Mercurial > hg > graal-compiler

comparison src/share/vm/opto/lcm.cpp @ 12171:4b078f877b56

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
8023988: Move local scheduling of nodes to the CFG creation and code motion phase (PhaseCFG) Summary: Moved local scheduling code from class Block to class PhaseCFG Reviewed-by: kvn, roland
author adlertz
date Sun, 01 Sep 2013 19:21:05 +0200
parents 650868c062a9
children a9a968364704
comparison
equal deleted inserted replaced
12169:29aa8936f03c 12171:4b078f877b56
56 // with suitable memory ops nearby. Use the memory op to do the NULL check. 56 // with suitable memory ops nearby. Use the memory op to do the NULL check.
57 // I can generate a memory op if there is not one nearby. 57 // I can generate a memory op if there is not one nearby.
58 // The proj is the control projection for the not-null case. 58 // The proj is the control projection for the not-null case.
59 // The val is the pointer being checked for nullness or 59 // The val is the pointer being checked for nullness or
60 // decodeHeapOop_not_null node if it did not fold into address. 60 // decodeHeapOop_not_null node if it did not fold into address.
61 void Block::implicit_null_check(PhaseCFG *cfg, Node *proj, Node *val, int allowed_reasons) { 61 void PhaseCFG::implicit_null_check(Block* block, Node *proj, Node *val, int allowed_reasons) {
62 // Assume if null check need for 0 offset then always needed 62 // Assume if null check need for 0 offset then always needed
63 // Intel solaris doesn't support any null checks yet and no 63 // Intel solaris doesn't support any null checks yet and no
64 // mechanism exists (yet) to set the switches at an os_cpu level 64 // mechanism exists (yet) to set the switches at an os_cpu level
65 if( !ImplicitNullChecks || MacroAssembler::needs_explicit_null_check(0)) return; 65 if( !ImplicitNullChecks || MacroAssembler::needs_explicit_null_check(0)) return;
66 66
67 // Make sure the ptr-is-null path appears to be uncommon! 67 // Make sure the ptr-is-null path appears to be uncommon!
68 float f = end()->as_MachIf()->_prob; 68 float f = block->end()->as_MachIf()->_prob;
69 if( proj->Opcode() == Op_IfTrue ) f = 1.0f - f; 69 if( proj->Opcode() == Op_IfTrue ) f = 1.0f - f;
70 if( f > PROB_UNLIKELY_MAG(4) ) return; 70 if( f > PROB_UNLIKELY_MAG(4) ) return;
71 71
72 uint bidx = 0; // Capture index of value into memop 72 uint bidx = 0; // Capture index of value into memop
73 bool was_store; // Memory op is a store op 73 bool was_store; // Memory op is a store op
74 74
75 // Get the successor block for if the test ptr is non-null 75 // Get the successor block for if the test ptr is non-null
76 Block* not_null_block; // this one goes with the proj 76 Block* not_null_block; // this one goes with the proj
77 Block* null_block; 77 Block* null_block;
78 if (get_node(number_of_nodes()-1) == proj) { 78 if (block->get_node(block->number_of_nodes()-1) == proj) {
79 null_block = _succs[0]; 79 null_block = block->_succs[0];
80 not_null_block = _succs[1]; 80 not_null_block = block->_succs[1];
81 } else { 81 } else {
82 assert(get_node(number_of_nodes()-2) == proj, "proj is one or the other"); 82 assert(block->get_node(block->number_of_nodes()-2) == proj, "proj is one or the other");
83 not_null_block = _succs[0]; 83 not_null_block = block->_succs[0];
84 null_block = _succs[1]; 84 null_block = block->_succs[1];
85 } 85 }
86 while (null_block->is_Empty() == Block::empty_with_goto) { 86 while (null_block->is_Empty() == Block::empty_with_goto) {
87 null_block = null_block->_succs[0]; 87 null_block = null_block->_succs[0];
88 } 88 }
89 89
91 // (See Parse::do_if and Parse::do_ifnull for the reason 91 // (See Parse::do_if and Parse::do_ifnull for the reason
92 // we need an uncommon trap. Briefly, we need a way to 92 // we need an uncommon trap. Briefly, we need a way to
93 // detect failure of this optimization, as in 6366351.) 93 // detect failure of this optimization, as in 6366351.)
94 { 94 {
95 bool found_trap = false; 95 bool found_trap = false;
96 for (uint i1 = 0; i1 < null_block->_nodes.size(); i1++) { 96 for (uint i1 = 0; i1 < null_block->number_of_nodes(); i1++) {
97 Node* nn = null_block->get_node(i1); 97 Node* nn = null_block->get_node(i1);
98 if (nn->is_MachCall() && 98 if (nn->is_MachCall() &&
99 nn->as_MachCall()->entry_point() == SharedRuntime::uncommon_trap_blob()->entry_point()) { 99 nn->as_MachCall()->entry_point() == SharedRuntime::uncommon_trap_blob()->entry_point()) {
100 const Type* trtype = nn->in(TypeFunc::Parms)->bottom_type(); 100 const Type* trtype = nn->in(TypeFunc::Parms)->bottom_type();
101 if (trtype->isa_int() && trtype->is_int()->is_con()) { 101 if (trtype->isa_int() && trtype->is_int()->is_con()) {
235 continue; // Give up is reference is beyond 4K page size 235 continue; // Give up is reference is beyond 4K page size
236 } 236 }
237 } 237 }
238 238
239 // Check ctrl input to see if the null-check dominates the memory op 239 // Check ctrl input to see if the null-check dominates the memory op
240 Block *cb = cfg->get_block_for_node(mach); 240 Block *cb = get_block_for_node(mach);
241 cb = cb->_idom; // Always hoist at least 1 block 241 cb = cb->_idom; // Always hoist at least 1 block
242 if( !was_store ) { // Stores can be hoisted only one block 242 if( !was_store ) { // Stores can be hoisted only one block
243 while( cb->_dom_depth > (_dom_depth + 1)) 243 while( cb->_dom_depth > (block->_dom_depth + 1))
244 cb = cb->_idom; // Hoist loads as far as we want 244 cb = cb->_idom; // Hoist loads as far as we want
245 // The non-null-block should dominate the memory op, too. Live 245 // The non-null-block should dominate the memory op, too. Live
246 // range spilling will insert a spill in the non-null-block if it is 246 // range spilling will insert a spill in the non-null-block if it is
247 // needs to spill the memory op for an implicit null check. 247 // needs to spill the memory op for an implicit null check.
248 if (cb->_dom_depth == (_dom_depth + 1)) { 248 if (cb->_dom_depth == (block->_dom_depth + 1)) {
249 if (cb != not_null_block) continue; 249 if (cb != not_null_block) continue;
250 cb = cb->_idom; 250 cb = cb->_idom;
251 } 251 }
252 } 252 }
253 if( cb != this ) continue; 253 if( cb != block ) continue;
254 254
255 // Found a memory user; see if it can be hoisted to check-block 255 // Found a memory user; see if it can be hoisted to check-block
256 uint vidx = 0; // Capture index of value into memop 256 uint vidx = 0; // Capture index of value into memop
257 uint j; 257 uint j;
258 for( j = mach->req()-1; j > 0; j-- ) { 258 for( j = mach->req()-1; j > 0; j-- ) {
260 vidx = j; 260 vidx = j;
261 // Ignore DecodeN val which could be hoisted to where needed. 261 // Ignore DecodeN val which could be hoisted to where needed.
262 if( is_decoden ) continue; 262 if( is_decoden ) continue;
263 } 263 }
264 // Block of memory-op input 264 // Block of memory-op input
265 Block *inb = cfg->get_block_for_node(mach->in(j)); 265 Block *inb = get_block_for_node(mach->in(j));
266 Block *b = this; // Start from nul check 266 Block *b = block; // Start from nul check
267 while( b != inb && b->_dom_depth > inb->_dom_depth ) 267 while( b != inb && b->_dom_depth > inb->_dom_depth )
268 b = b->_idom; // search upwards for input 268 b = b->_idom; // search upwards for input
269 // See if input dominates null check 269 // See if input dominates null check
270 if( b != inb ) 270 if( b != inb )
271 break; 271 break;
272 } 272 }
273 if( j > 0 ) 273 if( j > 0 )
274 continue; 274 continue;
275 Block *mb = cfg->get_block_for_node(mach); 275 Block *mb = get_block_for_node(mach);
276 // Hoisting stores requires more checks for the anti-dependence case. 276 // Hoisting stores requires more checks for the anti-dependence case.
277 // Give up hoisting if we have to move the store past any load. 277 // Give up hoisting if we have to move the store past any load.
278 if( was_store ) { 278 if( was_store ) {
279 Block *b = mb; // Start searching here for a local load 279 Block *b = mb; // Start searching here for a local load
280 // mach use (faulting) trying to hoist 280 // mach use (faulting) trying to hoist
281 // n might be blocker to hoisting 281 // n might be blocker to hoisting
282 while( b != this ) { 282 while( b != block ) {
283 uint k; 283 uint k;
284 for( k = 1; k < b->_nodes.size(); k++ ) { 284 for( k = 1; k < b->number_of_nodes(); k++ ) {
285 Node *n = b->get_node(k); 285 Node *n = b->get_node(k);
286 if( n->needs_anti_dependence_check() && 286 if( n->needs_anti_dependence_check() &&
287 n->in(LoadNode::Memory) == mach->in(StoreNode::Memory) ) 287 n->in(LoadNode::Memory) == mach->in(StoreNode::Memory) )
288 break; // Found anti-dependent load 288 break; // Found anti-dependent load
289 } 289 }
290 if( k < b->_nodes.size() ) 290 if( k < b->number_of_nodes() )
291 break; // Found anti-dependent load 291 break; // Found anti-dependent load
292 // Make sure control does not do a merge (would have to check allpaths) 292 // Make sure control does not do a merge (would have to check allpaths)
293 if( b->num_preds() != 2 ) break; 293 if( b->num_preds() != 2 ) break;
294 b = cfg->get_block_for_node(b->pred(1)); // Move up to predecessor block 294 b = get_block_for_node(b->pred(1)); // Move up to predecessor block
295 } 295 }
296 if( b != this ) continue; 296 if( b != block ) continue;
297 } 297 }
298 298
299 // Make sure this memory op is not already being used for a NullCheck 299 // Make sure this memory op is not already being used for a NullCheck
300 Node *e = mb->end(); 300 Node *e = mb->end();
301 if( e->is_MachNullCheck() && e->in(1) == mach ) 301 if( e->is_MachNullCheck() && e->in(1) == mach )
302 continue; // Already being used as a NULL check 302 continue; // Already being used as a NULL check
303 303
304 // Found a candidate! Pick one with least dom depth - the highest 304 // Found a candidate! Pick one with least dom depth - the highest
305 // in the dom tree should be closest to the null check. 305 // in the dom tree should be closest to the null check.
306 if (best == NULL || cfg->get_block_for_node(mach)->_dom_depth < cfg->get_block_for_node(best)->_dom_depth) { 306 if (best == NULL || get_block_for_node(mach)->_dom_depth < get_block_for_node(best)->_dom_depth) {
307 best = mach; 307 best = mach;
308 bidx = vidx; 308 bidx = vidx;
309 } 309 }
310 } 310 }
311 // No candidate! 311 // No candidate!
317 extern int implicit_null_checks; 317 extern int implicit_null_checks;
318 implicit_null_checks++; 318 implicit_null_checks++;
319 319
320 if( is_decoden ) { 320 if( is_decoden ) {
321 // Check if we need to hoist decodeHeapOop_not_null first. 321 // Check if we need to hoist decodeHeapOop_not_null first.
322 Block *valb = cfg->get_block_for_node(val); 322 Block *valb = get_block_for_node(val);
323 if( this != valb && this->_dom_depth < valb->_dom_depth ) { 323 if( block != valb && block->_dom_depth < valb->_dom_depth ) {
324 // Hoist it up to the end of the test block. 324 // Hoist it up to the end of the test block.
325 valb->find_remove(val); 325 valb->find_remove(val);
326 this->add_inst(val); 326 block->add_inst(val);
327 cfg->map_node_to_block(val, this); 327 map_node_to_block(val, block);
328 // DecodeN on x86 may kill flags. Check for flag-killing projections 328 // DecodeN on x86 may kill flags. Check for flag-killing projections
329 // that also need to be hoisted. 329 // that also need to be hoisted.
330 for (DUIterator_Fast jmax, j = val->fast_outs(jmax); j < jmax; j++) { 330 for (DUIterator_Fast jmax, j = val->fast_outs(jmax); j < jmax; j++) {
331 Node* n = val->fast_out(j); 331 Node* n = val->fast_out(j);
332 if( n->is_MachProj() ) { 332 if( n->is_MachProj() ) {
333 cfg->get_block_for_node(n)->find_remove(n); 333 get_block_for_node(n)->find_remove(n);
334 this->add_inst(n); 334 block->add_inst(n);
335 cfg->map_node_to_block(n, this); 335 map_node_to_block(n, block);
336 } 336 }
337 } 337 }
338 } 338 }
339 } 339 }
340 // Hoist the memory candidate up to the end of the test block. 340 // Hoist the memory candidate up to the end of the test block.
341 Block *old_block = cfg->get_block_for_node(best); 341 Block *old_block = get_block_for_node(best);
342 old_block->find_remove(best); 342 old_block->find_remove(best);
343 add_inst(best); 343 block->add_inst(best);
344 cfg->map_node_to_block(best, this); 344 map_node_to_block(best, block);
345 345
346 // Move the control dependence 346 // Move the control dependence
347 if (best->in(0) && best->in(0) == old_block->head()) 347 if (best->in(0) && best->in(0) == old_block->head())
348 best->set_req(0, head()); 348 best->set_req(0, block->head());
349 349
350 // Check for flag-killing projections that also need to be hoisted 350 // Check for flag-killing projections that also need to be hoisted
351 // Should be DU safe because no edge updates. 351 // Should be DU safe because no edge updates.
352 for (DUIterator_Fast jmax, j = best->fast_outs(jmax); j < jmax; j++) { 352 for (DUIterator_Fast jmax, j = best->fast_outs(jmax); j < jmax; j++) {
353 Node* n = best->fast_out(j); 353 Node* n = best->fast_out(j);
354 if( n->is_MachProj() ) { 354 if( n->is_MachProj() ) {
355 cfg->get_block_for_node(n)->find_remove(n); 355 get_block_for_node(n)->find_remove(n);
356 add_inst(n); 356 block->add_inst(n);
357 cfg->map_node_to_block(n, this); 357 map_node_to_block(n, block);
358 } 358 }
359 } 359 }
360 360
361 Compile *C = cfg->C;
362 // proj==Op_True --> ne test; proj==Op_False --> eq test. 361 // proj==Op_True --> ne test; proj==Op_False --> eq test.
363 // One of two graph shapes got matched: 362 // One of two graph shapes got matched:
364 // (IfTrue (If (Bool NE (CmpP ptr NULL)))) 363 // (IfTrue (If (Bool NE (CmpP ptr NULL))))
365 // (IfFalse (If (Bool EQ (CmpP ptr NULL)))) 364 // (IfFalse (If (Bool EQ (CmpP ptr NULL))))
366 // NULL checks are always branch-if-eq. If we see a IfTrue projection 365 // NULL checks are always branch-if-eq. If we see a IfTrue projection
367 // then we are replacing a 'ne' test with a 'eq' NULL check test. 366 // then we are replacing a 'ne' test with a 'eq' NULL check test.
368 // We need to flip the projections to keep the same semantics. 367 // We need to flip the projections to keep the same semantics.
369 if( proj->Opcode() == Op_IfTrue ) { 368 if( proj->Opcode() == Op_IfTrue ) {
370 // Swap order of projections in basic block to swap branch targets 369 // Swap order of projections in basic block to swap branch targets
371 Node *tmp1 = get_node(end_idx()+1); 370 Node *tmp1 = block->get_node(block->end_idx()+1);
372 Node *tmp2 = get_node(end_idx()+2); 371 Node *tmp2 = block->get_node(block->end_idx()+2);
373 _nodes.map(end_idx()+1, tmp2); 372 block->map_node(tmp2, block->end_idx()+1);
374 _nodes.map(end_idx()+2, tmp1); 373 block->map_node(tmp1, block->end_idx()+2);
375 Node *tmp = new (C) Node(C->top()); // Use not NULL input 374 Node *tmp = new (C) Node(C->top()); // Use not NULL input
376 tmp1->replace_by(tmp); 375 tmp1->replace_by(tmp);
377 tmp2->replace_by(tmp1); 376 tmp2->replace_by(tmp1);
378 tmp->replace_by(tmp2); 377 tmp->replace_by(tmp2);
379 tmp->destruct(); 378 tmp->destruct();
382 // Remove the existing null check; use a new implicit null check instead. 381 // Remove the existing null check; use a new implicit null check instead.
383 // Since schedule-local needs precise def-use info, we need to correct 382 // Since schedule-local needs precise def-use info, we need to correct
384 // it as well. 383 // it as well.
385 Node *old_tst = proj->in(0); 384 Node *old_tst = proj->in(0);
386 MachNode *nul_chk = new (C) MachNullCheckNode(old_tst->in(0),best,bidx); 385 MachNode *nul_chk = new (C) MachNullCheckNode(old_tst->in(0),best,bidx);
387 _nodes.map(end_idx(),nul_chk); 386 block->map_node(nul_chk, block->end_idx());
388 cfg->map_node_to_block(nul_chk, this); 387 map_node_to_block(nul_chk, block);
389 // Redirect users of old_test to nul_chk 388 // Redirect users of old_test to nul_chk
390 for (DUIterator_Last i2min, i2 = old_tst->last_outs(i2min); i2 >= i2min; --i2) 389 for (DUIterator_Last i2min, i2 = old_tst->last_outs(i2min); i2 >= i2min; --i2)
391 old_tst->last_out(i2)->set_req(0, nul_chk); 390 old_tst->last_out(i2)->set_req(0, nul_chk);
392 // Clean-up any dead code 391 // Clean-up any dead code
393 for (uint i3 = 0; i3 < old_tst->req(); i3++) 392 for (uint i3 = 0; i3 < old_tst->req(); i3++)
394 old_tst->set_req(i3, NULL); 393 old_tst->set_req(i3, NULL);
395 394
396 cfg->latency_from_uses(nul_chk); 395 latency_from_uses(nul_chk);
397 cfg->latency_from_uses(best); 396 latency_from_uses(best);
398 } 397 }
399 398
400 399
401 //------------------------------select----------------------------------------- 400 //------------------------------select-----------------------------------------
402 // Select a nice fellow from the worklist to schedule next. If there is only 401 // Select a nice fellow from the worklist to schedule next. If there is only
406 // These are chosen immediately. Some instructions are required to immediately 405 // These are chosen immediately. Some instructions are required to immediately
407 // precede the last instruction in the block, and these are taken last. Of the 406 // precede the last instruction in the block, and these are taken last. Of the
408 // remaining cases (most), choose the instruction with the greatest latency 407 // remaining cases (most), choose the instruction with the greatest latency
409 // (that is, the most number of pseudo-cycles required to the end of the 408 // (that is, the most number of pseudo-cycles required to the end of the
410 // routine). If there is a tie, choose the instruction with the most inputs. 409 // routine). If there is a tie, choose the instruction with the most inputs.
411 Node *Block::select(PhaseCFG *cfg, Node_List &worklist, GrowableArray<int> &ready_cnt, VectorSet &next_call, uint sched_slot) { 410 Node* PhaseCFG::select(Block* block, Node_List &worklist, GrowableArray<int> &ready_cnt, VectorSet &next_call, uint sched_slot) {
412 411
413 // If only a single entry on the stack, use it 412 // If only a single entry on the stack, use it
414 uint cnt = worklist.size(); 413 uint cnt = worklist.size();
415 if (cnt == 1) { 414 if (cnt == 1) {
416 Node *n = worklist[0]; 415 Node *n = worklist[0];
440 worklist.map(i,worklist.pop()); 439 worklist.map(i,worklist.pop());
441 return n; 440 return n;
442 } 441 }
443 442
444 // Final call in a block must be adjacent to 'catch' 443 // Final call in a block must be adjacent to 'catch'
445 Node *e = end(); 444 Node *e = block->end();
446 if( e->is_Catch() && e->in(0)->in(0) == n ) 445 if( e->is_Catch() && e->in(0)->in(0) == n )
447 continue; 446 continue;
448 447
449 // Memory op for an implicit null check has to be at the end of the block 448 // Memory op for an implicit null check has to be at the end of the block
450 if( e->is_MachNullCheck() && e->in(1) == n ) 449 if( e->is_MachNullCheck() && e->in(1) == n )
466 465
467 for (DUIterator_Fast jmax, j = n->fast_outs(jmax); j < jmax; j++) { 466 for (DUIterator_Fast jmax, j = n->fast_outs(jmax); j < jmax; j++) {
468 Node* use = n->fast_out(j); 467 Node* use = n->fast_out(j);
469 468
470 // The use is a conditional branch, make them adjacent 469 // The use is a conditional branch, make them adjacent
471 if (use->is_MachIf() && cfg->get_block_for_node(use) == this) { 470 if (use->is_MachIf() && get_block_for_node(use) == block) {
472 found_machif = true; 471 found_machif = true;
473 break; 472 break;
474 } 473 }
475 474
476 // More than this instruction pending for successor to be ready, 475 // More than this instruction pending for successor to be ready,
499 // MachTemps should be scheduled last so they are near their uses 498 // MachTemps should be scheduled last so they are near their uses
500 if (n->is_MachTemp()) { 499 if (n->is_MachTemp()) {
501 n_choice = 1; 500 n_choice = 1;
502 } 501 }
503 502
504 uint n_latency = cfg->get_latency_for_node(n); 503 uint n_latency = get_latency_for_node(n);
505 uint n_score = n->req(); // Many inputs get high score to break ties 504 uint n_score = n->req(); // Many inputs get high score to break ties
506 505
507 // Keep best latency found 506 // Keep best latency found
508 cand_cnt++; 507 cand_cnt++;
509 if (choice < n_choice || 508 if (choice < n_choice ||
527 return n; 526 return n;
528 } 527 }
529 528
530 529
531 //------------------------------set_next_call---------------------------------- 530 //------------------------------set_next_call----------------------------------
532 void Block::set_next_call( Node *n, VectorSet &next_call, PhaseCFG* cfg) { 531 void PhaseCFG::set_next_call(Block* block, Node* n, VectorSet& next_call) {
533 if( next_call.test_set(n->_idx) ) return; 532 if( next_call.test_set(n->_idx) ) return;
534 for( uint i=0; i<n->len(); i++ ) { 533 for( uint i=0; i<n->len(); i++ ) {
535 Node *m = n->in(i); 534 Node *m = n->in(i);
536 if( !m ) continue; // must see all nodes in block that precede call 535 if( !m ) continue; // must see all nodes in block that precede call
537 if (cfg->get_block_for_node(m) == this) { 536 if (get_block_for_node(m) == block) {
538 set_next_call(m, next_call, cfg); 537 set_next_call(block, m, next_call);
539 } 538 }
540 } 539 }
541 } 540 }
542 541
543 //------------------------------needed_for_next_call--------------------------- 542 //------------------------------needed_for_next_call---------------------------
544 // Set the flag 'next_call' for each Node that is needed for the next call to 543 // Set the flag 'next_call' for each Node that is needed for the next call to
545 // be scheduled. This flag lets me bias scheduling so Nodes needed for the 544 // be scheduled. This flag lets me bias scheduling so Nodes needed for the
546 // next subroutine call get priority - basically it moves things NOT needed 545 // next subroutine call get priority - basically it moves things NOT needed
547 // for the next call till after the call. This prevents me from trying to 546 // for the next call till after the call. This prevents me from trying to
548 // carry lots of stuff live across a call. 547 // carry lots of stuff live across a call.
549 void Block::needed_for_next_call(Node *this_call, VectorSet &next_call, PhaseCFG* cfg) { 548 void PhaseCFG::needed_for_next_call(Block* block, Node* this_call, VectorSet& next_call) {
550 // Find the next control-defining Node in this block 549 // Find the next control-defining Node in this block
551 Node* call = NULL; 550 Node* call = NULL;
552 for (DUIterator_Fast imax, i = this_call->fast_outs(imax); i < imax; i++) { 551 for (DUIterator_Fast imax, i = this_call->fast_outs(imax); i < imax; i++) {
553 Node* m = this_call->fast_out(i); 552 Node* m = this_call->fast_out(i);
554 if(cfg->get_block_for_node(m) == this && // Local-block user 553 if(get_block_for_node(m) == block && // Local-block user
555 m != this_call && // Not self-start node 554 m != this_call && // Not self-start node
556 m->is_MachCall() ) 555 m->is_MachCall() )
557 call = m; 556 call = m;
558 break; 557 break;
559 } 558 }
560 if (call == NULL) return; // No next call (e.g., block end is near) 559 if (call == NULL) return; // No next call (e.g., block end is near)
561 // Set next-call for all inputs to this call 560 // Set next-call for all inputs to this call
562 set_next_call(call, next_call, cfg); 561 set_next_call(block, call, next_call);
563 } 562 }
564 563
565 //------------------------------add_call_kills------------------------------------- 564 //------------------------------add_call_kills-------------------------------------
566 void Block::add_call_kills(MachProjNode *proj, RegMask& regs, const char* save_policy, bool exclude_soe) { 565 // helper function that adds caller save registers to MachProjNode
566 static void add_call_kills(MachProjNode *proj, RegMask& regs, const char* save_policy, bool exclude_soe) {
567 // Fill in the kill mask for the call 567 // Fill in the kill mask for the call
568 for( OptoReg::Name r = OptoReg::Name(0); r < _last_Mach_Reg; r=OptoReg::add(r,1) ) { 568 for( OptoReg::Name r = OptoReg::Name(0); r < _last_Mach_Reg; r=OptoReg::add(r,1) ) {
569 if( !regs.Member(r) ) { // Not already defined by the call 569 if( !regs.Member(r) ) { // Not already defined by the call
570 // Save-on-call register? 570 // Save-on-call register?
571 if ((save_policy[r] == 'C') || 571 if ((save_policy[r] == 'C') ||
577 } 577 }
578 } 578 }
579 579
580 580
581 //------------------------------sched_call------------------------------------- 581 //------------------------------sched_call-------------------------------------
582 uint Block::sched_call( Matcher &matcher, PhaseCFG* cfg, uint node_cnt, Node_List &worklist, GrowableArray<int> &ready_cnt, MachCallNode *mcall, VectorSet &next_call ) { 582 uint PhaseCFG::sched_call(Block* block, uint node_cnt, Node_List& worklist, GrowableArray<int>& ready_cnt, MachCallNode* mcall, VectorSet& next_call) {
583 RegMask regs; 583 RegMask regs;
584 584
585 // Schedule all the users of the call right now. All the users are 585 // Schedule all the users of the call right now. All the users are
586 // projection Nodes, so they must be scheduled next to the call. 586 // projection Nodes, so they must be scheduled next to the call.
587 // Collect all the defined registers. 587 // Collect all the defined registers.
590 assert( n->is_MachProj(), "" ); 590 assert( n->is_MachProj(), "" );
591 int n_cnt = ready_cnt.at(n->_idx)-1; 591 int n_cnt = ready_cnt.at(n->_idx)-1;
592 ready_cnt.at_put(n->_idx, n_cnt); 592 ready_cnt.at_put(n->_idx, n_cnt);
593 assert( n_cnt == 0, "" ); 593 assert( n_cnt == 0, "" );
594 // Schedule next to call 594 // Schedule next to call
595 _nodes.map(node_cnt++, n); 595 block->map_node(n, node_cnt++);
596 // Collect defined registers 596 // Collect defined registers
597 regs.OR(n->out_RegMask()); 597 regs.OR(n->out_RegMask());
598 // Check for scheduling the next control-definer 598 // Check for scheduling the next control-definer
599 if( n->bottom_type() == Type::CONTROL ) 599 if( n->bottom_type() == Type::CONTROL )
600 // Warm up next pile of heuristic bits 600 // Warm up next pile of heuristic bits
601 needed_for_next_call(n, next_call, cfg); 601 needed_for_next_call(block, n, next_call);
602 602
603 // Children of projections are now all ready 603 // Children of projections are now all ready
604 for (DUIterator_Fast jmax, j = n->fast_outs(jmax); j < jmax; j++) { 604 for (DUIterator_Fast jmax, j = n->fast_outs(jmax); j < jmax; j++) {
605 Node* m = n->fast_out(j); // Get user 605 Node* m = n->fast_out(j); // Get user
606 if(cfg->get_block_for_node(m) != this) { 606 if(get_block_for_node(m) != block) {
607 continue; 607 continue;
608 } 608 }
609 if( m->is_Phi() ) continue; 609 if( m->is_Phi() ) continue;
610 int m_cnt = ready_cnt.at(m->_idx)-1; 610 int m_cnt = ready_cnt.at(m->_idx)-1;
611 ready_cnt.at_put(m->_idx, m_cnt); 611 ready_cnt.at_put(m->_idx, m_cnt);
615 615
616 } 616 }
617 617
618 // Act as if the call defines the Frame Pointer. 618 // Act as if the call defines the Frame Pointer.
619 // Certainly the FP is alive and well after the call. 619 // Certainly the FP is alive and well after the call.
620 regs.Insert(matcher.c_frame_pointer()); 620 regs.Insert(_matcher.c_frame_pointer());
621 621
622 // Set all registers killed and not already defined by the call. 622 // Set all registers killed and not already defined by the call.
623 uint r_cnt = mcall->tf()->range()->cnt(); 623 uint r_cnt = mcall->tf()->range()->cnt();
624 int op = mcall->ideal_Opcode(); 624 int op = mcall->ideal_Opcode();
625 MachProjNode *proj = new (matcher.C) MachProjNode( mcall, r_cnt+1, RegMask::Empty, MachProjNode::fat_proj ); 625 MachProjNode *proj = new (C) MachProjNode( mcall, r_cnt+1, RegMask::Empty, MachProjNode::fat_proj );
626 cfg->map_node_to_block(proj, this); 626 map_node_to_block(proj, block);
627 insert_node(proj, node_cnt++); 627 block->insert_node(proj, node_cnt++);
628 628
629 // Select the right register save policy. 629 // Select the right register save policy.
630 const char * save_policy; 630 const char * save_policy;
631 switch (op) { 631 switch (op) {
632 case Op_CallRuntime: 632 case Op_CallRuntime:
633 case Op_CallLeaf: 633 case Op_CallLeaf:
634 case Op_CallLeafNoFP: 634 case Op_CallLeafNoFP:
635 // Calling C code so use C calling convention 635 // Calling C code so use C calling convention
636 save_policy = matcher._c_reg_save_policy; 636 save_policy = _matcher._c_reg_save_policy;
637 break; 637 break;
638 638
639 case Op_CallStaticJava: 639 case Op_CallStaticJava:
640 case Op_CallDynamicJava: 640 case Op_CallDynamicJava:
641 // Calling Java code so use Java calling convention 641 // Calling Java code so use Java calling convention
642 save_policy = matcher._register_save_policy; 642 save_policy = _matcher._register_save_policy;
643 break; 643 break;
644 644
645 default: 645 default:
646 ShouldNotReachHere(); 646 ShouldNotReachHere();
647 } 647 }
672 } 672 }
673 673
674 674
675 //------------------------------schedule_local--------------------------------- 675 //------------------------------schedule_local---------------------------------
676 // Topological sort within a block. Someday become a real scheduler. 676 // Topological sort within a block. Someday become a real scheduler.
677 bool Block::schedule_local(PhaseCFG *cfg, Matcher &matcher, GrowableArray<int> &ready_cnt, VectorSet &next_call) { 677 bool PhaseCFG::schedule_local(Block* block, GrowableArray<int>& ready_cnt, VectorSet& next_call) {
678 // Already "sorted" are the block start Node (as the first entry), and 678 // Already "sorted" are the block start Node (as the first entry), and
679 // the block-ending Node and any trailing control projections. We leave 679 // the block-ending Node and any trailing control projections. We leave
680 // these alone. PhiNodes and ParmNodes are made to follow the block start 680 // these alone. PhiNodes and ParmNodes are made to follow the block start
681 // Node. Everything else gets topo-sorted. 681 // Node. Everything else gets topo-sorted.
682 682
683 #ifndef PRODUCT 683 #ifndef PRODUCT
684 if (cfg->trace_opto_pipelining()) { 684 if (trace_opto_pipelining()) {
685 tty->print_cr("# --- schedule_local B%d, before: ---", _pre_order); 685 tty->print_cr("# --- schedule_local B%d, before: ---", block->_pre_order);
686 for (uint i = 0;i < _nodes.size();i++) { 686 for (uint i = 0;i < block->number_of_nodes(); i++) {
687 tty->print("# "); 687 tty->print("# ");
688 get_node(i)->fast_dump(); 688 block->get_node(i)->fast_dump();
689 } 689 }
690 tty->print_cr("#"); 690 tty->print_cr("#");
691 } 691 }
692 #endif 692 #endif
693 693
694 // RootNode is already sorted 694 // RootNode is already sorted
695 if( _nodes.size() == 1 ) return true; 695 if (block->number_of_nodes() == 1) {
696 return true;
697 }
696 698
697 // Move PhiNodes and ParmNodes from 1 to cnt up to the start 699 // Move PhiNodes and ParmNodes from 1 to cnt up to the start
698 uint node_cnt = end_idx(); 700 uint node_cnt = block->end_idx();
699 uint phi_cnt = 1; 701 uint phi_cnt = 1;
700 uint i; 702 uint i;
701 for( i = 1; i<node_cnt; i++ ) { // Scan for Phi 703 for( i = 1; i<node_cnt; i++ ) { // Scan for Phi
702 Node *n = get_node(i); 704 Node *n = block->get_node(i);
703 if( n->is_Phi() || // Found a PhiNode or ParmNode 705 if( n->is_Phi() || // Found a PhiNode or ParmNode
704 (n->is_Proj() && n->in(0) == head()) ) { 706 (n->is_Proj() && n->in(0) == block->head()) ) {
705 // Move guy at 'phi_cnt' to the end; makes a hole at phi_cnt 707 // Move guy at 'phi_cnt' to the end; makes a hole at phi_cnt
706 _nodes.map(i,get_node(phi_cnt)); 708 block->map_node(block->get_node(phi_cnt), i);
707 _nodes.map(phi_cnt++,n); // swap Phi/Parm up front 709 block->map_node(n, phi_cnt++); // swap Phi/Parm up front
708 } else { // All others 710 } else { // All others
709 // Count block-local inputs to 'n' 711 // Count block-local inputs to 'n'
710 uint cnt = n->len(); // Input count 712 uint cnt = n->len(); // Input count
711 uint local = 0; 713 uint local = 0;
712 for( uint j=0; j<cnt; j++ ) { 714 for( uint j=0; j<cnt; j++ ) {
713 Node *m = n->in(j); 715 Node *m = n->in(j);
714 if( m && cfg->get_block_for_node(m) == this && !m->is_top() ) 716 if( m && get_block_for_node(m) == block && !m->is_top() )
715 local++; // One more block-local input 717 local++; // One more block-local input
716 } 718 }
717 ready_cnt.at_put(n->_idx, local); // Count em up 719 ready_cnt.at_put(n->_idx, local); // Count em up
718 720
719 #ifdef ASSERT 721 #ifdef ASSERT
721 if( n->is_Mach() && n->as_Mach()->ideal_Opcode() == Op_StoreCM ) { 723 if( n->is_Mach() && n->as_Mach()->ideal_Opcode() == Op_StoreCM ) {
722 // Check the precedence edges 724 // Check the precedence edges
723 for (uint prec = n->req(); prec < n->len(); prec++) { 725 for (uint prec = n->req(); prec < n->len(); prec++) {
724 Node* oop_store = n->in(prec); 726 Node* oop_store = n->in(prec);
725 if (oop_store != NULL) { 727 if (oop_store != NULL) {
726 assert(cfg->get_block_for_node(oop_store)->_dom_depth <= this->_dom_depth, "oop_store must dominate card-mark"); 728 assert(get_block_for_node(oop_store)->_dom_depth <= block->_dom_depth, "oop_store must dominate card-mark");
727 } 729 }
728 } 730 }
729 } 731 }
730 } 732 }
731 #endif 733 #endif
745 n->del_req(TypeFunc::Parms); 747 n->del_req(TypeFunc::Parms);
746 n->add_prec(x); 748 n->add_prec(x);
747 } 749 }
748 } 750 }
749 } 751 }
750 for(uint i2=i; i2<_nodes.size(); i2++ ) // Trailing guys get zapped count 752 for(uint i2=i; i2< block->number_of_nodes(); i2++ ) // Trailing guys get zapped count
751 ready_cnt.at_put(get_node(i2)->_idx, 0); 753 ready_cnt.at_put(block->get_node(i2)->_idx, 0);
752 754
753 // All the prescheduled guys do not hold back internal nodes 755 // All the prescheduled guys do not hold back internal nodes
754 uint i3; 756 uint i3;
755 for(i3 = 0; i3<phi_cnt; i3++ ) { // For all pre-scheduled 757 for(i3 = 0; i3<phi_cnt; i3++ ) { // For all pre-scheduled
756 Node *n = get_node(i3); // Get pre-scheduled 758 Node *n = block->get_node(i3); // Get pre-scheduled
757 for (DUIterator_Fast jmax, j = n->fast_outs(jmax); j < jmax; j++) { 759 for (DUIterator_Fast jmax, j = n->fast_outs(jmax); j < jmax; j++) {
758 Node* m = n->fast_out(j); 760 Node* m = n->fast_out(j);
759 if (cfg->get_block_for_node(m) == this) { // Local-block user 761 if (get_block_for_node(m) == block) { // Local-block user
760 int m_cnt = ready_cnt.at(m->_idx)-1; 762 int m_cnt = ready_cnt.at(m->_idx)-1;
761 ready_cnt.at_put(m->_idx, m_cnt); // Fix ready count 763 ready_cnt.at_put(m->_idx, m_cnt); // Fix ready count
762 } 764 }
763 } 765 }
764 } 766 }
765 767
766 Node_List delay; 768 Node_List delay;
767 // Make a worklist 769 // Make a worklist
768 Node_List worklist; 770 Node_List worklist;
769 for(uint i4=i3; i4<node_cnt; i4++ ) { // Put ready guys on worklist 771 for(uint i4=i3; i4<node_cnt; i4++ ) { // Put ready guys on worklist
770 Node *m = get_node(i4); 772 Node *m = block->get_node(i4);
771 if( !ready_cnt.at(m->_idx) ) { // Zero ready count? 773 if( !ready_cnt.at(m->_idx) ) { // Zero ready count?
772 if (m->is_iteratively_computed()) { 774 if (m->is_iteratively_computed()) {
773 // Push induction variable increments last to allow other uses 775 // Push induction variable increments last to allow other uses
774 // of the phi to be scheduled first. The select() method breaks 776 // of the phi to be scheduled first. The select() method breaks
775 // ties in scheduling by worklist order. 777 // ties in scheduling by worklist order.
787 Node* d = delay.pop(); 789 Node* d = delay.pop();
788 worklist.push(d); 790 worklist.push(d);
789 } 791 }
790 792
791 // Warm up the 'next_call' heuristic bits 793 // Warm up the 'next_call' heuristic bits
792 needed_for_next_call(head(), next_call, cfg); 794 needed_for_next_call(block, block->head(), next_call);
793 795
794 #ifndef PRODUCT 796 #ifndef PRODUCT
795 if (cfg->trace_opto_pipelining()) { 797 if (trace_opto_pipelining()) {
796 for (uint j=0; j<_nodes.size(); j++) { 798 for (uint j=0; j< block->number_of_nodes(); j++) {
797 Node *n = get_node(j); 799 Node *n = block->get_node(j);
798 int idx = n->_idx; 800 int idx = n->_idx;
799 tty->print("# ready cnt:%3d ", ready_cnt.at(idx)); 801 tty->print("# ready cnt:%3d ", ready_cnt.at(idx));
800 tty->print("latency:%3d ", cfg->get_latency_for_node(n)); 802 tty->print("latency:%3d ", get_latency_for_node(n));
801 tty->print("%4d: %s\n", idx, n->Name()); 803 tty->print("%4d: %s\n", idx, n->Name());
802 } 804 }
803 } 805 }
804 #endif 806 #endif
805 807
806 uint max_idx = (uint)ready_cnt.length(); 808 uint max_idx = (uint)ready_cnt.length();
807 // Pull from worklist and schedule 809 // Pull from worklist and schedule
808 while( worklist.size() ) { // Worklist is not ready 810 while( worklist.size() ) { // Worklist is not ready
809 811
810 #ifndef PRODUCT 812 #ifndef PRODUCT
811 if (cfg->trace_opto_pipelining()) { 813 if (trace_opto_pipelining()) {
812 tty->print("# ready list:"); 814 tty->print("# ready list:");
813 for( uint i=0; i<worklist.size(); i++ ) { // Inspect entire worklist 815 for( uint i=0; i<worklist.size(); i++ ) { // Inspect entire worklist
814 Node *n = worklist[i]; // Get Node on worklist 816 Node *n = worklist[i]; // Get Node on worklist
815 tty->print(" %d", n->_idx); 817 tty->print(" %d", n->_idx);
816 } 818 }
817 tty->cr(); 819 tty->cr();
818 } 820 }
819 #endif 821 #endif
820 822
821 // Select and pop a ready guy from worklist 823 // Select and pop a ready guy from worklist
822 Node* n = select(cfg, worklist, ready_cnt, next_call, phi_cnt); 824 Node* n = select(block, worklist, ready_cnt, next_call, phi_cnt);
823 _nodes.map(phi_cnt++,n); // Schedule him next 825 block->map_node(n, phi_cnt++); // Schedule him next
824 826
825 #ifndef PRODUCT 827 #ifndef PRODUCT
826 if (cfg->trace_opto_pipelining()) { 828 if (trace_opto_pipelining()) {
827 tty->print("# select %d: %s", n->_idx, n->Name()); 829 tty->print("# select %d: %s", n->_idx, n->Name());
828 tty->print(", latency:%d", cfg->get_latency_for_node(n)); 830 tty->print(", latency:%d", get_latency_for_node(n));
829 n->dump(); 831 n->dump();
830 if (Verbose) { 832 if (Verbose) {
831 tty->print("# ready list:"); 833 tty->print("# ready list:");
832 for( uint i=0; i<worklist.size(); i++ ) { // Inspect entire worklist 834 for( uint i=0; i<worklist.size(); i++ ) { // Inspect entire worklist
833 Node *n = worklist[i]; // Get Node on worklist 835 Node *n = worklist[i]; // Get Node on worklist
838 } 840 }
839 841
840 #endif 842 #endif
841 if( n->is_MachCall() ) { 843 if( n->is_MachCall() ) {
842 MachCallNode *mcall = n->as_MachCall(); 844 MachCallNode *mcall = n->as_MachCall();
843 phi_cnt = sched_call(matcher, cfg, phi_cnt, worklist, ready_cnt, mcall, next_call); 845 phi_cnt = sched_call(block, phi_cnt, worklist, ready_cnt, mcall, next_call);
844 continue; 846 continue;
845 } 847 }
846 848
847 if (n->is_Mach() && n->as_Mach()->has_call()) { 849 if (n->is_Mach() && n->as_Mach()->has_call()) {
848 RegMask regs; 850 RegMask regs;
849 regs.Insert(matcher.c_frame_pointer()); 851 regs.Insert(_matcher.c_frame_pointer());
850 regs.OR(n->out_RegMask()); 852 regs.OR(n->out_RegMask());
851 853
852 MachProjNode *proj = new (matcher.C) MachProjNode( n, 1, RegMask::Empty, MachProjNode::fat_proj ); 854 MachProjNode *proj = new (C) MachProjNode( n, 1, RegMask::Empty, MachProjNode::fat_proj );
853 cfg->map_node_to_block(proj, this); 855 map_node_to_block(proj, block);
854 insert_node(proj, phi_cnt++); 856 block->insert_node(proj, phi_cnt++);
855 857
856 add_call_kills(proj, regs, matcher._c_reg_save_policy, false); 858 add_call_kills(proj, regs, _matcher._c_reg_save_policy, false);
857 } 859 }
858 860
859 // Children are now all ready 861 // Children are now all ready
860 for (DUIterator_Fast i5max, i5 = n->fast_outs(i5max); i5 < i5max; i5++) { 862 for (DUIterator_Fast i5max, i5 = n->fast_outs(i5max); i5 < i5max; i5++) {
861 Node* m = n->fast_out(i5); // Get user 863 Node* m = n->fast_out(i5); // Get user
862 if (cfg->get_block_for_node(m) != this) { 864 if (get_block_for_node(m) != block) {
863 continue; 865 continue;
864 } 866 }
865 if( m->is_Phi() ) continue; 867 if( m->is_Phi() ) continue;
866 if (m->_idx >= max_idx) { // new node, skip it 868 if (m->_idx >= max_idx) { // new node, skip it
867 assert(m->is_MachProj() && n->is_Mach() && n->as_Mach()->has_call(), "unexpected node types"); 869 assert(m->is_MachProj() && n->is_Mach() && n->as_Mach()->has_call(), "unexpected node types");
872 if( m_cnt == 0 ) 874 if( m_cnt == 0 )
873 worklist.push(m); 875 worklist.push(m);
874 } 876 }
875 } 877 }
876 878
877 if( phi_cnt != end_idx() ) { 879 if( phi_cnt != block->end_idx() ) {
878 // did not schedule all. Retry, Bailout, or Die 880 // did not schedule all. Retry, Bailout, or Die
879 Compile* C = matcher.C;
880 if (C->subsume_loads() == true && !C->failing()) { 881 if (C->subsume_loads() == true && !C->failing()) {
881 // Retry with subsume_loads == false 882 // Retry with subsume_loads == false
882 // If this is the first failure, the sentinel string will "stick" 883 // If this is the first failure, the sentinel string will "stick"
883 // to the Compile object, and the C2Compiler will see it and retry. 884 // to the Compile object, and the C2Compiler will see it and retry.
884 C->record_failure(C2Compiler::retry_no_subsuming_loads()); 885 C->record_failure(C2Compiler::retry_no_subsuming_loads());
886 // assert( phi_cnt == end_idx(), "did not schedule all" ); 887 // assert( phi_cnt == end_idx(), "did not schedule all" );
887 return false; 888 return false;
888 } 889 }
889 890
890 #ifndef PRODUCT 891 #ifndef PRODUCT
891 if (cfg->trace_opto_pipelining()) { 892 if (trace_opto_pipelining()) {
892 tty->print_cr("#"); 893 tty->print_cr("#");
893 tty->print_cr("# after schedule_local"); 894 tty->print_cr("# after schedule_local");
894 for (uint i = 0;i < _nodes.size();i++) { 895 for (uint i = 0;i < block->number_of_nodes();i++) {
895 tty->print("# "); 896 tty->print("# ");
896 get_node(i)->fast_dump(); 897 block->get_node(i)->fast_dump();
897 } 898 }
898 tty->cr(); 899 tty->cr();
899 } 900 }
900 #endif 901 #endif
901 902
917 } 918 }
918 } 919 }
919 } 920 }
920 921
921 //------------------------------catch_cleanup_find_cloned_def------------------ 922 //------------------------------catch_cleanup_find_cloned_def------------------
922 static Node *catch_cleanup_find_cloned_def(Block *use_blk, Node *def, Block *def_blk, PhaseCFG* cfg, int n_clone_idx) { 923 Node* PhaseCFG::catch_cleanup_find_cloned_def(Block *use_blk, Node *def, Block *def_blk, int n_clone_idx) {
923 assert( use_blk != def_blk, "Inter-block cleanup only"); 924 assert( use_blk != def_blk, "Inter-block cleanup only");
924 925
925 // The use is some block below the Catch. Find and return the clone of the def 926 // The use is some block below the Catch. Find and return the clone of the def
926 // that dominates the use. If there is no clone in a dominating block, then 927 // that dominates the use. If there is no clone in a dominating block, then
927 // create a phi for the def in a dominating block. 928 // create a phi for the def in a dominating block.
943 if( j == def_blk->_num_succs ) { 944 if( j == def_blk->_num_succs ) {
944 // Block at same level in dom-tree is not a successor. It needs a 945 // Block at same level in dom-tree is not a successor. It needs a
945 // PhiNode, the PhiNode uses from the def and IT's uses need fixup. 946 // PhiNode, the PhiNode uses from the def and IT's uses need fixup.
946 Node_Array inputs = new Node_List(Thread::current()->resource_area()); 947 Node_Array inputs = new Node_List(Thread::current()->resource_area());
947 for(uint k = 1; k < use_blk->num_preds(); k++) { 948 for(uint k = 1; k < use_blk->num_preds(); k++) {
948 Block* block = cfg->get_block_for_node(use_blk->pred(k)); 949 Block* block = get_block_for_node(use_blk->pred(k));
949 inputs.map(k, catch_cleanup_find_cloned_def(block, def, def_blk, cfg, n_clone_idx)); 950 inputs.map(k, catch_cleanup_find_cloned_def(block, def, def_blk, n_clone_idx));
950 } 951 }
951 952
952 // Check to see if the use_blk already has an identical phi inserted. 953 // Check to see if the use_blk already has an identical phi inserted.
953 // If it exists, it will be at the first position since all uses of a 954 // If it exists, it will be at the first position since all uses of a
954 // def are processed together. 955 // def are processed together.
966 967
967 // If an existing PhiNode was not found, make a new one. 968 // If an existing PhiNode was not found, make a new one.
968 if (fixup == NULL) { 969 if (fixup == NULL) {
969 Node *new_phi = PhiNode::make(use_blk->head(), def); 970 Node *new_phi = PhiNode::make(use_blk->head(), def);
970 use_blk->insert_node(new_phi, 1); 971 use_blk->insert_node(new_phi, 1);
971 cfg->map_node_to_block(new_phi, use_blk); 972 map_node_to_block(new_phi, use_blk);
972 for (uint k = 1; k < use_blk->num_preds(); k++) { 973 for (uint k = 1; k < use_blk->num_preds(); k++) {
973 new_phi->set_req(k, inputs[k]); 974 new_phi->set_req(k, inputs[k]);
974 } 975 }
975 fixup = new_phi; 976 fixup = new_phi;
976 } 977 }
1006 } 1007 }
1007 1008
1008 //------------------------------catch_cleanup_inter_block--------------------- 1009 //------------------------------catch_cleanup_inter_block---------------------
1009 // Fix all input edges in use that reference "def". The use is in a different 1010 // Fix all input edges in use that reference "def". The use is in a different
1010 // block than the def. 1011 // block than the def.
1011 static void catch_cleanup_inter_block(Node *use, Block *use_blk, Node *def, Block *def_blk, PhaseCFG* cfg, int n_clone_idx) { 1012 void PhaseCFG::catch_cleanup_inter_block(Node *use, Block *use_blk, Node *def, Block *def_blk, int n_clone_idx) {
1012 if( !use_blk ) return; // Can happen if the use is a precedence edge 1013 if( !use_blk ) return; // Can happen if the use is a precedence edge
1013 1014
1014 Node *new_def = catch_cleanup_find_cloned_def(use_blk, def, def_blk, cfg, n_clone_idx); 1015 Node *new_def = catch_cleanup_find_cloned_def(use_blk, def, def_blk, n_clone_idx);
1015 catch_cleanup_fix_all_inputs(use, def, new_def); 1016 catch_cleanup_fix_all_inputs(use, def, new_def);
1016 } 1017 }
1017 1018
1018 //------------------------------call_catch_cleanup----------------------------- 1019 //------------------------------call_catch_cleanup-----------------------------
1019 // If we inserted any instructions between a Call and his CatchNode, 1020 // If we inserted any instructions between a Call and his CatchNode,
1020 // clone the instructions on all paths below the Catch. 1021 // clone the instructions on all paths below the Catch.
1021 void Block::call_catch_cleanup(PhaseCFG* cfg, Compile* C) { 1022 void PhaseCFG::call_catch_cleanup(Block* block) {
1022 1023
1023 // End of region to clone 1024 // End of region to clone
1024 uint end = end_idx(); 1025 uint end = block->end_idx();
1025 if( !get_node(end)->is_Catch() ) return; 1026 if( !block->get_node(end)->is_Catch() ) return;
1026 // Start of region to clone 1027 // Start of region to clone
1027 uint beg = end; 1028 uint beg = end;
1028 while(!get_node(beg-1)->is_MachProj() || 1029 while(!block->get_node(beg-1)->is_MachProj() ||
1029 !get_node(beg-1)->in(0)->is_MachCall() ) { 1030 !block->get_node(beg-1)->in(0)->is_MachCall() ) {
1030 beg--; 1031 beg--;
1031 assert(beg > 0,"Catch cleanup walking beyond block boundary"); 1032 assert(beg > 0,"Catch cleanup walking beyond block boundary");
1032 } 1033 }
1033 // Range of inserted instructions is [beg, end) 1034 // Range of inserted instructions is [beg, end)
1034 if( beg == end ) return; 1035 if( beg == end ) return;
1035 1036
1036 // Clone along all Catch output paths. Clone area between the 'beg' and 1037 // Clone along all Catch output paths. Clone area between the 'beg' and
1037 // 'end' indices. 1038 // 'end' indices.
1038 for( uint i = 0; i < _num_succs; i++ ) { 1039 for( uint i = 0; i < block->_num_succs; i++ ) {
1039 Block *sb = _succs[i]; 1040 Block *sb = block->_succs[i];
1040 // Clone the entire area; ignoring the edge fixup for now. 1041 // Clone the entire area; ignoring the edge fixup for now.
1041 for( uint j = end; j > beg; j-- ) { 1042 for( uint j = end; j > beg; j-- ) {
1042 // It is safe here to clone a node with anti_dependence 1043 // It is safe here to clone a node with anti_dependence
1043 // since clones dominate on each path. 1044 // since clones dominate on each path.
1044 Node *clone = get_node(j-1)->clone(); 1045 Node *clone = block->get_node(j-1)->clone();
1045 sb->insert_node(clone, 1); 1046 sb->insert_node(clone, 1);
1046 cfg->map_node_to_block(clone, sb); 1047 map_node_to_block(clone, sb);
1047 } 1048 }
1048 } 1049 }
1049 1050
1050 1051
1051 // Fixup edges. Check the def-use info per cloned Node 1052 // Fixup edges. Check the def-use info per cloned Node
1052 for(uint i2 = beg; i2 < end; i2++ ) { 1053 for(uint i2 = beg; i2 < end; i2++ ) {
1053 uint n_clone_idx = i2-beg+1; // Index of clone of n in each successor block 1054 uint n_clone_idx = i2-beg+1; // Index of clone of n in each successor block
1054 Node *n = get_node(i2); // Node that got cloned 1055 Node *n = block->get_node(i2); // Node that got cloned
1055 // Need DU safe iterator because of edge manipulation in calls. 1056 // Need DU safe iterator because of edge manipulation in calls.
1056 Unique_Node_List *out = new Unique_Node_List(Thread::current()->resource_area()); 1057 Unique_Node_List *out = new Unique_Node_List(Thread::current()->resource_area());
1057 for (DUIterator_Fast j1max, j1 = n->fast_outs(j1max); j1 < j1max; j1++) { 1058 for (DUIterator_Fast j1max, j1 = n->fast_outs(j1max); j1 < j1max; j1++) {
1058 out->push(n->fast_out(j1)); 1059 out->push(n->fast_out(j1));
1059 } 1060 }
1060 uint max = out->size(); 1061 uint max = out->size();
1061 for (uint j = 0; j < max; j++) {// For all users 1062 for (uint j = 0; j < max; j++) {// For all users
1062 Node *use = out->pop(); 1063 Node *use = out->pop();
1063 Block *buse = cfg->get_block_for_node(use); 1064 Block *buse = get_block_for_node(use);
1064 if( use->is_Phi() ) { 1065 if( use->is_Phi() ) {
1065 for( uint k = 1; k < use->req(); k++ ) 1066 for( uint k = 1; k < use->req(); k++ )
1066 if( use->in(k) == n ) { 1067 if( use->in(k) == n ) {
1067 Block* block = cfg->get_block_for_node(buse->pred(k)); 1068 Block* b = get_block_for_node(buse->pred(k));
1068 Node *fixup = catch_cleanup_find_cloned_def(block, n, this, cfg, n_clone_idx); 1069 Node *fixup = catch_cleanup_find_cloned_def(b, n, block, n_clone_idx);
1069 use->set_req(k, fixup); 1070 use->set_req(k, fixup);
1070 } 1071 }
1071 } else { 1072 } else {
1072 if (this == buse) { 1073 if (block == buse) {
1073 catch_cleanup_intra_block(use, n, this, beg, n_clone_idx); 1074 catch_cleanup_intra_block(use, n, block, beg, n_clone_idx);
1074 } else { 1075 } else {
1075 catch_cleanup_inter_block(use, buse, n, this, cfg, n_clone_idx); 1076 catch_cleanup_inter_block(use, buse, n, block, n_clone_idx);
1076 } 1077 }
1077 } 1078 }
1078 } // End for all users 1079 } // End for all users
1079 1080
1080 } // End of for all Nodes in cloned area 1081 } // End of for all Nodes in cloned area
1081 1082
1082 // Remove the now-dead cloned ops 1083 // Remove the now-dead cloned ops
1083 for(uint i3 = beg; i3 < end; i3++ ) { 1084 for(uint i3 = beg; i3 < end; i3++ ) {
1084 get_node(beg)->disconnect_inputs(NULL, C); 1085 block->get_node(beg)->disconnect_inputs(NULL, C);
1085 remove_node(beg); 1086 block->remove_node(beg);
1086 } 1087 }
1087 1088
1088 // If the successor blocks have a CreateEx node, move it back to the top 1089 // If the successor blocks have a CreateEx node, move it back to the top
1089 for(uint i4 = 0; i4 < _num_succs; i4++ ) { 1090 for(uint i4 = 0; i4 < block->_num_succs; i4++ ) {
1090 Block *sb = _succs[i4]; 1091 Block *sb = block->_succs[i4];
1091 uint new_cnt = end - beg; 1092 uint new_cnt = end - beg;
1092 // Remove any newly created, but dead, nodes. 1093 // Remove any newly created, but dead, nodes.
1093 for( uint j = new_cnt; j > 0; j-- ) { 1094 for( uint j = new_cnt; j > 0; j-- ) {
1094 Node *n = sb->get_node(j); 1095 Node *n = sb->get_node(j);
1095 if (n->outcnt() == 0 && 1096 if (n->outcnt() == 0 &&