Such thing can happen, if global CSE chose a node from a dead block.
1715 * We move it simply to our block.
1716 * Note that neither Phi nor End nodes are floating, so we don't
1717 * need to handle them here.
1719 if (! in_dead_block) {
1720 if (get_irn_pinned(pred) == op_pin_state_floats &&
1721 is_Block_unreachable(get_irn_n(pred, -1)))
1722 set_nodes_block(pred, curr_block);
1724 place_floats_early(pred, worklist);
1728 * A node in the Bad block must stay in the bad block,
1729 * so don't compute a new block for it.
1734 /* Because all loops contain at least one op_pin_state_pinned node, now all
1735 our inputs are either op_pin_state_pinned or place_early() has already
1736 been finished on them. We do not have any unfinished inputs! */
1737 pred_block = get_irn_n(pred, -1);
1738 if ((!is_Block_dead(pred_block)) &&
1739 (get_Block_dom_depth(pred_block) > depth)) {
1741 depth = get_Block_dom_depth(pred_block);
1743 /* Avoid that the node is placed in the Start block */
1744 if ((depth == 1) && (get_Block_dom_depth(get_irn_n(n, -1)) > 1)
1745 && get_irg_phase_state(current_ir_graph) != phase_backend) {
1746 b = get_Block_cfg_out(get_irg_start_block(current_ir_graph), 0);
1747 assert(b != get_irg_start_block(current_ir_graph));
1752 set_nodes_block(n, b);
1756 * Add predecessors of non floating nodes and non-floating predecessors
1757 * of floating nodes to worklist and fix their blocks if the are in dead block.
1759 irn_arity = get_irn_arity(n);
1761 if (get_irn_op(n) == op_End) {
1763 * Simplest case: End node. Predecessors are keep-alives,
1764 * no need to move out of dead block.
1766 for (i = -1; i < irn_arity; ++i) {
1767 ir_node *pred = get_irn_n(n, i);
1768 if (irn_not_visited(pred))
1769 waitq_put(worklist, pred);
1771 } else if (is_Block(n)) {
1773 * Blocks: Predecessors are control flow, no need to move
1774 * them out of dead block.
1776 for (i = irn_arity - 1; i >= 0; --i) {
1777 ir_node *pred = get_irn_n(n, i);
1778 if (irn_not_visited(pred))
1779 waitq_put(worklist, pred);
1781 } else if (is_Phi(n)) {
1783 ir_node *curr_block = get_irn_n(n, -1);
1784 int in_dead_block = is_Block_unreachable(curr_block);
1787 * Phi nodes: move nodes from dead blocks into the effective use
1788 * of the Phi-input if the Phi is not in a bad block.
1790 pred = get_irn_n(n, -1);
1791 if (irn_not_visited(pred))
1792 waitq_put(worklist, pred);
1794 for (i = irn_arity - 1; i >= 0; --i) {
1795 ir_node *pred = get_irn_n(n, i);
1797 if (irn_not_visited(pred)) {
1798 if (! in_dead_block &&
1799 get_irn_pinned(pred) == op_pin_state_floats &&
1800 is_Block_unreachable(get_irn_n(pred, -1))) {
1801 set_nodes_block(pred, get_Block_cfgpred_block(curr_block, i));
1803 waitq_put(worklist, pred);
1808 ir_node *curr_block = get_irn_n(n, -1);
1809 int in_dead_block = is_Block_unreachable(curr_block);
1812 * All other nodes: move nodes from dead blocks into the same block.
1814 pred = get_irn_n(n, -1);
1815 if (irn_not_visited(pred))
1816 waitq_put(worklist, pred);
1818 for (i = irn_arity - 1; i >= 0; --i) {
1819 ir_node *pred = get_irn_n(n, i);
1821 if (irn_not_visited(pred)) {
1822 if (! in_dead_block &&
1823 get_irn_pinned(pred) == op_pin_state_floats &&
1824 is_Block_unreachable(get_irn_n(pred, -1))) {
1825 set_nodes_block(pred, curr_block);
1827 waitq_put(worklist, pred);
1834 * Floating nodes form subgraphs that begin at nodes as Const, Load,
1835 * Start, Call and that end at op_pin_state_pinned nodes as Store, Call. Place_early
1836 * places all floating nodes reachable from its argument through floating
1837 * nodes and adds all beginnings at op_pin_state_pinned nodes to the worklist.
1839 * @param worklist a worklist, used for the algorithm, empty on in/output
1841 static void place_early(waitq *worklist) {
1843 inc_irg_visited(current_ir_graph);
1845 /* this inits the worklist */
1846 place_floats_early(get_irg_end(current_ir_graph), worklist);
1848 /* Work the content of the worklist. */
1849 while (!waitq_empty(worklist)) {
1850 ir_node *n = waitq_get(worklist);
1851 if (irn_not_visited(n))
1852 place_floats_early(n, worklist);
1855 set_irg_outs_inconsistent(current_ir_graph);
1856 set_irg_pinned(current_ir_graph, op_pin_state_pinned);
1860 * Compute the deepest common ancestor of block and dca.
1862 static ir_node *calc_dca(ir_node *dca, ir_node *block) {
1865 /* we do not want to place nodes in dead blocks */
1866 if (is_Block_dead(block))
1869 /* We found a first legal placement. */
1870 if (!dca) return block;
1872 /* Find a placement that is dominates both, dca and block. */
1873 while (get_Block_dom_depth(block) > get_Block_dom_depth(dca))
1874 block = get_Block_idom(block);
1876 while (get_Block_dom_depth(dca) > get_Block_dom_depth(block)) {
1877 dca = get_Block_idom(dca);
1880 while (block != dca) {
1881 block = get_Block_idom(block); dca = get_Block_idom(dca);
1887 /** Deepest common dominance ancestor of DCA and CONSUMER of PRODUCER.
1888 * I.e., DCA is the block where we might place PRODUCER.
1889 * A data flow edge points from producer to consumer.
1892 consumer_dom_dca(ir_node *dca, ir_node *consumer, ir_node *producer) {
1893 ir_node *block = NULL;
1895 /* Compute the latest block into which we can place a node so that it is
1897 if (get_irn_op(consumer) == op_Phi) {
1898 /* our consumer is a Phi-node, the effective use is in all those
1899 blocks through which the Phi-node reaches producer */
1901 ir_node *phi_block = get_nodes_block(consumer);
1902 irn_arity = get_irn_arity(consumer);
1904 for (i = 0; i < irn_arity; i++) {
1905 if (get_irn_n(consumer, i) == producer) {
1906 ir_node *new_block = get_nodes_block(get_Block_cfgpred(phi_block, i));
1908 if (! is_Block_unreachable(new_block))
1909 block = calc_dca(block, new_block);
1914 block = get_irn_n(producer, -1);
1916 assert(is_no_Block(consumer));
1917 block = get_nodes_block(consumer);
1920 /* Compute the deepest common ancestor of block and dca. */
1921 return calc_dca(dca, block);
1924 /* FIXME: the name clashes here with the function from ana/field_temperature.c
1926 static INLINE int get_irn_loop_depth(ir_node *n) {
1927 return get_loop_depth(get_irn_loop(n));
1931 * Move n to a block with less loop depth than it's current block. The
1932 * new block must be dominated by early.
1934 * @param n the node that should be moved
1935 * @param early the earliest block we can n move to
1937 static void move_out_of_loops(ir_node *n, ir_node *early) {
1938 ir_node *best, *dca;
1942 /* Find the region deepest in the dominator tree dominating
1943 dca with the least loop nesting depth, but still dominated
1944 by our early placement. */
1945 dca = get_nodes_block(n);
1948 while (dca != early) {
1949 dca = get_Block_idom(dca);
1950 if (!dca || is_Bad(dca)) break; /* may be Bad if not reachable from Start */
1951 if (get_irn_loop_depth(dca) < get_irn_loop_depth(best)) {
1955 if (best != get_nodes_block(n)) {
1957 printf("Moving out of loop: "); DDMN(n);
1958 printf(" Outermost block: "); DDMN(early);
1959 printf(" Best block: "); DDMN(best);
1960 printf(" Innermost block: "); DDMN(get_nodes_block(n));
1962 set_nodes_block(n, best);
1967 * Find the latest legal block for N and place N into the
1968 * `optimal' Block between the latest and earliest legal block.
1969 * The `optimal' block is the dominance-deepest block of those
1970 * with the least loop-nesting-depth. This places N out of as many
1971 * loops as possible and then makes it as control dependent as
1974 * @param n the node to be placed
1975 * @param worklist a worklist, all successors of non-floating nodes are
1978 static void place_floats_late(ir_node *n, pdeq *worklist) {
1982 assert(irn_not_visited(n)); /* no multiple placement */
1984 mark_irn_visited(n);
1986 /* no need to place block nodes, control nodes are already placed. */
1987 if ((get_irn_op(n) != op_Block) &&
1989 (get_irn_mode(n) != mode_X)) {
1990 /* Remember the early_blk placement of this block to move it
1991 out of loop no further than the early_blk placement. */
1992 early_blk = get_irn_n(n, -1);
1995 * BEWARE: Here we also get code, that is live, but
1996 * was in a dead block. If the node is life, but because
1997 * of CSE in a dead block, we still might need it.
2000 /* Assure that our users are all placed, except the Phi-nodes.
2001 --- Each data flow cycle contains at least one Phi-node. We
2002 have to break the `user has to be placed before the
2003 producer' dependence cycle and the Phi-nodes are the
2004 place to do so, because we need to base our placement on the
2005 final region of our users, which is OK with Phi-nodes, as they
2006 are op_pin_state_pinned, and they never have to be placed after a
2007 producer of one of their inputs in the same block anyway. */
2008 for (i = get_irn_n_outs(n) - 1; i >= 0; --i) {
2009 ir_node *succ = get_irn_out(n, i);
2010 if (irn_not_visited(succ) && (get_irn_op(succ) != op_Phi))
2011 place_floats_late(succ, worklist);
2014 if (! is_Block_dead(early_blk)) {
2015 /* do only move things that where not dead */
2016 ir_op *op = get_irn_op(n);
2018 /* We have to determine the final block of this node... except for
2019 constants and Projs */
2020 if ((get_irn_pinned(n) == op_pin_state_floats) &&
2022 (op != op_SymConst) &&
2025 ir_node *dca = NULL; /* deepest common ancestor in the
2026 dominator tree of all nodes'
2027 blocks depending on us; our final
2028 placement has to dominate DCA. */
2029 for (i = get_irn_n_outs(n) - 1; i >= 0; --i) {
2030 ir_node *succ = get_irn_out(n, i);
2033 if (get_irn_op(succ) == op_End) {
2035 * This consumer is the End node, a keep alive edge.
2036 * This is not a real consumer, so we ignore it
2041 /* ignore if succ is in dead code */
2042 succ_blk = get_irn_n(succ, -1);
2043 if (is_Block_unreachable(succ_blk))
2045 dca = consumer_dom_dca(dca, succ, n);
2048 set_nodes_block(n, dca);
2049 move_out_of_loops(n, early_blk);
2055 /* Add successors of all non-floating nodes on list. (Those of floating
2056 nodes are placed already and therefore are marked.) */
2057 for (i = 0; i < get_irn_n_outs(n); i++) {
2058 ir_node *succ = get_irn_out(n, i);
2059 if (irn_not_visited(get_irn_out(n, i))) {
2060 pdeq_putr(worklist, succ);
2066 * Place floating nodes on the given worklist as late as possible using
2067 * the dominance tree.
2069 * @param worklist the worklist containing the nodes to place
2071 static void place_late(waitq *worklist) {
2073 inc_irg_visited(current_ir_graph);
2075 /* This fills the worklist initially. */
2076 place_floats_late(get_irg_start_block(current_ir_graph), worklist);
2078 /* And now empty the worklist again... */
2079 while (!waitq_empty(worklist)) {
2080 ir_node *n = waitq_get(worklist);
2081 if (irn_not_visited(n))
2082 place_floats_late(n, worklist);
2086 /* Code Placement. */
2087 void place_code(ir_graph *irg) {
2089 ir_graph *rem = current_ir_graph;
2091 current_ir_graph = irg;
2093 /* Handle graph state */
2094 assert(get_irg_phase_state(irg) != phase_building);
2097 if (1 || get_irg_loopinfo_state(irg) != loopinfo_consistent) {
2098 free_loop_information(irg);
2099 construct_backedges(irg);
2102 /* Place all floating nodes as early as possible. This guarantees
2103 a legal code placement. */
2104 worklist = new_waitq();
2105 place_early(worklist);
2107 /* place_early() invalidates the outs, place_late needs them. */
2108 compute_irg_outs(irg);
2110 /* Now move the nodes down in the dominator tree. This reduces the
2111 unnecessary executions of the node. */
2112 place_late(worklist);
2114 set_irg_outs_inconsistent(current_ir_graph);
2115 set_irg_loopinfo_inconsistent(current_ir_graph);
2116 del_waitq(worklist);
2117 current_ir_graph = rem;
2121 * Called by walker of remove_critical_cf_edges().
2123 * Place an empty block to an edge between a blocks of multiple
2124 * predecessors and a block of multiple successors.
2127 * @param env Environment of walker. The changed field.
2129 static void walk_critical_cf_edges(ir_node *n, void *env) {
2131 ir_node *pre, *block, *jmp;
2133 ir_graph *irg = get_irn_irg(n);
2135 /* Block has multiple predecessors */
2136 arity = get_irn_arity(n);
2138 if (n == get_irg_end_block(irg))
2139 return; /* No use to add a block here. */
2141 for (i = 0; i < arity; ++i) {
2144 pre = get_irn_n(n, i);
2145 cfop = get_irn_op(skip_Proj(pre));
2146 /* Predecessor has multiple successors. Insert new control flow edge but
2147 ignore exception edges. */
2148 if (! is_op_fragile(cfop) && is_op_forking(cfop)) {
2149 /* set predecessor of new block */
2150 block = new_r_Block(irg, 1, &pre);
2151 /* insert new jmp node to new block */
2152 jmp = new_r_Jmp(irg, block);
2153 /* set successor of new block */
2154 set_irn_n(n, i, jmp);
2156 } /* predecessor has multiple successors */
2157 } /* for all predecessors */
2158 } /* n is a multi-entry block */
2161 void remove_critical_cf_edges(ir_graph *irg) {
2164 irg_block_walk_graph(irg, NULL, walk_critical_cf_edges, &changed);
2166 /* control flow changed */
2167 set_irg_outs_inconsistent(irg);
2168 set_irg_extblk_inconsistent(irg);
2169 set_irg_doms_inconsistent(irg);
2170 set_irg_loopinfo_inconsistent(irg);