1 /* Coyright (C) 1998 - 2000 by Universitaet Karlsruhe
2 ** All rights reserved.
4 ** Author: Christian Schaefer
6 ** Optimizations for a whole ir graph, i.e., a procedure.
19 # include "irnode_t.h"
20 # include "irgraph_t.h"
28 # include "pdeq.h" /* Fuer code placement */
31 /* Defined in iropt.c */
32 pset *new_identities (void);
33 void del_identities (pset *value_table);
34 void add_identities (pset *value_table, ir_node *node);
36 /********************************************************************/
37 /* apply optimizations of iropt to all nodes. */
38 /********************************************************************/
40 void init_link (ir_node *n, void *env) {
41 set_irn_link(n, NULL);
45 optimize_in_place_wrapper (ir_node *n, void *env) {
49 for (i = 0; i < get_irn_arity(n); i++) {
50 optimized = optimize_in_place_2(get_irn_n(n, i));
51 set_irn_n(n, i, optimized);
54 if (get_irn_op(n) == op_Block) {
55 optimized = optimize_in_place_2(n);
56 if (optimized != n) exchange (n, optimized);
61 local_optimize_graph (ir_graph *irg) {
62 ir_graph *rem = current_ir_graph;
63 current_ir_graph = irg;
65 /* Handle graph state */
66 assert(get_irg_phase_state(irg) != phase_building);
67 if (get_opt_global_cse())
68 set_irg_pinned(current_ir_graph, floats);
69 if (get_irg_outs_state(current_ir_graph) == outs_consistent)
70 set_irg_outs_inconsistent(current_ir_graph);
71 if (get_irg_dom_state(current_ir_graph) == dom_consistent)
72 set_irg_dom_inconsistent(current_ir_graph);
74 /* Clean the value_table in irg for the cse. */
75 del_identities(irg->value_table);
76 irg->value_table = new_identities();
78 /* walk over the graph */
79 irg_walk(irg->end, init_link, optimize_in_place_wrapper, NULL);
81 current_ir_graph = rem;
84 /********************************************************************/
85 /* Routines for dead node elimination / copying garbage collection */
87 /********************************************************************/
89 /* Remeber the new node in the old node by using a field all nodes have. */
91 set_new_node (ir_node *old, ir_node *new)
96 /* Get this new node, before the old node is forgotton.*/
98 get_new_node (ir_node * n)
103 /* We use the block_visited flag to mark that we have computed the
104 number of useful predecessors for this block.
105 Further we encode the new arity in this flag in the old blocks.
106 Remembering the arity is useful, as it saves a lot of pointer
107 accesses. This function is called for all Phi and Block nodes
110 compute_new_arity(ir_node *b) {
114 irg_v = get_irg_block_visited(current_ir_graph);
115 block_v = get_Block_block_visited(b);
116 if (block_v >= irg_v) {
117 /* we computed the number of preds for this block and saved it in the
119 return block_v - irg_v;
121 /* compute the number of good predecessors */
122 res = get_irn_arity(b);
123 for (i = 0; i < get_irn_arity(b); i++)
124 if (get_irn_opcode(get_irn_n(b, i)) == iro_Bad) res--;
125 /* save it in the flag. */
126 set_Block_block_visited(b, irg_v + res);
131 /* Copies the node to the new obstack. The Ins of the new node point to
132 the predecessors on the old obstack. For block/phi nodes not all
133 predecessors might be copied. n->link points to the new node.
134 For Phi and Block nodes the function allocates in-arrays with an arity
135 only for useful predecessors. The arity is determined by counting
136 the non-bad predecessors of the block. */
138 copy_node (ir_node *n, void *env) {
142 if (get_irn_opcode(n) == iro_Block) {
144 new_arity = compute_new_arity(n);
146 block = get_nodes_Block(n);
147 if (get_irn_opcode(n) == iro_Phi) {
148 new_arity = compute_new_arity(block);
150 new_arity = get_irn_arity(n);
153 nn = new_ir_node(get_irn_dbg_info(n),
160 /* Copy the attributes. These might point to additional data. If this
161 was allocated on the old obstack the pointers now are dangling. This
162 frees e.g. the memory of the graph_arr allocated in new_immBlock. */
166 /* printf("\n old node: "); DDMSG2(n);
167 printf(" new node: "); DDMSG2(nn); */
171 /* Copies new predecessors of old node to new node remembered in link.
172 Spare the Bad predecessors of Phi and Block nodes. */
174 copy_preds (ir_node *n, void *env) {
175 ir_node *nn, *block, *on;
178 nn = get_new_node(n);
180 /* printf("\n old node: "); DDMSG2(n);
181 printf(" new node: "); DDMSG2(nn);
182 printf(" arities: old: %d, new: %d\n", get_irn_arity(n), get_irn_arity(nn)); */
184 if (get_irn_opcode(n) == iro_Block) {
185 /* Don't copy Bad nodes. */
187 for (i = 0; i < get_irn_arity(n); i++)
188 if (get_irn_opcode(get_irn_n(n, i)) != iro_Bad) {
189 set_irn_n (nn, j, get_new_node(get_irn_n(n, i)));
192 /* repair the block visited flag from above misuse. Repair it in both
193 graphs so that the old one can still be used. */
194 set_Block_block_visited(nn, 0);
195 set_Block_block_visited(n, 0);
196 /* Local optimization could not merge two subsequent blocks if
197 in array contained Bads. Now it's possible.
198 We don't call optimize_in_place as it requires
199 that the fields in ir_graph are set properly. */
200 if (get_Block_n_cfgpreds(nn) == 1
201 && get_irn_op(get_Block_cfgpred(nn, 0)) == op_Jmp)
202 exchange(nn, get_nodes_Block(get_Block_cfgpred(nn, 0)));
203 } else if (get_irn_opcode(n) == iro_Phi) {
204 /* Don't copy node if corresponding predecessor in block is Bad.
205 The Block itself should not be Bad. */
206 block = get_nodes_Block(n);
207 set_irn_n (nn, -1, get_new_node(block));
209 for (i = 0; i < get_irn_arity(n); i++)
210 if (get_irn_opcode(get_irn_n(block, i)) != iro_Bad) {
211 set_irn_n (nn, j, get_new_node(get_irn_n(n, i)));
214 /* If the pre walker reached this Phi after the post walker visited the
215 block block_visited is > 0. */
216 set_Block_block_visited(get_nodes_Block(n), 0);
217 /* Compacting the Phi's ins might generate Phis with only one
219 if (get_irn_arity(n) == 1)
220 exchange(n, get_irn_n(n, 0));
222 for (i = -1; i < get_irn_arity(n); i++)
223 set_irn_n (nn, i, get_new_node(get_irn_n(n, i)));
225 /* Now the new node is complete. We can add it to the hash table for cse.
226 @@@ inlinening aborts if we identify End. Why? */
227 if(get_irn_op(nn) != op_End)
228 add_identities (current_ir_graph->value_table, nn);
231 /* Copies the graph recursively, compacts the keepalive of the end node. */
234 ir_node *oe, *ne; /* old end, new end */
235 ir_node *ka; /* keep alive */
238 oe = get_irg_end(current_ir_graph);
239 /* copy the end node by hand, allocate dynamic in array! */
240 ne = new_ir_node(get_irn_dbg_info(oe),
247 /* Copy the attributes. Well, there might be some in the future... */
249 set_new_node(oe, ne);
251 /* copy the live nodes */
252 irg_walk(get_nodes_Block(oe), copy_node, copy_preds, NULL);
253 /* copy_preds for the end node ... */
254 set_nodes_Block(ne, get_new_node(get_nodes_Block(oe)));
256 /** ... and now the keep alives. **/
257 /* First pick the not marked block nodes and walk them. We must pick these
258 first as else we will oversee blocks reachable from Phis. */
259 for (i = 0; i < get_irn_arity(oe); i++) {
260 ka = get_irn_n(oe, i);
261 if ((get_irn_op(ka) == op_Block) &&
262 (get_irn_visited(ka) < get_irg_visited(current_ir_graph))) {
263 /* We must keep the block alive and copy everything reachable */
264 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
265 irg_walk(ka, copy_node, copy_preds, NULL);
266 add_End_keepalive(ne, get_new_node(ka));
270 /* Now pick the Phis. Here we will keep all! */
271 for (i = 0; i < get_irn_arity(oe); i++) {
272 ka = get_irn_n(oe, i);
273 if ((get_irn_op(ka) == op_Phi)) {
274 if (get_irn_visited(ka) < get_irg_visited(current_ir_graph)) {
275 /* We didn't copy the Phi yet. */
276 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
277 irg_walk(ka, copy_node, copy_preds, NULL);
279 add_End_keepalive(ne, get_new_node(ka));
284 /* Copies the graph reachable from current_ir_graph->end to the obstack
285 in current_ir_graph and fixes the environment.
286 Then fixes the fields in current_ir_graph containing nodes of the
290 /* Not all nodes remembered in current_ir_graph might be reachable
291 from the end node. Assure their link is set to NULL, so that
292 we can test whether new nodes have been computed. */
293 set_irn_link(get_irg_frame (current_ir_graph), NULL);
294 set_irn_link(get_irg_globals(current_ir_graph), NULL);
295 set_irn_link(get_irg_args (current_ir_graph), NULL);
297 /* we use the block walk flag for removing Bads from Blocks ins. */
298 inc_irg_block_visited(current_ir_graph);
303 /* fix the fields in current_ir_graph */
304 free_End(get_irg_end(current_ir_graph));
305 set_irg_end (current_ir_graph, get_new_node(get_irg_end(current_ir_graph)));
306 set_irg_end_block (current_ir_graph, get_new_node(get_irg_end_block(current_ir_graph)));
307 if (get_irn_link(get_irg_frame(current_ir_graph)) == NULL) {
308 copy_node (get_irg_frame(current_ir_graph), NULL);
309 copy_preds(get_irg_frame(current_ir_graph), NULL);
311 if (get_irn_link(get_irg_globals(current_ir_graph)) == NULL) {
312 copy_node (get_irg_globals(current_ir_graph), NULL);
313 copy_preds(get_irg_globals(current_ir_graph), NULL);
315 if (get_irn_link(get_irg_args(current_ir_graph)) == NULL) {
316 copy_node (get_irg_args(current_ir_graph), NULL);
317 copy_preds(get_irg_args(current_ir_graph), NULL);
319 set_irg_start (current_ir_graph, get_new_node(get_irg_start(current_ir_graph)));
321 set_irg_start_block(current_ir_graph,
322 get_new_node(get_irg_start_block(current_ir_graph)));
323 set_irg_frame (current_ir_graph, get_new_node(get_irg_frame(current_ir_graph)));
324 set_irg_globals(current_ir_graph, get_new_node(get_irg_globals(current_ir_graph)));
325 set_irg_args (current_ir_graph, get_new_node(get_irg_args(current_ir_graph)));
326 if (get_irn_link(get_irg_bad(current_ir_graph)) == NULL) {
327 copy_node(get_irg_bad(current_ir_graph), NULL);
328 copy_preds(get_irg_bad(current_ir_graph), NULL);
330 set_irg_bad(current_ir_graph, get_new_node(get_irg_bad(current_ir_graph)));
333 /* Copies all reachable nodes to a new obstack. Removes bad inputs
334 from block nodes and the corresponding inputs from Phi nodes.
335 Merges single exit blocks with single entry blocks and removes
337 Adds all new nodes to a new hash table for cse. Does not
338 perform cse, so the hash table might contain common subexpressions. */
339 /* Amroq call this emigrate() */
341 dead_node_elimination(ir_graph *irg) {
343 struct obstack *graveyard_obst = NULL;
344 struct obstack *rebirth_obst = NULL;
346 /* Remember external state of current_ir_graph. */
347 rem = current_ir_graph;
348 current_ir_graph = irg;
350 /* Handle graph state */
351 assert(get_irg_phase_state(current_ir_graph) != phase_building);
352 free_outs(current_ir_graph);
354 if (get_optimize() && get_opt_dead_node_elimination()) {
356 /* A quiet place, where the old obstack can rest in peace,
357 until it will be cremated. */
358 graveyard_obst = irg->obst;
360 /* A new obstack, where the reachable nodes will be copied to. */
361 rebirth_obst = (struct obstack *) xmalloc (sizeof (struct obstack));
362 current_ir_graph->obst = rebirth_obst;
363 obstack_init (current_ir_graph->obst);
365 /* We also need a new hash table for cse */
366 del_identities (irg->value_table);
367 irg->value_table = new_identities ();
369 /* Copy the graph from the old to the new obstack */
372 /* Free memory from old unoptimized obstack */
373 obstack_free(graveyard_obst, 0); /* First empty the obstack ... */
374 xfree (graveyard_obst); /* ... then free it. */
377 current_ir_graph = rem;
380 /**********************************************************************/
381 /* Funcionality for inlining */
382 /**********************************************************************/
384 /* Copy node for inlineing. Copies the node by calling copy_node and
385 then updates the entity if it's a local one. env must be a pointer
386 to the frame type of the procedure. The new entities must be in
387 the link field of the entities. */
389 copy_node_inline (ir_node *n, void *env) {
391 type *frame_tp = (type *)env;
394 if (get_irn_op(n) == op_Sel) {
395 new = get_new_node (n);
396 assert(get_irn_op(new) == op_Sel);
397 if (get_entity_owner(get_Sel_entity(n)) == frame_tp) {
398 set_Sel_entity(new, get_entity_link(get_Sel_entity(n)));
403 void inline_method(ir_node *call, ir_graph *called_graph) {
405 ir_node *post_call, *post_bl;
407 ir_node *end, *end_bl;
412 int arity, n_ret, n_exc, n_res, i, j, rem_opt;
413 type *called_frame, *caller_frame;
415 if (!get_optimize() || !get_opt_inline()) return;
416 /** Turn off optimizations, this can cause problems when allocating new nodes. **/
417 rem_opt = get_optimize();
420 /* Handle graph state */
421 assert(get_irg_phase_state(current_ir_graph) != phase_building);
422 assert(get_irg_pinned(current_ir_graph) == pinned);
423 assert(get_irg_pinned(called_graph) == pinned);
424 if (get_irg_outs_state(current_ir_graph) == outs_consistent)
425 set_irg_outs_inconsistent(current_ir_graph);
427 /** Check preconditions **/
428 assert(get_irn_op(call) == op_Call);
429 assert(get_Call_type(call) == get_entity_type(get_irg_ent(called_graph)));
430 assert(get_type_tpop(get_Call_type(call)) == type_method);
431 if (called_graph == current_ir_graph) return;
434 /** Part the Call node into two nodes. Pre_call collects the parameters of
435 the procedure and later replaces the Start node of the called graph.
436 Post_call is the old Call node and collects the results of the called
437 graph. Both will end up being a tuple. **/
438 post_bl = get_nodes_Block(call);
439 set_irg_current_block(current_ir_graph, post_bl);
440 /* XxMxPxP of Start + parameter of Call */
442 in[1] = get_Call_mem(call);
443 in[2] = get_irg_frame(current_ir_graph);
444 in[3] = get_irg_globals(current_ir_graph);
445 in[4] = new_Tuple (get_Call_n_params(call), get_Call_param_arr(call));
446 pre_call = new_Tuple(5, in);
449 /** Part the block of the Call node into two blocks.
450 The new block gets the ins of the old block, pre_call and all its
451 predecessors and all Phi nodes. **/
452 part_block(pre_call);
454 /** Prepare state for dead node elimination **/
455 /* Visited flags in calling irg must be >= flag in called irg.
456 Else walker and arity computation will not work. */
457 if (get_irg_visited(current_ir_graph) <= get_irg_visited(called_graph))
458 set_irg_visited(current_ir_graph, get_irg_visited(called_graph)+1); /***/
459 if (get_irg_block_visited(current_ir_graph)< get_irg_block_visited(called_graph))
460 set_irg_block_visited(current_ir_graph, get_irg_block_visited(called_graph));
461 /* Set pre_call as new Start node in link field of the start node of
462 calling graph and pre_calls block as new block for the start block
464 Further mark these nodes so that they are not visited by the
466 set_irn_link(get_irg_start(called_graph), pre_call);
467 set_irn_visited(get_irg_start(called_graph),
468 get_irg_visited(current_ir_graph));/***/
469 set_irn_link(get_irg_start_block(called_graph),
470 get_nodes_Block(pre_call));
471 set_irn_visited(get_irg_start_block(called_graph),
472 get_irg_visited(current_ir_graph)); /***/
474 /* Initialize for compaction of in arrays */
475 inc_irg_block_visited(current_ir_graph);
477 set_Block_block_visited(get_irg_start_block(called_graph),
478 get_irg_block_visited(current_ir_graph) +1 +1); /* count for self edge */
480 /*** Replicate local entities of the called_graph ***/
481 /* copy the entities. */
482 called_frame = get_irg_frame_type(called_graph);
483 for (i = 0; i < get_class_n_member(called_frame); i++) {
484 entity *new_ent, *old_ent;
485 old_ent = get_class_member(called_frame, i);
486 new_ent = copy_entity_own(old_ent, get_cur_frame_type());
487 set_entity_link(old_ent, new_ent);
490 /* visited is > than that of called graph. With this trick visited will
491 remain unchanged so that an outer walker, e.g., searching the call nodes
492 to inline, calling this inline will not visit the inlined nodes. */
493 set_irg_visited(current_ir_graph, get_irg_visited(current_ir_graph)-1);
495 /** Performing dead node elimination inlines the graph **/
496 /* Copies the nodes to the obstack of current_ir_graph. Updates links to new
498 /* @@@ endless loops are not copied!! */
499 irg_walk(get_irg_end(called_graph), copy_node_inline, copy_preds,
500 get_irg_frame_type(called_graph));
502 /* Repair called_graph */
503 set_irg_visited(called_graph, get_irg_visited(current_ir_graph));
504 set_irg_block_visited(called_graph, get_irg_block_visited(current_ir_graph));
505 set_Block_block_visited(get_irg_start_block(called_graph), 0);
507 /*** Merge the end of the inlined procedure with the call site ***/
508 /* We will turn the old Call node into a Tuple with the following
511 0: Phi of all Memories of Return statements.
512 1: Jmp from new Block that merges the control flow from all exception
513 predecessors of the old end block.
514 2: Tuple of all arguments.
515 3: Phi of Exception memories.
518 /** Precompute some values **/
519 end_bl = get_new_node(get_irg_end_block(called_graph));
520 end = get_new_node(get_irg_end(called_graph));
521 arity = get_irn_arity(end_bl); /* arity = n_exc + n_ret */
522 n_res = get_method_n_res(get_Call_type(call));
524 res_pred = (ir_node **) malloc (n_res * sizeof (ir_node *));
525 cf_pred = (ir_node **) malloc (arity * sizeof (ir_node *));
527 set_irg_current_block(current_ir_graph, post_bl); /* just to make sure */
529 /** archive keepalives **/
530 for (i = 0; i < get_irn_arity(end); i++)
531 add_End_keepalive(get_irg_end(current_ir_graph), get_irn_n(end, i));
532 /* The new end node will die, but the in array is not on the obstack ... */
535 /** Collect control flow from Return blocks to post_calls block. Replace
536 Return nodes by Jump nodes. **/
538 for (i = 0; i < arity; i++) {
540 ret = get_irn_n(end_bl, i);
541 if (get_irn_op(ret) == op_Return) {
542 cf_pred[n_ret] = new_r_Jmp(current_ir_graph, get_nodes_Block(ret));
546 set_irn_in(post_bl, n_ret, cf_pred);
548 /** Collect results from Return nodes to post_call. Post_call is
549 turned into a tuple. **/
550 turn_into_tuple(post_call, 4);
551 /* First the Memory-Phi */
553 for (i = 0; i < arity; i++) {
554 ret = get_irn_n(end_bl, i);
555 if (get_irn_op(ret) == op_Return) {
556 cf_pred[n_ret] = get_Return_mem(ret);
560 phi = new_Phi(n_ret, cf_pred, mode_M);
561 set_Tuple_pred(call, 0, phi);
562 set_irn_link(phi, get_irn_link(post_bl)); /* Conserve Phi-list for further inlinings */
563 set_irn_link(post_bl, phi);
564 /* Now the real results */
566 for (j = 0; j < n_res; j++) {
568 for (i = 0; i < arity; i++) {
569 ret = get_irn_n(end_bl, i);
570 if (get_irn_op(ret) == op_Return) {
571 cf_pred[n_ret] = get_Return_res(ret, j);
575 phi = new_Phi(n_ret, cf_pred, get_irn_mode(cf_pred[0]));
577 set_irn_link(phi, get_irn_link(post_bl)); /* Conserve Phi-list for further inlinings */
578 set_irn_link(post_bl, phi);
580 set_Tuple_pred(call, 2, new_Tuple(n_res, res_pred));
582 set_Tuple_pred(call, 2, new_Bad());
584 /* Finally the exception control flow. We need to add a Phi node to
585 collect the memory containing the exception objects. Further we need
586 to add another block to get a correct representation of this Phi. To
587 this block we add a Jmp that resolves into the X output of the Call
588 when the Call is turned into a tuple. */
590 for (i = 0; i < arity; i++) {
592 ret = get_irn_n(end_bl, i);
593 if (is_fragile_op(skip_Proj(ret)) || (get_irn_op(skip_Proj(ret)) == op_Raise)) {
594 cf_pred[n_exc] = ret;
599 new_Block(n_exc, cf_pred); /* whatch it: current_block is changed! */
600 set_Tuple_pred(call, 1, new_Jmp());
601 /* The Phi for the memories with the exception objects */
603 for (i = 0; i < arity; i++) {
605 ret = skip_Proj(get_irn_n(end_bl, i));
606 if (get_irn_op(ret) == op_Call) {
607 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_Block(ret), ret, mode_M, 3);
609 } else if (is_fragile_op(ret)) {
610 /* We rely that all cfops have the memory output at the same position. */
611 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_Block(ret), ret, mode_M, 0);
613 } else if (get_irn_op(ret) == op_Raise) {
614 cf_pred[n_exc] = new_r_Proj(current_ir_graph, get_nodes_Block(ret), ret, mode_M, 1);
618 set_Tuple_pred(call, 3, new_Phi(n_exc, cf_pred, mode_M));
620 set_Tuple_pred(call, 1, new_Bad());
621 set_Tuple_pred(call, 3, new_Bad());
626 /*** Correct the control flow to the end node.
627 If the exception control flow from the Call directly branched to the
628 end block we now have the following control flow predecessor pattern:
629 ProjX -> Tuple -> Jmp.
630 We must remove the Jmp along with it's empty block and add Jmp's
631 predecessors as predecessors of this end block. ***/
632 /* find the problematic predecessor of the end block. */
633 end_bl = get_irg_end_block(current_ir_graph);
634 for (i = 0; i < get_Block_n_cfgpreds(end_bl); i++) {
635 cf_op = get_Block_cfgpred(end_bl, i);
636 if (get_irn_op(cf_op) == op_Proj) {
637 cf_op = get_Proj_pred(cf_op);
638 if (get_irn_op(cf_op) == op_Tuple) {
639 cf_op = get_Tuple_pred(cf_op, 1);
640 assert(get_irn_op(cf_op) == op_Jmp);
646 if (i < get_Block_n_cfgpreds(end_bl)) {
647 bl = get_nodes_Block(cf_op);
648 arity = get_Block_n_cfgpreds(end_bl) + get_Block_n_cfgpreds(bl) - 1;
649 cf_pred = (ir_node **) malloc (arity * sizeof (ir_node *));
650 for (j = 0; j < i; j++)
651 cf_pred[j] = get_Block_cfgpred(end_bl, j);
652 for (j = j; j < i + get_Block_n_cfgpreds(bl); j++)
653 cf_pred[j] = get_Block_cfgpred(bl, j-i);
654 for (j = j; j < arity; j++)
655 cf_pred[j] = get_Block_cfgpred(end_bl, j-get_Block_n_cfgpreds(bl) +1);
656 set_irn_in(end_bl, arity, cf_pred);
660 /** Turn cse back on. **/
661 set_optimize(rem_opt);
664 /********************************************************************/
665 /* Apply inlineing to small methods. */
666 /********************************************************************/
670 /* It makes no sense to inline too many calls in one procedure. Anyways,
671 I didn't get a version with NEW_ARR_F to run. */
672 #define MAX_INLINE 1024
674 static void collect_calls(ir_node *call, void *env) {
675 ir_node **calls = (ir_node **)env;
678 ir_graph *called_irg;
680 if (get_irn_op(call) != op_Call) return;
682 addr = get_Call_ptr(call);
683 if (get_irn_op(addr) == op_Const) {
684 /* Check whether the constant is the pointer to a compiled entity. */
685 tv = get_Const_tarval(addr);
687 called_irg = get_entity_irg(tv->u.p.ent);
688 if (called_irg && pos < MAX_INLINE) {
689 /* The Call node calls a locally defined method. Remember to inline. */
698 /* Inlines all small methods at call sites where the called address comes
699 from a Const node that references the entity representing the called
701 The size argument is a rough measure for the code size of the method:
702 Methods where the obstack containing the firm graph is smaller than
704 void inline_small_irgs(ir_graph *irg, int size) {
706 ir_node *calls[MAX_INLINE];
707 ir_graph *rem = current_ir_graph;
709 if (!(get_optimize() && get_opt_inline())) return;
711 /*DDME(get_irg_ent(current_ir_graph));*/
713 current_ir_graph = irg;
714 /* Handle graph state */
715 assert(get_irg_phase_state(current_ir_graph) != phase_building);
717 /* Find Call nodes to inline.
718 (We can not inline during a walk of the graph, as inlineing the same
719 method several times changes the visited flag of the walked graph:
720 after the first inlineing visited of the callee equals visited of
721 the caller. With the next inlineing both are increased.) */
723 irg_walk(get_irg_end(irg), NULL, collect_calls, (void *) calls);
725 if ((pos > 0) && (pos < MAX_INLINE)) {
726 /* There are calls to inline */
727 collect_phiprojs(irg);
728 for (i = 0; i < pos; i++) {
731 tv = get_Const_tarval(get_Call_ptr(calls[i]));
732 callee = get_entity_irg(tv->u.p.ent);
733 if ((_obstack_memory_used(callee->obst) - obstack_room(callee->obst)) < size) {
734 /*printf(" inlineing "); DDME(tv->u.p.ent);*/
735 inline_method(calls[i], callee);
740 current_ir_graph = rem;
744 /********************************************************************/
745 /* Code Placement. Pinns all floating nodes to a block where they */
746 /* will be executed only if needed. */
747 /********************************************************************/
749 static pdeq *worklist; /* worklist of ir_node*s */
751 /* Find the earliest correct block for N. --- Place N into the
752 same Block as its dominance-deepest Input. */
754 place_floats_early (ir_node *n)
758 /* we must not run into an infinite loop */
759 assert (irn_not_visited(n));
762 /* Place floating nodes. */
763 if (get_op_pinned(get_irn_op(n)) == floats) {
765 ir_node *b = new_Bad(); /* The block to place this node in */
767 assert(get_irn_op(n) != op_Block);
769 if ((get_irn_op(n) == op_Const) ||
770 (get_irn_op(n) == op_SymConst) ||
772 /* These nodes will not be placed by the loop below. */
773 b = get_irg_start_block(current_ir_graph);
777 /* find the block for this node. */
778 for (i = 0; i < get_irn_arity(n); i++) {
779 ir_node *dep = get_irn_n(n, i);
781 if ((irn_not_visited(dep)) &&
782 (get_op_pinned(get_irn_op(dep)) == floats)) {
783 place_floats_early (dep);
785 /* Because all loops contain at least one pinned node, now all
786 our inputs are either pinned or place_early has already
787 been finished on them. We do not have any unfinished inputs! */
788 dep_block = get_nodes_Block(dep);
789 if ((!is_Bad(dep_block)) &&
790 (get_Block_dom_depth(dep_block) > depth)) {
792 depth = get_Block_dom_depth(dep_block);
794 /* Avoid that the node is placed in the Start block */
795 if ((depth == 1) && (get_Block_dom_depth(get_nodes_Block(n)) > 1)) {
796 b = get_Block_cfg_out(get_irg_start_block(current_ir_graph), 0);
797 assert(b != get_irg_start_block(current_ir_graph));
801 set_nodes_Block(n, b);
804 /* Add predecessors of non floating nodes on worklist. */
805 start = (get_irn_op(n) == op_Block) ? 0 : -1;
806 for (i = start; i < get_irn_arity(n); i++) {
807 ir_node *pred = get_irn_n(n, i);
808 if (irn_not_visited(pred)) {
809 pdeq_putr (worklist, pred);
814 /* Floating nodes form subgraphs that begin at nodes as Const, Load,
815 Start, Call and end at pinned nodes as Store, Call. Place_early
816 places all floating nodes reachable from its argument through floating
817 nodes and adds all beginnings at pinned nodes to the worklist. */
818 inline void place_early () {
823 inc_irg_visited(current_ir_graph);
825 /* this inits the worklist */
826 place_floats_early (get_irg_end(current_ir_graph));
828 /* Work the content of the worklist. */
829 while (!pdeq_empty (worklist)) {
830 ir_node *n = pdeq_getl (worklist);
831 if (irn_not_visited(n)) place_floats_early (n);
834 set_irg_outs_inconsistent(current_ir_graph);
835 current_ir_graph->pinned = pinned;
839 /* deepest common dominance ancestor of DCA and CONSUMER of PRODUCER */
841 consumer_dom_dca (ir_node *dca, ir_node *consumer, ir_node *producer)
845 /* Compute the latest block into which we can place a node so that it is
847 if (get_irn_op(consumer) == op_Phi) {
848 /* our comsumer is a Phi-node, the effective use is in all those
849 blocks through which the Phi-node reaches producer */
851 ir_node *phi_block = get_nodes_Block(consumer);
852 for (i = 0; i < get_irn_arity(consumer); i++) {
853 if (get_irn_n(consumer, i) == producer) {
854 block = get_nodes_Block(get_Block_cfgpred(phi_block, i));
858 assert(is_no_Block(consumer));
859 block = get_nodes_Block(consumer);
862 /* Compute the deepest common ancestor of block and dca. */
864 if (!dca) return block;
865 while (get_Block_dom_depth(block) > get_Block_dom_depth(dca))
866 block = get_Block_idom(block);
867 while (get_Block_dom_depth(dca) > get_Block_dom_depth(block))
868 dca = get_Block_idom(dca);
870 { block = get_Block_idom(block); dca = get_Block_idom(dca); }
876 /* @@@ Needs loop informations. Will implement later interprocedural. */
878 move_out_of_loops (ir_node *n, ir_node *dca)
882 /* Find the region deepest in the dominator tree dominating
883 dca with the least loop nesting depth, but still dominated
884 by our early placement. */
886 while (dca != get_nodes_Block(n)) {
887 dca = get_Block_idom(dca);
888 if (!dca) break; /* should we put assert(dca)? */
889 if (get_Block_loop_depth(dca) < get_Block_loop_depth(best)) {
893 if (get_Block_dom_depth(best) >= get_Block_dom_depth(get_nodes_Block(n)))
894 set_nodes_Block(n, best);
898 /* Find the latest legal block for N and place N into the
899 `optimal' Block between the latest and earliest legal block.
900 The `optimal' block is the dominance-deepest block of those
901 with the least loop-nesting-depth. This places N out of as many
902 loops as possible and then makes it as controldependant as
905 place_floats_late (ir_node *n)
909 assert (irn_not_visited(n)); /* no multiple placement */
911 /* no need to place block nodes, control nodes are already placed. */
912 if ((get_irn_op(n) != op_Block) && (!is_cfop(n)) && (get_irn_mode(n) != mode_X)) {
913 /* Assure that our users are all placed, except the Phi-nodes.
914 --- Each dataflow cycle contains at least one Phi-node. We
915 have to break the `user has to be placed before the
916 producer' dependance cycle and the Phi-nodes are the
917 place to do so, because we need to base our placement on the
918 final region of our users, which is OK with Phi-nodes, as they
919 are pinned, and they never have to be placed after a
920 producer of one of their inputs in the same block anyway. */
921 for (i = 0; i < get_irn_n_outs(n); i++) {
922 ir_node *succ = get_irn_out(n, i);
923 if (irn_not_visited(succ) && (get_irn_op(succ) != op_Phi))
924 place_floats_late (succ);
927 /* We have to determine the final block of this node... except for constants. */
928 if ((get_op_pinned(get_irn_op(n)) == floats) &&
929 (get_irn_op(n) != op_Const) &&
930 (get_irn_op(n) != op_SymConst)) {
931 ir_node *dca = NULL; /* deepest common ancestor in the
932 dominator tree of all nodes'
933 blocks depending on us; our final
934 placement has to dominate DCA. */
935 for (i = 0; i < get_irn_n_outs(n); i++) {
936 dca = consumer_dom_dca (dca, get_irn_out(n, i), n);
938 set_nodes_Block(n, dca);
940 move_out_of_loops (n, dca);
947 /* Add predecessors of all non-floating nodes on list. (Those of floating
948 nodes are placeded already and therefore are marked.) */
949 for (i = 0; i < get_irn_n_outs(n); i++) {
950 if (irn_not_visited(get_irn_out(n, i))) {
951 pdeq_putr (worklist, get_irn_out(n, i));
956 inline void place_late() {
958 inc_irg_visited(current_ir_graph);
960 /* This fills the worklist initially. */
961 place_floats_late(get_irg_start_block(current_ir_graph));
962 /* And now empty the worklist again... */
963 while (!pdeq_empty (worklist)) {
964 ir_node *n = pdeq_getl (worklist);
965 if (irn_not_visited(n)) place_floats_late(n);
969 void place_code(ir_graph *irg) {
970 ir_graph *rem = current_ir_graph;
971 current_ir_graph = irg;
973 if (!(get_optimize() && get_opt_global_cse())) return;
975 /* Handle graph state */
976 assert(get_irg_phase_state(irg) != phase_building);
977 if (get_irg_dom_state(irg) != dom_consistent)
980 /* Place all floating nodes as early as possible. This guarantees
981 a legal code placement. */
982 worklist = new_pdeq ();
985 /* place_early invalidates the outs, place_late needs them. */
987 /* Now move the nodes down in the dominator tree. This reduces the
988 unnecessary executions of the node. */
991 set_irg_outs_inconsistent(current_ir_graph);
993 current_ir_graph = rem;
998 /********************************************************************/
999 /* Control flow optimization. */
1000 /* Removes Bad control flow predecessors and empty blocks. A block */
1001 /* is empty if it contains only a Jmp node. */
1003 /********************************************************************/
1006 static void merge_blocks(ir_node *n, void *env) {
1008 set_irn_link(n, NULL);
1010 if (get_irn_op(n) == op_Block) {
1012 for (i = 0; i < get_Block_n_cfgpreds(n); i++)
1013 set_Block_cfgpred(n, i, skip_Tuple(get_Block_cfgpred(n, i)));
1014 } else if (get_irn_mode(n) == mode_X) {
1015 /* We will soon visit a block. Optimize it before visiting! */
1016 ir_node *b = get_nodes_Block(n);
1017 ir_node *new = equivalent_node(b);
1018 while (irn_not_visited(b) && (!is_Bad(new)) && (new != b)) {
1019 /* We would have to run gigo if new is bad. */
1020 if (get_optimize() && get_opt_control_flow()) exchange (b, new);
1022 new = equivalent_node(b);
1024 if (is_Bad(new)) exchange (n, new_Bad());
1028 static void collect_nodes(ir_node *n, void *env) {
1030 if (is_no_Block(n)) {
1031 ir_node *b = get_nodes_Block(n);
1033 if ((get_irn_op(n) == op_Phi)) {
1034 /* Collect Phi nodes to compact ins along with block's ins. */
1035 set_irn_link(n, get_irn_link(b));
1037 } else if (get_irn_op(n) != op_Jmp) { /* Check for non empty block. */
1038 mark_Block_block_visited(b);
1043 /* Returns true if pred is pred of block */
1044 int is_pred_of(ir_node *pred, ir_node *b) {
1046 for (i = 0; i < get_Block_n_cfgpreds(b); i++) {
1047 ir_node *b_pred = get_nodes_Block(get_Block_cfgpred(b, i));
1048 if (b_pred == pred) return 1;
1053 int test_whether_dispensable(ir_node *b, int pos) {
1054 int i, j, n_preds = 1;
1055 int dispensable = 1;
1056 ir_node *cfop = get_Block_cfgpred(b, pos);
1057 ir_node *pred = get_nodes_Block(cfop);
1059 if (get_Block_block_visited(pred) + 1
1060 < get_irg_block_visited(current_ir_graph)) {
1061 if (!get_optimize() || !get_opt_control_flow()) {
1062 /* Mark block so that is will not be removed. */
1063 set_Block_block_visited(pred, get_irg_block_visited(current_ir_graph)-1);
1066 /* Seems to be empty. */
1067 if (!get_irn_link(b)) {
1068 /* There are no Phi nodes ==> dispensable. */
1069 n_preds = get_Block_n_cfgpreds(pred);
1071 /* b's pred blocks and pred's pred blocks must be pairwise disjunct.
1072 Work preds < pos as if they were already removed. */
1073 for (i = 0; i < pos; i++) {
1074 ir_node *b_pred = get_nodes_Block(get_Block_cfgpred(b, i));
1075 if (get_Block_block_visited(b_pred) + 1
1076 < get_irg_block_visited(current_ir_graph)) {
1077 for (j = 0; j < get_Block_n_cfgpreds(b_pred); j++) {
1078 ir_node *b_pred_pred = get_nodes_Block(get_Block_cfgpred(b_pred, j));
1079 if (is_pred_of(b_pred_pred, pred)) dispensable = 0;
1082 if (is_pred_of(b_pred, pred)) dispensable = 0;
1085 for (i = pos +1; i < get_Block_n_cfgpreds(b); i++) {
1086 ir_node *b_pred = get_nodes_Block(get_Block_cfgpred(b, i));
1087 if (is_pred_of(b_pred, pred)) dispensable = 0;
1090 set_Block_block_visited(pred, get_irg_block_visited(current_ir_graph)-1);
1093 n_preds = get_Block_n_cfgpreds(pred);
1101 void optimize_blocks(ir_node *b, void *env) {
1102 int i, j, k, max_preds, n_preds;
1103 ir_node *pred, *phi;
1107 for (i = 0; i < get_Block_n_cfgpreds(b); i++) {
1108 pred = get_Block_cfgpred(b, i);
1109 max_preds += test_whether_dispensable(b, i);
1111 in = (ir_node **) malloc(max_preds * sizeof(ir_node *));
1115 printf(" working on "); DDMN(b);
1116 for (i = 0; i < get_Block_n_cfgpreds(b); i++) {
1117 pred = get_nodes_Block(get_Block_cfgpred(b, i));
1118 if (is_Bad(get_Block_cfgpred(b, i))) {
1119 printf(" removing Bad %i\n ", i);
1120 } else if (get_Block_block_visited(pred) +1
1121 < get_irg_block_visited(current_ir_graph)) {
1122 printf(" removing pred %i ", i); DDMN(pred);
1123 } else { printf(" Nothing to do for "); DDMN(pred); }
1125 ** end Debug output **/
1127 /** Fix the Phi nodes **/
1128 phi = get_irn_link(b);
1130 assert(get_irn_op(phi) == op_Phi);
1131 /* Find the new predecessors for the Phi */
1133 for (i = 0; i < get_Block_n_cfgpreds(b); i++) {
1134 pred = get_nodes_Block(get_Block_cfgpred(b, i));
1135 if (is_Bad(get_Block_cfgpred(b, i))) {
1137 } else if (get_Block_block_visited(pred) +1
1138 < get_irg_block_visited(current_ir_graph)) {
1139 /* It's an empty block and not yet visited. */
1140 ir_node *phi_pred = get_Phi_pred(phi, i);
1141 for (j = 0; j < get_Block_n_cfgpreds(pred); j++) {
1142 if (get_nodes_Block(phi_pred) == pred) {
1143 assert(get_irn_op(phi_pred) == op_Phi); /* Block is empty!! */
1144 in[n_preds] = get_Phi_pred(phi_pred, j);
1146 in[n_preds] = phi_pred;
1151 /* @@@ hier brauche ich schleifeninformation!!! Wenn keine Rueckwaertskante
1152 dann darfs auch keine Verwendung geben. */
1153 if (get_nodes_Block(phi_pred) == pred) {
1154 /* remove the Phi as it might be kept alive. Further there
1155 might be other users. */
1156 exchange(phi_pred, phi); /* geht, is aber doch semantisch falsch! */
1160 in[n_preds] = get_Phi_pred(phi, i);
1165 set_irn_in(phi, n_preds, in);
1166 phi = get_irn_link(phi);
1169 /** Move Phi nodes from removed blocks to this one.
1170 This happens only if merge between loop backedge and single loop entry. **/
1171 for (k = 0; k < get_Block_n_cfgpreds(b); k++) {
1172 pred = get_nodes_Block(get_Block_cfgpred(b, k));
1173 if (get_Block_block_visited(pred) +1
1174 < get_irg_block_visited(current_ir_graph)) {
1175 phi = get_irn_link(pred);
1177 if (get_irn_op(phi) == op_Phi) {
1178 set_nodes_Block(phi, b);
1181 for (i = 0; i < k; i++) {
1182 pred = get_nodes_Block(get_Block_cfgpred(b, i));
1183 if (is_Bad(get_Block_cfgpred(b, i))) {
1185 } else if (get_Block_block_visited(pred) +1
1186 < get_irg_block_visited(current_ir_graph)) {
1187 /* It's an empty block and not yet visited. */
1188 for (j = 0; j < get_Block_n_cfgpreds(pred); j++) {
1189 /* @@@ Hier brauche ich schleifeninformation!!! Kontrllflusskante
1190 muss Rueckwaertskante sein! (An allen vier in[n_preds] = phi
1191 Anweisungen.) Trotzdem tuts bisher!! */
1200 for (i = 0; i < get_Phi_n_preds(phi); i++) {
1201 in[n_preds] = get_Phi_pred(phi, i);
1204 for (i = k+1; i < get_Block_n_cfgpreds(b); i++) {
1205 pred = get_nodes_Block(get_Block_cfgpred(b, i));
1206 if (is_Bad(get_Block_cfgpred(b, i))) {
1208 } else if (get_Block_block_visited(pred) +1
1209 < get_irg_block_visited(current_ir_graph)) {
1210 /* It's an empty block and not yet visited. */
1211 for (j = 0; j < get_Block_n_cfgpreds(pred); j++) {
1220 set_irn_in(phi, n_preds, in);
1222 phi = get_irn_link(phi);
1227 /** Fix the block **/
1229 for (i = 0; i < get_Block_n_cfgpreds(b); i++) {
1230 pred = get_nodes_Block(get_Block_cfgpred(b, i));
1231 if (is_Bad(get_Block_cfgpred(b, i))) {
1233 } else if (get_Block_block_visited(pred) +1
1234 < get_irg_block_visited(current_ir_graph)) {
1235 /* It's an empty block and not yet visited. */
1236 assert(get_Block_n_cfgpreds(b) > 1);
1237 /* Else it should be optimized by equivalent_node. */
1238 for (j = 0; j < get_Block_n_cfgpreds(pred); j++) {
1239 in[n_preds] = get_Block_cfgpred(pred, j);
1242 /* Remove block as it might be kept alive. */
1243 exchange(pred, b/*new_Bad()*/);
1245 in[n_preds] = get_Block_cfgpred(b, i);
1249 set_irn_in(b, n_preds, in);
1253 void optimize_cf(ir_graph *irg) {
1256 ir_node *end = get_irg_end(irg);
1257 ir_graph *rem = current_ir_graph;
1258 current_ir_graph = irg;
1261 /* Handle graph state */
1262 assert(get_irg_phase_state(irg) != phase_building);
1263 if (get_irg_outs_state(current_ir_graph) == outs_consistent)
1264 set_irg_outs_inconsistent(current_ir_graph);
1265 if (get_irg_dom_state(current_ir_graph) == dom_consistent)
1266 set_irg_dom_inconsistent(current_ir_graph);
1268 //DDME(get_irg_ent(irg));
1270 /* Use block visited flag to mark non-empty blocks. */
1271 inc_irg_block_visited(irg);
1272 irg_walk(end, merge_blocks, collect_nodes, NULL);
1274 /* Optimize the standard code. */
1275 irg_block_walk(get_irg_end_block(irg), optimize_blocks, NULL, NULL);
1277 /* Walk all keep alives, optimize them if block, add to new in-array
1278 for end if useful. */
1279 in = NEW_ARR_F (ir_node *, 1);
1280 in[0] = get_nodes_Block(end);
1282 inc_irg_visited(current_ir_graph);
1283 for(i = 0; i < get_End_n_keepalives(end); i++) {
1284 ir_node *ka = get_End_keepalive(end, i);
1285 if (irn_not_visited(ka))
1286 if ((get_irn_op(ka) == op_Block) && Block_not_block_visited(ka)) {
1287 set_irg_block_visited(current_ir_graph, /* Don't walk all the way to Start. */
1288 get_irg_block_visited(current_ir_graph)-1);
1289 irg_block_walk(ka, optimize_blocks, NULL, NULL);
1290 mark_irn_visited(ka);
1291 ARR_APP1 (ir_node *, in, ka);
1292 } else if (get_irn_op(ka) == op_Phi) {
1293 mark_irn_visited(ka);
1294 ARR_APP1 (ir_node *, in, ka);
1297 /* DEL_ARR_F(end->in); GL @@@ tut nicht ! */
1300 current_ir_graph = rem;