2 * Copyright (C) 1995-2008 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 Dead node elimination and Procedure Inlining.
23 * @author Michael Beck, Goetz Lindenmaier
34 #include "irgraph_t.h"
37 #include "iroptimize.h"
44 #include "adt/array.h"
48 #include "adt/xmalloc.h"
52 #include "irbackedge_t.h"
53 #include "opt_inline_t.h"
58 #include "analyze_irg_args.h"
59 #include "iredges_t.h"
64 DEBUG_ONLY(static firm_dbg_module_t *dbg;)
66 /*------------------------------------------------------------------*/
67 /* Routines for dead node elimination / copying garbage collection */
69 /*------------------------------------------------------------------*/
72 * Remember the new node in the old node by using a field all nodes have.
74 #define set_new_node(oldn, newn) set_irn_link(oldn, newn)
77 * Get this new node, before the old node is forgotten.
79 #define get_new_node(oldn) get_irn_link(oldn)
82 * Check if a new node was set.
84 #define has_new_node(n) (get_new_node(n) != NULL)
87 * We use the block_visited flag to mark that we have computed the
88 * number of useful predecessors for this block.
89 * Further we encode the new arity in this flag in the old blocks.
90 * Remembering the arity is useful, as it saves a lot of pointer
91 * accesses. This function is called for all Phi and Block nodes
95 compute_new_arity(ir_node *b) {
96 int i, res, irn_arity;
99 irg_v = get_irg_block_visited(current_ir_graph);
100 block_v = get_Block_block_visited(b);
101 if (block_v >= irg_v) {
102 /* we computed the number of preds for this block and saved it in the
104 return block_v - irg_v;
106 /* compute the number of good predecessors */
107 res = irn_arity = get_irn_arity(b);
108 for (i = 0; i < irn_arity; i++)
109 if (is_Bad(get_irn_n(b, i))) res--;
110 /* save it in the flag. */
111 set_Block_block_visited(b, irg_v + res);
117 * Copies the node to the new obstack. The Ins of the new node point to
118 * the predecessors on the old obstack. For block/phi nodes not all
119 * predecessors might be copied. n->link points to the new node.
120 * For Phi and Block nodes the function allocates in-arrays with an arity
121 * only for useful predecessors. The arity is determined by counting
122 * the non-bad predecessors of the block.
124 * @param n The node to be copied
125 * @param env if non-NULL, the node number attribute will be copied to the new node
127 * Note: Also used for loop unrolling.
129 static void copy_node(ir_node *n, void *env) {
132 ir_op *op = get_irn_op(n);
136 /* node copied already */
138 } else if (op == op_Block) {
140 new_arity = compute_new_arity(n);
141 n->attr.block.graph_arr = NULL;
143 block = get_nodes_block(n);
145 new_arity = compute_new_arity(block);
147 new_arity = get_irn_arity(n);
150 nn = new_ir_node(get_irn_dbg_info(n),
157 /* Copy the attributes. These might point to additional data. If this
158 was allocated on the old obstack the pointers now are dangling. This
159 frees e.g. the memory of the graph_arr allocated in new_immBlock. */
160 if (op == op_Block) {
161 /* we cannot allow blocks WITHOUT macroblock input */
162 set_Block_MacroBlock(nn, get_Block_MacroBlock(n));
164 copy_node_attr(n, nn);
168 int copy_node_nr = env != NULL;
170 /* for easier debugging, we want to copy the node numbers too */
171 nn->node_nr = n->node_nr;
177 hook_dead_node_elim_subst(current_ir_graph, n, nn);
181 * Copies new predecessors of old node to new node remembered in link.
182 * Spare the Bad predecessors of Phi and Block nodes.
184 static void copy_preds(ir_node *n, void *env) {
189 nn = get_new_node(n);
192 /* copy the macro block header */
193 ir_node *mbh = get_Block_MacroBlock(n);
196 /* this block is a macroblock header */
197 set_Block_MacroBlock(nn, nn);
199 /* get the macro block header */
200 ir_node *nmbh = get_new_node(mbh);
201 assert(nmbh != NULL);
202 set_Block_MacroBlock(nn, nmbh);
205 /* Don't copy Bad nodes. */
207 irn_arity = get_irn_arity(n);
208 for (i = 0; i < irn_arity; i++) {
209 if (! is_Bad(get_irn_n(n, i))) {
210 ir_node *pred = get_irn_n(n, i);
211 set_irn_n(nn, j, get_new_node(pred));
215 /* repair the block visited flag from above misuse. Repair it in both
216 graphs so that the old one can still be used. */
217 set_Block_block_visited(nn, 0);
218 set_Block_block_visited(n, 0);
219 /* Local optimization could not merge two subsequent blocks if
220 in array contained Bads. Now it's possible.
221 We don't call optimize_in_place as it requires
222 that the fields in ir_graph are set properly. */
223 if ((get_opt_control_flow_straightening()) &&
224 (get_Block_n_cfgpreds(nn) == 1) &&
225 is_Jmp(get_Block_cfgpred(nn, 0))) {
226 ir_node *old = get_nodes_block(get_Block_cfgpred(nn, 0));
228 /* Jmp jumps into the block it is in -- deal self cycle. */
229 assert(is_Bad(get_new_node(get_irg_bad(current_ir_graph))));
230 exchange(nn, get_new_node(get_irg_bad(current_ir_graph)));
235 } else if (is_Phi(n) && get_irn_arity(n) > 0) {
236 /* Don't copy node if corresponding predecessor in block is Bad.
237 The Block itself should not be Bad. */
238 block = get_nodes_block(n);
239 set_nodes_block(nn, get_new_node(block));
241 irn_arity = get_irn_arity(n);
242 for (i = 0; i < irn_arity; i++) {
243 if (! is_Bad(get_irn_n(block, i))) {
244 ir_node *pred = get_irn_n(n, i);
245 set_irn_n(nn, j, get_new_node(pred));
246 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
250 /* If the pre walker reached this Phi after the post walker visited the
251 block block_visited is > 0. */
252 set_Block_block_visited(get_nodes_block(n), 0);
253 /* Compacting the Phi's ins might generate Phis with only one
255 if (get_irn_arity(nn) == 1)
256 exchange(nn, get_irn_n(nn, 0));
258 irn_arity = get_irn_arity(n);
259 for (i = -1; i < irn_arity; i++)
260 set_irn_n(nn, i, get_new_node(get_irn_n(n, i)));
262 /* Now the new node is complete. We can add it to the hash table for CSE.
263 @@@ inlining aborts if we identify End. Why? */
265 add_identities(current_ir_graph->value_table, nn);
269 * Copies the graph recursively, compacts the keep-alives of the end node.
271 * @param irg the graph to be copied
272 * @param copy_node_nr If non-zero, the node number will be copied
274 static void copy_graph(ir_graph *irg, int copy_node_nr) {
275 ir_node *oe, *ne, *ob, *nb, *om, *nm; /* old end, new end, old bad, new bad, old NoMem, new NoMem */
276 ir_node *ka; /* keep alive */
280 /* Some nodes must be copied by hand, sigh */
281 vfl = get_irg_visited(irg);
282 set_irg_visited(irg, vfl + 1);
284 oe = get_irg_end(irg);
285 mark_irn_visited(oe);
286 /* copy the end node by hand, allocate dynamic in array! */
287 ne = new_ir_node(get_irn_dbg_info(oe),
294 /* Copy the attributes. Well, there might be some in the future... */
295 copy_node_attr(oe, ne);
296 set_new_node(oe, ne);
298 /* copy the Bad node */
299 ob = get_irg_bad(irg);
300 mark_irn_visited(ob);
301 nb = new_ir_node(get_irn_dbg_info(ob),
308 copy_node_attr(ob, nb);
309 set_new_node(ob, nb);
311 /* copy the NoMem node */
312 om = get_irg_no_mem(irg);
313 mark_irn_visited(om);
314 nm = new_ir_node(get_irn_dbg_info(om),
321 copy_node_attr(om, nm);
322 set_new_node(om, nm);
324 /* copy the live nodes */
325 set_irg_visited(irg, vfl);
326 irg_walk(get_nodes_block(oe), copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
328 /* Note: from yet, the visited flag of the graph is equal to vfl + 1 */
330 /* visit the anchors as well */
331 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
332 ir_node *n = get_irg_anchor(irg, i);
334 if (n && (get_irn_visited(n) <= vfl)) {
335 set_irg_visited(irg, vfl);
336 irg_walk(n, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
340 /* copy_preds for the end node ... */
341 set_nodes_block(ne, get_new_node(get_nodes_block(oe)));
343 /*- ... and now the keep alives. -*/
344 /* First pick the not marked block nodes and walk them. We must pick these
345 first as else we will oversee blocks reachable from Phis. */
346 irn_arity = get_End_n_keepalives(oe);
347 for (i = 0; i < irn_arity; i++) {
348 ka = get_End_keepalive(oe, i);
350 if (get_irn_visited(ka) <= vfl) {
351 /* We must keep the block alive and copy everything reachable */
352 set_irg_visited(irg, vfl);
353 irg_walk(ka, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
355 add_End_keepalive(ne, get_new_node(ka));
359 /* Now pick other nodes. Here we will keep all! */
360 irn_arity = get_End_n_keepalives(oe);
361 for (i = 0; i < irn_arity; i++) {
362 ka = get_End_keepalive(oe, i);
364 if (get_irn_visited(ka) <= vfl) {
365 /* We didn't copy the node yet. */
366 set_irg_visited(irg, vfl);
367 irg_walk(ka, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
369 add_End_keepalive(ne, get_new_node(ka));
373 /* start block sometimes only reached after keep alives */
374 set_nodes_block(nb, get_new_node(get_nodes_block(ob)));
375 set_nodes_block(nm, get_new_node(get_nodes_block(om)));
379 * Copies the graph reachable from current_ir_graph->end to the obstack
380 * in current_ir_graph and fixes the environment.
381 * Then fixes the fields in current_ir_graph containing nodes of the
384 * @param copy_node_nr If non-zero, the node number will be copied
387 copy_graph_env(int copy_node_nr) {
388 ir_graph *irg = current_ir_graph;
389 ir_node *old_end, *new_anchor;
392 /* remove end_except and end_reg nodes */
393 old_end = get_irg_end(irg);
394 set_irg_end_except (irg, old_end);
395 set_irg_end_reg (irg, old_end);
397 /* Not all nodes remembered in irg might be reachable
398 from the end node. Assure their link is set to NULL, so that
399 we can test whether new nodes have been computed. */
400 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
401 ir_node *n = get_irg_anchor(irg, i);
403 set_new_node(n, NULL);
405 /* we use the block walk flag for removing Bads from Blocks ins. */
406 inc_irg_block_visited(irg);
409 copy_graph(irg, copy_node_nr);
412 old_end = get_irg_end(irg);
413 new_anchor = new_Anchor(irg);
415 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
416 ir_node *n = get_irg_anchor(irg, i);
418 set_irn_n(new_anchor, i, get_new_node(n));
421 irg->anchor = new_anchor;
423 /* ensure the new anchor is placed in the endblock */
424 set_nodes_block(new_anchor, get_irg_end_block(irg));
428 * Copies all reachable nodes to a new obstack. Removes bad inputs
429 * from block nodes and the corresponding inputs from Phi nodes.
430 * Merges single exit blocks with single entry blocks and removes
432 * Adds all new nodes to a new hash table for CSE. Does not
433 * perform CSE, so the hash table might contain common subexpressions.
435 void dead_node_elimination(ir_graph *irg) {
437 #ifdef INTERPROCEDURAL_VIEW
438 int rem_ipview = get_interprocedural_view();
440 struct obstack *graveyard_obst = NULL;
441 struct obstack *rebirth_obst = NULL;
442 assert(! edges_activated(irg) && "dead node elimination requires disabled edges");
444 /* inform statistics that we started a dead-node elimination run */
445 hook_dead_node_elim(irg, 1);
447 /* Remember external state of current_ir_graph. */
448 rem = current_ir_graph;
449 current_ir_graph = irg;
450 #ifdef INTERPROCEDURAL_VIEW
451 set_interprocedural_view(0);
454 assert(get_irg_phase_state(irg) != phase_building);
456 /* Handle graph state */
457 free_callee_info(irg);
461 /* @@@ so far we loose loops when copying */
462 free_loop_information(irg);
464 set_irg_doms_inconsistent(irg);
466 /* A quiet place, where the old obstack can rest in peace,
467 until it will be cremated. */
468 graveyard_obst = irg->obst;
470 /* A new obstack, where the reachable nodes will be copied to. */
471 rebirth_obst = xmalloc(sizeof(*rebirth_obst));
472 irg->obst = rebirth_obst;
473 obstack_init(irg->obst);
474 irg->last_node_idx = 0;
476 /* We also need a new value table for CSE */
477 del_identities(irg->value_table);
478 irg->value_table = new_identities();
480 /* Copy the graph from the old to the new obstack */
481 copy_graph_env(/*copy_node_nr=*/1);
483 /* Free memory from old unoptimized obstack */
484 obstack_free(graveyard_obst, 0); /* First empty the obstack ... */
485 xfree(graveyard_obst); /* ... then free it. */
487 /* inform statistics that the run is over */
488 hook_dead_node_elim(irg, 0);
490 current_ir_graph = rem;
491 #ifdef INTERPROCEDURAL_VIEW
492 set_interprocedural_view(rem_ipview);
497 * Relink bad predecessors of a block and store the old in array to the
498 * link field. This function is called by relink_bad_predecessors().
499 * The array of link field starts with the block operand at position 0.
500 * If block has bad predecessors, create a new in array without bad preds.
501 * Otherwise let in array untouched.
503 static void relink_bad_block_predecessors(ir_node *n, void *env) {
504 ir_node **new_in, *irn;
505 int i, new_irn_n, old_irn_arity, new_irn_arity = 0;
508 /* if link field of block is NULL, look for bad predecessors otherwise
509 this is already done */
510 if (is_Block(n) && get_irn_link(n) == NULL) {
511 /* save old predecessors in link field (position 0 is the block operand)*/
512 set_irn_link(n, get_irn_in(n));
514 /* count predecessors without bad nodes */
515 old_irn_arity = get_irn_arity(n);
516 for (i = 0; i < old_irn_arity; i++)
517 if (!is_Bad(get_irn_n(n, i)))
520 /* arity changing: set new predecessors without bad nodes */
521 if (new_irn_arity < old_irn_arity) {
522 /* Get new predecessor array. We do not resize the array, as we must
523 keep the old one to update Phis. */
524 new_in = NEW_ARR_D(ir_node *, current_ir_graph->obst, (new_irn_arity+1));
526 /* set new predecessors in array */
529 for (i = 0; i < old_irn_arity; i++) {
530 irn = get_irn_n(n, i);
532 new_in[new_irn_n] = irn;
533 is_backedge(n, i) ? set_backedge(n, new_irn_n-1) : set_not_backedge(n, new_irn_n-1);
537 /* ARR_SETLEN(int, n->attr.block.backedge, new_irn_arity); */
538 ARR_SHRINKLEN(n->attr.block.backedge, new_irn_arity);
540 } /* ir node has bad predecessors */
541 } /* Block is not relinked */
545 * Relinks Bad predecessors from Blocks and Phis called by walker
546 * remove_bad_predecesors(). If n is a Block, call
547 * relink_bad_block_redecessors(). If n is a Phi-node, call also the relinking
548 * function of Phi's Block. If this block has bad predecessors, relink preds
551 static void relink_bad_predecessors(ir_node *n, void *env) {
552 ir_node *block, **old_in;
553 int i, old_irn_arity, new_irn_arity;
555 /* relink bad predecessors of a block */
557 relink_bad_block_predecessors(n, env);
559 /* If Phi node relink its block and its predecessors */
561 /* Relink predecessors of phi's block */
562 block = get_nodes_block(n);
563 if (get_irn_link(block) == NULL)
564 relink_bad_block_predecessors(block, env);
566 old_in = (ir_node **)get_irn_link(block); /* Of Phi's Block */
567 old_irn_arity = ARR_LEN(old_in);
569 /* Relink Phi predecessors if count of predecessors changed */
570 if (old_irn_arity != ARR_LEN(get_irn_in(block))) {
571 /* set new predecessors in array
572 n->in[0] remains the same block */
574 for(i = 1; i < old_irn_arity; i++)
575 if (!is_Bad(old_in[i])) {
576 n->in[new_irn_arity] = n->in[i];
577 is_backedge(n, i) ? set_backedge(n, new_irn_arity) : set_not_backedge(n, new_irn_arity);
581 ARR_SETLEN(ir_node *, n->in, new_irn_arity);
582 ARR_SETLEN(int, n->attr.phi.u.backedge, new_irn_arity);
584 } /* n is a Phi node */
588 * Removes Bad Bad predecessors from Blocks and the corresponding
589 * inputs to Phi nodes as in dead_node_elimination but without
591 * On walking up set the link field to NULL, on walking down call
592 * relink_bad_predecessors() (This function stores the old in array
593 * to the link field and sets a new in array if arity of predecessors
596 void remove_bad_predecessors(ir_graph *irg) {
597 panic("Fix backedge handling first");
598 irg_walk_graph(irg, firm_clear_link, relink_bad_predecessors, NULL);
605 __)|_| | \_/ | \_/(/_ |_/\__|__
607 The following stuff implements a facility that automatically patches
608 registered ir_node pointers to the new node when a dead node elimination occurs.
611 struct _survive_dce_t {
615 hook_entry_t dead_node_elim;
616 hook_entry_t dead_node_elim_subst;
619 typedef struct _survive_dce_list_t {
620 struct _survive_dce_list_t *next;
622 } survive_dce_list_t;
624 static void dead_node_hook(void *context, ir_graph *irg, int start) {
625 survive_dce_t *sd = context;
628 /* Create a new map before the dead node elimination is performed. */
630 sd->new_places = pmap_create_ex(pmap_count(sd->places));
632 /* Patch back all nodes if dead node elimination is over and something is to be done. */
633 pmap_destroy(sd->places);
634 sd->places = sd->new_places;
635 sd->new_places = NULL;
640 * Hook called when dead node elimination replaces old by nw.
642 static void dead_node_subst_hook(void *context, ir_graph *irg, ir_node *old, ir_node *nw) {
643 survive_dce_t *sd = context;
644 survive_dce_list_t *list = pmap_get(sd->places, old);
647 /* If the node is to be patched back, write the new address to all registered locations. */
649 survive_dce_list_t *p;
651 for (p = list; p; p = p->next)
654 pmap_insert(sd->new_places, nw, list);
659 * Make a new Survive DCE environment.
661 survive_dce_t *new_survive_dce(void) {
662 survive_dce_t *res = xmalloc(sizeof(res[0]));
663 obstack_init(&res->obst);
664 res->places = pmap_create();
665 res->new_places = NULL;
667 res->dead_node_elim.hook._hook_dead_node_elim = dead_node_hook;
668 res->dead_node_elim.context = res;
669 res->dead_node_elim.next = NULL;
671 res->dead_node_elim_subst.hook._hook_dead_node_elim_subst = dead_node_subst_hook;
672 res->dead_node_elim_subst.context = res;
673 res->dead_node_elim_subst.next = NULL;
675 #ifndef FIRM_ENABLE_HOOKS
676 assert(0 && "need hooks enabled");
679 register_hook(hook_dead_node_elim, &res->dead_node_elim);
680 register_hook(hook_dead_node_elim_subst, &res->dead_node_elim_subst);
685 * Free a Survive DCE environment.
687 void free_survive_dce(survive_dce_t *sd) {
688 obstack_free(&sd->obst, NULL);
689 pmap_destroy(sd->places);
690 unregister_hook(hook_dead_node_elim, &sd->dead_node_elim);
691 unregister_hook(hook_dead_node_elim_subst, &sd->dead_node_elim_subst);
696 * Register a node pointer to be patched upon DCE.
697 * When DCE occurs, the node pointer specified by @p place will be
698 * patched to the new address of the node it is pointing to.
700 * @param sd The Survive DCE environment.
701 * @param place The address of the node pointer.
703 void survive_dce_register_irn(survive_dce_t *sd, ir_node **place) {
704 if (*place != NULL) {
705 ir_node *irn = *place;
706 survive_dce_list_t *curr = pmap_get(sd->places, irn);
707 survive_dce_list_t *nw = obstack_alloc(&sd->obst, sizeof(nw[0]));
712 pmap_insert(sd->places, irn, nw);
716 /*--------------------------------------------------------------------*/
717 /* Functionality for inlining */
718 /*--------------------------------------------------------------------*/
721 * Copy node for inlineing. Updates attributes that change when
722 * inlineing but not for dead node elimination.
724 * Copies the node by calling copy_node() and then updates the entity if
725 * it's a local one. env must be a pointer of the frame type of the
726 * inlined procedure. The new entities must be in the link field of
730 copy_node_inline(ir_node *n, void *env) {
732 ir_type *frame_tp = (ir_type *)env;
736 nn = get_new_node (n);
738 if (get_entity_owner(get_Sel_entity(n)) == frame_tp) {
739 set_Sel_entity(nn, get_entity_link(get_Sel_entity(n)));
741 } else if (is_Block(n)) {
742 nn = get_new_node (n);
743 nn->attr.block.irg = current_ir_graph;
748 * Walker: checks if P_value_arg_base is used.
750 static void find_addr(ir_node *node, void *env) {
751 int *allow_inline = env;
753 is_Start(get_Proj_pred(node)) &&
754 get_Proj_proj(node) == pn_Start_P_value_arg_base) {
760 * Check if we can inline a given call.
761 * Currently, we cannot inline two cases:
762 * - call with compound arguments
763 * - graphs that take the address of a parameter
765 * check these conditions here
767 static int can_inline(ir_node *call, ir_graph *called_graph) {
768 ir_type *call_type = get_Call_type(call);
769 int params, ress, i, res;
770 assert(is_Method_type(call_type));
772 params = get_method_n_params(call_type);
773 ress = get_method_n_ress(call_type);
775 /* check parameters for compound arguments */
776 for (i = 0; i < params; ++i) {
777 ir_type *p_type = get_method_param_type(call_type, i);
779 if (is_compound_type(p_type))
783 /* check results for compound arguments */
784 for (i = 0; i < ress; ++i) {
785 ir_type *r_type = get_method_res_type(call_type, i);
787 if (is_compound_type(r_type))
792 irg_walk_graph(called_graph, find_addr, NULL, &res);
798 exc_handler = 0, /**< There is a handler. */
799 exc_to_end = 1, /**< Branches to End. */
800 exc_no_handler = 2 /**< Exception handling not represented. */
803 /* Inlines a method at the given call site. */
804 int inline_method(ir_node *call, ir_graph *called_graph) {
806 ir_node *post_call, *post_bl;
807 ir_node *in[pn_Start_max];
808 ir_node *end, *end_bl;
812 int arity, n_ret, n_exc, n_res, i, n, j, rem_opt, irn_arity;
813 enum exc_mode exc_handling;
814 ir_type *called_frame, *curr_frame;
815 irg_inline_property prop = get_irg_inline_property(called_graph);
818 if (prop == irg_inline_forbidden)
821 ent = get_irg_entity(called_graph);
823 /* Do not inline variadic functions. */
824 if (get_method_variadicity(get_entity_type(ent)) == variadicity_variadic)
827 assert(get_method_n_params(get_entity_type(ent)) ==
828 get_method_n_params(get_Call_type(call)));
831 * We cannot inline a recursive call. The graph must be copied before
832 * the call the inline_method() using create_irg_copy().
834 if (called_graph == current_ir_graph)
838 * currently, we cannot inline two cases:
839 * - call with compound arguments
840 * - graphs that take the address of a parameter
842 if (! can_inline(call, called_graph))
845 DB((dbg, SET_LEVEL_1, "Inlining %+F(%+F) into %+F\n", call, called_graph, current_ir_graph));
847 /* -- Turn off optimizations, this can cause problems when allocating new nodes. -- */
848 rem_opt = get_opt_optimize();
851 /* Handle graph state */
852 assert(get_irg_phase_state(current_ir_graph) != phase_building);
853 assert(get_irg_pinned(current_ir_graph) == op_pin_state_pinned);
854 assert(get_irg_pinned(called_graph) == op_pin_state_pinned);
855 set_irg_outs_inconsistent(current_ir_graph);
856 set_irg_extblk_inconsistent(current_ir_graph);
857 set_irg_doms_inconsistent(current_ir_graph);
858 set_irg_loopinfo_inconsistent(current_ir_graph);
859 set_irg_callee_info_state(current_ir_graph, irg_callee_info_inconsistent);
861 /* -- Check preconditions -- */
862 assert(is_Call(call));
864 /* here we know we WILL inline, so inform the statistics */
865 hook_inline(call, called_graph);
867 /* -- Decide how to handle exception control flow: Is there a handler
868 for the Call node, or do we branch directly to End on an exception?
870 0 There is a handler.
872 2 Exception handling not represented in Firm. -- */
874 ir_node *proj, *Mproj = NULL, *Xproj = NULL;
875 for (proj = get_irn_link(call); proj; proj = get_irn_link(proj)) {
876 long proj_nr = get_Proj_proj(proj);
877 if (proj_nr == pn_Call_X_except) Xproj = proj;
878 if (proj_nr == pn_Call_M_except) Mproj = proj;
880 if (Mproj) { assert(Xproj); exc_handling = exc_handler; } /* Mproj */
881 else if (Xproj) { exc_handling = exc_to_end; } /* !Mproj && Xproj */
882 else { exc_handling = exc_no_handler; } /* !Mproj && !Xproj */
886 the procedure and later replaces the Start node of the called graph.
887 Post_call is the old Call node and collects the results of the called
888 graph. Both will end up being a tuple. -- */
889 post_bl = get_nodes_block(call);
890 set_irg_current_block(current_ir_graph, post_bl);
891 /* XxMxPxPxPxT of Start + parameter of Call */
892 in[pn_Start_X_initial_exec] = new_Jmp();
893 in[pn_Start_M] = get_Call_mem(call);
894 in[pn_Start_P_frame_base] = get_irg_frame(current_ir_graph);
895 in[pn_Start_P_globals] = get_irg_globals(current_ir_graph);
896 in[pn_Start_P_tls] = get_irg_tls(current_ir_graph);
897 in[pn_Start_T_args] = new_Tuple(get_Call_n_params(call), get_Call_param_arr(call));
898 /* in[pn_Start_P_value_arg_base] = ??? */
899 assert(pn_Start_P_value_arg_base == pn_Start_max - 1 && "pn_Start_P_value_arg_base not supported, fix");
900 pre_call = new_Tuple(pn_Start_max - 1, in);
904 The new block gets the ins of the old block, pre_call and all its
905 predecessors and all Phi nodes. -- */
906 part_block(pre_call);
908 /* -- Prepare state for dead node elimination -- */
909 /* Visited flags in calling irg must be >= flag in called irg.
910 Else walker and arity computation will not work. */
911 if (get_irg_visited(current_ir_graph) <= get_irg_visited(called_graph))
912 set_irg_visited(current_ir_graph, get_irg_visited(called_graph)+1);
913 if (get_irg_block_visited(current_ir_graph)< get_irg_block_visited(called_graph))
914 set_irg_block_visited(current_ir_graph, get_irg_block_visited(called_graph));
915 /* Set pre_call as new Start node in link field of the start node of
916 calling graph and pre_calls block as new block for the start block
918 Further mark these nodes so that they are not visited by the
920 set_irn_link(get_irg_start(called_graph), pre_call);
921 set_irn_visited(get_irg_start(called_graph), get_irg_visited(current_ir_graph));
922 set_irn_link(get_irg_start_block(called_graph), get_nodes_block(pre_call));
923 set_irn_visited(get_irg_start_block(called_graph), get_irg_visited(current_ir_graph));
924 set_irn_link(get_irg_bad(called_graph), get_irg_bad(current_ir_graph));
925 set_irn_visited(get_irg_bad(called_graph), get_irg_visited(current_ir_graph));
927 /* Initialize for compaction of in arrays */
928 inc_irg_block_visited(current_ir_graph);
930 /* -- Replicate local entities of the called_graph -- */
931 /* copy the entities. */
932 called_frame = get_irg_frame_type(called_graph);
933 curr_frame = get_irg_frame_type(current_ir_graph);
934 for (i = 0, n = get_class_n_members(called_frame); i < n; ++i) {
935 ir_entity *new_ent, *old_ent;
936 old_ent = get_class_member(called_frame, i);
937 new_ent = copy_entity_own(old_ent, curr_frame);
938 set_entity_link(old_ent, new_ent);
941 /* visited is > than that of called graph. With this trick visited will
942 remain unchanged so that an outer walker, e.g., searching the call nodes
943 to inline, calling this inline will not visit the inlined nodes. */
944 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
946 /* -- Performing dead node elimination inlines the graph -- */
947 /* Copies the nodes to the obstack of current_ir_graph. Updates links to new
949 irg_walk(get_irg_end(called_graph), copy_node_inline, copy_preds,
950 get_irg_frame_type(called_graph));
952 /* Repair called_graph */
953 set_irg_visited(called_graph, get_irg_visited(current_ir_graph));
954 set_irg_block_visited(called_graph, get_irg_block_visited(current_ir_graph));
955 set_Block_block_visited(get_irg_start_block(called_graph), 0);
957 /* -- Merge the end of the inlined procedure with the call site -- */
958 /* We will turn the old Call node into a Tuple with the following
961 0: Phi of all Memories of Return statements.
962 1: Jmp from new Block that merges the control flow from all exception
963 predecessors of the old end block.
964 2: Tuple of all arguments.
965 3: Phi of Exception memories.
966 In case the old Call directly branches to End on an exception we don't
967 need the block merging all exceptions nor the Phi of the exception
971 /* -- Precompute some values -- */
972 end_bl = get_new_node(get_irg_end_block(called_graph));
973 end = get_new_node(get_irg_end(called_graph));
974 arity = get_irn_arity(end_bl); /* arity = n_exc + n_ret */
975 n_res = get_method_n_ress(get_Call_type(call));
977 res_pred = xmalloc(n_res * sizeof(*res_pred));
978 cf_pred = xmalloc(arity * sizeof(*res_pred));
980 set_irg_current_block(current_ir_graph, post_bl); /* just to make sure */
982 /* -- archive keepalives -- */
983 irn_arity = get_irn_arity(end);
984 for (i = 0; i < irn_arity; i++) {
985 ir_node *ka = get_End_keepalive(end, i);
987 add_End_keepalive(get_irg_end(current_ir_graph), ka);
990 /* The new end node will die. We need not free as the in array is on the obstack:
991 copy_node() only generated 'D' arrays. */
993 /* -- Replace Return nodes by Jump nodes. -- */
995 for (i = 0; i < arity; i++) {
997 ret = get_irn_n(end_bl, i);
998 if (is_Return(ret)) {
999 cf_pred[n_ret] = new_r_Jmp(current_ir_graph, get_nodes_block(ret));
1003 set_irn_in(post_bl, n_ret, cf_pred);
1005 /* -- Build a Tuple for all results of the method.
1006 Add Phi node if there was more than one Return. -- */
1007 turn_into_tuple(post_call, pn_Call_max);
1008 /* First the Memory-Phi */
1010 for (i = 0; i < arity; i++) {
1011 ret = get_irn_n(end_bl, i);
1012 if (is_Return(ret)) {
1013 cf_pred[n_ret] = get_Return_mem(ret);
1017 phi = new_Phi(n_ret, cf_pred, mode_M);
1018 set_Tuple_pred(call, pn_Call_M_regular, phi);
1019 /* Conserve Phi-list for further inlinings -- but might be optimized */
1020 if (get_nodes_block(phi) == post_bl) {
1021 set_irn_link(phi, get_irn_link(post_bl));
1022 set_irn_link(post_bl, phi);
1024 /* Now the real results */
1026 for (j = 0; j < n_res; j++) {
1028 for (i = 0; i < arity; i++) {
1029 ret = get_irn_n(end_bl, i);
1030 if (is_Return(ret)) {
1031 cf_pred[n_ret] = get_Return_res(ret, j);
1036 phi = new_Phi(n_ret, cf_pred, get_irn_mode(cf_pred[0]));
1040 /* Conserve Phi-list for further inlinings -- but might be optimized */
1041 if (get_nodes_block(phi) == post_bl) {
1042 set_Phi_next(phi, get_Block_phis(post_bl));
1043 set_Block_phis(post_bl, phi);
1046 set_Tuple_pred(call, pn_Call_T_result, new_Tuple(n_res, res_pred));
1048 set_Tuple_pred(call, pn_Call_T_result, new_Bad());
1050 /* handle the regular call */
1051 set_Tuple_pred(call, pn_Call_X_regular, new_Jmp());
1053 /* For now, we cannot inline calls with value_base */
1054 set_Tuple_pred(call, pn_Call_P_value_res_base, new_Bad());
1056 /* Finally the exception control flow.
1057 We have two (three) possible situations:
1058 First if the Call branches to an exception handler: We need to add a Phi node to
1059 collect the memory containing the exception objects. Further we need
1060 to add another block to get a correct representation of this Phi. To
1061 this block we add a Jmp that resolves into the X output of the Call
1062 when the Call is turned into a tuple.
1063 Second the Call branches to End, the exception is not handled. Just
1064 add all inlined exception branches to the End node.
1065 Third: there is no Exception edge at all. Handle as case two. */
1066 if (exc_handling == exc_handler) {
1068 for (i = 0; i < arity; i++) {
1070 ret = get_irn_n(end_bl, i);
1071 irn = skip_Proj(ret);
1072 if (is_fragile_op(irn) || is_Raise(irn)) {
1073 cf_pred[n_exc] = ret;
1078 new_Block(n_exc, cf_pred); /* watch it: current_block is changed! */
1079 set_Tuple_pred(call, pn_Call_X_except, new_Jmp());
1080 /* The Phi for the memories with the exception objects */
1082 for (i = 0; i < arity; i++) {
1084 ret = skip_Proj(get_irn_n(end_bl, i));
1086 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 3);
1088 } else if (is_fragile_op(ret)) {
1089 /* We rely that all cfops have the memory output at the same position. */
1090 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 0);
1092 } else if (is_Raise(ret)) {
1093 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_block(ret), ret, mode_M, 1);
1097 set_Tuple_pred(call, pn_Call_M_except, new_Phi(n_exc, cf_pred, mode_M));
1099 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1100 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1103 ir_node *main_end_bl;
1104 int main_end_bl_arity;
1105 ir_node **end_preds;
1107 /* assert(exc_handling == 1 || no exceptions. ) */
1109 for (i = 0; i < arity; i++) {
1110 ir_node *ret = get_irn_n(end_bl, i);
1111 ir_node *irn = skip_Proj(ret);
1113 if (is_fragile_op(irn) || is_Raise(irn)) {
1114 cf_pred[n_exc] = ret;
1118 main_end_bl = get_irg_end_block(current_ir_graph);
1119 main_end_bl_arity = get_irn_arity(main_end_bl);
1120 end_preds = xmalloc((n_exc + main_end_bl_arity) * sizeof(*end_preds));
1122 for (i = 0; i < main_end_bl_arity; ++i)
1123 end_preds[i] = get_irn_n(main_end_bl, i);
1124 for (i = 0; i < n_exc; ++i)
1125 end_preds[main_end_bl_arity + i] = cf_pred[i];
1126 set_irn_in(main_end_bl, n_exc + main_end_bl_arity, end_preds);
1127 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1128 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1134 /* -- Turn CSE back on. -- */
1135 set_optimize(rem_opt);
1140 /********************************************************************/
1141 /* Apply inlineing to small methods. */
1142 /********************************************************************/
1144 static struct obstack temp_obst;
1146 /** Represents a possible inlinable call in a graph. */
1147 typedef struct _call_entry call_entry;
1148 struct _call_entry {
1149 ir_node *call; /**< the Call */
1150 ir_graph *callee; /**< the callee called here */
1151 call_entry *next; /**< for linking the next one */
1152 unsigned weight; /**< the weight of the call */
1156 * environment for inlining small irgs
1158 typedef struct _inline_env_t {
1159 struct obstack obst; /**< an obstack where call_entries are allocated on. */
1160 call_entry *head; /**< the head of the call entry list */
1161 call_entry *tail; /**< the tail of the call entry list */
1165 * Returns the irg called from a Call node. If the irg is not
1166 * known, NULL is returned.
1168 * @param call the call node
1170 static ir_graph *get_call_called_irg(ir_node *call) {
1173 addr = get_Call_ptr(call);
1174 if (is_Global(addr)) {
1175 ir_entity *ent = get_Global_entity(addr);
1176 return get_entity_irg(ent);
1183 * Walker: Collect all calls to known graphs inside a graph.
1185 static void collect_calls(ir_node *call, void *env) {
1186 if (is_Call(call)) {
1187 ir_graph *called_irg = get_call_called_irg(call);
1189 if (called_irg != NULL) {
1190 /* The Call node calls a locally defined method. Remember to inline. */
1191 inline_env_t *ienv = env;
1192 call_entry *entry = obstack_alloc(&ienv->obst, sizeof(*entry));
1194 entry->callee = called_irg;
1198 if (ienv->tail == NULL)
1201 ienv->tail->next = entry;
1208 * Inlines all small methods at call sites where the called address comes
1209 * from a Const node that references the entity representing the called
1211 * The size argument is a rough measure for the code size of the method:
1212 * Methods where the obstack containing the firm graph is smaller than
1215 void inline_small_irgs(ir_graph *irg, int size) {
1216 ir_graph *rem = current_ir_graph;
1220 current_ir_graph = irg;
1221 /* Handle graph state */
1222 assert(get_irg_phase_state(irg) != phase_building);
1223 free_callee_info(irg);
1225 /* Find Call nodes to inline.
1226 (We can not inline during a walk of the graph, as inlineing the same
1227 method several times changes the visited flag of the walked graph:
1228 after the first inlineing visited of the callee equals visited of
1229 the caller. With the next inlineing both are increased.) */
1230 obstack_init(&env.obst);
1231 env.head = env.tail = NULL;
1232 irg_walk_graph(irg, NULL, collect_calls, &env);
1234 if (env.head != NULL) {
1237 /* There are calls to inline */
1238 collect_phiprojs(irg);
1239 for (entry = env.head; entry != NULL; entry = entry->next) {
1240 ir_graph *callee = entry->callee;
1241 if (((_obstack_memory_used(callee->obst) - (int)obstack_room(callee->obst)) < size) ||
1242 (get_irg_inline_property(callee) >= irg_inline_forced)) {
1243 did_inline |= inline_method(entry->call, callee);
1246 if (did_inline != 0) {
1247 /* this irg got calls inlined */
1248 set_irg_outs_inconsistent(irg);
1249 set_irg_doms_inconsistent(irg);
1250 set_irg_loopinfo_inconsistent(irg);
1253 obstack_free(&env.obst, NULL);
1254 current_ir_graph = rem;
1258 * Environment for inlining irgs.
1261 int n_nodes; /**< Number of nodes in graph except Id, Tuple, Proj, Start, End. */
1262 int n_blocks; /**< Number of Blocks in graph without Start and End block. */
1263 int n_nodes_orig; /**< for statistics */
1264 call_entry *call_head; /**< The head of the list of all call nodes in this graph. */
1265 call_entry *call_tail; /**< The tail of the list of all call nodes in this graph .*/
1266 int n_call_nodes; /**< Number of Call nodes in the graph. */
1267 int n_call_nodes_orig; /**< for statistics */
1268 int n_callers; /**< Number of known graphs that call this graphs. */
1269 int n_callers_orig; /**< for statistics */
1270 unsigned got_inline:1; /**< Set, if at least one call inside this graph was inlined. */
1271 unsigned local_vars:1; /**< Set, if a inlined function gets the address of an inlined variable. */
1272 unsigned *local_weights; /**< Once allocated, the beneficial weight for transmitting local addresses. */
1276 * Allocate a new environment for inlining.
1278 static inline_irg_env *alloc_inline_irg_env(void) {
1279 inline_irg_env *env = obstack_alloc(&temp_obst, sizeof(*env));
1280 env->n_nodes = -2; /* do not count count Start, End */
1281 env->n_blocks = -2; /* do not count count Start, End Block */
1282 env->n_nodes_orig = -2; /* do not count Start, End */
1283 env->call_head = NULL;
1284 env->call_tail = NULL;
1285 env->n_call_nodes = 0;
1286 env->n_call_nodes_orig = 0;
1288 env->n_callers_orig = 0;
1289 env->got_inline = 0;
1290 env->local_vars = 0;
1291 env->local_weights = NULL;
1295 typedef struct walker_env {
1296 inline_irg_env *x; /**< the inline environment */
1297 char ignore_runtime; /**< the ignore runtime flag */
1298 char ignore_callers; /**< if set, do change callers data */
1302 * post-walker: collect all calls in the inline-environment
1303 * of a graph and sum some statistics.
1305 static void collect_calls2(ir_node *call, void *ctx) {
1307 inline_irg_env *x = env->x;
1308 ir_opcode code = get_irn_opcode(call);
1312 /* count meaningful nodes in irg */
1313 if (code != iro_Proj && code != iro_Tuple && code != iro_Sync) {
1314 if (code != iro_Block) {
1322 if (code != iro_Call) return;
1324 /* check, if it's a runtime call */
1325 if (env->ignore_runtime) {
1326 ir_node *symc = get_Call_ptr(call);
1328 if (is_Global(symc)) {
1329 ir_entity *ent = get_Global_entity(symc);
1331 if (get_entity_additional_properties(ent) & mtp_property_runtime)
1336 /* collect all call nodes */
1338 ++x->n_call_nodes_orig;
1340 callee = get_call_called_irg(call);
1341 if (callee != NULL) {
1342 if (! env->ignore_callers) {
1343 inline_irg_env *callee_env = get_irg_link(callee);
1344 /* count all static callers */
1345 ++callee_env->n_callers;
1346 ++callee_env->n_callers_orig;
1349 /* link it in the list of possible inlinable entries */
1350 entry = obstack_alloc(&temp_obst, sizeof(*entry));
1352 entry->callee = callee;
1354 if (x->call_tail == NULL)
1355 x->call_head = entry;
1357 x->call_tail->next = entry;
1358 x->call_tail = entry;
1363 * Returns TRUE if the number of callers is 0 in the irg's environment,
1364 * hence this irg is a leave.
1366 INLINE static int is_leave(ir_graph *irg) {
1367 inline_irg_env *env = get_irg_link(irg);
1368 return env->n_call_nodes == 0;
1372 * Returns TRUE if the number of nodes in the callee is
1373 * smaller then size in the irg's environment.
1375 INLINE static int is_smaller(ir_graph *callee, int size) {
1376 inline_irg_env *env = get_irg_link(callee);
1377 return env->n_nodes < size;
1381 * Append the nodes of the list src to the nodes of the list in environment dst.
1383 static void append_call_list(inline_irg_env *dst, call_entry *src) {
1384 call_entry *entry, *nentry;
1386 /* Note that the src list points to Call nodes in the inlined graph, but
1387 we need Call nodes in our graph. Luckily the inliner leaves this information
1388 in the link field. */
1389 for (entry = src; entry != NULL; entry = entry->next) {
1390 nentry = obstack_alloc(&temp_obst, sizeof(*nentry));
1391 nentry->call = get_irn_link(entry->call);
1392 nentry->callee = entry->callee;
1393 nentry->next = NULL;
1394 dst->call_tail->next = nentry;
1395 dst->call_tail = nentry;
1400 * Inlines small leave methods at call sites where the called address comes
1401 * from a Const node that references the entity representing the called
1403 * The size argument is a rough measure for the code size of the method:
1404 * Methods where the obstack containing the firm graph is smaller than
1407 void inline_leave_functions(int maxsize, int leavesize, int size, int ignore_runtime) {
1408 inline_irg_env *env;
1414 call_entry *entry, *tail;
1415 const call_entry *centry;
1416 pmap *copied_graphs;
1417 pmap_entry *pm_entry;
1419 rem = current_ir_graph;
1420 obstack_init(&temp_obst);
1422 /* a map for the copied graphs, used to inline recursive calls */
1423 copied_graphs = pmap_create();
1425 /* extend all irgs by a temporary data structure for inlining. */
1426 n_irgs = get_irp_n_irgs();
1427 for (i = 0; i < n_irgs; ++i)
1428 set_irg_link(get_irp_irg(i), alloc_inline_irg_env());
1430 /* Precompute information in temporary data structure. */
1431 wenv.ignore_runtime = ignore_runtime;
1432 wenv.ignore_callers = 0;
1433 for (i = 0; i < n_irgs; ++i) {
1434 ir_graph *irg = get_irp_irg(i);
1436 assert(get_irg_phase_state(irg) != phase_building);
1437 free_callee_info(irg);
1439 wenv.x = get_irg_link(irg);
1440 irg_walk_graph(irg, NULL, collect_calls2, &wenv);
1443 /* -- and now inline. -- */
1445 /* Inline leaves recursively -- we might construct new leaves. */
1449 for (i = 0; i < n_irgs; ++i) {
1451 int phiproj_computed = 0;
1453 current_ir_graph = get_irp_irg(i);
1454 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1457 for (entry = env->call_head; entry != NULL; entry = entry->next) {
1460 if (env->n_nodes > maxsize) break;
1463 callee = entry->callee;
1465 if (is_leave(callee) && (
1466 is_smaller(callee, leavesize) || (get_irg_inline_property(callee) >= irg_inline_forced))) {
1467 if (!phiproj_computed) {
1468 phiproj_computed = 1;
1469 collect_phiprojs(current_ir_graph);
1471 did_inline = inline_method(call, callee);
1474 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1476 /* was inlined, must be recomputed */
1477 phiproj_computed = 0;
1479 /* Do some statistics */
1480 env->got_inline = 1;
1481 --env->n_call_nodes;
1482 env->n_nodes += callee_env->n_nodes;
1483 --callee_env->n_callers;
1485 /* remove this call from the list */
1487 tail->next = entry->next;
1489 env->call_head = entry->next;
1495 env->call_tail = tail;
1497 } while (did_inline);
1499 /* inline other small functions. */
1500 for (i = 0; i < n_irgs; ++i) {
1502 int phiproj_computed = 0;
1504 current_ir_graph = get_irp_irg(i);
1505 env = (inline_irg_env *)get_irg_link(current_ir_graph);
1507 /* note that the list of possible calls is updated during the process */
1509 for (entry = env->call_head; entry != NULL; entry = entry->next) {
1514 callee = entry->callee;
1516 e = pmap_find(copied_graphs, callee);
1519 * Remap callee if we have a copy.
1520 * FIXME: Should we do this only for recursive Calls ?
1525 if (((is_smaller(callee, size) && (env->n_nodes < maxsize)) || /* small function */
1526 (get_irg_inline_property(callee) >= irg_inline_forced))) {
1527 if (current_ir_graph == callee) {
1529 * Recursive call: we cannot directly inline because we cannot walk
1530 * the graph and change it. So we have to make a copy of the graph
1534 inline_irg_env *callee_env;
1538 * No copy yet, create one.
1539 * Note that recursive methods are never leaves, so it is sufficient
1540 * to test this condition here.
1542 copy = create_irg_copy(callee);
1544 /* create_irg_copy() destroys the Proj links, recompute them */
1545 phiproj_computed = 0;
1547 /* allocate new environment */
1548 callee_env = alloc_inline_irg_env();
1549 set_irg_link(copy, callee_env);
1551 wenv.x = callee_env;
1552 wenv.ignore_callers = 1;
1553 irg_walk_graph(copy, NULL, collect_calls2, &wenv);
1556 * Enter the entity of the original graph. This is needed
1557 * for inline_method(). However, note that ent->irg still points
1558 * to callee, NOT to copy.
1560 set_irg_entity(copy, get_irg_entity(callee));
1562 pmap_insert(copied_graphs, callee, copy);
1565 /* we have only one caller: the original graph */
1566 callee_env->n_callers = 1;
1567 callee_env->n_callers_orig = 1;
1569 if (! phiproj_computed) {
1570 phiproj_computed = 1;
1571 collect_phiprojs(current_ir_graph);
1573 did_inline = inline_method(call, callee);
1575 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1577 /* was inlined, must be recomputed */
1578 phiproj_computed = 0;
1580 /* callee was inline. Append it's call list. */
1581 env->got_inline = 1;
1582 --env->n_call_nodes;
1583 append_call_list(env, callee_env->call_head);
1584 env->n_call_nodes += callee_env->n_call_nodes;
1585 env->n_nodes += callee_env->n_nodes;
1586 --callee_env->n_callers;
1588 /* after we have inlined callee, all called methods inside callee
1589 are now called once more */
1590 for (centry = callee_env->call_head; centry != NULL; centry = centry->next) {
1591 inline_irg_env *penv = get_irg_link(centry->callee);
1595 /* remove this call from the list */
1597 tail->next = entry->next;
1599 env->call_head = entry->next;
1605 env->call_tail = tail;
1608 for (i = 0; i < n_irgs; ++i) {
1609 irg = get_irp_irg(i);
1610 env = (inline_irg_env *)get_irg_link(irg);
1612 if (env->got_inline) {
1613 /* this irg got calls inlined */
1614 set_irg_outs_inconsistent(irg);
1615 set_irg_doms_inconsistent(irg);
1616 set_irg_loopinfo_inconsistent(irg);
1618 optimize_graph_df(irg);
1621 if (env->got_inline || (env->n_callers_orig != env->n_callers)) {
1622 DB((dbg, SET_LEVEL_1, "Nodes:%3d ->%3d, calls:%3d ->%3d, callers:%3d ->%3d, -- %s\n",
1623 env->n_nodes_orig, env->n_nodes, env->n_call_nodes_orig, env->n_call_nodes,
1624 env->n_callers_orig, env->n_callers,
1625 get_entity_name(get_irg_entity(irg))));
1629 /* kill the copied graphs: we don't need them anymore */
1630 foreach_pmap(copied_graphs, pm_entry) {
1631 ir_graph *copy = pm_entry->value;
1633 /* reset the entity, otherwise it will be deleted in the next step ... */
1634 set_irg_entity(copy, NULL);
1635 free_ir_graph(copy);
1637 pmap_destroy(copied_graphs);
1639 obstack_free(&temp_obst, NULL);
1640 current_ir_graph = rem;
1644 * Calculate the parameter weights for transmitting the address of a local variable.
1646 static unsigned calc_method_local_weight(ir_node *arg) {
1648 unsigned v, weight = 0;
1650 for (i = get_irn_n_outs(arg) - 1; i >= 0; --i) {
1651 ir_node *succ = get_irn_out(arg, i);
1653 switch (get_irn_opcode(succ)) {
1656 /* Loads and Store can be removed */
1660 /* check if all args are constant */
1661 for (j = get_Sel_n_indexs(succ) - 1; j >= 0; --j) {
1662 ir_node *idx = get_Sel_index(succ, j);
1663 if (! is_Const(idx))
1666 /* Check users on this Sel. Note: if a 0 is returned here, there was
1667 some unsupported node. */
1668 v = calc_method_local_weight(succ);
1671 /* we can kill one Sel with constant indexes, this is cheap */
1675 /* when looking backward we might find Id nodes */
1676 weight += calc_method_local_weight(succ);
1679 /* unoptimized tuple */
1680 for (j = get_Tuple_n_preds(succ) - 1; j >= 0; --j) {
1681 ir_node *pred = get_Tuple_pred(succ, j);
1683 /* look for Proj(j) */
1684 for (k = get_irn_n_outs(succ) - 1; k >= 0; --k) {
1685 ir_node *succ_succ = get_irn_out(succ, k);
1686 if (is_Proj(succ_succ)) {
1687 if (get_Proj_proj(succ_succ) == j) {
1689 weight += calc_method_local_weight(succ_succ);
1692 /* this should NOT happen */
1699 /* any other node: unsupported yet or bad. */
1707 * Calculate the parameter weights for transmitting the address of a local variable.
1709 static void analyze_irg_local_weights(inline_irg_env *env, ir_graph *irg) {
1710 ir_entity *ent = get_irg_entity(irg);
1712 int nparams, i, proj_nr;
1713 ir_node *irg_args, *arg;
1715 mtp = get_entity_type(ent);
1716 nparams = get_method_n_params(mtp);
1718 /* allocate a new array. currently used as 'analysed' flag */
1719 env->local_weights = NEW_ARR_D(unsigned, &temp_obst, nparams);
1721 /* If the method haven't parameters we have nothing to do. */
1725 assure_irg_outs(irg);
1726 irg_args = get_irg_args(irg);
1727 for (i = get_irn_n_outs(irg_args) - 1; i >= 0; --i) {
1728 arg = get_irn_out(irg_args, i);
1729 proj_nr = get_Proj_proj(arg);
1730 env->local_weights[proj_nr] = calc_method_local_weight(arg);
1735 * Calculate the benefice for transmitting an local variable address.
1736 * After inlining, the local variable might be transformed into a
1737 * SSA variable by scalar_replacement().
1739 static unsigned get_method_local_adress_weight(ir_graph *callee, int pos) {
1740 inline_irg_env *env = get_irg_link(callee);
1742 if (env->local_weights != NULL) {
1743 if (pos < ARR_LEN(env->local_weights))
1744 return env->local_weights[pos];
1748 analyze_irg_local_weights(env, callee);
1750 if (pos < ARR_LEN(env->local_weights))
1751 return env->local_weights[pos];
1756 * calculate a benefice value for inlining the given call.
1758 static int calc_inline_benefice(ir_node *call, ir_graph *callee, unsigned *local_adr) {
1759 ir_entity *ent = get_irg_entity(callee);
1766 inline_irg_env *curr_env, *callee_env;
1768 if (get_entity_additional_properties(ent) & mtp_property_noreturn) {
1769 /* do NOT inline noreturn calls */
1773 /* costs for every passed parameter */
1774 n_params = get_Call_n_params(call);
1775 mtp = get_entity_type(ent);
1776 cc = get_method_calling_convention(mtp);
1777 if (cc & cc_reg_param) {
1778 /* register parameter, smaller costs for register parameters */
1779 int max_regs = cc & ~cc_bits;
1781 if (max_regs < n_params)
1782 weight += max_regs * 2 + (n_params - max_regs) * 5;
1784 weight += n_params * 2;
1786 /* parameters are passed an stack */
1787 weight += 5 * n_params;
1790 /* constant parameters improve the benefice */
1791 frame_ptr = get_irg_frame(current_ir_graph);
1792 for (i = 0; i < n_params; ++i) {
1793 ir_node *param = get_Call_param(call, i);
1795 if (is_Const(param) || is_SymConst(param))
1796 weight += get_method_param_weight(ent, i);
1797 else if (is_Sel(param) && get_Sel_ptr(param) == frame_ptr) {
1799 * An address of a local variable is transmitted. After inlining,
1800 * scalar_replacement might be able to remove the local variable,
1803 v = get_method_local_adress_weight(callee, i);
1810 callee_env = get_irg_link(callee);
1811 if (get_entity_visibility(ent) == visibility_local &&
1812 callee_env->n_callers_orig == 1 &&
1813 callee != current_ir_graph) {
1814 /* we are the only caller, give big bonus */
1818 /* do not inline big functions */
1819 weight -= callee_env->n_nodes;
1821 /* reduce the benefice if the current function is already big */
1822 curr_env = get_irg_link(current_ir_graph);
1823 weight -= curr_env->n_nodes / 100;
1825 /* give a bonus for functions with one block */
1826 if (callee_env->n_blocks == 1)
1827 weight = weight * 3 / 2;
1833 * Heuristic inliner. Calculates a benifice value for every call and inlines
1834 * those calls with a value higher than the threshold.
1836 void inline_functions(int inline_threshold) {
1837 inline_irg_env *env;
1843 call_entry *entry, *tail;
1844 const call_entry *centry;
1845 pmap *copied_graphs;
1846 pmap_entry *pm_entry;
1848 rem = current_ir_graph;
1849 obstack_init(&temp_obst);
1851 /* a map for the copied graphs, used to inline recursive calls */
1852 copied_graphs = pmap_create();
1854 /* extend all irgs by a temporary data structure for inlining. */
1855 n_irgs = get_irp_n_irgs();
1856 for (i = 0; i < n_irgs; ++i)
1857 set_irg_link(get_irp_irg(i), alloc_inline_irg_env());
1859 /* Precompute information in temporary data structure. */
1860 wenv.ignore_runtime = 0;
1861 wenv.ignore_callers = 0;
1862 for (i = 0; i < n_irgs; ++i) {
1863 ir_graph *irg = get_irp_irg(i);
1865 assert(get_irg_phase_state(irg) != phase_building);
1866 free_callee_info(irg);
1868 wenv.x = get_irg_link(irg);
1869 irg_walk_graph(irg, NULL, collect_calls2, &wenv);
1872 /* -- and now inline. -- */
1873 for (i = 0; i < n_irgs; ++i) {
1875 int phiproj_computed = 0;
1877 current_ir_graph = get_irp_irg(i);
1878 env = get_irg_link(current_ir_graph);
1880 /* note that the list of possible calls is updated during the process */
1882 for (entry = env->call_head; entry != NULL; entry = entry->next) {
1889 callee = entry->callee;
1891 /* calculate the benifice on the original call to prevent excessive inlining */
1893 benefice = calc_inline_benefice(call, callee, &local_adr);
1894 DB((dbg, SET_LEVEL_2, "In %+F Call %+F has benefice %d\n", current_ir_graph, callee, benefice));
1896 e = pmap_find(copied_graphs, callee);
1899 * Remap callee if we have a copy.
1900 * FIXME: Should we do this only for recursive Calls ?
1905 if (benefice > -inline_threshold ||
1906 (get_irg_inline_property(callee) >= irg_inline_forced)) {
1907 if (current_ir_graph == callee) {
1909 * Recursive call: we cannot directly inline because we cannot walk
1910 * the graph and change it. So we have to make a copy of the graph
1914 inline_irg_env *callee_env;
1918 * No copy yet, create one.
1919 * Note that recursive methods are never leaves, so it is sufficient
1920 * to test this condition here.
1922 copy = create_irg_copy(callee);
1924 /* create_irg_copy() destroys the Proj links, recompute them */
1925 phiproj_computed = 0;
1927 /* allocate new environment */
1928 callee_env = alloc_inline_irg_env();
1929 set_irg_link(copy, callee_env);
1931 wenv.x = callee_env;
1932 wenv.ignore_callers = 1;
1933 irg_walk_graph(copy, NULL, collect_calls2, &wenv);
1936 * Enter the entity of the original graph. This is needed
1937 * for inline_method(). However, note that ent->irg still points
1938 * to callee, NOT to copy.
1940 set_irg_entity(copy, get_irg_entity(callee));
1942 pmap_insert(copied_graphs, callee, copy);
1945 /* we have only one caller: the original graph */
1946 callee_env->n_callers = 1;
1947 callee_env->n_callers_orig = 1;
1949 if (! phiproj_computed) {
1950 phiproj_computed = 1;
1951 collect_phiprojs(current_ir_graph);
1953 did_inline = inline_method(call, callee);
1955 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1957 /* was inlined, must be recomputed */
1958 phiproj_computed = 0;
1960 /* callee was inline. Append it's call list. */
1961 env->got_inline = 1;
1963 env->local_vars = 1;
1964 --env->n_call_nodes;
1965 append_call_list(env, callee_env->call_head);
1966 env->n_call_nodes += callee_env->n_call_nodes;
1967 env->n_nodes += callee_env->n_nodes;
1968 --callee_env->n_callers;
1970 /* after we have inlined callee, all called methods inside callee
1971 are now called once more */
1972 for (centry = callee_env->call_head; centry != NULL; centry = centry->next) {
1973 inline_irg_env *penv = get_irg_link(centry->callee);
1977 /* remove this call from the list */
1979 tail->next = entry->next;
1981 env->call_head = entry->next;
1987 env->call_tail = tail;
1990 for (i = 0; i < n_irgs; ++i) {
1991 irg = get_irp_irg(i);
1992 env = (inline_irg_env *)get_irg_link(irg);
1994 if (env->got_inline) {
1995 /* this irg got calls inlined */
1996 set_irg_outs_inconsistent(irg);
1997 set_irg_doms_inconsistent(irg);
1998 set_irg_loopinfo_inconsistent(irg);
2000 /* scalar replacement does not work well with Tuple nodes, so optimize them away */
2001 optimize_graph_df(irg);
2003 if (env->local_vars) {
2004 if (scalar_replacement_opt(irg))
2005 optimize_graph_df(irg);
2010 if (env->got_inline || (env->n_callers_orig != env->n_callers)) {
2011 DB((dbg, SET_LEVEL_1, "Nodes:%3d ->%3d, calls:%3d ->%3d, callers:%3d ->%3d, -- %s\n",
2012 env->n_nodes_orig, env->n_nodes, env->n_call_nodes_orig, env->n_call_nodes,
2013 env->n_callers_orig, env->n_callers,
2014 get_entity_name(get_irg_entity(irg))));
2018 /* kill the copied graphs: we don't need them anymore */
2019 foreach_pmap(copied_graphs, pm_entry) {
2020 ir_graph *copy = pm_entry->value;
2022 /* reset the entity, otherwise it will be deleted in the next step ... */
2023 set_irg_entity(copy, NULL);
2024 free_ir_graph(copy);
2026 pmap_destroy(copied_graphs);
2028 obstack_free(&temp_obst, NULL);
2029 current_ir_graph = rem;
2032 void firm_init_inline(void) {
2033 FIRM_DBG_REGISTER(dbg, "firm.opt.inline");