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
32 #include "irgraph_t.h"
35 #include "iroptimize.h"
52 #include "irbackedge_t.h"
53 #include "opt_inline_t.h"
58 #include "analyze_irg_args.h"
59 #include "iredges_t.h"
63 #include "iropt_dbg.h"
65 DEBUG_ONLY(static firm_dbg_module_t *dbg;)
67 /*------------------------------------------------------------------*/
68 /* Routines for dead node elimination / copying garbage collection */
70 /*------------------------------------------------------------------*/
73 * Remember the new node in the old node by using a field all nodes have.
75 #define set_new_node(oldn, newn) set_irn_link(oldn, newn)
78 * Get this new node, before the old node is forgotten.
80 #define get_new_node(oldn) get_irn_link(oldn)
83 * Check if a new node was set.
85 #define has_new_node(n) (get_new_node(n) != NULL)
88 * We use the block_visited flag to mark that we have computed the
89 * number of useful predecessors for this block.
90 * Further we encode the new arity in this flag in the old blocks.
91 * Remembering the arity is useful, as it saves a lot of pointer
92 * accesses. This function is called for all Phi and Block nodes
96 compute_new_arity(ir_node *b) {
97 int i, res, irn_arity;
100 irg_v = get_irg_block_visited(current_ir_graph);
101 block_v = get_Block_block_visited(b);
102 if (block_v >= irg_v) {
103 /* we computed the number of preds for this block and saved it in the
105 return block_v - irg_v;
107 /* compute the number of good predecessors */
108 res = irn_arity = get_irn_arity(b);
109 for (i = 0; i < irn_arity; i++)
110 if (is_Bad(get_irn_n(b, i))) res--;
111 /* save it in the flag. */
112 set_Block_block_visited(b, irg_v + res);
118 * Copies the node to the new obstack. The Ins of the new node point to
119 * the predecessors on the old obstack. For block/phi nodes not all
120 * predecessors might be copied. n->link points to the new node.
121 * For Phi and Block nodes the function allocates in-arrays with an arity
122 * only for useful predecessors. The arity is determined by counting
123 * the non-bad predecessors of the block.
125 * @param n The node to be copied
126 * @param env if non-NULL, the node number attribute will be copied to the new node
128 * Note: Also used for loop unrolling.
130 static void copy_node(ir_node *n, void *env) {
133 ir_op *op = get_irn_op(n);
137 /* node copied already */
139 } else if (op == op_Block) {
141 new_arity = compute_new_arity(n);
142 n->attr.block.graph_arr = NULL;
144 block = get_nodes_block(n);
146 new_arity = compute_new_arity(block);
148 new_arity = get_irn_arity(n);
151 nn = new_ir_node(get_irn_dbg_info(n),
158 /* Copy the attributes. These might point to additional data. If this
159 was allocated on the old obstack the pointers now are dangling. This
160 frees e.g. the memory of the graph_arr allocated in new_immBlock. */
161 if (op == op_Block) {
162 /* we cannot allow blocks WITHOUT macroblock input */
163 set_Block_MacroBlock(nn, get_Block_MacroBlock(n));
165 copy_node_attr(n, nn);
168 /* for easier debugging, we want to copy the node numbers too */
169 nn->node_nr = n->node_nr;
173 hook_dead_node_elim_subst(current_ir_graph, n, nn);
177 * Copies new predecessors of old node to new node remembered in link.
178 * Spare the Bad predecessors of Phi and Block nodes.
180 static void copy_preds(ir_node *n, void *env) {
185 nn = get_new_node(n);
188 /* copy the macro block header */
189 ir_node *mbh = get_Block_MacroBlock(n);
192 /* this block is a macroblock header */
193 set_Block_MacroBlock(nn, nn);
195 /* get the macro block header */
196 ir_node *nmbh = get_new_node(mbh);
197 assert(nmbh != NULL);
198 set_Block_MacroBlock(nn, nmbh);
201 /* Don't copy Bad nodes. */
203 irn_arity = get_irn_arity(n);
204 for (i = 0; i < irn_arity; i++) {
205 if (! is_Bad(get_irn_n(n, i))) {
206 ir_node *pred = get_irn_n(n, i);
207 set_irn_n(nn, j, get_new_node(pred));
211 /* repair the block visited flag from above misuse. Repair it in both
212 graphs so that the old one can still be used. */
213 set_Block_block_visited(nn, 0);
214 set_Block_block_visited(n, 0);
215 /* Local optimization could not merge two subsequent blocks if
216 in array contained Bads. Now it's possible.
217 We don't call optimize_in_place as it requires
218 that the fields in ir_graph are set properly. */
219 if (!has_Block_entity(nn) &&
220 get_opt_control_flow_straightening() &&
221 get_Block_n_cfgpreds(nn) == 1 &&
222 is_Jmp(get_Block_cfgpred(nn, 0))) {
223 ir_node *old = get_nodes_block(get_Block_cfgpred(nn, 0));
225 /* Jmp jumps into the block it is in -- deal self cycle. */
226 assert(is_Bad(get_new_node(get_irg_bad(current_ir_graph))));
227 exchange(nn, get_new_node(get_irg_bad(current_ir_graph)));
232 } else if (is_Phi(n) && get_irn_arity(n) > 0) {
233 /* Don't copy node if corresponding predecessor in block is Bad.
234 The Block itself should not be Bad. */
235 block = get_nodes_block(n);
236 set_nodes_block(nn, get_new_node(block));
238 irn_arity = get_irn_arity(n);
239 for (i = 0; i < irn_arity; i++) {
240 if (! is_Bad(get_irn_n(block, i))) {
241 ir_node *pred = get_irn_n(n, i);
242 set_irn_n(nn, j, get_new_node(pred));
243 /*if (is_backedge(n, i)) set_backedge(nn, j);*/
247 /* If the pre walker reached this Phi after the post walker visited the
248 block block_visited is > 0. */
249 set_Block_block_visited(get_nodes_block(n), 0);
250 /* Compacting the Phi's ins might generate Phis with only one
252 if (get_irn_arity(nn) == 1)
253 exchange(nn, get_irn_n(nn, 0));
255 irn_arity = get_irn_arity(n);
256 for (i = -1; i < irn_arity; i++)
257 set_irn_n(nn, i, get_new_node(get_irn_n(n, i)));
259 /* Now the new node is complete. We can add it to the hash table for CSE.
260 @@@ inlining aborts if we identify End. Why? */
262 add_identities(current_ir_graph->value_table, nn);
266 * Copies the graph recursively, compacts the keep-alives of the end node.
268 * @param irg the graph to be copied
269 * @param copy_node_nr If non-zero, the node number will be copied
271 static void copy_graph(ir_graph *irg, int copy_node_nr) {
272 ir_node *oe, *ne, *ob, *nb, *om, *nm; /* old end, new end, old bad, new bad, old NoMem, new NoMem */
273 ir_node *ka; /* keep alive */
277 /* Some nodes must be copied by hand, sigh */
278 vfl = get_irg_visited(irg);
279 set_irg_visited(irg, vfl + 1);
281 oe = get_irg_end(irg);
282 mark_irn_visited(oe);
283 /* copy the end node by hand, allocate dynamic in array! */
284 ne = new_ir_node(get_irn_dbg_info(oe),
291 /* Copy the attributes. Well, there might be some in the future... */
292 copy_node_attr(oe, ne);
293 set_new_node(oe, ne);
295 /* copy the Bad node */
296 ob = get_irg_bad(irg);
297 mark_irn_visited(ob);
298 nb = new_ir_node(get_irn_dbg_info(ob),
305 copy_node_attr(ob, nb);
306 set_new_node(ob, nb);
308 /* copy the NoMem node */
309 om = get_irg_no_mem(irg);
310 mark_irn_visited(om);
311 nm = new_ir_node(get_irn_dbg_info(om),
318 copy_node_attr(om, nm);
319 set_new_node(om, nm);
321 /* copy the live nodes */
322 set_irg_visited(irg, vfl);
323 irg_walk(get_nodes_block(oe), copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
325 /* Note: from yet, the visited flag of the graph is equal to vfl + 1 */
327 /* visit the anchors as well */
328 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
329 ir_node *n = get_irg_anchor(irg, i);
331 if (n && (get_irn_visited(n) <= vfl)) {
332 set_irg_visited(irg, vfl);
333 irg_walk(n, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
337 /* copy_preds for the end node ... */
338 set_nodes_block(ne, get_new_node(get_nodes_block(oe)));
340 /*- ... and now the keep alives. -*/
341 /* First pick the not marked block nodes and walk them. We must pick these
342 first as else we will oversee blocks reachable from Phis. */
343 irn_arity = get_End_n_keepalives(oe);
344 for (i = 0; i < irn_arity; i++) {
345 ka = get_End_keepalive(oe, i);
347 if (get_irn_visited(ka) <= vfl) {
348 /* We must keep the block alive and copy everything reachable */
349 set_irg_visited(irg, vfl);
350 irg_walk(ka, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
352 add_End_keepalive(ne, get_new_node(ka));
356 /* Now pick other nodes. Here we will keep all! */
357 irn_arity = get_End_n_keepalives(oe);
358 for (i = 0; i < irn_arity; i++) {
359 ka = get_End_keepalive(oe, i);
361 if (get_irn_visited(ka) <= vfl) {
362 /* We didn't copy the node yet. */
363 set_irg_visited(irg, vfl);
364 irg_walk(ka, copy_node, copy_preds, INT_TO_PTR(copy_node_nr));
366 add_End_keepalive(ne, get_new_node(ka));
370 /* start block sometimes only reached after keep alives */
371 set_nodes_block(nb, get_new_node(get_nodes_block(ob)));
372 set_nodes_block(nm, get_new_node(get_nodes_block(om)));
376 * Copies the graph reachable from current_ir_graph->end to the obstack
377 * in current_ir_graph and fixes the environment.
378 * Then fixes the fields in current_ir_graph containing nodes of the
381 * @param copy_node_nr If non-zero, the node number will be copied
384 copy_graph_env(int copy_node_nr) {
385 ir_graph *irg = current_ir_graph;
386 ir_node *old_end, *new_anchor;
389 /* remove end_except and end_reg nodes */
390 old_end = get_irg_end(irg);
391 set_irg_end_except (irg, old_end);
392 set_irg_end_reg (irg, old_end);
394 /* Not all nodes remembered in irg might be reachable
395 from the end node. Assure their link is set to NULL, so that
396 we can test whether new nodes have been computed. */
397 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
398 ir_node *n = get_irg_anchor(irg, i);
400 set_new_node(n, NULL);
402 /* we use the block walk flag for removing Bads from Blocks ins. */
403 inc_irg_block_visited(irg);
406 copy_graph(irg, copy_node_nr);
409 old_end = get_irg_end(irg);
410 new_anchor = new_Anchor(irg);
412 for (i = get_irg_n_anchors(irg) - 1; i >= 0; --i) {
413 ir_node *n = get_irg_anchor(irg, i);
415 set_irn_n(new_anchor, i, get_new_node(n));
418 irg->anchor = new_anchor;
420 /* ensure the new anchor is placed in the endblock */
421 set_nodes_block(new_anchor, get_irg_end_block(irg));
425 * Copies all reachable nodes to a new obstack. Removes bad inputs
426 * from block nodes and the corresponding inputs from Phi nodes.
427 * Merges single exit blocks with single entry blocks and removes
429 * Adds all new nodes to a new hash table for CSE. Does not
430 * perform CSE, so the hash table might contain common subexpressions.
432 void dead_node_elimination(ir_graph *irg) {
434 #ifdef INTERPROCEDURAL_VIEW
435 int rem_ipview = get_interprocedural_view();
437 struct obstack *graveyard_obst = NULL;
438 struct obstack *rebirth_obst = NULL;
440 edges_deactivate(irg);
442 /* inform statistics that we started a dead-node elimination run */
443 hook_dead_node_elim(irg, 1);
445 /* Remember external state of current_ir_graph. */
446 rem = current_ir_graph;
447 current_ir_graph = irg;
448 #ifdef INTERPROCEDURAL_VIEW
449 set_interprocedural_view(0);
452 assert(get_irg_phase_state(irg) != phase_building);
454 /* Handle graph state */
455 free_callee_info(irg);
459 /* @@@ so far we loose loops when copying */
460 free_loop_information(irg);
462 set_irg_doms_inconsistent(irg);
464 /* A quiet place, where the old obstack can rest in peace,
465 until it will be cremated. */
466 graveyard_obst = irg->obst;
468 /* A new obstack, where the reachable nodes will be copied to. */
469 rebirth_obst = XMALLOC(struct obstack);
470 irg->obst = rebirth_obst;
471 obstack_init(irg->obst);
472 irg->last_node_idx = 0;
474 /* We also need a new value table for CSE */
475 del_identities(irg->value_table);
476 irg->value_table = new_identities();
478 /* Copy the graph from the old to the new obstack */
479 copy_graph_env(/*copy_node_nr=*/1);
481 /* Free memory from old unoptimized obstack */
482 obstack_free(graveyard_obst, 0); /* First empty the obstack ... */
483 xfree(graveyard_obst); /* ... then free it. */
485 /* inform statistics that the run is over */
486 hook_dead_node_elim(irg, 0);
488 current_ir_graph = rem;
489 #ifdef INTERPROCEDURAL_VIEW
490 set_interprocedural_view(rem_ipview);
495 * Relink bad predecessors of a block and store the old in array to the
496 * link field. This function is called by relink_bad_predecessors().
497 * The array of link field starts with the block operand at position 0.
498 * If block has bad predecessors, create a new in array without bad preds.
499 * Otherwise let in array untouched.
501 static void relink_bad_block_predecessors(ir_node *n, void *env) {
502 ir_node **new_in, *irn;
503 int i, new_irn_n, old_irn_arity, new_irn_arity = 0;
506 /* if link field of block is NULL, look for bad predecessors otherwise
507 this is already done */
508 if (is_Block(n) && get_irn_link(n) == NULL) {
509 /* save old predecessors in link field (position 0 is the block operand)*/
510 set_irn_link(n, get_irn_in(n));
512 /* count predecessors without bad nodes */
513 old_irn_arity = get_irn_arity(n);
514 for (i = 0; i < old_irn_arity; i++)
515 if (!is_Bad(get_irn_n(n, i)))
518 /* arity changing: set new predecessors without bad nodes */
519 if (new_irn_arity < old_irn_arity) {
520 /* Get new predecessor array. We do not resize the array, as we must
521 keep the old one to update Phis. */
522 new_in = NEW_ARR_D(ir_node *, current_ir_graph->obst, (new_irn_arity+1));
524 /* set new predecessors in array */
527 for (i = 0; i < old_irn_arity; i++) {
528 irn = get_irn_n(n, i);
530 new_in[new_irn_n] = irn;
531 is_backedge(n, i) ? set_backedge(n, new_irn_n-1) : set_not_backedge(n, new_irn_n-1);
535 /* ARR_SETLEN(int, n->attr.block.backedge, new_irn_arity); */
536 ARR_SHRINKLEN(n->attr.block.backedge, new_irn_arity);
538 } /* ir node has bad predecessors */
539 } /* Block is not relinked */
543 * Relinks Bad predecessors from Blocks and Phis called by walker
544 * remove_bad_predecesors(). If n is a Block, call
545 * relink_bad_block_redecessors(). If n is a Phi-node, call also the relinking
546 * function of Phi's Block. If this block has bad predecessors, relink preds
549 static void relink_bad_predecessors(ir_node *n, void *env) {
550 ir_node *block, **old_in;
551 int i, old_irn_arity, new_irn_arity;
553 /* relink bad predecessors of a block */
555 relink_bad_block_predecessors(n, env);
557 /* If Phi node relink its block and its predecessors */
559 /* Relink predecessors of phi's block */
560 block = get_nodes_block(n);
561 if (get_irn_link(block) == NULL)
562 relink_bad_block_predecessors(block, env);
564 old_in = (ir_node **)get_irn_link(block); /* Of Phi's Block */
565 old_irn_arity = ARR_LEN(old_in);
567 /* Relink Phi predecessors if count of predecessors changed */
568 if (old_irn_arity != ARR_LEN(get_irn_in(block))) {
569 /* set new predecessors in array
570 n->in[0] remains the same block */
572 for(i = 1; i < old_irn_arity; i++)
573 if (!is_Bad(old_in[i])) {
574 n->in[new_irn_arity] = n->in[i];
575 is_backedge(n, i) ? set_backedge(n, new_irn_arity) : set_not_backedge(n, new_irn_arity);
579 ARR_SETLEN(ir_node *, n->in, new_irn_arity);
580 ARR_SETLEN(int, n->attr.phi.u.backedge, new_irn_arity);
582 } /* n is a Phi node */
586 * Removes Bad Bad predecessors from Blocks and the corresponding
587 * inputs to Phi nodes as in dead_node_elimination but without
589 * On walking up set the link field to NULL, on walking down call
590 * relink_bad_predecessors() (This function stores the old in array
591 * to the link field and sets a new in array if arity of predecessors
594 void remove_bad_predecessors(ir_graph *irg) {
595 panic("Fix backedge handling first");
596 irg_walk_graph(irg, firm_clear_link, relink_bad_predecessors, NULL);
603 __)|_| | \_/ | \_/(/_ |_/\__|__
605 The following stuff implements a facility that automatically patches
606 registered ir_node pointers to the new node when a dead node elimination occurs.
609 struct _survive_dce_t {
613 hook_entry_t dead_node_elim;
614 hook_entry_t dead_node_elim_subst;
617 typedef struct _survive_dce_list_t {
618 struct _survive_dce_list_t *next;
620 } survive_dce_list_t;
622 static void dead_node_hook(void *context, ir_graph *irg, int start) {
623 survive_dce_t *sd = context;
626 /* Create a new map before the dead node elimination is performed. */
628 sd->new_places = pmap_create_ex(pmap_count(sd->places));
630 /* Patch back all nodes if dead node elimination is over and something is to be done. */
631 pmap_destroy(sd->places);
632 sd->places = sd->new_places;
633 sd->new_places = NULL;
638 * Hook called when dead node elimination replaces old by nw.
640 static void dead_node_subst_hook(void *context, ir_graph *irg, ir_node *old, ir_node *nw) {
641 survive_dce_t *sd = context;
642 survive_dce_list_t *list = pmap_get(sd->places, old);
645 /* If the node is to be patched back, write the new address to all registered locations. */
647 survive_dce_list_t *p;
649 for (p = list; p; p = p->next)
652 pmap_insert(sd->new_places, nw, list);
657 * Make a new Survive DCE environment.
659 survive_dce_t *new_survive_dce(void) {
660 survive_dce_t *res = XMALLOC(survive_dce_t);
661 obstack_init(&res->obst);
662 res->places = pmap_create();
663 res->new_places = NULL;
665 res->dead_node_elim.hook._hook_dead_node_elim = dead_node_hook;
666 res->dead_node_elim.context = res;
667 res->dead_node_elim.next = NULL;
669 res->dead_node_elim_subst.hook._hook_dead_node_elim_subst = dead_node_subst_hook;
670 res->dead_node_elim_subst.context = res;
671 res->dead_node_elim_subst.next = NULL;
673 register_hook(hook_dead_node_elim, &res->dead_node_elim);
674 register_hook(hook_dead_node_elim_subst, &res->dead_node_elim_subst);
679 * Free a Survive DCE environment.
681 void free_survive_dce(survive_dce_t *sd) {
682 obstack_free(&sd->obst, NULL);
683 pmap_destroy(sd->places);
684 unregister_hook(hook_dead_node_elim, &sd->dead_node_elim);
685 unregister_hook(hook_dead_node_elim_subst, &sd->dead_node_elim_subst);
690 * Register a node pointer to be patched upon DCE.
691 * When DCE occurs, the node pointer specified by @p place will be
692 * patched to the new address of the node it is pointing to.
694 * @param sd The Survive DCE environment.
695 * @param place The address of the node pointer.
697 void survive_dce_register_irn(survive_dce_t *sd, ir_node **place) {
698 if (*place != NULL) {
699 ir_node *irn = *place;
700 survive_dce_list_t *curr = pmap_get(sd->places, irn);
701 survive_dce_list_t *nw = obstack_alloc(&sd->obst, sizeof(nw[0]));
706 pmap_insert(sd->places, irn, nw);
710 /*--------------------------------------------------------------------*/
711 /* Functionality for inlining */
712 /*--------------------------------------------------------------------*/
715 * Copy node for inlineing. Updates attributes that change when
716 * inlineing but not for dead node elimination.
718 * Copies the node by calling copy_node() and then updates the entity if
719 * it's a local one. env must be a pointer of the frame type of the
720 * inlined procedure. The new entities must be in the link field of
723 static void copy_node_inline(ir_node *n, void *env) {
725 ir_type *frame_tp = (ir_type *)env;
729 nn = get_new_node(n);
731 /* use copied entities from the new frame */
732 if (get_entity_owner(get_Sel_entity(n)) == frame_tp) {
733 set_Sel_entity(nn, get_entity_link(get_Sel_entity(n)));
735 } else if (is_Block(n)) {
736 nn = get_new_node(n);
737 nn->attr.block.irg.irg = current_ir_graph;
742 * Copies new predecessors of old node and move constants to
745 static void copy_preds_inline(ir_node *n, void *env) {
749 nn = skip_Id(get_new_node(n));
750 if (is_irn_constlike(nn)) {
751 /* move Constants into the start block */
752 set_nodes_block(nn, get_irg_start_block(current_ir_graph));
754 n = identify_remember(current_ir_graph->value_table, nn);
763 * Walker: checks if P_value_arg_base is used.
765 static void find_addr(ir_node *node, void *env) {
766 int *allow_inline = env;
768 ir_graph *irg = current_ir_graph;
769 if (get_Sel_ptr(node) == get_irg_frame(irg)) {
770 /* access to frame */
771 ir_entity *ent = get_Sel_entity(node);
772 if (get_entity_owner(ent) != get_irg_frame_type(irg)) {
773 /* access to value_type */
777 } else if (is_Alloc(node) && get_Alloc_where(node) == stack_alloc) {
779 * Refuse to inline alloca call unless user explicitly forced so as this
780 * may change program's memory overhead drastically when the function
781 * using alloca is called in loop. In GCC present in SPEC2000 inlining
782 * into schedule_block cause it to require 2GB of ram instead of 256MB.
784 * Sorrily this is true with our implementation also.
785 * Moreover, we cannot differentiate between alloca() and VLA yet, so this
786 * disables inlining of functions using VLA (with are completely save).
789 * - add a flag to the Alloc node for "real" alloca() calls
790 * - add a new Stack-Restore node at the end of a function using alloca()
797 * Check if we can inline a given call.
798 * Currently, we cannot inline two cases:
799 * - call with compound arguments
800 * - graphs that take the address of a parameter
802 * check these conditions here
804 static int can_inline(ir_node *call, ir_graph *called_graph) {
805 ir_type *call_type = get_Call_type(call);
806 int params, ress, i, res;
807 assert(is_Method_type(call_type));
809 params = get_method_n_params(call_type);
810 ress = get_method_n_ress(call_type);
812 /* check parameters for compound arguments */
813 for (i = 0; i < params; ++i) {
814 ir_type *p_type = get_method_param_type(call_type, i);
816 if (is_compound_type(p_type))
820 /* check results for compound arguments */
821 for (i = 0; i < ress; ++i) {
822 ir_type *r_type = get_method_res_type(call_type, i);
824 if (is_compound_type(r_type))
829 irg_walk_graph(called_graph, find_addr, NULL, &res);
835 exc_handler = 0, /**< There is a handler. */
836 exc_to_end = 1, /**< Branches to End. */
837 exc_no_handler = 2 /**< Exception handling not represented. */
840 /* Inlines a method at the given call site. */
841 int inline_method(ir_node *call, ir_graph *called_graph) {
843 ir_node *post_call, *post_bl;
844 ir_node *in[pn_Start_max];
845 ir_node *end, *end_bl, *block;
850 int arity, n_ret, n_exc, n_res, i, n, j, rem_opt, irn_arity, n_params;
851 enum exc_mode exc_handling;
852 ir_type *called_frame, *curr_frame, *mtp, *ctp;
855 irg_inline_property prop = get_irg_inline_property(called_graph);
856 unsigned long visited;
858 if (prop == irg_inline_forbidden)
861 ent = get_irg_entity(called_graph);
863 mtp = get_entity_type(ent);
864 ctp = get_Call_type(call);
865 n_params = get_method_n_params(mtp);
866 n_res = get_method_n_ress(mtp);
867 if (n_params > get_method_n_params(ctp)) {
868 /* this is a bad feature of C: without a prototype, we can
869 * call a function with less parameters than needed. Currently
870 * we don't support this, although we could use Unknown than. */
873 if (n_res != get_method_n_ress(ctp)) {
877 /* Argh, compiling C has some bad consequences:
878 * It is implementation dependent what happens in that case.
879 * We support inlining, if the bitsize of the types matches AND
880 * the same arithmetic is used. */
881 for (i = n_params - 1; i >= 0; --i) {
882 ir_type *param_tp = get_method_param_type(mtp, i);
883 ir_type *arg_tp = get_method_param_type(ctp, i);
885 if (param_tp != arg_tp) {
886 ir_mode *pmode = get_type_mode(param_tp);
887 ir_mode *amode = get_type_mode(arg_tp);
889 if (pmode == NULL || amode == NULL)
891 if (get_mode_size_bits(pmode) != get_mode_size_bits(amode))
893 if (get_mode_arithmetic(pmode) != get_mode_arithmetic(amode))
895 /* otherwise we can simply "reinterpret" the bits */
898 for (i = n_res - 1; i >= 0; --i) {
899 ir_type *decl_res_tp = get_method_res_type(mtp, i);
900 ir_type *used_res_tp = get_method_res_type(ctp, i);
902 if (decl_res_tp != used_res_tp) {
903 ir_mode *decl_mode = get_type_mode(decl_res_tp);
904 ir_mode *used_mode = get_type_mode(used_res_tp);
905 if (decl_mode == NULL || used_mode == NULL)
907 if (get_mode_size_bits(decl_mode) != get_mode_size_bits(used_mode))
909 if (get_mode_arithmetic(decl_mode) != get_mode_arithmetic(used_mode))
911 /* otherwise we can "reinterpret" the bits */
915 irg = get_irn_irg(call);
918 * We cannot inline a recursive call. The graph must be copied before
919 * the call the inline_method() using create_irg_copy().
921 if (called_graph == irg)
925 * currently, we cannot inline two cases:
926 * - call with compound arguments
927 * - graphs that take the address of a parameter
929 if (! can_inline(call, called_graph))
932 rem = current_ir_graph;
933 current_ir_graph = irg;
935 DB((dbg, LEVEL_1, "Inlining %+F(%+F) into %+F\n", call, called_graph, irg));
937 /* -- Turn off optimizations, this can cause problems when allocating new nodes. -- */
938 rem_opt = get_opt_optimize();
941 /* Handle graph state */
942 assert(get_irg_phase_state(irg) != phase_building);
943 assert(get_irg_pinned(irg) == op_pin_state_pinned);
944 assert(get_irg_pinned(called_graph) == op_pin_state_pinned);
945 set_irg_outs_inconsistent(irg);
946 set_irg_extblk_inconsistent(irg);
947 set_irg_doms_inconsistent(irg);
948 set_irg_loopinfo_inconsistent(irg);
949 set_irg_callee_info_state(irg, irg_callee_info_inconsistent);
950 set_irg_entity_usage_state(irg, ir_entity_usage_not_computed);
952 /* -- Check preconditions -- */
953 assert(is_Call(call));
955 /* here we know we WILL inline, so inform the statistics */
956 hook_inline(call, called_graph);
958 /* -- Decide how to handle exception control flow: Is there a handler
959 for the Call node, or do we branch directly to End on an exception?
961 0 There is a handler.
963 2 Exception handling not represented in Firm. -- */
965 ir_node *proj, *Mproj = NULL, *Xproj = NULL;
966 for (proj = get_irn_link(call); proj; proj = get_irn_link(proj)) {
967 long proj_nr = get_Proj_proj(proj);
968 if (proj_nr == pn_Call_X_except) Xproj = proj;
969 if (proj_nr == pn_Call_M_except) Mproj = proj;
971 if (Mproj) { assert(Xproj); exc_handling = exc_handler; } /* Mproj */
972 else if (Xproj) { exc_handling = exc_to_end; } /* !Mproj && Xproj */
973 else { exc_handling = exc_no_handler; } /* !Mproj && !Xproj */
976 /* create the argument tuple */
977 NEW_ARR_A(ir_type *, args_in, n_params);
979 block = get_nodes_block(call);
980 for (i = n_params - 1; i >= 0; --i) {
981 ir_node *arg = get_Call_param(call, i);
982 ir_type *param_tp = get_method_param_type(mtp, i);
983 ir_mode *mode = get_type_mode(param_tp);
985 if (mode != get_irn_mode(arg)) {
986 arg = new_r_Conv(block, arg, mode);
992 the procedure and later replaces the Start node of the called graph.
993 Post_call is the old Call node and collects the results of the called
994 graph. Both will end up being a tuple. -- */
995 post_bl = get_nodes_block(call);
996 set_irg_current_block(irg, post_bl);
997 /* XxMxPxPxPxT of Start + parameter of Call */
998 in[pn_Start_X_initial_exec] = new_Jmp();
999 in[pn_Start_M] = get_Call_mem(call);
1000 in[pn_Start_P_frame_base] = get_irg_frame(irg);
1001 in[pn_Start_P_tls] = get_irg_tls(irg);
1002 in[pn_Start_T_args] = new_Tuple(n_params, args_in);
1003 pre_call = new_Tuple(pn_Start_max, in);
1007 The new block gets the ins of the old block, pre_call and all its
1008 predecessors and all Phi nodes. -- */
1009 part_block(pre_call);
1011 /* -- Prepare state for dead node elimination -- */
1012 /* Visited flags in calling irg must be >= flag in called irg.
1013 Else walker and arity computation will not work. */
1014 if (get_irg_visited(irg) <= get_irg_visited(called_graph))
1015 set_irg_visited(irg, get_irg_visited(called_graph) + 1);
1016 if (get_irg_block_visited(irg) < get_irg_block_visited(called_graph))
1017 set_irg_block_visited(irg, get_irg_block_visited(called_graph));
1018 visited = get_irg_visited(irg);
1020 /* Set pre_call as new Start node in link field of the start node of
1021 calling graph and pre_calls block as new block for the start block
1023 Further mark these nodes so that they are not visited by the
1025 set_irn_link(get_irg_start(called_graph), pre_call);
1026 set_irn_visited(get_irg_start(called_graph), visited);
1027 set_irn_link(get_irg_start_block(called_graph), get_nodes_block(pre_call));
1028 set_irn_visited(get_irg_start_block(called_graph), visited);
1030 set_irn_link(get_irg_bad(called_graph), get_irg_bad(current_ir_graph));
1031 set_irn_visited(get_irg_bad(called_graph), visited);
1033 set_irn_link(get_irg_no_mem(called_graph), get_irg_no_mem(current_ir_graph));
1034 set_irn_visited(get_irg_no_mem(called_graph), visited);
1036 /* Initialize for compaction of in arrays */
1037 inc_irg_block_visited(irg);
1039 /* -- Replicate local entities of the called_graph -- */
1040 /* copy the entities. */
1041 irp_reserve_resources(irp, IR_RESOURCE_ENTITY_LINK);
1042 called_frame = get_irg_frame_type(called_graph);
1043 curr_frame = get_irg_frame_type(irg);
1044 for (i = 0, n = get_class_n_members(called_frame); i < n; ++i) {
1045 ir_entity *new_ent, *old_ent;
1046 old_ent = get_class_member(called_frame, i);
1047 new_ent = copy_entity_own(old_ent, curr_frame);
1048 set_entity_link(old_ent, new_ent);
1051 /* visited is > than that of called graph. With this trick visited will
1052 remain unchanged so that an outer walker, e.g., searching the call nodes
1053 to inline, calling this inline will not visit the inlined nodes. */
1054 set_irg_visited(irg, get_irg_visited(irg)-1);
1056 /* -- Performing dead node elimination inlines the graph -- */
1057 /* Copies the nodes to the obstack of current_ir_graph. Updates links to new
1059 irg_walk(get_irg_end(called_graph), copy_node_inline, copy_preds_inline,
1060 get_irg_frame_type(called_graph));
1062 irp_free_resources(irp, IR_RESOURCE_ENTITY_LINK);
1064 /* Repair called_graph */
1065 set_irg_visited(called_graph, get_irg_visited(irg));
1066 set_irg_block_visited(called_graph, get_irg_block_visited(irg));
1067 set_Block_block_visited(get_irg_start_block(called_graph), 0);
1069 /* -- Merge the end of the inlined procedure with the call site -- */
1070 /* We will turn the old Call node into a Tuple with the following
1073 0: Phi of all Memories of Return statements.
1074 1: Jmp from new Block that merges the control flow from all exception
1075 predecessors of the old end block.
1076 2: Tuple of all arguments.
1077 3: Phi of Exception memories.
1078 In case the old Call directly branches to End on an exception we don't
1079 need the block merging all exceptions nor the Phi of the exception
1083 /* -- Precompute some values -- */
1084 end_bl = get_new_node(get_irg_end_block(called_graph));
1085 end = get_new_node(get_irg_end(called_graph));
1086 arity = get_Block_n_cfgpreds(end_bl); /* arity = n_exc + n_ret */
1087 n_res = get_method_n_ress(get_Call_type(call));
1089 res_pred = XMALLOCN(ir_node*, n_res);
1090 cf_pred = XMALLOCN(ir_node*, arity);
1092 set_irg_current_block(irg, post_bl); /* just to make sure */
1094 /* -- archive keepalives -- */
1095 irn_arity = get_irn_arity(end);
1096 for (i = 0; i < irn_arity; i++) {
1097 ir_node *ka = get_End_keepalive(end, i);
1099 add_End_keepalive(get_irg_end(irg), ka);
1102 /* The new end node will die. We need not free as the in array is on the obstack:
1103 copy_node() only generated 'D' arrays. */
1105 /* -- Replace Return nodes by Jump nodes. -- */
1107 for (i = 0; i < arity; i++) {
1109 ret = get_Block_cfgpred(end_bl, i);
1110 if (is_Return(ret)) {
1111 cf_pred[n_ret] = new_r_Jmp(get_nodes_block(ret));
1115 set_irn_in(post_bl, n_ret, cf_pred);
1117 /* -- Build a Tuple for all results of the method.
1118 Add Phi node if there was more than one Return. -- */
1119 turn_into_tuple(post_call, pn_Call_max);
1120 /* First the Memory-Phi */
1122 for (i = 0; i < arity; i++) {
1123 ret = get_Block_cfgpred(end_bl, i);
1124 if (is_Return(ret)) {
1125 cf_pred[n_ret] = get_Return_mem(ret);
1129 phi = new_Phi(n_ret, cf_pred, mode_M);
1130 set_Tuple_pred(call, pn_Call_M_regular, phi);
1131 /* Conserve Phi-list for further inlinings -- but might be optimized */
1132 if (get_nodes_block(phi) == post_bl) {
1133 set_irn_link(phi, get_irn_link(post_bl));
1134 set_irn_link(post_bl, phi);
1136 /* Now the real results */
1138 for (j = 0; j < n_res; j++) {
1139 ir_type *res_type = get_method_res_type(ctp, j);
1140 ir_mode *res_mode = get_type_mode(res_type);
1142 for (i = 0; i < arity; i++) {
1143 ret = get_Block_cfgpred(end_bl, i);
1144 if (is_Return(ret)) {
1145 ir_node *res = get_Return_res(ret, j);
1146 if (get_irn_mode(res) != res_mode) {
1147 ir_node *block = get_nodes_block(res);
1148 res = new_r_Conv(block, res, res_mode);
1150 cf_pred[n_ret] = res;
1155 phi = new_Phi(n_ret, cf_pred, get_irn_mode(cf_pred[0]));
1159 /* Conserve Phi-list for further inlinings -- but might be optimized */
1160 if (get_nodes_block(phi) == post_bl) {
1161 set_Phi_next(phi, get_Block_phis(post_bl));
1162 set_Block_phis(post_bl, phi);
1165 set_Tuple_pred(call, pn_Call_T_result, new_Tuple(n_res, res_pred));
1167 set_Tuple_pred(call, pn_Call_T_result, new_Bad());
1169 /* handle the regular call */
1170 set_Tuple_pred(call, pn_Call_X_regular, new_Jmp());
1172 /* For now, we cannot inline calls with value_base */
1173 set_Tuple_pred(call, pn_Call_P_value_res_base, new_Bad());
1175 /* Finally the exception control flow.
1176 We have two (three) possible situations:
1177 First if the Call branches to an exception handler: We need to add a Phi node to
1178 collect the memory containing the exception objects. Further we need
1179 to add another block to get a correct representation of this Phi. To
1180 this block we add a Jmp that resolves into the X output of the Call
1181 when the Call is turned into a tuple.
1182 Second the Call branches to End, the exception is not handled. Just
1183 add all inlined exception branches to the End node.
1184 Third: there is no Exception edge at all. Handle as case two. */
1185 if (exc_handling == exc_handler) {
1187 for (i = 0; i < arity; i++) {
1189 ret = get_Block_cfgpred(end_bl, i);
1190 irn = skip_Proj(ret);
1191 if (is_fragile_op(irn) || is_Raise(irn)) {
1192 cf_pred[n_exc] = ret;
1197 ir_node *block = new_Block(n_exc, cf_pred);
1198 set_cur_block(block);
1200 set_Tuple_pred(call, pn_Call_X_except, new_Jmp());
1201 /* The Phi for the memories with the exception objects */
1203 for (i = 0; i < arity; i++) {
1205 ret = skip_Proj(get_Block_cfgpred(end_bl, i));
1207 cf_pred[n_exc] = new_r_Proj(get_nodes_block(ret), ret, mode_M, 3);
1209 } else if (is_fragile_op(ret)) {
1210 /* We rely that all cfops have the memory output at the same position. */
1211 cf_pred[n_exc] = new_r_Proj(get_nodes_block(ret), ret, mode_M, 0);
1213 } else if (is_Raise(ret)) {
1214 cf_pred[n_exc] = new_r_Proj(get_nodes_block(ret), ret, mode_M, 1);
1218 set_Tuple_pred(call, pn_Call_M_except, new_Phi(n_exc, cf_pred, mode_M));
1220 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1221 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1224 ir_node *main_end_bl;
1225 int main_end_bl_arity;
1226 ir_node **end_preds;
1228 /* assert(exc_handling == 1 || no exceptions. ) */
1230 for (i = 0; i < arity; i++) {
1231 ir_node *ret = get_Block_cfgpred(end_bl, i);
1232 ir_node *irn = skip_Proj(ret);
1234 if (is_fragile_op(irn) || is_Raise(irn)) {
1235 cf_pred[n_exc] = ret;
1239 main_end_bl = get_irg_end_block(irg);
1240 main_end_bl_arity = get_irn_arity(main_end_bl);
1241 end_preds = XMALLOCN(ir_node*, n_exc + main_end_bl_arity);
1243 for (i = 0; i < main_end_bl_arity; ++i)
1244 end_preds[i] = get_irn_n(main_end_bl, i);
1245 for (i = 0; i < n_exc; ++i)
1246 end_preds[main_end_bl_arity + i] = cf_pred[i];
1247 set_irn_in(main_end_bl, n_exc + main_end_bl_arity, end_preds);
1248 set_Tuple_pred(call, pn_Call_X_except, new_Bad());
1249 set_Tuple_pred(call, pn_Call_M_except, new_Bad());
1255 /* -- Turn CSE back on. -- */
1256 set_optimize(rem_opt);
1257 current_ir_graph = rem;
1262 /********************************************************************/
1263 /* Apply inlining to small methods. */
1264 /********************************************************************/
1266 static struct obstack temp_obst;
1268 /** Represents a possible inlinable call in a graph. */
1269 typedef struct _call_entry {
1270 ir_node *call; /**< The Call node. */
1271 ir_graph *callee; /**< The callee IR-graph. */
1272 list_head list; /**< List head for linking the next one. */
1273 int loop_depth; /**< The loop depth of this call. */
1274 int benefice; /**< The calculated benefice of this call. */
1275 unsigned local_adr:1; /**< Set if this call gets an address of a local variable. */
1276 unsigned all_const:1; /**< Set if this call has only constant parameters. */
1280 * environment for inlining small irgs
1282 typedef struct _inline_env_t {
1283 struct obstack obst; /**< An obstack where call_entries are allocated on. */
1284 list_head calls; /**< The call entry list. */
1288 * Returns the irg called from a Call node. If the irg is not
1289 * known, NULL is returned.
1291 * @param call the call node
1293 static ir_graph *get_call_called_irg(ir_node *call) {
1296 addr = get_Call_ptr(call);
1297 if (is_Global(addr)) {
1298 ir_entity *ent = get_Global_entity(addr);
1299 return get_entity_irg(ent);
1306 * Walker: Collect all calls to known graphs inside a graph.
1308 static void collect_calls(ir_node *call, void *env) {
1310 if (is_Call(call)) {
1311 ir_graph *called_irg = get_call_called_irg(call);
1313 if (called_irg != NULL) {
1314 /* The Call node calls a locally defined method. Remember to inline. */
1315 inline_env_t *ienv = env;
1316 call_entry *entry = obstack_alloc(&ienv->obst, sizeof(*entry));
1318 entry->callee = called_irg;
1319 entry->loop_depth = 0;
1320 entry->benefice = 0;
1321 entry->local_adr = 0;
1322 entry->all_const = 0;
1324 list_add_tail(&entry->list, &ienv->calls);
1330 * Inlines all small methods at call sites where the called address comes
1331 * from a Const node that references the entity representing the called
1333 * The size argument is a rough measure for the code size of the method:
1334 * Methods where the obstack containing the firm graph is smaller than
1337 void inline_small_irgs(ir_graph *irg, int size) {
1338 ir_graph *rem = current_ir_graph;
1342 current_ir_graph = irg;
1343 /* Handle graph state */
1344 assert(get_irg_phase_state(irg) != phase_building);
1345 free_callee_info(irg);
1347 /* Find Call nodes to inline.
1348 (We can not inline during a walk of the graph, as inlining the same
1349 method several times changes the visited flag of the walked graph:
1350 after the first inlining visited of the callee equals visited of
1351 the caller. With the next inlining both are increased.) */
1352 obstack_init(&env.obst);
1353 INIT_LIST_HEAD(&env.calls);
1354 irg_walk_graph(irg, NULL, collect_calls, &env);
1356 if (! list_empty(&env.calls)) {
1357 /* There are calls to inline */
1358 ir_reserve_resources(irg, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
1359 collect_phiprojs(irg);
1361 list_for_each_entry(call_entry, entry, &env.calls, list) {
1362 ir_graph *callee = entry->callee;
1363 irg_inline_property prop = get_irg_inline_property(callee);
1365 if (prop == irg_inline_forbidden || get_irg_additional_properties(callee) & mtp_property_weak) {
1366 /* do not inline forbidden / weak graphs */
1370 if (prop >= irg_inline_forced ||
1371 _obstack_memory_used(callee->obst) - (int)obstack_room(callee->obst) < size) {
1372 inline_method(entry->call, callee);
1375 ir_free_resources(irg, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
1377 obstack_free(&env.obst, NULL);
1378 current_ir_graph = rem;
1382 * Environment for inlining irgs.
1385 list_head calls; /**< List of of all call nodes in this graph. */
1386 unsigned *local_weights; /**< Once allocated, the beneficial weight for transmitting local addresses. */
1387 unsigned n_nodes; /**< Number of nodes in graph except Id, Tuple, Proj, Start, End. */
1388 unsigned n_blocks; /**< Number of Blocks in graph without Start and End block. */
1389 unsigned n_nodes_orig; /**< for statistics */
1390 unsigned n_call_nodes; /**< Number of Call nodes in the graph. */
1391 unsigned n_call_nodes_orig; /**< for statistics */
1392 unsigned n_callers; /**< Number of known graphs that call this graphs. */
1393 unsigned n_callers_orig; /**< for statistics */
1394 unsigned got_inline:1; /**< Set, if at least one call inside this graph was inlined. */
1395 unsigned local_vars:1; /**< Set, if an inlined function got the address of a local variable. */
1396 unsigned recursive:1; /**< Set, if this function is self recursive. */
1400 * Allocate a new environment for inlining.
1402 static inline_irg_env *alloc_inline_irg_env(void) {
1403 inline_irg_env *env = obstack_alloc(&temp_obst, sizeof(*env));
1404 INIT_LIST_HEAD(&env->calls);
1405 env->local_weights = NULL;
1406 env->n_nodes = -2; /* do not count count Start, End */
1407 env->n_blocks = -2; /* do not count count Start, End Block */
1408 env->n_nodes_orig = -2; /* do not count Start, End */
1409 env->n_call_nodes = 0;
1410 env->n_call_nodes_orig = 0;
1412 env->n_callers_orig = 0;
1413 env->got_inline = 0;
1414 env->local_vars = 0;
1419 typedef struct walker_env {
1420 inline_irg_env *x; /**< the inline environment */
1421 char ignore_runtime; /**< the ignore runtime flag */
1422 char ignore_callers; /**< if set, do change callers data */
1426 * post-walker: collect all calls in the inline-environment
1427 * of a graph and sum some statistics.
1429 static void collect_calls2(ir_node *call, void *ctx) {
1431 inline_irg_env *x = env->x;
1432 ir_opcode code = get_irn_opcode(call);
1436 /* count meaningful nodes in irg */
1437 if (code != iro_Proj && code != iro_Tuple && code != iro_Sync) {
1438 if (code != iro_Block) {
1446 if (code != iro_Call) return;
1448 /* check, if it's a runtime call */
1449 if (env->ignore_runtime) {
1450 ir_node *symc = get_Call_ptr(call);
1452 if (is_Global(symc)) {
1453 ir_entity *ent = get_Global_entity(symc);
1455 if (get_entity_additional_properties(ent) & mtp_property_runtime)
1460 /* collect all call nodes */
1462 ++x->n_call_nodes_orig;
1464 callee = get_call_called_irg(call);
1465 if (callee != NULL) {
1466 if (! env->ignore_callers) {
1467 inline_irg_env *callee_env = get_irg_link(callee);
1468 /* count all static callers */
1469 ++callee_env->n_callers;
1470 ++callee_env->n_callers_orig;
1472 if (callee == current_ir_graph)
1475 /* link it in the list of possible inlinable entries */
1476 entry = obstack_alloc(&temp_obst, sizeof(*entry));
1478 entry->callee = callee;
1479 entry->loop_depth = get_irn_loop(get_nodes_block(call))->depth;
1480 entry->benefice = 0;
1481 entry->local_adr = 0;
1482 entry->all_const = 0;
1484 list_add_tail(&entry->list, &x->calls);
1489 * Returns TRUE if the number of callers is 0 in the irg's environment,
1490 * hence this irg is a leave.
1492 inline static int is_leave(ir_graph *irg) {
1493 inline_irg_env *env = get_irg_link(irg);
1494 return env->n_call_nodes == 0;
1498 * Returns TRUE if the number of nodes in the callee is
1499 * smaller then size in the irg's environment.
1501 inline static int is_smaller(ir_graph *callee, unsigned size) {
1502 inline_irg_env *env = get_irg_link(callee);
1503 return env->n_nodes < size;
1507 * Duplicate a call entry.
1509 * @param entry the original entry to duplicate
1510 * @param new_call the new call node
1511 * @param loop_depth_delta
1512 * delta value for the loop depth
1514 static call_entry *duplicate_call_entry(const call_entry *entry,
1515 ir_node *new_call, int loop_depth_delta) {
1516 call_entry *nentry = obstack_alloc(&temp_obst, sizeof(*nentry));
1517 nentry->call = new_call;
1518 nentry->callee = entry->callee;
1519 nentry->benefice = entry->benefice;
1520 nentry->loop_depth = entry->loop_depth + loop_depth_delta;
1521 nentry->local_adr = entry->local_adr;
1522 nentry->all_const = entry->all_const;
1528 * Append all call nodes of the source environment to the nodes of in the destination
1531 * @param dst destination environment
1532 * @param src source environment
1533 * @param loop_depth the loop depth of the call that is replaced by the src list
1535 static void append_call_list(inline_irg_env *dst, inline_irg_env *src, int loop_depth) {
1536 call_entry *entry, *nentry;
1538 /* Note that the src list points to Call nodes in the inlined graph, but
1539 we need Call nodes in our graph. Luckily the inliner leaves this information
1540 in the link field. */
1541 list_for_each_entry(call_entry, entry, &src->calls, list) {
1542 nentry = duplicate_call_entry(entry, get_irn_link(entry->call), loop_depth);
1543 list_add_tail(&nentry->list, &dst->calls);
1545 dst->n_call_nodes += src->n_call_nodes;
1546 dst->n_nodes += src->n_nodes;
1550 * Inlines small leave methods at call sites where the called address comes
1551 * from a Const node that references the entity representing the called
1553 * The size argument is a rough measure for the code size of the method:
1554 * Methods where the obstack containing the firm graph is smaller than
1557 void inline_leave_functions(unsigned maxsize, unsigned leavesize,
1558 unsigned size, int ignore_runtime)
1560 inline_irg_env *env;
1566 call_entry *entry, *next;
1567 const call_entry *centry;
1568 pmap *copied_graphs;
1569 pmap_entry *pm_entry;
1571 rem = current_ir_graph;
1572 obstack_init(&temp_obst);
1574 /* a map for the copied graphs, used to inline recursive calls */
1575 copied_graphs = pmap_create();
1577 /* extend all irgs by a temporary data structure for inlining. */
1578 n_irgs = get_irp_n_irgs();
1579 for (i = 0; i < n_irgs; ++i)
1580 set_irg_link(get_irp_irg(i), alloc_inline_irg_env());
1582 /* Pre-compute information in temporary data structure. */
1583 wenv.ignore_runtime = ignore_runtime;
1584 wenv.ignore_callers = 0;
1585 for (i = 0; i < n_irgs; ++i) {
1586 ir_graph *irg = get_irp_irg(i);
1588 assert(get_irg_phase_state(irg) != phase_building);
1589 free_callee_info(irg);
1591 assure_cf_loop(irg);
1592 wenv.x = get_irg_link(irg);
1593 irg_walk_graph(irg, NULL, collect_calls2, &wenv);
1596 /* -- and now inline. -- */
1598 /* Inline leaves recursively -- we might construct new leaves. */
1602 for (i = 0; i < n_irgs; ++i) {
1604 int phiproj_computed = 0;
1606 current_ir_graph = get_irp_irg(i);
1607 env = get_irg_link(current_ir_graph);
1609 ir_reserve_resources(current_ir_graph, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
1610 list_for_each_entry_safe(call_entry, entry, next, &env->calls, list) {
1612 irg_inline_property prop;
1614 if (env->n_nodes > maxsize)
1618 callee = entry->callee;
1620 prop = get_irg_inline_property(callee);
1621 if (prop == irg_inline_forbidden || get_irg_additional_properties(callee) & mtp_property_weak) {
1622 /* do not inline forbidden / weak graphs */
1626 if (is_leave(callee) && (
1627 is_smaller(callee, leavesize) || prop >= irg_inline_forced)) {
1628 if (!phiproj_computed) {
1629 phiproj_computed = 1;
1630 collect_phiprojs(current_ir_graph);
1632 did_inline = inline_method(call, callee);
1635 inline_irg_env *callee_env = get_irg_link(callee);
1637 /* call was inlined, Phi/Projs for current graph must be recomputed */
1638 phiproj_computed = 0;
1640 /* Do some statistics */
1641 env->got_inline = 1;
1642 --env->n_call_nodes;
1643 env->n_nodes += callee_env->n_nodes;
1644 --callee_env->n_callers;
1646 /* remove this call from the list */
1647 list_del(&entry->list);
1652 ir_free_resources(current_ir_graph, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
1654 } while (did_inline);
1656 /* inline other small functions. */
1657 for (i = 0; i < n_irgs; ++i) {
1659 int phiproj_computed = 0;
1661 current_ir_graph = get_irp_irg(i);
1662 env = get_irg_link(current_ir_graph);
1664 ir_reserve_resources(current_ir_graph, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
1666 /* note that the list of possible calls is updated during the process */
1667 list_for_each_entry_safe(call_entry, entry, next, &env->calls, list) {
1668 irg_inline_property prop;
1673 callee = entry->callee;
1675 prop = get_irg_inline_property(callee);
1676 if (prop == irg_inline_forbidden || get_irg_additional_properties(callee) & mtp_property_weak) {
1677 /* do not inline forbidden / weak graphs */
1681 e = pmap_find(copied_graphs, callee);
1684 * Remap callee if we have a copy.
1685 * FIXME: Should we do this only for recursive Calls ?
1690 if (prop >= irg_inline_forced ||
1691 (is_smaller(callee, size) && env->n_nodes < maxsize) /* small function */) {
1692 if (current_ir_graph == callee) {
1694 * Recursive call: we cannot directly inline because we cannot walk
1695 * the graph and change it. So we have to make a copy of the graph
1699 inline_irg_env *callee_env;
1702 ir_free_resources(current_ir_graph, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
1705 * No copy yet, create one.
1706 * Note that recursive methods are never leaves, so it is sufficient
1707 * to test this condition here.
1709 copy = create_irg_copy(callee);
1711 /* create_irg_copy() destroys the Proj links, recompute them */
1712 phiproj_computed = 0;
1714 ir_reserve_resources(current_ir_graph, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
1716 /* allocate new environment */
1717 callee_env = alloc_inline_irg_env();
1718 set_irg_link(copy, callee_env);
1720 assure_cf_loop(copy);
1721 wenv.x = callee_env;
1722 wenv.ignore_callers = 1;
1723 irg_walk_graph(copy, NULL, collect_calls2, &wenv);
1726 * Enter the entity of the original graph. This is needed
1727 * for inline_method(). However, note that ent->irg still points
1728 * to callee, NOT to copy.
1730 set_irg_entity(copy, get_irg_entity(callee));
1732 pmap_insert(copied_graphs, callee, copy);
1735 /* we have only one caller: the original graph */
1736 callee_env->n_callers = 1;
1737 callee_env->n_callers_orig = 1;
1739 if (! phiproj_computed) {
1740 phiproj_computed = 1;
1741 collect_phiprojs(current_ir_graph);
1743 did_inline = inline_method(call, callee);
1745 inline_irg_env *callee_env = (inline_irg_env *)get_irg_link(callee);
1747 /* call was inlined, Phi/Projs for current graph must be recomputed */
1748 phiproj_computed = 0;
1750 /* callee was inline. Append it's call list. */
1751 env->got_inline = 1;
1752 --env->n_call_nodes;
1753 append_call_list(env, callee_env, entry->loop_depth);
1754 --callee_env->n_callers;
1756 /* after we have inlined callee, all called methods inside callee
1757 are now called once more */
1758 list_for_each_entry(call_entry, centry, &callee_env->calls, list) {
1759 inline_irg_env *penv = get_irg_link(centry->callee);
1763 /* remove this call from the list */
1764 list_del(&entry->list);
1769 ir_free_resources(current_ir_graph, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
1772 for (i = 0; i < n_irgs; ++i) {
1773 irg = get_irp_irg(i);
1774 env = get_irg_link(irg);
1776 if (env->got_inline) {
1777 optimize_graph_df(irg);
1780 if (env->got_inline || (env->n_callers_orig != env->n_callers)) {
1781 DB((dbg, LEVEL_1, "Nodes:%3d ->%3d, calls:%3d ->%3d, callers:%3d ->%3d, -- %s\n",
1782 env->n_nodes_orig, env->n_nodes, env->n_call_nodes_orig, env->n_call_nodes,
1783 env->n_callers_orig, env->n_callers,
1784 get_entity_name(get_irg_entity(irg))));
1788 /* kill the copied graphs: we don't need them anymore */
1789 foreach_pmap(copied_graphs, pm_entry) {
1790 ir_graph *copy = pm_entry->value;
1792 /* reset the entity, otherwise it will be deleted in the next step ... */
1793 set_irg_entity(copy, NULL);
1794 free_ir_graph(copy);
1796 pmap_destroy(copied_graphs);
1798 obstack_free(&temp_obst, NULL);
1799 current_ir_graph = rem;
1803 * Calculate the parameter weights for transmitting the address of a local variable.
1805 static unsigned calc_method_local_weight(ir_node *arg) {
1807 unsigned v, weight = 0;
1809 for (i = get_irn_n_outs(arg) - 1; i >= 0; --i) {
1810 ir_node *succ = get_irn_out(arg, i);
1812 switch (get_irn_opcode(succ)) {
1815 /* Loads and Store can be removed */
1819 /* check if all args are constant */
1820 for (j = get_Sel_n_indexs(succ) - 1; j >= 0; --j) {
1821 ir_node *idx = get_Sel_index(succ, j);
1822 if (! is_Const(idx))
1825 /* Check users on this Sel. Note: if a 0 is returned here, there was
1826 some unsupported node. */
1827 v = calc_method_local_weight(succ);
1830 /* we can kill one Sel with constant indexes, this is cheap */
1834 /* when looking backward we might find Id nodes */
1835 weight += calc_method_local_weight(succ);
1838 /* unoptimized tuple */
1839 for (j = get_Tuple_n_preds(succ) - 1; j >= 0; --j) {
1840 ir_node *pred = get_Tuple_pred(succ, j);
1842 /* look for Proj(j) */
1843 for (k = get_irn_n_outs(succ) - 1; k >= 0; --k) {
1844 ir_node *succ_succ = get_irn_out(succ, k);
1845 if (is_Proj(succ_succ)) {
1846 if (get_Proj_proj(succ_succ) == j) {
1848 weight += calc_method_local_weight(succ_succ);
1851 /* this should NOT happen */
1859 /* any other node: unsupported yet or bad. */
1867 * Calculate the parameter weights for transmitting the address of a local variable.
1869 static void analyze_irg_local_weights(inline_irg_env *env, ir_graph *irg) {
1870 ir_entity *ent = get_irg_entity(irg);
1872 int nparams, i, proj_nr;
1873 ir_node *irg_args, *arg;
1875 mtp = get_entity_type(ent);
1876 nparams = get_method_n_params(mtp);
1878 /* allocate a new array. currently used as 'analysed' flag */
1879 env->local_weights = NEW_ARR_D(unsigned, &temp_obst, nparams);
1881 /* If the method haven't parameters we have nothing to do. */
1885 assure_irg_outs(irg);
1886 irg_args = get_irg_args(irg);
1887 for (i = get_irn_n_outs(irg_args) - 1; i >= 0; --i) {
1888 arg = get_irn_out(irg_args, i);
1889 proj_nr = get_Proj_proj(arg);
1890 env->local_weights[proj_nr] = calc_method_local_weight(arg);
1895 * Calculate the benefice for transmitting an local variable address.
1896 * After inlining, the local variable might be transformed into a
1897 * SSA variable by scalar_replacement().
1899 static unsigned get_method_local_adress_weight(ir_graph *callee, int pos) {
1900 inline_irg_env *env = get_irg_link(callee);
1902 if (env->local_weights != NULL) {
1903 if (pos < ARR_LEN(env->local_weights))
1904 return env->local_weights[pos];
1908 analyze_irg_local_weights(env, callee);
1910 if (pos < ARR_LEN(env->local_weights))
1911 return env->local_weights[pos];
1916 * Calculate a benefice value for inlining the given call.
1918 * @param call the call node we have to inspect
1919 * @param callee the called graph
1921 static int calc_inline_benefice(call_entry *entry, ir_graph *callee)
1923 ir_node *call = entry->call;
1924 ir_entity *ent = get_irg_entity(callee);
1928 int i, n_params, all_const;
1930 irg_inline_property prop;
1932 inline_irg_env *callee_env;
1934 prop = get_irg_inline_property(callee);
1935 if (prop == irg_inline_forbidden) {
1936 DB((dbg, LEVEL_2, "In %+F Call to %+F: inlining forbidden\n",
1938 return entry->benefice = INT_MIN;
1941 if (get_irg_additional_properties(callee) & (mtp_property_noreturn | mtp_property_weak)) {
1942 DB((dbg, LEVEL_2, "In %+F Call to %+F: not inlining noreturn or weak\n",
1944 return entry->benefice = INT_MIN;
1947 /* costs for every passed parameter */
1948 n_params = get_Call_n_params(call);
1949 mtp = get_entity_type(ent);
1950 cc = get_method_calling_convention(mtp);
1951 if (cc & cc_reg_param) {
1952 /* register parameter, smaller costs for register parameters */
1953 int max_regs = cc & ~cc_bits;
1955 if (max_regs < n_params)
1956 weight += max_regs * 2 + (n_params - max_regs) * 5;
1958 weight += n_params * 2;
1960 /* parameters are passed an stack */
1961 weight += 5 * n_params;
1964 /* constant parameters improve the benefice */
1965 frame_ptr = get_irg_frame(current_ir_graph);
1967 for (i = 0; i < n_params; ++i) {
1968 ir_node *param = get_Call_param(call, i);
1970 if (is_Const(param)) {
1971 weight += get_method_param_weight(ent, i);
1974 if (is_SymConst(param))
1975 weight += get_method_param_weight(ent, i);
1976 else if (is_Sel(param) && get_Sel_ptr(param) == frame_ptr) {
1978 * An address of a local variable is transmitted. After
1979 * inlining, scalar_replacement might be able to remove the
1980 * local variable, so honor this.
1982 v = get_method_local_adress_weight(callee, i);
1985 entry->local_adr = 1;
1989 entry->all_const = all_const;
1991 callee_env = get_irg_link(callee);
1992 if (callee_env->n_callers == 1 &&
1993 callee != current_ir_graph &&
1994 get_entity_visibility(ent) == visibility_local) {
1998 /* give a bonus for functions with one block */
1999 if (callee_env->n_blocks == 1)
2000 weight = weight * 3 / 2;
2002 /* and one for small non-recursive functions: we want them to be inlined in mostly every case */
2003 if (callee_env->n_nodes < 30 && !callee_env->recursive)
2006 /* and finally for leaves: they do not increase the register pressure
2007 because of callee safe registers */
2008 if (callee_env->n_call_nodes == 0)
2011 /** it's important to inline inner loops first */
2012 if (entry->loop_depth > 30)
2013 weight += 30 * 1024;
2015 weight += entry->loop_depth * 1024;
2018 * All arguments constant is probably a good sign, give an extra bonus
2023 return entry->benefice = weight;
2026 static ir_graph **irgs;
2027 static int last_irg;
2030 * Callgraph walker, collect all visited graphs.
2032 static void callgraph_walker(ir_graph *irg, void *data) {
2034 irgs[last_irg++] = irg;
2038 * Creates an inline order for all graphs.
2040 * @return the list of graphs.
2042 static ir_graph **create_irg_list(void) {
2043 ir_entity **free_methods;
2045 int n_irgs = get_irp_n_irgs();
2047 cgana(&arr_len, &free_methods);
2048 xfree(free_methods);
2050 compute_callgraph();
2053 irgs = XMALLOCNZ(ir_graph*, n_irgs);
2055 callgraph_walk(NULL, callgraph_walker, NULL);
2056 assert(n_irgs == last_irg);
2062 * Push a call onto the priority list if its benefice is big enough.
2064 * @param pqueue the priority queue of calls
2065 * @param call the call entry
2066 * @param inlien_threshold
2067 * the threshold value
2069 static void maybe_push_call(pqueue_t *pqueue, call_entry *call,
2070 int inline_threshold)
2072 ir_graph *callee = call->callee;
2073 irg_inline_property prop = get_irg_inline_property(callee);
2074 int benefice = calc_inline_benefice(call, callee);
2076 DB((dbg, LEVEL_2, "In %+F Call %+F to %+F has benefice %d\n",
2077 get_irn_irg(call->call), call->call, callee, benefice));
2079 if (prop < irg_inline_forced && benefice < inline_threshold) {
2083 pqueue_put(pqueue, call, benefice);
2087 * Try to inline calls into a graph.
2089 * @param irg the graph into which we inline
2090 * @param maxsize do NOT inline if the size of irg gets
2091 * bigger than this amount
2092 * @param inline_threshold
2093 * threshold value for inline decision
2094 * @param copied_graphs
2095 * map containing copied of recursive graphs
2097 static void inline_into(ir_graph *irg, unsigned maxsize,
2098 int inline_threshold, pmap *copied_graphs)
2100 int phiproj_computed = 0;
2101 inline_irg_env *env = get_irg_link(irg);
2102 call_entry *curr_call;
2106 if (env->n_call_nodes == 0)
2109 if (env->n_nodes > maxsize) {
2110 DB((dbg, LEVEL_2, "%+F: too big (%d)\n", irg, env->n_nodes));
2114 current_ir_graph = irg;
2115 ir_reserve_resources(irg, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
2117 /* put irgs into the pqueue */
2118 pqueue = new_pqueue();
2120 list_for_each_entry(call_entry, curr_call, &env->calls, list) {
2121 assert(is_Call(curr_call->call));
2122 maybe_push_call(pqueue, curr_call, inline_threshold);
2125 /* note that the list of possible calls is updated during the process */
2126 while (!pqueue_empty(pqueue)) {
2128 call_entry *curr_call = pqueue_pop_front(pqueue);
2129 ir_graph *callee = curr_call->callee;
2130 ir_node *call_node = curr_call->call;
2131 inline_irg_env *callee_env = get_irg_link(callee);
2132 irg_inline_property prop = get_irg_inline_property(callee);
2134 const call_entry *centry;
2137 if ((prop < irg_inline_forced) && env->n_nodes + callee_env->n_nodes > maxsize) {
2138 DB((dbg, LEVEL_2, "%+F: too big (%d) + %+F (%d)\n", irg,
2139 env->n_nodes, callee, callee_env->n_nodes));
2143 e = pmap_find(copied_graphs, callee);
2145 int benefice = curr_call->benefice;
2147 * Reduce the weight for recursive function IFF not all arguments are const.
2148 * inlining recursive functions is rarely good.
2150 if (!curr_call->all_const)
2152 if (benefice < inline_threshold)
2156 * Remap callee if we have a copy.
2159 callee_env = get_irg_link(callee);
2162 if (current_ir_graph == callee) {
2164 * Recursive call: we cannot directly inline because we cannot
2165 * walk the graph and change it. So we have to make a copy of
2168 int benefice = curr_call->benefice;
2172 * Reduce the weight for recursive function IFF not all arguments are const.
2173 * inlining recursive functions is rarely good.
2175 if (!curr_call->all_const)
2177 if (benefice < inline_threshold)
2180 ir_free_resources(irg, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
2183 * No copy yet, create one.
2184 * Note that recursive methods are never leaves, so it is
2185 * sufficient to test this condition here.
2187 copy = create_irg_copy(callee);
2189 /* create_irg_copy() destroys the Proj links, recompute them */
2190 phiproj_computed = 0;
2192 ir_reserve_resources(irg, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
2194 /* allocate a new environment */
2195 callee_env = alloc_inline_irg_env();
2196 set_irg_link(copy, callee_env);
2198 assure_cf_loop(copy);
2199 wenv.x = callee_env;
2200 wenv.ignore_callers = 1;
2201 irg_walk_graph(copy, NULL, collect_calls2, &wenv);
2204 * Enter the entity of the original graph. This is needed
2205 * for inline_method(). However, note that ent->irg still points
2206 * to callee, NOT to copy.
2208 set_irg_entity(copy, get_irg_entity(callee));
2210 pmap_insert(copied_graphs, callee, copy);
2213 /* we have only one caller: the original graph */
2214 callee_env->n_callers = 1;
2215 callee_env->n_callers_orig = 1;
2217 if (! phiproj_computed) {
2218 phiproj_computed = 1;
2219 collect_phiprojs(current_ir_graph);
2221 did_inline = inline_method(call_node, callee);
2225 /* call was inlined, Phi/Projs for current graph must be recomputed */
2226 phiproj_computed = 0;
2228 /* remove it from the caller list */
2229 list_del(&curr_call->list);
2231 /* callee was inline. Append it's call list. */
2232 env->got_inline = 1;
2233 if (curr_call->local_adr)
2234 env->local_vars = 1;
2235 --env->n_call_nodes;
2237 /* we just generate a bunch of new calls */
2238 loop_depth = curr_call->loop_depth;
2239 list_for_each_entry(call_entry, centry, &callee_env->calls, list) {
2240 inline_irg_env *penv = get_irg_link(centry->callee);
2242 call_entry *new_entry;
2244 /* after we have inlined callee, all called methods inside
2245 * callee are now called once more */
2248 /* Note that the src list points to Call nodes in the inlined graph,
2249 * but we need Call nodes in our graph. Luckily the inliner leaves
2250 * this information in the link field. */
2251 new_call = get_irn_link(centry->call);
2252 assert(is_Call(new_call));
2254 new_entry = duplicate_call_entry(centry, new_call, loop_depth);
2255 list_add_tail(&new_entry->list, &env->calls);
2256 maybe_push_call(pqueue, new_entry, inline_threshold);
2259 env->n_call_nodes += callee_env->n_call_nodes;
2260 env->n_nodes += callee_env->n_nodes;
2261 --callee_env->n_callers;
2263 ir_free_resources(irg, IR_RESOURCE_IRN_LINK|IR_RESOURCE_PHI_LIST);
2268 * Heuristic inliner. Calculates a benefice value for every call and inlines
2269 * those calls with a value higher than the threshold.
2271 void inline_functions(unsigned maxsize, int inline_threshold) {
2272 inline_irg_env *env;
2276 pmap *copied_graphs;
2277 pmap_entry *pm_entry;
2280 rem = current_ir_graph;
2281 obstack_init(&temp_obst);
2283 irgs = create_irg_list();
2285 /* a map for the copied graphs, used to inline recursive calls */
2286 copied_graphs = pmap_create();
2288 /* extend all irgs by a temporary data structure for inlining. */
2289 n_irgs = get_irp_n_irgs();
2290 for (i = 0; i < n_irgs; ++i)
2291 set_irg_link(irgs[i], alloc_inline_irg_env());
2293 /* Pre-compute information in temporary data structure. */
2294 wenv.ignore_runtime = 0;
2295 wenv.ignore_callers = 0;
2296 for (i = 0; i < n_irgs; ++i) {
2297 ir_graph *irg = irgs[i];
2299 free_callee_info(irg);
2301 wenv.x = get_irg_link(irg);
2302 assure_cf_loop(irg);
2303 irg_walk_graph(irg, NULL, collect_calls2, &wenv);
2306 /* -- and now inline. -- */
2307 for (i = 0; i < n_irgs; ++i) {
2308 ir_graph *irg = irgs[i];
2310 inline_into(irg, maxsize, inline_threshold, copied_graphs);
2313 for (i = 0; i < n_irgs; ++i) {
2314 ir_graph *irg = irgs[i];
2316 env = get_irg_link(irg);
2317 if (env->got_inline) {
2318 /* this irg got calls inlined: optimize it */
2319 if (get_opt_combo()) {
2320 if (env->local_vars) {
2321 scalar_replacement_opt(irg);
2325 if (env->local_vars) {
2326 if (scalar_replacement_opt(irg)) {
2327 optimize_graph_df(irg);
2333 if (env->got_inline || (env->n_callers_orig != env->n_callers)) {
2334 DB((dbg, LEVEL_1, "Nodes:%3d ->%3d, calls:%3d ->%3d, callers:%3d ->%3d, -- %s\n",
2335 env->n_nodes_orig, env->n_nodes, env->n_call_nodes_orig, env->n_call_nodes,
2336 env->n_callers_orig, env->n_callers,
2337 get_entity_name(get_irg_entity(irg))));
2341 /* kill the copied graphs: we don't need them anymore */
2342 foreach_pmap(copied_graphs, pm_entry) {
2343 ir_graph *copy = pm_entry->value;
2345 /* reset the entity, otherwise it will be deleted in the next step ... */
2346 set_irg_entity(copy, NULL);
2347 free_ir_graph(copy);
2349 pmap_destroy(copied_graphs);
2353 obstack_free(&temp_obst, NULL);
2354 current_ir_graph = rem;
2357 void firm_init_inline(void) {
2358 FIRM_DBG_REGISTER(dbg, "firm.opt.inline");